Updated: Jun 09, 2017
  • Author: Leslie L Citrome, MD, MPH; Chief Editor: David Bienenfeld, MD  more...
  • Print


Most patients with mental disorders are not aggressive. Nonetheless, epidemiological evidence points to an increased risk for violence among individuals with a mental disorder compared with the general population. This article reviews this evidence and provides a framework for the assessment and treatment of these individuals.

Aggressive behavior in patients with psychiatric disorders has many possible causes [1] . Probably the most important causes are the presence of comorbid substance abuse, dependence, and intoxication. In addition, the disease process itself may produce hallucinations and delusions, which may provoke violence. Often, poor impulse control related to neuropsychiatric deficits may facilitate the discharge of aggressive tendencies. Finally, underlying personality characteristics, such as antisocial personality traits, also may influence the use of violent acts as a means to achieve certain goals. Environmental factors that are associated with aggressive behavior include a chaotic or unstable home or hospital situation, which may encourage maladaptive aggressive behaviors. Individuals may become aggressive for different reasons at different times.

Terms such as aggression, violence, crime, and hostility are observed in medical literature. Aggression is used for both humans and animals. In humans, aggression can denote verbal aggression, physical aggression against objects, or physical aggression against people. At times, aggression towards oneself (self-mutilation, suicidal gestures or acts) is included in the definition. Violence is used only when describing human behavior and denotes physical aggression by one person against another. Crime is defined as the intentional violation of criminal law. Hostility is a loosely defined term and can refer to aggression, irritability, suspicion, uncooperativeness, or jealousy.

This article mentions a number of medications, including new intramuscular formulations of second-generation antipsychotics, as well as an inhaled formulation of loxapine approved for the treatment of agitation in patients with bipolar disorder or schizophrenia. Full prescribing details, including precautions, adverse effects, use in pregnancy, and recommended dosing, can be found in the manufacturer's product information sheet and is available in the Physicians' Desk Reference.

Case study

Charles is a 35-year-old white male, diagnosed at different times as having schizophrenia, bipolar disorder, schizoaffective disorder, and posttraumatic stress disorder. His first psychiatric hospitalization was at age 25 when he was arrested for trespassing and was found to be incoherent. On the current admission, he presented to the emergency department stating that someone was following him. He became acutely agitated and was given an intramuscular injection of haloperidol 5 mg. He subsequently complained of stiffness in his neck and tongue and was drooling. He received diphenhydramine 50 mg intramuscularly for this dystonic reaction, followed by oral benztropine 2 mg. Once on the psychiatric inpatient unit, Charles refused all oral medications, saying he was allergic to Haldol. The next day Charles got into an argument with another patient about invading his personal space and shoved him away. This presented several problems:

1. What can Charles be given immediately to decrease his degree of agitation, without further alienating him?

2. What can Charles be given on a long-term basis to decrease the frequency and intensity of his episodes of agitation?



The probability of violent behavior among patients with mental disorders is greater than that for the general population. Although patients with mental disorders do not commit most violent crimes (a study from Sweden found that patients with severe mental illness commit 1 in 20 violent crimes [2] ), they are at an increased relative risk for committing them. This is especially true for patients with mental disorders and substance abuse problems. The epidemiological studies supporting this statement have been conducted around the world and include a variety of cultures.

Swinson et al studied homicide perpetrators in England and Wales from 1997-2006 and found that previous studies used court verdicts such as diminished responsibility as a proxy measure of mental disorder; the study found that this process used as a source of information to determine mental illness may not actually reflect an increase in homicide rates due to the prevalence of mental disorder. [3]

In the United States, self-reports of violent behavior were obtained in the Epidemiological Catchment Area (ECA) project. [4] The primary purpose of the project was to estimate the prevalence of untreated psychiatric disorders. Structured diagnostic interviews were administered to more than 20,000 persons residing in 5 areas of the country, both in the community and in institutions. In approximately 50% of this sample, data on violence were collected. The probabilities of violent behavior in male and female patients with schizophrenia were, respectively, 5.3 times and 5.9 times higher than in persons without any diagnosed mental disorder. The probabilities were also higher for patients diagnosed with affective disorder, but they were not higher for those diagnosed with anxiety disorder. Patients with a mental disorder and a comorbid substance abuse disorder had the highest probabilities of all—12.6 times higher for males and 9.1 times higher for females.

A US study made an attempt to compare rates of aggressive behavior among patients (N = 1136) with comparable controls (N = 519). Using self-reports, collateral informants, and police and hospital records, the authors found no significant difference between the prevalence of violence by patients without symptoms of substance abuse (4.3%) and the community controls who also were free of symptoms of substance abuse (3.3%). [5] Substance abuse symptoms significantly increased the rate of violence to 14.1% of the patient sample and 11.1% of the controls. Moreover, the targets of patient violence were more likely to be family members, friends, or acquaintances (89.3%) rather than strangers (10.7%). The data are hampered somewhat by methodological questions, including the possibility of not accounting for patients who may be repeatedly assaultive and treatment-resistant and those who are lost to follow-up.

A number of studies have been performed using data from Scandinavian countries. In these countries, registries of hospitalizations, arrests, and incarcerations are available that include the entire population. A study of an unselected Swedish birth cohort (N = 15,117) found that among persons without mental disorders, the conviction rate for violent offenses was 5.7% for men and 0.5% for women. Men with major mental disorders were approximately 4 times more likely to have been convicted of a violent offense than men without any mental disorder, and women with mental disorders were approximately 27 times more likely to be convicted than women without mental disorders. [6]

An even larger unselected cohort composed of 324,401 Danes found that both men and women hospitalized at least once with major mental disorders (eg, schizophrenia, manic-depressive psychosis, psychogenic psychosis), mental retardation, antisocial personality disorder, drug use disorders, alcohol use disorders, and other mental disorders were more likely to have been registered for at least 1 violent crime than persons with no history of psychiatric admissions. [7] For the period 1978-1990, for which computerized national police data were used in the analysis, men hospitalized with major mental disorders were approximately 9 times more likely to have been convicted of at least 1 violent offense than men with no history of psychiatric hospitalizations. For women, this relative risk was approximately 4.5.

An epidemiological study conducted in Finland found that among 11,017 people in one birth cohort, men who abused alcohol and were diagnosed with schizophrenia were found to be 25.2 (95% confidence interval, 6.1-97.5) times more likely to commit violent crimes than men who are mentally healthy. [8] Another Finnish study examining forensic patients (N = 693) found that the risk of committing a homicide was approximately 10 times greater for patients with schizophrenia than it was for the general population. [9] The risk increased to greater than 17 times for male patients with schizophrenia and coexisting alcoholism and greater than 80 times for female patients with schizophrenia and coexisting alcoholism. The actual number of positive cases in these large samples is small, however, and is dwarfed by the number of persons who were not mentally ill who committed violent crimes.

These findings are not limited to Scandinavia. Using case registries in Australia (N = 4156), the odds ratio for violent offenses was 2.4 for male individuals with schizophrenia with no substance abuse problems and 18.8 for schizophrenia complicated by substance abuse. [10] A review of data from 10 different countries reinforces the above findings. [11] Using self-reports, other informants, and clinic or hospital records, the authors concluded that the occurrence rate of assault in the cohort of 1017 patients with schizophrenia was 20.6%, with the rate 3 times higher in developing countries (31.5%) compared with developed countries (10.5%). Thus, this issue is relevant internationally and transcends many cultures and environments.

