Prescribing indications based on successful clinical trials in sepsis: A difficult exercise

Publisher Website
Google Scholar
Abstract
This is a “point of view” article commenting on the indications given several years ago to drotrecogin alpha (activated) for severe sepsis. This appears necessary following the recent publication of studies looking at subgroups within the initial PROWESS study (1–4), of the phase IV ENHANCE study (5), and of the results of two recent randomized clinical trials looking at the efficacy of this compound: the ADDRESS trial (6) (with an editorial (7)) and a pediatric study (not yet published), both recently stopped for futility. Sepsis is not a disease but an extremely heterogeneous syndrome that can stem from myriad infections with widely differing pathophysiological mechanisms and outcomes. The prognosis depends on many factors, including the health history of the patient (e.g., underlying condition, chronic comorbidities, ability to carry out activities of daily living, hospital or intensive care unit (ICU) stay before infection onset, intensity of the infectious insult, magnitude of the inflammatory response, and impact on vital organ functions). To reflect this multiplicity of prognostic factors, the latest international consensus conference (8) developed the “PIRO” grading system, where P stands for predisposing factors, I for the nature of the insult, R for the intensity of the response, and O for the number of organ dysfunctions. The site of infection and nature of the microorganisms also play a crucial role (9). The broad array of tools developed for assessing severity of illness of ICU patients includes global severity scores such as the Acute Physiology and Chronic Health Evaluation II (APACHE II) (10), organ dysfunction assessment methods such as those suggested at the 1991 American College of Chest Physicians/Society for Critical Care Medicine consensus conference, reported by Bone et al. (11), and more recent organ dysfunction scores such as the Sequential Organ Failure Assessment (SOFA) (12). SOFA was initially developed for sepsis but is basically a general organ dysfunction score. Although these tools work fairly well in very large populations (the databases used to develop APACHE II and the Severity Acute Physiology Score [SAPS II] (10, 13) contained data from 10,000 to 15,000 patients), they may be far less accurate for predicting mortality in smaller patient groups, such as those included in most of the recent sepsis trials, including the PROWESS trial (14). In addition, APACHE II and SAPS II have been validated for only the first 24 hours of the ICU stay. The patient’s history of health problems is taken into account only partially by severity scores yet constitutes a major and independent predictor for late mortality, as shown in several studies (15, 16). Unfortunately, most of the recent sepsis trials, including PROWESS failed to assess patient history in detail, in particular to ensure baseline comparability of the placebo and treated patients. A corollary of the complexity and heterogeneity of sepsis is that subgroup analyses may prove extremely misleading, as it is then impossible to ascertain that patients were similar at baseline in the two subgroups chosen for comparison. The optimal end point for assessing mortality in sepsis studies is a focus of active debate. Mortality at 28 days was used in most studies. This end point is open to criticism, however, since 30% of surviving sepsis patients are still hospitalized on day 28 (17). After a long and depressing series of negative trials, the PROWESS trial of recombinant activated protein C finally provided hope that a new drug improved outcomes for patients with severe sepsis (14). This drug, called activated drotrecogin alpha (Xigris, Lilly, Indianapolis, IN), is the first recombinant drug introduced on the market for sepsis.