Effect of obesity on perioperative outcomes following gastrointestinal surgery: meta-analysis
Carolyn Cullinane,Anna Fullard,Stefanie M Croghan,Jessie A Elliott,Christina A Fleming
DOI: https://doi.org/10.1093/bjsopen/zrad026
2023-07-10
BJS Open
Abstract:Background: Obesity can pose perioperative challenges related to obesity-associated co-morbidities and technical factors. However, the true impact of obesity on postoperative outcomes is not well established and reports are conflicting. The aim was to perform a systematic review and meta-analysis to explore the effect of obesity on perioperative outcomes for general surgery procedures in distinct obesity subtypes. Methods: A systematic review was performed for studies reporting postoperative outcomes in relation to BMI in upper gastrointestinal, hepatobiliary and colorectal based on an electronic search using the Cochrane Library, Science Direct, PubMed and Embase up to January 2022. The primary outcome was the incidence of 30-day postoperative mortality among patients with obesity undergoing general surgical procedures in comparison to patients with normal range BMI. Results: Sixty-two studies, including 1 886 326 patients, were eligible for inclusion. Overall, patients with obesity (including class I/II/II) had lower 30-day mortality rates in comparison to patients with a normal BMI (odds ratio (OR) 0.75, 95 per cent c.i. 0.66 to 0.86, P < 0.0001, I2 = 71 per cent); this was also observed specifically in emergency general surgery (OR 0.83, 95 per cent c.i. 0.79 to 0.87, P < 0.0000001, I2 = 7 per cent). Compared with normal BMI, obesity was positively associated with an increased risk of 30-day postoperative morbidity (OR 1.11, 95 per cent c.i. 1.04 to 1.19, P = 0.002, I2 = 85 per cent). However, there was no significant difference in postoperative morbidity rates between the cohorts of patients with a normal BMI and class I/II obesity (OR 0.98, 95 per cent c.i. 0.92 to 1.04, P = 0.542, I2 = 92 per cent). Overall, the cohort with obesity had a higher rate of postoperative wound infections compared with the non-obese group (OR 1.40, 95 per cent c.i. 1.24 to 1.59, P < 0.0001, I2 = 82 per cent). Conclusion: These data suggest a possible 'obesity paradox' and challenge the assumption that patients with obesity have higher postoperative mortality compared with patients with normal range BMI. Increased BMI alone is not associated with increased perioperative mortality in general surgery, highlighting the importance of more accurate body composition assessment, such as computed tomography anthropometrics, to support perioperative risk stratification and decision-making. Registration number: CRD42022337442 (PROSPERO https://www.crd.york.ac.uk/prospero/).