NeuroRehabilitation ELSEVIER

NeuroRehabilitation 5 (1995) 219-232

Current neuropharmacologic interventions for the management of brain injury agitation Todd Rowland a , Lawrence DePalmab a Department of Physical Medicine

and Rehabilitation, Rehabilitation and Research Center, Medical College of Virginia, Box 980661, Richmond, VA 23098-0661, USA bDepartment of Psychiatry, University ofAlabama, Birmingham, AL, USA Accepted 17 April 1995

Abstract

Agitation following brain injury is a significant and difficult problem; severe agitiation is most effectively treated by concurrent environmental, behavioral, and pharmacologic interventions. Delirium and agitation are briefly compared, however, a more thorough discussion of this topic appears in other articles within this issue. This article reviews the current literature in regard to practical pharmacologic interventions for agitation following brain injury and outlines short-term and long-term strategies. Common and serious side effects, as well as unique characteristics for each medication are highlighted. Keywords: Agitation; Brain injury; Delerium; Pharmacology; Management

1. Introduction This article focuses on perhaps more commonly used pharmacology and takes a practical approach to use of these agents. The review references well-documented sources and controlled studies when available in preference to anecdotal reports or single-case studies. Please also see the Cardenas and Bell article in this issue. It is extremely important to use the information within the context of non-pharmacologic approaches. 2. Differences between agitation and delirium Agitation as a clinical construct implies an increase in overall psychomotor activity that is in

some way inappropriate to the environment. Thus many different types of behavior may be labeled 'agitated,' regardless of the underlying etiology for such behavior. The schizophrenic patient pacing the floor under the influence of auditory hallucinations, the cocaine-intoxicated individual loudly threatening emergency room staff, the brain-injured person repeatedly telephoning relatives at frequent intervals, the elderly person with Alzheimer's dementia restlessly wandering at night; all of these may be described accurately as agitated despite the various mechanisms producing increased psychomotor activity in each case. Agitation is also frequently encountered in the setting of acute injury or medical illness, where it can be linked to global derangements of cerebral

1053-8135/95/$09.50 © 1995 Elsevier Science Ireland Ltd. All rights reserved. SSDII053-8135(95)OOI20-W

220

T. Rowland, L DePalma / NeuroRehavilitation 5 (1995) 219-232

functioning secondary to the underlying illness. In such cases, a diagnosis of delirium is made. The core feature of delirium is a clouding of consciousness, as manifested by impaired ability to focus, sustain, and shift attention, which characteristically follows a waxing and waning course [1]. Delirium is accompanied by disruption of the normal sleep-wake cycle as well as by alterations of psychomotor behavior [2,3]. This last feature is a source of frequent confusion to the non-psychiatric clinician: delirium may be manifested by either increased or decreased psychomotor activity (and often both in the same individual at different times). Thus the terms agitation and delirium are not synonymous; agitation is one possible manifestation of delirium, but as pointed out above, agitation frequently exists independently of delirium, so the delirious patient need not be agitated (quiet delirium). The etiology of delirium remains poorly understood, despite its frequent occurrence and significant prognostic implications [4]. There is a general consensus that delirium (or encephalopathy in the neurologic literature) represents global derangement of cerebral functioning at multiple levels. This derangement is reflected in the EEG by a pattern of generalized slowing of brain rhythms, and in functional brain imaging studies by generalized reduction of cerebral oxidative metabolism [2,5]. There are likely multiple perturbations of neurotransmitter balance occurring during delirium, though whether these are etiologically significant in themselves, or represent the measurable effects of some underlying causative process remains an open question. The relative balance of excitatory and inhibitory neurotransmitters has been a principal area of investigation, largely due to the clinical actions of pharmacologic agents. The observation that toxic doses of highly anticholinergic agents (e.g. atropine) induce a characteristic delirium, which may then be reversed by administering cholinesterase inhibitors (e.g. physostigmine), leads to the hypothesis that delirium represents a generalized reduction of cholinergic activity [6]. The effectiveness of dopamine-blocking agents (e.g. haloperidol) in reducing or eliminating manifestations of delirium (including anticholinergic

delirium) suggests that it is the relative balance of cholinergic and dopaminergic activity that is critical. Similarly, excessive activity of the excitatory neurotransmitter glutamate is postulated as etiologic in some forms of delirium, particularly that seen immediately following brain parenchymal damage by ischemia or trauma, based on the effectiveness of glutamate-blocking agents. But the effectiveness of GABA-enhancing agents (e.g. benzodiazepines) in such delirious states highlights the importance of the relative balance between the excitatory effects of glutamate and the inhibitory effects of GABA. As outlined in the discussion above, delirium manifests itself clinically as a spectrum of altered psychomotor behavior, ranging from quiet obtundation to severe agitation. While agitation is frequently encountered in the setting of delirium, it also exists independently of the fluctuating clouding of consciousness which is the core feature of delirium. A common clinical example of ongoing agitation occurring apart from a grossly altered sensorium is the excessive and inappropriate psychomotor activity of individuals with traumatic brain injury (TBI) many months following the injury, when attention, concentration, and memory deficits have stabilized. We suggest that such agitation may be qualitatively different from the agitation seen in the setting of acute delirium. As will be discussed later, there are clinical trials reporting efficacy for strongly anticholinergic agents in the management of agitation in TBI, a finding which runs counter to observations that anticholinergic agents aggravate delirium (see the article of Sandel et al.). It is imperative to evaluate and treat reversible etiologies of delirium other than brain injury prior to neuropharmacologic interventions. Please see the article of Mysiw et al. discussing the differential diagnosis of agitation. 3. Approaches to studying agitation The ideal method to study neuropharmacologic interventions for agitation would entail doubleblinded, placebo-controlled medication trials for rigorously defined target behavior (agitation) with concurrent use of both the prospective agitation

T. Rowland, L. DePalma / NeuroRehabilitation5 (1995) 219-232

monitoring system [7] and the prospective cognition/orientation monitoring system [8,9]. Unfortunately, such a study does not exist. The bulk of reports for neuropharmacologic interventions specific to agitation are case reports. In consideration of this situation there is limited data upon which to base firm recommendations. In addition to literature specific to brain injury patients, there is data relevant to analogous patient populations such as ADHD and organic brain syndrome [10]. 4. Strategies for short-term management of agitation (and breakthrough agitation) In discussing management of agitation it is useful to consider short-term and long-term management strategies separately. The goal of acute management of agitation is to gain immediate control of excessive psychomotor activity which represents a danger to the patient and others. A combination of environmental and pharmacologic restraint is typically employed to rapidly minimize the dangerous behavior, with the overriding concern being the immediate safety of the patient and others. In the rehabilitation setting, such measures obviously interfere with participation in therapies, and sedating agents directly impair cognition, so once acute agitation has been brought under control, the focus shifts to longterm management strategies for the agitated patient which will permit rehabilitation to proceed. The goal of long-term management of agitation is to achieve a balance between ongoing reduction of excessive psychomotor activity and possible deleterious side effects of the agents used. As with short-term management, a combination of environmental and pharmacologic techniques are used together, though in a more individually tailored fashion, often employing quite specific behavioral modification regimens. This last point is one that must be emphasized: though we focus here on pharmacologic management strategies, these are rarely successful in the absence of a coordinated behavioral management plan devised by the multidisciplinary treatment team. The short-term management strategies described here are intended to be used as tem-

