Cleveland Clinic Journal of Medicine

Acute myocardial infarction (MI) portends important and substantial consequences. Angioplasty or fibrinolytic therapy to open the blocked coronary artery is proven to improve the patient’s chances of surviving without consequent morbidity or death. But the diagnosis is not always straightforward. The presentation of acute MI can vary widely, and a number of other conditions—many of them equally serious emergencies—can mimic its symptoms, electrocardiographic signs, and biomarker patterns.

In an attempt to improve the accuracy of the diagnosis of MI, a multinational task force met in 1999 under the auspices of the European Society of Cardiology and the American College of Cardiology. The goal was to develop a simple, clinically oriented definition of MI that could be widely adopted. A document was created and published simultaneously in 2000 in the European Heart Journal and the Journal of the American College of Cardiology.¹ These organizations updated their paper in 2007 with a new definition of acute MI to account for advances in diagnosis and management.²

In this article we will review the new definition and how to make the diagnosis of acute MI today. Specifically, the updated definition includes:

• Subtypes of acute MI
• Imaging tests supporting the diagnosis
• Biomarker thresholds after percutaneous coronary intervention or bypass grafting.

TROPONIN: BETTER THAN CK, BUT NOT PERFECT

The original 2000 paper¹ and the 2007 update² featured the use of the cardiac biomarker troponin, which is considerably more sensitive and specific for heart damage than total creatine kinase (CK) or its isoform, CK-MB.

The new, more-sensitive biomarker-based definition of MI resulted in more cases of MI being diagnosed, and this has attracted the attention and scrutiny of many, especially population scientists and interventional cardiologists.³ This change has caused some controversy, especially when dealing with small rises in troponin following percutaneous coronary intervention.

In addition, some confusion over terminology remains. For example, the phrase “troponin leak” is often used to describe cases in which serum troponin levels rise but there is no MI. However, most experts believe that a rise and fall in troponin is due to true myocardial cell death. Troponin I and T are such large molecules that they cannot “leak” from a cardiac cell unless there has been irreparable cellular damage—that is, cell death.

Creatine kinase still has a role

In some cases, CK and CK-MB may be helpful in determining the acuity of myocardial necrosis, but their use will vary by institution. These biomarkers typically rise 2 to 4 hours after the initial event and fall within 24 to 48 hours, whereas troponin levels stay elevated for days or weeks. Thus, the presence of troponin without CK and CK-MB in the right clinical context may indicate a past MI that is no longer acute.

INFARCTION: CELL DEATH DUE TO ISCHEMIA

MI is myocardial cell death due to prolonged ischemia. Under the microscope, it can be categorized as coagulation necrosis in which ghost-like cell structures remain after hypoxic insult (typical of most MIs) or contraction band necrosis with amorphous cells that cannot contract anymore, the latter often a hallmark of excessive catecholamine damage or reperfusion injury. Apoptosis occurs in the heart but is technically not considered necrosis and is thought not to be associated with elevated troponin levels.^6,7

In experiments in animals, cell death can occur as little as 20 minutes after coronary artery occlusion, although completion of infarction is thought to take 2 to 4 hours. The time to infarct completion may be longer in patients with collateral circulation or when the culprit coronary artery has intermittent (“stuttering”) occlusion. Preconditioning of myocardial cells with intermittent ischemia can also influence the timing of myocardial necrosis by protecting against cell death to some extent. Alteration in myocardial demand can influence the time required for completion of infarction either favorably or unfavorably; hence, reducing myocardial demand is beneficial in acute MI.

Three pathologic phases of MI

MI can be categorized pathologically as acute, healing, or healed.

Acute MI. In the first 6 hours after coronary artery occlusion, coagulation necrosis can be seen with no cellular infiltration. After 6 hours, polymorphonuclear leukocytes infiltrate the infarcted area, and this may continue for up to 7 days if coronary perfusion does not increase or myocardial demand does not decrease.

Healing MI is characterized by mononuclear cells and fibroblasts and the absence of polymorphonuclear leukocytes. The entire healing process takes 5 to 6 weeks and can be altered by coronary reperfusion.

Healed MI refers to scar tissue without cellular infiltration.

CLINICAL FEATURES VARY WIDELY

Sir William Osler said, “Variability is the law of life, and as no two faces are the same, so no two bodies are alike, and no two individuals react alike and behave alike under the abnormal conditions which we know as disease.”⁸

Just so, patients with acute MI display a wide variety of presentations, from no symptoms (about 25%) to severe, crushing chest pain. Discomfort may occur in the upper back, neck, jaw, teeth, arms, wrist, and epigastrium. Shortness of breath, diaphoresis, nausea, vomiting, and even syncope may occur. Unlike in acute aortic dissection, the discomfort is not usually maximal at its onset: it builds up in a crescendo manner. It is not usually changed by position, but can lessen in intensity upon standing. The discomfort in the chest is deep and visceral, and typically not well localized. A pressure sensation, air hunger, or “gas buildup” can be described. The only symptom may be shortness of breath or severe diaphoresis. The symptoms can last from minutes to hours and can be relieved by sublingual nitroglycerin. Atypical or less-prominent symptoms may make the diagnosis more difficult in the elderly, patients with diabetes mellitus, and women.

The physical examination during acute MI usually finds no clear-cut distinguishing features. The patient may appear pale and diaphoretic, and the skin cool to the touch. Heart sounds are generally soft. A fourth heart sound may be audible. Blood pressure may be low, but it can vary widely. Tachycardia, particularly sinus tachycardia, and pulmonary edema are poor prognostic signs.

In view of the wide variation in presentations, the history and physical findings can raise the suspicion of acute MI, but sequential electrocardiograms and measurements of biomarkers (troponin) are always necessary.

ELECTROCARDIOGRAPHY: NECESSARY BUT NOT SUFFICIENT

ST-elevation MI vs non-ST-elevation MI

Changes in the ST segment can be very dynamic, making sequential tracings very useful. Rhythm disturbances and heart block are also more likely to be recorded when using sequential readings.

Pitfalls to electrocardiographic diagnosis

Prior MI. Q waves or QS complexes, when the Q wave is sufficiently wide (≥ 0.03 msec) or deep (≥ 1 mV), usually indicate a previous MI. However, many nuances that further raise or lower the suspicion for previous MI need to be considered. These are beyond the scope of this brief review but are available in the 2007 update.

Right ventricular infarction often requires the use of right-sided leads, which may reveal ST elevation in V₄R.

ECHOCARDIOGRAPHY IF THE DIAGNOSIS IS IN DOUBT

Echocardiography is an excellent way to assess wall-motion abnormalities. In the absence of any wall-motion abnormality, a large ST-elevation MI is unlikely. A large wall-motion abnormality would verify the probability of ongoing acute MI and thus would help with rapid decision-making.

Furthermore, echocardiography can help determine the likelihood that the patient has aortic dissection or pulmonary embolism, either of which can mimic acute MI but requires very different treatment.

CLINICAL CLASSIFICATION OF ACUTE MI

The new classification scheme of the different types of MI is shown in Table 4.

The new classification scheme does not include myocardial necrosis from mechanical manipulation of the heart during open heart surgery, from cardioversion, or from toxic drugs.

As clinicians are aware, it is not unusual to see elevated biomarker levels in a host of conditions unrelated to acute myocardial ischemia or MI. The new classification of acute MI is most helpful in this regard. It will likely be even more helpful in guiding treatment and management when new ultrasensitive troponin assays are widely introduced into clinical practice.

The new classification also negotiates the controversy regarding elevated biomarker levels following percutaneous coronary intervention. In brief, elevation of biomarkers is not entirely avoidable even with a successful percutaneous coronary intervention, and furthermore, there is no scientific cutoff for biomarker elevations. So, by arbitrary convention, the troponin level must rise to more than three times the 99th percentile upper reference limit to make the diagnosis of type 4a MI. A separate type 4b MI is ascribed to angiographic or autopsy-proven stent thrombosis.

The new guidelines also suggest that troponin values be more than five times the 99th percentile of the normal reference range during the first 72 hours following coronary artery bypass graft surgery (CABG) when considering a CABG-related MI (type 5). Whenever new pathologic Q waves appear in territories other than those identified before the procedure, MI should be considered, especially if associated with elevated biomarkers, new wall-motion abnormalities, or hemodynamic instability.

Thus, the diagnosis of acute MI now has widely accepted global criteria that distinguish various types of acute MI that occur under multiple circumstances. It is expected that describing the type of acute MI according to the new criteria will further enhance our understanding of the event, its proper management, and its prognosis.

Acute myocardial infarction (MI) portends important and substantial consequences. Angioplasty or fibrinolytic therapy to open the blocked coronary artery is proven to improve the patient’s chances of surviving without consequent morbidity or death. But the diagnosis is not always straightforward. The presentation of acute MI can vary widely, and a number of other conditions—many of them equally serious emergencies—can mimic its symptoms, electrocardiographic signs, and biomarker patterns.

In an attempt to improve the accuracy of the diagnosis of MI, a multinational task force met in 1999 under the auspices of the European Society of Cardiology and the American College of Cardiology. The goal was to develop a simple, clinically oriented definition of MI that could be widely adopted. A document was created and published simultaneously in 2000 in the European Heart Journal and the Journal of the American College of Cardiology.¹ These organizations updated their paper in 2007 with a new definition of acute MI to account for advances in diagnosis and management.²

In this article we will review the new definition and how to make the diagnosis of acute MI today. Specifically, the updated definition includes:

• Subtypes of acute MI
• Imaging tests supporting the diagnosis
• Biomarker thresholds after percutaneous coronary intervention or bypass grafting.

TROPONIN: BETTER THAN CK, BUT NOT PERFECT

The original 2000 paper¹ and the 2007 update² featured the use of the cardiac biomarker troponin, which is considerably more sensitive and specific for heart damage than total creatine kinase (CK) or its isoform, CK-MB.

The new, more-sensitive biomarker-based definition of MI resulted in more cases of MI being diagnosed, and this has attracted the attention and scrutiny of many, especially population scientists and interventional cardiologists.³ This change has caused some controversy, especially when dealing with small rises in troponin following percutaneous coronary intervention.

