The Comprehensive Complication Index

In Brief Objective: To test whether the newly developed comprehensive complication index (CCI) is more sensitive than traditional endpoints for detecting between-group differences in randomized controlled trials (RCTs). Background: A major challenge in RCTs is the choice of optimal endpoints to detect treatment effects. Mortality is no longer a sufficient marker in studies, and morbidity is often poorly defined. The CCI, integrating all complications including their severity in a linear scale ranging from 0 (no complication) to 100 (death), is a new tool, which may be more sensitive than other traditional endpoints to detect treatment effects on postoperative morbidity. Methods: The CCI was tested in 3 published RCTs from European centers evaluating pancreas, esophageal and colon resections. To compare the sensitivity of the CCI with traditional morbidity endpoints, for example, presence of any (yes/no) or only the most severe complications, all postoperative events were assessed, and the CCI calculated. Treatment effects and sample size calculations were compared using the CCI and traditional endpoints. Results: Although RCTs failed to show between-group differences using any or most severe complications, the CCI revealed significant differences between treatment groups in 2 RCTs—after pancreas (P = 0.009) and esophageal surgery (P = 0.014). The CCI in the RCT on colon resections confirmed the absence of between-group differences (P = 0.39). The required sample sizes in trials are up to 9 times lower for the CCI than for traditional morbidity endpoints. Conclusions: This study demonstrates superiority of the CCI to traditional endpoints. The CCI may serve as an appealing endpoint for future RCTs and may reduce the sample size. This study demonstrates superiority of the new comprehensive complication index (CCI) over traditional endpoints in detecting between-group differences in randomized control trials. The CCI may therefore serve as preferred endpoint in future RCTs and may help in designing RCTs with smaller sample sizes.