James Taylor, PharmD, CDE

For obese patients with type 2 diabetes (T2D) who do not tolerate other diabetes medications or for patients who need only a minimal reduction in glycosylated hemoglobin (A1c) to reach goal, bromocriptine may be a therapeutic agent to consider. Approved by the US Food and Drug Administration (FDA) in 2009, Cycloset is a quick-release formulation of bromocriptine mesylate, an ergot alkaloid dopamine D2 receptor agonist that has a faster onset of action than the customary formulation, Parlodel, which has been used to treat Parkinson’s disease, acromegaly, and hyperprolactinemia.¹ In addition to its modest benefit in improving glycemic control, Cyclocet avoids undesirable side effects such as hypoglycemia and weight gain.

VeroScience, LLC holds the new drug application and related technologies for Cycloset and partnered with Santarus, Inc. and S2 Therapeutics to market it in September 2010.²

Bromocriptine’s likely mechanism of action

Although its exact mechanism of action is unclear, bromocriptine does not stimulate insulin release, reduce hepatic glucose production, increase glucose transporter production, or increase or mimic glucagon-like peptide-1 activity as other T2D agents do.³ Its contribution to glycemic control in T2D has been hypothesized to be due to adjustments in the neural control of seasonal and diurnal patterns of food intake and nutrient storage.⁴

Early hunter-gatherers and farmers are thought to have benefited from a “thrifty genotype” that favored fat deposition when food was seasonally abundant. With food in western society available year-round and often energy dense in both fat and carbohydrates, this same gene may lead to obesity and noninsulin-dependent diabetes mellitus.⁵

The hypothesis assumes that circadian rhythm, photo- periodism, and seasonal factors play a role in insulin resistance, hepatic gluconeogenesis, and weight gain. In vertebrates, the neuroendocrine system plays an important role in synchronizing the animal with cyclic environmental changes. The hypothalamic suprachiasmatic nucleus (SCN) is known as the circadian pacemaker that maintains this rhythm. Oscillations in the SCN occur due to external cues such as changes in light or temperature. Circadian dopaminergic and serotonergic activities are likely responsible for modifying such oscillations, and neurotransmitters have been shown to regulate the dramatic seasonal alterations in body weight and body composition of all vertebrate classes.⁶ Bromocriptine can reverse metabolic alterations associated with insulin resistance and obesity by resetting central (hypothalamic) circadian organization of monoamine neuronal activities.⁷

Proven anti-T2D effects. When administered systemically or into the cerebral ventricle at first light, bromocriptine prevents or reverses seasonal fattening, insulin resistance, and hyperinsulinemia, and it decreases endogenous (hepatic) glucose production in mammals.^8-11 Bromocriptine also decreases both fasting and postprandial triglyceride and free fatty acid levels.¹

Clinical trials show modest benefit

Although bromocriptine has been studied since 1980 for its effects on hyperglycemia in T2D,¹² trials leading to the approval of the drug for clinical use with T2D have only been completed within the last 15 years. Randomized controlled trials of varying sizes and lasting from 6 to 52 weeks have shown absolute A1c reductions from 0.1% to 0.6%.^1,12-16 Compared with placebo, A1c reductions have ranged from 0.4% to 1.2% with monotherapy and in combination with other antidiabetes medications.^1,13-16

The manufacturer assessed bromocriptine in 4 studies involving patients with T2D. In all 4 studies, the bromocriptine dose was titrated to a maximum of 4.8 mg/d.¹⁶

One study involved 159 overweight subjects who were not meeting glycemic goals.¹⁶ Patients received either placebo or bromocriptine for 24 weeks in addition to diet and exercise. Mean baseline A1c was 9.0% in the bromocriptine group and 8.8% in the placebo group. After 24 weeks, A1c was reduced by 0.1% in the treatment group and increased by 0.3% in the placebo group. Mean fasting glucose was 215 mg/dL at baseline in the treatment group and was unchanged after 24 weeks. In the placebo group, fasting glucose increased from 205 to 228 mg/dL during the study. Weight increased by 0.2 kg in the treatment group and by 0.5 kg in the placebo group.

The next two 24-week manufacturer studies used similar designs to compare the addition of either bromocriptine or placebo to existing sulfonylurea therapy in patients with uncontrolled T2D.¹⁶ One study assigned 122 patients to bromocriptine and 127 to placebo. The bromocriptine group demonstrated mean reductions of 0.4% in A1c and 3 mg/dL in fasting glucose. In the placebo group, A1c increased by 0.3% and fasting glucose rose by 23 mg/dL.

The other study assigned 122 patients to bromocriptine and 123 to placebo. Adding bromocriptine reduced A1c, on average, by 0.1% and fasting glucose by 10 mg/dL. In the placebo group, A1c increased by 0.4% and fasting glucose increased by 28 mg/dL. All of these results were statistically significant.

The last manufacturer-reported study evaluated the addition of bromocriptine to other diabetes treatments (diet or up to 2 anti-diabetes medications).¹³ While the primary intent of this study was to evaluate safety, it also assessed efficacy. This was a 52-week, randomized placebo-controlled trial involving 3095 patients.

Overall, after 24 weeks there was no change in A1c levels after adding bromocriptine. However, most patients in this study were already at goal (A1c <7.0%). A subgroup analysis of those with an A1c level <7.5% while taking other agents did show some improvement with the addition of bromocriptine. Adding bromocriptine to metformin and a sulfonylurea significantly reduced A1c by 0.5%, on average. Similar results were seen in those who received other combinations of diabetes medications. After 52 weeks, 25% of those receiving bromocriptine who originally had an A1c level >7.5% achieved an A1c level <7.0%. Of the patients who received placebo, 9% obtained an A1c level <7%.

In a 24-week study, bromocriptine titrated up to 4.8 mg/d was given to patients either on no other diabetes medication or on a sulfonylurea.¹ In individuals not on any current treatment, A1c decreased by 0.2% in those who received bromocriptine. In patients already on a sulfonylurea, A1c declined by 0.1%. A1c increased by 0.3% in those receiving placebo.

Bromocriptine most effective when taken with food

When bromocriptine is taken orally, 65% to 95% of the dose is absorbed; however, only 7% reaches systemic circulation due to extensive hepatic extraction and first-pass metabolism.¹⁷ Bioavailability increases by 55% to 65% when the drug is taken with food, which is how it should be administered. The time to maximum plasma concentration is within an hour. With a high-fat meal, however, the time increases to 90 to 120 minutes. Bromocriptine is highly protein bound (90%-96%) and is metabolized extensively in the gastrointestinal (GI) tract and liver.¹⁷ CYP3A4 is the major metabolic pathway.^1,18 Most excretion of bromocriptine is through bile, with approximately 2% to 6% of an oral dose eliminated via urine. The elimination half-life is approximately 6 hours.^17,18

Dosing is once a day in the morning

Clinical trials investigating the use of bromocriptine in diabetes used doses ranging from 1.6 to 4.8 mg/d.^13-16,19 The FDA-approved dose range is 1.6 to 4.8 mg administered once daily with food, within 2 hours of waking in the morning.¹⁶ In healthy individuals, central nervous system (CNS) dopaminergic activity peaks in the early morning. Thus, morning dosing attempts to mimic dopaminergic activity and circadian rhythms in healthy lean individuals.⁶

Titrate to maximum dose. The product is available in a 0.8-mg tablet (TABLE). Titration to the maximum dose is recommended to reduce GI adverse effects, particularly nausea. Start treatment with 1 tablet (0.8 mg) and increase the dose by 1 tablet per week until the patient reaches a maximum tolerated dose or the maximum allowable daily dose of 4.8 mg (6 tablets).

Precautions with renal or hepatic impairment. No pharmacokinetic studies of bromocriptine have been conducted with patients who have renal impairment, and the kidney is a minor elimination pathway for bromocriptine. The package insert offers no specific dose recommendations for such patients, although it does recommend caution when using this product in patients with renal impairment. Studies of bromocriptine in patients with liver dysfunction are also lacking. However, as bromocriptine is predominately metabolized in the liver, use caution in patients with hepatic impairment.¹⁶

TABLE
Key prescribing information for bromocriptine¹⁶

Adverse effects are mostly GI related

In phase 3 clinical trials (bromocriptine n=2298; placebo n=1266), adverse events leading to drug discontinuation occurred in 539 (24%) of bromocriptine-treated patients and 118 (9%) placebo-treated patients.¹⁶ This difference was mostly driven by an increase in GI adverse events with bromocriptine, particularly nausea. The most commonly reported adverse events from bromocriptine (nausea, fatigue, vomiting, headache, and dizziness) lasted a median of 14 days and were more likely to occur during the initial titration period. None of the reports of nausea or vomiting was considered serious.

