Serum antinuclear antibody positivity (ANA) has been associated with elevated serum polychlorinated biphenyls (PCBs) among residents in PCB-polluted areas; however, associations in general populations have not been reported by congener type or with adjustment for mercury. Cross-sectional data on serum PCBs, total blood mercury, ANA, and potential confounders age, race, body mass index, menopausal status, and dietary eicosapentaenoic acid (EPA) were obtained from the 2003-2004 National Health and Nutrition Examination Survey (NHANES) for males and females aged 12-85. PCB congeners were summed separately for dioxin-like and nondioxin-like PCBs; the former were weighted for toxic equivalent factors. Total PCBs by congener type and mercury were analyzed as both continuous log-transformed variables and as categorical quintiles. Logistic regression models were stratified by sex. There were no associations between nondioxin-like PCBs or mercury and ANA among males or females. Among females (n=114 affected and 518 unaffected), adjusting for potential confounders, the prevalence odds for ANA positivity were significantly elevated per incremental increase in log-transformed dioxin-like PCBs (odds ratio {OR}=1.66; 95% confidence interval {CI}=1.24, 2.23); the highest dioxin-like PCB quintile (>0.00425-0.04339ng/g) was significantly associated with 4.04 (95% CI=2.43, 6.70) greater prevalence odds for ANA positivity relative to the lowest quintile (Ptrend<0.001). We present novel findings of an association between low-level dioxin-like PCBs and ANA among women. No associations were observed between mercury and ANA at mercury levels common to the U.S. population.