|
|
| |||||||||||||||||||||||||||||||||||||||||
One such target is the hemoglobin concentration. Anemia is commonly acquired in the ICU owing to hemodilution, blood loss, reduced red-cell production, and enhanced red-cell destruction. Furthermore, critical illness is characterized by decreased erythropoietin production, a blunted cellular response of erythropoietin, disordered iron metabolism, and nutritional deficiencies.
An important contribution to the literature appears in this issue of the Journal, with Corwin et al.1 posing this clinical question: What is the effect of weekly treatment with epoetin alfa on the percentage of critically ill patients receiving a red-cell transfusion? In this randomized, concealed, multicenter trial, 1460 medical, surgical, or trauma patients were assigned to receive epoetin alfa (40,000 U weekly for up to 3 weeks) or placebo between 48 and 96 hours after admission to the ICU. The primary outcome was the percentage of patients receiving any red-cell transfusion — an end point traditionally dependent on the transfusion threshold of individual physicians. Although more liberal transfusion of red-cell units contributed to higher transfusion rates among controls than among treated patients in previous trials involving critically ill patients,2 the trial by Corwin et al. used a target hemoglobin concentration of 7 g per deciliter to 9 g per deciliter, consistent with the restrictive transfusion strategy favored by the Transfusion Requirements in Critical Care (TRICC) Investigators.3 No difference was found in the percentage of patients who received a red-cell transfusion (46.0% in the epoetin alfa group and 48.3% in the placebo group, P=0.34) or in the number of red-cell units transfused per patient (4.5±4.6 and 4.3±4.8, respectively; P=0.42).
The transparent reporting of the recruitment and retention of patients in this trial is welcome. The reported data underscore the challenges of clinical research in critically ill patients. Of 6168 potentially eligible patients, 3587 (58.2%) patients or surrogates were approached to participate, of whom 2127 (59.3%) did not provide consent. Overall, 109 patients were lost to follow-up, 54 (7.4%) in the epoetin alfa group and 55 (7.6%) in the placebo group.
In this trial, approximately 95% of all patients experienced at least one adverse event, and approximately 44% experienced at least one serious adverse event. These rates reflect the maxim that critical illness is a series of established and acquired evolving, static, or resolving complications. However, such comprehensive reporting of adverse events, while fulfilling regulatory requirements, may induce either inappropriate alarm or a false impression of safety. From a clinical perspective, adverse events are interpreted in the context of the population enrolled. For example, since the physical examination is insensitive for identifying proximal deep-vein thrombosis and pulmonary embolism in semirecumbent patients receiving mechanical ventilation,4 venous thromboembolism will be underdiagnosed when objective screening tests are performed only on the basis of clinical suspicion.
As compared with the placebo group, the increased incidence of thrombotic vascular events in the epoetin alfa group (hazard ratio, 1.41; 95% confidence interval [CI], 1.06 to 1.86) was detected in the absence of screening tests and despite the exclusion of patients at high risk for this complication (those with chronic renal failure,5 previous venous thromboembolism,5 and cardiac ischemia). The 45 to 55% increase in the relative risk of vascular thrombotic events was independent of dose and is consistent with the higher rates of venous thromboembolism among patients with cancer treated with erythropoietin than among those treated with placebo.6 Greater rates of arteriovenous access thrombosis and mortality have been found among patients with chronic renal failure who had higher hemoglobin targets than among those who had lower targets (risk ratio for thrombosis, 1.34; 95% CI, 1.16 to 1.54; risk ratio for mortality, 1.17; 95% CI, 1.01 to 1.35).7 Recently, findings of increased numbers of thrombotic vascular events and deaths and concerns about tumor progression associated with the use of erythropoiesis-stimulating agents have led to black-box warnings from the Food and Drug Administration.
Corwin et al. reported similar mortality in the epoetin alfa group and the placebo group, both at 29 days and at 140 days, on the basis of a Cox regression analysis. This is consistent with a meta-analysis of nine trials involving critically ill patients that showed that the use of erythropoietin does not influence mortality (odds ratio among patients receiving erythropoietin vs. those receiving placebo, 0.86; 95% CI, 0.71 to 1.05).2 In the trial by Corwin et al., as is often the case in critical care research, it is challenging to balance the benefits and harms observed, given the diverse and complex courses of disease and treatment, potentially including vascular thrombotic events, death, transfusions, loss to follow-up, and discontinuation of the study drug (8.3% of patients in the epoetin alfa group and 10.9% of those in the placebo group).
The concept that erythropoietin saves the lives of trauma patients with hemoglobin concentrations below 12 g per deciliter who have been in the ICU for at least 48 hours and who do not have renal insufficiency or prior venous thromboembolism is tantalizing but premature. The interaction between the stratification variables of the admission group and the study group was not significant. In the study by Corwin et al., mortality data for the 793 trauma patients were adjusted for 10 covariates. Twelve more trauma patients had died in the placebo group than in the epoetin alfa group at both day 29 (26 vs. 14 [6.6% vs. 3.5%]) and day 140 (36 vs. 24 [9.2% vs. 6.0%]). This small absolute difference in the number of decedents is insufficient to provide support for the routine use of erythropoietin in practice. Trauma patients have a higher risk of venous thromboembolism than do other ICU patients, which could lower the number of patients "needed to harm." Inferences about subgroups are limited by unclear mechanisms of action, which appear to be exclusive to trauma patients and are not mediated by transfusions averted. Finally, the increased mortality in trials involving other populations receiving more than three doses of erythropoietin remains a concern.2,6,7 We trust that these hypothesis-generating subgroup results will be tested in a large, rigorous trial to carefully evaluate the use of erythropoietin in trauma patients.
As for other research directions, this intriguing trial should incite some investigators in bedside-to-bench research to explain the potential survival benefits of erythropoietin in trauma patients; others may seek understanding of unanticipated adverse events. Clinical investigators may reexamine the risk–benefit ratio for administering erythropoietin to patients with chronic renal failure in the ICU, given their high prevalence of vascular disease. Large observational studies are needed of anemia, erythropoietin, transfusions, and myocardial ischemia in patients in the ICU, in whom the biomarker troponin is commonly elevated, conferring an increased risk of death.8 Health services research on behavioral approaches to blood conservation and restrictive transfusion strategies will also advance this field.9
As this study illustrates, large, rigorous investigations involving vulnerable, critically ill patients are crucial to help inform clinicians about what to do, what to consider, and what to avoid. Without a clear indication for initiating erythropoietin in all critically ill patients, new prescriptions for this drug should be restricted to randomized trials with independent research oversight carefully examining fatal and nonfatal clinically important outcomes.
No potential conflict of interest relevant to this article was reported.
Source Information
From the Departments of Medicine (D.C., M.C.) and Clinical Epidemiology and Biostatistics (D.C.), St. Joseph's Healthcare, McMaster University, Hamilton, ON, Canada.
References
| |||||||||||||||||||||||||||||||||||||||||
This article has been cited by other articles:
HOME | SUBSCRIBE | SEARCH | CURRENT ISSUE | PAST ISSUES | COLLECTIONS | PRIVACY | HELP | beta.nejm.org Comments and questions? Please contact us. The New England Journal of Medicine is owned, published, and copyrighted © 2008 Massachusetts Medical Society. All rights reserved. |
Previous
