Since 1963, when Starzl et al performed the first successful liver transplantation,¹ outcomes of this life-saving procedure have continued to improve. Long-term survival rates have increased markedly: the current 5-year rate is 73.8% and the 10-year rate is 60%.²

This success means that internists will be caring for a greater number of liver transplant recipients and managing their long-term problems, such as hypertension, diabetes mellitus, dyslipidemia, obesity, metabolic syndrome, cardiovascular disease, renal insufficiency, osteoporosis, cancer, and gout.

This review will discuss these complications, focusing on the role the primary care physician assumes beyond the first year after transplantation.

ROLE OF THE PRIMARY CARE PHYSICIAN

Hepatologists, primary care physicians, and surgeons share the care of transplant recipients. The first several weeks after transplantation require close follow-up by the hepatologist and transplantation team, with particular attention paid to the patient’s overall health and well-being, medication compliance, and biochemical and immunosuppression monitoring.

After the first year, the primary care physician assumes a greater role, becoming the main provider of the patient’s care.^3,4 Good communication between the transplant center and the primary care physician should lead to a smooth transition.⁴ Although the hepatologist continues to manage immunosuppressive drugs, allograft rejections, and biliary complications, the primary care physician manages most of the long-term complications and thus needs to be aware of the common ones and feel comfortable managing them. Aims during visits are to screen for and detect common complications and manage them appropriately, in addition to performing annual physical examinations and routine health care. A reasonable interval for liver transplant recipients to visit their primary care physician is every 6 months.

IMMUNOSUPPRESSANT MEDICATIONS

Multiple agents are used for immunosuppression after liver transplantation:

• Calcineurin inhibitors (cyclosporine and tacrolimus)
• Antimetabolites (mycophenolate mofetil, azathioprine, and mycophenolate sodium)
• Mammalian target of rapamycin (mTOR) inhibitors (sirolimus and everolimus)
• Corticosteroids.

Table 1 lists their common side effects.

Most centers use a combination of two to four immunosuppressants as induction therapy in the immediate posttransplant period, then taper the doses and eliminate all but a calcineurin inhibitor and an antimetabolite. For example, some start with a combination of tacrolimus, mycophenolate mofetil, and a corticosteroid. The choice in the immediate posttransplant period is frequently made by the transplant center in cooperation with the hepatologist. By the time primary care physicians see these patients, they usually are on a calcineurin inhibitor alone or a calcineurin inhibitor plus mycophenolate mofetil.

Calcineurin inhibitors

Cyclosporine is metabolized by the cytochrome CYP3A4 pathway. With an average half-life of 15 hours, it is given orally, usually every 12 hours.

The dosage is adjusted according to the trough level. Higher levels are needed in the initial posttransplant period to prevent graft rejection, whereas lower levels are preferred later to decrease the occurrence and severity of adverse effects. Typical long-term trough levels are 50 to 100 ng/mL. Levels should be checked more often if an acute illness develops or the patient starts taking a potentially interfering drug.

Of importance: the dosage should be based on trough levels and not on random levels. Levels are often falsely high if blood samples are not drawn at the trough level. Repeating the measurement and making sure the sample is drawn at the trough level, ie, 12 hours after the last dose, is advised in this condition.

Cyclosporine causes widespread vasoconstriction resulting in decreased renal blood flow and systemic hypertension, often within a few days of starting it. Other important adverse effects include renal insufficiency, dyslipidemia, neurotoxicity (headache, tremor, seizure), and diabetes.

Tacrolimus is superior to cyclosporine in terms of survival, graft loss, acute rejection, and steroid-resistant rejection in the first year.⁵ Currently, it is the agent used most often for maintenance immunosuppression after liver transplantation.

Like cyclosporine, tacrolimus is metabolized in the liver by CYP3A4. Satisfactory trough levels after 1 year are 4 to 6 ng/mL.

The adverse effects of tacrolimus are similar to those of cyclosporine, but diabetes mellitus is more common with tacrolimus. Bone marrow suppression may occur more often with tacrolimus as well.

Antimetabolites

Antimetabolites are generally not potent enough to be used alone.

Mycophenolate mofetil causes adverse effects that include bone marrow suppression and gastrointestinal symptoms such as gastritis, diarrhea, and abdominal pain.

Azathioprine, infrequently used in transplantation in the United States, is nevertheless sometimes substituted for mycophenolate mofetil in pregnant women, as it seems safer for use in pregnancy.

Serum levels of azathioprine and mycophenolate mofetil are not routinely monitored.

mTOR inhibitors

Sirolimus and everolimus are mTOR inhibitors, inhibiting proliferation of lymphocytes.^6,7

Unlike calcineurin inhibitors, mTOR inhibitors are not associated with nephrotoxicity, neurotoxicity, renal dysfunction, hypertension, or diabetes. Sirolimus is considered an alternative to calcineurin inhibitors or, in some instances, used as add-on-therapy to lower the dose of the calcineurin inhibitor.

Sirolimus carries a black-box warning about hepatic artery thrombosis

However, sirolimus carries a potential risk of hepatic artery thrombosis, a life-threatening complication.⁸ This has led the US Food and Drug Administration (FDA) to require sirolimus to carry a black-box warning, and most transplant centers avoid using it in the first 30 days after transplantation.

Dyslipidemia is perhaps the most common adverse effect of sirolimus. Others include dose-related cytopenia and wound dehiscence.⁹

Everolimus has yet to be established for use in liver transplantation, although safety trials have been published.^10,11 The FDA currently recommends against using it in the first 30 days after liver transplantation.

Both sirolimus and everolimus are metabolized by CYP3A4, which is the same metabolic pathway used by cyclosporine and tacrolimus. Hence, drugs that inhibit CYP3A4 may significantly impair clearance of both sirolimus and everolimus.

Corticosteroids

Corticosteroids have been the cornerstone of immunosuppression and remain the first line of treatment for acute allograft rejection. High intravenous doses of corticosteroids are usually started in the peritransplant period and are then switched to oral doses, which are tapered and continued with a fixed dose such as 20 mg of prednisone daily for 3 to 6 months after transplantation. However, some transplant centers keep patients on prednisone 5 mg/day indefinitely.

Adverse effects of corticosteroids include diabetes, salt and fluid retention, hypertension, hyperlipidemia, cosmetic changes (acne, cervical fat pad or “buffalo hump”), delayed wound healing, susceptibility to infection, cataracts, osteopenia, and potential adrenal suppression.¹² There is concern that the use of these drugs may increase hepatitis C virus replication in patients who received a liver transplant for hepatitis C cirrhosis. Randomized trials have yielded conflicting results.^13–15

Drug interactions

Certain drugs can affect the metabolism of calcineurin inhibitors and mTOR inhibitors by inducing CYP3A4, which results in decreasing the levels of the immunosuppressive drugs, or by inhibiting CYP3A4, which has the opposite effect.

Medications that can decrease the levels of calcineurin inhibitors and mTOR inhibitors:

• Anticonvulsants (carbamazepine, phenobarbital, phenytoin)
• Antibiotics (rifampin, isoniazid)
• St John’s wort.

Medications that can increase the levels of calcineurin inhibitors and mTOR inhibitors:

• Antifungals (fluconazole, ketoconazole, itraconazole, voriconazole, aspofungin)
• Antibiotics (azithromycin, erythromycin, clarithromycin)
• Nondihydropyridine calcium channel blockers (diltiazem, verapamil).¹⁶

Selected antibiotics are generally well tolerated, such as penicillins, cephalosporins, quinolones, sulfonamides, and topical antifungal agents.

LONG-TERM COMPLICATIONS

Figure 1 summarizes the common long-term complications of liver transplantation.

Hypertension

The prevalence of hypertension after liver transplantation is 40% to 85%, which is markedly higher than in patients with chronic liver disease before liver transplantation.^17,18

One of the factors contributing to this increase is the use of immunosuppressive medications. Of these drugs, cyclosporine seems to be the one that most often causes an increase in both the incidence and the severity of hypertension, as it produces widespread vasoconstriction.¹⁹ Corticosteroids cause hypertension through their mineralocorticoid effects.

The diagnostic cutoffs for hypertension (ie, 140/90 mm Hg) and the treatment goals in posttransplant patients are similar to those in the general population. However, at our institution we target a blood pressure of less than 130/80 mm Hg in transplant patients because they have a high prevalence of other cardiovascular risk factors such as diabetes, obesity, and renal insufficiency.²⁰

Dihydropyridine calcium channel blockers such as amlodipine and nifedipine are considered the best first-line agents because they dilate renal afferent arterioles, an effect that may counteract the vasoconstriction mediated by calcineurin inhibitors. Nondihydropyridine calcium channel blockers such as diltiazem and verapamil tend to have more marked negative inotropic effects and are not recommended in liver transplant recipients because they increase the levels of calcineurin inhibitors.²¹

Diuretics (eg, furosemide) might be the second-line agents, especially in patients with peripheral edema.¹⁶ One should be vigilant for hyperuricemia if thiazide agents are used.

Angiotensin-converting-enzyme (ACE) inhibitors and angiotensin II receptor blockers (ARBs) are typically avoided in the early posttransplant period, but they can be started later and have additional benefits in patients with diabetes and congestive heart failure. Starting ACE inhibitors is acceptable in these patients unless there is a contraindication such as allergy to ACE inhibitors, hypotension, history of bilateral renal artery stenosis, significant hyperkalemia, or acute kidney injury. Monitor the serum potassium level closely for hyperkalemia in patients concurrently using calcineurin inhibitors.

Alpha-blockers and beta-blockers can be used as add-on therapy in patients with uncontrolled hypertension with the exception of carvedilol, because it increases the levels of calcineurin inhibitors.²²

Blood pressure monitoring by the primary care physician is recommended every 6 months after the early posttransplant period, or more frequently when changes in treatment are being considered.

If hypertension continues to be inadequately controlled despite treatment, changing the immunosuppressive drugs or decreasing the doses can be considered, but the transplant hepatologist must be involved in this decision.^23,24

Diabetes mellitus

The prevalence of diabetes mellitus is higher in liver transplant recipients than in the general population, reaching 30% to 40%.^17,25 In addition to preexisting diabetes, 15% of liver transplant recipients develop new-onset diabetes.^26,27

Risk factors for developing diabetes after liver transplantation include African American or Hispanic ethnicity, obesity, family history, pretransplant diabetes, hepatitis C virus infection, use of corticosteroids, and use of calcineurin inhibitors (tacrolimus more than cyclosporine) and sirolimus.²⁶

In addition to increasing the risk of cardiovascular disease and other diseases, diabetes decreases both patient and graft survival after liver transplantation.²⁸

The management of diabetes and the treatment target after transplantation should follow the American Diabetes Association guidelines for the treatment of type 2 diabetes mellitus.²⁹ Lifestyle modifications, diet, and exercise are as important for transplant patients as for the nontransplant population. Insulin therapy is usually needed in the early posttransplant period to control blood glucose levels well, especially with the high doses of corticosteroids used during the first few weeks. No trials to date have compared oral agents in posttransplant patients. Therefore, the choice of oral hypoglycemic agents should be individualized on the basis of the patient’s characteristics and comorbidities.

Screen all liver transplant recipients for diabetes regardless of their pretransplant status

We recommend that primary care providers screen all liver transplant recipients for diabetes regardless of their pretransplant status. This can be done by obtaining regular fasting blood glucose levels or a hemoglobin A_1c level every 6 months. Additionally, liver transplant recipients diagnosed with diabetes require annual eye examinations to look for cataracts and diabetes-related changes.³⁰

Dyslipidemia

On November 12, 2013, the American College of Cardiology and the American Heart Association (ACC/AHA) released new clinical practice guidelines for treating blood cholesterol levels.^31,32 According to these new guidelines, there are four groups of patients for whom treatment with statins is clearly indicated:

• Patients with cardiovascular disease
• Patients with low-density lipoprotein cholesterol (LDL-C) levels ≥ 190 mg/dL
• Patients 40 to 75 years old with type 2 diabetes
• Patients 40 to 75 years old with an estimated 10-year risk of cardiovascular disease of 7.5% or greater.

Liver transplant recipients should be evaluated on an individual basis to see if they fit in any of the four groups and if statin treatment therefore needs to be initiated.

A few things need to be kept in mind. First, the incidence of dyslipidemia after liver transplantation is estimated to be 45% to 69%. Risk factors include obesity, diabetes mellitus, cholestatic liver disease, and immunosuppressant medications.³³ Sirolimus has a significant and well-documented association with dyslipidemia. Cyclosporine and corticosteroids are also strongly associated with dyslipidemia. Tacrolimus has a minor effect, and mycophenolate mofetil and azathioprine have no significant effects on serum lipid levels.¹⁶

Second, of the seven currently marketed statins, pravastatin and fluvastatin are preferred in liver transplant recipients because they are not metabolized by the same cytochrome CYP3A4 pathway that metabolizes calcineurin inhibitors and sirolimus.³⁴ The doses of 40 to 80 mg daily of pravastatin or 40 mg twice daily of fluvastatin lower low-density lipoprotein cholesterol (LDL-C) levels by approximately 30% to 35%. However, these two agents are considered “moderate-intensity” statins according to the new ACC/AHA guidelines. The only two “high-intensity” statins are atorvastatin (40–80 mg) and rosuvastatin (20–40 mg), but they are both metabolized by CYP3A4. Therefore it is prudent to avoid them with the concurrent use of a calcineurin inhibitor or tacrolimus.

Gemfibrozil does not lower LDL-C and should not be used concomitantly with statins due to unacceptable risk of rhabdomyolysis and myopathy. Fenofibrates are usually avoided due to potential nephrotoxicity in patients receiving cyclosporine. Bile acid sequestrants (cholestyramine, colestipol, colesevelam) can decrease plasma mycophenolate mofetil levels by 35%.^16,35 Thus, these agents should be avoided if mycophenolate mofetil is used.

It is reasonable to screen all liver transplant recipients with a fasting lipid profile at 3, 6, and 12 months after transplantation and annually thereafter. Creatine kinase should be measured if the patient complains of severe muscle pain or weakness but not on a routine basis.

Obesity

Approximately one-third of patients who are of normal weight at the time of transplantation will become obese afterward.^18,25 Corticosteroid use is an important risk factor for posttransplant obesity, and tapering these drugs helps reduce weight.³⁶ Patients treated with cyclosporine are more likely to gain weight than those who receive tacrolimus.³⁷

Of importance: nonalcoholic fatty liver disease, currently the most common cause of chronic liver disease in adults, is rapidly increasing as an indication for liver transplantation. In fact, the proportion of liver transplantation procedures for nonalcoholic steatohepatitis-related cirrhosis increased from 1.2% in 2001 to 9.7% in 2009, and nonalcoholic steatohepatitis is expected to become the leading indication for liver transplantation in the next 20 years. And because nonalcoholic fatty liver disease is directly linked to obesity, the prevalence of obesity as a complication of liver transplantation will most likely increase in the near future.

Overweight liver transplant recipients may have great difficulty losing weight. Treatment starts with patient education on caloric restriction and exercise. If traditional measures fail to result in adequate weight loss, additional options include switching from cyclosporine to tacrolimus.²³

Bariatric surgery may become an option for posttransplant patients. In a recent case series from the University of Minnesota, Al-Nowaylati et al³⁸ reported their experience with seven patients who underwent orthotopic liver transplantation and then open Roux-en-Y gastric bypass. After bariatric surgery, the patients’ mean body mass index declined significantly, and glycemic control and high-density lipoprotein cholesterol (HDL-C) levels improved. However, one patient died of multiple organ failure, to which the bariatric surgery might have contributed.³⁸

Heimbach et al³⁹ conducted a study in patients referred for liver transplantation for whom a rigorous noninvasive weight-loss program before transplantation had failed. The researchers performed combined liver transplantation and sleeve gastrectomy in seven carefully selected patients who had failed to achieve weight loss to a body mass index less than 35 kg/m² before transplantation. All seven patients lost weight, decreasing their mean body mass index from 49 kg/m² before the procedure to 29 kg/m² at last follow-up, and none of them developed posttransplant diabetes or steatosis.

At this time, there is not enough evidence to recommend concurrent orthotopic liver transplantation plus bariatric surgery, or combined orthotopic liver transplantation and sleeve gastrectomy. More study is needed to further evaluate these advanced approaches.

Posttransplant metabolic syndrome

Metabolic syndrome is common after liver transplantation and is strongly associated with increased morbidity in this patient population.^40,41 The general definition of metabolic syndrome includes a combination of at least three of the following: hypertension, insulin resistance, hypertriglyceridemia, low HDL-C, and obesity.

The prevalence of metabolic syndrome is higher in patients after liver transplantation than in nontransplant patients. In a review of 252 liver transplant recipients, 52% were diagnosed with posttransplant metabolic syndrome, but only 5% had had it pretransplant.⁴²

Careful screening for posttransplant metabolic syndrome and early recognition of risk factors are important. Nevertheless, the treatment of this condition depends on treating its components according to recommended guidelines.⁴¹

Cardiovascular disease

The incidence of cardiovascular morbidity and death is increased after liver transplantation.²⁴ In addition, after liver transplantation, cardiovascular disease is a major cause of death unrelated to liver disease. It accounts for 12% to 16% of deaths and is the third most common cause of late mortality after liver transplantation.⁴³ Of note, a recent study by our group demonstrated that patients undergoing liver transplantation for nonalcoholic steatohepatitis had a significantly higher risk of a cardiovascular event during the 3 years after transplantation than patients undergoing liver transplantation for cholestatic liver disease.⁴⁴

Risk factors for cardiovascular disease after liver transplantation include older age at transplantation, male sex, posttransplant diabetes, posttransplant hypertension, and the use of mycophenolate mofetil.⁴⁴ Modifying the risk factors is essential in decreasing the risk of cardiovascular events.

It is reasonable to perform dobutamine stress testing every 3 to 5 years in patients with multiple risk factors for cardiovascular disease, or more frequently in those with preexisting coronary artery disease.^45,46

Malignancy

The risk of several malignancies increases after liver transplantation. Liver transplant recipients have an incidence of cancer 2.1 to 4.3 times greater than age- and sex-matched controls.^24,47–49

Skin cancers are the most common and account for almost 40% of malignancies in organ transplant recipients.⁵⁰ Whereas basal cell carcinoma is more common in the general population, squamous cell carcinoma is equally common in liver transplant recipients.

Multiple clinical studies have linked calcineurin inhibitors and azathioprine to the development of skin cancer. Annual skin examinations in addition to avoiding other risk factors such as smoking and sun exposure are generally recommended. Changing the immunosuppressants to sirolimus in high-risk patients may lower their chance of developing skin cancer.^51,52

Patients with ulcerative colitis who undergo liver transplantation because of sclerosing cholangitis are at higher risk of colon cancer and require annual colonoscopy with surveillance biopsies. Patients who undergo transplantation for alcoholic liver disease seem to have a higher risk of pulmonary and oropharyngeal cancers.^53,54

It is important that transplant patients adhere to recommended cancer screening guidelines, in view of their increased risk. Studies have shown improved overall survival in liver transplant recipients who underwent intensive cancer surveillance.⁵⁵

Renal insufficiency

Renal insufficiency is a well-recognized complication of liver transplantation and is associated with an increased long-term death rate.^56,57

The incidence of renal insufficiency increases dramatically over time. Ojo et al,⁵⁷ in a study of almost 37,000 liver transplant recipients, found that the incidence of chronic kidney disease (defined as an estimated glomerular filtration rate < 30 mL/min/1.73 m²) was 13.9% at 3 years, 18% at 5 years, and approximately 26% at 10 years.

Risk factors include the use of calcineurin inhibitors (both cyclosporine and tacrolimus), older age, female sex, lower pretransplant glomerular filtration rate, postoperative acute renal failure, diabetes, hypertension, hepatitis C virus infection, and transplantation before 1998.^58,59 Replacing a calcineurin inhibitor with mycophenolate mofetil or sirolimus may be considered with communication with the transplant center, as mycophenolate mofetil or sirolimus are associated with a lower risk of renal injury.^60–64

All liver transplant recipients should avoid nonsteroidal anti-inflammatory drugs and nephrotoxic medications

Starting 1 year after liver transplantation, primary care providers should screen for renal dysfunction by obtaining kidney function tests every 6 months, including urinalysis and microalbuminuria assessment. Equations for estimating the glomerular filtration rate used in practice, such as the Modification of Diet in Renal Disease Study equation, rely mainly on serum creatinine, which may lead to overestimating renal function in some circumstances. Therefore, other equations can be used to confirm the estimated glomerular filtration rate measured by creatinine clearance, and to more accurately evaluate kidney function. Calculators are available at www.kidney.org/professionals/kdoqi/gfr_calculator.

All liver transplant recipients should avoid nonsteroidal anti-inflammatory drugs and nephrotoxic medications, and should have their hypertension and diabetes adequately controlled.

Bone diseases

Osteopenia is another major complication of liver transplantation. One-third of liver transplant recipients have a bone mineral density below the fracture threshold.⁶⁵

Multiple factors contribute to increased bone loss after transplantation, including use of corticosteroids, use of calcineurin inhibitors (cyclosporine, tacrolimus), poor nutrition, vitamin D deficiency, immobility, sarcopenia (reduced muscle mass), hypogonadism, smoking, and alcohol abuse.⁶⁶ Even at low doses of less than 7.5 mg per day, corticosteroids inhibit osteoblast activity and increase bone resorption.

Studies have reported rapid bone loss at around 6 months after transplantation.^67–69 However, long-term follow-up of bone mineral density up to 15 years after transplantation revealed an improvement mainly in the 2nd postoperative year, with no deterioration afterward.⁶⁵

High-risk patients need to be identified early with appropriate screening and evaluation. Evaluation includes dual-energy x-ray absorptiometry and serum levels of calcium, phosphorous, parathyroid hormone, testosterone (men), estradiol (women), and alpha-25-hydroxyvitamin D. These tests are typically done before transplantation and then every other year afterward.

We recommend a daily dose of calcitriol and a calcium supplement to all our liver transplant patients.⁷⁰ If osteoporosis (a T-score 2.5 or more standard deviations below the mean) or a fragility fracture occurs, then the patient may benefit from an oral bisphosphonate. Calcitonin has also been shown to improve bone mineral density in patients with osteoporosis after liver transplantation.⁷¹

Hyperuricemia and gout

Although hyperuricemia is common in liver transplant recipients (reported in approximately 47%), the development of clinical gout is less common (6%).⁷² Asymptomatic hyperuricemia requires observation only and is not usually treated in liver transplant recipients.

Acute attacks of gout are typically managed with colchicine 0.6 mg every 2 hours, up to five doses. Prednisone can be considered if symptoms persist despite treatment with colchicine. Allopurinol in an initial dose of 100 mg daily is used as maintenance therapy to reduce production of uric acid.⁷³ However, because of the potential for drug interactions, the combination of azathioprine and allopurinol should be avoided.

Psychiatric complications and quality of life

Depression is common in liver transplant recipients, significantly more so in patients who received a transplant because of hepatitis C.⁷⁴ The type of immunosuppressant is not associated with the incidence of depression. When indicated, the internist may start the patient on a selective serotonin reuptake inhibitor such as citalopram 20 mg daily, as these medications are usually effective and well tolerated in liver transplant patients.⁷³

Liver transplantation has a major positive effect on quality of life. Most patients with end-stage liver disease have poor quality of life before transplantation, but this seems to improve notably afterward. A meta-analysis showed significant improvement in posttransplant physical health, sexual functioning, daily activities, and social functioning compared with before transplantation.⁷⁵

Alcohol abuse and smoking

Patients who underwent liver transplantation because of alcoholic liver disease should be advised to abstain from alcohol.¹⁹ Patients who underwent the procedure for a different indication are advised to avoid excessive alcohol intake, as it is proven to lower the survival rate.²⁰ Alcohol recidivism and smoking (including marijuana) are major problems, and internists are best positioned to address these issues and treat them.

Vaccinations

All liver transplant recipients should be vaccinated against influenza, pneumococcal infection, and tetanus. Hepatitis A and B vaccines are typically given before transplantation. In general, live vaccines such as measles-mumps-rubella and varicella are not recommended after any solid organ transplant.⁷⁶

A study in Germany showed that immunization rates were too low in solid-organ transplant recipients, and almost 90% of patients were not adequately informed about immunizations.⁷⁷ Hence, there may be room for improvement, and primary care providers should take the lead toward better outcomes in this regard.

Recurrence of the primary liver disease after transplantation

Different primary liver diseases recur with different frequencies.

Hepatitis C has the highest rate of recurrence of the liver diseases.^78,79 Reinfection with hepatitis C virus after liver transplantation is almost universal and can follow different patterns. One of the most aggressive patterns is fibrosing cholestatic hepatitis, which frequently leads to graft failure and death, and hence necessitates urgent detection and treatment.

Hepatocellular carcinoma also has a high recurrence rate.⁸⁰ Surveillance with liver ultrasonography or computed tomography is required every 6 months for the first 5 years after liver transplantation.

Other liver diseases. Nonalcoholic steatohepatitis, primary biliary cirrhosis, primary sclerosing cholangitis, autoimmune hepatitis, and hepatitis B infection also tend to recur after liver transplantation.^46,81 On the other hand, alpha-1 antitrypsin deficiency, Wilson disease, hemochromatosis, and metabolic disorders are “cured” after liver transplantation.

It is important to detect any increase in liver enzymes above baseline. An elevation of 1.5 times the upper limit of normal or more should trigger further investigation.

Allograft dysfunction

A number of complications can develop in the liver allograft and result in abnormal liver function tests and, if not treated, graft failure. The most common causes of late graft dysfunction include recurrence of primary liver disease, biliary complications, and chronic rejection.⁴⁶

Vascular complications include hepatic artery thrombosis and stenosis and are usually evaluated by liver ultrasonography and Doppler scan of the hepatic artery and venous structures.²⁴

Biliary strictures give a cholestatic picture, with elevated bilirubin and greater elevation of alkaline phosphatase than of alanine aminotransferase and aspartate aminotransferase. Strictures are usually treated by endoscopic dilation and stenting, but they may eventually require surgery.

Late acute cellular rejections occur in 10% to 20% of cases and are a risk factor for chronic rejection. Liver biopsy is needed to make the diagnosis, and pulsed doses of corticosteroids remain the backbone of treatment therapy.

Chronic rejection is not common, occurring in 3% to 4% of liver transplant recipients.⁴⁶ Treatment is based on increasing immunosuppression and ensuring compliance with prescribed medications. However, chronic rejection may not respond well, and repeat transplantation may be the last resort for some patients.

WHEN TO REFER TO THE HEPATOLOGIST

Some situations require referral to the hepatologist or the transplant center. In general, the following are best managed by a hepatologist: adjustment of immunosuppressive drugs and dosages, allograft dysfunction, vascular and biliary complications, progressing renal dysfunction, and recurrence of primary liver disease. Early communication with a hepatologist and the transplant center is recommended in these cases.

Since 1963, when Starzl et al performed the first successful liver transplantation,¹ outcomes of this life-saving procedure have continued to improve. Long-term survival rates have increased markedly: the current 5-year rate is 73.8% and the 10-year rate is 60%.²

This success means that internists will be caring for a greater number of liver transplant recipients and managing their long-term problems, such as hypertension, diabetes mellitus, dyslipidemia, obesity, metabolic syndrome, cardiovascular disease, renal insufficiency, osteoporosis, cancer, and gout.

This review will discuss these complications, focusing on the role the primary care physician assumes beyond the first year after transplantation.

ROLE OF THE PRIMARY CARE PHYSICIAN

Hepatologists, primary care physicians, and surgeons share the care of transplant recipients. The first several weeks after transplantation require close follow-up by the hepatologist and transplantation team, with particular attention paid to the patient’s overall health and well-being, medication compliance, and biochemical and immunosuppression monitoring.

After the first year, the primary care physician assumes a greater role, becoming the main provider of the patient’s care.^3,4 Good communication between the transplant center and the primary care physician should lead to a smooth transition.⁴ Although the hepatologist continues to manage immunosuppressive drugs, allograft rejections, and biliary complications, the primary care physician manages most of the long-term complications and thus needs to be aware of the common ones and feel comfortable managing them. Aims during visits are to screen for and detect common complications and manage them appropriately, in addition to performing annual physical examinations and routine health care. A reasonable interval for liver transplant recipients to visit their primary care physician is every 6 months.

IMMUNOSUPPRESSANT MEDICATIONS

Multiple agents are used for immunosuppression after liver transplantation:

• Calcineurin inhibitors (cyclosporine and tacrolimus)
• Antimetabolites (mycophenolate mofetil, azathioprine, and mycophenolate sodium)
• Mammalian target of rapamycin (mTOR) inhibitors (sirolimus and everolimus)
• Corticosteroids.

Table 1 lists their common side effects.

Most centers use a combination of two to four immunosuppressants as induction therapy in the immediate posttransplant period, then taper the doses and eliminate all but a calcineurin inhibitor and an antimetabolite. For example, some start with a combination of tacrolimus, mycophenolate mofetil, and a corticosteroid. The choice in the immediate posttransplant period is frequently made by the transplant center in cooperation with the hepatologist. By the time primary care physicians see these patients, they usually are on a calcineurin inhibitor alone or a calcineurin inhibitor plus mycophenolate mofetil.