There were no differences in the pattern of common adverse events across races or age groups (<65 vs >65 years old). Hypoglycemia occurred infrequently during the 52-week safety trial, with 6.9% of the bromocriptine patients and 5.3% of the placebo patients reporting an event.¹³ In this same safety trial, 1.6% of bromocriptine patients experienced syncope vs 0.7% of placebo-treated patients. CNS effects (somnolence and hypoesthesia) were minimal. Serious adverse events affected 8.5% of bromocriptine patients and 9.6% of placebo-treated patients (hazard ratio=1.02; 96% one-sided confidence interval, 1.27). Fewer people in the bromocriptine group reported a cardiovascular disease endpoint (composite of myocardial infarction, stroke, coronary revascularization, hospitalization for angina, and hospitalization for congestive heart failure) than did those in the placebo group (1.8% vs 3.2%, respectively).^13,16

Postmarketing data link bromocriptine with hallucinations, fibrotic complications, and psychotic disorders. However, these adverse reactions were found with the use of much higher doses (30-140 mg/d) and with other indications for bromocriptine. These reactions have not been reported in clinical trials of bromocriptine used to treat T2D.¹⁶

Drugs to avoid (or use cautiously) with bromocriptine

Because bromocriptine is highly bound to serum proteins, it may increase the unbound fraction of other highly protein-bound drugs (eg, salicylates, sulfonamides, chloramphenicol, probenecid), which could alter their effectiveness or risk for adverse effects. Because bromocriptine is a dopamine receptor agonist, concomitant use of dopamine antagonists such as neuroleptic agents (clozapine, olanzapine) or metoclopramide is not recommended.¹⁶

Combining bromocriptine with ergot-related drugs (eg, migraine therapies) may increase the occurrence of ergot-related adverse effects such as nausea, vomiting, and fatigue, and may diminish effectiveness of migraine therapies. Dosing of the 2 therapies should occur at least 6 hours apart.¹⁶

Bromocriptine is extensively metabolized via CYP3A4. Potent inhibitors of this enzyme (eg, azole antimycotics, HIV protease inhibitors) or inducers (eg, rifampin, carbamazepine, phenytoin, phenobarbital) should be used with caution. Clinical trial data are limited regarding the safety of sumatriptan (5-HT_1B agonist) used concurrently with bromocriptine, so it is prudent to avoid using them together.¹⁶

Not for breastfeeding moms, migraine sufferers

Bromocriptine is contraindicated for patients with syncopal migraine due to an increase in the likelihood of a hypotensive episode. It is also contraindicated for women who are breastfeeding due to its ability to inhibit lactation and to postmarketing reports of stroke in this population. Bromocriptine can lead to hypotension; monitor blood pressure during dose escalation and when a patient is taking antihypertensives.

Bromocriptine should not be used in patients with severe psychiatric disorders, as it may exacerbate their conditions or diminish the effectiveness of their treatment. Warn patients that somnolence can occur with bromocriptine, particularly during titration. No clinical studies have shown conclusive evidence of macrovascular risk reduction with bromocriptine or any other antidiabetic drug.¹⁶ But neither has bromocriptine increased risk for cardiovascular events.¹³

Putting bromocriptine’s usefulness into perspective

The larger studies of bromocriptine have shown absolute mean reductions in A1c of 0.1% to 0.6% and in fasting glucose of 0 to 10 mg/dL. When compared with placebo, mean A1c and fasting glucose differences were 0.4% to 1.2% and 23 to 38 mg/dL, respectively. While these findings were statistically significant when compared with placebo, they are clinically modest.

Although bromocriptine offers a few advantages, such as no weight gain, low risk of hypoglycemia, and possible beneficial effects on insulin resistance and triglyceride levels, its use should be limited at this time because it is less efficacious than other agents and long-term trials are lacking. Bromocriptine is not currently included in any treatment guidelines for the management of T2D. Cost is also a concern (TABLE). Because the medication is supplied only as 0.8-mg tablets, patients on the maximum dose would need to take 6 tablets once daily.

CORRESPONDENCE
Karen R. Sando, PharmD, CDE, University of Florida, College of Pharmacy, Department of Pharmacotherapy and Translational Research, 101 S. Newell Drive, HPNP Building, Room 3306, Gainesville, FL 32610; ksando@cop.ufl.edu

For obese patients with type 2 diabetes (T2D) who do not tolerate other diabetes medications or for patients who need only a minimal reduction in glycosylated hemoglobin (A1c) to reach goal, bromocriptine may be a therapeutic agent to consider. Approved by the US Food and Drug Administration (FDA) in 2009, Cycloset is a quick-release formulation of bromocriptine mesylate, an ergot alkaloid dopamine D2 receptor agonist that has a faster onset of action than the customary formulation, Parlodel, which has been used to treat Parkinson’s disease, acromegaly, and hyperprolactinemia.¹ In addition to its modest benefit in improving glycemic control, Cyclocet avoids undesirable side effects such as hypoglycemia and weight gain.

VeroScience, LLC holds the new drug application and related technologies for Cycloset and partnered with Santarus, Inc. and S2 Therapeutics to market it in September 2010.²

Bromocriptine’s likely mechanism of action

Although its exact mechanism of action is unclear, bromocriptine does not stimulate insulin release, reduce hepatic glucose production, increase glucose transporter production, or increase or mimic glucagon-like peptide-1 activity as other T2D agents do.³ Its contribution to glycemic control in T2D has been hypothesized to be due to adjustments in the neural control of seasonal and diurnal patterns of food intake and nutrient storage.⁴

Early hunter-gatherers and farmers are thought to have benefited from a “thrifty genotype” that favored fat deposition when food was seasonally abundant. With food in western society available year-round and often energy dense in both fat and carbohydrates, this same gene may lead to obesity and noninsulin-dependent diabetes mellitus.⁵

The hypothesis assumes that circadian rhythm, photo- periodism, and seasonal factors play a role in insulin resistance, hepatic gluconeogenesis, and weight gain. In vertebrates, the neuroendocrine system plays an important role in synchronizing the animal with cyclic environmental changes. The hypothalamic suprachiasmatic nucleus (SCN) is known as the circadian pacemaker that maintains this rhythm. Oscillations in the SCN occur due to external cues such as changes in light or temperature. Circadian dopaminergic and serotonergic activities are likely responsible for modifying such oscillations, and neurotransmitters have been shown to regulate the dramatic seasonal alterations in body weight and body composition of all vertebrate classes.⁶ Bromocriptine can reverse metabolic alterations associated with insulin resistance and obesity by resetting central (hypothalamic) circadian organization of monoamine neuronal activities.⁷

Proven anti-T2D effects. When administered systemically or into the cerebral ventricle at first light, bromocriptine prevents or reverses seasonal fattening, insulin resistance, and hyperinsulinemia, and it decreases endogenous (hepatic) glucose production in mammals.^8-11 Bromocriptine also decreases both fasting and postprandial triglyceride and free fatty acid levels.¹

Clinical trials show modest benefit

Although bromocriptine has been studied since 1980 for its effects on hyperglycemia in T2D,¹² trials leading to the approval of the drug for clinical use with T2D have only been completed within the last 15 years. Randomized controlled trials of varying sizes and lasting from 6 to 52 weeks have shown absolute A1c reductions from 0.1% to 0.6%.^1,12-16 Compared with placebo, A1c reductions have ranged from 0.4% to 1.2% with monotherapy and in combination with other antidiabetes medications.^1,13-16

The manufacturer assessed bromocriptine in 4 studies involving patients with T2D. In all 4 studies, the bromocriptine dose was titrated to a maximum of 4.8 mg/d.¹⁶

One study involved 159 overweight subjects who were not meeting glycemic goals.¹⁶ Patients received either placebo or bromocriptine for 24 weeks in addition to diet and exercise. Mean baseline A1c was 9.0% in the bromocriptine group and 8.8% in the placebo group. After 24 weeks, A1c was reduced by 0.1% in the treatment group and increased by 0.3% in the placebo group. Mean fasting glucose was 215 mg/dL at baseline in the treatment group and was unchanged after 24 weeks. In the placebo group, fasting glucose increased from 205 to 228 mg/dL during the study. Weight increased by 0.2 kg in the treatment group and by 0.5 kg in the placebo group.

The next two 24-week manufacturer studies used similar designs to compare the addition of either bromocriptine or placebo to existing sulfonylurea therapy in patients with uncontrolled T2D.¹⁶ One study assigned 122 patients to bromocriptine and 127 to placebo. The bromocriptine group demonstrated mean reductions of 0.4% in A1c and 3 mg/dL in fasting glucose. In the placebo group, A1c increased by 0.3% and fasting glucose rose by 23 mg/dL.

The other study assigned 122 patients to bromocriptine and 123 to placebo. Adding bromocriptine reduced A1c, on average, by 0.1% and fasting glucose by 10 mg/dL. In the placebo group, A1c increased by 0.4% and fasting glucose increased by 28 mg/dL. All of these results were statistically significant.

The last manufacturer-reported study evaluated the addition of bromocriptine to other diabetes treatments (diet or up to 2 anti-diabetes medications).¹³ While the primary intent of this study was to evaluate safety, it also assessed efficacy. This was a 52-week, randomized placebo-controlled trial involving 3095 patients.

Overall, after 24 weeks there was no change in A1c levels after adding bromocriptine. However, most patients in this study were already at goal (A1c <7.0%). A subgroup analysis of those with an A1c level <7.5% while taking other agents did show some improvement with the addition of bromocriptine. Adding bromocriptine to metformin and a sulfonylurea significantly reduced A1c by 0.5%, on average. Similar results were seen in those who received other combinations of diabetes medications. After 52 weeks, 25% of those receiving bromocriptine who originally had an A1c level >7.5% achieved an A1c level <7.0%. Of the patients who received placebo, 9% obtained an A1c level <7%.

In a 24-week study, bromocriptine titrated up to 4.8 mg/d was given to patients either on no other diabetes medication or on a sulfonylurea.¹ In individuals not on any current treatment, A1c decreased by 0.2% in those who received bromocriptine. In patients already on a sulfonylurea, A1c declined by 0.1%. A1c increased by 0.3% in those receiving placebo.

