 |
 |
|
|
|
Alzheimer’s disease (AD) is the most common type of dementia (1,2) and is characterized by cognitive deficits and behavioural and psychological symptoms. These behavioural and psychological symptoms of dementia (BPSD) include delusions, hallucinations, aggression, aberrant motor behaviour, sleep disruptions, agitation, depression, and apathy (3,4). Research in the pathobiology of AD has revealed a gross disruption of neurotransmitter systems (5), including the cholinergic (4) and serotonergic (6,7) systems in both cortical and subcortical areas of the brain. Although deficits in the cholinergic system have been associated with both cognitive changes (8) and BPSD (4,9), manipulation of the cholinergic system has limited effectiveness. Hence, attention has turned to other possible therapeutic targets for patients with AD. The GABA system is the major inhibitory system in the human brain (10,11). Beyond its acknowledged role in the pathophysiology of epilepsy (12), evidence suggests that GABA may play a supplementary role in other brain diseases by modulating dopamine and serotonin (13–16). GABA’s association with such neuropsychiatric symptoms as anxiety (17), aggression (17–24), and psychosis (17), as well as its ability to regulate acetylcholine, dopamine, and serotonin (25,26), make it a therapeutic target for controlling BPSD. Moreover, potentiating GABAergic inhibition can potentially counteract elevated glutamate excitation and decrease excitotoxicty in cortical circuits (27). We review the evidence supporting the putative role of GABA as a therapeutic target for controlling AD symptoms, emphasizing BPSD. After providing information on the GABAergic pharmacology, we review the evidence for GABA dysfunction among patients with AD and the clinical use of GABAergic agents. Section summaries of GABA Pharmacology and GABAergic Dysfunction in AD, as discussed in this paper, comprise the last paragraph and provide key information for the clinician. We retrieved articles for evidence of GABAergic disruption in AD from various sources, including electronic databases (for example, Medline and Embase) as well as cross-references from relevant articles. The following key words were used: gamma-aminobutyric acid, GABA, GABA receptors, neuropeptides, AD, dementia, behaviour, behavioural disorders, delusions, hallucinations, agitation, aggression, depression, cognition, and noncognitive. For articles on the clinical use of GABAergic agents among patients with AD, we searched each GABAergic agent individually, using the key words Alzheimer’s disease and dementia. GABA PharmacologyGABA is synthesized from glutamate by the enzyme glutamic acid decarboxylase (GAD) and is metabolized by the enzyme GABA-transaminase (GABA-T), whereby GABA released into nerve terminals is converted to succinic acid semi-aldehyde (28). The GABA receptor has been postulated to be a heterooligomeric glycoprotein with 5 or more trans-membrane units. To date, 3 subtypes of GABA-specific receptors have been identified: GABAA, GABAB, and GABAC (29). Both GABAA and GABAB receptors are localized in the central nervous system (CNS). GABAC is specific for the retina (30) and will not be discussed here. The GABAA receptor is the most extensively studied of the 3 receptor types. It comprises a combination of receptor subunits and is a heterooligomeric formation of alpha 1 to 6, beta 1 to 4, and gamma 1 to 4 (30); delta and epsilon (29,31); theta (32); or pi subunits (33). Therefore, an enormous array of subtype combinations may exist, but for a fully functional GABAA receptor, it appears that an alpha, a beta, and 1 other subunit type, such as gamma, delta, or epsilon, are required (29,33). The alpha–beta–gamma subunit containing receptors are categorized as GABAA1 to GABAA6, according to the alpha subunit present (30). Found on both pre- and postsynaptic neurons, the GABAA receptor is a ligand-gated ion channel that is connected to many other neurons, such as dopamine and serotonin. GABAA receptors containing the alpha1 subunit are responsive to all hypnotic drugs (34). When bound by GABA agonists, GABAA receptors increase the flux of chloride ions, which in turn causes hyperpolarization of neurons (26). In this way, GABAA receptors reduce synaptic transmission of dopamine, serotonin, and acetylcholine. In patients with AD, postsynaptic GABAA receptors elevate interstitial GABA caused by the conversion of glutamate to GABA by GAD, resulting in chronic depolarization and neuronal degeneration (35). In addition to a GABA-agonist binding site, the GABAA receptor has sites specific for benzodiazepines, barbiturates, steroids, and ethanol (26). The high-affinity binding of benzodiazepines, such as diazepam, is conferred by the 2 subunit and adjacent alpha1, alpha2, alpha3, or alpha5 subunits (33), with the alpha subunit being a key determinant of the benzodiazepine pharmacology (31,36). Evidence has shown that agonist binding to these other sites enhances GABA’s inhibitory effects. Therefore, any disruption in individual binding sites, or in the receptor as a whole, may also alter the regulatory effect that GABA may have on other neuronal systems. Each GABAA receptor subtype has specific properties, and research is beginning to illuminate the function of the various individual receptor subunits as well (Table 1). Research in this area is becoming increasingly complex because different combinations of receptor units produce receptor subtypes with varying characteristics. Much remains to be learned about the GABAA receptor subtype subunits.
