Drug Therapy

Most patients should receive it, with exceptions as noted below. Metformin is the cornerstone of diabetes therapy and should be considered in all patients with type 2 diabetes. Both the American Diabetes Association (ADA) and the American Association of Clinical Endocrinologists (AACE)^1,2 recommend it as first-line treatment for type 2 diabetes. It lowers blood glucose levels by inhibiting hepatic glucose production, and it does not tend to cause hypoglycemia.

However, metformin is underused. A 2012 study showed that only 50% to 70% of patients with type 2 diabetes treated with a sulfonylurea, dipeptidyl peptidase-4 (DPP-4) inhibitor, thiazolidinedione, or glucagon-like peptide-1 analogue also received metformin.³ This occurred despite guidelines recommending continuing metformin when starting other diabetes drugs.⁴

EVIDENCE METFORMIN IS EFFECTIVE

The United Kingdom Prospective Diabetes Study (UKPDS)⁵ found that metformin significantly reduced the incidence of:

• Any diabetes-related end point (hazard ratio [HR] 0.68, 95% confidence interval [CI] 0.53–0.87)
• Myocardial infarction (HR 0.61, 95% CI 0.41–0.89)
• Diabetes-related death (HR 0.58, 95% CI 0.37–0.91)
• All-cause mortality (HR 0.64; 95% CI 0.45–0.91).

The Hyperinsulinemia: Outcomes of Its Metabolic Effects (HOME) trial,⁶ a multicenter trial conducted in the Netherlands, evaluated the effect of adding  metformin (vs placebo) to existing insulin regimens. Metformin recipients had a significantly lower rate of macrovascular mortality (HR 0.61, 95% CI 0.40–0.94, P = .02), but not of the primary end point, an aggregate of microvascular and macrovascular morbidity and mortality.

The Study on the Prognosis and Effect of Antidiabetic Drugs on Type 2 Diabetes Mellitus With Coronary Artery Disease trial,⁷ a multicenter trial conducted in China, compared the effects of metformin vs glipizide on cardiovascular outcomes. At about 3 years of treatment, the metformin group had a significantly lower rate of the composite primary end point of recurrent cardiovascular events (HR 0.54, 95% CI 0.30–0.90). This end point included nonfatal myocardial infarction, nonfatal stroke, arterial revascularization by percutaneous transluminal coronary angioplasty or by coronary artery bypass graft, death from a cardiovascular cause, and death from any cause.

These studies prompted the ADA to emphasize that metformin can reduce the risk of cardiovascular events or death. Metformin also has been shown to be weight-neutral or to induce slight weight loss. Furthermore, it is inexpensive.

WHAT ABOUT THE RENAL EFFECTS?

Because metformin is renally cleared, it has caused some concern about nephrotoxicity, especially lactic acidosis, in patients with impaired renal function. But the most recent guidelines have relaxed the criteria for metformin use in this patient population.

Revised labeling

Metformin’s labeling,⁸ revised in 2016, states the following:

• If the estimated glomerular filtration rate (eGFR) is below 30 mL/min/1.73 m², metformin is contraindicated
• If the eGFR is between 30 and 45 mL/min/1.73 m², metformin is not recommended
• If the eGFR is below 45 mL/min/1.73 m² in a patient taking metformin, the risks and benefits of continuing treatment should be assessed, the dosage may need to be adjusted, and renal function should be monitored more frequently.⁸

These labeling revisions were based on a systematic review by Inzucchi et al⁹ that found metformin is not associated with increased rates of lactic acidosis in patients with mild to moderate kidney disease. Subsequently, an observational study published in 2018 by Lazarus et al¹⁰ showed that metformin increases the risk of acidosis only at eGFR levels below 30 mL/min/1.73 m². Also, a Cochrane review published in 2003 did not find a single case of lactic acidosis in 347 trials with 70,490 patient-years of metformin treatment.¹¹

Previous guidelines used serum creatinine levels, with metformin contraindicated at levels of 1.5 mg/dL or above for men and 1.4 mg/dL for women, or with abnormal creatinine clearance. The ADA and the AACE now use the eGFR^1,2 instead of the serum creatinine level to measure kidney function because it better accounts for factors such as the patient’s age, sex, race, and weight.

Despite the evidence, the common patient perception is that metformin is nephrotoxic, and it is important for practitioners to dispel this myth during clinic visits.

What about metformin use with contrast agents?

Labeling has a precautionary note stating that metformin should be held at the time of, or prior to, any imaging procedure involving iodinated contrast agents in patients with an eGFR between 30 and 60 mL/min/1.73 m²; in patients with a history of hepatic impairment, alcoholism, or heart failure; or in patients who will receive intra-arterial iodinated contrast. The eGFR should be reevaluated 48 hours after the imaging procedure.⁸

Additionally, if the iodinated contrast agent causes acute kidney injury, metformin could accumulate, with resultant lactate accumulation.

The American College of Radiology (ACR) has proposed less stringent guidelines for metformin during radiocontrast imaging studies. This change is based on evidence that lactic acidosis is rare­—about 10 cases per 100,000 patient-years—and that there are no reports of lactic acidosis after intravenously administered iodinated contrast in properly selected patients.^12,13

The ACR divides patients taking metformin into 2 categories:

• No evidence of acute kidney injury and eGFR greater than 30 mL/min/1.73 m²
• Either acute kidney injury or chronic kidney disease with eGFR below 30 mL/min/1.73 m² or undergoing arterial catheter studies with a high chance of embolization to the renal arteries.¹⁴

For the first group, they recommend against discontinuing metformin before or after giving iodinated contrast or checking kidney function after the procedure.

For the second group, they recommend holding metformin before and 48 hours after the procedure. It should not be restarted until renal function is confirmed to be normal.

 

 

METFORMIN AND INSULIN

The ADA recommends¹ continuing metformin after initiating insulin. However, in clinical practice, it is often not done.

Clinical trials have shown that combining metformin with insulin significantly improves glycemic control, prevents weight gain, and decreases insulin requirements.^15,16 One trial¹⁶ also looked at cardiovascular end points during a 4-year follow-up period;  combining metformin with insulin decreased the macrovascular disease-related event rate compared with insulin alone.

In the HOME trial,⁶ which added metformin to the existing insulin regimen, both groups gained weight, but the metformin group had gained about 3 kg less than the placebo group at the end of the 4.3-year trial. Metformin did not increase the risk of hypoglycemia, but it also did not reduce the risk of microvascular disease.

Concomitant metformin reduces costs

These days, practitioners can choose from a large selection of diabetes drugs. These include insulins with better pharmacokinetic profiles, as well as newer classes of noninsulin agents such as sodium-glucose cotransporter-2 inhibitors and glucagon-like peptide-1 analogues.

Metformin is less expensive than these newer drugs, and using it concomitantly with other diabetes drugs can decrease their dosage requirements, which in turn decreases their monthly costs.

GASTROINTESTINAL EFFECTS

Metformin’s gastrointestinal adverse effects such as diarrhea, flatulence, nausea, and vomiting are a barrier to its use. The actual incidence rate of diarrhea varies widely in randomized trials and observational studies, and gastrointestinal effects are worse in metformin-naive patients, as well as those who have chronic gastritis or Helicobacter pylori infection.¹⁷

We have found that starting metformin at a low dose and up-titrating it over several weeks increases tolerability. We often start patients at 500 mg/day and increase the dosage by 1 500-mg tablet every 1 to 2 weeks. Also, we have noticed that intolerance is more likely in patients who eat a high-carbohydrate diet, but there is no high-level evidence to back this up because patients in clinical trials all undergo nutrition counseling and are therefore more likely to adhere to the low-carbohydrate diet.

Also, the extended-release formulation is more tolerable than the immediate-release formulation and has similar glycemic efficacy. It may be an option as first-line therapy or for patients who have significant adverse effects from immediate-release metformin.¹⁸ For patients on the immediate-release formulation, taking it with meals helps lessen some gastrointestinal effects, and this should be emphasized at every visit.

Finally, we limit the metformin dose to 2,000 mg/day, rather than the 2,550 mg/day allowed on labeling. Garber et al¹⁹ found that the lower dosage still provides the maximum clinical efficacy.

OTHER CAUTIONS

Metformin should be avoided in patients with acute or unstable heart failure because of the increased risk of lactic acidosis.

It also should be avoided in patients with hepatic impairment, according to the labeling. But this remains controversial in practice. Zhang et al²⁰ showed that continuing metformin in patients with diabetes and cirrhosis decreases the mortality risk by 57% compared with those taken off metformin.

Diet and lifestyle measures need to be emphasized at each visit. Wing et al²¹ showed that calorie restriction regardless of weight loss is beneficial for glycemic control and insulin sensitivity in obese patients with diabetes.

TAKE-HOME POINTS

Metformin improves glycemic control without tending to cause weight gain or hypoglycemia. It may also have cardiovascular benefits. Metformin is an inexpensive agent that should be continued, if tolerated, in those who need additional agents for glycemic control. It should be considered in all adult patients with type 2 diabetes.   

Most patients should receive it, with exceptions as noted below. Metformin is the cornerstone of diabetes therapy and should be considered in all patients with type 2 diabetes. Both the American Diabetes Association (ADA) and the American Association of Clinical Endocrinologists (AACE)^1,2 recommend it as first-line treatment for type 2 diabetes. It lowers blood glucose levels by inhibiting hepatic glucose production, and it does not tend to cause hypoglycemia.

However, metformin is underused. A 2012 study showed that only 50% to 70% of patients with type 2 diabetes treated with a sulfonylurea, dipeptidyl peptidase-4 (DPP-4) inhibitor, thiazolidinedione, or glucagon-like peptide-1 analogue also received metformin.³ This occurred despite guidelines recommending continuing metformin when starting other diabetes drugs.⁴

EVIDENCE METFORMIN IS EFFECTIVE

The United Kingdom Prospective Diabetes Study (UKPDS)⁵ found that metformin significantly reduced the incidence of:

• Any diabetes-related end point (hazard ratio [HR] 0.68, 95% confidence interval [CI] 0.53–0.87)
• Myocardial infarction (HR 0.61, 95% CI 0.41–0.89)
• Diabetes-related death (HR 0.58, 95% CI 0.37–0.91)
• All-cause mortality (HR 0.64; 95% CI 0.45–0.91).

The Hyperinsulinemia: Outcomes of Its Metabolic Effects (HOME) trial,⁶ a multicenter trial conducted in the Netherlands, evaluated the effect of adding  metformin (vs placebo) to existing insulin regimens. Metformin recipients had a significantly lower rate of macrovascular mortality (HR 0.61, 95% CI 0.40–0.94, P = .02), but not of the primary end point, an aggregate of microvascular and macrovascular morbidity and mortality.

The Study on the Prognosis and Effect of Antidiabetic Drugs on Type 2 Diabetes Mellitus With Coronary Artery Disease trial,⁷ a multicenter trial conducted in China, compared the effects of metformin vs glipizide on cardiovascular outcomes. At about 3 years of treatment, the metformin group had a significantly lower rate of the composite primary end point of recurrent cardiovascular events (HR 0.54, 95% CI 0.30–0.90). This end point included nonfatal myocardial infarction, nonfatal stroke, arterial revascularization by percutaneous transluminal coronary angioplasty or by coronary artery bypass graft, death from a cardiovascular cause, and death from any cause.

These studies prompted the ADA to emphasize that metformin can reduce the risk of cardiovascular events or death. Metformin also has been shown to be weight-neutral or to induce slight weight loss. Furthermore, it is inexpensive.

WHAT ABOUT THE RENAL EFFECTS?

Because metformin is renally cleared, it has caused some concern about nephrotoxicity, especially lactic acidosis, in patients with impaired renal function. But the most recent guidelines have relaxed the criteria for metformin use in this patient population.

Revised labeling

Metformin’s labeling,⁸ revised in 2016, states the following:

• If the estimated glomerular filtration rate (eGFR) is below 30 mL/min/1.73 m², metformin is contraindicated
• If the eGFR is between 30 and 45 mL/min/1.73 m², metformin is not recommended
• If the eGFR is below 45 mL/min/1.73 m² in a patient taking metformin, the risks and benefits of continuing treatment should be assessed, the dosage may need to be adjusted, and renal function should be monitored more frequently.⁸

These labeling revisions were based on a systematic review by Inzucchi et al⁹ that found metformin is not associated with increased rates of lactic acidosis in patients with mild to moderate kidney disease. Subsequently, an observational study published in 2018 by Lazarus et al¹⁰ showed that metformin increases the risk of acidosis only at eGFR levels below 30 mL/min/1.73 m². Also, a Cochrane review published in 2003 did not find a single case of lactic acidosis in 347 trials with 70,490 patient-years of metformin treatment.¹¹

Previous guidelines used serum creatinine levels, with metformin contraindicated at levels of 1.5 mg/dL or above for men and 1.4 mg/dL for women, or with abnormal creatinine clearance. The ADA and the AACE now use the eGFR^1,2 instead of the serum creatinine level to measure kidney function because it better accounts for factors such as the patient’s age, sex, race, and weight.

Despite the evidence, the common patient perception is that metformin is nephrotoxic, and it is important for practitioners to dispel this myth during clinic visits.

What about metformin use with contrast agents?

Labeling has a precautionary note stating that metformin should be held at the time of, or prior to, any imaging procedure involving iodinated contrast agents in patients with an eGFR between 30 and 60 mL/min/1.73 m²; in patients with a history of hepatic impairment, alcoholism, or heart failure; or in patients who will receive intra-arterial iodinated contrast. The eGFR should be reevaluated 48 hours after the imaging procedure.⁸

Additionally, if the iodinated contrast agent causes acute kidney injury, metformin could accumulate, with resultant lactate accumulation.

The American College of Radiology (ACR) has proposed less stringent guidelines for metformin during radiocontrast imaging studies. This change is based on evidence that lactic acidosis is rare­—about 10 cases per 100,000 patient-years—and that there are no reports of lactic acidosis after intravenously administered iodinated contrast in properly selected patients.^12,13

The ACR divides patients taking metformin into 2 categories:

• No evidence of acute kidney injury and eGFR greater than 30 mL/min/1.73 m²
• Either acute kidney injury or chronic kidney disease with eGFR below 30 mL/min/1.73 m² or undergoing arterial catheter studies with a high chance of embolization to the renal arteries.¹⁴

For the first group, they recommend against discontinuing metformin before or after giving iodinated contrast or checking kidney function after the procedure.

For the second group, they recommend holding metformin before and 48 hours after the procedure. It should not be restarted until renal function is confirmed to be normal.

 

 

METFORMIN AND INSULIN

The ADA recommends¹ continuing metformin after initiating insulin. However, in clinical practice, it is often not done.

Clinical trials have shown that combining metformin with insulin significantly improves glycemic control, prevents weight gain, and decreases insulin requirements.^15,16 One trial¹⁶ also looked at cardiovascular end points during a 4-year follow-up period;  combining metformin with insulin decreased the macrovascular disease-related event rate compared with insulin alone.

In the HOME trial,⁶ which added metformin to the existing insulin regimen, both groups gained weight, but the metformin group had gained about 3 kg less than the placebo group at the end of the 4.3-year trial. Metformin did not increase the risk of hypoglycemia, but it also did not reduce the risk of microvascular disease.

Concomitant metformin reduces costs

These days, practitioners can choose from a large selection of diabetes drugs. These include insulins with better pharmacokinetic profiles, as well as newer classes of noninsulin agents such as sodium-glucose cotransporter-2 inhibitors and glucagon-like peptide-1 analogues.

Metformin is less expensive than these newer drugs, and using it concomitantly with other diabetes drugs can decrease their dosage requirements, which in turn decreases their monthly costs.

GASTROINTESTINAL EFFECTS

Metformin’s gastrointestinal adverse effects such as diarrhea, flatulence, nausea, and vomiting are a barrier to its use. The actual incidence rate of diarrhea varies widely in randomized trials and observational studies, and gastrointestinal effects are worse in metformin-naive patients, as well as those who have chronic gastritis or Helicobacter pylori infection.¹⁷

We have found that starting metformin at a low dose and up-titrating it over several weeks increases tolerability. We often start patients at 500 mg/day and increase the dosage by 1 500-mg tablet every 1 to 2 weeks. Also, we have noticed that intolerance is more likely in patients who eat a high-carbohydrate diet, but there is no high-level evidence to back this up because patients in clinical trials all undergo nutrition counseling and are therefore more likely to adhere to the low-carbohydrate diet.

Also, the extended-release formulation is more tolerable than the immediate-release formulation and has similar glycemic efficacy. It may be an option as first-line therapy or for patients who have significant adverse effects from immediate-release metformin.¹⁸ For patients on the immediate-release formulation, taking it with meals helps lessen some gastrointestinal effects, and this should be emphasized at every visit.

Finally, we limit the metformin dose to 2,000 mg/day, rather than the 2,550 mg/day allowed on labeling. Garber et al¹⁹ found that the lower dosage still provides the maximum clinical efficacy.

OTHER CAUTIONS

Metformin should be avoided in patients with acute or unstable heart failure because of the increased risk of lactic acidosis.

It also should be avoided in patients with hepatic impairment, according to the labeling. But this remains controversial in practice. Zhang et al²⁰ showed that continuing metformin in patients with diabetes and cirrhosis decreases the mortality risk by 57% compared with those taken off metformin.

Diet and lifestyle measures need to be emphasized at each visit. Wing et al²¹ showed that calorie restriction regardless of weight loss is beneficial for glycemic control and insulin sensitivity in obese patients with diabetes.

TAKE-HOME POINTS

Metformin improves glycemic control without tending to cause weight gain or hypoglycemia. It may also have cardiovascular benefits. Metformin is an inexpensive agent that should be continued, if tolerated, in those who need additional agents for glycemic control. It should be considered in all adult patients with type 2 diabetes.   

Most patients should receive it, with exceptions as noted below. Metformin is the cornerstone of diabetes therapy and should be considered in all patients with type 2 diabetes. Both the American Diabetes Association (ADA) and the American Association of Clinical Endocrinologists (AACE)^1,2 recommend it as first-line treatment for type 2 diabetes. It lowers blood glucose levels by inhibiting hepatic glucose production, and it does not tend to cause hypoglycemia.

However, metformin is underused. A 2012 study showed that only 50% to 70% of patients with type 2 diabetes treated with a sulfonylurea, dipeptidyl peptidase-4 (DPP-4) inhibitor, thiazolidinedione, or glucagon-like peptide-1 analogue also received metformin.³ This occurred despite guidelines recommending continuing metformin when starting other diabetes drugs.⁴

EVIDENCE METFORMIN IS EFFECTIVE

The United Kingdom Prospective Diabetes Study (UKPDS)⁵ found that metformin significantly reduced the incidence of:

• Any diabetes-related end point (hazard ratio [HR] 0.68, 95% confidence interval [CI] 0.53–0.87)
• Myocardial infarction (HR 0.61, 95% CI 0.41–0.89)
• Diabetes-related death (HR 0.58, 95% CI 0.37–0.91)
• All-cause mortality (HR 0.64; 95% CI 0.45–0.91).

The Hyperinsulinemia: Outcomes of Its Metabolic Effects (HOME) trial,⁶ a multicenter trial conducted in the Netherlands, evaluated the effect of adding  metformin (vs placebo) to existing insulin regimens. Metformin recipients had a significantly lower rate of macrovascular mortality (HR 0.61, 95% CI 0.40–0.94, P = .02), but not of the primary end point, an aggregate of microvascular and macrovascular morbidity and mortality.

The Study on the Prognosis and Effect of Antidiabetic Drugs on Type 2 Diabetes Mellitus With Coronary Artery Disease trial,⁷ a multicenter trial conducted in China, compared the effects of metformin vs glipizide on cardiovascular outcomes. At about 3 years of treatment, the metformin group had a significantly lower rate of the composite primary end point of recurrent cardiovascular events (HR 0.54, 95% CI 0.30–0.90). This end point included nonfatal myocardial infarction, nonfatal stroke, arterial revascularization by percutaneous transluminal coronary angioplasty or by coronary artery bypass graft, death from a cardiovascular cause, and death from any cause.

These studies prompted the ADA to emphasize that metformin can reduce the risk of cardiovascular events or death. Metformin also has been shown to be weight-neutral or to induce slight weight loss. Furthermore, it is inexpensive.

WHAT ABOUT THE RENAL EFFECTS?

Because metformin is renally cleared, it has caused some concern about nephrotoxicity, especially lactic acidosis, in patients with impaired renal function. But the most recent guidelines have relaxed the criteria for metformin use in this patient population.

Revised labeling

Metformin’s labeling,⁸ revised in 2016, states the following:

• If the estimated glomerular filtration rate (eGFR) is below 30 mL/min/1.73 m², metformin is contraindicated
• If the eGFR is between 30 and 45 mL/min/1.73 m², metformin is not recommended
• If the eGFR is below 45 mL/min/1.73 m² in a patient taking metformin, the risks and benefits of continuing treatment should be assessed, the dosage may need to be adjusted, and renal function should be monitored more frequently.⁸

These labeling revisions were based on a systematic review by Inzucchi et al⁹ that found metformin is not associated with increased rates of lactic acidosis in patients with mild to moderate kidney disease. Subsequently, an observational study published in 2018 by Lazarus et al¹⁰ showed that metformin increases the risk of acidosis only at eGFR levels below 30 mL/min/1.73 m². Also, a Cochrane review published in 2003 did not find a single case of lactic acidosis in 347 trials with 70,490 patient-years of metformin treatment.¹¹

Previous guidelines used serum creatinine levels, with metformin contraindicated at levels of 1.5 mg/dL or above for men and 1.4 mg/dL for women, or with abnormal creatinine clearance. The ADA and the AACE now use the eGFR^1,2 instead of the serum creatinine level to measure kidney function because it better accounts for factors such as the patient’s age, sex, race, and weight.

Despite the evidence, the common patient perception is that metformin is nephrotoxic, and it is important for practitioners to dispel this myth during clinic visits.

What about metformin use with contrast agents?

Labeling has a precautionary note stating that metformin should be held at the time of, or prior to, any imaging procedure involving iodinated contrast agents in patients with an eGFR between 30 and 60 mL/min/1.73 m²; in patients with a history of hepatic impairment, alcoholism, or heart failure; or in patients who will receive intra-arterial iodinated contrast. The eGFR should be reevaluated 48 hours after the imaging procedure.⁸

Additionally, if the iodinated contrast agent causes acute kidney injury, metformin could accumulate, with resultant lactate accumulation.

The American College of Radiology (ACR) has proposed less stringent guidelines for metformin during radiocontrast imaging studies. This change is based on evidence that lactic acidosis is rare­—about 10 cases per 100,000 patient-years—and that there are no reports of lactic acidosis after intravenously administered iodinated contrast in properly selected patients.^12,13

The ACR divides patients taking metformin into 2 categories:

• No evidence of acute kidney injury and eGFR greater than 30 mL/min/1.73 m²
• Either acute kidney injury or chronic kidney disease with eGFR below 30 mL/min/1.73 m² or undergoing arterial catheter studies with a high chance of embolization to the renal arteries.¹⁴

For the first group, they recommend against discontinuing metformin before or after giving iodinated contrast or checking kidney function after the procedure.

For the second group, they recommend holding metformin before and 48 hours after the procedure. It should not be restarted until renal function is confirmed to be normal.

 

 

METFORMIN AND INSULIN

The ADA recommends¹ continuing metformin after initiating insulin. However, in clinical practice, it is often not done.

Clinical trials have shown that combining metformin with insulin significantly improves glycemic control, prevents weight gain, and decreases insulin requirements.^15,16 One trial¹⁶ also looked at cardiovascular end points during a 4-year follow-up period;  combining metformin with insulin decreased the macrovascular disease-related event rate compared with insulin alone.

In the HOME trial,⁶ which added metformin to the existing insulin regimen, both groups gained weight, but the metformin group had gained about 3 kg less than the placebo group at the end of the 4.3-year trial. Metformin did not increase the risk of hypoglycemia, but it also did not reduce the risk of microvascular disease.

Concomitant metformin reduces costs

These days, practitioners can choose from a large selection of diabetes drugs. These include insulins with better pharmacokinetic profiles, as well as newer classes of noninsulin agents such as sodium-glucose cotransporter-2 inhibitors and glucagon-like peptide-1 analogues.

Metformin is less expensive than these newer drugs, and using it concomitantly with other diabetes drugs can decrease their dosage requirements, which in turn decreases their monthly costs.

GASTROINTESTINAL EFFECTS

Metformin’s gastrointestinal adverse effects such as diarrhea, flatulence, nausea, and vomiting are a barrier to its use. The actual incidence rate of diarrhea varies widely in randomized trials and observational studies, and gastrointestinal effects are worse in metformin-naive patients, as well as those who have chronic gastritis or Helicobacter pylori infection.¹⁷

We have found that starting metformin at a low dose and up-titrating it over several weeks increases tolerability. We often start patients at 500 mg/day and increase the dosage by 1 500-mg tablet every 1 to 2 weeks. Also, we have noticed that intolerance is more likely in patients who eat a high-carbohydrate diet, but there is no high-level evidence to back this up because patients in clinical trials all undergo nutrition counseling and are therefore more likely to adhere to the low-carbohydrate diet.

Also, the extended-release formulation is more tolerable than the immediate-release formulation and has similar glycemic efficacy. It may be an option as first-line therapy or for patients who have significant adverse effects from immediate-release metformin.¹⁸ For patients on the immediate-release formulation, taking it with meals helps lessen some gastrointestinal effects, and this should be emphasized at every visit.

Finally, we limit the metformin dose to 2,000 mg/day, rather than the 2,550 mg/day allowed on labeling. Garber et al¹⁹ found that the lower dosage still provides the maximum clinical efficacy.

OTHER CAUTIONS

Metformin should be avoided in patients with acute or unstable heart failure because of the increased risk of lactic acidosis.

It also should be avoided in patients with hepatic impairment, according to the labeling. But this remains controversial in practice. Zhang et al²⁰ showed that continuing metformin in patients with diabetes and cirrhosis decreases the mortality risk by 57% compared with those taken off metformin.

Diet and lifestyle measures need to be emphasized at each visit. Wing et al²¹ showed that calorie restriction regardless of weight loss is beneficial for glycemic control and insulin sensitivity in obese patients with diabetes.

TAKE-HOME POINTS

Metformin improves glycemic control without tending to cause weight gain or hypoglycemia. It may also have cardiovascular benefits. Metformin is an inexpensive agent that should be continued, if tolerated, in those who need additional agents for glycemic control. It should be considered in all adult patients with type 2 diabetes.   

While some classes of medications used to treat psychi­atric disorders, such as stimulants and benzodiazepines, are well-recognized as controlled substances and drugs of abuse, clinicians may be less familiar with the potential misuse/abuse of other psychiatric medications. This article reviews the evidence related to the misuse/abuse of anticholinergics, antidepressants, antipsychotics, and gabapentinoids.

The terms “misuse,” “abuse,” and “addiction” are used variably in the literature without standardized definitions. For this review, “misuse/abuse (M/A)” will be used to collectively describe self-administration that is recreational or otherwise inconsistent with legal or medical guidelines, unless a specific distinction is made. Whether or not the medications reviewed are truly “addictive” will be briefly discussed for each drug class, but the focus will be on clinically relevant aspects of M/A, including:

• excessive self-administration
• self-administration by non-oral routes
• co-administration with other drugs of abuse
• malingering of psychiatric symptoms to obtain prescriptions
• diversion for sale to third parties
• toxicity from overdose.

Anticholinergic medications

The first case describing the deliberate M/A of an anticholinergic medication for its euphoric effects was published in 1960.Further reportsfollowed in Europe before the M/A potential of prescription anticholinergic medications among psychiatric patients with an overdose syndrome characterized by atropinism and toxic psychosis was more widely recognized in the United States in the 1970s. Most reported cases of M/A to date have occurred among patients with psychiatric illness because anticholinergic medications, including trihexyphenidyl, benztropine, biperiden, procyclidine, and orphenadrine, were commonly prescribed for the management of first-generation and high dopamine D2-affinity antipsychotic-induced extrapyramidal symptoms (EPS). For example, one study of 234 consecutively hospitalized patients with schizophrenia noted an anticholinergic M/A incidence of 6.5%.¹

However, anticholinergic M/A is not limited to individuals with psychotic disorders. A UK study of 154 admissions to an inpatient unit specializing in behavioral disturbances found a 12-month trihexyphenidyl M/A incidence of 17%; the most common diagnosis among abusers was antisocial personality disorder.² Anticholinergic M/A has also been reported among patients with a primary diagnosis of substance use disorders (SUDs)³ as well as more indiscriminately in prison settings,⁴ with some inmates exchanging trihexyphenidyl as currency and using it recreationally by crushing it into powder and smoking it with tobacco.⁵ Others have noted that abusers sometimes take anticholinergics with alcohol in order to “potentiate” the effects of each substance.^6,7 Pullen et al⁸ described individuals with and without psychiatric illness who stole anticholinergic medications, purchased them from other patients, or bought them “on the street.” Malingering EPS in order to obtain anticholinergic medications has also been well documented.⁹ Clearly, anticholinergic M/A can occur in psychiatric and non-psychiatric populations, both within and outside of clinical settings. Although anticholinergic M/A appears to be less frequent in the United States now that second-generation antipsychotics (SGAs) are more frequently prescribed, M/A remains common in some settings outside of the United States.⁷

Among the various anticholinergic medications prescribed for EPS, trihexyphenidyl has been reported to have the greatest M/A potential, which has been attributed to its potency,¹⁰ its stimulating effects (whereas benztropine is more sedating),¹¹ and its former popularity among prescribers.⁸ Marken et al¹¹ published a review of 110 reports of M/A occurring in patients receiving anticholinergic medications as part of psychiatric treatment in which 69% of cases involved taking trihexyphenidyl 15 to 60 mg at a time (recommended dosing is 6 to 10 mg/d in divided doses).Most of these patients were prescribed anticholinergic medications for diagnostically appropriate reasons—only 7% were described as “true abusers” with no medical indication. Anticholinergic M/A was typically driven by a desire for euphoric and psychedelic/hallucinogenic effects, although in some cases, anticholinergic M/A was attributed to self-medication of EPS and depressive symptoms. These findings illustrate the blurred distinction between recreational use and perceived subjective benefit, and match those of a subsequent study of 50 psychiatric patients who reported anticholinergic M/A not only to “get high,” but to “decrease depression,” “increase energy,” and decrease antipsychotic adverse effects.¹² Once again, trihexyphenidyl was the most frequently misused anticholinergic in this sample.

Table 1^{2,3,7,8,10-15} outlines the subjective effects sought and experienced by anticholinergic abusers as well as potential toxic effects; there is the potential for overlap. Several authors have also described physiologic dependence with long-term trihexyphenidyl use, including tolerance and a withdrawal/abstinence syndrome.^7,16 In addition, there have been several reports of coma¹³ and death in the setting of intended suicide by overdose of anticholinergic medications.^14,15

Although anticholinergic M/A in the United States now appears to be less common, clinicians should remain aware of the M/A potential of anticholinergic medications prescribed for EPS. Management of M/A involves:

• detection
• reducing anticholinergic exposure by managing EPS with alternative strategies, such as switching or reducing the dose of the antipsychotic medication
• gradual tapering of anticholinergic medications to minimize withdrawal.¹¹

Continue to: Antidepressants

 

 

Antidepressants

Haddad¹⁷ published a review of 21 English-language case reports from 1966 to 1998 describing antidepressant use in which individuals met DSM-IV criteria for substance dependence to the medication. An additional 14 cases of antidepressant M/A were excluded based on insufficient details to support a diagnosis of dependence. The 21 reported cases involved:

• tranylcypromine (a monoamine oxidase inhibitor [MAOI])
• amitriptyline (a tricyclic antidepressant [TCA])
• fluoxetine (a selective serotonin reuptake inhibitor [SSRI])
• amineptine (a TCA previously available in France but removed from the market in 1999 in part due to its abuse potential)
• nomifensine (a norepinephrine/dopamine reuptake inhibitor previously available in the United Kingdom but removed in 1986 due to hemolytic anemia).

In 95% of cases, the antidepressants were prescribed for treatment of an affective disorder but were abused for stimulant effects or the perceived ability to lift mood, cause euphoria or a “high,” or to improve functioning. Two-thirds of cases involved patients with preexisting substance misuse. Placing the case reports in the context of the millions of patients prescribed antidepressants during this period, Haddad concluded the “incidence of [antidepressant] addiction [is] so low as to be clinically irrelevant.”¹⁷

Despite this conclusion, Haddad singled out amineptine and tranylcypromine as antidepressants with some evidence of true addictive potential.^17,18 A more recent case series described 14 patients who met DSM-IV criteria for substance abuse of tertiary amine TCAs (which have strong anticholinergic activity) and concluded that “misuse of [TCAs] is more common than generally appreciated.”¹⁹ In keeping with that claim, a study of 54 outpatients taking unspecified antidepressants found that up to 15% met DSM-III-R criteria for substance dependence (for the antidepressant) in the past year, although that rate was much lower than the rate of benzodiazepine dependence (47%) in a comparative sample.²⁰ Finally, a comprehensive review by Evans and Sullivan²¹ found anecdotal reports published before 2014 that detailed misuse, abuse, and dependence with MAOIs, TCAs, fluoxetine, venlafaxine, bupropion, tianeptine, and amineptine. Taken together, existing evidence indicates that select individuals—typically those with other SUD comorbidity—sometimes misuse antidepressants in a way that suggests addiction.

Still, while it is well known that abrupt cessation of antidepressants can result in a discontinuation syndrome characterized by flu-like symptoms, nausea, and dizziness,²² physiologic withdrawal effects must be distinguished from historical definitions of substance “abuse” and the broader concept of psychological “addiction” or drug dependence^18,23 now incorporated into the DSM-5 definition of SUDs.²⁴ Indeed, although withdrawal symptoms were reported by more than half of those who took antidepressants and responded to a recent online survey,²⁵ evidence to support the existence of significant antidepressant tolerance, craving, or compulsive use is lacking.^17,18 Antidepressants as a class do not appear to be significantly rewarding or reinforcing and, on the contrary, discontinuation by patients is common in clinical practice.²⁶ The popular claim that some individuals taking antidepressants “can’t quit”²⁷ must also be disentangled from loss of therapeutic effects upon cessation.

Bupropion. A more convincing argument for antidepressant addiction can be made for bupropion, a weak norepinephrine and dopamine reuptake inhibitor with an otherwise unclear mechanism of action.²⁸ In 2002, the first report of recreational bupropion M/A described a 13-year-old girl who took 2,400 mg orally (recommended maximum dose is 450 mg/d in divided doses) after being told it would give her “a better high than amphetamine.”²⁹ This was followed in the same year by the first report of recreational M/A of bupropion via nasal insufflation (snorting), resulting in a seizure,³⁰ and in 2013 by the first published case of M/A by IV self-administration.³¹

Continue to: The M/A potential of bupropion...

The M/A potential of bupropion, most commonly via intranasal administration, is now broadly recognized based on several case reports describing desired effects that include a euphoric high and a stimulating “buzz” similar to that of cocaine or methamphetamine but less intense.^29-36 Among recreational users, bupropion tablets are referred to as “welbys,” “wellies,” “dubs,” or “barnies.”³⁷ Media coverage of a 2013 outbreak of bupropion M/A in Toronto detailed administration by snorting, smoking, and injection, and described bupropion as “poor man’s cocaine.”³⁸ Between 2003 and 2016, 2,232 cases of bupropion misuse/abuse/dependence adverse drug reactions were reported to the European Monitoring Agency.³⁷ A review of intentional bupropion M/A reported to US Poison Control Centers between 2000 to 2013 found 975 such cases, with the yearly number tripling between 2000 and 2012.³⁹ In this sample, nearly half (45%) of the users were age 13 to 19, and 76% of cases involved oral ingestion. In addition to bupropion M/A among younger people, individuals who misuse bupropion often include those with existing SUDs but limited access to illicit stimulants and those trying to evade detection by urine toxicology screening.³³ For example, widespread use and diversion has been well documented within correctional settings, and as a result, many facilities have removed bupropion from their formularies.^{21,28,33,34,40}

Beyond desired effects, the most common adverse events associated with bupropion M/A are listed in Table 2,^{28,30,32-34,36,39} along with their incidence based on cases brought to the attention of US Poison Control Centers.³⁹ With relatively little evidence of a significant bupropion withdrawal syndrome,³⁷ the argument in favor of modeling bupropion as a truly addictive drug is limited to anecdotal reports of cravings and compulsive self-administration³⁵ and pro-dopaminergic activity (reuptake inhibition) that might provide a mechanism for potential rewarding and reinforcing effects.⁴⁰ While early preclinical studies of bupropion failed to provide evidence of amphetamine-like abuse potential,^41,42 non-oral administration in amounts well beyond therapeutic dosing could account for euphoric effects and a greater risk of psychological dependence and addiction.^21,28,40

Bupropion also has an FDA indication as an aid to smoking cessation treatment, and the medication demonstrated early promise in the pharmacologic treatment of psycho­stimulant use disorders, with reported improvements in cravings and other SUD outcomes.^43-45 However, subsequent randomized controlled trials (RCTs) failed to demonstrate a clear therapeutic role for bupropion in the treatment of cocaine^46,47 and methamphetamine use disorders (although some secondary analyses suggest possible therapeutic effects among non-daily stimulant users who are able to maintain good adherence with bupropion).^48-51 Given these overall discouraging results, the additive seizure risk of bupropion use with concomitant psychostimulant use, and the potential for M/A and diversion of bupropion (particularly among those with existing SUDs), the use of bupropion for the off-label treatment of stimulant use disorders is not advised.

Antipsychotics

As dopamine antagonists, antipsychotics are typically considered to have low potential for rewarding or reinforcing effects. Indeed, misuse of antipsychotics was a rarity in the first-generation era, with only a few published reports of haloperidol M/A within a small cluster of naïve young people who developed acute EPS,⁵² and a report of diversion in a prison with the “sadistic” intent of inflicting dystonic reactions on others.⁵³ A more recent report described 2additional cases of M/A involving haloperidol and trifluoperazine.⁵⁴ Some authors have described occasional drug-seeking behavior for low-potency D2 blockers such as chlorpromazine, presumably based on their M/A as anticholinergic medications.⁵⁵

The potential for antipsychotic M/A has gained wider recognition since the advent of the SGAs. Three cases of prescription olanzapine M/A have been published to date. One involved a man who malingered manic symptoms to obtain olanzapine, taking ≥40 mg at a time (beyond his prescribed dose of 20 mg twice daily) to get a “buzz,” and combining it with alcohol and benzodiazepines for additive effects or to “come down” from cocaine.⁵⁶ This patient noted that olanzapine was “a popular drug at parties” and was bought, sold, or traded among users, and occasionally administered intravenously. Two other cases described women who self-administered olanzapine, 40 to 50 mg/d, for euphoric and anxiolytic effects.^57,58 James et al⁵⁹ detailed a sample of 28 adults who reported “non-medical use” of olanzapine for anxiolytic effects, as a sleep aid, or to “escape from worries.”

Continue to: Quetiapine

Quetiapine. In contrast to some reports of olanzapine M/A in which the line between M/A and “self-medication” was blurred, quetiapine has become a more convincing example of clear recreational antipsychotic M/A. Since the first report of oral and intranasal quetiapine M/A in the Los Angeles County Jail published in 2004,⁵⁵ subsequent cases have detailed other novel methods of recreational self-administration^60-68 (Table 3^55,60-68), and additional reports have been published in non-English language journals.^69,70 Collectively, these case reports have detailed that quetiapine is:

• misused for primary subjective effects as well as to mitigate the unpleasant effects of other drugs^60,67
• referred to as “quell,”“Q,” “Susie-Q,” “squirrel,” and “baby heroin”^55,71,72
• often obtained by malingering psychiatric symptoms^55,61,63,65
• diverted/sold with “street value” both within and outside of psychiatric facilities and correctional settings.^{55,60-62,67,68,73}

These anecdotal accounts of quetiapine M/A have since been corroborated on a larger scale based on several retrospective studies. Although early reports of quetiapine M/A occurring in correctional settings have resulted in formulary removal,^71,74 quetiapine M/A is by no means limited to forensic populations and is especially common among those with comorbid SUDs. A survey of 74 patients enrolled in a Canadian methadone program reported that nearly 60% had misused quetiapine at some point.⁷⁵ Among an Australian sample of 868 individuals with active IV drug abuse, 31% reported having misused quetiapine.⁷⁶ Finally, within a small sample of patients with SUDs admitted to a detoxification unit in New York City, 17% reported M/A of SGAs.⁷⁷ In this study, SGAs were often taken in conjunction with other drugs of abuse in order to “recover” from or “enhance” the effects of other drugs or to “experiment.” Quetiapine was by far the most frequently abused SGA, reported in 96% of the sample; the most frequently reported SGA/drug combinations were quetiapine/alcohol/opioids, quetiapine/cocaine, and quetiapine/opioids.

Looking more broadly at poison center data, reports to the US National Poison Data System (NPDS) from 2005 to 2011 included 3,116 cases of quetiapine abuse (37.5%, defined as intentional recreational use in order to obtain a “high”) or misuse (62.5%, defined as improper use or dosing for non-recreational purposes).⁷⁸ A more recent analysis of NPDS reports from 2003 to 2013 found 2,118 cases of quetiapine abuse, representing 61% of all cases of reported SGA abuse.⁷⁹ An analysis of the European Medicines Agency Adverse Drug Database yielded 18,112 reports of quetia­pine misuse, abuse, dependence, and withdrawal for quetiapine (from 2005 to 2016) compared with 4,178 for olanzapine (from 2004 to 2016).⁸⁰ These reports identified 368 fatalities associated with quetiapine.

The rate of quetiapine M/A appears to be increasing sharply. Reports of quetiapine M/A to poison centers in Australia increased nearly 7-fold from 2006 to 2016.⁸¹ Based on reports to the Drug Abuse Warning System, US emergency department visits for M/A of quetiapine increased from 19,195 in 2005 to 32,024 in 2011 (an average of 27,114 visits/year), with 75% of cases involving quetiapine taken in combination with other prescription drugs, alcohol, or illicit drugs.⁸² Consistent with poison center data, M/A was reported for other antipsychotics, but none nearly as frequently as for quetiapine.

With increasingly frequent quetiapine M/A, clinicians should be vigilant in monitoring for medical morbidity related to quetiapine and cumulative toxicity with other drugs. The most frequent adverse events associated with quetiapine M/A reported to US Poison Control Centers are presented in Table 4.^78,79

Continue to: Unlike bupropion...

Unlike bupropion, quetiapine’s dopamine antagonism makes it unlikely to be a truly addictive drug, although this mechanism of action could mediate an increase in concurrent psychostimulant use.⁸³ A few case reports have described a quetiapine discontinuation syndrome similar to that of antidepressants,^60,65,84-88 but withdrawal symptoms suggestive of physiologic dependence may be mediated by non-dopaminergic effects through histamine and serotonin receptors.^84,89 Evidence for quetiapine misuse being associated with craving and compulsive use is lacking, and true quetiapine addiction is probably rare.