Bromocriptine most effective when taken with food

When bromocriptine is taken orally, 65% to 95% of the dose is absorbed; however, only 7% reaches systemic circulation due to extensive hepatic extraction and first-pass metabolism.¹⁷ Bioavailability increases by 55% to 65% when the drug is taken with food, which is how it should be administered. The time to maximum plasma concentration is within an hour. With a high-fat meal, however, the time increases to 90 to 120 minutes. Bromocriptine is highly protein bound (90%-96%) and is metabolized extensively in the gastrointestinal (GI) tract and liver.¹⁷ CYP3A4 is the major metabolic pathway.^1,18 Most excretion of bromocriptine is through bile, with approximately 2% to 6% of an oral dose eliminated via urine. The elimination half-life is approximately 6 hours.^17,18

Dosing is once a day in the morning

Clinical trials investigating the use of bromocriptine in diabetes used doses ranging from 1.6 to 4.8 mg/d.^13-16,19 The FDA-approved dose range is 1.6 to 4.8 mg administered once daily with food, within 2 hours of waking in the morning.¹⁶ In healthy individuals, central nervous system (CNS) dopaminergic activity peaks in the early morning. Thus, morning dosing attempts to mimic dopaminergic activity and circadian rhythms in healthy lean individuals.⁶

Titrate to maximum dose. The product is available in a 0.8-mg tablet (TABLE). Titration to the maximum dose is recommended to reduce GI adverse effects, particularly nausea. Start treatment with 1 tablet (0.8 mg) and increase the dose by 1 tablet per week until the patient reaches a maximum tolerated dose or the maximum allowable daily dose of 4.8 mg (6 tablets).

Precautions with renal or hepatic impairment. No pharmacokinetic studies of bromocriptine have been conducted with patients who have renal impairment, and the kidney is a minor elimination pathway for bromocriptine. The package insert offers no specific dose recommendations for such patients, although it does recommend caution when using this product in patients with renal impairment. Studies of bromocriptine in patients with liver dysfunction are also lacking. However, as bromocriptine is predominately metabolized in the liver, use caution in patients with hepatic impairment.¹⁶

TABLE
Key prescribing information for bromocriptine¹⁶

Adverse effects are mostly GI related

In phase 3 clinical trials (bromocriptine n=2298; placebo n=1266), adverse events leading to drug discontinuation occurred in 539 (24%) of bromocriptine-treated patients and 118 (9%) placebo-treated patients.¹⁶ This difference was mostly driven by an increase in GI adverse events with bromocriptine, particularly nausea. The most commonly reported adverse events from bromocriptine (nausea, fatigue, vomiting, headache, and dizziness) lasted a median of 14 days and were more likely to occur during the initial titration period. None of the reports of nausea or vomiting was considered serious.

There were no differences in the pattern of common adverse events across races or age groups (<65 vs >65 years old). Hypoglycemia occurred infrequently during the 52-week safety trial, with 6.9% of the bromocriptine patients and 5.3% of the placebo patients reporting an event.¹³ In this same safety trial, 1.6% of bromocriptine patients experienced syncope vs 0.7% of placebo-treated patients. CNS effects (somnolence and hypoesthesia) were minimal. Serious adverse events affected 8.5% of bromocriptine patients and 9.6% of placebo-treated patients (hazard ratio=1.02; 96% one-sided confidence interval, 1.27). Fewer people in the bromocriptine group reported a cardiovascular disease endpoint (composite of myocardial infarction, stroke, coronary revascularization, hospitalization for angina, and hospitalization for congestive heart failure) than did those in the placebo group (1.8% vs 3.2%, respectively).^13,16

Postmarketing data link bromocriptine with hallucinations, fibrotic complications, and psychotic disorders. However, these adverse reactions were found with the use of much higher doses (30-140 mg/d) and with other indications for bromocriptine. These reactions have not been reported in clinical trials of bromocriptine used to treat T2D.¹⁶

Drugs to avoid (or use cautiously) with bromocriptine

Because bromocriptine is highly bound to serum proteins, it may increase the unbound fraction of other highly protein-bound drugs (eg, salicylates, sulfonamides, chloramphenicol, probenecid), which could alter their effectiveness or risk for adverse effects. Because bromocriptine is a dopamine receptor agonist, concomitant use of dopamine antagonists such as neuroleptic agents (clozapine, olanzapine) or metoclopramide is not recommended.¹⁶

Combining bromocriptine with ergot-related drugs (eg, migraine therapies) may increase the occurrence of ergot-related adverse effects such as nausea, vomiting, and fatigue, and may diminish effectiveness of migraine therapies. Dosing of the 2 therapies should occur at least 6 hours apart.¹⁶

Bromocriptine is extensively metabolized via CYP3A4. Potent inhibitors of this enzyme (eg, azole antimycotics, HIV protease inhibitors) or inducers (eg, rifampin, carbamazepine, phenytoin, phenobarbital) should be used with caution. Clinical trial data are limited regarding the safety of sumatriptan (5-HT_1B agonist) used concurrently with bromocriptine, so it is prudent to avoid using them together.¹⁶

Not for breastfeeding moms, migraine sufferers

Bromocriptine is contraindicated for patients with syncopal migraine due to an increase in the likelihood of a hypotensive episode. It is also contraindicated for women who are breastfeeding due to its ability to inhibit lactation and to postmarketing reports of stroke in this population. Bromocriptine can lead to hypotension; monitor blood pressure during dose escalation and when a patient is taking antihypertensives.

Bromocriptine should not be used in patients with severe psychiatric disorders, as it may exacerbate their conditions or diminish the effectiveness of their treatment. Warn patients that somnolence can occur with bromocriptine, particularly during titration. No clinical studies have shown conclusive evidence of macrovascular risk reduction with bromocriptine or any other antidiabetic drug.¹⁶ But neither has bromocriptine increased risk for cardiovascular events.¹³

Putting bromocriptine’s usefulness into perspective

The larger studies of bromocriptine have shown absolute mean reductions in A1c of 0.1% to 0.6% and in fasting glucose of 0 to 10 mg/dL. When compared with placebo, mean A1c and fasting glucose differences were 0.4% to 1.2% and 23 to 38 mg/dL, respectively. While these findings were statistically significant when compared with placebo, they are clinically modest.

Although bromocriptine offers a few advantages, such as no weight gain, low risk of hypoglycemia, and possible beneficial effects on insulin resistance and triglyceride levels, its use should be limited at this time because it is less efficacious than other agents and long-term trials are lacking. Bromocriptine is not currently included in any treatment guidelines for the management of T2D. Cost is also a concern (TABLE). Because the medication is supplied only as 0.8-mg tablets, patients on the maximum dose would need to take 6 tablets once daily.

CORRESPONDENCE
Karen R. Sando, PharmD, CDE, University of Florida, College of Pharmacy, Department of Pharmacotherapy and Translational Research, 101 S. Newell Drive, HPNP Building, Room 3306, Gainesville, FL 32610; ksando@cop.ufl.edu

For obese patients with type 2 diabetes (T2D) who do not tolerate other diabetes medications or for patients who need only a minimal reduction in glycosylated hemoglobin (A1c) to reach goal, bromocriptine may be a therapeutic agent to consider. Approved by the US Food and Drug Administration (FDA) in 2009, Cycloset is a quick-release formulation of bromocriptine mesylate, an ergot alkaloid dopamine D2 receptor agonist that has a faster onset of action than the customary formulation, Parlodel, which has been used to treat Parkinson’s disease, acromegaly, and hyperprolactinemia.¹ In addition to its modest benefit in improving glycemic control, Cyclocet avoids undesirable side effects such as hypoglycemia and weight gain.

VeroScience, LLC holds the new drug application and related technologies for Cycloset and partnered with Santarus, Inc. and S2 Therapeutics to market it in September 2010.²

Bromocriptine’s likely mechanism of action

Although its exact mechanism of action is unclear, bromocriptine does not stimulate insulin release, reduce hepatic glucose production, increase glucose transporter production, or increase or mimic glucagon-like peptide-1 activity as other T2D agents do.³ Its contribution to glycemic control in T2D has been hypothesized to be due to adjustments in the neural control of seasonal and diurnal patterns of food intake and nutrient storage.⁴

Early hunter-gatherers and farmers are thought to have benefited from a “thrifty genotype” that favored fat deposition when food was seasonally abundant. With food in western society available year-round and often energy dense in both fat and carbohydrates, this same gene may lead to obesity and noninsulin-dependent diabetes mellitus.⁵

The hypothesis assumes that circadian rhythm, photo- periodism, and seasonal factors play a role in insulin resistance, hepatic gluconeogenesis, and weight gain. In vertebrates, the neuroendocrine system plays an important role in synchronizing the animal with cyclic environmental changes. The hypothalamic suprachiasmatic nucleus (SCN) is known as the circadian pacemaker that maintains this rhythm. Oscillations in the SCN occur due to external cues such as changes in light or temperature. Circadian dopaminergic and serotonergic activities are likely responsible for modifying such oscillations, and neurotransmitters have been shown to regulate the dramatic seasonal alterations in body weight and body composition of all vertebrate classes.⁶ Bromocriptine can reverse metabolic alterations associated with insulin resistance and obesity by resetting central (hypothalamic) circadian organization of monoamine neuronal activities.⁷

Proven anti-T2D effects. When administered systemically or into the cerebral ventricle at first light, bromocriptine prevents or reverses seasonal fattening, insulin resistance, and hyperinsulinemia, and it decreases endogenous (hepatic) glucose production in mammals.^8-11 Bromocriptine also decreases both fasting and postprandial triglyceride and free fatty acid levels.¹

Clinical trials show modest benefit

Although bromocriptine has been studied since 1980 for its effects on hyperglycemia in T2D,¹² trials leading to the approval of the drug for clinical use with T2D have only been completed within the last 15 years. Randomized controlled trials of varying sizes and lasting from 6 to 52 weeks have shown absolute A1c reductions from 0.1% to 0.6%.^1,12-16 Compared with placebo, A1c reductions have ranged from 0.4% to 1.2% with monotherapy and in combination with other antidiabetes medications.^1,13-16

The manufacturer assessed bromocriptine in 4 studies involving patients with T2D. In all 4 studies, the bromocriptine dose was titrated to a maximum of 4.8 mg/d.¹⁶

One study involved 159 overweight subjects who were not meeting glycemic goals.¹⁶ Patients received either placebo or bromocriptine for 24 weeks in addition to diet and exercise. Mean baseline A1c was 9.0% in the bromocriptine group and 8.8% in the placebo group. After 24 weeks, A1c was reduced by 0.1% in the treatment group and increased by 0.3% in the placebo group. Mean fasting glucose was 215 mg/dL at baseline in the treatment group and was unchanged after 24 weeks. In the placebo group, fasting glucose increased from 205 to 228 mg/dL during the study. Weight increased by 0.2 kg in the treatment group and by 0.5 kg in the placebo group.