Calcineurin inhibitors

Cyclosporine is metabolized by the cytochrome CYP3A4 pathway. With an average half-life of 15 hours, it is given orally, usually every 12 hours.

The dosage is adjusted according to the trough level. Higher levels are needed in the initial posttransplant period to prevent graft rejection, whereas lower levels are preferred later to decrease the occurrence and severity of adverse effects. Typical long-term trough levels are 50 to 100 ng/mL. Levels should be checked more often if an acute illness develops or the patient starts taking a potentially interfering drug.

Of importance: the dosage should be based on trough levels and not on random levels. Levels are often falsely high if blood samples are not drawn at the trough level. Repeating the measurement and making sure the sample is drawn at the trough level, ie, 12 hours after the last dose, is advised in this condition.

Cyclosporine causes widespread vasoconstriction resulting in decreased renal blood flow and systemic hypertension, often within a few days of starting it. Other important adverse effects include renal insufficiency, dyslipidemia, neurotoxicity (headache, tremor, seizure), and diabetes.

Tacrolimus is superior to cyclosporine in terms of survival, graft loss, acute rejection, and steroid-resistant rejection in the first year.⁵ Currently, it is the agent used most often for maintenance immunosuppression after liver transplantation.

Like cyclosporine, tacrolimus is metabolized in the liver by CYP3A4. Satisfactory trough levels after 1 year are 4 to 6 ng/mL.

The adverse effects of tacrolimus are similar to those of cyclosporine, but diabetes mellitus is more common with tacrolimus. Bone marrow suppression may occur more often with tacrolimus as well.

Antimetabolites

Antimetabolites are generally not potent enough to be used alone.

Mycophenolate mofetil causes adverse effects that include bone marrow suppression and gastrointestinal symptoms such as gastritis, diarrhea, and abdominal pain.

Azathioprine, infrequently used in transplantation in the United States, is nevertheless sometimes substituted for mycophenolate mofetil in pregnant women, as it seems safer for use in pregnancy.

Serum levels of azathioprine and mycophenolate mofetil are not routinely monitored.

mTOR inhibitors

Sirolimus and everolimus are mTOR inhibitors, inhibiting proliferation of lymphocytes.^6,7

Unlike calcineurin inhibitors, mTOR inhibitors are not associated with nephrotoxicity, neurotoxicity, renal dysfunction, hypertension, or diabetes. Sirolimus is considered an alternative to calcineurin inhibitors or, in some instances, used as add-on-therapy to lower the dose of the calcineurin inhibitor.

Sirolimus carries a black-box warning about hepatic artery thrombosis

However, sirolimus carries a potential risk of hepatic artery thrombosis, a life-threatening complication.⁸ This has led the US Food and Drug Administration (FDA) to require sirolimus to carry a black-box warning, and most transplant centers avoid using it in the first 30 days after transplantation.

Dyslipidemia is perhaps the most common adverse effect of sirolimus. Others include dose-related cytopenia and wound dehiscence.⁹

Everolimus has yet to be established for use in liver transplantation, although safety trials have been published.^10,11 The FDA currently recommends against using it in the first 30 days after liver transplantation.

Both sirolimus and everolimus are metabolized by CYP3A4, which is the same metabolic pathway used by cyclosporine and tacrolimus. Hence, drugs that inhibit CYP3A4 may significantly impair clearance of both sirolimus and everolimus.

Corticosteroids

Corticosteroids have been the cornerstone of immunosuppression and remain the first line of treatment for acute allograft rejection. High intravenous doses of corticosteroids are usually started in the peritransplant period and are then switched to oral doses, which are tapered and continued with a fixed dose such as 20 mg of prednisone daily for 3 to 6 months after transplantation. However, some transplant centers keep patients on prednisone 5 mg/day indefinitely.

Adverse effects of corticosteroids include diabetes, salt and fluid retention, hypertension, hyperlipidemia, cosmetic changes (acne, cervical fat pad or “buffalo hump”), delayed wound healing, susceptibility to infection, cataracts, osteopenia, and potential adrenal suppression.¹² There is concern that the use of these drugs may increase hepatitis C virus replication in patients who received a liver transplant for hepatitis C cirrhosis. Randomized trials have yielded conflicting results.^13–15

Drug interactions

Certain drugs can affect the metabolism of calcineurin inhibitors and mTOR inhibitors by inducing CYP3A4, which results in decreasing the levels of the immunosuppressive drugs, or by inhibiting CYP3A4, which has the opposite effect.

Medications that can decrease the levels of calcineurin inhibitors and mTOR inhibitors:

• Anticonvulsants (carbamazepine, phenobarbital, phenytoin)
• Antibiotics (rifampin, isoniazid)
• St John’s wort.

Medications that can increase the levels of calcineurin inhibitors and mTOR inhibitors:

• Antifungals (fluconazole, ketoconazole, itraconazole, voriconazole, aspofungin)
• Antibiotics (azithromycin, erythromycin, clarithromycin)
• Nondihydropyridine calcium channel blockers (diltiazem, verapamil).¹⁶

Selected antibiotics are generally well tolerated, such as penicillins, cephalosporins, quinolones, sulfonamides, and topical antifungal agents.

LONG-TERM COMPLICATIONS

Figure 1 summarizes the common long-term complications of liver transplantation.

Hypertension

The prevalence of hypertension after liver transplantation is 40% to 85%, which is markedly higher than in patients with chronic liver disease before liver transplantation.^17,18

One of the factors contributing to this increase is the use of immunosuppressive medications. Of these drugs, cyclosporine seems to be the one that most often causes an increase in both the incidence and the severity of hypertension, as it produces widespread vasoconstriction.¹⁹ Corticosteroids cause hypertension through their mineralocorticoid effects.

The diagnostic cutoffs for hypertension (ie, 140/90 mm Hg) and the treatment goals in posttransplant patients are similar to those in the general population. However, at our institution we target a blood pressure of less than 130/80 mm Hg in transplant patients because they have a high prevalence of other cardiovascular risk factors such as diabetes, obesity, and renal insufficiency.²⁰

Dihydropyridine calcium channel blockers such as amlodipine and nifedipine are considered the best first-line agents because they dilate renal afferent arterioles, an effect that may counteract the vasoconstriction mediated by calcineurin inhibitors. Nondihydropyridine calcium channel blockers such as diltiazem and verapamil tend to have more marked negative inotropic effects and are not recommended in liver transplant recipients because they increase the levels of calcineurin inhibitors.²¹

Diuretics (eg, furosemide) might be the second-line agents, especially in patients with peripheral edema.¹⁶ One should be vigilant for hyperuricemia if thiazide agents are used.

Angiotensin-converting-enzyme (ACE) inhibitors and angiotensin II receptor blockers (ARBs) are typically avoided in the early posttransplant period, but they can be started later and have additional benefits in patients with diabetes and congestive heart failure. Starting ACE inhibitors is acceptable in these patients unless there is a contraindication such as allergy to ACE inhibitors, hypotension, history of bilateral renal artery stenosis, significant hyperkalemia, or acute kidney injury. Monitor the serum potassium level closely for hyperkalemia in patients concurrently using calcineurin inhibitors.

Alpha-blockers and beta-blockers can be used as add-on therapy in patients with uncontrolled hypertension with the exception of carvedilol, because it increases the levels of calcineurin inhibitors.²²

Blood pressure monitoring by the primary care physician is recommended every 6 months after the early posttransplant period, or more frequently when changes in treatment are being considered.

If hypertension continues to be inadequately controlled despite treatment, changing the immunosuppressive drugs or decreasing the doses can be considered, but the transplant hepatologist must be involved in this decision.^23,24

Diabetes mellitus

The prevalence of diabetes mellitus is higher in liver transplant recipients than in the general population, reaching 30% to 40%.^17,25 In addition to preexisting diabetes, 15% of liver transplant recipients develop new-onset diabetes.^26,27

Risk factors for developing diabetes after liver transplantation include African American or Hispanic ethnicity, obesity, family history, pretransplant diabetes, hepatitis C virus infection, use of corticosteroids, and use of calcineurin inhibitors (tacrolimus more than cyclosporine) and sirolimus.²⁶

In addition to increasing the risk of cardiovascular disease and other diseases, diabetes decreases both patient and graft survival after liver transplantation.²⁸

The management of diabetes and the treatment target after transplantation should follow the American Diabetes Association guidelines for the treatment of type 2 diabetes mellitus.²⁹ Lifestyle modifications, diet, and exercise are as important for transplant patients as for the nontransplant population. Insulin therapy is usually needed in the early posttransplant period to control blood glucose levels well, especially with the high doses of corticosteroids used during the first few weeks. No trials to date have compared oral agents in posttransplant patients. Therefore, the choice of oral hypoglycemic agents should be individualized on the basis of the patient’s characteristics and comorbidities.

Screen all liver transplant recipients for diabetes regardless of their pretransplant status

We recommend that primary care providers screen all liver transplant recipients for diabetes regardless of their pretransplant status. This can be done by obtaining regular fasting blood glucose levels or a hemoglobin A_1c level every 6 months. Additionally, liver transplant recipients diagnosed with diabetes require annual eye examinations to look for cataracts and diabetes-related changes.³⁰

Dyslipidemia

On November 12, 2013, the American College of Cardiology and the American Heart Association (ACC/AHA) released new clinical practice guidelines for treating blood cholesterol levels.^31,32 According to these new guidelines, there are four groups of patients for whom treatment with statins is clearly indicated:

• Patients with cardiovascular disease
• Patients with low-density lipoprotein cholesterol (LDL-C) levels ≥ 190 mg/dL
• Patients 40 to 75 years old with type 2 diabetes
• Patients 40 to 75 years old with an estimated 10-year risk of cardiovascular disease of 7.5% or greater.

Liver transplant recipients should be evaluated on an individual basis to see if they fit in any of the four groups and if statin treatment therefore needs to be initiated.

A few things need to be kept in mind. First, the incidence of dyslipidemia after liver transplantation is estimated to be 45% to 69%. Risk factors include obesity, diabetes mellitus, cholestatic liver disease, and immunosuppressant medications.³³ Sirolimus has a significant and well-documented association with dyslipidemia. Cyclosporine and corticosteroids are also strongly associated with dyslipidemia. Tacrolimus has a minor effect, and mycophenolate mofetil and azathioprine have no significant effects on serum lipid levels.¹⁶

Second, of the seven currently marketed statins, pravastatin and fluvastatin are preferred in liver transplant recipients because they are not metabolized by the same cytochrome CYP3A4 pathway that metabolizes calcineurin inhibitors and sirolimus.³⁴ The doses of 40 to 80 mg daily of pravastatin or 40 mg twice daily of fluvastatin lower low-density lipoprotein cholesterol (LDL-C) levels by approximately 30% to 35%. However, these two agents are considered “moderate-intensity” statins according to the new ACC/AHA guidelines. The only two “high-intensity” statins are atorvastatin (40–80 mg) and rosuvastatin (20–40 mg), but they are both metabolized by CYP3A4. Therefore it is prudent to avoid them with the concurrent use of a calcineurin inhibitor or tacrolimus.

Gemfibrozil does not lower LDL-C and should not be used concomitantly with statins due to unacceptable risk of rhabdomyolysis and myopathy. Fenofibrates are usually avoided due to potential nephrotoxicity in patients receiving cyclosporine. Bile acid sequestrants (cholestyramine, colestipol, colesevelam) can decrease plasma mycophenolate mofetil levels by 35%.^16,35 Thus, these agents should be avoided if mycophenolate mofetil is used.

It is reasonable to screen all liver transplant recipients with a fasting lipid profile at 3, 6, and 12 months after transplantation and annually thereafter. Creatine kinase should be measured if the patient complains of severe muscle pain or weakness but not on a routine basis.

Obesity

Approximately one-third of patients who are of normal weight at the time of transplantation will become obese afterward.^18,25 Corticosteroid use is an important risk factor for posttransplant obesity, and tapering these drugs helps reduce weight.³⁶ Patients treated with cyclosporine are more likely to gain weight than those who receive tacrolimus.³⁷

Of importance: nonalcoholic fatty liver disease, currently the most common cause of chronic liver disease in adults, is rapidly increasing as an indication for liver transplantation. In fact, the proportion of liver transplantation procedures for nonalcoholic steatohepatitis-related cirrhosis increased from 1.2% in 2001 to 9.7% in 2009, and nonalcoholic steatohepatitis is expected to become the leading indication for liver transplantation in the next 20 years. And because nonalcoholic fatty liver disease is directly linked to obesity, the prevalence of obesity as a complication of liver transplantation will most likely increase in the near future.

Overweight liver transplant recipients may have great difficulty losing weight. Treatment starts with patient education on caloric restriction and exercise. If traditional measures fail to result in adequate weight loss, additional options include switching from cyclosporine to tacrolimus.²³

Bariatric surgery may become an option for posttransplant patients. In a recent case series from the University of Minnesota, Al-Nowaylati et al³⁸ reported their experience with seven patients who underwent orthotopic liver transplantation and then open Roux-en-Y gastric bypass. After bariatric surgery, the patients’ mean body mass index declined significantly, and glycemic control and high-density lipoprotein cholesterol (HDL-C) levels improved. However, one patient died of multiple organ failure, to which the bariatric surgery might have contributed.³⁸

Heimbach et al³⁹ conducted a study in patients referred for liver transplantation for whom a rigorous noninvasive weight-loss program before transplantation had failed. The researchers performed combined liver transplantation and sleeve gastrectomy in seven carefully selected patients who had failed to achieve weight loss to a body mass index less than 35 kg/m² before transplantation. All seven patients lost weight, decreasing their mean body mass index from 49 kg/m² before the procedure to 29 kg/m² at last follow-up, and none of them developed posttransplant diabetes or steatosis.

At this time, there is not enough evidence to recommend concurrent orthotopic liver transplantation plus bariatric surgery, or combined orthotopic liver transplantation and sleeve gastrectomy. More study is needed to further evaluate these advanced approaches.

Posttransplant metabolic syndrome

Metabolic syndrome is common after liver transplantation and is strongly associated with increased morbidity in this patient population.^40,41 The general definition of metabolic syndrome includes a combination of at least three of the following: hypertension, insulin resistance, hypertriglyceridemia, low HDL-C, and obesity.

The prevalence of metabolic syndrome is higher in patients after liver transplantation than in nontransplant patients. In a review of 252 liver transplant recipients, 52% were diagnosed with posttransplant metabolic syndrome, but only 5% had had it pretransplant.⁴²

Careful screening for posttransplant metabolic syndrome and early recognition of risk factors are important. Nevertheless, the treatment of this condition depends on treating its components according to recommended guidelines.⁴¹

Cardiovascular disease

The incidence of cardiovascular morbidity and death is increased after liver transplantation.²⁴ In addition, after liver transplantation, cardiovascular disease is a major cause of death unrelated to liver disease. It accounts for 12% to 16% of deaths and is the third most common cause of late mortality after liver transplantation.⁴³ Of note, a recent study by our group demonstrated that patients undergoing liver transplantation for nonalcoholic steatohepatitis had a significantly higher risk of a cardiovascular event during the 3 years after transplantation than patients undergoing liver transplantation for cholestatic liver disease.⁴⁴

Risk factors for cardiovascular disease after liver transplantation include older age at transplantation, male sex, posttransplant diabetes, posttransplant hypertension, and the use of mycophenolate mofetil.⁴⁴ Modifying the risk factors is essential in decreasing the risk of cardiovascular events.

It is reasonable to perform dobutamine stress testing every 3 to 5 years in patients with multiple risk factors for cardiovascular disease, or more frequently in those with preexisting coronary artery disease.^45,46

Malignancy

The risk of several malignancies increases after liver transplantation. Liver transplant recipients have an incidence of cancer 2.1 to 4.3 times greater than age- and sex-matched controls.^24,47–49

Skin cancers are the most common and account for almost 40% of malignancies in organ transplant recipients.⁵⁰ Whereas basal cell carcinoma is more common in the general population, squamous cell carcinoma is equally common in liver transplant recipients.

Multiple clinical studies have linked calcineurin inhibitors and azathioprine to the development of skin cancer. Annual skin examinations in addition to avoiding other risk factors such as smoking and sun exposure are generally recommended. Changing the immunosuppressants to sirolimus in high-risk patients may lower their chance of developing skin cancer.^51,52

Patients with ulcerative colitis who undergo liver transplantation because of sclerosing cholangitis are at higher risk of colon cancer and require annual colonoscopy with surveillance biopsies. Patients who undergo transplantation for alcoholic liver disease seem to have a higher risk of pulmonary and oropharyngeal cancers.^53,54

It is important that transplant patients adhere to recommended cancer screening guidelines, in view of their increased risk. Studies have shown improved overall survival in liver transplant recipients who underwent intensive cancer surveillance.⁵⁵

Renal insufficiency

Renal insufficiency is a well-recognized complication of liver transplantation and is associated with an increased long-term death rate.^56,57

The incidence of renal insufficiency increases dramatically over time. Ojo et al,⁵⁷ in a study of almost 37,000 liver transplant recipients, found that the incidence of chronic kidney disease (defined as an estimated glomerular filtration rate < 30 mL/min/1.73 m²) was 13.9% at 3 years, 18% at 5 years, and approximately 26% at 10 years.

Risk factors include the use of calcineurin inhibitors (both cyclosporine and tacrolimus), older age, female sex, lower pretransplant glomerular filtration rate, postoperative acute renal failure, diabetes, hypertension, hepatitis C virus infection, and transplantation before 1998.^58,59 Replacing a calcineurin inhibitor with mycophenolate mofetil or sirolimus may be considered with communication with the transplant center, as mycophenolate mofetil or sirolimus are associated with a lower risk of renal injury.^60–64

All liver transplant recipients should avoid nonsteroidal anti-inflammatory drugs and nephrotoxic medications

Starting 1 year after liver transplantation, primary care providers should screen for renal dysfunction by obtaining kidney function tests every 6 months, including urinalysis and microalbuminuria assessment. Equations for estimating the glomerular filtration rate used in practice, such as the Modification of Diet in Renal Disease Study equation, rely mainly on serum creatinine, which may lead to overestimating renal function in some circumstances. Therefore, other equations can be used to confirm the estimated glomerular filtration rate measured by creatinine clearance, and to more accurately evaluate kidney function. Calculators are available at www.kidney.org/professionals/kdoqi/gfr_calculator.

All liver transplant recipients should avoid nonsteroidal anti-inflammatory drugs and nephrotoxic medications, and should have their hypertension and diabetes adequately controlled.

Bone diseases

Osteopenia is another major complication of liver transplantation. One-third of liver transplant recipients have a bone mineral density below the fracture threshold.⁶⁵

Multiple factors contribute to increased bone loss after transplantation, including use of corticosteroids, use of calcineurin inhibitors (cyclosporine, tacrolimus), poor nutrition, vitamin D deficiency, immobility, sarcopenia (reduced muscle mass), hypogonadism, smoking, and alcohol abuse.⁶⁶ Even at low doses of less than 7.5 mg per day, corticosteroids inhibit osteoblast activity and increase bone resorption.

Studies have reported rapid bone loss at around 6 months after transplantation.^67–69 However, long-term follow-up of bone mineral density up to 15 years after transplantation revealed an improvement mainly in the 2nd postoperative year, with no deterioration afterward.⁶⁵

High-risk patients need to be identified early with appropriate screening and evaluation. Evaluation includes dual-energy x-ray absorptiometry and serum levels of calcium, phosphorous, parathyroid hormone, testosterone (men), estradiol (women), and alpha-25-hydroxyvitamin D. These tests are typically done before transplantation and then every other year afterward.

We recommend a daily dose of calcitriol and a calcium supplement to all our liver transplant patients.⁷⁰ If osteoporosis (a T-score 2.5 or more standard deviations below the mean) or a fragility fracture occurs, then the patient may benefit from an oral bisphosphonate. Calcitonin has also been shown to improve bone mineral density in patients with osteoporosis after liver transplantation.⁷¹

Hyperuricemia and gout

Although hyperuricemia is common in liver transplant recipients (reported in approximately 47%), the development of clinical gout is less common (6%).⁷² Asymptomatic hyperuricemia requires observation only and is not usually treated in liver transplant recipients.

Acute attacks of gout are typically managed with colchicine 0.6 mg every 2 hours, up to five doses. Prednisone can be considered if symptoms persist despite treatment with colchicine. Allopurinol in an initial dose of 100 mg daily is used as maintenance therapy to reduce production of uric acid.⁷³ However, because of the potential for drug interactions, the combination of azathioprine and allopurinol should be avoided.

Psychiatric complications and quality of life

Depression is common in liver transplant recipients, significantly more so in patients who received a transplant because of hepatitis C.⁷⁴ The type of immunosuppressant is not associated with the incidence of depression. When indicated, the internist may start the patient on a selective serotonin reuptake inhibitor such as citalopram 20 mg daily, as these medications are usually effective and well tolerated in liver transplant patients.⁷³

Liver transplantation has a major positive effect on quality of life. Most patients with end-stage liver disease have poor quality of life before transplantation, but this seems to improve notably afterward. A meta-analysis showed significant improvement in posttransplant physical health, sexual functioning, daily activities, and social functioning compared with before transplantation.⁷⁵

Alcohol abuse and smoking

Patients who underwent liver transplantation because of alcoholic liver disease should be advised to abstain from alcohol.¹⁹ Patients who underwent the procedure for a different indication are advised to avoid excessive alcohol intake, as it is proven to lower the survival rate.²⁰ Alcohol recidivism and smoking (including marijuana) are major problems, and internists are best positioned to address these issues and treat them.

Vaccinations

All liver transplant recipients should be vaccinated against influenza, pneumococcal infection, and tetanus. Hepatitis A and B vaccines are typically given before transplantation. In general, live vaccines such as measles-mumps-rubella and varicella are not recommended after any solid organ transplant.⁷⁶

A study in Germany showed that immunization rates were too low in solid-organ transplant recipients, and almost 90% of patients were not adequately informed about immunizations.⁷⁷ Hence, there may be room for improvement, and primary care providers should take the lead toward better outcomes in this regard.

Recurrence of the primary liver disease after transplantation

Different primary liver diseases recur with different frequencies.

Hepatitis C has the highest rate of recurrence of the liver diseases.^78,79 Reinfection with hepatitis C virus after liver transplantation is almost universal and can follow different patterns. One of the most aggressive patterns is fibrosing cholestatic hepatitis, which frequently leads to graft failure and death, and hence necessitates urgent detection and treatment.

Hepatocellular carcinoma also has a high recurrence rate.⁸⁰ Surveillance with liver ultrasonography or computed tomography is required every 6 months for the first 5 years after liver transplantation.

Other liver diseases. Nonalcoholic steatohepatitis, primary biliary cirrhosis, primary sclerosing cholangitis, autoimmune hepatitis, and hepatitis B infection also tend to recur after liver transplantation.^46,81 On the other hand, alpha-1 antitrypsin deficiency, Wilson disease, hemochromatosis, and metabolic disorders are “cured” after liver transplantation.

It is important to detect any increase in liver enzymes above baseline. An elevation of 1.5 times the upper limit of normal or more should trigger further investigation.

Allograft dysfunction

A number of complications can develop in the liver allograft and result in abnormal liver function tests and, if not treated, graft failure. The most common causes of late graft dysfunction include recurrence of primary liver disease, biliary complications, and chronic rejection.⁴⁶

Vascular complications include hepatic artery thrombosis and stenosis and are usually evaluated by liver ultrasonography and Doppler scan of the hepatic artery and venous structures.²⁴

Biliary strictures give a cholestatic picture, with elevated bilirubin and greater elevation of alkaline phosphatase than of alanine aminotransferase and aspartate aminotransferase. Strictures are usually treated by endoscopic dilation and stenting, but they may eventually require surgery.

Late acute cellular rejections occur in 10% to 20% of cases and are a risk factor for chronic rejection. Liver biopsy is needed to make the diagnosis, and pulsed doses of corticosteroids remain the backbone of treatment therapy.

Chronic rejection is not common, occurring in 3% to 4% of liver transplant recipients.⁴⁶ Treatment is based on increasing immunosuppression and ensuring compliance with prescribed medications. However, chronic rejection may not respond well, and repeat transplantation may be the last resort for some patients.

WHEN TO REFER TO THE HEPATOLOGIST

Some situations require referral to the hepatologist or the transplant center. In general, the following are best managed by a hepatologist: adjustment of immunosuppressive drugs and dosages, allograft dysfunction, vascular and biliary complications, progressing renal dysfunction, and recurrence of primary liver disease. Early communication with a hepatologist and the transplant center is recommended in these cases.

Since 1963, when Starzl et al performed the first successful liver transplantation,¹ outcomes of this life-saving procedure have continued to improve. Long-term survival rates have increased markedly: the current 5-year rate is 73.8% and the 10-year rate is 60%.²

This success means that internists will be caring for a greater number of liver transplant recipients and managing their long-term problems, such as hypertension, diabetes mellitus, dyslipidemia, obesity, metabolic syndrome, cardiovascular disease, renal insufficiency, osteoporosis, cancer, and gout.

This review will discuss these complications, focusing on the role the primary care physician assumes beyond the first year after transplantation.

ROLE OF THE PRIMARY CARE PHYSICIAN

Hepatologists, primary care physicians, and surgeons share the care of transplant recipients. The first several weeks after transplantation require close follow-up by the hepatologist and transplantation team, with particular attention paid to the patient’s overall health and well-being, medication compliance, and biochemical and immunosuppression monitoring.

After the first year, the primary care physician assumes a greater role, becoming the main provider of the patient’s care.^3,4 Good communication between the transplant center and the primary care physician should lead to a smooth transition.⁴ Although the hepatologist continues to manage immunosuppressive drugs, allograft rejections, and biliary complications, the primary care physician manages most of the long-term complications and thus needs to be aware of the common ones and feel comfortable managing them. Aims during visits are to screen for and detect common complications and manage them appropriately, in addition to performing annual physical examinations and routine health care. A reasonable interval for liver transplant recipients to visit their primary care physician is every 6 months.

IMMUNOSUPPRESSANT MEDICATIONS

Multiple agents are used for immunosuppression after liver transplantation:

• Calcineurin inhibitors (cyclosporine and tacrolimus)
• Antimetabolites (mycophenolate mofetil, azathioprine, and mycophenolate sodium)
• Mammalian target of rapamycin (mTOR) inhibitors (sirolimus and everolimus)
• Corticosteroids.

Table 1 lists their common side effects.

Most centers use a combination of two to four immunosuppressants as induction therapy in the immediate posttransplant period, then taper the doses and eliminate all but a calcineurin inhibitor and an antimetabolite. For example, some start with a combination of tacrolimus, mycophenolate mofetil, and a corticosteroid. The choice in the immediate posttransplant period is frequently made by the transplant center in cooperation with the hepatologist. By the time primary care physicians see these patients, they usually are on a calcineurin inhibitor alone or a calcineurin inhibitor plus mycophenolate mofetil.

Calcineurin inhibitors

Cyclosporine is metabolized by the cytochrome CYP3A4 pathway. With an average half-life of 15 hours, it is given orally, usually every 12 hours.

The dosage is adjusted according to the trough level. Higher levels are needed in the initial posttransplant period to prevent graft rejection, whereas lower levels are preferred later to decrease the occurrence and severity of adverse effects. Typical long-term trough levels are 50 to 100 ng/mL. Levels should be checked more often if an acute illness develops or the patient starts taking a potentially interfering drug.

Of importance: the dosage should be based on trough levels and not on random levels. Levels are often falsely high if blood samples are not drawn at the trough level. Repeating the measurement and making sure the sample is drawn at the trough level, ie, 12 hours after the last dose, is advised in this condition.

Cyclosporine causes widespread vasoconstriction resulting in decreased renal blood flow and systemic hypertension, often within a few days of starting it. Other important adverse effects include renal insufficiency, dyslipidemia, neurotoxicity (headache, tremor, seizure), and diabetes.

Tacrolimus is superior to cyclosporine in terms of survival, graft loss, acute rejection, and steroid-resistant rejection in the first year.⁵ Currently, it is the agent used most often for maintenance immunosuppression after liver transplantation.

Like cyclosporine, tacrolimus is metabolized in the liver by CYP3A4. Satisfactory trough levels after 1 year are 4 to 6 ng/mL.

The adverse effects of tacrolimus are similar to those of cyclosporine, but diabetes mellitus is more common with tacrolimus. Bone marrow suppression may occur more often with tacrolimus as well.

Antimetabolites

Antimetabolites are generally not potent enough to be used alone.

Mycophenolate mofetil causes adverse effects that include bone marrow suppression and gastrointestinal symptoms such as gastritis, diarrhea, and abdominal pain.

Azathioprine, infrequently used in transplantation in the United States, is nevertheless sometimes substituted for mycophenolate mofetil in pregnant women, as it seems safer for use in pregnancy.

Serum levels of azathioprine and mycophenolate mofetil are not routinely monitored.

mTOR inhibitors

Sirolimus and everolimus are mTOR inhibitors, inhibiting proliferation of lymphocytes.^6,7

Unlike calcineurin inhibitors, mTOR inhibitors are not associated with nephrotoxicity, neurotoxicity, renal dysfunction, hypertension, or diabetes. Sirolimus is considered an alternative to calcineurin inhibitors or, in some instances, used as add-on-therapy to lower the dose of the calcineurin inhibitor.

