Preoperative Skeletal Muscle Index as an Independent Predictor of Anastomotic Leakage After Colorectal Cancer Surgery: A Retrospective Cohort Study

Background: Anastomotic leakage remains the most feared complication after colorectal cancer surgery, and reliable preoperative risk stratification is still evolving. The skeletal muscle index (SMI), derived from routine computed tomography, offers an objective, widely available measure of muscle mass. This study evaluated whether a low preoperative SMI predicts anastomotic leakage after colorectal cancer resection. Methods: We retrospectively screened 550 consecutive patients who underwent colorectal cancer resection with primary anastomosis at a single tertiary referral centre between 2018 and 2024, of whom 468 satisfied the eligibility criteria and formed the analytic cohort. Skeletal muscle area was measured on preoperative axial computed tomography at the third lumbar vertebra and normalised to the height squared to obtain the SMI. Anastomotic leakage was defined according to the International Study Group of Rectal Cancer criteria. Receiver operating characteristic analysis identified a single pragmatic SMI cut-off. Group comparisons used Mann–Whitney U and chi-square tests, and independent predictors were assessed by multivariable logistic regression. Results: Anastomotic leakage occurred in 33 of 468 patients (7.1%). Patients who developed leakage had a significantly lower mean SMI than those who did not (40.0 cm2/m2 vs. 46.8 cm2/m2, p = 0.005). An SMI below 48.6 cm2/m2 predicted leakage with 90.9% sensitivity, 38.2% specificity, and a 98.2% negative predictive value (area under the curve 0.646). Because specificity (38.2%) and positive predictive value (10.0%) were low, the index performed as a rule-out marker: a preserved SMI identified a low-risk subgroup, whereas a low-SMI had limited positive predictive value. The leakage rate reached 10.0% in the low SMI group versus 1.8% in the high-SMI group (p = 0.0008). Following multivariable analysis, low SMI was the only independent predictor of anastomotic leakage (adjusted odds ratio: 5.67, 95% confidence interval: 1.69 to 19.01, p = 0.005). Conclusions: A low preoperative SMI independently increased the odds of anastomotic leakage almost six-fold after colorectal cancer surgery, while its high negative predictive value identified a large group of patients at very low risk. Routine SMI assessment on existing staging computed tomography may strengthen preoperative risk stratification at no additional cost or radiation.