Design: Randomised, open-label trial at two high-volume centres
Population: 158 adults undergoing pancreatoduodenectomy, France
Primary outcome: Delayed gastric emptying within 90 days
Effect: 38.0% side-to-side vs 30.4% end-to-side; RR 1.25 (95% CI 0.81–1.94)

The IPAD trial randomised adults having pancreatoduodenectomy at two high-volume French pancreatic centres, during the operation, to side-to-side or end-to-side gastrojejunostomy. The primary endpoint was delayed gastric emptying within 90 days.

Among 158 patients, delayed gastric emptying occurred in 38.0% after side-to-side and 30.4% after end-to-side reconstruction (risk ratio 1.25, 95% CI 0.81–1.94). Pancreatic and biliary fistula, reintervention, mortality, nutrition and quality of life at 90 days were similar.

Retrospective series had suggested side-to-side reconstruction reduced delayed emptying. This first randomised test did not support it, and the point estimate leaned the other way. The trial was modest and open-label, so a small difference cannot be excluded.

In practice, the gastrojejunostomy configuration is not the lever for delayed gastric emptying. Surgeons can use the technique they do reliably, and look to other factors — fistula prevention, early feeding protocols — instead.

Use the gastrojejunostomy configuration you perform reliably; neither reduced delayed emptying.
Do not change a stable technique on the strength of retrospective series.
Look to pancreatic fistula prevention and feeding protocols to reduce delayed gastric emptying.
Audit delayed gastric emptying using the international (ISGPS) definition.

Why it matters

It removes a technical choice that retrospective data had made seem important.

The statistics, in plain English

The confidence interval runs from 0.81 to 1.94, so the trial cannot exclude a small benefit for either technique, but it makes a large benefit for side-to-side reconstruction unlikely.