Correspondence
Recombinant Human Activated Protein C for Severe Sepsis
N Engl J Med 2001; 345:219-221July 19, 2001
- Article
To the Editor:
Bernard and coworkers (March 8 issue)1 conclude that recombinant human activated protein C (drotrecogin alfa [activated]) is efficacious in patients with sepsis. The placebo and treatment groups appeared to be closely matched, with one important exception: the interval between the diagnosis of sepsis and the administration of appropriate antibiotic therapy. The authors state only that appropriate antibiotic therapy was started within 48 hours of the diagnosis of severe sepsis.
It is axiomatic that the more severe the sepsis, the more important the requirement for prompt administration of appropriate antibiotic therapy. In a study of gram-negative bacterial sepsis in mice, we demonstrated that each hour that appropriate antibiotic therapy was delayed resulted in progressive increases in mortality.2 After a delay of three to eight hours, depending on the bacterial species, mortality by day 4 was no longer reducible. Windows of antibiotic effectiveness can also close so rapidly during severe sepsis in humans that even prompt initiation of antibiotic therapy may not reduce mortality significantly.3 We agree that difficulty determining the time of onset of sepsis is a major impediment to progress in clinical trials, whereas animal models do not pose this problem.4 Nevertheless, because each hour of delay in initiating appropriate antibiotic therapy in patients with severe sepsis may affect the outcome, the greatest possible precision must be used in showing that this factor is similar in the study groups before any differences in outcome are attributed to other antisepsis treatments. In a previous trial involving sepsis, the investigators applied this principle, performing a statistical analysis to compare the study groups.5
5 ReferencesSheldon E. Greisman, M.D.
University of Maryland School of Medicine, Baltimore, MD 21201Curtis A. Johnston, M.D.
Lawrence and Memorial Hospital, New London, CT 06320Mark S. Gosnell, M.D.
University of Maryland School of Medicine, Baltimore, MD 212011
Bernard GR ,Vincent J-L ,Laterre P-F , et al. Efficacy and safety of recombinant human activated protein C for severe sepsis. N Engl J Med 2001;344:699-709
Full Text | Web of Science | Medline2
Johnston CA ,Greisman SE . Endotoxemia induced by antibiotic therapy: a mechanism for adrenal corticosteroid protection in gram-negative sepsis. Trans Assoc Am Physicians 1984;97:172-181
Medline3
Radetsky M . The timing of antimicrobial therapy and outcome in serious bacterial infection. Curr Opin Infect Dis 1994;7:341-344
CrossRef | Web of Science4
Dhainaut JFA ,Tenaillon A ,Hemmer M , et al. Confirmatory platelet-activating factor receptor antagonist trial in patients with severe Gram-negative bacterial sepsis: a phase III, randomized, double-blind, placebo-controlled, multicenter trial. Crit Care Med 1998;26:1963-1971
CrossRef | Web of Science | Medline5
Ziegler EJ ,Fisher CJ Jr ,Sprung CL , et al. Treatment of gram-negative bacteremia and septic shock with HA-1A human monoclonal antibody against endotoxin. N Engl J Med 1991;324:429-436
Full Text | Web of Science | Medline
To the Editor:
Bernard et al. report the efficacy and safety of recombinant human activated protein C for the treatment of severe sepsis. The authors state, “reductions in the relative risk of death were observed regardless of whether the patients had a deficiency of protein C at base line.” They also state that “measurements of protein C are not necessary to identify which patients would benefit from treatment with drotrecogin alfa activated.” The results of a stratified analysis, presented in Table 4 of their article, indicate that in the group of patients who did not have a deficiency of protein C at base line, the reduction in the risk of death with the use of drotrecogin alfa activated was not statistically significant.
The incidence of serious bleeding was higher in the group assigned to drotrecogin alfa activated than in the placebo group (3.5 percent vs. 2.0 percent). Serious bleeding occurred primarily in patients with an identifiable predisposition to bleeding. It would be interesting to know whether the authors stratified the risk of bleeding in relation to the presence or absence of protein C deficiency at base line as well. This type of analysis might be valuable because it would allow a more complete risk–benefit analysis for this subgroup of patients.
Sandeep Kapur, M.D.
Yizhak Kupfer, M.D.
Sidney Tessler, M.D.
Maimonides Medical Center, Brooklyn, NY 11219
stessler@maimonidesmed.org To the Editor:
Bernard et al. conclude that drotrecogin alfa activated significantly reduced mortality in patients with severe sepsis. However, the randomization scheme evidently failed, and the authors did not adjust for this in their analyses. Nor did they or the editorialist1 discuss possible selection bias. Table 1 of their article does not give P values, but it shows that the patients in the placebo group were more fragile and ill than those in the group assigned to drotrecogin alfa activated. Of the nine categories of prior preexisting conditions listed, eight were more prevalent in the placebo group than in the treatment group; the difference was significant for congestive cardiomyopathy (P=0.038) and it approached significance for chronic obstructive pulmonary disease and cancer (P=0.065 and P=0.062, respectively). More patients in the placebo group underwent mechanical ventilation (P=0.039), and more patients in this group were in shock and needed vasopressors (P=0.066). A remarkable error is the enrollment of one patient in the treatment group who did not meet the inclusion criterion of a dysfunctional organ system.
