Pharmacologic Management of Organic Agitation

Original Authors: Paul Perry, Ph.D, BCPP, Dan Smith, Pharm.D., BCPP
Latest Reviser: Vicki Ellingrod, Pharm.D., BCPP
Creation Date: 1996
Last Revision Date: September 2001
Peer Review Status: Internally Peer Reviewed

This section will review the pharmacologic management of behavioral disturbance (BD) associated primarily with Alzheimer's dementia (DAT) as well as BD secondary to organic causes and delirium. Drug therapy may be necessary when the patient's behavior becomes agitated, aggressive, hostile, or otherwise disruptive and dangerous to themselves or those around them. Unless otherwise noted, BD will be used generically in reference to this constellation of symptoms. The recommendations in this review assume that the underlying cause of the patient's BD has been diagnosed and that non-pharmacologic approaches have been exhausted. The BDs associated with personality disorders are not addressed.

The handout will discuss antipsychotic and non-antipsychotic treatments. The review of non-antipsychotic treatments will only analyze prospective evaluations of drug therapies in BD. Only studies with lithium, _-blockers (propranolol and pindolol), carbamazepine, benzodiazepines (diazepam, chlordiazepoxide and oxazepam), buspirone, and naltrexone met this criteria. Retrospective studies and case series were excluded from this analysis to avoid the biases inherent in reports of these types.

Behavioral Disturbances in Alzheimer's Dementia
In Alzheimer's disease at least 70% of patients will experience symptoms of psychosis within 7 years of diagnosis. It is there behavioral manifestations of dementia which are of tremendous importance due to the amount of stress they can cause the patient as well as the primary caregiver (US DHHS, 1984). In addition, it is these symptoms that often lead to the frequent institutionalization of DAT patients (Ferris et al 1985).

These symptoms can be classified in three categories; Psychosis, which consists of paranoia, delusions, and hallucinations, Non-aggressive behavioral, which consists of wanders, pacing, bossiness, repetitive behaviors, complaining and whining, verbal interruptions, and attention-seeking behaviors, and Aggressive behavioral, which consists of hitting, pushing, scratching, biting, screaming, and kicking.

In terms of treating these behavioral manifestations of DAT, four classes of medication have been used the most. These consist of the antipshchotics, antidepressants, anxiolytics, and antiepileptics. The following is a review of their use.

ANTIPSYCHOTICS

Antipsychotics have long been the primary pharmacologic approach in the treatment of BD in DAT. Guidelines for antipsychotic use in long-term care facilities were updated in the Omnibus Budget Reconciliation Act (OBRA) by the Health Care Financing Administration (HCFA), a government agency which regulates Medicare and Medicaid recipients (Druckenrod et al 1993). These changes went into effect October 1, 1990. The antipsychotics were defined by HCFA as the phenothiazine, butyrophenone, and thioxanthene chemical classes and the individual agents, molindone and loxapine. Atypical antipsychotics were not originally included in this designation, but recently guidelines have included these agents (Table X). This legislation was passed in part due to the risk of adverse effects associated with these drugs including pseudoparkinsonism, dystonias, akathisia, tardive dyskinesia, orthostatic hypotension, and anticholinergic delirium.

OBRA regulations indicated that appropriate indications for antipsychotics would include acute psychotic episodes, atypical psychoses, brief reaction psychosis, delusional disorder, Huntington's disease, psychotic mood disorders (mania and depression with psychotic features), schizoaffective disorder, schizophrenia, schizophreniform disorder, Tourette's syndrome, symptomatic (_7 days) treatment of hiccups, nausea, vomiting, or pruritus. Dementia and delirium with associated psychotic and/or agitated behaviors that (1) are quantitatively and objectively documented; (2) are not caused by preventable reasons; and (3) are causing the resident to present danger to himself or to others or experience psychotic symptoms not exhibited as dangerous, but causes the patient distress or impairment in functional capacity are appropriate indications for antipsychotics (Druckenrod et al 1993). Any other specific condition not mentioned is not an appropriate use of an antipsychotic.

Pharmacologic management of agitation associated with dementia has been the subject of several reviews (Risse and Barnes 1986, Yudofsky et al 1990, Kunik et al 1994). Recently, clinical guidelines on the treatment of agitation in older patients with dementia were published (Anonymous 1998). Recommendations for the first line treatment of these behavioral disturbances are included in Figure 1.

Typical Antipsychotics
Schneider et al (1990) performed a meta-analysis of 18 controlled, blinded studies of antipsychotic treatment of inpatients with severe dementia. In seven studies that compared an antipsychotic to placebo, the antipsychotic produced an average improvement of 59%. This compared to an average placebo improvement rate of 41% (range, 0 to 67%). Although the meta-analysis could not determine individual symptom responsiveness to antipsychotics, author's assessment of individual studies indicated that agitation, uncooperativeness, and hallucinations tended to improve with medication compared to placebo. In six and five studies, respectively, comparing thioridazine and haloperidol to other antipsychotics, all antipsychotics were equally effective in the symptomatic treatment of dementia. The daily doses expressed in chlorpromazine equivalents ranged from 66 to 267 mg/d or approximately 1 to 4 mg/d of haloperidol (see antipsychotic dosing chapter for dosage equivalents) and duration of treatment in the studies varied from three to 18 weeks, with most lasting from four to eight weeks. It is possible that higher daily doses and/or a longer treatment duration may have increased the drug/placebo response difference. In addition, more mildly ill patients with dementia might respond at a higher rate to antipsychotics.

In 1997 Sultzer et al conducted a study in 29 subjects with DAT. This was a randomized, double-blind study comparing haloperidol with Trazodone for the treatment of agitation. The mean dosage of haloperidol was 2.5 mg ± 1.7 mg/day and for Trazodone was 218 ± 65 mg/day. Overall the response rate was around 64% with 57% of the haloperidol patients and 71% of the Trazodone patients having an endpoint Clinical Global Impressions (CGI) score of 1 or 2. In terms of side effects, 50% of those receiving haloperidol and 14% of those receiving Trazodone complained of an adverse event. Thus, both haloperidol and Trazodone are effect for reducing agitated symptoms in DAT, but the use of haloperidol may be associated with more adverse events.

Though as-needed dosing is commonly used to manage patients with dementia that are agitated, there is no study concerning the efficacy of this treatment (Druckenbrod et al 1993). It is recommended that "prn for agitation" orders should specify severity of behavior, quality, and duration before as-needed medication is administered. Orders should specify maximum daily doses and management of adverse effects.

Atypical Antipsychotics
With the availability of the newer atypical antipsychotics the use of typical agents has diminished. Although, they are FDA approved for the treatment of schizophrenia, many if not all of the newer agents are being used for the treatment of agitation in elderly patients residing in skilled care facilities or nursing homes. Based on their current usage, research is now being done to determine their effectiveness on agitation as well as their effect on cognition.

Clozapine: The first atypical to be studied was clozapine. Although this medication is very effective for treating schizophrenia, its use in dementia is somewhat limited due to its side effect profile. In a retrospective study by Oberholzer et al (1992), 18 patients (mean age 81.5 years old) were given daily doses of clozapine ranging from 12. 5 - 200 m/day for 0.2 - 22 months for the treatment of agitated dementia. Of the subjects, 4 of the 18 discontinued the clozapine a few days after it was started due to confusion, somnolence, motor restlessness, and lack of therapeutic effect, while 14 of the 18 experienced no problems. Seven of these patients were classified as having severe dementia of the Alzheimer's type (DAT) and were assessed using the Sandoz Clinical Assessment Scale (SCAS) and the Nurses Observation Scale for Inpatient Evaluation (NOISE). In all of these patients there were statistically significant increases in irritability (p< 0.0004), decreases in social competency (p = 0.023), and decreases in social interest (p = 0.008) when the clozapine was stopped. After reinstitution of clozapine these variables went back to baseline. Thus, although clozapine may be useful in the treatment of agitation associated with dementia and DAT, significant side effects may limit its use.

Risperidone: In an open label study by Jeste et al (1996), 14 agitated subjects (age range 54 - 100 yo) were given risperidone (mean dose 1.7 ± 1.9 mg/day) for an average of 13 ± 12.3 weeks. Of these 14 patients, 2 were considered treatment failures, 6 showed marked improvement, and 6 showed moderate improvements. In examining the dosage of risperidone used, the authors found that patients with dementia used lower doses (1 mg/day) compared to non-demented patients (2.3 mg/day).

In a second open trial by the same authors, 19 subjects with a mean age 64.7 ± 14.4 years were given risperidone for a mean treatment time of 10.6 ± 7 weeks. In looking at their mini mental status scores (MMSE) before and after treatment the authors found a statistically significant increase after treatment (p = 0.005). Before treatment the mean MMSE was 24.2 ± 5.5 and after this increased to 28.2 ± 2.5. Although it is not know if this improvement is due to a direct cognitive effect or a reduction in psychopathology, these results are promising.

In 1999 De Deyn et al published a randomized placebo controlled trial comparing risperidone to haloperidol in the treatment of agitation in DAT. This study included 334 subjects. The mean dosage of risperidone was 1.1 mg/day for haloperidol was 1.2 mg/day. The primary outcome measure was the BEHAVE-AD scale. In looking at the completer analysis at 12 weeks, 72% of the risperidone treated patients, 69% of the haloperidol treated patients, and 61% of the placebo patients had greater than a 30% reduction in the BEHAVE-AD score. Thus, there were no statistical differences between the three groups in terms of efficacy. The drop out rate for this study was relatively high at 35% and as such may have affected these results.

Olanzapine: The first study of olazapine was an 8 week trial published by Satterlee et al (1995). As part of this double blind trial, 238 patients (³ 65 years old) with psychotic or behavioral symptoms of DAT were given 1 - 8 mg/day of olanzapine or placebo. Sixty-nine of these patients were receiving < 5 mg/day. No differences were found between olanzapine and placebo in terms of efficacy, hepatic transmaminases, leukopenia, EPS, or orthostatic hypotension.

