""" After adjusting for possible confounding variables, BMI was an independent risk factor for low testosterone level """

So after adjusting for BMI, sugary beverage consumption had nothing to do with T. But BMI is correlated with T which is (very) old news, and sugary drink consumption is presumably correlated with BMI. Yawn.