The next two 24-week manufacturer studies used similar designs to compare the addition of either bromocriptine or placebo to existing sulfonylurea therapy in patients with uncontrolled T2D.¹⁶ One study assigned 122 patients to bromocriptine and 127 to placebo. The bromocriptine group demonstrated mean reductions of 0.4% in A1c and 3 mg/dL in fasting glucose. In the placebo group, A1c increased by 0.3% and fasting glucose rose by 23 mg/dL.

The other study assigned 122 patients to bromocriptine and 123 to placebo. Adding bromocriptine reduced A1c, on average, by 0.1% and fasting glucose by 10 mg/dL. In the placebo group, A1c increased by 0.4% and fasting glucose increased by 28 mg/dL. All of these results were statistically significant.

The last manufacturer-reported study evaluated the addition of bromocriptine to other diabetes treatments (diet or up to 2 anti-diabetes medications).¹³ While the primary intent of this study was to evaluate safety, it also assessed efficacy. This was a 52-week, randomized placebo-controlled trial involving 3095 patients.

Overall, after 24 weeks there was no change in A1c levels after adding bromocriptine. However, most patients in this study were already at goal (A1c <7.0%). A subgroup analysis of those with an A1c level <7.5% while taking other agents did show some improvement with the addition of bromocriptine. Adding bromocriptine to metformin and a sulfonylurea significantly reduced A1c by 0.5%, on average. Similar results were seen in those who received other combinations of diabetes medications. After 52 weeks, 25% of those receiving bromocriptine who originally had an A1c level >7.5% achieved an A1c level <7.0%. Of the patients who received placebo, 9% obtained an A1c level <7%.

In a 24-week study, bromocriptine titrated up to 4.8 mg/d was given to patients either on no other diabetes medication or on a sulfonylurea.¹ In individuals not on any current treatment, A1c decreased by 0.2% in those who received bromocriptine. In patients already on a sulfonylurea, A1c declined by 0.1%. A1c increased by 0.3% in those receiving placebo.

Bromocriptine most effective when taken with food

When bromocriptine is taken orally, 65% to 95% of the dose is absorbed; however, only 7% reaches systemic circulation due to extensive hepatic extraction and first-pass metabolism.¹⁷ Bioavailability increases by 55% to 65% when the drug is taken with food, which is how it should be administered. The time to maximum plasma concentration is within an hour. With a high-fat meal, however, the time increases to 90 to 120 minutes. Bromocriptine is highly protein bound (90%-96%) and is metabolized extensively in the gastrointestinal (GI) tract and liver.¹⁷ CYP3A4 is the major metabolic pathway.^1,18 Most excretion of bromocriptine is through bile, with approximately 2% to 6% of an oral dose eliminated via urine. The elimination half-life is approximately 6 hours.^17,18

Dosing is once a day in the morning

Clinical trials investigating the use of bromocriptine in diabetes used doses ranging from 1.6 to 4.8 mg/d.^13-16,19 The FDA-approved dose range is 1.6 to 4.8 mg administered once daily with food, within 2 hours of waking in the morning.¹⁶ In healthy individuals, central nervous system (CNS) dopaminergic activity peaks in the early morning. Thus, morning dosing attempts to mimic dopaminergic activity and circadian rhythms in healthy lean individuals.⁶

Titrate to maximum dose. The product is available in a 0.8-mg tablet (TABLE). Titration to the maximum dose is recommended to reduce GI adverse effects, particularly nausea. Start treatment with 1 tablet (0.8 mg) and increase the dose by 1 tablet per week until the patient reaches a maximum tolerated dose or the maximum allowable daily dose of 4.8 mg (6 tablets).

Precautions with renal or hepatic impairment. No pharmacokinetic studies of bromocriptine have been conducted with patients who have renal impairment, and the kidney is a minor elimination pathway for bromocriptine. The package insert offers no specific dose recommendations for such patients, although it does recommend caution when using this product in patients with renal impairment. Studies of bromocriptine in patients with liver dysfunction are also lacking. However, as bromocriptine is predominately metabolized in the liver, use caution in patients with hepatic impairment.¹⁶

TABLE
Key prescribing information for bromocriptine¹⁶

Adverse effects are mostly GI related

In phase 3 clinical trials (bromocriptine n=2298; placebo n=1266), adverse events leading to drug discontinuation occurred in 539 (24%) of bromocriptine-treated patients and 118 (9%) placebo-treated patients.¹⁶ This difference was mostly driven by an increase in GI adverse events with bromocriptine, particularly nausea. The most commonly reported adverse events from bromocriptine (nausea, fatigue, vomiting, headache, and dizziness) lasted a median of 14 days and were more likely to occur during the initial titration period. None of the reports of nausea or vomiting was considered serious.

There were no differences in the pattern of common adverse events across races or age groups (<65 vs >65 years old). Hypoglycemia occurred infrequently during the 52-week safety trial, with 6.9% of the bromocriptine patients and 5.3% of the placebo patients reporting an event.¹³ In this same safety trial, 1.6% of bromocriptine patients experienced syncope vs 0.7% of placebo-treated patients. CNS effects (somnolence and hypoesthesia) were minimal. Serious adverse events affected 8.5% of bromocriptine patients and 9.6% of placebo-treated patients (hazard ratio=1.02; 96% one-sided confidence interval, 1.27). Fewer people in the bromocriptine group reported a cardiovascular disease endpoint (composite of myocardial infarction, stroke, coronary revascularization, hospitalization for angina, and hospitalization for congestive heart failure) than did those in the placebo group (1.8% vs 3.2%, respectively).^13,16

Postmarketing data link bromocriptine with hallucinations, fibrotic complications, and psychotic disorders. However, these adverse reactions were found with the use of much higher doses (30-140 mg/d) and with other indications for bromocriptine. These reactions have not been reported in clinical trials of bromocriptine used to treat T2D.¹⁶

Drugs to avoid (or use cautiously) with bromocriptine

Because bromocriptine is highly bound to serum proteins, it may increase the unbound fraction of other highly protein-bound drugs (eg, salicylates, sulfonamides, chloramphenicol, probenecid), which could alter their effectiveness or risk for adverse effects. Because bromocriptine is a dopamine receptor agonist, concomitant use of dopamine antagonists such as neuroleptic agents (clozapine, olanzapine) or metoclopramide is not recommended.¹⁶

Combining bromocriptine with ergot-related drugs (eg, migraine therapies) may increase the occurrence of ergot-related adverse effects such as nausea, vomiting, and fatigue, and may diminish effectiveness of migraine therapies. Dosing of the 2 therapies should occur at least 6 hours apart.¹⁶

Bromocriptine is extensively metabolized via CYP3A4. Potent inhibitors of this enzyme (eg, azole antimycotics, HIV protease inhibitors) or inducers (eg, rifampin, carbamazepine, phenytoin, phenobarbital) should be used with caution. Clinical trial data are limited regarding the safety of sumatriptan (5-HT_1B agonist) used concurrently with bromocriptine, so it is prudent to avoid using them together.¹⁶

Not for breastfeeding moms, migraine sufferers

Bromocriptine is contraindicated for patients with syncopal migraine due to an increase in the likelihood of a hypotensive episode. It is also contraindicated for women who are breastfeeding due to its ability to inhibit lactation and to postmarketing reports of stroke in this population. Bromocriptine can lead to hypotension; monitor blood pressure during dose escalation and when a patient is taking antihypertensives.

Bromocriptine should not be used in patients with severe psychiatric disorders, as it may exacerbate their conditions or diminish the effectiveness of their treatment. Warn patients that somnolence can occur with bromocriptine, particularly during titration. No clinical studies have shown conclusive evidence of macrovascular risk reduction with bromocriptine or any other antidiabetic drug.¹⁶ But neither has bromocriptine increased risk for cardiovascular events.¹³

Putting bromocriptine’s usefulness into perspective

The larger studies of bromocriptine have shown absolute mean reductions in A1c of 0.1% to 0.6% and in fasting glucose of 0 to 10 mg/dL. When compared with placebo, mean A1c and fasting glucose differences were 0.4% to 1.2% and 23 to 38 mg/dL, respectively. While these findings were statistically significant when compared with placebo, they are clinically modest.

Although bromocriptine offers a few advantages, such as no weight gain, low risk of hypoglycemia, and possible beneficial effects on insulin resistance and triglyceride levels, its use should be limited at this time because it is less efficacious than other agents and long-term trials are lacking. Bromocriptine is not currently included in any treatment guidelines for the management of T2D. Cost is also a concern (TABLE). Because the medication is supplied only as 0.8-mg tablets, patients on the maximum dose would need to take 6 tablets once daily.

CORRESPONDENCE
Karen R. Sando, PharmD, CDE, University of Florida, College of Pharmacy, Department of Pharmacotherapy and Translational Research, 101 S. Newell Drive, HPNP Building, Room 3306, Gainesville, FL 32610; ksando@cop.ufl.edu

We all know that the key to optimal diabetes management is tight glucose control, which can be achieved with multiple daily fingersticks, good record keeping of the results, and appropriate modification of the medication regimen, diet, and exercise schedule.

But patients find the routine burdensome, and many skip fingersticks or abandon the process entirely. And even those who follow the program faithfully may find that it fails to protect them from unpleasant and potentially dangerous episodes of hyper- and hypoglycemia.