More recently, Street et al (2000) reported the results of a multi-center, double blind, placebo controlled study in which 206 nursing home patients with DAT were randomized to receive either placebo, 5, 10, or 15mg of olazapine for 6 weeks. The efficacy measures used in this study were the neuropsychiatric inventory - Nursing home version (NPI-NHV) that rates agitation, delusions, and hallucinations. Overall, more olazapine patients than placebo patients showed a ³ 50% decrease in the total NPI-NHV scores. This improvement was seen in 65% of patients receiving 5mg, 57% of patients receiving 10mg and 34% of patients receiving placebo. Those in the 15mg group did not experience a significant improvement. Additionally, the olazapine groups were equal to the placebo groups on both the Simpson Angus scores and the Abnormal Involuntary Movement Scale (AIMS) scores. Subjects receiving olanzapine though had a higher incidence of somnolence and abnormal gait, compared to placebo. Although MMSE were not looked at, olazapine may represent another treatment option for agitation in DAT.

Quetiapine: Very little is known about the use of quetiapine in the elderly with agitation. In 1999 Schnieder reported on the results of a one year open label study in 78 patients with DAT. The mean dose used was 100 mg/day. The total BPRS score and hostility score were significantly reduced in both groups (p< 0.05). Subjects categorized as hostile showed a greater improvement in hostility which was disproportionate to their reduction in negative symptoms. Thus, quetiapine may have an independent effect on hostility, but further studies need to be done.

Antipsychotic Summary
The use of antipsychotics for agitation in patients with DAT has become very commonplace. In looking at the literature all agents have been shown to be effective, and as such the choice for which of these agents to use should be based on the patient factors, potential for drug interactions and adverse events, and prescriber experience.

Recently the atypical antipsychotic ziprasidone has become available, although no information regarding its use in the population is available. One of the concerns about using these medications is their risk for prolongation of the QTc interval on an electrocardiogram. This prolongation of the QTc may put patients at risk for the development of Torsades de Pointes (TdP). There are some risk factors for TdP that are not related to the use of antipsychotics, but must be taken into consideration when using these medications in this group of patients. These include, the presence of electrolyte disturbances (i.e. potassium, magnesium, sodium), the use of medications that block the potassium channels (i.e. Tricyclic antidepressants, Fluoxetine, etc…), bradycardia, congential long QT syndrome, congestive heart failure or cardiovascular disease, syncope, female gender, and drug interactions through CYP3A4 and 2D6 which potential increase serum concentrations of the antipsychotics. Thus, any patient being prescribed an antipsychotic needs to be evaluated for the presence of these risk factors as well as for the evaluation of potential drug interactions. A good web site to refer to is www.torsades.org which has an updated listing of medications that can cause TdP. If patients have any of these risk factors than an ECG should be obtained at baseline, as well as the midpoint in therapy and once the patient is at steady state. In patients with a baseline QTc > 450 msec extreme caution, including close monitoring of heart rate and ECG, should be taken when using these medications. A patient who's QTc goes higher than 500 msec should be taken off the medication.

In elderly patients with agitation and dementia, several guidelines on the dosing of antipsychotics exist. Initially start with dosages that are one third to one half of the usage dose for younger patients. If this is ineffective, combining with a very low dose benzodiazepines may be another therapeutic option. When monitoring therapy, watch closely for adverse drug reactions, specifically anticholinergic delirium, movement disorders, and orthostatic hypotension.

Anxiolytics

Benzodiazepines
Though many benzodiazepines are available for the management of agitation, comparative studies to determine efficacy differences between agents are few.

Placebo-Controlled Studies: In the first study, Sanders (1965) conducted an 8 week, double-blind study comparing oxazepam and placebo in 94 chronically institutionalized "emotionally disturbed" geriatric patients. The patients' diagnoses, by current standards, were unclear. The study's target symptoms were anxiety, tension, dysphoria, agitation, obsessive thinking, compulsive behavior, irritability, lethargy, insomnia, conversion, phobia, and autonomic reactions. The patients were randomized to receive individually titrated oxazepam doses 20-40 mg/d or placebo. Other psychotropic medications were discontinued two weeks before the study. The oxazepam group included 30 women and 20 men (mean age = 81). The placebo group included 25 women and 19 men (mean age = 79). The physician graded each target symptom using a 4 point scale where 4 = persistent discomfort and 1 = no problem. A global rating was used to measure overall function using a 5 point scale of pronounced improvement, moderate improvement, slight improvement, none, or worse overall condition. The reduction in all target symptoms was significantly greater for the oxazepam group when assessed at weeks four and six. At week 8 adequate comparisons were impossible because 52% of the patients in the control group required active treatment while on placebo. Global scores were also better for the oxazepam group. The target symptoms most responsive to oxazepam treatment included anxiety, tension, dysphoria, lethargy, agitation and obsessive thinking. Side effects observed in seven of the oxazepam patients included nausea, headache and dizziness.

In a nearly identical study, Beber (1965) evaluated oxazepam in 100 institutionalized patients (mean age = 79 years) diagnosed with chronic brain syndrome, anxiety, and depression. The target symptoms were the same as those in Sanders' 1965 study. The patients were randomized into two groups of 50 patients of similar age and sex distribution. One group was titrated with oxazepam up to 40 mg/day while the others received placebo. At each of 8 weekly observations, significantly greater improvement, calculated as percent of original status, was observed in the oxazepam group versus the placebo group for all of the target symptoms. Improvement was reported to be most pronounced for the symptoms of anxiety, tension, irritability, insomnia and agitation. The reliability of this informal rating scale is questionable. Globally, 68% of the oxazepam patients achieved moderate to notable improvement. All symptoms were eliminated in 10 patients. None of the oxazepam treated patients had signs of deterioration. Similar improvement was seen in 18% of the placebo patients with complete remission in one patient. Deterioration occurred in 12% of the placebo group. Sixteen patients treated with oxazepam and eight patients in the placebo group experienced adverse effects including drowsiness, dizziness and nausea.

A series of 148, randomized, chronically ill, elderly patients (98 men and 50 women) were studied by Chesrow et al (1965). Of these, 139 had cerebral arteriosclerosis with "psychoneurosis." The drug dose was titrated using identical tablets of oxazepam 10 mg, chlordiazepoxide 10 mg, or placebo. The average dose between treatments was roughly equivalent (35 mg/d for oxazepam and 30 mg/d for chlordiazepoxide). Concomitant psychotropic drug use was unclear. The grading system for target symptoms was 3 = severe to 0 = not present. Patients were rated at the beginning and the end of the study. Oxazepam was administered for 1-26 weeks, chlordiazepoxide 1-20 weeks, and placebo 1-13 weeks. Globally, marked or moderate improvement was seen in 87% of the oxazepam group, 52% of the chlordiazepoxide group and 0% of the placebo group. Placebo treatment was dropped midway through the study due to breakthrough agitation, irritability, and insomnia. Drowsiness, slurred speech, and other adverse effects occurred in 29 of 71 chlordiazepoxide patients. Side effects occurred in only 3 of the oxazepam patients. This finding is not surprising since oxazepam is metabolized to an inactive conjugate while chlordiazepoxide is metabolized to active metabolites that accumulate in the elderly.

De Lemos et al (1965) compared the effectiveness of diazepam and placebo in the treatment of geriatric patients suffering from dementia or chronic brain syndrome. Following a two week stabilization period, patients received a week of placebo in place of their pre-existing medications with subsequent randomization to diazepam 7.5 mg/d (n=14) or placebo (n=15) for the following two weeks. Ten of the diazepam patients and two of the placebo patients experienced improvement of their symptoms of anxiety, confusion, daytime hyperactivity, nocturnal awakenings, restlessness, agitation, and floor-walking (diazepam vs. placebo p < 0.004). Twelve of the 14 diazepam patients experienced drowsiness necessitating dosage reduction. With such a short trial, it is unlikely that diazepam serum concentrations ever reached steady-state. It is unknown if these positive results would be maintained long-term or if diazepam toxicity would have become more pronounced at steady-state serum concentrations.

Benzodiazepine vs. Antipsychotic Comparison Studies: Kirven et al (1973) conducted a 4-week, double-blind, parallel group study to compare the efficacy of thioridazine and diazepam in 56 hospitalized geriatric patients with dementia. The 56 patients' target symptoms were anxiety, agitation, tension, or sleep disturbances. The patients were randomly assigned to receive either thioridazine 10 mg or diazepam 2 mg. The doses were titrated to a daily maximum of thioridazine 200 mg or diazepam 40 mg. The treatment groups were of similar sex, age, and duration of hospitalization. Two objective rating scales, the modified Hamilton Anxiety Scale (HAM-A) and a modified Nurses' Observation Scale for Inpatient Evaluation (NOSIE), and two global evaluations were used. The global evaluations consisted of the patient's overall degree of mental illness and overall change in patients condition. The mean doses for diazepam and thioridazine were 9 mg/d and 39 mg/d, respectively. On the HAM-A, none of the differences between the two drugs on the individual items were significant at the end of the trial although symptoms of anxious mood, tension, depressed mood, and agitation improved significantly for both drugs. The NOSIE data demonstrated that 47% of the thioridazine-treated and 39% of diazepam-treated patients were improved, again a non-significant finding. Drowsiness occurred in two of the thioridazine and four of the diazepam patients.

In a study by Cervera (1974), 60 patients with non-psychotic dementia were randomized in a parallel manner to receive thioridazine (mean dose = 79 mg/d) or diazepam (mean dose = 12 mg/d) for four weeks. For the 56 patients who completed the study, HAM-A and NOSIE scores improved significantly from baseline for both agents. There were no differences between the drugs on the HAM-A measure of agitation measure or the NOSIE subscale measure of irritability. Thioridazine produced superior improvement to diazepam on most HAM-A and NOSIE subscales, but these differences were not statistically significant. Sedation was reported in three diazepam patients.

Likewise, Covington (1975), in a 4 week, double-blind study compared thioridazine and diazepam in 40 geriatric patients with symptoms of dementia. All had varying degrees of confusion, lack of alertness, agitation, anxiety, insomnia, depressed mood, and irritability. The mean dose was 32.9 mg/d and 7.2 mg/d for thioridazine and diazepam, respectively. Twenty patients received each drug. Patients were rated using a modified HAM-A, NOSIE, and 2 global assessments. On the HAM-A, the thioridazine group experienced a reduction (p<0.05) in severity on 4 of 8 factors including anxious mood, fears, agitation, and interview behavior. In the diazepam patients, none of these 8 factors decreased significantly. On the NOSIE, the patients also improved more on thioridazine overall than on diazepam. In the diazepam patients, the NOSIE scores actually became worse on 6 of 8 factors. Globally, 65% of the thioridazine patient had some improvement while only 30% of the diazepam patients had a similar degree of improvement. This is the only study that suggests that low dose antipsychotic treatment is more effective than benzodiazepines in the treatment of agitation. No serious side effects were noted in these patients.

