OBJECTIVE To examine changes in plasma lipids and lipoproteins after 51 months of reduced energy intake and sustained weight loss. METHODS One-hundred patients were randomized to one of two dietary interventions for 3 months (weight loss period). Groups A and B received an energy-restricted diet plan of 5.2-6.3 MJ/day but group B was further instructed to replace two of three meals with a nutrient-fortified liquid meal replacement (MR). Upon completion of the weight loss period, all patients were given the same instructions regarding energy intake and were advised to use one MR daily. Body weight and 7 day food diaries were measured monthly or bimonthly and blood lipids at baseline, 3, 9 and 51 months. RESULTS Of the original 100 patients 75 had completed 4 y. Of those 75, 73 had complete lipid records. Baseline body weights of Groups A and B were 90.7+/-14.0 and 91.6+/-9.8 kg, respectively. The percentage change in total cholesterol (%DeltaTC) decreased in a linear fashion with increasing weight loss, when all data was combined, but did not approach statistical significance (P< or =0.26, r=0.02). Further regression analysis found a significant negative linear relationship (P< or =0.0001, r=0.69) between initial total cholesterol (TC) concentrations and %DeltaTC. Hence, data from 27 of the 73 completers who exhibited an elevated serum total cholesterol (> or =6.2 mmol/l) were isolated and analyzed further. Baseline TC was 6.75+/-0.64, 5.85+/-0.63 at 9 months (P<0.05) and 5.76+/-0.52 mmol/l at 51 months (P<0.05). Similar values for VLDL-cholesterol were 1.33+/-0.80, 0.74+/-0.24 and 0.66+/-0.21 mmol/l by 51 months (P<0.05). Weight decreased by 5.2+/-5.1, 7.6+/-4.9 and 6.7+/-4.6% at 3, 9 and 51 months, respectively. CONCLUSIONS Continuous energy restriction associated with a clinically meaningful weight loss significantly improved the lipid profile of high-risk patients. Similar weight and diet changes occurring in patients with normal plasma cholesterol were either increased or without affect.