Individual patient data meta-analysis of hydroxyethyl starch 130/0.4-0.42 versus crystalloid for fluid resuscitation in patients with severe sepsis: a statistical analysis plan.

BACKGROUND

The Crystalloid versus Hydroxyethyl Starch Trial (CHEST) and the Scandinavian Starch in Severe Sepsis/ Septic Shock (6S) trial reported that 6% hydroxyethyl starch (HES) is associated with increased use of renal replacement therapy and death in critically ill patients. Data collection was harmonised between the two trials in order to facilitate a preplanned individual patient data meta-analysis (IPDMA) of patients with severe sepsis.

OBJECTIVES AND RATIONALE: To publish a statistical analysis plan (SAP) for an IPDMA of patients with severe sepsis enrolled in the 6S trial and the CHEST.

METHODS AND OUTCOMES: The SAP is described in broad detail with specific information regarding baseline characteristics and process of care. The outcomes for the trial have been described and are presented as primary, secondary and exploratory outcomes with appropriate comparisons between groups detailed. Subgroups have been defined based on pre-randomisation variables.

CONCLUSION

We developed a preanalysis SAP to combine data on patients with severe sepsis from the 6S trial and the CHEST. Prepublication of our SAP will reduce the risk of bias in the reporting of the results and improve confidence in the estimates of effects, allowing comparisons with conventional meta-analyses and assisting in the translation of research findings into clinical practice.