Male gender is associated with deficit schizophrenia: a meta-analysis

Abstract

An association between deficit schizophrenia and male gender could be expected, since male schizophrenic subjects have been repeatedly found more severe than females on several dimensions of severity. Surprisingly, very few studies have confirmed such an association. We performed a more definitive test of this association using a meta-analysis. A pooled odds ratio was computed based on the 23 studies that reported the gender ratio in deficit vs. non-deficit schizophrenia. We tested for the heterogeneity of the association and examined the potential impact of the sampling method, the method used to assess the deficit syndrome, the breadth of diagnoses included and the mean duration of illness. A highly significant association between male gender and deficit schizophrenia was observed (pooled odds ratio=1.75). There was no definitive evidence that differences across studies in sampling methods, breadth of diagnoses included, mean duration of illness and methods to assess the deficit syndrome affected the strength of the association. However, the studies using the ‘Proxy Deficit Syndrome’ method to assess the deficit syndrome yielded qualitatively weaker evidence. This significant association between male gender and deficit schizophrenia may reflect the influence of a gender related factor (e.g. sexual hormones) or gender differences in the liability to different etiologies of schizophrenia. The role of gender as a potential confounder must be closely examined in studies comparing deficit and non-deficit SZ .

Introduction

Carpenter et al (Carpenter et al., 1988, Kirkpatrick et al., 1989) defined deficit schizophrenia (SZ) by the presence of at least two out of six negative symptoms for at least 12 months. In addition, these symptoms could not be explained by anxiety, psychosis, neuroleptic side-effects, concurrent depression or other causes. A very high degree of temporal stability of this distinction has been documented in at least two studies (Amador et al., 1999, Fenton and McGlashan, 1994), and it has been found not to be confounded by neuroleptic side-effects (Bustillo et al., 1995). Among several differences that have been observed between deficit and non-deficit SZ, the poorer outcome and the poorer premorbid adjustment in deficit SZ are probably the best replicated (see Roy et al., in press, for a review). Since male SZ subjects have been repeatedly found to have a poorer outcome and poorer premorbid adjustment than their female counterparts (Castle and Murray, 1991, Hafner and An Der Helden, 1997), an association between deficit SZ and male gender could be speculated. Surprisingly, in a recent review on the validity of deficit vs. non-deficit SZ subtypes (Roy et al., in press), we found no consistent evidence across studies for such an association, although several studies reported an overproportion of males in deficit SZ that did not reach statistical significance in most instances. Consequently, it is possible that these studies did not have sufficient statistical power individually to yield significant evidence for such an association. The present paper used meta-analytic techniques to determine whether pooling these studies would yield significant gender ratio differences in deficit vs. non-deficit SZ.