Response

We read with interest the article by Moncrieff et al, recently published in your Journal. ^{Reference Moncrieff, Pillai, Marston, Lewis, Barnes and Johnson1} The authors investigated the association between relapse and the outcome of schizophrenia and recurrent psychotic disorders over a 24-month follow-up period. The study findings showed no difference between those who relapsed and those who did not on changes in social functioning, quality of life, symptom scores or overall employment rates between baseline and follow-up. The authors concluded that their data provide little evidence that relapse has a detrimental effect in the long term on people with schizophrenia and recurrent psychosis.

In our opinion, the finding of little evidence for an association between relapse and the outcome of recurrent non-affective psychotic disorders deserves further discussion. In particular, we see two concerns.

First, the authors compared patients who relapsed and those who did not. They missed the opportunity to consider in their analysis the duration of relapses (i.e. time spent in relapse) as a prognostic factor. For instance, the findings from one of the largest prospective longitudinal studies in a cohort of people with recent-onset psychoses (IOWA Longitudinal Study) showed an association between relapse duration and neurobiological indices, such as general and regional brain measures, supporting the notion that relapses might have ‘neurotoxic’ effects. ^{Reference Andreasen, Liu, Ziebell, Vora and Ho2} In the same study, simply counting the number of relapses, on the other hand, had non-predictive value on such outcome measures. Thus, it could be pointed out that a more detailed investigation of relapse characteristics might have shown a different impact of such a variable on the outcome measures of the study.

Second, the mean age of patients was 44.25 (s.d. = 12.82) years for those who relapsed and 47.30 (s.d = 11.21) for those who did not. Both mean ages represented in the study refer to a chronic condition over the course of schizophrenia. This may explain why the paired analysis comparing individuals’ 24-month outcome measures with their baseline scores showed no statistically significant changes. Both the baseline scores in a sample of patients with a chronic condition and the duration of follow-up pose the question to the reader as to whether the study design was capable of detecting clinically meaningful changes.

Even though it remains possible that relapse is detrimental to prognosis early on in the course of schizophrenia specifically, ^{Reference Vita, De Peri and Borgwardt3} more refined prospective longitudinal studies are warranted to answer the question as to whether relapse also affects the outcome of recurrent non-affective psychotic disorders, in order to guide clinicians` informed decision-making about the use of antipsychotic treatment in individuals with long-term conditions.