Author + information
- Vera Bittner, MD, MSPH∗ ()
- Division of Cardiovascular Disease, Department of Medicine, University of Alabama at Birmingham, Birmingham, Alabama
- ↵∗Reprint requests and correspondence:
Dr. Vera Bittner, University of Alabama Birmingham, 701 19th Street South, LHRB 310, Birmingham, Alabama 35294.
Statins are effective in the primary and secondary prevention of coronary heart disease and stroke. Although many would agree that primary prevention of atherosclerotic cardiovascular disease (ASCVD) is preferable to secondary prevention, no consensus exists on how to best identify individuals at risk for the disease, when to commence screening and risk assessment, at what age to start (or stop) treatment, how to treat, and how intensively to treat. The 2013 American College of Cardiology/American Heart Association (ACC/AHA) cholesterol guidelines concluded that net benefit was sufficient to recommend statin treatment for patients with primary low-density lipoprotein (LDL) cholesterol elevations ≥190 mg/dl and for those age 40 to 75 years with an LDL cholesterol level between 70 and 189 mg/dl and ≥7.5% 10-year ASCVD risk derived from the pooled cohort risk equation (1). The guideline authors further emphasized that risk stratification should be followed by a clinician–patient discussion and shared decision-making. Although several investigator groups have reported good calibration and case discrimination with the pooled cohort risk equation, others found overestimation of ASCVD risk.
Concerns regarding potential overtreatment thus persist, especially among older subjects whose primary determinant of risk in the equation is age. Alternate approaches to identifying subjects likely to benefit from primary prevention with statins have thus been proposed. Ridker and Wilson (2) described an approach using entry criteria from primary prevention statin trials and, given null results, limiting statin therapy in subjects with heart failure or undergoing hemodialysis. More recently, Ridker et al. (3) proposed a hybrid approach using both trial entry criteria and global risk prediction. How does the practicing clinician choose among these approaches?
In this issue of the Journal, Mortensen et al. (4) describe a head-to-head comparison of the 2013 ACC/AHA guideline approach, a trial-based approach, and a hybrid approach. They applied these approaches to the Copenhagen General Population Study, a contemporary population-based cohort of Danish men and women. After excluding patients with diabetes, ASCVD, statin use, or missing information at baseline, as well as those patients who were outside the age range of 40 to 75 years (thus including only the age range recommended in the ACC/AHA guidelines), the authors followed up 37,892 patients prospectively for 5 years for the occurrence of “hard” ASVCD events. Statin eligibility varied greatly among the 3 approaches: 42% with the 2013 ACC/AHA approach, 56% with the trial-based approach, and 21% with the hybrid approach. Importantly, these approaches identified different individuals in the population, a phenomenon more apparent in women than in men.
Using methods identical to those of Muntner et al. (5), Mortensen et al. (4) found that the 2013 ACC/AHA algorithm was well calibrated in the 7.5% risk range in this population. The predicted event rate (per 1,000 patient-years) was highest with the hybrid approach (11.2), intermediate with the 2013 ACC/AHA algorithm (9.8), and lowest with the trial-based approach (6.8). Discrimination was moderate with all approaches, highest with the 2013 ACC/AHA approach, intermediate with the hybrid approach, and lowest with the trial-based approach (area under the receiver-operating characteristic curves of 0.676, 0.613, and 0.572, respectively). Compared with the 2013 ACC/AHA approach, the trial-based and hybrid approaches had negative net reclassification indexes. The authors thus concluded that the guidelines-recommended approach will prevent more ASCVD events and, compared with the trial-based approach, will achieve this goal by treating fewer people.