Sirolimus carries a black-box warning about hepatic artery thrombosis

However, sirolimus carries a potential risk of hepatic artery thrombosis, a life-threatening complication.⁸ This has led the US Food and Drug Administration (FDA) to require sirolimus to carry a black-box warning, and most transplant centers avoid using it in the first 30 days after transplantation.

Dyslipidemia is perhaps the most common adverse effect of sirolimus. Others include dose-related cytopenia and wound dehiscence.⁹

Everolimus has yet to be established for use in liver transplantation, although safety trials have been published.^10,11 The FDA currently recommends against using it in the first 30 days after liver transplantation.

Both sirolimus and everolimus are metabolized by CYP3A4, which is the same metabolic pathway used by cyclosporine and tacrolimus. Hence, drugs that inhibit CYP3A4 may significantly impair clearance of both sirolimus and everolimus.

Corticosteroids

Corticosteroids have been the cornerstone of immunosuppression and remain the first line of treatment for acute allograft rejection. High intravenous doses of corticosteroids are usually started in the peritransplant period and are then switched to oral doses, which are tapered and continued with a fixed dose such as 20 mg of prednisone daily for 3 to 6 months after transplantation. However, some transplant centers keep patients on prednisone 5 mg/day indefinitely.

Adverse effects of corticosteroids include diabetes, salt and fluid retention, hypertension, hyperlipidemia, cosmetic changes (acne, cervical fat pad or “buffalo hump”), delayed wound healing, susceptibility to infection, cataracts, osteopenia, and potential adrenal suppression.¹² There is concern that the use of these drugs may increase hepatitis C virus replication in patients who received a liver transplant for hepatitis C cirrhosis. Randomized trials have yielded conflicting results.^13–15

Drug interactions

Certain drugs can affect the metabolism of calcineurin inhibitors and mTOR inhibitors by inducing CYP3A4, which results in decreasing the levels of the immunosuppressive drugs, or by inhibiting CYP3A4, which has the opposite effect.

Medications that can decrease the levels of calcineurin inhibitors and mTOR inhibitors:

• Anticonvulsants (carbamazepine, phenobarbital, phenytoin)
• Antibiotics (rifampin, isoniazid)
• St John’s wort.

Medications that can increase the levels of calcineurin inhibitors and mTOR inhibitors:

• Antifungals (fluconazole, ketoconazole, itraconazole, voriconazole, aspofungin)
• Antibiotics (azithromycin, erythromycin, clarithromycin)
• Nondihydropyridine calcium channel blockers (diltiazem, verapamil).¹⁶

Selected antibiotics are generally well tolerated, such as penicillins, cephalosporins, quinolones, sulfonamides, and topical antifungal agents.

LONG-TERM COMPLICATIONS

Figure 1 summarizes the common long-term complications of liver transplantation.

Hypertension

The prevalence of hypertension after liver transplantation is 40% to 85%, which is markedly higher than in patients with chronic liver disease before liver transplantation.^17,18

One of the factors contributing to this increase is the use of immunosuppressive medications. Of these drugs, cyclosporine seems to be the one that most often causes an increase in both the incidence and the severity of hypertension, as it produces widespread vasoconstriction.¹⁹ Corticosteroids cause hypertension through their mineralocorticoid effects.

The diagnostic cutoffs for hypertension (ie, 140/90 mm Hg) and the treatment goals in posttransplant patients are similar to those in the general population. However, at our institution we target a blood pressure of less than 130/80 mm Hg in transplant patients because they have a high prevalence of other cardiovascular risk factors such as diabetes, obesity, and renal insufficiency.²⁰

Dihydropyridine calcium channel blockers such as amlodipine and nifedipine are considered the best first-line agents because they dilate renal afferent arterioles, an effect that may counteract the vasoconstriction mediated by calcineurin inhibitors. Nondihydropyridine calcium channel blockers such as diltiazem and verapamil tend to have more marked negative inotropic effects and are not recommended in liver transplant recipients because they increase the levels of calcineurin inhibitors.²¹

Diuretics (eg, furosemide) might be the second-line agents, especially in patients with peripheral edema.¹⁶ One should be vigilant for hyperuricemia if thiazide agents are used.

Angiotensin-converting-enzyme (ACE) inhibitors and angiotensin II receptor blockers (ARBs) are typically avoided in the early posttransplant period, but they can be started later and have additional benefits in patients with diabetes and congestive heart failure. Starting ACE inhibitors is acceptable in these patients unless there is a contraindication such as allergy to ACE inhibitors, hypotension, history of bilateral renal artery stenosis, significant hyperkalemia, or acute kidney injury. Monitor the serum potassium level closely for hyperkalemia in patients concurrently using calcineurin inhibitors.

Alpha-blockers and beta-blockers can be used as add-on therapy in patients with uncontrolled hypertension with the exception of carvedilol, because it increases the levels of calcineurin inhibitors.²²

Blood pressure monitoring by the primary care physician is recommended every 6 months after the early posttransplant period, or more frequently when changes in treatment are being considered.

If hypertension continues to be inadequately controlled despite treatment, changing the immunosuppressive drugs or decreasing the doses can be considered, but the transplant hepatologist must be involved in this decision.^23,24

Diabetes mellitus

The prevalence of diabetes mellitus is higher in liver transplant recipients than in the general population, reaching 30% to 40%.^17,25 In addition to preexisting diabetes, 15% of liver transplant recipients develop new-onset diabetes.^26,27

Risk factors for developing diabetes after liver transplantation include African American or Hispanic ethnicity, obesity, family history, pretransplant diabetes, hepatitis C virus infection, use of corticosteroids, and use of calcineurin inhibitors (tacrolimus more than cyclosporine) and sirolimus.²⁶

In addition to increasing the risk of cardiovascular disease and other diseases, diabetes decreases both patient and graft survival after liver transplantation.²⁸

The management of diabetes and the treatment target after transplantation should follow the American Diabetes Association guidelines for the treatment of type 2 diabetes mellitus.²⁹ Lifestyle modifications, diet, and exercise are as important for transplant patients as for the nontransplant population. Insulin therapy is usually needed in the early posttransplant period to control blood glucose levels well, especially with the high doses of corticosteroids used during the first few weeks. No trials to date have compared oral agents in posttransplant patients. Therefore, the choice of oral hypoglycemic agents should be individualized on the basis of the patient’s characteristics and comorbidities.

Screen all liver transplant recipients for diabetes regardless of their pretransplant status

We recommend that primary care providers screen all liver transplant recipients for diabetes regardless of their pretransplant status. This can be done by obtaining regular fasting blood glucose levels or a hemoglobin A_1c level every 6 months. Additionally, liver transplant recipients diagnosed with diabetes require annual eye examinations to look for cataracts and diabetes-related changes.³⁰

Dyslipidemia

On November 12, 2013, the American College of Cardiology and the American Heart Association (ACC/AHA) released new clinical practice guidelines for treating blood cholesterol levels.^31,32 According to these new guidelines, there are four groups of patients for whom treatment with statins is clearly indicated:

• Patients with cardiovascular disease
• Patients with low-density lipoprotein cholesterol (LDL-C) levels ≥ 190 mg/dL
• Patients 40 to 75 years old with type 2 diabetes
• Patients 40 to 75 years old with an estimated 10-year risk of cardiovascular disease of 7.5% or greater.

Liver transplant recipients should be evaluated on an individual basis to see if they fit in any of the four groups and if statin treatment therefore needs to be initiated.

A few things need to be kept in mind. First, the incidence of dyslipidemia after liver transplantation is estimated to be 45% to 69%. Risk factors include obesity, diabetes mellitus, cholestatic liver disease, and immunosuppressant medications.³³ Sirolimus has a significant and well-documented association with dyslipidemia. Cyclosporine and corticosteroids are also strongly associated with dyslipidemia. Tacrolimus has a minor effect, and mycophenolate mofetil and azathioprine have no significant effects on serum lipid levels.¹⁶

Second, of the seven currently marketed statins, pravastatin and fluvastatin are preferred in liver transplant recipients because they are not metabolized by the same cytochrome CYP3A4 pathway that metabolizes calcineurin inhibitors and sirolimus.³⁴ The doses of 40 to 80 mg daily of pravastatin or 40 mg twice daily of fluvastatin lower low-density lipoprotein cholesterol (LDL-C) levels by approximately 30% to 35%. However, these two agents are considered “moderate-intensity” statins according to the new ACC/AHA guidelines. The only two “high-intensity” statins are atorvastatin (40–80 mg) and rosuvastatin (20–40 mg), but they are both metabolized by CYP3A4. Therefore it is prudent to avoid them with the concurrent use of a calcineurin inhibitor or tacrolimus.

Gemfibrozil does not lower LDL-C and should not be used concomitantly with statins due to unacceptable risk of rhabdomyolysis and myopathy. Fenofibrates are usually avoided due to potential nephrotoxicity in patients receiving cyclosporine. Bile acid sequestrants (cholestyramine, colestipol, colesevelam) can decrease plasma mycophenolate mofetil levels by 35%.^16,35 Thus, these agents should be avoided if mycophenolate mofetil is used.

It is reasonable to screen all liver transplant recipients with a fasting lipid profile at 3, 6, and 12 months after transplantation and annually thereafter. Creatine kinase should be measured if the patient complains of severe muscle pain or weakness but not on a routine basis.

Obesity

Approximately one-third of patients who are of normal weight at the time of transplantation will become obese afterward.^18,25 Corticosteroid use is an important risk factor for posttransplant obesity, and tapering these drugs helps reduce weight.³⁶ Patients treated with cyclosporine are more likely to gain weight than those who receive tacrolimus.³⁷

Of importance: nonalcoholic fatty liver disease, currently the most common cause of chronic liver disease in adults, is rapidly increasing as an indication for liver transplantation. In fact, the proportion of liver transplantation procedures for nonalcoholic steatohepatitis-related cirrhosis increased from 1.2% in 2001 to 9.7% in 2009, and nonalcoholic steatohepatitis is expected to become the leading indication for liver transplantation in the next 20 years. And because nonalcoholic fatty liver disease is directly linked to obesity, the prevalence of obesity as a complication of liver transplantation will most likely increase in the near future.

Overweight liver transplant recipients may have great difficulty losing weight. Treatment starts with patient education on caloric restriction and exercise. If traditional measures fail to result in adequate weight loss, additional options include switching from cyclosporine to tacrolimus.²³

Bariatric surgery may become an option for posttransplant patients. In a recent case series from the University of Minnesota, Al-Nowaylati et al³⁸ reported their experience with seven patients who underwent orthotopic liver transplantation and then open Roux-en-Y gastric bypass. After bariatric surgery, the patients’ mean body mass index declined significantly, and glycemic control and high-density lipoprotein cholesterol (HDL-C) levels improved. However, one patient died of multiple organ failure, to which the bariatric surgery might have contributed.³⁸

Heimbach et al³⁹ conducted a study in patients referred for liver transplantation for whom a rigorous noninvasive weight-loss program before transplantation had failed. The researchers performed combined liver transplantation and sleeve gastrectomy in seven carefully selected patients who had failed to achieve weight loss to a body mass index less than 35 kg/m² before transplantation. All seven patients lost weight, decreasing their mean body mass index from 49 kg/m² before the procedure to 29 kg/m² at last follow-up, and none of them developed posttransplant diabetes or steatosis.

At this time, there is not enough evidence to recommend concurrent orthotopic liver transplantation plus bariatric surgery, or combined orthotopic liver transplantation and sleeve gastrectomy. More study is needed to further evaluate these advanced approaches.

Posttransplant metabolic syndrome

Metabolic syndrome is common after liver transplantation and is strongly associated with increased morbidity in this patient population.^40,41 The general definition of metabolic syndrome includes a combination of at least three of the following: hypertension, insulin resistance, hypertriglyceridemia, low HDL-C, and obesity.

The prevalence of metabolic syndrome is higher in patients after liver transplantation than in nontransplant patients. In a review of 252 liver transplant recipients, 52% were diagnosed with posttransplant metabolic syndrome, but only 5% had had it pretransplant.⁴²

Careful screening for posttransplant metabolic syndrome and early recognition of risk factors are important. Nevertheless, the treatment of this condition depends on treating its components according to recommended guidelines.⁴¹

Cardiovascular disease

The incidence of cardiovascular morbidity and death is increased after liver transplantation.²⁴ In addition, after liver transplantation, cardiovascular disease is a major cause of death unrelated to liver disease. It accounts for 12% to 16% of deaths and is the third most common cause of late mortality after liver transplantation.⁴³ Of note, a recent study by our group demonstrated that patients undergoing liver transplantation for nonalcoholic steatohepatitis had a significantly higher risk of a cardiovascular event during the 3 years after transplantation than patients undergoing liver transplantation for cholestatic liver disease.⁴⁴

Risk factors for cardiovascular disease after liver transplantation include older age at transplantation, male sex, posttransplant diabetes, posttransplant hypertension, and the use of mycophenolate mofetil.⁴⁴ Modifying the risk factors is essential in decreasing the risk of cardiovascular events.

It is reasonable to perform dobutamine stress testing every 3 to 5 years in patients with multiple risk factors for cardiovascular disease, or more frequently in those with preexisting coronary artery disease.^45,46

Malignancy

The risk of several malignancies increases after liver transplantation. Liver transplant recipients have an incidence of cancer 2.1 to 4.3 times greater than age- and sex-matched controls.^24,47–49

Skin cancers are the most common and account for almost 40% of malignancies in organ transplant recipients.⁵⁰ Whereas basal cell carcinoma is more common in the general population, squamous cell carcinoma is equally common in liver transplant recipients.

Multiple clinical studies have linked calcineurin inhibitors and azathioprine to the development of skin cancer. Annual skin examinations in addition to avoiding other risk factors such as smoking and sun exposure are generally recommended. Changing the immunosuppressants to sirolimus in high-risk patients may lower their chance of developing skin cancer.^51,52

Patients with ulcerative colitis who undergo liver transplantation because of sclerosing cholangitis are at higher risk of colon cancer and require annual colonoscopy with surveillance biopsies. Patients who undergo transplantation for alcoholic liver disease seem to have a higher risk of pulmonary and oropharyngeal cancers.^53,54

It is important that transplant patients adhere to recommended cancer screening guidelines, in view of their increased risk. Studies have shown improved overall survival in liver transplant recipients who underwent intensive cancer surveillance.⁵⁵

Renal insufficiency

Renal insufficiency is a well-recognized complication of liver transplantation and is associated with an increased long-term death rate.^56,57

The incidence of renal insufficiency increases dramatically over time. Ojo et al,⁵⁷ in a study of almost 37,000 liver transplant recipients, found that the incidence of chronic kidney disease (defined as an estimated glomerular filtration rate < 30 mL/min/1.73 m²) was 13.9% at 3 years, 18% at 5 years, and approximately 26% at 10 years.

Risk factors include the use of calcineurin inhibitors (both cyclosporine and tacrolimus), older age, female sex, lower pretransplant glomerular filtration rate, postoperative acute renal failure, diabetes, hypertension, hepatitis C virus infection, and transplantation before 1998.^58,59 Replacing a calcineurin inhibitor with mycophenolate mofetil or sirolimus may be considered with communication with the transplant center, as mycophenolate mofetil or sirolimus are associated with a lower risk of renal injury.^60–64

All liver transplant recipients should avoid nonsteroidal anti-inflammatory drugs and nephrotoxic medications

Starting 1 year after liver transplantation, primary care providers should screen for renal dysfunction by obtaining kidney function tests every 6 months, including urinalysis and microalbuminuria assessment. Equations for estimating the glomerular filtration rate used in practice, such as the Modification of Diet in Renal Disease Study equation, rely mainly on serum creatinine, which may lead to overestimating renal function in some circumstances. Therefore, other equations can be used to confirm the estimated glomerular filtration rate measured by creatinine clearance, and to more accurately evaluate kidney function. Calculators are available at www.kidney.org/professionals/kdoqi/gfr_calculator.

All liver transplant recipients should avoid nonsteroidal anti-inflammatory drugs and nephrotoxic medications, and should have their hypertension and diabetes adequately controlled.

Bone diseases

Osteopenia is another major complication of liver transplantation. One-third of liver transplant recipients have a bone mineral density below the fracture threshold.⁶⁵

Multiple factors contribute to increased bone loss after transplantation, including use of corticosteroids, use of calcineurin inhibitors (cyclosporine, tacrolimus), poor nutrition, vitamin D deficiency, immobility, sarcopenia (reduced muscle mass), hypogonadism, smoking, and alcohol abuse.⁶⁶ Even at low doses of less than 7.5 mg per day, corticosteroids inhibit osteoblast activity and increase bone resorption.

Studies have reported rapid bone loss at around 6 months after transplantation.^67–69 However, long-term follow-up of bone mineral density up to 15 years after transplantation revealed an improvement mainly in the 2nd postoperative year, with no deterioration afterward.⁶⁵

High-risk patients need to be identified early with appropriate screening and evaluation. Evaluation includes dual-energy x-ray absorptiometry and serum levels of calcium, phosphorous, parathyroid hormone, testosterone (men), estradiol (women), and alpha-25-hydroxyvitamin D. These tests are typically done before transplantation and then every other year afterward.

We recommend a daily dose of calcitriol and a calcium supplement to all our liver transplant patients.⁷⁰ If osteoporosis (a T-score 2.5 or more standard deviations below the mean) or a fragility fracture occurs, then the patient may benefit from an oral bisphosphonate. Calcitonin has also been shown to improve bone mineral density in patients with osteoporosis after liver transplantation.⁷¹

Hyperuricemia and gout

Although hyperuricemia is common in liver transplant recipients (reported in approximately 47%), the development of clinical gout is less common (6%).⁷² Asymptomatic hyperuricemia requires observation only and is not usually treated in liver transplant recipients.

Acute attacks of gout are typically managed with colchicine 0.6 mg every 2 hours, up to five doses. Prednisone can be considered if symptoms persist despite treatment with colchicine. Allopurinol in an initial dose of 100 mg daily is used as maintenance therapy to reduce production of uric acid.⁷³ However, because of the potential for drug interactions, the combination of azathioprine and allopurinol should be avoided.

Psychiatric complications and quality of life

Depression is common in liver transplant recipients, significantly more so in patients who received a transplant because of hepatitis C.⁷⁴ The type of immunosuppressant is not associated with the incidence of depression. When indicated, the internist may start the patient on a selective serotonin reuptake inhibitor such as citalopram 20 mg daily, as these medications are usually effective and well tolerated in liver transplant patients.⁷³

Liver transplantation has a major positive effect on quality of life. Most patients with end-stage liver disease have poor quality of life before transplantation, but this seems to improve notably afterward. A meta-analysis showed significant improvement in posttransplant physical health, sexual functioning, daily activities, and social functioning compared with before transplantation.⁷⁵

Alcohol abuse and smoking

Patients who underwent liver transplantation because of alcoholic liver disease should be advised to abstain from alcohol.¹⁹ Patients who underwent the procedure for a different indication are advised to avoid excessive alcohol intake, as it is proven to lower the survival rate.²⁰ Alcohol recidivism and smoking (including marijuana) are major problems, and internists are best positioned to address these issues and treat them.

Vaccinations

All liver transplant recipients should be vaccinated against influenza, pneumococcal infection, and tetanus. Hepatitis A and B vaccines are typically given before transplantation. In general, live vaccines such as measles-mumps-rubella and varicella are not recommended after any solid organ transplant.⁷⁶

A study in Germany showed that immunization rates were too low in solid-organ transplant recipients, and almost 90% of patients were not adequately informed about immunizations.⁷⁷ Hence, there may be room for improvement, and primary care providers should take the lead toward better outcomes in this regard.

Recurrence of the primary liver disease after transplantation

Different primary liver diseases recur with different frequencies.

Hepatitis C has the highest rate of recurrence of the liver diseases.^78,79 Reinfection with hepatitis C virus after liver transplantation is almost universal and can follow different patterns. One of the most aggressive patterns is fibrosing cholestatic hepatitis, which frequently leads to graft failure and death, and hence necessitates urgent detection and treatment.

Hepatocellular carcinoma also has a high recurrence rate.⁸⁰ Surveillance with liver ultrasonography or computed tomography is required every 6 months for the first 5 years after liver transplantation.

Other liver diseases. Nonalcoholic steatohepatitis, primary biliary cirrhosis, primary sclerosing cholangitis, autoimmune hepatitis, and hepatitis B infection also tend to recur after liver transplantation.^46,81 On the other hand, alpha-1 antitrypsin deficiency, Wilson disease, hemochromatosis, and metabolic disorders are “cured” after liver transplantation.

It is important to detect any increase in liver enzymes above baseline. An elevation of 1.5 times the upper limit of normal or more should trigger further investigation.

Allograft dysfunction

A number of complications can develop in the liver allograft and result in abnormal liver function tests and, if not treated, graft failure. The most common causes of late graft dysfunction include recurrence of primary liver disease, biliary complications, and chronic rejection.⁴⁶

Vascular complications include hepatic artery thrombosis and stenosis and are usually evaluated by liver ultrasonography and Doppler scan of the hepatic artery and venous structures.²⁴

Biliary strictures give a cholestatic picture, with elevated bilirubin and greater elevation of alkaline phosphatase than of alanine aminotransferase and aspartate aminotransferase. Strictures are usually treated by endoscopic dilation and stenting, but they may eventually require surgery.

Late acute cellular rejections occur in 10% to 20% of cases and are a risk factor for chronic rejection. Liver biopsy is needed to make the diagnosis, and pulsed doses of corticosteroids remain the backbone of treatment therapy.

Chronic rejection is not common, occurring in 3% to 4% of liver transplant recipients.⁴⁶ Treatment is based on increasing immunosuppression and ensuring compliance with prescribed medications. However, chronic rejection may not respond well, and repeat transplantation may be the last resort for some patients.

WHEN TO REFER TO THE HEPATOLOGIST

Some situations require referral to the hepatologist or the transplant center. In general, the following are best managed by a hepatologist: adjustment of immunosuppressive drugs and dosages, allograft dysfunction, vascular and biliary complications, progressing renal dysfunction, and recurrence of primary liver disease. Early communication with a hepatologist and the transplant center is recommended in these cases.

A 40-year-old man with end-stage renal disease on intermittent hemodialysis presented to the emergency department with a 1-week history of pain affecting his left lower back, left flank, and left lower abdomen, diagnosed as zoster prodrome.

Of note, both corneas were cloudy, most severely in the limbus (Figure 1). His visual acuity and findings on funduscopic examination were normal.

CORNEAL OPACITY

The finding of corneal opacity should prompt an immediate ophthalmologic evaluation by the internist as well as an ophthalmologist. The initial examination should include visual acuity testing; gross examination with the naked eye; penlight examination of the pupil, conjunctiva, and anterior chamber; funduscopic examination to at least confirm a red reflex; and fluorescein examination of the cornea. Fluorescein testing is done last, as the dye may interfere with the other initial tests.¹

A number of causes of opacity

A number of conditions can cause corneal opacity. Several genetic conditions can cause developmental anomalies of the cornea, leading to corneal defects present at birth.² Causes of secondary corneal opacity in early infancy include infections such as herpes, iatrogenic injury during amniocentesis or forceps delivery, and infantile congenital glaucoma.²

Later in life, causes of corneal opacity include cataract, glaucoma, chemical exposure, foreign body injury, irradiation, infection (eg, syphilis, herpes, chlamydia), endophthalmitis, and metabolic genetic disorders such as Fabry disease, trisomy 18 syndrome, and lecithin-cholesterol acyltransferase (LCAT) deficiency.³

LCAT DEFICIENCY

LCAT is a key protein in reverse transport of cholesterol from the systemic circulation to the liver for excretion into the bile. Its deficiency results in low serum concentrations of high-density lipoprotein cholesterol (HDL-C).⁴ About 80 different mutations in the LCAT gene have been linked to LCAT deficiency.⁵

LCAT deficiency varies in severity. Patients with complete deficiency can have nearly undetectable levels of HDL-C, eruptive xanthoma, hepatosplenomegaly, and premature coronary artery disease (ie, by age 40).5,6 Features of coronary atherosclerosis can be lacking in patients with partial deficiency.

Regardless of the degree of LCAT deficiency, most patients have corneal opacification that is most severe near the limbus (thus, the term “fish eye syndrome”) and anemia.⁷ Although corneal opacification presents early in life and persists, it does not seem to affect vision.⁵ The anemia is associated with enhanced fractional clearance of red blood cells secondary to hypersplenism.⁸

LCAT deficiency and the kidneys

LCAT deficiency has its most devastating effect on the kidney. Renal disease begins early in life with mild proteinuria and microscopic hematuria. With increasing age, renal function deteriorates and proteinuria and hematuria worsen.⁹

Renal biopsy study may reveal foam cells in the glomerular tufts, arterioles with thickened intima and narrowed lumens, and subendothelial deposits of lipids in the renal arteries and arterioles.¹⁰ Some studies have suggested that kidney disease is most likely initiated by lipid deposition or cellular uptake of lipoproteins in the glomerular basement membrane, mesangium, and capillary subendothelium.

Treatment

There are few treatment options for patients with LCAT deficiency. Control of hypertension, if present, may halt or slow renal deterioration.⁵ Many patients eventually require dialysis, and some undergo renal transplantation, but the renal disease can recur.

OUR PATIENT

Our patient had a known diagnosis of LCAT deficiency. Five years before this presentation at our emergency department, he developed malignant hypertension, followed shortly by renal disease. Over the next 4 years, his kidney function deteriorated, culminating in the need for dialysis; his corneal opacities manifested and gradually worsened; and after extensive studies including kidney biopsies, he was finally diagnosed with LCAT deficiency.

He also exhibited a chronically low level of HDL-C (2 to 5 mg/dL) and significant coronary artery disease. Although unrelated, his zoster pain was treated with renally dosed acyclovir and gabapentin. He never demonstrated the characteristic rash, and his pain improved significantly within 5 days of treatment.

A 40-year-old man with end-stage renal disease on intermittent hemodialysis presented to the emergency department with a 1-week history of pain affecting his left lower back, left flank, and left lower abdomen, diagnosed as zoster prodrome.

Of note, both corneas were cloudy, most severely in the limbus (Figure 1). His visual acuity and findings on funduscopic examination were normal.

CORNEAL OPACITY

The finding of corneal opacity should prompt an immediate ophthalmologic evaluation by the internist as well as an ophthalmologist. The initial examination should include visual acuity testing; gross examination with the naked eye; penlight examination of the pupil, conjunctiva, and anterior chamber; funduscopic examination to at least confirm a red reflex; and fluorescein examination of the cornea. Fluorescein testing is done last, as the dye may interfere with the other initial tests.¹

A number of causes of opacity

A number of conditions can cause corneal opacity. Several genetic conditions can cause developmental anomalies of the cornea, leading to corneal defects present at birth.² Causes of secondary corneal opacity in early infancy include infections such as herpes, iatrogenic injury during amniocentesis or forceps delivery, and infantile congenital glaucoma.²

Later in life, causes of corneal opacity include cataract, glaucoma, chemical exposure, foreign body injury, irradiation, infection (eg, syphilis, herpes, chlamydia), endophthalmitis, and metabolic genetic disorders such as Fabry disease, trisomy 18 syndrome, and lecithin-cholesterol acyltransferase (LCAT) deficiency.³

LCAT DEFICIENCY

LCAT is a key protein in reverse transport of cholesterol from the systemic circulation to the liver for excretion into the bile. Its deficiency results in low serum concentrations of high-density lipoprotein cholesterol (HDL-C).⁴ About 80 different mutations in the LCAT gene have been linked to LCAT deficiency.⁵

LCAT deficiency varies in severity. Patients with complete deficiency can have nearly undetectable levels of HDL-C, eruptive xanthoma, hepatosplenomegaly, and premature coronary artery disease (ie, by age 40).5,6 Features of coronary atherosclerosis can be lacking in patients with partial deficiency.

Regardless of the degree of LCAT deficiency, most patients have corneal opacification that is most severe near the limbus (thus, the term “fish eye syndrome”) and anemia.⁷ Although corneal opacification presents early in life and persists, it does not seem to affect vision.⁵ The anemia is associated with enhanced fractional clearance of red blood cells secondary to hypersplenism.⁸

LCAT deficiency and the kidneys

LCAT deficiency has its most devastating effect on the kidney. Renal disease begins early in life with mild proteinuria and microscopic hematuria. With increasing age, renal function deteriorates and proteinuria and hematuria worsen.⁹

Renal biopsy study may reveal foam cells in the glomerular tufts, arterioles with thickened intima and narrowed lumens, and subendothelial deposits of lipids in the renal arteries and arterioles.¹⁰ Some studies have suggested that kidney disease is most likely initiated by lipid deposition or cellular uptake of lipoproteins in the glomerular basement membrane, mesangium, and capillary subendothelium.

Treatment

There are few treatment options for patients with LCAT deficiency. Control of hypertension, if present, may halt or slow renal deterioration.⁵ Many patients eventually require dialysis, and some undergo renal transplantation, but the renal disease can recur.

