Current practices in managing acutely disturbed patients at three hospitals in Rio de Janeiro-Brazil: a prevalence study

Background The medical management of aggressive and violent behaviour is a critical situation for which there is little evidence. In order to prepare for a randomised trial, due to start in the psychiatric emergency rooms of Rio de Janeiro in 2001, a survey of current practice was necessary. Methods A seven day survey of pharmacological management of aggressive people with psychosis in the emergency rooms of all four public psychiatric hospitals in Rio de Janeiro, Brazil. Results In one hospital data were not available. Of the 764 people with psychosis attending these ERs, 74 were given IM medication for rapid tranquillisation (9.7%, 2.1/week/100,000). A haloperidol-promethazine mix (with or without other drugs) was used for the majority of patients (83%). Conclusion The haloperidol-promethazine mix, given intramuscularly for rapid tranquilization, is prevalent in Rio, where it is considered both safe and efficient. However, scientific evaluation of all pharmacological approaches to rapid tranquilization of psychotic people is inadequate or incomplete and a randomized trial of IM haloperidol-promethazine is overdue.


Background
Agitated or violent patients constitute 10% of all emergency psychiatric treatment [8]. The majority of these people have severe psychiatric problems such as schizophrenia, affective disorder or substance abuse [8]. Less frequently, organic illness or serious psychological stresses underlie the aggression. The medical management of aggressive and violent behavior by people with schizophrenia or other serious mental illnesses is of continuing central concern to clinicians [2].
In order to gain an overview of the evidence underlying current practice, the authors first undertook a search of the Cochrane Library. The search strategies used for all data-bases mentioned are available from the authors. Relevant randomised clinical trials seem few, small and of debatable clinical utility [19]. For example, we found only one randomised trial (n = 40) comparing haloperidol (IV) with placebo for acutely disturbed psychiatric patients [14]. Only one statistically significant difference was found (RR unimproved by 2 hours 0.19 95% CI 0.04 to 0.9). A very small (n = 12) randomised trial compared benzodiazepines with placebo for acutely disturbed people [11], and, unsurprisingly, found no clear differences. Seven trials (total n = 206), compared benzodiazepines versus typical antipsychotics [9]. Benzodiazepines are statistically significantly more likely to produce an 'improvement' by 1.5 hours (RR 1.6 95% CI 1.02 to 2.5), but patients may also be at greater risk of needing additional injections (RR 0.66 95% CI 0.42 to 1.02). On the other hand, more patients given haloperidol are asleep by three hours than those allocated to benzodiazepines (RR 1.6 95% CI 0.99 to 2.5). Ninety-six people were randomised to trials investigating the value of a benzodiazepine-haloperidol mix versus haloperidol alone for acutely disturbed people [9]. The combination, largely with lorazepam, is no better than haloperidol for all the outcomes measured (eg. unimproved by 1.5 hours RR 0.7 95% CI 0.3 to 1.7), except for being asleep by 3 hours, favouring the mix (RR 2.0 95% CI 1.1 to 3.5). A systematic review of droperidol for acutely disturbed people, now withdrawn from use because of cardiac problems with prolonged use, shows it to be of unclear advantage when compared with placebo (1 RCT, n = 41, RR needing additional injection by 30 minutes 0.46 95% CI 0.2 to 1.2) or haloperidol (1RCT, n = 27, RR needing additional injection by 30 minutes 0.45 95% CI 0.2 to 1.01) [5]. A systematic review of zuclopenthixol acetate (5 RCTs, total n = 413), found no statistically significant differences for outcomes such as 'not sedated by two hours' (RR 0.6 95% CI 0.3 to 1.3) and 'needing another injection' (RR 1.5 95% CI 0.8 to 2.9), when compared to haloperidol, chlorpromazine or clothiapine [6].
With such limited evidence it is understandable that doctors may differ in their choice of drugs. We then sought surveys of the opinion of clinicians in MEDLINE and Psy-cINFO. Groups of doctors in the UK and USA had been asked to list their preferred pharmacological management of acutely aggressive patients. In the USA [3], medical directors of 20 psychiatric emergency rooms were asked about their management of agitated patients. In the UK [4], 28 psychiatrists were asked to give their opinion on how a violent patient with a diagnosis of acute paranoid psychosis should be pharmacologically managed.
The results of these surveys are shown in Table 1 and suggest some similar patterns of drug preference but no consistent favourite regimen.
Management preferred by clinicians, however, may not reflect real-world practice. We searched MEDLINE to identify surveys of Emergency Room practice. In 1989, a prospective survey over 160 days was conducted in a major psychiatric centre located in south London [12]. Patients were included if they had received an injection as an acute treatment and during this period, 102 incidents were reported. In France a similar survey was conducted during 9 months in 1997 [10]. Physical restraint by nurses was used for 64% of people in the UK survey and for 86% of agitated/violent attendees at the French emergency room in Rouen. The results of these two surveys are presented in Table 2.
In the UK, intravenous treatments were common, and the doses employed, high. In France, where aggression due to intoxication was common, loxapine IM was largely used.
Other less informative surveys were identified [15].
About one percent of people suffer from schizophrenia [7]. Over 80% of those people live in low to middle income countries [13]. In global terms, therefore, by far the greatest burden of care of people with schizophrenia falls to those in low and middle-income countries [20]. People whose illness precipitates aggressive or violent behavior are unlikely to be very rare in these societies, so it is reasonable to assume that most psychosis-related aggression is being managed in the developing world. Recognizing this, and that the evidence-base of management is weak, the authors undertook a survey of practice as a preliminary phase of a large, pragmatic randomised trial. This paper reports a survey of drug management of severely disturbed behavior, undertaken in one week in March 2000 in the Psychiatric Emergency Units of Rio de Janeiro, Brazil.