Benzodiazepine Summary
Patient Population: In general, each benzodiazepine study reviewed used an elderly dementia patient population with poorly defined, non-reproducible endpoints. The diagnoses of these patients, by current standards, were also poorly defined. The results of these studies should not be extrapolated to other BD populations such as mentally retarded patients.

Efficacy: Overall, if all benzodiazepine studies are combined, the response rate was 59% (167/285). The placebo response rate was 16% (20/125), and the antipsychotic (thioridazine) response rate was 54% (26/48). The efficacy reported in the diazepam studies is unclear because of the length of these studies (DeLemos et al 1965, Kirven and Montero 1973, Cervera 1974, Covington 1975). The 2-4 week study design used in many studies is probably not long enough for diazepam to reach steady-state serum concentrations in the elderly. As a result, full drug effect or adverse effects may not have become completely apparent by study end.

It is important to note that patients with dementia are often psychotic and that antipsychotics should be considered alone or in addition to a benzodiazepine. If the studies comparing antipsychotics and benzodiazepines in management of the chronically agitated patient with dementia are considered separately, then generally antipsychotics are considered more effective than benzodiazepines used alone in severely agitated patients (Risse and Barnes 1986, Kunik et al 1994).

Adverse Effects: Typical adverse effects of benzodiazepines include primarily central nervous system effects (i.e., drowsiness, confusion, fatigue, and ataxia) and may lead to impaired psychomotor performance and falls (Kruse 1990). The risk of falls with benzodiazepines has been extensively studied. A definite relationship between benzodiazepine use and falls has not been proven (Kruse 1990). However, several studies have reported that falls in the elderly may occur more often early in treatment, occur more frequently with long-acting benzodiazepines, and may be dose-related. Oxazepam or lorazepam, which are less likely to accumulate in the elderly, should be considered as primary choices. The issue of benzodiazepine-induced disinhibition was not addressed in these studies but it should also be considered as a potential adverse effect, particularly at lower doses.

Buspirone
The role of buspirone in patients with DAT and agitation has been examined in a few case reports in the literature (reference). The mean dosage of Buspirone ranged from 15-60 mg/day for up to 8 weeks. Overall a response rate of approximately 40% has been seen. The most common adverse drug reactions include headache, nervousness, and dizziness.

A more recently controlled trial (reference) compared haloperidol (1.5 mg/day) to Buspirone (15 mg/day) for agitation in DAT. Both of the groups showed improvements in behavior, with an additional reduction in tension and anxiety being seen in the busprione group.

Due to busirone's better side effect profile compared to the benzodiazpine's its use may be warrant in patients experiencing high anxiety leading to the behavioral disturbance.

Antidepressants

Trazodone
The antidepressant most commonly used in patients with DAT and agitation is Trazodone, although others have also been used. When looking at the literature, the use of Trazodone has been primarily reports in case reports and limited controlled trials. The usually dosage of trazodone ranges from 150-400mg/day, usually given at bedtime due to the sedative effects of the drug. According to the case reports available, the response rate seen with Trazodone is approximately 50%. The symptoms most commonly reported as improved include irritability, anxiety, restlessness, affective disturbances, and sleeplessness. The most common adverse drug reactions to Trazodone include orthostatic hypotension, sedation, and delirium. Thus, close monitoring of blood pressure, and mental status is needed during therapy.

Selective Serotonin Reuptake Inhibitors
Due to the favorable side effect profile seen with the selective serotonin reuptake inhibitors (SSRI), the use of other antidepressants such as the tricyclics have significantly decreased and are often considered contraindicated in elderly patients.

Most of the information on the SSRIs comes from case reports in the literature. Citalopram, Fluoxetine, Fluvoxamine, and Sertraline are the four SSRIs that have been used. When looking at these cases, only Citalopram and Sertraline have been reported as being effective for the treatment of behavioral disturbances in DAT. The range of dosages used for Citalopram was 20-30mg/day. The primary symptoms that improved were emotional bluntness, confusion, irritability, anxiety, fear/panic, depressed mood, and restlessness. For Sertraline approximately a 60% response rate has been seen with irritability, depression, and attention all noted as symptoms that improved. For Fluoxetine and fluvoxamine no specific effect of medication has been noted. In the Fluoxetine cases 20-40mg of was used for a total of 4 weeks. Based on the extended half-life of Fluoxetine in older subjects this is not long enough for the drug to reach a steady state level, which may have effect the study results. Additionally, the dosage of Fluoxetine used is rather high for geriatric patients and as such may also have contributed to the high side effect profile of confusion, agitation, dizziness, GI upset, and insomnia. Thus, similar to the antipsychotics, if an antidepressant is to be used in this population, careful consideration of dosage as well as duration needs to be taken. It is best to use the lowest dosage possible, and to titrate up from this starting dose slowly, to avoid the occurrence of adverse drug reactions, which may lead to a negative outcome.

Anticonvulsants
Although the majority of anticonvulsants do not have formal indications for the treatment of psychiatric conditions, their use is often employed due to their mood stabilizing effects. The majority of the literature supporting their use in behavioral disturbances in DAT comes from case reports and a few smaller controlled trials. Valproate, carbamazepine, and gabapentin have all been reported as being helpful in the reduction of agitation in DAT.

Carbamazepine has been studied in over 4 open trials, and two controlled trials. The range of dosages used is between 100-300mg/day for 4-5 weeks. The serum concentrations associated with these dosages has ranged from 3-4 mcg/ml to 4.5-5.7 mcg/ml. There have been no studies which equate serum concentrations of carbamazpine with therapeutic response. Overall, response rates for these reports range from 0-64%, with the higher response rates being associated with dosages around 300 mg/day. The most common adverse drug reactions seen were ataxia, dizziness, rashes, drowsiness, GI upset, diplopia, and confusion. Additionally, the use of carbamzapine in this older population carries with it the risk of seeing the syndrome of inappropriate antidiuretic hormone (SIADH) and as such monitoring of serum sodium may be appropriate. Tariot et al (1994) evaluated the effectiveness of carbamazepine in the treatment of agitated demented inpatients. Using a 5-week placebo controlled double-blind cross-over design 25 patients were randomized to either carbamazepine or placebo for 5 weeks, washed-out for 2 weeks and then switched to the other treatment. According to the CGI, 64% (16) of the carbamazepine patients were rated as improved while only 16% (4) of the patients improved while on placebo. The change in the BPRS scores were significant greater for carbamazepine than placebo. One subject developed carbamazepine-induced tics, and one died from pneumonia. The authors concluded that carbamazepine in low doses can reduce agitation in some but not all patients with limited ADRs. The usual starting dose of carmabazpine to be used is around 50 mg Q12h. This may be increased to 200 mg Q12h based on tolerability.

Valproate has also been examined for the treatment of BD in DAT in at least four open trials. The range of dosages used is 240 &endash; 2500 mg/day with serum levels ranging from 6.2 &endash; 54.7 mcg/ml. Similar to carbamazpine there been no studies which equate serum concentrations of valproate with therapeutic response. The response rates reported range from 50-100% and the most common adverse drug reactions include, nausea, vomiting, diarrhea, sedation, dizziness, tremors, transient alopecia, and weight gain. Lott et al (1995) reported a study in which 10 nursing home patients (71 &endash; 94 years old) were given open label valproate for the treatment of dementia and behavioral agitation. The dosage ranged from 375 &endash; 750 mg/day for 4 &endash; 34 weeks. Agitation was rated by the nursing staff members who found that 8 of the 10 patients had ³ a 50% decrease in the frequency of episodes. This response was maintained in all patients and side effects were minimal. The usual starting dose should be 125 mg/day, which may be increased to 125 mg/day every 3-6 days if tolerated.

Lastly, there is a case report using gabapentin for agitation in dementia. Gleason and Schneider (1990) treated 9 agitated antipsychotic-refractory Alzheimer's outpatients with CBZ 200-1000 mg/d (mean = 578 mg/d). The patients received no other psychotropic medications during the study. Behavioral agitation measured by the Brief Psychiatric Rating Scale (BPRS) decreased from 45 before CBZ to 36 during CBZ treatment. Five of the nine patients exhibited global improvement. Ataxia was reported in three patients.

Behavioral Disturbances in Patients with Developmental Delays

Lithium
Open Prospective Studies: Dostal and Zvolsky (1970) studied 14 phenothiazine-resistant severely mentally retarded aggressive adolescents. Lithium sulfate was titrated slowly over the course of 4 months to a mean dose of 45 mEq/d (equivalent to 1,667 mg of lithium carbonate). Concomitant antipsychotics were discontinued three months following the initiation of lithium treatment. After eight months of treatment, significant improvement compared to baseline was noted in 11 of the 14 patients' affect, aggressiveness, psychomotor activity, and undisciplined behavior. Patient response was poor for the first four months until they had been titrated to full doses of lithium with serum concentrations between 0.9-0.95 mEq/L. This therapeutic effect reportedly disappeared within two months of lithium discontinuation. Nine of the patients experienced significant polyuria resulting in frequent bedwetting, an adverse effect that may limit the usefulness of lithium in this population.

Similarly, Micev and Lynch (1979) treated 10 mentally retarded patients with aggressive and self-mutilating behavior with lithium doses that were titrated to concentrations in the range of 0.6-1.4 mEq/l in addition to their usual medications. Globally, five of the nine patients with aggressive tendencies experienced marked improvement, three experienced slight improvement and the ninth experienced no change. Six of the eight self-mutilating patients ceased this behavior while receiving lithium, one improved mildly, and one had no response. Interpretation of this study is difficult since the authors did not indicate if the doses of the patients' pre-existing medications were lowered or discontinued as a result of lithium treatment. Lithium adverse effects were also not discussed.