Similar to bupropion, preliminary findings have suggested promise for quetiapine as a putative therapy for other SUDs.^90-93 However, subsequent RCTs have failed to demonstrate a therapeutic effect for alcohol and cocaine use disorders.^94-96 Given these negative results and the clear M/A potential of quetiapine, off-label use of quetiapine for the treatment of SUDs and psychiatric symptoms among those with SUDs must be considered judiciously, with an eye towards possible diversion and avoiding the substitution of one drug of abuse for another.

Gabapentinoids

In 1997, the first published case report of gabapentin M/A described a woman who self-administered her husband’s gabapentin to reduce cravings for and withdrawal from cocaine.⁹⁷ The authors highlighted the possible therapeutic benefit of gabapentin in this regard rather than raising concerns about diversion and M/A. By 2004, however, reports of recreational gabapentin M/A emerged among inmates incarcerated within Florida correctional facilities who self-administered intranasal gabapentin to achieve a “high” that was “reminiscent of prior effects from intranasal ingestion of cocaine powder.”⁹⁸ In 2007, a single case of gabapentin misuse up to 7,200 mg/d (recommended dosing is ≤3,600 mg/d) was reported, with documentation of both tolerance and withdrawal symptoms.⁹⁹ As of 2017, a total of 36 cases of gabapentin M/A and 19 cases of pregabalin M/A have been published.¹⁰⁰

In the past decade, anecdotal reports have given way to larger-scale epidemiologic data painting a clear picture of the now-widespread M/A of gabapentin and other gabapentinoids. For example, a study of online descriptions of gabapentin and pregabalin M/A from 2008 to 2010 documented:

• oral and IM use (gabapentin)
• IV and rectal (“plugging”) use (pregabalin)
• “parachuting” (emptying the contents of capsules for a larger dose) (pregabalin)
• euphoric, entactogenic, stimulant, calming/anxiolytic, and dissociative subjective effects (gabapentin/pregabalin)
• rapid development of tolerance to euphoric effects leading to self-administration of increasing doses (gabapentin/pregabalin)
• frequent co-administration with other drugs of abuse, including alcohol, benzodiazepines, cannabis, stimulants, opiates, hallucinogens, gamma-hydroxybutyrate, mephedrone, and Salvia divinorum (gabapentin/pregabalin)¹⁰¹

Several systematic reviews of both anecdotal reports and epidemiologic studies published in the past few years provide additional evidence of the above, such as:

• excessive dosing with self-administration
• intranasal and inhaled routes of administration
• diversion and “street value”
• greater M/A potential of pregabalin than gabapentin
• the presence of gabapentinoids in postmortem toxicology analyses, suggesting a role in overdose fatalities when combined with other drugs.^100,102,103

Continue to: The European Medicine Agency's EudraVigilance database...

The European Medicine Agency’s EudraVigilance database included 4,301 reports of gabapentin misuse, abuse, or dependence, and 7,639 such reports for pregabalin, from 2006 to 2015 (rising sharply after 2012), with 86 gabapentin-related and 27 pregabalin-related fatalities.¹⁰⁴ Data from the Drug Diversion Program of the Researched Abuse, Diversion, and Addiction-Related Surveillance System from 2002 to 2015 have likewise revealed that gabapentin diversion increased significantly in 2013.¹⁰⁵

While the prevalence of gabapentinoid M/A is not known, rates appear to be significantly lower than for traditional drugs of abuse such as cannabis, cocaine, 3,4-methylenedioxymethamphetamine (MDMA), and opioids.^106,107 However, gabapentin and pregabalin M/A appears to be increasingly common among individuals with SUDs and in particular among those with opioid use disorders (OUDs). For example, a 2015 report indicated that 15% of an adult cohort in Appalachian Kentucky with nonmedical use of diverted prescription opioids reported gabapentin M/A, an increase of nearly 3,000% since 2008.¹⁰⁸ Based on data from a US insurance enrollment and claims database, researchers found that the rate of gabapentin overuse among those also overusing opioids was 12% compared with only 2% for those using gabapentin alone.¹⁰⁹ It has also been reported that gabapentin is sometimes used as a “cutting agent” for heroin.¹¹⁰

Those who use gabapentinoids together with opioids report that gabapentin and pregabalin potentiate the euphoric effects of methadone¹¹¹ and endorse specific beliefs that pregabalin increases both the desired effects of heroin as well as negative effects such as “blackouts,” loss of control, and risk of overdose.¹¹² Indeed, sustained M/A of gabapentin and opioids together has been found to increase emergency department utilization, drug-related hospitalization, and respiratory depression.¹¹³ Based on a case-control study of opioid users in Canada, co-prescription of gabapentin and opioids was associated with a 50% increase in death from opioid-related causes compared with prescription of opioids alone.¹¹⁴

Case reports documenting tolerance, withdrawal, craving, and loss of control suggest a true addictive potential for gabapentinoids, but Bonnet and Sherbaum¹⁰⁰ concluded that while there is robust evidence of abusers “liking” gabapentin and pregabalin (eg, reward), evidence of “wanting” them (eg, psychological dependence) in the absence of other SUDs has been limited to only a few anecdotal reports with pregabalin. Accordingly, the risk of true addiction to gabapentinoids by those without preexisting SUDs appears to be low. Nonetheless, the M/A potential of both gabapentin and pregabalin is clear and in the context of a nationwide opioid epidemic, the increased morbidity/mortality risk related to combined use of gabapentinoids and opioids is both striking and concerning. Consequently, the state of Kentucky recently recognized the M/A potential of gabapentin by designating it a Schedule V controlled substance (pregabalin is already a Schedule V drug according to the US Drug Enforcement Agency),^103,113 and several other states now mandate the reporting of gabapentin prescriptions to prescription drug monitoring programs.¹¹⁵

Following a similar pattern to antidepressants and antipsychotics, a potential role for gabapentin in the treatment of cocaine use disorders was supported in preliminary studies,^116-118 but not in subsequent RCTs.^119-121 However, there is evidence from RCTs to support the use of gabapentin and pregabalin in the treatment of alcohol use disorders.^122-124 Gabapentin was also found to significantly reduce cannabis use and withdrawal symptoms in patients compared with placebo in an RCT of individuals with cannabis use disorders.¹²⁵ The perceived safety of gabapentinoids by clinicians, their subjective desirability by patients with SUDs, and efficacy data supporting a therapeutic role in SUDs must be balanced with recognition that approximately 80% of gabapentin prescriptions are written for off-label indications for which there is little supporting evidence,¹⁰⁹ such as low back pain.¹²⁶ Clinicians considering prescribing gabapentinoids to manage psychiatric symptoms, such as anxiety and insomnia, should carefully consider the risk of M/A and other potential morbidities, especially in the setting of SUDs and OUD in particular.

Continue to: Problematic, even if not addictive

Problematic, even if not addictive

It is sometimes claimed that “addiction” to psychiatric medications is not limited to stimulants and benzodiazepines.^27,127 Although anticholinergics, antidepressants, antipsychotics, and gabapentinoids can be drugs of abuse, with some users reporting physiologic withdrawal upon discontinuation, there is only limited evidence that the M/A of these psychiatric medications is associated with the characteristic features of a more complete definition of “addiction,” which may include:

• inability to consistently abstain
• impairment in behavioral control
• diminished recognition of significant problems associated with use
• a dysfunctional emotional response to chronic use.¹²⁸

Nonetheless, the literature documenting anticholinergic, antidepressant, antipsychotic, and gabapentinoid M/A includes several common features, including:

• initial reports among those with limited access to illicit drugs (eg, young people and incarcerated individuals) and subsequent spread to a wider population with more unconventional routes of administration
• use for recreational purposes and other subjective pseudo-therapeutic effects, often in combination with alcohol and illicit drugs
• greater M/A potential of certain medications within each of these drug classes (eg, trihexyphenidyl, bupropion, quetiapine)
• malingering psychiatric symptoms in order to obtain medications from prescribers and diversion for black market sale
• observations that medications might constitute therapy for SUDs that were not supported in subsequent RCTs (with the exception of gabapentin for alcohol and cannabis use disorders)
• increasing evidence of toxicity related to M/A, which suggests that prescription by clinicians has limited benefit and high risk for patients with SUDs.

Bottom Line

Some psychiatric medications are taken as drugs of abuse. Clinicians should be particularly aware of the misuse/abuse potential of anticholinergics, antidepressants, antipsychotics, and gabapentinoids, and use them cautiously, if at all, when treating patients with existing substance use disorders.

Related Resources

• Substance Abuse and Mental Health Services Administration. Prescription drug misuse and abuse. https://www.samhsa.gov/topics/prescription-drug-misuse-abuse.
• Substance Abuse and Mental Health Services Administration. Types of commonly misused or abused drugs. https://www.samhsa.gov/prescription-drug-misuse-abuse/types.
• National Institute on Drug Abuse. Misuse of prescription drugs. https://www.drugabuse.gov/publications/research-reports/misuse-prescription-drugs/summary.
• National Institute on Drug Abuse. New clinician screening tool available for substance use. https://www.drugabuse.gov/news-events/news-releases/2018/06/newclinician-screening-tool-available-substance-use.

Drug Brand Names

Amitriptyline • Elavil, Endep
Benztropine • Cogentin
Biperiden • Akineton
Bupropion • Wellbutrin, Zyban
Chlorpromazine • Thorzine
Fluoxetine • Prozac
Haloperidol • Haldol
Olanzapine • Zyprexa
Orphenadrine • Disipal, Norflex
Pregabalin • Lyrica, Lyrica CR
Procyclidine • Kemadrin
Quetiapine • Seroquel
Tianeptine • Coaxil, Stablon
Tranylcypromine • Parnate
Trifluoperazine • Stelazine
Trihexyphenidyl • Artane, Tremin
Venlafaxine • Effexor

While some classes of medications used to treat psychi­atric disorders, such as stimulants and benzodiazepines, are well-recognized as controlled substances and drugs of abuse, clinicians may be less familiar with the potential misuse/abuse of other psychiatric medications. This article reviews the evidence related to the misuse/abuse of anticholinergics, antidepressants, antipsychotics, and gabapentinoids.

The terms “misuse,” “abuse,” and “addiction” are used variably in the literature without standardized definitions. For this review, “misuse/abuse (M/A)” will be used to collectively describe self-administration that is recreational or otherwise inconsistent with legal or medical guidelines, unless a specific distinction is made. Whether or not the medications reviewed are truly “addictive” will be briefly discussed for each drug class, but the focus will be on clinically relevant aspects of M/A, including:

• excessive self-administration
• self-administration by non-oral routes
• co-administration with other drugs of abuse
• malingering of psychiatric symptoms to obtain prescriptions
• diversion for sale to third parties
• toxicity from overdose.

Anticholinergic medications

The first case describing the deliberate M/A of an anticholinergic medication for its euphoric effects was published in 1960.Further reportsfollowed in Europe before the M/A potential of prescription anticholinergic medications among psychiatric patients with an overdose syndrome characterized by atropinism and toxic psychosis was more widely recognized in the United States in the 1970s. Most reported cases of M/A to date have occurred among patients with psychiatric illness because anticholinergic medications, including trihexyphenidyl, benztropine, biperiden, procyclidine, and orphenadrine, were commonly prescribed for the management of first-generation and high dopamine D2-affinity antipsychotic-induced extrapyramidal symptoms (EPS). For example, one study of 234 consecutively hospitalized patients with schizophrenia noted an anticholinergic M/A incidence of 6.5%.¹

However, anticholinergic M/A is not limited to individuals with psychotic disorders. A UK study of 154 admissions to an inpatient unit specializing in behavioral disturbances found a 12-month trihexyphenidyl M/A incidence of 17%; the most common diagnosis among abusers was antisocial personality disorder.² Anticholinergic M/A has also been reported among patients with a primary diagnosis of substance use disorders (SUDs)³ as well as more indiscriminately in prison settings,⁴ with some inmates exchanging trihexyphenidyl as currency and using it recreationally by crushing it into powder and smoking it with tobacco.⁵ Others have noted that abusers sometimes take anticholinergics with alcohol in order to “potentiate” the effects of each substance.^6,7 Pullen et al⁸ described individuals with and without psychiatric illness who stole anticholinergic medications, purchased them from other patients, or bought them “on the street.” Malingering EPS in order to obtain anticholinergic medications has also been well documented.⁹ Clearly, anticholinergic M/A can occur in psychiatric and non-psychiatric populations, both within and outside of clinical settings. Although anticholinergic M/A appears to be less frequent in the United States now that second-generation antipsychotics (SGAs) are more frequently prescribed, M/A remains common in some settings outside of the United States.⁷

Among the various anticholinergic medications prescribed for EPS, trihexyphenidyl has been reported to have the greatest M/A potential, which has been attributed to its potency,¹⁰ its stimulating effects (whereas benztropine is more sedating),¹¹ and its former popularity among prescribers.⁸ Marken et al¹¹ published a review of 110 reports of M/A occurring in patients receiving anticholinergic medications as part of psychiatric treatment in which 69% of cases involved taking trihexyphenidyl 15 to 60 mg at a time (recommended dosing is 6 to 10 mg/d in divided doses).Most of these patients were prescribed anticholinergic medications for diagnostically appropriate reasons—only 7% were described as “true abusers” with no medical indication. Anticholinergic M/A was typically driven by a desire for euphoric and psychedelic/hallucinogenic effects, although in some cases, anticholinergic M/A was attributed to self-medication of EPS and depressive symptoms. These findings illustrate the blurred distinction between recreational use and perceived subjective benefit, and match those of a subsequent study of 50 psychiatric patients who reported anticholinergic M/A not only to “get high,” but to “decrease depression,” “increase energy,” and decrease antipsychotic adverse effects.¹² Once again, trihexyphenidyl was the most frequently misused anticholinergic in this sample.

Table 1^{2,3,7,8,10-15} outlines the subjective effects sought and experienced by anticholinergic abusers as well as potential toxic effects; there is the potential for overlap. Several authors have also described physiologic dependence with long-term trihexyphenidyl use, including tolerance and a withdrawal/abstinence syndrome.^7,16 In addition, there have been several reports of coma¹³ and death in the setting of intended suicide by overdose of anticholinergic medications.^14,15

Although anticholinergic M/A in the United States now appears to be less common, clinicians should remain aware of the M/A potential of anticholinergic medications prescribed for EPS. Management of M/A involves:

• detection
• reducing anticholinergic exposure by managing EPS with alternative strategies, such as switching or reducing the dose of the antipsychotic medication
• gradual tapering of anticholinergic medications to minimize withdrawal.¹¹

Continue to: Antidepressants

 

 

Antidepressants

Haddad¹⁷ published a review of 21 English-language case reports from 1966 to 1998 describing antidepressant use in which individuals met DSM-IV criteria for substance dependence to the medication. An additional 14 cases of antidepressant M/A were excluded based on insufficient details to support a diagnosis of dependence. The 21 reported cases involved:

• tranylcypromine (a monoamine oxidase inhibitor [MAOI])
• amitriptyline (a tricyclic antidepressant [TCA])
• fluoxetine (a selective serotonin reuptake inhibitor [SSRI])
• amineptine (a TCA previously available in France but removed from the market in 1999 in part due to its abuse potential)
• nomifensine (a norepinephrine/dopamine reuptake inhibitor previously available in the United Kingdom but removed in 1986 due to hemolytic anemia).

In 95% of cases, the antidepressants were prescribed for treatment of an affective disorder but were abused for stimulant effects or the perceived ability to lift mood, cause euphoria or a “high,” or to improve functioning. Two-thirds of cases involved patients with preexisting substance misuse. Placing the case reports in the context of the millions of patients prescribed antidepressants during this period, Haddad concluded the “incidence of [antidepressant] addiction [is] so low as to be clinically irrelevant.”¹⁷

Despite this conclusion, Haddad singled out amineptine and tranylcypromine as antidepressants with some evidence of true addictive potential.^17,18 A more recent case series described 14 patients who met DSM-IV criteria for substance abuse of tertiary amine TCAs (which have strong anticholinergic activity) and concluded that “misuse of [TCAs] is more common than generally appreciated.”¹⁹ In keeping with that claim, a study of 54 outpatients taking unspecified antidepressants found that up to 15% met DSM-III-R criteria for substance dependence (for the antidepressant) in the past year, although that rate was much lower than the rate of benzodiazepine dependence (47%) in a comparative sample.²⁰ Finally, a comprehensive review by Evans and Sullivan²¹ found anecdotal reports published before 2014 that detailed misuse, abuse, and dependence with MAOIs, TCAs, fluoxetine, venlafaxine, bupropion, tianeptine, and amineptine. Taken together, existing evidence indicates that select individuals—typically those with other SUD comorbidity—sometimes misuse antidepressants in a way that suggests addiction.

Still, while it is well known that abrupt cessation of antidepressants can result in a discontinuation syndrome characterized by flu-like symptoms, nausea, and dizziness,²² physiologic withdrawal effects must be distinguished from historical definitions of substance “abuse” and the broader concept of psychological “addiction” or drug dependence^18,23 now incorporated into the DSM-5 definition of SUDs.²⁴ Indeed, although withdrawal symptoms were reported by more than half of those who took antidepressants and responded to a recent online survey,²⁵ evidence to support the existence of significant antidepressant tolerance, craving, or compulsive use is lacking.^17,18 Antidepressants as a class do not appear to be significantly rewarding or reinforcing and, on the contrary, discontinuation by patients is common in clinical practice.²⁶ The popular claim that some individuals taking antidepressants “can’t quit”²⁷ must also be disentangled from loss of therapeutic effects upon cessation.

Bupropion. A more convincing argument for antidepressant addiction can be made for bupropion, a weak norepinephrine and dopamine reuptake inhibitor with an otherwise unclear mechanism of action.²⁸ In 2002, the first report of recreational bupropion M/A described a 13-year-old girl who took 2,400 mg orally (recommended maximum dose is 450 mg/d in divided doses) after being told it would give her “a better high than amphetamine.”²⁹ This was followed in the same year by the first report of recreational M/A of bupropion via nasal insufflation (snorting), resulting in a seizure,³⁰ and in 2013 by the first published case of M/A by IV self-administration.³¹

Continue to: The M/A potential of bupropion...

The M/A potential of bupropion, most commonly via intranasal administration, is now broadly recognized based on several case reports describing desired effects that include a euphoric high and a stimulating “buzz” similar to that of cocaine or methamphetamine but less intense.^29-36 Among recreational users, bupropion tablets are referred to as “welbys,” “wellies,” “dubs,” or “barnies.”³⁷ Media coverage of a 2013 outbreak of bupropion M/A in Toronto detailed administration by snorting, smoking, and injection, and described bupropion as “poor man’s cocaine.”³⁸ Between 2003 and 2016, 2,232 cases of bupropion misuse/abuse/dependence adverse drug reactions were reported to the European Monitoring Agency.³⁷ A review of intentional bupropion M/A reported to US Poison Control Centers between 2000 to 2013 found 975 such cases, with the yearly number tripling between 2000 and 2012.³⁹ In this sample, nearly half (45%) of the users were age 13 to 19, and 76% of cases involved oral ingestion. In addition to bupropion M/A among younger people, individuals who misuse bupropion often include those with existing SUDs but limited access to illicit stimulants and those trying to evade detection by urine toxicology screening.³³ For example, widespread use and diversion has been well documented within correctional settings, and as a result, many facilities have removed bupropion from their formularies.^{21,28,33,34,40}

Beyond desired effects, the most common adverse events associated with bupropion M/A are listed in Table 2,^{28,30,32-34,36,39} along with their incidence based on cases brought to the attention of US Poison Control Centers.³⁹ With relatively little evidence of a significant bupropion withdrawal syndrome,³⁷ the argument in favor of modeling bupropion as a truly addictive drug is limited to anecdotal reports of cravings and compulsive self-administration³⁵ and pro-dopaminergic activity (reuptake inhibition) that might provide a mechanism for potential rewarding and reinforcing effects.⁴⁰ While early preclinical studies of bupropion failed to provide evidence of amphetamine-like abuse potential,^41,42 non-oral administration in amounts well beyond therapeutic dosing could account for euphoric effects and a greater risk of psychological dependence and addiction.^21,28,40

Bupropion also has an FDA indication as an aid to smoking cessation treatment, and the medication demonstrated early promise in the pharmacologic treatment of psycho­stimulant use disorders, with reported improvements in cravings and other SUD outcomes.^43-45 However, subsequent randomized controlled trials (RCTs) failed to demonstrate a clear therapeutic role for bupropion in the treatment of cocaine^46,47 and methamphetamine use disorders (although some secondary analyses suggest possible therapeutic effects among non-daily stimulant users who are able to maintain good adherence with bupropion).^48-51 Given these overall discouraging results, the additive seizure risk of bupropion use with concomitant psychostimulant use, and the potential for M/A and diversion of bupropion (particularly among those with existing SUDs), the use of bupropion for the off-label treatment of stimulant use disorders is not advised.

Antipsychotics

As dopamine antagonists, antipsychotics are typically considered to have low potential for rewarding or reinforcing effects. Indeed, misuse of antipsychotics was a rarity in the first-generation era, with only a few published reports of haloperidol M/A within a small cluster of naïve young people who developed acute EPS,⁵² and a report of diversion in a prison with the “sadistic” intent of inflicting dystonic reactions on others.⁵³ A more recent report described 2additional cases of M/A involving haloperidol and trifluoperazine.⁵⁴ Some authors have described occasional drug-seeking behavior for low-potency D2 blockers such as chlorpromazine, presumably based on their M/A as anticholinergic medications.⁵⁵

The potential for antipsychotic M/A has gained wider recognition since the advent of the SGAs. Three cases of prescription olanzapine M/A have been published to date. One involved a man who malingered manic symptoms to obtain olanzapine, taking ≥40 mg at a time (beyond his prescribed dose of 20 mg twice daily) to get a “buzz,” and combining it with alcohol and benzodiazepines for additive effects or to “come down” from cocaine.⁵⁶ This patient noted that olanzapine was “a popular drug at parties” and was bought, sold, or traded among users, and occasionally administered intravenously. Two other cases described women who self-administered olanzapine, 40 to 50 mg/d, for euphoric and anxiolytic effects.^57,58 James et al⁵⁹ detailed a sample of 28 adults who reported “non-medical use” of olanzapine for anxiolytic effects, as a sleep aid, or to “escape from worries.”

Continue to: Quetiapine

Quetiapine. In contrast to some reports of olanzapine M/A in which the line between M/A and “self-medication” was blurred, quetiapine has become a more convincing example of clear recreational antipsychotic M/A. Since the first report of oral and intranasal quetiapine M/A in the Los Angeles County Jail published in 2004,⁵⁵ subsequent cases have detailed other novel methods of recreational self-administration^60-68 (Table 3^55,60-68), and additional reports have been published in non-English language journals.^69,70 Collectively, these case reports have detailed that quetiapine is:

• misused for primary subjective effects as well as to mitigate the unpleasant effects of other drugs^60,67
• referred to as “quell,”“Q,” “Susie-Q,” “squirrel,” and “baby heroin”^55,71,72
• often obtained by malingering psychiatric symptoms^55,61,63,65
• diverted/sold with “street value” both within and outside of psychiatric facilities and correctional settings.^{55,60-62,67,68,73}

These anecdotal accounts of quetiapine M/A have since been corroborated on a larger scale based on several retrospective studies. Although early reports of quetiapine M/A occurring in correctional settings have resulted in formulary removal,^71,74 quetiapine M/A is by no means limited to forensic populations and is especially common among those with comorbid SUDs. A survey of 74 patients enrolled in a Canadian methadone program reported that nearly 60% had misused quetiapine at some point.⁷⁵ Among an Australian sample of 868 individuals with active IV drug abuse, 31% reported having misused quetiapine.⁷⁶ Finally, within a small sample of patients with SUDs admitted to a detoxification unit in New York City, 17% reported M/A of SGAs.⁷⁷ In this study, SGAs were often taken in conjunction with other drugs of abuse in order to “recover” from or “enhance” the effects of other drugs or to “experiment.” Quetiapine was by far the most frequently abused SGA, reported in 96% of the sample; the most frequently reported SGA/drug combinations were quetiapine/alcohol/opioids, quetiapine/cocaine, and quetiapine/opioids.

Looking more broadly at poison center data, reports to the US National Poison Data System (NPDS) from 2005 to 2011 included 3,116 cases of quetiapine abuse (37.5%, defined as intentional recreational use in order to obtain a “high”) or misuse (62.5%, defined as improper use or dosing for non-recreational purposes).⁷⁸ A more recent analysis of NPDS reports from 2003 to 2013 found 2,118 cases of quetiapine abuse, representing 61% of all cases of reported SGA abuse.⁷⁹ An analysis of the European Medicines Agency Adverse Drug Database yielded 18,112 reports of quetia­pine misuse, abuse, dependence, and withdrawal for quetiapine (from 2005 to 2016) compared with 4,178 for olanzapine (from 2004 to 2016).⁸⁰ These reports identified 368 fatalities associated with quetiapine.

The rate of quetiapine M/A appears to be increasing sharply. Reports of quetiapine M/A to poison centers in Australia increased nearly 7-fold from 2006 to 2016.⁸¹ Based on reports to the Drug Abuse Warning System, US emergency department visits for M/A of quetiapine increased from 19,195 in 2005 to 32,024 in 2011 (an average of 27,114 visits/year), with 75% of cases involving quetiapine taken in combination with other prescription drugs, alcohol, or illicit drugs.⁸² Consistent with poison center data, M/A was reported for other antipsychotics, but none nearly as frequently as for quetiapine.

With increasingly frequent quetiapine M/A, clinicians should be vigilant in monitoring for medical morbidity related to quetiapine and cumulative toxicity with other drugs. The most frequent adverse events associated with quetiapine M/A reported to US Poison Control Centers are presented in Table 4.^78,79

Continue to: Unlike bupropion...

Unlike bupropion, quetiapine’s dopamine antagonism makes it unlikely to be a truly addictive drug, although this mechanism of action could mediate an increase in concurrent psychostimulant use.⁸³ A few case reports have described a quetiapine discontinuation syndrome similar to that of antidepressants,^60,65,84-88 but withdrawal symptoms suggestive of physiologic dependence may be mediated by non-dopaminergic effects through histamine and serotonin receptors.^84,89 Evidence for quetiapine misuse being associated with craving and compulsive use is lacking, and true quetiapine addiction is probably rare.

Similar to bupropion, preliminary findings have suggested promise for quetiapine as a putative therapy for other SUDs.^90-93 However, subsequent RCTs have failed to demonstrate a therapeutic effect for alcohol and cocaine use disorders.^94-96 Given these negative results and the clear M/A potential of quetiapine, off-label use of quetiapine for the treatment of SUDs and psychiatric symptoms among those with SUDs must be considered judiciously, with an eye towards possible diversion and avoiding the substitution of one drug of abuse for another.

Gabapentinoids

In 1997, the first published case report of gabapentin M/A described a woman who self-administered her husband’s gabapentin to reduce cravings for and withdrawal from cocaine.⁹⁷ The authors highlighted the possible therapeutic benefit of gabapentin in this regard rather than raising concerns about diversion and M/A. By 2004, however, reports of recreational gabapentin M/A emerged among inmates incarcerated within Florida correctional facilities who self-administered intranasal gabapentin to achieve a “high” that was “reminiscent of prior effects from intranasal ingestion of cocaine powder.”⁹⁸ In 2007, a single case of gabapentin misuse up to 7,200 mg/d (recommended dosing is ≤3,600 mg/d) was reported, with documentation of both tolerance and withdrawal symptoms.⁹⁹ As of 2017, a total of 36 cases of gabapentin M/A and 19 cases of pregabalin M/A have been published.¹⁰⁰

In the past decade, anecdotal reports have given way to larger-scale epidemiologic data painting a clear picture of the now-widespread M/A of gabapentin and other gabapentinoids. For example, a study of online descriptions of gabapentin and pregabalin M/A from 2008 to 2010 documented:

• oral and IM use (gabapentin)
• IV and rectal (“plugging”) use (pregabalin)
• “parachuting” (emptying the contents of capsules for a larger dose) (pregabalin)
• euphoric, entactogenic, stimulant, calming/anxiolytic, and dissociative subjective effects (gabapentin/pregabalin)
• rapid development of tolerance to euphoric effects leading to self-administration of increasing doses (gabapentin/pregabalin)
• frequent co-administration with other drugs of abuse, including alcohol, benzodiazepines, cannabis, stimulants, opiates, hallucinogens, gamma-hydroxybutyrate, mephedrone, and Salvia divinorum (gabapentin/pregabalin)¹⁰¹

Several systematic reviews of both anecdotal reports and epidemiologic studies published in the past few years provide additional evidence of the above, such as:

• excessive dosing with self-administration
• intranasal and inhaled routes of administration
• diversion and “street value”
• greater M/A potential of pregabalin than gabapentin
• the presence of gabapentinoids in postmortem toxicology analyses, suggesting a role in overdose fatalities when combined with other drugs.^100,102,103

Continue to: The European Medicine Agency's EudraVigilance database...

The European Medicine Agency’s EudraVigilance database included 4,301 reports of gabapentin misuse, abuse, or dependence, and 7,639 such reports for pregabalin, from 2006 to 2015 (rising sharply after 2012), with 86 gabapentin-related and 27 pregabalin-related fatalities.¹⁰⁴ Data from the Drug Diversion Program of the Researched Abuse, Diversion, and Addiction-Related Surveillance System from 2002 to 2015 have likewise revealed that gabapentin diversion increased significantly in 2013.¹⁰⁵

While the prevalence of gabapentinoid M/A is not known, rates appear to be significantly lower than for traditional drugs of abuse such as cannabis, cocaine, 3,4-methylenedioxymethamphetamine (MDMA), and opioids.^106,107 However, gabapentin and pregabalin M/A appears to be increasingly common among individuals with SUDs and in particular among those with opioid use disorders (OUDs). For example, a 2015 report indicated that 15% of an adult cohort in Appalachian Kentucky with nonmedical use of diverted prescription opioids reported gabapentin M/A, an increase of nearly 3,000% since 2008.¹⁰⁸ Based on data from a US insurance enrollment and claims database, researchers found that the rate of gabapentin overuse among those also overusing opioids was 12% compared with only 2% for those using gabapentin alone.¹⁰⁹ It has also been reported that gabapentin is sometimes used as a “cutting agent” for heroin.¹¹⁰

Those who use gabapentinoids together with opioids report that gabapentin and pregabalin potentiate the euphoric effects of methadone¹¹¹ and endorse specific beliefs that pregabalin increases both the desired effects of heroin as well as negative effects such as “blackouts,” loss of control, and risk of overdose.¹¹² Indeed, sustained M/A of gabapentin and opioids together has been found to increase emergency department utilization, drug-related hospitalization, and respiratory depression.¹¹³ Based on a case-control study of opioid users in Canada, co-prescription of gabapentin and opioids was associated with a 50% increase in death from opioid-related causes compared with prescription of opioids alone.¹¹⁴

Case reports documenting tolerance, withdrawal, craving, and loss of control suggest a true addictive potential for gabapentinoids, but Bonnet and Sherbaum¹⁰⁰ concluded that while there is robust evidence of abusers “liking” gabapentin and pregabalin (eg, reward), evidence of “wanting” them (eg, psychological dependence) in the absence of other SUDs has been limited to only a few anecdotal reports with pregabalin. Accordingly, the risk of true addiction to gabapentinoids by those without preexisting SUDs appears to be low. Nonetheless, the M/A potential of both gabapentin and pregabalin is clear and in the context of a nationwide opioid epidemic, the increased morbidity/mortality risk related to combined use of gabapentinoids and opioids is both striking and concerning. Consequently, the state of Kentucky recently recognized the M/A potential of gabapentin by designating it a Schedule V controlled substance (pregabalin is already a Schedule V drug according to the US Drug Enforcement Agency),^103,113 and several other states now mandate the reporting of gabapentin prescriptions to prescription drug monitoring programs.¹¹⁵

Following a similar pattern to antidepressants and antipsychotics, a potential role for gabapentin in the treatment of cocaine use disorders was supported in preliminary studies,^116-118 but not in subsequent RCTs.^119-121 However, there is evidence from RCTs to support the use of gabapentin and pregabalin in the treatment of alcohol use disorders.^122-124 Gabapentin was also found to significantly reduce cannabis use and withdrawal symptoms in patients compared with placebo in an RCT of individuals with cannabis use disorders.¹²⁵ The perceived safety of gabapentinoids by clinicians, their subjective desirability by patients with SUDs, and efficacy data supporting a therapeutic role in SUDs must be balanced with recognition that approximately 80% of gabapentin prescriptions are written for off-label indications for which there is little supporting evidence,¹⁰⁹ such as low back pain.¹²⁶ Clinicians considering prescribing gabapentinoids to manage psychiatric symptoms, such as anxiety and insomnia, should carefully consider the risk of M/A and other potential morbidities, especially in the setting of SUDs and OUD in particular.

Continue to: Problematic, even if not addictive

Problematic, even if not addictive

It is sometimes claimed that “addiction” to psychiatric medications is not limited to stimulants and benzodiazepines.^27,127 Although anticholinergics, antidepressants, antipsychotics, and gabapentinoids can be drugs of abuse, with some users reporting physiologic withdrawal upon discontinuation, there is only limited evidence that the M/A of these psychiatric medications is associated with the characteristic features of a more complete definition of “addiction,” which may include:

• inability to consistently abstain
• impairment in behavioral control
• diminished recognition of significant problems associated with use
• a dysfunctional emotional response to chronic use.¹²⁸

Nonetheless, the literature documenting anticholinergic, antidepressant, antipsychotic, and gabapentinoid M/A includes several common features, including:

• initial reports among those with limited access to illicit drugs (eg, young people and incarcerated individuals) and subsequent spread to a wider population with more unconventional routes of administration
• use for recreational purposes and other subjective pseudo-therapeutic effects, often in combination with alcohol and illicit drugs
• greater M/A potential of certain medications within each of these drug classes (eg, trihexyphenidyl, bupropion, quetiapine)
• malingering psychiatric symptoms in order to obtain medications from prescribers and diversion for black market sale
• observations that medications might constitute therapy for SUDs that were not supported in subsequent RCTs (with the exception of gabapentin for alcohol and cannabis use disorders)
• increasing evidence of toxicity related to M/A, which suggests that prescription by clinicians has limited benefit and high risk for patients with SUDs.

Bottom Line

Some psychiatric medications are taken as drugs of abuse. Clinicians should be particularly aware of the misuse/abuse potential of anticholinergics, antidepressants, antipsychotics, and gabapentinoids, and use them cautiously, if at all, when treating patients with existing substance use disorders.

Related Resources

• Substance Abuse and Mental Health Services Administration. Prescription drug misuse and abuse. https://www.samhsa.gov/topics/prescription-drug-misuse-abuse.
• Substance Abuse and Mental Health Services Administration. Types of commonly misused or abused drugs. https://www.samhsa.gov/prescription-drug-misuse-abuse/types.
• National Institute on Drug Abuse. Misuse of prescription drugs. https://www.drugabuse.gov/publications/research-reports/misuse-prescription-drugs/summary.
• National Institute on Drug Abuse. New clinician screening tool available for substance use. https://www.drugabuse.gov/news-events/news-releases/2018/06/newclinician-screening-tool-available-substance-use.

Drug Brand Names

Amitriptyline • Elavil, Endep
Benztropine • Cogentin
Biperiden • Akineton
Bupropion • Wellbutrin, Zyban
Chlorpromazine • Thorzine
Fluoxetine • Prozac
Haloperidol • Haldol
Olanzapine • Zyprexa
Orphenadrine • Disipal, Norflex
Pregabalin • Lyrica, Lyrica CR
Procyclidine • Kemadrin
Quetiapine • Seroquel
Tianeptine • Coaxil, Stablon
Tranylcypromine • Parnate
Trifluoperazine • Stelazine
Trihexyphenidyl • Artane, Tremin
Venlafaxine • Effexor

While some classes of medications used to treat psychi­atric disorders, such as stimulants and benzodiazepines, are well-recognized as controlled substances and drugs of abuse, clinicians may be less familiar with the potential misuse/abuse of other psychiatric medications. This article reviews the evidence related to the misuse/abuse of anticholinergics, antidepressants, antipsychotics, and gabapentinoids.

The terms “misuse,” “abuse,” and “addiction” are used variably in the literature without standardized definitions. For this review, “misuse/abuse (M/A)” will be used to collectively describe self-administration that is recreational or otherwise inconsistent with legal or medical guidelines, unless a specific distinction is made. Whether or not the medications reviewed are truly “addictive” will be briefly discussed for each drug class, but the focus will be on clinically relevant aspects of M/A, including:

• excessive self-administration
• self-administration by non-oral routes
• co-administration with other drugs of abuse
• malingering of psychiatric symptoms to obtain prescriptions
• diversion for sale to third parties
• toxicity from overdose.

Anticholinergic medications

The first case describing the deliberate M/A of an anticholinergic medication for its euphoric effects was published in 1960.Further reportsfollowed in Europe before the M/A potential of prescription anticholinergic medications among psychiatric patients with an overdose syndrome characterized by atropinism and toxic psychosis was more widely recognized in the United States in the 1970s. Most reported cases of M/A to date have occurred among patients with psychiatric illness because anticholinergic medications, including trihexyphenidyl, benztropine, biperiden, procyclidine, and orphenadrine, were commonly prescribed for the management of first-generation and high dopamine D2-affinity antipsychotic-induced extrapyramidal symptoms (EPS). For example, one study of 234 consecutively hospitalized patients with schizophrenia noted an anticholinergic M/A incidence of 6.5%.¹

However, anticholinergic M/A is not limited to individuals with psychotic disorders. A UK study of 154 admissions to an inpatient unit specializing in behavioral disturbances found a 12-month trihexyphenidyl M/A incidence of 17%; the most common diagnosis among abusers was antisocial personality disorder.² Anticholinergic M/A has also been reported among patients with a primary diagnosis of substance use disorders (SUDs)³ as well as more indiscriminately in prison settings,⁴ with some inmates exchanging trihexyphenidyl as currency and using it recreationally by crushing it into powder and smoking it with tobacco.⁵ Others have noted that abusers sometimes take anticholinergics with alcohol in order to “potentiate” the effects of each substance.^6,7 Pullen et al⁸ described individuals with and without psychiatric illness who stole anticholinergic medications, purchased them from other patients, or bought them “on the street.” Malingering EPS in order to obtain anticholinergic medications has also been well documented.⁹ Clearly, anticholinergic M/A can occur in psychiatric and non-psychiatric populations, both within and outside of clinical settings. Although anticholinergic M/A appears to be less frequent in the United States now that second-generation antipsychotics (SGAs) are more frequently prescribed, M/A remains common in some settings outside of the United States.⁷

Among the various anticholinergic medications prescribed for EPS, trihexyphenidyl has been reported to have the greatest M/A potential, which has been attributed to its potency,¹⁰ its stimulating effects (whereas benztropine is more sedating),¹¹ and its former popularity among prescribers.⁸ Marken et al¹¹ published a review of 110 reports of M/A occurring in patients receiving anticholinergic medications as part of psychiatric treatment in which 69% of cases involved taking trihexyphenidyl 15 to 60 mg at a time (recommended dosing is 6 to 10 mg/d in divided doses).Most of these patients were prescribed anticholinergic medications for diagnostically appropriate reasons—only 7% were described as “true abusers” with no medical indication. Anticholinergic M/A was typically driven by a desire for euphoric and psychedelic/hallucinogenic effects, although in some cases, anticholinergic M/A was attributed to self-medication of EPS and depressive symptoms. These findings illustrate the blurred distinction between recreational use and perceived subjective benefit, and match those of a subsequent study of 50 psychiatric patients who reported anticholinergic M/A not only to “get high,” but to “decrease depression,” “increase energy,” and decrease antipsychotic adverse effects.¹² Once again, trihexyphenidyl was the most frequently misused anticholinergic in this sample.

Table 1^{2,3,7,8,10-15} outlines the subjective effects sought and experienced by anticholinergic abusers as well as potential toxic effects; there is the potential for overlap. Several authors have also described physiologic dependence with long-term trihexyphenidyl use, including tolerance and a withdrawal/abstinence syndrome.^7,16 In addition, there have been several reports of coma¹³ and death in the setting of intended suicide by overdose of anticholinergic medications.^14,15

Although anticholinergic M/A in the United States now appears to be less common, clinicians should remain aware of the M/A potential of anticholinergic medications prescribed for EPS. Management of M/A involves:

• detection
• reducing anticholinergic exposure by managing EPS with alternative strategies, such as switching or reducing the dose of the antipsychotic medication
• gradual tapering of anticholinergic medications to minimize withdrawal.¹¹

Continue to: Antidepressants

 

 

Antidepressants

Haddad¹⁷ published a review of 21 English-language case reports from 1966 to 1998 describing antidepressant use in which individuals met DSM-IV criteria for substance dependence to the medication. An additional 14 cases of antidepressant M/A were excluded based on insufficient details to support a diagnosis of dependence. The 21 reported cases involved:

• tranylcypromine (a monoamine oxidase inhibitor [MAOI])
• amitriptyline (a tricyclic antidepressant [TCA])
• fluoxetine (a selective serotonin reuptake inhibitor [SSRI])
• amineptine (a TCA previously available in France but removed from the market in 1999 in part due to its abuse potential)
• nomifensine (a norepinephrine/dopamine reuptake inhibitor previously available in the United Kingdom but removed in 1986 due to hemolytic anemia).

In 95% of cases, the antidepressants were prescribed for treatment of an affective disorder but were abused for stimulant effects or the perceived ability to lift mood, cause euphoria or a “high,” or to improve functioning. Two-thirds of cases involved patients with preexisting substance misuse. Placing the case reports in the context of the millions of patients prescribed antidepressants during this period, Haddad concluded the “incidence of [antidepressant] addiction [is] so low as to be clinically irrelevant.”¹⁷

Despite this conclusion, Haddad singled out amineptine and tranylcypromine as antidepressants with some evidence of true addictive potential.^17,18 A more recent case series described 14 patients who met DSM-IV criteria for substance abuse of tertiary amine TCAs (which have strong anticholinergic activity) and concluded that “misuse of [TCAs] is more common than generally appreciated.”¹⁹ In keeping with that claim, a study of 54 outpatients taking unspecified antidepressants found that up to 15% met DSM-III-R criteria for substance dependence (for the antidepressant) in the past year, although that rate was much lower than the rate of benzodiazepine dependence (47%) in a comparative sample.²⁰ Finally, a comprehensive review by Evans and Sullivan²¹ found anecdotal reports published before 2014 that detailed misuse, abuse, and dependence with MAOIs, TCAs, fluoxetine, venlafaxine, bupropion, tianeptine, and amineptine. Taken together, existing evidence indicates that select individuals—typically those with other SUD comorbidity—sometimes misuse antidepressants in a way that suggests addiction.