The newer technology of continuous glucose monitoring systems (CGMS) offers the promise of overcoming these limitations. But how do these new systems work and what does the evidence tell us about their potential benefits and remaining uncertainties? Read on.

The old way: Take a snapshot
The variables that affect blood glucose levels—meals and snacks, exercise or the lack of it, dosages and timing of medication, and stress, among others—keep changing throughout the day and night. The impact of these variables cannot be adequately captured in snapshot blood glucose levels taken at isolated moments in the patient’s day. Achieving glycemic control with blood glucose monitors can be difficult for some patients, especially since the data generated are dependent on the patient’s willingness and ability to self-monitor his or her glucose levels.

The new way: Monitor continuously

CGMS measure the amount of glucose in the interstitial fluid—not in the blood. These measurements are taken every 5 minutes or so, depending on the system. Each system consists of a sensor, transmitter, and receiver. The sensor is a fine wire—about the diameter of 2 human hairs—that sticks into the skin of the abdomen or upper arm and is kept in place by an adhesive pad. The transmitter fits on the sensor pad and sends information to the receiver via radio waves. Sensors are disposable; they last for 3 to 5 days and then must be replaced. The system is wireless, so your patient isn’t tethered to the equipment.

Calibration with a glucose meter is still necessary. To be sure that interstitial glucose measurements reflect actual blood glucose levels, currently available systems require daily calibration with conventional blood glucose monitors. Patients will still have to do fingersticks, but far less frequently. The FDA has approved CGMS for use only in conjunction with conventional glucose testing. Traditional glucose self-monitoring may also be necessary when CGM results do not correspond to symptoms patients are experiencing.

Receiver displays data, can set off an alarm. Glucose measurements from the CGMS are displayed and stored in the receiver, and the data can be downloaded to a computer using the manufacturer’s data software. Continuous readings over a 24-hour period for up to 7 days allow the user to detect variation and identify trends. High and low glucose value thresholds can be customized for the individual patient and fed into the system. When these thresholds are exceeded, an alarm will sound. The receiver displays directional arrows showing the rate of change in glucose levels, allowing the patient to predict—and possibly prevent—hypoglycemic episodes.

Impact of events can be noted. The systems also allow for input of additional information about events that may affect blood glucose levels, such as medication, exercise, and food intake. Patients can use information about how these events affect their glucose levels to adjust the prandial or basal insulin dose, modify the insulin correction algorithm, or adjust their diet. Patients can bring computer-generated charts and graphs to office visits as a basis for joint decision-making about their care. Short-term, periodic use of a CGMS in patients with type 2 diabetes may identify times when patients need more frequent self-testing or guide further therapy selection.

These systems are available now

The systems available in the United States include:

• the iPro Continuous Glucose Monitor, Guardian Real-Time System, and Mini Med Paradigm Real-Time System—all from Medtronic, Inc.
• the SEVEN PLUS, from DexCom
• the FreeStyle Navigator, from Abbott.^1-3

The SEVEN PLUS and the FreeStyle Navigator are FDA approved for adults only. Pediatric versions of Medtronic’s MiniMed Paradigm and Guardian systems are approved for use in patients ages 7 to 17. All these systems require a prescription. For detailed comparisons of the features of these systems, see the TABLE.

TABLE
Continuous glucose monitoring systems: The options

	SEVEN PLUS	FreeStyle Navigator	Guardian Real-Time System	MiniMed Paradigm Real-Time System	iPro Continuous Glucose Monitor*
Manufacturer	DexCom	Abbott	Medtronic, Inc.	Medtronic, Inc.	Medtronic, Inc.
URL	www.dexcom.com	www.freestylenavigator.com	www.minimed.com	www.minimed.com	www.minimed.com
Price	$799 for system; $399 for 4 sensors; $79 for software	$1250 for system; $450 for 6 sensors	$1350 for system, including 4 sensors; $350 for 10 sensors	$999, plus cost of insulin pump; $35 per sensor	$1090 for start-up; $350 for 10 sensors
Receiver range	5 feet	10 feet	6 feet	6 feet
Sensor life	Up to 7 days	Up to 5 days	Up to 3 days	Up to 3 days	Up to 3 days
Calibration	2 hours after insertion, then every 12 hours	At least 4 times over a 5-day period at 10, 12, 24, and 72 hours after insertion	2 hours after insertion, again within 6 hours, then every 12 hours	2 hours after insertion, then within next 6 hours, then every 12 hours
User-set alarm for highs/lows	Yes, plus factory alarm at 55 mg/dL that can’t be disabled	Yes	Yes	Yes
Glucose reading display frequency	Every 5 minutes	Once every minute	Measures every minute, displays an average of every 5 minutes	Measures every minute, displays an average of every 5 minutes
Displays directional trends	Yes	Yes	Yes	Yes
Sources: Diabetes Network. Diabetes technology. Available at: www.diabetesnet.com/diabetes_technology/continuous_monitoring.php. Accessed January 6, 2010.
DexCom. Available at: www.dexcom.com. Accessed January 6, 2010.
FreeStyle Navigator. Available at: www.freestylenavigator.com. Accessed January 6, 2010.
Medtronic. Available at: www.minimed.com. Accessed January 6, 2010.
Conversations with Robert Sala, sales representative, DexCom, on May 1 and May 8, 2009.
* iPro consists of sensor and transmitter only; no receiver. Sensor is inserted by provider; data are uploaded in provider’s office to help guide therapeutic decision-making.

Patients with severe diabetes benefit most
Patients with type 1 diabetes who use an insulin pump or are being switched from multiple injections to pumps, and patients who have problems with hypoglycemia are good candidates for CGMS. The latter group includes those who are not aware of their hypoglycemic state, those who have nighttime hypoglycemia, and those who experience severe episodes of hypoglycemia. The category also includes patients who keep their blood glucose levels higher than appropriate goals would indicate, because of their fear of hypoglycemia.

An additional group of patients who might benefit, although they do not fit currently approved indications for these devices, are pregnant women who should maintain tight glucose control. Other patients who might find CGMS useful are those with glycemic variability or those who have not achieved their A1C goal and want to be proactive.

Your letter of medical necessity can qualify patients like these for Medicare or private insurance reimbursement for the CGMS and for ongoing sensor supplies. You may also choose to purchase a system yourself for patients to use, and bill the patient’s insurance company for the service.

Accuracy continues to be a concern
Currently available systems are more accurate than the first generation of these devices. When glucose is rapidly changing, users need to be aware that there may be a time lag before the interstitial glucose reaches the same level as the blood glucose. So, while medication changes can be made based on CGMS, values should be confirmed with a fingerstick.

SEVEN and Navigator are comparable
A number of studies have confirmed the accuracy of CGMS.^4-7 A study by Garg and colleagues compared the accuracy of the DexCom SEVEN and the FreeStyle Navigator.⁶ Fourteen patients wore sensors from both systems for 3 consecutive, 5-day periods. Laboratory reference measurements of venous blood glucose were taken every 15 minutes through an 8-hour period on days 5, 10, and 15 in clinic using the YSI STAT Plus Glucose Analyzer. Sensors were replaced at the end of the clinic day on days 5 and 10, and the sensors were removed at the end of day 15. The mean absolute relative difference for CGM compared with laboratory glucose measures was 16.8% for the SEVEN and 16.1% for the Navigator (P=.38), an insignificant difference between the 2 systems.

The 2 systems were also compared using continuous glucose error grid analysis, which evaluates how accurately CGM data lead to an appropriate clinical response by the patient. The error grid is divided into 5 zones and superimposed on the correlation plot. Plots in Zone A are a perfect fit and plots in Zone B are “benign error” that does not result in an inaccurate clinical response. The percentage of data points in Zones A or B was 94.8% for the SEVEN and 93.2% for the Navigator. The SEVEN provided better agreement with laboratory glucose measures for the range 40 to 80 mg/dL (P<.001).

Guardian evaluation has similar results
A similar study done by Medtronic in 2004 evaluated the Guardian RT, an earlier version of the Guardian, in 16 patients.⁷ Values from the Guardian RT were compared with reference YSI STAT Plus glucose analyzer glucose values taken every 30 minutes in clinic. The mean absolute percent difference was 19.7%±18.4%. Of the 3941 total paired glucose measurements, 96% fell in the clinically acceptable error grid Zones A or B. For low glucose values between 40 and 80 mg/dL, 76.1% of readings fell in Zones A or B; for high values, over 240 mg/dL, 86.8% of readings fell in Zones A or B. Accuracy in the hypoglycemic ranges declined as the time increased from insertion of the sensor.

Safety risks are few, minor

Insertion of the sensor can pose minor safety risks, including infection, inflammation, and bleeding. Adverse events reported in 1 study consisted mainly of mild sensor site reactions such as blisters, bullae, edema, and erythema, none of which required treatment.⁶ The CGMS must be removed prior to magnetic imaging studies and the devices are not approved for use on airplanes. When the FreeStyle Navigator sensor is removed, a portion of the membrane polymer is left in the skin. The company reports no health effects in clinical studies, aside from sensor site reactions mentioned above, but long-term effects of sensor membrane fragments remaining in the skin are unknown.⁸

CGMS have the potential to reduce diabetic complications

Glycemic fluctuations that occur throughout the day may be an independent risk factor in the development of diabetic complications.^9-11 Continuous monitoring that can detect such fluctuations could, potentially, reduce complications, but further studies are needed to determine whether CGMS users actually experience fewer complications. Several studies have shown a relationship between postpran-dial glucose fluctuations and macrovascular disease.^12-14 An analysis of data from the Diabetes Control and Complications Trial (DCCT) showed that A1C, mean blood glucose, and glycemic variability were independent risk factors for severe hypoglycemia.¹⁵ Reducing glycemic fluctuations may, therefore, reduce the risk of severe hypoglycemia.

