Forensic schizophrenia patients are often treated with higher doses and with more than one antipsychotic.Reference Morrissette and Stahl ^{1 –} Reference Connolly and Taylor ⁵ It is assumed that the fundaments of this practice are the lack of responses to the usual doses and the attempts to control violent behavior, though the approach is assumed to be often unnecessary and supporting evidence for its use is weak.Reference Morrissette and Stahl ¹

The main topic of “forensic psychopharmacology” is the treatment of aggression. However, the questions of a standardized definition and, consequently, clinical approaches to aggression remain unresolved. Moreover, typical forensic schizophrenia inpatients (at least in Croatia) are too treatment-resistant to have good insight into their condition and the need for continuous treatment, but they take medications and are rarely violent.

The situation is further complicated by the fact that violence may be differently motivated, and that approaches to different types of aggression should differ.Reference Stahl ⁶ For instance, psychotically motivated offense is probably the main reason for hospitalization of forensic schizophrenia patients, while studies imply that impulsive violence is the main type exhibited by hospitalized patients.Reference Stahl ^{6 ,} Reference Quanbeck ⁷ Do we then treat current aggression, reasons for past aggression, or “future danger to self or others”? This unresolved question is what motivates clinicians to prescribe certain dosages.

Thus far, we have reported that forensic patients’ histories of mechanical restraint are associated with more intensive positive symptoms and also with the characteristics of personality that may predispose the patients to impulsive reactions.Reference Margetić, Aukst Margetić and Ivanec ⁸ It is worth noting that the great majority of these patients had been restrained in previous years and none in the 3 months prior to the study. As the use of restraints is dominantly linked with aggression, the findings suggest that the patients’ current doses of antipsychotics may be related to aggressive behavior that took place in the past, and which was not only a consequence of symptoms but also associated with the characteristics of their personality.Reference Margetić, Aukst Margetić and Ivanec ⁸ Moreover, the findings are in accordance with some previous findings that indicated that violence is generally more common in treatment-resistant psychotic patients.Reference Ehmann, Smith and Yamamoto ^{9 –} Reference Conley and Buchanan ¹¹ It is logical to expect that those patients “need” higher doses.

However, we had conducted a larger study, and one of the aims was to assess the relationships between treatment strategies and different behaviors in this population. There are no reports related to a possible association of the doses with the seriousness of offense or suicidality that occurs more frequently in aggressive schizophrenia patients.Reference Apter, Kotler and Sevy ^{12 –} Reference McGirr and Turecki ¹⁴ Thus, it would be valuable to present data related to the issues, while the more detailed descriptions of the selection of the subjects and the instruments used are present elsewhere.Reference Margetić, Aukst Margetić and Ivanec ^{8 ,} Reference Margetić, Aukst Margetić and Ivanec ¹⁵

The sample in our study comprised 54 forensic male patients with schizophrenia or schizoaffective disorder, but without comorbid antisocial personality disorder or substance dependence.

The mean age of participants was 44.7 (±8.39) years. The duration of institutionalization was 6.6 (±5.08) years. The predominant diagnostic subgroup was paranoid schizophrenia (N=42; 77.8%). In the “homicidal” group, 4 of the 22 patients attempted homicide, while others committed homicide. In the “other offenses group” (N=32; 59.3% of all patients), the predominant act was assault on others (N=27; 84.4% of the group). Fifteen patients attempted suicide. All attempts had taken place a number of years before the study, and all but 2 were committed before hospitalization. The Positive and Negative Syndrome Scale (PANSS) score was 85.4 (±15.66); the positive symptoms subscale score was 19.4 (±4.52), the negative symptoms subscale score was 23.2 (±5.23), and the general psychopathology symptoms subscale score was 42.7 (±8.61). The mean chlorpromazine equivalent dose, calculated according to recommendations,Reference Gardner, Murphy and O’Donnell ¹⁶ was 913.3 (±489.04) CPZeq/day. The recommended dosage is usually below 600 CZPeq/day, and the maximal approved dosage in many countries is 1000 CZPeq/day.Reference Hasan, Falkai and Wobrock ¹⁷ In this study, 15 patients were given the dose of up to 600 CZPeq/day, and 15 were given 1000 CZPeq/day or higher. Accordingly, there is no doubt that the majority of patients were treated with higher than usual doses, and a substantial proportion of them with even higher than maximum doses. Consequently, we conducted a more detailed analysis of the motivation for prescribing these higher doses.