Still, while it is well known that abrupt cessation of antidepressants can result in a discontinuation syndrome characterized by flu-like symptoms, nausea, and dizziness,²² physiologic withdrawal effects must be distinguished from historical definitions of substance “abuse” and the broader concept of psychological “addiction” or drug dependence^18,23 now incorporated into the DSM-5 definition of SUDs.²⁴ Indeed, although withdrawal symptoms were reported by more than half of those who took antidepressants and responded to a recent online survey,²⁵ evidence to support the existence of significant antidepressant tolerance, craving, or compulsive use is lacking.^17,18 Antidepressants as a class do not appear to be significantly rewarding or reinforcing and, on the contrary, discontinuation by patients is common in clinical practice.²⁶ The popular claim that some individuals taking antidepressants “can’t quit”²⁷ must also be disentangled from loss of therapeutic effects upon cessation.

Bupropion. A more convincing argument for antidepressant addiction can be made for bupropion, a weak norepinephrine and dopamine reuptake inhibitor with an otherwise unclear mechanism of action.²⁸ In 2002, the first report of recreational bupropion M/A described a 13-year-old girl who took 2,400 mg orally (recommended maximum dose is 450 mg/d in divided doses) after being told it would give her “a better high than amphetamine.”²⁹ This was followed in the same year by the first report of recreational M/A of bupropion via nasal insufflation (snorting), resulting in a seizure,³⁰ and in 2013 by the first published case of M/A by IV self-administration.³¹

Continue to: The M/A potential of bupropion...

The M/A potential of bupropion, most commonly via intranasal administration, is now broadly recognized based on several case reports describing desired effects that include a euphoric high and a stimulating “buzz” similar to that of cocaine or methamphetamine but less intense.^29-36 Among recreational users, bupropion tablets are referred to as “welbys,” “wellies,” “dubs,” or “barnies.”³⁷ Media coverage of a 2013 outbreak of bupropion M/A in Toronto detailed administration by snorting, smoking, and injection, and described bupropion as “poor man’s cocaine.”³⁸ Between 2003 and 2016, 2,232 cases of bupropion misuse/abuse/dependence adverse drug reactions were reported to the European Monitoring Agency.³⁷ A review of intentional bupropion M/A reported to US Poison Control Centers between 2000 to 2013 found 975 such cases, with the yearly number tripling between 2000 and 2012.³⁹ In this sample, nearly half (45%) of the users were age 13 to 19, and 76% of cases involved oral ingestion. In addition to bupropion M/A among younger people, individuals who misuse bupropion often include those with existing SUDs but limited access to illicit stimulants and those trying to evade detection by urine toxicology screening.³³ For example, widespread use and diversion has been well documented within correctional settings, and as a result, many facilities have removed bupropion from their formularies.^{21,28,33,34,40}

Beyond desired effects, the most common adverse events associated with bupropion M/A are listed in Table 2,^{28,30,32-34,36,39} along with their incidence based on cases brought to the attention of US Poison Control Centers.³⁹ With relatively little evidence of a significant bupropion withdrawal syndrome,³⁷ the argument in favor of modeling bupropion as a truly addictive drug is limited to anecdotal reports of cravings and compulsive self-administration³⁵ and pro-dopaminergic activity (reuptake inhibition) that might provide a mechanism for potential rewarding and reinforcing effects.⁴⁰ While early preclinical studies of bupropion failed to provide evidence of amphetamine-like abuse potential,^41,42 non-oral administration in amounts well beyond therapeutic dosing could account for euphoric effects and a greater risk of psychological dependence and addiction.^21,28,40

Bupropion also has an FDA indication as an aid to smoking cessation treatment, and the medication demonstrated early promise in the pharmacologic treatment of psycho­stimulant use disorders, with reported improvements in cravings and other SUD outcomes.^43-45 However, subsequent randomized controlled trials (RCTs) failed to demonstrate a clear therapeutic role for bupropion in the treatment of cocaine^46,47 and methamphetamine use disorders (although some secondary analyses suggest possible therapeutic effects among non-daily stimulant users who are able to maintain good adherence with bupropion).^48-51 Given these overall discouraging results, the additive seizure risk of bupropion use with concomitant psychostimulant use, and the potential for M/A and diversion of bupropion (particularly among those with existing SUDs), the use of bupropion for the off-label treatment of stimulant use disorders is not advised.

Antipsychotics

As dopamine antagonists, antipsychotics are typically considered to have low potential for rewarding or reinforcing effects. Indeed, misuse of antipsychotics was a rarity in the first-generation era, with only a few published reports of haloperidol M/A within a small cluster of naïve young people who developed acute EPS,⁵² and a report of diversion in a prison with the “sadistic” intent of inflicting dystonic reactions on others.⁵³ A more recent report described 2additional cases of M/A involving haloperidol and trifluoperazine.⁵⁴ Some authors have described occasional drug-seeking behavior for low-potency D2 blockers such as chlorpromazine, presumably based on their M/A as anticholinergic medications.⁵⁵

The potential for antipsychotic M/A has gained wider recognition since the advent of the SGAs. Three cases of prescription olanzapine M/A have been published to date. One involved a man who malingered manic symptoms to obtain olanzapine, taking ≥40 mg at a time (beyond his prescribed dose of 20 mg twice daily) to get a “buzz,” and combining it with alcohol and benzodiazepines for additive effects or to “come down” from cocaine.⁵⁶ This patient noted that olanzapine was “a popular drug at parties” and was bought, sold, or traded among users, and occasionally administered intravenously. Two other cases described women who self-administered olanzapine, 40 to 50 mg/d, for euphoric and anxiolytic effects.^57,58 James et al⁵⁹ detailed a sample of 28 adults who reported “non-medical use” of olanzapine for anxiolytic effects, as a sleep aid, or to “escape from worries.”

Continue to: Quetiapine

Quetiapine. In contrast to some reports of olanzapine M/A in which the line between M/A and “self-medication” was blurred, quetiapine has become a more convincing example of clear recreational antipsychotic M/A. Since the first report of oral and intranasal quetiapine M/A in the Los Angeles County Jail published in 2004,⁵⁵ subsequent cases have detailed other novel methods of recreational self-administration^60-68 (Table 3^55,60-68), and additional reports have been published in non-English language journals.^69,70 Collectively, these case reports have detailed that quetiapine is:

• misused for primary subjective effects as well as to mitigate the unpleasant effects of other drugs^60,67
• referred to as “quell,”“Q,” “Susie-Q,” “squirrel,” and “baby heroin”^55,71,72
• often obtained by malingering psychiatric symptoms^55,61,63,65
• diverted/sold with “street value” both within and outside of psychiatric facilities and correctional settings.^{55,60-62,67,68,73}

These anecdotal accounts of quetiapine M/A have since been corroborated on a larger scale based on several retrospective studies. Although early reports of quetiapine M/A occurring in correctional settings have resulted in formulary removal,^71,74 quetiapine M/A is by no means limited to forensic populations and is especially common among those with comorbid SUDs. A survey of 74 patients enrolled in a Canadian methadone program reported that nearly 60% had misused quetiapine at some point.⁷⁵ Among an Australian sample of 868 individuals with active IV drug abuse, 31% reported having misused quetiapine.⁷⁶ Finally, within a small sample of patients with SUDs admitted to a detoxification unit in New York City, 17% reported M/A of SGAs.⁷⁷ In this study, SGAs were often taken in conjunction with other drugs of abuse in order to “recover” from or “enhance” the effects of other drugs or to “experiment.” Quetiapine was by far the most frequently abused SGA, reported in 96% of the sample; the most frequently reported SGA/drug combinations were quetiapine/alcohol/opioids, quetiapine/cocaine, and quetiapine/opioids.

Looking more broadly at poison center data, reports to the US National Poison Data System (NPDS) from 2005 to 2011 included 3,116 cases of quetiapine abuse (37.5%, defined as intentional recreational use in order to obtain a “high”) or misuse (62.5%, defined as improper use or dosing for non-recreational purposes).⁷⁸ A more recent analysis of NPDS reports from 2003 to 2013 found 2,118 cases of quetiapine abuse, representing 61% of all cases of reported SGA abuse.⁷⁹ An analysis of the European Medicines Agency Adverse Drug Database yielded 18,112 reports of quetia­pine misuse, abuse, dependence, and withdrawal for quetiapine (from 2005 to 2016) compared with 4,178 for olanzapine (from 2004 to 2016).⁸⁰ These reports identified 368 fatalities associated with quetiapine.

The rate of quetiapine M/A appears to be increasing sharply. Reports of quetiapine M/A to poison centers in Australia increased nearly 7-fold from 2006 to 2016.⁸¹ Based on reports to the Drug Abuse Warning System, US emergency department visits for M/A of quetiapine increased from 19,195 in 2005 to 32,024 in 2011 (an average of 27,114 visits/year), with 75% of cases involving quetiapine taken in combination with other prescription drugs, alcohol, or illicit drugs.⁸² Consistent with poison center data, M/A was reported for other antipsychotics, but none nearly as frequently as for quetiapine.

With increasingly frequent quetiapine M/A, clinicians should be vigilant in monitoring for medical morbidity related to quetiapine and cumulative toxicity with other drugs. The most frequent adverse events associated with quetiapine M/A reported to US Poison Control Centers are presented in Table 4.^78,79

Continue to: Unlike bupropion...

Unlike bupropion, quetiapine’s dopamine antagonism makes it unlikely to be a truly addictive drug, although this mechanism of action could mediate an increase in concurrent psychostimulant use.⁸³ A few case reports have described a quetiapine discontinuation syndrome similar to that of antidepressants,^60,65,84-88 but withdrawal symptoms suggestive of physiologic dependence may be mediated by non-dopaminergic effects through histamine and serotonin receptors.^84,89 Evidence for quetiapine misuse being associated with craving and compulsive use is lacking, and true quetiapine addiction is probably rare.

Similar to bupropion, preliminary findings have suggested promise for quetiapine as a putative therapy for other SUDs.^90-93 However, subsequent RCTs have failed to demonstrate a therapeutic effect for alcohol and cocaine use disorders.^94-96 Given these negative results and the clear M/A potential of quetiapine, off-label use of quetiapine for the treatment of SUDs and psychiatric symptoms among those with SUDs must be considered judiciously, with an eye towards possible diversion and avoiding the substitution of one drug of abuse for another.

Gabapentinoids

In 1997, the first published case report of gabapentin M/A described a woman who self-administered her husband’s gabapentin to reduce cravings for and withdrawal from cocaine.⁹⁷ The authors highlighted the possible therapeutic benefit of gabapentin in this regard rather than raising concerns about diversion and M/A. By 2004, however, reports of recreational gabapentin M/A emerged among inmates incarcerated within Florida correctional facilities who self-administered intranasal gabapentin to achieve a “high” that was “reminiscent of prior effects from intranasal ingestion of cocaine powder.”⁹⁸ In 2007, a single case of gabapentin misuse up to 7,200 mg/d (recommended dosing is ≤3,600 mg/d) was reported, with documentation of both tolerance and withdrawal symptoms.⁹⁹ As of 2017, a total of 36 cases of gabapentin M/A and 19 cases of pregabalin M/A have been published.¹⁰⁰

In the past decade, anecdotal reports have given way to larger-scale epidemiologic data painting a clear picture of the now-widespread M/A of gabapentin and other gabapentinoids. For example, a study of online descriptions of gabapentin and pregabalin M/A from 2008 to 2010 documented:

• oral and IM use (gabapentin)
• IV and rectal (“plugging”) use (pregabalin)
• “parachuting” (emptying the contents of capsules for a larger dose) (pregabalin)
• euphoric, entactogenic, stimulant, calming/anxiolytic, and dissociative subjective effects (gabapentin/pregabalin)
• rapid development of tolerance to euphoric effects leading to self-administration of increasing doses (gabapentin/pregabalin)
• frequent co-administration with other drugs of abuse, including alcohol, benzodiazepines, cannabis, stimulants, opiates, hallucinogens, gamma-hydroxybutyrate, mephedrone, and Salvia divinorum (gabapentin/pregabalin)¹⁰¹

Several systematic reviews of both anecdotal reports and epidemiologic studies published in the past few years provide additional evidence of the above, such as:

• excessive dosing with self-administration
• intranasal and inhaled routes of administration
• diversion and “street value”
• greater M/A potential of pregabalin than gabapentin
• the presence of gabapentinoids in postmortem toxicology analyses, suggesting a role in overdose fatalities when combined with other drugs.^100,102,103

Continue to: The European Medicine Agency's EudraVigilance database...

The European Medicine Agency’s EudraVigilance database included 4,301 reports of gabapentin misuse, abuse, or dependence, and 7,639 such reports for pregabalin, from 2006 to 2015 (rising sharply after 2012), with 86 gabapentin-related and 27 pregabalin-related fatalities.¹⁰⁴ Data from the Drug Diversion Program of the Researched Abuse, Diversion, and Addiction-Related Surveillance System from 2002 to 2015 have likewise revealed that gabapentin diversion increased significantly in 2013.¹⁰⁵

While the prevalence of gabapentinoid M/A is not known, rates appear to be significantly lower than for traditional drugs of abuse such as cannabis, cocaine, 3,4-methylenedioxymethamphetamine (MDMA), and opioids.^106,107 However, gabapentin and pregabalin M/A appears to be increasingly common among individuals with SUDs and in particular among those with opioid use disorders (OUDs). For example, a 2015 report indicated that 15% of an adult cohort in Appalachian Kentucky with nonmedical use of diverted prescription opioids reported gabapentin M/A, an increase of nearly 3,000% since 2008.¹⁰⁸ Based on data from a US insurance enrollment and claims database, researchers found that the rate of gabapentin overuse among those also overusing opioids was 12% compared with only 2% for those using gabapentin alone.¹⁰⁹ It has also been reported that gabapentin is sometimes used as a “cutting agent” for heroin.¹¹⁰

Those who use gabapentinoids together with opioids report that gabapentin and pregabalin potentiate the euphoric effects of methadone¹¹¹ and endorse specific beliefs that pregabalin increases both the desired effects of heroin as well as negative effects such as “blackouts,” loss of control, and risk of overdose.¹¹² Indeed, sustained M/A of gabapentin and opioids together has been found to increase emergency department utilization, drug-related hospitalization, and respiratory depression.¹¹³ Based on a case-control study of opioid users in Canada, co-prescription of gabapentin and opioids was associated with a 50% increase in death from opioid-related causes compared with prescription of opioids alone.¹¹⁴

Case reports documenting tolerance, withdrawal, craving, and loss of control suggest a true addictive potential for gabapentinoids, but Bonnet and Sherbaum¹⁰⁰ concluded that while there is robust evidence of abusers “liking” gabapentin and pregabalin (eg, reward), evidence of “wanting” them (eg, psychological dependence) in the absence of other SUDs has been limited to only a few anecdotal reports with pregabalin. Accordingly, the risk of true addiction to gabapentinoids by those without preexisting SUDs appears to be low. Nonetheless, the M/A potential of both gabapentin and pregabalin is clear and in the context of a nationwide opioid epidemic, the increased morbidity/mortality risk related to combined use of gabapentinoids and opioids is both striking and concerning. Consequently, the state of Kentucky recently recognized the M/A potential of gabapentin by designating it a Schedule V controlled substance (pregabalin is already a Schedule V drug according to the US Drug Enforcement Agency),^103,113 and several other states now mandate the reporting of gabapentin prescriptions to prescription drug monitoring programs.¹¹⁵

Following a similar pattern to antidepressants and antipsychotics, a potential role for gabapentin in the treatment of cocaine use disorders was supported in preliminary studies,^116-118 but not in subsequent RCTs.^119-121 However, there is evidence from RCTs to support the use of gabapentin and pregabalin in the treatment of alcohol use disorders.^122-124 Gabapentin was also found to significantly reduce cannabis use and withdrawal symptoms in patients compared with placebo in an RCT of individuals with cannabis use disorders.¹²⁵ The perceived safety of gabapentinoids by clinicians, their subjective desirability by patients with SUDs, and efficacy data supporting a therapeutic role in SUDs must be balanced with recognition that approximately 80% of gabapentin prescriptions are written for off-label indications for which there is little supporting evidence,¹⁰⁹ such as low back pain.¹²⁶ Clinicians considering prescribing gabapentinoids to manage psychiatric symptoms, such as anxiety and insomnia, should carefully consider the risk of M/A and other potential morbidities, especially in the setting of SUDs and OUD in particular.

Continue to: Problematic, even if not addictive

Problematic, even if not addictive

It is sometimes claimed that “addiction” to psychiatric medications is not limited to stimulants and benzodiazepines.^27,127 Although anticholinergics, antidepressants, antipsychotics, and gabapentinoids can be drugs of abuse, with some users reporting physiologic withdrawal upon discontinuation, there is only limited evidence that the M/A of these psychiatric medications is associated with the characteristic features of a more complete definition of “addiction,” which may include:

• inability to consistently abstain
• impairment in behavioral control
• diminished recognition of significant problems associated with use
• a dysfunctional emotional response to chronic use.¹²⁸

Nonetheless, the literature documenting anticholinergic, antidepressant, antipsychotic, and gabapentinoid M/A includes several common features, including:

• initial reports among those with limited access to illicit drugs (eg, young people and incarcerated individuals) and subsequent spread to a wider population with more unconventional routes of administration
• use for recreational purposes and other subjective pseudo-therapeutic effects, often in combination with alcohol and illicit drugs
• greater M/A potential of certain medications within each of these drug classes (eg, trihexyphenidyl, bupropion, quetiapine)
• malingering psychiatric symptoms in order to obtain medications from prescribers and diversion for black market sale
• observations that medications might constitute therapy for SUDs that were not supported in subsequent RCTs (with the exception of gabapentin for alcohol and cannabis use disorders)
• increasing evidence of toxicity related to M/A, which suggests that prescription by clinicians has limited benefit and high risk for patients with SUDs.

Bottom Line

Some psychiatric medications are taken as drugs of abuse. Clinicians should be particularly aware of the misuse/abuse potential of anticholinergics, antidepressants, antipsychotics, and gabapentinoids, and use them cautiously, if at all, when treating patients with existing substance use disorders.

Related Resources

• Substance Abuse and Mental Health Services Administration. Prescription drug misuse and abuse. https://www.samhsa.gov/topics/prescription-drug-misuse-abuse.
• Substance Abuse and Mental Health Services Administration. Types of commonly misused or abused drugs. https://www.samhsa.gov/prescription-drug-misuse-abuse/types.
• National Institute on Drug Abuse. Misuse of prescription drugs. https://www.drugabuse.gov/publications/research-reports/misuse-prescription-drugs/summary.
• National Institute on Drug Abuse. New clinician screening tool available for substance use. https://www.drugabuse.gov/news-events/news-releases/2018/06/newclinician-screening-tool-available-substance-use.

Drug Brand Names

Amitriptyline • Elavil, Endep
Benztropine • Cogentin
Biperiden • Akineton
Bupropion • Wellbutrin, Zyban
Chlorpromazine • Thorzine
Fluoxetine • Prozac
Haloperidol • Haldol
Olanzapine • Zyprexa
Orphenadrine • Disipal, Norflex
Pregabalin • Lyrica, Lyrica CR
Procyclidine • Kemadrin
Quetiapine • Seroquel
Tianeptine • Coaxil, Stablon
Tranylcypromine • Parnate
Trifluoperazine • Stelazine
Trihexyphenidyl • Artane, Tremin
Venlafaxine • Effexor

CHICAGO – A physician’s baseline opioid prescribing rate strongly predicts future chronic opioid use in rheumatoid arthritis patients, an analysis of data from the Consortium of Rheumatology Researchers of North America (Corrona) Rheumatoid Arthritis Registry suggests.

sdominick/iStock/Getty Images

The baseline 12-month opioid prescribing rate of 148 physicians in the initial cohort varied widely from 0% to 70% (median, 27%), and among 9,337 patients in the registry beyond the baseline 12 months, physician opioid prescribing rates during the baseline period were significantly associated with risk for chronic opioid use, Yvonne C. Lee, MD, reported at the annual meeting of the American College of Rheumatology. She and her colleagues defined chronic opioid use as any opioid use during at least two consecutive study visits.

“It is important to understand the relative contributions of patient vs. physician characteristics on chronic opioid use,” said Dr. Lee of Northwestern University, Chicago. She that the goals of the current study were to identify the extent to which rheumatologists in the United States varied in baseline opioid prescribing rates and to determine the implications of baseline prescribing rates with respect to future chronic opioid use.

Compared with the lowest quartile of baseline opioid prescribing (rate of 18% or less), the second, third, and fourth quartiles of prescribing were associated with increasing odds of chronic opioid use (odds ratios of 1.16, 1.89, and 2.01 for the quartiles, respectively) during the study period, she said.

The researchers saw similar relationships when they used a stricter definition of opioid use and when they extended the cutoff between the baseline and study periods to 18 months. The relationships persisted after adjusting for numerous patient characteristics, such as age, sex, race, insurance status, RA duration, and treatments used, she said.

Subgroup analyses were also conducted to examine heterogeneity across clinical characteristics, including Clinical Disease Activity Index score (10 or less vs. greater than 10), pain intensity (scores of 40 or less, greater than 40 to 60, and greater than 60 out of 100), and use vs. nonuse of antidepressant medication. The relationships between physician baseline prescribing and chronic opioid use were similar across subgroups, she noted.

The findings help to characterize the role of rheumatologists’ prescribing rates in the ongoing opioid crisis even though the conclusions that can be reached are limited by the fact that some patients may receive opioid prescriptions from physicians outside the registry, by a lack of data on specific opioid types and doses, and by a lack of detailed information about physician characteristics, Dr. Lee said.

Physicians were included in the analysis only if they had contributed at least 10 RA patients to the registry within their first year of participation, and patients were included if they were patients of those physicians, if they had at least 12 months of follow-up data available, and if they were not prevalent opioid users at study entry.

A long-term goal is to target interventions to appropriate subgroups, she said, noting that 21%-29% of patients who are prescribed opioids for chronic pain misuse them, and more than 33,000 Americans die of opioid overdoses each year.

“Implications [of the findings] are that, in addition to targeting patients, we may also really want to consider interventions that target high-intensity prescribers. This may be useful for helping to decrease chronic opioid use in patients,” she concluded.

Dr. Lee has an investigator-initiated grant from Pfizer and owns stock in Express Scripts.

sworcester@mdedge.com

SOURCE: Lee Y et al. Arthritis Rheumatol. 2018;70(Suppl 10), Abstract 1917

CHICAGO – A physician’s baseline opioid prescribing rate strongly predicts future chronic opioid use in rheumatoid arthritis patients, an analysis of data from the Consortium of Rheumatology Researchers of North America (Corrona) Rheumatoid Arthritis Registry suggests.

sdominick/iStock/Getty Images

The baseline 12-month opioid prescribing rate of 148 physicians in the initial cohort varied widely from 0% to 70% (median, 27%), and among 9,337 patients in the registry beyond the baseline 12 months, physician opioid prescribing rates during the baseline period were significantly associated with risk for chronic opioid use, Yvonne C. Lee, MD, reported at the annual meeting of the American College of Rheumatology. She and her colleagues defined chronic opioid use as any opioid use during at least two consecutive study visits.

“It is important to understand the relative contributions of patient vs. physician characteristics on chronic opioid use,” said Dr. Lee of Northwestern University, Chicago. She that the goals of the current study were to identify the extent to which rheumatologists in the United States varied in baseline opioid prescribing rates and to determine the implications of baseline prescribing rates with respect to future chronic opioid use.

Compared with the lowest quartile of baseline opioid prescribing (rate of 18% or less), the second, third, and fourth quartiles of prescribing were associated with increasing odds of chronic opioid use (odds ratios of 1.16, 1.89, and 2.01 for the quartiles, respectively) during the study period, she said.

The researchers saw similar relationships when they used a stricter definition of opioid use and when they extended the cutoff between the baseline and study periods to 18 months. The relationships persisted after adjusting for numerous patient characteristics, such as age, sex, race, insurance status, RA duration, and treatments used, she said.

Subgroup analyses were also conducted to examine heterogeneity across clinical characteristics, including Clinical Disease Activity Index score (10 or less vs. greater than 10), pain intensity (scores of 40 or less, greater than 40 to 60, and greater than 60 out of 100), and use vs. nonuse of antidepressant medication. The relationships between physician baseline prescribing and chronic opioid use were similar across subgroups, she noted.

The findings help to characterize the role of rheumatologists’ prescribing rates in the ongoing opioid crisis even though the conclusions that can be reached are limited by the fact that some patients may receive opioid prescriptions from physicians outside the registry, by a lack of data on specific opioid types and doses, and by a lack of detailed information about physician characteristics, Dr. Lee said.

Physicians were included in the analysis only if they had contributed at least 10 RA patients to the registry within their first year of participation, and patients were included if they were patients of those physicians, if they had at least 12 months of follow-up data available, and if they were not prevalent opioid users at study entry.

A long-term goal is to target interventions to appropriate subgroups, she said, noting that 21%-29% of patients who are prescribed opioids for chronic pain misuse them, and more than 33,000 Americans die of opioid overdoses each year.

“Implications [of the findings] are that, in addition to targeting patients, we may also really want to consider interventions that target high-intensity prescribers. This may be useful for helping to decrease chronic opioid use in patients,” she concluded.

Dr. Lee has an investigator-initiated grant from Pfizer and owns stock in Express Scripts.

sworcester@mdedge.com

SOURCE: Lee Y et al. Arthritis Rheumatol. 2018;70(Suppl 10), Abstract 1917

CHICAGO – A physician’s baseline opioid prescribing rate strongly predicts future chronic opioid use in rheumatoid arthritis patients, an analysis of data from the Consortium of Rheumatology Researchers of North America (Corrona) Rheumatoid Arthritis Registry suggests.

sdominick/iStock/Getty Images

The baseline 12-month opioid prescribing rate of 148 physicians in the initial cohort varied widely from 0% to 70% (median, 27%), and among 9,337 patients in the registry beyond the baseline 12 months, physician opioid prescribing rates during the baseline period were significantly associated with risk for chronic opioid use, Yvonne C. Lee, MD, reported at the annual meeting of the American College of Rheumatology. She and her colleagues defined chronic opioid use as any opioid use during at least two consecutive study visits.

“It is important to understand the relative contributions of patient vs. physician characteristics on chronic opioid use,” said Dr. Lee of Northwestern University, Chicago. She that the goals of the current study were to identify the extent to which rheumatologists in the United States varied in baseline opioid prescribing rates and to determine the implications of baseline prescribing rates with respect to future chronic opioid use.

Compared with the lowest quartile of baseline opioid prescribing (rate of 18% or less), the second, third, and fourth quartiles of prescribing were associated with increasing odds of chronic opioid use (odds ratios of 1.16, 1.89, and 2.01 for the quartiles, respectively) during the study period, she said.

The researchers saw similar relationships when they used a stricter definition of opioid use and when they extended the cutoff between the baseline and study periods to 18 months. The relationships persisted after adjusting for numerous patient characteristics, such as age, sex, race, insurance status, RA duration, and treatments used, she said.

Subgroup analyses were also conducted to examine heterogeneity across clinical characteristics, including Clinical Disease Activity Index score (10 or less vs. greater than 10), pain intensity (scores of 40 or less, greater than 40 to 60, and greater than 60 out of 100), and use vs. nonuse of antidepressant medication. The relationships between physician baseline prescribing and chronic opioid use were similar across subgroups, she noted.

The findings help to characterize the role of rheumatologists’ prescribing rates in the ongoing opioid crisis even though the conclusions that can be reached are limited by the fact that some patients may receive opioid prescriptions from physicians outside the registry, by a lack of data on specific opioid types and doses, and by a lack of detailed information about physician characteristics, Dr. Lee said.

Physicians were included in the analysis only if they had contributed at least 10 RA patients to the registry within their first year of participation, and patients were included if they were patients of those physicians, if they had at least 12 months of follow-up data available, and if they were not prevalent opioid users at study entry.

A long-term goal is to target interventions to appropriate subgroups, she said, noting that 21%-29% of patients who are prescribed opioids for chronic pain misuse them, and more than 33,000 Americans die of opioid overdoses each year.

“Implications [of the findings] are that, in addition to targeting patients, we may also really want to consider interventions that target high-intensity prescribers. This may be useful for helping to decrease chronic opioid use in patients,” she concluded.

Dr. Lee has an investigator-initiated grant from Pfizer and owns stock in Express Scripts.

sworcester@mdedge.com

SOURCE: Lee Y et al. Arthritis Rheumatol. 2018;70(Suppl 10), Abstract 1917

SAN DIEGO – Apixaban is as safe as, and more effective than, dalteparin for patients with cancer-associated venous thromboembolism (VTE), according to the Phase 3 ADAM VTE trial.

The rates of major bleeding and clinically relevant nonmajor bleeding in patients who received apixaban were similar to those in patients who received dalteparin. However, the rate of VTE recurrence was significantly lower with apixaban than it was with dalteparin.

“[A]pixaban was associated with very low bleeding rates and venous thrombosis recurrence rates compared to dalteparin,” said Robert D. McBane II, MD, of the Mayo Clinic in Rochester, Minn., at the annual meeting of the American Society of Hematology.

The trial included 300 adults (aged 18 years and older) with active cancer and acute VTE who were randomized to receive apixaban (n = 150) or dalteparin (n = 150). The dose and schedule for oral apixaban was 10 mg twice daily for 7 days followed by 5 mg twice daily for 6 months. Dalteparin was given subcutaneously at 200 IU/kg per day for 1 month followed by 150 IU/kg daily for 6 months. Among the patients in the study, 145 patients in the apixaban arm and 142 in the dalteparin arm ultimately received their assigned treatment.

Every month, patients completed an anticoagulation satisfaction survey and bruise survey (a modification of the Duke Anticoagulation Satisfaction Scale). They also underwent lab testing (complete blood count, liver and renal function testing) and were assessed for outcomes, medication reconciliation, drug compliance, and ECOG status on a monthly basis.
 

Patient characteristics

Baseline characteristics were similar between the treatment arms. The mean age was 64 years in both arms, and roughly half of patients in both arms were female. Hematologic malignancies were present in 9% of patients in the apixaban arm and 11% in the dalteparin arm. Others included lung, colorectal,

pancreatic/hepatobiliary, gynecologic, breast, genitourinary, upper gastrointestinal, and brain cancers.

Of patients in the study, 65% of those in the apixaban arm and 66% in the dalteparin arm had distant metastasis, and 74% of patients in both arms were receiving chemotherapy while on study.

Patients had the following qualifying thrombotic events:

• Any pulmonary embolism (PE) – 55% of patients in the apixaban arm and 51% in the dalteparin arm
• Any deep vein thrombosis (DVT) – 48% and 47%, respectively
• PE only – 44% and 39%, respectively
• PE with DVT – 12% in both arms
• DVT only – 37% and 35%, respectively
• Lower extremity DVT – 31% and 34%, respectively
• Upper extremity DVT – 17% and 14%, respectively
• Cerebral venous thrombosis (VT) – 1% and 0%, respectively
• Splanchnic VT – 8% and 18%, respectively.

Bleeding, thrombosis, and death

The study’s primary endpoint was major bleeding, which did not occur in any of the apixaban-treated patients. However, major bleeding did occur in two (1.4%) patients in the dalteparin arm (P = .14).

A secondary endpoint was major bleeding plus clinically relevant nonmajor bleeding. This occurred in nine (6.2%) patients in the apixaban arm and nine (6.3%) in the dalteparin arm (P = .88).

The researchers also assessed VTE recurrence. One patient in the apixaban arm (0.7%) and nine in the dalteparin arm (6.3%) had VTE recurrence (P = .03).

The patient in the apixaban arm experienced cerebral VT, and the patients with recurrence in the dalteparin arm had leg (n = 4) or arm (n = 2) VTE, PE (n = 1), or splanchnic VT (n = 2).

One patient in each arm (0.7%) had arterial thrombosis.

There was no significant difference in cumulative mortality between the treatment arms (hazard ratio, 1.40; P = .3078).
 

Satisfaction and discontinuation

Overall, apixaban fared better than dalteparin in the monthly patient satisfaction surveys. At various time points, apixaban-treated patients were significantly less likely to be concerned about excessive bruising, find anticoagulant treatment a burden or difficult to carry out, or say anticoagulant treatment added stress to their lives, negatively impacted their quality of life, or caused them “a great deal” of worry, irritation, or frustration.

However, apixaban-treated patients were less likely than dalteparin recipients to have confidence that their drug protected them from VTE recurrence, while the apixaban recipients were more likely than the dalteparin group to report overall satisfaction with their treatment.

In addition, premature treatment discontinuation was more common in the dalteparin group than in the apixaban group – 15% and 4%, respectively (P = .0012).

“Apixaban was well tolerated with superior patient safety satisfaction, as well as significantly fewer study drug discontinuations compared to dalteparin,” Dr. McBane said. “I believe that these data support the use of apixaban for the acute treatment of cancer-associated venous thromboembolism.”

This study was funded by BMS/Pfizer Alliance. Dr. McBane declared no other conflicts of interest.

jensmith@mdedge.com

SOURCE: McBane RD et al. ASH 2018, Abstract 421.

SAN DIEGO – Apixaban is as safe as, and more effective than, dalteparin for patients with cancer-associated venous thromboembolism (VTE), according to the Phase 3 ADAM VTE trial.

The rates of major bleeding and clinically relevant nonmajor bleeding in patients who received apixaban were similar to those in patients who received dalteparin. However, the rate of VTE recurrence was significantly lower with apixaban than it was with dalteparin.

“[A]pixaban was associated with very low bleeding rates and venous thrombosis recurrence rates compared to dalteparin,” said Robert D. McBane II, MD, of the Mayo Clinic in Rochester, Minn., at the annual meeting of the American Society of Hematology.

The trial included 300 adults (aged 18 years and older) with active cancer and acute VTE who were randomized to receive apixaban (n = 150) or dalteparin (n = 150). The dose and schedule for oral apixaban was 10 mg twice daily for 7 days followed by 5 mg twice daily for 6 months. Dalteparin was given subcutaneously at 200 IU/kg per day for 1 month followed by 150 IU/kg daily for 6 months. Among the patients in the study, 145 patients in the apixaban arm and 142 in the dalteparin arm ultimately received their assigned treatment.

Every month, patients completed an anticoagulation satisfaction survey and bruise survey (a modification of the Duke Anticoagulation Satisfaction Scale). They also underwent lab testing (complete blood count, liver and renal function testing) and were assessed for outcomes, medication reconciliation, drug compliance, and ECOG status on a monthly basis.
 

Patient characteristics

Baseline characteristics were similar between the treatment arms. The mean age was 64 years in both arms, and roughly half of patients in both arms were female. Hematologic malignancies were present in 9% of patients in the apixaban arm and 11% in the dalteparin arm. Others included lung, colorectal,

pancreatic/hepatobiliary, gynecologic, breast, genitourinary, upper gastrointestinal, and brain cancers.

Of patients in the study, 65% of those in the apixaban arm and 66% in the dalteparin arm had distant metastasis, and 74% of patients in both arms were receiving chemotherapy while on study.

Patients had the following qualifying thrombotic events:

• Any pulmonary embolism (PE) – 55% of patients in the apixaban arm and 51% in the dalteparin arm
• Any deep vein thrombosis (DVT) – 48% and 47%, respectively
• PE only – 44% and 39%, respectively
• PE with DVT – 12% in both arms
• DVT only – 37% and 35%, respectively
• Lower extremity DVT – 31% and 34%, respectively
• Upper extremity DVT – 17% and 14%, respectively
• Cerebral venous thrombosis (VT) – 1% and 0%, respectively
• Splanchnic VT – 8% and 18%, respectively.

Bleeding, thrombosis, and death

The study’s primary endpoint was major bleeding, which did not occur in any of the apixaban-treated patients. However, major bleeding did occur in two (1.4%) patients in the dalteparin arm (P = .14).

A secondary endpoint was major bleeding plus clinically relevant nonmajor bleeding. This occurred in nine (6.2%) patients in the apixaban arm and nine (6.3%) in the dalteparin arm (P = .88).

The researchers also assessed VTE recurrence. One patient in the apixaban arm (0.7%) and nine in the dalteparin arm (6.3%) had VTE recurrence (P = .03).

The patient in the apixaban arm experienced cerebral VT, and the patients with recurrence in the dalteparin arm had leg (n = 4) or arm (n = 2) VTE, PE (n = 1), or splanchnic VT (n = 2).

One patient in each arm (0.7%) had arterial thrombosis.

There was no significant difference in cumulative mortality between the treatment arms (hazard ratio, 1.40; P = .3078).
 

Satisfaction and discontinuation

Overall, apixaban fared better than dalteparin in the monthly patient satisfaction surveys. At various time points, apixaban-treated patients were significantly less likely to be concerned about excessive bruising, find anticoagulant treatment a burden or difficult to carry out, or say anticoagulant treatment added stress to their lives, negatively impacted their quality of life, or caused them “a great deal” of worry, irritation, or frustration.

However, apixaban-treated patients were less likely than dalteparin recipients to have confidence that their drug protected them from VTE recurrence, while the apixaban recipients were more likely than the dalteparin group to report overall satisfaction with their treatment.

In addition, premature treatment discontinuation was more common in the dalteparin group than in the apixaban group – 15% and 4%, respectively (P = .0012).

“Apixaban was well tolerated with superior patient safety satisfaction, as well as significantly fewer study drug discontinuations compared to dalteparin,” Dr. McBane said. “I believe that these data support the use of apixaban for the acute treatment of cancer-associated venous thromboembolism.”

This study was funded by BMS/Pfizer Alliance. Dr. McBane declared no other conflicts of interest.

jensmith@mdedge.com

SOURCE: McBane RD et al. ASH 2018, Abstract 421.

SAN DIEGO – Apixaban is as safe as, and more effective than, dalteparin for patients with cancer-associated venous thromboembolism (VTE), according to the Phase 3 ADAM VTE trial.

The rates of major bleeding and clinically relevant nonmajor bleeding in patients who received apixaban were similar to those in patients who received dalteparin. However, the rate of VTE recurrence was significantly lower with apixaban than it was with dalteparin.

“[A]pixaban was associated with very low bleeding rates and venous thrombosis recurrence rates compared to dalteparin,” said Robert D. McBane II, MD, of the Mayo Clinic in Rochester, Minn., at the annual meeting of the American Society of Hematology.

The trial included 300 adults (aged 18 years and older) with active cancer and acute VTE who were randomized to receive apixaban (n = 150) or dalteparin (n = 150). The dose and schedule for oral apixaban was 10 mg twice daily for 7 days followed by 5 mg twice daily for 6 months. Dalteparin was given subcutaneously at 200 IU/kg per day for 1 month followed by 150 IU/kg daily for 6 months. Among the patients in the study, 145 patients in the apixaban arm and 142 in the dalteparin arm ultimately received their assigned treatment.

Every month, patients completed an anticoagulation satisfaction survey and bruise survey (a modification of the Duke Anticoagulation Satisfaction Scale). They also underwent lab testing (complete blood count, liver and renal function testing) and were assessed for outcomes, medication reconciliation, drug compliance, and ECOG status on a monthly basis.
 

Patient characteristics

Baseline characteristics were similar between the treatment arms. The mean age was 64 years in both arms, and roughly half of patients in both arms were female. Hematologic malignancies were present in 9% of patients in the apixaban arm and 11% in the dalteparin arm. Others included lung, colorectal,

pancreatic/hepatobiliary, gynecologic, breast, genitourinary, upper gastrointestinal, and brain cancers.

Of patients in the study, 65% of those in the apixaban arm and 66% in the dalteparin arm had distant metastasis, and 74% of patients in both arms were receiving chemotherapy while on study.

Patients had the following qualifying thrombotic events:

• Any pulmonary embolism (PE) – 55% of patients in the apixaban arm and 51% in the dalteparin arm
• Any deep vein thrombosis (DVT) – 48% and 47%, respectively
• PE only – 44% and 39%, respectively
• PE with DVT – 12% in both arms
• DVT only – 37% and 35%, respectively
• Lower extremity DVT – 31% and 34%, respectively
• Upper extremity DVT – 17% and 14%, respectively
• Cerebral venous thrombosis (VT) – 1% and 0%, respectively
• Splanchnic VT – 8% and 18%, respectively.

Bleeding, thrombosis, and death

The study’s primary endpoint was major bleeding, which did not occur in any of the apixaban-treated patients. However, major bleeding did occur in two (1.4%) patients in the dalteparin arm (P = .14).

A secondary endpoint was major bleeding plus clinically relevant nonmajor bleeding. This occurred in nine (6.2%) patients in the apixaban arm and nine (6.3%) in the dalteparin arm (P = .88).

The researchers also assessed VTE recurrence. One patient in the apixaban arm (0.7%) and nine in the dalteparin arm (6.3%) had VTE recurrence (P = .03).

The patient in the apixaban arm experienced cerebral VT, and the patients with recurrence in the dalteparin arm had leg (n = 4) or arm (n = 2) VTE, PE (n = 1), or splanchnic VT (n = 2).

One patient in each arm (0.7%) had arterial thrombosis.

There was no significant difference in cumulative mortality between the treatment arms (hazard ratio, 1.40; P = .3078).
 

Satisfaction and discontinuation

Overall, apixaban fared better than dalteparin in the monthly patient satisfaction surveys. At various time points, apixaban-treated patients were significantly less likely to be concerned about excessive bruising, find anticoagulant treatment a burden or difficult to carry out, or say anticoagulant treatment added stress to their lives, negatively impacted their quality of life, or caused them “a great deal” of worry, irritation, or frustration.

However, apixaban-treated patients were less likely than dalteparin recipients to have confidence that their drug protected them from VTE recurrence, while the apixaban recipients were more likely than the dalteparin group to report overall satisfaction with their treatment.

In addition, premature treatment discontinuation was more common in the dalteparin group than in the apixaban group – 15% and 4%, respectively (P = .0012).

“Apixaban was well tolerated with superior patient safety satisfaction, as well as significantly fewer study drug discontinuations compared to dalteparin,” Dr. McBane said. “I believe that these data support the use of apixaban for the acute treatment of cancer-associated venous thromboembolism.”

This study was funded by BMS/Pfizer Alliance. Dr. McBane declared no other conflicts of interest.

jensmith@mdedge.com

SOURCE: McBane RD et al. ASH 2018, Abstract 421.

BONITA SPRINGS, FLA. – Clinician shortages and alarming opioid overdose rates are prompting rural health care centers to turn to telemedicine for delivering treatments for opioid use disorder (OUD). Several studies suggest that these treatments are being delivered effectively, experts said at the annual meeting of the American Academy of Addiction Psychiatry.

In both Maryland and West Virginia, for example, success using a telehealth approach has been reported recently, said David Moore, MD, PhD, assistant professor of psychiatry at Yale University, New Haven, Conn.

Specifically, in rural Maryland, physicians used telemedicine to provide buprenorphine treatment for OUD at a treatment center in August 2015. Researchers at the University of Maryland, Baltimore, looked at the first 177 of the patients treated with the approach. They found that retention in treatment was 91% at 1 month and 57% at 3 months. Of patients still in treatment at 3 months, 86% had urine that was opioid negative, researchers said (Am J Addict. 2018 Dec;27[8]:612-17).

