Mood Stabilizers in Dementia: An Interactive Grand Rounds

An audience-interactive case study session followed didactic presentations on the history and current data regarding mood stabilizers for the treatment of behavioral disturbances of dementia. This symposium was held in association with the 15th Annual Meeting of the AAGP.

This program was supported by an unrestricted educational grant from Abbott Laboratories.

Introduction

Anticonvulsant mood stabilizers were first developed for the treatment of epilepsy, but clinical studies soon revealed their broader efficacy, such as in bipolar illness and migraine. Today, new information at the subcellular level suggests that the ultimate potential of some mood stabilizing drugs may not yet be fully realized.

New information, coupled with an expanding understanding of the pathogenesis and pathophysiology of the behavioral and psychiatric symptoms associated with dementia, have illuminated new ways to optimize the therapeutic effects of mood stabilizing drugs.

The Expanding Role of Mood Stabilizers

Lithium
The history of mood stabilizers begins with lithium. First developed as an anticonvulsant for epilepsy, lithium was soon found to be more widely effective, based on uncontrolled studies into the treatment of mania in elderly persons. According to J. Craig Nelson, MD, Program Chair and Professor of Psychiatry at Yale University School of Medicine, four large open series of case studies published in 1980 documented the efficacy of lithium in older patients (Himmelhoch Am J Psychiatry 1980).

In elderly people overall, lithium is a very difficult drug to use. It is well known to have a narrow therapeutic index. Safe lithium use depends on normal renal clearance, and in elderly individuals, age-related reductions in renal clearance, combined with diuretic and NSAID use, salt-restricted diets, poor hydration, and other factors, may increase the risk of toxicity. In addition, older people have increased CNS sensitivity and even at normal therapeutic doses, lithium can on occasion induce tremors, gait disturbances, and delirium.

Carbamazepine
Carbamazepine was developed in the 1950s. By 1963, its anticonvulsant properties were described; later, serendipitous observation of its beneficial psychotropic effects led to studies that resulted in its FDA approval as an anticonvulsant in 1974. Its anti-manic properties were described in 1978, and these led to reports of use for mania, then agitation. Of three placebo-controlled studies with carbamazepine for behavior disturbance, two showed a benefit over placebo at a dose of about 300 mg/d (Tariot PN et al. Am Geriatr Soc 1994; Tariot PN et al. Am Geriatr Soc 1998).

The mechanism explaining the effect of carbamazepine on mania or in the management of agitation has still not been fully ascertained, said Dr. Nelson. Proposed mechanisms are anti-kindling effects, effects on neuronal ion channels to reduce high-frequency repetitive firing; and effects on synaptic and post-synaptic transmission. Of these, anti-kindling effects especially have received support from the scientific community at several junctures.

Valproate
Valproate was synthesized in 1882, and until the early 1960s was used as an organic solvent. Its anticonvulsant activity was noted in 1963, and in 1966, its anti-manic activity was reported in a French publication. In 1967, valproate was introduced as an anti-epileptic agent in France. It was first used as an anticonvulsant in the United States in 1978.

The anti-agitation effects of valproate were observed in 1992. Between the years of 1991 and 1994, a number of placebo-controlled studies in mania were conducted, and these joined an extensive literature of open studies to describe the use of valproate in dementia and in younger patients with various organic syndromes associated with behavioral problems. The overall response rate of about 60% was effectively identical to response rates with neuroleptics for the management of agitation.

Two studies of divalproex have recently been published (Porsteinsson et al. Am J Geriatr Psychiatry 2001; Tariot PN et al. J Am Geriatr Soc 2001). The Porsteinsson study, a model for the design of larger studies and a dosing study, showed a nearly significant P-value favoring divalproex for agitation (p=0.07). (The average dose was 826 mg/d.) The larger of the two studies, by Tariot (N=175), was designed to evaluate the efficacy of divalproex at aggressive dosing levels of 20 mg/kg in patients with behavioral disorders with manic features. Of two scales used by the investigators to evaluate response, the mania scale, typically used for bipolar patients, showed no evidence of superiority of drug over placebo. However, the Cohen-Mansfield scale, which is commonly used for agitation, showed a very significant effect for divalproex over placebo.

Possible mechanisms of action of divalproex are enhancement of central GABA transmission; direct neuronal effects on Na and K; inhibition of limbic kindling; neurotransmitter effects on dopamine, serotonin, aspartate, and somatostatin, and neuroprotection.