221

porary measures to gain rapid control of dangerously agitated patients. Thus, they may be employed in the setting of acute illness or injury, or for control of breakthrough agitation in the chronic setting, when long-term agents are already in use. The division of agents into the short and long-term categories presented here is certainly not intended to be absolute, but is a convenient format for discussion; there are roles for more extended use of selected agents described for short-term management of agitation, and some of the long-term agents presented are useful as well in acute management. Analogous to the pharmacologic control of seizures, monotherapy is preferred in order to improve compliance and reduce potential side-effects of treatment, however, in practice it is often necessary to employ a combination of several agents to achieve satisfactory control of agitation. In addition, the general principle of 'start low and go slow' developed in geriatric medicine is applicable as well to patients with traumatic brain injury, since they tend to be more sensitive, both to beneficial and adverse effects of centrally-acting medications [12,13)' 4.1. Benzodiazepines (BZDs)

The benzodiazepines (BZDs) are a large class of structurally similar agents, varying widely in half-life and potency, which all act pharmacologically by enhancing central GABA activity (recall that GABA is the most abundant eNS inhibitory neurotransmitter). The BZDs bind directly to all subtypes of the GABA receptor complex at a unique site and cause protein conformational changes which facilitate GABA binding, thus leading to increased channel-open frequency. The resultant overall increase in central inhibitory tone produces the well-characterized clinical effects of anxiolysis, skeletal muscle relaxation and sedation [14,15]. The BZDs are frequently the first-choice agent for acute control of agitation in the emergency department and intensive-care settings, largely due to their safety profile and rapid onset of action. The BZDs have also been shown useful in both animal and human studies in long-term reduction of aggressive behavior, though the human studies have been performed primarily in demen-

222

T. Rowland, L. DePalma / NeuroRehavilitation 5 (1995) 219-232

tia patients [10,11]. Short-acting BZDs may be preferable in many situations of acute agitation, particularly when the etiology of the agitation is poorly understood and rapid discontinuation of the drug may be necessary. In general, the shortacting BZDs also have the most rapid onset of action, making them especially useful when rapid control of dangerous behavior is required. On the neurotrauma rehabilitation unit, shortacting BZDs have a discrete utility in the shortterm management of acute agitation, but the same features which make these agents so useful also lead to a number of distinct drawbacks in the TBI patient. Of foremost concern from the standpoint of rehabilitation are the well-characterized adverse effects of BZDs on cognition, chiefiy learning and memory [16,17]. The amnestic effects of the short-acting BZDs are used to advantage in minor surgical procedures with conscious sedation, but only add to the difficulties caused by memory impairment in the TBI patient. Also of concern, the BZDs may give rise to paradoxical agitation, particularly in the elderly, which can lead to an escalating cycle of increasing doses of the agent which is worsening the agitation. On the other extreme, oversedation to the point of obtundation is clearly not beneficial for ongoing rehabilitation. Agents which potentiate GABAergic activity have been shown to delay sensorimotor neural recovery in animal models following experimental brain injury, and there is retrospective human data showing a similar delay in motor recovery following stroke [18,19]. There is currently no human data on the impact of BZDs on cognitive recovery following TBI, though extrapolating from available data with normal volunteers it seems evident that BZDs should be used with caution in TBI patients. As with all CNS depressants, respiratory depression is a potential life-threatening adverse effect of BZDs, though it is usually severe only when other CNS depressants are used concomitantly. With the recent introduction of fiumazenil, a BZD antagonist, concern over possible central respiratory depression has become less critical. Finally, BZDs produce physiologic tolerance when used continuously, which often limits their effectiveness and

necessitates careful discontinuation protocols to avoid withdrawal syndromes, including seizures. In our experience, lorazepam is a useful agent for control of acute agitation in the neurotrauma rehabilitation setting. It has a relatively short half-life (10-20 h) and rapid onset of action (within 30 min with p.o. dosing, much less with i.m. or i.v. administration). Like all BZDs, lorazepam is hepatically metabolized, but it is one of the few BZDs with no active metabolites, and thus is tolerated well by medically ill and elderly patients [20]. We recommend doses of 0.5-1 mg q 8 h, titrating upwards to maximal doses of 8 to 12 mg per 24 h. For rapid sedation, lorazepam may be given 1-2 mg q 1 h until adequate sedation is achieved, though we have rarely found this necessary. Generally we start administering agents for long-term management of agitation concomitantly and the lorazepam is withdrawn as soon as possible in order to minimize its impact on cognitive recovery. Other benzodiazepines may be useful but are not available in p.o., i.m. and i.v. forms as is lorazepam. Clonazepam is a long-acting benzodiazepine (half-life of approximately 30 h) which may be useful in the long-term management of agitation, but there are, to date, no controlled studies of this agent's impact on agitation. 4.2. Neuro[eptics The traditional neuroleptics, also known as antipsychotics, comprise a structurally diverse class of agents which share the common pharmacologic action of blockade of the central dopamine type 2 receptor (D2), which is hypothesized to be the basis for their primary clinical effects [21]. A pharmacologically distinct class of recently-developed neuroleptics, termed the atypical neuroleptics, act at a different subset of dopamine receptors and at serotonin receptors (see the article of Bell and Cardenas). While a complete discussion of neuroleptic pharmacology is beyond the scope of this article, rational use of neuroleptics requires a basic understanding of the differences in nondopaminergic blockade among the various agents. The traditional neuroleptics may be all placed on

