Design: Randomised, open-label trial, 2 high-volume French centres
Population: 158 adults undergoing pancreatoduodenectomy for benign or malignant disease
Primary outcome: Delayed gastric emptying within 90 days
Effect: 38.0% vs 30.4%; RR 1.25 (95% CI 0.81 to 1.94), P = 0.31

Delayed gastric emptying is among the commonest problems after pancreatoduodenectomy, and retrospective series had suggested the shape of the gastrojejunostomy matters. The IPAD trial tested this at two high-volume French centres, randomising 158 patients in theatre to a side-to-side or an end-to-side anastomosis.

Delayed gastric emptying occurred in 38.0% after side-to-side and 30.4% after end-to-side, a difference that was not statistically significant and, if anything, pointed the other way from the hypothesis. Pancreatic and biliary fistula, reintervention, mortality, nutrition and quality of life at 90 days were all similar.

The practical message is that surgeons can keep the configuration they do most reliably. Effort to reduce delayed gastric emptying is better spent on fistula prevention, early recognition of intra-abdominal collections and nutrition. The trial was open-label and moderately sized, and it was run in high-volume units.

Use the gastrojejunostomy configuration your unit performs most consistently; this trial found no reason to switch.
When delayed gastric emptying appears, look for a pancreatic fistula or collection before attributing it to the anastomosis.
Plan nutritional support, including feeding jejunostomy or nasojejunal access where your unit uses it, for patients at high risk of slow emptying.
Record delayed gastric emptying with the ISGPS grading so your own rates can be audited against published figures.

Why it matters

It removes one technical variable from the search for a fix to the commonest morbidity after Whipple.

The statistics, in plain English

A risk ratio of 1.25 with a confidence interval from 0.81 to 1.94 crosses 1.0, so the trial cannot say either technique is better. The interval is wide enough that a modest difference in either direction is still possible; it rules out a large benefit for side-to-side.