Colorectal Surgery

Preoperative PCSK9 inhibitors associated with lower 30-day MACEs and mortality versus statins

In a 1:1 propensity-matched cohort of 35,923 pairs, preoperative PCSK9 inhibitor use had lower 30-day MACEs 6.4% vs 9.6% (RR 0.67, NNT 31). All-cause mortality, AKI, respiratory infections, delirium, and elevated liver enzymes were also less frequent with PCSK9 inhibitors. The association persisted across time intervals and therapy durations in the TriNetX multicenter dataset. Results are observational; randomized trials are needed before adopting routine preoperative PCSK9 therapy. 1

Multicenter MySurgeryRisk model predicts major postoperative complications with high AUROCs

Across 508,097 encounters (366,875 patients), the MySurgeryRisk XGBoost models achieved AUROCs of 0.92–0.95 for AKI, ICU admission, MV, and mortality. Procedure code and clinician-specific factors were the dominant predictors of risk across 14 institutions. The model maintained discrimination when trained on 2012–2020 data and validated on 2020–2023 encounters. Before clinical deployment, locally test and recalibrate the model and assess workflow integration and alert fatigue. 2

Four multidomain postoperative recovery endotypes link to symptoms and disease-free survival

In 292 patients after curative colorectal resection, multitrajectory modeling identified four recovery endotypes with distinct inflammation, autonomic, sleep, and activity profiles. Major complication rates and clinically important fatigue and poor sleep increased across endotypes (from 10.6% to 28.6%). Three-year disease-free survival differed across endotypes (86.5% to 61.3%), though survival analyses were exploratory. Multidomain phenotyping may help target symptom management and monitoring, but requires external validation before routine use. 3