OUR PATIENT

Our patient had a known diagnosis of LCAT deficiency. Five years before this presentation at our emergency department, he developed malignant hypertension, followed shortly by renal disease. Over the next 4 years, his kidney function deteriorated, culminating in the need for dialysis; his corneal opacities manifested and gradually worsened; and after extensive studies including kidney biopsies, he was finally diagnosed with LCAT deficiency.

He also exhibited a chronically low level of HDL-C (2 to 5 mg/dL) and significant coronary artery disease. Although unrelated, his zoster pain was treated with renally dosed acyclovir and gabapentin. He never demonstrated the characteristic rash, and his pain improved significantly within 5 days of treatment.

A 40-year-old man with end-stage renal disease on intermittent hemodialysis presented to the emergency department with a 1-week history of pain affecting his left lower back, left flank, and left lower abdomen, diagnosed as zoster prodrome.

Of note, both corneas were cloudy, most severely in the limbus (Figure 1). His visual acuity and findings on funduscopic examination were normal.

CORNEAL OPACITY

The finding of corneal opacity should prompt an immediate ophthalmologic evaluation by the internist as well as an ophthalmologist. The initial examination should include visual acuity testing; gross examination with the naked eye; penlight examination of the pupil, conjunctiva, and anterior chamber; funduscopic examination to at least confirm a red reflex; and fluorescein examination of the cornea. Fluorescein testing is done last, as the dye may interfere with the other initial tests.¹

A number of causes of opacity

A number of conditions can cause corneal opacity. Several genetic conditions can cause developmental anomalies of the cornea, leading to corneal defects present at birth.² Causes of secondary corneal opacity in early infancy include infections such as herpes, iatrogenic injury during amniocentesis or forceps delivery, and infantile congenital glaucoma.²

Later in life, causes of corneal opacity include cataract, glaucoma, chemical exposure, foreign body injury, irradiation, infection (eg, syphilis, herpes, chlamydia), endophthalmitis, and metabolic genetic disorders such as Fabry disease, trisomy 18 syndrome, and lecithin-cholesterol acyltransferase (LCAT) deficiency.³

LCAT DEFICIENCY

LCAT is a key protein in reverse transport of cholesterol from the systemic circulation to the liver for excretion into the bile. Its deficiency results in low serum concentrations of high-density lipoprotein cholesterol (HDL-C).⁴ About 80 different mutations in the LCAT gene have been linked to LCAT deficiency.⁵

LCAT deficiency varies in severity. Patients with complete deficiency can have nearly undetectable levels of HDL-C, eruptive xanthoma, hepatosplenomegaly, and premature coronary artery disease (ie, by age 40).5,6 Features of coronary atherosclerosis can be lacking in patients with partial deficiency.

Regardless of the degree of LCAT deficiency, most patients have corneal opacification that is most severe near the limbus (thus, the term “fish eye syndrome”) and anemia.⁷ Although corneal opacification presents early in life and persists, it does not seem to affect vision.⁵ The anemia is associated with enhanced fractional clearance of red blood cells secondary to hypersplenism.⁸

LCAT deficiency and the kidneys

LCAT deficiency has its most devastating effect on the kidney. Renal disease begins early in life with mild proteinuria and microscopic hematuria. With increasing age, renal function deteriorates and proteinuria and hematuria worsen.⁹

Renal biopsy study may reveal foam cells in the glomerular tufts, arterioles with thickened intima and narrowed lumens, and subendothelial deposits of lipids in the renal arteries and arterioles.¹⁰ Some studies have suggested that kidney disease is most likely initiated by lipid deposition or cellular uptake of lipoproteins in the glomerular basement membrane, mesangium, and capillary subendothelium.

Treatment

There are few treatment options for patients with LCAT deficiency. Control of hypertension, if present, may halt or slow renal deterioration.⁵ Many patients eventually require dialysis, and some undergo renal transplantation, but the renal disease can recur.

OUR PATIENT

Our patient had a known diagnosis of LCAT deficiency. Five years before this presentation at our emergency department, he developed malignant hypertension, followed shortly by renal disease. Over the next 4 years, his kidney function deteriorated, culminating in the need for dialysis; his corneal opacities manifested and gradually worsened; and after extensive studies including kidney biopsies, he was finally diagnosed with LCAT deficiency.

He also exhibited a chronically low level of HDL-C (2 to 5 mg/dL) and significant coronary artery disease. Although unrelated, his zoster pain was treated with renally dosed acyclovir and gabapentin. He never demonstrated the characteristic rash, and his pain improved significantly within 5 days of treatment.

A 24-year-old woman from northern India came to our medical center because of lower back pain for the past 2 years. The pain was initially a dull, continuous ache and did not radiate. She had no fever, night sweats, weight loss, or other constitutional symptoms.

In addition, she had seen her local practitioner 1 year earlier because of burning during urination and occasional frequency. She had been found to have an 8-mm calculus in the lower calyx of the left kidney, for which she underwent two sessions of shock-wave lithotripsy, but she did not pass any stone fragments. Because her back pain continued, she sought medical treatment at our center.

On evaluation at our facility, she was found to have paraspinal muscle spasm and scoliosis. Her gait was antalgic. Sensations were normal over both lower limbs in all dermatomes. Power was grade 5 throughout, and deep tendon reflexes were normal. The straight-leg-raising test was positive for reproducible pain in the lower back and sciatic pain radiating down the back of both legs.

Laboratory testing showed that her hemoglobin was low at 9.7 g/dL (reference range 11.5–15.5), but the rest of the complete blood cell count was within normal limits. C-reactive protein was elevated at 70.7 mg/L (reference range < 6 mg/L). An enzyme-linked immunosorbent assay was negative for human immunodeficiency virus (HIV).

Nonenhanced computed tomography of the abdomen revealed destruction of vertebral body end plates and disks from the L2 lower end plate to the L5 superior end plate. The left transverse processes of the L3, L4, and L5 vertebral bodies were also destroyed. The scan also revealed bilateral psoas abscesses larger than 10 by 10 cm (Figure 1), with the right side larger than the left, and confirmed a stone in the left lower renal calyx (Figure 2).

She underwent bilateral ultrasonographically guided drainage of the abscesses. Culture of the thick pus that was aspirated grew Mycobacterium tuberculosis. Tuberculosis therapy was started with isoniazid, rifampicin, pyrazinamide, and ethambutol. Her condition improved rapidly over the next 2 to 3 months. She completed 18 months of tuberculosis therapy.

Because her spine was stable, with no collapse of vertebrae, she did not require orthopedic intervention.

SPINAL TUBERCULOSIS

Spinal tuberculosis, or Pott disease, is still a common cause of back pain in areas where the infection is rampant, such as northern India.¹Mycobacterium infections continue to be a problem, especially coexisting with HIV infection.^2,3 In fact, the World Health Organization and the United States Agency for International Development have referred to this as a twin epidemic.⁴

Early diagnosis and prompt treatment can prevent or minimize spinal deformity and permanent neurologic disability.⁵ Rapidly progressive and significant neurologic involvement requires surgical management. On the other hand, Patil et al⁶ reported a series of 50 cases in which early radiologic evidence of spinal cord compression from tuberculosis was managed nonoperatively.

When evaluating back pain, symptoms that should ring the alarm include weight loss, constitutional symptoms, no change in pain status after 6 weeks of treatment with a nonsteroidal anti-inflammatory drug, pain at night or at rest, and neurologic symptoms. Our patient had no relief of pain and thus sought treatment.

An important take-home message is that small, nonobstructive renal calculi almost never cause back pain, and when incidentally detected, as in this patient, should not be considered the cause of back pain.

A 24-year-old woman from northern India came to our medical center because of lower back pain for the past 2 years. The pain was initially a dull, continuous ache and did not radiate. She had no fever, night sweats, weight loss, or other constitutional symptoms.

In addition, she had seen her local practitioner 1 year earlier because of burning during urination and occasional frequency. She had been found to have an 8-mm calculus in the lower calyx of the left kidney, for which she underwent two sessions of shock-wave lithotripsy, but she did not pass any stone fragments. Because her back pain continued, she sought medical treatment at our center.

On evaluation at our facility, she was found to have paraspinal muscle spasm and scoliosis. Her gait was antalgic. Sensations were normal over both lower limbs in all dermatomes. Power was grade 5 throughout, and deep tendon reflexes were normal. The straight-leg-raising test was positive for reproducible pain in the lower back and sciatic pain radiating down the back of both legs.

Laboratory testing showed that her hemoglobin was low at 9.7 g/dL (reference range 11.5–15.5), but the rest of the complete blood cell count was within normal limits. C-reactive protein was elevated at 70.7 mg/L (reference range < 6 mg/L). An enzyme-linked immunosorbent assay was negative for human immunodeficiency virus (HIV).

Nonenhanced computed tomography of the abdomen revealed destruction of vertebral body end plates and disks from the L2 lower end plate to the L5 superior end plate. The left transverse processes of the L3, L4, and L5 vertebral bodies were also destroyed. The scan also revealed bilateral psoas abscesses larger than 10 by 10 cm (Figure 1), with the right side larger than the left, and confirmed a stone in the left lower renal calyx (Figure 2).

She underwent bilateral ultrasonographically guided drainage of the abscesses. Culture of the thick pus that was aspirated grew Mycobacterium tuberculosis. Tuberculosis therapy was started with isoniazid, rifampicin, pyrazinamide, and ethambutol. Her condition improved rapidly over the next 2 to 3 months. She completed 18 months of tuberculosis therapy.

Because her spine was stable, with no collapse of vertebrae, she did not require orthopedic intervention.

SPINAL TUBERCULOSIS

Spinal tuberculosis, or Pott disease, is still a common cause of back pain in areas where the infection is rampant, such as northern India.¹Mycobacterium infections continue to be a problem, especially coexisting with HIV infection.^2,3 In fact, the World Health Organization and the United States Agency for International Development have referred to this as a twin epidemic.⁴

Early diagnosis and prompt treatment can prevent or minimize spinal deformity and permanent neurologic disability.⁵ Rapidly progressive and significant neurologic involvement requires surgical management. On the other hand, Patil et al⁶ reported a series of 50 cases in which early radiologic evidence of spinal cord compression from tuberculosis was managed nonoperatively.

When evaluating back pain, symptoms that should ring the alarm include weight loss, constitutional symptoms, no change in pain status after 6 weeks of treatment with a nonsteroidal anti-inflammatory drug, pain at night or at rest, and neurologic symptoms. Our patient had no relief of pain and thus sought treatment.

An important take-home message is that small, nonobstructive renal calculi almost never cause back pain, and when incidentally detected, as in this patient, should not be considered the cause of back pain.

A 24-year-old woman from northern India came to our medical center because of lower back pain for the past 2 years. The pain was initially a dull, continuous ache and did not radiate. She had no fever, night sweats, weight loss, or other constitutional symptoms.

In addition, she had seen her local practitioner 1 year earlier because of burning during urination and occasional frequency. She had been found to have an 8-mm calculus in the lower calyx of the left kidney, for which she underwent two sessions of shock-wave lithotripsy, but she did not pass any stone fragments. Because her back pain continued, she sought medical treatment at our center.

On evaluation at our facility, she was found to have paraspinal muscle spasm and scoliosis. Her gait was antalgic. Sensations were normal over both lower limbs in all dermatomes. Power was grade 5 throughout, and deep tendon reflexes were normal. The straight-leg-raising test was positive for reproducible pain in the lower back and sciatic pain radiating down the back of both legs.

Laboratory testing showed that her hemoglobin was low at 9.7 g/dL (reference range 11.5–15.5), but the rest of the complete blood cell count was within normal limits. C-reactive protein was elevated at 70.7 mg/L (reference range < 6 mg/L). An enzyme-linked immunosorbent assay was negative for human immunodeficiency virus (HIV).

Nonenhanced computed tomography of the abdomen revealed destruction of vertebral body end plates and disks from the L2 lower end plate to the L5 superior end plate. The left transverse processes of the L3, L4, and L5 vertebral bodies were also destroyed. The scan also revealed bilateral psoas abscesses larger than 10 by 10 cm (Figure 1), with the right side larger than the left, and confirmed a stone in the left lower renal calyx (Figure 2).

She underwent bilateral ultrasonographically guided drainage of the abscesses. Culture of the thick pus that was aspirated grew Mycobacterium tuberculosis. Tuberculosis therapy was started with isoniazid, rifampicin, pyrazinamide, and ethambutol. Her condition improved rapidly over the next 2 to 3 months. She completed 18 months of tuberculosis therapy.

Because her spine was stable, with no collapse of vertebrae, she did not require orthopedic intervention.

SPINAL TUBERCULOSIS

Spinal tuberculosis, or Pott disease, is still a common cause of back pain in areas where the infection is rampant, such as northern India.¹Mycobacterium infections continue to be a problem, especially coexisting with HIV infection.^2,3 In fact, the World Health Organization and the United States Agency for International Development have referred to this as a twin epidemic.⁴

Early diagnosis and prompt treatment can prevent or minimize spinal deformity and permanent neurologic disability.⁵ Rapidly progressive and significant neurologic involvement requires surgical management. On the other hand, Patil et al⁶ reported a series of 50 cases in which early radiologic evidence of spinal cord compression from tuberculosis was managed nonoperatively.

When evaluating back pain, symptoms that should ring the alarm include weight loss, constitutional symptoms, no change in pain status after 6 weeks of treatment with a nonsteroidal anti-inflammatory drug, pain at night or at rest, and neurologic symptoms. Our patient had no relief of pain and thus sought treatment.

An important take-home message is that small, nonobstructive renal calculi almost never cause back pain, and when incidentally detected, as in this patient, should not be considered the cause of back pain.

Any time patients enter or leave the hospital, they risk being harmed by errors in their medications.¹ Adverse events from medication errors during transitions of care are common but often preventable. One key approach is to systematically review every patient’s medication list on admission and discharge and resolve any discrepancies. These transitions are also an opportunity to address other medication-related problems, such as adherence, drug interactions, and clinical appropriateness.

This article summarizes the types and prevalence of medication problems that occur during hospital-based transitions of care, and suggests strategies to decrease the risk of medication errors, focusing on medication reconciliation and related interventions that clinicians can use at the bedside to improve medication safety.

DEFINING TERMS

A medication discrepancy is any variance noted in a patient’s documented medication regimen across different medication lists or sites of care. While some differences reflect intentional and clinically appropriate changes to the regimen, others are unintentional and reflect inaccurate or incomplete information. These unintentional discrepancies are medication errors.

Depending on the clinical circumstances and medications involved, such errors may lead to an adverse drug event (ADE), defined as actual harm or injury resulting from a medication. Sometimes a medication error does not cause harm immediately but could if left uncorrected; this is called a potential ADE.

An important goal during transitions of care is to reduce unintentional medication discrepancies, thereby reducing potential and actual ADEs.

ERRORS ARISE AT DISCHARGE—AND EVEN MORE AT ADMISSION

Hospital discharge is a widely recognized transition in which patient harm occurs. As many as 70% of patients may have an unintentional medication discrepancy at hospital discharge, with many of those discrepancies having potential for harm.² Indeed, during the first few weeks after discharge, 50% of patients have a clinically important medication error,³ and 20% experience an adverse event, most commonly an ADE.⁴ ADEs are associated with excess health care utilization,^5–7 and many are preventable through strategies such as medication reconciliation.^5,8

Importantly, more errors arise at hospital admission than at other times.^9,10

Errors in medication histories are the most common source of discrepancies, affecting up to two-thirds of admitted patients.^11,12 More than one-quarter of hospital prescribing errors can be attributed to incomplete medication histories at the time of admission,¹³ and nearly three times as many clinically important medication discrepancies are related to history-taking errors on admission rather than reconciliation errors at discharge.⁹

Most discrepancies occurring at the time of admission have the potential to cause harm, particularly if the errors persist beyond discharge.¹⁴ Therefore, taking a complete and accurate medication history on hospital admission is critical to ensuring safe care transitions.

MEDICATION RECONCILIATION: BARRIERS AND FACILITATORS

Medication reconciliation is a strategy for reducing medication discrepancies in patients moving across care settings. Simply put, it is the process by which a patient’s medication list is obtained, compared, and clarified across different sites of care.¹⁵ It has consistently been shown to decrease medication errors compared with usual care,¹⁶ and it is strongly supported by national and international organizations.^17–21

In clinical practice, many physicians and institutions have found medication reconciliation difficult to implement, owing to barriers at the level of the patient, provider, and system (Table  1). In response to these challenges, two initiatives have synthesized best practices and offer toolkits that hospitals and clinicians can use: the Medications at Transitions and Clinical Handoffs (MATCH) program²² and the Multi-Center Medication Reconciliation Quality Improvement Study (MARQUIS).²³

Lack of resources is a widely acknowledged challenge. Thus, the MARQUIS investigators²³ suggested focusing on the admission history, where most errors occur, and applying the most resource-intensive interventions in patients at highest risk of ADEs, ie, those who are elderly, have multiple comorbid conditions, or take numerous medications.¹⁶

Although the risk of an ADE increases with the number of medications a patient takes,⁴ the exact number of drugs that defines high risk has not been well established. Targeting patients who take 10 or more maintenance medications is a reasonable initial approach,²⁴ but institutions should tailor risk stratification to their patient populations and available resources. Patients taking high-risk medications such as anticoagulants and insulin could also be prioritized for review, since these medications are more likely to cause serious patient harm when used without appropriate clinical oversight.⁷

Using pharmacy staff to perform medication history-taking, reconciliation, and patient counseling has been shown to produce favorable patient outcomes, particularly for higher-risk patients.^16,23

The MARQUIS investigators found they could boost the chances of success by sharing stories of patient harm to foster “buy-in” among frontline staff, providing formal training to clinicians on how to take a medication history, and obtaining the support of nursing leaders to champion improvement efforts.²³

Additionally, patients should be empowered to maintain an accurate medication list. We address strategies for improving patient engagement and adherence in a later section.

BEST PRACTICES FOR IMPROVING MEDICATION SAFETY

Medication reconciliation is one of several measures necessary to optimize medication safety during transitions of care. It typically includes the following actions:

• Interview the patient or caregiver to determine the list of medications the patient is currently taking (or supposed to be taking); consult other sources if needed.
• List medications that are being ordered during the clinical encounter.
• Compare these two lists, making note of medications that are stopped, changed, or newly prescribed.
• Resolve any discrepancies.
• Communicate the reconciled list to the patient, appropriate caregivers, and providers of follow-up care.

At a rudimentary level, medication reconciliation encompasses medication list management along the continuum of care. However, we recommend leveraging medication reconciliation as an opportunity to further enhance medication safety by reviewing the appropriateness of each medication, seizing opportunities to streamline or simplify the regimen, assessing patient and caregiver understanding of medication instructions and potential ADEs, and delivering appropriate counseling to enhance medication use. Table 2 outlines our framework for medication management during hospital-based transitions.

STEP 1: OBTAIN A COMPLETE PREADMISSION MEDICATION LIST

The “best-possible medication history” is obtained in a systematic process of interviewing the patient or caregiver plus reviewing at least one other reliable source.²³ The resulting list should include all medications the patient is taking (prescription and nonprescription), doses, directions for use, formulations if applicable, indications, start and stop dates, and medication allergies and reactions.

Review existing information. Before eliciting a history from the patient, review his or her recorded medical history and existing medication lists (eg, prior discharge summaries, records from other facilities, records from outpatient visits, pharmacy fill data). This will provide context about the regimen and help identify issues and questions that can be addressed during the history-taking.

Ask open-ended questions. Instead of just asking the patient to confirm the accuracy of the existing medication list, we recommend actively obtaining the full medication list from the patient or caregiver. The conversation should begin with an open-ended question such as, “What medications do you take at home?” This approach will also allow the clinician to gauge the patient’s level of understanding of each medication’s indication and dosing instructions. Using a series of prompts such as those recommended in Table 3 will elicit a best-possible medical history, while verifying all of the medications on the existing list.

Clarify discrepancies. Resolve differences between the existing medication lists and the patient’s or caregiver’s report during the preadmission interview. Examples include errors of omission (a medication is missing), errors of commission (an additional medication is present), and discrepancies in the strength, formulation, dosing instructions, and indications for the drugs. If necessary, other sources of information should be consulted, such as the patient’s medication bottles, pharmacy or pharmacies, primary care physician, and a family member or caregiver.

Assess adherence. The extent to which patients take their medications as directed is an important component of the history, but is often left out. Medication nonadherence rates in the United States are 40% to 70%,²⁵ contributing to poor patient outcomes and imposing extraordinary costs on the health care system.²⁶

Asking open-ended, nonjudgmental questions at the time of hospital admission will help to uncover medication-taking behaviors as well as barriers to adherence (Table 3). The patient’s responses should be taken into account when determining the treatment plan.

Document your findings. After completing the medication history and clarifying discrepancies, document the preadmission list in the medical record. All members of the health care team should have access to view and update the same list, as new information about the preadmission medications may be uncovered after the initial history.

Make clinical decisions. Complete the admission medication reconciliation by deciding whether each medication on the list should be continued, changed, held, or discontinued on the basis of the patient’s clinical condition. Well-designed information technology applications enable the provider to document each action and the rationale for it, as well as carry that information into the order-entry system. Medications marked as held or discontinued on admission should be revisited as the patient’s clinical condition changes and at discharge.

STEP 2: AVOID RECONCILIATION ERRORS

Reconciliation errors reflect discrepancies between the medication history and the medications that are ordered after admission.

Reconciliation errors are less common than medication history errors, accounting for approximately one-third of potentially harmful medication errors in hospitalized medical patients.⁹ These include errors of omission (a medication is omitted from the orders), errors of commission (a medication is prescribed with no indication for continuation), and therapeutic duplication.

Preventing errors of omission

Medications are often held at transition points for appropriate clinical reasons. Examples include holding anticoagulants and antiplatelet agents in patients who have gastrointestinal bleeding or an upcoming procedure, antihypertensives in patients with hemodynamic instability, and other chronic medications in patients with an acute illness.

Poor documentation and communication of these decisions can lead to a failure to resume the medications—an error of omission—at hospital discharge.

Hospitalized patients are at risk of unintentional discontinuation of their chronic medications, including antiplatelet drugs, anticoagulants, statins, and thyroid replacement, particularly if admitted to the intensive care unit.¹² These errors can be minimized by a standardized medication reconciliation process at each transition and clear documentation of the medication plan.

Communication among providers can be improved if the admitting clinician documents clearly whether each preadmission medication is being continued, changed, or stopped, along with the reason for doing so, and makes this information available throughout the hospital stay. Upon transfer to another unit and at discharge, the physician should review each; preadmission medication that was held and the patient’s current clinical status and, based on that information, decide whether medications that were held should be resumed. If a medication will be restarted later, specific instructions should be documented and communicated to the patient and the physicians who are taking over his or her care.

Preventing errors of commission

Failure to perform a complete reconciliation at each transition of care and match each medication with an appropriate indication can lead to errors of commission.

One study showed that 44% of patients were prescribed at least one unnecessary drug at hospital discharge, one-fourth of which were started during the hospitalization.²⁷ Commonly prescribed unnecessarily were gastrointestinal agents, central nervous system drugs, nutrients, and supplements.

It is critical to assess each medication’s ongoing need, appropriateness, and risk-benefit ratio at every transition. Medications no longer indicated should be discontinued in order to simplify the regimen, avoid unnecessary drug exposure, and prevent ADEs.

For example, proton pump inhibitors or histamine 2 receptor blockers are often started in the hospital for stress ulcer prophylaxis. One-third of patients are then discharged home on the medication, and 6 months later half of those patients are still taking the unnecessary drug.²⁸ This situation can be avoided by limiting use of these medications to appropriate circumstances, clearly marking the indication as stress ulcer prophylaxis (as opposed to an ongoing condition that will require continuing it after discharge), and discontinuing the agent when appropriate.

All drugs, even common and seemingly benign ones, carry some risk and should be discontinued when no longer needed. Thus, medications added during the hospitalization to control acute symptoms should also be reviewed at each transition to prevent inappropriate continuation when symptoms have resolved.

One study, for example, found that many patients were discharged with inappropriate prescriptions for atypical antipsychotics after receiving them in the intensive care unit, likely for delirium.²⁹ Documenting that an acute issue such as delirium has resolved should prompt the discontinuation of therapy.

Preventing therapeutic duplication

Therapeutic duplication occurs in about 8% of discharges.¹ These errors often result from formulary substitutions or altering the dosage form in the acute setting. For example, patients who receive a prescription for the substituted agent at discharge and also resume their prehospitalization medications end up with duplicate therapy.

Therapeutic substitution is common at the time of admission to the hospital as a result of formulary restrictions. Drug classes that are frequently substituted include statins, antihypertensives, urinary antispasmodics, and proton pump inhibitors. Physicians should be familiar with the preferred agents on the hospital formulary and make careful note when a substitution occurs. Furthermore, hospital systems should be developed to remind the physician to switch back to the outpatient medication at discharge.

Similar problems occur when home medications are replaced with different dosage forms with different pharmacokinetic properties. For example, a long-acting medication may be temporarily replaced with an intravenous solution or immediate-release tablet for several reasons, including nothing-by-mouth status, unstable clinical condition, need for titration, and need to crush the tablet to give the drug per tube. The differing formulations must be reconciled throughout the patient’s hospital course and at discharge to avoid therapeutic duplication and serious medication errors. Deliberate changes to the dosage form should be clearly communicated in the discharge medication list so that patients and other clinicians are aware.

Hospital systems should also have the capability to identify duplications in the medication list and to warn prescribers of these errors. The ability to group medications by drug class or sort the medication list alphabetically by generic name can help uncover duplication errors.

STEP 3: REVIEW THE LIST IN VIEW OF THE CLINICAL PICTURE

Transitions of care should prompt providers to review the medication list for possible drug-disease interactions, confirm compliance with evidence-based guidelines, and evaluate the risks and benefits of each medication in the context of the patient’s age and acute and chronic medical issues. This is also an opportunity to screen the full list for potentially inappropriate medications and high-alert drugs such as insulin or anticoagulants, which are more likely than other drugs to cause severe harm when used in error.

Acute kidney injury. New drug-disease interactions can arise during a hospitalization and can affect dosing and the choice of drug. The onset of acute kidney injury, for example, often necessitates adjusting or discontinuing nephrotoxic and renally excreted medications. ADEs or potential ADEs have been reported in 43% of hospitalized patients with acute kidney injury.³⁰

Because acute kidney injury is often transient, medications may need to be held or adjusted several times until renal function stabilizes. This can be challenging across the continuum of care and requires close monitoring of the serum creatinine level and associated drug doses and levels, if applicable. Well-designed clinical decision support tools can integrate laboratory data and alert the prescriber to a clinically important increase or decrease in serum creatinine that may warrant a change in therapy. Modifications to the regimen and a plan for timely follow-up of the serum creatinine level should be clearly documented in the discharge plan.

Liver disease. Similar attention should be given to drugs that are hepatically metabolized if a patient has acute or chronic liver impairment.

Geriatric patients, particularly those who present with altered mental status, falls, or urinary retention, should have their medication list reviewed for potentially inappropriate medications, which are drugs that pose increased risk of poor outcomes in older adults.^31,32 Patients and providers may have been willing to accept the risk of medications such as anticholinergics or sedative-hypnotics when the drugs were initiated, but circumstances can change over time, especially in this patient population. Hospitalization is a prime opportunity to screen for medications that meet the Beers criteria³¹ for agents to avoid or use with caution in older adults.

As-needed medications. Medications prescribed on an as-needed basis in the hospital should be reviewed for continuation or discontinuation at discharge. How often the medication was given can inform this decision.

For example, if as-needed opioids were used frequently, failure to develop a plan of care for pain can lead to persistent symptoms and, possibly, to readmission.^33,34 Similar scenarios occur with use of as-needed blood pressure medications, laxatives, and correction-dose insulin.

If an as-needed medication was used consistently during hospitalization, the physician should consider whether a regularly scheduled medication is needed. Conversely, if the medication was not used during the inpatient admission, it can likely be discontinued.

STEP 4: PREPARE THE PATIENT AND FOLLOW-UP PROVIDER

Once a clinician has performed medication reconciliation, including obtaining a best-possible medical history and carefully reviewing the medication list and orders for errors and clinical appropriateness, the next steps are to ensure the patient understands what he or she needs to do and to confirm that suitable follow-up plans are in place. These measures should be taken at all transitions of care but are critically important at hospital discharge.

Preparing the patient and caregiver

An accurate, reconciled medication list should be given to the patient, caregiver, or both, and should be reviewed before discharge.¹⁷

Approximately one-third of Americans have low health literacy skills, so medication lists and associated materials should be easy to understand.³⁵ Medication lists should be written in plain language and formatted for optimal readability (Table 4), clearly stating which medications to continue, change, hold temporarily, and stop.

Patients recall and comprehend about half of the information provided during a medical encounter.³⁶ Thus, medication teaching should focus on key points including changes or additions to the regimen, specific instructions for follow-up and monitoring, and how to handle common and serious side effects.

To confirm patient understanding, clinicians should use “teach-back,” ie, provide the patient with information and then ask him or her to repeat back key points.^37,38 The patient and family should also be encouraged to ask questions before discharge.