CGMS data can change behavior, reduce hypoglycemia. The data a CGMS generates could be used to adjust medications or diet on the basis of real-time glucose levels, identify glucose trends, and aid in pattern management by providing retrospective, nearly continuous glucose values. One study evaluated the benefit of using a CGM in 90 type 1 and type 2 patients receiving insulin.⁴ All patients wore the monitor at home and at work during daily activities. Patients were randomized to a control group that was blinded to their glucose data and an experimental group that saw the display readings, could review trends, and received alerts and alarms from the system.

The results showed that the group that saw the display spent 21% less time in a hypoglycemic state and 26% more time in the target glycemic range than the control group. Nocturnal hypoglycemia was also significantly reduced in the group that had access to the display. These improvements were seen even though no prescribed plan to adjust therapy on the basis of glucose readings was in place, and must therefore have been the result of diet or insulin changes patients made on their own initiative in response to their CGM readings. Thus, in this study, providing more frequent glucose readings to patients improved safety of insulin and glycemic control.

Studies have also been done comparing the efficacy of CGM and traditional monitoring systems on hemoglobin A1C.¹⁶ These studies revealed a trend toward lower A1C with the use of CGMS, but the results were not statistically significant (0.22%; 95% confidence interval, -0.439% to 0.004%; P=.055).

Crossing the barriers to adoption
Before CGMS can become widespread in the primary care setting, barriers to their adoption must be addressed. Some clinicians continue to be dubious about the accuracy of the readings because CGMS measure interstitial glucose levels, rather than blood glucose. As we have seen, studies have been published that indicate a high level of accuracy for CGM readings, but more research needs to be done.

In the real world of the caregiver’s office, physicians and patients will have much to learn before CGMS come into widespread acceptance. Patients and providers both need to learn to use the new equipment and how to apply the data it provides. Physicians and patients will need to take account of the time lag before a CGMS reading catches up with a standard reading, and check with a standard blood glucose meter before making medication adjustments. Patients will need to understand the time to onset and peak of their insulins so that they can make appropriate adjustments.

Providers will have to find ways to incorporate the technology into their already busy clinical practice. Integrating CGMS data into electronic medical records or downloading data before scheduled office visits may streamline the process.

So where does this leave you, the busy family physician?
CGMS can provide useful information to select patients, making it possible for them to alter their diet and lifestyle choices and make better insulin treatment decisions. Although CGMS may not be able to eliminate the need for traditional self-monitoring of blood glucose entirely, using the 2 methods together does offer additional advantages. These new devices may help prevent hypo- and hyperglycemic episodes, improve patients’ quality of life, and potentially reduce the likelihood that complications will develop. Long-term studies will be necessary to confirm these potential benefits.

CORRESPONDENCE Rachel B. Hrabchak, PharmD, Clinical Assistant Professor, AHEC Pharmacy Coordinator, Department of Pharmacotherapy and Translational Research, University of Florida College of Pharmacy, 101 South Newell Drive, HPNP Building 212, Room 3309, Gainesville, FL 32611; hrabchak@cop.ufl.edu

We all know that the key to optimal diabetes management is tight glucose control, which can be achieved with multiple daily fingersticks, good record keeping of the results, and appropriate modification of the medication regimen, diet, and exercise schedule.

But patients find the routine burdensome, and many skip fingersticks or abandon the process entirely. And even those who follow the program faithfully may find that it fails to protect them from unpleasant and potentially dangerous episodes of hyper- and hypoglycemia.

The newer technology of continuous glucose monitoring systems (CGMS) offers the promise of overcoming these limitations. But how do these new systems work and what does the evidence tell us about their potential benefits and remaining uncertainties? Read on.

The old way: Take a snapshot
The variables that affect blood glucose levels—meals and snacks, exercise or the lack of it, dosages and timing of medication, and stress, among others—keep changing throughout the day and night. The impact of these variables cannot be adequately captured in snapshot blood glucose levels taken at isolated moments in the patient’s day. Achieving glycemic control with blood glucose monitors can be difficult for some patients, especially since the data generated are dependent on the patient’s willingness and ability to self-monitor his or her glucose levels.

The new way: Monitor continuously

CGMS measure the amount of glucose in the interstitial fluid—not in the blood. These measurements are taken every 5 minutes or so, depending on the system. Each system consists of a sensor, transmitter, and receiver. The sensor is a fine wire—about the diameter of 2 human hairs—that sticks into the skin of the abdomen or upper arm and is kept in place by an adhesive pad. The transmitter fits on the sensor pad and sends information to the receiver via radio waves. Sensors are disposable; they last for 3 to 5 days and then must be replaced. The system is wireless, so your patient isn’t tethered to the equipment.

Calibration with a glucose meter is still necessary. To be sure that interstitial glucose measurements reflect actual blood glucose levels, currently available systems require daily calibration with conventional blood glucose monitors. Patients will still have to do fingersticks, but far less frequently. The FDA has approved CGMS for use only in conjunction with conventional glucose testing. Traditional glucose self-monitoring may also be necessary when CGM results do not correspond to symptoms patients are experiencing.

Receiver displays data, can set off an alarm. Glucose measurements from the CGMS are displayed and stored in the receiver, and the data can be downloaded to a computer using the manufacturer’s data software. Continuous readings over a 24-hour period for up to 7 days allow the user to detect variation and identify trends. High and low glucose value thresholds can be customized for the individual patient and fed into the system. When these thresholds are exceeded, an alarm will sound. The receiver displays directional arrows showing the rate of change in glucose levels, allowing the patient to predict—and possibly prevent—hypoglycemic episodes.

Impact of events can be noted. The systems also allow for input of additional information about events that may affect blood glucose levels, such as medication, exercise, and food intake. Patients can use information about how these events affect their glucose levels to adjust the prandial or basal insulin dose, modify the insulin correction algorithm, or adjust their diet. Patients can bring computer-generated charts and graphs to office visits as a basis for joint decision-making about their care. Short-term, periodic use of a CGMS in patients with type 2 diabetes may identify times when patients need more frequent self-testing or guide further therapy selection.

These systems are available now

The systems available in the United States include:

• the iPro Continuous Glucose Monitor, Guardian Real-Time System, and Mini Med Paradigm Real-Time System—all from Medtronic, Inc.
• the SEVEN PLUS, from DexCom
• the FreeStyle Navigator, from Abbott.^1-3

The SEVEN PLUS and the FreeStyle Navigator are FDA approved for adults only. Pediatric versions of Medtronic’s MiniMed Paradigm and Guardian systems are approved for use in patients ages 7 to 17. All these systems require a prescription. For detailed comparisons of the features of these systems, see the TABLE.

TABLE
Continuous glucose monitoring systems: The options

	SEVEN PLUS	FreeStyle Navigator	Guardian Real-Time System	MiniMed Paradigm Real-Time System	iPro Continuous Glucose Monitor*
Manufacturer	DexCom	Abbott	Medtronic, Inc.	Medtronic, Inc.	Medtronic, Inc.
URL	www.dexcom.com	www.freestylenavigator.com	www.minimed.com	www.minimed.com	www.minimed.com
Price	$799 for system; $399 for 4 sensors; $79 for software	$1250 for system; $450 for 6 sensors	$1350 for system, including 4 sensors; $350 for 10 sensors	$999, plus cost of insulin pump; $35 per sensor	$1090 for start-up; $350 for 10 sensors
Receiver range	5 feet	10 feet	6 feet	6 feet
Sensor life	Up to 7 days	Up to 5 days	Up to 3 days	Up to 3 days	Up to 3 days
Calibration	2 hours after insertion, then every 12 hours	At least 4 times over a 5-day period at 10, 12, 24, and 72 hours after insertion	2 hours after insertion, again within 6 hours, then every 12 hours	2 hours after insertion, then within next 6 hours, then every 12 hours
User-set alarm for highs/lows	Yes, plus factory alarm at 55 mg/dL that can’t be disabled	Yes	Yes	Yes
Glucose reading display frequency	Every 5 minutes	Once every minute	Measures every minute, displays an average of every 5 minutes	Measures every minute, displays an average of every 5 minutes
Displays directional trends	Yes	Yes	Yes	Yes
Sources: Diabetes Network. Diabetes technology. Available at: www.diabetesnet.com/diabetes_technology/continuous_monitoring.php. Accessed January 6, 2010.
DexCom. Available at: www.dexcom.com. Accessed January 6, 2010.
FreeStyle Navigator. Available at: www.freestylenavigator.com. Accessed January 6, 2010.
Medtronic. Available at: www.minimed.com. Accessed January 6, 2010.
Conversations with Robert Sala, sales representative, DexCom, on May 1 and May 8, 2009.
* iPro consists of sensor and transmitter only; no receiver. Sensor is inserted by provider; data are uploaded in provider’s office to help guide therapeutic decision-making.

Patients with severe diabetes benefit most
Patients with type 1 diabetes who use an insulin pump or are being switched from multiple injections to pumps, and patients who have problems with hypoglycemia are good candidates for CGMS. The latter group includes those who are not aware of their hypoglycemic state, those who have nighttime hypoglycemia, and those who experience severe episodes of hypoglycemia. The category also includes patients who keep their blood glucose levels higher than appropriate goals would indicate, because of their fear of hypoglycemia.

An additional group of patients who might benefit, although they do not fit currently approved indications for these devices, are pregnant women who should maintain tight glucose control. Other patients who might find CGMS useful are those with glycemic variability or those who have not achieved their A1C goal and want to be proactive.

Your letter of medical necessity can qualify patients like these for Medicare or private insurance reimbursement for the CGMS and for ongoing sensor supplies. You may also choose to purchase a system yourself for patients to use, and bill the patient’s insurance company for the service.

Accuracy continues to be a concern
Currently available systems are more accurate than the first generation of these devices. When glucose is rapidly changing, users need to be aware that there may be a time lag before the interstitial glucose reaches the same level as the blood glucose. So, while medication changes can be made based on CGMS, values should be confirmed with a fingerstick.