Double-Blind, Placebo Controlled Studies: Craft et al (1987) studied 42 hospitalized mentally handicapped patients aged 19-65 years with a history of aggression severe enough to require routine pharmacotherapy. Aggression levels were rated by the nursing staff using a 5-point scale. (1 = well behaved, 2 = mood uncertain, 3 = overt aggression or attempted aggression, 4 = additional medication required to control patient; and 5 = seclusion required). No further definition of aggression was given. Patients were randomly assigned to two parallel treatment groups, placebo (n=20) and lithium (n=22) for 12 weeks. The lithium treatment group received lithium carbonate titrated from a starting dose of 800 mg/d to maintain the serum concentration above 0.7 mEq/L. Lithium concentrations as high as 1.25 mEq/l occurred. The patients' pre-existing regimens, which included anti-convulsants and "tranquilizers" (presumably antipsychotics), were continued unchanged during the trial. A significant reduction in aggression was reported in 73% (16/22) of patients treated with lithium in contrast to 30% (6/20) of those treated with placebo. There were significant differences (p < 0.05) in mean weekly aggression scores (baseline score @ 2 for both groups; 12 week score @ 1.5 for lithium; the placebo group did not change). The frequency of aggressive episodes was also significantly lower in the lithium group although specific values were not reported. Three patients in the lithium group and six in the placebo group required additional, unspecified medication to control aggression. Lithium adverse effects of polydipsia, polyuria, tremor, drowsiness, incoordination and vomiting were reported in 36% of the lithium patients and 20% of the controls. The authors noted empirically that 6-8 weeks of lithium treatment was often necessary to see an antiaggressive effect.

Worrall et al (1975) treated eight chronically hospitalized mentally retarded female patients, 33-57 years old, with a history of frequent aggressive behavior. Two of the patients suffered from phenylketonuria, one from encephalitis, one from Down's syndrome, and four had idiopathic mental retardation. Nursing staff assessed the aggressive behavior on a 7 point scale with a reported high level of inter-rater reliability. Patients received placebo or lithium carbonate titrated to concentrations between 0.6-1.4 mEq/L for 4 week intervals over 16 weeks in a crossover fashion. The lack of washout between phases may have falsely elevated aggression scores early in the placebo phases and falsely depressed the scores early in the lithium phases. Concomitant psychotropic medications were unchanged during the study. Two patients with serum lithium concentrations of 1.2 and 1.5 mEq/l were withdrawn from the study because of ataxia, anorexia, and insomnia. Displays of aggression decreased significantly during lithium treatment in the remaining six patients. Three patients (50%) were less aggressive on lithium and were considered responders. The authors felt that two weeks were necessary before a response was likely. Placebo response was not reported.

In a double-blind crossover trial, Tyrer et al (1984) treated 25 persistently aggressive mentally handicapped subjects 14-50 years old with either lithium or placebo. Patients received placebo or sufficient lithium carbonate to produce serum concentrations between 0.5-0.8 mEq/L for two months in a crossover design with a one month washout between treatments. Assessment of the patients was performed by the nursing staff based upon 20 behavioral items in areas of aggression, hyperactivity, antisocial behavior and destructiveness. Pre-existing psychotropic medications were continued unchanged during the study. Sixty-eight percent of the patients were significantly less aggressive during the lithium phase compared to placebo. No episodes of toxicity or adverse effects were observed. The placebo response rate was not reported. The authors concluded that the results indicate that lithium is effective in controlling BD in mentally handicapped patients.

Lithium Summary
Patient Population: All of the lithium studies used mentally retarded patients as the study population. The efficacy of lithium in other populations is unknown.

Efficacy: From these studies, the overall improvement rates for lithium treatment of BD is: open prospective, 83% (19/23) and double-blind, placebo-controlled, 68% (36/53). A 30% placebo response rate (6/20) can be estimated only from the Craft et al study. Interpretation of these studies is clearly hindered by the fact that antipsychotics and anticonvulsants were administered in combination with lithium during the studies. It is not known if lithium alone will produce a similar response. Despite the flaws in these data (summarized in Table 1), a 2-6 week trial of lithium carbonate in sufficient doses to produce serum levels between 0.5 and 1.0 mEq/L is a reasonable clinical approach to treat BD in mentally retarded patients. Attempts to lower the antipsychotic dose should be delayed until the patient is stabilized on lithium.

Adverse Effects: Lithium concentrations greater than 1.0 mEq/l were associated with signs and symptoms of toxicity in some patients (Craft et al 1987, Worrall et al 1975). At lower concentrations, polyuria with subsequent enuresis was a significant problem in some studies (Dostal and Zvolsky 1970, Craft et al 1987). The potential for severe polyuria leading to dehydration with subsequent lithium intoxication and renal damage may be increased in this population of patients with poor communication skills. Typical lithium adverse effects, including gastrointestinal distress, hypothyroidism, and weight gain, should also be anticipated.

Anticonvulsants

Carbamazepine (CBZ)
Open Prospective Studies: Patterson reported the first two studies documenting the effectiveness of CBZ in the treatment of BD (Patterson 1987, Patterson 1988). Both reports studied CBZ titrated to antiepileptic blood concentrations (8-12 mg/ml) in the treatment of repeatedly assaultive patients with degenerative or traumatic dementias. The author failed to indicate whether the patients were receiving concomitant psychotropic medications. The patients' BDs were quantified by recording each episode of assaultive behavior. In the first study, 8 patients were observed for two 7-day periods before the start of CBZ treatment and then for two more 7-day treatments during CBZ treatment (Patterson 1987). The mean number of assaults/week decreased significantly from 3.4/week before CBZ to 1.6/week during CBZ treatment. Two patients each experienced the adverse effects of diplopia and ataxia. The second study of 13 patients was identical to the first except that there were also two additional 7-day observation periods after CBZ treatment (Patterson 1988). The mean number of assaults/week decreased significantly from 6.1/week before CBZ to 3.3 during CBZ treatment. In the two weeks following CBZ, assaults increased to 6.3/week. All 13 patients had fewer assaults while receiving CBZ. Two patients in this study also experienced ataxia and diplopia.

Double-Blind, Placebo Controlled Studies: In a seven month double-blind placebo controlled crossover trial, Reid et al (1981) studied 12 severely and profoundly mentally retarded patients with severe refractory behavioral problems. The patients were divided into two groups. Group A consisted of six patients in whom hyperactivity was the predominant BD. The remaining six patients (group B) were also hyperactive, but this symptom was part of a wider spectrum of BD including: eating disorder, irritability, self-injury, stereotypy, stripping, and noisiness. In the first 3 months, six randomly selected patients received CBZ titrated to blood concentrations of 6-10 mg/L while the other six received placebo. The medications were reversed after three months. Throughout the two 3 month periods the nursing staff rated the patients. The scale went from +3 "best I have known" to -3 "worst I have known." Zero was normal for the patient. The behavior ratings for each patient were averaged for each 3 month period and compared by a paired rank-sum test. One patient dropped out of the study due to medical illness. Overall, CBZ was no better than placebo in the remaining eleven patients. Three patients in Group A and one in Group B were considered significantly better while receiving CBZ compared with placebo. One patient was considered better on placebo. This response, however, was less than impressive. Average patient improvement was less than +1 (a little better than normal) in all eleven evaluable patients. Likewise, average deterioration was never less than -1. Changes of this magnitude, although statistically significant, are not clinically significant. Small sample size and three protocol violations in three patients also minimizes the results of this study. The authors did not report any serious adverse effects to CBZ.

Carbamazepine Summary
Patient Population: The patient populations of the carbamazepine studies had diagnoses of dementia and mental retardation.

Efficacy: From these studies, overall response rates to CBZ treatment is estimated as follows: open prospective, 80% (24/30) and double-blind, placebo-controlled, 56% (24/30). There was one placebo responder (9%, 1/11). Currently because of less available data, we recommend a therapeutic CBZ trial of at least 2 weeks is justified to treat BD if the patient had failed trials on lithium (for mentally retarded patients) and a b-blocker.

Adverse Effects: The only adverse effects mentioned in these studies were mild diplopia, tics and ataxia. There were no hematological complications.

B-Blockers
Uncontrolled Prospective Studies: Ratey et al (1986) utilized propranolol in the first prospective study of b-blockers in the treatment of assaultive or self-abusive behavior in 19 chronically hospitalized mentally retarded patients (mean age = 33 years) with IQs < 50 who were refractory to various drug, educational, and behavioral regimens. Eighteen of these patients were receiving antipsychotics before the study (mean dose = 723 mg chlorpromazine-equivalents). The propranolol dose was increased biweekly until either a behavioral endpoint was attained or hypotension and/or bradycardia necessitated termination of the dose titration. The effective dose range was 40 to 240 mg/day (mean = 120 mg/d). At an average of 6.6 months follow-up, eleven of 19 (58%) subjects had substantial improvement in the areas of self-abusive and aggressive behavior. An additional five patients improved moderately (overall improvement rate = 84%). Some of the primary behaviors that were diminished or completely extinguished included abusiveness (verbal and physical), assaultiveness (verbal and physical), hollering, restlessness, and gross avoidant behavior (fleeing contact). The average antipsychotic dose decreased 44% to 402 mg/d (chlorpromazine equivalents) in the 18 patients who were receiving antipsychotics at the beginning of the study. Four of these patients completely discontinued antipsychotics while receiving propranolol. In contrast to other similar studies, the authors commented that lower than expected effective doses of propranolol were possible because they were willing to wait longer for the anticipated endpoints. Hypotension and/or bradycardia occurred in 37% (7/19) of the patients of sufficient magnitude that they were unable to tolerate doses above 100 mg/day. Two of the three nonresponders were in this dose range. The authors hypothesized that the b-blockers' behavioral effects result from a lowering of the patient's level of arousal thereby lessening somatic anxiety with a subsequent decrease in aggression, regardless of the patient's diagnosis.

The lowered arousal hypothesis was supported in another uncontrolled study conducted by Ratey et al (1987) in which propranolol or nadolol were used to treat the impulsive or aggressive behaviors of eight autistic adults (mean age = 32 years). The b-blockers were added to the patient's current medications at 40 mg/day and raised to an average dose of 225 mg/day (range = 100-420 mg/d). Seven of the patients were also receiving antipsychotics at the beginning of the study (mean dose = 1074 mg/d chlorpromazine equivalents). All eight of the patients had increased attention span accompanied by decreased impulsive, aggressive, and ritualized behaviors. Six patients also had improved social skills and sought more human contact. At an average of 13 months follow-up, the average antipsychotic dose decreased 66% to 368 mg/d in the seven antipsychotic users. Antipsychotic discontinuation was achieved in one of these patients. Two of the patients had difficulty tolerating the drug. One dropped out of the study. Both of these studies suggest subjectively reductions in aggressive behavior and the antipsychotic dose in patients receiving _-blockers. However, interpretation of these results is difficult the data were not statistically analyzed.