In addition, some confusion over terminology remains. For example, the phrase “troponin leak” is often used to describe cases in which serum troponin levels rise but there is no MI. However, most experts believe that a rise and fall in troponin is due to true myocardial cell death. Troponin I and T are such large molecules that they cannot “leak” from a cardiac cell unless there has been irreparable cellular damage—that is, cell death.

Creatine kinase still has a role

In some cases, CK and CK-MB may be helpful in determining the acuity of myocardial necrosis, but their use will vary by institution. These biomarkers typically rise 2 to 4 hours after the initial event and fall within 24 to 48 hours, whereas troponin levels stay elevated for days or weeks. Thus, the presence of troponin without CK and CK-MB in the right clinical context may indicate a past MI that is no longer acute.

INFARCTION: CELL DEATH DUE TO ISCHEMIA

MI is myocardial cell death due to prolonged ischemia. Under the microscope, it can be categorized as coagulation necrosis in which ghost-like cell structures remain after hypoxic insult (typical of most MIs) or contraction band necrosis with amorphous cells that cannot contract anymore, the latter often a hallmark of excessive catecholamine damage or reperfusion injury. Apoptosis occurs in the heart but is technically not considered necrosis and is thought not to be associated with elevated troponin levels.^6,7

In experiments in animals, cell death can occur as little as 20 minutes after coronary artery occlusion, although completion of infarction is thought to take 2 to 4 hours. The time to infarct completion may be longer in patients with collateral circulation or when the culprit coronary artery has intermittent (“stuttering”) occlusion. Preconditioning of myocardial cells with intermittent ischemia can also influence the timing of myocardial necrosis by protecting against cell death to some extent. Alteration in myocardial demand can influence the time required for completion of infarction either favorably or unfavorably; hence, reducing myocardial demand is beneficial in acute MI.

Three pathologic phases of MI

MI can be categorized pathologically as acute, healing, or healed.

Acute MI. In the first 6 hours after coronary artery occlusion, coagulation necrosis can be seen with no cellular infiltration. After 6 hours, polymorphonuclear leukocytes infiltrate the infarcted area, and this may continue for up to 7 days if coronary perfusion does not increase or myocardial demand does not decrease.

Healing MI is characterized by mononuclear cells and fibroblasts and the absence of polymorphonuclear leukocytes. The entire healing process takes 5 to 6 weeks and can be altered by coronary reperfusion.

Healed MI refers to scar tissue without cellular infiltration.

CLINICAL FEATURES VARY WIDELY

Sir William Osler said, “Variability is the law of life, and as no two faces are the same, so no two bodies are alike, and no two individuals react alike and behave alike under the abnormal conditions which we know as disease.”⁸

Just so, patients with acute MI display a wide variety of presentations, from no symptoms (about 25%) to severe, crushing chest pain. Discomfort may occur in the upper back, neck, jaw, teeth, arms, wrist, and epigastrium. Shortness of breath, diaphoresis, nausea, vomiting, and even syncope may occur. Unlike in acute aortic dissection, the discomfort is not usually maximal at its onset: it builds up in a crescendo manner. It is not usually changed by position, but can lessen in intensity upon standing. The discomfort in the chest is deep and visceral, and typically not well localized. A pressure sensation, air hunger, or “gas buildup” can be described. The only symptom may be shortness of breath or severe diaphoresis. The symptoms can last from minutes to hours and can be relieved by sublingual nitroglycerin. Atypical or less-prominent symptoms may make the diagnosis more difficult in the elderly, patients with diabetes mellitus, and women.

The physical examination during acute MI usually finds no clear-cut distinguishing features. The patient may appear pale and diaphoretic, and the skin cool to the touch. Heart sounds are generally soft. A fourth heart sound may be audible. Blood pressure may be low, but it can vary widely. Tachycardia, particularly sinus tachycardia, and pulmonary edema are poor prognostic signs.

In view of the wide variation in presentations, the history and physical findings can raise the suspicion of acute MI, but sequential electrocardiograms and measurements of biomarkers (troponin) are always necessary.

ELECTROCARDIOGRAPHY: NECESSARY BUT NOT SUFFICIENT

ST-elevation MI vs non-ST-elevation MI

Changes in the ST segment can be very dynamic, making sequential tracings very useful. Rhythm disturbances and heart block are also more likely to be recorded when using sequential readings.

Pitfalls to electrocardiographic diagnosis

Prior MI. Q waves or QS complexes, when the Q wave is sufficiently wide (≥ 0.03 msec) or deep (≥ 1 mV), usually indicate a previous MI. However, many nuances that further raise or lower the suspicion for previous MI need to be considered. These are beyond the scope of this brief review but are available in the 2007 update.

Right ventricular infarction often requires the use of right-sided leads, which may reveal ST elevation in V₄R.

ECHOCARDIOGRAPHY IF THE DIAGNOSIS IS IN DOUBT

Echocardiography is an excellent way to assess wall-motion abnormalities. In the absence of any wall-motion abnormality, a large ST-elevation MI is unlikely. A large wall-motion abnormality would verify the probability of ongoing acute MI and thus would help with rapid decision-making.

Furthermore, echocardiography can help determine the likelihood that the patient has aortic dissection or pulmonary embolism, either of which can mimic acute MI but requires very different treatment.

CLINICAL CLASSIFICATION OF ACUTE MI

The new classification scheme of the different types of MI is shown in Table 4.

The new classification scheme does not include myocardial necrosis from mechanical manipulation of the heart during open heart surgery, from cardioversion, or from toxic drugs.

As clinicians are aware, it is not unusual to see elevated biomarker levels in a host of conditions unrelated to acute myocardial ischemia or MI. The new classification of acute MI is most helpful in this regard. It will likely be even more helpful in guiding treatment and management when new ultrasensitive troponin assays are widely introduced into clinical practice.

The new classification also negotiates the controversy regarding elevated biomarker levels following percutaneous coronary intervention. In brief, elevation of biomarkers is not entirely avoidable even with a successful percutaneous coronary intervention, and furthermore, there is no scientific cutoff for biomarker elevations. So, by arbitrary convention, the troponin level must rise to more than three times the 99th percentile upper reference limit to make the diagnosis of type 4a MI. A separate type 4b MI is ascribed to angiographic or autopsy-proven stent thrombosis.

The new guidelines also suggest that troponin values be more than five times the 99th percentile of the normal reference range during the first 72 hours following coronary artery bypass graft surgery (CABG) when considering a CABG-related MI (type 5). Whenever new pathologic Q waves appear in territories other than those identified before the procedure, MI should be considered, especially if associated with elevated biomarkers, new wall-motion abnormalities, or hemodynamic instability.

Thus, the diagnosis of acute MI now has widely accepted global criteria that distinguish various types of acute MI that occur under multiple circumstances. It is expected that describing the type of acute MI according to the new criteria will further enhance our understanding of the event, its proper management, and its prognosis.

Acute myocardial infarction (MI) portends important and substantial consequences. Angioplasty or fibrinolytic therapy to open the blocked coronary artery is proven to improve the patient’s chances of surviving without consequent morbidity or death. But the diagnosis is not always straightforward. The presentation of acute MI can vary widely, and a number of other conditions—many of them equally serious emergencies—can mimic its symptoms, electrocardiographic signs, and biomarker patterns.

In an attempt to improve the accuracy of the diagnosis of MI, a multinational task force met in 1999 under the auspices of the European Society of Cardiology and the American College of Cardiology. The goal was to develop a simple, clinically oriented definition of MI that could be widely adopted. A document was created and published simultaneously in 2000 in the European Heart Journal and the Journal of the American College of Cardiology.¹ These organizations updated their paper in 2007 with a new definition of acute MI to account for advances in diagnosis and management.²

In this article we will review the new definition and how to make the diagnosis of acute MI today. Specifically, the updated definition includes:

• Subtypes of acute MI
• Imaging tests supporting the diagnosis
• Biomarker thresholds after percutaneous coronary intervention or bypass grafting.

TROPONIN: BETTER THAN CK, BUT NOT PERFECT

The original 2000 paper¹ and the 2007 update² featured the use of the cardiac biomarker troponin, which is considerably more sensitive and specific for heart damage than total creatine kinase (CK) or its isoform, CK-MB.

The new, more-sensitive biomarker-based definition of MI resulted in more cases of MI being diagnosed, and this has attracted the attention and scrutiny of many, especially population scientists and interventional cardiologists.³ This change has caused some controversy, especially when dealing with small rises in troponin following percutaneous coronary intervention.

In addition, some confusion over terminology remains. For example, the phrase “troponin leak” is often used to describe cases in which serum troponin levels rise but there is no MI. However, most experts believe that a rise and fall in troponin is due to true myocardial cell death. Troponin I and T are such large molecules that they cannot “leak” from a cardiac cell unless there has been irreparable cellular damage—that is, cell death.

Creatine kinase still has a role

In some cases, CK and CK-MB may be helpful in determining the acuity of myocardial necrosis, but their use will vary by institution. These biomarkers typically rise 2 to 4 hours after the initial event and fall within 24 to 48 hours, whereas troponin levels stay elevated for days or weeks. Thus, the presence of troponin without CK and CK-MB in the right clinical context may indicate a past MI that is no longer acute.

INFARCTION: CELL DEATH DUE TO ISCHEMIA

MI is myocardial cell death due to prolonged ischemia. Under the microscope, it can be categorized as coagulation necrosis in which ghost-like cell structures remain after hypoxic insult (typical of most MIs) or contraction band necrosis with amorphous cells that cannot contract anymore, the latter often a hallmark of excessive catecholamine damage or reperfusion injury. Apoptosis occurs in the heart but is technically not considered necrosis and is thought not to be associated with elevated troponin levels.^6,7

In experiments in animals, cell death can occur as little as 20 minutes after coronary artery occlusion, although completion of infarction is thought to take 2 to 4 hours. The time to infarct completion may be longer in patients with collateral circulation or when the culprit coronary artery has intermittent (“stuttering”) occlusion. Preconditioning of myocardial cells with intermittent ischemia can also influence the timing of myocardial necrosis by protecting against cell death to some extent. Alteration in myocardial demand can influence the time required for completion of infarction either favorably or unfavorably; hence, reducing myocardial demand is beneficial in acute MI.

