Proton pump inhibitors in older adults

Drug-class library

Proton pump inhibitors are the clearest example on this site of a gap between what observational studies suggested and what randomized evidence found. They are also, by some distance, the drug most often taken in a long-term care setting without a reason anyone can name.

How common is use without an indication

Among Medicare Part A skilled-nursing admissions to 22 urban, suburban and rural Midwestern US facilities between January 2010 and May 2011, 79.7% were prescribed a proton pump inhibitor and 65.3% of those had no appropriate diagnostic code for it. Reflux was the blanket diagnosis most often recorded, usually without documented follow-up or symptomatic evidence of active disease. Even after counting chronic NSAID, aspirin and warfarin use as valid indications, 24% of all admissions received one with no relevant gastrointestinal diagnosis.

A multicenter observational study of 2,579 residents across 27 Italian nursing homes found 45.6% on a proton pump inhibitor and only 50.7% of those with an evidence-based indication. Facility-level inappropriate use ranged from 22% to 63%, and residents on 13 or more drugs had roughly ten times the odds of receiving one compared with those on four or fewer — a pattern that tracks polypharmacy far more than illness.

A large part of it starts upstream. Of 24,751 intensive care admissions across nine affiliated ICUs, 4,127 were newly started on a proton pump inhibitor and 2,467 (60%) had no long-term indication. Of those, 1,122 (45%) were continued after transfer to the floor and 668 (27%) were discharged on it. Discharge to a rehabilitation facility was itself a risk factor (adjusted odds ratio 2.29, 95% CI 1.62 to 3.24), as was discharge to a nursing home (1.43, 95% CI 1.04 to 1.96). Stress-ulcer prophylaxis becomes a standing prescription in transit.

The harms — and which kind of evidence each rests on

What the randomized evidence found

The COMPASS trial is the only adequately powered randomized safety assessment of long-term proton pump inhibitor therapy. It randomized 17,598 participants with stable cardiovascular and peripheral artery disease to pantoprazole 40 mg daily or placebo, with a median follow-up of 3.01 years and 53,152 patient-years, collecting pneumonia, C. difficile and other enteric infections, fractures, gastric atrophy, chronic kidney disease, diabetes, COPD, dementia, cardiovascular disease, cancer, hospitalizations and all-cause mortality every six months.

There was no statistically significant difference between groups on any safety outcome except enteric infections — 1.4% against 1.0%, odds ratio 1.33 (95% CI 1.01 to 1.75). For C. difficile specifically, events were roughly twice as common on pantoprazole but there were only 13 events in total and the difference was not statistically significant. The authors concluded that pantoprazole “is not associated with any adverse event when used for 3 years, with the possible exception of an increased risk of enteric infections.”

The limit that goes with it: median follow-up was three years, so the trial cannot exclude harms with a longer latency — which is the most relevant caveat for dementia and kidney disease.

What the observational studies reported

These are associations. None of them establishes causation, and the effect sizes are mostly small.

Reported harm	Design	Association
Hip fracture	Nested case-control, 13,556 cases / 135,386 controls aged over 50	More than one year of use: adjusted odds ratio 1.44 (95% CI 1.30–1.59); long-term high-dose 2.65 (1.80–3.90)
C. difficile infection	Meta-analysis of 42 observational studies, ~313,000 participants	Odds ratio 1.74 (95% CI 1.47–2.85), I² = 85%
Community-acquired pneumonia	Meta-analysis of 26 studies, 226,769 cases	Odds ratio 1.49 (95% CI 1.16–1.92), I² = 99.2%; highest in the first 30 days
Hypomagnesemia	Meta-analysis of 9 observational studies, 109,798 patients	Relative risk 1.43 (95% CI 1.08–1.88)
Vitamin B12 deficiency	Case-control, 25,956 cases / 184,199 controls	Two or more years of use: odds ratio 1.65 (95% CI 1.58–1.73)
Chronic kidney disease	Two population cohorts (n = 10,482 and 248,751)	Adjusted hazard ratio 1.50 (95% CI 1.14–1.96); versus an active comparator 1.39 (1.01–1.91)
Dementia	Claims cohort, 73,679 participants aged 75 and over	Hazard ratio 1.44 (95% CI 1.36–1.52) — but only 4% were regular users, and they were older

