Adding and Switching Antipsychotics: In Reply

In Reply: Dr. Hovens and Dr. Loonen raise several important points regarding the methods and results of our study of adding or switching antipsychotic medications for treatment-resistant positive psychotic symptoms. As they correctly note, the design of our study did not permit us to follow patients for whom antipsychotic treatment was discontinued. We specifically asked psychiatrists to report on the response to a medication change (defined as either a switch to or an addition of another medication, and not a change in dosage) of one of their patients who was exhibiting treatment-refractory psychotic symptoms. Although we cannot know for certain, it is unlikely that discontinuing antipsychotic treatment would have led to clinical outcomes similar to those observed in the chronically ill, treatment-resistant sample of patients who switched antipsychotic medications in our study, as suggested by Hovens and Loonen. A clinical trial of antipsychotic discontinuation may be warranted in a limited number of cases—for example, after sustained symptom remission following a single psychotic episode. However, clinical experience suggests and guidelines for the treatment of schizophrenia ( 1 ) recommend continuous antipsychotic therapy to prevent symptom relapse among persons who have experienced multiple episodes of psychotic symptoms.

Hovens and Loonen correctly point out that we did not query psychiatrists in detail about the extent to which their patients adhered to the change in antipsychotic treatment. Because more complicated treatment regimens, such as antipsychotic polypharmacy, have been shown to increase the likelihood of medication nonadherence ( 2 ), it is possible, as suggested by Hovens and Loonen, that the lack of effectiveness of antipsychotic polypharmacy reported by psychiatrists in our study may have been a result of patients' nonadherence to the treatment rather than to its inherent lack of efficacy. However, several randomized controlled trials of antipsychotic combinations have not produced favorable results ( 3 ). Second, although we did not report it in our paper, 28% of psychiatrists in the study who added an antipsychotic and 34% who switched agents agreed that patients' refusal to accept optimal medication management was a barrier to effective treatment. As such, medication nonadherence may have been a complicating factor in both groups of patients, and as suggested by Hovens and Loonen, improving the therapeutic alliance by asking patients about their treatment preferences is one possible strategy for providers to enhance patients' acceptance of evidence-based treatments and improve outcomes.

With respect to evidence-based treatments, we echo the sentiment of Hovens and Loonen regarding the exceedingly low use of clozapine in the sample of treatment-resistant patients in our study. Despite strong evidence for its efficacy, clozapine remains underused in the United States, probably because of administrative barriers ( 4 ) and concerns about agranulocytosis and other admittedly troublesome potential metabolic side effects ( 5 ). However, serious discussion is warranted among clinicians, researchers, and patients affected by medication management decisions to determine whether these concerns about clozapine justify the use of treatment strategies, such as antipsychotic polypharmacy, with little evidence of efficacy and demonstrated safety concerns of their own.