What does the research show about proton pump inhibitors?

Proton pump inhibitors

Trials found proton pump inhibitors relieved heartburn and healed erosive esophagitis more often than placebo or H2 blockers, cut rebleeding from ulcers, and prevented NSAID ulcers. They did not improve chronic cough or asthma. Long-term use was linked in observational studies to infections, fractures, kidney disease and low magnesium, but a 3-year trial found only more gut infections.

Add to Bio HackChecked 2026-11-13

Status: Prescription drugs in the US, and omeprazole and esomeprazole are also sold over the counter for frequent heartburn as a 14-day course. FDA-approved uses on the labels cited here are healing and maintenance of erosive esophagitis, heartburn from reflux disease, lowering the risk of gastric ulcers in people taking NSAIDs, Helicobacter pylori eradication with antibiotics, and Zollinger-Ellison syndrome.
Also called: PPIs, PPI, omeprazole, Prilosec, Prilosec OTC, esomeprazole, Nexium, Nexium 24HR, pantoprazole, Protonix, lansoprazole, Prevacid, rabeprazole, Aciphex, acid reducers
Taken for: heartburn, acid reflux, GERD, erosive esophagitis, stomach ulcers, ulcer bleeding, H. pylori, indigestion, stomach protection with NSAIDs or aspirin, stress ulcers in intensive care, throat reflux, chronic cough, asthma with reflux, infant reflux and crying

Strong several reviews agree · Moderate one review or consistent trials · Limited one or a few small studies · Untested no study in people, listed with what it rests on

Shown to help9

ModerateShown to relieve heartburn more often than placebo or H2 blockers in adults with reflux symptoms, with a larger effect when treated without an endoscopy first.
ModerateShown to heal erosive esophagitis more often than placebo or H2 blockers. In one label trial, 92.6% healed on pantoprazole 40 mg by 8 weeks against 39.7% on placebo.Very common
ModerateShown to keep healed esophagitis from returning more often than ranitidine, with 83% to 86% still healed at 12 months on pantoprazole 40 mg against 35% to 37%.Very common
ModerateShown to heal slightly more erosive esophagitis with esomeprazole than with other proton pump inhibitors, about 1 extra healed for every 25 treated, more in severe disease.
ModerateShown to cut rebleeding, surgery and repeat endoscopy after ulcer bleeding, about 1 rebleed prevented per 13 treated, with no overall change in deaths.
ModerateShown to prevent stomach and duodenal ulcers seen on endoscopy in people taking NSAIDs, with gastric ulcers in about 5% on esomeprazole against 12% to 17% on placebo.Common
ModerateShown to lower clinically important upper gut bleeding in critically ill adults, 1.0% against 3.5% in one trial, with no difference in deaths at 90 days.
ModerateShown, as part of eradication therapy with antibiotics, to heal more duodenal ulcers than ulcer-healing drugs alone, with 12.4% of ulcers persisting against 18.7%.
ModerateShown to relieve overall indigestion symptoms slightly more than placebo in functional dyspepsia, with about 11 treated for one extra person helped.

