Tough Questions About the Value of Statin Therapy for Primary Prevention: Did JUPITER Miss the Moon?

An unusually rancorous debate that was recently described by the president of the American Heart Association as “a statistical tug-of-war” has raised more questions than answers about the risks and benefits of statins in primary prevention. Spending money to save money seemed like a reasonable strategy in lowering serum lipids in primary prevention of adverse cardiovascular outcomes, particularly when pravastatin and simvastatin became available by generic name in April and June 2006, respectively, soon thereafter permitting the treatment of several patients for the same drug cost as treating 1 patient with either brand pravastatin (Pravachol) or brand simvastatin (Zocor). Nevertheless, it has long been recognized that the small effect sizes associated with avoidance of adverse cardiovascular outcomes make the use of statins expensive even in secondary prevention, up to $1.1 million in drug cost in 2004 dollars to prevent 1 nonfatal stroke. In an assessment of the cost-effectiveness of primary prevention, Pletcher et al. (2009) used Markov modeling to estimate the economic and clinical effects of bringing all adults in the United States into compliance with Adult Treatment Panel III (ATP III) guidelines, finding that 11.1 million adults without coronary heart disease (CHD) would undergo newly initiated (9.7 million) or intensified (1.4 million) statin treatment. The net cost, after accounting for medical cost offsets due to avoided cardiovascular events (20,000 myocardial infarctions [MIs] and 10,000 cardiovascular deaths annually), would be $3.6 billion per year, or $42,000 per quality-adjusted life year over a 30-year time horizon. Although the current (July 2010) market price of generic simvastatin is approximately 60% lower than the $2.11 per tablet assumed in the model’s base-case analysis, and discounts off consumer cash price are commonly taken by health plans, Pletcher et al. note that treating all patients with low-density lipoprotein cholesterol (LDL-C) exceeding 130 milligrams per deciliter (mg per dL) would yield net cost savings only if statins cost less than $0.10 per pill. The investment strategy for the use of statins in primary prevention dimmed further with publication of a meta-analysis by Ray et al. in June 2010. The combined results of 11 randomized controlled trials (RCTs) involving 65,229 persons with intermediate to high risk of a cardiovascular outcome but without cardiovascular disease at baseline showed that statins were not associated with reduction in the risk of all-cause mortality over 244,000 person-years of follow-up. There were 1,447 all-cause deaths among 32,606 patients who received placebo (4.4%) versus 1,346 deaths among 32,623 patients who received statins (4.1%, risk ratio = 0.91, 95% confidence interval [CI] = 0.83-1.01). Across the 11 studies, the mortality rate for placebo ranged from 3.6 to 26.0 per 1,000 person-years versus a range from 2.4 to 27.2 per 1,000 person-years for statins, and participant age at baseline accounted for an estimated 66% of the variation in mortality rates. There was lack of significant effect on mortality despite evidence of LDL-C reduction; during a mean of 3.7 years follow-up, the mean LDL-C levels for placebo-treated and statin-treated patients were 134 mg per dL and 94 mg per dL, respectively. The results of this metaanalysis were bolstered by the absence of evidence of statistical heterogeneity among the 11 RCTs despite heterogeneity in the demographic and clinical characteristics of the study samples. The research by Ray et al. is also compelling because it is the first meta-analysis to exclude entirely the effect of statins in patients with known CHD. A previous meta-analysis of 10 RCTs in primary prevention with statins performed by Brugts et al. (2009) had found modest effects on all-cause mortality over an average 4.1 years of follow-up (rates of 5.1% and 5.7% for statinand placebo-treated patients, respectively, odds ratio [OR] = 0.88, 95% CI = 0.81-0.96), but that meta-analysis included 4,445 participants (6.3%) with a prior history of cardiovascular disease.9 With the circumspect work of Ray et al., we now have additional confidence in examining the value for money in primary prevention of cardiovascular events with statins. If the 0.3% absolute difference in the mortality rate for statin therapy (4.1%, weighted mean of 10.7 per 1,000 person-years) versus placebo (4.4%, weighted mean of 11.4 per 1,000 person years) was statistically significant, which it was not, the effect of statin treatment was an estimated 7 fewer deaths per 10,000 person-years of treatment. At current (July 2010) realworld discounted drug prices, preventing 1 all-cause death in primary prevention would require about $103,000 of generic simvastatin ($72 per year times 10,000 patient-years to prevent 7 deaths), $137,000 of generic pravastatin, or $2 million of rosuvastatin (Crestor). Or, one can use the results from Brugts et al. to calculate the direct rosuvastatin drug cost to prevent 1 all-cause death, approximately $975,000 (number needed to EDITORIAL