Greendyke et al (1984)8 studied eight patients (mean age = 56 years) with organic brain disease of various etiologies including anoxia, trauma, stroke, senile dementia, and Huntington's disease. All had a history of physical attacks which were not responsive to treatment with antipsychotics, lithium, or sedatives. Propranolol was started at a dose of 80 mg/d and titrated upward at a rate of 40 mg every 3 days to a maximum dose of 520 mg/d (mean dose = 465 mg/d) in five of the eight patients. Decreased assaultive behavior occurred in seven patients. For the six patients in whom adequate data were collected the number of assaults noted per 20 day observation period decreased significantly from 6.8 to 1.1 per patient. Four of these patients had been maintained on antipsychotics before the study (mean dose in chlorpromazine equivalents = 931 mg/d). The antipsychotic dose decreased 79% in these four patients by the end of the study to 193 mg/d. The authors felt that improvement usually required one month of propranolol treatment to improve behavior. The propranolol dose was above 300 mg/d before improvement occurred in all seven patients who could tolerate the drug. One patient was unable to tolerate the hypotensive effects of the drug. In the remaining seven patients, a total of 151 of 2172 propranolol doses were withheld during the study due to bradycardia or hypotension.

Placebo-Controlled Studies: In a double-blind placebo controlled crossover study, Greendyke et al (1986) utilized propranolol to treat ten patients suffering from severe dementia of various etiologies who exhibited a lack of impulse control and frequent violent behavior. Sustained-release propranolol was titrated over 15-20 days to a dose of 520 mg/d. All scheduled psychotropic medications were discontinued two weeks before the propranolol was started. Oral paraldehyde or IM phenobarbital were used on a prn basis as "rescue medications" throughout the study. Nine patients completed both 11 week segments of the study (mean age = 52 years). One patient with uncontrollable behavior dropped out during placebo treatment and never received active treatment. Five patients improved markedly (55%), two improved moderately (22%), and two had little or no improvement in assaultive behavior (22%, p<0.05). The number of assaults and attempted assaults was significantly less during propranolol treatment compared to placebo. In the seven responders the total number of assaults fell from 88 during placebo treatment to 52 during propranolol treatment. Assault frequency in the non-responders was not reported. It was also noted that the intensity and duration of the outbursts appeared to decrease during propranolol treatment. Social interest, irritability, psychomotor retardation, and rescue medication use did not differ between placebo and propranolol. Seven (78%) of the patients manifested objective signs of hypotension and bradycardia. The authors concluded that propranolol was effective in controlling violent behavior in demented patients without causing sedation.

In a similar study, the same group studied pindolol in the treatment of 11 treatment refractory assaultive patients refractory to conventional pharmacotherapy suffering from organic brain syndrome (mean age 52 years) (1986). The protocol was identical to that of the previous study Greendyke et al 1986), except pindolol was used instead of propranolol. Pindolol was chosen for its partial agonist effect (intrinsic sympathomimetic activity) to theoretically minimize hypotension and bradycardia. The pindolol was titrated over 18-24 days to a dose of 60 mg/d and then maintained at this dose for 10 days. After this period, the pindolol dose was titrated to 100 mg/d to study the effects of higher doses. Patients were crossed over after pindolol was tapered in the first treatment group. The patients were rated on a subjective scale of 1 to 5 scale for behaviors including hostility, uncommunicativeness, uncooperativeness, and stereotypic behavior. The majority of the patients experienced a significant improvement in the areas of hostility (82%), uncommunicativeness (82%), uncooperativeness (78%), and repetitive behavior (82%). Compared to the authors' previous study, pindolol appeared to have a more rapid onset of action than propranolol with beneficial effects occurring within two weeks of treatment. The optimal dose range was thought to be 40-60 mg/day. At doses above this, some patients were reportedly "overstimulated." The pindolol group also required less rescue medication compared to placebo. In contrast to previous studies with propranolol, no episodes of bradycardia or hypotension occurred during pindolol treatment.

Lastly, Greendyke et al (1989) studied 15 organic patients with behavioral disturbances that included aggressiveness, resistance to care, and sexually inappropriate behavior. These patients were less severely disturbed than those in previous studies from this center. All antipsychotics and antidepressants were discontinued over a two week period. Using a cross-over design with a 1-week washout between treatments, patients received pindolol 40 mg/d and placebo for 10 weeks each. Of the 13 completers, insignificant decreases were seen in the four rating scales used to assess their behavior. On the Overt Aggression Scale, scores decreased in six of 13 patients, an insignificant finding. The suggested reason for this unimpressive finding was that the patients had less severe BD than those in their other studies. Nevertheless, the authors felt that eight of the 13 were improved during an open follow-up period, and ten of the 13 were placed at a lower level of care than before pindolol treatment. The authors noted that psychotropic medications were restarted during the follow-up period although lower doses were necessary than before pindolol treatment. They also noted that in their experience with pindolol in over 100 patients there have been no episodes of hypotension or bradycardia.

B-Blocker Summary
Patient Population: Of all of the drugs reviewed in this paper, b-blockers have been studied and found effective in more different patient populations than any other drug. These populations include patients with mental retardation, autism, dementia, and organic brain syndrome.

Efficacy: From these studies, overall response rates to b-blocker treatment of BD was calculated as follows: open prospective, 81% (39/48) and double-blind, placebo-controlled, 67% (22/33). A placebo response rate cannot be determined from the data available. Although they were not directly compared, propranolol and pindolol appeared to be comparable in efficacy in the treatment of BD. Other lipophilic b-blockers are also probably useful. Hydrophilic b-blockers, however, are not yet studied well enough to recommend their use. Unless adverse effects are an issue, a two to four week trial of propranolol, titrated until a clinical effect or adverse effects occur, should be the first choice since it is less expensive than pindolol. Otherwise, pindolol 40-60 mg/d may be used.

Adverse Effects: In the studies reviewed, propranolol caused clinically significant bradycardia or hypotension in 13-78% of patients (Ratey et al 1986, Greendyke et al 1989). This may limit propranolol's utility, particularly in older patients. Pindolol, however, was not associated with any cardiovascular effects and may be a reasonable alternative in patients who are unable to tolerate propranolol. Patients should also be monitored for other adverse effects, notably b-blocker-induced psychosis or mood changes.

Buspirone
Buspirone is a non-sedating anxiolytic which has recently become a popular treatment in many psychiatric disorders, including BD. Well-designed studies, however, have not yet been performed to support this indication. In an open study, Ratey et al (1989) reported behavioral improvement in nine of 14 developmentally disabled patients with mixtures of various diagnoses including alcoholism, autism, seizure disorder, phenylketonuria, and anxiety disorder who received buspirone 15 to 45 mg/d. In a second study, Ratey et al (1991) reported five of six mentally retarded patients improved behaviorally following treatment with buspirone at doses up to 45 mg/d. Although this study was blinded with a placebo baseline, the utility of this is doubtful since the researchers were unblinded during the middle phases of the 18-week study. The response rate combining these two studies was 70%. Until adequate studies are performed, however, buspirone is not recommended as a first-line treatment for agitation.

Naltrexone
In the late 1970s, it was noted that neonatal rats and chicks exposed to high levels of opiates showed autistic-like withdrawal after they were born. Opiate treated animals exhibit unusual motor flurries much like autistic children's hyperactivity. Additionally they exhibit other unusual postures and preservative behaviors and fail to evince normal separation anxiety when removed from their mothers. Likewise as previously noted by Ross et al (1987) there is an increase in CSF beta-endorphin (BE) activity in autistic patients. Researchers have noted similar behavior pattern between autistics and opiate addicts such as social withdrawal, self-stimulation, and high levels of pain tolerance. It is speculated that disturbances in brain opioid levels may block psychosocial development at its earliest stages leading to failures in language acquisition and other idiosyncrasies in learning. There are two theories as to why abnormal opoid levels cause self-injury, and why opiate antagonist drugs may reduce it. The pain theory suggests that self-injurious behavior is a form of self-stimulation, possibly elicited in response to reduced sensory stimulation, and that opiate blockers may attenuate self-injurious behavior by enhancing the feeling of pain. The addiction theory suggests the purpose of self-injurious behavior is to promote pain-induced release of opiates to achieve a 'high' that can be reversed by opiate antagonists. Thus the use of naltrexone in the treatment of autism is reasonable since it antagonizes endogenous opiate receptor activity. There are a large number of uncontrolled reports supporting the effectiveness of naltrexone in the treatment of autism. However, only two controlled trials are available and neither results are very encouraging.

Campbell et al (1990) conducted a double-blind, placebo-controlled study designed to assess critically the effects of naltrexone on behavioral symptoms and learning in autistic children, and its safety in 18 children (14 males, 4 females), ages 3-8 years. Subjects were randomly assigned to naltrexone (0.5-1.0 mg/kg/d) or placebo and received daily doses over a period of 21 days. Naltrexone was superior to placebo according to blind Clinical Global Consensus Ratings (unpublished scale). However, other behavioral rating measures did not confirm this result. There was only a suggestion that naltrexone reduced fidgety and hyperactive behavior and tended to alleviate overall symptomatology in older children. The explanation between the discrepancy between the CPRS/CGI measures and the Clinical Global Consensus Ratings was that the most dramatic and consistent changes during naltrexone administration in each child were decreases of withdrawal and increases of verbal production and communicative speech. It was felt that the Consensus rating took these symptoms into account whereas the CPRS and CGI did not. Naltrexone did not appear to affect discrimination learning. Results are preliminary and, owing to the small sample size, can be considered only suggestive until this study is completed or replication is obtained from independent research.

Scifo et al (1991) studied the effects of naltrexone in treatment of 12 autistic (DSM-III-R) children (7-16 years, 10 males). Doses of 0.5, 1.0, 1.5 mg/kg, and placebo were administered once every 48 hours in a single evening dose for 5 weeks each according to a randomized double-blind cross-over design. The behavioral evaluation was conducted using the specific Behavioral Summarized Evaluation and Childhood Autism Rating scales. Significant reductions of autistic symptomatology were observed in seven (58%) of the children. There was no correlation between the plasma _-endorphin levels and the clinical condition.

