This meta-analysis attempted to derive pooled estimates for the putative association between echocardiographic or electrocardiographic left ventricular hypertrophy and the deletion/insertion (D/I) polymorphism of the angiotensin-I converting enzyme. Casecontrol studies were combined, using the Mantel and Haenszel approach. Joint P-values for continuous variables were calculated by Stouffer's method. Continuous measurements of left ventricular mass, which were reported in different units, were expressed on a percentage scale using the within-study mean of the II genotype as the denominator. The computerised database used for this analysis, included 28 reports with an overall sample size of 6638 subjects. The prevalence of the D allele was significantly lower in Japanese (37.2%) than in Caucasians (56.2%). A funnel plot including 12 casecontrol studies (4094 subjects) suggested that no publication bias was present. Overall, left ventricular hypertrophy was not associated with the D allele. Compared with the II genotype, the excess risks of left ventricular hypertrophy associated with DD and DI geno-