Three pathologic phases of MI

MI can be categorized pathologically as acute, healing, or healed.

Acute MI. In the first 6 hours after coronary artery occlusion, coagulation necrosis can be seen with no cellular infiltration. After 6 hours, polymorphonuclear leukocytes infiltrate the infarcted area, and this may continue for up to 7 days if coronary perfusion does not increase or myocardial demand does not decrease.

Healing MI is characterized by mononuclear cells and fibroblasts and the absence of polymorphonuclear leukocytes. The entire healing process takes 5 to 6 weeks and can be altered by coronary reperfusion.

Healed MI refers to scar tissue without cellular infiltration.

CLINICAL FEATURES VARY WIDELY

Sir William Osler said, “Variability is the law of life, and as no two faces are the same, so no two bodies are alike, and no two individuals react alike and behave alike under the abnormal conditions which we know as disease.”⁸

Just so, patients with acute MI display a wide variety of presentations, from no symptoms (about 25%) to severe, crushing chest pain. Discomfort may occur in the upper back, neck, jaw, teeth, arms, wrist, and epigastrium. Shortness of breath, diaphoresis, nausea, vomiting, and even syncope may occur. Unlike in acute aortic dissection, the discomfort is not usually maximal at its onset: it builds up in a crescendo manner. It is not usually changed by position, but can lessen in intensity upon standing. The discomfort in the chest is deep and visceral, and typically not well localized. A pressure sensation, air hunger, or “gas buildup” can be described. The only symptom may be shortness of breath or severe diaphoresis. The symptoms can last from minutes to hours and can be relieved by sublingual nitroglycerin. Atypical or less-prominent symptoms may make the diagnosis more difficult in the elderly, patients with diabetes mellitus, and women.

The physical examination during acute MI usually finds no clear-cut distinguishing features. The patient may appear pale and diaphoretic, and the skin cool to the touch. Heart sounds are generally soft. A fourth heart sound may be audible. Blood pressure may be low, but it can vary widely. Tachycardia, particularly sinus tachycardia, and pulmonary edema are poor prognostic signs.

In view of the wide variation in presentations, the history and physical findings can raise the suspicion of acute MI, but sequential electrocardiograms and measurements of biomarkers (troponin) are always necessary.

ELECTROCARDIOGRAPHY: NECESSARY BUT NOT SUFFICIENT

ST-elevation MI vs non-ST-elevation MI

Changes in the ST segment can be very dynamic, making sequential tracings very useful. Rhythm disturbances and heart block are also more likely to be recorded when using sequential readings.

Pitfalls to electrocardiographic diagnosis

Prior MI. Q waves or QS complexes, when the Q wave is sufficiently wide (≥ 0.03 msec) or deep (≥ 1 mV), usually indicate a previous MI. However, many nuances that further raise or lower the suspicion for previous MI need to be considered. These are beyond the scope of this brief review but are available in the 2007 update.

Right ventricular infarction often requires the use of right-sided leads, which may reveal ST elevation in V₄R.

ECHOCARDIOGRAPHY IF THE DIAGNOSIS IS IN DOUBT

Echocardiography is an excellent way to assess wall-motion abnormalities. In the absence of any wall-motion abnormality, a large ST-elevation MI is unlikely. A large wall-motion abnormality would verify the probability of ongoing acute MI and thus would help with rapid decision-making.

Furthermore, echocardiography can help determine the likelihood that the patient has aortic dissection or pulmonary embolism, either of which can mimic acute MI but requires very different treatment.

CLINICAL CLASSIFICATION OF ACUTE MI

The new classification scheme of the different types of MI is shown in Table 4.

The new classification scheme does not include myocardial necrosis from mechanical manipulation of the heart during open heart surgery, from cardioversion, or from toxic drugs.

As clinicians are aware, it is not unusual to see elevated biomarker levels in a host of conditions unrelated to acute myocardial ischemia or MI. The new classification of acute MI is most helpful in this regard. It will likely be even more helpful in guiding treatment and management when new ultrasensitive troponin assays are widely introduced into clinical practice.

The new classification also negotiates the controversy regarding elevated biomarker levels following percutaneous coronary intervention. In brief, elevation of biomarkers is not entirely avoidable even with a successful percutaneous coronary intervention, and furthermore, there is no scientific cutoff for biomarker elevations. So, by arbitrary convention, the troponin level must rise to more than three times the 99th percentile upper reference limit to make the diagnosis of type 4a MI. A separate type 4b MI is ascribed to angiographic or autopsy-proven stent thrombosis.

The new guidelines also suggest that troponin values be more than five times the 99th percentile of the normal reference range during the first 72 hours following coronary artery bypass graft surgery (CABG) when considering a CABG-related MI (type 5). Whenever new pathologic Q waves appear in territories other than those identified before the procedure, MI should be considered, especially if associated with elevated biomarkers, new wall-motion abnormalities, or hemodynamic instability.

Thus, the diagnosis of acute MI now has widely accepted global criteria that distinguish various types of acute MI that occur under multiple circumstances. It is expected that describing the type of acute MI according to the new criteria will further enhance our understanding of the event, its proper management, and its prognosis.

In Reply: We appreciate the interest and comments of Dr Hirsch. Due to space limitations of the Journal’s 1-Minute Consult format, we were unable to elaborate on the cost and reimbursement. Since shingles vaccine is not covered by Medicare part B, reimbursement and administration of this vaccine remains challenging, and resources like eDispense are helpful tools for the physician to simplify the process of reimbursement—with no charge to the physician.

In Reply: We appreciate the interest and comments of Dr Hirsch. Due to space limitations of the Journal’s 1-Minute Consult format, we were unable to elaborate on the cost and reimbursement. Since shingles vaccine is not covered by Medicare part B, reimbursement and administration of this vaccine remains challenging, and resources like eDispense are helpful tools for the physician to simplify the process of reimbursement—with no charge to the physician.

In Reply: We appreciate the interest and comments of Dr Hirsch. Due to space limitations of the Journal’s 1-Minute Consult format, we were unable to elaborate on the cost and reimbursement. Since shingles vaccine is not covered by Medicare part B, reimbursement and administration of this vaccine remains challenging, and resources like eDispense are helpful tools for the physician to simplify the process of reimbursement—with no charge to the physician.

To the Editor: In their 1-Minute Consult, Drs. Singh and Englund give a thorough review of Zostavax, the vaccine to prevent shingles.¹ Unfortunately, the information they provided on cost and reimbursement is incomplete.

As they noted, this vaccine is not covered by Medicare part B and is mandated to be covered by Medicare part D as a “prescription drug.” Furthermore, the vaccine administration charge cannot be billed to Medicare part B. Since physician offices do not bill prescription drug plans, physicians are permitted to administer the vaccine, charge the patient for the vaccine and administration, and have the patient submit the receipt to his or her prescription drug provider for reimbursement. There is no fee schedule for this vaccine, so physicians are free to charge a fee that they deem reasonable. For patients without part D, it is reasonable to ask them to call their prescription provider and inquire about coverage before vaccination, since many commercial plans will not cover the vaccine, and the $200 or more price may be unaffordable for many.

Alternatively, physician offices may enroll with a private vendor, eDispense Vaccine Manager, at enroll.edispense.com, and submit charges for Zostavax electronically to the patient’s Medicare part D provider; eDispense is contracted with most of the large part D providers. This service allows the physician to input the patient’s demographics and get an immediate response, showing the patient’s coverage and copayment, and allowing the physician to submit the claim electronically. There is no charge to the physician, and the reimbursement covers the cost of the vaccine, the administration cost, and a small profit. If the patient wishes to pay for the vaccine, the system can produce a receipt containing all the information needed for submission by the patient to the insurer. (Note: I have no financial or ownership interest in eDispense.com or Merck.)

Since recommending and administering Zostavax is soon to become the standard of care, the availability of these options will provide better care than the authors’ recommendation that patients pick up the vaccine and transport it back to the physician’s office on ice, which risks defrosting and inactivating the vaccine, or leaving patients to find a vaccine provider on their own.

To the Editor: In their 1-Minute Consult, Drs. Singh and Englund give a thorough review of Zostavax, the vaccine to prevent shingles.¹ Unfortunately, the information they provided on cost and reimbursement is incomplete.

As they noted, this vaccine is not covered by Medicare part B and is mandated to be covered by Medicare part D as a “prescription drug.” Furthermore, the vaccine administration charge cannot be billed to Medicare part B. Since physician offices do not bill prescription drug plans, physicians are permitted to administer the vaccine, charge the patient for the vaccine and administration, and have the patient submit the receipt to his or her prescription drug provider for reimbursement. There is no fee schedule for this vaccine, so physicians are free to charge a fee that they deem reasonable. For patients without part D, it is reasonable to ask them to call their prescription provider and inquire about coverage before vaccination, since many commercial plans will not cover the vaccine, and the $200 or more price may be unaffordable for many.

Alternatively, physician offices may enroll with a private vendor, eDispense Vaccine Manager, at enroll.edispense.com, and submit charges for Zostavax electronically to the patient’s Medicare part D provider; eDispense is contracted with most of the large part D providers. This service allows the physician to input the patient’s demographics and get an immediate response, showing the patient’s coverage and copayment, and allowing the physician to submit the claim electronically. There is no charge to the physician, and the reimbursement covers the cost of the vaccine, the administration cost, and a small profit. If the patient wishes to pay for the vaccine, the system can produce a receipt containing all the information needed for submission by the patient to the insurer. (Note: I have no financial or ownership interest in eDispense.com or Merck.)

Since recommending and administering Zostavax is soon to become the standard of care, the availability of these options will provide better care than the authors’ recommendation that patients pick up the vaccine and transport it back to the physician’s office on ice, which risks defrosting and inactivating the vaccine, or leaving patients to find a vaccine provider on their own.

To the Editor: In their 1-Minute Consult, Drs. Singh and Englund give a thorough review of Zostavax, the vaccine to prevent shingles.¹ Unfortunately, the information they provided on cost and reimbursement is incomplete.