Zingarelli et al. (1992) examined the clinical effects of naltrexone on autistic behavior in a group of eight patients, 5 males and 3 females with an age range of 19 to 39 years old and an IQ ranging from 9 to 30. Using a double-blind crossover placebo controlled randomized assignment design, one group received naltrexone during the first and third treatment periods while the second group naltrexone during the second and fourth treatment periods. Each treatment period was 3 weeks long and after each period there was a one week washout placebo period. Naltrexone and its major metabolite 6-_-naltrexol have short half-lives of approximately 10 and 14 hours respectively. Naltrexone 50 mg/d po (0.6-1.1 mg/kg/d) was administered. The number of episodes of target symptoms were recorded at the end of each 8 hour A.M. and P.M. work shift and an observer time-sampled each patient for 10 minutes twice a week. No consistent patterns due to drug treatment were seen. One patient showed a decrease in all four of his time-sampled behaviors during active treatment. It was concluded that the drug has no clinical effect on the self-injurious behaviors and maladaptive idiosyncratic mannerisms which were monitored. However, if the results of the plasma BE concentrations are considered these findings are not surprising BE concentrations actually increased significantly with naltrexone treatment. Ross et al (1987) noted in their fenfluramine study that morphine pretreatment potentiates and naloxone antagonizes fenfluramine-induce depletion of striatal and hypothalamic serotonin stores. Thus it would seem that instead of an opiate antagonist, an opiate agonist ought to be administered to these patient to obtain a therapeutic effect.

ACUTE AGITATION/DELIRIUM

It has been reported that agitation and delirium affects almost 60% of patients with a critical illness (Fraser et al 1994). Rapid control of these patients is usually accomplished through the use of IV antipsychotics, typically haloperidol (Santos et al 1992). Unlike the low-potency antipsychotics (e.g., chlorpromazine, thioridazine), haloperidol has minimal effects on hemodynamic and respiratory function and is unlikely to cause or aggravate delirium (Fraser et al 1994). Though the IV route of administration is not FDA-approved it has become an accepted pharmacotherapeutic practice as >700 reports involving >2,000 patients have been published (Fraser et al 1994). Droperidol, a butyrophenone analog of haloperidol, has also been utilized clinically in the care of the agitated medical/surgical patient. However, it produces sedation, tachycardia, and orthostatic hypotension in greater than 10% of patients which makes it a less desirable agent.

Haloperidol
Tesar et al (1985) noted that haloperidol doses of 1-2 mg q4h were often times ineffective in the treatment of agitation in cardiac patients. He described four cases in which large doses of haloperidol were required for control of agitation. The effective IV doses in the four cases were 140 mg/d; 20 mg every 4 hours; 485 mg over 8 hours; and daily doses of 80, 285, 130, 460, and 530 mg. No cardiovascular or neurologic ADRs were observed.

Moulaert (1989) administered IV haloperidol to six acutely agitated postoperative (vascular surgery) patients. The initial dose was 5 mg IV. The dose was doubled every 30 minutes until the patient was sedated. The protocol was repeated if the patients became agitated again. Twelve protocols were required for the six patients. The mean dose of haloperidol was 38 mg/patients or 19 mg/protocol. Overall, it usually required 60 minutes to observe significant improvement in the patient's agitation. Increased respiratory rate, pulse, and blood pressure returned to normal after administration of the haloperidol. No side effects were noted.

More recently Tesar et al (1991) have recommended an initial haloperidol dose of 0.5 to 10 mg and doubling the dose every 15 to 20 minutes until agitation is controlled. They suggest that individual bolus doses >50 mg are rarely necessary.

Another group suggested that an acceptable initial haloperidol dose is 5 mg administered as either an IV bolus given slow push over 1 minute or an IV drip run in over 10 minutes (Santos et al 1992). IV haloperidol's onset of action ranges from 10-30 minutes. The IV line should be flushed after administration since physical incompatibilities can result in the precipitation of drugs such as haloperidol and phenytoin. If the patient remains agitated after 20-30 minutes, it is recommended to double the dose every 20 minutes until the patient is sedated. The maximum recommend dose is 75 mg per hour. Once the effective bolus dose has been established, the haloperidol can be safely administered by an IV "drip" infusion rate of 10-12 mg/h, though doses as high as 18 mg/h have been used (Riker et al 1994). Published doses of continuous infusion haloperidol have ranged from 1 mg/h to 40 mg/h for as long as 12 days (Fraser et al 1994).

Haloperidol is rarely associated with cardiac effects such as conduction delays and arrhythmias (Fraser et al 1994). However, Metzer and Friedman (1993) have reported three cases of torsades de pointes that developed during treatment with IV haloperidol. Total haloperidol doses of 115, 490, and 825 mg were associated with an 28%, 31%, and 22% increase, respectively in the QTc interval. All three patients had cardiac disease and dilated ventricles by echocardiogram. There are also five additional cases of Torsades de Pointes associated with the use of oral haloperidol. Patients with torsades de pointes may develop either ventricular fibrillation or tachycardia.

Additionally, the blood pressure should be monitored routinely. Although rare, there are patients who will become hypotensive after receiving IV haloperidol. After the blood pressure returns to baseline, a rechallenge with a smaller dose may be attempted. If hypotension still occurs then IV lorazepam alone might be considered.

Adverse reactions such as antipsychotic malignant syndrome and extrapyramidal side effects (EPS) are reportedly less common with IV haloperidol in contrast to oral and IM administration in critically-ill patients (Menza et al 1987, Fraser et al 1994). Menza et al (1987) measured EPS intensity in four patients receiving intravenous haloperidol and six patients receiving oral haloperidol. The patients receiving intravenous haloperidol experienced significantly less intense EPS than did the patients receiving oral haloperidol. Some tenuous theories proposed to explain the low rate of EPS include IV administration produces lower metabolites levels that may be responsible for EPS, that critically ill patients have low levels of acetylcholine and the dopamine blocking properties reestablishes a relative balance, and that the use of benzodiazepines in combination reduces dystonia and akathisia. However, in a prospective study of 38 AIDS patients treated with IV haloperidol and lorazepam, a 50% rate of extrapyramidal side effects was detected (Fernandez et al 1989).

Recommended concentration limits of haloperidol lactate admixtures is based solely upon visual incompatibility studies (Fraser et al 1994). The recommended concentrations 3 mg/ml for 5% dextrose, 1 mg/ml for lactated ringer's, 0.46% sodium chloride, 5% dextrose with 0.22% sodium chloride, and 0.75 mg/ml for 0.9% sodium chloride.

Droperidol
Droperidol is a butyrophenone analog of haloperdiol that often is used in the management of acute agitation due to its sedative properties. In a study by Rosen et al 1997, 46 consecutive combative/agitated patients " in the field" were randomized to receive either droperidol 5mg IV or placebo for their agitation. Assessments were done at baseline, 5 minutes, and 10 minutes consisting of a 5 point agitation scale (1 = none and 5 = violent), the Glasgow Coma Score, and vital signs. Only 5% of the patients included in this study were psychotic. In the others, agitation was contributed to trauma and medical emergencies. The mean baseline agitation scores ranged from 3.61 &endash; 3.65. These scores dropped to 1.04 in the droperidol group and 2.57 in the placebo group at the 10 minutes evaluation. This difference was statistically significant (p< 0.001). Interesting, there were no statistical differences in mean systolic blood pressure between the two groups after droperidol administration. Thus, using droperidol may be a relatively safe option in agitated patients.

Benzodiazepines
Because of the need for parenteral dosage forms, this limits the number of available benzodiazepines for the management of agitation (i.e., lorazepam, diazepam, midazolam). Most studies use lorazepam as the benzodiazepine of choice.

Yudofsky et al (1990) has recommended a 6-week schedule for the use of lorazepam in elderly patients with dementia. A recommended schedule is 1 to 2 mg po or IM q1h until the patient is sedated. If IV is required, then 2 mg/min IV push is repeated q30m until sedated. The recommended maintenance dose is 2 mg po or IM three-times-a-day. If the patient is not agitated for 48 hours then the total daily dose should be tapered at a rate of 10% per day. If agitation resumes, then consider etiological causes. After 6 weeks, if lorazepam can not be discontinued then consider an agent for managing chronic agitation.

Salzman et al (1991) compared the effectiveness of single doses of lorazepam 2 mg IM stat versus haloperidol 5 mg IM stat for the treatment of acute assaultive and/or aggressive behavior in psychotic inpatients. Each treatment group was composed of 30 subjects most of whom were already receiving antipsychotics for diagnoses that included schizophrenia, bipolar affective disorder, schizoaffective disorder, or organic psychosis. According to evaluation at 2 hours and 24 hours after the injection, there were no differences in the changes in the BPRS, CGI, and Overt Aggression Scale. The effect of lorazepam was independent of sedation. Predictably, the haloperidol patients experienced more EPS than the lorazepam patients during the 24 hour follow-up period.

Antipsychotic / Benzodiazepine Combination
Most literature on the combination treatment of acute agitation in delirium uses haloperidol and lorazepam due to their parenteral availability.

Adams et al (1986) described a group of 25 acutely ill delirious cancer patients (22-75 years) who had been referred to his group for evaluation of acute behavioral changes. In all cases the EEG confirmed the clinical finding of moderate to sever diffuse cerebral slowing. Sixteen of the patients had been intubated and were on mechanical ventilators at the time of referral. The patients' agitated behaviors were treated with a combination of IV haloperidol and lorazepam. Total IV 24-hour doses ranged from 100-480 mg for haloperidol and 36-480 mg for lorazepam. The highest prolonged combination dosage of the two drugs was 10 mg each hourly for 15 consecutive days. Mild to moderate levels of sedation were achieved in 24 of the 25 patients within the first 90 minutes. No patent experienced any respiratory, cardiac, or hemodynamic ADRs during treatment. Continuous monitoring showed than in most cases blood pressures and pulses returned normal limits within after the first one or two doses of the combination. Arterial blood gases and tidal volume improved significantly in seven respirator assisted patients once they were adequately sedated. Interestingly, the authors reported that patients who were deeply sedated with haloperidol and lorazepam revived spontaneously if they experience pulmonary deterioration. The authors noted that the combination did not depress the respiratory drive nor mask sudden pulmonary failure.

