To test whether the diurnal rhythm of cholesterol synthesis in humans is entrained to meal timing, the effect of a 6.5-h delay of mealtimes was investigated in four normal lipidemic male subjects. Cholesterol fractional synthetic rate was measured by deuterium incorporation from body water using blood sampling every 2 h. The baseline was a 24-h control period in which three Western-style meals were consumed at 0700, 1150, and 1640, followed by 3 days in which meals were delayed by 6.5 h, i.e., meals consumed at 1330, 1820, and 2310 without changing the sleep-wake and light-dark cycles. Cholesterol synthesis was maximal at 2200 +/- 0200 and minimal at 1130 +/- 0050 on the baseline day. On day 1 of the shifted meals, the maximum was delayed 6.0 +/- 0.5 h and the nadir was not changed. On day 3, the maximum was delayed 8.6 +/- 3.7 h and the minimum was delayed 6.5 +/- 2.4 h from baseline. The mean amplitude of the cholesterol rhythm was significantly greater on day 3,233 +/- 35%, compared with baseline which was 109 +/- 15%. A strong negative correlation (r = -0.66 +/- 0.10) was found between the rhythms of cholesterol synthesis and cortisol during the baseline day, but there was a phase delay in the rhythm of cholesterol synthesis relative to cortisol on day 1 and day 3. Findings indicate that the 24-h variation in cholesterol synthesis is strongly dependent on meal timing.