|
|
| |||||||||||||||||||||||||||||||||||||||
It is impossible to tell how many of the patients were receiving "optimal therapy" at baseline, defined as meeting current targets in all three of the following areas: the use of aspirin when indicated, hypertension treated to national goals, and no tobacco use. Future substudies should examine the number of patients who would need to be treated to prevent one cardiovascular event in the subgroup of patients who were already receiving optimal therapy at baseline. Public health might be better served by improving compliance with existing standards.
Raymond J. Gibbons, M.D.
Mayo Clinic
Rochester, MN 55905
gibbons.raymond{at}mayo.edu
Dr. Gibbons reports receiving a research grant from King Pharmaceuticals. No other potential conflict of interest relevant to this letter was reported.
References
Although Ridker et al. state that "by design, the study population was diverse," no Asian countries are included among the study sites, and the demographic breakdown of the 17,802 patients according to race or ethnic group includes only white, black, Hispanic, and "other or unknown." It would appear that there may have been a conscious choice to exclude people of Asian descent from JUPITER.
Adriane Fugh-Berman, M.D.
Georgetown University Medical Center
Washington, DC 20007
ajf29{at}georgetown.edu
References
On the basis of our calculations, in the rosuvastatin group, as compared with the placebo group, the number of deaths from cardiovascular causes was not significantly reduced (31 vs. 37 deaths), although the number of deaths from any cause was significantly reduced (167 vs. 210 deaths). This finding is at odds with extensive data from previous statin trials. In addition, the authors suggest that their results support treating patients on the basis of elevations in C-reactive protein. However, they provide no results showing that C-reactive protein is an independent predictor of the relative or absolute benefit of therapy, since the treatment effects seen with rosuvastatin could have been mediated by reductions in low-density lipoprotein (LDL) cholesterol. Multivariable models that adjust for baseline levels of LDL cholesterol and changes in LDL cholesterol over time would further clarify the role of C-reactive protein.
Paul S. Chan, M.D., M.Sc.
Mid America Heart Institute
Kansas City, MO 64112
paulchan{at}umich.edu
Brahmajee K. Nallamothu, M.D., M.P.H.
University of Michigan Medical School
Ann Arbor, MI 48109
Rodney A. Hayward, M.D., M.P.H.
VA Ann Arbor Health Services Research and Development Center of Excellence
Ann Arbor, MI 48104
Could the authors provide data showing whether there was a gradient of risk for cardiovascular events and death according to baseline levels of C-reactive protein or a gradient of benefit from rosuvastatin according to the extent of the baseline elevation? Furthermore, could they reassure clinicians that there was no incremental risk among patients with the lowest baseline cholesterol levels who were treated with a lipid-lowering statin? Was the clinical benefit explained by changes in levels of C-reactive protein, and how could clinicians monitor the intervention in practice in order to achieve a clinical benefit?
Elizabeth R. Jenny-Avital, M.D.
Jacobi Medical Center
Bronx, NY 10461
jennyavita{at}earthlink.net
Among 59,006 patients, the risk of diabetes for patients receiving a statin was similar to that for patients receiving placebo (relative risk, 1.06; 95% CI, 0.91 to 1.23). The risk of diabetes appears to increase with increased potency of the lipid-lowering agent. For the two large, placebo-controlled trials of pravastatin, the West of Scotland Coronary Prevention Study (WOSCOPS) and the Long-Term Intervention with Pravastatin in Ischemic Disease (LIPID) study, the relative risk of diabetes in the pravastatin group was 0.81 (95% CI, 0.64 to 1.02). For the two large, placebo-controlled trials of rosuvastatin, JUPITER and the Controlled Rosuvastatin Multinational Trial in Heart Failure (CORONA), the relative risk of diabetes in the rosuvastatin group was 1.22 (95% CI, 1.05 to 1.42). For drugs with intermediate potency, simvastatin and atorvastatin, the values fell in between these extremes.
Koon-Hou Mak, M.D.
Gleneagles Medical Centre
Singapore 258499, Singapore
makheart{at}gmail.com
Edwin S.-Y. Chan, B.V.M.S., Ph.D.
Singapore Clinical Research Institute
Singapore 138669, Singapore
References
The majority of randomized clinical trials that are stopped early because of an observed benefit of the treatment under investigation are industry-funded drug trials that are stopped at the first interim analysis, with the results published in a high-impact medical journal. The hazard ratio of 0.56 for the primary end point in JUPITER is close to the median risk ratio of 0.53 among 143 truncated randomized trials.1 Truncated trials overestimate the treatment effect.2 This factor was important in the Candesartan in Heart Failure: Assessment of Reduction in Mortality and Morbidity (CHARM) study, in which early stopping was resisted.3 Because rosuvastatin would be given long-term for primary prevention, the JUPITER study investigators should have continued follow-up to determine whether the positive results would have continued or would have declined to a more modest effect.
