by Charles Bankhead
Reanalysis of a landmark cholesterol-lowering trial of people typically considered at low risk for heart attacks indicated that the results are flawed — and do not support the primary-prevention benefits that made headlines, authors of the review asserted.
The reanalysis of the massive JUPITER trial involving almost 18,000 people with low or normal cholesterol but elevated levels of the inflammatory biomarker C-reactive protein (CRP) — turned up no evidence of the “striking decrease in coronary heart disease complications” reported by the trial investigators. Instead, the reanalysis has called into question the involvement of drug companies in such clinical trials, according to an article in the June 28 issue of the Archives of Internal Medicine.
Moreover, Michel de Lorgeril, MD, of Joseph Fourier University and the National Center of Scientific Research in Grenoble, France, and co-authors argue that a major discrepancy exists between the JUPITER trial’s report of significant reductions in nonfatal stroke and myocardial infarction (MI) but a lack of effects on fatal stroke or MI. Moreover, cardiovascular mortality and the case-fatality rate for MI fell far below predicted rates.
“The JUPITER data set appears biased,” de Lorgeril and co-authors concluded.
Paul Ridker, MD, of Harvard Medical School and Brigham and Women’s Hospital in Boston, a major proponent of high-sensitivity CRP (hsCRP) as a marker of coronary risk, dismissed de Lorgeril’s criticisms. Ridker initially reported the JUPITER results at the American Heart Association meeting in 2008.
In an e-mail to MedPage Today, Ridker said that JUPITER data “overwhelmingly stand for themselves. Among a group of individuals with low levels of cholesterol, we clearly demonstrate that those with elevated levels of hsCRP are in fact a high risk population, and that using statin therapy in this group cuts event rates for myocardial infarction and stroke in half.”
And Ridker pointed out that the “FDA has extensively reviewed these data, found the trial to be well conducted, and recently provided a new indication for the use of statins in primary prevention on the basis of the JUPITER data.”
AstraZeneca, the maker of rosuvastatin (Crestor), also defended the JUPITER results and the way in which the study was conducted.
Donna Huang, an AstraZeneca spokesperson, told MedPage Today in an e-mail that the study “was undertaken with a fully independent steering committee, data and safety monitoring board, and academic study statistician.”
And, she pointed out that Ridker and his co-investigators controlled all the data. “AstraZeneca played no role in conducting data analyses and had no access to unblinded trial data,” she wrote.
But the authors of reanalysis say that the links between study sponsor and study investigators are too strong to ignore.
“Three other trials involving rosuvastatin (Crestor) therapy in high-risk patients did not show any protection. The authors of the JUPITER study fail to comment on these negative trials but go on to report secondary endpoint and subgroup analyses that appear to support the efficacy and safety of rosuvastatin therapy,” de Lorgeril and colleagues wrote.
Moreover, de Lorgeril and co-authors point out that nine of 14 authors of the JUPITER article have financial relationships with AstraZeneca, which sponsored the trial. First author Ridker has a patent interest in the assay for CRP, an inflammation biomarker evaluated in all JUPITER trial participants.
“The sponsor’s pervasive role is clearly described in the second paragraph of the ‘Methods’ section of the report: ‘the sponsor collected the trial data and monitored the study sites,'” the authors wrote.
De Lorgeril and co-authors concluded that “the results of the JUPITER trial are clinically inconsistent and therefore should not change medical practice or clinical guidelines. The results of the JUPITER trial support concerns that commercially sponsored clinical trials are at risk of poor quality and bias.”
Adding to the controversy, authors of another article in the same issue of Archives reported that a meta-analysis of 11 large primary-prevention trials showed no effect of statin therapy on all-cause mortality in high-risk patients.
The JUPITER trial (Justification for the Use of Statins in Primary Prevention) has stood alone in its finding of a significant benefit in patients with no evidence of CHD (N Engl J Med 2008; 170: 1032-36). The trial examined the effect of rosuvastatin in patients with normal or low cholesterol levels but elevated levels of CRP.
Investigators randomized 17,802 apparently healthy men and women to rosuvastatin or placebo. The primary endpoint was the composite of MI, stroke, arterial revascularization, hospitalization for unstable angina, and cardiovascular mortality. The trial ended prematurely after a median follow-up of 1.9 years, when an interim analysis showed a 44% reduction in the composite endpoint in the rosuvastatin group.
De Lorgeril and co-authors cited the early termination as one of several methodologic problems with JUPITER. Although prespecified early stopping points are a well-accepted feature of clinical trials, the rules for stopping should be clearly described. That was not the case in the published description of the JUPITER protocol.