Adams (1988) recommended starting treatment with haloperidol 5 mg and lorazepam 0.5 mg given over 1 minute. If the patient shows an unacceptable response within 20 minutes, then haloperidol 5 mg and lorazepam 0.5-2.0 mg are administered. Little or no response again within 20 minutes should be treated with haloperidol 10 mg and lorazepam 2-10 mg every half hour until the patient is soundly sedated (e.g., unresponsive to verbal stimuli, but responsive to vigorous flexor-extensor movements of the arms). When the patient is sedated the lorazepam is discontinued and the haloperidol dose is reduced by 50%, and the time between injections is doubled. If the patient becomes restless, then haloperidol and lorazepam are restarted at the previous highest effective dose and administered every one to three hours for the next 10 to 12 hours, at which time the previously-described tapering procedure can be performed. He indicated that most patients respond to treatment within 10 to 60 minutes of initiation of treatment. However, in severe agitation, haloperidol and lorazepam each were administered in doses of 240 mg/d for over two weeks without adverse effects. The author notes that typically less than 100 mg/d of each drug is necessary. In his experience doses >10 mg/hr of each drug do not increase the response rate.

However, there are two problems with benzodiazepine use in the ICU patient. First excess sedation may be a problem if too large a dose is used, thereby resulting in the patient being less arousable. Finally tolerance eventually develops to the sedative effect.

Garza-Trevino 1989 concluded that haloperidol 5 mg and lorazepam 4 mg IM in combination were more effective than either drug alone. The effect was most obvious within the first 30-60 minutes of treatment.

Haloperidol and lorazepam remain stable in a IV bag for 4-6 hours and may be mixed in the same syringe (Santos et al 1992).

Thus, in the treatment of acute organic agitation both the antipsychotics and benzodiazepines have been found effective. A third approach to this problem is the combination of these two treatments, which often results in the utilization of lower dosages of each. Thus, if the use of one treatment does not result in control of the agitated behaviors, use of the second is warranted. Once patients have been on a stable dose of either agent alone or in combination for at least 48 hours, a prudent taper of approximately 10% per day should be undertaken. If behaviors reappear, then the dose of each medications should be titrated up until the desired result is achieved. This process should then be repeated, if possible, until patients are successfully taken off all agitation medications.

CONCLUSIONS - NON-ANTIPSYCHOTICS

Interpretation of the studies presented in this review is hampered by poor study design in nearly all of the studies. DerSimonian et al (1982) have published guidelines for appropriate reporting of study methods and results to allow for the adequate further analysis of clinical trials by the reader. Items that need to be clearly defined include 1) reporting of eligibility criteria for the study; 2) entry into the study before allocation to a treatment group; 3) random allocation to treatment groups; 4) the method of randomization; 5) methods of blinding the patients and the investigators; 6) a clear description of treatment complications; 7) itemization and explanation of study drop-outs; 8) statistical analyses and methods, and 9) power calculations. As a rule, the articles cited in this review fail to report their handling of many of these items. The long-term efficacy of these drugs is also unclear. Many of the studies are small and open, therefore possibly overestimating the efficacy of the drugs reviewed. Future studies will benefit from the use of a homogenous patient population, preferably using well-accepted diagnostic criteria to make the diagnoses. Many of the reports studied poorly defined populations with behavioral disorders of diverse etiologies. Future studies should study narrow the diagnostic populations as much as possible, for example, autism, Alzheimer's disease, or alcoholic dementia. In some cases, such as with mental retardation, which has many causes, this will be difficult. Furthermore, adequate endpoint measurements should be employed to prevent clouding of the results. Use of a validated psychometric rating scale such as the Overt Aggression Scale, is best suited for these studies (Yudofsky et al 1986). In determining response rates, an arbitrary measure of response should be agreed upon before the study is commenced to allow for objective determination of response. Inter-rater reliability, concomitant drug use, dropout rate, and placebo response rate should also be reported.

Since the studies reviewed generally fail to meet the above criteria, the reader is strongly cautioned to view the response rates found in this review with skepticism. The intent of this review was to propose pharmacotherapeutic alternatives to the antipsychotics in the treatment of agitation in behaviorally disturbed patients. Objective measures of improvement (such as the Overt Aggression Scale) should be used as the primary assessment of change in behavior.

Of the reviewed drugs, lithium and lipid soluble b-blockers are the best studied. Water soluble b-blockers are not studied well enough to recommend their use. CBZ, in particular, has been poorly studied and should be reserved primarily for those agitated patients with concomitant seizure disorders. The benzodiazepine studies are old, but oxazepam or lorazepam may be useful. Sedation was also a recurrent problem in these studies. Buspirone should not be prescribed until controlled studies with positive findings are available in the literature.

The studies summarized provide alternatives to antipsychotics in the treatment of BD. It may be that the use of these drugs may at least allow for reduced antipsychotic dosages. Well designed studies are clearly indicated to clarify the extent of the utility of these drugs.

Table 1. Controlled and uncontrolled studies examining the efficacy of non-neuroleptics in the treatment of aggression in non-affectively ill patients.
STUDY	DIAGNOSIS	TREATMENT	DESIGN	OUTCOME
Dostal and Zvolsky, 1970	14 mentally retarded age = 11-17 years (mean = 14 years)	Li = 0.3-0.95 mEq/L	open 8 months	79% (11/14) decreased aggression
Micev and Lynch, 1974	10 mentally retarded age = 16-58 years	Li = 0.6-1.4 mEq/L	open 12 weeks	89% (8/9) decreased aggression 75% 6/8 decreased self-mutilation
Craft et al, 1987	42 mentally handicapped age = 19-65 years (mean = 33 years)	Li = 0.7-1.0 mEq/L	DB, PB* parallel groups 16 weeks	Li = 73% (16/22) placebo = 30% (6/20) decreased aggression
Worral et al, 1975	8 mentally deficient age = 33-57 years (mean = 44 years)	Li = 0.6-1.4 mEq/L	DB, PB, crossover 16 weeks	Li = 50% (3/6) placebo = ?% decreased aggression
Tyrer et al, 1984	25 mentally handicapped age = 14-50 (mean = 27 years)	Li = 0.5-0.8 mEq/L	DB, PB, crossover 5 months	68% (17/25) placebo = ?% decreased aggression
Ratey et al, 1986	19 mentally retarded age = 22-49 years (mean = 33 years)	propranolol 40-240 mg/d female mean = 116 mg/d male mean = 130 mg/d	open 6.6 months	84% (16/19) moderate to pronounced improvement
Ratey et al, 1987	8 autism age = 25-50 years (mean = 32 years)	propranolol or nadolol 100-420 mg/d mean = 225 mg/d	open 13 months	100% (8/8) decreased aggression Allowed lower neuroleptic doses.
Greendyke et al, 1984	8 organic brain syndrome age = 43-79 years (mean = 56 years)	propranolol 200 to 520 mg/d	open 6 months	87% (7/8) decreased aggression Cardiovascular side effects limited dosage in some
Greendyke et al, 1986	9 dementia age = 27-75 years (mean = 52 years)	propranolol 520 mg/d x 11 wk	DB, PB, crossover 6 months	Improvement 55% (5/9) marked 22% (2/9) moderate
Greendyke et al, 1986	11 dementia age = 28-76 years (mean = 52 years)	pindolol 10-100 mg/d optimal = 40-60 mg/d	DB, PB, crossover @ 3 months	Improvement 82% (9/11) decreased hostility
Greendyke et al, 1989	15 organic brain syndrome age =38-81 years (mean = 64 years)	pindolol 40 mg/d	DB, PB, crossover. Open follow-up 8 months	Improvement Controlled phase: 45% (6/13) N.S.** Open phase: 62% (8/13)
Patterson, 1987	8 degenerative or traumatic dementia age = 37-66 years (mean = 52 years)	carbamazepine 8-12 mg/L	open 1 month	75% (6/8) decreased assaultiveness
Patterson, 1988	13 dementia age = 48-69 years (mean = ? years)	carbamazepine 800 mg/d 8-12 mg/ml	open 6 weeks	100% (13/13) decreased assaultiveness
Gleason and Schneider, 1990	9 Alzheimer's Disease age = 60-85 years (mean = 72 years)	carbamazepine 200-1000 mg/d	open length = ?	56% (5/9) decreased agitation
Reid et al, 1981	12 mentally retarded age = 14-43 years (mean = 30 years)	carbamazepine 5.9-9.9 mg/ml	DB, PB cross-over 7 months	carbamazepine 36% (4/11) placebo 9% (1/11)
Sanders, 1965	94 dementia age = 54-94 years (mean = 80 years)	oxazepam 10-40 mg/d	DB, PB parallel groups 2 months	> fair improvement @ 4 weeks oxazepam = 47% (22/47) placebo = 24% (9/38)
Beber, 1965	100 dementia age = 67-94 years (mean = 79 years)	oxazepam 10-40 mg/d	DB, PB parallel groups 2 months	> moderate improvement oxazepam = 68% (34/50) placebo = 18% (9/50)
Chesrow et al, 1965	148 dementia age = 33-97 (mean = ?)	oxazepam 35 mg/d chlordiazepoxide 30 mg/d	DB, PB parallel groups 6 months	> moderate improvement oxazepam = 87% (48/55) chlordiazepoxide = 51% (36/71) placebo = 0% (0/22)
De Lemos et al 1965.	29 chronic brain syndrome or dementia age = 51-96 years (mean = 66 years)	diazepam 7.5 mg/d	DB, PB parallel groups 1 month	Improvement diazepam = 71% (10/14) placebo = 13% (2/15)
Kirven et al, 1973	56 dementia age = "elderly"	diazepam mean = 9 mg/d thioridazine mean = 39 mg/d	DB parallel groups 1 month	diazepam 39% (11/28) thioridazine 47% (13/28)
Cervera, 1974	56 dementia age = "elderly"	diazepam mean = 12 mg/d thioridazine mean = 79 mg/d	DB parallel groups 1 month	Improvement on HAM-A and NOSIE scales.+
Covington, 1975	40 dementia mean age = 80 years	thioridazine 10-80 mg/d diazepam 4-18 mg/d	DB parallel groups 1 month	Improvement thioridazine = 65% (13/20) diazepam = 30% (6/20)
Ratey et al, 1989	14 developmental disability age = 23-63 years (mean = 35 years)	buspirone 15-45 mg/d	open length = ?	Improvement 65% (9/14)
Ratey et al, 1991	6 mentally retarded age = 21-42 years (mean = 29 years)	buspirone 15-45 mg/d	partially blinded 18 weeks	Improvement 83% (5/6)

*DB = Double-blind; PB = Placebo
**N.S. = Not significant
#OAS = Overt Aggression Scale
+HAM-A = Hamilton Anxiety Scale; NOSIE = Nurses' Observation Scale for Inpatient Evaluation

REFERENCES

Adams F, Fernandez F, Anersson BS (1986). Emergency pharmacotherapy of delirium in the critically ill cancer patient. Psychosom 27 (suppl 1):33-7.