Materials and Methods
The period of the survey covered emergency consultations from Saturday 25 th March 2000 to Friday 31 st March 2000, inclusive.
Data collection is difficult because information is recorded differently in each institution and electronic data are not available. The principle researcher repeatedly visited the emergency rooms and short stay wards of the hospitals and collected data on everyone who had presented for an emergency consultation during the prevalence period. Emergency room notes were also inspected and medical records sought for additional information on use of emergency intramuscular sedation. The main focus of this work the management of disturbed people who were likely to suffer from a psychotic illness. Therefore, whenever a primary diagnosis of substance abuse had been made, data were not recorded.

Results
It was impossible to be confident that the records of use of intramuscular (IM) medication for the 133 people who attended the emergency room of Hospital Jurandir Manfredini were accurate, so data are not presented. Between 3% and 15% of medical notes were not available in the other hospitals, so it was not possible to know if those people were also given emergency sedation (Table 3). During the seven-day period 764 patients attended the emergency room in the three hospitals and at least 74 received emergency sedative intramuscular drugs for psychi-atric conditions that were not clearly induced by abuse of substances. All 74 patients were thought to be suffering from a psychotic illness at the time of sedation but follow up was not possible to confirm the initial impression. No patient received intravenous sedation and no patient received extra doses. The total proportion of those receiving IM sedation was similar among the three hospitals (8-11%). Data on the use of physical restraints were not recorded but it is likely that a high proportion of people were subject to four-point restraint, as this is standard practice in these hospitals. Seclusion was not an option and attempts at talking down were attempted for everyone.
The drugs of choice for IM sedation in the three hospitals in Rio de Janeiro are shown in Table 4. A haloperidol-promethazine mix was the most popular combination and was used in over 60% of episodes. This proportion reached 83% if we include haloperidol-promethazine plus other drugs.

Discussion
Surveys of the emergency sedation of acutely disturbed people with serious mental illness are rare. They are difficult to undertake involve piecing together particularly fragmented records and combining these data with the recent memories of the relevant health professionals. In a Developing World environment this process may be even more problematic with vastly under-resourced services and fully paper-based record keeping of varying standards.
For a population of about 3.5 m covered by the three hospitals, we estimate that at least 74 people per week are sedated for disturbed behavior thought to be the result of serious mental illness. Eventually, a proportion of this group may not have been diagnosed as having had psychotic illness. The authors suspect, that however, because the emergency rooms are only used for psychiatric crises, and as the authors tried to screen out those suspected of abusing substances, the proportion not suffering from a psychotic illness would be very small.
According to these figures, acute sedation for this group in Rio de Janeiro seems to be less prevalent than in central London (Brazil 2.1/week/100,000 vs London 3.3/week/ 100,000) [12]. There may be many reasons for this discrepancy and the authors would not wish to read too much into these findings.
Emergency treatment, however, does seem more consistent in Rio de Janeiro than in London. Eighty three percent of patients receiving any drug for sedation in the emergency room of these hospitals were given promethazine in addition to haloperidol. Brazilian clinicians consider this approach both safe and efficient. The rational for this combination lies in the will to swiftly instigate antipsychotic drugs, and in the sedative and antimuscarinic properties of promethazine. Doses of promethazine are usually between 25-50 mg but, as adjunctive sedative for emergency use, could reach 100 mg IM. The onset of action is about 1-2 hours after intramuscular administration and half-life is 5-14 hours. The main adverse reactions of promethazine are gastrointestinal disturbances, dry mouth and blurred vision. [18] Paradoxical reactions such as CNS stimulation and extrapyramidal symptoms have been reported, but are exceedingly rare.
Despite the high prevalence of use in Brazil, and we have now been alerted to wide use in the Indian Sub-continent, the authors have been unable to find any randomised studies evaluating the use of haloperidol-promethazine mix IM in the psychiatric emergency.

Conclusions
From the global perspective, current use of drugs in the psychiatric emergency is varied [1]. Expert consensus differs, and systematic reviews of the best evidence produce equivocal results [6]. Brazilian psychiatrists are consistent in favouring use of IM haloperidol-promethazine for people who are dangerously aggressive due to psychotic illnesses, but recognize that authority to recommend the use of this combination has to be supported by well-designed, conducted and reported randomized trials. The TREC study (Tranquilização Rápida-Ensaio Clínico), comparing haloperidol-promethazine IM with a benzodiazepine, now running in Brazil and India, is one such trial (protocol available from corresponding author).