Control groups in recent septic shock trials: a systematic review.

Abstract

The interpretation of septic shock trial data is profoundly affected by patients, control intervention, co-interventions and selected outcome measures. We evaluated the reporting of control groups in recent septic shock trials. We searched for original articles presenting randomized clinical trials (RCTs) in adult septic shock patients from 2006 to 2016. We included RCTs focusing on septic shock patients with at least two parallel groups and at least 50 patients in the control group. We selected and evaluated data items regarding patients, control group characteristics, and mortality outcomes, and calculated a data completeness score to provide an overall view of quality of reporting. A total of 24 RCTs were included (mean n=287 patients and 71% of eligible patients were randomized). Of the 24 studies, 14 (58%) presented baseline data on vasopressors and 58% the proportion of patients with elevated lactate values. Five studies (21%) provided data to estimate the proportion of septic shock patients fulfilling the Sepsis-3 definition. The mean data completeness score was 19 out of 36 (range 8-32). Of 18 predefined control group characteristics, a mean of 8 (range 2-17) were reported. Only 2 (8%) trials provided adequate data to confirm that their control group treatment represented usual care. Recent trials in septic shock provide inadequate data on the control group treatment and hemodynamic values. We propose a standardized trial dataset to be created and validated, comprising characteristics of patient population, interventions administered, hemodynamic values achieved, surrogate organ dysfunction, and mortality outcomes, to allow better analysis and interpretation of future trial results.