Objective: To compare nine different algorithms for determining the necessity for drug treatment of hyperlipidaemia in patients without cardiovascular disease.
Design: Risk-factor data on 400 patients were obtained. Algorithms assessed included the Framingham model, National Cholesterol Education Program, European, British, New Zealand, British Regional Heart Study, Sheffield, Munster Heart Study and Dundee guidelines and a local general rule for treatment of atheroma (General Rule to Enable Atheroma Treatment).
Setting: Three hospital lipid clinics.
Results: Patients were 56% male, aged 53.8 +/- 12.3 years, 38% smoked, 55% had a family history of disease (37% familial hypercholesterolaemia, 33% hypertension and 5.2% left ventricular hypertrophy); 31% were moderately obese and 15% were diabetic. The average cholesterol level was 7.9 +/- 2.6 mmol/l. Median Framingham risk was 1.66%/year and this risk was > 2%/year for 37% and > 3%/year for 17.5%. Treatment was required for 86% by National Cholesterol Education Program, 70% by General Rule to Enable Atheroma Treatment and European, 61% by New Zealand, 58% by British, 42% by Dundee, 40% by Munster Heart Study, 25% by British Regional Heart Study and 16% by Sheffield guidelines. Algorithms varied regarding recommendations for treatment of men, women, patients with familial hypercholesterolaemia and diabetes. Confounders included risk levels, definitions and weightings for additional risk factors.
Conclusions: Different algorithms vary widely in their predictions of need for drug treatment of this high-risk population with hyperlipidaemia. Authors of guidelines for cardiovascular disease need to unify risk thresholds and definitions and audit the effects of additional weighting factors so that patients can be treated consistently.