Reported risks16

ModerateShown in label trials to raise headache, diarrhea and dizziness against placebo, with headache in 12.2% on pantoprazole against 8.5% on placebo.Very common
ModerateLinked to low blood magnesium in observational studies, in 19.4% of users against 13.5% of non-users. The label warns of rare serious cases, mostly after a year of use.Very common
ModerateLinked to about twice the odds of small intestinal bacterial overgrowth, found in 36.8% of users studied against 19.9% of controls, rising with longer use.Very common
LimitedOne trial found 44% of healthy volunteers had acid-related symptoms after stopping 8 weeks of a proton pump inhibitor, against 15% after placebo.Very common
ModerateLinked to about 1.7 times the odds of C. difficile infection in observational studies. In a 3-year trial, enteric infections were 1.4% on pantoprazole against 1.0% on placebo.Common
StrongLinked to about 1.3 times the hip fracture risk in observational studies. A 3-year randomized trial found no difference in fractures. The label warns of fracture risk with long-term use.How often is not known
ModerateMixed results: observational studies linked proton pump inhibitors to about 27% higher odds of pneumonia, while a 3-year randomized trial found no difference from placebo.How often is not known
ModerateLinked to about 1.36 times the risk of chronic kidney disease in observational studies, in low-quality evidence. A 3-year randomized trial found no difference from placebo.How often is not known
ModerateMixed results: one 2026 meta-analysis linked use to about 18% higher dementia risk in older adults, two others found no significant link, and a 3-year trial found no difference.How often is not known
ModerateLinked to higher odds of a vitamin B12 deficiency diagnosis with 2 or more years of use, while measured B12 levels mostly did not differ from non-users.How often is not known
ModerateLinked to more fundic gland polyps with use of a year or longer, and to about 1.4 times the risk of stomach cancer in a few observational studies.How often is not known
LimitedThe label warns that acute tubulointerstitial nephritis, a kidney inflammation, has occurred with proton pump inhibitors at any point during use. It gives no rate.How often is not known
LimitedMixed results: a veterans cohort linked proton pump inhibitor use to about 25% higher risk of death than H2 blocker use, while a 3-year randomized trial found no difference.How often is not known
LimitedThe esomeprazole label warns against use with clopidogrel because it reduces clopidogrel's antiplatelet activity. One trial found no rise in heart events with omeprazole.How often is not known
LimitedThe labels warn that proton pump inhibitors can raise warfarin's effect and methotrexate levels and change the levels of some HIV drugs. They give no rate.How often is not known
LimitedThe label warns of severe skin reactions and of cutaneous and systemic lupus reported with proton pump inhibitors, mostly cutaneous lupus. It gives no rate.How often is not known

What people take it for, and what was found

Shown to help9

Moderate

Heartburn relief in reflux disease

Shown to relieve heartburn more often than placebo or H2 blockers in adults with reflux symptoms, with a larger effect when treated without an endoscopy first.

Compared with placebo [1]; H2 blockers and prokinetic drugs [1]

The numbers and sources

Effect: In people treated for reflux symptoms without an endoscopy first, the risk ratio for heartburn remission against placebo was 0.37 (95% CI 0.32 to 0.44) for proton pump inhibitors and 0.77 (95% CI 0.60 to 0.99) for H2 blockers, a value below 1 favoring the drug. Against H2 blockers directly, proton pump inhibitors had a risk ratio of 0.66 (95% CI 0.60 to 0.73). In people with heartburn and a normal endoscopy, the risk ratio against placebo was 0.71 (95% CI 0.65 to 0.78) [1].
Population: 34 randomized trials of short-term treatment in adults with reflux symptoms, 15 treating without endoscopy, 15 in people with a normal endoscopy and 4 in both [1].
Sources: [1]
Caveats: Short-term trials only. The placebo comparison in people treated without endoscopy rests on two trials [1].

ModerateVery common

Healing erosive esophagitis

Shown to heal erosive esophagitis more often than placebo or H2 blockers. In one label trial, 92.6% healed on pantoprazole 40 mg by 8 weeks against 39.7% on placebo.

Compared with placebo [2][3]; H2 blockers [2]; nizatidine [3]

