Antipsychotic polypharmacy: never say never, but never say always

Combining medications for the treatment of schizophrenia is the rule rather than the exception today (1). However, it is likely also the most prevalent, least evidence-based practice in psychopharmacology today, especially antipsychotic polypharmacy (combining of two or more antipsychotics). The merits and pitfalls of this approach have been much debated by this editorialist (2–7) and within the pages of this journal (8, 9) and also by many others [e.g. 10–16; see also references in (1)]. Antipsychotic polypharmacy/combination treatments are associated with higher medication costs, more side effects, no clearly documented improvement in efficacy, and worse outcomes (e.g. 1). So, why do we keep doing it and in fact why are we increasing our use of antipsychotic polypharmacy/combinations? The easiest thing to take away from studies of antipsychotic monotherapy versus polypharmacy/combinations is that the usual patient with schizophrenia should start on one antipsychotic and receive an adequate dose and duration of it prior to switching to another monotherapy, and to do this prior to resorting to giving two drugs (1–16). One thing that polypharmacy/combination studies have consistently shown is that adding a second drug will not improve psychosis symptoms beyond the expected 20–30% improvement in the ‘median’ patient in studies of schizophrenia, despite the wish to reach for more efficacy by adding another drug (1–16). Thus, many experts and payors have been opposed to this practice. Current treatment guidelines thus strongly support monotherapy for schizophrenia (1–16). However, it may be unrealistic and indeed poor clinical management to always say ‘never’ to polypharmacy/combinations. Rank and file psychiatrists seem to have ignored the call for ‘always say monotherapy’ as the data show that antipsychotic polypharmacy/combination use is growing and is higher than ever, overtaking monotherapy (11), and seemingly impervious to educational and administrative interventions to reduce polypharmacy (12, 17). Although some believe these trends reflect substandard care, the shift toward more antipsychotic polypharmacy/combinations may actually be due to something else, namely, the growing gap between evidence-based practice and practice-based evidence, at least for certain patient populations receiving antipsychotics. That is, while evidence-based practice indeed does support monotherapy of antipsychotics as the best trade-off between risks and benefits, a close look at the specific patient populations from these studies shows that they include patients with schizophrenia who are generally not too disabled to give informed consent for treatment; not so violent and aggressive to be unable to cooperate; not active current substance abusers; and without a long history of failing to respond to numerous monotherapies at adequate doses and duration. Real world psychiatrists, however, have to treat schizophrenic patients who do not respond to numerous monotherapies, who have uncontrolled impulsivity, violence, aggression, and active substance abuse, who are increasingly housed long term on forensic units following commission of violent crimes, and/or who may have borderline, affective, and even antisocial features. What is one to do for these patients when monotherapies fail and there is no evidence for them? The answer seems increasingly to be coming from practice-based evidence. That is, empiric observations among clinicians caring for difficult patients are that some patients who fail standard doses of antipsychotic monotherapies do better with high monotherapy dosing, with two antipsychotics or with an antipsychotic plus another agent. I must admit, I was originally skeptical of this approach, and my initial writings have all criticized antipsychotic polypharmacy as expensive, unproven, and causing more side effects than therapeutic benefits (2–9). However, in recent years, I have become impressed with the therapeutic benefits of antipsychotic high-dose polypharmacy/combinations in selected cases and have even included a few such cases from a forensic psychiatry facility where I consult (Patton State Hospital in San Bernardino, California) and who had good outcomes from high-dose polypharmacy/combination treatments in a recent case book of mine (18–21). Indeed, evidence does support the possibility that there are differences between patients who receive polypharmacy and those who do not, and between patients who benefit from polypharmacy and those who benefit from monotherapy (e.g. 1–16). Did the patients in the study of Langle et al. (1) fare poorer with polypharmacy/augmentation because they received this treatment, or did they receive this treatment because they fared even more poorly on monotherapy? Does the study of Correll et al. (13) show that when patients do not fare well even on clozapine that they respond to polypharmacy? When Essock et al. (16) switched patients off polypharmacy and onto monotherapy, about a third of them relapsed and had to go back to polypharmacy. Who were these patients? Interestingly, in their study population, none of their patients had any significant symptoms of ‘uncontrolled hostility’, so one wonders whether the switches to monotherapy would have failed even more often in a setting where there patients have a great many symptoms of uncontrolled hostility. The paradigm for treating schizophrenia has obviously shifted from using monotherapy while discouraging all uses of polypharmacy/combinations to determining instead who should get polypharmacy/combinations. Treatment guidelines are now needed for this. It would be ideal to have more studies of polypharmacy/combinations to help develop these guidelines, but it is not likely that there will ever be any randomized, multicenter placebo controlled trials for the most difficult patients because of ethical and pragmatic barriers to their study. Perhaps developing an expert consensus from case-based medicine and practice-based evidence is the best approach for now. If so, the existing literature combined with known psychopharmacologic principles already suggest a way forward: No antipsychotic (except perhaps clozapine) has shown to be effective if fewer than 60% of dopamine D2 receptors are blocked. In the absence of clinically available positron emission tomography (PET) scans for individual patients, the next best thing is to rule out pharmacokinetic failures in monotherapy patients not responding to standard doses of antipsychotics by using therapeutic drug monitoring much more frequently than is typical in contemporary psychiatry practice (22, 23). In cases of low plasma drug levels with adequate doses, high doses are called for and in this context are not really high, but merely adequate for that patient. Consider purposely blocking more than 60% of D2 receptors, especially for cases of violence, aggression, and treatment resistant positive symptoms. Empirically, some patients respond to high doses that target higher than the traditional degree of D2 receptor occupancy (e.g. 18–21), and this can be implemented either by high-dose monotherapies or by using two antipsychotics. Demonstrating clear clinical improvement without unacceptable side effects should be a criterion for continuing this approach. For patients with affective symptoms, consider an antidepressant, mood stabilizer, lithium, or even a short-term benzodiazepine and document clear improvement without unacceptable side effects as a criterion for continuing this approach as well. Although it is tempting to leave well enough alone if one of these heroic and possibly controversial high dose or polypharmacy interventions is selected, it may be prudent nevertheless to attempt a trial at conversion to monotherapy just to prove that long-term high doses or polypharmacy/combinations are necessary. In the Essock et al. (16) study, for example, two-third of patients successfully switched to monotherapy, thus documenting the need for return to long-term polypharmacy in the other one-third. We should grab this opportunity to improve the treatment of schizophrenia by defining who are candidates for polypharmacy/augmentation, rather than continuing to try to deny this approach to everyone.