T. Rowland, L. DePalma / NeuroRehabilitation5 (J995) 219-232

a continuum between high-potency agents and low-potency agents; haloperidol, the prototypical high-potency neuroleptic, is a relatively 'clean' dopamine-blocker, having little affinity for other CNS receptors, while chlorpromazine, the prototypical low-potency neuroleptic, is a rather 'dirty' dopamine-blocker, having strong affinities for a variety of other CNS receptors, particularly cholinergic receptors. The other neuroleptics fall in intermediate positions between these two prototypical agents [22]. The clinical significance of this distinction is that one can tailor neuroleptic treatment based on the desired side-effect profile of a specific agent using receptor affinity tables. For example, rapid sedation of a highly agitated psychotic patient may be accomplished by i.m. chlorpromazine by virtue not only of its dopamine-blocking effects, but also its anticholinergic and antihistaminic effects. The trade-off in this case is that central alpha receptor blockade by chlorpromazine can produce dangerous hypotension, particularly in the elderly or medically ill. Since dopamine-blockade alone is sufficient to lyse agitation in most cases, the clinically preferred agent for control of acute agitation in the medical setting is a high-potency neuroleptic such as haloperidol. In very severe cases of agitation, haloperidol is used in an alternating fashion with the short-acting BZD lorazepam [23]. This concomitant use of a neuroleptic and a benzodiazepine is a rational choice in the management of acute agitation in that it synergistically combines pharmacologically-distinct mechanisms of action. The use of neuroleptics in the context of traumatic brain injury remains a rather controversial subject, perhaps due in part to the overly-extensive use of these agents following their initial introduction. There is substantial evidence in the psychiatric literature documenting the effectiveness of neuroleptics in both the short and longterm management of agitation and aggressive behavior [21]. Of concern to the rehabilitation professional, however, is laboratory evidence that neuroleptics may compromise cognitive recovery following traumatic brain injury. This concern is founded on the 1982 study of Feeney et al. in which rodents with experimental head injuries,

223

placed on haloperidol, demonstrated poorer relearning of a motor task when compared to nontreated controls [24]. This data has not since been replicated, and there are no similar controlled studies in primates or humans, so these conclusions must be interpreted with caution. There are a number of studies in humans, however, documenting a variety of neuroleptic-induced cognitive-behavioral changes, including anhedonia [25], dysphoria [26], and cognitive/motor impairment [16]. Like the benzodiazepines, neuroleptics clearly have the potential to impair cognitive functioning in the normal individual, but there are no studies documenting risk/benefit ratios when these agents are used for the short-term management of severe agitation in TBI patients. It may well be that allowing severe agitation to disrupt the rehabilitation process impairs cognitive recovery to a greater extent than any intrinsic effect of neuroleptics. This is clearly a question of clinical significance which requires further study. What then are the clinical indications for use of neuroleptics in the neurotrauma rehabilitation setting? The most obvious indication is agitation due to or accompanied by psychotic symptoms such as delusions, paranoia, or hallucinations. High-potency neuroleptics rapidly abolish such psychotic phenomena with little to no sedation when judiciously titrated. Extreme aggression or rage is another indication for initiation of neuroleptic treatment, although one must take care to rule-out episodic rage due to complex seizures, which is more appropriately treated with an anticonvulsant. Mania following TBI, particularly when grandiose delusions are present, is appropriately treated by initiation of a neuroleptic in conjunction with an anti-manic agent such as lithium. The neuroleptic is then withdrawn as elevated or the irritable mood subsides in response to the anti-manic agent. Finally, the occurrence of Tourette's syndrome, either pre-existing or new-onset (rare), in the rehabilitation setting is appropriately managed using high-potency neuroleptics, especially pimozide and haloperidol [27]. The use of neuroleptics in the neurotrauma rehabilitation setting is guided by several fairly simple principles: as with any centrally-acting agent in the brain-injured patient, the minimal

224

T Rowland, L. DePalma / NeuroRehauilitation 5 (1995) 219-232

effective dose of neuroleptic can be determined by starting at a low dose, titrating upwards slowly, and clearly documenting response to treatment, preferably using a standardized scale [12,13]. If rapid increases in neuroleptic doses are required to gain control of severe agitation or psychosis, then a minimal baseline standing dose can be ordered to insure that some neuroleptic is administered at predictable intervals, and PRN doses used to determine total daily neuroleptic requirement. This total dosage can then be divided and given on a standing basis, with tapering of the total dose starting soon after control is achieved. This last point is crucial in the rehabilitation setting if possible cognitive side effects are to be minimized; neuroleptics should not be considered long-term maintenance agents for patients with brain injury unless there are ongoing psychotic symptoms which necessitate their use [12]. The common adverse effects of neuroleptics are directly related to their pharmacologic action, especially as they affect the central balance of dopaminergic and cholinergic activity. Parkinsonian symptoms of bradykinesia and cogwheeling rigidity, also commonly known as extrapyramidal symptoms (EPS), are due to dopamine blockade in the basal ganglia, and are a reliable indicator that the drug has crossed the blood-brain barrier and is acting centrally. High-potency neuroleptics such as haloperidol are more likely to cause EPS than low-potency neuroleptics, as they have little intrinsic anticholinergic activity to counterbalance the reduction in dopaminergic neurotransmission. More dramatic forms of EPS, the acute dystonias, arise early in neuroleptic therapy and can present as truncal extensor spasm (opisthotonos), oro-lingual spasm and oculo-gyric crisis. The acute dystoni as, while frightening in appearance, are rapidly resolved by the administration of an anticholinergic agent such as benztropine. Tardive dyskinesias, a late complication of neuroleptic treatment, present as involuntary choreoathetoid movements typically beginning with oro-lingualbuccal dyskinesias and then progressing from the distal to proximal extremities to involve axial musculature in severe cases. There is no entirely effective treatment for tardive dyskinesia, although the atypical antipsychotics can often re-

duce or abolish the abnormal movements [28,29]. Finally, a very common side effect of neuroleptics is akathisia, an extremely uncomfortable internal sensation that one must move about. The patient with akathisia appears restless and agitated, which often leads to the erroneous conclusion that more neuroleptic is required. If escalating doses of a neuroleptic seem to be worsening agitation, suspect akathisia, and reduce rather than increase the dose. Concomitant use of a J3-blocker or benzodiazepine will often prove effective in alleviating akathisia if symptoms do not permit reduction of the neuroleptic dosage [30]. In our experience, haloperidol is a reliable neuroleptic agent for control of acute agitation with psychosis in the neurotrauma rehabilitation setting. It may be delivered p.o., i.m., or Lv. and has a rapid onset of action (within minutes when given Lv.). Like all the neuroleptics, haloperidol is hepatically metabolized and highly protein-bound, with a serum half-life ranging from 12 to 24 h [15]. We usually initiate haloperidol at standing doses of 0.5-1 mg q 12 h, with 0.5-1 mg p.r.n. q 6-8 h. Rarely has any patient in our experience required more than 10 mg haloperidol in 24 h to control agitation due to psychotic symptoms. We begin tapering the total haloperidol dose within several days after symptoms have resolved, leaving p.r.n. doses for breakthrough symptoms, and nearly all patients are off haloperidol at the time of discharge. 4.3. Antihistamines Antihistamines, such as diphenhydramine hydrochloride, are used in some centers for shortterm control of agitation. Diphenhydramine's central effects are mediated through blockade of central HI receptors and inhibition of histamine N-methyltransferase. Antagonism of other central nervous system receptor sites, such as those for serotonin, acetylcholine, and a-adrenergic stimulation may also be involved in its sedative /hypnotic effects [31]. There are no studies of the efficacy of diphenhydramine for control of agitation. Diphenhydramine hydrochloride has a peak plasma concentration of 2-3 h and a plasma t1/2 of 4-6 h. We recommend an initial dose of 25 mg p.o./Lm. q 12 h, with maximal dosing up to 50 mg