SEVEN and Navigator are comparable
A number of studies have confirmed the accuracy of CGMS.^4-7 A study by Garg and colleagues compared the accuracy of the DexCom SEVEN and the FreeStyle Navigator.⁶ Fourteen patients wore sensors from both systems for 3 consecutive, 5-day periods. Laboratory reference measurements of venous blood glucose were taken every 15 minutes through an 8-hour period on days 5, 10, and 15 in clinic using the YSI STAT Plus Glucose Analyzer. Sensors were replaced at the end of the clinic day on days 5 and 10, and the sensors were removed at the end of day 15. The mean absolute relative difference for CGM compared with laboratory glucose measures was 16.8% for the SEVEN and 16.1% for the Navigator (P=.38), an insignificant difference between the 2 systems.

The 2 systems were also compared using continuous glucose error grid analysis, which evaluates how accurately CGM data lead to an appropriate clinical response by the patient. The error grid is divided into 5 zones and superimposed on the correlation plot. Plots in Zone A are a perfect fit and plots in Zone B are “benign error” that does not result in an inaccurate clinical response. The percentage of data points in Zones A or B was 94.8% for the SEVEN and 93.2% for the Navigator. The SEVEN provided better agreement with laboratory glucose measures for the range 40 to 80 mg/dL (P<.001).

Guardian evaluation has similar results
A similar study done by Medtronic in 2004 evaluated the Guardian RT, an earlier version of the Guardian, in 16 patients.⁷ Values from the Guardian RT were compared with reference YSI STAT Plus glucose analyzer glucose values taken every 30 minutes in clinic. The mean absolute percent difference was 19.7%±18.4%. Of the 3941 total paired glucose measurements, 96% fell in the clinically acceptable error grid Zones A or B. For low glucose values between 40 and 80 mg/dL, 76.1% of readings fell in Zones A or B; for high values, over 240 mg/dL, 86.8% of readings fell in Zones A or B. Accuracy in the hypoglycemic ranges declined as the time increased from insertion of the sensor.

Safety risks are few, minor

Insertion of the sensor can pose minor safety risks, including infection, inflammation, and bleeding. Adverse events reported in 1 study consisted mainly of mild sensor site reactions such as blisters, bullae, edema, and erythema, none of which required treatment.⁶ The CGMS must be removed prior to magnetic imaging studies and the devices are not approved for use on airplanes. When the FreeStyle Navigator sensor is removed, a portion of the membrane polymer is left in the skin. The company reports no health effects in clinical studies, aside from sensor site reactions mentioned above, but long-term effects of sensor membrane fragments remaining in the skin are unknown.⁸

CGMS have the potential to reduce diabetic complications

Glycemic fluctuations that occur throughout the day may be an independent risk factor in the development of diabetic complications.^9-11 Continuous monitoring that can detect such fluctuations could, potentially, reduce complications, but further studies are needed to determine whether CGMS users actually experience fewer complications. Several studies have shown a relationship between postpran-dial glucose fluctuations and macrovascular disease.^12-14 An analysis of data from the Diabetes Control and Complications Trial (DCCT) showed that A1C, mean blood glucose, and glycemic variability were independent risk factors for severe hypoglycemia.¹⁵ Reducing glycemic fluctuations may, therefore, reduce the risk of severe hypoglycemia.

CGMS data can change behavior, reduce hypoglycemia. The data a CGMS generates could be used to adjust medications or diet on the basis of real-time glucose levels, identify glucose trends, and aid in pattern management by providing retrospective, nearly continuous glucose values. One study evaluated the benefit of using a CGM in 90 type 1 and type 2 patients receiving insulin.⁴ All patients wore the monitor at home and at work during daily activities. Patients were randomized to a control group that was blinded to their glucose data and an experimental group that saw the display readings, could review trends, and received alerts and alarms from the system.

The results showed that the group that saw the display spent 21% less time in a hypoglycemic state and 26% more time in the target glycemic range than the control group. Nocturnal hypoglycemia was also significantly reduced in the group that had access to the display. These improvements were seen even though no prescribed plan to adjust therapy on the basis of glucose readings was in place, and must therefore have been the result of diet or insulin changes patients made on their own initiative in response to their CGM readings. Thus, in this study, providing more frequent glucose readings to patients improved safety of insulin and glycemic control.

Studies have also been done comparing the efficacy of CGM and traditional monitoring systems on hemoglobin A1C.¹⁶ These studies revealed a trend toward lower A1C with the use of CGMS, but the results were not statistically significant (0.22%; 95% confidence interval, -0.439% to 0.004%; P=.055).

Crossing the barriers to adoption
Before CGMS can become widespread in the primary care setting, barriers to their adoption must be addressed. Some clinicians continue to be dubious about the accuracy of the readings because CGMS measure interstitial glucose levels, rather than blood glucose. As we have seen, studies have been published that indicate a high level of accuracy for CGM readings, but more research needs to be done.

In the real world of the caregiver’s office, physicians and patients will have much to learn before CGMS come into widespread acceptance. Patients and providers both need to learn to use the new equipment and how to apply the data it provides. Physicians and patients will need to take account of the time lag before a CGMS reading catches up with a standard reading, and check with a standard blood glucose meter before making medication adjustments. Patients will need to understand the time to onset and peak of their insulins so that they can make appropriate adjustments.

Providers will have to find ways to incorporate the technology into their already busy clinical practice. Integrating CGMS data into electronic medical records or downloading data before scheduled office visits may streamline the process.

So where does this leave you, the busy family physician?
CGMS can provide useful information to select patients, making it possible for them to alter their diet and lifestyle choices and make better insulin treatment decisions. Although CGMS may not be able to eliminate the need for traditional self-monitoring of blood glucose entirely, using the 2 methods together does offer additional advantages. These new devices may help prevent hypo- and hyperglycemic episodes, improve patients’ quality of life, and potentially reduce the likelihood that complications will develop. Long-term studies will be necessary to confirm these potential benefits.

CORRESPONDENCE Rachel B. Hrabchak, PharmD, Clinical Assistant Professor, AHEC Pharmacy Coordinator, Department of Pharmacotherapy and Translational Research, University of Florida College of Pharmacy, 101 South Newell Drive, HPNP Building 212, Room 3309, Gainesville, FL 32611; hrabchak@cop.ufl.edu

We all know that the key to optimal diabetes management is tight glucose control, which can be achieved with multiple daily fingersticks, good record keeping of the results, and appropriate modification of the medication regimen, diet, and exercise schedule.

But patients find the routine burdensome, and many skip fingersticks or abandon the process entirely. And even those who follow the program faithfully may find that it fails to protect them from unpleasant and potentially dangerous episodes of hyper- and hypoglycemia.

The newer technology of continuous glucose monitoring systems (CGMS) offers the promise of overcoming these limitations. But how do these new systems work and what does the evidence tell us about their potential benefits and remaining uncertainties? Read on.

The old way: Take a snapshot
The variables that affect blood glucose levels—meals and snacks, exercise or the lack of it, dosages and timing of medication, and stress, among others—keep changing throughout the day and night. The impact of these variables cannot be adequately captured in snapshot blood glucose levels taken at isolated moments in the patient’s day. Achieving glycemic control with blood glucose monitors can be difficult for some patients, especially since the data generated are dependent on the patient’s willingness and ability to self-monitor his or her glucose levels.

The new way: Monitor continuously

CGMS measure the amount of glucose in the interstitial fluid—not in the blood. These measurements are taken every 5 minutes or so, depending on the system. Each system consists of a sensor, transmitter, and receiver. The sensor is a fine wire—about the diameter of 2 human hairs—that sticks into the skin of the abdomen or upper arm and is kept in place by an adhesive pad. The transmitter fits on the sensor pad and sends information to the receiver via radio waves. Sensors are disposable; they last for 3 to 5 days and then must be replaced. The system is wireless, so your patient isn’t tethered to the equipment.

Calibration with a glucose meter is still necessary. To be sure that interstitial glucose measurements reflect actual blood glucose levels, currently available systems require daily calibration with conventional blood glucose monitors. Patients will still have to do fingersticks, but far less frequently. The FDA has approved CGMS for use only in conjunction with conventional glucose testing. Traditional glucose self-monitoring may also be necessary when CGM results do not correspond to symptoms patients are experiencing.

Receiver displays data, can set off an alarm. Glucose measurements from the CGMS are displayed and stored in the receiver, and the data can be downloaded to a computer using the manufacturer’s data software. Continuous readings over a 24-hour period for up to 7 days allow the user to detect variation and identify trends. High and low glucose value thresholds can be customized for the individual patient and fed into the system. When these thresholds are exceeded, an alarm will sound. The receiver displays directional arrows showing the rate of change in glucose levels, allowing the patient to predict—and possibly prevent—hypoglycemic episodes.

Impact of events can be noted. The systems also allow for input of additional information about events that may affect blood glucose levels, such as medication, exercise, and food intake. Patients can use information about how these events affect their glucose levels to adjust the prandial or basal insulin dose, modify the insulin correction algorithm, or adjust their diet. Patients can bring computer-generated charts and graphs to office visits as a basis for joint decision-making about their care. Short-term, periodic use of a CGMS in patients with type 2 diabetes may identify times when patients need more frequent self-testing or guide further therapy selection.

These systems are available now

The systems available in the United States include:

• the iPro Continuous Glucose Monitor, Guardian Real-Time System, and Mini Med Paradigm Real-Time System—all from Medtronic, Inc.
• the SEVEN PLUS, from DexCom
• the FreeStyle Navigator, from Abbott.^1-3

The SEVEN PLUS and the FreeStyle Navigator are FDA approved for adults only. Pediatric versions of Medtronic’s MiniMed Paradigm and Guardian systems are approved for use in patients ages 7 to 17. All these systems require a prescription. For detailed comparisons of the features of these systems, see the TABLE.