As they noted, this vaccine is not covered by Medicare part B and is mandated to be covered by Medicare part D as a “prescription drug.” Furthermore, the vaccine administration charge cannot be billed to Medicare part B. Since physician offices do not bill prescription drug plans, physicians are permitted to administer the vaccine, charge the patient for the vaccine and administration, and have the patient submit the receipt to his or her prescription drug provider for reimbursement. There is no fee schedule for this vaccine, so physicians are free to charge a fee that they deem reasonable. For patients without part D, it is reasonable to ask them to call their prescription provider and inquire about coverage before vaccination, since many commercial plans will not cover the vaccine, and the $200 or more price may be unaffordable for many.

Alternatively, physician offices may enroll with a private vendor, eDispense Vaccine Manager, at enroll.edispense.com, and submit charges for Zostavax electronically to the patient’s Medicare part D provider; eDispense is contracted with most of the large part D providers. This service allows the physician to input the patient’s demographics and get an immediate response, showing the patient’s coverage and copayment, and allowing the physician to submit the claim electronically. There is no charge to the physician, and the reimbursement covers the cost of the vaccine, the administration cost, and a small profit. If the patient wishes to pay for the vaccine, the system can produce a receipt containing all the information needed for submission by the patient to the insurer. (Note: I have no financial or ownership interest in eDispense.com or Merck.)

Since recommending and administering Zostavax is soon to become the standard of care, the availability of these options will provide better care than the authors’ recommendation that patients pick up the vaccine and transport it back to the physician’s office on ice, which risks defrosting and inactivating the vaccine, or leaving patients to find a vaccine provider on their own.

To the Editor: I read with interest the article by Perkins and colleagues, “A young pregnant woman with shortness of breath” on pages 788–792 of the November 2008 issue of the Cleveland Clinic Journal of Medicine. An incorrect meaning occurs in the article. Thyrotoxicosis is the state of symptomatic thyroid hormone excess, of both endogenous and exogenous cause. It is not synonymous with hyperthyroidism, which is the result of excessive thyroid function.

To the Editor: I read with interest the article by Perkins and colleagues, “A young pregnant woman with shortness of breath” on pages 788–792 of the November 2008 issue of the Cleveland Clinic Journal of Medicine. An incorrect meaning occurs in the article. Thyrotoxicosis is the state of symptomatic thyroid hormone excess, of both endogenous and exogenous cause. It is not synonymous with hyperthyroidism, which is the result of excessive thyroid function.

To the Editor: I read with interest the article by Perkins and colleagues, “A young pregnant woman with shortness of breath” on pages 788–792 of the November 2008 issue of the Cleveland Clinic Journal of Medicine. An incorrect meaning occurs in the article. Thyrotoxicosis is the state of symptomatic thyroid hormone excess, of both endogenous and exogenous cause. It is not synonymous with hyperthyroidism, which is the result of excessive thyroid function.

Your 84-year-old patient's son is distraught. “I know Mom has dementia, but I don’t understand why she cannot relax. She is busy all night long, taking out the silverware, packing her clothes, and trying to leave the house. Sometimes she tells me that there are little children in the room. These hallucinations scare me, although they do not seem to bother her very much. She keeps me awake; I’m often late to work because I’m up much of the night. I’m afraid I’m going to lose my job; and I don’t want to put Mom into a nursing home. Please give her a medication for this behavior.”

Another of your patients, an 82-year-old man, is admitted to a nursing home after an emergency hospitalization in the geriatric psychiatry unit. His daughter left him alone with her boyfriend one morning while she went to work. Not recognizing him, your patient attacked the young man with a kitchen knife. The police initially arrested your patient and then had him admitted to the psychiatric unit. He is discharged 2 weeks later to the nursing home.

Can anything be done for these patients?

A GROWING PROBLEM

Dementia is a growing problem with the aging of the population. At the time of the 2000 census there were 4.5 million people in the United States with Alzheimer disease, the most common type of dementia,¹ and the prevalence is expected to increase to 13.2 million by the year 2050.¹

CONSERVATIVE MEASURES ARE THE MAINSTAY OF TREATMENT

As for offending drugs, removing an antimuscarinic or anticholinergic drug may resolve hallucinations; stopping propoxyphene (Darvon) may improve sleep.

No drugs are approved for treating hallucinations, agitation, or other distressing behavior in neurodegenerative diseases such as Alzheimer dementia. Rather, the mainstay of treatment is behavioral and environmental modification.⁷ In an environment optimized to maximize comfort, reduce stress, and permit safe wandering, behavioral medications may be unnecessary.

Nevertheless, environments are not always optimal, and physicians may offer medications to treat behavioral symptoms to improve quality of life and to let patients keep living in the community.

Below, we discuss the drugs used to treat behavioral problems in dementia, evidence for the efficacy of these drugs, and their potential for adverse effects.

ANTIPSYCHOTIC DRUGS: SMALL BENEFIT, BIG RISK

Although antipsychotic drugs, both typical and atypical, are often used to treat dementia- related behaviors, their beneficial effects are minimal and adverse effects are common.^8,9

Aggression has been considered a symptom that might respond to an atypical antipsychotic drug.¹⁰ However, the Clinical Antipsychotic Trials of Intervention Effectiveness—Alzheimer’s Disease (CATIE-AD) trial¹¹ found no differences in efficacy between placebo and the atypical antipsychotics olanzapine (Zyprexa), quetiapine (Seroquel), and risperidone (Risperdal) in treating psychosis, aggression, and agitation in dementia. In that study, rates of drug discontinuation due to adverse effects ranged from 5% for placebo to 24% for olanzapine. Overall, 82% of the patients stopped taking their initially assigned medications during the 36-week period of the trial.¹¹

Antipsychotic drugs may cause more adverse effects in patients with Parkinson disease or dementia with Lewy bodies, and medications with the least dopamine D2 receptor blockade are chosen to reduce the impact on the parkinsonism. Patients with movement disorders were excluded from the CATIE-AD study, and data on this topic are very limited. Quetiapine and olanzapine are often used as alternatives to clozapine (Clozaril) for treating psychosis in Parkinson disease and may have a role in dementia with Lewy bodies.^12,13

Atypical antipsychotics carry significant risk of illness and even death. The US Food and Drug Administration (FDA) has published advisories about hyperglycemia, cerebrovascular events, and death.¹⁴ Returning to the older, “typical” antipsychotics is not a solution either, given their high incidence of extrapyramidal symptoms¹⁵ and potentially higher risk of death.^16,17

Even if effective, try stopping the drug

Even in the few situations in dementia in which antipsychotics prove efficacious, a trial of dose-reduction and possible discontinuation is a part of the appropriate plan of care. Symptoms such as aggression and delusions may decrease as the underlying dementia progresses.² A consensus statement on antipsychotic drug use in the elderly¹⁸ recommended stopping antipsychotic drugs as follows:

• If given for delirium—discontinue the drug after 1 week
• For agitated dementia—taper within 3 to 6 months to determine the lowest effective maintenance dose
• For psychotic major depression—discontinue after 6 months
• For mania with psychosis—discontinue after 3 months.¹⁸

Disorders for which antipsychotics are not recommended are irritability, hostility, generalized anxiety, and insomnia. In contrast with recommendations for dementia-related behaviors, the psychosis of schizophrenia is treated lifelong at the lowest effective dose of medication.

ANTIDEPRESSANTS: MANY CHOICES, LITTLE EVIDENCE

Depression is hard to assess in a patient with dementia, particularly since apathy is a common symptom in both dementia and depression and may confuse the presentation. Additionally, screening tests for depression have not been validated in the demented elderly.

Depression in dementia is associated with poorer quality of life, greater disability in activities of daily living, a faster cognitive decline, a high rate of nursing home placement, a higher death rate, and a higher frequency of depression and burden in caregivers.¹⁹ Quality of life may improve with antidepressant treatment even if the patient does not meet all the criteria for a major depressive disorder. Provisional recommendations for diagnosing depression in dementia suggest using three (instead of five) or more criteria, and include irritability or social isolation as additional criteria.²⁰

Choosing an antidepressant

Only a few randomized controlled trials of antidepressants for depression with dementia have been completed, each with a small number of patients.

Mirtazapine (Remeron) is what we recommend to improve sleep and appetite and restore lost weight.²¹ It can be used in patients with Parkinson disease or parkinsonian symptoms who experience increased tremors or bradykinesia with selective serotonin reuptake inhibitors (SSRIs). On the other hand, it may not be the best option for those with diabetes mellitus, metabolic syndrome, hyperlipidemia, or obesity. It may rarely also cause a reversible agranulocytosis.

Venlafaxine (Effexor) and duloxetine (Cymbalta) are serotonin-norepinephrine reuptake inhibitors (SNRIs) and may help in concomitant pain syndromes.²² Either drug can cause anorexia at any dose and can elevate blood pressure at higher doses. Venlafaxine may also cause insomnia in some patients.

Bupropion (Wellbutrin) can be difficult to titrate to an effective dose in an older person with unsuspected renal insufficiency, and it may interact at the P450 complex.²³ The risk of seizures is greater at higher bupropion serum levels. There is also a high incidence of weight loss. Frail elderly patients, those with hypertension, and those vulnerable to hallucinations will likely do better with another drug.

Nefazodone is a third- or fourth-line antidepressive choice because of the risk of hepatic failure. However, it can help reduce disabling anxiety associated with depression. The FDA requires periodic liver function testing if this drug is used.

Trazodone in low doses (≤ 100 mg) each evening may help with sleep, but it cannot be titrated to antidepressive doses in older adults because of orthostatic effects.

Nortriptyline is recommended by some geriatricians for depression or pathologic crying in patients with mixed vascular dementia. However, it often causes cardiac conduction delays with reflex sympathetic tachycardia and anticholinergic side effects.

Monoamine oxidase inhibitors interact with many foods and drugs, limiting their use in older adults.

Methylphenidate (Ritalin) at low doses is used off-label for depression in palliative care, with noted rapid improvements in mood and appetite.²⁴ Monitoring for increases in blood pressure, heart rate, and respiratory rate is essential if this stimulant is chosen. Patients who respond may make a transition to other traditional drugs after 2 to 4 weeks.

