and Squires, 1977; Möhler and Okada, 1977). This receptor has been localized on certain subunits of the GABA_A-benzodiazepine receptor-chloride channel complex (e.g., Schwartz, 1988; Haefely, 1990). One may anticipate the identification of endogenous chemical factors that activate or inhibit benzodiazepine receptors and play a significant role in the neural mechanisms of intense affect, extending also to aggressive and violent behavior. At present, a wide range of substances has been synthesized that alter the activity of benzodiazepine receptors and have considerable therapeutic promise in the acute as well as the long-term treatment of violent and aggressive individuals.

One of the first identified major properties of benzodiazepines was their taming and antiaggressive effect on wild and domesticated animals ranging from large primates to various zoo and laboratory animals (Table 10A; e.g., Heise and Boff, 1961; Heuschele, 1961; Langfeldt and Ursin, 1971). In the past 30 years, benzodiazepines' effects on many types of aggressive behavior in animals, as well as on hostile, aggressive and violent behavior in clinical populations, have been assessed (for reviews, see DiMascio, 1973; Tupin, 1985; Miczek, 1987; Brizer, 1988). Several important pharmacologic, clinical, and behavioral insights about benzodiazepines and aggressive or violent behavior have emerged.

Benzodiazepines effectively reduce aggressive behavior in animals and humans that is primarily of a defensive nature. Consistent evidence documents that chlordiazepoxide, diazepam, oxazepam, and other benzodiazepines effectively reduce retaliatory, defensive, and flight reactions in various animal preparations. As detailed in Table 10A, when feral animals are provoked by an approaching experimenter, or when laboratory rats or cats are exposed to painful electric shock pulses or electrical stimulation of limbic and hypothalamic structures, they react with retaliative and defensive acts (e.g., Christmas and Maxwell, 1970; Malick, 1970; Langfeldt and Ursin, 1971; Fukuda and Tsumagari, 1983; Blanchard et al., 1989; Sulcova and Krsiak, 1989; Kalin and Shelton, 1989). Benzodiazepines decrease these defensive reactions on acute administration in a dose range that is lower than that sufficient to cause sedation or muscle relaxation.

Offensive and charging aggressive behavior in animals as well as assaultive and combative behavior in humans may also be reduced

by benzodiazepines, as demonstrated by the early studies with chlordiazepoxide and diazepam, and more recently with lorazepam, oxazepam, and midazolam (e.g., Lion, 1979; Bond et al., 1989; see Tables 10A and 10B, 11A and 11B). For example in animal studies, chlordiazepoxide, diazepam, alprazolam, or other benzodiazepines cause isolated mice to pursue, threaten, and bite opponents less frequently (e.g., Da Vanzo et al., 1966; Poshivalov et al., 1987; Krsiak and Sulcova, 1990), and reduce the aggression of dominant rhesus monkeys toward group members (e.g., Delgado et al., 1976). In clinical studies, different kinds of benzodiazepines reduce various types of aggressive and violent behavior ranging from from assaults, homicidal violence, and destructive or abusive behavior, to temper tantrums, postictal aggressive outbursts, agitation, and feelings of hostility and irritability (see Tables 11A and 11B; e.g., Tobin and Lewis, 1960; Monroe and Dale, 1967; Rickels and Downing, 1974; Lion, 1979; Keats and Mukherjee, 1988).

Shortly after the introduction of chlordiazepoxide (Librium®) and diazepam (Valium®), these substances were evaluated in the medication of patient inmates in penal institutions and juvenile detention centers (Kalina, 1964; Gleser et al., 1965). Hostility, abusiveness, belligerence, destructiveness, and assaultiveness were found to be markedly reduced in these open, nonplacebo-controlled trials with both benzodiazepines. Even on acute administration, midazolam stopped temper tantrums, assaults, and self-injurious behavior in certain mentally retarded patients (Bond et al., 1989).

A significant limitation of the antiaggressive effects of acutely administered benzodiazepines is their strongly sedating and muscle-relaxant side effects. Several extensive dose-effect studies in animal preparations indicate that the dose range for the antiaggressive effects overlaps considerably that for sedation (see Table 10A). Although these side effects are acceptable in an emergency situation with a violent individual, they indicate that the antiaggressive effects of acutely administered benzodiazepines cannot be considered behaviorally specific. With repeated administration of benzodiazepines, substantial tolerance to the sedative and muscle-relaxant effects is seen (File, 1985; Sepinwall et al., 1978).

The most problematic feature of benzodiazepines and aggressive behavior is their potential to increase this behavior under several conditions in a considerable proportion of animals and in humans (see Tables 10 and 11). Starting in the 1960s a series of case reports, as well as experimentally well-controlled double-blind studies, alerted to the paradoxical aggression-heightening effects of benzodiazepines in certain individuals (Lion et al., 1975a;

Salzman et al., 1969, 1974; DiMascio et al., 1969; Kochansky et al., 1975, 1977; Griffiths et al., 1983; Lipman et al., 1986). For example, clonazepam treatment had to be discontinued in epileptic patients due to the emergence of aggressive outbursts and temper tantrums (Guldenpfennig, 1973). Also, several case reports indicate that alprazolam may induce hostile, agitated, irritable, and physically assaultive behavior in panic disorder patients (e.g., Rosenbaum et al., 1984; Strahan et al., 1985; Gardner and Cowdry, 1985; Pecknold and Fleury, 1986).

Every review during the past two decades points to the increased hostility and violent episodes (paradoxical rage) in a certain proportion of individuals treated with benzodiazepines, although opinions differ as to the frequency or rarity of these "paradoxical" responses (see, Tables 11A and 11B; e.g., Rickels and Downing, 1974; Zisook et al., 1978; Lipman et al., 1986; Dietch and Jennings, 1988). Violent outbursts are more likely in certain individuals as a result of toxic reactions to diazepam, chlordiazepoxide, clonazepam, and alprazolam, but are apparently absent with oxazepam (Bond and Lader, 1979; Sheard, 1983; Eichelman, 1987; Dietch and Jennings, 1988). The benzodiazepine dose is a critical determinant of whether or not irritability, hostility, or violent acts may occur (e.g., Azcarate, 1975). Systematic dose-effect determinations in animal aggression preparations demonstrate the aggression-increasing effects of lower diazepam doses and the opposite, aggression-decreasing effects at higher doses (Table 10A). There is also evidence that during the course of chronic benzodiazepine treatment, and on withdrawal from prolonged treatment, increased irritability and aggressive behavior may emerge in a certain proportion of individuals (Table 10B and 11B; e.g., File, 1986a,b; Yoshimura et al., 1987; Yudofsky et al., 1987).

Because benzodiazepines are the most widely prescribed psychoactive drugs worldwide, an unambiguous assessment of their violence-controlling efficacy as well as their so-called paradoxical aggression-enhancing effects is urgently needed. Of particular urgency is the improved psychiatric and neurologic diagnosis of those patients who are prone to exhibit increased hostility and violent outbursts. It is feasible that the regulation of the GABA_A-benzodiazepine receptor is critically determined by genetic predispositions and modulated by life experiences with affective aggressive behaviors. Pharmacologic probes of the GABA_A-benzodiazepine receptor complex should be investigated for its utility as a diagnostic tool in individuals with various anxiety disorders and also in those with a propensity for violent outbursts.

In recent years, the range of substances activating and blocking