Context: Low testosterone (T) has been associated with incident metabolic syndrome (MetS), but it remains unclear if this association is independent of sex hormone binding globulin (SHBG). Estradiol (E2) may also be associated with MetS, but few studies have investigated this. Objective: To study the association between baseline sex steroids and the development of incident MetSandto investigate the influence of SHBG, bodymassindex (BMI)andinsulin resistanceonthis risk. Methods: Three thousand three hundred sixty nine community-dwelling men aged 40-79 years were recruited for participation in EMAS. MetS was defined by the updated NCEP ATP III criteria. Testosterone and E2 levels were measured by liquid and gas chromatography/mass spectrometry, respectively. Logistic regression was used to assess the association between sex steroids and incident MetS. Results: One thousand six hundred fifty one men without MetS at baseline were identified. During follow-up, 289 men developed incident MetS, while 1362 men did not develop MetS. Men with lower baseline total T levels were at higher risk for developing MetS [odds ratio (OR) = 1.72, P <.001), even after adjustment for SHBG (OR = 1.43, P =.001), BMI (OR = 1.44, P <.001) or homeostasis model assessment of insulin resistance (HOMA-IR) (OR=1.64, P<.001). E2 was not associated with development of MetS (OR = 1.04; P =.56). However, a lower E2/T ratio was associated with a lower risk of incident MetS (OR = 0.38; P <.001), even after adjustment for SHBG (OR = 0.48; P <.001), BMI (OR = 0.60; P =.001) or HOMA-IR (OR = 0.41; P <.001). Conclusions: Inmen, lowerTlevels, but not E2, are linked withanincreased risk of developing MetS, independent of SHBG, BMI or insulin resistance. A lower E2/T ratio may be protective against developing MetS.