If not already addressed during the hospital stay, barriers to medication adherence and ability to obtain the medications should be attended to at this time (Table 5). Also, the plan to pick up the medications should be verified with the patient and caregiver. Verify that there is transportation to a particular pharmacy that is open at the time of discharge, and that the patient can afford the medications.

Ensuring appropriate follow-up

Studies have shown that timely in-home or telephone follow-up after discharge can decrease adverse events and postdischarge health care utilization.^39,40 Telephone follow-up that includes thorough medication reconciliation can help detect and resolve medication issues early after discharge and can close gaps related to monitoring and follow-up.

Medication reconciliation by telephone can be time-consuming. Depending on the number of medications that need to be reviewed, calls can take between 10 and 60 minutes. Postdischarge phone calls should be performed by clinical personnel who are able to identify medication-related problems. A pharmacist should be an available resource to assist with complex regimens, to help resolve medication discrepancies, and to address patient concerns. Table 6 provides tips for conducting follow-up phone calls.

Resolving discrepancies identified during follow-up calls can be difficult, as changes to the medication regimen are often not communicated effectively to other members of the care team. Physicians should document the complete medication list and plan in the discharge summary, and there should be a method for the caller to record updates to the medication list in the medical record so that they are apparent at the outpatient follow-up visit.

An additional challenge is that it is frequently unclear which physician “owns” which medications. Therefore, designating a contact person for each medication until follow-up can be very valuable. At a minimum, a “physician owner” for high-alert medications such as insulin, anticoagulants, and diuretics should be identified to provide close follow-up, titration, and monitoring.

There should also be a plan for the patient to obtain refills of essential long-term medications, such as antiplatelet agents following stent placement.

SUMMARY AND RECOMMENDATIONS

Medication-related problems during hospital admission and discharge are common and range from minor discrepancies in the medication list to errors in history-taking, prescribing, and reconciliation that can lead to potential or actual patient harm. Putting systems in place to facilitate medication reconciliation can decrease the occurrence of medication discrepancies and ADEs, thereby improving patient safety during these critical transitions between care settings and providers. Institutional medication reconciliation programs should focus resources on the admission history-taking step, target the highest-risk patients for the most intensive interventions, and involve pharmacy personnel when possible.

On an individual level, clinicians can incorporate additional interventions into their workflows to optimize medication safety for hospitalized patients. Using a structured approach to obtain a complete and accurate medication list at the time of hospital admission will help providers identify medication-related problems and prevent the propagation of errors throughout the hospital stay and at discharge. Focusing additional time and effort on a comprehensive review of the medication list for errors of omission and commission, patient-specific needs, and high-alert drugs will further decrease the risk of medication errors. Finally, providing discharge counseling targeting patient barriers to adherence and ensuring a proper handover of medication information and rationale for medication changes to outpatient providers will improve the chances of a safe transition.

Any time patients enter or leave the hospital, they risk being harmed by errors in their medications.¹ Adverse events from medication errors during transitions of care are common but often preventable. One key approach is to systematically review every patient’s medication list on admission and discharge and resolve any discrepancies. These transitions are also an opportunity to address other medication-related problems, such as adherence, drug interactions, and clinical appropriateness.

This article summarizes the types and prevalence of medication problems that occur during hospital-based transitions of care, and suggests strategies to decrease the risk of medication errors, focusing on medication reconciliation and related interventions that clinicians can use at the bedside to improve medication safety.

DEFINING TERMS

A medication discrepancy is any variance noted in a patient’s documented medication regimen across different medication lists or sites of care. While some differences reflect intentional and clinically appropriate changes to the regimen, others are unintentional and reflect inaccurate or incomplete information. These unintentional discrepancies are medication errors.

Depending on the clinical circumstances and medications involved, such errors may lead to an adverse drug event (ADE), defined as actual harm or injury resulting from a medication. Sometimes a medication error does not cause harm immediately but could if left uncorrected; this is called a potential ADE.

An important goal during transitions of care is to reduce unintentional medication discrepancies, thereby reducing potential and actual ADEs.

ERRORS ARISE AT DISCHARGE—AND EVEN MORE AT ADMISSION

Hospital discharge is a widely recognized transition in which patient harm occurs. As many as 70% of patients may have an unintentional medication discrepancy at hospital discharge, with many of those discrepancies having potential for harm.² Indeed, during the first few weeks after discharge, 50% of patients have a clinically important medication error,³ and 20% experience an adverse event, most commonly an ADE.⁴ ADEs are associated with excess health care utilization,^5–7 and many are preventable through strategies such as medication reconciliation.^5,8

Importantly, more errors arise at hospital admission than at other times.^9,10

Errors in medication histories are the most common source of discrepancies, affecting up to two-thirds of admitted patients.^11,12 More than one-quarter of hospital prescribing errors can be attributed to incomplete medication histories at the time of admission,¹³ and nearly three times as many clinically important medication discrepancies are related to history-taking errors on admission rather than reconciliation errors at discharge.⁹

Most discrepancies occurring at the time of admission have the potential to cause harm, particularly if the errors persist beyond discharge.¹⁴ Therefore, taking a complete and accurate medication history on hospital admission is critical to ensuring safe care transitions.

MEDICATION RECONCILIATION: BARRIERS AND FACILITATORS

Medication reconciliation is a strategy for reducing medication discrepancies in patients moving across care settings. Simply put, it is the process by which a patient’s medication list is obtained, compared, and clarified across different sites of care.¹⁵ It has consistently been shown to decrease medication errors compared with usual care,¹⁶ and it is strongly supported by national and international organizations.^17–21

In clinical practice, many physicians and institutions have found medication reconciliation difficult to implement, owing to barriers at the level of the patient, provider, and system (Table  1). In response to these challenges, two initiatives have synthesized best practices and offer toolkits that hospitals and clinicians can use: the Medications at Transitions and Clinical Handoffs (MATCH) program²² and the Multi-Center Medication Reconciliation Quality Improvement Study (MARQUIS).²³

Lack of resources is a widely acknowledged challenge. Thus, the MARQUIS investigators²³ suggested focusing on the admission history, where most errors occur, and applying the most resource-intensive interventions in patients at highest risk of ADEs, ie, those who are elderly, have multiple comorbid conditions, or take numerous medications.¹⁶

Although the risk of an ADE increases with the number of medications a patient takes,⁴ the exact number of drugs that defines high risk has not been well established. Targeting patients who take 10 or more maintenance medications is a reasonable initial approach,²⁴ but institutions should tailor risk stratification to their patient populations and available resources. Patients taking high-risk medications such as anticoagulants and insulin could also be prioritized for review, since these medications are more likely to cause serious patient harm when used without appropriate clinical oversight.⁷

Using pharmacy staff to perform medication history-taking, reconciliation, and patient counseling has been shown to produce favorable patient outcomes, particularly for higher-risk patients.^16,23

The MARQUIS investigators found they could boost the chances of success by sharing stories of patient harm to foster “buy-in” among frontline staff, providing formal training to clinicians on how to take a medication history, and obtaining the support of nursing leaders to champion improvement efforts.²³

Additionally, patients should be empowered to maintain an accurate medication list. We address strategies for improving patient engagement and adherence in a later section.

BEST PRACTICES FOR IMPROVING MEDICATION SAFETY

Medication reconciliation is one of several measures necessary to optimize medication safety during transitions of care. It typically includes the following actions:

• Interview the patient or caregiver to determine the list of medications the patient is currently taking (or supposed to be taking); consult other sources if needed.
• List medications that are being ordered during the clinical encounter.
• Compare these two lists, making note of medications that are stopped, changed, or newly prescribed.
• Resolve any discrepancies.
• Communicate the reconciled list to the patient, appropriate caregivers, and providers of follow-up care.

At a rudimentary level, medication reconciliation encompasses medication list management along the continuum of care. However, we recommend leveraging medication reconciliation as an opportunity to further enhance medication safety by reviewing the appropriateness of each medication, seizing opportunities to streamline or simplify the regimen, assessing patient and caregiver understanding of medication instructions and potential ADEs, and delivering appropriate counseling to enhance medication use. Table 2 outlines our framework for medication management during hospital-based transitions.

STEP 1: OBTAIN A COMPLETE PREADMISSION MEDICATION LIST

The “best-possible medication history” is obtained in a systematic process of interviewing the patient or caregiver plus reviewing at least one other reliable source.²³ The resulting list should include all medications the patient is taking (prescription and nonprescription), doses, directions for use, formulations if applicable, indications, start and stop dates, and medication allergies and reactions.

Review existing information. Before eliciting a history from the patient, review his or her recorded medical history and existing medication lists (eg, prior discharge summaries, records from other facilities, records from outpatient visits, pharmacy fill data). This will provide context about the regimen and help identify issues and questions that can be addressed during the history-taking.

Ask open-ended questions. Instead of just asking the patient to confirm the accuracy of the existing medication list, we recommend actively obtaining the full medication list from the patient or caregiver. The conversation should begin with an open-ended question such as, “What medications do you take at home?” This approach will also allow the clinician to gauge the patient’s level of understanding of each medication’s indication and dosing instructions. Using a series of prompts such as those recommended in Table 3 will elicit a best-possible medical history, while verifying all of the medications on the existing list.

Clarify discrepancies. Resolve differences between the existing medication lists and the patient’s or caregiver’s report during the preadmission interview. Examples include errors of omission (a medication is missing), errors of commission (an additional medication is present), and discrepancies in the strength, formulation, dosing instructions, and indications for the drugs. If necessary, other sources of information should be consulted, such as the patient’s medication bottles, pharmacy or pharmacies, primary care physician, and a family member or caregiver.

Assess adherence. The extent to which patients take their medications as directed is an important component of the history, but is often left out. Medication nonadherence rates in the United States are 40% to 70%,²⁵ contributing to poor patient outcomes and imposing extraordinary costs on the health care system.²⁶

Asking open-ended, nonjudgmental questions at the time of hospital admission will help to uncover medication-taking behaviors as well as barriers to adherence (Table 3). The patient’s responses should be taken into account when determining the treatment plan.

Document your findings. After completing the medication history and clarifying discrepancies, document the preadmission list in the medical record. All members of the health care team should have access to view and update the same list, as new information about the preadmission medications may be uncovered after the initial history.

Make clinical decisions. Complete the admission medication reconciliation by deciding whether each medication on the list should be continued, changed, held, or discontinued on the basis of the patient’s clinical condition. Well-designed information technology applications enable the provider to document each action and the rationale for it, as well as carry that information into the order-entry system. Medications marked as held or discontinued on admission should be revisited as the patient’s clinical condition changes and at discharge.

STEP 2: AVOID RECONCILIATION ERRORS

Reconciliation errors reflect discrepancies between the medication history and the medications that are ordered after admission.

Reconciliation errors are less common than medication history errors, accounting for approximately one-third of potentially harmful medication errors in hospitalized medical patients.⁹ These include errors of omission (a medication is omitted from the orders), errors of commission (a medication is prescribed with no indication for continuation), and therapeutic duplication.

Preventing errors of omission

Medications are often held at transition points for appropriate clinical reasons. Examples include holding anticoagulants and antiplatelet agents in patients who have gastrointestinal bleeding or an upcoming procedure, antihypertensives in patients with hemodynamic instability, and other chronic medications in patients with an acute illness.

Poor documentation and communication of these decisions can lead to a failure to resume the medications—an error of omission—at hospital discharge.

Hospitalized patients are at risk of unintentional discontinuation of their chronic medications, including antiplatelet drugs, anticoagulants, statins, and thyroid replacement, particularly if admitted to the intensive care unit.¹² These errors can be minimized by a standardized medication reconciliation process at each transition and clear documentation of the medication plan.

Communication among providers can be improved if the admitting clinician documents clearly whether each preadmission medication is being continued, changed, or stopped, along with the reason for doing so, and makes this information available throughout the hospital stay. Upon transfer to another unit and at discharge, the physician should review each; preadmission medication that was held and the patient’s current clinical status and, based on that information, decide whether medications that were held should be resumed. If a medication will be restarted later, specific instructions should be documented and communicated to the patient and the physicians who are taking over his or her care.

Preventing errors of commission

Failure to perform a complete reconciliation at each transition of care and match each medication with an appropriate indication can lead to errors of commission.

One study showed that 44% of patients were prescribed at least one unnecessary drug at hospital discharge, one-fourth of which were started during the hospitalization.²⁷ Commonly prescribed unnecessarily were gastrointestinal agents, central nervous system drugs, nutrients, and supplements.

It is critical to assess each medication’s ongoing need, appropriateness, and risk-benefit ratio at every transition. Medications no longer indicated should be discontinued in order to simplify the regimen, avoid unnecessary drug exposure, and prevent ADEs.

For example, proton pump inhibitors or histamine 2 receptor blockers are often started in the hospital for stress ulcer prophylaxis. One-third of patients are then discharged home on the medication, and 6 months later half of those patients are still taking the unnecessary drug.²⁸ This situation can be avoided by limiting use of these medications to appropriate circumstances, clearly marking the indication as stress ulcer prophylaxis (as opposed to an ongoing condition that will require continuing it after discharge), and discontinuing the agent when appropriate.

All drugs, even common and seemingly benign ones, carry some risk and should be discontinued when no longer needed. Thus, medications added during the hospitalization to control acute symptoms should also be reviewed at each transition to prevent inappropriate continuation when symptoms have resolved.

One study, for example, found that many patients were discharged with inappropriate prescriptions for atypical antipsychotics after receiving them in the intensive care unit, likely for delirium.²⁹ Documenting that an acute issue such as delirium has resolved should prompt the discontinuation of therapy.

Preventing therapeutic duplication

Therapeutic duplication occurs in about 8% of discharges.¹ These errors often result from formulary substitutions or altering the dosage form in the acute setting. For example, patients who receive a prescription for the substituted agent at discharge and also resume their prehospitalization medications end up with duplicate therapy.

Therapeutic substitution is common at the time of admission to the hospital as a result of formulary restrictions. Drug classes that are frequently substituted include statins, antihypertensives, urinary antispasmodics, and proton pump inhibitors. Physicians should be familiar with the preferred agents on the hospital formulary and make careful note when a substitution occurs. Furthermore, hospital systems should be developed to remind the physician to switch back to the outpatient medication at discharge.

Similar problems occur when home medications are replaced with different dosage forms with different pharmacokinetic properties. For example, a long-acting medication may be temporarily replaced with an intravenous solution or immediate-release tablet for several reasons, including nothing-by-mouth status, unstable clinical condition, need for titration, and need to crush the tablet to give the drug per tube. The differing formulations must be reconciled throughout the patient’s hospital course and at discharge to avoid therapeutic duplication and serious medication errors. Deliberate changes to the dosage form should be clearly communicated in the discharge medication list so that patients and other clinicians are aware.

Hospital systems should also have the capability to identify duplications in the medication list and to warn prescribers of these errors. The ability to group medications by drug class or sort the medication list alphabetically by generic name can help uncover duplication errors.

STEP 3: REVIEW THE LIST IN VIEW OF THE CLINICAL PICTURE

Transitions of care should prompt providers to review the medication list for possible drug-disease interactions, confirm compliance with evidence-based guidelines, and evaluate the risks and benefits of each medication in the context of the patient’s age and acute and chronic medical issues. This is also an opportunity to screen the full list for potentially inappropriate medications and high-alert drugs such as insulin or anticoagulants, which are more likely than other drugs to cause severe harm when used in error.

Acute kidney injury. New drug-disease interactions can arise during a hospitalization and can affect dosing and the choice of drug. The onset of acute kidney injury, for example, often necessitates adjusting or discontinuing nephrotoxic and renally excreted medications. ADEs or potential ADEs have been reported in 43% of hospitalized patients with acute kidney injury.³⁰

Because acute kidney injury is often transient, medications may need to be held or adjusted several times until renal function stabilizes. This can be challenging across the continuum of care and requires close monitoring of the serum creatinine level and associated drug doses and levels, if applicable. Well-designed clinical decision support tools can integrate laboratory data and alert the prescriber to a clinically important increase or decrease in serum creatinine that may warrant a change in therapy. Modifications to the regimen and a plan for timely follow-up of the serum creatinine level should be clearly documented in the discharge plan.

Liver disease. Similar attention should be given to drugs that are hepatically metabolized if a patient has acute or chronic liver impairment.

Geriatric patients, particularly those who present with altered mental status, falls, or urinary retention, should have their medication list reviewed for potentially inappropriate medications, which are drugs that pose increased risk of poor outcomes in older adults.^31,32 Patients and providers may have been willing to accept the risk of medications such as anticholinergics or sedative-hypnotics when the drugs were initiated, but circumstances can change over time, especially in this patient population. Hospitalization is a prime opportunity to screen for medications that meet the Beers criteria³¹ for agents to avoid or use with caution in older adults.

As-needed medications. Medications prescribed on an as-needed basis in the hospital should be reviewed for continuation or discontinuation at discharge. How often the medication was given can inform this decision.

For example, if as-needed opioids were used frequently, failure to develop a plan of care for pain can lead to persistent symptoms and, possibly, to readmission.^33,34 Similar scenarios occur with use of as-needed blood pressure medications, laxatives, and correction-dose insulin.

If an as-needed medication was used consistently during hospitalization, the physician should consider whether a regularly scheduled medication is needed. Conversely, if the medication was not used during the inpatient admission, it can likely be discontinued.

STEP 4: PREPARE THE PATIENT AND FOLLOW-UP PROVIDER

Once a clinician has performed medication reconciliation, including obtaining a best-possible medical history and carefully reviewing the medication list and orders for errors and clinical appropriateness, the next steps are to ensure the patient understands what he or she needs to do and to confirm that suitable follow-up plans are in place. These measures should be taken at all transitions of care but are critically important at hospital discharge.

Preparing the patient and caregiver

An accurate, reconciled medication list should be given to the patient, caregiver, or both, and should be reviewed before discharge.¹⁷

Approximately one-third of Americans have low health literacy skills, so medication lists and associated materials should be easy to understand.³⁵ Medication lists should be written in plain language and formatted for optimal readability (Table 4), clearly stating which medications to continue, change, hold temporarily, and stop.

Patients recall and comprehend about half of the information provided during a medical encounter.³⁶ Thus, medication teaching should focus on key points including changes or additions to the regimen, specific instructions for follow-up and monitoring, and how to handle common and serious side effects.

To confirm patient understanding, clinicians should use “teach-back,” ie, provide the patient with information and then ask him or her to repeat back key points.^37,38 The patient and family should also be encouraged to ask questions before discharge.

If not already addressed during the hospital stay, barriers to medication adherence and ability to obtain the medications should be attended to at this time (Table 5). Also, the plan to pick up the medications should be verified with the patient and caregiver. Verify that there is transportation to a particular pharmacy that is open at the time of discharge, and that the patient can afford the medications.

Ensuring appropriate follow-up

Studies have shown that timely in-home or telephone follow-up after discharge can decrease adverse events and postdischarge health care utilization.^39,40 Telephone follow-up that includes thorough medication reconciliation can help detect and resolve medication issues early after discharge and can close gaps related to monitoring and follow-up.

Medication reconciliation by telephone can be time-consuming. Depending on the number of medications that need to be reviewed, calls can take between 10 and 60 minutes. Postdischarge phone calls should be performed by clinical personnel who are able to identify medication-related problems. A pharmacist should be an available resource to assist with complex regimens, to help resolve medication discrepancies, and to address patient concerns. Table 6 provides tips for conducting follow-up phone calls.

Resolving discrepancies identified during follow-up calls can be difficult, as changes to the medication regimen are often not communicated effectively to other members of the care team. Physicians should document the complete medication list and plan in the discharge summary, and there should be a method for the caller to record updates to the medication list in the medical record so that they are apparent at the outpatient follow-up visit.

An additional challenge is that it is frequently unclear which physician “owns” which medications. Therefore, designating a contact person for each medication until follow-up can be very valuable. At a minimum, a “physician owner” for high-alert medications such as insulin, anticoagulants, and diuretics should be identified to provide close follow-up, titration, and monitoring.

There should also be a plan for the patient to obtain refills of essential long-term medications, such as antiplatelet agents following stent placement.

SUMMARY AND RECOMMENDATIONS

Medication-related problems during hospital admission and discharge are common and range from minor discrepancies in the medication list to errors in history-taking, prescribing, and reconciliation that can lead to potential or actual patient harm. Putting systems in place to facilitate medication reconciliation can decrease the occurrence of medication discrepancies and ADEs, thereby improving patient safety during these critical transitions between care settings and providers. Institutional medication reconciliation programs should focus resources on the admission history-taking step, target the highest-risk patients for the most intensive interventions, and involve pharmacy personnel when possible.

On an individual level, clinicians can incorporate additional interventions into their workflows to optimize medication safety for hospitalized patients. Using a structured approach to obtain a complete and accurate medication list at the time of hospital admission will help providers identify medication-related problems and prevent the propagation of errors throughout the hospital stay and at discharge. Focusing additional time and effort on a comprehensive review of the medication list for errors of omission and commission, patient-specific needs, and high-alert drugs will further decrease the risk of medication errors. Finally, providing discharge counseling targeting patient barriers to adherence and ensuring a proper handover of medication information and rationale for medication changes to outpatient providers will improve the chances of a safe transition.

Any time patients enter or leave the hospital, they risk being harmed by errors in their medications.¹ Adverse events from medication errors during transitions of care are common but often preventable. One key approach is to systematically review every patient’s medication list on admission and discharge and resolve any discrepancies. These transitions are also an opportunity to address other medication-related problems, such as adherence, drug interactions, and clinical appropriateness.

This article summarizes the types and prevalence of medication problems that occur during hospital-based transitions of care, and suggests strategies to decrease the risk of medication errors, focusing on medication reconciliation and related interventions that clinicians can use at the bedside to improve medication safety.

DEFINING TERMS

A medication discrepancy is any variance noted in a patient’s documented medication regimen across different medication lists or sites of care. While some differences reflect intentional and clinically appropriate changes to the regimen, others are unintentional and reflect inaccurate or incomplete information. These unintentional discrepancies are medication errors.

Depending on the clinical circumstances and medications involved, such errors may lead to an adverse drug event (ADE), defined as actual harm or injury resulting from a medication. Sometimes a medication error does not cause harm immediately but could if left uncorrected; this is called a potential ADE.

An important goal during transitions of care is to reduce unintentional medication discrepancies, thereby reducing potential and actual ADEs.

ERRORS ARISE AT DISCHARGE—AND EVEN MORE AT ADMISSION

Hospital discharge is a widely recognized transition in which patient harm occurs. As many as 70% of patients may have an unintentional medication discrepancy at hospital discharge, with many of those discrepancies having potential for harm.² Indeed, during the first few weeks after discharge, 50% of patients have a clinically important medication error,³ and 20% experience an adverse event, most commonly an ADE.⁴ ADEs are associated with excess health care utilization,^5–7 and many are preventable through strategies such as medication reconciliation.^5,8

Importantly, more errors arise at hospital admission than at other times.^9,10

Errors in medication histories are the most common source of discrepancies, affecting up to two-thirds of admitted patients.^11,12 More than one-quarter of hospital prescribing errors can be attributed to incomplete medication histories at the time of admission,¹³ and nearly three times as many clinically important medication discrepancies are related to history-taking errors on admission rather than reconciliation errors at discharge.⁹

Most discrepancies occurring at the time of admission have the potential to cause harm, particularly if the errors persist beyond discharge.¹⁴ Therefore, taking a complete and accurate medication history on hospital admission is critical to ensuring safe care transitions.

MEDICATION RECONCILIATION: BARRIERS AND FACILITATORS

Medication reconciliation is a strategy for reducing medication discrepancies in patients moving across care settings. Simply put, it is the process by which a patient’s medication list is obtained, compared, and clarified across different sites of care.¹⁵ It has consistently been shown to decrease medication errors compared with usual care,¹⁶ and it is strongly supported by national and international organizations.^17–21

In clinical practice, many physicians and institutions have found medication reconciliation difficult to implement, owing to barriers at the level of the patient, provider, and system (Table  1). In response to these challenges, two initiatives have synthesized best practices and offer toolkits that hospitals and clinicians can use: the Medications at Transitions and Clinical Handoffs (MATCH) program²² and the Multi-Center Medication Reconciliation Quality Improvement Study (MARQUIS).²³

Lack of resources is a widely acknowledged challenge. Thus, the MARQUIS investigators²³ suggested focusing on the admission history, where most errors occur, and applying the most resource-intensive interventions in patients at highest risk of ADEs, ie, those who are elderly, have multiple comorbid conditions, or take numerous medications.¹⁶

Although the risk of an ADE increases with the number of medications a patient takes,⁴ the exact number of drugs that defines high risk has not been well established. Targeting patients who take 10 or more maintenance medications is a reasonable initial approach,²⁴ but institutions should tailor risk stratification to their patient populations and available resources. Patients taking high-risk medications such as anticoagulants and insulin could also be prioritized for review, since these medications are more likely to cause serious patient harm when used without appropriate clinical oversight.⁷

Using pharmacy staff to perform medication history-taking, reconciliation, and patient counseling has been shown to produce favorable patient outcomes, particularly for higher-risk patients.^16,23

The MARQUIS investigators found they could boost the chances of success by sharing stories of patient harm to foster “buy-in” among frontline staff, providing formal training to clinicians on how to take a medication history, and obtaining the support of nursing leaders to champion improvement efforts.²³

Additionally, patients should be empowered to maintain an accurate medication list. We address strategies for improving patient engagement and adherence in a later section.

BEST PRACTICES FOR IMPROVING MEDICATION SAFETY

Medication reconciliation is one of several measures necessary to optimize medication safety during transitions of care. It typically includes the following actions:

• Interview the patient or caregiver to determine the list of medications the patient is currently taking (or supposed to be taking); consult other sources if needed.
• List medications that are being ordered during the clinical encounter.
• Compare these two lists, making note of medications that are stopped, changed, or newly prescribed.
• Resolve any discrepancies.
• Communicate the reconciled list to the patient, appropriate caregivers, and providers of follow-up care.

At a rudimentary level, medication reconciliation encompasses medication list management along the continuum of care. However, we recommend leveraging medication reconciliation as an opportunity to further enhance medication safety by reviewing the appropriateness of each medication, seizing opportunities to streamline or simplify the regimen, assessing patient and caregiver understanding of medication instructions and potential ADEs, and delivering appropriate counseling to enhance medication use. Table 2 outlines our framework for medication management during hospital-based transitions.

STEP 1: OBTAIN A COMPLETE PREADMISSION MEDICATION LIST

The “best-possible medication history” is obtained in a systematic process of interviewing the patient or caregiver plus reviewing at least one other reliable source.²³ The resulting list should include all medications the patient is taking (prescription and nonprescription), doses, directions for use, formulations if applicable, indications, start and stop dates, and medication allergies and reactions.

Review existing information. Before eliciting a history from the patient, review his or her recorded medical history and existing medication lists (eg, prior discharge summaries, records from other facilities, records from outpatient visits, pharmacy fill data). This will provide context about the regimen and help identify issues and questions that can be addressed during the history-taking.

Ask open-ended questions. Instead of just asking the patient to confirm the accuracy of the existing medication list, we recommend actively obtaining the full medication list from the patient or caregiver. The conversation should begin with an open-ended question such as, “What medications do you take at home?” This approach will also allow the clinician to gauge the patient’s level of understanding of each medication’s indication and dosing instructions. Using a series of prompts such as those recommended in Table 3 will elicit a best-possible medical history, while verifying all of the medications on the existing list.

Clarify discrepancies. Resolve differences between the existing medication lists and the patient’s or caregiver’s report during the preadmission interview. Examples include errors of omission (a medication is missing), errors of commission (an additional medication is present), and discrepancies in the strength, formulation, dosing instructions, and indications for the drugs. If necessary, other sources of information should be consulted, such as the patient’s medication bottles, pharmacy or pharmacies, primary care physician, and a family member or caregiver.

Assess adherence. The extent to which patients take their medications as directed is an important component of the history, but is often left out. Medication nonadherence rates in the United States are 40% to 70%,²⁵ contributing to poor patient outcomes and imposing extraordinary costs on the health care system.²⁶

Asking open-ended, nonjudgmental questions at the time of hospital admission will help to uncover medication-taking behaviors as well as barriers to adherence (Table 3). The patient’s responses should be taken into account when determining the treatment plan.

Document your findings. After completing the medication history and clarifying discrepancies, document the preadmission list in the medical record. All members of the health care team should have access to view and update the same list, as new information about the preadmission medications may be uncovered after the initial history.

Make clinical decisions. Complete the admission medication reconciliation by deciding whether each medication on the list should be continued, changed, held, or discontinued on the basis of the patient’s clinical condition. Well-designed information technology applications enable the provider to document each action and the rationale for it, as well as carry that information into the order-entry system. Medications marked as held or discontinued on admission should be revisited as the patient’s clinical condition changes and at discharge.

STEP 2: AVOID RECONCILIATION ERRORS

Reconciliation errors reflect discrepancies between the medication history and the medications that are ordered after admission.

Reconciliation errors are less common than medication history errors, accounting for approximately one-third of potentially harmful medication errors in hospitalized medical patients.⁹ These include errors of omission (a medication is omitted from the orders), errors of commission (a medication is prescribed with no indication for continuation), and therapeutic duplication.