Do these results resolve the dilemma for the practicing clinician on whom to treat and how to treat? Among the 3 options, the 2013 ACC/AHA guidelines approach seemed to be best for identifying subjects destined to have a future event (4). However, is this approach necessarily best at identifying subjects at risk who will benefit from treatment? In clinical trials of lipid-lowering therapy, relative risk reductions are generally in the 30% range (6). In a trial-based strategy, there is an inherent assumption that subjects who meet the same entry characteristics as subjects enrolled in the trials would experience this same relative risk reduction, assuming similar adherence to therapy. However, the trial-based strategy does not try to match exclusion criteria, thus identifying a substantially larger group of subjects than would have been eligible to participate in the trials. In addition, each decision about trial participation is based not only on inclusion and exclusion criteria but also on unmeasured patient and investigator characteristics that would be impossible to match; these characteristics, however, could exert a major impact on ultimate treatment benefit.
For a risk-based approach, we have no prospective data on expected benefit from treatment and have to extrapolate that the benefit would be similar to that observed in the clinical trials. Some argue that risk driven by age and nonlipid risk factors would be less likely to benefit from lipid-lowering therapy. Others might argue that lipid-lowering therapy will be effective: after all, subjects with lifelong lower LDL cholesterol levels through proprotein convertase subtilisin/kexin type 9 (PCSK9) mutations seem to be protected from ASCVD even in the presence of major nonlipid risk factors (7). The answer could only come from a randomized trial, but such a trial would be complex and expensive and is unlikely to be funded.
All 3 approaches suffer from the limited data we have on nonwhite subjects. The 2013 ACC/AHA calculator provides estimates for white and black subjects but does not have specific estimates for other ethnicities. Trial and hybrid approaches are inherently hampered by the types of participants in the trials and, potentially, by the limited applicability of trials conducted in specific populations in other geographic regions, risk environments, or ethnicities.
Last, but not least: what about the 8% of subjects not identified by any of the approaches but who nevertheless experienced events? What did we miss? As one would expect, these subjects had a global risk below the threshold and did not meet typical trial entry criteria. However, one intriguing observation emerges from Online Tables 2 to 4 accompanying Mortensen et al. (4): such subjects had substantially higher lipoprotein(a) (Lp[a]) levels. Should we adopt Lp(a) more widely as a marker to identify subjects at risk and use Lp(a) levels as a treatment criterion? Or should we go further and make Lp(a) a treatment target (8)? Clearly, additional research is needed in this area, and ongoing trials with PCSK9 inhibitors will provide important insights in the near future.
So what do I tell my primary prevention patient in the clinic tomorrow? Quantification of risk is just a first step. For now, the 2013 ACC/AHA risk calculator is our best tool to accomplish this goal and, in the absence of contraindications or patient wishes, we should treat those who fall into the primary prevention statin benefit groups. However, it is worth remembering that, in the absence of end-stage renal disease and heart failure, we have no data confirming that subjects who fall outside the age and lipid thresholds that define the statin benefit groups would not benefit from treatment. A thorough evaluation of each subject’s circumstances, a detailed clinician–patient discussion, and shared decision making are critical. Let us remember, too, that an exclusive “lipocentric approach” to ASCVD risk reduction begets large residual risk, as documented by the results of our lipid-lowering clinical trials (6).
Atherosclerosis is a multifactorial disease and requires a multifactorial approach with smoking cessation, dietary modification and weight management, regular physical activity, attention to psychosocial risk factors, and pharmacological therapy of lipid and nonlipid risk factors. Comprehensive risk factor control is associated with improved prognosis, and our challenge is to develop care models that will allow us to achieve such control.
↵∗ Editorials published in the Journal of the American College of Cardiology reflect the views of the authors and do not necessarily represent the views of JACC or the American College of Cardiology.
Dr. Bittner has received research support from Amgen, AstraZeneca, Bayer HealthCare, Janssen, Pfizer, and Sanofi; and has served on advisory panels for Amgen and Eli Lilly.
- 2015 American College of Cardiology Foundation
- Stone N.J.,
- Robinson J.G.,
- et al.
- Ridker P.M.,
- Rose L.,
- Cook N.R.
- Mortensen M.B.,
- Afzal S.,
- Nordestgaard B.G.,
- Falk E.