Several reports demonstrate association between variants of the cytotoxic T lymphocyte antigen-4 (CTLA-4) and autoimmune diseases. CTLA-4 may generate autoimmunity by immune dysregulation, making CTLA-4 an attractive candidate gene for systemic lupus erythematosus (SLE) susceptibility. Previous CTLA-4 association studies with SLE, however, have produced inconsistent results. We have performed a meta-analysis to better assess the purported associations. A total of 14 independent studies (to July 2004) testing association between one or more CTLA-4 polymorphisms and SLE were used in this analysis. We have compared allele and genotype frequencies at four polymorphic sites found in exon-1 (at +49), the promoter region (at -318 and -1722), and the 3' untranslated region (3'UTR) (dinucleotide repeat). We have evaluated both fixed and random effect models, depending on the presence of between-study heterogeneity. The data demonstrate that the exon-1 +49 polymorphism is significantly associated with SLE susceptibility. The overall risk, measured by odds ratio (OR), for exon-1 +49 GG genotype is 1.287 [95% confidence interval (CI)=1.031-1.562, P=0.011]. Stratification by ethnicity indicates the exon-1 +49 GG genotype is associated with SLE, at least in Asians (OR=1.293, 95% CI=1.031-1.620, P=0.026). European-derived populations have an effect of similar magnitude (OR=1.268, 95% CI=0.860-1.870, P=0.230), though not significant. Similar trends are found in allele-specific risk estimates and disease association. The OR for the exon-1 +49 risk allele (G) in Asians is 1.246 (95% CI=1.057-1.469, P=0.009), while Europeans have no evidence of allelic association (OR=0.978, 95% CI=0.833-1.148, P=0.780). In conclusion, this meta-analysis supports the CTLA-4 exon-1 +49 (A/G) polymorphism influencing the risk for developing SLE, especially in Asians.