GSK-3 and BCL-2
Neuroprotective effects, such as those identified with valproate, are the focus of much scientific attention, for implications relating to distinguishing properties and benefits of individual drugs. Specific neuroprotective actions may arise from drug effects on GSK-3 beta (the “death enzyme”), or Bcl-2, itself a major neuropathic and neuroprotective protein.

In animal studies, Bcl-2 has been found to help to protect against focal ischemia and traumatic brain injury, among a number of mechanisms of neuroprotection. Bcl-2 over-expression actually promotes regeneration of axons in the mammalian central nervous system, Dr. Nelson said. Studies have shown that lithium and valproate increase Bcl-2 levels in the rat frontal cortex (Chen G et al. Psychosom Med 1999), and that valproate enhances neurite outgrowth and growth cone formation (Manji HK et al. Biol Psychiatry 2000). Lithium, according to researchers Moore and colleagues, actually increases gray matter (Moore GJ et al. Lancet 2000).

Effective Treatment of Behavioral Disturbances in the Elderly

Among the first principles of successful disease management of patients with behavioral disturbances are avoiding drugs that may worsen symptoms or affect the progression of dementia, said Larry E. Tune, MD, Chief of the Division of Geriatric Psychiatry at the Wesley Woods Geriatric Center at Emory University. In this regard, anticholinergics are prime offenders, and are among the most commonly prescribed drugs in the elderly.

Optimal management of behavioral symptoms also requires an interdisciplinary team approach, with both non-pharmacological and pharmacological interventions. The selection of pharmacologic agents may be best informed by new antipsychotic studies that suggest target doses to optimize the risks and benefits.

BPSD (Behavioral and Psychological Symptoms of Dementia)
An approach to tailoring medications to symptoms or complexes of symptoms has been promulgated by Alistair Burns and Bryan Lawlor in the UK. BPSD, Behavioral and Psychological Symptoms of Dementia, is a term used to describe a heterogeneous range of psychological reactions, psychiatric symptoms and behaviors occurring in people with dementia of any etiology. BPSD “clusters” may be regarded as sub-syndromes under the general umbrella of dementia manifestations, Dr. Tune explained. Each cluster is organized by a common involved neurotransmitter(s). Symptoms within each cluster may thus be expected to respond best to one or more specific drug classes. Dr. Tune presented the example of psychosis, which has an underlying acetylcholine, serotonin, and dopamine basis, and may be expected to respond to antipsychotics and AChE inhibitors.

Dr. Tune suggested that integrating BPSD into the biopsychosocial approach to behavioral disturbances would first require ruling out medical, and treatable, causes for agitation or delirium, such as dementia in which anticholinergic polypharmacy is a contributing factor. The focus on acetylcholine is rational, Dr. Tune said: acetylcholine levels decline with age; acetylcholine may be a central “lesion” in dementia and delirium; and acetylcholine decreases in Alzheimer’s disease and other dementias, and with anticholinergic use.

Evidence that anticholinergics may aggravate delirium was provided by a nursing home intervention study in which 28 patients taking anticholinergic medications were randomly assigned to either continue their same anticholinergic medication, or receive an anticholinergic dose reduced by 25%. After a 2-week period, patients taking the reduced anticholinergic dose were found to be less delirious, based on results from the Saskatoon Delirium Checklist.

Cholinesterase Inhibitors
Each of the three leading cholinesterase inhibitors—galantamine, rivastigmine, and donepezil—has demonstrated significant effects on cognition function, and behavior.

An important benefit of cholinester-ase inhibitor use is preservation of a patient’s at-home functioning, and a resulting delay in nursing home placement. This was specifically demonstrated in a study with donepezil, in which 36 weeks of treatment resulted in as much as a 30-month delay in nursing home admission.

Divalproex in Dementia
Dr. Tune remarked that divalproex is effective as monotherapy for agitation and aggression, and in combination therapy when symptoms of agitation are not responsive or are worsened by antipsychotics or benzodiazepines. This agent may enable the use of target doses of antipsychotics, and may allow for greater disease specificity in treatment regimens. A trial that was revealing of the potentials of divalproex was a 6-week, double blind, placebo-controlled trial of divalproex versus placebo followed by a 12-week open label trial. Patients were started at a dose of 125 mg BID and titrated aggressively to 20 mg/kg (30 mg/kg max). Divalproex-treated patients were significantly improved, based on change in the Cohen- Mansfield Agitation Inventory total score and verbal aggression subsection.