Caveats with SSRIs

• Despite the safety profile of SSRIs in older adults, care must be taken when prescribing them to frail elderly patients, given recent data associating SSRIs with falls and fragility fractures^25,26 and urinary incontinence.²⁷
• SSRIs may decrease appetite during initial treatment.
• Sertraline (Zoloft) may have to be started at a very low dose to decrease possible adverse gastrointestinal symptoms, such as diarrhea.
• Paroxetine (Paxil) has multiple interactions at the cytochrome P450 complex and has the most anticholinergic properties of the SSRIs, rendering it more likely to cause adverse drug reactions, constipation, and delirium.
• Daily fluoxetine (Prozac) may not be appropriate in older adults because of its long half-life and the risk of insomnia and agitation.²⁸
• Tremors can emerge with all the SSRIs; akathisia, dystonia, and parkinsonism are also possible.²⁹
• Hyponatremia, bruising, and increased bleeding time can occur with any SSRI.
• Abrupt cessation of any SSRI except fluoxetine (due to its long half-life) or of SNRIs may cause a very unpleasant flu-like withdrawal syndrome.
• Apathy can be a reversible, dose-dependent adverse effect of SSRIs in young persons³⁰; there are no data on the dose at which this adverse effect might emerge in demented elderly patients.

In a systematic review, Sink et al³¹ found citalopram (Celexa) to help reduce nondepressive agitation.

How long should depression be treated?

Antidepressant treatment is typically for 6 to 12 months. However, the optimal duration in an older adult with dementia is not known and is not addressed in either the American Psychiatric Association practice guideline on dementia³² or the position statement of the American Association for Geriatric Psychiatry.³³

Patients with executive dysfunction, particularly those with perseveration and diminished inhibition, may be less likely to respond to antidepressants, and the symptoms are more likely to recur if they do respond.³⁴ It may be appropriate to treat them for a year and then withdraw the drug and monitor for recurrence. Some patients may need indefinite treatment.

Apathy in elderly patients with dementia is common. It is found in nearly half of elderly patients with mild dementia and in nearly all of those with severe dementia. If accompanied by depressive symptoms such as sadness, guilt, feelings of worthlessness, passive or active death wish, changes in sleep or appetite, or tearfulness, apathy and other depressive symptoms may respond to antidepressive treatment—both behavioral and pharmacologic. When present in dementia without depressive symptomatology, apathy is unlikely to respond to antidepressants. In particular, SSRIs may actually induce or worsen apathy through their effect on the angular gyrus. Apathy can be very frustrating to family members but not troublesome at all to the patient.

No medication carries an indication for apathy in dementia. Although stimulants such as methylphenidate and modafinil (Provigil) have been used, there is no evidence to date from any controlled study of efficacy and safety in this population.

Try nondrug measures concomitantly

Given the limited evidence of efficacy of antidepressive therapy in demented elderly patients, nonpharmacologic therapy should be offered concomitantly.

Evidence-based nonpharmacologic treatment for depression in dementia includes increasing enjoyable activities and socialization with people and pets, reducing the need to perform frustrating activities, redirecting perseverative behaviors and speech, and addressing caregiver needs.³⁴ Exercise may improve physical functioning in depression with dementia.³⁵ A comprehensive sleep program may improve associated sleep disorders.³⁶

An intensive collaborative-care intervention³⁷ resulted in more demented elderly patients in the intervention group receiving a cholinesterase inhibitor and an antidepressive than in the usual-care group. Outcomes included fewer behavioral symptoms, less caregiver distress, and less caregiver depression.

So far, no randomized trial has shown electroconvulsive therapy to be effective in elderly patients with depression and dementia.³⁸

ANTICONVULSANT DRUGS MAY STABILIZE MOOD

On the basis of small studies with some contradictory outcomes,³⁹ both older and newer anticonvulsants have been used in nonpsychotic agitation, aggression, and impulsivity in a variety of psychiatric disorders, brain injury, and dementia.⁴⁰ Most of the data are on the older drugs such as valproic acid and carbamazepine (Tegretol).

Valproic acid is associated with an adverse metabolic profile (hyperglycemia, weight gain, and hyperlipidemia),^41,42 dose-related orthostasis, sedation, and worsening cognitive performance. In addition, the possibility of thrombocytopenia and blood level fluctuations requires monitoring. Older adults may tolerate 250 to 500 mg/day with minimal adverse effects.

Carbamazepine reduced aggression in a blinded, placebo-controlled study in nursing home patients.⁴³ Use of carbamazepine requires monitoring of hematologic and liver profiles, alters the metabolism of itself and other drugs, and is associated with dose-related sedation.

Lamotrigine (Lamictal) takes a long time to titrate but may help with nonpsychotic agitation and impulsivity; it is a relatively new drug, and there are limited data to support its use at this time in the elderly.

Gabapentin (Gabarone), in case reports at doses primarily from 600 to 1,200 mg/day, reduced behavioral and psychological problems of patients with dementia and with good renal clearance.⁴⁴ Some patients may experience tremors or oversedation.

Phenytoin (Dilantin) is not a good choice for behavioral problems because of unwanted effects on teeth, bones, and balance.

Levetiracetam (Keppra) may cause behavioral disturbances to emerge or worsen.⁴⁵

Emerging evidence suggests that all anticonvulsants may also be associated with an increased risk of depressive symptoms.

COGNITIVE ENHANCERS MAY IMPROVE BEHAVIOR

Acetylcholinesterase inhibitors may improve some behavioral symptoms of dementia. In an open-label retrospective trial, delusionality, irritability, anxiety, disinhibition, and agitation improved in some patients on these drugs.⁴⁶ Patients most likely to respond were those with the most impairment from these behaviors and those with depressive or apathetic symptoms.⁴⁶ A Cochrane review found a modest beneficial effect on behavior.⁴⁷

Acetylcholinesterase inhibitors may reduce symptoms of apathy. Additionally, they actually improve depressive symptoms in mild to moderate dementia independent of any effect on cognition.⁴⁸

Memantine (Namenda), approved for the treatment of moderate to severe dementia, may reduce the prevalence and incidence of agitation, particularly in more advanced dementia.⁴⁹

The cognitive enhancers all require several weeks for titration and are not helpful for the acute management of behavioral or depressive symptoms.

OTHER DRUGS

Beta-blockers⁵⁰ and estrogen⁵¹ have been studied as off-label, nonneuroleptic treatments for male aggression. Use of progesterone in men with inappropriate sexual behavior⁵² may have benefit; further interventions are reviewed by Srinivasan and Weinberg.⁵³ These recommendations are based on small case series. In addition, the hormonal treatments may carry significant morbidity.

Sedative hypnotics were evaluated for sleep difficulties in demented patients in a meta-analysis by Glass et al,⁵⁴ who found adverse cognitive events, psychomotor events, and daytime fatigue more common (5, 2.6, and 3.8 times, respectively) in the sedative group than in the placebo group.

For agitation in delirium, haloperidol (Haldol) is preferable to benzodiazepines, based on studies from the 1970s.⁵⁵ Although benzodiazepines carry an indication for anxiety, newly prescribed benzodiazepines and those with a longer half-life are associated with hip fractures in older adults,⁵⁶ possibly from sedation.

WHAT TO DO FOR YOUR PATIENTS

The 84-year-old woman

For the 84-year-old woman who is keeping her son awake all night, recommend making the environment safe for her to wander, including placing a bolt on the doors leading to the basement and outdoors and moving the knives to an area that she cannot reach, to avoid accidents. Recommend that she be given things to do that are repetitive, such as folding towels and arranging drawers. Referring her to day care may improve socialization and increase physical activity during the day, possibly improving her sleep time at night.

The 82-year-old man

Let’s assume the 82-year-old man arrested and then hospitalized is placed on risperidone 1 mg twice daily prior to discharge to the nursing home. In the nursing home, he becomes irritable with any change in his routine: the door has to be open by exactly 6 inches; his meals have to be identical and served on time; the newspaper needs to arrive by 8 AM. Since routine is paramount in the nursing home, the staff accommodates his need for a very regular schedule. Donepezil (Aricept) and memantine can be added as cognitive enhancers, and citalopram can be added for possible depression and obsessive features. The daughter should then be approached about reducing the risperidone dose and, hopefully, discontinuing it in the future.

Comment. A stable, routine environment is the most important intervention for managing this aggressive resident’s behavior, although he may have been helped to some degree by the adjunct medications. Once he is stable, the daughter may be able to bring him home for weekends and holidays, as long as she is advised never to surprise him with an unexpected visit or to bring home unexpected guests.

Your 84-year-old patient's son is distraught. “I know Mom has dementia, but I don’t understand why she cannot relax. She is busy all night long, taking out the silverware, packing her clothes, and trying to leave the house. Sometimes she tells me that there are little children in the room. These hallucinations scare me, although they do not seem to bother her very much. She keeps me awake; I’m often late to work because I’m up much of the night. I’m afraid I’m going to lose my job; and I don’t want to put Mom into a nursing home. Please give her a medication for this behavior.”

Another of your patients, an 82-year-old man, is admitted to a nursing home after an emergency hospitalization in the geriatric psychiatry unit. His daughter left him alone with her boyfriend one morning while she went to work. Not recognizing him, your patient attacked the young man with a kitchen knife. The police initially arrested your patient and then had him admitted to the psychiatric unit. He is discharged 2 weeks later to the nursing home.

Can anything be done for these patients?

A GROWING PROBLEM

Dementia is a growing problem with the aging of the population. At the time of the 2000 census there were 4.5 million people in the United States with Alzheimer disease, the most common type of dementia,¹ and the prevalence is expected to increase to 13.2 million by the year 2050.¹

CONSERVATIVE MEASURES ARE THE MAINSTAY OF TREATMENT

As for offending drugs, removing an antimuscarinic or anticholinergic drug may resolve hallucinations; stopping propoxyphene (Darvon) may improve sleep.

