We evaluated dietary counselling by dietitians in hypercholesterolaemic hypertensives over 6 months. A total of 141 patients were randomly assigned to intensive advice or usual care. Body weight fell significantly in the intervention group, but did not in the controls. There was a modest but significant fall in total cholesterol from 7.1 to 6.8 and 7.1 to 6.9 mmol/l in the diet and the control groups, respectively (4 and 3%). A similar pattern emerged from the triglyceride measurements. Low-density lipoprotein fell in both groups, but only achieved significance (P less than 0.05) in the intervention group. High-density lipoprotein did not change. There was a more marked change in cholesterol when serum levels during the study were compared with the previous annual review. These falls occurred after selection for the study but before random allocation to groups. They are unlikely to reflect regression to the mean as the lipids were stable for 2 or more years before the study, but may reflect spontaneous changes in diet after the patients were labelled hypercholesterolaemic. Dietary advice can lower total cholesterol but the magnitude of this decrease is small. Additional approaches are likely to be required to reduce plasma cholesterol to a normal range.