Pharmacotherapy of Behavioral and Psychiatric Symptoms: Minimizing ADR Risk

Psychoactive drug therapy, especially in an elderly and frail population, may be associated with grave negative outcomes such as falls, said speaker Lori Daiello, Pharm D, Geriatric Psychopharmacology Specialist and President of Pharmacotherapy Solutions.

Selected Drug-Related Problems
SSRIs
Hyponatremia is one of the most common adverse events associated with SSRI antidepressant use, Dr. Daiello said. Although specific incidence data are lacking, hyponatremia can result in hospitalization and may be an undiagnosed source of confusion and lethargy. Patients at the greatest risk for SSRI-induced hyponatremia are older, female, and have comorbidities. Studies have suggested that concomitant use of antihypertensives, diuretics, ACE inhibitors, and calcium channel blockers may increase the risk of SSRI-associated hyponatremia. Another area of potential concern with the use of SSRIs within complex drug regimens is CYP450 pharmacokinetic drug interactions; paroxetine, fluoxetine and fluvoxamine can inhibit the hepatic metabolism of many medications, such as antihypertensives, cholesterol-lowering drug, and other psychoactive medications.

Benzodiazepines
Benzodiazepines have a well-known negative impact on short-term memory. Substantiating this was a small case series examining the effect of withdrawing sedative hypnotics from patients. Even those with pre-existing cognitive impairment had small improvements in short-term memory function once the drug was withdrawn (Salzman C et al. J Geriatr Psychiatry 1992).

Falls are a perennial concern with benzodiazepines, and conventional wisdom has held that the shorter the drug half-life the safer it may be for patients at risk for falls. One study from a large nursing home cohort involving 2,000 patients, found a 44% increased rate of falls in patients using benzodiazepines compared to those not taking the drugs (Ray WA et al. J Am Geriatr Soc 2000). Drug dosage was also a factor in both daytime and nighttime falls, as was the use of long-acting benzodiazepines. However, and surprisingly, short-acting benzodiazepines such as oxazepam were associated with a slight decrease in daytime falls, compared with controls. As Dr. Daiello interpreted this data, “…when you treat very anxious, psychomotor-agitated patients [who are active in daytime] with these drugs, you may have an impact by actually decreasing their falls.” However, short-acting benzodiazepine hypnotics such as zolpidem were shown to significantly increase the risk of nighttime falls.

Anticonvulsant Mood Stabilizers
Data in dementia-related behavioral disorders indicate that anticonvulsant mood stabilizers are associated with sedation that may vary depending on dose and titration rate. When dosed properly Dr. Daiello said, studies have shown that long-lasting sedation is not an issue with these agents. This finding has been borne out by studies with carbamazepine and valproate, showing no significant change in MMSE scores between treated and non-treated patients.

Chronic therapy gives rise to adverse effects specific to each anticonvulsant agent. With carbamazepine, a generally benign leukopenia may occur. Other less common adverse effects associated with long-term carbamazepine therapy are hyponatremia, bradycardia and AV block. Dr. Daiello emphasized that this agent should not be the drug of choice in patients with conduction disorders. Significant and numerous drug interactions may also occur with carbamazepine. “It has been said that carbamazepine participates in at least 120 different drug interactions—either decreased serum levels through its auto-induction of cytochrome P450 (warfarin, theophylline, many antipsychotics, valproate, zaleplon, nefazodone, alprazolam), or having its levels increased by some other common medications, notably antibiotics,” she said (Grossman F. Pharmacotherapy 1998).

Mild thrombocytopenia may occur with chronic valproate therapy (Grossman F. Pharmacotherapy 1998). Dr. Daiello cited a retrospective review of the use of valproate in a psychiatric population (N=264), with findings of a 12% prevalence of thrombocytopenia in the general population, and a 22% prevalence in a cohort of persons over age 65. However, in this population, there were no clinically significant events associated with the platelet drop and no cases of severe thrombocytopenia (Conley EL et al. Pharmacotherapy 2001;21:1325-1330). Dr. Daiello commented that there has been very little problem with thrombocytopenia within the entire data set in dementia patients, probably reflective of the use of lower doses.