No drugs are approved for treating hallucinations, agitation, or other distressing behavior in neurodegenerative diseases such as Alzheimer dementia. Rather, the mainstay of treatment is behavioral and environmental modification.⁷ In an environment optimized to maximize comfort, reduce stress, and permit safe wandering, behavioral medications may be unnecessary.

Nevertheless, environments are not always optimal, and physicians may offer medications to treat behavioral symptoms to improve quality of life and to let patients keep living in the community.

Below, we discuss the drugs used to treat behavioral problems in dementia, evidence for the efficacy of these drugs, and their potential for adverse effects.

ANTIPSYCHOTIC DRUGS: SMALL BENEFIT, BIG RISK

Although antipsychotic drugs, both typical and atypical, are often used to treat dementia- related behaviors, their beneficial effects are minimal and adverse effects are common.^8,9

Aggression has been considered a symptom that might respond to an atypical antipsychotic drug.¹⁰ However, the Clinical Antipsychotic Trials of Intervention Effectiveness—Alzheimer’s Disease (CATIE-AD) trial¹¹ found no differences in efficacy between placebo and the atypical antipsychotics olanzapine (Zyprexa), quetiapine (Seroquel), and risperidone (Risperdal) in treating psychosis, aggression, and agitation in dementia. In that study, rates of drug discontinuation due to adverse effects ranged from 5% for placebo to 24% for olanzapine. Overall, 82% of the patients stopped taking their initially assigned medications during the 36-week period of the trial.¹¹

Antipsychotic drugs may cause more adverse effects in patients with Parkinson disease or dementia with Lewy bodies, and medications with the least dopamine D2 receptor blockade are chosen to reduce the impact on the parkinsonism. Patients with movement disorders were excluded from the CATIE-AD study, and data on this topic are very limited. Quetiapine and olanzapine are often used as alternatives to clozapine (Clozaril) for treating psychosis in Parkinson disease and may have a role in dementia with Lewy bodies.^12,13

Atypical antipsychotics carry significant risk of illness and even death. The US Food and Drug Administration (FDA) has published advisories about hyperglycemia, cerebrovascular events, and death.¹⁴ Returning to the older, “typical” antipsychotics is not a solution either, given their high incidence of extrapyramidal symptoms¹⁵ and potentially higher risk of death.^16,17

Even if effective, try stopping the drug

Even in the few situations in dementia in which antipsychotics prove efficacious, a trial of dose-reduction and possible discontinuation is a part of the appropriate plan of care. Symptoms such as aggression and delusions may decrease as the underlying dementia progresses.² A consensus statement on antipsychotic drug use in the elderly¹⁸ recommended stopping antipsychotic drugs as follows:

• If given for delirium—discontinue the drug after 1 week
• For agitated dementia—taper within 3 to 6 months to determine the lowest effective maintenance dose
• For psychotic major depression—discontinue after 6 months
• For mania with psychosis—discontinue after 3 months.¹⁸

Disorders for which antipsychotics are not recommended are irritability, hostility, generalized anxiety, and insomnia. In contrast with recommendations for dementia-related behaviors, the psychosis of schizophrenia is treated lifelong at the lowest effective dose of medication.

ANTIDEPRESSANTS: MANY CHOICES, LITTLE EVIDENCE

Depression is hard to assess in a patient with dementia, particularly since apathy is a common symptom in both dementia and depression and may confuse the presentation. Additionally, screening tests for depression have not been validated in the demented elderly.

Depression in dementia is associated with poorer quality of life, greater disability in activities of daily living, a faster cognitive decline, a high rate of nursing home placement, a higher death rate, and a higher frequency of depression and burden in caregivers.¹⁹ Quality of life may improve with antidepressant treatment even if the patient does not meet all the criteria for a major depressive disorder. Provisional recommendations for diagnosing depression in dementia suggest using three (instead of five) or more criteria, and include irritability or social isolation as additional criteria.²⁰

Choosing an antidepressant

Only a few randomized controlled trials of antidepressants for depression with dementia have been completed, each with a small number of patients.

Mirtazapine (Remeron) is what we recommend to improve sleep and appetite and restore lost weight.²¹ It can be used in patients with Parkinson disease or parkinsonian symptoms who experience increased tremors or bradykinesia with selective serotonin reuptake inhibitors (SSRIs). On the other hand, it may not be the best option for those with diabetes mellitus, metabolic syndrome, hyperlipidemia, or obesity. It may rarely also cause a reversible agranulocytosis.

Venlafaxine (Effexor) and duloxetine (Cymbalta) are serotonin-norepinephrine reuptake inhibitors (SNRIs) and may help in concomitant pain syndromes.²² Either drug can cause anorexia at any dose and can elevate blood pressure at higher doses. Venlafaxine may also cause insomnia in some patients.

Bupropion (Wellbutrin) can be difficult to titrate to an effective dose in an older person with unsuspected renal insufficiency, and it may interact at the P450 complex.²³ The risk of seizures is greater at higher bupropion serum levels. There is also a high incidence of weight loss. Frail elderly patients, those with hypertension, and those vulnerable to hallucinations will likely do better with another drug.

Nefazodone is a third- or fourth-line antidepressive choice because of the risk of hepatic failure. However, it can help reduce disabling anxiety associated with depression. The FDA requires periodic liver function testing if this drug is used.

Trazodone in low doses (≤ 100 mg) each evening may help with sleep, but it cannot be titrated to antidepressive doses in older adults because of orthostatic effects.

Nortriptyline is recommended by some geriatricians for depression or pathologic crying in patients with mixed vascular dementia. However, it often causes cardiac conduction delays with reflex sympathetic tachycardia and anticholinergic side effects.

Monoamine oxidase inhibitors interact with many foods and drugs, limiting their use in older adults.

Methylphenidate (Ritalin) at low doses is used off-label for depression in palliative care, with noted rapid improvements in mood and appetite.²⁴ Monitoring for increases in blood pressure, heart rate, and respiratory rate is essential if this stimulant is chosen. Patients who respond may make a transition to other traditional drugs after 2 to 4 weeks.

Caveats with SSRIs

• Despite the safety profile of SSRIs in older adults, care must be taken when prescribing them to frail elderly patients, given recent data associating SSRIs with falls and fragility fractures^25,26 and urinary incontinence.²⁷
• SSRIs may decrease appetite during initial treatment.
• Sertraline (Zoloft) may have to be started at a very low dose to decrease possible adverse gastrointestinal symptoms, such as diarrhea.
• Paroxetine (Paxil) has multiple interactions at the cytochrome P450 complex and has the most anticholinergic properties of the SSRIs, rendering it more likely to cause adverse drug reactions, constipation, and delirium.
• Daily fluoxetine (Prozac) may not be appropriate in older adults because of its long half-life and the risk of insomnia and agitation.²⁸
• Tremors can emerge with all the SSRIs; akathisia, dystonia, and parkinsonism are also possible.²⁹
• Hyponatremia, bruising, and increased bleeding time can occur with any SSRI.
• Abrupt cessation of any SSRI except fluoxetine (due to its long half-life) or of SNRIs may cause a very unpleasant flu-like withdrawal syndrome.
• Apathy can be a reversible, dose-dependent adverse effect of SSRIs in young persons³⁰; there are no data on the dose at which this adverse effect might emerge in demented elderly patients.

In a systematic review, Sink et al³¹ found citalopram (Celexa) to help reduce nondepressive agitation.

How long should depression be treated?

Antidepressant treatment is typically for 6 to 12 months. However, the optimal duration in an older adult with dementia is not known and is not addressed in either the American Psychiatric Association practice guideline on dementia³² or the position statement of the American Association for Geriatric Psychiatry.³³

Patients with executive dysfunction, particularly those with perseveration and diminished inhibition, may be less likely to respond to antidepressants, and the symptoms are more likely to recur if they do respond.³⁴ It may be appropriate to treat them for a year and then withdraw the drug and monitor for recurrence. Some patients may need indefinite treatment.

Apathy in elderly patients with dementia is common. It is found in nearly half of elderly patients with mild dementia and in nearly all of those with severe dementia. If accompanied by depressive symptoms such as sadness, guilt, feelings of worthlessness, passive or active death wish, changes in sleep or appetite, or tearfulness, apathy and other depressive symptoms may respond to antidepressive treatment—both behavioral and pharmacologic. When present in dementia without depressive symptomatology, apathy is unlikely to respond to antidepressants. In particular, SSRIs may actually induce or worsen apathy through their effect on the angular gyrus. Apathy can be very frustrating to family members but not troublesome at all to the patient.

No medication carries an indication for apathy in dementia. Although stimulants such as methylphenidate and modafinil (Provigil) have been used, there is no evidence to date from any controlled study of efficacy and safety in this population.

Try nondrug measures concomitantly

Given the limited evidence of efficacy of antidepressive therapy in demented elderly patients, nonpharmacologic therapy should be offered concomitantly.

Evidence-based nonpharmacologic treatment for depression in dementia includes increasing enjoyable activities and socialization with people and pets, reducing the need to perform frustrating activities, redirecting perseverative behaviors and speech, and addressing caregiver needs.³⁴ Exercise may improve physical functioning in depression with dementia.³⁵ A comprehensive sleep program may improve associated sleep disorders.³⁶

An intensive collaborative-care intervention³⁷ resulted in more demented elderly patients in the intervention group receiving a cholinesterase inhibitor and an antidepressive than in the usual-care group. Outcomes included fewer behavioral symptoms, less caregiver distress, and less caregiver depression.

So far, no randomized trial has shown electroconvulsive therapy to be effective in elderly patients with depression and dementia.³⁸

ANTICONVULSANT DRUGS MAY STABILIZE MOOD

On the basis of small studies with some contradictory outcomes,³⁹ both older and newer anticonvulsants have been used in nonpsychotic agitation, aggression, and impulsivity in a variety of psychiatric disorders, brain injury, and dementia.⁴⁰ Most of the data are on the older drugs such as valproic acid and carbamazepine (Tegretol).

Valproic acid is associated with an adverse metabolic profile (hyperglycemia, weight gain, and hyperlipidemia),^41,42 dose-related orthostasis, sedation, and worsening cognitive performance. In addition, the possibility of thrombocytopenia and blood level fluctuations requires monitoring. Older adults may tolerate 250 to 500 mg/day with minimal adverse effects.