There is an additional reason to doubt the validity of the trial. All base-line data on coagulation and inflammation were missing more often in the placebo group than in the treatment group (Table 3), and the difference in plasma D-dimer levels was significant (P=0.028). Did the patients in the placebo group initially receive less attention, care, or both? Taken together, these discrepancies between the study groups may account for the observed difference in survival after 28 days. A 6.1 percent absolute reduction in mortality from severe sepsis with the use of a single drug would be a striking achievement, even if this reduction occurred in a highly selected group of patients. However, we are not convinced that the use of drotrecogin alfa activated really represents such a breakthrough.
1 ReferencesAlewijn Ott, M.D., Ph.D.
Henri A. Verbrugh, M.D., Ph.D.
Erasmus Medical Center Rotterdam, 3000 CA Rotterdam, the Netherlands
ott@bacl.azr. nl 1
Mathay MA . Severe sepsis -- a new treatment with both anticoagulant and antiinflammatory properties. N Engl J Med 2001;344:759-762
Full Text | Web of Science | Medline
To the Editor:
Bernard et al. state, “In the United States, approximately 750,000 cases of sepsis occur each year, at least 225,000 of which are fatal.” If septicemia were responsible for this number of deaths, it would be the third leading cause of death in the United States, after diseases of the heart and cancers, and would account for 9.72 percent of all reported deaths in 1997.
The Statistical Abstract of the United States, 2000 edition, provides data on deaths and death rates according to selected causes. In Table 126, septicemia is reported to have been the cause of 23,600 deaths in 1998, with a crude death rate of 8.7 per 100,000 population.1
1 ReferencesPeter E. O'Connor
1460 Bayhead Dr., Apt. 1409, Virginia Beach, VA 23456Author/Editor Response
The authors reply:
To the Editor: Greisman and colleagues are concerned about the interval between the onset of sepsis and the administration of appropriate antibiotic therapy.1,2 According to the judgment of a blinded clinical-evaluation committee, 89.3 percent of the patients in the drotrecogin alfa activated group and 89.1 percent of those in the placebo group received appropriate antibiotics within 24 hours of the onset of severe sepsis.
Kapur et al. ask about treatment effects in the patients who did not have a deficiency of protein C. In this subgroup of 195 patients, the relative reduction in the risk of death associated with the administration of drotrecogin alfa activated was 41.7 percent (P=0.06). Two of 90 patients who received drotrecogin alfa activated (2.2 percent) had a serious bleeding event — an event rate similar to that in the overall study population.
Ott and Verbrugh suggest that a maldistribution of patients with certain prior or preexisting conditions may have confounded the analysis. However, the misuse of significance testing to determine potentially relevant imbalances in base-line characteristics is well documented.3 A more appropriate approach is to use analyses of mortality from all causes at 28 days, stratified according to clinically relevant covariates selected a priori, as we did in our trial; the results are shown in Table 4 of our article. For the covariates discussed by Ott and Verbrugh, the adjusted relative reduction in the risk of death was as follows: congestive cardiomyopathy, 18.9 percent (P=0.008); chronic obstructive pulmonary disease, 19.5 percent (P=0.007); cancer, 19.8 percent (P=0.006); mechanical ventilation, 18.5 percent (P=0.009); shock, 19.8 percent (P=0.006); and use of vasopressors, 18.7 percent (P=0.008). Thus, formal adjustment of these covariates produced results similar to those for the overall trial (relative reduction in the risk of death, 19.4 percent; P=0.005).
Ott and Verbrugh also suggest that less attention, care, or both might have been given to patients in the placebo group in whom base-line measurements of coagulation and inflammation were not obtained. Actually, in both study groups, these measurements were obtained in more than 90 percent of patients, indicating that a very committed group of investigators participated in this double-blind, multinational study.
O'Connor asks about the data we cited on the incidence of severe sepsis and mortality attributable to it. We acknowledge that there is a dearth of reliable statistics on this question. For our report, we relied on recent work by Angus et al., whom we believe have performed the most robust exploration of this question to date.4 They report that death associated with severe sepsis occurs as often as death associated with acute myocardial infarction, just as O'Connor suggests. In any case, there is little doubt that severe sepsis is a huge threat to public health. Furthermore, considering the striking age-related incidence and mortality rates, the burden of severe sepsis will probably become an even larger problem as our population ages.
4 ReferencesGordon R. Bernard, M.D.
E. Wesley Ely, M.D., M.P.H.
Vanderbilt University School of Medicine, Nashville, TN 37232
gordon.bernard@mcmail. vanderbilt. edu Jeffrey D. Helterbrand, Ph.D.
Lilly Research Laboratories, Indianapolis, IN 462851
Kollef MH ,Sherman G ,Ward S ,Fraser VJ . Inadequate antimicrobial treatment of infections: a risk factor for hospital mortality among critically ill patients. Chest 1999;115:462-474
CrossRef | Web of Science | Medline2
Meehan TP ,Fine MJ ,Krumholz HM , et al. Quality of care, process, and outcomes in elderly patients with pneumonia. JAMA 1997;278:2080-2084
CrossRef | Web of Science | Medline3
Senn S. Statistical issues in drug development. Chichester, England: John Wiley, 1997:98.
4
Angus DC, Linde-Zwerble WT, Lidicker J, Clermont J, Carcillo J, Pinsky PR. National incidence, cost and outcome of severe sepsis in the US. Crit Care Med (in press).
- Citing Articles (2)
Citing Articles
