Epidemiological studies have shown that low testosterone is associated with metabolic syndrome (MetS) in Caucasian men. We investigated whether testosterone level is related to the prevalence of MetS in middle-aged Japanese men. A cross-sectional survey was conducted in 194 men aged 30-64 years (49 ± 9). Blood sampling was performed in the morning after a 12-h fast, and the relationship between plasma hormone and MetS was analyzed. Low total testosterone was associated with MetS according to the Japanese criteria (HRs of 2.02 by quartile of testosterone; 95% CI¼1.43-2.87) and the International Diabetes Federation criteria (HRs of 1.68 by quartile of testosterone; 95% CI¼1.25-2.25). Age-adjusted regression analyses revealed that testosterone was significantly related to the MetS parameters of obesity (b¼À0.365 and À0.343 for waist circumference and body mass index, respectively), hypertension (b¼À0.278 and À0.157 for systolic and diastolic blood pressure, respectively), dyslipidemia (b¼À0.242 and 0.228 for triglycerides and high-density lipoprotein cholesterol, respectively), insulin resistance (b¼À0.253 and À0.333 for fasting plasma glucose and homeostasis model assessment of insulin resistance, respectively) and adiponectin (b¼0.216). Inclusion of waist circumference into the model largely weakened the association of testosterone with other metabolic risk factors. In contrast, high estradiol was associated with MetS and its parameters, mostly attributing to the positive correlation between estradiol and obesity. Dehydroepiandrosterone sulfate was not associated with MetS or its parameters. These results suggest that low testosterone is associated with MetS and its parameters in middle-aged Japanese men. The association between estradiol and MetS needs further investigation.