Data from randomized clinical trials have become the foundation for clinical practice guidelines, including recommendation statements from the US Preventive Services Task Force (USPSTF) on statin therapy for primary prevention.1,2 As new trial data have become available, treatment recommendations have evolved to incorporate these new insights. Yet randomized clinical trials can neither represent the full spectrum of patients nor evaluate all questions and decisions faced in clinical practice. Guideline writers must therefore decide whether and how to generalize from incomplete evidence to care recommendations. Two hypothetical patient scenarios illustrate how far statin therapy guidelines have come and what important gaps remain.
Mr Young is a 41-year-old man with a history of hypertension (blood pressure of 128/70 mm Hg with treatment) but no history of diabetes or tobacco use. His total cholesterol level is 245 mg/dL; low-density lipoprotein cholesterol (LDL-C) level, 155 mg/dL; and high-density lipoprotein cholesterol level, 50 mg/dL. His estimated 10-year cardiovascular disease (CVD) risk is 2%,3 with an estimated lifetime CVD risk of 50%.4
Mr Smith is a 63-year-old man with a risk factor profile similar to that of Mr Young, except that his total cholesterol level is 160 mg/dL; LDL-C level, 80 mg/dL; and high-density lipoprotein cholesterol level, 50 mg/dL. His estimated 10-year CVD risk is 10%,3 and his estimated lifetime CVD risk is 50%.4
Based on these patients’ respective 10-year CVD risk estimates, the USPSTF recommendation statement would not suggest statin treatment for Mr Young despite his elevated LDL-C level but would recommend therapy for Mr Smith despite his “normal” LDL-C level. How did the USPSTF arrive at these apparently counterintuitive conclusions?
The first statin trials in primary prevention focused on adults with extremely high cholesterol levels.5 Subsequent trials, such as WOSCOPS6 and AFCAPS,7 confirmed the benefit of statins among those with moderately elevated levels of LDL-C. The JUPITER study8 shifted the treatment approach entirely by finding that statins were beneficial when used by adults with normal LDL-C levels but with high levels of C-reactive protein. HOPE-3 recently provided further evidence for a CVD risk–based treatment strategy by demonstrating the effectiveness of statins for older adults with at least 1 CVD risk factor, regardless of baseline LDL-C level.9 Combined, these trials suggested that statin recommendations could be based primarily on a patient’s underlying CVD risk rather than on his or her cholesterol level. Thus, the conclusion that Mr Smith (who has elevated 10-year CVD risk but a normal LDL-C level) would benefit from statin therapy appears supported by trial evidence.
Disseminating a treatment strategy based largely on CVD risk alone has been a difficult message for the clinical community to accept and implement. Nearly a generation of physicians has considered high cholesterol levels, rather than generalized CVD risk, the target for statin treatment. Only a minority of physicians consistently use complex, risk-based probabilistic calculations to determine therapy. Furthermore, the management approach for other CVD risk factors such as hypertension and diabetes has focused on reaching certain thresholds, regardless of CVD risk. Why should cholesterol be different? Several key questions deserve careful consideration.
First, should LDL-C be considered in treatment recommendations beyond CVD risk? The decision to use absolute risk to guide statin recommendations is based on the finding that the relative risk reduction seen with statin therapy is independent of baseline risk; thus, those with the highest absolute baseline CVD risk experience the greatest reduction in CVD events.10 However, the relative risk reduction of lipid-lowering therapy is also proportional to mmol/dL reduction in LDL-C level.11,12 This supports the contention that those with the highest baseline LDL-C levels should benefit the most from treatment because they have the most potential decline in LDL-C with intervention. One way to reconcile these findings is to incorporate both LDL-C levels and CVD risk into treatment recommendations, as has been done in the European guidelines.13 This approach recognizes that the relative benefit of statins is proportional to LDL-C lowering but that the absolute treatment benefit is largely driven by baseline risk.