TABLE
Continuous glucose monitoring systems: The options

	SEVEN PLUS	FreeStyle Navigator	Guardian Real-Time System	MiniMed Paradigm Real-Time System	iPro Continuous Glucose Monitor*
Manufacturer	DexCom	Abbott	Medtronic, Inc.	Medtronic, Inc.	Medtronic, Inc.
URL	www.dexcom.com	www.freestylenavigator.com	www.minimed.com	www.minimed.com	www.minimed.com
Price	$799 for system; $399 for 4 sensors; $79 for software	$1250 for system; $450 for 6 sensors	$1350 for system, including 4 sensors; $350 for 10 sensors	$999, plus cost of insulin pump; $35 per sensor	$1090 for start-up; $350 for 10 sensors
Receiver range	5 feet	10 feet	6 feet	6 feet
Sensor life	Up to 7 days	Up to 5 days	Up to 3 days	Up to 3 days	Up to 3 days
Calibration	2 hours after insertion, then every 12 hours	At least 4 times over a 5-day period at 10, 12, 24, and 72 hours after insertion	2 hours after insertion, again within 6 hours, then every 12 hours	2 hours after insertion, then within next 6 hours, then every 12 hours
User-set alarm for highs/lows	Yes, plus factory alarm at 55 mg/dL that can’t be disabled	Yes	Yes	Yes
Glucose reading display frequency	Every 5 minutes	Once every minute	Measures every minute, displays an average of every 5 minutes	Measures every minute, displays an average of every 5 minutes
Displays directional trends	Yes	Yes	Yes	Yes
Sources: Diabetes Network. Diabetes technology. Available at: www.diabetesnet.com/diabetes_technology/continuous_monitoring.php. Accessed January 6, 2010.
DexCom. Available at: www.dexcom.com. Accessed January 6, 2010.
FreeStyle Navigator. Available at: www.freestylenavigator.com. Accessed January 6, 2010.
Medtronic. Available at: www.minimed.com. Accessed January 6, 2010.
Conversations with Robert Sala, sales representative, DexCom, on May 1 and May 8, 2009.
* iPro consists of sensor and transmitter only; no receiver. Sensor is inserted by provider; data are uploaded in provider’s office to help guide therapeutic decision-making.

Patients with severe diabetes benefit most
Patients with type 1 diabetes who use an insulin pump or are being switched from multiple injections to pumps, and patients who have problems with hypoglycemia are good candidates for CGMS. The latter group includes those who are not aware of their hypoglycemic state, those who have nighttime hypoglycemia, and those who experience severe episodes of hypoglycemia. The category also includes patients who keep their blood glucose levels higher than appropriate goals would indicate, because of their fear of hypoglycemia.

An additional group of patients who might benefit, although they do not fit currently approved indications for these devices, are pregnant women who should maintain tight glucose control. Other patients who might find CGMS useful are those with glycemic variability or those who have not achieved their A1C goal and want to be proactive.

Your letter of medical necessity can qualify patients like these for Medicare or private insurance reimbursement for the CGMS and for ongoing sensor supplies. You may also choose to purchase a system yourself for patients to use, and bill the patient’s insurance company for the service.

Accuracy continues to be a concern
Currently available systems are more accurate than the first generation of these devices. When glucose is rapidly changing, users need to be aware that there may be a time lag before the interstitial glucose reaches the same level as the blood glucose. So, while medication changes can be made based on CGMS, values should be confirmed with a fingerstick.

SEVEN and Navigator are comparable
A number of studies have confirmed the accuracy of CGMS.^4-7 A study by Garg and colleagues compared the accuracy of the DexCom SEVEN and the FreeStyle Navigator.⁶ Fourteen patients wore sensors from both systems for 3 consecutive, 5-day periods. Laboratory reference measurements of venous blood glucose were taken every 15 minutes through an 8-hour period on days 5, 10, and 15 in clinic using the YSI STAT Plus Glucose Analyzer. Sensors were replaced at the end of the clinic day on days 5 and 10, and the sensors were removed at the end of day 15. The mean absolute relative difference for CGM compared with laboratory glucose measures was 16.8% for the SEVEN and 16.1% for the Navigator (P=.38), an insignificant difference between the 2 systems.

The 2 systems were also compared using continuous glucose error grid analysis, which evaluates how accurately CGM data lead to an appropriate clinical response by the patient. The error grid is divided into 5 zones and superimposed on the correlation plot. Plots in Zone A are a perfect fit and plots in Zone B are “benign error” that does not result in an inaccurate clinical response. The percentage of data points in Zones A or B was 94.8% for the SEVEN and 93.2% for the Navigator. The SEVEN provided better agreement with laboratory glucose measures for the range 40 to 80 mg/dL (P<.001).

Guardian evaluation has similar results
A similar study done by Medtronic in 2004 evaluated the Guardian RT, an earlier version of the Guardian, in 16 patients.⁷ Values from the Guardian RT were compared with reference YSI STAT Plus glucose analyzer glucose values taken every 30 minutes in clinic. The mean absolute percent difference was 19.7%±18.4%. Of the 3941 total paired glucose measurements, 96% fell in the clinically acceptable error grid Zones A or B. For low glucose values between 40 and 80 mg/dL, 76.1% of readings fell in Zones A or B; for high values, over 240 mg/dL, 86.8% of readings fell in Zones A or B. Accuracy in the hypoglycemic ranges declined as the time increased from insertion of the sensor.

Safety risks are few, minor

Insertion of the sensor can pose minor safety risks, including infection, inflammation, and bleeding. Adverse events reported in 1 study consisted mainly of mild sensor site reactions such as blisters, bullae, edema, and erythema, none of which required treatment.⁶ The CGMS must be removed prior to magnetic imaging studies and the devices are not approved for use on airplanes. When the FreeStyle Navigator sensor is removed, a portion of the membrane polymer is left in the skin. The company reports no health effects in clinical studies, aside from sensor site reactions mentioned above, but long-term effects of sensor membrane fragments remaining in the skin are unknown.⁸

CGMS have the potential to reduce diabetic complications

Glycemic fluctuations that occur throughout the day may be an independent risk factor in the development of diabetic complications.^9-11 Continuous monitoring that can detect such fluctuations could, potentially, reduce complications, but further studies are needed to determine whether CGMS users actually experience fewer complications. Several studies have shown a relationship between postpran-dial glucose fluctuations and macrovascular disease.^12-14 An analysis of data from the Diabetes Control and Complications Trial (DCCT) showed that A1C, mean blood glucose, and glycemic variability were independent risk factors for severe hypoglycemia.¹⁵ Reducing glycemic fluctuations may, therefore, reduce the risk of severe hypoglycemia.

CGMS data can change behavior, reduce hypoglycemia. The data a CGMS generates could be used to adjust medications or diet on the basis of real-time glucose levels, identify glucose trends, and aid in pattern management by providing retrospective, nearly continuous glucose values. One study evaluated the benefit of using a CGM in 90 type 1 and type 2 patients receiving insulin.⁴ All patients wore the monitor at home and at work during daily activities. Patients were randomized to a control group that was blinded to their glucose data and an experimental group that saw the display readings, could review trends, and received alerts and alarms from the system.

The results showed that the group that saw the display spent 21% less time in a hypoglycemic state and 26% more time in the target glycemic range than the control group. Nocturnal hypoglycemia was also significantly reduced in the group that had access to the display. These improvements were seen even though no prescribed plan to adjust therapy on the basis of glucose readings was in place, and must therefore have been the result of diet or insulin changes patients made on their own initiative in response to their CGM readings. Thus, in this study, providing more frequent glucose readings to patients improved safety of insulin and glycemic control.

Studies have also been done comparing the efficacy of CGM and traditional monitoring systems on hemoglobin A1C.¹⁶ These studies revealed a trend toward lower A1C with the use of CGMS, but the results were not statistically significant (0.22%; 95% confidence interval, -0.439% to 0.004%; P=.055).

Crossing the barriers to adoption
Before CGMS can become widespread in the primary care setting, barriers to their adoption must be addressed. Some clinicians continue to be dubious about the accuracy of the readings because CGMS measure interstitial glucose levels, rather than blood glucose. As we have seen, studies have been published that indicate a high level of accuracy for CGM readings, but more research needs to be done.

In the real world of the caregiver’s office, physicians and patients will have much to learn before CGMS come into widespread acceptance. Patients and providers both need to learn to use the new equipment and how to apply the data it provides. Physicians and patients will need to take account of the time lag before a CGMS reading catches up with a standard reading, and check with a standard blood glucose meter before making medication adjustments. Patients will need to understand the time to onset and peak of their insulins so that they can make appropriate adjustments.

Providers will have to find ways to incorporate the technology into their already busy clinical practice. Integrating CGMS data into electronic medical records or downloading data before scheduled office visits may streamline the process.

So where does this leave you, the busy family physician?
CGMS can provide useful information to select patients, making it possible for them to alter their diet and lifestyle choices and make better insulin treatment decisions. Although CGMS may not be able to eliminate the need for traditional self-monitoring of blood glucose entirely, using the 2 methods together does offer additional advantages. These new devices may help prevent hypo- and hyperglycemic episodes, improve patients’ quality of life, and potentially reduce the likelihood that complications will develop. Long-term studies will be necessary to confirm these potential benefits.

CORRESPONDENCE Rachel B. Hrabchak, PharmD, Clinical Assistant Professor, AHEC Pharmacy Coordinator, Department of Pharmacotherapy and Translational Research, University of Florida College of Pharmacy, 101 South Newell Drive, HPNP Building 212, Room 3309, Gainesville, FL 32611; hrabchak@cop.ufl.edu