The French National Registry of HA-1A (Centoxin) in septic shock. A cohort study of 600 patients. The National Committee for the Evaluation of Centoxin

Abstract

Background: Antiendotoxin monoclonal antibodies are under investigation as adjunctive therapy for severe sepsis. One of these antibodies (the human IgM monoclonal antibody HA-1A) was marketed in Europe during 1992 for treatment of patients with severe sepsis, especially those with shock, and probable gram-negative bacilli (GNB) bacteremia, after a placebo-controlled trial showed improved outcome in such patients. This cohort study was designed to assess characteristics and determinants of outcome of patients through a nationwide registry of all adult patients hospitalized in intensive care units of French hospitals and treated with HA-1A.

Methods: Main outcome measures included (1) survival at 14 days, 28 days, and hospital discharge, with comparison of observed with expected hospital mortality according to the Acute Physiology and Chronic Health Evaluation II (APACHE II) score, and (2) analysis of risk factors for death using a Cox proportional hazards model.

Results: Of 600 patients who received HA-1A in a 1-year period, 75% had documented GNB infection and 39% had GNB bacteremia; 94% had hypotension and 54% had refractory shock. Crude overall mortality rates were 49% and 61.3% (P < 10(-4)) at day 14 after therapy and hospital discharge, respectively. Hospital mortality tended to be higher than expected from the APACHE II score (61.3% vs 56%, P = .06), especially in patients without GNB infection (68.1% vs 58.3%). Independent risk factors for death were the APACHE II score (P < 10(-4)), the number of organ system failures (P = 10(-4)), the prognosis of underlying disease (P = .001), and non-GNB infection (relative risk, 1.32 [95% confidence interval, 1.05 to 1.66]; P = .02).

Conclusions: This cohort study did not confirm decreased mortality in patients with GNB bacteremia, as reported in the first placebo-controlled HA-1A trial, and it further suggests excess mortality in the subgroup of patients with non-GNB infection. In clinical trials of patients with severe sepsis, controlling for both the severity of acute illness score and the number of organ system failures, as well as for the severity of underlying disease and the source of infection, should be considered, especially when subgroups are analyzed.