Effectiveness of lipid lowering drugs in general practice

Editor—The article by Hippisley-Cox et al illustrates a major problem of describing a recommended cholesterol concentration as a target value—scatter around a bullseye will always ensure at least 50% of values above the target.¹ What was most interesting about their data was the dispersion of last recorded serum cholesterol concentrations about the means. This was small for simvastatin and atorvastatin, indicating that cholesterol values were close to recommended values even for those > 5 mmol/l, and as Marshall has implied,² many of these patients may have concentrations ≤5 mmol/l on remeasurement.

Although only one trial has compared five statins in a single study,³ several paired comparisons of the efficacy of the statins⁴ and of statins versus fibrates⁵ have been undertaken. The data of Hippisley-Cox et al are consistent with these.

However, the statement “Statins reduce lipid levels better than fibrates” is at best misleading. Fibrates are often used in diabetic patients and other patients with an atherogenic lipoprotein phenotype (raised triglyceride, low high density lipoprotein, and mildly raised low density lipoprotein cholesterol concentrations), in whom cholesterol lowering is not the only consideration. Fibrates are also often used in combination treatment or in patients intolerant or poorly responsive to statins. Hence Marshall may be wrong to dismiss selection bias as a confounding problem. It was unclear whether patients receiving combined treatment were excluded from the analysis, and if included, which starting cholesterol concentrations were chosen. The lack of dosing data also makes it difficult to assess the validity of the statement that a target value of ≤5 mmol/l is unrealistic.