T. Rowland, L. DePalma / NeuroRehabilitation5 (1995) 219-232

p.o.ji.m. q 6 h. In consideration of experimental evidence of a hypocholinergic etiology for delirium [2,3], it is difficult to strongly recommend extensive use of diphenhydramine for control of agitation. 4.4. Barbiturates

The barbiturates represent a large class of structurally similar agents, varying widely in halflife and potency, which all act pharmacologically by enhancing central GABA activity. Their mechanism of action is distinct from that of the benzodiazepines, in that the barbiturates bind to a different site on the GABA receptor complex and prolong chloride channel-open time, rather than increasing channel-open frequency. The resultant effect is similar to that of the BZDs, with an overall increase in central inhibitory tone producing anxiolysis, skeletal muscle relaxation and sedation [15]. The clinical use of barbiturates has been largely superseded by the safer BZDs, except in the instance of phenobarbital used as an anticonvulsant, and the use of ultra-short-acting barbiturates in anesthesia. The drawbacks of the barbiturates are many: they share all the adverse effects of BZDs on cognition, plus they have a very low therapeutic index (lethal doses are often less than three times therapeutic doses); greater respiratory depressant effects than BZDs; more rapid development of tolerance than BZDs; and a host of drug-drug interactions secondary to hepatic microsomal enzyme induction. Despite these overall drawbacks, phenobarbital may play a useful role in the TBI patient with seizures and agitation; however, phenobarbital has been frequently found to have the least favorable cognitive side effect profile of commonly used anticonvulsants [32,33]. 4.5. Chloral derivative hypnotic agents

Chloral derivative hypnotic agents, as typified by chloral hydrate (the active ingredient in the legendary 'Mickey Finn'), act as CNS depressants by enhancing central GABA activity through poorly-understood mechanisms. While these agents have been largely superseded by safer and more effective hypnotics, they remain clinically

225

useful for rapid, short-term sedation. All the chloral derivative hypnotics are hepatically reduced to the pharmacologically active intermediate trichloroethanol within minutes of absorption. The serum half-life of trichloroethanol is between 4 and 12 h. Chloral hydrate is generally tolerated well by the elderly when used as a hypnotic, though its unpleasant taste and tendency to produce gastric irritation often limit its use [15]. 5. Strategies for long-term management of agitation Long-term, management of agitation is focused upon optimizing monotherapy if possible. Often, several potentially beneficial agents are used in combination at a suboptimal dosage. It is extremely difficult to determine the efficacy of any single agent in this setting. We strongly recommend a single case study approach with each long-term agent. 5.1. Beta blockers

Propranolol is one of the better studied agents for the treatment of brain injury agitation [34]. Brooke et al. [34] prospectively studied 25 severely brain-injured patients who displayed agitated behavior and used a placebo control design to evaluate the efficacy of propranolol. They found that the treatment group displayed lower Overt Aggression Scale (OAS) scores over a 7-week period as compared to the placebo group; there was no difference in the number of episodes between the two groups. The limitations of this study include the use of OAS, small number of patients, the limited scope of the definition of agitation, and its descriptive nature. The OAS was initially designed to measure aggression in a psychiatric population [35] and was not specifically designed to measure agitation - it's reliability and validity in this regard are unproven. Initial doses of propranolol LA in the Brooke study were 60 mg; a well tolerated alternative is 20 mg p.o. TID, although the LA form is often not completely equivalent and retitration may be necessary [36]. Peak serum levels with propranolol LA occur at about 6 h and serum t1/2 is 10 h; the plasma concentration is more unpre-

226

T. Rowland, L. DePalma / NeuroRehavilitation 5 (1995) 219-232

dictable with the LA formulation than with divided doses of propranolol. The peripheral sympathetic blockade side effects (hypotension, bradycardia) are more minimal after 120 mg/ day dosage is reached, as the majority of peripheral receptors are blocked, although sympathetic tone varies widely between individuals [36]. Titration of dosage is empiric and rate of titration is based primarily upon toleration of side effects. Significant side effects include hypotension, bradycardia, depression, sexual dysfunction and sedation. Propranolol is contraindicated with congestive heart failure and relatively contraindicated in the setting of IDDM, as it blunts the patient's ability to perceive hypoglycemic episodes. Brooke's protocol [34] is a rational approach to the clinical use of propranolol: initial dosing of 60 mg propranolol LA and increasing by 60 mg every third day to a maximal dose of 420 mg/day. It is certain other investigators report patients tolerating higher dosages, up to 600 mg/day [37]. It is unclear if other f3-blockers are useful in the control of brain injury agitation; propranolol is certainly the most lipophilic and therefore more readily crosses the blood-brain barrier. Propranolol, particularly the long-acting formulation (LA) may have a role in treatment of episodic dyscontrol or agitation, which occurs episodically during the period when attention, concentration and memory deficits have stabilized [10,27,38].

5.2. Tricyclics Tricyclic antidepressants (TCAs) may be useful in the long-term control of brain injury agitation; animal studies provide evidence of the roles of serotonin (5HT), norepinephrine (NE), and acetylcholine (ACH) in aggression [38,39,40] and certainly TCAs affect these receptors to various degrees. Amitriptyline is the only antidepressant that has been prospectively studied for the control of brain injury agitation. Mysiw et al., 1988 [41] found that 13 of 20 severely brain-injured patients who were significantly agitated appeared to have a decrease in their agitation while receiving amitriptyline, six of 20 showed no improvement, and one patient's mental status deteriorated during the study. In particular, Mysiw

