There is strong evidence from both animal- and in vitro-models that paraoxonase (PON1) is involved in the onset of cardiovascular disease. In humans there is no consensus on this issue and therefore we investigated the effect of PON1 genotype and activity on the incidence of coronary heart disease (CHD) and acute myocardial infarction (AMI) in a large prospective cohort of 17,357 middle-aged women. We applied a case-cohort design using the CHD (n=211) and AMI cases (n=71) and a random sample from the baseline cohort (n=1527). A weighted Cox proportional hazards model was used to estimate age- and multivariate-adjusted hazard ratios (HR) for the PON1 genetic variants (192Q > R and -107C > T) and tertiles of the PON1 arylesterase- and paraoxonase activities. Neither the PON1 genetic variants, nor the PON1 activities affected the incidence of CHD in general, but, an increased paraoxonase activity was associated with a higher risk of AMI: the second and third tertile HR were 1.31 and 2.07, respectively (P-trend=0.029, multivariate model). In the subgroup of never-smokers, paraoxonase activity was associated with an increased risk for AMI: the second and third tertile HR were 4.1 and 4.7, respectively (P-trend=0.009, multivariate model). Additionally, when compared to the lowest paraoxonase tertile in never-smokers, the highest paraoxonase tertile in current-smokers showed a 19.2-fold higher risk for AMI (95%CI: 5.3-69.5, P < 0.0001, multivariate model). In conclusion, this study shows that in middle-aged women paraoxonase activity was associated with an increased risk for AMI and that the risk was modified by the effects of smoking.