OBJECTIVE : Whether endogenous testosterone concentrations are associated with atrial fibrillation (AF) development is not well established. We assessed the association between plasma total testosterone concentrations and incident AF in a population-based longitudinal study. METHODS : Using data from the prospective Atherosclerosis Risk in Communities (ARIC) study, we identified incident AF among 9282 participants who had plasma total testosterone measured by liquid chromatography tandem mass spectrometry at Visit 4 (1996-1998). METHODS : AF cases were identified by electrocardiograms performed during study visits, hospital records/discharge codes, and death certificates through 2013. We estimated hazard ratios (HRs) and 95% confidence intervals (95% CIs) for incident AF across quartiles of plasma total testosterone, stratified by sex, with multivariable Cox models. RESULTS : The mean age of the participant sample at ARIC Visit 4 was 63 years (range 52-75); 54.5% were women. Mean (SD) plasma total testosterone levels were 537 ng/dL (213) for men and 27.6 ng/dL (34.7) for women. Over a mean of 13.7 years of follow-up, 1664 incident cases of AF were identified. Comparing those in the highest quartile of plasma total testosterone concentration to those in the lowest quartile and after adjustment for potential confounding variables, there was a positive association between plasma total testosterone and incident AF in men (HR 1.33, 95% CI 1.07, 1.66), but no such association in women (HR 0.99, 95% CI 0.80, 1.22). Conclusion A higher plasma total testosterone concentration was associated with a modestly greater incidence rate of AF in men.