noted that a higher percentage of RLAS IV patients responded to the amitriptyline than did the higher-level, RLAS VI, patients. The strengths of this study include its prospective design, concurrent use of OGMS [8], and exclusion of other etiologies of delirium; however, no objective measure of agitation was utilized and it was not placebo controlled. This study is intriguing in light of the proposed hypocholinergic etiology of delirium; amitriptyline is the most anticholinergic of the TCAs [31] and is often implicated in the exacerbation of delirium with hospitalized patients. These results suggest a difference between experimentally induced delirium and brain injury agitation. Amitriptyline and TCAs have a serum t1/2 of 10-30 h and take about 7 days to reach steady state. Mysiw used up to 125 mg/day of amitriptyline for the agitated patients and up to 150 mg/ day is utilized in the treatment of depression. We recommend initial use of 25 mg p.o. q hand increases of 25 mg/day until total dosage of 100 mg/ day is reached; measurement of plasma levels is not recommended to determine efficacy, but can be useful in suspected cases of toxicity. The majority of amitriptyline's side effects are anticholinergic in nature, these include: dry mouth, dry eyes, sedation, constipation and urinary retention. In addition to its anticholinergic side effects, amitriptyline is a potent sympathetic (alpha) blocker and antihistamine. The alpha blockade can cause orthostatic hypotension, while the antihistaminergic activity contributes to sedation. Amitriptyline's anticholinergic side effects are additive with other medications, and amitriptyline should be given with caution to patients receiving antihistamines (diphenhydramine). TCAs, in particular amitriptyline, are relatively contraindicated in patients with CAD and history of cardiac arrthymias. Secondary amine TCAs, such as nortriptyline may be used in the treatment of agitation but are unstudied.

5.3. Novel antidepressants Certain novel antidepressants, such as trazodone, appear logical in the treatment of brain injury agitation. Trazodone has very little anticholinergic activity and its mechanism appears to

T. Rowland, L. DePalma / NeuroRehabilitation 5 (1995) 219-232

be largely related to its effect on serotonin; it has been shown to selectively block the reuptake of serotonin at the presynaptic neuronal membrane and thus the effects of serotonin are potentiated [42-44]. In addition to its serotonergic activity, trazodone has central alpha1-adrenergic blocking activity and some histamine blocking activity [42]. Trazodone for control of brain injury agitation has been retrospectively studied and appeared to improve significant agitation in six of ten severely brain-injured patients; one of the ten experienced a mental status decline during the trial of trazodone [45]. This case series sample size precluded appropriate statistical analysis; however, this series is intriguing and suggests further prospective evaluation of trazodone for brain injury agitation may be fruitful. Trazodone is often used for agitated brain injury patients who display a disrupted sleep-wake cycle; this is quite logical given the primary mechanism of serotonin reuptake blockade. In the setting of agitation with sleep-wake cycle disruption, we recommend trazodone 50 mg p.o. q h with increases of 50 mg up to a total dose of 400 mg/ day (divided doses) as necessary. A number of case reports detail the successful treatment of agitation with organic brain syndrome with trazodone and L-tryptophan [46-48]. Trazodone is well tolerated and its most common side effects are sedation and mild hypotension. Priapism is a relatively uncommon, but serious side effect (1:6000), and can require surgical treatment [49]. In the majority of patients who experience priapism, discontinuing trazodone is sufficient. 5.4. Lithium Lithium, a simple monovalent cation, has been widely and effectively used for the treatment of bipolar affective disorder since its general clinical introduction in the early 1970s, and has shown promise for the management of aggression and disorders of impulse-control [21,38]. Lithium acts pharmacologically by inhibiting phosphatidylinositol second-messenger systems in a variety of neurons, although its exact mechanism of mood stabilization remains unclear [50]. Lithium is clinically indicated for use in bipolar and unipolar affective disorders, disorders of

227

episodic dyscontrol, and syndromes of aggression and impulsivity arising from neurologic damage, whether congenital or acquired [21]. While there is a wealth of anecdotal evidence for the efficacy of lithium in the short and long-term management of agitation and aggressive behavior, there is a paucity of well-designed studies. The available experimental data on lithium efficacy in agitation have been obtained in mentally retarded populations in two open and three blinded, placebo-controlled trials: response rates varied from 50 to 89% (in the largest study, with n = 42, response rate to lithium was 73% vs. 30% in controls) [10]. There are no controlled population studies of lithium use for agitation in traumatic brain injury, but at least one single-case design has shown dramatic efficacy in this setting: Haas and colleagues described the use of lithium in a severely agitated TBI patient who had been unresponsive to trials of multiple other medications, including haloperidol, thioridazine, diazepam, propranolol, chloral hydrate, benztropine, and methylphenidate [51]. Within 48 h after introduction of lithium, agitation that had persisted for the previous 26 days rapidly resolved. Lithium is administered enterally as lithium carbonate, a solid, or lithium citrate, a liquid preparation, but cannot be administered i.m. or i.v. Sustained-release oral preparations are available and tend to minimize gastric irritation. Lithium is rapidly absorbed from the GI tract, and has a high volume of distribution. The serum peak occurs 0.5-2 h after ingestion, or about 4 h for sustained-release preparations. Lithium is unique among psychotropic agents in that its metabolism is entirely renal, with a half-life of approximately 20 h in those with normal renal function. Lithium is effectively cleared by dialysis, which becomes clinically significant in situations of toxicity and in patients with minimal renal function [15,21]. The adverse effects of lithium which are most troubling to patients are gastrointestinal, neurologic, and dermatologic, but lithium has a wide range of effects on various organ systems [15,21]. GI side effects include gastric distress and diarrhea, usually when beginning therapy, but persisting in some patients. Giving lithium with meals,

228

T. Rowland, L. DePalma / NeuroRehavilitation 5 (1995) 219-232

in smaller doses, or in sustained-release preparations will reduce or eliminate GI complaints. CNS side effects can include somnolence, cognitive slowing, fine tremor, and motor incoordination, all of which become more pronounced with increasing serum levels and are valuable clinical indicators of possible toxicity. Lithium often causes acneiform dermatologic eruptions, which may be treated topically. The most common renal side effect of lithium is polyuria with associated polydipsia due to decreased distal tubule fluid resorption. Polyuria can progress to nephrogenic diabetes insipidus which is unresponsive to vasopressin, but can be ameliorated by thiazide diuretics. Lithium can suppress thyroid function, causing a generally transient hypothyroidism, but if decreased thyroid levels persist, it is appropriate to initiate thyroid supplementation. Cardiac side effects of lithium consist primarily of benign EKG changes (T wave flattening or inversion), though lithium does suppress sinus node activity to some extent. Lithium has no hematologic toxicity, but does cause a benign leukocytosis early in the course of treatment. Finally, and for unclear reasons, lithium frequently causes carbohydrate cravings with associated weight gain which is quite distressing to some patients. Routine monitoring of lithium levels, EKG, electrolytes, renal function and thyroid function is considered the standard of care, and will alert the clinician to potential problems before they become serious. Current recommendations include a pretreatment evaluation consisting of serum electrolytes with BUN and creatinine, thyroid function tests (T4 , free T4 , and TSH), complete blood count, urinalysis, EKG, and a pregnancy test (if there is any possibility the patient could be pregnant). Serum lithium levels should be obtained approximately 12 h after the most recent dose, and are recommended every 4-8 weeks once a stable level has been achieved. This frequency can gradually be reduced in reliable patients. Renal and thyroid function should be reassessed every 6 months, and the remainder of the initial evaluation labs are repeated yearly [52]. Rational treatment of agitation using lithium is guided both by the clinical response and serum levels with well-established toxic ranges. Thera-