Preventing errors of omission

Medications are often held at transition points for appropriate clinical reasons. Examples include holding anticoagulants and antiplatelet agents in patients who have gastrointestinal bleeding or an upcoming procedure, antihypertensives in patients with hemodynamic instability, and other chronic medications in patients with an acute illness.

Poor documentation and communication of these decisions can lead to a failure to resume the medications—an error of omission—at hospital discharge.

Hospitalized patients are at risk of unintentional discontinuation of their chronic medications, including antiplatelet drugs, anticoagulants, statins, and thyroid replacement, particularly if admitted to the intensive care unit.¹² These errors can be minimized by a standardized medication reconciliation process at each transition and clear documentation of the medication plan.

Communication among providers can be improved if the admitting clinician documents clearly whether each preadmission medication is being continued, changed, or stopped, along with the reason for doing so, and makes this information available throughout the hospital stay. Upon transfer to another unit and at discharge, the physician should review each; preadmission medication that was held and the patient’s current clinical status and, based on that information, decide whether medications that were held should be resumed. If a medication will be restarted later, specific instructions should be documented and communicated to the patient and the physicians who are taking over his or her care.

Preventing errors of commission

Failure to perform a complete reconciliation at each transition of care and match each medication with an appropriate indication can lead to errors of commission.

One study showed that 44% of patients were prescribed at least one unnecessary drug at hospital discharge, one-fourth of which were started during the hospitalization.²⁷ Commonly prescribed unnecessarily were gastrointestinal agents, central nervous system drugs, nutrients, and supplements.

It is critical to assess each medication’s ongoing need, appropriateness, and risk-benefit ratio at every transition. Medications no longer indicated should be discontinued in order to simplify the regimen, avoid unnecessary drug exposure, and prevent ADEs.

For example, proton pump inhibitors or histamine 2 receptor blockers are often started in the hospital for stress ulcer prophylaxis. One-third of patients are then discharged home on the medication, and 6 months later half of those patients are still taking the unnecessary drug.²⁸ This situation can be avoided by limiting use of these medications to appropriate circumstances, clearly marking the indication as stress ulcer prophylaxis (as opposed to an ongoing condition that will require continuing it after discharge), and discontinuing the agent when appropriate.

All drugs, even common and seemingly benign ones, carry some risk and should be discontinued when no longer needed. Thus, medications added during the hospitalization to control acute symptoms should also be reviewed at each transition to prevent inappropriate continuation when symptoms have resolved.

One study, for example, found that many patients were discharged with inappropriate prescriptions for atypical antipsychotics after receiving them in the intensive care unit, likely for delirium.²⁹ Documenting that an acute issue such as delirium has resolved should prompt the discontinuation of therapy.

Preventing therapeutic duplication

Therapeutic duplication occurs in about 8% of discharges.¹ These errors often result from formulary substitutions or altering the dosage form in the acute setting. For example, patients who receive a prescription for the substituted agent at discharge and also resume their prehospitalization medications end up with duplicate therapy.

Therapeutic substitution is common at the time of admission to the hospital as a result of formulary restrictions. Drug classes that are frequently substituted include statins, antihypertensives, urinary antispasmodics, and proton pump inhibitors. Physicians should be familiar with the preferred agents on the hospital formulary and make careful note when a substitution occurs. Furthermore, hospital systems should be developed to remind the physician to switch back to the outpatient medication at discharge.

Similar problems occur when home medications are replaced with different dosage forms with different pharmacokinetic properties. For example, a long-acting medication may be temporarily replaced with an intravenous solution or immediate-release tablet for several reasons, including nothing-by-mouth status, unstable clinical condition, need for titration, and need to crush the tablet to give the drug per tube. The differing formulations must be reconciled throughout the patient’s hospital course and at discharge to avoid therapeutic duplication and serious medication errors. Deliberate changes to the dosage form should be clearly communicated in the discharge medication list so that patients and other clinicians are aware.

Hospital systems should also have the capability to identify duplications in the medication list and to warn prescribers of these errors. The ability to group medications by drug class or sort the medication list alphabetically by generic name can help uncover duplication errors.

STEP 3: REVIEW THE LIST IN VIEW OF THE CLINICAL PICTURE

Transitions of care should prompt providers to review the medication list for possible drug-disease interactions, confirm compliance with evidence-based guidelines, and evaluate the risks and benefits of each medication in the context of the patient’s age and acute and chronic medical issues. This is also an opportunity to screen the full list for potentially inappropriate medications and high-alert drugs such as insulin or anticoagulants, which are more likely than other drugs to cause severe harm when used in error.

Acute kidney injury. New drug-disease interactions can arise during a hospitalization and can affect dosing and the choice of drug. The onset of acute kidney injury, for example, often necessitates adjusting or discontinuing nephrotoxic and renally excreted medications. ADEs or potential ADEs have been reported in 43% of hospitalized patients with acute kidney injury.³⁰

Because acute kidney injury is often transient, medications may need to be held or adjusted several times until renal function stabilizes. This can be challenging across the continuum of care and requires close monitoring of the serum creatinine level and associated drug doses and levels, if applicable. Well-designed clinical decision support tools can integrate laboratory data and alert the prescriber to a clinically important increase or decrease in serum creatinine that may warrant a change in therapy. Modifications to the regimen and a plan for timely follow-up of the serum creatinine level should be clearly documented in the discharge plan.

Liver disease. Similar attention should be given to drugs that are hepatically metabolized if a patient has acute or chronic liver impairment.

Geriatric patients, particularly those who present with altered mental status, falls, or urinary retention, should have their medication list reviewed for potentially inappropriate medications, which are drugs that pose increased risk of poor outcomes in older adults.^31,32 Patients and providers may have been willing to accept the risk of medications such as anticholinergics or sedative-hypnotics when the drugs were initiated, but circumstances can change over time, especially in this patient population. Hospitalization is a prime opportunity to screen for medications that meet the Beers criteria³¹ for agents to avoid or use with caution in older adults.

As-needed medications. Medications prescribed on an as-needed basis in the hospital should be reviewed for continuation or discontinuation at discharge. How often the medication was given can inform this decision.

For example, if as-needed opioids were used frequently, failure to develop a plan of care for pain can lead to persistent symptoms and, possibly, to readmission.^33,34 Similar scenarios occur with use of as-needed blood pressure medications, laxatives, and correction-dose insulin.

If an as-needed medication was used consistently during hospitalization, the physician should consider whether a regularly scheduled medication is needed. Conversely, if the medication was not used during the inpatient admission, it can likely be discontinued.

STEP 4: PREPARE THE PATIENT AND FOLLOW-UP PROVIDER

Once a clinician has performed medication reconciliation, including obtaining a best-possible medical history and carefully reviewing the medication list and orders for errors and clinical appropriateness, the next steps are to ensure the patient understands what he or she needs to do and to confirm that suitable follow-up plans are in place. These measures should be taken at all transitions of care but are critically important at hospital discharge.

Preparing the patient and caregiver

An accurate, reconciled medication list should be given to the patient, caregiver, or both, and should be reviewed before discharge.¹⁷

Approximately one-third of Americans have low health literacy skills, so medication lists and associated materials should be easy to understand.³⁵ Medication lists should be written in plain language and formatted for optimal readability (Table 4), clearly stating which medications to continue, change, hold temporarily, and stop.

Patients recall and comprehend about half of the information provided during a medical encounter.³⁶ Thus, medication teaching should focus on key points including changes or additions to the regimen, specific instructions for follow-up and monitoring, and how to handle common and serious side effects.

To confirm patient understanding, clinicians should use “teach-back,” ie, provide the patient with information and then ask him or her to repeat back key points.^37,38 The patient and family should also be encouraged to ask questions before discharge.

If not already addressed during the hospital stay, barriers to medication adherence and ability to obtain the medications should be attended to at this time (Table 5). Also, the plan to pick up the medications should be verified with the patient and caregiver. Verify that there is transportation to a particular pharmacy that is open at the time of discharge, and that the patient can afford the medications.

Ensuring appropriate follow-up

Studies have shown that timely in-home or telephone follow-up after discharge can decrease adverse events and postdischarge health care utilization.^39,40 Telephone follow-up that includes thorough medication reconciliation can help detect and resolve medication issues early after discharge and can close gaps related to monitoring and follow-up.

Medication reconciliation by telephone can be time-consuming. Depending on the number of medications that need to be reviewed, calls can take between 10 and 60 minutes. Postdischarge phone calls should be performed by clinical personnel who are able to identify medication-related problems. A pharmacist should be an available resource to assist with complex regimens, to help resolve medication discrepancies, and to address patient concerns. Table 6 provides tips for conducting follow-up phone calls.

Resolving discrepancies identified during follow-up calls can be difficult, as changes to the medication regimen are often not communicated effectively to other members of the care team. Physicians should document the complete medication list and plan in the discharge summary, and there should be a method for the caller to record updates to the medication list in the medical record so that they are apparent at the outpatient follow-up visit.

An additional challenge is that it is frequently unclear which physician “owns” which medications. Therefore, designating a contact person for each medication until follow-up can be very valuable. At a minimum, a “physician owner” for high-alert medications such as insulin, anticoagulants, and diuretics should be identified to provide close follow-up, titration, and monitoring.

There should also be a plan for the patient to obtain refills of essential long-term medications, such as antiplatelet agents following stent placement.

SUMMARY AND RECOMMENDATIONS

Medication-related problems during hospital admission and discharge are common and range from minor discrepancies in the medication list to errors in history-taking, prescribing, and reconciliation that can lead to potential or actual patient harm. Putting systems in place to facilitate medication reconciliation can decrease the occurrence of medication discrepancies and ADEs, thereby improving patient safety during these critical transitions between care settings and providers. Institutional medication reconciliation programs should focus resources on the admission history-taking step, target the highest-risk patients for the most intensive interventions, and involve pharmacy personnel when possible.

On an individual level, clinicians can incorporate additional interventions into their workflows to optimize medication safety for hospitalized patients. Using a structured approach to obtain a complete and accurate medication list at the time of hospital admission will help providers identify medication-related problems and prevent the propagation of errors throughout the hospital stay and at discharge. Focusing additional time and effort on a comprehensive review of the medication list for errors of omission and commission, patient-specific needs, and high-alert drugs will further decrease the risk of medication errors. Finally, providing discharge counseling targeting patient barriers to adherence and ensuring a proper handover of medication information and rationale for medication changes to outpatient providers will improve the chances of a safe transition.

I’m not sure if I recall seeing kids in long lines outside of school waiting to receive the polio vaccine, or if these are just memories from movie film clips. I’ve never seen a patient with an active polio infection, and I’ve seen only a few with postpolio syndromes. I’ve never seen a patient with tetanus, smallpox, diphtheria, or typical measles. I’ve seen three cases of pertussis that I know of, and the long delay in diagnosing the first one (my wife) was clearly because at that time clinicians caring for adults were not attuned to a disease that had virtually disappeared from the American landscape. Once I was sensitized to its presence, it was far easier to make the diagnosis in the second case I encountered (myself). The list of infectious diseases that have almost vanished in the last 75 years with the development of specific vaccines is not long, but it is striking. We can easily lose sight of that when focusing on the less-than-perfect effectiveness of the pneumococcal and annual influenza vaccines.

My message in recounting these observations is that, growing up in the traditional Western medical establishment, I find it hard from a historical perspective to view vaccines as anything but a positive contribution to our public and personal health. And yet a vocal minority, generally outside the medical establishment, maintains that vaccination is a potentially dangerous practice to be avoided whenever possible. Their biological arguments are tenuous and rarely supported by controlled clinical outcomes or observational data. The elimination of trace amounts of mercury-containing preservatives from some vaccines has done little to dampen their concerns. The arguments against routine vaccination and mandated vaccination of schoolchildren to maintain herd immunity have acquired a libertarian tone. While I may share the philosophy behind their perspective—for example, I wear my seat belt while driving, but I don’t think I should be fined if I don’t—my not wearing a seat belt does not increase the chance that those who encounter me on a plane, in a movie theater, or at an amusement park will die when subsequently driving their car.

In all likelihood, I will retire from medicine before I ever see a case of typical diphtheria. I don’t think that is an accident of nature or the effect of better hygiene. I’m hoping that the generation of physicians to follow will see far less cervical cancer, and that physicians in Asia will see far less hepatitis B-associated hepatocellular carcinoma as a result of effective vaccination against the viruses associated with these cancers.

As Drs. Faria Farhat and Glenn Wortmann and Dr. Atul Khasnis discuss in their papers in this issue of the Journal, we have more to learn about how to most effectively use vaccines in special populations. It is clearly not a one-strategy-fits-all world. The decision to vaccinate these patients is usually less about public health than about the health of the individual patient.

The real-world effectiveness of many vaccines is less than it appeared to be in controlled clinical trials. Unfortunately, the patients who most need protection against infections, the immunosuppressed, have a blunted response to many vaccines and perhaps should not receive live vaccines. But we have too little evidence on how and when to optimally vaccinate these patients. It still feels a bit like a casino, not a clinic, when I discuss with a modestly immunosuppressed patient whether he or she should be vaccinated with a live vaccine to reduce the risk of shingles and postherpetic neuralgia.

If we have the opportunity, vaccinating before starting immunosuppressive drugs (or before splenectomy) makes sense. But often that is not an option. We are frequently faced with the need to extrapolate efficacy and safety experiences from clinical trials of vaccines that are conducted with healthier patients and with relatively short follow-up. The two vaccination papers in this issue of the Journal provide us with useful information about immunologic and other issues involved when making the decision to vaccinate special patient populations.

Buckle up wisely.

I’m not sure if I recall seeing kids in long lines outside of school waiting to receive the polio vaccine, or if these are just memories from movie film clips. I’ve never seen a patient with an active polio infection, and I’ve seen only a few with postpolio syndromes. I’ve never seen a patient with tetanus, smallpox, diphtheria, or typical measles. I’ve seen three cases of pertussis that I know of, and the long delay in diagnosing the first one (my wife) was clearly because at that time clinicians caring for adults were not attuned to a disease that had virtually disappeared from the American landscape. Once I was sensitized to its presence, it was far easier to make the diagnosis in the second case I encountered (myself). The list of infectious diseases that have almost vanished in the last 75 years with the development of specific vaccines is not long, but it is striking. We can easily lose sight of that when focusing on the less-than-perfect effectiveness of the pneumococcal and annual influenza vaccines.

My message in recounting these observations is that, growing up in the traditional Western medical establishment, I find it hard from a historical perspective to view vaccines as anything but a positive contribution to our public and personal health. And yet a vocal minority, generally outside the medical establishment, maintains that vaccination is a potentially dangerous practice to be avoided whenever possible. Their biological arguments are tenuous and rarely supported by controlled clinical outcomes or observational data. The elimination of trace amounts of mercury-containing preservatives from some vaccines has done little to dampen their concerns. The arguments against routine vaccination and mandated vaccination of schoolchildren to maintain herd immunity have acquired a libertarian tone. While I may share the philosophy behind their perspective—for example, I wear my seat belt while driving, but I don’t think I should be fined if I don’t—my not wearing a seat belt does not increase the chance that those who encounter me on a plane, in a movie theater, or at an amusement park will die when subsequently driving their car.

In all likelihood, I will retire from medicine before I ever see a case of typical diphtheria. I don’t think that is an accident of nature or the effect of better hygiene. I’m hoping that the generation of physicians to follow will see far less cervical cancer, and that physicians in Asia will see far less hepatitis B-associated hepatocellular carcinoma as a result of effective vaccination against the viruses associated with these cancers.

As Drs. Faria Farhat and Glenn Wortmann and Dr. Atul Khasnis discuss in their papers in this issue of the Journal, we have more to learn about how to most effectively use vaccines in special populations. It is clearly not a one-strategy-fits-all world. The decision to vaccinate these patients is usually less about public health than about the health of the individual patient.

The real-world effectiveness of many vaccines is less than it appeared to be in controlled clinical trials. Unfortunately, the patients who most need protection against infections, the immunosuppressed, have a blunted response to many vaccines and perhaps should not receive live vaccines. But we have too little evidence on how and when to optimally vaccinate these patients. It still feels a bit like a casino, not a clinic, when I discuss with a modestly immunosuppressed patient whether he or she should be vaccinated with a live vaccine to reduce the risk of shingles and postherpetic neuralgia.

If we have the opportunity, vaccinating before starting immunosuppressive drugs (or before splenectomy) makes sense. But often that is not an option. We are frequently faced with the need to extrapolate efficacy and safety experiences from clinical trials of vaccines that are conducted with healthier patients and with relatively short follow-up. The two vaccination papers in this issue of the Journal provide us with useful information about immunologic and other issues involved when making the decision to vaccinate special patient populations.

Buckle up wisely.

I’m not sure if I recall seeing kids in long lines outside of school waiting to receive the polio vaccine, or if these are just memories from movie film clips. I’ve never seen a patient with an active polio infection, and I’ve seen only a few with postpolio syndromes. I’ve never seen a patient with tetanus, smallpox, diphtheria, or typical measles. I’ve seen three cases of pertussis that I know of, and the long delay in diagnosing the first one (my wife) was clearly because at that time clinicians caring for adults were not attuned to a disease that had virtually disappeared from the American landscape. Once I was sensitized to its presence, it was far easier to make the diagnosis in the second case I encountered (myself). The list of infectious diseases that have almost vanished in the last 75 years with the development of specific vaccines is not long, but it is striking. We can easily lose sight of that when focusing on the less-than-perfect effectiveness of the pneumococcal and annual influenza vaccines.

My message in recounting these observations is that, growing up in the traditional Western medical establishment, I find it hard from a historical perspective to view vaccines as anything but a positive contribution to our public and personal health. And yet a vocal minority, generally outside the medical establishment, maintains that vaccination is a potentially dangerous practice to be avoided whenever possible. Their biological arguments are tenuous and rarely supported by controlled clinical outcomes or observational data. The elimination of trace amounts of mercury-containing preservatives from some vaccines has done little to dampen their concerns. The arguments against routine vaccination and mandated vaccination of schoolchildren to maintain herd immunity have acquired a libertarian tone. While I may share the philosophy behind their perspective—for example, I wear my seat belt while driving, but I don’t think I should be fined if I don’t—my not wearing a seat belt does not increase the chance that those who encounter me on a plane, in a movie theater, or at an amusement park will die when subsequently driving their car.

In all likelihood, I will retire from medicine before I ever see a case of typical diphtheria. I don’t think that is an accident of nature or the effect of better hygiene. I’m hoping that the generation of physicians to follow will see far less cervical cancer, and that physicians in Asia will see far less hepatitis B-associated hepatocellular carcinoma as a result of effective vaccination against the viruses associated with these cancers.

As Drs. Faria Farhat and Glenn Wortmann and Dr. Atul Khasnis discuss in their papers in this issue of the Journal, we have more to learn about how to most effectively use vaccines in special populations. It is clearly not a one-strategy-fits-all world. The decision to vaccinate these patients is usually less about public health than about the health of the individual patient.

The real-world effectiveness of many vaccines is less than it appeared to be in controlled clinical trials. Unfortunately, the patients who most need protection against infections, the immunosuppressed, have a blunted response to many vaccines and perhaps should not receive live vaccines. But we have too little evidence on how and when to optimally vaccinate these patients. It still feels a bit like a casino, not a clinic, when I discuss with a modestly immunosuppressed patient whether he or she should be vaccinated with a live vaccine to reduce the risk of shingles and postherpetic neuralgia.

If we have the opportunity, vaccinating before starting immunosuppressive drugs (or before splenectomy) makes sense. But often that is not an option. We are frequently faced with the need to extrapolate efficacy and safety experiences from clinical trials of vaccines that are conducted with healthier patients and with relatively short follow-up. The two vaccination papers in this issue of the Journal provide us with useful information about immunologic and other issues involved when making the decision to vaccinate special patient populations.

Buckle up wisely.

Vaccination is the standard of care as part of health maintenance for healthy people and for patients with myriad medical conditions. In an article in this issue of Cleveland Clinic Journal of Medicine, Drs. Faria Farhat and Glenn Wortmann¹ make recommendations about vaccinations in special populations, including people with a disordered immune system or who are otherwise at heightened risk of infection (eg, because of older age, international travel, comorbidities, and medications).

See related article

But special populations are not all the same in their responses to vaccination. For example, patients with systemic autoimmune diseases are a heterogeneous group with disease-specific immunologic perturbations and immunosuppression that vary by medication and dose, all affecting the response to vaccination. Also, two or more “special” situations may coexist in the same patient.

AREAS OF UNCERTAINTY FOR CLINICAL PRACTICE

Several groups have issued guidelines and recommendations about vaccination in immunocompromised patients, but many areas of uncertainty exist in clinical practice. Most of these arise from a lack of data on immunogenicity and outcomes.

For example, although two pneumococcal vaccines are used in adults, no studies have compared the immunogenicity of the 13-valent pneumococcal conjugate vaccine (PCV13, which is T–cell-dependent) with that of the 23-valent pneumococcal polysaccharide vaccine (PPSV23, which is T–cell-independent) in immunocompromised adults.

Also, whether to use zoster vaccine in immunosuppressed patients ages 50 through 59 is debated. The vaccine is approved by the US Food and Drug Administration (FDA) for this age group, but the Advisory Committee on Immunization Practices (ACIP) does not recommend it, and the Infectious Diseases Society of America (IDSA) suggests that it be “considered” in patients on low-intensity immunosuppressive treatment.²

Testing to ensure optimal response to vaccination has been recommended in several articles and guidelines. However, antibody titers do not necessarily correlate with protection, and at this time no consensus exists about the timing of or need for testing for the response to immunization, the methods to use, how to interpret the results in terms of an adequate or inadequate response, or the role of booster immunization in routine clinical practice.

IS VACCINATION COST-EFFECTIVE?

Also relevant is whether vaccination is cost-effective.

Although vaccination with the PCV13 vaccine is possibly more cost-effective than PPSV23 in US adults based on a model that included immunosuppressed patients, the results of this study were sensitive to several assumptions, and the authors did not extrapolate their conclusion to immunocompromised individuals.³

In another cost-effectiveness analysis, immunocompromised patients were vaccinated with PCV13 at diagnosis and, starting a year later, were followed according to current PPSV23 vaccination guidelines. The PCV13 vaccine’s efficacy against invasive pneumococcal disease and pneumonia based on the modeled program led to cost savings, added quality-adjusted life-years, and prevented invasive pneumococcal disease, mostly in patients with human immunodeficiency virus  (HIV) infection and those on dialysis (unpublished data cited in a 2012 US Centers for Disease Control and Prevention [CDC] report).⁴

No cost-effectiveness studies of influenza vaccination in immunocompromised adults have been conducted in the United States.

The various recommendations by the FDA, the ACIP, and the IDSA regarding the appropriate age at which to give zoster vaccine may also have been influenced by cost-effectiveness studies. These have reported mixed results and have not specifically focused on special populations.⁵

IMPROVING VACCINATION RATES

Rates of vaccination in special populations are suboptimal, and remedial measures to improve coverage have been proposed. One-page vaccine questionnaires or handouts for patients as well as “pop-up” reminders for vaccination in the electronic health record for physicians have resulted in higher rates of indicated vaccinations. Both the CDC and the American College of Physicians (ACP) offer downloadable tools—the CDC Vaccine Schedules App⁶ and the ACP Immunization Advisor,⁷ respectively—that are based on the 2012 ACIP guidelines and are extremely useful for busy practitioners. The CDC also offers patients a vaccine questionnaire⁸ that allows them to determine which vaccinations they may need and also to learn about those vaccines.

MOVING AHEAD

The development of vaccines is ongoing and will be driven by identification of new molecular targets. Advances in therapies for immunocompromised patients such as those with HIV infection will, we hope, decrease the risk of opportunistic infections. The list of vaccine-preventable diseases may continue to grow, as may the list of special populations. Optimal vaccination and outcomes may emerge from expected improved vaccine coverage as a result of the increased health insurance coverage resulting from the much-maligned Patient Protection and Affordable Care Act and ongoing studies regarding efficacy, safety, and cost-effectiveness, especially pertaining to specific patient populations.

Vaccination is the standard of care as part of health maintenance for healthy people and for patients with myriad medical conditions. In an article in this issue of Cleveland Clinic Journal of Medicine, Drs. Faria Farhat and Glenn Wortmann¹ make recommendations about vaccinations in special populations, including people with a disordered immune system or who are otherwise at heightened risk of infection (eg, because of older age, international travel, comorbidities, and medications).

See related article

But special populations are not all the same in their responses to vaccination. For example, patients with systemic autoimmune diseases are a heterogeneous group with disease-specific immunologic perturbations and immunosuppression that vary by medication and dose, all affecting the response to vaccination. Also, two or more “special” situations may coexist in the same patient.

AREAS OF UNCERTAINTY FOR CLINICAL PRACTICE

Several groups have issued guidelines and recommendations about vaccination in immunocompromised patients, but many areas of uncertainty exist in clinical practice. Most of these arise from a lack of data on immunogenicity and outcomes.

For example, although two pneumococcal vaccines are used in adults, no studies have compared the immunogenicity of the 13-valent pneumococcal conjugate vaccine (PCV13, which is T–cell-dependent) with that of the 23-valent pneumococcal polysaccharide vaccine (PPSV23, which is T–cell-independent) in immunocompromised adults.

Also, whether to use zoster vaccine in immunosuppressed patients ages 50 through 59 is debated. The vaccine is approved by the US Food and Drug Administration (FDA) for this age group, but the Advisory Committee on Immunization Practices (ACIP) does not recommend it, and the Infectious Diseases Society of America (IDSA) suggests that it be “considered” in patients on low-intensity immunosuppressive treatment.²

Testing to ensure optimal response to vaccination has been recommended in several articles and guidelines. However, antibody titers do not necessarily correlate with protection, and at this time no consensus exists about the timing of or need for testing for the response to immunization, the methods to use, how to interpret the results in terms of an adequate or inadequate response, or the role of booster immunization in routine clinical practice.

IS VACCINATION COST-EFFECTIVE?

Also relevant is whether vaccination is cost-effective.

Although vaccination with the PCV13 vaccine is possibly more cost-effective than PPSV23 in US adults based on a model that included immunosuppressed patients, the results of this study were sensitive to several assumptions, and the authors did not extrapolate their conclusion to immunocompromised individuals.³

In another cost-effectiveness analysis, immunocompromised patients were vaccinated with PCV13 at diagnosis and, starting a year later, were followed according to current PPSV23 vaccination guidelines. The PCV13 vaccine’s efficacy against invasive pneumococcal disease and pneumonia based on the modeled program led to cost savings, added quality-adjusted life-years, and prevented invasive pneumococcal disease, mostly in patients with human immunodeficiency virus  (HIV) infection and those on dialysis (unpublished data cited in a 2012 US Centers for Disease Control and Prevention [CDC] report).⁴

No cost-effectiveness studies of influenza vaccination in immunocompromised adults have been conducted in the United States.

The various recommendations by the FDA, the ACIP, and the IDSA regarding the appropriate age at which to give zoster vaccine may also have been influenced by cost-effectiveness studies. These have reported mixed results and have not specifically focused on special populations.⁵

IMPROVING VACCINATION RATES

Rates of vaccination in special populations are suboptimal, and remedial measures to improve coverage have been proposed. One-page vaccine questionnaires or handouts for patients as well as “pop-up” reminders for vaccination in the electronic health record for physicians have resulted in higher rates of indicated vaccinations. Both the CDC and the American College of Physicians (ACP) offer downloadable tools—the CDC Vaccine Schedules App⁶ and the ACP Immunization Advisor,⁷ respectively—that are based on the 2012 ACIP guidelines and are extremely useful for busy practitioners. The CDC also offers patients a vaccine questionnaire⁸ that allows them to determine which vaccinations they may need and also to learn about those vaccines.

MOVING AHEAD

The development of vaccines is ongoing and will be driven by identification of new molecular targets. Advances in therapies for immunocompromised patients such as those with HIV infection will, we hope, decrease the risk of opportunistic infections. The list of vaccine-preventable diseases may continue to grow, as may the list of special populations. Optimal vaccination and outcomes may emerge from expected improved vaccine coverage as a result of the increased health insurance coverage resulting from the much-maligned Patient Protection and Affordable Care Act and ongoing studies regarding efficacy, safety, and cost-effectiveness, especially pertaining to specific patient populations.

Vaccination is the standard of care as part of health maintenance for healthy people and for patients with myriad medical conditions. In an article in this issue of Cleveland Clinic Journal of Medicine, Drs. Faria Farhat and Glenn Wortmann¹ make recommendations about vaccinations in special populations, including people with a disordered immune system or who are otherwise at heightened risk of infection (eg, because of older age, international travel, comorbidities, and medications).

See related article

But special populations are not all the same in their responses to vaccination. For example, patients with systemic autoimmune diseases are a heterogeneous group with disease-specific immunologic perturbations and immunosuppression that vary by medication and dose, all affecting the response to vaccination. Also, two or more “special” situations may coexist in the same patient.

AREAS OF UNCERTAINTY FOR CLINICAL PRACTICE

Several groups have issued guidelines and recommendations about vaccination in immunocompromised patients, but many areas of uncertainty exist in clinical practice. Most of these arise from a lack of data on immunogenicity and outcomes.

