The JUPITER trial and statins in primary prevention: Unequivocal benefit, or still a debatable issue?
Click Here to Manage Email Alerts
The clinical utility of statins for mortality reduction in primary prevention has again become a “hot-button issue” in light of four recent articles from the Archives of Internal Medicine questioning their presumed benefit. Several of the articles conclude that lipid-lowering with statin therapy in the primary prevention setting is not advisable since meta-analyses did not show a reduction in total mortality. The authors neglect the major decreases in MI, stroke, revascularization and HF. Their logic, however, appears fractured.
In large part, these articles take aim at the Justification for the Use of Statins in Prevention: an Intervention Trial Evaluating Rosuvastatin (JUPITER) trial, which demonstrated that primary prevention patients with “normal” LDL levels and increased subclinical inflammation gain a mortality benefit from taking rosuvastatin (Crestor, AstraZeneca). Even before JUPITER, a 2008 meta-analysis of prior trials demonstrated that “statins have a clear role in prevention of CVD mortality.”
Currently, some groups argue for even earlier and more aggressive treatment with statins to arrest atherosclerosis, but there are physicians who continue to categorically question their benefit. How can we explain such divergent thought?
The answer lies largely within the time frame in question — do we use statins to treat a lifelong disease (atherosclerosis), or do we simply use statins to improve near-term survival? The answer also has profound implications, not only in terms of clinical guidelines, but also in terms of the cost-effectiveness of both statins and primary prevention screening programs (eg, with high sensitivity CRP [hsCRP] or coronary artery calcium).
Looking at the data
Ray and colleagues published a 2010 meta-analysis of 11 trials and 65,229 patients apparently showing that statins are of no significant benefit in the high-risk primary prevention population because there was no clear decrease in total short-term mortality. Of the 11 trials included, two studied diabetic patients exclusively. Statistical analysis of mortality rate revealed a borderline significant RR of 0.92 (95% CI, 0.84-1.02), even after exclusion of these two studies. The study by Ray and colleagues appears to contradict three other recent large meta-analyses highlighting a significant mortality benefit and reduction in CV events with statins.
For example, an analysis published by Mills and colleagues included 20 randomized clinical trials and found an RR of all-cause mortality of 0.93 (95% CI, 0.87-0.99), besides a significant reduction in CV deaths (RR=0.89; 95% CI, 0.81-0.98) and MI (RR=0.77; 95% CI, 0.63-0.95). Brugts and colleagues studied 10 randomized controlled trials and noted that statins significantly reduced the risk of all-cause mortality (OR=0.88; 95% CI, 0.81-0.96), major coronary events (OR=0.70; 95% CI, 0.61-0.81) and major cerebrovascular events (OR=0.81; 95% CI, 0.71-0.93). It is unclear whether the disparity between the latter two meta-analyses and the report by Ray and colleagues may have been caused by differences in trial selection or statistical analysis.
It is difficult to defend such a generalized statement regarding the presumed lack of benefit of statins in primary prevention when there is evidence for clear benefit in certain subgroups under the primary prevention umbrella. Women, for example, have been definitively shown to have a reduction in CV events with statin use. One superb meta-analysis evaluating six primary prevention trials for statin benefit in women determined that the RR for any CHD event was 0.78 (95% CI, 0.64-0.96).
A secondary analysis of JUPITER concluded that the 5,695 elderly patients (>70 years of age) accrued a significant reduction in CV events during the study period. Moreover, there was a significant 20% reduction in mortality in the JUPITER trial. Overall, the present weight of evidence appears to refute the notion that statins are of no proven benefit in primary prevention.
A primary prevention benefit in women?
The recent comments by Rita Redberg, MD, editor of the Archives of Internal Medicine, on The New England Journal of Medicine’s Cardio-Exchange website about the lack of evidence for statin primary prevention benefit in women is contradicted by sub-analyses of prior trials, including JUPITER. Her argument that “an asymptomatic, healthy patient” with high LDL requires therapeutic lifestyle changes alone, without statin therapy, must be balanced by the knowledge that about one-third of patients presenting with an MI are completely asymptomatic before their event. It is important to reflect on the fact that lifestyle improvements alone often do not reduce LDL, HDL, glucose or BP to guideline-mandated levels, necessitating appropriate pharmacological therapy for optimal risk factor reduction.
