ObjectiveThis study examined the risk factors associated with failure of enhanced recovery protocol after major hepatobiliary and pancreatic (HBP) surgery.Setting and participantsA retrospective cohort of 194 adult patients undergoing major HBP surgery at a university hospital in Hong Kong was followed up for 30 days. The patients were from a larger cohort study of 736 consecutive adults with preoperative urinary cotinine concentration to examine the association between passive smoking and risk of perioperative respiratory complications and postoperative morbidities.Outcome measuresThe primary outcome was failure of enhanced recovery protocol. This was defined as a composite measure of the following events: intensive care unit (ICU) stay more than 24 h after surgery, unplanned admission to ICU within 30 days after surgery, hospital readmission, reoperation and mortality.ResultsThere were 25 failures of enhanced recovery after HBP surgery (12.9%, 95% CI 8.5% to 18.4%). After adjusting for elective ICU admission, smokers (relative risk (RR ) 2.21, 95% CI 1.10 to 4.46), high preoperative alanine transaminase/glutamic-pyruvic transaminase (RR 3.55,95% CI 1.68 to 7.49) and postoperative morbidities (RR 2.69, 95% CI 1.30 to 5.56) were associated with failures of enhanced recovery in the generalised estimating equation risk model. Compared with those managed successfully, failures stayed longer in ICU (median 19 vs 25 h, p<0.001) and in hospital for postoperative care (median 7 vs 13 days, p=0.003).ConclusionsSmokers and patients having high preoperative alanine transaminase/glutamic-pyruvic transaminase concentration or have a high risk of postoperative morbidities are likely to fail enhanced recovery protocol in HBP surgery programmes.