And in West Virginia, researchers reviewed the records of 100 patients receiving buprenorphine treatment to compare outcomes of those treated with telemedicine and those treated face-to-face. They found no significant differences between the groups in additional substance use, time to achieve 30 days and 90 days of abstinence, or retention rates at 3 months and 1 year (J Addict Med. 2017 Mar-Apr;11[2]:138-44).

In addition, Dr. Moore said, he has had success with home inductions in the northern reaches of Maine. His first home induction involved a 55-year-old veteran with a history of oxycodone and hydrocodone use who used illicit buprenorphine when he could. Dr. Moore said the man was referred to him on a Monday. He called in a prescription for the drug the next day and gave the patient a handout on how to do a home induction. “Then he had a phone check-in, and we followed up on Thursday. It actually worked really well,” Dr. Moore said.

The dearth of buprenorphine providers in northern Maine makes those kinds of arrangements attractive, he said. In Maine’s Piscataquis County, he said, there is one buprenorphine provider for every 2,000 square miles. “We have about 1 in every 5 square miles in New Haven,” he said. “Thinking about the distance you have to travel, it gets to be pretty daunting.”

The ability of clinicians to provide buprenorphine with telemedicine varies by state. Among the resources needed to provide telemedicine services are reliable Internet access and an ability for a patient to consent to the treatment.

Nationwide, 56.3% of rural counties have no buprenorphine provider, according to a recent study, said Lewei (Allison) Lin, MD, assistant professor of psychiatry at the University of Michigan, Ann Arbor. A survey of 1,100 rural providers, results of which were included in that study, found that 48% of them said concerns about substance diversion or misuse were a barrier to providing buprenorphine, and 44% cited a lack of mental health or psychosocial support services.

“Although this country still has a major issue with access to treatment, we see that the access problem is about double in rural areas,” she said. “If you add on the distance issue and time, this becomes an even greater challenge.”

Nurse practitioners and physician assistants are now able to obtain a Drug Enforcement Administration waiver that will allow them to prescribe OUD, thanks to the Comprehensive Addiction and Recovery Act of 2016.

BONITA SPRINGS, FLA. – Clinician shortages and alarming opioid overdose rates are prompting rural health care centers to turn to telemedicine for delivering treatments for opioid use disorder (OUD). Several studies suggest that these treatments are being delivered effectively, experts said at the annual meeting of the American Academy of Addiction Psychiatry.

In both Maryland and West Virginia, for example, success using a telehealth approach has been reported recently, said David Moore, MD, PhD, assistant professor of psychiatry at Yale University, New Haven, Conn.

Specifically, in rural Maryland, physicians used telemedicine to provide buprenorphine treatment for OUD at a treatment center in August 2015. Researchers at the University of Maryland, Baltimore, looked at the first 177 of the patients treated with the approach. They found that retention in treatment was 91% at 1 month and 57% at 3 months. Of patients still in treatment at 3 months, 86% had urine that was opioid negative, researchers said (Am J Addict. 2018 Dec;27[8]:612-17).

And in West Virginia, researchers reviewed the records of 100 patients receiving buprenorphine treatment to compare outcomes of those treated with telemedicine and those treated face-to-face. They found no significant differences between the groups in additional substance use, time to achieve 30 days and 90 days of abstinence, or retention rates at 3 months and 1 year (J Addict Med. 2017 Mar-Apr;11[2]:138-44).

In addition, Dr. Moore said, he has had success with home inductions in the northern reaches of Maine. His first home induction involved a 55-year-old veteran with a history of oxycodone and hydrocodone use who used illicit buprenorphine when he could. Dr. Moore said the man was referred to him on a Monday. He called in a prescription for the drug the next day and gave the patient a handout on how to do a home induction. “Then he had a phone check-in, and we followed up on Thursday. It actually worked really well,” Dr. Moore said.

The dearth of buprenorphine providers in northern Maine makes those kinds of arrangements attractive, he said. In Maine’s Piscataquis County, he said, there is one buprenorphine provider for every 2,000 square miles. “We have about 1 in every 5 square miles in New Haven,” he said. “Thinking about the distance you have to travel, it gets to be pretty daunting.”

The ability of clinicians to provide buprenorphine with telemedicine varies by state. Among the resources needed to provide telemedicine services are reliable Internet access and an ability for a patient to consent to the treatment.

Nationwide, 56.3% of rural counties have no buprenorphine provider, according to a recent study, said Lewei (Allison) Lin, MD, assistant professor of psychiatry at the University of Michigan, Ann Arbor. A survey of 1,100 rural providers, results of which were included in that study, found that 48% of them said concerns about substance diversion or misuse were a barrier to providing buprenorphine, and 44% cited a lack of mental health or psychosocial support services.

“Although this country still has a major issue with access to treatment, we see that the access problem is about double in rural areas,” she said. “If you add on the distance issue and time, this becomes an even greater challenge.”

Nurse practitioners and physician assistants are now able to obtain a Drug Enforcement Administration waiver that will allow them to prescribe OUD, thanks to the Comprehensive Addiction and Recovery Act of 2016.

BONITA SPRINGS, FLA. – Clinician shortages and alarming opioid overdose rates are prompting rural health care centers to turn to telemedicine for delivering treatments for opioid use disorder (OUD). Several studies suggest that these treatments are being delivered effectively, experts said at the annual meeting of the American Academy of Addiction Psychiatry.

In both Maryland and West Virginia, for example, success using a telehealth approach has been reported recently, said David Moore, MD, PhD, assistant professor of psychiatry at Yale University, New Haven, Conn.

Specifically, in rural Maryland, physicians used telemedicine to provide buprenorphine treatment for OUD at a treatment center in August 2015. Researchers at the University of Maryland, Baltimore, looked at the first 177 of the patients treated with the approach. They found that retention in treatment was 91% at 1 month and 57% at 3 months. Of patients still in treatment at 3 months, 86% had urine that was opioid negative, researchers said (Am J Addict. 2018 Dec;27[8]:612-17).

And in West Virginia, researchers reviewed the records of 100 patients receiving buprenorphine treatment to compare outcomes of those treated with telemedicine and those treated face-to-face. They found no significant differences between the groups in additional substance use, time to achieve 30 days and 90 days of abstinence, or retention rates at 3 months and 1 year (J Addict Med. 2017 Mar-Apr;11[2]:138-44).

In addition, Dr. Moore said, he has had success with home inductions in the northern reaches of Maine. His first home induction involved a 55-year-old veteran with a history of oxycodone and hydrocodone use who used illicit buprenorphine when he could. Dr. Moore said the man was referred to him on a Monday. He called in a prescription for the drug the next day and gave the patient a handout on how to do a home induction. “Then he had a phone check-in, and we followed up on Thursday. It actually worked really well,” Dr. Moore said.

The dearth of buprenorphine providers in northern Maine makes those kinds of arrangements attractive, he said. In Maine’s Piscataquis County, he said, there is one buprenorphine provider for every 2,000 square miles. “We have about 1 in every 5 square miles in New Haven,” he said. “Thinking about the distance you have to travel, it gets to be pretty daunting.”

The ability of clinicians to provide buprenorphine with telemedicine varies by state. Among the resources needed to provide telemedicine services are reliable Internet access and an ability for a patient to consent to the treatment.

Nationwide, 56.3% of rural counties have no buprenorphine provider, according to a recent study, said Lewei (Allison) Lin, MD, assistant professor of psychiatry at the University of Michigan, Ann Arbor. A survey of 1,100 rural providers, results of which were included in that study, found that 48% of them said concerns about substance diversion or misuse were a barrier to providing buprenorphine, and 44% cited a lack of mental health or psychosocial support services.

“Although this country still has a major issue with access to treatment, we see that the access problem is about double in rural areas,” she said. “If you add on the distance issue and time, this becomes an even greater challenge.”

Nurse practitioners and physician assistants are now able to obtain a Drug Enforcement Administration waiver that will allow them to prescribe OUD, thanks to the Comprehensive Addiction and Recovery Act of 2016.

The benefits of statins highly offset the associated risks in appropriate patients, according to a scientific statement issued by the American Heart Association.

“The review covers the general patient population, as well as demographic subgroups, including the elderly, children, pregnant women, East Asians, and patients with specific conditions.” wrote Connie B. Newman, MD, of New York University, together with her colleagues. The report is in Arteriosclerosis, Thrombosis, and Vascular Biology.

After an extensive review of the literature pertaining to statin safety and tolerability, Dr. Newman and her colleagues reported the compiled findings from several randomized controlled trials, in addition to observational data, where required. They found that the risk of serious muscle complications, such as rhabdomyolysis, attributable to statin use was less than 0.1%. Furthermore, they noted that the risk of serious hepatotoxicity was even less likely, occurring in about 1 in 10,000 patients treated with therapy.

“There is no convincing evidence for a causal relationship between statins and cancer, cataracts, cognitive dysfunction, peripheral neuropathy, erectile dysfunction, or tendinitis,” the experts wrote. “In U.S. clinical practices, roughly 10% of patients stop taking a statin because of subjective complaints, most commonly muscle symptoms without raised creatine kinase,” they further reported.

Contrastingly, data from randomized trials have shown that the change in the incidence of muscle-related symptoms in patients treated with statins versus placebo is less than 1%. Moreover, the incidence is even lower, with an estimated rate of 0.1%, in those who stopped statin therapy because of these symptoms. Given these results, Dr. Newman and her colleagues said that muscle-related symptoms among statin-treated patients are not due to the pharmacological activity of the statin.

“Restarting statin therapy in these patients can be challenging, but it is important, especially in patients at high risk of cardiovascular events, for whom prevention of these events is a priority,” they added.

A large proportion of the population takes statin therapy to lower the risk of major cardiovascular events, including ischemic stroke, myocardial infarction, and other adverse effects of cardiovascular disease. At maximal doses, statins may decrease LDL-cholesterol levels by roughly 55%-60%. In addition, given the multitude of available generics, statins are an economical treatment option for most patients.

However, Dr. Newman and her colleagues suggested that, when considering statin therapy in special populations, particularly in patients with end-stage renal failure or severe hepatic disease, commencing treatment is not recommended.

“The lack of proof of cardiovascular benefit in patients with end-stage renal disease suggests that initiating statin treatment in these patients is generally not warranted,” the experts wrote. “Data on safety in people with more serious liver disease are insufficient, and statin treatment is generally discouraged,” they added.

With respect to statin-induced adverse effects, they are usually reversible upon discontinuation of therapy, with the exception of hemorrhagic stroke. However, damage from an ischemic stroke or myocardial infarction may result in death. As a result, in patients who would benefit from statin therapy, based on most recent guidelines, cardiovascular benefits greatly exceed potential safety concerns.

Dr. Newman and her coauthors disclosed financial affiliations with Amgen, Kowa, Regeneron, Sanofi, and others.

SOURCE: Newman CB et al. Arterioscler Thromb Vasc Biol. 2018 Dec 10. doi: 10.1161/ATV.0000000000000073

The benefits of statins highly offset the associated risks in appropriate patients, according to a scientific statement issued by the American Heart Association.

“The review covers the general patient population, as well as demographic subgroups, including the elderly, children, pregnant women, East Asians, and patients with specific conditions.” wrote Connie B. Newman, MD, of New York University, together with her colleagues. The report is in Arteriosclerosis, Thrombosis, and Vascular Biology.

After an extensive review of the literature pertaining to statin safety and tolerability, Dr. Newman and her colleagues reported the compiled findings from several randomized controlled trials, in addition to observational data, where required. They found that the risk of serious muscle complications, such as rhabdomyolysis, attributable to statin use was less than 0.1%. Furthermore, they noted that the risk of serious hepatotoxicity was even less likely, occurring in about 1 in 10,000 patients treated with therapy.

“There is no convincing evidence for a causal relationship between statins and cancer, cataracts, cognitive dysfunction, peripheral neuropathy, erectile dysfunction, or tendinitis,” the experts wrote. “In U.S. clinical practices, roughly 10% of patients stop taking a statin because of subjective complaints, most commonly muscle symptoms without raised creatine kinase,” they further reported.

Contrastingly, data from randomized trials have shown that the change in the incidence of muscle-related symptoms in patients treated with statins versus placebo is less than 1%. Moreover, the incidence is even lower, with an estimated rate of 0.1%, in those who stopped statin therapy because of these symptoms. Given these results, Dr. Newman and her colleagues said that muscle-related symptoms among statin-treated patients are not due to the pharmacological activity of the statin.

“Restarting statin therapy in these patients can be challenging, but it is important, especially in patients at high risk of cardiovascular events, for whom prevention of these events is a priority,” they added.

A large proportion of the population takes statin therapy to lower the risk of major cardiovascular events, including ischemic stroke, myocardial infarction, and other adverse effects of cardiovascular disease. At maximal doses, statins may decrease LDL-cholesterol levels by roughly 55%-60%. In addition, given the multitude of available generics, statins are an economical treatment option for most patients.

However, Dr. Newman and her colleagues suggested that, when considering statin therapy in special populations, particularly in patients with end-stage renal failure or severe hepatic disease, commencing treatment is not recommended.

“The lack of proof of cardiovascular benefit in patients with end-stage renal disease suggests that initiating statin treatment in these patients is generally not warranted,” the experts wrote. “Data on safety in people with more serious liver disease are insufficient, and statin treatment is generally discouraged,” they added.

With respect to statin-induced adverse effects, they are usually reversible upon discontinuation of therapy, with the exception of hemorrhagic stroke. However, damage from an ischemic stroke or myocardial infarction may result in death. As a result, in patients who would benefit from statin therapy, based on most recent guidelines, cardiovascular benefits greatly exceed potential safety concerns.

Dr. Newman and her coauthors disclosed financial affiliations with Amgen, Kowa, Regeneron, Sanofi, and others.

SOURCE: Newman CB et al. Arterioscler Thromb Vasc Biol. 2018 Dec 10. doi: 10.1161/ATV.0000000000000073

The benefits of statins highly offset the associated risks in appropriate patients, according to a scientific statement issued by the American Heart Association.

“The review covers the general patient population, as well as demographic subgroups, including the elderly, children, pregnant women, East Asians, and patients with specific conditions.” wrote Connie B. Newman, MD, of New York University, together with her colleagues. The report is in Arteriosclerosis, Thrombosis, and Vascular Biology.

After an extensive review of the literature pertaining to statin safety and tolerability, Dr. Newman and her colleagues reported the compiled findings from several randomized controlled trials, in addition to observational data, where required. They found that the risk of serious muscle complications, such as rhabdomyolysis, attributable to statin use was less than 0.1%. Furthermore, they noted that the risk of serious hepatotoxicity was even less likely, occurring in about 1 in 10,000 patients treated with therapy.

“There is no convincing evidence for a causal relationship between statins and cancer, cataracts, cognitive dysfunction, peripheral neuropathy, erectile dysfunction, or tendinitis,” the experts wrote. “In U.S. clinical practices, roughly 10% of patients stop taking a statin because of subjective complaints, most commonly muscle symptoms without raised creatine kinase,” they further reported.

Contrastingly, data from randomized trials have shown that the change in the incidence of muscle-related symptoms in patients treated with statins versus placebo is less than 1%. Moreover, the incidence is even lower, with an estimated rate of 0.1%, in those who stopped statin therapy because of these symptoms. Given these results, Dr. Newman and her colleagues said that muscle-related symptoms among statin-treated patients are not due to the pharmacological activity of the statin.

“Restarting statin therapy in these patients can be challenging, but it is important, especially in patients at high risk of cardiovascular events, for whom prevention of these events is a priority,” they added.

A large proportion of the population takes statin therapy to lower the risk of major cardiovascular events, including ischemic stroke, myocardial infarction, and other adverse effects of cardiovascular disease. At maximal doses, statins may decrease LDL-cholesterol levels by roughly 55%-60%. In addition, given the multitude of available generics, statins are an economical treatment option for most patients.

However, Dr. Newman and her colleagues suggested that, when considering statin therapy in special populations, particularly in patients with end-stage renal failure or severe hepatic disease, commencing treatment is not recommended.

“The lack of proof of cardiovascular benefit in patients with end-stage renal disease suggests that initiating statin treatment in these patients is generally not warranted,” the experts wrote. “Data on safety in people with more serious liver disease are insufficient, and statin treatment is generally discouraged,” they added.

With respect to statin-induced adverse effects, they are usually reversible upon discontinuation of therapy, with the exception of hemorrhagic stroke. However, damage from an ischemic stroke or myocardial infarction may result in death. As a result, in patients who would benefit from statin therapy, based on most recent guidelines, cardiovascular benefits greatly exceed potential safety concerns.

Dr. Newman and her coauthors disclosed financial affiliations with Amgen, Kowa, Regeneron, Sanofi, and others.

SOURCE: Newman CB et al. Arterioscler Thromb Vasc Biol. 2018 Dec 10. doi: 10.1161/ATV.0000000000000073

With intensive insulin regimens and home blood glucose monitoring, patients with type 1 diabetes are controlling their blood glucose better than in the past. Nevertheless, glucose regulation is still imperfect and tedious, and striving for tight glycemic control poses the risk of hypoglycemia.

Prominent among the challenges are the sheer numbers involved. Some 1.25 million Americans have type 1 diabetes, and another 30 million have type 2, but only about 7,000 to 8,000 pancreases are available for transplant each year.¹ While awaiting a breakthrough—perhaps involving stem cells, perhaps involving organs obtained from animals—an insulin pump may offer better diabetes control for many. Another possibility is a closed-loop system with a continuous glucose monitor that drives a dual-infusion pump, delivering insulin when glucose levels rise too high, and glucagon when they dip too low.

DIABETES WAS KNOWN IN ANCIENT TIMES

About 3,000 years ago, Egyptians described the syndrome of thirst, emaciation, and sweet urine that attracted ants. The term diabetes (Greek for siphon) was first recorded in 1425; mellitus (Latin for sweet with honey) was not added until 1675.

In 1857, Bernard hypothesized that diabetes was caused by overproduction of glucose in the liver. This idea was replaced in 1889, when Mering and Minkowski proposed the dysfunctional pancreas theory that eventually led to the discovery of the beta cell.²

In 1921, Banting and Best isolated insulin, and for the past 100 years subcutaneous insulin replacement has been the mainstay of treatment. But starting about 50 years ago, researchers have been looking for safe and long-lasting ways to replace beta cells and eliminate the need for exogenous insulin replacement.

TRANSPLANTING THE WHOLE PANCREAS

The first whole-pancreas transplant was performed in 1966 by Kelly et al,³ followed by 13 more by 1973.⁴ These first transplant grafts were short-lived, with only 1 graft surviving longer than 1 year. Since then, more than 12,000 pancreases have been transplanted worldwide, as refinements in surgical techniques and immunosuppressive therapies have improved patient and graft survival rates.⁴

Today, most pancreas transplants are in patients who have both type 1 diabetes and end-stage renal disease due to diabetic neph­ropathy, and most receive both a kidney and a pancreas at the same time. Far fewer patients receive a pancreas after previously receiving a kidney, or receive a pancreas alone.

The bile duct of the transplanted pancreas is usually routed into the patient’s small intestine, as nature intended, and less often into the bladder. Although bladder drainage is associated with urinary complications, it has the advantage of allowing measurement of pancreatic amylase levels in the urine to monitor for graft rejection. With simultaneous pancreas and kidney transplant, the serum creatinine concentration can also be monitored for rejection of the kidney graft.

Current immunosuppressive regimens vary but generally consist of anti-T-cell antibodies at the time of surgery, followed by lifelong treatment with the combination of a calcineurin inhibitor (cyclosporine or tacrolimus) and an antimetabolite (mycophenolate mofetil or azathioprine).

Outcomes are good. The rates of patient and graft survival are highest with simultaneous pancreas-kidney transplant, and somewhat lower with pancreas-after-kidney and pancreas-alone transplant.

Benefits of pancreas transplant

Most recipients can stop taking insulin immediately after the procedure, and their hemoglobin A_1c levels normalize and stay low for the life of the graft. Lipid levels also decrease, although this has not been directly correlated with lower risk of vascular disease.⁴

Transplant also reduces or eliminates some complications of diabetes, including retinopathy, nephropathy, cardiomyopathy, and gastropathy.

For example, in patients undergoing simultaneous pancreas-kidney transplant, diabetic nephropathy does not recur in the new kidney. Fioretto et al⁵ reported that nephropathy lesions reversed during the 10 years after pancreas transplant.

Kennedy et al^6,7 found that preexisting diabetic neuropathy improved slightly (although neurologic status did not completely return to normal) over a period of up to 42 months in a group of patients who received a pancreas transplant, whereas it tended to worsen in a control group. Both groups were assessed at baseline and at 12 and 24 months, with a subgroup followed through 42 months, and they underwent testing of motor, sensory, and autonomic function.^6,7

Disadvantages of pancreas transplant

Disadvantages of whole-pancreas transplant include hypoglycemia (usually mild), adverse effects of immunosuppression, potential for surgical complications including an increased rate of death in the first 90 days after the procedure, and cost.

In an analysis comparing the 5-year estimated costs of dialysis, kidney transplant alone from cadavers or live donors, or simultaneous pancreas-kidney transplant for diabetic patients with end-stage renal disease, the least expensive option was kidney transplant from a live donor.⁸ The most expensive option was simultaneous pancreas-kidney transplant, but quality of life was better with this option. The analysis did not consider the potential cost of long-term treatments for complications related to diabetes that could be saved with a pancreas transplant.

Data conflict regarding the risk of death with different types of pancreas transplants. A retrospective cohort study of data from 124 US transplant centers reported in 2003 found higher mortality rates in pancreas-alone transplant recipients than in patients on a transplant waiting list receiving conventional therapy.⁹ In contrast, a 2004 study reported that after the first 90 days, when the risk of death was clearly higher, mortality rates were lower after simultaneous pancreas-kidney transplant and pancreas-after-kidney transplant.¹⁰ After pancreas-alone transplant, however, mortality rates were higher than with exogenous insulin therapy.

Although outcomes have improved, fewer patients with type 1 diabetes are undergoing pancreas transplant in recent years.

Interestingly, more simultaneous pancreas-kidney transplants are being successfully performed in patients with type 2 diabetes, who now account for 8% of all simultaneous pancreas-kidney transplant recipients.¹¹ Outcomes of pancreas transplant appear to be similar regardless of diabetes type.

Bottom line

Pancreas transplant is a viable option for certain cases of complicated diabetes.

TRANSPLANTING ISLET CELLS

Despite its successes, pancreas transplant is major surgery and requires lifetime immunosuppression. Research is ongoing into a less-invasive procedure that, it is hoped, would require less immunosuppression: transplanting islets by themselves.

Islet autotransplant after pancreatectomy

For some patients with chronic pancreatitis, the only option to relieve chronic pain, narcotic dependence, and poor quality of life is to remove the pancreas. In the past, this desperate measure would instantly and inevitably cause diabetes, but not anymore.

Alpha cells and glucagon are a different story; a complication of islet transplant is hypoglycemia. In 2016, Lin et al¹² reported spontaneous hypoglycemia in 6 of 12 patients who maintained insulin independence after autotransplant of islets. Although the transplanted islets had functional alpha cells that could in theory produce glucagon, as well as beta cells that produce insulin and C-peptide, apparently the alpha cells were not secreting glucagon in response to the hypoglycemia.

Location may matter. Gupta et al,¹³ in a 1997 study in dogs, found that more hypoglycemia occurs if islets are autotransplanted into the liver than if they are transplanted into the peritoneal cavity. A possible explanation may have to do with the glycemic environment of the liver.

Islet allotransplant

Islets can also be taken from cadaver donors and transplanted into patients with type 1 diabetes, who do not have enough working beta cells.

Success of allotransplant increased after the publication of observational data from the program in Edmonton in Canada, in which 7 consecutive patients with type 1 diabetes achieved initial insulin independence after islet allotransplant using steroid-free immunosuppression.¹⁴ Six recipients required islets from 2 donors, and 1 required islets from 4 donors, so they all received large volumes of at least 11,000 islet equivalents (IEQ) per kilogram of body weight.

In a subsequent report from the same team,¹⁵ 16 (44%) of 36 patients remained insulin-free at 1 year, and C-peptide secretion was detectable in 70% at 2 years. But despite the elevated C-peptide levels, only 5 patients remained insulin-independent by 2 years. Lower hemoglobin A_1c levels and decreases in hypoglycemic events from baseline also were noted.

The Clinical Islet Transplantation Consortium (CITC)¹⁶ and Collaborative Islet Transplant Registry (CITR)¹⁷ were established in 2004 to combine data and resources from centers around the world, including several that specialize in islet isolation and purification. Currently, more than 80 studies are being conducted.

The CITC and CITR now have data on more than 1,000 allogeneic islet transplant recipients (islet transplant alone, after kidney transplant, or simultaneous with it). The primary outcomes are hemoglobin A_1c levels below 7% fasting C-peptide levels 0.3 ng/mL or higher, and fasting blood glucose of 60 to 140 mg/dL with no severe hypoglycemic events. The best results for islet-alone transplant have been in recipients over age 35 who received at least 325,000 IEQs with use of tumor necrosis factor antagonists for induction and calcineurin inhibitors or mammalian target of rapamycin (mTOR) inhibitors for maintenance.¹⁷

The best success for islet-after-kidney transplant was achieved with the same protocol but with insulin given to the donor during hospitalization before pancreas procurement. For participants with favorable factors, a hemoglobin A_1c at or below 6.5% was achieved in about 80% at 1 year after last infusion, with more than 80% maintaining their fasting blood glucose level goals. About 70% of these patients were insulin-independent at 1 year. Hypoglycemia unawareness resolved in these patients even 5 years after infusion. Although there were no deaths or disabilities related to these transplants, bleeding occurred in 1 of 15 procedures. There was also a notable decline in estimated glomerular filtration rates with calcineurin inhibitor-based immunosuppression.¹⁷

Making islets go farther

One of the greatest challenges to islet transplant is the need for multiple donors to provide enough islet cells to overcome the loss of cells during transplant. Pancreases are already in short supply, and if each recipient needs more than 1, this makes the shortage worse. Some centers have achieved transplant with fewer donors,^18,19 possibly by selecting pancreases from young donors who had a high body mass index and more islet cells, and harvesting and using them with a shorter cold ischemic time.

The number of viable, functioning islet cells drastically decreases after transplant, especially when transplanted into the portal system. This phenomenon is linked to an instant, blood-mediated inflammatory reaction involving antibody binding, complement and coagulation cascade activation, and platelet aggregation. The reaction, part of the innate immune system, damages the islet cells and leads to insulin dumping and early graft loss in studies in vitro and in vivo. Another factor affecting the survival of the graft cells is the low oxygen tension in the portal system.

For this reason, sites such as the pancreas, gastric submucosa, genitourinary tract, muscle, omentum, bone marrow, kidney capsule, peritoneum, anterior eye chamber, testis, and thymus are being explored.²⁰

To create a more supportive environment for the transplanted cells, biotechnicians are trying to encapsulate islets in a semipermeable membrane that would protect them from the immune system while still allowing oxygen, nutrients, waste products, and, critically, insulin to diffuse in and out. Currently, no site or encapsulated product has been more successful than the current practice of implanting naked islets in the portal system.²⁰

Bottom line

Without advances in transplant sites or increasing the yield of islet cells to allow single-donor transplants, islet cell allotransplant will not be feasible for most patients with type 1 diabetes.

Xenotransplant: Can pig cells make up the shortage?

Use of animal kidneys (xenotransplant) is a potential solution to the shortage of human organs for transplant.

In theory, pigs could be a source. Porcine insulin is similar to human insulin (differing by only 1 amino acid), and it should be possible to breed “knockout” pigs that lack the antigens responsible for acute humoral rejection.²¹

On the other hand, transplant of porcine islets poses several immunologic, physiologic, ethical, legal, and infectious concerns. For example, porcine tissue could carry pig viruses, such as porcine endogenous retroviruses.²¹ And even if the pigs are genetically modified, patients will still require immunosuppressive therapy.

A review of 17 studies of pig islet xenotransplant into nonhuman primates found that in 5 of the studies (4 using diabetic primates) the grafts survived at least 3 months.²² Of these, 1 study used encapsulation, and the rest used intensive and toxic immunosuppression.

More research is needed to make xenotransplant a clinical option.

Transplanting stem cells or beta cells grown from stem cells

Stem cells provide an exciting potential alternative to the limited donor pool. During the past decade, several studies have shown success using human pluripotent stem cells (embryonic stem cells and human-induced pluripotent stem cells), mesenchymal stem cells isolated from adult tissues, and directly programmed somatic cells. Researchers have created stable cultures of pluripotent stem cells from embryonic stem cells, which could possibly be produced on a large scale and banked.²³

Human pluripotent stem cells derived from pancreatic progenitors have been shown to mature into more functional, islet-like structures in vivo. They transform into subtypes of islet cells including alpha, beta, and delta cells, ghrelin-producing cells, and pancreatic polypeptide hormone-producing cells. This process takes 2 to 6 weeks. In mice, these cells have been shown to maintain glucose homeostasis.²⁴ Phase 1 and 2 trials in humans are now being conducted.

Pagliuca et al²⁵ generated functional human pancreatic beta cells in vitro from embryonic stem cells. Rezania et al²⁴ reversed diabetes with insulin-producing cells derived in vitro from human pluripotent stem cells. The techniques used in these studies contributed to the success of a study by Vegas et al,²⁶ who achieved successful long-term glycemic control in mice using polymer-encapsulated human stem cell-derived beta cells.

Reversal of autoimmunity is an important step that needs to be overcome in stem cell transplant for type 1 diabetes. Nikolic et al²⁷ have achieved mixed allogeneic chimerism across major histocompatibility complex barriers with nonmyeloablative conditioning in advanced-diabetic nonobese diabetic mice. However, conditioning alone (ie, without bone marrow transplant) does not permit acceptance of allogeneic islets and does not reverse autoimmunity or allow islet regeneration.²⁸ Adding allogeneic bone marrow transplant to conditioned nonobese diabetic mice leads to tolerance to the donor and reverses autoimmunity.

THE ‘BIONIC’ PANCREAS

While we wait for advances in islet cell transplant, improved insulin pumps hold promise.

One such experimental device, the iLet (Beta Bionics, Boston, MA), designed by Damiano et al, consists of 2 infusion pumps (1 for insulin, 1 for glucagon) linked to a continuous glucose monitor via a smartphone app.

The monitor measures the glucose level every 5 minutes and transmits the information wirelessly to the phone app, which calculates the amount of insulin and glucagon required to stabilize the blood glucose: more insulin if too high, more glucagon if too low. The phone transmits this information to the pumps.

Dubbed the “bionic” pancreas, this closed-loop system frees patients from the tasks of measuring their glucose multiple times a day, calculating the appropriate dose, and giving multiple insulin injections.

The 2016 summer camp study²⁹ followed 19 preteens wearing the bionic pancreas for 5 days. During this time, the patients had lower mean glucose levels and less hypoglycemia than during control periods. No episodes of severe hypoglycemia were recorded.

El-Khatib et al³⁰ randomly assigned 43 patients to treatment with either the bihormonal bionic pancreas or usual care (a conventional insulin pump or a sensor-augmented insulin pump) for 11 days, followed by 11 days of the opposite treatment. All participants continued their normal activities. The bionic pancreas system was superior to the insulin pump in terms of the mean glucose concentration and mean time in the hypoglycemic range (P < .0001 for both results).

Bottom line

As the search continues for better solutions, advances in technology such as the bionic pancreas could provide a safer (ie, less hypoglycemic) and more successful alternative for insulin replacement in the near future.

With intensive insulin regimens and home blood glucose monitoring, patients with type 1 diabetes are controlling their blood glucose better than in the past. Nevertheless, glucose regulation is still imperfect and tedious, and striving for tight glycemic control poses the risk of hypoglycemia.

Prominent among the challenges are the sheer numbers involved. Some 1.25 million Americans have type 1 diabetes, and another 30 million have type 2, but only about 7,000 to 8,000 pancreases are available for transplant each year.¹ While awaiting a breakthrough—perhaps involving stem cells, perhaps involving organs obtained from animals—an insulin pump may offer better diabetes control for many. Another possibility is a closed-loop system with a continuous glucose monitor that drives a dual-infusion pump, delivering insulin when glucose levels rise too high, and glucagon when they dip too low.

DIABETES WAS KNOWN IN ANCIENT TIMES

About 3,000 years ago, Egyptians described the syndrome of thirst, emaciation, and sweet urine that attracted ants. The term diabetes (Greek for siphon) was first recorded in 1425; mellitus (Latin for sweet with honey) was not added until 1675.

In 1857, Bernard hypothesized that diabetes was caused by overproduction of glucose in the liver. This idea was replaced in 1889, when Mering and Minkowski proposed the dysfunctional pancreas theory that eventually led to the discovery of the beta cell.²

In 1921, Banting and Best isolated insulin, and for the past 100 years subcutaneous insulin replacement has been the mainstay of treatment. But starting about 50 years ago, researchers have been looking for safe and long-lasting ways to replace beta cells and eliminate the need for exogenous insulin replacement.

TRANSPLANTING THE WHOLE PANCREAS

The first whole-pancreas transplant was performed in 1966 by Kelly et al,³ followed by 13 more by 1973.⁴ These first transplant grafts were short-lived, with only 1 graft surviving longer than 1 year. Since then, more than 12,000 pancreases have been transplanted worldwide, as refinements in surgical techniques and immunosuppressive therapies have improved patient and graft survival rates.⁴

Today, most pancreas transplants are in patients who have both type 1 diabetes and end-stage renal disease due to diabetic neph­ropathy, and most receive both a kidney and a pancreas at the same time. Far fewer patients receive a pancreas after previously receiving a kidney, or receive a pancreas alone.

The bile duct of the transplanted pancreas is usually routed into the patient’s small intestine, as nature intended, and less often into the bladder. Although bladder drainage is associated with urinary complications, it has the advantage of allowing measurement of pancreatic amylase levels in the urine to monitor for graft rejection. With simultaneous pancreas and kidney transplant, the serum creatinine concentration can also be monitored for rejection of the kidney graft.

Current immunosuppressive regimens vary but generally consist of anti-T-cell antibodies at the time of surgery, followed by lifelong treatment with the combination of a calcineurin inhibitor (cyclosporine or tacrolimus) and an antimetabolite (mycophenolate mofetil or azathioprine).

Outcomes are good. The rates of patient and graft survival are highest with simultaneous pancreas-kidney transplant, and somewhat lower with pancreas-after-kidney and pancreas-alone transplant.

Benefits of pancreas transplant

Most recipients can stop taking insulin immediately after the procedure, and their hemoglobin A_1c levels normalize and stay low for the life of the graft. Lipid levels also decrease, although this has not been directly correlated with lower risk of vascular disease.⁴

Transplant also reduces or eliminates some complications of diabetes, including retinopathy, nephropathy, cardiomyopathy, and gastropathy.

For example, in patients undergoing simultaneous pancreas-kidney transplant, diabetic nephropathy does not recur in the new kidney. Fioretto et al⁵ reported that nephropathy lesions reversed during the 10 years after pancreas transplant.

Kennedy et al^6,7 found that preexisting diabetic neuropathy improved slightly (although neurologic status did not completely return to normal) over a period of up to 42 months in a group of patients who received a pancreas transplant, whereas it tended to worsen in a control group. Both groups were assessed at baseline and at 12 and 24 months, with a subgroup followed through 42 months, and they underwent testing of motor, sensory, and autonomic function.^6,7

Disadvantages of pancreas transplant

Disadvantages of whole-pancreas transplant include hypoglycemia (usually mild), adverse effects of immunosuppression, potential for surgical complications including an increased rate of death in the first 90 days after the procedure, and cost.

In an analysis comparing the 5-year estimated costs of dialysis, kidney transplant alone from cadavers or live donors, or simultaneous pancreas-kidney transplant for diabetic patients with end-stage renal disease, the least expensive option was kidney transplant from a live donor.⁸ The most expensive option was simultaneous pancreas-kidney transplant, but quality of life was better with this option. The analysis did not consider the potential cost of long-term treatments for complications related to diabetes that could be saved with a pancreas transplant.

Data conflict regarding the risk of death with different types of pancreas transplants. A retrospective cohort study of data from 124 US transplant centers reported in 2003 found higher mortality rates in pancreas-alone transplant recipients than in patients on a transplant waiting list receiving conventional therapy.⁹ In contrast, a 2004 study reported that after the first 90 days, when the risk of death was clearly higher, mortality rates were lower after simultaneous pancreas-kidney transplant and pancreas-after-kidney transplant.¹⁰ After pancreas-alone transplant, however, mortality rates were higher than with exogenous insulin therapy.

Although outcomes have improved, fewer patients with type 1 diabetes are undergoing pancreas transplant in recent years.

Interestingly, more simultaneous pancreas-kidney transplants are being successfully performed in patients with type 2 diabetes, who now account for 8% of all simultaneous pancreas-kidney transplant recipients.¹¹ Outcomes of pancreas transplant appear to be similar regardless of diabetes type.

Bottom line

Pancreas transplant is a viable option for certain cases of complicated diabetes.

TRANSPLANTING ISLET CELLS

Despite its successes, pancreas transplant is major surgery and requires lifetime immunosuppression. Research is ongoing into a less-invasive procedure that, it is hoped, would require less immunosuppression: transplanting islets by themselves.

Islet autotransplant after pancreatectomy

For some patients with chronic pancreatitis, the only option to relieve chronic pain, narcotic dependence, and poor quality of life is to remove the pancreas. In the past, this desperate measure would instantly and inevitably cause diabetes, but not anymore.

Alpha cells and glucagon are a different story; a complication of islet transplant is hypoglycemia. In 2016, Lin et al¹² reported spontaneous hypoglycemia in 6 of 12 patients who maintained insulin independence after autotransplant of islets. Although the transplanted islets had functional alpha cells that could in theory produce glucagon, as well as beta cells that produce insulin and C-peptide, apparently the alpha cells were not secreting glucagon in response to the hypoglycemia.

Location may matter. Gupta et al,¹³ in a 1997 study in dogs, found that more hypoglycemia occurs if islets are autotransplanted into the liver than if they are transplanted into the peritoneal cavity. A possible explanation may have to do with the glycemic environment of the liver.

Islet allotransplant

Islets can also be taken from cadaver donors and transplanted into patients with type 1 diabetes, who do not have enough working beta cells.

Success of allotransplant increased after the publication of observational data from the program in Edmonton in Canada, in which 7 consecutive patients with type 1 diabetes achieved initial insulin independence after islet allotransplant using steroid-free immunosuppression.¹⁴ Six recipients required islets from 2 donors, and 1 required islets from 4 donors, so they all received large volumes of at least 11,000 islet equivalents (IEQ) per kilogram of body weight.

In a subsequent report from the same team,¹⁵ 16 (44%) of 36 patients remained insulin-free at 1 year, and C-peptide secretion was detectable in 70% at 2 years. But despite the elevated C-peptide levels, only 5 patients remained insulin-independent by 2 years. Lower hemoglobin A_1c levels and decreases in hypoglycemic events from baseline also were noted.

The Clinical Islet Transplantation Consortium (CITC)¹⁶ and Collaborative Islet Transplant Registry (CITR)¹⁷ were established in 2004 to combine data and resources from centers around the world, including several that specialize in islet isolation and purification. Currently, more than 80 studies are being conducted.

The CITC and CITR now have data on more than 1,000 allogeneic islet transplant recipients (islet transplant alone, after kidney transplant, or simultaneous with it). The primary outcomes are hemoglobin A_1c levels below 7% fasting C-peptide levels 0.3 ng/mL or higher, and fasting blood glucose of 60 to 140 mg/dL with no severe hypoglycemic events. The best results for islet-alone transplant have been in recipients over age 35 who received at least 325,000 IEQs with use of tumor necrosis factor antagonists for induction and calcineurin inhibitors or mammalian target of rapamycin (mTOR) inhibitors for maintenance.¹⁷

The best success for islet-after-kidney transplant was achieved with the same protocol but with insulin given to the donor during hospitalization before pancreas procurement. For participants with favorable factors, a hemoglobin A_1c at or below 6.5% was achieved in about 80% at 1 year after last infusion, with more than 80% maintaining their fasting blood glucose level goals. About 70% of these patients were insulin-independent at 1 year. Hypoglycemia unawareness resolved in these patients even 5 years after infusion. Although there were no deaths or disabilities related to these transplants, bleeding occurred in 1 of 15 procedures. There was also a notable decline in estimated glomerular filtration rates with calcineurin inhibitor-based immunosuppression.¹⁷

Making islets go farther

One of the greatest challenges to islet transplant is the need for multiple donors to provide enough islet cells to overcome the loss of cells during transplant. Pancreases are already in short supply, and if each recipient needs more than 1, this makes the shortage worse. Some centers have achieved transplant with fewer donors,^18,19 possibly by selecting pancreases from young donors who had a high body mass index and more islet cells, and harvesting and using them with a shorter cold ischemic time.

The number of viable, functioning islet cells drastically decreases after transplant, especially when transplanted into the portal system. This phenomenon is linked to an instant, blood-mediated inflammatory reaction involving antibody binding, complement and coagulation cascade activation, and platelet aggregation. The reaction, part of the innate immune system, damages the islet cells and leads to insulin dumping and early graft loss in studies in vitro and in vivo. Another factor affecting the survival of the graft cells is the low oxygen tension in the portal system.

For this reason, sites such as the pancreas, gastric submucosa, genitourinary tract, muscle, omentum, bone marrow, kidney capsule, peritoneum, anterior eye chamber, testis, and thymus are being explored.²⁰

To create a more supportive environment for the transplanted cells, biotechnicians are trying to encapsulate islets in a semipermeable membrane that would protect them from the immune system while still allowing oxygen, nutrients, waste products, and, critically, insulin to diffuse in and out. Currently, no site or encapsulated product has been more successful than the current practice of implanting naked islets in the portal system.²⁰

Bottom line

Without advances in transplant sites or increasing the yield of islet cells to allow single-donor transplants, islet cell allotransplant will not be feasible for most patients with type 1 diabetes.

Xenotransplant: Can pig cells make up the shortage?

Use of animal kidneys (xenotransplant) is a potential solution to the shortage of human organs for transplant.

In theory, pigs could be a source. Porcine insulin is similar to human insulin (differing by only 1 amino acid), and it should be possible to breed “knockout” pigs that lack the antigens responsible for acute humoral rejection.²¹

On the other hand, transplant of porcine islets poses several immunologic, physiologic, ethical, legal, and infectious concerns. For example, porcine tissue could carry pig viruses, such as porcine endogenous retroviruses.²¹ And even if the pigs are genetically modified, patients will still require immunosuppressive therapy.

A review of 17 studies of pig islet xenotransplant into nonhuman primates found that in 5 of the studies (4 using diabetic primates) the grafts survived at least 3 months.²² Of these, 1 study used encapsulation, and the rest used intensive and toxic immunosuppression.

More research is needed to make xenotransplant a clinical option.

