Should we reduce blood cholesterol to prevent cardiovascular disease among patients with chronic renal failure?

There is general agreement that, among individuals without overt renal impairment, blood cholesterol is an important cause of coronary artery disease (CAD). Prospective observational studies among such individuals have shown that the relationship between CAD risk (plotted on a doubling scale) and blood total cholesterol is roughly linear, and that within the range studied there is no threshold below which a lower cholesterol level is not associated with a lower risk of CAD. Associations between CAD risk and blood cholesterol appear to be similarly ‘log‐linear’ in high‐risk populations, including those with established cardiovascular disease [1]. Large‐scale randomized trials are required to assess how quickly an excess risk of CAD is reversed by treatments that lower cholesterol. Several trials of the effects of statin therapy on CAD morbidity and mortality among individuals with a prior history of CAD [2], and among those at low risk of CAD, have now been reported. These have shown that reductions of 1–2 mmol/l in blood total cholesterol (and LDL‐cholesterol) produce reductions of about 25–40% in the risk of a CAD event. Since patients with chronic renal failure (CRF) are, collectively, at high risk of cardiovascular disease [3], should they all receive routine statin therapy? The absolute reduction in the risk of CAD can be calculated as the product of two variables: (a) the absolute risk of a CAD event and (b) the proportional reduction in that risk associated with treatment (which appears to be linearly related to the absolute reduction in blood cholesterol). If it is assumed, for the sake of the argument, that these data can be applied to individuals with renal impairment, then what are the expected benefits in such patients? In the general population, cardiac disease is usually caused by atheromatous CAD. However, among patients with CRF the most prevalent cardiac abnormality—found in about three‐quarters of patients commencing dialysis [4]—is left ventricular hypertrophy. The true prevalence of atheromatous CAD in CRF is difficult to determine reliably, since myocardial ischaemia may occur in the absence of significant coronary atheroma [5], but symptoms or signs of CAD are present in about 25% of dialysis patients [6]. In the US, approximately one‐quarter of cardiac deaths among individuals on dialysis programmes are attributed to myocardial infarction (MI), whilst the remaining three‐quarters are categorized as sudden or arrhythmic, or of some other type [6]. Routine cause of death data may be subject to substantial misclassification errors [7], however, and some sudden or arrhythmic deaths may well be due to undocumented MI. Nevertheless, it does appear that structural myocardial abnormalities account for a higher proportion of cardiac mortality in CRF than in the general population. There are at least three reasons for uncertainty about the overall effects of lowering cholesterol on the risk of a cardiac event among patients with chronic renal failure. (i) The overall effects of lowering cholesterol on the risk of a cardiac event will be determined by its effects on non‐fatal events (i.e. non‐fatal MI) and on fatal events (i.e. fatal MI, arrhythmic or sudden death, fatal heart failure, and other cardiac causes of death). All other things being equal, for a given statin drug dose and a particular LDL‐cholesterol concentration, it should be possible to estimate the expected proportional reduction in the risk of MI among particular patients with chronic renal failure. Among those patients at high absolute risk of atheromatous CAD, such reductions may be clinically worthwhile. However, the effects of a statin on those cardiac deaths that are not definitely attributable to MI is much less certain, and some patients with CRF may be at high absolute risk of arrhythmic cardiac death but relatively low absolute risk of MI. (ii) Whilst hypercholesterolaemia may develop in the nephrotic syndrome and following renal transplantation, many patients with CRF have average (or even below average) plasma LDL‐cholesterol levels [8]. This is relevant because, despite the log‐linear associations between CAD risk and blood cholesterol among individuals without renal failure, the absolute reductions in blood cholesterol (and hence the proportional reductions in CAD events) produced by a fixed statin dose are smaller at lower levels of cholesterol. Hence there is uncertainty as to whether it is worthwhile lowering LDL‐cholesterol below about 3.5 mmol/l [2]. (iii) An additional reason for uncertainty, when assessing whether reducing blood cholesterol might help to avoid cardiac disease among patients with renal failure, is the absence of any reliable direct evidence on long‐term associations between blood cholesterol and the risk of cardiac disease in such patients. Several short‐term registry‐based observational studies have demonstrated an inverse association between total cholesterol and all‐cause mortality [9], or cardiovascular mortality [10]. These results may be unreliable, however, because it was not possible to correct for ‘confounding by disease’. Such confounding occurs because a persisting acute‐phase response, often associated with protein‐calorie malnutrition, results in a compensatory reduction in hepatic cholesterol synthesis among the sickest individuals [11]. Thus those individuals with the most severe renal or comorbid disease will tend both to have lower cholesterol levels and, independently, to be at highest risk of death. In summary therefore, uncertainties remain about the effects of cholesterol‐lowering therapy on the overall risk of cardiac disease among particular patients with CRF. As a result, several randomized trials are currently assessing the effects of cholesterol‐lowering in such patients. The Assessment of Lescol in Renal Transplantation (ALERT) study is comparing fluvastatin 40 mg daily versus placebo among over 2000 stable transplant recipients, and the Die Deutsche Diabetes Dialyse (4D) study plans to randomize 1200 type 2 diabetic patients on dialysis between atorvastatin 20 mg daily and matching placebo [12]. The UK‐Heart and Renal Protection (UK‐HARP) pilot study is assessing the biochemical efficacy and safety of simvastatin 20 mg daily versus placebo (as well as aspirin 100 mg daily versus placebo) in 600 patients who either have elevated serum creatinine (pre‐dialysis), or are already on dialysis, or have a functioning transplant. The information from this study will then be used to prepare for a larger study of the effects of treatment on clinical events. These trials will provide information about the effects of cholesterol‐lowering on cardiac disease among particular patients with CRF, and about the risks of hepatotoxicity or myopathy, as well as helping to assess, among patients with progressive renal failure, whether lowering cholesterol can delay loss of renal function [13]. Although at present cholesterol‐lowering treatment may well be appropriate for patients with CRF who are reliably known to be at high absolute risk of atheromatous CAD, for other patients the results of the ongoing randomized trials will be needed to resolve the current uncertainties about the balance of benefits and risks of cholesterol‐lowering therapy. Correspondence and offprint requests to: Dr Colin Baigent, Clinical Trial Service Unit and Epidemiological Studies Unit, Harkness Building, Radcliffe Infirmary, Woodstock Road, Oxford OX2 6HE, UK.