J R Geddes1, S M Lawrie. 1. University Department of Psychiatry, Warneford Hospital, Oxford.
Abstract
BACKGROUND: This systematic review of the association between schizophrenia and obstetric complications (OCs) was performed to assess the degree of heterogeneity between studies and to quantify the increased risk of schizophrenia in exposed subjects. METHOD: Twenty case-control, one prospective cohort and two historical cohort studies were identified using a MEDLINE search supplemented by a manual search and direct communication with other researchers. Individual odds ratios were calculated for each study and a pooled estimate produced. The effect of methodological variation between studies was further assessed by partitioning studies according to study characteristics. RESULTS: Overall, there was no significant heterogeneity of effect between studies, and the pooled odds ratio for the exposure to OCs on subsequent development of schizophrenia was 2.0 (95% CI 1.6-2.4). However, partitioning according to study design revealed significant heterogeneity (z = -2.60, P < 0.01) between the pooled estimate from the case-control studies and that from the historical cohort studies. There was a gap on the funnel plot in the region of small studies finding no effect. CONCLUSIONS: Although the result indicates that subjects exposed to OCs are twice as likely to develop schizophrenia, the analysis suggests that the observed association could be inflated by both selection and publication biases.
BACKGROUND: This systematic review of the association between schizophrenia and obstetric complications (OCs) was performed to assess the degree of heterogeneity between studies and to quantify the increased risk of schizophrenia in exposed subjects. METHOD: Twenty case-control, one prospective cohort and two historical cohort studies were identified using a MEDLINE search supplemented by a manual search and direct communication with other researchers. Individual odds ratios were calculated for each study and a pooled estimate produced. The effect of methodological variation between studies was further assessed by partitioning studies according to study characteristics. RESULTS: Overall, there was no significant heterogeneity of effect between studies, and the pooled odds ratio for the exposure to OCs on subsequent development of schizophrenia was 2.0 (95% CI 1.6-2.4). However, partitioning according to study design revealed significant heterogeneity (z = -2.60, P < 0.01) between the pooled estimate from the case-control studies and that from the historical cohort studies. There was a gap on the funnel plot in the region of small studies finding no effect. CONCLUSIONS: Although the result indicates that subjects exposed to OCs are twice as likely to develop schizophrenia, the analysis suggests that the observed association could be inflated by both selection and publication biases.