Transplanting stem cells or beta cells grown from stem cells

Stem cells provide an exciting potential alternative to the limited donor pool. During the past decade, several studies have shown success using human pluripotent stem cells (embryonic stem cells and human-induced pluripotent stem cells), mesenchymal stem cells isolated from adult tissues, and directly programmed somatic cells. Researchers have created stable cultures of pluripotent stem cells from embryonic stem cells, which could possibly be produced on a large scale and banked.²³

Human pluripotent stem cells derived from pancreatic progenitors have been shown to mature into more functional, islet-like structures in vivo. They transform into subtypes of islet cells including alpha, beta, and delta cells, ghrelin-producing cells, and pancreatic polypeptide hormone-producing cells. This process takes 2 to 6 weeks. In mice, these cells have been shown to maintain glucose homeostasis.²⁴ Phase 1 and 2 trials in humans are now being conducted.

Pagliuca et al²⁵ generated functional human pancreatic beta cells in vitro from embryonic stem cells. Rezania et al²⁴ reversed diabetes with insulin-producing cells derived in vitro from human pluripotent stem cells. The techniques used in these studies contributed to the success of a study by Vegas et al,²⁶ who achieved successful long-term glycemic control in mice using polymer-encapsulated human stem cell-derived beta cells.

Reversal of autoimmunity is an important step that needs to be overcome in stem cell transplant for type 1 diabetes. Nikolic et al²⁷ have achieved mixed allogeneic chimerism across major histocompatibility complex barriers with nonmyeloablative conditioning in advanced-diabetic nonobese diabetic mice. However, conditioning alone (ie, without bone marrow transplant) does not permit acceptance of allogeneic islets and does not reverse autoimmunity or allow islet regeneration.²⁸ Adding allogeneic bone marrow transplant to conditioned nonobese diabetic mice leads to tolerance to the donor and reverses autoimmunity.

THE ‘BIONIC’ PANCREAS

While we wait for advances in islet cell transplant, improved insulin pumps hold promise.

One such experimental device, the iLet (Beta Bionics, Boston, MA), designed by Damiano et al, consists of 2 infusion pumps (1 for insulin, 1 for glucagon) linked to a continuous glucose monitor via a smartphone app.

The monitor measures the glucose level every 5 minutes and transmits the information wirelessly to the phone app, which calculates the amount of insulin and glucagon required to stabilize the blood glucose: more insulin if too high, more glucagon if too low. The phone transmits this information to the pumps.

Dubbed the “bionic” pancreas, this closed-loop system frees patients from the tasks of measuring their glucose multiple times a day, calculating the appropriate dose, and giving multiple insulin injections.

The 2016 summer camp study²⁹ followed 19 preteens wearing the bionic pancreas for 5 days. During this time, the patients had lower mean glucose levels and less hypoglycemia than during control periods. No episodes of severe hypoglycemia were recorded.

El-Khatib et al³⁰ randomly assigned 43 patients to treatment with either the bihormonal bionic pancreas or usual care (a conventional insulin pump or a sensor-augmented insulin pump) for 11 days, followed by 11 days of the opposite treatment. All participants continued their normal activities. The bionic pancreas system was superior to the insulin pump in terms of the mean glucose concentration and mean time in the hypoglycemic range (P < .0001 for both results).

Bottom line

As the search continues for better solutions, advances in technology such as the bionic pancreas could provide a safer (ie, less hypoglycemic) and more successful alternative for insulin replacement in the near future.

With intensive insulin regimens and home blood glucose monitoring, patients with type 1 diabetes are controlling their blood glucose better than in the past. Nevertheless, glucose regulation is still imperfect and tedious, and striving for tight glycemic control poses the risk of hypoglycemia.

Prominent among the challenges are the sheer numbers involved. Some 1.25 million Americans have type 1 diabetes, and another 30 million have type 2, but only about 7,000 to 8,000 pancreases are available for transplant each year.¹ While awaiting a breakthrough—perhaps involving stem cells, perhaps involving organs obtained from animals—an insulin pump may offer better diabetes control for many. Another possibility is a closed-loop system with a continuous glucose monitor that drives a dual-infusion pump, delivering insulin when glucose levels rise too high, and glucagon when they dip too low.

DIABETES WAS KNOWN IN ANCIENT TIMES

About 3,000 years ago, Egyptians described the syndrome of thirst, emaciation, and sweet urine that attracted ants. The term diabetes (Greek for siphon) was first recorded in 1425; mellitus (Latin for sweet with honey) was not added until 1675.

In 1857, Bernard hypothesized that diabetes was caused by overproduction of glucose in the liver. This idea was replaced in 1889, when Mering and Minkowski proposed the dysfunctional pancreas theory that eventually led to the discovery of the beta cell.²

In 1921, Banting and Best isolated insulin, and for the past 100 years subcutaneous insulin replacement has been the mainstay of treatment. But starting about 50 years ago, researchers have been looking for safe and long-lasting ways to replace beta cells and eliminate the need for exogenous insulin replacement.

TRANSPLANTING THE WHOLE PANCREAS

The first whole-pancreas transplant was performed in 1966 by Kelly et al,³ followed by 13 more by 1973.⁴ These first transplant grafts were short-lived, with only 1 graft surviving longer than 1 year. Since then, more than 12,000 pancreases have been transplanted worldwide, as refinements in surgical techniques and immunosuppressive therapies have improved patient and graft survival rates.⁴

Today, most pancreas transplants are in patients who have both type 1 diabetes and end-stage renal disease due to diabetic neph­ropathy, and most receive both a kidney and a pancreas at the same time. Far fewer patients receive a pancreas after previously receiving a kidney, or receive a pancreas alone.

The bile duct of the transplanted pancreas is usually routed into the patient’s small intestine, as nature intended, and less often into the bladder. Although bladder drainage is associated with urinary complications, it has the advantage of allowing measurement of pancreatic amylase levels in the urine to monitor for graft rejection. With simultaneous pancreas and kidney transplant, the serum creatinine concentration can also be monitored for rejection of the kidney graft.

Current immunosuppressive regimens vary but generally consist of anti-T-cell antibodies at the time of surgery, followed by lifelong treatment with the combination of a calcineurin inhibitor (cyclosporine or tacrolimus) and an antimetabolite (mycophenolate mofetil or azathioprine).

Outcomes are good. The rates of patient and graft survival are highest with simultaneous pancreas-kidney transplant, and somewhat lower with pancreas-after-kidney and pancreas-alone transplant.

Benefits of pancreas transplant

Most recipients can stop taking insulin immediately after the procedure, and their hemoglobin A_1c levels normalize and stay low for the life of the graft. Lipid levels also decrease, although this has not been directly correlated with lower risk of vascular disease.⁴

Transplant also reduces or eliminates some complications of diabetes, including retinopathy, nephropathy, cardiomyopathy, and gastropathy.

For example, in patients undergoing simultaneous pancreas-kidney transplant, diabetic nephropathy does not recur in the new kidney. Fioretto et al⁵ reported that nephropathy lesions reversed during the 10 years after pancreas transplant.

Kennedy et al^6,7 found that preexisting diabetic neuropathy improved slightly (although neurologic status did not completely return to normal) over a period of up to 42 months in a group of patients who received a pancreas transplant, whereas it tended to worsen in a control group. Both groups were assessed at baseline and at 12 and 24 months, with a subgroup followed through 42 months, and they underwent testing of motor, sensory, and autonomic function.^6,7

Disadvantages of pancreas transplant

Disadvantages of whole-pancreas transplant include hypoglycemia (usually mild), adverse effects of immunosuppression, potential for surgical complications including an increased rate of death in the first 90 days after the procedure, and cost.

In an analysis comparing the 5-year estimated costs of dialysis, kidney transplant alone from cadavers or live donors, or simultaneous pancreas-kidney transplant for diabetic patients with end-stage renal disease, the least expensive option was kidney transplant from a live donor.⁸ The most expensive option was simultaneous pancreas-kidney transplant, but quality of life was better with this option. The analysis did not consider the potential cost of long-term treatments for complications related to diabetes that could be saved with a pancreas transplant.

Data conflict regarding the risk of death with different types of pancreas transplants. A retrospective cohort study of data from 124 US transplant centers reported in 2003 found higher mortality rates in pancreas-alone transplant recipients than in patients on a transplant waiting list receiving conventional therapy.⁹ In contrast, a 2004 study reported that after the first 90 days, when the risk of death was clearly higher, mortality rates were lower after simultaneous pancreas-kidney transplant and pancreas-after-kidney transplant.¹⁰ After pancreas-alone transplant, however, mortality rates were higher than with exogenous insulin therapy.

Although outcomes have improved, fewer patients with type 1 diabetes are undergoing pancreas transplant in recent years.

Interestingly, more simultaneous pancreas-kidney transplants are being successfully performed in patients with type 2 diabetes, who now account for 8% of all simultaneous pancreas-kidney transplant recipients.¹¹ Outcomes of pancreas transplant appear to be similar regardless of diabetes type.

Bottom line

Pancreas transplant is a viable option for certain cases of complicated diabetes.

TRANSPLANTING ISLET CELLS

Despite its successes, pancreas transplant is major surgery and requires lifetime immunosuppression. Research is ongoing into a less-invasive procedure that, it is hoped, would require less immunosuppression: transplanting islets by themselves.

Islet autotransplant after pancreatectomy

For some patients with chronic pancreatitis, the only option to relieve chronic pain, narcotic dependence, and poor quality of life is to remove the pancreas. In the past, this desperate measure would instantly and inevitably cause diabetes, but not anymore.

Alpha cells and glucagon are a different story; a complication of islet transplant is hypoglycemia. In 2016, Lin et al¹² reported spontaneous hypoglycemia in 6 of 12 patients who maintained insulin independence after autotransplant of islets. Although the transplanted islets had functional alpha cells that could in theory produce glucagon, as well as beta cells that produce insulin and C-peptide, apparently the alpha cells were not secreting glucagon in response to the hypoglycemia.

Location may matter. Gupta et al,¹³ in a 1997 study in dogs, found that more hypoglycemia occurs if islets are autotransplanted into the liver than if they are transplanted into the peritoneal cavity. A possible explanation may have to do with the glycemic environment of the liver.

Islet allotransplant

Islets can also be taken from cadaver donors and transplanted into patients with type 1 diabetes, who do not have enough working beta cells.

Success of allotransplant increased after the publication of observational data from the program in Edmonton in Canada, in which 7 consecutive patients with type 1 diabetes achieved initial insulin independence after islet allotransplant using steroid-free immunosuppression.¹⁴ Six recipients required islets from 2 donors, and 1 required islets from 4 donors, so they all received large volumes of at least 11,000 islet equivalents (IEQ) per kilogram of body weight.

In a subsequent report from the same team,¹⁵ 16 (44%) of 36 patients remained insulin-free at 1 year, and C-peptide secretion was detectable in 70% at 2 years. But despite the elevated C-peptide levels, only 5 patients remained insulin-independent by 2 years. Lower hemoglobin A_1c levels and decreases in hypoglycemic events from baseline also were noted.

The Clinical Islet Transplantation Consortium (CITC)¹⁶ and Collaborative Islet Transplant Registry (CITR)¹⁷ were established in 2004 to combine data and resources from centers around the world, including several that specialize in islet isolation and purification. Currently, more than 80 studies are being conducted.

The CITC and CITR now have data on more than 1,000 allogeneic islet transplant recipients (islet transplant alone, after kidney transplant, or simultaneous with it). The primary outcomes are hemoglobin A_1c levels below 7% fasting C-peptide levels 0.3 ng/mL or higher, and fasting blood glucose of 60 to 140 mg/dL with no severe hypoglycemic events. The best results for islet-alone transplant have been in recipients over age 35 who received at least 325,000 IEQs with use of tumor necrosis factor antagonists for induction and calcineurin inhibitors or mammalian target of rapamycin (mTOR) inhibitors for maintenance.¹⁷

The best success for islet-after-kidney transplant was achieved with the same protocol but with insulin given to the donor during hospitalization before pancreas procurement. For participants with favorable factors, a hemoglobin A_1c at or below 6.5% was achieved in about 80% at 1 year after last infusion, with more than 80% maintaining their fasting blood glucose level goals. About 70% of these patients were insulin-independent at 1 year. Hypoglycemia unawareness resolved in these patients even 5 years after infusion. Although there were no deaths or disabilities related to these transplants, bleeding occurred in 1 of 15 procedures. There was also a notable decline in estimated glomerular filtration rates with calcineurin inhibitor-based immunosuppression.¹⁷

Making islets go farther

One of the greatest challenges to islet transplant is the need for multiple donors to provide enough islet cells to overcome the loss of cells during transplant. Pancreases are already in short supply, and if each recipient needs more than 1, this makes the shortage worse. Some centers have achieved transplant with fewer donors,^18,19 possibly by selecting pancreases from young donors who had a high body mass index and more islet cells, and harvesting and using them with a shorter cold ischemic time.

The number of viable, functioning islet cells drastically decreases after transplant, especially when transplanted into the portal system. This phenomenon is linked to an instant, blood-mediated inflammatory reaction involving antibody binding, complement and coagulation cascade activation, and platelet aggregation. The reaction, part of the innate immune system, damages the islet cells and leads to insulin dumping and early graft loss in studies in vitro and in vivo. Another factor affecting the survival of the graft cells is the low oxygen tension in the portal system.

For this reason, sites such as the pancreas, gastric submucosa, genitourinary tract, muscle, omentum, bone marrow, kidney capsule, peritoneum, anterior eye chamber, testis, and thymus are being explored.²⁰

To create a more supportive environment for the transplanted cells, biotechnicians are trying to encapsulate islets in a semipermeable membrane that would protect them from the immune system while still allowing oxygen, nutrients, waste products, and, critically, insulin to diffuse in and out. Currently, no site or encapsulated product has been more successful than the current practice of implanting naked islets in the portal system.²⁰

Bottom line

Without advances in transplant sites or increasing the yield of islet cells to allow single-donor transplants, islet cell allotransplant will not be feasible for most patients with type 1 diabetes.

Xenotransplant: Can pig cells make up the shortage?

Use of animal kidneys (xenotransplant) is a potential solution to the shortage of human organs for transplant.

In theory, pigs could be a source. Porcine insulin is similar to human insulin (differing by only 1 amino acid), and it should be possible to breed “knockout” pigs that lack the antigens responsible for acute humoral rejection.²¹

On the other hand, transplant of porcine islets poses several immunologic, physiologic, ethical, legal, and infectious concerns. For example, porcine tissue could carry pig viruses, such as porcine endogenous retroviruses.²¹ And even if the pigs are genetically modified, patients will still require immunosuppressive therapy.

A review of 17 studies of pig islet xenotransplant into nonhuman primates found that in 5 of the studies (4 using diabetic primates) the grafts survived at least 3 months.²² Of these, 1 study used encapsulation, and the rest used intensive and toxic immunosuppression.

More research is needed to make xenotransplant a clinical option.

Transplanting stem cells or beta cells grown from stem cells

Stem cells provide an exciting potential alternative to the limited donor pool. During the past decade, several studies have shown success using human pluripotent stem cells (embryonic stem cells and human-induced pluripotent stem cells), mesenchymal stem cells isolated from adult tissues, and directly programmed somatic cells. Researchers have created stable cultures of pluripotent stem cells from embryonic stem cells, which could possibly be produced on a large scale and banked.²³

Human pluripotent stem cells derived from pancreatic progenitors have been shown to mature into more functional, islet-like structures in vivo. They transform into subtypes of islet cells including alpha, beta, and delta cells, ghrelin-producing cells, and pancreatic polypeptide hormone-producing cells. This process takes 2 to 6 weeks. In mice, these cells have been shown to maintain glucose homeostasis.²⁴ Phase 1 and 2 trials in humans are now being conducted.

Pagliuca et al²⁵ generated functional human pancreatic beta cells in vitro from embryonic stem cells. Rezania et al²⁴ reversed diabetes with insulin-producing cells derived in vitro from human pluripotent stem cells. The techniques used in these studies contributed to the success of a study by Vegas et al,²⁶ who achieved successful long-term glycemic control in mice using polymer-encapsulated human stem cell-derived beta cells.

Reversal of autoimmunity is an important step that needs to be overcome in stem cell transplant for type 1 diabetes. Nikolic et al²⁷ have achieved mixed allogeneic chimerism across major histocompatibility complex barriers with nonmyeloablative conditioning in advanced-diabetic nonobese diabetic mice. However, conditioning alone (ie, without bone marrow transplant) does not permit acceptance of allogeneic islets and does not reverse autoimmunity or allow islet regeneration.²⁸ Adding allogeneic bone marrow transplant to conditioned nonobese diabetic mice leads to tolerance to the donor and reverses autoimmunity.

THE ‘BIONIC’ PANCREAS

While we wait for advances in islet cell transplant, improved insulin pumps hold promise.

One such experimental device, the iLet (Beta Bionics, Boston, MA), designed by Damiano et al, consists of 2 infusion pumps (1 for insulin, 1 for glucagon) linked to a continuous glucose monitor via a smartphone app.

The monitor measures the glucose level every 5 minutes and transmits the information wirelessly to the phone app, which calculates the amount of insulin and glucagon required to stabilize the blood glucose: more insulin if too high, more glucagon if too low. The phone transmits this information to the pumps.

Dubbed the “bionic” pancreas, this closed-loop system frees patients from the tasks of measuring their glucose multiple times a day, calculating the appropriate dose, and giving multiple insulin injections.

The 2016 summer camp study²⁹ followed 19 preteens wearing the bionic pancreas for 5 days. During this time, the patients had lower mean glucose levels and less hypoglycemia than during control periods. No episodes of severe hypoglycemia were recorded.

El-Khatib et al³⁰ randomly assigned 43 patients to treatment with either the bihormonal bionic pancreas or usual care (a conventional insulin pump or a sensor-augmented insulin pump) for 11 days, followed by 11 days of the opposite treatment. All participants continued their normal activities. The bionic pancreas system was superior to the insulin pump in terms of the mean glucose concentration and mean time in the hypoglycemic range (P < .0001 for both results).

Bottom line

As the search continues for better solutions, advances in technology such as the bionic pancreas could provide a safer (ie, less hypoglycemic) and more successful alternative for insulin replacement in the near future.

Marijuana, which is still illegal under federal law but legal in 30 states for medical purposes as of this writing, has shown promising results for treating peripheral neuropathy. Studies suggest that cannabis may be an option for patients whose pain responds poorly to standard treatments; however, its use may be restricted by cognitive and psychiatric adverse effects, particularly at high doses.¹

See related editorial

In this article, we discuss the basic pharmacology of cannabis and how it may affect neuropathic pain. We review clinical trials on its use for peripheral neuropathy and provide guidance for its use.

PERIPHERAL NEUROPATHY IS COMMON AND COMPLEX

An estimated 20 million people in the United States suffer from neuropathic pain. The prevalence is higher in certain populations, with 26% of people over age 65 and 30% of patients with diabetes mellitus affected.^2–4

Peripheral neuropathy is a complex, chronic state that occurs when nerve fibers are damaged, dysfunctional, or injured, sending incorrect signals to pain centers in the central nervous system.⁵ It is characterized by weakness, pain, and paresthesias that typically begin in the hands or feet and progress proximally.⁴ Symptoms depend on the number and types of nerves affected.

In many cases, peripheral neuropathy is idiopathic, but common causes include diabetes, alcoholism, human immunodeficiency virus (HIV) infection, and autoimmune disease. Others include toxicity from chemotherapy and heavy metals.

Peripheral neuropathy significantly worsens quality of life and function. Many patients experience emotional, cognitive, and functional problems, resulting in high rates of medical and psychiatric comorbidities and occupational impairment.^4,6,7 Yet despite its clinical and epidemiologic significance, it is often undertreated.⁸

STANDARD TREATMENTS INADEQUATE

Peripheral neuropathy occurs in patients with a wide range of comorbidities and is especially difficult to treat. Mainstays of therapy include anticonvulsants, tricyclic antidepressants, and serotonin-norepinephrine reuptake inhibitors.⁹ A more invasive option is spinal cord stimulation.

These treatments can have considerable adverse effects, and response rates remain suboptimal, with pain relief insufficient to improve quality of life for many patients.^9,10 Better treatments are needed to improve clinical outcomes and patient experience.¹¹

CANNABIS: A MIX OF COMPOUNDS

Cannabis sativa has been used as an analgesic for centuries. The plant contains more than 400 chemical compounds and is often used for its euphoric properties. Long-term use may lead to addiction and cognitive impairment.^12,13

Tetrahydrocannabinol (THC) and cannabidiol (CBD) are the main components and the 2 best-studied cannabinoids with analgesic effects.

THC is the primary psychoactive component of cannabis. Its effects include relaxation, altered perception, heightened sensations, increased libido, and perceptual distortions of time and space. Temporary effects may include decreased short-term memory, dry mouth, impaired motor function, conjunctival injection, paranoia, and anxiety.

CBD is nonpsychoactive and has anti-inflammatory and antioxidant properties. It has been shown to reduce pain and inflammation without the effects of THC.¹⁴

Other compounds in the cannabis plant include phytocannabinoids, flavonoids, and tapenoids, which may produce individual, interactive, or synergistic effects.¹⁵ Different strains of cannabis have varying amounts of the individual components, making comparisons among clinical studies difficult.

THE ENDOCANNABINOID SYSTEM

The endogenous mammalian cannabinoid system plays a regulatory role in the development, homeostasis, and neuroplasticity of the central nervous system. It is also involved in modulating pain transmission in the nociceptive pathway.

Two of the most abundant cannabinoid endogenous ligands are anandamide and 2-arachidonylglycerol.⁹ These endocanna­b­inoids are produced on demand in the central nervous system to reduce pain by acting as a circuit breaker.^16–18 They target the G protein-coupled cannabinoid receptors CB1 and CB2, located throughout the central and peripheral nervous system and in organs and tissues.¹²

CB1 receptors are found primarily in the central nervous system, specifically in areas involved in movement, such as the basal ganglia and cerebellum, as well as in areas involved in memory, such as the hippocampus.¹² They are also abundant in brain regions implicated in conducting and modulating pain signals, including the periaqueductal gray and the dorsal horn of the spinal cord.^16–20

CB2 receptors are mostly found in peripheral tissues and organs, mainly those involved in the immune system, including splenic, tonsillar, and hematopoietic cells.¹² They help regulate inflammation, allodynia, and hyperalgesia.¹⁷

Modifying response to injury

Following a nerve injury, neurons along the nociceptive pathway may become more reactive and responsive in a process known as sensitization.²¹ The process involves a cascade of cellular events that result in sprouting of pain-sensitive nerve endings.^21,22

Cannabinoids are thought to reduce pain by modifying these cellular events. They also inhibit nociceptive conduction in the dorsal horn of the spinal cord and in the ascending spinothalamic tract.²⁰ CB1 receptors found in nociceptive terminals along the peripheral nervous system impede pain conduction, while activation of CB2 receptors in immune cells decreases the release of nociceptive agents.

 

 

STUDIES OF CANNABIS FOR NEUROPATHIC PAIN

A number of studies have evaluated cannabis for treating neuropathic pain. Overall, available data support the efficacy of smoked or inhaled cannabis in its flower form when used as monotherapy or adjunctive therapy for relief of neuropathic pain of various etiologies. Many studies also report secondary benefits, including better sleep and functional improvement.^23,24

However, adverse effects are common, especially at high doses, and include difficulty concentrating, lightheadedness, fatigue, and tachycardia. More serious reported adverse effects include anxiety, paranoia, and psychosis.

Wilsey et al, 2008: Neuropathic pain reduced

Wilsey et al²⁵ conducted a double-blind, placebo-controlled crossover study that assessed the effects of smoking cannabis in 38 patients with central or peripheral neuropathic pain. Participants were assigned to smoke either high- or low-dose cannabis (7% or 3.5% delta-9-THC) or placebo cigarettes. Cigarettes were smoked during treatment sessions using the following regimen: 2 puffs at 60 minutes from baseline, 3 puffs at 120 minutes, and 4 puffs at 180 minutes. Patients were assessed after each set of puffs and for 2 hours afterwards. The primary outcome was spontaneous relief of pain as measured by a visual analog scale.

Pain intensity was comparable and significantly reduced in both treatment groups compared with placebo. At the high dose, some participants experienced neurocognitive impairment in attention, learning, memory, and psychomotor speed; only learning and memory declined at the low dose.

Ellis et al, 2009: Pain reduction in HIV neuropathy

Ellis et al²³ conducted a double-blind, placebo-controlled crossover trial in patients with HIV neuropathy that was unresponsive to at least 2 analgesics with different modes of action. During each treatment week, participants were randomly assigned to smoke either active cannabis or placebo, while continuing their standard therapy. Titration started at 4% THC and was adjusted based on tolerability and efficacy. Twenty-eight of the 34 enrolled patients completed both cannabis and placebo treatments. The principal outcome was change in pain intensity from baseline at the end of each week, using the Descriptor Differential Scale of Pain Intensity.

Of the 28 patients, 46% achieved an average pain reduction of 3.3 points (30%). One patient experienced cannabis-induced psychosis, and another developed an intractable cough, which resolved with smoking cessation.

Ware et al, 2010: Reduced posttraumatic or postsurgical neuropathic pain

Ware et al²⁴ performed a randomized crossover trial in 21 patients with posttraumatic or postsurgical neuropathic pain. Participants inhaled 4 different formulations of cannabis (containing 0%, 2.5%, 6.0%, and 9.4% THC) during 4 14-day periods. They inhaled a 25-mg dose through a pipe 3 times a day for the first 5 days of each cycle, followed by a 9-day washout period. Daily average pain intensity was measured using a numeric rating scale. The investigators also assessed mood, sleep, quality of life, and adverse effects.

Patients in the 9.4% THC group reported significantly less pain and better sleep, with average pain scores decreasing from 6.1 to 5.4 on an 11-point scale. Although the benefit was modest, the authors noted that the pain had been refractory to standard treatments.

The number of reported adverse events increased with greater potency and were most commonly throat irritation, burning sensation, headache, dizziness, and fatigue. This study suggests that THC potency affects tolerability, with higher doses eliciting clinically important adverse effects, some of which may reduce the ability to perform activities of daily living, such as driving.

Wilsey et al, 2013: Use in resistant neuropathic pain

Wilsey et al²⁶ conducted another double-blind, placebo-controlled crossover study assessing the effect of vaporized cannabis on central and peripheral neuropathic pain resistant to first-line pharmacotherapies. Dose-effect relationships were explored using medium-dose (3.5%), low-dose (1.3%), and placebo cannabis. The primary outcome measure was a 30% reduction in pain intensity based on a visual analog scale.

In the placebo group, 26% of patients achieved this vs 57% of the low-dose cannabis group and 61% of those receiving the medium dose. No significant difference was found between the 2 active doses in reducing neuropathic pain, and both were more effective than placebo. The number needed to treat to achieve a 30% reduction in pain was about 3 for both cannabis groups compared with placebo. Psychoactive effects were minimal, of short duration, and reversible.

Wallace et al, 2015: Use in diabetic peripheral neuropathy

Wallace et al²⁷ conducted a randomized, double-blind, placebo-controlled crossover study evaluating cannabis for diabetic peripheral neuropathy in 16 patients. Each had experienced at least 6 months of neuropathic pain in their feet. The participants inhaled a single dose of 1%, 4%, or 7% THC cannabis or placebo. Spontaneous pain was reported with a visual analog scale and also tested with a foam brush and von Frey filament at intervals until 4 hours after treatment.

Pain scores were lower with treatment compared with placebo, with high-dose cannabis having the greatest analgesic effect. Pain reduction lasted for the full duration of the test. Cannabis recipients had declines in attention and working memory, with the high-dose group experiencing the greatest impact 15 minutes after treatment. High-dose recipients also had poorer scores on testing of quick task-switching, with the greatest effect at 2 hours.²⁷

Research and market cannabis are not equal

Results of US studies must be qualified. Most have used cannabis provided by the National Institute of Drug Abuse (NIDA),^23–26 which differs in potency from commercially available preparations. This limits the clinical usefulness of the analysis of benefits and risks.

Vergara et al²⁸ found that NIDA varieties contained much lower THC levels and as much as 23 times the cannabinol content as cannabis in state-legalized markets.

Studies based on NIDA varieties likely underestimate the risks of consumer-purchased cannabis, as THC is believed to be most responsible for the risk of psychosis and impaired driving and cognition.^24,28

 

 

CBD MAY PROTECT AGAINST ADVERSE EFFECTS

Studies of CBD alone are limited to preclinical data.²⁹ Evidence suggests that CBD alone or combined with THC can suppress chronic neuropathic pain, and that CBD may have a protective effect after nerve injury.³⁰

Nabiximols, an oromucosal spray preparation with equal amounts of THC and CBD, has been approved in Canada as well as in European countries including the United Kingdom. Although its use has not been associated with many of the adverse effects of inhaled cannabis,^30–32 evidence of efficacy from clinical trials has been mixed.

Lynch et al,³¹ in a 2014 randomized, double-blind, placebo-controlled crossover pilot study³¹ evaluated nabiximols in 16 patients with neuropathic pain related to chemotherapy. No statistically significant difference was found between treatment and placebo. However, the trial was underpowered.

Serpell et al,³² in a 2014 European randomized, placebo-controlled parallel-group study, evaluated 246 patients with peripheral neuropathy with allodynia, with 128 receiving active treatment (THC-CBD oromucosal spray) and 118 receiving placebo. Over the 15-week study, participants continued their current analgesic treatments.

Pain was reduced in the treatment group, but the difference from placebo was not statistically significant. However, the treatment group reported significantly better sleep quality and Patient Global Impression of Change measures (reflecting a patient’s belief of treatment efficacy).

META-ANALYSES CONFIRM EFFECT

Three meta-analyses of available studies of the effects of cannabis on neuropathic pain have been completed.

Andreae et al, 2015: 5 trials, 178 patients

Andreae et al¹ evaluated 5 randomized controlled trials in 178 patients in North America. All had had neuropathy for at least 3 months, with a pain level of at least about 3 on a scale of 10. Two studies had patients with HIV-related neuropathy; the other 3 involved patients with neuropathy related to trauma, diabetes, complex regional pain syndrome, or spinal cord injury. All trials used whole cannabis plant provided by NIDA, and the main outcomes were patient-reported pain scales. No study evaluated pain beyond 2 weeks after trial termination.

They found that 1 of every 5 to 6 patients treated with cannabis had at least a 30% pain reduction.

Nugent et al, 2017: 13 trials, 246 patients

Nugent et al³³ reviewed 13 trials in 246 patients that evaluated the effects of different cannabis-based preparations on either central or peripheral neuropathic pain from various conditions. Actively treated patients were more likely to report a 30% improvement in neuropathic pain. Again, studies tended to be small and brief.

Cochrane review, 2018: 16 trials, 1,750 patients

A Cochrane review³⁴ analyzed 16 trials (in 1,750 patients) lasting 2 to 26 weeks. Treatments included an oromucosal spray with a plant-derived combination of THC and CBD, nabilone, inhaled herbal cannabis, and plant-derived THC.

With cannabis-based treatments, significantly more people achieved 50% or greater pain relief than with placebo (21% vs 17%, number needed to treat 20); 30% pain reduction was achieved in 39% of treated patients vs 33% of patients taking placebo (number needed to treat 11).

On the other hand, significantly more participants withdrew from studies because of adverse events with cannabis-based treatments than placebo (10% vs 5%), with psychiatric disorders occurring in 17% of patients receiving active treatment vs 5% of those receiving placebo (number needed to harm 10). 

The primary studies suffered from methodologic limitations including small size, short duration, and inconsistency of formulations and study designs. Further evaluation of long-term efficacy, tolerability, and addiction potential is critical to determine the risk-benefit ratio.

RISKS OF CANNABIS USE

Like any drug therapy, cannabis has effects that may limit its use. Cannabis can affect a person’s psyche, physiology, and lifestyle.

Impaired attention, task speed

Neurocognitive changes associated with cannabis use—especially dizziness, fatigue, and slowed task-switching—could affect driving and other complex tasks. Evidence indicates that such activities should be avoided in the hours after treatment.^26,27,32,33

Concern over brain development

Most worrisome is the effect of long-term cannabis use on brain development in young adults. Regular use of cannabis at an early age is associated with lower IQ, decline in school performance, and lower rates of high school graduation.³⁵

Avoid in psychiatric patients

It is unlikely that cannabis can be safely used in patients with psychiatric illnesses. Anxiety, depression, and psychotic disorders can be exacerbated by the regular use of cannabis, and the risk of developing these conditions is increased while using cannabis.^36,37

High concentrations of THC (the highest concentration used in the above studies was 9.5%) can cause anxiety, paranoia, and psychosis.

Respiratory effects

Long-term cannabis smoking may cause wheezing, cough, dyspnea, and exacerbations of chronic bronchitis. There is some evidence that symptoms improve after stopping smoking.^33,38

SHOULD WE RECOMMEND CANNABIS?

Where cannabis can be legally used, doctors should be familiar with the literature and its limitations so that they can counsel patients on the best use and potential risks and benefits of cannabis treatment.

A recent conceptualization of pain suggests that a pain score reflects a composite of sensory factors (eg, tissue damage), cognitive factors (eg, beliefs about pain), and affective factors (eg, anxiety, depression).³⁹ Physicians should keep this in mind when evaluating patients to better assess the risks and benefits of cannabis. While pharmacotherapy may address sensory factors, cognitive behavioral therapy may help alter beliefs about the pain as well as anxiety and depressive symptoms that might influence subjective reports.

Ideally, patients being considered for cannabis treatment would have a type of neuropathic pain proven to respond to cannabis in randomized, controlled studies, as well as evidence of failed first-line treatments.

Relative contraindications include depression, anxiety, substance use, psychotic disorders, and respiratory conditions, and these should also be considered.

Although current research shows an analgesic benefit of cannabis on neuropathic pain comparable to that of gabapentin,⁴⁰ further investigation is needed to better evaluate long-term safety, efficacy, and interactions with standard therapies. Until we have a more complete picture, we should use the current literature, along with a thorough knowledge of each patient, to determine if the benefits of cannabis therapy outweigh the risks.

Acknowledgments: We thank Camillo Ferrari, BS, and Christina McMahon, BA, for their helpful comments.

Marijuana, which is still illegal under federal law but legal in 30 states for medical purposes as of this writing, has shown promising results for treating peripheral neuropathy. Studies suggest that cannabis may be an option for patients whose pain responds poorly to standard treatments; however, its use may be restricted by cognitive and psychiatric adverse effects, particularly at high doses.¹

See related editorial

In this article, we discuss the basic pharmacology of cannabis and how it may affect neuropathic pain. We review clinical trials on its use for peripheral neuropathy and provide guidance for its use.

PERIPHERAL NEUROPATHY IS COMMON AND COMPLEX

An estimated 20 million people in the United States suffer from neuropathic pain. The prevalence is higher in certain populations, with 26% of people over age 65 and 30% of patients with diabetes mellitus affected.^2–4

Peripheral neuropathy is a complex, chronic state that occurs when nerve fibers are damaged, dysfunctional, or injured, sending incorrect signals to pain centers in the central nervous system.⁵ It is characterized by weakness, pain, and paresthesias that typically begin in the hands or feet and progress proximally.⁴ Symptoms depend on the number and types of nerves affected.

In many cases, peripheral neuropathy is idiopathic, but common causes include diabetes, alcoholism, human immunodeficiency virus (HIV) infection, and autoimmune disease. Others include toxicity from chemotherapy and heavy metals.

Peripheral neuropathy significantly worsens quality of life and function. Many patients experience emotional, cognitive, and functional problems, resulting in high rates of medical and psychiatric comorbidities and occupational impairment.^4,6,7 Yet despite its clinical and epidemiologic significance, it is often undertreated.⁸

STANDARD TREATMENTS INADEQUATE

Peripheral neuropathy occurs in patients with a wide range of comorbidities and is especially difficult to treat. Mainstays of therapy include anticonvulsants, tricyclic antidepressants, and serotonin-norepinephrine reuptake inhibitors.⁹ A more invasive option is spinal cord stimulation.

These treatments can have considerable adverse effects, and response rates remain suboptimal, with pain relief insufficient to improve quality of life for many patients.^9,10 Better treatments are needed to improve clinical outcomes and patient experience.¹¹

CANNABIS: A MIX OF COMPOUNDS

Cannabis sativa has been used as an analgesic for centuries. The plant contains more than 400 chemical compounds and is often used for its euphoric properties. Long-term use may lead to addiction and cognitive impairment.^12,13

Tetrahydrocannabinol (THC) and cannabidiol (CBD) are the main components and the 2 best-studied cannabinoids with analgesic effects.

THC is the primary psychoactive component of cannabis. Its effects include relaxation, altered perception, heightened sensations, increased libido, and perceptual distortions of time and space. Temporary effects may include decreased short-term memory, dry mouth, impaired motor function, conjunctival injection, paranoia, and anxiety.

CBD is nonpsychoactive and has anti-inflammatory and antioxidant properties. It has been shown to reduce pain and inflammation without the effects of THC.¹⁴

Other compounds in the cannabis plant include phytocannabinoids, flavonoids, and tapenoids, which may produce individual, interactive, or synergistic effects.¹⁵ Different strains of cannabis have varying amounts of the individual components, making comparisons among clinical studies difficult.

THE ENDOCANNABINOID SYSTEM

The endogenous mammalian cannabinoid system plays a regulatory role in the development, homeostasis, and neuroplasticity of the central nervous system. It is also involved in modulating pain transmission in the nociceptive pathway.

Two of the most abundant cannabinoid endogenous ligands are anandamide and 2-arachidonylglycerol.⁹ These endocanna­b­inoids are produced on demand in the central nervous system to reduce pain by acting as a circuit breaker.^16–18 They target the G protein-coupled cannabinoid receptors CB1 and CB2, located throughout the central and peripheral nervous system and in organs and tissues.¹²

CB1 receptors are found primarily in the central nervous system, specifically in areas involved in movement, such as the basal ganglia and cerebellum, as well as in areas involved in memory, such as the hippocampus.¹² They are also abundant in brain regions implicated in conducting and modulating pain signals, including the periaqueductal gray and the dorsal horn of the spinal cord.^16–20

CB2 receptors are mostly found in peripheral tissues and organs, mainly those involved in the immune system, including splenic, tonsillar, and hematopoietic cells.¹² They help regulate inflammation, allodynia, and hyperalgesia.¹⁷

Modifying response to injury

Following a nerve injury, neurons along the nociceptive pathway may become more reactive and responsive in a process known as sensitization.²¹ The process involves a cascade of cellular events that result in sprouting of pain-sensitive nerve endings.^21,22

Cannabinoids are thought to reduce pain by modifying these cellular events. They also inhibit nociceptive conduction in the dorsal horn of the spinal cord and in the ascending spinothalamic tract.²⁰ CB1 receptors found in nociceptive terminals along the peripheral nervous system impede pain conduction, while activation of CB2 receptors in immune cells decreases the release of nociceptive agents.

 

 

STUDIES OF CANNABIS FOR NEUROPATHIC PAIN

A number of studies have evaluated cannabis for treating neuropathic pain. Overall, available data support the efficacy of smoked or inhaled cannabis in its flower form when used as monotherapy or adjunctive therapy for relief of neuropathic pain of various etiologies. Many studies also report secondary benefits, including better sleep and functional improvement.^23,24

However, adverse effects are common, especially at high doses, and include difficulty concentrating, lightheadedness, fatigue, and tachycardia. More serious reported adverse effects include anxiety, paranoia, and psychosis.

Wilsey et al, 2008: Neuropathic pain reduced

Wilsey et al²⁵ conducted a double-blind, placebo-controlled crossover study that assessed the effects of smoking cannabis in 38 patients with central or peripheral neuropathic pain. Participants were assigned to smoke either high- or low-dose cannabis (7% or 3.5% delta-9-THC) or placebo cigarettes. Cigarettes were smoked during treatment sessions using the following regimen: 2 puffs at 60 minutes from baseline, 3 puffs at 120 minutes, and 4 puffs at 180 minutes. Patients were assessed after each set of puffs and for 2 hours afterwards. The primary outcome was spontaneous relief of pain as measured by a visual analog scale.

Pain intensity was comparable and significantly reduced in both treatment groups compared with placebo. At the high dose, some participants experienced neurocognitive impairment in attention, learning, memory, and psychomotor speed; only learning and memory declined at the low dose.

Ellis et al, 2009: Pain reduction in HIV neuropathy

Ellis et al²³ conducted a double-blind, placebo-controlled crossover trial in patients with HIV neuropathy that was unresponsive to at least 2 analgesics with different modes of action. During each treatment week, participants were randomly assigned to smoke either active cannabis or placebo, while continuing their standard therapy. Titration started at 4% THC and was adjusted based on tolerability and efficacy. Twenty-eight of the 34 enrolled patients completed both cannabis and placebo treatments. The principal outcome was change in pain intensity from baseline at the end of each week, using the Descriptor Differential Scale of Pain Intensity.

Of the 28 patients, 46% achieved an average pain reduction of 3.3 points (30%). One patient experienced cannabis-induced psychosis, and another developed an intractable cough, which resolved with smoking cessation.

Ware et al, 2010: Reduced posttraumatic or postsurgical neuropathic pain

Ware et al²⁴ performed a randomized crossover trial in 21 patients with posttraumatic or postsurgical neuropathic pain. Participants inhaled 4 different formulations of cannabis (containing 0%, 2.5%, 6.0%, and 9.4% THC) during 4 14-day periods. They inhaled a 25-mg dose through a pipe 3 times a day for the first 5 days of each cycle, followed by a 9-day washout period. Daily average pain intensity was measured using a numeric rating scale. The investigators also assessed mood, sleep, quality of life, and adverse effects.

Patients in the 9.4% THC group reported significantly less pain and better sleep, with average pain scores decreasing from 6.1 to 5.4 on an 11-point scale. Although the benefit was modest, the authors noted that the pain had been refractory to standard treatments.

The number of reported adverse events increased with greater potency and were most commonly throat irritation, burning sensation, headache, dizziness, and fatigue. This study suggests that THC potency affects tolerability, with higher doses eliciting clinically important adverse effects, some of which may reduce the ability to perform activities of daily living, such as driving.

Wilsey et al, 2013: Use in resistant neuropathic pain

Wilsey et al²⁶ conducted another double-blind, placebo-controlled crossover study assessing the effect of vaporized cannabis on central and peripheral neuropathic pain resistant to first-line pharmacotherapies. Dose-effect relationships were explored using medium-dose (3.5%), low-dose (1.3%), and placebo cannabis. The primary outcome measure was a 30% reduction in pain intensity based on a visual analog scale.

In the placebo group, 26% of patients achieved this vs 57% of the low-dose cannabis group and 61% of those receiving the medium dose. No significant difference was found between the 2 active doses in reducing neuropathic pain, and both were more effective than placebo. The number needed to treat to achieve a 30% reduction in pain was about 3 for both cannabis groups compared with placebo. Psychoactive effects were minimal, of short duration, and reversible.