peutic lithium levels for the treatment of bipolar disorder range from 0.5 to 1.2 mEq/l, but lithium is often effective at much lower levels in TBI patients (0.3 to 0.6 mEq/I). A reasonable initial lithium dose in the brain-injured individual is 150-300 mg p.o. q day or divided bj.d. The dose can be titrated upwards in increments of 150 or 300 mg per day, with the maintenance dose determined by clinical response and monitoring serum levels and/or toxicity. Lithium intoxication is generally observed at serum levels exceeding 1.5 mEq/l, with levels above 2.5 mEq/1 considered life-threatening. Toxicity consists of severe manifestations of lithium's typical adverse effects, with particularly prominent CNS manifestations that can progress from stupor to coma to seizures and ultimately death if not adequately treated. Outpatients taking lithium for chronic control of aggression and impulsivity must be informed that sodium restriction is hazardous, and that adequate fluid intake during exercise or hot weather is essential to prevent elevated lithium levels. The treating physician should also be aware that several commonly used medications, including NSAIDs, diuretics, and tetracycline, can increase lithium levels [21,22]. 5.5. Anticonvulsants

The anticonvulsant agents carbamazepine and valproic acid are rapidly gaining acceptance in the field of psychiatry for treatment of a variety of affective disorders, as well as disorders of episodic dyscontrol and aggression [21]. Initially, these agents were used as augmentation to lithium therapy, or second-line monotherapy agents when lithium could not be tolerated. Within the past several years, however, clinical evidence has suggested that the anticonvulsants may be first-line choices for a number of disorders, particularly those involving impulsivity and aggression [21,53]. This knowledge is now being applied in the neurotrauma rehabilitation setting, with favorable clinical results based so far only on anecdotal evidence. As with much of the research on pharmacologic management of agitation, the few available controlled studies using anticonvulsants have been performed in populations with demen-

T. Rowland, L. DePalma / NeuroRehabilitation 5 (1995) 219-232

tia and mental retardation, rather than traumatic brain injury [10]. Carbamazepine. Carbamazepine is structurally similar to the tricyclic antidepressant imipramine, and hence shares some common properties with the tricyclic antidepressants, most notably cardiotoxicity in overdose and utility in the treatment of neuropathic pain. The precise mechanisms of carbamazepine's anticonvulsant, antimanic, and anti-aggressive actions are not clearly understood. Central pharmacologic actions of carbamazepine may include stabilization of neuronal membranes via decreased sodium flux, potentiation of alpha2-adrenergic receptors, and regulatory effects on calcium channels via a subtype of the benzodiazepine receptor [15,21]. The efficacy of carbamazepine in the treatment of temporal lobe epilepsy, which is frequently accompanied by neuropsychiatric symptoms and behavioral disturbance, suggests that carbamazepine may be acting to suppress aberrant limbic system activity which is hypothesized to be causal in episodic dyscontrol syndromes [54,55]. Data on efficacy of carbamazepine for long-term control of agitation is limited; an overall response rate of 80% was found in three small open trials (total n = 30), while the response rate was 36% in a double-blind, placebo-controlled study of 11 mentally retarded patients with severely agitated behavior [10,56]. The pharmacokinetics of carbamazepine are quite complex, which occasionally leads to difficulties in regulating the serum levels of the drug. Available only in oral preparations (tablets or suspension), gastrointestinal absorption of carbamazepine is slow and erratic, with peak serum concentrations typically occurring 4-8 h following ingestion, but occasionally up to 24 h after ingestion. Carbamazepine undergoes hepatic metabolism with a half-life of 10-20 h, achieving steady-state concentrations in 4-5 days. Since carbamazepine induces its own metabolism, however, serum concentrations can be expected to fall approximately 3-4 weeks after initiation of therapy, often necessitating an increased dose to maintain therapeutic serum levels. Steady-state carbamazepine serum levels cannot be predicted from the oral dose.

229

Therapeutic carbamazepine serum levels for the control of affective and behavioral disorders (including chronic agitation) range from 8 to 12 mcg/ml, which is higher than the usual therapeutic range for the treatment of temporal lobe epilepsy (4-10 mcg/mO [20]. For the long-term management of agitation in TBI patients we usually initiate carbamazepine at doses of 100-200 mg p.o. b.Ld. to avoid potential CNS side effects which could interfere with participation in therapies. Clinical response is documented and a steady-state serum level is drawn prior to any dosage increases, which are generally accomplished gradually (200 mg increases every 2-3 days). The final maintenance dose is determined by both clinical response and therapeutic serum levels, and in our experience range from 400 to 1600 mg per day, divided t.i.d. to q.i.d. at higher doses to minimize serum level fluctuations with attendant side effects. The most common adverse effects of carbamazepine are CNS and GI effects: sedation, dizziness, motor incoordination, ataxia, diplopia; and nausea, vomiting, dyspepsia, and diarrhea [15]. These typically occur with initiation of therapy or dosage increases, may be reduced by gradual titration of dosage, and usually resolve with ongoing therapy. Several rare but quite serious adverse reactions may occur with carbamazepine treatment: an acute chemical hepatitis (1 in 25000) [57] or cholestasis which may be fatal can occur early in treatment; severe bone marrow suppression with aplastic anemia and/or agranulocytosis (1.4 per million) [57] can occur at any time during treatment; and finally, Stevens-Johnson syndrome of exfoliative dermatitis (5.75 per million) [57] can occur with carbamazepine, though ordinary drug hypersensitivity reactions are more commonly seen. The feared complications of hepatitis and agranulocytosis dictate that the CBC and LFTs be monitored closely, particUlarly during initiation of carbamazepine treatment. Baseline liver function tests, electrolytes, and a CBC with platelets and differential should be obtained before starting the medication. The LFTs should be checked after 1 month of treatment, then quarterly for the first year, and then annually thereafter unless changes arc made in

