Fenofibrate increases plasma homocysteine. Because the concentration of plasma homocysteine depends on renal function, we postulate that increases in plasma homocysteine are a result of the known impairment of renal function caused by fenofibrate. Gemfibrozil, another fibrate, does not affect renal function. In a crossover study we tested whether gemfibrozil would raise homocysteine. 22 patients who had hypertriglyceridaemia were given 900 mg gemfibrozil or 200 mg fenofibrate daily for 6 weeks. Lipids were altered similarly, but homocysteine, creatinine, and cystatin C were raised by fenofibrate but not by gemfibrozil (p for differences between treatment effects: 0.007, 0.006, and 0.040, respectively). We propose gemfibrozil should be the fibrate of choice.