News that there may be errors in the algorithms used to calculate statin prescription will leave cardiology clinics and surgeries fighting through the mist of the latest statin controversy. In 25 years of medical practice, the class of drugs known as statins seems consistently to have provoked more controversy than any other. Why?
Epidemiological studies since the 1970s have shown clearly that the higher the blood cholesterol, the higher the risks of heart attack. However, even a series of such observations is not sufficient to show a causative relation between blood cholesterol and cardiovascular risk; still less that therapeutically lowering cholesterol would translate into any outcome benefit. That would require evidence from a randomised placebo-controlled clinical trial, and ideally more than one.
So how does the biomedical community (scientists, clinicians and the pharmaceutical industry) approach the challenge of showing that a treatment works? To test the effects of a drug requires a defined population and an outcome that can be quantified in a practical time frame to allow statistical comparisons between those receiving placebo and those receiving the active drug. As in the case of statins, it is logical, therefore, to design the first trials to treat patients who are perceived to be at high risk of the measurable event; not least because, from a statistical point of view, the size (and therefore cost) of the trial will depend on (a) the frequency of the measured event and (b) the magnitude of the treatment effect — so focusing on high-risk patients makes sense.
I remember clearly hearing the results of the landmark Scandinavian Simvastatin Survival Study as a trainee cardiologist in London in 1994. This clinical trial studied 4,444 patients with both prior heart attack and high cholesterol (even after lifestyle advice and modification). The results were startling: over the five and a half years median follow-up period, simvastatin lowered total cholesterol by 25 per cent. In the placebo group 256 patients (12 per cent) died compared with 182 (8 per cent) in the simvastatin group. So, the risk of death was reduced by about a third. In other words, for every 100 patients treated for about five years after a heart attack, four people who would otherwise have died did not. The probability that this was a ‘chance’ result (ie wrong) was less than one in 1,000.
Death is obviously the most robust quantifiable end point of all — but there were commensurate reductions in heart attack risk and the need for heart operations. Whether one thinks that these results are numerically impressive or not is to some extent subjective. In most areas of medicine, a mortality reduction of around one third with a daily tablet would be considered unusually effective.
Over the next 20 or so years a series of trials sequentially extended the demonstration of statin efficacy to populations that were at lower baseline risk and had lower starting levels of cholesterol. Overall, the results were clear and consistent: lowering LDL-cholesterol with statins reduces cardiovascular risk, but the magnitude of the reduction depends on the patient’s risk — the higher the risk, the more there is to gain, which is not surprising.
Accordingly, pretty much anyone who has had a heart attack or who has coronary artery disease will find themselves in a category of patients for which statin efficacy has been shown in trials. So how come the controversy? As more trials emerged, the indications for statins increased. The groups considered eligible for prescriptions of statins expanded and the seemingly ever-changing goal posts may have raised suspicions about the motivation of both the pharmaceutical industry and of the prescribing physicians. Indeed, the medical profession has not been good at communicating with the public about how clinical trials work, the inevitably changing landscape of clinical evidence, the nature and magnitude of the projected benefits (or lack of them) and, crucially, where the uncertainties lie.
So what about the uncertainties? For the most part these start where the margins of the clear evidence base start to fray and partially informed (albeit well-intentioned) speculation begins. In part, this is where risk calculators (like QRISK2) and guidelines come in — to plug the gap where trial data are lacking. Current algorithms such as QRISK2 combine a number of variables including age, cholesterol, blood pressure, smoking history, family history etc and compute a 10-year risk of cardiovascular events. Equipped with this type of individualised assessment of risk, it is possible for bodies such as NICE to make recommendations for treatment with carefully considered but ultimately arbitrary thresholds.
An important topic of debate in this area is the question of ‘lifetime risk’. We know that atherosclerosis, the process of artery ‘furring-up’ that underlies heart attack and many strokes, starts in early life (certainly by the 20s) and progresses over many years. We know this from postmortem examination of the hearts of people who died prematurely eg in road accidents. So, extrapolating from the trial evidence, one line of thought is to make interventions that delay progression of atherosclerosis from a relatively early stage in order to ameliorate lifetime risk of cardiovascular disease. However, since clinical trial data require measurable events in a manageable time frame, a long-term benefit of this approach will be difficult to show.
So what to do? Given a high lifetime risk of coronary disease and the early onset of the process, one could take a view that an individual, even in early life, should be able to prospectively survey their future risk and make a decision on their approach to interventions that may ameliorate that risk before it reaches a high level. This would include lifestyle modification but could also include preventive treatment, eg with statins. Clearly healthy, symptom-free patients taking drugs over long periods requires very careful consideration of the potential risks and benefits.
My larger points are that: (1) a given individual, properly equipped with knowledge of the risks and benefits of a treatment, and in consultation with their physician, should be able to make informed choices (and to revisit and revise them as the evidence accumulates) and (2) a more robust understanding of both the processes of medical decision-making and the extent and nature of the body of evidence, including its gaps and deficiencies, will leave patients less susceptible to the shock waves of new or unexpected findings and less exposed to the manipulative sensationalism that characterises some lay reporting, particularly, it seems, when it come to statins.
Dr Robin Choudhury is professor of cardiovascular medicine at the University of Oxford, consultant cardiologist at the John Radcliffe hospital and clinical director of the Oxford Acute Vascular Imaging Centre.