The GABAB receptor is not widely distributed in the brain, compared with GABAA (26). GABAB is linked to G proteins and a second-messenger system that mediates calcium or potassium channels, producing slow inhibitory postsynaptic potentials in several brain regions (29). GABAB receptors are heterooligomeric and are made up of a GABABR1a or GABABR1b subunit and a GABABR2 subunit (29). Although both receptors bind GABA, GABAB is not modulated by benzodiazepines, barbiturates, or steroids, unlike GABAA (29). Instead, it is distinguished from GABAA through selective affinity for baclofen (37). Recent evidence in animal studies suggests that GABAB may be linked to cognition (38). In summary, there are 2 main types of GABA receptors in the CNS: GABAA and GABAB, both of which have many subtypes. The GABAA receptor is both pre- and postsynaptic; is widely distributed; and has modulatory sites specific for benzodiazepines, barbiturates, steroids, and ethanol. GABAB is also found both presynaptically (that is, on autoreceptors and glutamate neurons, which inhibit neurotransmitter release) and postsynaptically (for example, in the hippocampus), which decrease neuronal excitability and may be linked to cognition. Evidence for GABAergic Dysfunction in ADThere are 4 lines of evidence that evaluate the possibility of disruptions in the GABAergic system in AD: postmortem studies, antemortem studies, neuroimaging studies, and markers of CNS GABA. Although riddled with limitations, postmortem studies in AD comprise the bulk of evidence attempting to establish altered GABA in patients with AD. There are 22 studies in the medical literature examining GABA concentration and GABA benzodiazepine binding in AD patients, compared with control subjects (39–60). These postmortem investigations include, first, studies measuring GABA concentrations (39 49,59,60); and second, GABA receptor ligand-binding studies that use GABA-specific ligands (61–64), the GABAA specific agonist muscimol (51,55,65), or benzodiazepines (51–58). Postmortem studies on cortical areas have, for the most part, shown that the frontal (24% to 29%), temporal (19% to 47%), and parietal (21% to 47%) regions may have reduced GABA concentration in AD (Table 2). Binding studies using GABA or benzodiazepines (Table 3) suggest that the temporal region is affected in AD patients. The limbic regions that appear to have reduced GABA in patients with AD include the cingulate (26% to 36%), amygdala (17% to 28%), and thalamus (28% to 36%) (Table 2). GABA in the hippocampus, caudate, putamen, and nucleus accumbens does not appear to be affected by AD, though there have been some positive results found within the hippocampus and caudate (Tables 2 and 3). GAD and GABA-T have also been measured centrally postmortem. GAD is located only within neurons and can therefore be used as a marker for GABA localization in the brain. GABA-T is the enzyme responsible for GABA degradation in the brain. Potential alterations in GABA-T in AD patients may reflect differences in the rate at which GABA undergoes catabolism. Five studies have examined GAD activity in AD patients postmortem (66–70). Despite differing methodologies and design issues, studies suggest decreased GAD activity with compensatory increases in mRNA expression. Three studies that have examined whether GABA-T activity is altered in AD patients show a possible disease-dependent effect of AD on GABA-T activity (71–73). Most postmortem studies are limited by differing postmortem delays, lack of control for AD severity, and concomitant use of psychotropic medications. Overall, the postmortem studies in AD suggest that GABA concentrations are decreased in cortical areas of the brain with decreasing GAD activity. In addition, limbic areas are varyingly affected by a decrease in GAD activity.
Given the limitations of interpreting postmortem tissue analysis, antemortem studies should be of greater value in determining neurotransmitter concentration among patients with AD. Two of the 3 studies are negative (41,74), and the third reports increased GABA in the frontal cortex (42). The ability to draw conclusions from the antemortem data is limited by the small number of patients, the absence of disease severity reports, and confounding by concomitant medications. Imaging techniques, such as positron emission tomography (PET) and single photon emission tomography (SPECT), provide a visual means of identifying cortical and subcortical GABA deficiencies. To date, 5 studies have used neuroimaging to identify potential changes in GABA neurons in AD patients (Table 4) (75–79). There were no changes in flumazenil binding (77,78) with PET, despite demonstrated glucose hypometabolism in the parietal cortex. However, SPECT with 123I-iomazenil (a high-affinity benzodiazepine antagonist) found significant reductions in binding in the parietal cortex (76,79) in AD patients, compared with non-AD patients.
