Failing to Gain: Another Explanation of Cognitive Change in Schizophrenia and Other Psychoses in the Decade Following the First Episode

to the editor: The decade-long follow-up study of cognitive ability in first-episode psychosis by Zanelli and colleagues (1), published in the October 2019 issue of the Journal, represents one of the largest and longest such studies to date. Significant declines were reported in IQ, verbal knowledge, and memory across the follow-up, while performance on measures of processing speed and executive function remained stable. This was interpreted as reflecting a longitudinal decrement in crystallized cognitive abilities, a finding contradictory to available evidence (2). We believe that a failure to make age-appropriate gains is a more likely interpretation, particularly given that fluid intellectual abilities that support crystallized abilities remained stable. We suggest caution regarding the authors’ interpretation of a possible pathological process, given two key limitations.

First, standardization of cognitive test scores in the whole sample was based only on the normative sample data. Should crystallized intelligence in the normative sample increase over time, as is expected over the life course, but remain unchanged in patients, then this apparent stability would appear as a decrement. This was demonstrated in a longitudinal study of childhood schizophrenia, whereby age-scaled scores show a degenerative pattern, while raw scores indicated that this pattern resulted from failure to make age-appropriate gains (3). Analysis of raw scores is necessary to truly establish the pattern of performance and confirm a deteriorating or unchanging pattern over time.

Second, the authors adjusted for baseline education level, while follow-up education levels were not considered. We hypothesize that raw scores on measures of crystallized intelligence remain stable in participants with schizophrenia (and other psychoses) over time rather than decline. This may result from the illness or limited social and educational opportunities, which would adversely affect acquisition of skills. Thus, follow-up education levels would better index skill development over the life course.

A further potential limitation is that change trajectories were examined in a subset of healthy individuals with low IQ compared with those with higher IQ; the same analysis was not performed in the schizophrenia group. Our recent cross-sectional study (4) of patients with established schizophrenia identified a putative decline in fluid cognition only in patients with low crystallized abilities. We suggest that a comparison of schizophrenia patients with lower and higher IQ is necessary to understand trajectories.