[Skip to Content]
[Skip to Content Landing]
May 21, 1997

p55 Tumor Necrosis Factor Receptor Fusion Protein in the Treatment of Patients With Severe Sepsis and Septic Shock: A Randomized Controlled Multicenter Trial

Author Affiliations

for the Ro 45-2081 Study Group
From the Division of Pulmonary Sciences and Critical Care Medicine, University of Colorado Health Sciences Center, Denver (Dr Abraham); Division of infectious Diseases, Centre Hospitalier Universitaire Vaudois, Lausanne, Switzerland (Dr Glauser); Division of Infectious Diseases, Texas Tech University Health Sciences Center, Lubbock (Dr Butler), Division of Infectious Diseases, Hôpital Cantonal Universitaire, Geneva, Switzerland (Dr Garbino); Division of Infectious Diseases, University of Southern California, Los Angeles (Dr Gelmont); Intensive Care Service, Hôpital St-Luc, Brussels, Belgium (Dr Laterre); Department of Surgery, University of Tennessee, Memphis (Dr Kudsk); Department of Surgery, Academic Hospital Rotterdam, Rotterdam, the Netherlands (Dr Bruining); Department of Anesthiology, University of Arizona Health Sciences Center, Tucson (Dr Otto); Division of Infectious Diseases, Albany Medical Center, Albany (Dr Tobin); and F. Hoffmann-La Roche Ltd, Basel, Switzerland (Drs Zwingelstein, Lesslauer, and Leighton).

JAMA. 1997;277(19):1531-1538. doi:10.1001/jama.1997.03540430043031

Objective.  —To evaluate the safety and efficacy of p55 tumor necrosis factor receptor fusion protein, a recombinant chimeric protein of human p55 (type I) tumor necrosis factor receptor (CD120a) extracellular domain and lgG1 sequences (referred to as p55-lgG), in the treatment of patients with severe sepsis or septic shock.

Design.  —Randomized, prospective, multicenter, double-blind, placebo-controlled clinical trial.

Setting.  —Forty-four community and university-affiliated hospitals in the United States and Europe.

patients.  —There were 498 patients enrolled in this clinical trial.

Intervention.  —Patients prospectively stratified within each site into refractory shock or severe sepsis groups were randomized to receive a single infusion of p55IgG, 0.083 mg/kg, 0.042 mg/kg, or 0.008 mg/kg, or placebo. Patients received standard aggressive medical/surgical care during the 28-day postinfusion period.

Outcome Measure.  —Twenty-eight—day all-cause mortality.

Results.  —The distribution of variables describing demographics, organ system dysfunction or failure, infecting microorganisms, predicted mortality, plasma interleukin 6 levels, and plasma tumor necrosis factor α (TNF-α) levels were similar among patients in the p55-lgG and placebo treatment arms. A planned interim analysis was performed after 201 patients were enrolled. Because a statistically nonsignificant trend toward increased mortality was present in patients who had received 0.008 mg/kg, this treatment arm was discontinued, and the study continued with 3 arms. Among all infused patients, there was a statistically nonsignificant trend toward reduced 28-day all-cause mortality in those who received p55-lgG compared with placebo-treated patients (5% reduction, 0.042 mg/kg vs placebo; 15% reduction, 0.083 mg/kg vs placebo; P=.30). However, in patients with severe sepsis and early septic shock (n=247), therapy with p55-lgG, 0.083 mg/kg, was associated with a 36% reduction in 28-day all-cause mortality compared with placebo (P=.07): 20 (23%) of 87 patients died among those treated with p55-lgG, 0.083 mg/kg; 30 (37%) of 82 among those treated with p55-lgG, 0.042 mg/kg; and 28 (36%) of 78 in the placebo group. A prospectively planned logistic regression analysis to assess treatment effect on 28-day all-cause mortality by means of predicted mortality and serum interleukin 6 levels as continuous covariates demonstrated a significant improvement in outcome for the patients with severe sepsis treated with p55-lgG, 0.083 mg/kg, compared with placebo (P=.01). Serious adverse events, including death and the development of new organ system dysfunction, were reported in 65% of patients infused with placebo, with no increased frequency (56%) present in the 2 p55-lgG treatment arms. There were no reports of immediate hypersensitivity reactions caused by p55-lgG.

Conclusions.  —In this dose-finding study, there was no decrease in mortality between placebo and p55-lgG in all infused patients. In the prospectively defined population of patients with severe sepsis who received p55-lgG, 0.083 mg/kg, there was a trend toward reduced mortality at day 28 that became significant when predicted mortality and plasma interleukin 6 levels were included in a logistic regression analysis.