Schizophrenia and bipolar disorder both show some evidence for genetic anticipation. In addition, significant expansion of anonymous CAG repeats throughout the genome has been detected in both of these disorders. The gene KCNN3, which codes for a small/intermediate conductance, calcium-regulated potassium channel, contains a highly polymorphic CAG-repeat array in exon 1. Initial evidence for association of both schizophrenia and bipolar disorder with increased CAG-repeat length of KCNN3 has not been consistently replicated. In the present study, we performed several meta-analyses to evaluate the pooled evidence for association with CAG-repeat length of KCNN3 derived from case-control and family-based studies of both disorders. Each group of studies was analyzed under two models, including a test for direct association with repeat length, and a test for association with dichotomized repeat-length groups. No evidence for a linear relationship between disease risk and repeat length was observed, as all pooled odds ratios approximated 1.0. Results of dichotomized allele-group analyses were more variable, especially for schizophrenia, where case-control studies found a significant association with longer repeats but family-based studies implicated shorter alleles. The results of these meta-analyses demonstrate that the risks for both schizophrenia and bipolar disorder are largely, if not entirely, independent of CAG-repeat length in exon 1 of KCNN3. This study cannot exclude the possibility that some aspect of this polymorphism, such as repeat-length disparity in heterozygotes, influences risk for these disorders. Further, it remains unknown if this polymorphism, or one in linkage disequilibrium with it, contributes to some distinct feature of the disorder, such as symptom severity or anticipation.
Copyright 2003 Wiley-Liss, Inc.