230

T. Rowland, L. DePalma / NeuroRehavilitation 5 (J995) 219-232

the treatment regimen (e.g. other medications with potential hepatotoxicity are introduced). Current recommendations suggest repeating the CBC and a serum carbamazepine level every 2 weeks for the first 2 months of treatment, and every 3 months thereafter. In addition, patients should be informed of the symptoms of leukopenia and hepatic dysfunction, and told to contact the physician immediately should any of these occur [21]. Valproic acid. Valproic acid is a simple branched-chain carboxylic acid that apparently acts at neuronal sodium channels and increases central GABA concentrations (via inhibiting GABA catabolism). It is an effective anticonvulsant, but recently has also been found as effective as lithium in the control of mania [58]. Like carbamazepine, valproic acid is increasingly being used as a first-line agent for a variety of affective disorders, as well as disorders of episodic dyscontrol and aggression, though evidence supporting its use in behavioral dyscontrol remains largely anecdotal. There is only one published prospective study documenting the efficacy of valproic acid for long-term management of agitation, and this was in four subjects with senile dementia, rather than TBI [59]. The pharmacokinetics of valproic acid are more simple than those of carbamazepine. Valproic acid is available in oral preparations as valproic acid and sodium valproate (Depakene) or divalproex sodium (Depakote). Dosing is equivalent with either preparation. Valproic acid is rapidly and completely absorbed from the GI tract, with a serum peak occurring in 1-4 h (absorption is slightly delayed when the drug is taken with food). Valproic acid undergoes hepatic metabolism with a half-life of between 8 and 15 h [15]. There are no established therapeutic serum valproic acid levels for the management of affective disorders or behavioral dyscontrol. At present, most clinicians treating these conditions use the recommended serum levels for seizure control, 50-100 mcg/ml, as a guideline. The initial valproic acid dose for the treatment of seizures is 15 mg/kg, divided t.i.d. to q.i.d., but in patients with traumatic brain injury we generally initiate treatment at lower doses (e.g. 250 mg b.i.d.) to

avoid possible adverse CNS effects. The dosage may be titrated upward by 250 mg every 2-3 days. Clinical response is documented and a steadystate serum level is drawn prior to any dosage increases. The final maintenance dose, which is determined both by clinical response and by therapeutic serum levels, in our experience ranges from usually 1000 to 2500 mg per day, divided t.i.d. to q.i.d. Valproic acid has a more benign side effect profile than carbamazepine, though by no means is it without risk. The most common adverse effects of valproic acid are sedation, nausea, and vomiting, all of which are generally transient and can be avoided by dosing at mealtimes and gradual upward titration. Rarely, valproic acid can cause ataxia, asterixis, tremor, or headaches. The most serious possible adverse reactions to valproic acid are pancreatitis and hepatitis, usually occurring in the first 6 months of treatment. Valproic acid causes a transient, asymptomatic elevation of LFfs in up to 40% of patients, but rarely a fulminant chemical hepatitis can occur; the incidence of severe hepatitis is about 1 in 50000, mostly in children younger than 10 years-old being treated for seizures with multiple medications [20]. Early detection of these possibly severe reactions is accomplished by carefully monitoring LFfs and amylase, especially during the first 6 months of treatment. 6. Summary There is very little specific information available relevant to neuropharmacologic interventions with brain injury agitation; therefore, it is challenging to make clearcut recommendations or prioritize medications. It is important to understand the most common side effects and contraindications for these medications as this often plays a large role in the choice and sequence of therapy. This article is not exhaustive, and is intended to highlight the more commonly used and studied agitation medications. The article of Drs Cardenas and Bell will discuss other possible interventions, particularly the newer pharmacologic agents.

T. Rowland, L. DePalma / NeuroRehabilitation 5 (1995) 219-232

Acknowledgements

The authors thank Pam Slaughter and Laura Cole for their assistance in the preparation of the manuscript. References [I] American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders, 4th Ed (DSM-IV). Washington, DC, American Psychiatric Press, 1994. [2] Trzepacz, PT. The neuropathogenesis of delirium: A need to focus our research. Psychosomatics 1994;35:374-391. [3] Lipowski ZJ. State of the art: Delirium (acute confusional state). JAmMed Assoc 1987;258(13):1789-1793. [4] Trzepacz PT. Delirium. In: Silver JM, Hales RM and Yudofsky SC, eds. Neuropsychiatry of Traumatic Brain Injury. Washington, DC: American Psychiatric Press 1994;189-218. [5] Slaby AE, Erie SR. Dementia and delirium. In: Fogel BS and Stoudemire A, eds. Psychiatric Care of the Medical Patient. New York: Oxford University Press 1993;415-453. [6] Itil T, Fink M. Anticholinergic drug-induced delirium: Experimental modification, quantitative EEG and behavioral correlations. J Nerv Ment Dis 1966;143:492-507. [7] Corrigan JD. Development of a scale of assessment of agitation following traumatic brain injury. J Clin Exp Neuropsychol 1989;11(2):261-277. [8] Corrigan JD, Arnett JA, Houck U et al. Reality orientation for brain injured patients: group treatment and monitoring of recovery. Arch Phys Med Rehabil 1985;66:626-630. [9] Levin H et aI. Galveston orientation and amnesia test: a practical scale to assess cognition after head injury. J Nerv Ment Dis 1979;167:675-684. [10] Smith DA, Perry PJ. Non-neuroleptic treatment of disruptive behavior in organic mental syndromes. Ann Pharmacother 1992;26: 1400-1408. [11] Schneider LS, Sobin PB. Non-neuroleptic treatment of behavioral symptoms and agitation in Alzheimer's disease and other dementia. Psychopharmacol Bull 1992;28(1):71-79. [12] Gualtieri CT. Traumatic brain injury. In: Fogel BS and Stoudemire A, eds. Psychiatric Care of the Medical Patient. New York: Oxford University Press 1993;517-535. [13] Silver JM, Yudofsky SC. Pharmacologic treatment of neuropsychiatric disorders. NeuroRehabilitation 1993;3:15-25. [14] Zorumski CF, Isenberg KE. Insights into the structure and function of GABA-Benzodiazepine receptors: Ion channels and psychiatry. American Journal of Psychiatry 1991;148:162-173.

