Association between anaesthetic technique and unplanned admission to intensive care after thoracic lung resection surgery: the second Association of Cardiothoracic Anaesthesia and Critical Care (ACTACC) National Audit

Unplanned intensive care admission is a devastating complication of lung resection and is associated with significantly increased mortality. We carried out a two-year retrospective national multicentre cohort study to investigate the influence of anaesthetic and analgesic technique on the need for unplanned postoperative intensive care admission. All patients undergoing lung resection surgery in 16 thoracic surgical centres in the UK in the calendar years 2013 and 2014 were included. We defined critical care admission as the unplanned need for either tracheal intubation and mechanical ventilation or renal replacement therapy, and sought an association between mode of anaesthesia (total intravenous anaesthesia vs. volatile) and analgesic technique (epidural vs. paravertebral) and need for intensive care admission. A total of 253 out of 11,208 patients undergoing lung resection in the study period had an unplanned admission to intensive care in the postoperative period, giving an incidence of intensive care unit admission of 2.3% (95%CI 2.0-2.6%). Patients who had an unplanned admission to intensive care unit had a higher mortality (29.00% vs. 0.03%, p < 0.001), and hospital length of stay was increased (26 vs. 6 days, p < 0.001). Across univariate, complete case and multiple imputation (multivariate) models, there was a strong and significant effect of both anaesthetic and analgesic technique on the need for intensive care admission. Patients receiving total intravenous anaesthesia (OR 0.50 (95%CI 0.34-0.70)), and patients receiving epidural analgesia (OR 0.56 (95%CI 0.41-0.78)) were less likely to have an unplanned admission to intensive care after thoracic surgery. This large retrospective study suggests a significant effect of both anaesthetic and analgesic technique on outcome in patients undergoing lung resection. We must emphasise that the observed association does not directly imply causation, and suggest that well-conducted, large-scale randomised controlled trials are required to address these fundamental questions.