Violent behavior in hospitals has also been studied. Violent or threatening behavior is a frequent reason for admission to a psychiatric inpatient facility, and that behavior may continue after the admission. During the first 24 hours after the admission to a psychiatric inpatient unit, 33 (13%) of 253 patients physically attacked another person. [12] Patients who were manic were the most likely diagnostic group to be assaultive during the initial phase of hospitalization, with 12 (26.1%) of 46 patients attacking another person, compared with 9 (10.3%) of 87 diagnosed with schizophrenia and 12 (10%) of 120 diagnosed with mental illness other than schizophrenia or mania. Another group reported that in the first 8 days after hospitalization, 25 (8.7%) of 289 patients who were schizophrenic or schizoaffective assaulted someone at least once. [13]

Long-term hospitalization does not eliminate violent behavior. A study of 5164 long-term patients indicated that 7% of these patients physically attacked another person at least once during a 3-month period. [14] Patients who were assaultive were more likely than patients who were nonassaultive to have a primary diagnosis of nonparanoid schizophrenia, psychotic organic brain syndrome, mental retardation, or personality disorder.

Many studies do not distinguish between recidivistic and transient assaultiveness, but such distinction is important. Persistent (recidivistic) and transient violence may differ in cause, management, and social consequences. A small number of patients may be responsible for the bulk of all incidents. This is supported by 2 studies.

The first study, conducted in the United States, was a 6-month study of 1552 inpatients with various diagnoses that detected 576 violent incidents; a small group of recidivistic patients (5%) caused 53% of the incidents. [15] The diagnosis of schizophrenia was slightly overrepresented among the recidivists who were male; personality and impulse disorder diagnoses were frequent among the male and female recidivists.

The second study was conducted in Australia. [16] The patients who were recidivistic (12%) accounted for 69% of the 752 violent incidents identified. The males were more likely to have an organic brain syndrome, and the females were more likely to have a personality disorder.


Assessment and Differential Diagnosis

Patient Assessment

Patient assessment involves (1) the gathering of information about past and current behavior from the patient, health care providers, family, and friends; (2) a review of past treatment (successful and unsuccessful); and (3) a clinical examination of the patient over time.

In the assessment of a patient who is acutely agitated and whose history is unknown, attempts are made to rule out somatic conditions that require emergency treatment. Delirium is a medical emergency. Once the patient is under behavioral control, further medical and psychiatric workups can be accomplished. Mechanical restraints may be necessary to prevent agitated patients from injuring themselves or others while the medical workup is being conducted. For patients who are acutely agitated and for whom the episode is one of many, the acute episode is managed, and, subsequently, time is devoted to strategies designed to reduce the intensity and frequency of episodes.

Included in a physical examination should be a thorough mental status examination. Key elements include an assessment of affect and thought content, especially hallucinations, delusions, suicidal ideation, and homicidal ideation. An assessment of orientation and memory is also crucial for establishing a differential diagnosis. Disorientation may be the first clue that an underlying somatic condition is altering the patient's mental status.

Care must be taken to not miss comorbid conditions that may manifest as acute intoxication or withdrawal, such as alcohol or sedative abuse or dependence. A concomitant seizure disorder may complicate the clinical picture, especially if neuroleptic therapy appears to worsen the condition. Adverse drug effects, such as akathisia, may serve as stimuli for striking out. Antisocial personality traits may be the most important factor in some instances of patient violence in which goal-directed behavior, such as extortion of money or cigarettes, is evident.

Comorbidities are common and it is often difficult to ascertain the principal cause of the aggressive behaviors. Aggressive behaviors are heterogeneous in origin. [1] Moreover, additional differential diagnostic considerations should also include borderline personality disorder, obsessive-compulsive disorder, generalized anxiety disorders, and neurodevelopmental disorders. Patients may be diagnosed as having mania when they actually have these other chronic conditions (which can be comorbid) that are left untreated if unrecognized.

Differential Diagnosis

Differentiating patterns of violence central to the development of a differential diagnosis is achieved by analyzing the pattern of the violence. Whether aggressive episodes are singular or repetitive, with low or high potential for actual injury, helps guide the clinician in formulating immediate management plans, a provisional diagnosis, and a long-term strategy. Some patients are violent only when in a chaotic environment; others are persistently violent regardless of the milieu. Patients who were persistently violent were found to be more likely to have impairments in stereognosis, graphesthesia, tandem walk, and walking-associated movements and selective impairment in visual-spatial functioning as determined by neuropsychological testing. More detailed discussions regarding the neurology of aggression in general are beyond the scope of this article.

In contrast to the patient who is persistently violent, those who are transiently violent respond to the introduction of a new, structured environment. Environmental factors leading to increased aggressive behavior in a psychiatric ward include crowding and, possibly, an overauthoritative attitude by nursing staff and underinvolvement of medical staff with regard to ward activities. Time of day may be a factor, with a peak problem period of 7:00-9:00 AM reported in one facility. Apparently, patients who are transiently violent are more responsive to typical neuroleptic medication and have less neurological impairment than those who are persistently violent.


Patients with schizophrenia living in the community usually would not fall into the persistently violent category, but they may present acutely with aggressive and violent behavior. This may be due to acute decompensation secondary to covert or overt noncompliance with psychotropic medication therapy. Decompensation also may be due to a failure of the current medication regimen. The clinical features expected would be a worsening of psychotic symptoms and, possibly, command hallucinations, although the importance of the latter in violent behavior is in dispute. Studies report that 24-44% of aggressive acts committed by individuals with schizophrenia occur during an acute phase of the illness. Neuroleptic blood levels have been found to be inversely correlated with danger-related events in a group of recently hospitalized male patients who are schizophrenic.

See Medscape's article Schizophrenia.

Substance abuse, intoxication, and withdrawal

Some of the violent behavior occurring among psychiatric patients is attributable to comorbidity with substance abuse. Aggressive and violent behavior in patients can be precipitated by alcohol, cocaine, phencyclidine (PCP), or amphetamine intoxication. Inpatients can also be abusing illicit substances because access to drugs and alcohol, although difficult, is not impossible. Caffeine intoxication, water intoxication, antihistamine intoxication, and the ingestion of deodorants and aerosols have also been described in inpatients. Withdrawal from substances can lead to aggressive behavior, sometimes in relation to goal-directed drug-seeking behavior and, at other times, in relation to paranoia and extreme anxiety.

See Medscape's article Alcoholism.

Medical and neurological conditions

Conditions such as brain injuries, brain tumors, or metabolic disturbances may precipitate aggressive behavior in patients not normally known to be violent. For patients already in the hospital, regular and routine physical assessments, including laboratory tests, can be expected to screen for most problems. For patients with a history of seizure disorder, recent evidence suggests that interictal violence is associated more with psychopathology and mental retardation than with epileptiform activity or other seizure variables.

Dementia/Alzheimer disease

Patients with dementia may be emotionally labile or prone to poor impulse control. Although some may appear frail, serious injury to self and others can result from a rage reaction to a perceived threat. Usually, a history of cognitive decline clarifies the diagnosis, but this history may not always be available in an emergency department. In a nursing home, the issue of behavioral dyscontrol has led to the use of mechanical and chemical restraints and has resulted in the development of federal regulations in the United States to curb their widespread use.

See Medscape's articles Alzheimer Disease and Vascular Dementia.

Antisocial personality disorders

Antisocial personality disorders may coexist with other psychiatric disorders, including schizophrenia. Antisocial personality traits may be present even if the full disorder cannot be diagnosed. Violence secondary to the antisocial personality disorder/traits may be evaluated by examining the context of the aggressive incident. Intimidation of patients and staff or material gain may be factors in violent behavior. For example, fighting over money, cigarettes, and access to sexual partners and attacks on caregivers who deny a patient's request or try to set limits to patient behavior (eg, enforcing a smoking ban) may occur.