Putative markers of CNS GABA include cerebrospinal fluid (CSF) GABA and plasma GABA. CSF is in direct contact with the brain and, therefore, may contain endogenous chemical constituents found in cortical and subcortical areas. Of 12 published studies (Table 5) (80–91), only 4 found significant reductions (40% to 77%) in CSF GABA among AD patients, compared with control subjects (88–91). Although GABA cannot cross the blood–brain barrier, some investigators have considered whether GABA concentrations are altered in the plasma of patients with AD. Of 3 studies in the medical literature, 1 found AD patients to have a significant reduction in GABA (51%) (80), and 2 (81,82) were unable to detect a significant difference. The validity of these markers remains uncertain.
Of the studies mentioned, 2 described BPSD. Procter and colleagues matched for age, autopsy delay, and brain storage time 17 patients with AD and 18 patients with no dementia (50). Areas of interest included the frontal, temporal, and parietal cortices in both groups. Aggression was the behaviour of interest and was determined qualitatively through retrospective chart reviews and correspondence with caregivers up to 6 months antemortem. The authors reported no significant differences in GABA concentration between AD patients and control subjects with or without behaviours in all brain areas. However, only 1 or 2 patients with aggression for each brain area were analyzed. It must be noted that, owing to the effect of the agonal state, the authors used samples from patients with protracted illness (that is, 7 AD patients and 10 control patients). Wyper and colleagues provided the single imaging study that reported on behaviours among the patients studied. Of the 6 patients, 2 were qualitatively determined to have ideational and dressing apraxis, paranoid delusions, and irritability, respectively (79). Unfortunately, owing to the small sample size, the authors were unable to correlate behaviours with affected regions. In summary, 4 lines of evidence evaluate the possibility of disruptions in the GABAergic system in AD: postmortem studies, antemortem studies, neuroimaging studies, and markers of CNS GABA. Neuroimaging techniques are consistent with postmortem studies in finding large variations in the presence and extent of changes in GABA. In general, more extensive cortical involvement was demonstrated in the postmortem analyses (likely showing predominantly advanced AD), and more limited cortical involvement was demonstrated in the neuroimaging studies (likely showing earlier AD). Involvement of the limbic system was more variable, which is consistent with its role in BPSD. Disruptions in the GABAergic system have not yet been linked with specific BPSD. Clinical Use of GABAergic Agents in Dementia GABA Receptor Ligands Benzodiazepine Receptor Ligands Alprazolam, a short- to intermediate-acting benzodiazepine agonist, has not been extensively studied among the psychogeriatric population. In a randomized trial, Ancill and colleagues (93) compared alprazolam to lorazepam in dementia patients with agitation. The authors did not find significant differences in the 2 medication groups with respect to efficacy, but they did report that lorazepam was associated with more adverse events. In a multicentre, randomized, controlled crossover trial, Christensen and colleagues compared alprazolam to haloperidol in treating agitation and behavioural disorders associated with dementia (94). Forty-eight patients who completed the full protocol were found to be no different, whether they received alprazolam or haloperidol, regarding the number of behavioural episodes, activities of daily living, or clinical global impression (CGI). These results did not change when patients were stratified for severity of cognitive impairment and behavioural episodes. Although the number of dropouts in each group did not differ, one-third of the participants did not complete the full protocol. Clonazepam is a high-potency benzodiazepine with specific affinity for the benzodiazepine receptor. To date, there are no randomized controlled trials (RCTs) examining clonazepam effectiveness in treating BPSD. Smeraski reported on a case wherein clonazepam was used in conjunction with thioridazine to treat agitation, hostility, and combativeness in a 63-year-old inpatient diagnosed with multiple-infarct dementia (95). Within 24 hours of clonazepam administration, the patient had improvement in agitation and combativeness, which was maintained on 4 mg daily of clonazepam. In another case report by Freinhar and Alvarez, the authors found successful treatment of violent behaviour and agitation in 2 elderly patients with organic brain syndromes (96). Ginsburg reported on the cases of 6 consecutive AD patients presenting with agitated and aggressive behaviour who were treated with low-dosage clonazepam (97). Four of the 6 patients showed improvement in behaviour, and sedation was the most common dose-limiting adverse event. In a prospective observational cohort study, Calkin and colleagues identified 24 inpatients treated with clonazepam (mean dosage 1.2 mg) over a 21-month evaluation period to determine whether the drug improved behaviours and was associated with greater severity of adverse events (98). Patients diagnosed with dementia (n = 16) showed no significant improvement in