Abstract Background Recent evidence has suggested an association between postoperative non-steroidal anti-inflammatory drugs (NSAIDs) and increased operation-specific complications. This study aimed to determine the safety profile following gastrointestinal surgery across a multicentre setting in the UK. Methods This multicentre study was carried out during a 2-week interval in September–October 2013. Consecutive adults undergoing elective or emergency gastrointestinal resection were included. The study was powered to detect a 10 per cent increase in major complications (grade III–V according to the Dindo–Clavien classification). The effect of administration of NSAIDs on the day of surgery or the following 2 days was risk-adjusted using propensity score matching and multivariable logistic regression to produce adjusted odds ratios (ORs). The type of NSAID and the dose were registered. Results Across 109 centres, early postoperative NSAIDs were administered to 242 (16·1 per cent) of 1503 patients. Complications occurred in 981 patients (65·3 per cent), which were major in 257 (17·1 per cent) and minor (Dindo–Clavien grade I–II) in 724 (48·2 per cent). Propensity score matching created well balanced groups. Treatment with NSAIDs was associated with a reduction in overall complications (OR 0·72, 95 per cent confidence interval 0·52 to 0·99; P = 0·041). This effect predominately comprised a reduction in minor complications with high-dose NSAIDs (OR 0·57, 0·39 to 0·89; P = 0·009). Conclusion Early use of NSAIDs is associated with a reduction in postoperative adverse events following major gastrointestinal surgery.