Wallace et al, 2015: Use in diabetic peripheral neuropathy

Wallace et al²⁷ conducted a randomized, double-blind, placebo-controlled crossover study evaluating cannabis for diabetic peripheral neuropathy in 16 patients. Each had experienced at least 6 months of neuropathic pain in their feet. The participants inhaled a single dose of 1%, 4%, or 7% THC cannabis or placebo. Spontaneous pain was reported with a visual analog scale and also tested with a foam brush and von Frey filament at intervals until 4 hours after treatment.

Pain scores were lower with treatment compared with placebo, with high-dose cannabis having the greatest analgesic effect. Pain reduction lasted for the full duration of the test. Cannabis recipients had declines in attention and working memory, with the high-dose group experiencing the greatest impact 15 minutes after treatment. High-dose recipients also had poorer scores on testing of quick task-switching, with the greatest effect at 2 hours.²⁷

Research and market cannabis are not equal

Results of US studies must be qualified. Most have used cannabis provided by the National Institute of Drug Abuse (NIDA),^23–26 which differs in potency from commercially available preparations. This limits the clinical usefulness of the analysis of benefits and risks.

Vergara et al²⁸ found that NIDA varieties contained much lower THC levels and as much as 23 times the cannabinol content as cannabis in state-legalized markets.

Studies based on NIDA varieties likely underestimate the risks of consumer-purchased cannabis, as THC is believed to be most responsible for the risk of psychosis and impaired driving and cognition.^24,28

 

 

CBD MAY PROTECT AGAINST ADVERSE EFFECTS

Studies of CBD alone are limited to preclinical data.²⁹ Evidence suggests that CBD alone or combined with THC can suppress chronic neuropathic pain, and that CBD may have a protective effect after nerve injury.³⁰

Nabiximols, an oromucosal spray preparation with equal amounts of THC and CBD, has been approved in Canada as well as in European countries including the United Kingdom. Although its use has not been associated with many of the adverse effects of inhaled cannabis,^30–32 evidence of efficacy from clinical trials has been mixed.

Lynch et al,³¹ in a 2014 randomized, double-blind, placebo-controlled crossover pilot study³¹ evaluated nabiximols in 16 patients with neuropathic pain related to chemotherapy. No statistically significant difference was found between treatment and placebo. However, the trial was underpowered.

Serpell et al,³² in a 2014 European randomized, placebo-controlled parallel-group study, evaluated 246 patients with peripheral neuropathy with allodynia, with 128 receiving active treatment (THC-CBD oromucosal spray) and 118 receiving placebo. Over the 15-week study, participants continued their current analgesic treatments.

Pain was reduced in the treatment group, but the difference from placebo was not statistically significant. However, the treatment group reported significantly better sleep quality and Patient Global Impression of Change measures (reflecting a patient’s belief of treatment efficacy).

META-ANALYSES CONFIRM EFFECT

Three meta-analyses of available studies of the effects of cannabis on neuropathic pain have been completed.

Andreae et al, 2015: 5 trials, 178 patients

Andreae et al¹ evaluated 5 randomized controlled trials in 178 patients in North America. All had had neuropathy for at least 3 months, with a pain level of at least about 3 on a scale of 10. Two studies had patients with HIV-related neuropathy; the other 3 involved patients with neuropathy related to trauma, diabetes, complex regional pain syndrome, or spinal cord injury. All trials used whole cannabis plant provided by NIDA, and the main outcomes were patient-reported pain scales. No study evaluated pain beyond 2 weeks after trial termination.

They found that 1 of every 5 to 6 patients treated with cannabis had at least a 30% pain reduction.

Nugent et al, 2017: 13 trials, 246 patients

Nugent et al³³ reviewed 13 trials in 246 patients that evaluated the effects of different cannabis-based preparations on either central or peripheral neuropathic pain from various conditions. Actively treated patients were more likely to report a 30% improvement in neuropathic pain. Again, studies tended to be small and brief.

Cochrane review, 2018: 16 trials, 1,750 patients

A Cochrane review³⁴ analyzed 16 trials (in 1,750 patients) lasting 2 to 26 weeks. Treatments included an oromucosal spray with a plant-derived combination of THC and CBD, nabilone, inhaled herbal cannabis, and plant-derived THC.

With cannabis-based treatments, significantly more people achieved 50% or greater pain relief than with placebo (21% vs 17%, number needed to treat 20); 30% pain reduction was achieved in 39% of treated patients vs 33% of patients taking placebo (number needed to treat 11).

On the other hand, significantly more participants withdrew from studies because of adverse events with cannabis-based treatments than placebo (10% vs 5%), with psychiatric disorders occurring in 17% of patients receiving active treatment vs 5% of those receiving placebo (number needed to harm 10). 

The primary studies suffered from methodologic limitations including small size, short duration, and inconsistency of formulations and study designs. Further evaluation of long-term efficacy, tolerability, and addiction potential is critical to determine the risk-benefit ratio.

RISKS OF CANNABIS USE

Like any drug therapy, cannabis has effects that may limit its use. Cannabis can affect a person’s psyche, physiology, and lifestyle.

Impaired attention, task speed

Neurocognitive changes associated with cannabis use—especially dizziness, fatigue, and slowed task-switching—could affect driving and other complex tasks. Evidence indicates that such activities should be avoided in the hours after treatment.^26,27,32,33

Concern over brain development

Most worrisome is the effect of long-term cannabis use on brain development in young adults. Regular use of cannabis at an early age is associated with lower IQ, decline in school performance, and lower rates of high school graduation.³⁵

Avoid in psychiatric patients

It is unlikely that cannabis can be safely used in patients with psychiatric illnesses. Anxiety, depression, and psychotic disorders can be exacerbated by the regular use of cannabis, and the risk of developing these conditions is increased while using cannabis.^36,37

High concentrations of THC (the highest concentration used in the above studies was 9.5%) can cause anxiety, paranoia, and psychosis.

Respiratory effects

Long-term cannabis smoking may cause wheezing, cough, dyspnea, and exacerbations of chronic bronchitis. There is some evidence that symptoms improve after stopping smoking.^33,38

SHOULD WE RECOMMEND CANNABIS?

Where cannabis can be legally used, doctors should be familiar with the literature and its limitations so that they can counsel patients on the best use and potential risks and benefits of cannabis treatment.

A recent conceptualization of pain suggests that a pain score reflects a composite of sensory factors (eg, tissue damage), cognitive factors (eg, beliefs about pain), and affective factors (eg, anxiety, depression).³⁹ Physicians should keep this in mind when evaluating patients to better assess the risks and benefits of cannabis. While pharmacotherapy may address sensory factors, cognitive behavioral therapy may help alter beliefs about the pain as well as anxiety and depressive symptoms that might influence subjective reports.

Ideally, patients being considered for cannabis treatment would have a type of neuropathic pain proven to respond to cannabis in randomized, controlled studies, as well as evidence of failed first-line treatments.

Relative contraindications include depression, anxiety, substance use, psychotic disorders, and respiratory conditions, and these should also be considered.

Although current research shows an analgesic benefit of cannabis on neuropathic pain comparable to that of gabapentin,⁴⁰ further investigation is needed to better evaluate long-term safety, efficacy, and interactions with standard therapies. Until we have a more complete picture, we should use the current literature, along with a thorough knowledge of each patient, to determine if the benefits of cannabis therapy outweigh the risks.

Acknowledgments: We thank Camillo Ferrari, BS, and Christina McMahon, BA, for their helpful comments.

Marijuana, which is still illegal under federal law but legal in 30 states for medical purposes as of this writing, has shown promising results for treating peripheral neuropathy. Studies suggest that cannabis may be an option for patients whose pain responds poorly to standard treatments; however, its use may be restricted by cognitive and psychiatric adverse effects, particularly at high doses.¹

See related editorial

In this article, we discuss the basic pharmacology of cannabis and how it may affect neuropathic pain. We review clinical trials on its use for peripheral neuropathy and provide guidance for its use.

PERIPHERAL NEUROPATHY IS COMMON AND COMPLEX

An estimated 20 million people in the United States suffer from neuropathic pain. The prevalence is higher in certain populations, with 26% of people over age 65 and 30% of patients with diabetes mellitus affected.^2–4

Peripheral neuropathy is a complex, chronic state that occurs when nerve fibers are damaged, dysfunctional, or injured, sending incorrect signals to pain centers in the central nervous system.⁵ It is characterized by weakness, pain, and paresthesias that typically begin in the hands or feet and progress proximally.⁴ Symptoms depend on the number and types of nerves affected.

In many cases, peripheral neuropathy is idiopathic, but common causes include diabetes, alcoholism, human immunodeficiency virus (HIV) infection, and autoimmune disease. Others include toxicity from chemotherapy and heavy metals.

Peripheral neuropathy significantly worsens quality of life and function. Many patients experience emotional, cognitive, and functional problems, resulting in high rates of medical and psychiatric comorbidities and occupational impairment.^4,6,7 Yet despite its clinical and epidemiologic significance, it is often undertreated.⁸

STANDARD TREATMENTS INADEQUATE

Peripheral neuropathy occurs in patients with a wide range of comorbidities and is especially difficult to treat. Mainstays of therapy include anticonvulsants, tricyclic antidepressants, and serotonin-norepinephrine reuptake inhibitors.⁹ A more invasive option is spinal cord stimulation.

These treatments can have considerable adverse effects, and response rates remain suboptimal, with pain relief insufficient to improve quality of life for many patients.^9,10 Better treatments are needed to improve clinical outcomes and patient experience.¹¹

CANNABIS: A MIX OF COMPOUNDS

Cannabis sativa has been used as an analgesic for centuries. The plant contains more than 400 chemical compounds and is often used for its euphoric properties. Long-term use may lead to addiction and cognitive impairment.^12,13

Tetrahydrocannabinol (THC) and cannabidiol (CBD) are the main components and the 2 best-studied cannabinoids with analgesic effects.

THC is the primary psychoactive component of cannabis. Its effects include relaxation, altered perception, heightened sensations, increased libido, and perceptual distortions of time and space. Temporary effects may include decreased short-term memory, dry mouth, impaired motor function, conjunctival injection, paranoia, and anxiety.

CBD is nonpsychoactive and has anti-inflammatory and antioxidant properties. It has been shown to reduce pain and inflammation without the effects of THC.¹⁴

Other compounds in the cannabis plant include phytocannabinoids, flavonoids, and tapenoids, which may produce individual, interactive, or synergistic effects.¹⁵ Different strains of cannabis have varying amounts of the individual components, making comparisons among clinical studies difficult.

THE ENDOCANNABINOID SYSTEM

The endogenous mammalian cannabinoid system plays a regulatory role in the development, homeostasis, and neuroplasticity of the central nervous system. It is also involved in modulating pain transmission in the nociceptive pathway.

Two of the most abundant cannabinoid endogenous ligands are anandamide and 2-arachidonylglycerol.⁹ These endocanna­b­inoids are produced on demand in the central nervous system to reduce pain by acting as a circuit breaker.^16–18 They target the G protein-coupled cannabinoid receptors CB1 and CB2, located throughout the central and peripheral nervous system and in organs and tissues.¹²

CB1 receptors are found primarily in the central nervous system, specifically in areas involved in movement, such as the basal ganglia and cerebellum, as well as in areas involved in memory, such as the hippocampus.¹² They are also abundant in brain regions implicated in conducting and modulating pain signals, including the periaqueductal gray and the dorsal horn of the spinal cord.^16–20

CB2 receptors are mostly found in peripheral tissues and organs, mainly those involved in the immune system, including splenic, tonsillar, and hematopoietic cells.¹² They help regulate inflammation, allodynia, and hyperalgesia.¹⁷

Modifying response to injury

Following a nerve injury, neurons along the nociceptive pathway may become more reactive and responsive in a process known as sensitization.²¹ The process involves a cascade of cellular events that result in sprouting of pain-sensitive nerve endings.^21,22

Cannabinoids are thought to reduce pain by modifying these cellular events. They also inhibit nociceptive conduction in the dorsal horn of the spinal cord and in the ascending spinothalamic tract.²⁰ CB1 receptors found in nociceptive terminals along the peripheral nervous system impede pain conduction, while activation of CB2 receptors in immune cells decreases the release of nociceptive agents.

 

 

STUDIES OF CANNABIS FOR NEUROPATHIC PAIN

A number of studies have evaluated cannabis for treating neuropathic pain. Overall, available data support the efficacy of smoked or inhaled cannabis in its flower form when used as monotherapy or adjunctive therapy for relief of neuropathic pain of various etiologies. Many studies also report secondary benefits, including better sleep and functional improvement.^23,24

However, adverse effects are common, especially at high doses, and include difficulty concentrating, lightheadedness, fatigue, and tachycardia. More serious reported adverse effects include anxiety, paranoia, and psychosis.

Wilsey et al, 2008: Neuropathic pain reduced

Wilsey et al²⁵ conducted a double-blind, placebo-controlled crossover study that assessed the effects of smoking cannabis in 38 patients with central or peripheral neuropathic pain. Participants were assigned to smoke either high- or low-dose cannabis (7% or 3.5% delta-9-THC) or placebo cigarettes. Cigarettes were smoked during treatment sessions using the following regimen: 2 puffs at 60 minutes from baseline, 3 puffs at 120 minutes, and 4 puffs at 180 minutes. Patients were assessed after each set of puffs and for 2 hours afterwards. The primary outcome was spontaneous relief of pain as measured by a visual analog scale.

Pain intensity was comparable and significantly reduced in both treatment groups compared with placebo. At the high dose, some participants experienced neurocognitive impairment in attention, learning, memory, and psychomotor speed; only learning and memory declined at the low dose.

Ellis et al, 2009: Pain reduction in HIV neuropathy

Ellis et al²³ conducted a double-blind, placebo-controlled crossover trial in patients with HIV neuropathy that was unresponsive to at least 2 analgesics with different modes of action. During each treatment week, participants were randomly assigned to smoke either active cannabis or placebo, while continuing their standard therapy. Titration started at 4% THC and was adjusted based on tolerability and efficacy. Twenty-eight of the 34 enrolled patients completed both cannabis and placebo treatments. The principal outcome was change in pain intensity from baseline at the end of each week, using the Descriptor Differential Scale of Pain Intensity.

Of the 28 patients, 46% achieved an average pain reduction of 3.3 points (30%). One patient experienced cannabis-induced psychosis, and another developed an intractable cough, which resolved with smoking cessation.

Ware et al, 2010: Reduced posttraumatic or postsurgical neuropathic pain

Ware et al²⁴ performed a randomized crossover trial in 21 patients with posttraumatic or postsurgical neuropathic pain. Participants inhaled 4 different formulations of cannabis (containing 0%, 2.5%, 6.0%, and 9.4% THC) during 4 14-day periods. They inhaled a 25-mg dose through a pipe 3 times a day for the first 5 days of each cycle, followed by a 9-day washout period. Daily average pain intensity was measured using a numeric rating scale. The investigators also assessed mood, sleep, quality of life, and adverse effects.

Patients in the 9.4% THC group reported significantly less pain and better sleep, with average pain scores decreasing from 6.1 to 5.4 on an 11-point scale. Although the benefit was modest, the authors noted that the pain had been refractory to standard treatments.

The number of reported adverse events increased with greater potency and were most commonly throat irritation, burning sensation, headache, dizziness, and fatigue. This study suggests that THC potency affects tolerability, with higher doses eliciting clinically important adverse effects, some of which may reduce the ability to perform activities of daily living, such as driving.

Wilsey et al, 2013: Use in resistant neuropathic pain

Wilsey et al²⁶ conducted another double-blind, placebo-controlled crossover study assessing the effect of vaporized cannabis on central and peripheral neuropathic pain resistant to first-line pharmacotherapies. Dose-effect relationships were explored using medium-dose (3.5%), low-dose (1.3%), and placebo cannabis. The primary outcome measure was a 30% reduction in pain intensity based on a visual analog scale.

In the placebo group, 26% of patients achieved this vs 57% of the low-dose cannabis group and 61% of those receiving the medium dose. No significant difference was found between the 2 active doses in reducing neuropathic pain, and both were more effective than placebo. The number needed to treat to achieve a 30% reduction in pain was about 3 for both cannabis groups compared with placebo. Psychoactive effects were minimal, of short duration, and reversible.

Wallace et al, 2015: Use in diabetic peripheral neuropathy

Wallace et al²⁷ conducted a randomized, double-blind, placebo-controlled crossover study evaluating cannabis for diabetic peripheral neuropathy in 16 patients. Each had experienced at least 6 months of neuropathic pain in their feet. The participants inhaled a single dose of 1%, 4%, or 7% THC cannabis or placebo. Spontaneous pain was reported with a visual analog scale and also tested with a foam brush and von Frey filament at intervals until 4 hours after treatment.

Pain scores were lower with treatment compared with placebo, with high-dose cannabis having the greatest analgesic effect. Pain reduction lasted for the full duration of the test. Cannabis recipients had declines in attention and working memory, with the high-dose group experiencing the greatest impact 15 minutes after treatment. High-dose recipients also had poorer scores on testing of quick task-switching, with the greatest effect at 2 hours.²⁷

Research and market cannabis are not equal

Results of US studies must be qualified. Most have used cannabis provided by the National Institute of Drug Abuse (NIDA),^23–26 which differs in potency from commercially available preparations. This limits the clinical usefulness of the analysis of benefits and risks.

Vergara et al²⁸ found that NIDA varieties contained much lower THC levels and as much as 23 times the cannabinol content as cannabis in state-legalized markets.

Studies based on NIDA varieties likely underestimate the risks of consumer-purchased cannabis, as THC is believed to be most responsible for the risk of psychosis and impaired driving and cognition.^24,28

 

 

CBD MAY PROTECT AGAINST ADVERSE EFFECTS

Studies of CBD alone are limited to preclinical data.²⁹ Evidence suggests that CBD alone or combined with THC can suppress chronic neuropathic pain, and that CBD may have a protective effect after nerve injury.³⁰

Nabiximols, an oromucosal spray preparation with equal amounts of THC and CBD, has been approved in Canada as well as in European countries including the United Kingdom. Although its use has not been associated with many of the adverse effects of inhaled cannabis,^30–32 evidence of efficacy from clinical trials has been mixed.

Lynch et al,³¹ in a 2014 randomized, double-blind, placebo-controlled crossover pilot study³¹ evaluated nabiximols in 16 patients with neuropathic pain related to chemotherapy. No statistically significant difference was found between treatment and placebo. However, the trial was underpowered.

Serpell et al,³² in a 2014 European randomized, placebo-controlled parallel-group study, evaluated 246 patients with peripheral neuropathy with allodynia, with 128 receiving active treatment (THC-CBD oromucosal spray) and 118 receiving placebo. Over the 15-week study, participants continued their current analgesic treatments.

Pain was reduced in the treatment group, but the difference from placebo was not statistically significant. However, the treatment group reported significantly better sleep quality and Patient Global Impression of Change measures (reflecting a patient’s belief of treatment efficacy).

META-ANALYSES CONFIRM EFFECT

Three meta-analyses of available studies of the effects of cannabis on neuropathic pain have been completed.

Andreae et al, 2015: 5 trials, 178 patients

Andreae et al¹ evaluated 5 randomized controlled trials in 178 patients in North America. All had had neuropathy for at least 3 months, with a pain level of at least about 3 on a scale of 10. Two studies had patients with HIV-related neuropathy; the other 3 involved patients with neuropathy related to trauma, diabetes, complex regional pain syndrome, or spinal cord injury. All trials used whole cannabis plant provided by NIDA, and the main outcomes were patient-reported pain scales. No study evaluated pain beyond 2 weeks after trial termination.

They found that 1 of every 5 to 6 patients treated with cannabis had at least a 30% pain reduction.

Nugent et al, 2017: 13 trials, 246 patients

Nugent et al³³ reviewed 13 trials in 246 patients that evaluated the effects of different cannabis-based preparations on either central or peripheral neuropathic pain from various conditions. Actively treated patients were more likely to report a 30% improvement in neuropathic pain. Again, studies tended to be small and brief.

Cochrane review, 2018: 16 trials, 1,750 patients

A Cochrane review³⁴ analyzed 16 trials (in 1,750 patients) lasting 2 to 26 weeks. Treatments included an oromucosal spray with a plant-derived combination of THC and CBD, nabilone, inhaled herbal cannabis, and plant-derived THC.

With cannabis-based treatments, significantly more people achieved 50% or greater pain relief than with placebo (21% vs 17%, number needed to treat 20); 30% pain reduction was achieved in 39% of treated patients vs 33% of patients taking placebo (number needed to treat 11).

On the other hand, significantly more participants withdrew from studies because of adverse events with cannabis-based treatments than placebo (10% vs 5%), with psychiatric disorders occurring in 17% of patients receiving active treatment vs 5% of those receiving placebo (number needed to harm 10). 

The primary studies suffered from methodologic limitations including small size, short duration, and inconsistency of formulations and study designs. Further evaluation of long-term efficacy, tolerability, and addiction potential is critical to determine the risk-benefit ratio.

RISKS OF CANNABIS USE

Like any drug therapy, cannabis has effects that may limit its use. Cannabis can affect a person’s psyche, physiology, and lifestyle.

Impaired attention, task speed

Neurocognitive changes associated with cannabis use—especially dizziness, fatigue, and slowed task-switching—could affect driving and other complex tasks. Evidence indicates that such activities should be avoided in the hours after treatment.^26,27,32,33

Concern over brain development

Most worrisome is the effect of long-term cannabis use on brain development in young adults. Regular use of cannabis at an early age is associated with lower IQ, decline in school performance, and lower rates of high school graduation.³⁵

Avoid in psychiatric patients

It is unlikely that cannabis can be safely used in patients with psychiatric illnesses. Anxiety, depression, and psychotic disorders can be exacerbated by the regular use of cannabis, and the risk of developing these conditions is increased while using cannabis.^36,37

High concentrations of THC (the highest concentration used in the above studies was 9.5%) can cause anxiety, paranoia, and psychosis.

Respiratory effects

Long-term cannabis smoking may cause wheezing, cough, dyspnea, and exacerbations of chronic bronchitis. There is some evidence that symptoms improve after stopping smoking.^33,38

SHOULD WE RECOMMEND CANNABIS?

Where cannabis can be legally used, doctors should be familiar with the literature and its limitations so that they can counsel patients on the best use and potential risks and benefits of cannabis treatment.

A recent conceptualization of pain suggests that a pain score reflects a composite of sensory factors (eg, tissue damage), cognitive factors (eg, beliefs about pain), and affective factors (eg, anxiety, depression).³⁹ Physicians should keep this in mind when evaluating patients to better assess the risks and benefits of cannabis. While pharmacotherapy may address sensory factors, cognitive behavioral therapy may help alter beliefs about the pain as well as anxiety and depressive symptoms that might influence subjective reports.

Ideally, patients being considered for cannabis treatment would have a type of neuropathic pain proven to respond to cannabis in randomized, controlled studies, as well as evidence of failed first-line treatments.

Relative contraindications include depression, anxiety, substance use, psychotic disorders, and respiratory conditions, and these should also be considered.

Although current research shows an analgesic benefit of cannabis on neuropathic pain comparable to that of gabapentin,⁴⁰ further investigation is needed to better evaluate long-term safety, efficacy, and interactions with standard therapies. Until we have a more complete picture, we should use the current literature, along with a thorough knowledge of each patient, to determine if the benefits of cannabis therapy outweigh the risks.

Acknowledgments: We thank Camillo Ferrari, BS, and Christina McMahon, BA, for their helpful comments.

Unfortunately, recent research has not unveiled a breakthrough for preventing or treating cognitive impairment or Alzheimer disease. But several studies from the last 2 years are helping to drive the field of geriatrics forward, providing evidence of what does and does not help a variety of issues specific to the elderly. 

Based on a search of the 2017 and 2018 literature, this article presents new evidence on preventing and treating cognitive impairment, managing dementia-associated behavioral disturbances and delirium, preventing falls, and improving inpatient mobility and posthospital care transitions.

COGNITIVE IMPAIRMENT, DEMENTIA: STILL NO SILVER BULLET

With the exception of oral anticoagulation treatment for atrial fibrillation, there is little evidence that pharmacologic or nonpharmacologic interventions slow the onset or progression of Alzheimer disease.

Nonpharmacologic interventions

Home occupational therapy. A 2-year home-based occupational therapy intervention¹ showed no evidence of slowing functional decline in patients with Alzheimer disease. The randomized controlled trial involving 180 participants consisted of monthly sessions of an intensive, well-established collaborative-care management model that included fall prevention and other safety strategies, personalized training in activities of daily living, exercise, and education. Outcome measures for activities of daily living did not differ significantly between the treatment and control groups.¹

Physical activity. Whether physical activity interventions slow cognitive decline and prevent dementia in cognitively intact adults was examined in a systematic review of 32 trials.² Most of the trials followed patients for 6 months; a few stretched for 1 or 2 years.

Evidence was insufficient to prove cognitive benefit for short-term, single-component or multicomponent physical activity interventions. However, a multidomain physical activity intervention that also included dietary modifications and cognitive training did show a delay in cognitive decline, but only “low-strength” evidence.²

Nutritional supplements. The antioxidants vitamin E and selenium were studied for their possible cognitive benefit in the double-blind randomized Prevention of Alzheimer Disease by Vitamin E and Selenium trial³ in 3,786 asymptomatic men ages 60 and older. Neither supplement was found to prevent dementia over a 7-year follow-up period.

A review of 38 trials⁴ evaluated the effects on cognition of omega-3 fatty acids, soy, ginkgo biloba, B vitamins, vitamin D plus calcium, vitamin C, beta-carotene, and multi-ingredient supplements. It found insufficient evidence to recommend any over-the-counter supplement for cognitive protection in adults with normal cognition or mild cognitive impairment.

Pharmacologic treatments

Testosterone supplementation. The Testosterone Trials tested the effects of testosterone gel vs placebo for 1 year on 493 men over age 65 with low testosterone (< 275 ng/mL) and with subjective memory complaints and objective memory performance deficits. Treatment was not associated with improved memory or other cognitive functions compared with placebo.⁵

Antiamyloid drugs. A randomized, double-blind, placebo-controlled trial in nearly 2,000 patients evaluated verubecestat, an oral beta-site amyloid precursor protein-cleaving enzyme-1 inhibitor that reduces the amyloid-beta level in cerebrospinal fluid.⁶

Verubecestat did not reduce cognitive or functional decline in patients with mild-to-moderate Alzheimer disease, while adverse events including rashes, falls, injuries, sleep disturbances, suicidal ideation, weight loss, and hair color change were more common in the treatment groups. The trial was terminated early because of futility at 50 months.

And in a placebo-controlled trial of solanezumab, a monoclonal antibody directed against the amyloid beta peptide, no benefit was demonstrated at 80 weeks in more than 2,000 patients with Alzheimer disease.⁷

Multiple common agents. A well-conducted systematic review⁸ of 51 trials of at least a 6-month duration did not support the use of antihypertensive agents, diabetes medications, nonsteroidal anti-inflammatory drugs, aspirin, hormones, or lipid-lowering drugs for cognitive protection for people with normal cognition or mild cognitive impairment.

However, some studies found reassuring evidence that standard therapies for other conditions do not worsen cognitive decline and are protective for atrial fibrillation.⁸

Proton-pump inhibitors. Concern exists for a potential link between dementia risk and proton-pump inhibitors, which are widely used to treat acid-related gastrointestinal disorders.⁹

A prospective population-based cohort study¹⁰ of nearly 3,500 people ages 65 and older without baseline dementia screened participants for dementia every 2 years over a mean period of 7.5 years and provided further evaluation for those who screened positive. Use of proton-pump inhibitors was not found to be associated with dementia risk, even with high cumulative exposure.

Results from this study do not support avoiding proton-pump inhibitors out of concern for dementia risk, although long-term use is associated with other safety concerns.

Oral anticoagulation. The increased risk of dementia with atrial fibrillation is well documented.¹¹

A retrospective study¹² based on a Swedish health registry and using more than 444,000 patients covering more than 1.5 million years at risk found that oral anticoagulant treatment at baseline conferred a 29% lower risk of dementia in an intention-to-treat analysis and a 48% lower risk in on-treatment analysis compared with no oral anticoagulation therapy. No difference was found between new oral anticoagulants and warfarin.

Transcatheter aortic valve implantation is not associated with cognitive decline

For patients with severe aortic stenosis who are not surgical candidates, transcatheter aortic valve implantation is superior to standard medical therapy,¹³ but there are concerns of neurologic and cognitive changes after the procedure.¹⁴ A meta-analysis of 18 studies assessing cognitive performance in more than 1,000 patients (average age ≥ 80) after undergoing the procedure for severe aortic stenosis found no significant cognitive performance changes from baseline perioperatively or 3 or 6 months later.¹⁵

 

 

TREATING DEMENTIA-ASSOCIATED BEHAVIORAL DISTURBANCES

Behavioral and psychiatric symptoms often accompany dementia, but no drugs have yet been approved by the US Food and Drug Administration (FDA) to address them in this population. Nonpharmacologic interventions are recommended as first-line therapy.

Antipsychotics are not recommended

Antipsychotics are often prescribed,¹⁶ although they are associated with metabolic syndrome¹⁷ and increased risks of stroke and death.¹⁸ The FDA has issued black box warnings against using antipsychotics for behavioral management in patients with dementia. Further, the American Geriatrics Society and the American Psychiatric Association do not endorse using them as initial therapy for behavioral and psychological symptoms of dementia.^16,19

The Centers for Medicare and Medicaid Services partnered with nursing homes to improve the quality of care for patients with dementia, with results measured as the rate of prescribing antipsychotic medications. Although the use of psychotropic medications declined after initiating the partnership, the use of mood stabilizers increased, possibly as a substitute for antipsychotics.²⁰

Dextromethorphan-quinidine use is up, despite modest evidence of benefit

A consumer news report in 2017 stated that the use of dextromethorphan-quinidine in long-term care facilities increased by nearly 400% between 2012 and 2016.²¹

Evidence for its benefits comes from a 10-week, phase 2, randomized controlled trial conducted at 42 US study sites with 194 patients with probable Alzheimer disease. Compared with the placebo group, the active treatment group had mildly reduced agitation but an increased risk of falls, dizziness, and diarrhea. However, rates of adverse effects were low, and the authors concluded that treatment was generally well tolerated.²²

Pimavanserin: No long-term benefit for psychosis

In a phase 2, randomized, double-blind, placebo-controlled trial in 181 patients with possible or probable Alzheimer disease and psychotic symptoms, pimavanserin was associated with improved symptoms as measured by the Neuropsychiatric Inventory–Nursing Home Version psychosis score at 6 weeks, but no difference was found compared with placebo at 12 weeks. The treatment group had more adverse events, including agitation, aggression, peripheral edema, anxiety, and symptoms of dementia, although the differences were not statistically significant.²³               

DELIRIUM: AVOID ANTIPSYCHOTICS

Delirium is common in hospitalized older adults, especially those who have baseline cognitive or functional impairment and are exposed to precipitating factors such as treatment with anticholinergic or narcotic medications, infection, surgery, or admission to an intensive care unit.²⁴

Delirium at discharge predicts poor outcomes

In a prospective study of 152 hospitalized patients with delirium, those who either did not recover from delirium or had only partially recovered at discharge were more likely to visit the emergency department, be rehospitalized, or die during the subsequent 3 months than those who had fully recovered from delirium at discharge.²⁵

Multicomponent, patient-centered approach can help

A randomized trial in 377 patients in Taiwan evaluated the use of a modified Hospital Elder Life Program, consisting of 3 protocols focused on orienting communication, oral and nutritional assistance, and early mobilization. Patients were at least 65 years old and undergoing elective abdominal surgery with expected length of hospital stay longer than 6 days. The program, administered daily during hospitalization, significantly lowered postoperative delirium by 56% and hospital stay by 2 days compared with usual care.²⁶

Prophylactic haloperidol does not improve outcomes

In a multicenter randomized, double-blind, placebo-controlled trial, van den Boogaard et al studied prophylactic intravenous haloperidol in nearly 1,800 critically ill patients at high risk of delirium.²⁷ Haloperidol did not improve survival at 28 days compared with placebo. For secondary outcomes, including delirium incidence, delirium-free and coma-free days, duration of mechanical ventilation, and hospital and intensive care department length of stay, treatment was not found to differ statistically from placebo.

Antipsychotics may worsen delirium

A double-blind, parallel-arm, dose-titrated randomized trial, conducted at 11 Australian hospices or hospitals with palliative care services, administered oral risperidone, haloperidol, or placebo to 247 patients with life-limiting illness and delirium. Both treatment groups had higher delirium symptom scores than the placebo group.²⁸

In addition, a systematic review and meta-analysis of 19 studies found no benefit of antipsychotic medications for preventing or treating delirium in hospitalized adults.²⁹

Antipsychotics are often continued indefinitely

A retrospective chart review at a US academic health system found³⁰ that among 487 patients with a new antipsychotic medication prescribed during hospitalization, 147 (30.2%) were discharged on an antipsychotic. Of these, 121 (82.3%) had a diagnosis of delirium. Only 15 (12.4%) had discharge summaries that included instructions for discontinuing the drug.

Another US health system retrospectively reviewed antipsychotic use and found³¹ that out of 260 patients who were newly exposed to an antipsychotic drug during hospitalization, 146 (56.2%) were discharged on an antipsychotic drug, and 65% of these patients were still on the drug at the time of the next hospital admission.

 

 

EXERCISE, EXERCISE, EXERCISE

Exercise recommended, but not vitamin D, to prevent falls

In 2018, the US Preventive Services Task Force updated its recommendations for preventing falls in community-dwelling older adults.³² Based on the findings of several trials, the task force recommends exercise interventions for adults age 65 and older who are at increased risk for falls. Gait, balance, and functional training were studied in 17 trials, resistance training in 13, flexibility in 8, endurance training in 5, and tai chi in 3, with 5 studies including general physical activity. Exercise interventions most commonly took place for 3 sessions per week for 12 months (range 2–42 months).

The task force also recommends against vitamin D supplementation for fall prevention in community-dwelling adults age 65 or older who are not known to have osteoporosis or vitamin D deficiency.

Early mobilization helps inpatients

Hospitalized older adults usually spend most of their time in bed. Forty-five previously ambulatory patients (age ≥ 65 without dementia or delirium) in a Veterans Affairs hospital were monitored with wireless accelerometers and were found to spend, on average, 83% of the measured hospital stay in bed. Standing or walking time ranged from 0.2% to 21%, with a median of only 3% (43 minutes a day).³³

Since falls with injury became a Centers for Medicare and Medicaid Services nonreimbursable hospital-acquired condition, tension has arisen between promoting mobility and preventing falls.³⁴ Two studies evaluating the adoption of mobility-restricting approaches such as bed-alarms, “fall-alert” signs, supervision of patients in the bathroom, and ensuring patients’ walking aids are within reach, did not find a significant reduction in falls or fall-related injuries.^35,36

A clinically significant loss of community mobility is common after hospitalization in older adults.³⁷ Older adults who developed mobility impairment during hospitalization had a higher risk of death in a large, retrospective study.³⁸ A large Canadian multisite intervention trial³⁹ that promoted early mobilization in older patients who were admitted to general medical wards resulted in increased mobilization and significantly shorter hospital stays.

POSTHOSPITAL CARE NEEDS IMPROVEMENT

After hospitalization, older adults who have difficulty with activities of daily living or complex medical needs often require continued care.

About 20% of hospitalized Medicare beneficiaries in the United States are discharged to skilled nursing facilities.⁴⁰ This is often a stressful transition, and most people have little guidance on selecting a facility and simply choose one based on its proximity to home.⁴¹

A program of frequent visits by hospital-employed physicians and advanced practice professionals at skilled nursing facilities resulted in a significantly lower 30-day readmission rate compared with nonparticipating skilled nursing facilities in the same geographic area.⁴²

Home healthcare is recommended after hospital discharge at a rapidly increasing rate. Overall referral rates increased from 8.6% to 14.1% between 2001 and 2012, and from 14.3% to 24.0% for patients with heart failure.⁴³ A qualitative study of home healthcare nurses found a need for improved care coordination between home healthcare agencies and discharging hospitals, including defining accountability for orders and enhancing communication.⁴⁴

Unfortunately, recent research has not unveiled a breakthrough for preventing or treating cognitive impairment or Alzheimer disease. But several studies from the last 2 years are helping to drive the field of geriatrics forward, providing evidence of what does and does not help a variety of issues specific to the elderly. 

Based on a search of the 2017 and 2018 literature, this article presents new evidence on preventing and treating cognitive impairment, managing dementia-associated behavioral disturbances and delirium, preventing falls, and improving inpatient mobility and posthospital care transitions.

COGNITIVE IMPAIRMENT, DEMENTIA: STILL NO SILVER BULLET

With the exception of oral anticoagulation treatment for atrial fibrillation, there is little evidence that pharmacologic or nonpharmacologic interventions slow the onset or progression of Alzheimer disease.

Nonpharmacologic interventions

Home occupational therapy. A 2-year home-based occupational therapy intervention¹ showed no evidence of slowing functional decline in patients with Alzheimer disease. The randomized controlled trial involving 180 participants consisted of monthly sessions of an intensive, well-established collaborative-care management model that included fall prevention and other safety strategies, personalized training in activities of daily living, exercise, and education. Outcome measures for activities of daily living did not differ significantly between the treatment and control groups.¹

Physical activity. Whether physical activity interventions slow cognitive decline and prevent dementia in cognitively intact adults was examined in a systematic review of 32 trials.² Most of the trials followed patients for 6 months; a few stretched for 1 or 2 years.

Evidence was insufficient to prove cognitive benefit for short-term, single-component or multicomponent physical activity interventions. However, a multidomain physical activity intervention that also included dietary modifications and cognitive training did show a delay in cognitive decline, but only “low-strength” evidence.²

Nutritional supplements. The antioxidants vitamin E and selenium were studied for their possible cognitive benefit in the double-blind randomized Prevention of Alzheimer Disease by Vitamin E and Selenium trial³ in 3,786 asymptomatic men ages 60 and older. Neither supplement was found to prevent dementia over a 7-year follow-up period.

A review of 38 trials⁴ evaluated the effects on cognition of omega-3 fatty acids, soy, ginkgo biloba, B vitamins, vitamin D plus calcium, vitamin C, beta-carotene, and multi-ingredient supplements. It found insufficient evidence to recommend any over-the-counter supplement for cognitive protection in adults with normal cognition or mild cognitive impairment.

Pharmacologic treatments

Testosterone supplementation. The Testosterone Trials tested the effects of testosterone gel vs placebo for 1 year on 493 men over age 65 with low testosterone (< 275 ng/mL) and with subjective memory complaints and objective memory performance deficits. Treatment was not associated with improved memory or other cognitive functions compared with placebo.⁵

Antiamyloid drugs. A randomized, double-blind, placebo-controlled trial in nearly 2,000 patients evaluated verubecestat, an oral beta-site amyloid precursor protein-cleaving enzyme-1 inhibitor that reduces the amyloid-beta level in cerebrospinal fluid.⁶

Verubecestat did not reduce cognitive or functional decline in patients with mild-to-moderate Alzheimer disease, while adverse events including rashes, falls, injuries, sleep disturbances, suicidal ideation, weight loss, and hair color change were more common in the treatment groups. The trial was terminated early because of futility at 50 months.

And in a placebo-controlled trial of solanezumab, a monoclonal antibody directed against the amyloid beta peptide, no benefit was demonstrated at 80 weeks in more than 2,000 patients with Alzheimer disease.⁷

Multiple common agents. A well-conducted systematic review⁸ of 51 trials of at least a 6-month duration did not support the use of antihypertensive agents, diabetes medications, nonsteroidal anti-inflammatory drugs, aspirin, hormones, or lipid-lowering drugs for cognitive protection for people with normal cognition or mild cognitive impairment.

However, some studies found reassuring evidence that standard therapies for other conditions do not worsen cognitive decline and are protective for atrial fibrillation.⁸

Proton-pump inhibitors. Concern exists for a potential link between dementia risk and proton-pump inhibitors, which are widely used to treat acid-related gastrointestinal disorders.⁹

A prospective population-based cohort study¹⁰ of nearly 3,500 people ages 65 and older without baseline dementia screened participants for dementia every 2 years over a mean period of 7.5 years and provided further evaluation for those who screened positive. Use of proton-pump inhibitors was not found to be associated with dementia risk, even with high cumulative exposure.

Results from this study do not support avoiding proton-pump inhibitors out of concern for dementia risk, although long-term use is associated with other safety concerns.

Oral anticoagulation. The increased risk of dementia with atrial fibrillation is well documented.¹¹

A retrospective study¹² based on a Swedish health registry and using more than 444,000 patients covering more than 1.5 million years at risk found that oral anticoagulant treatment at baseline conferred a 29% lower risk of dementia in an intention-to-treat analysis and a 48% lower risk in on-treatment analysis compared with no oral anticoagulation therapy. No difference was found between new oral anticoagulants and warfarin.

Transcatheter aortic valve implantation is not associated with cognitive decline

For patients with severe aortic stenosis who are not surgical candidates, transcatheter aortic valve implantation is superior to standard medical therapy,¹³ but there are concerns of neurologic and cognitive changes after the procedure.¹⁴ A meta-analysis of 18 studies assessing cognitive performance in more than 1,000 patients (average age ≥ 80) after undergoing the procedure for severe aortic stenosis found no significant cognitive performance changes from baseline perioperatively or 3 or 6 months later.¹⁵

 

 

TREATING DEMENTIA-ASSOCIATED BEHAVIORAL DISTURBANCES

Behavioral and psychiatric symptoms often accompany dementia, but no drugs have yet been approved by the US Food and Drug Administration (FDA) to address them in this population. Nonpharmacologic interventions are recommended as first-line therapy.