There is an enormous body of evidence supporting the lipid hypothesis of atherosclerosis. High LDL in the primary prevention setting should be regarded as a snapshot of a decade’s worth of atherosclerotic progression from a fatty streak to an unstable lipid-rich plaque, often compounded by significant subclinical inflammation. Even before the advent of JUPITER, a large 2005 meta-analysis published in the Lancet involving 90,056 people performed by the Cholesterol Treatment Trialists’ Collaborators showed irrefutable evidence that lowering LDL with statins decreases the incidence of major CV events, both in primary and secondary prevention. More stringent treatment of LDL in a primary prevention setting will most certainly be reflected in the new Adult Treatment Panel guidelines set for release in 2011.
Recent appraisals and criticisms of JUPITER
The results of the JUPITER trial have clearly changed the definition of “at-risk” and have revolutionized our management of patients without documented CVD. The FDA has endorsed and expanded the use of statin therapy for patients with elevated hsCRP, normal LDL and at least one additional risk factor. The Canadian Cardiovascular Society has already incorporated JUPITER’s results into its updated guidelines with respect to preventive therapy for CVD. Despite its widespread acceptance, JUPITER’s results have also sparked a raging controversy for a variety of reasons. De Lorgeril and colleagues and Kaul and colleagues both raised pertinent questions regarding JUPITER’s statistical analysis and early stoppage.
One of the concerns raised stems from the premature stopping of the JUPITER trial, a tactic that has been found to possibly overestimate treatment effect. Conversely, the philosophy behind stopping a trial early is ethical in nature. We must, as researchers, strive to prevent a clear injustice to any group, either by withholding an appropriate treatment or unacceptable risk or side effect of a treatment. In JUPITER’s case, as outlined by Ridker in his recent editorial (a piece that the editorial board of the Archives of Internal Medicine refused to publish), a clearly defined prespecified endpoint had been reached that provided “proof beyond reasonable doubt” of rosuvastatin’s CV benefit in primary prevention.
De Lorgeril and colleagues said the statistical analysis leaves several unanswered questions regarding the supposed discrepancy between total mortality and CV mortality in the trial. However, the strict documentation employed before classifying a death as CV led to only a proportion of probable CV deaths being definitively adjudicated as CV. Total mortality was therefore used as one of the primary endpoints (HR=0.80; 95% CI, 0.67-0.97). Subsequent publication of CV mortality data revealed a similar HR when compared with total mortality (0.82; 95% CI, 0.52-1.2). This negates the concerns about lack of proper documentation of CV mortality and provides a reasonable explanation for the disparity in numbers between CV and total mortality.
Conclusion
Statins form part of the standard of care for secondary prevention of CHD, and it is difficult to imagine that they do not also benefit patients earlier in the disease course. The JUPITER trial has clearly extended the therapeutic arsenal available to preventive cardiologists, but has been met with skepticism by some researchers, particularly cost-conscious internists. The trial’s secondary analyses have provided the requisite details to answer all major concerns with regards to the methodology and statistical analysis.
Although Ray and colleagues’ meta-analysis has shown no benefit of statins on mortality in primary prevention, several others have demonstrated significant benefit with respect to mortality and CV events. The authors of the Archives of Internal Medicine articles said lipid-lowering in people with multiple risks is not advisable because most of the articles studied did not show a reduction in total mortality. They ignore statistically significant reductions in the incidence of MI, HF or the need for primary coronary intervention or CV surgery.
Special consideration must also be given to treating women and the elderly within the primary prevention population, subgroups for whom research has shown a clear benefit from statin treatment. In this day and age, how could someone treat osteoporosis with a bisphosphonate to prevent a fracture in 10 to 15 years, but not treat atherosclerosis with a generic drug that has been shown to potentially cause regression of the disease and markedly reduce CV morbidity?
Rajesh Tota-Maharaj, MD, is a doctor of internal medicine at the Greater Baltimore Medical Center. Michael J. Blaha, MD, is a cardiologist at the Ciccarone Center for the Prevention of Heart Disease at Johns Hopkins University in Baltimore.
Roger S. Blumenthal, MD, is a professor of medicine at Johns Hopkins University, and is the section editor of the CHD/Prevention section of the Cardiology Today Editorial Board.
For more information:
- Balgent C. Lancet. 2005;366:1267-1278.
- Brugts J. BMJ. 2009;338:b2376.
- de Lorgeril M. Arch Intern Med. 2010;170:1032-1036.
- Green L. Arch Intern Med. 2010;170:1007-1008.
- Kaul S. Arch Intern Med. 2010;170:1073-1077.
- Mills E. J Am Coll Cardiol. 2008;52:1769-1781.
- Ray K. Arch Intern Med. 2010;170:1024-1031.