A formal review applying causation criteria to this literature concluded that “evidence is inadequate to establish causal relationships between PPI therapy and many of the proposed associations,” and that “residual confounding related to study design and the overextrapolation of quantitatively small estimates of effect size have probably led to much of the current controversy about PPI safety” — causing “unnecessary concern among patients and prescribers.” The authors add that patients with a proven indication “should continue to receive it in the lowest effective dose.”

What the FDA-approved label says

Notably, the label itself describes the fracture evidence as observational: “Several published observational studies suggest that PPI therapy may be associated with an increased risk for osteoporosis-related fractures of the hip, wrist, or spine… Patients should use the lowest dose and shortest duration of PPI therapy appropriate to the condition being treated.”

It also states that daily long-term use beyond about three years may lead to cyanocobalamin (vitamin B12) malabsorption or deficiency; that hypomagnesemia “has been reported rarely in patients treated with PPIs for at least three months, and in most cases after a year of therapy,” with serious events including tetany, arrhythmias and seizures; and that fundic gland polyp risk increases with use beyond one year.

When a proton pump inhibitor should be continued

This is the half of the subject that gets lost. The following are sourced continuation indications, not exceptions to be argued around.

• Erosive esophagitis, LA grade C or D: the American College of Gastroenterology recommends maintenance therapy indefinitely or antireflux surgery — a strong recommendation on moderate evidence.
• Barrett’s esophagus: long-term therapy if symptomatic; consider long-term therapy even if asymptomatic.
• Reflux disease with acid-related complications — erosive esophagitis or peptic stricture — for short-term healing, maintenance of healing and long-term symptom control.
• Continuing NSAIDs in a patient at high risk of ulcer-related bleeding.
• A documented history of bleeding gastrointestinal ulcer — an explicit exclusion from the deprescribing guideline.

Two further points from the same guidance are useful in a facility. In patients on clopidogrel with grade C or D esophagitis or inadequately controlled symptoms, “the established benefits of PPI treatment outweigh their proposed but highly questionable cardiovascular risks.” And both the gastroenterology societies advise against routine monitoring driven by the harm literature — no routine screening of bone mineral density, creatinine, magnesium or B12, and no routine supplementation beyond the recommended dietary allowance, absent other risk factors.

The deprescribing guideline, and what it actually says

The 2017 evidence-based guideline covers adults over 18 with upper gastrointestinal symptoms who have completed at least a four-week course resulting in resolution of those symptoms. For them it makes a strong recommendation on low-quality evidence to decrease the daily dose or stop and change to on-demand use, and a weak recommendation to consider an H2 receptor antagonist as an alternative.

It also names situations where deprescribing follows completion of a short-term indication: H. pylori eradication, ICU stress-ulcer prophylaxis, and uncomplicated peptic ulcer disease without ongoing chronic NSAID use.

The reason the first recommendation is strong despite low-quality evidence is specific: low-dose therapy did not produce significantly more relapses than standard dose (relative risk 1.16, 95% CI 0.93 to 1.44), while on-demand use and stepping down to an H2 blocker did increase relapse risk.

Rebound — real, and more qualified than usually stated

A randomized, double-blind, placebo-controlled trial of 120 healthy volunteers — not patients — found that after eight weeks of esomeprazole followed by four weeks of placebo, 44% reported at least one clinically relevant acid-related symptom in weeks 9 to 12, against 15% of those on placebo throughout.

But the deprescribing guideline itself is careful about what that means clinically: “Differentiating ‘rebound hypersecretion’ from symptoms of an underlying disorder such as GERD is challenging. While studies of healthy volunteers taking PPIs have resulted in acid-related symptoms following deprescribing, the clinical significance remains unknown.” On tapering it says its search “did not identify trials that adequately addressed optimal tapering approaches,” that there is “very low-quality evidence” that abrupt discontinuation increases symptom relapse, and that it “might be prudent” to reduce to the lowest effective dose first.