Mood disorders

Using a sample of 1140 recently incarcerated male felons, evidence was found of a direct relationship between a lifetime diagnosis of dysthymia and an arrest or incarceration history for robbery, as well as with multiple incidents of fighting since age 18 years. The manic state has been associated with violent behavior among 40 male psychiatric inpatients diagnosed with bipolar disorder. In a sample of 20 inpatients with mania and 856 with other diagnoses, assaultiveness was not present in the manic group, but agitation was observed more frequently when compared with all other diagnostic categories.

Posttraumatic stress disorder

Posttraumatic stress disorder (PTSD) has been associated with anger, hostility, and violence, although the presence of comorbid conditions, such as mood disorders and substance abuse disorders, may be confounding factors. These confounding factors were controlled for in a study of 27 outpatients who were Vietnam War veterans with PTSD and 15 controls who were Vietnam War combat veterans without PTSD. Subjects with PTSD scored significantly higher than subjects without PTSD on measures of hostility and violence.

Panic disorder

Panic disorder with aggressive thoughts and behaviors in association with panic symptoms has been reported in a series of 3 patients. Although each patient was interviewed carefully using standard diagnostic criteria, the case descriptions hint at significant disturbances of mood and impulse control, the presence of substance abuse, and a history of trauma.

See Medscape's articles Posttraumatic Stress Disorder and Panic Disorder.


Treatment - Management of Acute Aggression

Numerous books and articles offer good practical advice on training issues and on handling patients who are agitated. Others focus on restraint and seclusion and on psychodynamic strategies. Behavioral, psychological, and pharmacological interventions are used simultaneously. Organizations (eg, Non-Abusive Psychological & Physical Intervention) are available to train hospital and clinic staff in methods of assessing, preventing, and physically managing dangerous behavior. Typically, the staff is trained as a team that includes physicians, nurses, therapy aides, social workers, psychologists, security personnel, and others who might have patient contact.

The principal elements of the nonpharmacological management of aggressive behavior include the following:

  • Assess the environment for potential dangers (eg, objects that can be thrown or used as a weapon).
  • Assess the physical demeanor of the patient (eg, many patients make a fist before punching or kicking).
  • Know where the patient is at all times (eg, do not turn your back to the patient; do not leave the patient alone and therefore unobserved).
  • Take verbal threats seriously.
  • Remain several feet away to avoid crowding the patient.
  • Clear the area of other patients.
  • Summon additional help (a "show of force" or a "show of concern"); this is not a time for heroics.
  • Remain calm, maintain a confident and competent demeanor, and attempt to deescalate by engaging the patient in conversation.
  • Avoid arguments between staff members in front of the patient.
  • If restraints are necessary, have at least 4 people available.

Medications are often used to manage agitated behavior [17] . These include benzodiazepines (eg, lorazepam) and antipsychotics (eg, haloperidol). The US Food and Drug Administration (FDA) has approved new intramuscular formulations of second-generation antipsychotics that can be used to calm agitated patients [18] . Ziprasidone intramuscular is approved for use in agitation associated with schizophrenia, and olanzapine intramuscular and aripiprazole intramuscular are approved for use in agitation associated with either schizophrenia or bipolar mania. In general, intramuscular injection has a faster onset of action than oral administration; however, a patient may calm down readily after an oral dose because he or she realizes that action has been taken and help is being provided. Sublingual administration may have a faster onset of action than oral ingestion and has the added advantage of distracting the agitated patient while the pill is dissolving.


Lorazepam appears to be a good rational choice when treating an acute episode of agitation, especially when the etiology is not clear. Lorazepam is a nonspecific sedating benzodiazepine. A brief but influential report compared the antiaggressive effects of a single dose of haloperidol (10 mg) with a single dose of lorazepam (2 mg) in violent psychiatric patients at a crisis unit. Lorazepam was at least as effective as haloperidol in controlling violent behavior. Of all the benzodiazepines available, lorazepam is the only one reliably absorbed when administered intramuscularly. Its half-life is relatively short (10-20 h), and it has no active metabolites. The usual dosage of 0.5-2 mg every 1-6 hours may be administered orally, sublingually, intramuscularly, or intravenously. Caution is required when respiratory depression is a possibility, but this is less of a problem than withamobarbital (Amytal), an agent popular before the advent of lorazepam.

An interaction with benzodiazepines and clozapine has been reported, producing (sometimes fatal) respiratory depression and marked sedation, excessive sialorrhea, and ataxia, but this combination has been used successfully by a number of practitioners. Paradoxical reactions to benzodiazepines, as exhibited by hostility or violence, have been an area of concern, but the evidence is not convincing. Disinhibition with benzodiazepines is, in any event, uncommon and even more unlikely to occur when benzodiazepines are administered within the context of single or limited doses in a crisis situation.

In current clinical practice, benzodiazepines are not widely used long term to control violence because of practical and ethical problems with tolerance and dependence. However, lorazepam is an excellent tool for short-term, quick reduction of violent behavior. The possibility of alcohol or sedative withdrawal as a cause of agitation is another point in favor of using lorazepam. The use of a neuroleptic in this instance is suboptimal and may lower the seizure threshold. If delirium tremens is present, full supportive medical care must be available; when medical complications arise, the mortality rate has been reported to be as high as 20%.

First-Generation Antipsychotics

First-generation antipsychotics are discussed below.


Neuroleptics universally cause sedation when administered at a high enough dose. Haloperidol, a butyrophenone, has been used frequently as an intramuscular medication as needed for agitation and aggressive behavior, including in an emergency department setting, for a wide variety of patients. The advantage of haloperidol over the low-potency neuroleptics (eg, chlorpromazine) is that it causes less hypotension, has fewer anticholinergic adverse effects, and causes less of a decrease in the seizure threshold.

Despite the advantages of haloperidol over chlorpromazine, clinicians may desire a more sedating agent, so the low-potency neuroleptics continue to be used. In addition to this nonspecific sedation, another benefit is its antipsychotic effect, but this becomes evident only after the acute episode of agitation has subsided. Although once popular, rapid tranquilization with injectable haloperidol does not lead to a more extensive or rapid alleviation of psychotic symptoms, as was observed in a group of patients who are mainly schizophrenic.

Neuroleptics may have a longer-lasting effect on the reduction of agitation by treating the underlying psychosis. On the other hand, high doses of neuroleptics may lead to more adverse effects, including akathisia. Akathisia increases irritability and has been reported to be associated with violence in several cases. High-dose haloperidol treatment of chronic schizophrenia (60 mg/d) has been associated with violence; akathisia might have been the mediating variable. For patients who are resistant to typical neuroleptics, no benefits exist beyond the acute sedative effect.

Acute mania, frequently associated with assaultiveness, can be controlled effectively and quickly by antipsychotics. Low doses of haloperidol or other neuroleptics are used to manage aggression in elderly patients. Neuroleptics are frequently prescribed to reduce aggressive behavior in persons with mental retardation, but their effectiveness for this indication is doubtful.

Small studies of various neuroleptics reported some success, but a large placebo-controlled study demonstrated a significant (but unexplained) increase of aggressiveness with thioridazine (200 mg/d). These investigators suggested that the removal of physical restraints resulted in a decrease of aggressiveness that was independent of psychopharmacological treatment. The administration of neuroleptics to control behavior in individuals who are not psychotic and who are mentally retarded is common, but the effectiveness and safety of this treatment have not been demonstrated clearly, and the practice remains controversial. Haloperidol was used successfully in a sample of 11 patients who developed violent, aggressive behavior after a severe closed head injury.