Antipsychotics are not recommended

Antipsychotics are often prescribed,¹⁶ although they are associated with metabolic syndrome¹⁷ and increased risks of stroke and death.¹⁸ The FDA has issued black box warnings against using antipsychotics for behavioral management in patients with dementia. Further, the American Geriatrics Society and the American Psychiatric Association do not endorse using them as initial therapy for behavioral and psychological symptoms of dementia.^16,19

The Centers for Medicare and Medicaid Services partnered with nursing homes to improve the quality of care for patients with dementia, with results measured as the rate of prescribing antipsychotic medications. Although the use of psychotropic medications declined after initiating the partnership, the use of mood stabilizers increased, possibly as a substitute for antipsychotics.²⁰

Dextromethorphan-quinidine use is up, despite modest evidence of benefit

A consumer news report in 2017 stated that the use of dextromethorphan-quinidine in long-term care facilities increased by nearly 400% between 2012 and 2016.²¹

Evidence for its benefits comes from a 10-week, phase 2, randomized controlled trial conducted at 42 US study sites with 194 patients with probable Alzheimer disease. Compared with the placebo group, the active treatment group had mildly reduced agitation but an increased risk of falls, dizziness, and diarrhea. However, rates of adverse effects were low, and the authors concluded that treatment was generally well tolerated.²²

Pimavanserin: No long-term benefit for psychosis

In a phase 2, randomized, double-blind, placebo-controlled trial in 181 patients with possible or probable Alzheimer disease and psychotic symptoms, pimavanserin was associated with improved symptoms as measured by the Neuropsychiatric Inventory–Nursing Home Version psychosis score at 6 weeks, but no difference was found compared with placebo at 12 weeks. The treatment group had more adverse events, including agitation, aggression, peripheral edema, anxiety, and symptoms of dementia, although the differences were not statistically significant.²³               

DELIRIUM: AVOID ANTIPSYCHOTICS

Delirium is common in hospitalized older adults, especially those who have baseline cognitive or functional impairment and are exposed to precipitating factors such as treatment with anticholinergic or narcotic medications, infection, surgery, or admission to an intensive care unit.²⁴

Delirium at discharge predicts poor outcomes

In a prospective study of 152 hospitalized patients with delirium, those who either did not recover from delirium or had only partially recovered at discharge were more likely to visit the emergency department, be rehospitalized, or die during the subsequent 3 months than those who had fully recovered from delirium at discharge.²⁵

Multicomponent, patient-centered approach can help

A randomized trial in 377 patients in Taiwan evaluated the use of a modified Hospital Elder Life Program, consisting of 3 protocols focused on orienting communication, oral and nutritional assistance, and early mobilization. Patients were at least 65 years old and undergoing elective abdominal surgery with expected length of hospital stay longer than 6 days. The program, administered daily during hospitalization, significantly lowered postoperative delirium by 56% and hospital stay by 2 days compared with usual care.²⁶

Prophylactic haloperidol does not improve outcomes

In a multicenter randomized, double-blind, placebo-controlled trial, van den Boogaard et al studied prophylactic intravenous haloperidol in nearly 1,800 critically ill patients at high risk of delirium.²⁷ Haloperidol did not improve survival at 28 days compared with placebo. For secondary outcomes, including delirium incidence, delirium-free and coma-free days, duration of mechanical ventilation, and hospital and intensive care department length of stay, treatment was not found to differ statistically from placebo.

Antipsychotics may worsen delirium

A double-blind, parallel-arm, dose-titrated randomized trial, conducted at 11 Australian hospices or hospitals with palliative care services, administered oral risperidone, haloperidol, or placebo to 247 patients with life-limiting illness and delirium. Both treatment groups had higher delirium symptom scores than the placebo group.²⁸

In addition, a systematic review and meta-analysis of 19 studies found no benefit of antipsychotic medications for preventing or treating delirium in hospitalized adults.²⁹

Antipsychotics are often continued indefinitely

A retrospective chart review at a US academic health system found³⁰ that among 487 patients with a new antipsychotic medication prescribed during hospitalization, 147 (30.2%) were discharged on an antipsychotic. Of these, 121 (82.3%) had a diagnosis of delirium. Only 15 (12.4%) had discharge summaries that included instructions for discontinuing the drug.

Another US health system retrospectively reviewed antipsychotic use and found³¹ that out of 260 patients who were newly exposed to an antipsychotic drug during hospitalization, 146 (56.2%) were discharged on an antipsychotic drug, and 65% of these patients were still on the drug at the time of the next hospital admission.

 

 

EXERCISE, EXERCISE, EXERCISE

Exercise recommended, but not vitamin D, to prevent falls

In 2018, the US Preventive Services Task Force updated its recommendations for preventing falls in community-dwelling older adults.³² Based on the findings of several trials, the task force recommends exercise interventions for adults age 65 and older who are at increased risk for falls. Gait, balance, and functional training were studied in 17 trials, resistance training in 13, flexibility in 8, endurance training in 5, and tai chi in 3, with 5 studies including general physical activity. Exercise interventions most commonly took place for 3 sessions per week for 12 months (range 2–42 months).

The task force also recommends against vitamin D supplementation for fall prevention in community-dwelling adults age 65 or older who are not known to have osteoporosis or vitamin D deficiency.

Early mobilization helps inpatients

Hospitalized older adults usually spend most of their time in bed. Forty-five previously ambulatory patients (age ≥ 65 without dementia or delirium) in a Veterans Affairs hospital were monitored with wireless accelerometers and were found to spend, on average, 83% of the measured hospital stay in bed. Standing or walking time ranged from 0.2% to 21%, with a median of only 3% (43 minutes a day).³³

Since falls with injury became a Centers for Medicare and Medicaid Services nonreimbursable hospital-acquired condition, tension has arisen between promoting mobility and preventing falls.³⁴ Two studies evaluating the adoption of mobility-restricting approaches such as bed-alarms, “fall-alert” signs, supervision of patients in the bathroom, and ensuring patients’ walking aids are within reach, did not find a significant reduction in falls or fall-related injuries.^35,36

A clinically significant loss of community mobility is common after hospitalization in older adults.³⁷ Older adults who developed mobility impairment during hospitalization had a higher risk of death in a large, retrospective study.³⁸ A large Canadian multisite intervention trial³⁹ that promoted early mobilization in older patients who were admitted to general medical wards resulted in increased mobilization and significantly shorter hospital stays.

POSTHOSPITAL CARE NEEDS IMPROVEMENT

After hospitalization, older adults who have difficulty with activities of daily living or complex medical needs often require continued care.

About 20% of hospitalized Medicare beneficiaries in the United States are discharged to skilled nursing facilities.⁴⁰ This is often a stressful transition, and most people have little guidance on selecting a facility and simply choose one based on its proximity to home.⁴¹

A program of frequent visits by hospital-employed physicians and advanced practice professionals at skilled nursing facilities resulted in a significantly lower 30-day readmission rate compared with nonparticipating skilled nursing facilities in the same geographic area.⁴²

Home healthcare is recommended after hospital discharge at a rapidly increasing rate. Overall referral rates increased from 8.6% to 14.1% between 2001 and 2012, and from 14.3% to 24.0% for patients with heart failure.⁴³ A qualitative study of home healthcare nurses found a need for improved care coordination between home healthcare agencies and discharging hospitals, including defining accountability for orders and enhancing communication.⁴⁴

Unfortunately, recent research has not unveiled a breakthrough for preventing or treating cognitive impairment or Alzheimer disease. But several studies from the last 2 years are helping to drive the field of geriatrics forward, providing evidence of what does and does not help a variety of issues specific to the elderly. 

Based on a search of the 2017 and 2018 literature, this article presents new evidence on preventing and treating cognitive impairment, managing dementia-associated behavioral disturbances and delirium, preventing falls, and improving inpatient mobility and posthospital care transitions.

COGNITIVE IMPAIRMENT, DEMENTIA: STILL NO SILVER BULLET

With the exception of oral anticoagulation treatment for atrial fibrillation, there is little evidence that pharmacologic or nonpharmacologic interventions slow the onset or progression of Alzheimer disease.

Nonpharmacologic interventions

Home occupational therapy. A 2-year home-based occupational therapy intervention¹ showed no evidence of slowing functional decline in patients with Alzheimer disease. The randomized controlled trial involving 180 participants consisted of monthly sessions of an intensive, well-established collaborative-care management model that included fall prevention and other safety strategies, personalized training in activities of daily living, exercise, and education. Outcome measures for activities of daily living did not differ significantly between the treatment and control groups.¹

Physical activity. Whether physical activity interventions slow cognitive decline and prevent dementia in cognitively intact adults was examined in a systematic review of 32 trials.² Most of the trials followed patients for 6 months; a few stretched for 1 or 2 years.

Evidence was insufficient to prove cognitive benefit for short-term, single-component or multicomponent physical activity interventions. However, a multidomain physical activity intervention that also included dietary modifications and cognitive training did show a delay in cognitive decline, but only “low-strength” evidence.²

Nutritional supplements. The antioxidants vitamin E and selenium were studied for their possible cognitive benefit in the double-blind randomized Prevention of Alzheimer Disease by Vitamin E and Selenium trial³ in 3,786 asymptomatic men ages 60 and older. Neither supplement was found to prevent dementia over a 7-year follow-up period.

A review of 38 trials⁴ evaluated the effects on cognition of omega-3 fatty acids, soy, ginkgo biloba, B vitamins, vitamin D plus calcium, vitamin C, beta-carotene, and multi-ingredient supplements. It found insufficient evidence to recommend any over-the-counter supplement for cognitive protection in adults with normal cognition or mild cognitive impairment.

Pharmacologic treatments

Testosterone supplementation. The Testosterone Trials tested the effects of testosterone gel vs placebo for 1 year on 493 men over age 65 with low testosterone (< 275 ng/mL) and with subjective memory complaints and objective memory performance deficits. Treatment was not associated with improved memory or other cognitive functions compared with placebo.⁵

Antiamyloid drugs. A randomized, double-blind, placebo-controlled trial in nearly 2,000 patients evaluated verubecestat, an oral beta-site amyloid precursor protein-cleaving enzyme-1 inhibitor that reduces the amyloid-beta level in cerebrospinal fluid.⁶

Verubecestat did not reduce cognitive or functional decline in patients with mild-to-moderate Alzheimer disease, while adverse events including rashes, falls, injuries, sleep disturbances, suicidal ideation, weight loss, and hair color change were more common in the treatment groups. The trial was terminated early because of futility at 50 months.

And in a placebo-controlled trial of solanezumab, a monoclonal antibody directed against the amyloid beta peptide, no benefit was demonstrated at 80 weeks in more than 2,000 patients with Alzheimer disease.⁷

Multiple common agents. A well-conducted systematic review⁸ of 51 trials of at least a 6-month duration did not support the use of antihypertensive agents, diabetes medications, nonsteroidal anti-inflammatory drugs, aspirin, hormones, or lipid-lowering drugs for cognitive protection for people with normal cognition or mild cognitive impairment.

However, some studies found reassuring evidence that standard therapies for other conditions do not worsen cognitive decline and are protective for atrial fibrillation.⁸

Proton-pump inhibitors. Concern exists for a potential link between dementia risk and proton-pump inhibitors, which are widely used to treat acid-related gastrointestinal disorders.⁹

A prospective population-based cohort study¹⁰ of nearly 3,500 people ages 65 and older without baseline dementia screened participants for dementia every 2 years over a mean period of 7.5 years and provided further evaluation for those who screened positive. Use of proton-pump inhibitors was not found to be associated with dementia risk, even with high cumulative exposure.

Results from this study do not support avoiding proton-pump inhibitors out of concern for dementia risk, although long-term use is associated with other safety concerns.

Oral anticoagulation. The increased risk of dementia with atrial fibrillation is well documented.¹¹

A retrospective study¹² based on a Swedish health registry and using more than 444,000 patients covering more than 1.5 million years at risk found that oral anticoagulant treatment at baseline conferred a 29% lower risk of dementia in an intention-to-treat analysis and a 48% lower risk in on-treatment analysis compared with no oral anticoagulation therapy. No difference was found between new oral anticoagulants and warfarin.

Transcatheter aortic valve implantation is not associated with cognitive decline

For patients with severe aortic stenosis who are not surgical candidates, transcatheter aortic valve implantation is superior to standard medical therapy,¹³ but there are concerns of neurologic and cognitive changes after the procedure.¹⁴ A meta-analysis of 18 studies assessing cognitive performance in more than 1,000 patients (average age ≥ 80) after undergoing the procedure for severe aortic stenosis found no significant cognitive performance changes from baseline perioperatively or 3 or 6 months later.¹⁵

 

 

TREATING DEMENTIA-ASSOCIATED BEHAVIORAL DISTURBANCES

Behavioral and psychiatric symptoms often accompany dementia, but no drugs have yet been approved by the US Food and Drug Administration (FDA) to address them in this population. Nonpharmacologic interventions are recommended as first-line therapy.

Antipsychotics are not recommended

Antipsychotics are often prescribed,¹⁶ although they are associated with metabolic syndrome¹⁷ and increased risks of stroke and death.¹⁸ The FDA has issued black box warnings against using antipsychotics for behavioral management in patients with dementia. Further, the American Geriatrics Society and the American Psychiatric Association do not endorse using them as initial therapy for behavioral and psychological symptoms of dementia.^16,19

The Centers for Medicare and Medicaid Services partnered with nursing homes to improve the quality of care for patients with dementia, with results measured as the rate of prescribing antipsychotic medications. Although the use of psychotropic medications declined after initiating the partnership, the use of mood stabilizers increased, possibly as a substitute for antipsychotics.²⁰

Dextromethorphan-quinidine use is up, despite modest evidence of benefit

A consumer news report in 2017 stated that the use of dextromethorphan-quinidine in long-term care facilities increased by nearly 400% between 2012 and 2016.²¹

Evidence for its benefits comes from a 10-week, phase 2, randomized controlled trial conducted at 42 US study sites with 194 patients with probable Alzheimer disease. Compared with the placebo group, the active treatment group had mildly reduced agitation but an increased risk of falls, dizziness, and diarrhea. However, rates of adverse effects were low, and the authors concluded that treatment was generally well tolerated.²²

Pimavanserin: No long-term benefit for psychosis

In a phase 2, randomized, double-blind, placebo-controlled trial in 181 patients with possible or probable Alzheimer disease and psychotic symptoms, pimavanserin was associated with improved symptoms as measured by the Neuropsychiatric Inventory–Nursing Home Version psychosis score at 6 weeks, but no difference was found compared with placebo at 12 weeks. The treatment group had more adverse events, including agitation, aggression, peripheral edema, anxiety, and symptoms of dementia, although the differences were not statistically significant.²³               

DELIRIUM: AVOID ANTIPSYCHOTICS

Delirium is common in hospitalized older adults, especially those who have baseline cognitive or functional impairment and are exposed to precipitating factors such as treatment with anticholinergic or narcotic medications, infection, surgery, or admission to an intensive care unit.²⁴

Delirium at discharge predicts poor outcomes

In a prospective study of 152 hospitalized patients with delirium, those who either did not recover from delirium or had only partially recovered at discharge were more likely to visit the emergency department, be rehospitalized, or die during the subsequent 3 months than those who had fully recovered from delirium at discharge.²⁵

Multicomponent, patient-centered approach can help

A randomized trial in 377 patients in Taiwan evaluated the use of a modified Hospital Elder Life Program, consisting of 3 protocols focused on orienting communication, oral and nutritional assistance, and early mobilization. Patients were at least 65 years old and undergoing elective abdominal surgery with expected length of hospital stay longer than 6 days. The program, administered daily during hospitalization, significantly lowered postoperative delirium by 56% and hospital stay by 2 days compared with usual care.²⁶

Prophylactic haloperidol does not improve outcomes

In a multicenter randomized, double-blind, placebo-controlled trial, van den Boogaard et al studied prophylactic intravenous haloperidol in nearly 1,800 critically ill patients at high risk of delirium.²⁷ Haloperidol did not improve survival at 28 days compared with placebo. For secondary outcomes, including delirium incidence, delirium-free and coma-free days, duration of mechanical ventilation, and hospital and intensive care department length of stay, treatment was not found to differ statistically from placebo.

Antipsychotics may worsen delirium

A double-blind, parallel-arm, dose-titrated randomized trial, conducted at 11 Australian hospices or hospitals with palliative care services, administered oral risperidone, haloperidol, or placebo to 247 patients with life-limiting illness and delirium. Both treatment groups had higher delirium symptom scores than the placebo group.²⁸

In addition, a systematic review and meta-analysis of 19 studies found no benefit of antipsychotic medications for preventing or treating delirium in hospitalized adults.²⁹

Antipsychotics are often continued indefinitely

A retrospective chart review at a US academic health system found³⁰ that among 487 patients with a new antipsychotic medication prescribed during hospitalization, 147 (30.2%) were discharged on an antipsychotic. Of these, 121 (82.3%) had a diagnosis of delirium. Only 15 (12.4%) had discharge summaries that included instructions for discontinuing the drug.

Another US health system retrospectively reviewed antipsychotic use and found³¹ that out of 260 patients who were newly exposed to an antipsychotic drug during hospitalization, 146 (56.2%) were discharged on an antipsychotic drug, and 65% of these patients were still on the drug at the time of the next hospital admission.

 

 

EXERCISE, EXERCISE, EXERCISE

Exercise recommended, but not vitamin D, to prevent falls

In 2018, the US Preventive Services Task Force updated its recommendations for preventing falls in community-dwelling older adults.³² Based on the findings of several trials, the task force recommends exercise interventions for adults age 65 and older who are at increased risk for falls. Gait, balance, and functional training were studied in 17 trials, resistance training in 13, flexibility in 8, endurance training in 5, and tai chi in 3, with 5 studies including general physical activity. Exercise interventions most commonly took place for 3 sessions per week for 12 months (range 2–42 months).

The task force also recommends against vitamin D supplementation for fall prevention in community-dwelling adults age 65 or older who are not known to have osteoporosis or vitamin D deficiency.

Early mobilization helps inpatients

Hospitalized older adults usually spend most of their time in bed. Forty-five previously ambulatory patients (age ≥ 65 without dementia or delirium) in a Veterans Affairs hospital were monitored with wireless accelerometers and were found to spend, on average, 83% of the measured hospital stay in bed. Standing or walking time ranged from 0.2% to 21%, with a median of only 3% (43 minutes a day).³³

Since falls with injury became a Centers for Medicare and Medicaid Services nonreimbursable hospital-acquired condition, tension has arisen between promoting mobility and preventing falls.³⁴ Two studies evaluating the adoption of mobility-restricting approaches such as bed-alarms, “fall-alert” signs, supervision of patients in the bathroom, and ensuring patients’ walking aids are within reach, did not find a significant reduction in falls or fall-related injuries.^35,36

A clinically significant loss of community mobility is common after hospitalization in older adults.³⁷ Older adults who developed mobility impairment during hospitalization had a higher risk of death in a large, retrospective study.³⁸ A large Canadian multisite intervention trial³⁹ that promoted early mobilization in older patients who were admitted to general medical wards resulted in increased mobilization and significantly shorter hospital stays.

POSTHOSPITAL CARE NEEDS IMPROVEMENT

After hospitalization, older adults who have difficulty with activities of daily living or complex medical needs often require continued care.

About 20% of hospitalized Medicare beneficiaries in the United States are discharged to skilled nursing facilities.⁴⁰ This is often a stressful transition, and most people have little guidance on selecting a facility and simply choose one based on its proximity to home.⁴¹

A program of frequent visits by hospital-employed physicians and advanced practice professionals at skilled nursing facilities resulted in a significantly lower 30-day readmission rate compared with nonparticipating skilled nursing facilities in the same geographic area.⁴²

Home healthcare is recommended after hospital discharge at a rapidly increasing rate. Overall referral rates increased from 8.6% to 14.1% between 2001 and 2012, and from 14.3% to 24.0% for patients with heart failure.⁴³ A qualitative study of home healthcare nurses found a need for improved care coordination between home healthcare agencies and discharging hospitals, including defining accountability for orders and enhancing communication.⁴⁴

Narcolepsy was originally described in the late 1800s by the French physician Jean-Baptiste-Edouard Gélineau, who reported the case of a wine merchant suffering from somnolence. In this first description, he coined the term narcolepsie by joining the Greek words narke (numbness or stupor) and lepsis (attack).¹

Since then, the disorder has been further characterized, and some insight into its biological underpinnings has been established. Importantly, treatments have improved and expanded, facilitating its management and thereby improving quality of life for those with the disorder.

This review focuses on clinically relevant features of the disorder and proposes management strategies.

CLINICAL FEATURES

Narcolepsy is characterized by instability of sleep-wake transitions.

Daytime sleepiness

Clinically, narcolepsy manifests with excessive daytime sleepiness that can be personally and socially disabling. Cataplexy, sleep paralysis, and hypnagogic or hypnopompic hallucinations can also be present,^2,3 but they are not necessary for diagnosis. In fact, a minority of patients with narcolepsy have all these symptoms.⁴ Narcolepsy is divided into type 1 (with cataplexy) and type 2 (without cataplexy).²

Sleepiness tends to be worse with inactivity, and sleep can often be irresistible. Sleep attacks can come on suddenly and may be brief enough to manifest as a lapse in consciousness.

Short naps tend to be refreshing. Rapid eye movement (REM) latency—the interval between falling asleep and the onset of the REM sleep—is short in narcolepsy, and since the REM stage is when dreaming occurs, naps often include dreaming. Therefore, when taking a history, it is worthwhile to ask patients whether they dream during naps; a yes answer supports the diagnosis of narcolepsy.⁵

In children, sleepiness can manifest in reduced concentration and behavioral issues.⁶ Napping after age 5 or 6 is considered abnormal and may reflect pathologic sleepiness.⁷

Cataplexy

Cataplexy—transient muscle weakness triggered by emotion—is a specific feature of narcolepsy type 1. It often begins in the facial muscles and can manifest with slackening of the jaw or brief dropping of the head. However, episodes can be more dramatic and, if the trunk and limb muscles are affected, can result in collapsing to the ground.

Cataplexy usually has its onset at about the same time as the sleepiness associated with narcolepsy, but it can arise even years later.⁸ Episodes can last from a few seconds to 2 minutes. Consciousness is always preserved. A range of emotions can trigger cataplexy, but typically the emotion is a positive one such as laughter or excitement.⁹ Deep tendon reflexes disappear in cataplexy, so checking reflexes during a witnessed episode can be clinically valuable.²

Cataplexy can worsen with stress and insufficient sleep, occasionally with “status cataplecticus,” in which repeated, persistent episodes of cataplexy occur over several hours.⁸ Status cataplecticus can be spontaneous or an effect of withdrawal from anticataplectic medications.²

Cataplexy is thought to represent intrusion of REM sleep and its associated muscle atonia during wakefulness.

Sleep paralysis, hallucinations

Sleep paralysis and hallucinations are other features of narcolepsy that reflect this REM dissociation from sleep.

Sleep paralysis occurs most commonly upon awakening, but sometimes just before sleep onset. In most cases, it is manifested by inability to move the limbs or speak, lasting several seconds or, in rare cases, minutes at a time. Sleep paralysis can be associated with a sensation of fear or suffocation, especially when initially experienced.⁸

Hypnopompic hallucinations, occurring upon awakening, are more common than hypnagogic hallucinations, which are experienced before falling asleep. The hallucinations are often vivid and usually visual, although other types of hallucinations are possible. Unlike those that occur in psychotic disorders, the hallucinations tend to be associated with preserved insight that they are not real.¹⁰

Notably, both sleep paralysis and hallucinations are nonspecific symptoms that are common in the general population.^8,11,12

Fragmented sleep

Although they are very sleepy, people with narcolepsy generally cannot stay asleep for very long. Their sleep tends to be extremely fragmented, and they often wake up several times a night.²

This sleep pattern reflects the inherent instability of sleep-wake transitions in narcolepsy. In fact, over a 24-hour period, adults with narcolepsy have a normal amount of sleep.¹³ In children, however, when narcolepsy first arises, the 24-hour sleep time can increase abruptly and can sometimes be associated with persistent cataplexy that can manifest as a clumsy gait.¹⁴

Weight gain, obstructive sleep apnea

Weight gain is common, particularly after symptom onset, and especially in children. As a result, obesity is a frequent comorbidity.¹⁵ Because obstructive sleep apnea can consequently develop, all patients with narcolepsy require screening for sleep-disordered breathing.

Other sleep disorders often accompany narcolepsy and are more common than in the general population.¹⁶ In a study incorporating both clinical and polysomnographic data of 100 patients with narcolepsy, insomnia was the most common comorbid sleep disorder, with a prevalence of 28%; others were REM sleep behavior disorder (24%), restless legs syndrome (24%), obstructive sleep apnea (21%), and non-REM parasomnias.¹⁷

 

 

PSYCHOSOCIAL CONSEQUENCES

Narcolepsy has significant psychosocial consequences. As a result of their symptoms, people with narcolepsy may not be able to meet academic or work-related demands.

Additionally, their risk of a motor vehicle accident is 3 to 4 times higher than in the general population, and more than one-third of patients have been in an accident due to sleepiness.¹⁸ There is some evidence to show that treatment eliminates this risk.¹⁹

Few systematic studies have examined mood disorders in narcolepsy. However, studies tend to show a higher prevalence of psychiatric disorders than in the general population, with depression and anxiety the most com-mon.^20,21

DIAGNOSIS IS OFTEN DELAYED

The prevalence of narcolepsy type 1 is between 25 and 100 per 100,000 people.²² In a Mayo Clinic study,²³ the incidence of narcolepsy type 1 was estimated to be 0.74 per 100,000 person-years. Epidemiologic data on narcolepsy type 2 are sparse, but patients with narcolepsy without cataplexy are thought to represent only 36% of all narcolepsy patients.²³

Diagnosis is often delayed, with the average time between the onset of symptoms and the diagnosis ranging from 8 to 22 years. With increasing awareness, the efficiency of the diagnostic process is improving, and this delay is expected to lessen accordingly.²⁴

Symptoms most commonly arise in the second decade; but the age at onset ranges significantly, between the first and fifth decades. Narcolepsy has a bimodal distribution in incidence, with the biggest peak at approximately age 15 and second smaller peak in the mid-30s. Some studies have suggested a slight male predominance.^23,25

DIAGNOSIS

History is key

The history should include specific questions about the hallmark features of narcolepsy, including cataplexy, sleep paralysis, and sleep-related hallucinations. For individual assessment of subjective sleepiness, the Epworth Sleepiness Scale or Pediatric Daytime Sleepiness Scale can be administered quickly in the office setting.^26,27

The Epworth score is calculated from the self-rated likelihood of falling asleep in 8 different situations, with possible scores of 0 (would never doze) to 3 (high chance of dozing) on each question, for a total possible score of 0 to 24. Normal total scores are between 0 and 10, while scores greater than 10 reflect pathologic sleepiness. Scores on the Epworth Sleepiness Scale in those with narcolepsy tend to reflect moderate to severe sleepiness, or at least 13, as opposed to patients with obstructive sleep apnea, whose scores commonly reflect milder sleepiness.²⁸

Testing with actigraphy and polysomnography

It is imperative to rule out insufficient sleep and other sleep disorders as a cause of daytime sleepiness. This can be done with a careful clinical history, actigraphy with sleep logs, and polysomnography.

In the 2 to 4 weeks before actigraphy and subsequent testing, all medications with alerting or sedating properties (including anti­depressants) should be tapered off to prevent influence on the results of the study.

Delayed sleep-phase disorder presents at a similar age as narcolepsy and can be associated with similar degrees of sleepiness. However, individuals with delayed sleep phase disorder have an inappropriately timed sleep-wake cycle so that there is a shift in their desired sleep onset and awakening times. It is common—prevalence estimates vary but average about 1% in the general population.²⁹

Insufficient sleep syndrome is even more common, especially in teenagers and young adults, with increasing family, social, and academic demands. Sleep needs vary across the life span. A teenager needs 8 to 10 hours of sleep per night, and a young adult needs 7 to 9 hours. A study of 1,285 high school students found that 10.4% were not getting enough sleep.³⁰

If actigraphy data suggest a circadian rhythm disorder or insufficient sleep that could explain the symptoms of sleepiness, then further testing should be halted and these specific issues should be addressed. In these cases, working with the patient toward maintaining a regular sleep-wake schedule with 7 to 8 hours of nightly sleep will often resolve symptoms.

If actigraphy demonstrates the patient is maintaining a regular sleep schedule and allowing adequate time for nightly sleep, the next step is polysomnography.

Polysomnography is performed to detect other disorders that can disrupt sleep, such as sleep-disordered breathing or periodic limb movement disorder.^2,5 In addition, polysomnography can provide assurance that adequate sleep was obtained prior to the next step in testing.

Multiple sleep latency test

If sufficient sleep is obtained on polysomnograpy (at least 6 hours for an adult) and no other sleep disorder is identified, a multiple sleep latency test is performed. A urine toxicology screen is typically performed on the day of the test to ensure that drugs are not affecting the results.

The multiple sleep latency test consists of 4 to 5 nap opportunities at 2-hour intervals in a quiet dark room conducive to sleep, during which both sleep and REM latency are recorded. The sleep latency of those with narcolepsy is significantly shortened, and the diagnosis of narcolepsy requires an average sleep latency of less than 8 minutes.

Given the propensity for REM sleep in narcolepsy, another essential feature for diagnosis is the sleep-onset REM period (SOREMP). A SOREMP is defined as a REM latency of less than 15 minutes. A diagnosis of narcolepsy re-quires a SOREMP in at least 2 of the naps in a multiple sleep latency test (or 1 nap if the shortened REM latency is seen during polysomnography).³¹

The multiple sleep latency test has an imperfect sensitivity, though, and should be repeated when there is a high suspicion of narcolepsy.^32–34 It is not completely specific either, and false-positive results occur. In fact, SOREMPs can be seen in the general population, particularly in those with a circadian rhythm disorder, insufficient sleep, or sleep-disordered breathing. Two or more SOREMPs in an multiple sleep latency test can be seen in a small proportion of the general population.³⁵ The results of a multiple sleep latency test should be interpreted in the clinical context.

Differential diagnosis

Narcolepsy type 1 is distinguished from type 2 by the presence of cataplexy. A cerebrospinal fluid hypocretin 1 level of 110 pg/mL or less, or less than one-third of the mean value obtained in normal individuals, can substitute for the multiple sleep latency test in diagnosing narcolepsy type 1.³¹ Currently, hypocretin testing is generally not performed in clinical practice, although it may become a routine part of the narcolepsy evaluation in the future.

Thus, according to the International Classification of Sleep Disorders, 3rd edition,³¹ the diagnosis of narcolepsy type 1 requires excessive daytime sleepiness for at least 3 months that cannot be explained by another sleep disorder, medical or neurologic disorder, mental disorder, medication use, or substance use disorder, and at least 1 of the following:

• Cataplexy and mean sleep latency of 8 minutes or less with at least 2 SOREMPs on multiple sleep latency testing (1 of which can be on the preceding night’s polysomography)
• Cerebrospinal fluid hypocretin 1 levels less than 110 pg/mL or one-third the baseline normal levels and mean sleep latency ≤ 8 minutes with ≥ 2 SOREMPs on multiple sleep latency testing.

Similarly, the diagnosis of narcolepsy type 2 requires excessive daytime sleepiness for at least 3 months that cannot be explained by another sleep disorder, medical or neurological disorder, mental disorder, medication use, or substance use disorder, plus:

• Mean sleep latency of 8 minutes or less with at least 2 SOREMPs on multiple sleep latency testing.

Idiopathic hypersomnia, another disorder of central hypersomnolence, is also characterized by disabling sleepiness. It is diagnostically differentiated from narcolepsy, as there are fewer than 2 SOREMPs. As opposed to narcolepsy, in which naps tend to be refreshing, even prolonged naps in idiopathic hypersomnia are often not helpful in restoring wakefulness. In idiopathic hypersomnia, sleep is usually not fragmented, and there are few nocturnal arousals. Sleep times can often be prolonged as well, whereas in narcolepsy total sleep time through the day may not be increased but is not consolidated.

Kleine-Levin syndrome is a rarer disorder of hypersomnia. It is episodic compared with the relatively persistent sleepiness in narcolepsy and idiopathic hypersomnia. Periods of hypersomnia occur intermittently for days to weeks and are accompanied by cognitive and behavioral changes including hyperphagia and hypersexuality.⁴

LINKED TO HYPOCRETIN DEFICIENCY

Over the past 2 decades, the underlying pathophysiology of narcolepsy type 1 has been better characterized.

Narcolepsy type 1 has been linked to a deficiency in hypocretin in the central nervous system.³⁶ Hypocretin (also known as orexin) is a hormone produced in the hypothalamus that acts on multiple brain regions and maintains alertness. For unclear reasons, hypothalamic neurons producing hypocretin are selectively reduced in narcolepsy type 1. Hypocretin also stabilizes wakefulness and inhibits REM sleep; therefore, hypocretin deficiency can lead to inappropriate intrusions of REM sleep onto wakefulness, leading to the hallmark features of narcolepsy—cataplexy, sleep-related hallucinations, and sleep paralysis.³⁷ According to one theory, cataplexy is triggered by emotional stimuli because of a pathway between the medial prefrontal cortex and the amygdala to the pons.³⁸

Cerebrospinal fluid levels of hypocretin in patients with narcolepsy type 2 tend to be normal, and the biologic underpinnings of narcolepsy type 2 remain mysterious. However, in the subgroup of those with narcolepsy type 2 in which hypocretin is low, many individuals go on to develop cataplexy, thereby evolving to narcolepsy type 1.³⁶

POSSIBLE AUTOIMMUNE BASIS

Narcolepsy is typically a sporadic disorder, although familial cases have been described. The risk of a parent with narcolepsy having a child who is affected is approximately 1%.⁵

Narcolepsy type 1 is strongly associated with HLA-DQB1*0602, with up to 95% of those affected having at least one allele.³⁹ Having 2 copies of the allele further increases the risk of developing narcolepsy.⁴⁰ However, this allele is far from specific for narcolepsy with cataplexy, as it occurs in 12% to 38% of the general population.⁴¹ Therefore, HLA typing currently has limited clinical utility. The exact cause is as yet unknown, but substantial literature proposes an autoimmune basis of the disorder, given the strong association with the HLA subtype.^42–44

After the 2009 H1N1 influenza pandemic, there was a significant increase in the incidence of narcolepsy with cataplexy, which again sparked interest in an autoimmune etiology underlying the disorder. Pandemrix, an H1N1 vaccine produced as a result of the 2009 pandemic, appeared to also be associated with an increase in the incidence of narcolepsy. An association with other upper respiratory infections has also been noted, further supporting a possible autoimmune basis.

A few studies have looked for serum autoantibodies involved in the pathogenesis of narcolepsy. Thus far, only one has been identified, an antibody to Tribbles homolog 2, found in 20% to 40% of those with new onset of nar-colepsy.^42–44

TREATMENTS FOR DAYTIME SLEEPINESS

As with many chronic disorders, the treatment of narcolepsy consists of symptomatic rather than curative management, which can be done through both pharmacologic and nonpharmacologic means.

Nondrug measures

Scheduled naps lasting 15 to 20 minutes can help improve alertness.⁴⁵ A consistent sleep schedule with good sleep hygiene, ensuring sufficient nightly sleep, is also important. In one study, the combination of scheduled naps and regular nocturnal sleep times reduced the level of daytime sleepiness and unintentional daytime sleep. Daytime naps were most helpful for those with the highest degree of daytime sleepiness.⁴⁵

Strategic use of caffeine can be helpful and can reduce dependence on pharmacologic treatment.

Screening should be performed routinely for other sleep disorders, such as sleep-disordered breathing, which should be treated if identified.^5,18 When being treated for other medical conditions, individuals with narcolepsy should avoid medications that can cause sedation, such as opiates or barbiturates; alcohol should be minimized or avoided.

Networking with other individuals with narcolepsy through support groups such as Narcolepsy Network can be valuable for learning coping skills and connecting with community resources. Psychological counseling for the patient, and sometimes the family, can also be useful. School-age children may need special accommodations such as schedule adjustments to allow for scheduled naps or frequent breaks to maintain alertness.

People with narcolepsy tend to function better in careers that do not require long periods of sitting, as sleepiness tends to be worse, but instead offer flexibility and require higher levels of activity that tend to combat sleepiness. They should not work as commercial drivers.¹⁸

 

 

Medications

While behavioral interventions in narcolepsy are vital, they are rarely sufficient, and drugs that promote daytime wakefulness are used as an adjunct (Table 2).⁴⁶

Realistic expectations should be established when starting, as some degree of residual sleepiness usually remains even with optimal medical therapy. Medications should be strategically scheduled to maximize alertness during necessary times such as at work or school or during driving. Patients should specifically be counseled to avoid driving if sleepy.^18,47

Modafinil is often used as a first-line therapy, given its favorable side-effect profile and low potential for abuse. Its pharmacologic action has been debated but it probably acts as a selective dopamine reuptake inhibitor. It is typically taken twice daily (upon waking and early afternoon) and is usually well tolerated.

Potential side effects include headache, nausea, dry mouth, anorexia, diarrhea, and, rarely, Stevens-Johnson syndrome. Cardiovascular side effects are minimal, making it a favorable choice in older patients.^18,48

A trial in 283 patients showed significantly lower levels of sleepiness in patients taking modafinil 200 mg or 400 mg than in a control group. Other trials have supported these findings and showed improved driving performance on modafinil.¹⁸

Notably, modafinil can increase the metabolism of oral contraceptives, thereby reducing their efficacy. Women of childbearing age should be warned about this interaction and should be transitioned to nonhormonal forms of contraception.^2,47

Armodafinil, a purified R-isomer of modafinil, has a longer half-life and requires only once-daily dosing.⁵

If modafinil or armodafinil fails to optimally manage daytime sleepiness, a traditional stimulant such as methylphenidate or an amphetamine is often used.

Methylphenidate and amphetamines primarily inhibit the reuptake and increase the release of the monoamines, mainly dopamine, and to a lesser degree serotonin and norepinephrine.

These drugs have more significant adverse effects that can involve the cardiovascular system, causing hypertension and arrhythmias. Anorexia, weight loss, and, particularly with high doses, psychosis can occur.⁴⁹

These drugs should be avoided in patients with a history of significant cardiovascular disease. Before starting stimulant therapy, a thorough cardiovascular examination should be done, often including electrocardiography to ensure there is no baseline arrhythmia.

Patients on these medications should be followed closely to ensure that blood pressure, pulse, and weight are not negatively affected.^18,50 Addiction and tolerance can develop with these drugs, and follow-up should include assessment for dependence. Some states may require prescription drug monitoring to ensure the drugs are not being abused or diverted.

Short- and long-acting formulations of both methylphenidate and amphetamines are available, and a long-acting form is often used in conjunction with a short-acting form as needed.¹⁸

Addiction and drug-seeking behavior can develop but are unusual in those taking stimulants to treat narcolepsy.⁴⁹

Follow-up

Residual daytime sleepiness can be measured subjectively through the Epworth Sleepiness Scale on follow-up. If necessary, a maintenance-of-wakefulness test can provide an objective assessment of treatment efficacy.¹⁸

As narcolepsy is a chronic disorder, treatment should evolve with time. Most medications that treat narcolepsy are categorized by the US Food and Drug Administration as pregnancy category C, as we do not have adequate studies in human pregnancies to evaluate their effects. When a patient with narcolepsy becomes pregnant, she should be counseled about the risks and benefits of remaining on therapy. Treatment should balance the risks of sleepiness with the potential risks of remaining on medications.⁵⁰ In the elderly, as cardiovascular comorbidities tend to increase, the risks and benefits of therapy should be routinely reevaluated.

For cataplexy

Sodium oxybate,^51–53 the most potent anticataplectic drug, is the sodium salt of gamma hydroxybutyrate, a metabolite of gamma-aminobutyric acid. Sodium oxybate can be prescribed in the United States, Canada, and Europe. The American Academy of Sleep Medicine recommends sodium oxybate for cataplexy, daytime sleepiness, and disrupted sleep based on 3 level-1 studies and 2 level-4 studies.⁴⁶

Sodium oxybate increases slow-wave sleep, improves sleep continuity, and often helps to mitigate daytime sleepiness. Due to its short half-life, its administration is unusual: the first dose is taken before bedtime and the second dose 2.5 to 4 hours later. Some patients set an alarm clock to take the second dose, while others awaken spontaneously to take the second dose. Most patients find that with adherence to dosing and safety instructions, sodium oxybate can serve as a highly effective form of treatment of both excessive sleepiness and cataplexy and may reduce the need for stimulant-based therapies.

The most common adverse effects are nausea, mood swings, and enuresis. Occasionally, psychosis can result and limit use of the drug. Obstructive sleep apnea can also develop or worsen.⁵² Because of its high salt content, sodium oxybate should be used with caution in those with heart failure, hypertension, or renal impairment. Its relative, gamma hydroxybutyrate, causes rapid sedation and has been notorious for illegal use as a date rape drug.

In the United States, sodium oxybate is distributed only through a central pharmacy to mitigate potential abuse. Due to this system, the rates of diversion are extremely low, estimated in a postmarketing analysis to be 1 instance per 5,200 patients treated. In the same study, abuse and dependence were both rare as well, about 1 case for every 2,600 and 6,500 patients treated.^6,18,52,53

Antidepressants promote the action of norepinephrine and, to a lesser degree, serotonin, thereby suppressing REM sleep.

Venlafaxine, a serotonin-norepinephrine reuptake inhibitor, is often used as a first-line treatment for cataplexy. Selective serotonin reuptake inhibitors such as fluoxetine are also used with success. Tricyclic antidepressants such as protriptyline or clomipramine are extremely effective for cataplexy, but are rarely used due to their adverse effects.^2,47

FUTURE WORK

While our understanding of narcolepsy has advanced, there are still gaps in our knowledge of the disorder—namely, the specific trigger for the loss of hypocretin neurons in type 1 narcolepsy and the underlying pathophysiology of type 2.

A number of emerging therapies target the hypocretin system, including peptide replacement, neuronal transplant, and immunotherapy preventing hypocretin neuronal cell death.^50,54,55 Additional drugs designed to improve alertness that do not involve the hypocretin system are also being developed, including a histamine inverse agonist.^50,56 Sodium oxybate and modafinil, although currently approved for use in adults, are still off-label in pediatric practice. Studies of the safety and efficacy of these medications in children are needed.^7,57

Narcolepsy was originally described in the late 1800s by the French physician Jean-Baptiste-Edouard Gélineau, who reported the case of a wine merchant suffering from somnolence. In this first description, he coined the term narcolepsie by joining the Greek words narke (numbness or stupor) and lepsis (attack).¹

Since then, the disorder has been further characterized, and some insight into its biological underpinnings has been established. Importantly, treatments have improved and expanded, facilitating its management and thereby improving quality of life for those with the disorder.

This review focuses on clinically relevant features of the disorder and proposes management strategies.

CLINICAL FEATURES

Narcolepsy is characterized by instability of sleep-wake transitions.

Daytime sleepiness

Clinically, narcolepsy manifests with excessive daytime sleepiness that can be personally and socially disabling. Cataplexy, sleep paralysis, and hypnagogic or hypnopompic hallucinations can also be present,^2,3 but they are not necessary for diagnosis. In fact, a minority of patients with narcolepsy have all these symptoms.⁴ Narcolepsy is divided into type 1 (with cataplexy) and type 2 (without cataplexy).²

Sleepiness tends to be worse with inactivity, and sleep can often be irresistible. Sleep attacks can come on suddenly and may be brief enough to manifest as a lapse in consciousness.

Short naps tend to be refreshing. Rapid eye movement (REM) latency—the interval between falling asleep and the onset of the REM sleep—is short in narcolepsy, and since the REM stage is when dreaming occurs, naps often include dreaming. Therefore, when taking a history, it is worthwhile to ask patients whether they dream during naps; a yes answer supports the diagnosis of narcolepsy.⁵

In children, sleepiness can manifest in reduced concentration and behavioral issues.⁶ Napping after age 5 or 6 is considered abnormal and may reflect pathologic sleepiness.⁷

Cataplexy

Cataplexy—transient muscle weakness triggered by emotion—is a specific feature of narcolepsy type 1. It often begins in the facial muscles and can manifest with slackening of the jaw or brief dropping of the head. However, episodes can be more dramatic and, if the trunk and limb muscles are affected, can result in collapsing to the ground.