Carbamazepine reduced aggression in a blinded, placebo-controlled study in nursing home patients.⁴³ Use of carbamazepine requires monitoring of hematologic and liver profiles, alters the metabolism of itself and other drugs, and is associated with dose-related sedation.

Lamotrigine (Lamictal) takes a long time to titrate but may help with nonpsychotic agitation and impulsivity; it is a relatively new drug, and there are limited data to support its use at this time in the elderly.

Gabapentin (Gabarone), in case reports at doses primarily from 600 to 1,200 mg/day, reduced behavioral and psychological problems of patients with dementia and with good renal clearance.⁴⁴ Some patients may experience tremors or oversedation.

Phenytoin (Dilantin) is not a good choice for behavioral problems because of unwanted effects on teeth, bones, and balance.

Levetiracetam (Keppra) may cause behavioral disturbances to emerge or worsen.⁴⁵

Emerging evidence suggests that all anticonvulsants may also be associated with an increased risk of depressive symptoms.

COGNITIVE ENHANCERS MAY IMPROVE BEHAVIOR

Acetylcholinesterase inhibitors may improve some behavioral symptoms of dementia. In an open-label retrospective trial, delusionality, irritability, anxiety, disinhibition, and agitation improved in some patients on these drugs.⁴⁶ Patients most likely to respond were those with the most impairment from these behaviors and those with depressive or apathetic symptoms.⁴⁶ A Cochrane review found a modest beneficial effect on behavior.⁴⁷

Acetylcholinesterase inhibitors may reduce symptoms of apathy. Additionally, they actually improve depressive symptoms in mild to moderate dementia independent of any effect on cognition.⁴⁸

Memantine (Namenda), approved for the treatment of moderate to severe dementia, may reduce the prevalence and incidence of agitation, particularly in more advanced dementia.⁴⁹

The cognitive enhancers all require several weeks for titration and are not helpful for the acute management of behavioral or depressive symptoms.

OTHER DRUGS

Beta-blockers⁵⁰ and estrogen⁵¹ have been studied as off-label, nonneuroleptic treatments for male aggression. Use of progesterone in men with inappropriate sexual behavior⁵² may have benefit; further interventions are reviewed by Srinivasan and Weinberg.⁵³ These recommendations are based on small case series. In addition, the hormonal treatments may carry significant morbidity.

Sedative hypnotics were evaluated for sleep difficulties in demented patients in a meta-analysis by Glass et al,⁵⁴ who found adverse cognitive events, psychomotor events, and daytime fatigue more common (5, 2.6, and 3.8 times, respectively) in the sedative group than in the placebo group.

For agitation in delirium, haloperidol (Haldol) is preferable to benzodiazepines, based on studies from the 1970s.⁵⁵ Although benzodiazepines carry an indication for anxiety, newly prescribed benzodiazepines and those with a longer half-life are associated with hip fractures in older adults,⁵⁶ possibly from sedation.

WHAT TO DO FOR YOUR PATIENTS

The 84-year-old woman

For the 84-year-old woman who is keeping her son awake all night, recommend making the environment safe for her to wander, including placing a bolt on the doors leading to the basement and outdoors and moving the knives to an area that she cannot reach, to avoid accidents. Recommend that she be given things to do that are repetitive, such as folding towels and arranging drawers. Referring her to day care may improve socialization and increase physical activity during the day, possibly improving her sleep time at night.

The 82-year-old man

Let’s assume the 82-year-old man arrested and then hospitalized is placed on risperidone 1 mg twice daily prior to discharge to the nursing home. In the nursing home, he becomes irritable with any change in his routine: the door has to be open by exactly 6 inches; his meals have to be identical and served on time; the newspaper needs to arrive by 8 AM. Since routine is paramount in the nursing home, the staff accommodates his need for a very regular schedule. Donepezil (Aricept) and memantine can be added as cognitive enhancers, and citalopram can be added for possible depression and obsessive features. The daughter should then be approached about reducing the risperidone dose and, hopefully, discontinuing it in the future.

Comment. A stable, routine environment is the most important intervention for managing this aggressive resident’s behavior, although he may have been helped to some degree by the adjunct medications. Once he is stable, the daughter may be able to bring him home for weekends and holidays, as long as she is advised never to surprise him with an unexpected visit or to bring home unexpected guests.

Your 84-year-old patient's son is distraught. “I know Mom has dementia, but I don’t understand why she cannot relax. She is busy all night long, taking out the silverware, packing her clothes, and trying to leave the house. Sometimes she tells me that there are little children in the room. These hallucinations scare me, although they do not seem to bother her very much. She keeps me awake; I’m often late to work because I’m up much of the night. I’m afraid I’m going to lose my job; and I don’t want to put Mom into a nursing home. Please give her a medication for this behavior.”

Another of your patients, an 82-year-old man, is admitted to a nursing home after an emergency hospitalization in the geriatric psychiatry unit. His daughter left him alone with her boyfriend one morning while she went to work. Not recognizing him, your patient attacked the young man with a kitchen knife. The police initially arrested your patient and then had him admitted to the psychiatric unit. He is discharged 2 weeks later to the nursing home.

Can anything be done for these patients?

A GROWING PROBLEM

Dementia is a growing problem with the aging of the population. At the time of the 2000 census there were 4.5 million people in the United States with Alzheimer disease, the most common type of dementia,¹ and the prevalence is expected to increase to 13.2 million by the year 2050.¹

CONSERVATIVE MEASURES ARE THE MAINSTAY OF TREATMENT

As for offending drugs, removing an antimuscarinic or anticholinergic drug may resolve hallucinations; stopping propoxyphene (Darvon) may improve sleep.

No drugs are approved for treating hallucinations, agitation, or other distressing behavior in neurodegenerative diseases such as Alzheimer dementia. Rather, the mainstay of treatment is behavioral and environmental modification.⁷ In an environment optimized to maximize comfort, reduce stress, and permit safe wandering, behavioral medications may be unnecessary.

Nevertheless, environments are not always optimal, and physicians may offer medications to treat behavioral symptoms to improve quality of life and to let patients keep living in the community.

Below, we discuss the drugs used to treat behavioral problems in dementia, evidence for the efficacy of these drugs, and their potential for adverse effects.

ANTIPSYCHOTIC DRUGS: SMALL BENEFIT, BIG RISK

Although antipsychotic drugs, both typical and atypical, are often used to treat dementia- related behaviors, their beneficial effects are minimal and adverse effects are common.^8,9

Aggression has been considered a symptom that might respond to an atypical antipsychotic drug.¹⁰ However, the Clinical Antipsychotic Trials of Intervention Effectiveness—Alzheimer’s Disease (CATIE-AD) trial¹¹ found no differences in efficacy between placebo and the atypical antipsychotics olanzapine (Zyprexa), quetiapine (Seroquel), and risperidone (Risperdal) in treating psychosis, aggression, and agitation in dementia. In that study, rates of drug discontinuation due to adverse effects ranged from 5% for placebo to 24% for olanzapine. Overall, 82% of the patients stopped taking their initially assigned medications during the 36-week period of the trial.¹¹

Antipsychotic drugs may cause more adverse effects in patients with Parkinson disease or dementia with Lewy bodies, and medications with the least dopamine D2 receptor blockade are chosen to reduce the impact on the parkinsonism. Patients with movement disorders were excluded from the CATIE-AD study, and data on this topic are very limited. Quetiapine and olanzapine are often used as alternatives to clozapine (Clozaril) for treating psychosis in Parkinson disease and may have a role in dementia with Lewy bodies.^12,13

Atypical antipsychotics carry significant risk of illness and even death. The US Food and Drug Administration (FDA) has published advisories about hyperglycemia, cerebrovascular events, and death.¹⁴ Returning to the older, “typical” antipsychotics is not a solution either, given their high incidence of extrapyramidal symptoms¹⁵ and potentially higher risk of death.^16,17

Even if effective, try stopping the drug

Even in the few situations in dementia in which antipsychotics prove efficacious, a trial of dose-reduction and possible discontinuation is a part of the appropriate plan of care. Symptoms such as aggression and delusions may decrease as the underlying dementia progresses.² A consensus statement on antipsychotic drug use in the elderly¹⁸ recommended stopping antipsychotic drugs as follows:

• If given for delirium—discontinue the drug after 1 week
• For agitated dementia—taper within 3 to 6 months to determine the lowest effective maintenance dose
• For psychotic major depression—discontinue after 6 months
• For mania with psychosis—discontinue after 3 months.¹⁸

Disorders for which antipsychotics are not recommended are irritability, hostility, generalized anxiety, and insomnia. In contrast with recommendations for dementia-related behaviors, the psychosis of schizophrenia is treated lifelong at the lowest effective dose of medication.

ANTIDEPRESSANTS: MANY CHOICES, LITTLE EVIDENCE

Depression is hard to assess in a patient with dementia, particularly since apathy is a common symptom in both dementia and depression and may confuse the presentation. Additionally, screening tests for depression have not been validated in the demented elderly.

Depression in dementia is associated with poorer quality of life, greater disability in activities of daily living, a faster cognitive decline, a high rate of nursing home placement, a higher death rate, and a higher frequency of depression and burden in caregivers.¹⁹ Quality of life may improve with antidepressant treatment even if the patient does not meet all the criteria for a major depressive disorder. Provisional recommendations for diagnosing depression in dementia suggest using three (instead of five) or more criteria, and include irritability or social isolation as additional criteria.²⁰

Choosing an antidepressant

Only a few randomized controlled trials of antidepressants for depression with dementia have been completed, each with a small number of patients.

Mirtazapine (Remeron) is what we recommend to improve sleep and appetite and restore lost weight.²¹ It can be used in patients with Parkinson disease or parkinsonian symptoms who experience increased tremors or bradykinesia with selective serotonin reuptake inhibitors (SSRIs). On the other hand, it may not be the best option for those with diabetes mellitus, metabolic syndrome, hyperlipidemia, or obesity. It may rarely also cause a reversible agranulocytosis.

