I'm quoting the abstract, but it seems a rather weak result.
1,410 men aged 20 + years who attended the morning examination session of the Third National Health and Nutrition Examination Survey (1988–1991) were included in the analysis. Coffee and soft drink consumption was assessed using a food frequency questionnaire. Daily caffeine intake was estimated by multiplying caffeine content per cup times the daily frequency of coffee, tea, or soft drink consumption. Serum levels of hormones and SHBG were measured by immunoassay. Associations of frequency of beverage consumption or estimated caffeine intake with hormone levels were examined using multivariable linear regression.
Coffee consumption was positively associated with SHBG concentration (p = 0.045) taking lifestyle factors into account, but mutually adjusting for testosterone and estradiol attenuated the association; no association with SHBG was observed for soft drink consumption or caffeine intake. No associations between caffeinated beverage consumption and androgen or estrogen concentrations were observed.