So: rebound is a real, randomized finding in healthy volunteers, and the claim that a taper is required to avoid it in long-term patients rests on very low-quality evidence. Saying that accurately is better than overstating it.

It is a trade-off, not a free win

The Cochrane review behind the guideline pooled six trials covering 1,758 participants. On-demand use may increase lack of symptom control — relative risk 1.71 (95% CI 1.31 to 2.21), low certainty, favoring continuous therapy — while producing a clinically significant reduction in pill burden of 3.79 fewer tablets per week (95% CI −4.73 to −2.84), moderate certainty. There was low-certainty evidence of reduced patient satisfaction with on-demand use.

One further caution about applying that evidence here: participants in five of the six trials were in their late forties to fifties. Only one trial had a mean age of 73. This evidence is only indirectly applicable to the frail older adults this site is about, and that limitation should not be quietly dropped.

The guideline does supply a follow-up schedule: reassess symptom control at four weeks after deprescribing, and at twelve weeks assess symptoms, frequency of on-demand use, and whether further investigation or a return to continuous treatment is needed.

Common questions

Are proton pump inhibitors dangerous?

The large randomized safety trial — 17,598 participants over a median of three years — found no significant excess of fracture, pneumonia, kidney disease, dementia, diabetes, cancer or death, with a modest excess of enteric infections as the only signal. The alarming figures in circulation come from observational studies with substantial confounding. The stronger argument for review is not danger; it is that a majority of people on one in long-term care have no recorded reason to be. Polypharmacy in long-term care.

Who should stay on one?

People with grade C or D erosive esophagitis, Barrett’s esophagus, a documented history of bleeding gastrointestinal ulcer, peptic stricture, or continuing NSAID use with high bleeding risk. Those are sourced continuation indications, and the deprescribing guideline explicitly excludes several of them from its scope. The medication review.

Back to the drug-class library.

