Patient-level predictors of safety and recurrence after cold versus hot snare resection of large, non-pedunculated colorectal polyps: an individual patient data meta-analysis of four randomised-controlled trials.

BACKGROUND: Cold resection is standard for small colorectal polyps and sessile serrated lesions (SSLs). For large (≥20 mm) non-pedunculated adenomas, it appears safer than hot snare endoscopic mucosal resection (EMR) but is associated with higher rates of residual/recurrent adenoma (RRA). OBJECTIVE: To pool individual patient data from all available randomised controlled trials (RCTs) and determine whether RRA differs across polyp subgroups. DESIGN: International, multicentre post hoc individual patient data meta-analysis of four RCTs. Primary outcomes were rates of major adverse events (AE) (AE=deep mural injury III-V/perforation and/or delayed bleeding) and RRA. Subgroup analyses assessed outcomes by location, size, morphology and histology. RESULTS: 1509 polyps in 1423 patients (mean age 66.1 years; mean size 31.3 mm) from 50 centres were included. Major AE occurred in 1.7% of cold versus 5.9% of hot snare resections (OR 0.26; p<0.001). Cold snare was the only factor to reduce major AE in multivariate analysis (OR 0.21; 95% CI 0.07 to 0.60; p=0.004). RRA was higher after cold resection (26.5% vs 12.8%; OR 2.61; p<0.001), with a significant interaction by lesion diameter (p=0.035). By histology, differences were the smallest in SSLs (15.0% vs 8.1%; OR 2.13; p=0.053) and greatest in adenomas with high-grade dysplasia (40.8% vs 18.0%; OR 2.88; p=0.002). CONCLUSIONS: Cold resection is safer than hot EMR but results in more than double the RRA in all adenoma subgroups, especially those with high-grade dysplasia. It may be appropriate for SSLs between 2 cm and 4 cm and selected cases under anticoagulant/antiplatelet therapy or multimorbid patients, for which further studies are needed.