OBJECTIVE In this meta-analysis, we attempted to derive pooled estimates for the putative associations between various cardiovascular-renal disorders and the M235T polymorphism of the angiotensinogen gene. METHODS Case-control studies were combined, using the Mantel and Haenszel approach. Joint P values for continuous variables were calculated by Stouffer's method. Continuous measurements reported in different units were expressed on a percentage scale using the intrastudy mean of the MM genotype as denominator. RESULTS The computerized database used for this analysis included 69 reports with an overall sample size of 27,906 subjects. Overall, possession of the T allele was associated with an increased risk of hypertension. In comparison with the MM reference group (number of studies, n = 32), the excess risk was 31% (P = 0.001) in TT homozygotes and 11% (P = 0.03) in TM heterozygotes. The sensitivity analysis showed that this association was present only in whites (T allelic frequency, f = 42.2%), but not in blacks (f = 77.0%) or Asians (f = 78.0%). Atherosclerotic complications (n = 12), renal microvascular disorders (n = 13), cardiomyopathy (n = 2) or diabetic retinopathy (n = 3) were not correlated with the M235T polymorphism. Publication bias was observed for hypertension, but not for coronary heart disease, including myocardial infarction, and for microvascular nephropathy. Furthermore, in comparison with the MM control group, the circulating angiotensinogen levels (n = 8) were raised by 11 and 7% (P = 0.01) in TT and TM subjects, respectively. In contrast, plasma levels of the angiotensin I converting enzyme (n = 4) and body mass index (n = 15) were not associated with the T allele. CONCLUSIONS The T allele encoding angiotensinogen is not associated with atherosclerotic or microvascular complications, but in Caucasians behaves as a marker for hypertension. This association, which may have been inflated by publication bias, does not necessarily imply causality.