Cataplexy usually has its onset at about the same time as the sleepiness associated with narcolepsy, but it can arise even years later.⁸ Episodes can last from a few seconds to 2 minutes. Consciousness is always preserved. A range of emotions can trigger cataplexy, but typically the emotion is a positive one such as laughter or excitement.⁹ Deep tendon reflexes disappear in cataplexy, so checking reflexes during a witnessed episode can be clinically valuable.²

Cataplexy can worsen with stress and insufficient sleep, occasionally with “status cataplecticus,” in which repeated, persistent episodes of cataplexy occur over several hours.⁸ Status cataplecticus can be spontaneous or an effect of withdrawal from anticataplectic medications.²

Cataplexy is thought to represent intrusion of REM sleep and its associated muscle atonia during wakefulness.

Sleep paralysis, hallucinations

Sleep paralysis and hallucinations are other features of narcolepsy that reflect this REM dissociation from sleep.

Sleep paralysis occurs most commonly upon awakening, but sometimes just before sleep onset. In most cases, it is manifested by inability to move the limbs or speak, lasting several seconds or, in rare cases, minutes at a time. Sleep paralysis can be associated with a sensation of fear or suffocation, especially when initially experienced.⁸

Hypnopompic hallucinations, occurring upon awakening, are more common than hypnagogic hallucinations, which are experienced before falling asleep. The hallucinations are often vivid and usually visual, although other types of hallucinations are possible. Unlike those that occur in psychotic disorders, the hallucinations tend to be associated with preserved insight that they are not real.¹⁰

Notably, both sleep paralysis and hallucinations are nonspecific symptoms that are common in the general population.^8,11,12

Fragmented sleep

Although they are very sleepy, people with narcolepsy generally cannot stay asleep for very long. Their sleep tends to be extremely fragmented, and they often wake up several times a night.²

This sleep pattern reflects the inherent instability of sleep-wake transitions in narcolepsy. In fact, over a 24-hour period, adults with narcolepsy have a normal amount of sleep.¹³ In children, however, when narcolepsy first arises, the 24-hour sleep time can increase abruptly and can sometimes be associated with persistent cataplexy that can manifest as a clumsy gait.¹⁴

Weight gain, obstructive sleep apnea

Weight gain is common, particularly after symptom onset, and especially in children. As a result, obesity is a frequent comorbidity.¹⁵ Because obstructive sleep apnea can consequently develop, all patients with narcolepsy require screening for sleep-disordered breathing.

Other sleep disorders often accompany narcolepsy and are more common than in the general population.¹⁶ In a study incorporating both clinical and polysomnographic data of 100 patients with narcolepsy, insomnia was the most common comorbid sleep disorder, with a prevalence of 28%; others were REM sleep behavior disorder (24%), restless legs syndrome (24%), obstructive sleep apnea (21%), and non-REM parasomnias.¹⁷

 

 

PSYCHOSOCIAL CONSEQUENCES

Narcolepsy has significant psychosocial consequences. As a result of their symptoms, people with narcolepsy may not be able to meet academic or work-related demands.

Additionally, their risk of a motor vehicle accident is 3 to 4 times higher than in the general population, and more than one-third of patients have been in an accident due to sleepiness.¹⁸ There is some evidence to show that treatment eliminates this risk.¹⁹

Few systematic studies have examined mood disorders in narcolepsy. However, studies tend to show a higher prevalence of psychiatric disorders than in the general population, with depression and anxiety the most com-mon.^20,21

DIAGNOSIS IS OFTEN DELAYED

The prevalence of narcolepsy type 1 is between 25 and 100 per 100,000 people.²² In a Mayo Clinic study,²³ the incidence of narcolepsy type 1 was estimated to be 0.74 per 100,000 person-years. Epidemiologic data on narcolepsy type 2 are sparse, but patients with narcolepsy without cataplexy are thought to represent only 36% of all narcolepsy patients.²³

Diagnosis is often delayed, with the average time between the onset of symptoms and the diagnosis ranging from 8 to 22 years. With increasing awareness, the efficiency of the diagnostic process is improving, and this delay is expected to lessen accordingly.²⁴

Symptoms most commonly arise in the second decade; but the age at onset ranges significantly, between the first and fifth decades. Narcolepsy has a bimodal distribution in incidence, with the biggest peak at approximately age 15 and second smaller peak in the mid-30s. Some studies have suggested a slight male predominance.^23,25

DIAGNOSIS

History is key

The history should include specific questions about the hallmark features of narcolepsy, including cataplexy, sleep paralysis, and sleep-related hallucinations. For individual assessment of subjective sleepiness, the Epworth Sleepiness Scale or Pediatric Daytime Sleepiness Scale can be administered quickly in the office setting.^26,27

The Epworth score is calculated from the self-rated likelihood of falling asleep in 8 different situations, with possible scores of 0 (would never doze) to 3 (high chance of dozing) on each question, for a total possible score of 0 to 24. Normal total scores are between 0 and 10, while scores greater than 10 reflect pathologic sleepiness. Scores on the Epworth Sleepiness Scale in those with narcolepsy tend to reflect moderate to severe sleepiness, or at least 13, as opposed to patients with obstructive sleep apnea, whose scores commonly reflect milder sleepiness.²⁸

Testing with actigraphy and polysomnography

It is imperative to rule out insufficient sleep and other sleep disorders as a cause of daytime sleepiness. This can be done with a careful clinical history, actigraphy with sleep logs, and polysomnography.

In the 2 to 4 weeks before actigraphy and subsequent testing, all medications with alerting or sedating properties (including anti­depressants) should be tapered off to prevent influence on the results of the study.

Delayed sleep-phase disorder presents at a similar age as narcolepsy and can be associated with similar degrees of sleepiness. However, individuals with delayed sleep phase disorder have an inappropriately timed sleep-wake cycle so that there is a shift in their desired sleep onset and awakening times. It is common—prevalence estimates vary but average about 1% in the general population.²⁹

Insufficient sleep syndrome is even more common, especially in teenagers and young adults, with increasing family, social, and academic demands. Sleep needs vary across the life span. A teenager needs 8 to 10 hours of sleep per night, and a young adult needs 7 to 9 hours. A study of 1,285 high school students found that 10.4% were not getting enough sleep.³⁰

If actigraphy data suggest a circadian rhythm disorder or insufficient sleep that could explain the symptoms of sleepiness, then further testing should be halted and these specific issues should be addressed. In these cases, working with the patient toward maintaining a regular sleep-wake schedule with 7 to 8 hours of nightly sleep will often resolve symptoms.

If actigraphy demonstrates the patient is maintaining a regular sleep schedule and allowing adequate time for nightly sleep, the next step is polysomnography.

Polysomnography is performed to detect other disorders that can disrupt sleep, such as sleep-disordered breathing or periodic limb movement disorder.^2,5 In addition, polysomnography can provide assurance that adequate sleep was obtained prior to the next step in testing.

Multiple sleep latency test

If sufficient sleep is obtained on polysomnograpy (at least 6 hours for an adult) and no other sleep disorder is identified, a multiple sleep latency test is performed. A urine toxicology screen is typically performed on the day of the test to ensure that drugs are not affecting the results.

The multiple sleep latency test consists of 4 to 5 nap opportunities at 2-hour intervals in a quiet dark room conducive to sleep, during which both sleep and REM latency are recorded. The sleep latency of those with narcolepsy is significantly shortened, and the diagnosis of narcolepsy requires an average sleep latency of less than 8 minutes.

Given the propensity for REM sleep in narcolepsy, another essential feature for diagnosis is the sleep-onset REM period (SOREMP). A SOREMP is defined as a REM latency of less than 15 minutes. A diagnosis of narcolepsy re-quires a SOREMP in at least 2 of the naps in a multiple sleep latency test (or 1 nap if the shortened REM latency is seen during polysomnography).³¹

The multiple sleep latency test has an imperfect sensitivity, though, and should be repeated when there is a high suspicion of narcolepsy.^32–34 It is not completely specific either, and false-positive results occur. In fact, SOREMPs can be seen in the general population, particularly in those with a circadian rhythm disorder, insufficient sleep, or sleep-disordered breathing. Two or more SOREMPs in an multiple sleep latency test can be seen in a small proportion of the general population.³⁵ The results of a multiple sleep latency test should be interpreted in the clinical context.

Differential diagnosis

Narcolepsy type 1 is distinguished from type 2 by the presence of cataplexy. A cerebrospinal fluid hypocretin 1 level of 110 pg/mL or less, or less than one-third of the mean value obtained in normal individuals, can substitute for the multiple sleep latency test in diagnosing narcolepsy type 1.³¹ Currently, hypocretin testing is generally not performed in clinical practice, although it may become a routine part of the narcolepsy evaluation in the future.

Thus, according to the International Classification of Sleep Disorders, 3rd edition,³¹ the diagnosis of narcolepsy type 1 requires excessive daytime sleepiness for at least 3 months that cannot be explained by another sleep disorder, medical or neurologic disorder, mental disorder, medication use, or substance use disorder, and at least 1 of the following:

• Cataplexy and mean sleep latency of 8 minutes or less with at least 2 SOREMPs on multiple sleep latency testing (1 of which can be on the preceding night’s polysomography)
• Cerebrospinal fluid hypocretin 1 levels less than 110 pg/mL or one-third the baseline normal levels and mean sleep latency ≤ 8 minutes with ≥ 2 SOREMPs on multiple sleep latency testing.

Similarly, the diagnosis of narcolepsy type 2 requires excessive daytime sleepiness for at least 3 months that cannot be explained by another sleep disorder, medical or neurological disorder, mental disorder, medication use, or substance use disorder, plus:

• Mean sleep latency of 8 minutes or less with at least 2 SOREMPs on multiple sleep latency testing.

Idiopathic hypersomnia, another disorder of central hypersomnolence, is also characterized by disabling sleepiness. It is diagnostically differentiated from narcolepsy, as there are fewer than 2 SOREMPs. As opposed to narcolepsy, in which naps tend to be refreshing, even prolonged naps in idiopathic hypersomnia are often not helpful in restoring wakefulness. In idiopathic hypersomnia, sleep is usually not fragmented, and there are few nocturnal arousals. Sleep times can often be prolonged as well, whereas in narcolepsy total sleep time through the day may not be increased but is not consolidated.

Kleine-Levin syndrome is a rarer disorder of hypersomnia. It is episodic compared with the relatively persistent sleepiness in narcolepsy and idiopathic hypersomnia. Periods of hypersomnia occur intermittently for days to weeks and are accompanied by cognitive and behavioral changes including hyperphagia and hypersexuality.⁴

LINKED TO HYPOCRETIN DEFICIENCY

Over the past 2 decades, the underlying pathophysiology of narcolepsy type 1 has been better characterized.

Narcolepsy type 1 has been linked to a deficiency in hypocretin in the central nervous system.³⁶ Hypocretin (also known as orexin) is a hormone produced in the hypothalamus that acts on multiple brain regions and maintains alertness. For unclear reasons, hypothalamic neurons producing hypocretin are selectively reduced in narcolepsy type 1. Hypocretin also stabilizes wakefulness and inhibits REM sleep; therefore, hypocretin deficiency can lead to inappropriate intrusions of REM sleep onto wakefulness, leading to the hallmark features of narcolepsy—cataplexy, sleep-related hallucinations, and sleep paralysis.³⁷ According to one theory, cataplexy is triggered by emotional stimuli because of a pathway between the medial prefrontal cortex and the amygdala to the pons.³⁸

Cerebrospinal fluid levels of hypocretin in patients with narcolepsy type 2 tend to be normal, and the biologic underpinnings of narcolepsy type 2 remain mysterious. However, in the subgroup of those with narcolepsy type 2 in which hypocretin is low, many individuals go on to develop cataplexy, thereby evolving to narcolepsy type 1.³⁶

POSSIBLE AUTOIMMUNE BASIS

Narcolepsy is typically a sporadic disorder, although familial cases have been described. The risk of a parent with narcolepsy having a child who is affected is approximately 1%.⁵

Narcolepsy type 1 is strongly associated with HLA-DQB1*0602, with up to 95% of those affected having at least one allele.³⁹ Having 2 copies of the allele further increases the risk of developing narcolepsy.⁴⁰ However, this allele is far from specific for narcolepsy with cataplexy, as it occurs in 12% to 38% of the general population.⁴¹ Therefore, HLA typing currently has limited clinical utility. The exact cause is as yet unknown, but substantial literature proposes an autoimmune basis of the disorder, given the strong association with the HLA subtype.^42–44

After the 2009 H1N1 influenza pandemic, there was a significant increase in the incidence of narcolepsy with cataplexy, which again sparked interest in an autoimmune etiology underlying the disorder. Pandemrix, an H1N1 vaccine produced as a result of the 2009 pandemic, appeared to also be associated with an increase in the incidence of narcolepsy. An association with other upper respiratory infections has also been noted, further supporting a possible autoimmune basis.

A few studies have looked for serum autoantibodies involved in the pathogenesis of narcolepsy. Thus far, only one has been identified, an antibody to Tribbles homolog 2, found in 20% to 40% of those with new onset of nar-colepsy.^42–44

TREATMENTS FOR DAYTIME SLEEPINESS

As with many chronic disorders, the treatment of narcolepsy consists of symptomatic rather than curative management, which can be done through both pharmacologic and nonpharmacologic means.

Nondrug measures

Scheduled naps lasting 15 to 20 minutes can help improve alertness.⁴⁵ A consistent sleep schedule with good sleep hygiene, ensuring sufficient nightly sleep, is also important. In one study, the combination of scheduled naps and regular nocturnal sleep times reduced the level of daytime sleepiness and unintentional daytime sleep. Daytime naps were most helpful for those with the highest degree of daytime sleepiness.⁴⁵

Strategic use of caffeine can be helpful and can reduce dependence on pharmacologic treatment.

Screening should be performed routinely for other sleep disorders, such as sleep-disordered breathing, which should be treated if identified.^5,18 When being treated for other medical conditions, individuals with narcolepsy should avoid medications that can cause sedation, such as opiates or barbiturates; alcohol should be minimized or avoided.

Networking with other individuals with narcolepsy through support groups such as Narcolepsy Network can be valuable for learning coping skills and connecting with community resources. Psychological counseling for the patient, and sometimes the family, can also be useful. School-age children may need special accommodations such as schedule adjustments to allow for scheduled naps or frequent breaks to maintain alertness.

People with narcolepsy tend to function better in careers that do not require long periods of sitting, as sleepiness tends to be worse, but instead offer flexibility and require higher levels of activity that tend to combat sleepiness. They should not work as commercial drivers.¹⁸

 

 

Medications

While behavioral interventions in narcolepsy are vital, they are rarely sufficient, and drugs that promote daytime wakefulness are used as an adjunct (Table 2).⁴⁶

Realistic expectations should be established when starting, as some degree of residual sleepiness usually remains even with optimal medical therapy. Medications should be strategically scheduled to maximize alertness during necessary times such as at work or school or during driving. Patients should specifically be counseled to avoid driving if sleepy.^18,47

Modafinil is often used as a first-line therapy, given its favorable side-effect profile and low potential for abuse. Its pharmacologic action has been debated but it probably acts as a selective dopamine reuptake inhibitor. It is typically taken twice daily (upon waking and early afternoon) and is usually well tolerated.

Potential side effects include headache, nausea, dry mouth, anorexia, diarrhea, and, rarely, Stevens-Johnson syndrome. Cardiovascular side effects are minimal, making it a favorable choice in older patients.^18,48

A trial in 283 patients showed significantly lower levels of sleepiness in patients taking modafinil 200 mg or 400 mg than in a control group. Other trials have supported these findings and showed improved driving performance on modafinil.¹⁸

Notably, modafinil can increase the metabolism of oral contraceptives, thereby reducing their efficacy. Women of childbearing age should be warned about this interaction and should be transitioned to nonhormonal forms of contraception.^2,47

Armodafinil, a purified R-isomer of modafinil, has a longer half-life and requires only once-daily dosing.⁵

If modafinil or armodafinil fails to optimally manage daytime sleepiness, a traditional stimulant such as methylphenidate or an amphetamine is often used.

Methylphenidate and amphetamines primarily inhibit the reuptake and increase the release of the monoamines, mainly dopamine, and to a lesser degree serotonin and norepinephrine.

These drugs have more significant adverse effects that can involve the cardiovascular system, causing hypertension and arrhythmias. Anorexia, weight loss, and, particularly with high doses, psychosis can occur.⁴⁹

These drugs should be avoided in patients with a history of significant cardiovascular disease. Before starting stimulant therapy, a thorough cardiovascular examination should be done, often including electrocardiography to ensure there is no baseline arrhythmia.

Patients on these medications should be followed closely to ensure that blood pressure, pulse, and weight are not negatively affected.^18,50 Addiction and tolerance can develop with these drugs, and follow-up should include assessment for dependence. Some states may require prescription drug monitoring to ensure the drugs are not being abused or diverted.

Short- and long-acting formulations of both methylphenidate and amphetamines are available, and a long-acting form is often used in conjunction with a short-acting form as needed.¹⁸

Addiction and drug-seeking behavior can develop but are unusual in those taking stimulants to treat narcolepsy.⁴⁹

Follow-up

Residual daytime sleepiness can be measured subjectively through the Epworth Sleepiness Scale on follow-up. If necessary, a maintenance-of-wakefulness test can provide an objective assessment of treatment efficacy.¹⁸

As narcolepsy is a chronic disorder, treatment should evolve with time. Most medications that treat narcolepsy are categorized by the US Food and Drug Administration as pregnancy category C, as we do not have adequate studies in human pregnancies to evaluate their effects. When a patient with narcolepsy becomes pregnant, she should be counseled about the risks and benefits of remaining on therapy. Treatment should balance the risks of sleepiness with the potential risks of remaining on medications.⁵⁰ In the elderly, as cardiovascular comorbidities tend to increase, the risks and benefits of therapy should be routinely reevaluated.

For cataplexy

Sodium oxybate,^51–53 the most potent anticataplectic drug, is the sodium salt of gamma hydroxybutyrate, a metabolite of gamma-aminobutyric acid. Sodium oxybate can be prescribed in the United States, Canada, and Europe. The American Academy of Sleep Medicine recommends sodium oxybate for cataplexy, daytime sleepiness, and disrupted sleep based on 3 level-1 studies and 2 level-4 studies.⁴⁶

Sodium oxybate increases slow-wave sleep, improves sleep continuity, and often helps to mitigate daytime sleepiness. Due to its short half-life, its administration is unusual: the first dose is taken before bedtime and the second dose 2.5 to 4 hours later. Some patients set an alarm clock to take the second dose, while others awaken spontaneously to take the second dose. Most patients find that with adherence to dosing and safety instructions, sodium oxybate can serve as a highly effective form of treatment of both excessive sleepiness and cataplexy and may reduce the need for stimulant-based therapies.

The most common adverse effects are nausea, mood swings, and enuresis. Occasionally, psychosis can result and limit use of the drug. Obstructive sleep apnea can also develop or worsen.⁵² Because of its high salt content, sodium oxybate should be used with caution in those with heart failure, hypertension, or renal impairment. Its relative, gamma hydroxybutyrate, causes rapid sedation and has been notorious for illegal use as a date rape drug.

In the United States, sodium oxybate is distributed only through a central pharmacy to mitigate potential abuse. Due to this system, the rates of diversion are extremely low, estimated in a postmarketing analysis to be 1 instance per 5,200 patients treated. In the same study, abuse and dependence were both rare as well, about 1 case for every 2,600 and 6,500 patients treated.^6,18,52,53

Antidepressants promote the action of norepinephrine and, to a lesser degree, serotonin, thereby suppressing REM sleep.

Venlafaxine, a serotonin-norepinephrine reuptake inhibitor, is often used as a first-line treatment for cataplexy. Selective serotonin reuptake inhibitors such as fluoxetine are also used with success. Tricyclic antidepressants such as protriptyline or clomipramine are extremely effective for cataplexy, but are rarely used due to their adverse effects.^2,47

FUTURE WORK

While our understanding of narcolepsy has advanced, there are still gaps in our knowledge of the disorder—namely, the specific trigger for the loss of hypocretin neurons in type 1 narcolepsy and the underlying pathophysiology of type 2.

A number of emerging therapies target the hypocretin system, including peptide replacement, neuronal transplant, and immunotherapy preventing hypocretin neuronal cell death.^50,54,55 Additional drugs designed to improve alertness that do not involve the hypocretin system are also being developed, including a histamine inverse agonist.^50,56 Sodium oxybate and modafinil, although currently approved for use in adults, are still off-label in pediatric practice. Studies of the safety and efficacy of these medications in children are needed.^7,57

Narcolepsy was originally described in the late 1800s by the French physician Jean-Baptiste-Edouard Gélineau, who reported the case of a wine merchant suffering from somnolence. In this first description, he coined the term narcolepsie by joining the Greek words narke (numbness or stupor) and lepsis (attack).¹

Since then, the disorder has been further characterized, and some insight into its biological underpinnings has been established. Importantly, treatments have improved and expanded, facilitating its management and thereby improving quality of life for those with the disorder.

This review focuses on clinically relevant features of the disorder and proposes management strategies.

CLINICAL FEATURES

Narcolepsy is characterized by instability of sleep-wake transitions.

Daytime sleepiness

Clinically, narcolepsy manifests with excessive daytime sleepiness that can be personally and socially disabling. Cataplexy, sleep paralysis, and hypnagogic or hypnopompic hallucinations can also be present,^2,3 but they are not necessary for diagnosis. In fact, a minority of patients with narcolepsy have all these symptoms.⁴ Narcolepsy is divided into type 1 (with cataplexy) and type 2 (without cataplexy).²

Sleepiness tends to be worse with inactivity, and sleep can often be irresistible. Sleep attacks can come on suddenly and may be brief enough to manifest as a lapse in consciousness.

Short naps tend to be refreshing. Rapid eye movement (REM) latency—the interval between falling asleep and the onset of the REM sleep—is short in narcolepsy, and since the REM stage is when dreaming occurs, naps often include dreaming. Therefore, when taking a history, it is worthwhile to ask patients whether they dream during naps; a yes answer supports the diagnosis of narcolepsy.⁵

In children, sleepiness can manifest in reduced concentration and behavioral issues.⁶ Napping after age 5 or 6 is considered abnormal and may reflect pathologic sleepiness.⁷

Cataplexy

Cataplexy—transient muscle weakness triggered by emotion—is a specific feature of narcolepsy type 1. It often begins in the facial muscles and can manifest with slackening of the jaw or brief dropping of the head. However, episodes can be more dramatic and, if the trunk and limb muscles are affected, can result in collapsing to the ground.

Cataplexy usually has its onset at about the same time as the sleepiness associated with narcolepsy, but it can arise even years later.⁸ Episodes can last from a few seconds to 2 minutes. Consciousness is always preserved. A range of emotions can trigger cataplexy, but typically the emotion is a positive one such as laughter or excitement.⁹ Deep tendon reflexes disappear in cataplexy, so checking reflexes during a witnessed episode can be clinically valuable.²

Cataplexy can worsen with stress and insufficient sleep, occasionally with “status cataplecticus,” in which repeated, persistent episodes of cataplexy occur over several hours.⁸ Status cataplecticus can be spontaneous or an effect of withdrawal from anticataplectic medications.²

Cataplexy is thought to represent intrusion of REM sleep and its associated muscle atonia during wakefulness.

Sleep paralysis, hallucinations

Sleep paralysis and hallucinations are other features of narcolepsy that reflect this REM dissociation from sleep.

Sleep paralysis occurs most commonly upon awakening, but sometimes just before sleep onset. In most cases, it is manifested by inability to move the limbs or speak, lasting several seconds or, in rare cases, minutes at a time. Sleep paralysis can be associated with a sensation of fear or suffocation, especially when initially experienced.⁸

Hypnopompic hallucinations, occurring upon awakening, are more common than hypnagogic hallucinations, which are experienced before falling asleep. The hallucinations are often vivid and usually visual, although other types of hallucinations are possible. Unlike those that occur in psychotic disorders, the hallucinations tend to be associated with preserved insight that they are not real.¹⁰

Notably, both sleep paralysis and hallucinations are nonspecific symptoms that are common in the general population.^8,11,12

Fragmented sleep

Although they are very sleepy, people with narcolepsy generally cannot stay asleep for very long. Their sleep tends to be extremely fragmented, and they often wake up several times a night.²

This sleep pattern reflects the inherent instability of sleep-wake transitions in narcolepsy. In fact, over a 24-hour period, adults with narcolepsy have a normal amount of sleep.¹³ In children, however, when narcolepsy first arises, the 24-hour sleep time can increase abruptly and can sometimes be associated with persistent cataplexy that can manifest as a clumsy gait.¹⁴

Weight gain, obstructive sleep apnea

Weight gain is common, particularly after symptom onset, and especially in children. As a result, obesity is a frequent comorbidity.¹⁵ Because obstructive sleep apnea can consequently develop, all patients with narcolepsy require screening for sleep-disordered breathing.

Other sleep disorders often accompany narcolepsy and are more common than in the general population.¹⁶ In a study incorporating both clinical and polysomnographic data of 100 patients with narcolepsy, insomnia was the most common comorbid sleep disorder, with a prevalence of 28%; others were REM sleep behavior disorder (24%), restless legs syndrome (24%), obstructive sleep apnea (21%), and non-REM parasomnias.¹⁷

 

 

PSYCHOSOCIAL CONSEQUENCES

Narcolepsy has significant psychosocial consequences. As a result of their symptoms, people with narcolepsy may not be able to meet academic or work-related demands.

Additionally, their risk of a motor vehicle accident is 3 to 4 times higher than in the general population, and more than one-third of patients have been in an accident due to sleepiness.¹⁸ There is some evidence to show that treatment eliminates this risk.¹⁹

Few systematic studies have examined mood disorders in narcolepsy. However, studies tend to show a higher prevalence of psychiatric disorders than in the general population, with depression and anxiety the most com-mon.^20,21

DIAGNOSIS IS OFTEN DELAYED

The prevalence of narcolepsy type 1 is between 25 and 100 per 100,000 people.²² In a Mayo Clinic study,²³ the incidence of narcolepsy type 1 was estimated to be 0.74 per 100,000 person-years. Epidemiologic data on narcolepsy type 2 are sparse, but patients with narcolepsy without cataplexy are thought to represent only 36% of all narcolepsy patients.²³

Diagnosis is often delayed, with the average time between the onset of symptoms and the diagnosis ranging from 8 to 22 years. With increasing awareness, the efficiency of the diagnostic process is improving, and this delay is expected to lessen accordingly.²⁴

Symptoms most commonly arise in the second decade; but the age at onset ranges significantly, between the first and fifth decades. Narcolepsy has a bimodal distribution in incidence, with the biggest peak at approximately age 15 and second smaller peak in the mid-30s. Some studies have suggested a slight male predominance.^23,25

DIAGNOSIS

History is key

The history should include specific questions about the hallmark features of narcolepsy, including cataplexy, sleep paralysis, and sleep-related hallucinations. For individual assessment of subjective sleepiness, the Epworth Sleepiness Scale or Pediatric Daytime Sleepiness Scale can be administered quickly in the office setting.^26,27

The Epworth score is calculated from the self-rated likelihood of falling asleep in 8 different situations, with possible scores of 0 (would never doze) to 3 (high chance of dozing) on each question, for a total possible score of 0 to 24. Normal total scores are between 0 and 10, while scores greater than 10 reflect pathologic sleepiness. Scores on the Epworth Sleepiness Scale in those with narcolepsy tend to reflect moderate to severe sleepiness, or at least 13, as opposed to patients with obstructive sleep apnea, whose scores commonly reflect milder sleepiness.²⁸

Testing with actigraphy and polysomnography

It is imperative to rule out insufficient sleep and other sleep disorders as a cause of daytime sleepiness. This can be done with a careful clinical history, actigraphy with sleep logs, and polysomnography.

In the 2 to 4 weeks before actigraphy and subsequent testing, all medications with alerting or sedating properties (including anti­depressants) should be tapered off to prevent influence on the results of the study.

Delayed sleep-phase disorder presents at a similar age as narcolepsy and can be associated with similar degrees of sleepiness. However, individuals with delayed sleep phase disorder have an inappropriately timed sleep-wake cycle so that there is a shift in their desired sleep onset and awakening times. It is common—prevalence estimates vary but average about 1% in the general population.²⁹

Insufficient sleep syndrome is even more common, especially in teenagers and young adults, with increasing family, social, and academic demands. Sleep needs vary across the life span. A teenager needs 8 to 10 hours of sleep per night, and a young adult needs 7 to 9 hours. A study of 1,285 high school students found that 10.4% were not getting enough sleep.³⁰

If actigraphy data suggest a circadian rhythm disorder or insufficient sleep that could explain the symptoms of sleepiness, then further testing should be halted and these specific issues should be addressed. In these cases, working with the patient toward maintaining a regular sleep-wake schedule with 7 to 8 hours of nightly sleep will often resolve symptoms.

If actigraphy demonstrates the patient is maintaining a regular sleep schedule and allowing adequate time for nightly sleep, the next step is polysomnography.

Polysomnography is performed to detect other disorders that can disrupt sleep, such as sleep-disordered breathing or periodic limb movement disorder.^2,5 In addition, polysomnography can provide assurance that adequate sleep was obtained prior to the next step in testing.

Multiple sleep latency test

If sufficient sleep is obtained on polysomnograpy (at least 6 hours for an adult) and no other sleep disorder is identified, a multiple sleep latency test is performed. A urine toxicology screen is typically performed on the day of the test to ensure that drugs are not affecting the results.

The multiple sleep latency test consists of 4 to 5 nap opportunities at 2-hour intervals in a quiet dark room conducive to sleep, during which both sleep and REM latency are recorded. The sleep latency of those with narcolepsy is significantly shortened, and the diagnosis of narcolepsy requires an average sleep latency of less than 8 minutes.

Given the propensity for REM sleep in narcolepsy, another essential feature for diagnosis is the sleep-onset REM period (SOREMP). A SOREMP is defined as a REM latency of less than 15 minutes. A diagnosis of narcolepsy re-quires a SOREMP in at least 2 of the naps in a multiple sleep latency test (or 1 nap if the shortened REM latency is seen during polysomnography).³¹

The multiple sleep latency test has an imperfect sensitivity, though, and should be repeated when there is a high suspicion of narcolepsy.^32–34 It is not completely specific either, and false-positive results occur. In fact, SOREMPs can be seen in the general population, particularly in those with a circadian rhythm disorder, insufficient sleep, or sleep-disordered breathing. Two or more SOREMPs in an multiple sleep latency test can be seen in a small proportion of the general population.³⁵ The results of a multiple sleep latency test should be interpreted in the clinical context.

Differential diagnosis

Narcolepsy type 1 is distinguished from type 2 by the presence of cataplexy. A cerebrospinal fluid hypocretin 1 level of 110 pg/mL or less, or less than one-third of the mean value obtained in normal individuals, can substitute for the multiple sleep latency test in diagnosing narcolepsy type 1.³¹ Currently, hypocretin testing is generally not performed in clinical practice, although it may become a routine part of the narcolepsy evaluation in the future.

Thus, according to the International Classification of Sleep Disorders, 3rd edition,³¹ the diagnosis of narcolepsy type 1 requires excessive daytime sleepiness for at least 3 months that cannot be explained by another sleep disorder, medical or neurologic disorder, mental disorder, medication use, or substance use disorder, and at least 1 of the following:

• Cataplexy and mean sleep latency of 8 minutes or less with at least 2 SOREMPs on multiple sleep latency testing (1 of which can be on the preceding night’s polysomography)
• Cerebrospinal fluid hypocretin 1 levels less than 110 pg/mL or one-third the baseline normal levels and mean sleep latency ≤ 8 minutes with ≥ 2 SOREMPs on multiple sleep latency testing.

Similarly, the diagnosis of narcolepsy type 2 requires excessive daytime sleepiness for at least 3 months that cannot be explained by another sleep disorder, medical or neurological disorder, mental disorder, medication use, or substance use disorder, plus:

• Mean sleep latency of 8 minutes or less with at least 2 SOREMPs on multiple sleep latency testing.

Idiopathic hypersomnia, another disorder of central hypersomnolence, is also characterized by disabling sleepiness. It is diagnostically differentiated from narcolepsy, as there are fewer than 2 SOREMPs. As opposed to narcolepsy, in which naps tend to be refreshing, even prolonged naps in idiopathic hypersomnia are often not helpful in restoring wakefulness. In idiopathic hypersomnia, sleep is usually not fragmented, and there are few nocturnal arousals. Sleep times can often be prolonged as well, whereas in narcolepsy total sleep time through the day may not be increased but is not consolidated.

Kleine-Levin syndrome is a rarer disorder of hypersomnia. It is episodic compared with the relatively persistent sleepiness in narcolepsy and idiopathic hypersomnia. Periods of hypersomnia occur intermittently for days to weeks and are accompanied by cognitive and behavioral changes including hyperphagia and hypersexuality.⁴

LINKED TO HYPOCRETIN DEFICIENCY

Over the past 2 decades, the underlying pathophysiology of narcolepsy type 1 has been better characterized.

Narcolepsy type 1 has been linked to a deficiency in hypocretin in the central nervous system.³⁶ Hypocretin (also known as orexin) is a hormone produced in the hypothalamus that acts on multiple brain regions and maintains alertness. For unclear reasons, hypothalamic neurons producing hypocretin are selectively reduced in narcolepsy type 1. Hypocretin also stabilizes wakefulness and inhibits REM sleep; therefore, hypocretin deficiency can lead to inappropriate intrusions of REM sleep onto wakefulness, leading to the hallmark features of narcolepsy—cataplexy, sleep-related hallucinations, and sleep paralysis.³⁷ According to one theory, cataplexy is triggered by emotional stimuli because of a pathway between the medial prefrontal cortex and the amygdala to the pons.³⁸

Cerebrospinal fluid levels of hypocretin in patients with narcolepsy type 2 tend to be normal, and the biologic underpinnings of narcolepsy type 2 remain mysterious. However, in the subgroup of those with narcolepsy type 2 in which hypocretin is low, many individuals go on to develop cataplexy, thereby evolving to narcolepsy type 1.³⁶

POSSIBLE AUTOIMMUNE BASIS

Narcolepsy is typically a sporadic disorder, although familial cases have been described. The risk of a parent with narcolepsy having a child who is affected is approximately 1%.⁵

Narcolepsy type 1 is strongly associated with HLA-DQB1*0602, with up to 95% of those affected having at least one allele.³⁹ Having 2 copies of the allele further increases the risk of developing narcolepsy.⁴⁰ However, this allele is far from specific for narcolepsy with cataplexy, as it occurs in 12% to 38% of the general population.⁴¹ Therefore, HLA typing currently has limited clinical utility. The exact cause is as yet unknown, but substantial literature proposes an autoimmune basis of the disorder, given the strong association with the HLA subtype.^42–44

After the 2009 H1N1 influenza pandemic, there was a significant increase in the incidence of narcolepsy with cataplexy, which again sparked interest in an autoimmune etiology underlying the disorder. Pandemrix, an H1N1 vaccine produced as a result of the 2009 pandemic, appeared to also be associated with an increase in the incidence of narcolepsy. An association with other upper respiratory infections has also been noted, further supporting a possible autoimmune basis.

A few studies have looked for serum autoantibodies involved in the pathogenesis of narcolepsy. Thus far, only one has been identified, an antibody to Tribbles homolog 2, found in 20% to 40% of those with new onset of nar-colepsy.^42–44

TREATMENTS FOR DAYTIME SLEEPINESS

As with many chronic disorders, the treatment of narcolepsy consists of symptomatic rather than curative management, which can be done through both pharmacologic and nonpharmacologic means.

Nondrug measures

Scheduled naps lasting 15 to 20 minutes can help improve alertness.⁴⁵ A consistent sleep schedule with good sleep hygiene, ensuring sufficient nightly sleep, is also important. In one study, the combination of scheduled naps and regular nocturnal sleep times reduced the level of daytime sleepiness and unintentional daytime sleep. Daytime naps were most helpful for those with the highest degree of daytime sleepiness.⁴⁵

Strategic use of caffeine can be helpful and can reduce dependence on pharmacologic treatment.

Screening should be performed routinely for other sleep disorders, such as sleep-disordered breathing, which should be treated if identified.^5,18 When being treated for other medical conditions, individuals with narcolepsy should avoid medications that can cause sedation, such as opiates or barbiturates; alcohol should be minimized or avoided.

Networking with other individuals with narcolepsy through support groups such as Narcolepsy Network can be valuable for learning coping skills and connecting with community resources. Psychological counseling for the patient, and sometimes the family, can also be useful. School-age children may need special accommodations such as schedule adjustments to allow for scheduled naps or frequent breaks to maintain alertness.

People with narcolepsy tend to function better in careers that do not require long periods of sitting, as sleepiness tends to be worse, but instead offer flexibility and require higher levels of activity that tend to combat sleepiness. They should not work as commercial drivers.¹⁸

 

 

Medications

While behavioral interventions in narcolepsy are vital, they are rarely sufficient, and drugs that promote daytime wakefulness are used as an adjunct (Table 2).⁴⁶

Realistic expectations should be established when starting, as some degree of residual sleepiness usually remains even with optimal medical therapy. Medications should be strategically scheduled to maximize alertness during necessary times such as at work or school or during driving. Patients should specifically be counseled to avoid driving if sleepy.^18,47

Modafinil is often used as a first-line therapy, given its favorable side-effect profile and low potential for abuse. Its pharmacologic action has been debated but it probably acts as a selective dopamine reuptake inhibitor. It is typically taken twice daily (upon waking and early afternoon) and is usually well tolerated.

Potential side effects include headache, nausea, dry mouth, anorexia, diarrhea, and, rarely, Stevens-Johnson syndrome. Cardiovascular side effects are minimal, making it a favorable choice in older patients.^18,48

A trial in 283 patients showed significantly lower levels of sleepiness in patients taking modafinil 200 mg or 400 mg than in a control group. Other trials have supported these findings and showed improved driving performance on modafinil.¹⁸

Notably, modafinil can increase the metabolism of oral contraceptives, thereby reducing their efficacy. Women of childbearing age should be warned about this interaction and should be transitioned to nonhormonal forms of contraception.^2,47

Armodafinil, a purified R-isomer of modafinil, has a longer half-life and requires only once-daily dosing.⁵

If modafinil or armodafinil fails to optimally manage daytime sleepiness, a traditional stimulant such as methylphenidate or an amphetamine is often used.

Methylphenidate and amphetamines primarily inhibit the reuptake and increase the release of the monoamines, mainly dopamine, and to a lesser degree serotonin and norepinephrine.

These drugs have more significant adverse effects that can involve the cardiovascular system, causing hypertension and arrhythmias. Anorexia, weight loss, and, particularly with high doses, psychosis can occur.⁴⁹

These drugs should be avoided in patients with a history of significant cardiovascular disease. Before starting stimulant therapy, a thorough cardiovascular examination should be done, often including electrocardiography to ensure there is no baseline arrhythmia.

Patients on these medications should be followed closely to ensure that blood pressure, pulse, and weight are not negatively affected.^18,50 Addiction and tolerance can develop with these drugs, and follow-up should include assessment for dependence. Some states may require prescription drug monitoring to ensure the drugs are not being abused or diverted.

Short- and long-acting formulations of both methylphenidate and amphetamines are available, and a long-acting form is often used in conjunction with a short-acting form as needed.¹⁸

Addiction and drug-seeking behavior can develop but are unusual in those taking stimulants to treat narcolepsy.⁴⁹

Follow-up

Residual daytime sleepiness can be measured subjectively through the Epworth Sleepiness Scale on follow-up. If necessary, a maintenance-of-wakefulness test can provide an objective assessment of treatment efficacy.¹⁸

As narcolepsy is a chronic disorder, treatment should evolve with time. Most medications that treat narcolepsy are categorized by the US Food and Drug Administration as pregnancy category C, as we do not have adequate studies in human pregnancies to evaluate their effects. When a patient with narcolepsy becomes pregnant, she should be counseled about the risks and benefits of remaining on therapy. Treatment should balance the risks of sleepiness with the potential risks of remaining on medications.⁵⁰ In the elderly, as cardiovascular comorbidities tend to increase, the risks and benefits of therapy should be routinely reevaluated.

For cataplexy

Sodium oxybate,^51–53 the most potent anticataplectic drug, is the sodium salt of gamma hydroxybutyrate, a metabolite of gamma-aminobutyric acid. Sodium oxybate can be prescribed in the United States, Canada, and Europe. The American Academy of Sleep Medicine recommends sodium oxybate for cataplexy, daytime sleepiness, and disrupted sleep based on 3 level-1 studies and 2 level-4 studies.⁴⁶

Sodium oxybate increases slow-wave sleep, improves sleep continuity, and often helps to mitigate daytime sleepiness. Due to its short half-life, its administration is unusual: the first dose is taken before bedtime and the second dose 2.5 to 4 hours later. Some patients set an alarm clock to take the second dose, while others awaken spontaneously to take the second dose. Most patients find that with adherence to dosing and safety instructions, sodium oxybate can serve as a highly effective form of treatment of both excessive sleepiness and cataplexy and may reduce the need for stimulant-based therapies.

The most common adverse effects are nausea, mood swings, and enuresis. Occasionally, psychosis can result and limit use of the drug. Obstructive sleep apnea can also develop or worsen.⁵² Because of its high salt content, sodium oxybate should be used with caution in those with heart failure, hypertension, or renal impairment. Its relative, gamma hydroxybutyrate, causes rapid sedation and has been notorious for illegal use as a date rape drug.

In the United States, sodium oxybate is distributed only through a central pharmacy to mitigate potential abuse. Due to this system, the rates of diversion are extremely low, estimated in a postmarketing analysis to be 1 instance per 5,200 patients treated. In the same study, abuse and dependence were both rare as well, about 1 case for every 2,600 and 6,500 patients treated.^6,18,52,53

Antidepressants promote the action of norepinephrine and, to a lesser degree, serotonin, thereby suppressing REM sleep.

Venlafaxine, a serotonin-norepinephrine reuptake inhibitor, is often used as a first-line treatment for cataplexy. Selective serotonin reuptake inhibitors such as fluoxetine are also used with success. Tricyclic antidepressants such as protriptyline or clomipramine are extremely effective for cataplexy, but are rarely used due to their adverse effects.^2,47

FUTURE WORK

While our understanding of narcolepsy has advanced, there are still gaps in our knowledge of the disorder—namely, the specific trigger for the loss of hypocretin neurons in type 1 narcolepsy and the underlying pathophysiology of type 2.

A number of emerging therapies target the hypocretin system, including peptide replacement, neuronal transplant, and immunotherapy preventing hypocretin neuronal cell death.^50,54,55 Additional drugs designed to improve alertness that do not involve the hypocretin system are also being developed, including a histamine inverse agonist.^50,56 Sodium oxybate and modafinil, although currently approved for use in adults, are still off-label in pediatric practice. Studies of the safety and efficacy of these medications in children are needed.^7,57