Droperidol, another neuroleptic in the butyrophenone class, is used most often for induction of anesthesia. The medication is not approved by the FDA for psychiatric conditions but has been used for sedating agitated patients in an emergency department setting. It is no longer available in the UK and has a black box warning in the United States because of the possibility of prolongation of the QT interval and subsequent risk of fatal arrhythmia (torsade de pointes).

Second-Generation (Atypical) Antipsychotics

The new second-generation antipsychotics are emerging as important options in the management of acute agitation in persons with psychosis. The atypical antipsychotics have several advantages over typical antipsychotics. These advantages include a lower propensity for extrapyramidal adverse effects (including akathisia), a lower propensity towards tardive dyskinesia, and, perhaps, a specific antiaggressive effect over time, as will be described later.


Oral ziprasidone was commercially released in 2001, and the intramuscular formulation became generally available in 2002. Premarketing clinical trials of intramuscular ziprasidone demonstrated efficacy of both a 10- and a 20-mg dose in reducing acute agitation in approximately 200 patients with schizophrenia and schizoaffective disorder. Rather than using placebo or an active comparator, these 2 studies compared a therapeutic amount of ziprasidone with a 2-mg injection.

Peak plasma concentrations after an intramuscular dose of ziprasidone are achieved in 30-45 minutes, compared with 8 hours (with half-life of 3.8 h) with oral dosing. The Cmax was higher with a 20-mg injection of ziprasidone compared with a 10-mg injection, probably explaining the greater efficacy of the higher dose. Significant improvement with 20 mg was observed at 30 minutes postinjection. The improvement with the 20-mg dose increased until 2 hours and was maintained until at least 4 hours postdosing. The number of 20-mg injections needed to achieve the desired outcome was 1 (42%) or 2 (37%). Overall, intramuscular ziprasidone was well tolerated and almost free of extrapyramidal adverse effects. The most common treatment-emergent adverse events at the 10- and 20-mg doses were nausea, headache, and dizziness.

Per the product package insert, the recommended dose is 10-20 mg administered as required, not to exceed 40 mg/d. Doses of 10 mg may be administered every 2 hours; doses of 20 mg may be administered every 4 hours. Coadministering oral and intramuscular ziprasidone is not recommended. The magnitude of QTc increases with intramuscular ziprasidone is comparable to that described for oral ziprasidone. Because of this effect, the product package insert contains a warning that ziprasidone should not be used with other drugs that are known to prolong the QTc interval, in patients with congenital long QT syndrome, and in patients with a history of cardiac arrhythmias. However, prolongation of the QT interval with intramuscular ziprasidone is similar to that seen with intramuscular haloperidol, and has not been a clinical obstacle to its use.


Oral olanzapine was commercially released in 1996, and the intramuscular formulation is also now available. Four premarketing clinical trials demonstrated the efficacy of olanzapine in reducing acute agitation in approximately 1050 patients with schizophrenia, schizoaffective disorder, bipolar disorder, or dementia. Peak plasma concentrations after an intramuscular dose of olanzapine are achieved in 15-30 minutes, compared with 3-6 hours with oral dosing.

Each of these double-blinded randomized clinical trials included a comparator medication (either haloperidol or lorazepam) and placebo. Agitation was measured by the Positive and Negative Syndrome Scale Excited-Component (PANSS-EC) and the Agitation-Calmness Rating Scale. The onset of action of intramuscular olanzapine (10 mg) is rapid, occurring as early as 15 minutes, and is superior to intramuscular haloperidol (7.5 mg) among patients with schizophrenia.

An analogous study with bipolar mania patients revealed similar findings, with the onset of action occurring as early as 30 minutes, with superiority to intramuscular lorazepam (2 mg). The presence of psychotic symptoms was not necessary for a treatment response; response was observed for both manic patients with psychotic features and for manic patients without psychosis.

The dementia study used lower doses of olanzapine (2.5 mg and 5 mg) and used a lower dose of lorazepam (1 mg) as the active comparator. Both doses of olanzapine were as efficacious as the dose of lorazepam. However, intramuscular olanzapine is not FDA approved for use in agitation associated with dementia.

Intramuscular olanzapine exhibited less extrapyramidal adverse effects compared with intramuscular haloperidol. No adverse event was significantly more frequent for intramuscular olanzapine than intramuscular haloperidol or intramuscular lorazepam. Usually (76% of the time), only one 10-mg olanzapine injection was required to achieve the desired outcome. Intramuscular olanzapine was not associated with excessive or undesirable sedation. Product labeling recommends a dose of 10 mg per injection for adults with schizophrenia or bipolar disorder and lower doses for medically compromised or elderly patients.

In contrast to ziprasidone, no evidence indicates that significant QTc interval prolongations will occur with olanzapine. Of interest is that bradycardia was observed in approximately one third of the nonagitated healthy volunteers who received olanzapine in the intramuscular olanzapine clinical trials, whereas less than 5% of agitated patients experienced sinus bradycardia. Most cases of bradycardia were observed to be associated with a decrement in resting blood pressure or an orthostatic drop, consistent with a benign and self-limited neurally mediated reflex bradycardia (vasovagal) and likely related to the alpha-1 antagonism of olanzapine.


Oral aripiprazole was commercially released in 2002. Aripiprazole intramuscular was approved by the FDA in late 2006 for the indication of agitation associated with schizophrenia or bipolar mania. Three one-day, placebo-controlled inpatient trials with active comparators were conducted, and trial designs were similar to the registration studies conducted for olanzapine intramuscular (similar primary outcome measures, including the PANSS-EC).

In a study comparing intramuscular aripiprazole 9.75 mg with intramuscular haloperidol 6.5 mg, analysis indicated that aripiprazole was noninferior to haloperidol, and both aripiprazole and haloperidol were superior to placebo. Aripiprazole was also tested in agitated patients with bipolar disorder, with statistically superior improvement in agitation seen for aripiprazole 9.75 mg and 15 mg and lorazepam 2 mg compared with placebo. In product labeling, the usual recommended dose is 9.75 mg. Safety concerns noted in product labeling include greater sedation and orthostatic hypotension with the combination of lorazepam and aripiprazole as compared with that observed with aripiprazole alone. No evidence shows any clinically significant prolongation of the QT interval with aripiprazole.

Are the newer IM antipsychotics better?

The availability of intramuscular formulations of ziprasidone, olanzapine, and aripiprazole begs the question whether they provide any substantial advantages over older agents such as haloperidol. This can be answered by examining these agents using the tools of evidence-based medicine, namely number needed to treat (NNT) and number needed to harm (NNH). [18, 19]

If we define response as a set amount of reduction of agitation at 2 hours after the first injection, NNT for response versus placebo (or placebo equivalent) in treating agitation in the registration studies at the recommended dose of ziprasidone 10–20 mg was 3 (95% CI, 2-4), for olanzapine 10 mg was 3 (95% CI, 2-3), and for aripiprazole 9.75 mg was 5 (95% CI, 4-8). This means, for example, that for every 5 patients with agitation who were given aripiprazole versus placebo, one additional responder was encountered. The confidence intervals for all 3 second-generation antipsychotics overlap with that for haloperidol where the NNT was 4 (95% CI, 3-5). Thus, we can conclude that we can anticipate similar efficacy when using the newer antipsychotics compared with haloperidol when treating agitation.

However, the major advantage of using the newer agents is their acute tolerability — they are less likely to be associated with extra-pyramidal adverse effects. For example, for olanzapine 10 mg IM versus haloperidol 7.5 mg IM in one of the studies of patients with agitation associated with schizophrenia, NNH to encounter one additional case of requiring anticholinergic medication was 7 (95% CI, 5-13).

