Introduction: The obesity epidemic coupled with epidemiologic evidence of the link between pancreatic cancer and obesity has raised the interest in the impact of body mass index (BMI) on outcomes for resected pancreatic cancer.

Methods: All patients who underwent pancreatoduodenectomy (PD) for pancreatic adenocarcinoma from 1981 to 2007 were categorized into four groups according to their BMI (<25, 25 to <30, 30 to <35, and ≥35). Associations of these BMI groups with perioperative (operating time, blood loss, complications, in-hospital mortality), pathologic (tumor diameter, tumor stage, differentiation, lymph node status, R0 status) features and long-term patient outcome were evaluated using Kruskal-Wallis and chi-square tests, logistic regression, and Cox proportional hazards regression. A second set of analyses were performed by dichotomizing patients into morbidly obese (BMI ≥ 35) in comparison to the rest.

Results: Of the 586 consecutive patients studied, there were 232 (39.6%) with BMI <25, 232 (39.6%) with BMI 25 to <30, 89 (15.2%) with BMI 30 to <35, and 33 (5.6%) with BMI ≥ 35. Operating time (P = 0.003) and intraoperative blood loss (P < 0.001) increased with BMI, although none of the remaining perioperative features differed significantly among the BMI groups. Similarly, there were no significant associations between BMI group and the pathological features studied, particularly lymph node status (P= 0.98). BMI was not associated with lymph node status even after adjusting for tumor diameter. All analyses were repeated for the morbidly obese. Cox regression did not demonstrate an impact of BMI or morbid obesity on overall or disease-free survival.

Conclusions: BMI (and morbid obesity) does not appear to influence long-term outcomes for patients undergoing PD. Surgeons should be vigilant of the greater risk of perioperative blood loss with increasing BMI.