Although erectile dysfunction is frequently seen in patients with manifestations of arteriosclerotic disease, the independent contribution of serum cholesterol in predicting erectile dysfunction is unclear. The aim of this study was to examine the relation between serum cholesterol and erectile dysfunction. Medical histories, physical examinations, and blood tests were obtained at Cooper Clinic, Dallas, Texas, from 3,250 men aged 26-83 years (mean, 51 years) without erectile dysfunction at their first visit, who had one more clinic visit, all between 1987 and 1991. These men were followed 6-48 months after the first clinic visit (mean, 22 months). Erectile dysfunction was reported in 71 men (2.2%) during follow-up. Every mmol/liter of increase in total cholesterol was associated with 1.32 times the risk of erectile dysfunction (95% confidence interval 1.04-1.68), while every mmol/liter of increase in high density lipoprotein cholesterol was associated with 0.38 times the risk (95% confidence interval 0.18-0.80). Men with a high density lipoprotein cholesterol measurement over 1.55 mmol/liter (60 mg/dl) had 0.30 times the risk (95% confidence interval 0.09-1.03) as did men with less than 0.78 mmol/liter (30 mg/dl). Men with total cholesterol over 6.21 mmol/liter (240 mg/dl) had 1.83 times the risk (95% confidence interval 1.00-3.37) as did men with less than 4.65 mmol/liter (180 mg/dl). Those differences remained essentially unchanged after adjustment for other potential confounders. The authors conclude that a high level of total cholesterol and a low level of high density lipoprotein cholesterol are important risk factors for erectile dysfunction.