Luc A. Pierard, M.D., Ph.D.
University Hospital Sart Tilman
B 4000 Liege, Belgium
lpierard{at}chu.ulg.ac.be
Dr. Pierard reports receiving lecture fees from AstraZeneca, Merck–Schering-Plough, and Pfizer and grant support from AstraZeneca and Schering-Plough, and serving on advisory boards for AstraZeneca and Merck–Schering-Plough. No other potential conflict of interest relevant to this letter was reported.
References
The correct P value for the sequential analysis, as conducted, is P<0.05, not P<0.00001, as reported. In addition, the point estimate of the treatment effect from a trial that was terminated early for efficacy is biased in favor of the treatment.1 Thus, although it can be agreed that rosuvastatin lowered the risk of cardiovascular disease in this study, the methods used to report the results overestimate the strength of the association.
Clarence E. Davis, Ph.D.
University of North Carolina
Chapel Hill, NC 27599
References
Michael T. Koller, M.D., M.Sc.
Heiner C. Bucher, M.D., M.P.H.
Basel Institute for Clinical Epidemiology and Biostatistics
CH-4031 Basel, Switzerland
kollerm{at}uhbs.ch
Ewout W. Steyerberg, Ph.D.
Erasmus University
3000 Rotterdam, the Netherlands
References
Full prescribing data for rosuvastatin among Asians were not available in 2002. Thus, Fugh-Berman is correct that Asian participation was marginal. The safety of rosuvastatin has subsequently been established and the 20-mg dose approved for patients of Asian descent.
The calculations by Chan et al. are incorrect partly because they do not account for deaths from vascular causes, such as aneurysm rupture. Furthermore, because we prespecified very strict confirmation criteria, many out-of-hospital deaths from cardiovascular causes were classified as being from noncardiovascular causes for trial purposes. On the basis of these strict criteria, the numbers of confirmed deaths from cardiovascular causes were 35 in the rosuvastatin group and 43 in the placebo group, with a hazard ratio in the rosuvastatin group of 0.82 (95% CI, 0.52 to 1.27), which was similar to the reported hazard ratio for death from any cause of 0.80 (95% CI, 0.67 to 0.97). Our trial is consistent with the notion that achieving very low levels of high-sensitivity C-reactive protein and LDL cholesterol can enhance statin benefits1,2 — analyses that will interest Chan et al., along with Jenny-Avital. As anticipated, the absolute risk of a cardiovascular event increased with increased levels of high-sensitivity C-reactive protein and decreased with decreased levels.
We partially disagree with Mak and Chan. If the "protective" effect on diabetes incidence reported in WOSCOPS is treated as hypothesis-generating, then a summary of published hypothesis-testing trials demonstrates that all statins modestly increase the risk of diabetes, with no heterogeneity according to potency. In our study, many of the patients in whom diabetes developed were obese or had an impaired fasting glucose level, groups in which large reductions in vascular events were associated with rosuvastatin.
The independent data and safety monitoring board for our trial followed rigorous principles3 in its prespecification that early termination of the study because of an observed benefit would require proof beyond a reasonable doubt. Members of the board were experienced in monitoring publicly and privately funded trials and viewed the trial's prespecified statistical boundary as only one component required for proof. Although the formal statistical boundary was conservative and evaluated only after accrual of ample data, the board elected to continue the trial for an additional 6 months after the boundary was crossed. Data that were accrued thereafter independently confirmed both the magnitude and statistical significance of the apparent benefit. We thus respectfully disagree with Pierard and Davis. The board appropriately protected the interests of society and the trial participants and provided a valid estimate of the treatment effect.4
The evaluation by Koller et al. ignores the significant reduction in death from any cause that we observed. If death from any cause is added to our primary composite outcome (a standard approach to account for competing risks), then the absolute risk difference increases and the number needed to treat declines.
Paul M Ridker, M.D.
Robert J. Glynn, Sc.D.
Brigham and Women's Hospital
Boston, MA 02115
pridker{at}partners.org
References
| |||||||||||||||||||||||||||||||||||||||
HOME | SUBSCRIBE | SEARCH | CURRENT ISSUE | PAST ISSUES | COLLECTIONS | PRIVACY | TERMS OF USE | HELP | beta.nejm.org Comments and questions? Please contact us. The New England Journal of Medicine is owned, published, and copyrighted © 2009 Massachusetts Medical Society. All rights reserved. |
Previous
