Background: Over the last 25 years, a growing number of clinical trials have evaluated novel sepsis therapies. To promote uniformity in inclusion criteria for patient enrollment, the American College of Chest Physicians and Society of Critical Care Medicine first published consensus conference definitions for sepsis in 1992.
Study objectives: To characterize (1) the utilization of specific criteria for patient enrollment in sepsis clinical trials and (2) the impact that the consensus conference definitions have had on these criteria.
Design: We used MEDLINE to identify clinical trials in sepsis from 1976 to 2001. Clinical trials published after the consensus conference (ACC; from 1993 to 2001) were compared with trials published before the consensus conference (BCC; from 1976 to 1992).
Results: We identified 176 clinical trials (ACC, 119 trials; BCC, 57 trials). Clinical trials published ACC were more likely to utilize or reference a previously published standard for inclusion criteria (65% vs 11%, respectively; p < 0.001). The consensus conference definitions were the standards used in 69% of these trials. The utilization of specified values for WBC count, temperature (T), heart rate (HR), and respiratory rate (RR) was significantly increased in the ACC group compared to the BCC group, as follows: WBC count, 62% vs 26%, respectively (p < 0.001); T, 89% vs 56%, respectively (p < 0.001); HR, 77% vs 26%, respectively (p < 0.001); and RR, respectively 76% vs 28% (p < 0.001). ACC, clinical trials were less likely to require blood culture positivity (4 of 119 trials [3%] vs 9 of 57 trials [16%], respectively; p < 0.006) and were more likely to incorporate markers of acute organ dysfunction (81 of 119 trials [68%] vs 28 of 57 trials [49%], respectively; p < 0.03) in the inclusion criteria.
Conclusions: (1) Since 1992 there has been a significant increase in the utilization of predefined sepsis criteria for patient enrollment in clinical trials, and this increase can be attributed to the existence of consensus conference definitions. (2) Compared to inclusion criteria BCC, inclusion criteria ACC were less reliant on blood culture positivity and were more likely to incorporate markers of organ dysfunction.