In a genetically heterogeneous group of 109 patients with a clinical diagnosis of heterozygous familial hypercholesterolaemia (FH), the influence of gender, apolipoprotein (apo) E genotype and the type of molecular defect in the LDL-receptor (LDLR) gene on the reduction of plasma LDL-cholesterol levels to treatment with a HMG-CoA reductase inhibitor (simvastatin) were studied. Response was determined as the percentage fall in LDL-cholesterol from untreated levels and as the proportion of patients where levels fell below 4.9 or 4.1 mmol/l. Of the patients, 86 individuals had tendon xanthomata (TX+) and a diagnosis of 'definite' FH and these individuals presented with a significantly higher untreated LDL-cholesterol compared to the 23 individuals who did not have xanthomas (TX-) and a diagnosis of 'probable' FH (8.14+/-0.19 vs. 6.81+/-0.25, P= 0.001). Overall, HMG-CoA reductase inhibitor doses of 10, 20 or 40 mg/day resulted in a significant fall of LDL-cholesterol levels of 29, 39 and 49%, but at all doses those with TX had significantly higher levels than those without, and significantly fewer TX + patients achieved LDL-cholesterol levels below 4.9 or 4.1 mmol/l than the TX - group (P < 0.05 at each dose). In the TX+ group the response to treatment was of similar magnitude in men and women and in patients with different apoE genotype. In the 'probable' FH probands only three mutations were identified (detection rate 13%), one in the LDLR gene and two in the APOB gene, a detection rate significantly lower (P= 0.02) than in the 'definite' FH probands where 28 mutations were detected (detection rate 37%). In the TX + patients where no mutation was detected, treatment resulted in a greater proportion achieving LDL-cholesterol levels below 4.9 and 4.1 mmol/l compared to those with any LDLR mutation, this difference was close to statistical significance at the 4.9 mmol/l threshold at 10 mg/day (41 vs. 13%, P = 0.058). For the 14 patients with an LDLR mutation that was predicted to be 'severe', fewer achieved LDL-cholesterol levels below 4.9 or 4.1 mmol/l at each dosage compared to the 16 individuals with 'mild' mutations, and this difference was statistically significant at the maximal dosage of 40 mg/day (P = 0.018). Thus although characterisation of the molecular defect in FH patients may not be relevant to their immediate clinical management, those with a particular mutation may need more aggressive lipid-lowering treatment to reach LDL-cholesterol levels recommended to reduce the risk of coronary heart disease (CHD).