Fifteen Years of Reflux Control Against an Unexplained Decline
A woman's CKD keeps progressing without a clear driver, while she's taken the same reflux medication for fifteen years. Newer analysis has reopened whether that medication is a real, modifiable contributor — but no trial has tested stopping it to find out.
A question her primary care physician raised before today's visit, not one Beatrice O. brought herself, is what this conversation is actually about. She is a 68-year-old woman who has taught watercolor classes out of a converted sunroom for over a decade, work she picked up after her reflux — bad enough in her fifties to wake her most nights — finally settled once omeprazole entered her routine fifteen years ago and never left it. Her CKD, stage 3b, has been under nephrology's watch for three years, and her eGFR has continued a slow decline from 42 to 33 over the past two years without a clear driver — her diabetes is well-controlled, her blood pressure is at goal, and no other cause has surfaced on repeated workup. Her primary care physician, reviewing the chart before today's visit, raised a question she hadn't considered before: whether the omeprazole itself might be contributing.
That question has real literature behind it, evolving rather than settled. Lazarus and colleagues' 2016 cohort analysis first linked PPI use to increased CKD incidence and progression, association data that drew real methodological criticism at the time for potential confounding by indication — sicker patients with more comorbidities are simply more likely to be on a PPI for other reasons. Klatte and colleagues answered that criticism directly in the Swedish SCREAM cohort by comparing new PPI users against new users of an H2 blocker — people being treated for the same problem, so the indication itself no longer separates the groups — and still found PPI initiation associated with faster CKD progression, though they were careful to call the effect modest and residual confounding still possible. Where that leaves Beatrice specifically is narrower than the headline suggests. Almabruk and colleagues' 2026 meta-analysis, pooling fifteen of these studies, found the association strongest for incident CKD in people who did not have it before, weaker for progression to end-stage disease, and — on progression of established disease, which is the only endpoint that describes her — not statistically significant at all. The question her physician is raising rests on the part of this literature that speaks least directly to a woman who already has stage 3b disease.
Fifteen years of one drug, and a decline nothing else explains
I'd deprescribe the omeprazole. Lazarus's original cohort data raised this association, and Klatte's SCREAM analysis answered the obvious objection to it — comparing new PPI users against new H2-blocker users, so both groups are people someone thought needed acid suppression — and still found faster CKD progression in the PPI arm. Her decline has no other explanation despite good diabetes and blood pressure control — removing a plausible, modifiable contributor seems like the reasonable next step.
I'd want real caution before stopping fifteen years of what has genuinely worked — her GERD was severe enough to disrupt her sleep for years before this drug controlled it. An active-comparator design fixes confounding by indication; it doesn't make the study a trial, and no randomized trial has actually tested stopping a PPI to see whether CKD progression slows. I'd also press on which endpoint we're borrowing from: Almabruk and colleagues' pooled 2026 meta-analysis finds the signal for incident CKD, and loses significance on progression of established disease — which is the only thing Beatrice has. We'd be trading guaranteed symptom recurrence for the weakest cell in that table.
I take the strengthened association seriously — my hesitation is specifically about acting on it as though it were interventional evidence, which it still isn't.
We don't have to choose between full continuation and stopping outright. Step her down to the lowest effective dose, or trial an H2 blocker in its place, and track both her reflux symptoms and her eGFR trajectory directly over the next several months. That actually tests the hypothesis for Beatrice specifically, rather than deciding from population data alone in either direction — if her decline slows, that's real individual evidence; if her symptoms return badly, we have a clear reason to go back.
Agreed: omeprazole stopped and famotidine started as a structured trial, with GERD symptoms tracked weekly by patient report and eGFR rechecked at three months to see whether her trajectory actually changes.
Not resolved, and explicitly acknowledged as unresolvable by this one case: whether the population-level PPI-CKD association genuinely reflects causation for patients broadly remains an open scientific question this individual trial cannot settle, even if it does inform Beatrice's own care directly.