Chronic arsenic poisoning is a major worldwide public health problem. Recently, we had reported chronic arsenic poisoning was dose-dependently associated with ventricular abnormalities quantified by electrocardiographic QT prolongation linking to atherosclerotic diseases. An association of chronic arsenic poisoning with ventricular repolarization inhomogeneity quantified by QT dispersion (QTD) is of particular interest from a theoretical and practical perspective. We aimed to further elucidate (1) the association of chronic arsenic exposure with ventricular abnormalities quantified by QTD, (2) the association of QTD with atherosclerotic diseases and (3) the predictability of QTD for long-term mortality in subjects with chronic arsenic poisoning. We followed up 280 men and 355 women living in arseniasis-endemic area in southwestern coast in Taiwan for 17 years. QTD in electrocardiogram and carotid intima-media thickness by ultrasonography were measured. Coronary artery disease was diagnosed by an abnormal electrocardiogram and a definite history. Cumulative arsenic exposure was significantly associated with QTD showing a dose-response relationship (P < 0.001). Significant associations of the QTD with coronary artery disease and carotid atherosclerosis existed after adjustment for potential confounders in the multiple linear regression analysis (all P values < 0.05). In the multivariate Cox regression analyses, the hazard ratios (95% confidence interval, P value) of cumulative cardiovascular and all-cause mortality were 3.9 (2.1-6.2, P = 0.002) and 1.4 (0.9-2.3, P = 0.10), respectively, for QTD > or = 65 ms compared with QTD < 65. QTD may be indicated as an early biomarker for atherosclerotic diseases and a significant and strong predictor of cardiovascular mortality in population with chronic arsenic exposure.