These newer agents are not devoid of acute adverse effects, but they are encountered more seldomly than a therapeutic reduction in agitation compared with placebo. For example, treatment-emergent adverse events occurring during the pivotal trials revealed statistically significant NNH versus placebo (or placebo equivalent) for aripiprazole for headache (NNH = 20; 95% CI, 11-170) and nausea (NNH = 17; 95% CI, 11-38), for ziprasidone in the treatment of headache (NNH = 15; 95% CI = 8-703), and for olanzapine in treatment of emergent hypotension (NNH = 50; 95% CI, 30-154).

Inhaled loxapine and sublingual asenapine

Inhaled loxapine received approval in 2013 for the acute treatment of agitation associated with schizophrenia or bipolar mania on the basis of double-blind, placebo-controlled, randomized trials in patients with schizophrenia and bipolar mania. NNT for response for 10 mg versus placebo at 2 hours after administration was 3 (95% CI 3–4), placing it on par with IM options. Statistically significant separation from placebo occurred as early as 10 minutes after the administration of the medication. In the United States, the recommended dose is 10 mg with only a single dose within a 24-hour period permitted, and because of the potential risk of bronchospasm, inhaled loxapine can be administered only by a healthcare professional in a healthcare facility that is enrolled in the Risk Evaluation and Mitigation Strategy program for the product. In the pivotal clinical trials, dysgeusia (an unpleasant taste in the mouth) was the spontaneously reported adverse event whose incidence was ≥ 5% or higher and greater than twice than that for placebo, resulting in a NNH versus placebo of 11 (95% 7-23). [20]

Although sublingual asenapine has not been approved for the treatment of agitation, in a double-blind, placebo-controlled, randomized study of agitated adults (any diagnosis) in an emergency department, sublingual asenapine 10 mg was efficacious with an effect size comparable to that observed for IM medications; the NNT versus placebo for response at 2 hours after administration was 3 (95% CI 2–4). In routine use, asenapine has been associated with oral hypoesthesia and dysgeusia. [21]

Antipsychotics - A Caveat

Generally, antipsychotic agents are not recommended for aggressive patients who do not have a diagnosis of a psychotic disorder or bipolar mania. Nonspecific sedating agents (eg, lorazepam) are preferred. The use of antipsychotics such as haloperidol places the patient at risk for adverse effects, such as acute dystonia in the short-term, or tardive dyskinesia in the longer term. Moreover, that antipsychotic-induced akathisia can result in further agitation cannot be overemphasized. The newer second-generation antipsychotics have a lower propensity to cause these adverse effects, but the risk is not completely eliminated.

Acute Aggressive Behavior

In summary, patients with aggressive behavior first must be assessed for the possibility of comorbid conditions. Medical conditions, including acute withdrawal from alcohol or sedatives, should be excluded. Beyond the acute management of an aggressive episode, long-term management depends on the pattern of violence, either transient or persistent.

Short-term sedation with lorazepam is safe and effective for acute episodes. Use of first-generation antipsychotics such as haloperidol may lead to adverse effects, such as akathisia, which may, in turn, precipitate further agitation. Second-generation antipsychotics such as ziprasidone, olanzapine, and aripiprazole are now available as short-acting intramuscular agents and are specifically FDA-approved for agitation associated with schizophrenia (ziprasidone, olanzapine, and aripiprazole) and bipolar mania (olanzapine and aripiprazole). They are relatively free of extra-pyramidal side effects, making them a dominant choice over the older agents.


The Next Step after the Acute Aggressive Episode

Once the acute agitation is managed, longer-term strategies are required. As useful as lorazepam is as an acute intervention, it is not recommended as a long-term solution. Specialized units, such as secure or psychiatric intensive care units, provide a structured environment that optimizes staff and patient safety. In general, these specialized units are staffed with persons trained in interacting with patients who are volatile and difficult to manage [22] . Preventive aggression devices are a form of ambulatory restraints that can be used as an alternative to seclusion. Patients in these wrist-to-belt and/or ankle-to-ankle restraints can remain with their peers on the ward, eat their meals, and interact, yet they are prevented from striking out and injuring others. When used in combination with a comprehensive behavior modification program, these patients can be weaned off ambulatory restraints.

Pharmacotherapy for the longer-term management of violent behavior depends on the individual patient's underlying clinical problem. Theoretical rationales for treatment strategies have included the serotonin hypothesis. The atypical antipsychotics (eg, clozapine, risperidone), beta-adrenergic blockers, mood stabilizers (eg, lithium, carbamazepine, valproate), antidepressants, and buspirone have been used.

Treatment of the underlying disorder is key. Often, when the primary psychiatric problem is treated successfully, the associated aggressive behavior is reduced. Unfortunately, perhaps one third of patients with schizophrenia do not respond to antipsychotic treatment or respond only partially. Chronically violent patients with schizophrenia may receive higher doses of neuroleptics without clear evidence that this reduces the frequency of violent behavior. This actually may worsen akathisia, further increasing the risk of aggressive behavior, as discussed earlier. Other complicating factors are comorbid conditions, such as substance abuse or antisocial personality disorder. Because of the above considerations, several pharmacological agents have been studied with regard to specific antiaggressive activity, with the goal being to use them to specifically target aggressive behavior.

A common theme among many agents proposed as antiaggressive drugs is their effect on the serotonin neurotransmitter system. In many species, the serotonergic system is involved in the modulation of aggressive behavior. A disturbance of this system has been implicated in impulsive violence in humans. Impulsive violent behavior may be directed against others or against oneself. These issues have been reviewed extensively elsewhere. In humans, disturbance of the serotonergic system has been inferred from low levels of 5-hydroxyindoleacetic acid in the cerebrospinal fluid or from a blunted response to neuroendocrine challenges. This work was conducted largely with aggressive patients with personality disorders and substance abuse disorders. Drugs that affect the serotonin neurotransmitter system include antipsychotics (eg, clozapine, risperidone), antidepressants (eg, fluoxetine, citalopram), and anxiolytics (eg, buspirone). These and other medications are described in the following sections.

Special mention should be made regarding the legal implications of treating a patient who has aggressive behavior, specifically regarding any threats this person may make towards somebody else. Extensive debate has occurred regarding the duty to warn and the duty to protect (the most famous case being Tarasoff v The Regents of the University of California). In these matters, the issue of patient confidentiality may be overruled by safety concerns. Although discussion of these issues is beyond the scope of this article, clinicians must be mindful of any specific threat made by a patient. Hospitalization of the patient may satisfy the immediate need to protect a potential victim, but further consultation is highly recommended.


Antipsychotics for Persistent Aggressive Behavior


The atypical antipsychotic clozapine, in addition to being an effective treatment in patients refractory to typical neuroleptics, may have specific antiaggressive effects. This may be due to the effects of clozapine on the serotonin system and to its selective affinity for the limbic system. One retrospective study in a state hospital found that the number of violent episodes among inpatients with schizophrenia decreased after they began clozapine treatment. Similar results were found among 5 patients with schizophrenia in a specialized unit for severely aggressive patients and in 2 of 8 patients with brain injuries who exhibited chronic posttraumatic aggressive behavior.

In a program spanning 21 state hospitals in New York, 223 patients with schizophrenia who received clozapine had the Brief Psychiatric Rating Scale completed at baseline, 6 weeks, 12 weeks, and at the end point. A selective effect of clozapine on hostility was noted. Another study reported a reduction in the use of seclusion and restraints in a group of patients in a state hospital who were schizophrenic and schizoaffective who received clozapine.