“Indeed, we still do not know which endpoint was used to define [the rules for stopping], or which level of benefits — unexpected on the basis of the a priori calculated hypothesis — was required to justify early termination,” de Lorgeril and co-authors wrote.
The authors also expressed concern that the trial ended early despite the fact that the data were not consistent with a large difference between rosuvastatin and placebo. Considering only “hard endpoints,” such as fatal and nonfatal MI and stroke, the trial ended after 240 events, a small number given the size of the trial.
“Moreover, a close examination of the all-cause mortality curves shows that the curves were actually converging when the trial ended, suggesting that the borderline significant difference between groups may have disappeared in the case of a slightly longer follow-up,” the authors wrote.
De Lorgeril and co-authors also could not resolve what they considered implausible differences in hard endpoints. Performing their own calculations based on data reported in the JUPITER published article, the authors concluded that total cardiovascular mortality was identical in the rosuvastatin and placebo groups. Despite the JUPITER investigators’ statement of an “unequivocal reduction in cardiovascular mortality” as the justification for stopping the trial, the published article included no cardiovascular mortality data.
Another inconsistency related to the mix of fatal and nonfatal cardiovascular endpoints.
“The ratio of fatal myocardial infarction (nine for rosuvastatin and six for placebo) to nonfatal myocardial infarction (22 and 62) is incredibly low, especially in the placebo group,” de Lorgeril and co-authors wrote. “Mortality from acute myocardial infarction is a very important issue in cardiology. The data would suggest that the hearts of the JUPITER patients were unexpectedly — and inexplicably — highly resistant to acute ischemic and infarction.”
De Lorgeril and co-authors also take JUPITER investigators to task for failure to describe the criteria used to define cardiovascular death and for omission of any discussion of sudden cardiac death (SCD).
“It is therefore very surprising that no SCD is reported in the JUPITER trial because SCD usually represents about 65% to 70% of total cardiac mortality,” the authors of the reanalysis observed.
On the basis of their review, de Lorgeril and co-authors concluded that “the time has come for a critical reappraisal of cholesterol-lowering and statin treatments for the prevention of CHD complications. The emphasis on pharmaceuticals for the prevention of CHD diverts individual and public health attention away from the proven efficacy of adopting a healthy lifestyle, including regular physical activity, not smoking, and a Mediterranean-style diet.”
The meta-analysis reported in the same issue of the journal encompassed 11 randomized clinical trials involving 65,229 patients and 244,000 person-years of follow-up. The primary objective was to determine whether statin therapy reduces all-cause mortality among intermediate and high-risk people with no history of cardiovascular disease, wrote Kausik K. Ray, MD, of the University of Cambridge in England, and co-authors.
Collectively, the trials documented a total of 2,793 cardiovascular deaths. Analysis of all-cause mortality showed a nonsignificant 9% reduction in relative risk in patients treated with statins (95% CI 0.83 to 1.01). The analysis also showed no statistical heterogeneity among the studies.
The findings call into question current clinical guidelines regarding use of statins’ cardiovascular prevention.
“Current prevention guidelines endorse statin therapy for subjects at high global risk of incident CVD as a means to reduce fatal and nonfatal events,” Ray and co-authors wrote. “Due consideration is needed in applying statin therapy in lower-risk primary prevention populations.”
In an editorial that accompanied the two articles, Lee A. Green, MD, of the University of Michigan in Ann Arbor, said the de Lorgeril and Kay studies add fuel to a high-stakes debate.
“Most patients who have major coronary events do not have previously known disease, so primary prevention could deliver large outcome benefits,” wrote Green.
Extending the discussion of large risks and benefits, Green added that “three quarters of the patients who take statins are taking them for primary prevention, so enormous expenditures (from payors’ perspectives) or revenues (from industry’s perspective) are at stake.”
The uncertainty surrounding the outcomes of such clinical trials also is high, Green continued. Given the short-term nature of clinical trials versus the need for long-term treatment in clinical practice, “no data settle the increasing-returns versus the diminishing-returns extrapolations.”
“Ray and colleagues’ meta-analysis makes it clear that in the short term, for true primary prevention, the benefit, if any, is very small,” Green wrote. “In the long term, although sincere advocates on both sides will try to convince us otherwise, we really must admit that we do not know. We need good research to find out, and, as de Lorgeril and colleagues point out, that search must be free of incentives to find any particular desired answer.”
Charles Bankhead is a MedPage Today staff writer.