Mohammed Nabil Quraishi, Maryam A AlAhmad, Alaa El-Hussuna, Badr Al-Bawardy
Although there was a signal for association of advanced-therapy with certain complications, the signal is small, heterogeneous, of very low certainty, and absent in drug-level-confirmed evidence. These data are not robust evidence of independent harm and do not justify routine preoperative discontinuation or delay of necessary surgery.
INTRODUCTION: Whether preoperative advanced therapy increases postoperative complications in inflammatory bowel disease (IBD) is contested. We conducted a cross-class systematic review distinguishing crude from confounding-adjusted estimates.
METHODS: Five databases were searched to May 2026 for studies of patients with IBD undergoing abdominal surgery. Random-effects meta-analysis estimated comparative odds ratios (ORs) with 95% confidence intervals (CIs) versus unexposed patients across 13 pre-specified outcomes. Confounding-adjusted estimates were pooled separately by inverse-variance, contrasted with crude, meta-regressed on adjustment quality, and rated by GRADE.
RESULTS: Ninety-four publications were included. Across the pooled comparisons, advanced therapies were not associated with a uniform excess of postoperative complications; an isolated deep vein thrombosis signal (OR 2.62) most plausibly reflects disease-severity confounding. Crude anti-tumour necrosis factor (anti-TNF) exposure was associated with higher overall complications (OR 1.14, 95% CI 1.02-1.28) and, most reproducibly, intra-abdominal abscess (1.18, 1.01-1.39; 1.34 good-quality); infliximab with sepsis (2.38, 1.30-4.38). Infectious complications, SSI and anastomotic leak were non-significant, and signals were confined to Crohn's disease. After confounding adjustment, anti-TNF odds remained elevated (overall 1.53, 1.12-2.08, k = 18; infectious 1.51, 1.18-1.92) but were heterogeneous, attenuated with adjustment depth to non-significance in the most-adjusted stratum, and null in the only study confirming exposure by measured serum drug levels rather than a preoperative calendar window (infection 1.05). Vedolizumab showed no complication increase but higher readmission (1.70) and reoperation (1.65), consistent with channelling bias.
CONCLUSIONS: Although there was a signal for association of advanced-therapy with certain complications, the signal is small, heterogeneous, of very low certainty, and absent in drug-level-confirmed evidence. These data are not robust evidence of independent harm and do not justify routine preoperative discontinuation or delay of necessary surgery.
TRIAL REGISTRATION: This systematic review was prospectively registered with the International Prospective Register of Systematic Reviews (PROSPERO; CRD420261323118).