231

[15] Gilman AG, Goodman LS, Rail TW, Murad F, eds. The Pharmacological Basis of Therapeutics, 7th Ed. New York, Macmillan, 1990. [16] Killian GA, Holzman PS, Davis JM et al. Effects of psychotropic medication on selected cognitive and perceptual measures. J Abnorm Psychol 1984;93(1):58-70. [17] Barbee JG. Memory, benzodiazepines and anxiety: integration of theoretical and clinical perspectives. J Clin Psychiatry 1993;54(Suppi):86-97. [18] Goldstein LB. Pharmacologic modulation of recovery after stroke: clinical data. J Neuro Rehabil 1991;5:129-140. [19] Goldstein LB. Basic and clinical studies of pharmacologic effects on recovery from brain injury. J Neural Transplant Plast 1993;4:175-192. [20] Kaplan HI, Sadock BJ. Pocket Handbook of Psychiatric Drug Treatment. Baltimore, Williams & Wilkins, 1993. [21] Kaplan HI, Sadock BJ, Grebb JA Kaplan and Sadock's Synopsis of Psychiatry: behavioral sciences, clinical psychiatry, 7th Ed. Baltimore, Williams & Wilkins, 1994. [22] Jenkins SC, Gibbs TP, Szymanski SR. A Pocket Reference for Psychiatrists. Washington, DC, American Psychiatric Press, 1990. [23] Goldstein MG, Haltzman SD. Intensive care. In: Fogel BS and Stoudemire A, eds. Psychiatric Care of the Medical Patient. New York: Oxford University Press 1993;241-265. [24] Feeney DM, Gonzalez A, Law WA Amphetamine, haloperidol, and experience interact to affect rate of recovery after motor cortex injury. Science 1982;217:855-857. [25] Wise RA. Neuroleptics and operant behavior: The anhedonia hypothesis. Behav Brain Sci 1982;5:39-82. [26] Carie ED, Polinsky RJ. Haloperidol induced dysphoria in patients with Tourette's syndrome. Am J Psychiatry 1979;136(9):1216-1217. [27] Gualtieri CT. Pharmacotherapy and the neurobehavioral sequelae of traumatic brain injury. Brain Inj 1988;2(2):101-129. [28] Meltzer HY. Dimensions of outcome with ciozapine. Br J Psychiatry 1991;17(Suppi):46-53. [29] Lieberman JA, Saltz BL, Johns CA et al. The effects of clozapine on tardive dyskinesia. Br J Psychiatry 1991; 158:503-510. [30] Gershanik OS. Drug-induced movement disorders. Curr Opin Neurol Neurosurg 1993;6:369-376. [31] USP Drug Information for the Health Care Professional, Statref, May 1994. [32] Vining EPG et al. Psycologic and behavioral effects of antiepileptic drugs in children: a double-blind comparison between phenobarbital and valproic acid. Pediatrics 1987;80:165-174. [33] Meador KJ et al. Comparative cognitive effects of anticonvulsants. Neurology 1990;40:391-394. [34] Brooke MM, Patterson D, Questad K et aI. The treatment of agitation during initial hospitalization after

232

[35] [36] [37]

[38]

[39] [40] [41] [42] [43] [44]

[45] [46]

T. Rowland, L. DePalma / NeuroRehavilitation 5 (1995) 219-232

traumatic brain injury. Arch Phys Med Rehabil 1992;73:917-92l. Yudofsky SC, Silver JM, Jackson W et al. The overt aggression scale for the objective rating of verbal and physical aggression. Am J Psychiatry 1986;143:35-39. Physicians GenRx, The Complete Drug Reference, Statref, May 1994. Yudofsky S, Williams D, Gorman J. Propranolol in treatment of rage and violent behavior in patients with chronic brain syndromes. Am J Psychiatry 1981;138:218-220. Eichelman B. Neurochemical and psychopharmacologic aspects of aggressive behavior. In: Meltzer HY, ed. Psychopharmacology: The Third Generation of Progress. New York: Raven Press 1987;697-704. Bradley PB. Introduction to Neuropharmacology. Boston, Wright, 1989. Mandel P, Mack G, Kempf E. Molecular basis of some models of aggressive behavior. In: Psychopharmacology of Aggression. New York: Raven Press, 1979. Mysiw WJ, Jackson RD, Corrigan JD. Amitriptyline for posttraumatic agitation. Am J Phys Med Rehabil 1988. AHFS Drug Information, Statref, May 1994. Riblet LA, Taylor DP. Pharmacology and neurochemistry of trazodone. J Clin Psychopharmacol 1981;l(Suppl):17-22. Bossier JR, Portmann-Cristesco E, Soubrie P et al. Pharmacological and biochemical features of trazodone. In: Ban TA, Silvestrini B, eds. Trazodone: Modern Problems of Pharmacopsychiatry. Basel: Karger, 1974;9:18-28. Rowland T, Mysiw WJ, Bogner J et al. Trazodone for posttraumatic agitation (abstract). Arch Phys Med Rehabil 1992;73:963. Simpson D, Foster D. Improvement in organically disturbed behavior with trazodone treatment. J Clin Psychiatry 1986;47(4):191-193.

[47] O'Neil M, Page N, Adkins WN et al. Tryptophan-trazodone treatment of aggressive behavior. Lancet 1986;Oct11:859-860. [48] Wilcock G, Stevens J, Perkins A Trazodone/tryptophan for aggressive behaviour. Lancet 1987;Apr-18:929-930. [49] AMA Drug Evaluations, Statref, May 1994. [50] Baraban JM, Worley PF, Snyder SH. Second messenger systems and psychoactive drug action: Focus on the phosphoinositide system and lithium. Am J Psychiatry 1989;146:1251-1260. [51] Haas JF, Cope DN. Neuropharmacologic management of behavior sequelae in head injury: A case report. Arch Phys Med Rehabil 1985;66:472-474. [52] Goodwin FK, Jamison KR. (1990). Manic Depressive Illness. New York, Oxford University Press, 1990. [53] Monroe RR. Dyscontrol syndrome: long-term follow-up. Compr Psychiatry 1989;30:489-497. [54] Leiderman DB, Csernansky JG, Moses JA Jr. Neuroendocrinology and limbic epilepsy: relationships to psychopathology, seizure variables, and neuropsychological function. Epilepsia 1990;31:270-274. [55] Adamec RE. Partial kindling of the ventral hippocampus: identification of changes in limbic physiology which accompany changes in feline aggression and defense. Physiol Behav 1991;49:443-453. [56] Reid AH, Naylor OJ, Kay DS. A double-blind, placebo controlled, crossover trial of carbamazepine in overactive, severely mentally handicapped patients. Psychol Med 1981;11:109-113. [57] Pellock JM. Carbemazepine side effects in children and adults. Epilepsia 1987;28(Suppl 3):S64-S70. [58] Bowden CL, Brugger AM, Swann AC et al. Efficacy of divalproex vs lithium and placebo in the treatment of mania. JAmMed Assoc 1994;271:918-924. [59] Mellow AM, Solano-Lopez C, Davis S. Sodium valproate in the treatment of behavioral disturbance in dementia. J Geriatr Psychiatry NeuroI1993;6(4):205-209.