References

1. Patterson Burdsall D, Flores HC, Krueger J, et al. Use of proton pump inhibitors with lack of diagnostic indications in 22 Midwestern US skilled nursing facilities. Journal of the American Medical Directors Association. 2013;14(6):429–432. PMID 23583000 · DOI 10.1016/j.jamda.2013.01.021.
2. Pasina L, Novella A, Elli C, Nobili A, Ianes A. Overuse of proton pump inhibitors in nursing homes: An Italian multicenter observational study. Pharmacoepidemiology and Drug Safety. 2020;29(4):461–466. PMID 31990131 · DOI 10.1002/pds.4963.
3. Blackett JW, Faye AS, Phipps M, Li J, Lebwohl B, Freedberg DE. Prevalence and Risk Factors for Inappropriate Continuation of Proton Pump Inhibitors After Discharge From the Intensive Care Unit. Mayo Clinic Proceedings. 2021;96(10):2550–2560. PMID 33308869 · DOI 10.1016/j.mayocp.2020.07.038.
4. Moayyedi P, Eikelboom JW, Bosch J, et al. Safety of Proton Pump Inhibitors Based on a Large, Multi-Year, Randomized Trial of Patients Receiving Rivaroxaban or Aspirin. Gastroenterology. 2019;157(3):682–691.e2. PMID 31152740 · DOI 10.1053/j.gastro.2019.05.056.
5. Yang YX, Lewis JD, Epstein S, Metz DC. Long-term proton pump inhibitor therapy and risk of hip fracture. JAMA. 2006;296(24):2947–2953. PMID 17190895 · DOI 10.1001/jama.296.24.2947.
6. Kwok CS, Arthur AK, Anibueze CI, Singh S, Cavallazzi R, Loke YK. Risk of Clostridium difficile infection with acid suppressing drugs and antibiotics: meta-analysis. The American Journal of Gastroenterology. 2012;107(7):1011–1019. PMID 22525304 · DOI 10.1038/ajg.2012.108.
7. Lambert AA, Lam JO, Paik JJ, Ugarte-Gil C, Drummond MB, Crowell TA. Risk of community-acquired pneumonia with outpatient proton-pump inhibitor therapy: a systematic review and meta-analysis. PLoS One. 2015;10(6):e0128004. PMID 26042842 · DOI 10.1371/journal.pone.0128004.
8. Lam JR, Schneider JL, Zhao W, Corley DA. Proton pump inhibitor and histamine 2 receptor antagonist use and vitamin B12 deficiency. JAMA. 2013;310(22):2435–2442. PMID 24327038 · DOI 10.1001/jama.2013.280490.
9. Lazarus B, Chen Y, Wilson FP, et al. Proton Pump Inhibitor Use and the Risk of Chronic Kidney Disease. JAMA Internal Medicine. 2016;176(2):238–246. PMID 26752337 · DOI 10.1001/jamainternmed.2015.7193.
10. Gomm W, von Holt K, Thomé F, et al. Association of Proton Pump Inhibitors With Risk of Dementia: A Pharmacoepidemiological Claims Data Analysis. JAMA Neurology. 2016;73(4):410–416. PMID 26882076 · DOI 10.1001/jamaneurol.2015.4791.
11. Vaezi MF, Yang YX, Howden CW. Complications of Proton Pump Inhibitor Therapy. Gastroenterology. 2017;153(1):35–48. PMID 28528705 · DOI 10.1053/j.gastro.2017.04.047.
12. Wyeth Pharmaceuticals LLC. PROTONIX (pantoprazole sodium) delayed-release tablets — FDA-approved prescribing information. DailyMed, U.S. National Library of Medicine. 2026. https://dailymed.nlm.nih.gov/dailymed/lookup.cfm?setid=08098cb2-c048-4640-f387-6beec4a38936.
13. Farrell B, Pottie K, Thompson W, et al. Deprescribing proton pump inhibitors: Evidence-based clinical practice guideline. Canadian Family Physician. 2017;63(5):354–364. PMID 28500192. No DOI is assigned to this article; cite by PMID/PMCID.
14. Reimer C, Søndergaard B, Hilsted L, Bytzer P. Proton-pump inhibitor therapy induces acid-related symptoms in healthy volunteers after withdrawal of therapy. Gastroenterology. 2009;137(1):80–87. PMID 19362552 · DOI 10.1053/j.gastro.2009.03.058.
15. Boghossian TA, Rashid FJ, Thompson W, et al. Deprescribing versus continuation of chronic proton pump inhibitor use in adults. Cochrane Database of Systematic Reviews. 2017;3(3):CD011969. PMID 28301676 · DOI 10.1002/14651858.CD011969.pub2.
16. Freedberg DE, Kim LS, Yang YX. The Risks and Benefits of Long-term Use of Proton Pump Inhibitors: Expert Review and Best Practice Advice From the American Gastroenterological Association. Gastroenterology. 2017;152(4):706–715. PMID 28257716 · DOI 10.1053/j.gastro.2017.01.031.
17. Katz PO, Dunbar KB, Schnoll-Sussman FH, Greer KB, Yadlapati R, Spechler SJ. ACG Clinical Guideline for the Diagnosis and Management of Gastroesophageal Reflux Disease. The American Journal of Gastroenterology. 2022;117(1):27–56. PMID 34807007 · DOI 10.14309/ajg.0000000000001538.
18. Lavan AH, O’Mahony D, Gallagher P. STOPPFrail (Screening Tool of Older Persons’ Prescriptions in Frail adults with a limited life expectancy) criteria: application to a representative population awaiting long-term nursing care. European Journal of Clinical Pharmacology. 2019;75(5):723–731. PMID 30685856 · DOI 10.1007/s00228-019-02630-3.

Bibliographic records on this page were retrieved from PubMed. Every reference links to its DOI or PubMed record.

Written and medically reviewed by Dr. Gurpreet Singh Padda, MD, MBA, MHP, founder and medical director of DWARAA. Last reviewed August 21, 2026. This page is educational. DWARAA does not prescribe, deprescribe, diagnose or treat.