The numbers and sources

How common: Very common: more than 1 in 10 people.
Effect: Standard-dose proton pump inhibitors healed esophagitis more often than placebo (RR 0.22, 95% CI 0.15 to 0.31, a value below 1 favoring the drug) and more often than H2 blockers alone or with prokinetics (RR 0.51, 95% CI 0.44 to 0.59). H2 blockers also beat placebo (RR 0.74) [2]. In a US placebo-controlled trial, 8-week healing was 66.0% on pantoprazole 10 mg, 83.5% on 20 mg and 92.6% on 40 mg against 39.7% on placebo. Against nizatidine 150 mg twice daily, 8-week healing was 79.2% and 82.9% on pantoprazole 20 and 40 mg and 41.4% on nizatidine [3].
Population: 134 trials with 35,978 people with esophagitis, of which 5 trials with 965 people compared a standard dose with placebo [2]; 603 adults with grade 2 or higher esophagitis in the placebo trial and 243 in the nizatidine trial [3].
Sources: [2], [3]
Caveats: Cochrane withdrew the 2007 review in 2011 [2]. The label trials were analyzed per protocol, and the placebo group had 68 people [3].

ModerateVery common

Keeping esophagitis healed

Shown to keep healed esophagitis from returning more often than ranitidine, with 83% to 86% still healed at 12 months on pantoprazole 40 mg against 35% to 37%.

Compared with ranitidine 150 mg twice daily [3]

The numbers and sources

How common: Very common: more than 1 in 10 people.
Effect: In two identical 12-month trials, the share still healed at 12 months was 86% and 83% on pantoprazole 40 mg, 70% and 72% on 20 mg, and 35% and 37% on ranitidine. Daytime heartburn episodes in month 12 averaged 2.9 on pantoprazole 40 mg against 17.5 on ranitidine [3].
Population: 386 and 404 adults with endoscopically healed erosive esophagitis [3].
Sources: [3]
Caveats: The comparison was an H2 blocker, not placebo. Controlled studies did not extend beyond 12 months [3].

Moderate

Esomeprazole against other proton pump inhibitors

Shown to heal slightly more erosive esophagitis with esomeprazole than with other proton pump inhibitors, about 1 extra healed for every 25 treated, more in severe disease.

Compared with other proton pump inhibitors [4]

The numbers and sources

Effect: At 8 weeks, esomeprazole raised the chance of healing by 5% relative to other proton pump inhibitors (RR 1.05, 95% CI 1.02 to 1.08), an absolute difference of 4% and a number needed to treat of 25. By severity grade A to D, the numbers needed to treat were 50, 33, 14 and 8. Symptom relief at 4 weeks was 8% more likely (RR 1.08, 95% CI 1.05 to 1.11) [4].
Population: 10 randomized trials with 15,316 people with erosive esophagitis [4].
Sources: [4]
Caveats: The abstract does not name the comparator drugs or doses one by one. The trials were published from 1995 to 2005. The gap was small in milder disease [4].

Moderate

Rebleeding from stomach and duodenal ulcers

Shown to cut rebleeding, surgery and repeat endoscopy after ulcer bleeding, about 1 rebleed prevented per 13 treated, with no overall change in deaths.

Compared with placebo or H2 blockers [5]

The numbers and sources

Effect: Proton pump inhibitors cut rebleeding (OR 0.49, 95% CI 0.37 to 0.65, number needed to treat 13), surgery (OR 0.61, 95% CI 0.48 to 0.78, number needed to treat 34) and repeat endoscopic treatment (OR 0.32, 95% CI 0.20 to 0.51). Deaths did not differ overall (OR 1.01, 95% CI 0.74 to 1.40). Deaths fell in Asian trials (OR 0.35) and in people with active bleeding or a visible vessel (OR 0.53) [5].
Population: 24 randomized trials with 4,373 people with bleeding peptic ulcers [5].
Sources: [5]
Caveats: Heterogeneity was present for rebleeding. The mortality benefit appeared only in subgroups [5].

ModerateCommon

Preventing ulcers in people taking NSAIDs

Shown to prevent stomach and duodenal ulcers seen on endoscopy in people taking NSAIDs, with gastric ulcers in about 5% on esomeprazole against 12% to 17% on placebo.

