PANCREATODUODENECTOMY FOLLOWING ROUX-EN-Y GASTRIC BYPASS: OPERATIVE CONSIDERATIONS AND OUTCOMES
Maxwell T. Trudeau*1,2, Laura Maggino2, Brett L. Ecker3, Charles Vollmer3, The Pancreatoduodenectomy after Roux-en-Y Gastric Bypass Study Group2
1Department of Surgery, Perelman School of Medicine, Philadelphia, PA; 2Department of Surgery, University of Verona, Verona, Italy; 3Department of Surgery, Hospital of the University of Pennsylvania, Philadelphia, PA
Intro:
Roux-en-Y bypass (RYGB) has historically been the most applied operation for morbid obesity. Some patients will ultimately require pancreatoduodenectomy (PD) for a spectrum of periampullary pathologies. It is suspected that the altered anatomy of RYGB reconstruction influences intraoperative decisions, and contributes to worse outcomes in patients who subsequently require PD.
Methods:
A multi-national (4), multi-center (28) collaborative of 55 pancreatic surgeons who have performed PD following RYGB (2005-2018) was created. Demographics, operative details and perioperative outcomes (including Postoperative Morbidity Index-PMI) from this cohort were analyzed and compared in a propensity-scored matched analysis (Sex, Age, BMI, ASA Class) to a multi-center cohort of 5,533 PDs without prior RYGB. Multivariate regression identified Roux anatomy factors that dictated PD reconstructive variations.
Results:
94 open PDs were performed on RYGB patients (52% MIS; 70% stomach transected from pouch; median-10 years prior/100 lb. weight loss). Pathologic indications between the RYGB anatomy and normal anatomy cohorts did not differ (p=0.133; malignancy ≈ 65%). Of the14 distinct reconstructions employed (of 32 options), 4 accounted for 64%, and none demonstrated superior outcomes. The most common reconstruction (25%) removed the remnant stomach, used the original biliopancreatic (BP) limb, and attached the PJ most distally, followed by the HJ. Two-thirds of the cases used the original BP limb for reconstruction - most commonly placed trans-mesenteric (69%). There were no appreciable outcome benefits vs. those where a secondary Roux limb was created for BP reconstitution. The only significant factor driving surgeons to pursue reconstruction requiring a new Roux was a short original BP limb (OR 8.17; 95% CI 1.68-39.60). Remnant stomachs were removed in 45% - paradoxically more often when still in continuity with the Roux pouch (p=0.046), with no outcome differences between resected and retained stomachs. Venting G-tubes were used in 36% of retained stomachs without outcome benefits. J-tubes were used infrequently (12%). In the risk-matched analysis, RYGB patients showed no significant differences in major postop outcomes (Table) including: Any/Severe complications, mortality, length of stay, pancreatic fistula, DGE, reoperation and postop disposition. The PMI was equivalent for all patients (0.175 RYGB vs. 0.204 traditional PD; p=0.255), and for those suffering complications (0.288 vs 0.327; p=0.173). However, ICU use, transfusions, and readmission rates were higher in the RYGB cohort.
Conclusion:
PD after RYGB is an infrequently encountered, unique and challenging scenario for any given surgeon. These patients do not suffer higher morbidity than those with unaltered anatomy. Various technical reconstructive options do not appear to confer distinct benefits.
A Risk Matched Comparison of Clinical Postoperative Course between RYGB Anatomy and Normal Anatomy for Pancreatoduodenectomy
| RYGB Anatomy (N=94) | Normal Anatomy (N=93) | ||||
| Clinical Outcome Metric | Frequency | % | Frequency | % | P-Value |
| Operative Time (min), median (IQR) | 353 | (258.50-439.50) | 334 | (285.00-403.00) | 0.200 |
| Estimated Blood Loss (ml), median (IQR) | 400 | (250.00-700.00) | 300 | (188.00-600.00) | 0.266 |
| Any complication (Accordion ≥1) | 57 | (60.6) | 58 | (62.4) | 0.808 |
| Severe complication (Accordion ≥3) | 15 | (16) | 21 | (22.6) | 0.251 |
| Clinically Relevant Pancreatic Fistula | 3 | (3.2) | 9 | (9.7) | 0.070 |
| Clinically Relevant Delayed Gastric Emptying | 5 | (5.4) | 2 | (12.5) | 0.283 |
| Mortality, 90 days | 3 | (3.2) | 2 | (2.2) | 0.369 |
| Postoperative Morbidity Index, mean (SD) | 0.175 | (0.22) | 0.204 | (0.23) | 0.255 |
| Length of Stay, median (IQR) | 8.00 | (6.00-11.00) | 8.00 | (7.00-10.00) | 0.402 |
| Discharge Disposition to Home | 65 | (69.9) | 47 | (77) | 0.329 |
| Readmission | 23 | (24.5) | 10 | (11.1) | 0.018 |
| Reoperation | 2 | (2.1) | 3 | (3.2) | 0.642 |
| Intensive care unit used | 44 | (46.8) | 8 | (12.7) | <0.001 |
| TEN or TPN Use | 24 | (25.5) | 15 | (23.8) | 0.807 |
| Transfusion (Intra-op or Post-op) | 32 | (34) | 12 | (19.4) | 0.046 |
| Percutaneous Drain Placement | 13 | (13.8) | 8 | (8.6) | 0.258 |
Back to 2019 Abstracts