Adams F (1988). Emergency intravenous sedation of the delirious, medically ill patient. J Clin Psychiatry 49(Suppl):22-6.

Beber CR (1965). Management of behavior in the institutionalized aged. Dis Nerv Sys 26:591-595.

Campbell M, Anderson LT, Small AM, et al (1990). Naltrexone in autistic children: a double-blind and placebo controlled study. Psychopharmacol Bull 26:130-5.

Cervera AA (1974). Psychoactive drug therapy in the senile patient: controlled comparison of thioridazine and diazepam. Psychiatry Digest 1974;35:15-21.

Chesrow RJ, Kaplitz SE, Vetra H, Breme JT, Marquardt GH (1965). Blind study of oxazepam in the management of geriatric patients with behavioral problems. Clin Med 13:1001-5.

Covington J S (1975). Alleviating agitation, apprehension, and related symptoms in geriatric patients: A double-blind comparison of a phenothiazine and a benzodiazepine. South Med J 68: 719-724.

Craft M, Ismail A, Krishnamurti D, Mathews J, Regan A, Seth RV, et al (1987). Lithium in the treatment of aggression in mentally handicapped patients. A double blind trial. Br J Psychiatry 150: 685-9.

De Lemos GP, Clement WR, Nickels E (1965). Effect of diazepam suspension in geriatric patients hospitalized for psychiatric illnesses. J Am Geriatric Soc 13:355-9.

DerSimonian R, Charette LJ, McPeek B, Mosteller F (1982). Reporting on methods in clinical trials. N Engl J Med 306:1332-7.

Dostal T, Zvolsky P (1970). Antiaggressive effect of lithium salts in treatment of severe mentally retarded adolescents. Int Pharmacopsychiat 5:203-7.

Druckenbrod RW, Rosen J, Cluxton RJ (1993). As-needed dosing of antipsychotic drugs: Limitations and guidelines for use in the elderly agitated patient. Ann Pharmacother 27:645-8.

Fernandez F, Levy JK, Mansell PWA (1989). Management of delirium in terminally ill AIDS patients. In J Psychiatry Med 19:165-72.

Fraser GL, Riker RR (1994). Controlling severe agitation in the critically ill with intravenous haloperidol. Hosp Pharm 29;689-21

Gleason RP, Schneider LS (1990). Carbamazepine treatment of agitation in Alzheimer's outpatients refractory to antipsychotics. J Clin Psychiatry 51:115-8.

Greendyke RM, Berkner, Webster JC, Gulya A (1989). Treatment of behavioral problems with pindolol. Psychosomatics 30:161-5.

Greendyke RM, Kanter DR (1986). Therapeutic effects of pindolol on behavioral disturbances associated with organic brain disease: a double blind study. J Clin Psychiatry 47: 423-6.

Greendyke RM, Kanter DR, Schuster DB, Verstreate S, Wootton J (1986). Propranolol treatment of assaultive patients with organic brain disease. J Nerv Ment Dis 174: 290-4.

Greendyke RM, Schuester DB, Wooton JA (1984). Propranolol in the treatment of assaultive patient with organic brain disease. J Clin Psychopharmacol 4:282-5.

Jeste DV, Eatham JH, Lacro JP et al (1996). Management of late-life psychosis. J Clin Psychiatry; 57 (suppl3): 39 &endash; 45.

Kirven LE, Montero EF (1973). Comparison of thioridazine and diazepam in the control of nonpsychotic symptoms associated with senility: double-blind study. J Am Geriatr Soc 21:546-551.

Kruse WHH (1990). Problems and pitfalls in the use of benzodiazepines in the elderly. Drug Safety 5:328-44.

Kunik ME, Yudosky SC, Silver JM, Hales RE (1994). Pharmacologic approach to management of agitation associated with dementia. J Clin Psychiatry 55(Suppl):13-7.

Menza MA, Murray GB, Holmes VF, Fafuls WA (1987). Decreased extrapyramidal symptoms with IV haloperidol. J Clin Psychiatry 48:278-80.

Metzger E, Friedman R (1993). Prolongation of the corrected QT and torsades de pointes cardiac arrhythmia associated with IV haloperidol in the medically ill. J Clin Psychopharmacol 13:128-32.

Micev V, Lynch DM (1979). Effect of lithium on disturbed severely mentally retarded patients (letter). Br J Psychiatry 125:110

Moulaert P (1989). Treatment of acute nonspecific delirium with IV haloperidol in surgical intensive care patients. Acta Anaesthesiol Belg 40:183-6.

Oberholzer AF, Henricksen C, Monsch AU, et al (1992). Safety and effectivenss of low dose clozapine in psychogeriatric patients: a preliminary study. Int Psychogeriatr; 4: 187 &endash; 195.

Patterson J F (1988). A preliminary study of carbamazepine in the treatment of assaultive patients with dementia. J Geriatr Psychiatry Neurol 1:21-23.

Patterson JF (1987). Carbamazepine for assaultive patients with organic brain disease. Psychosomatics 28:579-81.

Ratey J, Sovner R, Parks A, Rogentine K (1991). Buspirone treatment of aggression and anxiety in mentally retarded patients: a multiple-baseline, placebo lead-in study. J Clin Psychiatry 52:159-62.

Ratey JJ, Bemporad J, Sorgi P, Bick P, Polakoff S, O'Driscoll G, et al (1987). Brief report: open trial effects of beta-blockers on speech and social behaviors in 8 autistic adults. J Autism Dev Disord 7: 439-46.

Ratey JJ, Mikkelsen EJ, Smith B, Upadhyaya A, Zuckerman S, Martell D et al (1986). _-Blockers in the severely and profoundly mentally retarded. J Clin Psychopharmacol 6:103-7.

Ratey JJ, Sovner R, Mikkelsen E, Chmielinski HE (1989). Buspirone therapy for maladaptive behavior and anxiety in developmentally disabled persons. J Clin Psychiatry 50:382-4.

Reid AH, Naylor GJ, Kay DS (1981). A double-blind, placebo c controlled, crossover trial of carbamazepine in overactive, severely mentally handicapped patients. Psychol Med 11:109-113.

Riker RR, Fraser GL, Cox PM (1994). Continuous infusion haloperidol controls agitation in critically ill patients. Crit Care Med :433-40.

Risse SC, Barnes R (1986). Pharmacologic treatment of agitation associated with dementia. J Am Geriatr Soc 34:368-76.

Rosen CL, Ratliff AF, Wolfe RE, et al (1997). The efficacy of intravenous droperidol in the prehospital setting. J Emer Med 15:13-7.

Ross DL, Klykylo WM, Hitzemann R (1987). Reduction of elevated CSF beta-endorphin by fenfluramine in infantile autism. Pediatr Neurol 3:83-6.

Salzman C, Solomon D, Miyawaki E, et al (1991). Parenteral lorazepam versus parenteral haloperidol for the control of psychotic disruptive behavior. J Clin Psychiatry 52:177-80.

Sanders JF (1965). Evaluation of oxazepam and placebo in emotionally disturbed aged patients. Geriatrics 20:739-746.

Santos AB, Wohlreich MM, Pinosky ST (1992). Managing agitation in the critical care setting. J S C Med Assoc 88:386-91.

Satterlee WG, Reams SG, Burns PR, Hamilton S, Tran PV, and Tollefson GD (1995). A clinical update on olanzapine treatment in schizophrenia and in elderly Alzheimer's disease patients (abstract). Psychopharmacol Bull; 31: 534.

Schneider LS, Pollock VE, Lyness SA (1990). A metaanalysis of controlled trials of antipsychotic treatment in dementia. J Am Geriatr Soc 38:553-63.

Schneider L, Yeung P, Schweitzer D et al (1999). Quetiapine may reduced hostility in patietns with psychoses relation to Alzheimer's disease (abstract). Schizophrenia Res; 36 295.

Scifo R, Batticane N, Quattropani, MC et al (1991). A double-blind trial with naltrexone in autism. Brain Dysfunct 4:310-7.

Street J, Clark WS, Gannon KS et al (1999). Olanzapine in the treatment of psychosis and behavioral disturbances associated with Alzheimer's disease (abstract). Schizophrenia Res; 36: 298.

Tariot PN, Erb R, Leibovici, et al. (1994). Carbamazepine treatment of agitation in nursing home patients with dementia: a preliminary study. J Am Geriatr Soc 42:1160-6.

Tesar GE, Murray GB, Cassem NH (1985). Use of high-dose intravenous haloperidol in the treatment of agitated cardiac patients. J Clin Psychopharmacol 5:344-7.

Tesar GE, Stern TA (1991). The diagnosis and treatment of agitation and delirium in the ICU patient. In Rippe JM, Irwin RS, Alpert JS, Find MD, eds. Intensive Care Medicine, 2nd ed, Boston, Little Brown & Co 1865-75.

Tyrer SP, Walsh A, Edwards DE, Berney TP, Stephens DA (1984). Factors associated with a good response to lithium in aggressive mentally handicapped subjects. Prog Neuropsychopharmacol Biol Psychiatry 8: 751-5.

Worrall EP, Moody JP, Naylor GJ (1975).. Lithium in non-manic depressives: antiaggressive effect and red blood cell lithium values. Br J Psychiatry 126: 464-8.

Yudofsky SC, Silver JM, Jackson W, Endicott J, Williams D (1986). The overt aggression scale for the objective rating of verbal and physical aggression. Am J Psychiatry 143:35-9.

Yudofsky SC, Silver JM, Hales RE (1990). Pharmacologic management of aggression in the elderly. J Clin Psychiatry 51 (Suppl):22-8.

Zingarelli G, Ellman G, Hom A, et al (1992). Clinical effects of naltrexone on autistic behavior. Am J Ment Retard 97:57-63.

Title Page