For example, although two pneumococcal vaccines are used in adults, no studies have compared the immunogenicity of the 13-valent pneumococcal conjugate vaccine (PCV13, which is T–cell-dependent) with that of the 23-valent pneumococcal polysaccharide vaccine (PPSV23, which is T–cell-independent) in immunocompromised adults.

Also, whether to use zoster vaccine in immunosuppressed patients ages 50 through 59 is debated. The vaccine is approved by the US Food and Drug Administration (FDA) for this age group, but the Advisory Committee on Immunization Practices (ACIP) does not recommend it, and the Infectious Diseases Society of America (IDSA) suggests that it be “considered” in patients on low-intensity immunosuppressive treatment.²

Testing to ensure optimal response to vaccination has been recommended in several articles and guidelines. However, antibody titers do not necessarily correlate with protection, and at this time no consensus exists about the timing of or need for testing for the response to immunization, the methods to use, how to interpret the results in terms of an adequate or inadequate response, or the role of booster immunization in routine clinical practice.

IS VACCINATION COST-EFFECTIVE?

Also relevant is whether vaccination is cost-effective.

Although vaccination with the PCV13 vaccine is possibly more cost-effective than PPSV23 in US adults based on a model that included immunosuppressed patients, the results of this study were sensitive to several assumptions, and the authors did not extrapolate their conclusion to immunocompromised individuals.³

In another cost-effectiveness analysis, immunocompromised patients were vaccinated with PCV13 at diagnosis and, starting a year later, were followed according to current PPSV23 vaccination guidelines. The PCV13 vaccine’s efficacy against invasive pneumococcal disease and pneumonia based on the modeled program led to cost savings, added quality-adjusted life-years, and prevented invasive pneumococcal disease, mostly in patients with human immunodeficiency virus  (HIV) infection and those on dialysis (unpublished data cited in a 2012 US Centers for Disease Control and Prevention [CDC] report).⁴

No cost-effectiveness studies of influenza vaccination in immunocompromised adults have been conducted in the United States.

The various recommendations by the FDA, the ACIP, and the IDSA regarding the appropriate age at which to give zoster vaccine may also have been influenced by cost-effectiveness studies. These have reported mixed results and have not specifically focused on special populations.⁵

IMPROVING VACCINATION RATES

Rates of vaccination in special populations are suboptimal, and remedial measures to improve coverage have been proposed. One-page vaccine questionnaires or handouts for patients as well as “pop-up” reminders for vaccination in the electronic health record for physicians have resulted in higher rates of indicated vaccinations. Both the CDC and the American College of Physicians (ACP) offer downloadable tools—the CDC Vaccine Schedules App⁶ and the ACP Immunization Advisor,⁷ respectively—that are based on the 2012 ACIP guidelines and are extremely useful for busy practitioners. The CDC also offers patients a vaccine questionnaire⁸ that allows them to determine which vaccinations they may need and also to learn about those vaccines.

MOVING AHEAD

The development of vaccines is ongoing and will be driven by identification of new molecular targets. Advances in therapies for immunocompromised patients such as those with HIV infection will, we hope, decrease the risk of opportunistic infections. The list of vaccine-preventable diseases may continue to grow, as may the list of special populations. Optimal vaccination and outcomes may emerge from expected improved vaccine coverage as a result of the increased health insurance coverage resulting from the much-maligned Patient Protection and Affordable Care Act and ongoing studies regarding efficacy, safety, and cost-effectiveness, especially pertaining to specific patient populations.

Most vaccinations are given during childhood, but some require boosting during adulthood or are indicated for specific patient populations such as international travelers or those with certain medical conditions. Although generally safe, some vaccines contain live, attenuated organisms that can cause disease in immunocompromised patients. Thus, knowledge of the indications for and contraindications to specific vaccinations is critical to protect adults in special circumstances who are at risk.

See related commentary

Vaccines have helped eliminate or significantly reduce the burden of more than a dozen illnesses.^1–3 The Advisory Committee on Immunization Practices (ACIP) of the US Centers for Disease Control and Prevention  (CDC) makes recommendations about vaccinations for normal adults and children as well as for certain groups at high risk of vaccine-preventable infections.⁴ Tables 1 and 2 summarize the recommendations for vaccination by medical condition.⁴  In addition, several applications are available online, including downloadable apps from the (www.cdc.gov/vaccines/schedules/Schedulers/adult-scheduler.html) and the American College of Physicians (http://immunization.acponline.org/app/).

HUMANITY’S GREATEST ADVANCES IN PREVENTING INFECTIOUS DISEASE

Immunization and improved sanitation are humanity’s greatest advances in preventing sickness and death from infectious diseases. Since Jenner’s discovery in 1796 that milkmaids who had contracted cowpox (vaccinia) were immune to smallpox, vaccination has eliminated smallpox, markedly decreased the incidence of many infectious diseases, and, most recently, shown efficacy in preventing cervical cancer (with the human papillomavirus vaccine) and hepatocellular cancer (with the hepatitis B vaccine).^1–3

Unfortunately, vaccination rates remain low for most routine vaccinations indicated for adults. For example, about 60% of adults over age 65 receive pneumococcal vaccination, and fewer than 10% of black patients over age 60 receive zoster vaccination.⁵ Various factors may account for these low rates, including financial disincentives.⁶

Nevertheless, vaccination remains one of medicine’s most effective defenses against infectious diseases and is especially important in the special populations discussed below. By being steadfast proponents of vaccination, especially for the most vulnerable patients, physicians can help ensure the optimum protection for their patients.

VACCINATING PREGNANT PATIENTS

When considering vaccination during pregnancy, one must consider the risk and benefit of the vaccine and the risk of the disease in both the mother and the child.

In general, if a pregnant woman is at high risk of exposure to a particular infection, the benefits of vaccinating her against it outweigh the risks. Vaccinating the mother can also protect against certain infections in early infancy through transfer of vaccine-induced immunoglobin G (IgG) across the placenta.⁷ In general, inactivated vaccines are considered safe in pregnancy, while live-attenuated vaccines are contraindicated.⁴ Special considerations for pregnant women include:

Tetanus, diphtheria, and acellular pertussis (Tdap). One dose of Tdap vaccine should be given during each pregnancy, preferably at 27 to 36 weeks of gestation, regardless of when the patient received a previous dose.⁸

Inactivated influenza vaccine should be given as early as possible during the influenza season (October to March) to all pregnant women, regardless of trimester.

Inactivated polio vaccine may be considered for pregnant women with known exposure to polio or travel to endemic areas.

Hepatitis A, hepatitis B, pneumococcal polysaccharide, meningococcal conjugate, and meningococcal polysaccharide vaccines can be given to women at risk of these infections. If a pregnant patient requires pneumococcal polysaccharide vaccine, it should be given during the second or third trimester, as the safety of this vaccine during the first trimester has not been established.⁹

Smallpox, measles-mumps-rubella, and varicella-containing vaccines are contraindicated in pregnancy. Household contacts of a pregnant woman should not receive smallpox vaccine, as it is the only vaccine known to cause harm to the fetus.¹⁰

Human papillomavirus vaccination is not recommended during pregnancy.

Yellow fever live-attenuated vaccine. The safety of this vaccine during pregnancy has not been established, and it is in the US Food and Drug Administration (FDA) pregnancy category C. However, this vaccine is required for entry into certain countries, and it may be offered if the patient is truly at risk of contracting yellow fever. Because pregnancy may affect immunologic response, serologic testing is recommended to document an immune response. If the patient’s itinerary puts her at low risk of yellow fever, then writing her a vaccine waiver letter can be considered.¹¹

VACCINATING IMMUNOCOMPROMISED PATIENTS (NON-HIV)

People who do not have human immunodeficiency virus (HIV) but have a condition such as functional asplenia (sickle cell disease), anatomic asplenia, or complement component deficiency are at higher risk of infection with the encapsulated bacteria Streptococcus pneumoniae, Neisseria meningitidis, and Haemophilus influenzae type b.

Corticosteroids, chemotherapy, radiation for hematologic or solid-organ malignancies, and immune modulators can alter the immune system and pose a risk with the use of live-attenuated vaccines. A corticosteroid dosage equivalent to 2 mg/kg of body weight per day or higher or 20 mg/day of prednisone or higher is generally considered immunosuppressive.

Candidates for organ transplantation should receive vaccinations as early as possible during the disease course leading to transplantation. Vaccinations should be given as soon as the decision is made that the patient is a candidate for transplantation, which could be years or months before the patient actually receives the transplant. In addition to reviewing previously administered vaccinations, pretransplant serologic testing for hepatitis B, varicella, measles, mumps, and rubella antibodies helps to evaluate the need for vaccination.¹²

Recipients of hematopoietic stem cell transplantation are at risk of infections with encapsulated bacteria and certain other vaccine-preventable infections. Antibody titers are significantly reduced after stem cell transplantation because of ablation of bone marrow, and thus certain vaccines should be readministered 3 to 6 months after transplantation (eg, influenza, pneumococcal, and H influenzae vaccines). If the recipient is presumed to be immunocompetent, then varicella or measles-mumps-rubella vaccine can be given 24 months after transplantation.¹³

Apart from adhering to the routine vaccination schedule and avoiding live-attenuated vaccines, specific recommendations apply to persons with immunocompromising conditions¹⁴:

Quadrivalent meningococcal conjugate vaccine should be given to adults of all ages with asplenia or complement component deficiency. The schedule includes two doses at least 2 months apart initially and then revaccination every 5 years.

H influenzae type b vaccine should be given to people with asplenia and recipients of hematopoietic stem cells. One dose is recommended for those with asplenia (functional, anatomic, or elective splenectomy) or sickle cell disease if they have not already received it. A three-dose schedule is considered for hematopoietic stem cell transplant recipients 6 to 12 months after successful transplantation.

Pneumococcal conjugate (PCV13) and pneumococcal polysaccharide (PPSV23) vaccinations are recommended for people who have immunocompromising conditions. PCV13, the newer pneumococcal vaccine, was approved by the FDA in 2010 for use in children and was recommended by the ACIP in 2012 for adults age 19 and older with immunocompromising conditions.

People who have not previously received either of these vaccines and are age 19 or older with immunocompromising conditions including asplenia, chronic renal failure, nephrotic syndrome, cerebrospinal fluid leakage, or cochlear implant should receive a single dose of PCV13 followed by a dose of PPSV23 at least 8 weeks later. One-time revaccination 5 years after the first dose of PPSV23 is recommended for patients with immunocompromising conditions.

For those who have previously been vaccinated with PPSV23, a dose of PCV13 can be given 1 or more years after the last dose of PPSV23. These dosing intervals are important, as lower opsonophagocytic antibody responses have been noted if repeat doses of either pneumococcal vaccine are given sooner than the recommended interval.¹⁵

Inactivated influenza vaccine is recommended annually, except for patients who are unlikely to respond or those who have received anti-B-cell antibodies within 6 months. Live-attenuated influenza vaccine should not be given to immunocompromised patients.

VACCINATING PATIENTS WHO HAVE HIV

People with HIV should be routinely screened for immunity against certain infections and should be offered vaccination if not immune. The response to vaccines may vary depending on the CD4 count, with a good response in patients whose infection is well controlled with antiretroviral agents and with a preserved CD4 count.¹⁶ Special considerations for HIV patients include the following:

Hepatitis A vaccine may be offered to all HIV patients who have no evidence of immunity against hepatitis A, with negative antihepatitis A total and IgG antibodies.

Human papillomavirus vaccine is recommended for men and women with HIV through age 26.

Varicella and measles-mumps-rubella are live-attenuated vaccines and may be considered in patients who are nonimmune and with CD4 counts of 200 cells/µL or higher. However, the ACIP does not make a recommendation regarding the zoster vaccine in HIV patients with CD4 cell counts of 200 cells/µL or higher. In general, live-attenuated vaccines should be avoided in patients with CD4 counts less than 200 or with severe immunocompromised status because of risk of acquiring severe, life-threatening infections.

Pneumococcal vaccine should be given to HIV patients if they have not received it before. The schedule is one dose of PCV13, followed by a dose of PPSV23 at least 8 weeks later. If a patient has been previously vaccinated with PPSV23, then PCV13 is recommended at least 1 year after PPSV23.

Inactivated influenza vaccine is recommended annually. Live-attenuated influenza vaccine should not be given.

Hepatitis B vaccine should be given to nonimmune patients without past or present hepatitis B infection. These patients require higher doses of hepatitis B vaccine (40 μg/mL) than immunocompetent patients, who receive 20 μg/mL. The options include Recombivax HB 40 μg/mL given on a three-dose schedule at 0, 1, and 6 months, and Engerix B, two 20-μg/mL injections given simultaneously on a four-dose schedule at 0, 1, 2, and 6 months.

Meningococcal vaccine. HIV infection is not an indication for meningococcal vaccine unless the patient has other risk factors, such as anatomic or functional asplenia, persistent complement component deficiency, occupational exposure, and travel to endemic areas.

VACCINATING PATIENTS WHO ARE OLDER THAN 60

The immune system deteriorates with age, as does immunity gained from previous vaccinations. Vaccination in this age group reduces the risk of illness and death.¹⁷

Zoster vaccine should be offered to people age 60 and older regardless of previous episodes of herpes zoster unless there is a contraindication such as severe immunodeficiency. The zoster vaccine can reduce the incidence of postherpetic neuralgia by 66.5% and herpes zoster by 51% in patients over age 60.¹⁸

Pneumococcal conjugate vaccine. PCV13 should be offered to all adults age 65 or older. If a person age 65 or older has not received any pneumococcal vaccine before then, PCV13 should be given first, followed by a dose of PPSV23 at least 6 to 12 months after PCV13.

Pneumococcal polysaccharide vaccine. If PPSV23 was given before age 65 for another indication, a dose of PCV 13 should be given at age 65 or later, as long as 6 to 12 months have passed since the previous dose of PPSV 23. The patient should receive the last dose of PPSV23 vaccine 5 years after the first dose of PPSV23.⁴

Influenza vaccine. People 65 or older are at higher risk of complications from influenza, and vaccine should be offered annually. High-dose inactivated influenza vaccine can be used in this age group.⁴

Tdap. If never given before, Tdap is recommended regardless of the interval since the most recent Td vaccination, followed by a Td booster every 10 years.

VACCINATING PATIENTS WHO HAVE CHRONIC KIDNEY DISEASE

Patients with chronic kidney disease are at risk of certain infections, so vaccination is an important preventive measure.¹⁹ Immunizations should be offered to all patients with chronic kidney disease regardless of the disease stage, but they are recommended during the early stages of progressive renal disease to increase the likelihood of vaccine-induced immunity.²⁰

Pneumococcal conjugate vaccine. PCV13 is recommended for adults 19 or older with chronic renal disease or nephrotic syndrome. One dose of PCV13 should be given, followed by a dose of PPSV23 at least 8 weeks later. If the patient has been previously vaccinated with PPSV23, then PCV13 at least 1 year after PPSV23 is recommended.

Hepatitis B vaccine should be given to nonimmune patients without past or present hepatitis B infection. Adult patients on hemodialysis require higher doses of hepatitis B vaccine. The options include Recombivax HB 40 μg/mL given on a three-dose schedule at 0, 1, and 6 months, and Engerix B, two 20-μg/mL injections given simultaneously on a four-dose schedule at 0, 1, 2, and 6 months.

Influenza vaccine should be offered annually to patients with chronic kidney disease.

VACCINATING IMMUNOCOMPROMISED INTERNATIONAL TRAVELERS

International travel for business or pleasure is increasingly common, and immunocompromised patients require specific attention as they may face unanticipated pathogens or have special requirements. Transplant recipients should ideally receive routine and travel-related vaccines as early as possible before transplantation. Vaccination is generally avoided in the first 6 months after organ transplantation to avoid confusion with early graft dysfunction or rejection.²¹ However, it should be considered as soon as a patient develops an illness that might lead to transplantation.

Evaluation of patients for vaccination should include an assessment of the travel-specific epidemiologic risk, the nature of the vaccine (live-attenuated or other), and the immune status. As discussed above, live-attenuated vaccines should be avoided in immunocompromised patients, and thus the injectable typhoid vaccine should be given in lieu of the attenuated oral vaccine.

Yellow fever vaccine is required before entrance to certain countries but should not be given to immunocompromised patients, although it can probably be given to asymptomatic HIV-infected adults with a CD4 count higher than 200 cells/μL who are exposed to substantial risk.²² For patients who cannot receive the vaccine, some governments will accept a physician’s letter stating the patient has a contraindication to vaccination.

VACCINATING HOUSEHOLD MEMBERS OF IMMUNOCOMPROMISED PATIENTS

Protecting immunocompromised patients from infectious diseases involves vaccinating not only the patient but also household members so that they do not acquire infections and then bring them into the household. Immunocompetent members of a household can receive inactivated vaccines based on the recommended ACIP schedule.

Annual inactivated influenza vaccination is recommended, although the live-attenuated influenza virus vaccine can be substituted if the immunocompromised patient is not within 2 months of hematopoietic stem cell transplantation, does not have graft-vs-host disease, and does not have severe combined immune deficiency.

Other live-attenuated vaccines can usually be given if indicated, including measles-mumps-rubella vaccine, rotavirus vaccine in infants, varicella vaccine, and zoster vaccine.¹⁴

Most vaccinations are given during childhood, but some require boosting during adulthood or are indicated for specific patient populations such as international travelers or those with certain medical conditions. Although generally safe, some vaccines contain live, attenuated organisms that can cause disease in immunocompromised patients. Thus, knowledge of the indications for and contraindications to specific vaccinations is critical to protect adults in special circumstances who are at risk.

See related commentary

Vaccines have helped eliminate or significantly reduce the burden of more than a dozen illnesses.^1–3 The Advisory Committee on Immunization Practices (ACIP) of the US Centers for Disease Control and Prevention  (CDC) makes recommendations about vaccinations for normal adults and children as well as for certain groups at high risk of vaccine-preventable infections.⁴ Tables 1 and 2 summarize the recommendations for vaccination by medical condition.⁴  In addition, several applications are available online, including downloadable apps from the (www.cdc.gov/vaccines/schedules/Schedulers/adult-scheduler.html) and the American College of Physicians (http://immunization.acponline.org/app/).

HUMANITY’S GREATEST ADVANCES IN PREVENTING INFECTIOUS DISEASE

Immunization and improved sanitation are humanity’s greatest advances in preventing sickness and death from infectious diseases. Since Jenner’s discovery in 1796 that milkmaids who had contracted cowpox (vaccinia) were immune to smallpox, vaccination has eliminated smallpox, markedly decreased the incidence of many infectious diseases, and, most recently, shown efficacy in preventing cervical cancer (with the human papillomavirus vaccine) and hepatocellular cancer (with the hepatitis B vaccine).^1–3

Unfortunately, vaccination rates remain low for most routine vaccinations indicated for adults. For example, about 60% of adults over age 65 receive pneumococcal vaccination, and fewer than 10% of black patients over age 60 receive zoster vaccination.⁵ Various factors may account for these low rates, including financial disincentives.⁶

Nevertheless, vaccination remains one of medicine’s most effective defenses against infectious diseases and is especially important in the special populations discussed below. By being steadfast proponents of vaccination, especially for the most vulnerable patients, physicians can help ensure the optimum protection for their patients.

VACCINATING PREGNANT PATIENTS

When considering vaccination during pregnancy, one must consider the risk and benefit of the vaccine and the risk of the disease in both the mother and the child.

In general, if a pregnant woman is at high risk of exposure to a particular infection, the benefits of vaccinating her against it outweigh the risks. Vaccinating the mother can also protect against certain infections in early infancy through transfer of vaccine-induced immunoglobin G (IgG) across the placenta.⁷ In general, inactivated vaccines are considered safe in pregnancy, while live-attenuated vaccines are contraindicated.⁴ Special considerations for pregnant women include:

Tetanus, diphtheria, and acellular pertussis (Tdap). One dose of Tdap vaccine should be given during each pregnancy, preferably at 27 to 36 weeks of gestation, regardless of when the patient received a previous dose.⁸

Inactivated influenza vaccine should be given as early as possible during the influenza season (October to March) to all pregnant women, regardless of trimester.

Inactivated polio vaccine may be considered for pregnant women with known exposure to polio or travel to endemic areas.

Hepatitis A, hepatitis B, pneumococcal polysaccharide, meningococcal conjugate, and meningococcal polysaccharide vaccines can be given to women at risk of these infections. If a pregnant patient requires pneumococcal polysaccharide vaccine, it should be given during the second or third trimester, as the safety of this vaccine during the first trimester has not been established.⁹

Smallpox, measles-mumps-rubella, and varicella-containing vaccines are contraindicated in pregnancy. Household contacts of a pregnant woman should not receive smallpox vaccine, as it is the only vaccine known to cause harm to the fetus.¹⁰

Human papillomavirus vaccination is not recommended during pregnancy.

Yellow fever live-attenuated vaccine. The safety of this vaccine during pregnancy has not been established, and it is in the US Food and Drug Administration (FDA) pregnancy category C. However, this vaccine is required for entry into certain countries, and it may be offered if the patient is truly at risk of contracting yellow fever. Because pregnancy may affect immunologic response, serologic testing is recommended to document an immune response. If the patient’s itinerary puts her at low risk of yellow fever, then writing her a vaccine waiver letter can be considered.¹¹

VACCINATING IMMUNOCOMPROMISED PATIENTS (NON-HIV)

People who do not have human immunodeficiency virus (HIV) but have a condition such as functional asplenia (sickle cell disease), anatomic asplenia, or complement component deficiency are at higher risk of infection with the encapsulated bacteria Streptococcus pneumoniae, Neisseria meningitidis, and Haemophilus influenzae type b.

Corticosteroids, chemotherapy, radiation for hematologic or solid-organ malignancies, and immune modulators can alter the immune system and pose a risk with the use of live-attenuated vaccines. A corticosteroid dosage equivalent to 2 mg/kg of body weight per day or higher or 20 mg/day of prednisone or higher is generally considered immunosuppressive.

Candidates for organ transplantation should receive vaccinations as early as possible during the disease course leading to transplantation. Vaccinations should be given as soon as the decision is made that the patient is a candidate for transplantation, which could be years or months before the patient actually receives the transplant. In addition to reviewing previously administered vaccinations, pretransplant serologic testing for hepatitis B, varicella, measles, mumps, and rubella antibodies helps to evaluate the need for vaccination.¹²

Recipients of hematopoietic stem cell transplantation are at risk of infections with encapsulated bacteria and certain other vaccine-preventable infections. Antibody titers are significantly reduced after stem cell transplantation because of ablation of bone marrow, and thus certain vaccines should be readministered 3 to 6 months after transplantation (eg, influenza, pneumococcal, and H influenzae vaccines). If the recipient is presumed to be immunocompetent, then varicella or measles-mumps-rubella vaccine can be given 24 months after transplantation.¹³

Apart from adhering to the routine vaccination schedule and avoiding live-attenuated vaccines, specific recommendations apply to persons with immunocompromising conditions¹⁴:

Quadrivalent meningococcal conjugate vaccine should be given to adults of all ages with asplenia or complement component deficiency. The schedule includes two doses at least 2 months apart initially and then revaccination every 5 years.

H influenzae type b vaccine should be given to people with asplenia and recipients of hematopoietic stem cells. One dose is recommended for those with asplenia (functional, anatomic, or elective splenectomy) or sickle cell disease if they have not already received it. A three-dose schedule is considered for hematopoietic stem cell transplant recipients 6 to 12 months after successful transplantation.

Pneumococcal conjugate (PCV13) and pneumococcal polysaccharide (PPSV23) vaccinations are recommended for people who have immunocompromising conditions. PCV13, the newer pneumococcal vaccine, was approved by the FDA in 2010 for use in children and was recommended by the ACIP in 2012 for adults age 19 and older with immunocompromising conditions.

People who have not previously received either of these vaccines and are age 19 or older with immunocompromising conditions including asplenia, chronic renal failure, nephrotic syndrome, cerebrospinal fluid leakage, or cochlear implant should receive a single dose of PCV13 followed by a dose of PPSV23 at least 8 weeks later. One-time revaccination 5 years after the first dose of PPSV23 is recommended for patients with immunocompromising conditions.

For those who have previously been vaccinated with PPSV23, a dose of PCV13 can be given 1 or more years after the last dose of PPSV23. These dosing intervals are important, as lower opsonophagocytic antibody responses have been noted if repeat doses of either pneumococcal vaccine are given sooner than the recommended interval.¹⁵

Inactivated influenza vaccine is recommended annually, except for patients who are unlikely to respond or those who have received anti-B-cell antibodies within 6 months. Live-attenuated influenza vaccine should not be given to immunocompromised patients.

VACCINATING PATIENTS WHO HAVE HIV

People with HIV should be routinely screened for immunity against certain infections and should be offered vaccination if not immune. The response to vaccines may vary depending on the CD4 count, with a good response in patients whose infection is well controlled with antiretroviral agents and with a preserved CD4 count.¹⁶ Special considerations for HIV patients include the following:

Hepatitis A vaccine may be offered to all HIV patients who have no evidence of immunity against hepatitis A, with negative antihepatitis A total and IgG antibodies.

Human papillomavirus vaccine is recommended for men and women with HIV through age 26.

Varicella and measles-mumps-rubella are live-attenuated vaccines and may be considered in patients who are nonimmune and with CD4 counts of 200 cells/µL or higher. However, the ACIP does not make a recommendation regarding the zoster vaccine in HIV patients with CD4 cell counts of 200 cells/µL or higher. In general, live-attenuated vaccines should be avoided in patients with CD4 counts less than 200 or with severe immunocompromised status because of risk of acquiring severe, life-threatening infections.

Pneumococcal vaccine should be given to HIV patients if they have not received it before. The schedule is one dose of PCV13, followed by a dose of PPSV23 at least 8 weeks later. If a patient has been previously vaccinated with PPSV23, then PCV13 is recommended at least 1 year after PPSV23.

Inactivated influenza vaccine is recommended annually. Live-attenuated influenza vaccine should not be given.

Hepatitis B vaccine should be given to nonimmune patients without past or present hepatitis B infection. These patients require higher doses of hepatitis B vaccine (40 μg/mL) than immunocompetent patients, who receive 20 μg/mL. The options include Recombivax HB 40 μg/mL given on a three-dose schedule at 0, 1, and 6 months, and Engerix B, two 20-μg/mL injections given simultaneously on a four-dose schedule at 0, 1, 2, and 6 months.

Meningococcal vaccine. HIV infection is not an indication for meningococcal vaccine unless the patient has other risk factors, such as anatomic or functional asplenia, persistent complement component deficiency, occupational exposure, and travel to endemic areas.

VACCINATING PATIENTS WHO ARE OLDER THAN 60

The immune system deteriorates with age, as does immunity gained from previous vaccinations. Vaccination in this age group reduces the risk of illness and death.¹⁷

Zoster vaccine should be offered to people age 60 and older regardless of previous episodes of herpes zoster unless there is a contraindication such as severe immunodeficiency. The zoster vaccine can reduce the incidence of postherpetic neuralgia by 66.5% and herpes zoster by 51% in patients over age 60.¹⁸

Pneumococcal conjugate vaccine. PCV13 should be offered to all adults age 65 or older. If a person age 65 or older has not received any pneumococcal vaccine before then, PCV13 should be given first, followed by a dose of PPSV23 at least 6 to 12 months after PCV13.

Pneumococcal polysaccharide vaccine. If PPSV23 was given before age 65 for another indication, a dose of PCV 13 should be given at age 65 or later, as long as 6 to 12 months have passed since the previous dose of PPSV 23. The patient should receive the last dose of PPSV23 vaccine 5 years after the first dose of PPSV23.⁴

Influenza vaccine. People 65 or older are at higher risk of complications from influenza, and vaccine should be offered annually. High-dose inactivated influenza vaccine can be used in this age group.⁴

Tdap. If never given before, Tdap is recommended regardless of the interval since the most recent Td vaccination, followed by a Td booster every 10 years.

VACCINATING PATIENTS WHO HAVE CHRONIC KIDNEY DISEASE

Patients with chronic kidney disease are at risk of certain infections, so vaccination is an important preventive measure.¹⁹ Immunizations should be offered to all patients with chronic kidney disease regardless of the disease stage, but they are recommended during the early stages of progressive renal disease to increase the likelihood of vaccine-induced immunity.²⁰

Pneumococcal conjugate vaccine. PCV13 is recommended for adults 19 or older with chronic renal disease or nephrotic syndrome. One dose of PCV13 should be given, followed by a dose of PPSV23 at least 8 weeks later. If the patient has been previously vaccinated with PPSV23, then PCV13 at least 1 year after PPSV23 is recommended.

Hepatitis B vaccine should be given to nonimmune patients without past or present hepatitis B infection. Adult patients on hemodialysis require higher doses of hepatitis B vaccine. The options include Recombivax HB 40 μg/mL given on a three-dose schedule at 0, 1, and 6 months, and Engerix B, two 20-μg/mL injections given simultaneously on a four-dose schedule at 0, 1, 2, and 6 months.

Influenza vaccine should be offered annually to patients with chronic kidney disease.

