AIM: Anastomotic leakage (AL) is a common and serious complication following sphincter-preserving surgery for rectal cancer. Early detection and intervention can improve clinical outcomes. The aim of this prospective cohort study was to compare intraperitoneal microdialysis with a clinical scoring system for early detection of AL. METHOD: A microdialysis catheter was anchored near the anastomosis at low anterior resection (LAR) for rectal cancer. Peritoneal fluid samples were analysed (lactate, pyruvate, glucose and glycerol concentration) 4-hourly and compared with a daily clinical leak score (DULK = Dutch leakage). At day 7 a pelvic CT with rectal contrast enema was performed to establish if there had been a radiological leak. RESULTS: In this two-centre study, 129 patients [median age 65 (26-82) years; 60.5% male] underwent LAR. The leak rate was 27% (grade A, n = 11; grade B, n = 12; grade C, n = 12). Receiver operator characteristic analysis demonstrated a lactate cut-off value of 9.8 mm and had 77% sensitivity, 82% specificity, 78% accuracy, a positive predictive value (PPV) of 58, a negative predictive value (NPV) of 88 (CI 79-94) and an area under the curve (AUC) of 0.9 for AL. This compared with a clinical score ≥ 4, which had 57% sensitivity, 79% specificity, 71% accuracy, a PPV of 46, a NPV of 82 and an AUC of 0.7 for AL. The mean day for a positive test when using delta lactate ≥ 6.3 mm was 1.6 days and for leak score ≥ 4 it was 3.3 days (NS). CONCLUSION: When AL occurs, intraperitoneal lactate concentration increases over time, and at a certain cut-off has a higher sensitivity, specificity, accuracy, PPV and NPV than a clinical scoring system.