These uncontrolled studies provided a strong signal evidencing an advantage of clozapine in the management of persistent aggressive behavior, and these results led to a randomized double-blinded clinical trial that yielded results supporting this conclusion [23] . This 14-week study compared the specific antihostility effects of clozapine with those of olanzapine, risperidone, or haloperidol in 157 inpatients with schizophrenia or schizoaffective disorder and found that clozapine had significantly greater antihostility effects than haloperidol or risperidone. These effects on hostility were specific: they were independent of antipsychotic effects on delusional thinking, formal thought disorder, or hallucinations and were independent of sedation.

The above advantage for clozapine was confirmed in a second double-blind randomized clinical trial that specifically enrolled patients (N=110) who had exhibited physically aggressive behavior. [24] Patients were randomized to receive clozapine, olanzapine, or haloperidol double blind for up to 12 weeks. Patients randomized to clozapine had statistically significant lower endpoint aggression scores than patients randomized to either olanzapine or haloperidol. Patients randomized to olanzapine had statistically significant lower endpoint aggression scores than patients randomized to haloperidol. No differences were seen among the 3 groups in terms of reduction of psychopathology as measured by the total Positive and Negative Syndrome Scale score.


The atypical neuroleptic risperidone also affects the serotonin system and has also been demonstrated to have a selective effect on hostility. The authors had access to the database generated by a multicenter, double-blinded, placebo-controlled, parallel-group, 9-week clinical trial of risperidone versus haloperidol. Hostility was measured by the Positive and Negative Syndrome Scale.

One hundred thirty-nine patients had a baseline hostility score of at least mild, and all carried the diagnosis of schizophrenia. Risperidone was found to have a greater selective effect on hostility than haloperidol. Although this was supported by a review of the impact of risperidone on seclusion and restraint in 74 patients receiving risperidone (average risperidone dose, 5.1 mg/d) at a state psychiatric hospital, this effect was not evident in a retrospective case-controlled study of 27 patients with schizophrenia or schizoaffective disorder (average risperidone dose, 6.8 mg/d) or in another negative report of a group of 20 forensic patients with chronic schizophrenia (minimum risperidone dose, 6 mg/d, average dose not reported).


Olanzapine was also found to be superior to haloperidol on measures of agitation, as determined by a posthoc analysis of a phase 3 randomized clinical registration trial in patients with schizophrenia, but selectivity of effect was not reported.

More recently, olanzapine’s superiority over other first-line antipsychotics (typical and atypical) for specific anti-hostility effect was demonstrated in post hoc analyses of patients with chronic schizophrenia and among first-episode patients with schizophrenia. [25, 26]


Quetiapine may preferentially reduce hostility and aggression. Both quetiapine and haloperidol were superior to placebo in reducing positive symptoms of schizophrenia, but only quetiapine was superior to placebo in measures of aggression and hostility as determined by a posthoc analysis of a phase 3 randomized clinical registration trial. This is supported by 2 case reports, one describing a dramatic response to quetiapine monotherapy in a persistently aggressive patient with schizoaffective disorder who did not respond to numerous other medications and one describing therapy with a combination of quetiapine and valproate in a patient with treatment-resistant schizophrenia.

Ziprasidone, Aripiprazole, Cariprazine

Published post hoc analyses have demonstrated some degree of selective anti-hostility effect for ziprasidone (versus haloperidol), aripiprazole (both for oral and the lauroxil long-acting formulation, versus placebo), and cariprazine (versus placebo). [27, 28]


At this time, the weight of the evidence favors clozapine as a specific antiaggressive treatment, with demonstrated superiority to haloperidol, olanzapine and risperidone in patients with schizophrenia or schizoaffective disorder. A reasonable second choice would be olanzapine. More research is needed to compare the other atypical antipsychotics with clozapine for this indication. Ideally, such studies should be performed double-blinded with subjects specially selected because of their aggressive behavior. This is difficult because of a number of factors, including the relative rarity of aggressive events and consequent need for a large sample size and lengthy baseline and trial periods, as well as selection/consent bias.


Beta-blockers and Mood Stabilizers for Treatment of Persistent Aggression


Beta-adrenergic blockers, in particular propranolol, have been used in the treatment of aggressive behavior. Several propranolol trials have been published regarding patients with organic brain disease. Subject age ranged from 7-86 years. The diagnoses included mental retardation; autism; posttraumatic organic brain syndromes; various types of dementia; Huntington disease; Wilson disease; postencephalitic psychosis; and apparent or presumed chronic central nervous system dysfunction inferred from soft neurological signs, abnormal EEG findings, or clinical seizures.

The time of onset of the antiaggressive effect of propranolol is difficult to glean from the reports; it appeared to range from 12 hours to 2 months. Current literature suggests that onset of the antiaggressive effects can be observed 4-8 weeks after the effective dose is reached. The effective doses can be reached only after a gradual buildup. The dose of propranolol perceived as effective by the researchers was remarkably variable. Doses higher than 600 mg/d are used only exceptionally. The dose-response relationship has not been studied systematically.

Propranolol has been used as an adjunctive treatment for schizophrenia, and a reduction in symptoms, including aggression, was found. A chart review of chronically assaultive patients with schizophrenia receiving nadolol or propranolol revealed a 70% decrease in actual assaults for 4 of the 7 patients. A double-blinded placebo-controlled study of adjunctive nadolol in 41 patients, 29 of whom were schizophrenic, found a decline in the frequency of aggression compared with controls.

Another double-blinded placebo-controlled study of adjunctive nadolol in 30 violent inpatients, of whom 23 were schizophrenic, found a trend towards lower hostility for the active treatment group. Both nadolol studies involved dosages of as high as 120 mg/d. Nadolol was as effective as propranolol in suppressing aggressive behavior in a patient who was retarded who received a separate trial of each of these 2 agents. Metoprolol, a selective beta-adrenergic receptor blocker, was effective in 2 cases of intermittent explosive disorder and 2 cases of mental retardation.

The main practical problem with the administration of propranolol for aggression is the high frequency of adverse cardiovascular effects (reduction of pulse rate and blood pressure). Pindolol, besides its nonselective beta-adrenergic blocking activity, has a partial agonist effect (intrinsic sympathomimetic activity). This accounts for the reduction of adverse cardiovascular effects due to pindolol in a normotensive population, making it a safer and more attractive compound for psychiatric indications than propranolol. Pindolol was tested in a double-blinded placebo-controlled crossover study using 11 patients with dementias of varying origin. It reduced assaultiveness and hostility without producing lethargy.

Mood Stabilizers

Mood stabilizers, such as lithium, valproate, and carbamazepine, are used as the primary medical treatment for bipolar disorder and as adjuncts to antipsychotic treatment for patients with schizophrenia [29] . Although they are not prototypical antiaggressive drugs, they are commonly used for this purpose in clinical practice. A survey of a state psychiatric hospital in New York revealed that 25% of the patients were receiving lithium, carbamazepine, or valproate. Diagnoses were schizoaffective disorder (44%), schizophrenia (28%), bipolar disorder (19%), or other diagnoses (9%). The indications for the mood stabilizers, as documented in the clinical records, were impulse control (46%), assaultive or aggressive behavior (44%), or other indications (10%). The adjunctive use of mood stabilizers (especially valproate) in schizophrenic patients has increased markedly in the past 5 years.

