Small Bowel Injury on a Drug That’s Working: The NSAID Enteropathy Nobody Was Watching For
A single patient, iron-deficient and asymptomatic, with a small bowel that’s been quietly injured by exactly the drug her stomach was already being protected from. The disagreement is about where the protection actually needs to be aimed.
Doris L., a 74-year-old avid gardener, has kept a quarter-acre vegetable plot going every summer since her husband passed six years ago, kneeling in the dirt most mornings despite knee osteoarthritis that has needed daily diclofenac for the better part of a decade. Her rheumatologist added a PPI alongside it years ago, standard practice for chronic NSAID use, and she has had no heartburn, no dyspepsia, nothing that would have flagged a problem. What brought her in was a routine blood count showing a hemoglobin of 9.8 and a ferritin low enough to confirm iron-deficiency anemia she had no symptoms to explain — no visible bleeding, no change in bowel habits, nothing on colonoscopy or upper endoscopy either.
Capsule endoscopy, ordered specifically because the standard two-scope workup came back clean, found what neither scope could reach: multiple small bowel erosions and two circumferential “diaphragm” strictures in the mid-jejunum, the signature lesion of chronic NSAID small bowel injury. That finding reframes what her PPI has actually been doing for her all these years — protecting her stomach lining from exactly the injury it was designed to prevent, while leaving the small bowel, which prostaglandin-mediated PPI protection was never built to cover, exposed the whole time. Real capsule-endoscopy data comparing regimens head-to-head has found meaningfully less small bowel mucosal injury with a COX-2-selective agent alone than with a nonselective NSAID paired with a PPI: Goldstein and colleagues randomized healthy volunteers to celecoxib, naproxen plus omeprazole, or placebo and counted mucosal breaks at two weeks, finding roughly 3.0 per subject on naproxen-plus-omeprazole against 0.3 on celecoxib. Doris is not that population and it matters here rather than as a footnote — that trial ran two weeks in volunteers with clean baseline capsule studies, while she has taken diclofenac daily for a decade and already has diaphragm strictures, and in chronic users studied past three months Maiden and colleagues found no significant difference in small bowel injury between COX-2-selective and nonselective users at all. Separately, Wallace and colleagues showed PPI co-administration worsens NSAID small-bowel injury by inducing dysbiosis, which suggests her PPI pairing may not have been neutral for the jejunum it never covered. Her anemia is the first hard evidence that whatever has protected her stomach for ten years has not been protecting the rest of her gut at all.
GI clinic, capsule endoscopy results review
I want to switch her to celecoxib and stop the diclofenac. The endoscopic comparison data on this is real and specific to the exact injury we’re looking at — Goldstein’s capsule endoscopy trial showed roughly ten times fewer small bowel mucosal breaks on celecoxib than on naproxen plus omeprazole — and there’s real reason to think her PPI hasn’t just failed to protect her small bowel, it may have made the local injury environment worse by disrupting the small-intestinal microbiome, which is what Wallace showed.
I don’t want to give up a pain regimen that’s genuinely worked for ten years without weighing what we’re trading it for. Diclofenac keeps her out in that garden every morning; I don’t know yet that celecoxib will do the same for her specifically, and at 74 I want her cardiovascular risk on the table too, even though it’s currently low. Misoprostol has real randomized evidence for reducing NSAID small bowel injury directly — Fujimori’s pilot trial, read by capsule endoscopy — and it lets her keep the drug we already know works for her pain.
And I want to be blunt about which trial population that is: Goldstein randomized healthy volunteers with clean baseline capsule studies for two weeks. Doris has been on diclofenac for ten years and already has strictures. Maiden looked at people who’d been on these drugs longer than three months — the group she actually belongs to — and found no significant difference in small bowel injury between COX-2-selective and nonselective users. So "less injury at two weeks in volunteers" and "the right trade for this specific patient, whose pain control already works" are different claims, and misoprostol answers the injury question without forcing the second one.
Both of you are choosing between two ways of adding or changing a drug, and neither of you has asked what her diclofenac dose actually is, or whether she could get the same pain control on less. Chronic daily use at a fixed dose for a decade is exactly the exposure pattern that produces this kind of cumulative injury, and dose reduction is a real lever that doesn’t require picking a side in the switch-versus-add-on debate.
I’d want her actual dose and any history of dose titration in front of us before either plan proceeds — if she’s been on a higher dose than her current pain level actually requires, trialing a reduction, alone or alongside either of your proposals, could lower her ongoing exposure regardless of which drug or add-on she ends up on.
Agreed: pull her full diclofenac dosing history and trial a reduction to the lowest dose that controls her OA symptoms before finalizing either the misoprostol add-on or the celecoxib switch; iron repletion starts today regardless of that decision.
Not agreed: which of the two options is the better fallback if dose reduction alone doesn’t resolve her injury risk. The gastroenterologist still favors the COX-2 switch on the endoscopic-injury data; the primary care physician still favors misoprostol to preserve her working pain regimen.