Despite extensive knowledge about abnormal lipid patterns in patients with end-stage renal disease, the association between cholesterol and the development of renal dysfunction is unclear. We evaluated this association in a prospective cohort study among 4,483 initially healthy men participating in the Physicians' Health Study who provided blood samples in 1982 and 1996. Main outcome measures were elevated creatinine, defined as >/= 1.5 mg/dl (133 micromol/L), and reduced estimated creatinine clearance, defined as </=55 ml/min. Cholesterol parameters included total cholesterol (<200, 200 to 239, and >/= 240 mg/dl), HDL (<40 or >/= 40 mg/dl), total non-HDL cholesterol, and the ratio of total cholesterol to HDL. We used logistic regression to calculate age- and multivariable adjusted odds ratios as a measure for the relative risk. After 14 yr, 134 men (3.0%) had elevated creatinine and 244 (5.4%) had reduced creatinine clearance. The multivariable relative risk for elevated creatinine was 1.77 (95% confidence interval [CI], 1.10 to 2.86) for total cholesterol >/= 240 mg/dl, 2.16 (95% CI, 1.42 to 3.27) for HDL <40 mg/dl, 2.34 (95% CI, 1.34 to 4.07) for the highest quartile of total cholesterol/HDL ratio (>/= >6.8), and 2.16 (95% CI, 1.22 to 3.80) for the highest quartile of non-HDL cholesterol (>/= 196.1). Similar although smaller associations were observed between cholesterol parameters and reduced creatinine clearance. Elevated total cholesterol, high non-HDL cholesterol, a high ratio of total cholesterol/HDL, and low HDL in particular were significantly associated with an increased risk of developing renal dysfunction in men with an initial creatinine <1.5 mg/dl.