Second, is 10 years the ideal horizon for all adults? The time horizon chosen to identify those at high risk for CVD has important implications for who is selected for therapy. Mr Young has a high cholesterol level and hypertension but is relatively protected in the short term from cardiac events by his younger age. Thus, statin therapy was not recommended by the USPSTF. The USPSTF decision to concentrate on a 10-year risk horizon for primary prevention was largely arbitrary. To date, no statin trial has enrolled patients based solely on 10-year risk. Additionally, for younger adults, a decade-wide time window ignores the pathobiology of atherosclerosis, a progressive lifetime process of cholesterol-rich plaque development in arteries that begins in the teenage years. Epidemiologic evidence from the Framingham Heart Study demonstrates that prolonged exposure to high LDL-C levels is associated with long-term risk of heart disease in a dose-responsive fashion, similar to the way that pack-years of smoking increase CVD risk.14
Waiting to treat Mr Young until he is old enough to reach a 7.5% or higher 10-year CVD risk threshold potentially exposes him to many years of elevated cholesterol levels. Long-term follow-up of the WOSCOPS trial provides compelling evidence that early treatment can delay the onset of CVD later in life among adults with elevated cholesterol levels.15 Thus, it is plausible that because prolonged elevation of LDL-C levels increases CVD risk, prolonged reductions in LDL-C with statin use could similarly lead to cumulative benefit. Considering long-term risk of CVD in younger adults offers an alternative to 10-year risk for identifying those who may benefit from early, aggressive therapy.
Third, what additional research is needed? Clinical trials have helped transform guidelines from recommending statins to treat a cholesterol number to recommendations that treat generalized CVD risk. But more evidence is still needed.
As noted by the USPSTF, persons older than 75 years have been understudied in statin trials. Only PROSPER specifically studied statin use among older adults.16 Although the overall trial demonstrated a modest benefit from statin therapy, the study combined both primary and secondary prevention populations. Age alone confers a substantial cardiovascular risk, but can statins reduce age-related CVD risk in the absence of other risk factors? There may be a point at which it is too late for statin therapy. Statins were not shown to benefit adults with high-risk disease states such as heart failure or end-stage renal disease with dialysis.17,18 Thus, it will be important to establish the benefits of statins in high-risk elderly persons before widespread adoption of this therapy.
In addition, among adults aged 40 to 75 years, additional clarification is needed to better identify who benefits the most from statins. Should this be based on individual CVD risk factors (HOPE-3), an inflammatory biomarker (JUPITER), a composite CVD risk score, a cholesterol threshold, or some combination of these? There also is a need to evaluate whether there is a risk “floor” for statin benefits. Clinical trials progressively lowered the bar for eligibility, yet even in these lower-risk populations, statins have continued to be shown to be effective. With large enough trials conducted for sufficient duration, would even low-risk individuals derive significant benefit from intervention? Dare cardiovascular experts consider a future in which there are near-universal statin recommendations for middle aged adults?
Also, and perhaps most important, among persons younger than 40 years, the long-term risks of statin therapy and potential cumulative benefits from long-term treatment need to be understood. Although long-term trials are expensive, at a population level, the potential public health benefits of early prevention are substantial. Even though the incidence of CVD is highest in older adults, because of their relative numbers in the population, nearly half of CVD events occur in adults younger than 65 years.19
Fourth, in the absence of clear answers, should the default strategy be reconsidered? When younger adults like Mr Young do not meet a specific guideline recommendation, should the default be to not offer treatment? While the absence of evidence does not equate to the evidence of absence, a failure to recommend treatment usually will result in a failure to treat. As an alternative, it may be reasonable to consider offering therapy to younger populations even before the benefits are fully confirmed. This consideration is based on an evaluation of the potential benefits, risk, and costs of statin treatment.
In almost every trial to date, statins have been found efficacious in reducing CVD risk.2 The USPSTF has concluded that the risks of statins—one of the most-studied therapeutic classes of medicine—in trials were minimal (if any).2 Furthermore, because statins have been used broadly in community practice for decades, pharmacoepidemiologic data support their long-term safety. From the resource perspective, the vast majority of statins are now available as generic products and require limited monitoring, leading to quite modest therapeutic costs.
For patients in the gray area not covered by the guidelines, clinicians should be cautioned against adopting either a “treat none” or a “treat all” strategy. Rather, gaps in the evidence provide opportunities for clinicians to practice the art of medicine and engage with patients in shared decision making regarding strategies for CVD prevention.
Corresponding Author: Eric D. Peterson, MD, Duke Clinical Research Institute, 2400 Pratt St, Durham, NC 27705 (firstname.lastname@example.org).
Conflict of Interest Disclosures: The authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Dr Navar reported receiving research support to her institution and research consulting fees from Regeneron and Sanofi. Dr Peterson reported receiving research grants and consulting fees from Merck, Sanofi, Regeneron, and AstraZeneca.
Navar AM, Peterson ED. Evolving Approaches for Statins in Primary PreventionProgress, but Questions Remain. JAMA. 2016;316(19):1981-1983. doi:10.1001/jama.2016.15094