BACKGROUND: Anaphylaxis during anaesthesia is a serious complication for patients and anaesthetists. METHODS: The 6th National Audit Project (NAP6) on perioperative anaphylaxis collected and reviewed 266 reports of Grades 3-5 anaphylaxis over 1 yr from all NHS hospitals in the UK. RESULTS: The estimated incidence was ≈1:10 000 anaesthetics. Case exclusion because of reporting delays or incomplete data means true incidence might be ≈70% higher. The distribution of 199 identified culprit agents included antibiotics (94), neuromuscular blocking agents (65), chlorhexidine (18), and Patent Blue dye (9). Teicoplanin comprised 12% of antibiotic exposures, but caused 38% of antibiotic-induced anaphylaxis. Eighteen patients reacted to an antibiotic test dose. Succinylcholine-induced anaphylaxis, mainly presenting with bronchospasm, was two-fold more likely than other neuromuscular blocking agents. Atracurium-induced anaphylaxis mainly presented with hypotension. Non-depolarising neuromuscular blocking agents had similar incidences to each other. There were no reports of local anaesthetic or latex-induced anaphylaxis. The commonest presenting features were hypotension (46%), bronchospasm (18%), tachycardia (9.8%), oxygen desaturation (4.7%), bradycardia (3%), and reduced/absent capnography trace (2.3%). All patients were hypotensive during the episode. Onset was rapid for neuromuscular blocking agents and antibiotics, but delayed with chlorhexidine and Patent Blue dye. There were 10 deaths and 40 cardiac arrests. Pulseless electrical activity was the usual type of cardiac arrest, often with bradycardia. Poor outcomes were associated with increased ASA, obesity, beta blocker, and angiotensin-converting enzyme inhibitor medication. Seventy per cent of cases were reported to the hospital incident reporting system, and only 24% to Medicines and Healthcare products Regulatory Agency via the Yellow Card Scheme. CONCLUSIONS: The overall incidence of perioperative anaphylaxis was estimated to be 1 in 10 000 anaesthetics.