With the increasing number of pancreaticoduodenectomies performed for benign or low-grade indications, the population of long-term survivors continues to grow, making this clinical situation increasingly common. Pancreaticoduodenectomy (PD) leads to chronic biliary reflux into the stomach. We hypothesised that this chronic exposure may induce sustained gastric inflammation and, over time, promote gastric carcinogenesis.

Aim: To determine whether pancreaticoduodenectomy is a risk factor for gastric adenocarcinoma.

This study had two complementary components:

-Multicenter ambispective endoscopic cohort:All adults who had undergone PD more than 10 years earlier systematically underwent upper gastrointestinal endoscopy.In the prospective cohort, endoscopy was performed as part of the study and included bile reflux, chromoendoscopy, and standardized, staged gastric biopsies with centralized pathological review.In the retrospective cohort, we extracted data from any upper endoscopy performed >10 years after PD, including detailed endoscopic findings and corresponding histopathology.

-Nationwide population-based analysis: using the French nationwide claims database, to estimate the incidence of gastric adenocarcinoma in patients ≥5 years after PD.

The primary endpoint was the diagnosis of gastric adenocarcinoma >10 years after PD. Secondary endpoints were the presence of preneoplastic gastric lesions and the identification of associated risk factors.

A total of 143 patients were included: 50% were male; age at PD was 53±13.3 years; mean follow-up was 16.1±5.4 years. Endoscopic evaluation was prospective in 68 patients (48%) and retrospective in 75 (52%). Indications for PD were malignant tumors in 50% (including 17% neuroendocrine tumors) and benign/low-grade lesions with malignant potential in 40% (including 30% IPMN).

Among them, 10 developed gastric adenocarcinoma (7%), with a median time to diagnosis of 13.5 years. Tumors were located on or adjacent to the gastrojejunal anastomosis and were predominantly poorly differentiated. Symptoms associated with cancer were persistent nausea (p=0.008) and gastrointestinal bleeding (p=0.018), including iron-deficiency anemia.

Preneoplastic lesions were identified in 16% of patients with gastric biopsies (n=126): high-grade dysplasia 1%, low-grade dysplasia 8%, and intestinal metaplasia 7%. These patients were significantly more likely to have hereditary cancer predispositions (Lynch, FAP, MEN, or other; p=0.002) and to report persistent nausea (p=0.01). No significant associations were observed with smoking status, Helicobacter pylori infection, type of pancreatic anastomosis, biliary reflux or pylorus preservation.

In the population-based national analysis, among 9,957 patients who had undergone PD more than 5 years earlier (2009–2018), 0.7% developed gastric cancer (71 patients). All cancers occurred more than 8 years after PD. Two-thirds of the cohort were between 5 and 10 years post-surgery at the time of analysis.

Gastric adenocarcinoma after PD is not a rare event. While the risk appears negligible within the first 10 years, our large cohort shows that beyond 10 years the risk rises sufficiently to justify endoscopic surveillance with systematic staged biopsies. Patients should be informed of key alarm symptoms (persistent nausea, anemia, or gastrointestinal bleeding) which warrant urgent upper endoscopy.