Paul M Ridker, MD, MPH; Peter W. F. Wilson, MD
When the last set of US lipid-lowering guidelines was updated in 2004,1 placebo-controlled trials had proven statin therapy to be effective for the secondary prevention of cardiovascular disease. At that time, the ability of these agents to reduce rates of vascular events in primary prevention remained controversial, long-term safety with regard to cancer was unclear, and pharmacologic therapy was relatively expensive.
In 2004, faced with limited data, the guideline authors recommended that physicians follow a 2-step process that was based largely on an estimate of absolute risk for coronary heart disease. In step 1, using an epidemiologic global risk algorithm, physicians were advised to stratify patients into lower-, intermediate-, and higher-risk subgroups calculated over a 10-year time frame. Then, to balance cost, adverse effects, and efficacy, a second step suggested that physicians recommend lifestyle interventions for patients at lower risk and limit pharmacologic intervention with statins to those at higher risk. The use of statins for patients at intermediate risk was encouraged in many instances, even though data supporting pharmacotherapy for patients in this category were modest. Some experts in vascular disease prevention now advocate for the use of imaging tests and advanced biomarker testing to refine absolute risk estimates.2
In the absence of definitive trial data, the logic of this risk-based approach to statin therapy was understandable even if it had never been formally tested. Under the assumption that statins provide a similar relative risk reduction for all patients, the prevailing view that the most judicious way to use these agents in clinical practice was to limit their prescription to patients at greatest absolute risk for atherosclerotic disease in an effort to achieve the highest absolute risk reductions was reasonable.
However, in 2013, given the large number of randomized trials of statin therapy published during the past 2 decades (see Supplement), risk-based statin allocation may no longer be optimal for patient care, and alternative approaches for prevention of initial and recurrent vascular disease are being considered.3 Several factors support an approach that directly relies on clinical trial data for the formulation of statin guidelines.4
First, no global-risk algorithm or any vascular imaging modality has been used as an enrollment criterion for any major statin trial. Basing statin prescription on either of these metrics is difficult to defend in an evidence-based environment. Second, smoking and hypertension act synergistically with other factors to increase global risk, yet for these conditions the first intervention should be smoking cessation and blood pressure reduction, not necessarily lipid-lowering therapy. Third, risk prediction models are less accurate in ethnic minorities and women, groups for whom systematic undertreatment is a common problem. Fourth, in primary prevention, most vascular events occur among individuals with relatively modest 10-year risk estimates. As such, limiting pharmacologic intervention to those with highest absolute risk misses large opportunities for prevention. Fifth, and perhaps most important, randomized trials completed since 2004 directly contradict the simple notion that statin therapy delivers a constant relative risk reduction across all patient groups. This is a major issue because the assumption of similar efficacy for all treated patients is the underlying premise for arguments to base prescription on absolute risk.
For instance, the CORONA, AURORA, 4-D, and GISSI-HF trials included a total of 13 613 patients with either congestive heart failure or end-stage renal failure and were reported between 2005 and 2009. All 4 trials enrolled patients with very high absolute risk who would be anticipated to have substantial coronary artery calcification. Yet despite achieving large low-density lipoprotein reductions with statin therapy, the patients in these 4 trials obtained no significant clinical benefit in terms of event reduction. By contrast, the SHARP trial published in 2011 demonstrated that simvastatin plus ezetimibe, compared with placebo, reduced the risk of vascular events among patients with chronic renal insufficiency who were not undergoing dialysis. Thus, in 2013, trial data demonstrate the difficulty of recommending statins to all high-risk patients without regard for underlying clinical conditions.
In addition, the WOSCOPS, AFCAPS/TexCAPS, ASCOT-LLA, CARDS, MEGA, and JUPITER trials included 55 097 primary prevention patients and were reported between 1995 and 2008. These 6 trials enrolled participants with much lower absolute risk. In contrast to 4 of the 5 high-population risk trials noted previously, all 6 of these low-population risk trials demonstrated benefits with statin therapy. As confirmed in a recent meta-analysis,5 there appears to be a statistically significant trend for greater relative risk reductions with statin therapy for individuals at lower levels of absolute risk than for those at higher levels of absolute risk. This trial-based observation is consistent with the biological concept that early use of statins can slow plaque growth and potentially provide a more beneficial clinical strategy compared with initiation of statins later in the disease process (when significant coronary calcification is present). Thus, in 2013, trial data contradict the recommendation that lower-risk populations should be recommended lifestyle interventions only.
Given this accumulating evidence, instead of extrapolating from observational estimates of absolute risk, another approach may be to prescribe statins to patient populations for whom clinical trials have demonstrated benefit (Table). An evidence-based interpretation of available trial data would be to endorse statin use for secondary prevention (4S, CARE, LIPID, HPS, PROSPER) and for patients with chronic renal insufficiency (SHARP), but to limit statin use among those undergoing hemodialysis (SHARP, 4-D, AURORA) or with isolated heart failure (CORONA, GISSI-HF). Further, at least in trials of secondary prevention, higher-intensity statin regimens have been shown to deliver greater relative risk reductions than lower-intensity statin regimens (PROVE-IT, A to Z, IDEAL, TNT, SEARCH).
Similarly, a trial-based recommendation in primary prevention would be to consider statin use as an adjunct to lifestyle interventions among middle-aged and older individuals with elevated low-density lipoprotein levels (WOSCOPS, MEGA), low high-density lipoprotein levels (AFCAPS/TexCAPS), elevated C-reactive protein level (JUPITER), or multiple risk factors inclusive of diabetes (CARDS) and hypertension (ASCOT-LLA). In this setting, higher-intensity statins also may provide greater relative risk reductions, but this benefit should be balanced against modestly higher relative risks for developing diabetes.6- 7
A trial-based approach could provide simplicity for the clinician and would set a clear precedent for how guidelines will evolve as more information accrues. If appropriately designed clinical trials based on imaging tests are completed and demonstrate efficacy, then trial-based guidelines can easily be updated to incorporate such advances. Similarly, if novel lipid-lowering agents currently under investigation demonstrate event reduction in addition to statin therapy (or as monotherapy for those intolerant of statins), then such trial data can be rapidly incorporated into recommendations.
For individual patients, physicians still need to consider issues such as number-needed-to-treat and remember that any use of statin therapy should be an adjunct to lifestyle interventions such as diet, exercise, and smoking cessation, not a replacement. Extrapolation of findings from “efficacy” populations in clinical trials to “effectiveness” populations must be made with care as expected benefits and risks, particularly adverse effects, may differ between patients in clinical trials and individual patients in real-world practice.
In summary, in 2013, the continued use of absolute risk assessment to guide statin prescription will most likely overtreat many who do not benefit and undertreat many for whom efficacy is certain. Data from statin trials are now robust enough to allow guideline authors to directly answer the critical questions of what works and in whom without estimating cardiovascular risk or using radiologic imaging tests.4
The cardiovascular prevention community has a low-cost treatment proven in randomized clinical trials to reduce rates of myocardial infarction, stroke, and cardiovascular death in multiple population groups. Moving forward, trial-based guidelines would provide an evidence-based approach for the use of statins in the primary and secondary prevention of cardiovascular disease.
