The PAncreatic Surgery Composite Endpoint (PACE): Development and Validation of a Clinically Relevant Endpoint Requiring Lower Sample Sizes.

Abstract

Objective: To provide a composite endpoint in pancreatic surgery.

Background: Single endpoints in prospective and randomized studies have become impractical due to their low frequency and the marginal benefit of new interventions.

Methods: Data from prospective studies were used to develop (n=1273) and validate (n=544) a composite endpoint based on postoperative pancreatic fistula, postpancreatectomy hemorrhage, as well as reoperation and reinterventions. All patients had pancreatectomies of different extents. The association of the developed PAncreatic surgery Composite Endpoint (PACE) with prolonged length of hospital stay >75th percentile and mortality was assessed. A single-institution database was used for external validation (n=2666). Sample size calculations were made for single outcomes and the composite endpoint.

Results: In the internal validation cohort, the PACE demonstrated an area under the curve of 78.0%, a sensitivity of 90.4%, and a specificity of 67.6% in predicting a prolonged length of hospital stay. In the external cohort, the area under the curve was 76.9%, a sensitivity of 73.8%, and a specificity of 80.1%. The 90-day mortality rate was significantly different for patients with a positive versus a negative PACE both in the development and internal validation cohort (5.1% vs 0.9%; P < 0.001), as well as in the external validation cohort (8.5% vs 1.2%, P < 0.001). The PACE enabled sample size reductions of up to 80.5% compared to single outcomes.

Conclusions: The PACE performed well in predicting prolonged hospital stays and can be used as a standardized and clinically relevant endpoint for future prospective trials enabling lower sample sizes and therefore improved feasibility compared to single outcome parameters.