In terms of drug interactions, it is significant to note that aspirin can decrease the beta-oxidation of valproate and may increase valproic acid levels. Overall, however, Dr. Daiello said, drug interactions are a lesser issue with valproate than with carbamazepine.

Gabapentin has not been evaluated in placebo-controlled trials for the treatment of dementia-related aggression or agitation but has gained interest for this use because it is not metabolized through the liver; instead, it is excreted by the kidneys unchanged, thus removing any concerns of liver-related drug interactions. Case reports in the psychiatric literature of gabapentin use in behavioral disorders of dementia number only a total of 23, and a wide range of doses have been reported in these case studies, ranging from 200 mg per day to 2400 mg/day.

In gabapentin-treated patients referred to Dr. Daiello for evaluation of the source of gait instability, she advises an immediate check of renal function. Because gabapentin is primarily excreted by the kidneys, it is “vastly affected by decreases in renal function.” Many geriatric patients may require a lower dose of gabapentin to avoid adverse events.

Dr. Daiello concluded with a set of best practice suggestions for ADR risk reduction in the frail elderly. Team with a geriatric-trained pharmacist, she suggested. In these patients, many or most of whom have comorbid disease, consider the effect of the disease on organ function and the effect of concomitant drug prior to selecting therapy. Set parameters for monitoring of therapeutic and adverse events. Minimize polypharmacy and aim for the minimum effective maintenance dose of medications.

Clinical Applications: A Case Study Report

An audience-interactive case study segment was monitored and chaired by Anton P. Porsteinsson, MD, Assistant Professor of Psychiatry at the University of Rochester School of Medicine.

Case: Polly. Advanced Alzheimer’s disease with verbal agitation
Polly is introduced as an 89-year-old Caucasian woman with advanced Alzheimer’s disease. Her mini-mental score is below 5. She is difficult to medicate and bathe and is socially isolated and keeps to herself. Interaction on a one-on-one basis calms her, but her most troublesome behavioral problem is verbal agitation. She has received haloperidol and various other antipsychotics and has a history of psychosis. (Table 1)

Dr. Porsteinsson commented that this patient reflects the reality in inpatient psychiatric services and nursing homes: advanced dementia, social isolation, limited family involvement. Lack of details about her past history prevents caregivers from having insight into her pre-morbid personality. Her behavioral disturbance has been resistant to treatment.

What percentage of patients like Polly respond to a first medication choice?
In a meta-analysis of studies with conventional antipsychotics (Schneider LS. J Am Geriatrics Soc 1990;38:553-563), about 57% had a favorable response. Based upon all clinical trials, Dr. Porsteinsson said, between 50% and 60% of patients will have a positive response to the initial medication.

If there is no or a suboptimal response to a first agent, which choice is preferred: a different medication in the same class, a medication from a different class, or augment with a different class of medication?

To answer this question, based on trial data, Dr. Porsteinsson pointed to the CATIE study, which suggests selecting a different agent within the same class (atypical antipsychotics) and proceeding to a possible inclusion of an SSRI.

What would be the next step if there were a partial response to an optimum dose: switch to a different medication in the same class; switch to a different medication class; or augment with a different class of medications?

Standard practice in psychiatry would recommend augmentation when there is a partial response with one agent, if the dose of the first medication has been optimized , Dr. Porsteinsson said.

What is the preferred combination strategy for agitation that surfaces in a nursing home patient with dementia: an atypical antipsychotic and an antidepressant; an atypical antipsychotic and an anticonvulsant; an anticonvulsant and an antidepressant or a cholinester-ase inhibitor and an anticonvulsant?

The most common combination therapy in dementia, for any kind of patient, is a cholinesterase inhibitor and an atypical antipsychotic. Another option might be an atypical antipsychotic and an anticonvulsant, but there are minimal data to support this combination.

Dr. Porsteinsson presented a second patient with mild Alzheimer’s disease that is progressing to the moderate stage. His questions focused on eliciting audience views regarding data that some treatments, notably cholinesterase inhibitors, might prevent behavioral disturbances from surfacing if they are initiated early in the course of the disease. The audience response suggested that they would act on these early data in terms of clinical decisions, but, as Dr. Porsteinsson said “..we have a long way to go in convincing a skeptical audience [that prevention of the emergence of behavioral disturbances can actually be achieved.].