Generally, mood stabilizers work better than placebo for reducing aggressive behavior, especially carbamazepine/oxcarbazepine, phenytoin, and lithium. However, further randomized controlled trials are recommended. [30]


Aggressivity is observed frequently during manic episodes. Reduction of aggressivity is an important part of the therapeutic effect of lithium in patients who are manic. The antiaggressive properties of lithium in people who are mentally retarded were first demonstrated in an open study in 1970 that demonstrated that the effect of this treatment was not limited specifically to aggressive behavior; additional reductions of restlessness and hyperactivity occurred.

The antiaggressive effects of lithium in people who are mentally retarded were confirmed in another open trial in 1984, a retrospective study in 1989, and a double-blinded placebo-controlled trial in 1987. The recommended serum lithium concentration to achieve an antiaggressive effect was 0.7-1 mEq/L or 0.5-0.8 mEq/L. Although the original report mentioned uncontrollable polydipsia and enuresis, these adverse effects of lithium treatment were not prominent in a larger placebo-controlled trial or in another open study using slightly lower lithium serum levels.

A small retrospective uncontrolled study confirmed the antiaggressive effects of lithium in patients who were mentally retarded; self-injurious behavior exhibited by these aggressive individuals was also reduced by the treatment. Belligerence in a patient recovering from a head injury failed to respond to propranolol or haloperidol, but it disappeared when the patient was administered lithium; the effective levels were 0.4-0.8 mEq/L. Maintaining the plasma levels relatively low may be important because a severe closed head injury was reported to reduce tolerance to the adverse neurocognitive effects of lithium in a patient who was bipolar and was maintained on lithium.

Lithium treatment reduced the number of infractions involving violent behavior of recurrently violent prisoners in open trials and in a double-blinded placebo-controlled study. Lithium plasma levels in these trials were mostly below 1 mEq/L; such relatively low levels generally are not associated with lethargy or general suppression of motor activity in adult patients. Most prisoners in the open trial had various organic mental syndromes or schizophrenia, whereas the subjects in the controlled trial were described as having nonpsychotic personality disorders.

A paradoxical increase of aggressivity during lithium treatment has been described in one patient. This patient had a temporal lobe EEG spike focus. His EEG abnormality and aggressivity increased simultaneously after the start of lithium treatment and then improved after the treatment was stopped. However, a study reporting the beneficial effects of lithium included at least 4 patients with epilepsy (1 had temporal lobe epilepsy) and another study observed 8 patients with epilepsy, only 1 of whom demonstrated a slight increase of seizure frequency during the lithium treatment.

The effectiveness of lithium therapy for schizophrenia is not established. Active affective symptoms, previous affective episodes, and a family history of affective disorder may predict a favorable response to lithium but also may provide clues that the diagnosis could be something other than schizophrenia. A double-blinded, placebo-controlled, parallel-design clinical trial involving seriously ill state hospital patients with schizophrenia who had not responded to prior trials of typical neuroleptics demonstrated no advantage of lithium combined with haloperidol over haloperidol alone.

When lithium was added to neuroleptics for the treatment of patients with resistant schizophrenia who were classified as dangerous, violent, or criminal, no benefits were observed after 4 weeks of adjunctive lithium. However, case reports mention patients with paranoid schizophrenia with aggressive or disorderly behaviors who have responded to the addition of lithium to their neuroleptic treatment and who deteriorated after the lithium was discontinued but who subsequently improved when it was reinstituted.


Valproate for the control of aggressivity has been described to date in approximately 20 published reports comprising approximately 200 patients. Diagnoses in these reports cover the spectrum of psychiatric conditions, including dementia, borderline personality disorder, organic mood syndrome, bipolar disorder, schizophrenia, mental retardation, and schizoaffective disorder. The methodologies of these studies included case reports, retrospective chart reviews, and open-label designs. Only one double-blinded study has been reported, in which 16 patients with borderline personality disorder were enrolled and improvement was noted.

A comparison of valproate and carbamazepine in hospitalized patients (diagnosis not reported) revealed a decrease in the number of hours spent in mechanical restraints for both groups, with valproic acid being more effective than carbamazepine. An open-design study involving 35 individuals randomly selected from consecutive admissions to a large state hospital setting found that divalproex sodium reduced agitation in patients with bipolar disorder or borderline personality disorder, as measured by the number of hours per week spent in seclusion.

Although commonly used in patients with schizophrenia, valproate has not been studied adequately in this population. Some evidence is favorable; although not assessing hostility directly, positive symptoms were reduced with adjunctive valproate in a 28-day double-blinded randomized study with olanzapine and risperidone among patients with an acute episode of schizophrenia (N = 242). The mean dose of adjunctive valproate was approximately 2300 mg/d and was well tolerated, with few differences in adverse events between monotherapy with an antipsychotic or combination therapy with valproate. Posthoc analysis revealed a specific antihostility effect in the first 7 days of combination treatment.

In a recently reported 8-week open-label randomized parallel group clinical trial in hospitalized adults (N=33) diagnosed with schizophrenia and with hostile behavior, patients were randomly assigned to receive risperidone alone or risperidone plus valproate. Although patients receiving combination treatment were more likely to complete the study, a meaningful advantage for combination therapy, as evidenced by rating scales, was not observed [31] .


Studies examining the usefulness of carbamazepine in the control of aggressive behavior are few. Eight women diagnosed with schizophrenia (some with EEG abnormalities) and exhibiting violent behavior were administered carbamazepine in addition to their regular neuroleptic medication, with good results. In a retrospective study of violent patients, mostly with schizophrenia and with and without EEG abnormalities, significant reductions in aggressivity were noted with carbamazepine, regardless of EEG status. A brief letter reported that carbamazepine was administered in an open uncontrolled trial to 34 adult inpatients with rage outbursts of various etiologies. Most patients (N = 19) had intermittent explosive disorder. Significant improvement was noted in the severity of physical assaults.

A multifacility double-blinded study comparing the effect of a 4-week trial of adjunctive carbamazepine versus placebo with standard neuroleptic treatment was conducted on 162 patients with a Diagnostic and Statistical Manual of Mental Disorders, Third Edition diagnosis of schizophrenia or schizoaffective disorder, all with excited states or aggressive/violent behavior that responded unsatisfactorily to neuroleptic treatment. No statistically significant difference in response occurred among the patients receiving either adjunctive carbamazepine therapy or placebo, but a trend towards moderate improvement with carbamazepine was noted (P < .10).

A 15-week, double-blinded, randomized within-patient crossover study assessing adjunctive carbamazepine in chronic inpatients without epilepsy and with EEG temporal lobe abnormalities was performed on 13 patients, 9 of whom had schizophrenia and of whom 7 completed the study. Overt aggression was rated as twice as severe and 1.5 times as common with placebo compared with carbamazepine. Responses were not correlated with EEG deterioration or improvement.

With rare exceptions, the published studies of carbamazepine and aggressivity are not blinded and are uncontrolled for placebo effect. In addition, plasma levels of concomitant neuroleptics are not measured, leaving open the possibility for undetected pharmacokinetic interactions. Despite these limitations, carbamazepine appears to be a useful adjunct to neuroleptic therapy and may lower aggression in a broad spectrum of disorders.

Thus, the available evidence indicates that mood stabilizers may reduce aggressive behavior in patients with a broad spectrum of diagnoses. Mechanisms of the antiaggressive action of mood stabilizers are unclear. Because most of the reported trials used mood stabilizers as adjunctive treatment, some of the effects possibly depend on pharmacodynamic and pharmacokinetic interactions with antipsychotics or other concomitant medications.


