Preoperative Predictors for Complications After Pancreaticoduodenectomy: Impact of BMI and Visceral Fat Mass
Michael G. House*, Yuman Fong, Mladjan Protic, Rohit Sharma, Mithat Gonen, Sara Olson, Robert C. Kurtz, Murray F. Brennan, Peter J. Allen
Surgery, Memorial Sloan-Kettering Cancer Center, New York, NY
Background: Approximately one-half of patients undergoing pancreaticoduodenectomy (PD) experience postoperative complications. We examined preoperative patient and radiographic factors for association with operative morbidity.
Methods: Retrospective review of a prospective pancreatic cancer database identified 356 consecutive patients who underwent PD for pancreatic adenocarcinoma between 2000 and 2005. Complication data were maintained according to a previously described grading/scoring system. Patient co-morbidities were identified from patient questionnaire and chart review. Blinded review of all preoperative cross-sectional abdominal imaging was conducted for pancreatic duct size and pancreatic atrophy. Retrorenal visceral fat thickness (VF) was measured as an indicator of central obesity. VF was defined as the distance between the posterior capsule of the left kidney and the innermost layer of the posterior abdominal wall at the renal hilum. Patient and radiographic characteristics, and their association with postoperative complications were assessed using univariate and multivariate analysis.
Results: Postoperative complications developed in 135 patients (38%), pancreatic fistula/abscess (15%), wound infection (14%), and delayed gastric emptying (4%). On multivariate analysis, the only patient or radiographic characteristics associated with postoperative complications were BMI, pancreatic atrophy, and the extent of central obesity (VF) (Table). Obesity, defined as BMI ≥ 30 kg/m2, was present in 21% of patients. Wound infections occurred in 21% of the obese patients compared to 12% of the non-obese patients, p=0.041. BMI was not associated with overall complications or pancreatic fistula after PD. VF greater than 2 cm was present in 123 patients (35%). Overall complications occurred in 51% of patients with VF ≥ 2 cm , compared to 31% of patients with VF < 2 cm, p<0.001. Postop pancreatic fistula developed in 24% of patients with VF ≥ 2 cm and only 10% of patients with VF < 2 cm, p=0.01. Pancreatic duct size and the presence of gland atrophy on preoperative CT or MRI scan did not correlate with postop pancreatic fistula.
Conclusions: Generalized obesity was associated with significantly higher rates of postoperative wound infections following PD, and patients with central obesity experienced significantly higher rates of overall complications and pancreatic fistula.
BMI (kg/m2) | p-value | Visceral fat (cm) | p-value | Panc atrophy (% of pts) | p-value | |
Any comp Yes (135) No (221) | 26.9 ± 4.8 25.9 ± 4.9 | 0.07 | 2.1 ± 1.2 1.6 ± 1.0 | 0.001 | 47.5 59.2 | 0.05 |
Panc fist Yes (54) No (302) | 27.3 ± 5.3 26.1 ± 4.8 | 0.1 | 2.1 ± 1.0 1.7 ± 1.1 | 0.018 | 45.8 56.3 | 0.2 |
Wound infn Yes (49) No (307) | 27.7 ± 4.9 26.1 ± 4.8 | 0.04 | 2.1 ± 1.3 1.8 ± 1.0 | 0.14 | 55.8 54.5 | 0.9 |
DGE Yes (14) No (352) | 26.1 ± 2.7 26.3 ± 4.9 | 0.8 | 1.7 ± 0.8 1.8 ± 1.1 | 0.7 | 50.0 54.8 | 0.7 |
2007 Program and Abstracts | 2007 Posters