Compared with placebo [6][7]

The numbers and sources

How common: Common: between 1 in 100 and 1 in 10 people.
Effect: Proton pump inhibitors lowered the risk of endoscopic gastric ulcers (RR 0.40, 95% CI 0.32 to 0.51) and duodenal ulcers in long-term NSAID users. The review found ulcer complications such as bleeding were shown to fall only with high-dose misoprostol [6]. In two 26-week trials in people aged 60 or older or with a past ulcer, 95.4% and 94.7% stayed free of gastric ulcers on esomeprazole 20 mg against 88.2% and 83.3% on placebo [7].
Population: 40 randomized trials of drugs to prevent NSAID ulcers [6]; 1,429 adults taking NSAIDs, median age 66, in two trials [7].
Sources: [6], [7]
Caveats: Endoscopic ulcers are a stand-in for bleeding ulcers. The esomeprazole trials did not show a drop in duodenal ulcers because there were too few [7].

Moderate

Stress ulcer bleeding in intensive care

Shown to lower clinically important upper gut bleeding in critically ill adults, 1.0% against 3.5% in one trial, with no difference in deaths at 90 days.

Compared with placebo [10][11]

The numbers and sources

Effect: In ventilated adults, clinically important upper gut bleeding occurred in 1.0% on pantoprazole and 3.5% on placebo (HR 0.30, 95% CI 0.19 to 0.47). Deaths at 90 days were 29.1% and 30.9% (HR 0.94, 95% CI 0.85 to 1.04) [10]. In adults with an unplanned ICU admission, clinically important bleeding occurred in 2.5% on pantoprazole and 4.2% on placebo. Deaths at 90 days were 31.1% and 30.4% (RR 1.02, 95% CI 0.91 to 1.13), and a composite of bleeding, pneumonia, C. difficile infection and heart ischemia did not differ [11].
Population: 4,821 adults on invasive ventilation in 68 ICUs [10]; 3,298 adults at risk of gut bleeding in European ICUs [11].
Sources: [10], [11]
Caveats: Both trials used intravenous pantoprazole in critically ill adults. The results do not apply to people outside intensive care.

Moderate

Helicobacter pylori ulcers, as part of eradication therapy

Shown, as part of eradication therapy with antibiotics, to heal more duodenal ulcers than ulcer-healing drugs alone, with 12.4% of ulcers persisting against 18.7%.

Compared with ulcer-healing drugs alone or no treatment [12]

The numbers and sources

Effect: In duodenal ulcer healing, eradication therapy plus an ulcer-healing drug left ulcers persisting in 12.4% against 18.7% on the ulcer-healing drug alone (RR 0.66, 95% CI 0.58 to 0.76), and in 21.7% against 58.5% with no treatment (RR 0.37). Differences for gastric ulcer healing were imprecise (RR 1.23, 95% CI 0.90 to 1.68) [12].
Population: 55 randomized trials in people with peptic ulcer and H. pylori infection [12].
Sources: [12]
Caveats: Eradication regimens combine an acid reducer with antibiotics, so this block does not isolate the proton pump inhibitor. The evidence was rated low to very low quality [12].

Moderate

Functional dyspepsia

Shown to relieve overall indigestion symptoms slightly more than placebo in functional dyspepsia, with about 11 treated for one extra person helped.

Compared with placebo [13]; H2 blockers and prokinetic drugs [13]

The numbers and sources

Effect: Proton pump inhibitors lowered the risk of persisting dyspepsia symptoms against placebo (RR 0.88, 95% CI 0.82 to 0.94), a number needed to treat of 11. Against H2 blockers the difference was not clear (RR 0.88, 95% CI 0.74 to 1.04). Low and standard doses performed similarly [13].
Population: 25 randomized trials with 8,453 participants, 18 trials with 6,172 people for the placebo comparison [13].
Sources: [13]
Caveats: Moderate-quality evidence. Results did not differ by H. pylori status [13].