behaviours (as measured by the Brief Psychiatric Rating Scale [BPRS] and the Cohen–Mansfield Agitation Inventory), but were found to have less frequent side effects and showed global improvements. Importantly, clonazepam was not associated with significant decline in cognitive functions, as determined by Mini-Mental State Exam (MMSE) scores. Because of the various active metabolites produced, diazepam is a long-acting benzodiazepine. Only 2 clinical trials have been conducted using diazepam to treat BPSD. In both reports, diazepam was compared with thioridazine in a double-blind, randomized fashion (99,100). Both investigators reported significant improvement in behavioural symptoms for both agents, though thioridazine produced slightly greater improvement and diazepam was found to produce tolerance over time. Lorazepam is a commonly used benzodiazepine among the geriatric population (101). Fritz and Stewart reported on the successful use of lorazepam for the treatment of resistive aggression in 2 patients diagnosed with dementia (102). Except for the aforementioned study comparing lorazepam with alprazolam (93), only 2 other studies exist in the medical literature. In a randomized, placebo-controlled trial examining the effectiveness of lorazepam in managing chronic behaviours among patients with probable mild-to-moderate AD (n = 10), the AD patients showed a significant trend toward BPSD improvement with lorazepam treatment, compared with age-matched subjects with no AD or BPSD (n = 10) (103). Meehan and colleagues reported a significant improvement in acute BPSD with intramuscular lorazepam treatment, compared with those treated with placebo in a double-blind, randomized study. This effect became apparent 60 minutes following lorazepam administration and was maintained for 2 hours. After 24 hours, the change in the lorazepam group was no longer significant (104). The only significant adverse events noted with lorazepam use among AD patients were drowsiness and dry mouth. Oxazepam, a short-acting active metabolite of diazepam, has been used in several investigations to control BPSD. The earliest studies were published in the mid-1960s by 3 independent authors examining the effectiveness of oxazepam, compared with placebo, in over 300 elderly patients (105–107). The authors found oxazepam to be superior to placebo; however, such study limitations as lack of diagnostic criteria and lack of qualitative efficacy reports make it difficult to draw definitive conclusions. In an 8-week RCT, Coccaro and colleagues compared oxazepam (n = 19, mean dosage 30 mg) with haloperidol (n = 20, mean dosage 1.5 mg) and diphenhydramine (n = 20, mean dosage 81.3 mg) among patients diagnosed with dementia (108). Although all 3 groups showed significant improvements in BPSD and activities of daily living, both haloperidol and diphenhydramine had a greater nonsignificant trend toward improvement in BPSD, compared with oxazepam. The study was limited by the absence of a placebo-control group and enrollment of patients with moderate-to-severe cognitive impairment. Herz and others published the only other study using oxazepam to control behaviours in patients with dementia (109). The authors used a single-case study approach to determine the efficacy of various agents in managing resistive behaviours in patients diagnosed with probable AD. Ten patients were randomly administered each of the following 4 trial medications in double-blind fashion for 19 weeks: 10 mg of oxazepam 3 times daily, 1 mg of thiothixene 3 times daily, 25 mg of diphenhydramine 3 times daily, and placebo. Ratings of effectiveness in managing resistance determined whether a patient was switched to a new medication. Of the patients who showed improvement in resistive score, thiothixene was effective in 5 patients, oxazepam in 2 patients, and diphenhydramine in 1 patient. The study was limited by the fact that the treatment strategy did not consider changes in the patients’ disease over time and did not examine the effect of the medications on potential concurrent behaviours. Zolpidem, a nonbenzodiazepine hypnotic, binds specifically to the benzodiazepine alpha1 receptor and, therefore, has anxiolytic properties. Shelton and Hocking (110) reported on 2 cases of patients diagnosed with AD who were treated with zolpidem for severe sleep disturbance and nighttime wandering. Doble reported that receptors containing the alpha1 subunit only responded to zolpidem and that any other alpha subtypes do not (34). Both patients, who were previously nonresponsive to antipsychotics, showed positive improvements in sleep promotion without untoward effects. Shelton and colleagues reported on zolpidem being used to restore normal sleep patterns among patients with dementia, and slight behavioural improvement was also noted (110). No other data have been published regarding zolpidem use for managing BPSD in patients with AD. In summary, though many of the benzodiazepines lack controlled clinical trials, from the evidence to date, benzodiazepines seem to be useful in managing BPSD. The primary concern with these agents is the potential for adverse events, which may be augmented in an elderly population and are owing to pharmacologic changes. A recent study by Fastbom and colleagues suggests that benzodiazepines may have | |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||