The doses correlated with PANSS scores (r=355; p=0.008), positive symptoms (r=0.479, p<0.001), and symptoms of general psychopathology (r=0.324, p=0.017), but not with negative symptoms (r=0.118, p=0.397). All items from the positive subscale, except “hostility” (P7; r=–0.027, p=0.846), correlated with the antipsychotic dosage. The most prominent correlations between positive symptoms items and dosage were for “delusions” (P1; r=0.455, p=0.001) and “conceptual disorganization” (P2; r=0.403, p=0.003). It is worth noting that the mean “hostility” item score was 1.78 (±0.984). There was no significant correlation with any item from the negative subscale. A number of general symptoms subscale items showed significant correlation, which was highest with “unusual thought content” (G9; r=0.429, p=0,001). However, significant correlations were not found with “guilt feelings” (G3; r=0.169, p=0.222) or “depression” (G6; r=0.155, p=0.263).

This cross-sectional study provides only more or less speculative explanations. It has been suggested that forensic patients are treated with higher doses due to violence. Nevertheless, all patients who took part in the study were on a stable medication regime, and their mean score in the “hostility” item was rather low. Thus, one may suppose that the “hostility” item is the target, and that the dosages should increase until side effects or the scores are as low as they are in this study. Here, we emphasize that the aim of forensic treatment is not only to decrease the intensity of schizophrenia symptoms, but also to decrease the risk of violent reoffending and, ultimately, to protect society.

Another reason that higher doses are used is not because the patients are violent, but rather because they “must remain unaggressive.” Here, a number of unresolved questions appear, and one seems to be especially intriguing. If schizophrenia per se does not explain the development of violence, and if the target of treatment with antipsychotics is aggression, does it mean that the dose for treatment of the “must remain unaggressive” condition should be the same as for schizophrenia?

On the other hand, psychotically motivated violent offenses are characteristically dominantly motivated by positive symptoms, among them delusions. The clinician’s intention to deal with these symptoms in patients with long-lasting mean PANSS scores above 80 seems to be an acceptable explanation. Accordingly, if the positive symptoms are intensive, and/or refractory to the treatment, the doses are high because of the clinician’s fear or insecurity. This is in accordance with the findings by Fazel et al,Reference Fazel, Zetterqvist and Larsson ¹⁸ who found that in differently diagnosed persons with criminal convictions, the rate of violence reduction was stronger if the prescribed doses of antipsychotics were higher.

Probably the most interesting finding is that there was no significant difference between the doses of patients who had committed or attempted homicide and those who had committed other offenses, but there was a difference between those who had and those who had not attempted suicide (Table 1). The effect size for the analysis was large (Cohen’s d=0.96). As shown in the table, there were no significant differences in psychopathology scores regardless of whether the patients were dichotomized according to a type of offense or previous suicide attempt.

Table 1 Differences between groups regarding age, antipsychotic dosing, and severity of psychopathology tested with t-tests

Dose: chlorpromazine equivalent daily dose (mg); PANSS: Positive and Negative Syndrome Scale scores; Pos: positive symptoms subscale scores; Neg: negative symptoms subscale scores; Gen: general psychopathology symptoms subscale scores.

In fact, this could be the first description of a possible association between doses of antipsychotics and previous suicidal attempts in forensic schizophrenia patients. Reports with some similarities exist, but the most recent review regarding the association between antipsychotic dosage and suicidal behavior in schizophrenia showed mixed results and provided no clear conclusion.Reference Kasckow, Felmet and Zisook ¹⁹ Furthermore, previous studies used different methodologies, and their comparability is questionable.

For example, Taiminen and KujariReference Taiminen and Kujari ²⁰ found lower neuroleptic doses in suicidal patients, but suggested that results were probably a consequence of the differences in the symptom profile. Hettige et al Reference Hettige, Kennedy and De Luca ²¹ did not confirm the hypothesis that the doses are associated with the history of suicide attempts in patients with schizophrenia. In a recent article related to the topic, Reutfors et al Reference Reutfors, Clapham and Bahmanyar ²² reported a lower suicide risk in schizophrenia patients with a history of extrapyramidal symptoms, and suggested that this “could reflect higher antipsychotic adherence, exposure to higher dosage, or polypharmacy among these patients (p. 341).”

However, the main differences between this and other studies are that the subjects in this study were only male patients with a history of serious, mostly violent, offenses. As mentioned above, aggressive schizophrenia patients are at increased risk for suicide.Reference Apter, Kotler and Sevy ^{12 –} Reference McGirr and Turecki ¹⁴

Taken together, our findings suggest that forensic patients are not treated with higher doses because they are currently violent or have been violent in recent months. We would rather assume that prescribing patterns are influenced by current psychopathology and also by the patients’ previous actions, which in the majority of subjects in this study took place years ago.

Nevertheless, without prospective studies that would follow a larger number of variables including psychopathology, and specific and important types of behavior and personality (which are interconnected with both psychopathology and behavior), we do not understand the clinicians’ motives for differences in prescribing patterns of antipsychotics in this population.