Other Agents and Conclusion

Benzodiazepine - Clonazepam

Clonazepam, a high-potency benzodiazepine, had been reported to be useful as an antiaggressive agent in a case report of a patient with schizophrenia and a seizure disorder. In this particular case, both phenytoin and carbamazepine were ineffective treatments. The patient's treatment was most successful with a combination of haloperidol and clonazepam. This result is in contrast to a double-blinded placebo-controlled trial of adjunctive clonazepam in 13 patients with schizophrenia who were receiving neuroleptics. In that study, no additional therapeutic benefit was observed, and, in fact, 4 patients demonstrated violent behavior during the course of clonazepam treatment.

Antidepressant - Amitriptyline

The value of amitriptyline for the treatment of posttraumatic agitation (including some threatening behavior) was observed in a case report of a patient who developed agitated, physically threatening behavior after a closed head injury; this behavior was witnessed in additional patients. The current interest in the role of certain antidepressants in aggression treatment is based on the crucial role of serotonergic regulation of impulsive aggression against self and others. Antidepressants with specific effects on serotonin receptors are now available.

Antidepressant - Trazodone

Trazodone, a selective serotonin receptor 2 antagonist, reduced violent outbursts in an open trial of 4 patients with dementia who showed no clinical signs of depression; these observations were partially replicated in another sample of 7 patients with dementia who were aggressive, 3 of whom improved. A combination of trazodone and tryptophan (a serotonin precursor) was used successfully to reduce aggressive behavior and temper tantrums in a patient who was mentally retarded and in several elderly patients with dementia. A similar effect was reported in an adult patient with Down syndrome when trazodone was administered in combination with a serotonin-enhancing diet.

Antidepressant - Fluoxetine

Fluoxetine, a selective serotonin reuptake inhibitor, decreased impulsive aggression (including assaultiveness) in 3 patients with personality disorders; this was an open trial. Patients with chronic schizophrenia who remained symptomatic while maintained on antipsychotics were administered adjunctive fluoxetine; violent incidents decreased in 4 cases and increased in 1 case in this retrospective uncontrolled study. Fluoxetine was effective against anger attacks in patients with unipolar depression who were irritable and hostile. In one study, 21 persons who ranged from severely to profoundly mentally retarded and who also displayed aggression and self-injurious behavior were treated with fluoxetine on an open-label basis; 19 demonstrated a positive outcome, and no worsening attributable to fluoxetine was reported in the remaining 2 patients. In a preliminary medication trial with 5 patients withrefractoryborderlinedisorder, fluoxetine led to a decrease in impulsiveness.

Fluoxetine has been blamed for inducing murder or suicide and has been used as a legal defense and as a plaintiff's argument in seeking compensatory and punitive damages in a variety of court cases. In early 1990, a case report examined 6 patients who were depressed but free of recent serious suicidal ideation who developed intense and violent suicidal preoccupation after starting fluoxetine treatment. The authors recommended that fluoxetine be used cautiously. In another study, increased aggression was observed in a group of 19 adult inpatients who were mentally retarded and had epilepsy and a history of current or recent aggressive behavior. Each was taking other psychotropic medication concurrently, with wide individual differences in fluoxetine drug response.

The Lilly research laboratories have published several meta-analyses concluding that fluoxetine is not associated with increased suicidal ideation in treatment for obsessive-compulsive disorder, increased suicidal ideation during pharmacotherapy for depression, or increased aggression. Specifically, a relative risk analysis of 3992 subjects enrolled in several double-blinded placebo-controlled clinical trials covering multiple indications (eg, depression, obesity, bulimia nervosa, obsessive-compulsive disorder, smoking cessation, alcoholism) demonstrated a 4-fold higher likelihood for an aggressive event for patients receiving placebo compared with fluoxetine. A thoughtful review and reinterpretation of the emergence of suicidal and aggressive behavior during antidepressant treatment places this clinical issue in perspective.

Antidepressant - Citalopram

The antidepressant citalopram, a selective serotonin reuptake inhibitor newly available in the United States, was tested for antiaggressive effects in 15 patients who were chronically violent and hospitalized with schizophrenia. This was a double-blinded placebo-controlled crossover study that lasted 48 weeks (24 wk on active citalopram, 24 wk on citalopram placebo). The trial medication was added to the concurrent neuroleptic treatment. The number of aggressive incidents decreased during the active citalopram treatment.

Anxiolytic - Buspirone

Buspirone, an anxiolytic with antidepressant properties and a serotonin receptor 1A agonist, was reported to improve aggressive and self-injurious behavior in 9 of 14 patients who were mentally retarded; this effect was confirmed in a controlled study by the same group. Several case studies reporting similar results were published. A case of posttraumatic belligerence (patient described as "extremely hostile, argumentative, impulsive, and threatening") was successfully managed with buspirone, which was started several months after the head injury.


Treatment of the underlying psychiatric disorder is essential to reducing violence. A careful assessment, leading to a correct diagnosis, optimizes treatment selection. Clozapine appears to be more effective than other antipsychotics in reducing aggressivity in patients with schizophrenia or schizoaffective disorder.

Beta-blockers, well studied in the treatment of aggressive behavior in patients with brain injury, also may be helpful as an adjunctive agent to antipsychotics for aggression and schizophrenia. Mood stabilizers, such as carbamazepine or valproate are also used in patients with schizophrenia, typically as adjuncts to antipsychotic treatment to decrease the intensity and frequency of agitation and poor impulse control. However, they have not been studied extensively under double-blinded placebo-controlled conditions.

Lithium certainly is useful as an antimanic medication and reduces agitation and aggression in patients with bipolar disorder. Lithium has been reported as being useful in reducing aggression in people who are mentally retarded and in reducing the number of infractions involving violent behavior in a study of prisoners who are recurrently violent. The use of lithium for schizophrenia and aggression has not been adequately investigated.


Patient and Family Education

Although this article has focused on the assessment and management of aggression associated with major mental disorders, the practitioner is at times faced with evaluating victims of domestic violence. A review of domestic violence in the United Kingdom reveals some sobering statistics: every year 1 in 10 women in the UK is physically abused by a partner or ex-partner, and 1 in 4 experience abuse during their lifetime. [32] Spousal abuse has devastating consequences to the entire family unit, as children are often witnesses to these events. Perpetrator programs have focused on ensuring safety and eliminating denial.

In the United States, resources and laws vary by state. For example, in Texas, the law requires doctors to provide safety and shelter information to patients with injuries believed to be caused by family violence and to document this in the medical record (see Interpersonal Violence in Texas: A Physician's Role). In New York, the Office for the Prevention of Domestic Violence was created in 1992 as an executive level state agency and offers community coordination and assistance at the local level, training and technical assistance, and public education and outreach.

The American Psychological Association has information on Controlling Anger; this topic has been popularized in the news and has provided the inspiration for a plethora of lay and self-help groups and courses. Teaching anger management in the community involves both didactic and experiential components, as well as homework to reinforce the points made.

A history of ADHD and conduct disorder are possible risk factors for intimate partner violence. This suggests the need for services and treatment strategies that promote healthy intimate relationships and specifically address the risks for violence in youths with ADHD and conduct disorder. [33] The Center for Education and Research on Mental Health Therapeutics (CERTs) released the Treatment of Maladaptive Aggression in Youth (T-MAY) guidelines in May 2012. The guidelines recommend various treatment principles, including the use of psychosocial interventions as first-line treatment of maladaptive aggression in youth. [34] Aggression Replacement Training (ART) is a cognitive behavioral intervention program to help children and adolescents improve social skill competence and moral reasoning, better manage anger, and reduce aggressive behavior (see Fact Sheet from the NationalCenter for Mental Health Promotion and Youth Violence Prevention).

Other patient and family education Web sites include the following: