International INtegrated Database for the Evaluation of severe sePsis and drotrecogin alfa (activated) THerapy: component trials and statistical methods for INDEPTH

ABSTRACT

Objectives: To better understand the effects of drotrecogin alfa (activated) (DrotAA) in severe sepsis patients, and the natural progression of severe sepsis, by creating a database of severe sepsis patients using the appropriate statistical analysis methods to integrate data from various trials.

Patients and methods: Patient-level data from five severe sepsis trials, conducted by the same sponsor (Eli Lilly and Company, Indianapolis, IN, USA), were combined in an integrated database. Patients from various studies were included and received either DrotAA at 24 μg/kg/h for 96 hours ( n = 3228) or placebo ( n = 1231), in addition to standard supportive care. The following adjustments to the analyses were made to allow for the combined, and thus non-randomized, nature of the data: (1) differences in observed outcomes between studies were investigated to assess the extent of study-to-study variation before combining study-level data across trials for statistical analysis; (2) random study effects were included in models for patient-level data to capture potential extraneous study-to-study variation; and (3) propensity scores were computed and included as covariates in models for patient-level data to adjust for the non-randomized nature of the data.

Results: Baseline characteristics were similar across the studies, supporting the combination of study-level data across trials. Comparing aggregate event rates between the two treatment arms yielded a relative risk for mortality (DrotAA versus placebo) of 0.79 (95% confidence interval [CI] 0.71–0.88), p < 0.0001. For patient-level analyses, after adjustment for 13 independent variables and random study effects, the odds ratio for mortality in the DrotAA versus placebo patients was 0.71 (95% CI 0.59–0.86), p = 0.0003. With adjustment for 13 independent variables and propensity score, the odds ratio was 0.79 (95% CI 0.67–0.93), p = 0.006. Limitations of this integrated database include the modest total number of the trials in the database and the fact that only one component trial in the database contributed data from both placebo and DrotAA-treated patients.

Summary: A robust severe sepsis database was developed which will be suitable for future studies on the progression of severe sepsis and the mechanism of action of DrotAA. Initial analysis of data from INDEPTH provides additional evidence that treatment of severe sepsis patients with DrotAA is associated with a sustained survival advantage throughout 28‐day follow-up.

Key words: :

• Activated protein C
• Combined databases
• Drotrecogin
• INDEPTH
• Propensity score
• Random effects
• Severe sepsis
• Xigris