Venlafaxine (Effexor) and duloxetine (Cymbalta) are serotonin-norepinephrine reuptake inhibitors (SNRIs) and may help in concomitant pain syndromes.²² Either drug can cause anorexia at any dose and can elevate blood pressure at higher doses. Venlafaxine may also cause insomnia in some patients.

Bupropion (Wellbutrin) can be difficult to titrate to an effective dose in an older person with unsuspected renal insufficiency, and it may interact at the P450 complex.²³ The risk of seizures is greater at higher bupropion serum levels. There is also a high incidence of weight loss. Frail elderly patients, those with hypertension, and those vulnerable to hallucinations will likely do better with another drug.

Nefazodone is a third- or fourth-line antidepressive choice because of the risk of hepatic failure. However, it can help reduce disabling anxiety associated with depression. The FDA requires periodic liver function testing if this drug is used.

Trazodone in low doses (≤ 100 mg) each evening may help with sleep, but it cannot be titrated to antidepressive doses in older adults because of orthostatic effects.

Nortriptyline is recommended by some geriatricians for depression or pathologic crying in patients with mixed vascular dementia. However, it often causes cardiac conduction delays with reflex sympathetic tachycardia and anticholinergic side effects.

Monoamine oxidase inhibitors interact with many foods and drugs, limiting their use in older adults.

Methylphenidate (Ritalin) at low doses is used off-label for depression in palliative care, with noted rapid improvements in mood and appetite.²⁴ Monitoring for increases in blood pressure, heart rate, and respiratory rate is essential if this stimulant is chosen. Patients who respond may make a transition to other traditional drugs after 2 to 4 weeks.

Caveats with SSRIs

• Despite the safety profile of SSRIs in older adults, care must be taken when prescribing them to frail elderly patients, given recent data associating SSRIs with falls and fragility fractures^25,26 and urinary incontinence.²⁷
• SSRIs may decrease appetite during initial treatment.
• Sertraline (Zoloft) may have to be started at a very low dose to decrease possible adverse gastrointestinal symptoms, such as diarrhea.
• Paroxetine (Paxil) has multiple interactions at the cytochrome P450 complex and has the most anticholinergic properties of the SSRIs, rendering it more likely to cause adverse drug reactions, constipation, and delirium.
• Daily fluoxetine (Prozac) may not be appropriate in older adults because of its long half-life and the risk of insomnia and agitation.²⁸
• Tremors can emerge with all the SSRIs; akathisia, dystonia, and parkinsonism are also possible.²⁹
• Hyponatremia, bruising, and increased bleeding time can occur with any SSRI.
• Abrupt cessation of any SSRI except fluoxetine (due to its long half-life) or of SNRIs may cause a very unpleasant flu-like withdrawal syndrome.
• Apathy can be a reversible, dose-dependent adverse effect of SSRIs in young persons³⁰; there are no data on the dose at which this adverse effect might emerge in demented elderly patients.

In a systematic review, Sink et al³¹ found citalopram (Celexa) to help reduce nondepressive agitation.

How long should depression be treated?

Antidepressant treatment is typically for 6 to 12 months. However, the optimal duration in an older adult with dementia is not known and is not addressed in either the American Psychiatric Association practice guideline on dementia³² or the position statement of the American Association for Geriatric Psychiatry.³³

Patients with executive dysfunction, particularly those with perseveration and diminished inhibition, may be less likely to respond to antidepressants, and the symptoms are more likely to recur if they do respond.³⁴ It may be appropriate to treat them for a year and then withdraw the drug and monitor for recurrence. Some patients may need indefinite treatment.

Apathy in elderly patients with dementia is common. It is found in nearly half of elderly patients with mild dementia and in nearly all of those with severe dementia. If accompanied by depressive symptoms such as sadness, guilt, feelings of worthlessness, passive or active death wish, changes in sleep or appetite, or tearfulness, apathy and other depressive symptoms may respond to antidepressive treatment—both behavioral and pharmacologic. When present in dementia without depressive symptomatology, apathy is unlikely to respond to antidepressants. In particular, SSRIs may actually induce or worsen apathy through their effect on the angular gyrus. Apathy can be very frustrating to family members but not troublesome at all to the patient.

No medication carries an indication for apathy in dementia. Although stimulants such as methylphenidate and modafinil (Provigil) have been used, there is no evidence to date from any controlled study of efficacy and safety in this population.

Try nondrug measures concomitantly

Given the limited evidence of efficacy of antidepressive therapy in demented elderly patients, nonpharmacologic therapy should be offered concomitantly.

Evidence-based nonpharmacologic treatment for depression in dementia includes increasing enjoyable activities and socialization with people and pets, reducing the need to perform frustrating activities, redirecting perseverative behaviors and speech, and addressing caregiver needs.³⁴ Exercise may improve physical functioning in depression with dementia.³⁵ A comprehensive sleep program may improve associated sleep disorders.³⁶

An intensive collaborative-care intervention³⁷ resulted in more demented elderly patients in the intervention group receiving a cholinesterase inhibitor and an antidepressive than in the usual-care group. Outcomes included fewer behavioral symptoms, less caregiver distress, and less caregiver depression.

So far, no randomized trial has shown electroconvulsive therapy to be effective in elderly patients with depression and dementia.³⁸

ANTICONVULSANT DRUGS MAY STABILIZE MOOD

On the basis of small studies with some contradictory outcomes,³⁹ both older and newer anticonvulsants have been used in nonpsychotic agitation, aggression, and impulsivity in a variety of psychiatric disorders, brain injury, and dementia.⁴⁰ Most of the data are on the older drugs such as valproic acid and carbamazepine (Tegretol).

Valproic acid is associated with an adverse metabolic profile (hyperglycemia, weight gain, and hyperlipidemia),^41,42 dose-related orthostasis, sedation, and worsening cognitive performance. In addition, the possibility of thrombocytopenia and blood level fluctuations requires monitoring. Older adults may tolerate 250 to 500 mg/day with minimal adverse effects.

Carbamazepine reduced aggression in a blinded, placebo-controlled study in nursing home patients.⁴³ Use of carbamazepine requires monitoring of hematologic and liver profiles, alters the metabolism of itself and other drugs, and is associated with dose-related sedation.

Lamotrigine (Lamictal) takes a long time to titrate but may help with nonpsychotic agitation and impulsivity; it is a relatively new drug, and there are limited data to support its use at this time in the elderly.

Gabapentin (Gabarone), in case reports at doses primarily from 600 to 1,200 mg/day, reduced behavioral and psychological problems of patients with dementia and with good renal clearance.⁴⁴ Some patients may experience tremors or oversedation.

Phenytoin (Dilantin) is not a good choice for behavioral problems because of unwanted effects on teeth, bones, and balance.

Levetiracetam (Keppra) may cause behavioral disturbances to emerge or worsen.⁴⁵

Emerging evidence suggests that all anticonvulsants may also be associated with an increased risk of depressive symptoms.

COGNITIVE ENHANCERS MAY IMPROVE BEHAVIOR

Acetylcholinesterase inhibitors may improve some behavioral symptoms of dementia. In an open-label retrospective trial, delusionality, irritability, anxiety, disinhibition, and agitation improved in some patients on these drugs.⁴⁶ Patients most likely to respond were those with the most impairment from these behaviors and those with depressive or apathetic symptoms.⁴⁶ A Cochrane review found a modest beneficial effect on behavior.⁴⁷

Acetylcholinesterase inhibitors may reduce symptoms of apathy. Additionally, they actually improve depressive symptoms in mild to moderate dementia independent of any effect on cognition.⁴⁸

Memantine (Namenda), approved for the treatment of moderate to severe dementia, may reduce the prevalence and incidence of agitation, particularly in more advanced dementia.⁴⁹

The cognitive enhancers all require several weeks for titration and are not helpful for the acute management of behavioral or depressive symptoms.

OTHER DRUGS

Beta-blockers⁵⁰ and estrogen⁵¹ have been studied as off-label, nonneuroleptic treatments for male aggression. Use of progesterone in men with inappropriate sexual behavior⁵² may have benefit; further interventions are reviewed by Srinivasan and Weinberg.⁵³ These recommendations are based on small case series. In addition, the hormonal treatments may carry significant morbidity.

Sedative hypnotics were evaluated for sleep difficulties in demented patients in a meta-analysis by Glass et al,⁵⁴ who found adverse cognitive events, psychomotor events, and daytime fatigue more common (5, 2.6, and 3.8 times, respectively) in the sedative group than in the placebo group.

For agitation in delirium, haloperidol (Haldol) is preferable to benzodiazepines, based on studies from the 1970s.⁵⁵ Although benzodiazepines carry an indication for anxiety, newly prescribed benzodiazepines and those with a longer half-life are associated with hip fractures in older adults,⁵⁶ possibly from sedation.

WHAT TO DO FOR YOUR PATIENTS

The 84-year-old woman

For the 84-year-old woman who is keeping her son awake all night, recommend making the environment safe for her to wander, including placing a bolt on the doors leading to the basement and outdoors and moving the knives to an area that she cannot reach, to avoid accidents. Recommend that she be given things to do that are repetitive, such as folding towels and arranging drawers. Referring her to day care may improve socialization and increase physical activity during the day, possibly improving her sleep time at night.

The 82-year-old man

Let’s assume the 82-year-old man arrested and then hospitalized is placed on risperidone 1 mg twice daily prior to discharge to the nursing home. In the nursing home, he becomes irritable with any change in his routine: the door has to be open by exactly 6 inches; his meals have to be identical and served on time; the newspaper needs to arrive by 8 AM. Since routine is paramount in the nursing home, the staff accommodates his need for a very regular schedule. Donepezil (Aricept) and memantine can be added as cognitive enhancers, and citalopram can be added for possible depression and obsessive features. The daughter should then be approached about reducing the risperidone dose and, hopefully, discontinuing it in the future.

Comment. A stable, routine environment is the most important intervention for managing this aggressive resident’s behavior, although he may have been helped to some degree by the adjunct medications. Once he is stable, the daughter may be able to bring him home for weekends and holidays, as long as she is advised never to surprise him with an unexpected visit or to bring home unexpected guests.