Two Routes Through a Bypassed Gut, and No Evidence Yet to Choose Between Them

A stone lodged in the common bile duct is one of the more routine emergencies in gastroenterology. In most adults it is resolved in twenty minutes with a duodenoscope, a wire and a small cut through the sphincter. In a patient who had a Roux-en-Y gastric bypass a decade earlier, none of that applies: the stomach is a narrow pouch, the duodenum is disconnected, and the papilla sits at the end of a limb of small bowel the standard scope cannot follow.

Getting there means inventing a route. The most widely adopted is EDGE — endoscopic ultrasound-directed transgastric ERCP — in which the endosonographer punctures from the gastric remnant into the excluded stomach, holds the tract open with a lumen-apposing metal stent, and passes a duodenoscope through it. What has never been settled is sequencing. The tract can be created and used the same day, or allowed to mature over weeks as a deliberate two-stage procedure. A meta-analysis published on 19 October in Gastrointestinal Endoscopy, the journal of the American Society for Gastrointestinal Endoscopy, set out to decide between them and found what a field this young usually finds: no clear winner.

Six outcomes, six intervals that all cross one

The authors pooled seven observational studies covering 346 patients, 47 per cent treated in a single session. Mean age was 58, roughly a quarter were men, and about 80 per cent of procedures were for biliary rather than pancreatic disease.

Nothing separated the two approaches. Technical success came out at an odds ratio of 0.34, clinical success at 1.88, stent dislodgement or migration at 1.70, overall adverse events at 1.45, a persistent fistula at 1.00 and weight gain at 1.25. Every one of those intervals spans 1, from 0.03–3.53 for technical success to 0.38–9.17 for clinical success. The point estimates do not even point the same way: single-session EDGE looks worse on technical success, better on clinical success, worse on migration.

Weight gain is the outcome that belongs specifically to this population: a fistula left patent gives ingested food a second route into the excluded stomach and duodenum, which is why the review counted it at all. The pooled estimate is a shrug — 1.25, with an interval running from a quarter the odds to nearly five times the odds. That is not evidence that regain is a non-issue; it is evidence that seven small series could not measure it.

The alternative is a long, slow fishing expedition

EDGE is not the first thing tried everywhere. Double-balloon-assisted ERCP uses the existing anatomy instead of making a new hole, and its limits were quantified prospectively at one US tertiary centre this year. Of 94 bypass patients undergoing ERCP, 42 were analysed after triage and 31 of those — 73.8 per cent — were completed. Failure was predictable in real time: more than 23 minutes to reach the jejuno-jejunal anastomosis, unusual resistance to scope advancement, failure to reach the ampulla. With two or more present, the model discriminated failure with an area under the curve of 0.93 — the authors read that as a trigger to cross over to EDGE mid-case.

The menu is longer than two items. A narrative review in Diagnostics in June listed the rest: laparoscopic transgastric ERCP, rendezvous procedures, common bile duct exploration, cholangioscopy. All clear the duct more than 80 per cent of the time. Adverse events run from 8 to 22 per cent, length of stay from one to seventeen days, operating time from 100 to 350 minutes. The reviewers’ conclusion was that no accepted algorithm exists — the choice is made by local expertise and by who is on call, not by evidence.

The evidence level matters. The meta-analysis pools case series, not randomised trials. The double-balloon study is a single-centre prospective cohort with a triaged population. The narrative review is the weakest layer — a curated argument, not a systematic search.

A population that keeps growing

None of this would matter if the numbers were static. The NIDDK’s patient information on gallstones states the link directly: obesity raises the risk of stones, fast weight loss raises it further, and weight-loss surgery “may lead to fast weight loss and higher risk of gallstones.” The same material notes that an untreated blockage of the bile or pancreatic duct can be fatal — these patients arrive as emergencies, not elective referrals. The population is also ageing into its complication years: the long horizon the LABS cohort, funded by NIDDK, was built to follow through its study archive, and the reason endoscopy questions like the ones in our piece on reflux and procedure choice keep resurfacing long after the operation.

What the null licenses

A pooled odds ratio of 1.00 for persistent fistula across 346 patients is not reassurance. It is a statement that the comparison has not really been made. Single-session EDGE is faster and cheaper in anaesthetics; two-stage EDGE is gentler on a fresh tract and lets a team close the fistula in a planned second act. Both are defensible. Which one a patient gets is settled by institution, not by data.

The study that would settle it is not exotic: a registry with standard definitions of “single session”, of fistula persistence, of weight change at one and two years, and follow-up long enough to catch a leak that declares itself late. Until then, the honest reading is that the field built an ingenious workaround for an anatomy it created itself, and has not yet done the bookkeeping to know how best to use it.