TO THE EDITOR: A key issue in understanding schizophrenia and in developing better drugs for the disease is whether any improvement in the negative symptoms associated with the use of neuroleptics reflects improvement not only in secondary negative symptoms but in primary negative symptoms as well. On the basis of data from a double-blind, controlled study comparing the effects of olanzapine to those of haloperidol and placebo in exacerbated schizophrenic patients (R1558281B), Tollefson and Sanger intended subsequently to determine to what extent the superior total effect on negative symptoms was direct or indirect (R1558282B). The authors defined the direct treatment effect of olanzapine as the additional improvement in negative symptoms remaining after they had corrected for changes in positive symptoms, depressive symptoms, and extrapyramidal symptoms. They found that olanzapine had a greater direct effect than placebo and haloperidol on negative symptoms, hypothesizing that this finding represented an improvement in primary negative symptoms. Some comments, however, are warranted.

The original study was designed not for evaluating the effect on primary negative symptoms but for evaluating the effect on positive symptoms. Tollefson and Sanger state that it was not known what the patients' negative symptom histories were; in turn, it was not known whether they had exhibited a chronic deficit state before the index admission (R1558282B). The authors ignore this limitation, however, when they state that "it is likely that these results could be generalized to patients in a chronic deficit state who are not in an acute exacerbation of schizophrenia." That the benefits of olanzapine on negative symptoms were replicable in a subgroup of patients with prominent negative symptoms was not surprising because this subgroup of patients was a subset of the study group from which the results to be replicated were generated.

The authors do not mention the issue of dealing with the psychic side effects of haloperidol such as reduced speed of thinking, lack of energy, flat affect, and anhedonia (R1558283B). These symptoms are phenomenologically indistinguishable from a range of negative symptoms measured on the Scale for the Assessment of Negative Symptoms, and the number of such symptoms may not necessarily correlate with the number of extrapyramidal side effects. Therefore, the psychic side effects of haloperidol may contribute to the magnitude of the superior direct effect of olanzapine when compared with haloperidol. Improvement in negative symptoms due to changes in positive symptoms not captured by the Brief Psychiatric Rating Scale may be another source of bias.