A meta-analysis of P50 studies in patients with schizophrenia and relatives: Differences in methodology between research groups

Objective: To determine whether patients with schizophrenia as well as their relatives show deficits in sensory gating reflected by an abnormal P50 ratio and to quantify the differences from controls. Methods: A systematic search on articles published between 1982 and 2006 was conducted. 28 patient studies that were suitable for analysis including 891 patients and 686 controls were retrieved. Six studies on P50 of relatives of schizophrenic patients were identified, including 317 relatives and 294 controls. Results: In the patient studies we found an P50 effect size of 1.28 (SD=0.72). We confirmed high variability in outcomes across studies. Almost half of the studies included where published by one laboratory of the University of Colorado and these results differed significantly from the results found in studies performed in other laboratories. We found correlations between effect size outcome and sound intensity, filter settings and subjects' position which could be explained by differences between the Colorado laboratory and the other groups. In the relative studies we found a mean P50 effect size of 0.85 (+/-0.42). Conclusions: The differences in methodology and lack of reported demographics and methodology including raters blinding in some studies makes it hard to compare results across studies and to evaluate the validity and reliability of P50 as a candidate endophenotype for schizophrenia. There are large differences in outcomes from Colorado studies and non-Colorado studies. In contrast to the Colorado studies in the non-Colorado studies P50 suppression would not qualify as an endophenotype for schizophrenia. These differences might be explained by the differences in methodology e.g. lower levels of sound intensity, differences in filter settings and subjects' position. Finally we make some recommendations for future research based on the outcomes of this meta-analysis. (c) 2007 Elsevier B.V. All rights reserved.