ObjectiveTo investigate the association between serum total testosterone (TT) levels and metabolic syndrome (MetS) or its components among adult women.Methods2,678 women from NHANES 2011-2016 were included in this cross-sectional study. MetS was determined according to the National Cholesterol Education Program Adult Treatment Panel III guidelines. The association between serum TT levels and MetS was evaluated by two logistics regression models and the adjusted restricted cubic spline (RCS). Stratified analysis and sensitive analysis were also conducted.ResultsContinuous TT levels were negatively associated with the occurrence of MetS, and the ORs associated with per SD increase in ln TT were 0.70 (95%CI: 0.58-0.85) in 2011-2014 and 0.56 (95%CI: 0.39-0.79) in 2015-2016 in Model A. High TT group were less likely to have MetS (OR=0.60, 95%CI: 0.45-0.80 in 2011-2014 and OR=0.50, 95%CI: 0.32-0.78 in 2015-2016) when compared to the low TT group. When TT levels were divided into quartiles, TT levels were negatively correlated with the incidence of MetS (p for trend < 0.001). Similar trend was observed in Model B. Multivariate-adjusted logistic regression with RCS exhibited that TT had a L-shaped dose–response association with MetS or its components. Interaction analyses revealed that women who were less than 50 years old (OR=0.37, 95%CI: 0.22, 0.63), with depression (OR=0.50, 95%CI: 0.29, 0.87) or being smokers (OR=0.37, 95%CI: 0.23, 0.54) showed lower ORs than those who were over 50 years old (OR=0.66, 95%CI: 0.40, 1.09), without depression (OR=0.59, 95%CI: 0.41, 0.85) or non-smokers (OR=0.59, 95%CI: 0.39, 0.89) when measure the association between ln TT and the occurrence of MetS.ConclusionsOur study indicated that TT levels are negatively correlated with the occurrence of MetS, with interaction effects of age, smoke behaviors, and depressive status.