VACCINATING IMMUNOCOMPROMISED INTERNATIONAL TRAVELERS

International travel for business or pleasure is increasingly common, and immunocompromised patients require specific attention as they may face unanticipated pathogens or have special requirements. Transplant recipients should ideally receive routine and travel-related vaccines as early as possible before transplantation. Vaccination is generally avoided in the first 6 months after organ transplantation to avoid confusion with early graft dysfunction or rejection.²¹ However, it should be considered as soon as a patient develops an illness that might lead to transplantation.

Evaluation of patients for vaccination should include an assessment of the travel-specific epidemiologic risk, the nature of the vaccine (live-attenuated or other), and the immune status. As discussed above, live-attenuated vaccines should be avoided in immunocompromised patients, and thus the injectable typhoid vaccine should be given in lieu of the attenuated oral vaccine.

Yellow fever vaccine is required before entrance to certain countries but should not be given to immunocompromised patients, although it can probably be given to asymptomatic HIV-infected adults with a CD4 count higher than 200 cells/μL who are exposed to substantial risk.²² For patients who cannot receive the vaccine, some governments will accept a physician’s letter stating the patient has a contraindication to vaccination.

VACCINATING HOUSEHOLD MEMBERS OF IMMUNOCOMPROMISED PATIENTS

Protecting immunocompromised patients from infectious diseases involves vaccinating not only the patient but also household members so that they do not acquire infections and then bring them into the household. Immunocompetent members of a household can receive inactivated vaccines based on the recommended ACIP schedule.

Annual inactivated influenza vaccination is recommended, although the live-attenuated influenza virus vaccine can be substituted if the immunocompromised patient is not within 2 months of hematopoietic stem cell transplantation, does not have graft-vs-host disease, and does not have severe combined immune deficiency.

Other live-attenuated vaccines can usually be given if indicated, including measles-mumps-rubella vaccine, rotavirus vaccine in infants, varicella vaccine, and zoster vaccine.¹⁴

Most vaccinations are given during childhood, but some require boosting during adulthood or are indicated for specific patient populations such as international travelers or those with certain medical conditions. Although generally safe, some vaccines contain live, attenuated organisms that can cause disease in immunocompromised patients. Thus, knowledge of the indications for and contraindications to specific vaccinations is critical to protect adults in special circumstances who are at risk.

See related commentary

Vaccines have helped eliminate or significantly reduce the burden of more than a dozen illnesses.^1–3 The Advisory Committee on Immunization Practices (ACIP) of the US Centers for Disease Control and Prevention  (CDC) makes recommendations about vaccinations for normal adults and children as well as for certain groups at high risk of vaccine-preventable infections.⁴ Tables 1 and 2 summarize the recommendations for vaccination by medical condition.⁴  In addition, several applications are available online, including downloadable apps from the (www.cdc.gov/vaccines/schedules/Schedulers/adult-scheduler.html) and the American College of Physicians (http://immunization.acponline.org/app/).

HUMANITY’S GREATEST ADVANCES IN PREVENTING INFECTIOUS DISEASE

Immunization and improved sanitation are humanity’s greatest advances in preventing sickness and death from infectious diseases. Since Jenner’s discovery in 1796 that milkmaids who had contracted cowpox (vaccinia) were immune to smallpox, vaccination has eliminated smallpox, markedly decreased the incidence of many infectious diseases, and, most recently, shown efficacy in preventing cervical cancer (with the human papillomavirus vaccine) and hepatocellular cancer (with the hepatitis B vaccine).^1–3

Unfortunately, vaccination rates remain low for most routine vaccinations indicated for adults. For example, about 60% of adults over age 65 receive pneumococcal vaccination, and fewer than 10% of black patients over age 60 receive zoster vaccination.⁵ Various factors may account for these low rates, including financial disincentives.⁶

Nevertheless, vaccination remains one of medicine’s most effective defenses against infectious diseases and is especially important in the special populations discussed below. By being steadfast proponents of vaccination, especially for the most vulnerable patients, physicians can help ensure the optimum protection for their patients.

VACCINATING PREGNANT PATIENTS

When considering vaccination during pregnancy, one must consider the risk and benefit of the vaccine and the risk of the disease in both the mother and the child.

In general, if a pregnant woman is at high risk of exposure to a particular infection, the benefits of vaccinating her against it outweigh the risks. Vaccinating the mother can also protect against certain infections in early infancy through transfer of vaccine-induced immunoglobin G (IgG) across the placenta.⁷ In general, inactivated vaccines are considered safe in pregnancy, while live-attenuated vaccines are contraindicated.⁴ Special considerations for pregnant women include:

Tetanus, diphtheria, and acellular pertussis (Tdap). One dose of Tdap vaccine should be given during each pregnancy, preferably at 27 to 36 weeks of gestation, regardless of when the patient received a previous dose.⁸

Inactivated influenza vaccine should be given as early as possible during the influenza season (October to March) to all pregnant women, regardless of trimester.

Inactivated polio vaccine may be considered for pregnant women with known exposure to polio or travel to endemic areas.

Hepatitis A, hepatitis B, pneumococcal polysaccharide, meningococcal conjugate, and meningococcal polysaccharide vaccines can be given to women at risk of these infections. If a pregnant patient requires pneumococcal polysaccharide vaccine, it should be given during the second or third trimester, as the safety of this vaccine during the first trimester has not been established.⁹

Smallpox, measles-mumps-rubella, and varicella-containing vaccines are contraindicated in pregnancy. Household contacts of a pregnant woman should not receive smallpox vaccine, as it is the only vaccine known to cause harm to the fetus.¹⁰

Human papillomavirus vaccination is not recommended during pregnancy.

Yellow fever live-attenuated vaccine. The safety of this vaccine during pregnancy has not been established, and it is in the US Food and Drug Administration (FDA) pregnancy category C. However, this vaccine is required for entry into certain countries, and it may be offered if the patient is truly at risk of contracting yellow fever. Because pregnancy may affect immunologic response, serologic testing is recommended to document an immune response. If the patient’s itinerary puts her at low risk of yellow fever, then writing her a vaccine waiver letter can be considered.¹¹

VACCINATING IMMUNOCOMPROMISED PATIENTS (NON-HIV)

People who do not have human immunodeficiency virus (HIV) but have a condition such as functional asplenia (sickle cell disease), anatomic asplenia, or complement component deficiency are at higher risk of infection with the encapsulated bacteria Streptococcus pneumoniae, Neisseria meningitidis, and Haemophilus influenzae type b.

Corticosteroids, chemotherapy, radiation for hematologic or solid-organ malignancies, and immune modulators can alter the immune system and pose a risk with the use of live-attenuated vaccines. A corticosteroid dosage equivalent to 2 mg/kg of body weight per day or higher or 20 mg/day of prednisone or higher is generally considered immunosuppressive.

Candidates for organ transplantation should receive vaccinations as early as possible during the disease course leading to transplantation. Vaccinations should be given as soon as the decision is made that the patient is a candidate for transplantation, which could be years or months before the patient actually receives the transplant. In addition to reviewing previously administered vaccinations, pretransplant serologic testing for hepatitis B, varicella, measles, mumps, and rubella antibodies helps to evaluate the need for vaccination.¹²

Recipients of hematopoietic stem cell transplantation are at risk of infections with encapsulated bacteria and certain other vaccine-preventable infections. Antibody titers are significantly reduced after stem cell transplantation because of ablation of bone marrow, and thus certain vaccines should be readministered 3 to 6 months after transplantation (eg, influenza, pneumococcal, and H influenzae vaccines). If the recipient is presumed to be immunocompetent, then varicella or measles-mumps-rubella vaccine can be given 24 months after transplantation.¹³

Apart from adhering to the routine vaccination schedule and avoiding live-attenuated vaccines, specific recommendations apply to persons with immunocompromising conditions¹⁴:

Quadrivalent meningococcal conjugate vaccine should be given to adults of all ages with asplenia or complement component deficiency. The schedule includes two doses at least 2 months apart initially and then revaccination every 5 years.

H influenzae type b vaccine should be given to people with asplenia and recipients of hematopoietic stem cells. One dose is recommended for those with asplenia (functional, anatomic, or elective splenectomy) or sickle cell disease if they have not already received it. A three-dose schedule is considered for hematopoietic stem cell transplant recipients 6 to 12 months after successful transplantation.

Pneumococcal conjugate (PCV13) and pneumococcal polysaccharide (PPSV23) vaccinations are recommended for people who have immunocompromising conditions. PCV13, the newer pneumococcal vaccine, was approved by the FDA in 2010 for use in children and was recommended by the ACIP in 2012 for adults age 19 and older with immunocompromising conditions.

People who have not previously received either of these vaccines and are age 19 or older with immunocompromising conditions including asplenia, chronic renal failure, nephrotic syndrome, cerebrospinal fluid leakage, or cochlear implant should receive a single dose of PCV13 followed by a dose of PPSV23 at least 8 weeks later. One-time revaccination 5 years after the first dose of PPSV23 is recommended for patients with immunocompromising conditions.

For those who have previously been vaccinated with PPSV23, a dose of PCV13 can be given 1 or more years after the last dose of PPSV23. These dosing intervals are important, as lower opsonophagocytic antibody responses have been noted if repeat doses of either pneumococcal vaccine are given sooner than the recommended interval.¹⁵

Inactivated influenza vaccine is recommended annually, except for patients who are unlikely to respond or those who have received anti-B-cell antibodies within 6 months. Live-attenuated influenza vaccine should not be given to immunocompromised patients.

VACCINATING PATIENTS WHO HAVE HIV

People with HIV should be routinely screened for immunity against certain infections and should be offered vaccination if not immune. The response to vaccines may vary depending on the CD4 count, with a good response in patients whose infection is well controlled with antiretroviral agents and with a preserved CD4 count.¹⁶ Special considerations for HIV patients include the following:

Hepatitis A vaccine may be offered to all HIV patients who have no evidence of immunity against hepatitis A, with negative antihepatitis A total and IgG antibodies.

Human papillomavirus vaccine is recommended for men and women with HIV through age 26.

Varicella and measles-mumps-rubella are live-attenuated vaccines and may be considered in patients who are nonimmune and with CD4 counts of 200 cells/µL or higher. However, the ACIP does not make a recommendation regarding the zoster vaccine in HIV patients with CD4 cell counts of 200 cells/µL or higher. In general, live-attenuated vaccines should be avoided in patients with CD4 counts less than 200 or with severe immunocompromised status because of risk of acquiring severe, life-threatening infections.

Pneumococcal vaccine should be given to HIV patients if they have not received it before. The schedule is one dose of PCV13, followed by a dose of PPSV23 at least 8 weeks later. If a patient has been previously vaccinated with PPSV23, then PCV13 is recommended at least 1 year after PPSV23.

Inactivated influenza vaccine is recommended annually. Live-attenuated influenza vaccine should not be given.

Hepatitis B vaccine should be given to nonimmune patients without past or present hepatitis B infection. These patients require higher doses of hepatitis B vaccine (40 μg/mL) than immunocompetent patients, who receive 20 μg/mL. The options include Recombivax HB 40 μg/mL given on a three-dose schedule at 0, 1, and 6 months, and Engerix B, two 20-μg/mL injections given simultaneously on a four-dose schedule at 0, 1, 2, and 6 months.

Meningococcal vaccine. HIV infection is not an indication for meningococcal vaccine unless the patient has other risk factors, such as anatomic or functional asplenia, persistent complement component deficiency, occupational exposure, and travel to endemic areas.

VACCINATING PATIENTS WHO ARE OLDER THAN 60

The immune system deteriorates with age, as does immunity gained from previous vaccinations. Vaccination in this age group reduces the risk of illness and death.¹⁷

Zoster vaccine should be offered to people age 60 and older regardless of previous episodes of herpes zoster unless there is a contraindication such as severe immunodeficiency. The zoster vaccine can reduce the incidence of postherpetic neuralgia by 66.5% and herpes zoster by 51% in patients over age 60.¹⁸

Pneumococcal conjugate vaccine. PCV13 should be offered to all adults age 65 or older. If a person age 65 or older has not received any pneumococcal vaccine before then, PCV13 should be given first, followed by a dose of PPSV23 at least 6 to 12 months after PCV13.

Pneumococcal polysaccharide vaccine. If PPSV23 was given before age 65 for another indication, a dose of PCV 13 should be given at age 65 or later, as long as 6 to 12 months have passed since the previous dose of PPSV 23. The patient should receive the last dose of PPSV23 vaccine 5 years after the first dose of PPSV23.⁴

Influenza vaccine. People 65 or older are at higher risk of complications from influenza, and vaccine should be offered annually. High-dose inactivated influenza vaccine can be used in this age group.⁴

Tdap. If never given before, Tdap is recommended regardless of the interval since the most recent Td vaccination, followed by a Td booster every 10 years.

VACCINATING PATIENTS WHO HAVE CHRONIC KIDNEY DISEASE

Patients with chronic kidney disease are at risk of certain infections, so vaccination is an important preventive measure.¹⁹ Immunizations should be offered to all patients with chronic kidney disease regardless of the disease stage, but they are recommended during the early stages of progressive renal disease to increase the likelihood of vaccine-induced immunity.²⁰

Pneumococcal conjugate vaccine. PCV13 is recommended for adults 19 or older with chronic renal disease or nephrotic syndrome. One dose of PCV13 should be given, followed by a dose of PPSV23 at least 8 weeks later. If the patient has been previously vaccinated with PPSV23, then PCV13 at least 1 year after PPSV23 is recommended.

Hepatitis B vaccine should be given to nonimmune patients without past or present hepatitis B infection. Adult patients on hemodialysis require higher doses of hepatitis B vaccine. The options include Recombivax HB 40 μg/mL given on a three-dose schedule at 0, 1, and 6 months, and Engerix B, two 20-μg/mL injections given simultaneously on a four-dose schedule at 0, 1, 2, and 6 months.

Influenza vaccine should be offered annually to patients with chronic kidney disease.

VACCINATING IMMUNOCOMPROMISED INTERNATIONAL TRAVELERS

International travel for business or pleasure is increasingly common, and immunocompromised patients require specific attention as they may face unanticipated pathogens or have special requirements. Transplant recipients should ideally receive routine and travel-related vaccines as early as possible before transplantation. Vaccination is generally avoided in the first 6 months after organ transplantation to avoid confusion with early graft dysfunction or rejection.²¹ However, it should be considered as soon as a patient develops an illness that might lead to transplantation.

Evaluation of patients for vaccination should include an assessment of the travel-specific epidemiologic risk, the nature of the vaccine (live-attenuated or other), and the immune status. As discussed above, live-attenuated vaccines should be avoided in immunocompromised patients, and thus the injectable typhoid vaccine should be given in lieu of the attenuated oral vaccine.

Yellow fever vaccine is required before entrance to certain countries but should not be given to immunocompromised patients, although it can probably be given to asymptomatic HIV-infected adults with a CD4 count higher than 200 cells/μL who are exposed to substantial risk.²² For patients who cannot receive the vaccine, some governments will accept a physician’s letter stating the patient has a contraindication to vaccination.

VACCINATING HOUSEHOLD MEMBERS OF IMMUNOCOMPROMISED PATIENTS

Protecting immunocompromised patients from infectious diseases involves vaccinating not only the patient but also household members so that they do not acquire infections and then bring them into the household. Immunocompetent members of a household can receive inactivated vaccines based on the recommended ACIP schedule.

Annual inactivated influenza vaccination is recommended, although the live-attenuated influenza virus vaccine can be substituted if the immunocompromised patient is not within 2 months of hematopoietic stem cell transplantation, does not have graft-vs-host disease, and does not have severe combined immune deficiency.

Other live-attenuated vaccines can usually be given if indicated, including measles-mumps-rubella vaccine, rotavirus vaccine in infants, varicella vaccine, and zoster vaccine.¹⁴

In the April 2015 issue, on page 214 in the article by Jin XW, McKenzie ML, Yen-Lieberman B, “Can the test for human papillomavirus DNA be used as a stand-alone, first-line screening test for cervical cancer?”, the source for the information on predictive values was not cited. The final bulleted item should have read as follows:

• HPV testing by itself performed better than Pap-HPV cotesting, with positive predictive values of 12.25% vs 11.04% and negative predictive values of 99.58% vs 99.52% (data presented to the FDA Medical Devices Advisory Committee, Microbiology Panel. March 12, 2014. FDA Executive Summary).

This oversight has been corrected in the online version of the article.

In the April 2015 issue, on page 214 in the article by Jin XW, McKenzie ML, Yen-Lieberman B, “Can the test for human papillomavirus DNA be used as a stand-alone, first-line screening test for cervical cancer?”, the source for the information on predictive values was not cited. The final bulleted item should have read as follows:

• HPV testing by itself performed better than Pap-HPV cotesting, with positive predictive values of 12.25% vs 11.04% and negative predictive values of 99.58% vs 99.52% (data presented to the FDA Medical Devices Advisory Committee, Microbiology Panel. March 12, 2014. FDA Executive Summary).

This oversight has been corrected in the online version of the article.

In the April 2015 issue, on page 214 in the article by Jin XW, McKenzie ML, Yen-Lieberman B, “Can the test for human papillomavirus DNA be used as a stand-alone, first-line screening test for cervical cancer?”, the source for the information on predictive values was not cited. The final bulleted item should have read as follows:

• HPV testing by itself performed better than Pap-HPV cotesting, with positive predictive values of 12.25% vs 11.04% and negative predictive values of 99.58% vs 99.52% (data presented to the FDA Medical Devices Advisory Committee, Microbiology Panel. March 12, 2014. FDA Executive Summary).

This oversight has been corrected in the online version of the article.

The Dartmouth Institute for Health Policy and Clinical Practice and other institutions have long tried to quantify the prevalence and geographic variation of low-value care. Researchers have roughly defined low-value care as tests and procedures for which the potential benefit does not outweigh the potential harm, though the calculus can change significantly from patient to patient.

William Schpero, a former health policy fellow at the Dartmouth Institute and now a PhD student in health policy and management at Yale University, says he and colleagues initially identified three theoretical reasons for the geographical variation. An increase in the use of low-value care, they reasoned, might be driven by patients demanding more intensive treatments, by financial incentives to providers, or by providers supplying more services.

After adjusting for differences in the health status of patient populations, however, the Dartmouth Institute’s work consistently revealed large unexplained geographical variations, a finding that also held true for patients in the last six months of life. These variations, Schpero says, suggest that patient demand is not a major driver. A collaborative Dartmouth and Harvard study, released by the National Bureau of Economic Research in 2013, instead pointed to a more likely rationale, at least during the last two years of a patient’s life.

By linking patient and physician surveys to Medicare claims data, the report examined how physician and patient preferences affected overall healthcare spending in different geographic regions. The paper identified “cowboys,” or physicians who used treatments that were more intensive than those recommended by guidelines, and “comforters,” or physicians who recommended more low-cost treatment protocols.

Older physicians and smaller practices, the study suggested, were more likely to recommend higher levels of follow-up care and fall into the “cowboys” category.

It was this difference in physician culture and beliefs about effective treatment, not patient preferences, that drove most of the variation in healthcare spending. Monetary incentives, meanwhile, had only a marginal effect. If all physicians were to follow professional guidelines for effective care and not exceed recommended treatments, the report suggested, Medicare spending for end-of-life care could be reduced by 36 percent, “which is a huge, huge number,” Schpero says.

The Dartmouth Institute for Health Policy and Clinical Practice and other institutions have long tried to quantify the prevalence and geographic variation of low-value care. Researchers have roughly defined low-value care as tests and procedures for which the potential benefit does not outweigh the potential harm, though the calculus can change significantly from patient to patient.

William Schpero, a former health policy fellow at the Dartmouth Institute and now a PhD student in health policy and management at Yale University, says he and colleagues initially identified three theoretical reasons for the geographical variation. An increase in the use of low-value care, they reasoned, might be driven by patients demanding more intensive treatments, by financial incentives to providers, or by providers supplying more services.

After adjusting for differences in the health status of patient populations, however, the Dartmouth Institute’s work consistently revealed large unexplained geographical variations, a finding that also held true for patients in the last six months of life. These variations, Schpero says, suggest that patient demand is not a major driver. A collaborative Dartmouth and Harvard study, released by the National Bureau of Economic Research in 2013, instead pointed to a more likely rationale, at least during the last two years of a patient’s life.

By linking patient and physician surveys to Medicare claims data, the report examined how physician and patient preferences affected overall healthcare spending in different geographic regions. The paper identified “cowboys,” or physicians who used treatments that were more intensive than those recommended by guidelines, and “comforters,” or physicians who recommended more low-cost treatment protocols.

Older physicians and smaller practices, the study suggested, were more likely to recommend higher levels of follow-up care and fall into the “cowboys” category.

It was this difference in physician culture and beliefs about effective treatment, not patient preferences, that drove most of the variation in healthcare spending. Monetary incentives, meanwhile, had only a marginal effect. If all physicians were to follow professional guidelines for effective care and not exceed recommended treatments, the report suggested, Medicare spending for end-of-life care could be reduced by 36 percent, “which is a huge, huge number,” Schpero says.

The Dartmouth Institute for Health Policy and Clinical Practice and other institutions have long tried to quantify the prevalence and geographic variation of low-value care. Researchers have roughly defined low-value care as tests and procedures for which the potential benefit does not outweigh the potential harm, though the calculus can change significantly from patient to patient.

William Schpero, a former health policy fellow at the Dartmouth Institute and now a PhD student in health policy and management at Yale University, says he and colleagues initially identified three theoretical reasons for the geographical variation. An increase in the use of low-value care, they reasoned, might be driven by patients demanding more intensive treatments, by financial incentives to providers, or by providers supplying more services.

After adjusting for differences in the health status of patient populations, however, the Dartmouth Institute’s work consistently revealed large unexplained geographical variations, a finding that also held true for patients in the last six months of life. These variations, Schpero says, suggest that patient demand is not a major driver. A collaborative Dartmouth and Harvard study, released by the National Bureau of Economic Research in 2013, instead pointed to a more likely rationale, at least during the last two years of a patient’s life.

By linking patient and physician surveys to Medicare claims data, the report examined how physician and patient preferences affected overall healthcare spending in different geographic regions. The paper identified “cowboys,” or physicians who used treatments that were more intensive than those recommended by guidelines, and “comforters,” or physicians who recommended more low-cost treatment protocols.

Older physicians and smaller practices, the study suggested, were more likely to recommend higher levels of follow-up care and fall into the “cowboys” category.

It was this difference in physician culture and beliefs about effective treatment, not patient preferences, that drove most of the variation in healthcare spending. Monetary incentives, meanwhile, had only a marginal effect. If all physicians were to follow professional guidelines for effective care and not exceed recommended treatments, the report suggested, Medicare spending for end-of-life care could be reduced by 36 percent, “which is a huge, huge number,” Schpero says.

Sometimes, simply knowing what a test costs can make all the difference.

Many physicians have sheepishly admitted that they know little about the price tags attached to the procedures and tests they order on a routine basis—or how that might impact their patients financially. In Medscape’s Physician Compensation Report for 2012, only 38% of surveyed doctors said they regularly discussed the cost of treatment with their patients. The following year, the rate had dropped to 30%.

One medical resident, Neel Shah, MD, MPP discovered how important those discussions can be.

After a woman admitted to the ED tested positive on a pregnancy test, a follow-up hormone test warned of potential trouble with her pregnancy, and Dr. Shah asked her to return to the hospital for an ultrasound.

“She refused to come in until I could tell her how much the ultrasound would cost,” he recalls. Other providers had told him that bringing up costs with patients would decrease their trust, because they didn’t want doctors to focus on anything but providing care. “With her, it was very clear that my inability to tell her what things cost actually eroded her trust in me and, in her mind, she was being reasonable,” he says.

“If you’re trying to tell doctors or clinicians in general that we ought to be doing things differently, as a young person, it’s not a great way to make friends.”

–Neel Shah, MD, MPP

Dr. Shah had already grown disillusioned in medical school, watching providers around him make clinical decisions without regard to the cost for patients, and he took a hiatus to study politics at Harvard’s Kennedy School of Government. When he and a collaborator subsequently launched the nonprofit organization Costs of Care to point out the downsides of that lack of transparency, however, they received a less-than-enthusiastic reception from some quarters.

“If you’re trying to tell doctors or clinicians in general that we ought to be doing things differently, as a young person, it’s not a great way to make friends,” says

Dr. Shah, who is now an assistant professor at Harvard Medical School in Boston.

The group gained traction as its cost-awareness manifesto began to resonate with the public, however, and the essays on its site have been picked up by multiple media groups. Dr. Shah’s own experience with his pregnant patient, however, made one of the strongest impressions on him.

Getting an answer to her about the cost of an ultrasound took nearly 24 hours, he recalls, “because nobody around me knew.” In the interim, he fretted that his patient might have an ectopic pregnancy and bleed to death. She didn’t, but the outcome could have been very different, he says.

“That really struck home for me, for sure,” he adds. “I think about that all the time.”

Sometimes, simply knowing what a test costs can make all the difference.

Many physicians have sheepishly admitted that they know little about the price tags attached to the procedures and tests they order on a routine basis—or how that might impact their patients financially. In Medscape’s Physician Compensation Report for 2012, only 38% of surveyed doctors said they regularly discussed the cost of treatment with their patients. The following year, the rate had dropped to 30%.

One medical resident, Neel Shah, MD, MPP discovered how important those discussions can be.

After a woman admitted to the ED tested positive on a pregnancy test, a follow-up hormone test warned of potential trouble with her pregnancy, and Dr. Shah asked her to return to the hospital for an ultrasound.

“She refused to come in until I could tell her how much the ultrasound would cost,” he recalls. Other providers had told him that bringing up costs with patients would decrease their trust, because they didn’t want doctors to focus on anything but providing care. “With her, it was very clear that my inability to tell her what things cost actually eroded her trust in me and, in her mind, she was being reasonable,” he says.

“If you’re trying to tell doctors or clinicians in general that we ought to be doing things differently, as a young person, it’s not a great way to make friends.”

–Neel Shah, MD, MPP

Dr. Shah had already grown disillusioned in medical school, watching providers around him make clinical decisions without regard to the cost for patients, and he took a hiatus to study politics at Harvard’s Kennedy School of Government. When he and a collaborator subsequently launched the nonprofit organization Costs of Care to point out the downsides of that lack of transparency, however, they received a less-than-enthusiastic reception from some quarters.

“If you’re trying to tell doctors or clinicians in general that we ought to be doing things differently, as a young person, it’s not a great way to make friends,” says

Dr. Shah, who is now an assistant professor at Harvard Medical School in Boston.

The group gained traction as its cost-awareness manifesto began to resonate with the public, however, and the essays on its site have been picked up by multiple media groups. Dr. Shah’s own experience with his pregnant patient, however, made one of the strongest impressions on him.

Getting an answer to her about the cost of an ultrasound took nearly 24 hours, he recalls, “because nobody around me knew.” In the interim, he fretted that his patient might have an ectopic pregnancy and bleed to death. She didn’t, but the outcome could have been very different, he says.

“That really struck home for me, for sure,” he adds. “I think about that all the time.”

Sometimes, simply knowing what a test costs can make all the difference.

Many physicians have sheepishly admitted that they know little about the price tags attached to the procedures and tests they order on a routine basis—or how that might impact their patients financially. In Medscape’s Physician Compensation Report for 2012, only 38% of surveyed doctors said they regularly discussed the cost of treatment with their patients. The following year, the rate had dropped to 30%.

One medical resident, Neel Shah, MD, MPP discovered how important those discussions can be.

After a woman admitted to the ED tested positive on a pregnancy test, a follow-up hormone test warned of potential trouble with her pregnancy, and Dr. Shah asked her to return to the hospital for an ultrasound.

“She refused to come in until I could tell her how much the ultrasound would cost,” he recalls. Other providers had told him that bringing up costs with patients would decrease their trust, because they didn’t want doctors to focus on anything but providing care. “With her, it was very clear that my inability to tell her what things cost actually eroded her trust in me and, in her mind, she was being reasonable,” he says.

“If you’re trying to tell doctors or clinicians in general that we ought to be doing things differently, as a young person, it’s not a great way to make friends.”

–Neel Shah, MD, MPP

Dr. Shah had already grown disillusioned in medical school, watching providers around him make clinical decisions without regard to the cost for patients, and he took a hiatus to study politics at Harvard’s Kennedy School of Government. When he and a collaborator subsequently launched the nonprofit organization Costs of Care to point out the downsides of that lack of transparency, however, they received a less-than-enthusiastic reception from some quarters.

“If you’re trying to tell doctors or clinicians in general that we ought to be doing things differently, as a young person, it’s not a great way to make friends,” says

Dr. Shah, who is now an assistant professor at Harvard Medical School in Boston.

The group gained traction as its cost-awareness manifesto began to resonate with the public, however, and the essays on its site have been picked up by multiple media groups. Dr. Shah’s own experience with his pregnant patient, however, made one of the strongest impressions on him.

Getting an answer to her about the cost of an ultrasound took nearly 24 hours, he recalls, “because nobody around me knew.” In the interim, he fretted that his patient might have an ectopic pregnancy and bleed to death. She didn’t, but the outcome could have been very different, he says.

“That really struck home for me, for sure,” he adds. “I think about that all the time.”