Mixed results2

Moderate

Stomach protection with aspirin, clopidogrel or blood thinners

Mixed results: omeprazole cut stomach bleeding events in people on aspirin and clopidogrel, while pantoprazole did not cut overall upper gut events in people on aspirin or rivaroxaban.

Compared with placebo [8][9]

The numbers and sources

Effect: In people on aspirin and clopidogrel, upper gut events over 180 days were 1.1% on omeprazole and 2.9% on placebo (HR 0.34, 95% CI 0.18 to 0.63), and overt upper gut bleeding fell (HR 0.13) [8]. In people with stable heart or artery disease on low-dose rivaroxaban, aspirin or both, upper gut events did not differ (102 of 8,791 on pantoprazole against 116 of 8,807 on placebo, HR 0.88, 95% CI 0.67 to 1.15). Bleeding from stomach and duodenal lesions fell (HR 0.52, 95% CI 0.28 to 0.94), with 982 needing treatment to prevent one [9].
Population: 3,761 patients with coronary disease on dual antiplatelet therapy [8]; 17,598 patients with stable cardiovascular or peripheral artery disease, median 3 years [9].
Sources: [8], [9]
Caveats: The first trial ended early with fewer people than planned [8]. The two trials enrolled different populations at different bleeding risk.

Moderate

Throat reflux (laryngopharyngeal reflux)

Mixed results: two meta-analyses found throat reflux symptom scores improved more than on placebo, but disagreed on response rates, and throat exam findings did not differ.

Compared with placebo [14][15]

The numbers and sources

Effect: One meta-analysis found a higher response rate on a proton pump inhibitor (risk difference 0.15, 95% CI 0.01 to 0.30) and better reflux symptom index scores (SMD 1.65), with no difference in the reflux finding score seen on laryngoscopy (SMD 0.62, 95% CI -0.96 to 2.19) [14]. A second found better symptom index scores (SMD 3.65, 95% CI 1.56 to 5.75) but no significant difference in response rate or the reflux finding score [15].
Population: 14 randomized trials with 771 participants [14]; 13 randomized trials with 831 participants [15].
Sources: [14], [15]
Caveats: Small trials with differing definitions of response. The two analyses overlap in the trials they include.

Not shown to help2

Moderate

Chronic cough

Not shown to resolve long-lasting cough in adults more than placebo overall. In infants, more side effects occurred on a proton pump inhibitor.

Compared with placebo [16]

The numbers and sources

Effect: In adults, pooled data showed no significant difference in total cough resolution (OR 0.46, 95% CI 0.19 to 1.15). Cough scores improved only in sensitivity analyses (SMD -0.41). In one infant trial, cough did not improve and adverse events rose (OR 5.56, 95% CI 1.18 to 26.25), one extra event for every 11 infants treated for 4 weeks [16].
Population: 19 studies, 13 in adults, of which 9 compared 2 to 3 months of a proton pump inhibitor with placebo [16].
Sources: [16]
Caveats: Trials differed in how they measured cough. Children's studies could not be combined [16].

Moderate

Asthma in people with reflux

Not shown to consistently improve lung function, asthma symptoms or asthma medicine use in people with asthma and reflux.

Compared with control groups the abstract does not describe [17]

The numbers and sources

Effect: Anti-reflux treatment did not consistently improve lung function, asthma symptoms, night-time asthma or use of asthma medicines. The authors suggested some subgroups may benefit but could not predict which [17].
Population: 12 trials, 6 of proton pump inhibitors, lasting 1 week to 6 months, in adults and children with asthma and reflux [17].
Sources: [17]
Caveats: The review pooled proton pump inhibitors with H2 blockers, surgery and lifestyle measures, and dates from 2003 [17].

Doses studied

The pantoprazole label trials used 10, 20 or 40 mg once daily for up to 8 weeks to heal esophagitis and 20 or 40 mg once daily for up to 12 months to keep it healed. For Zollinger-Ellison syndrome the label lists 40 mg twice daily, with doses up to 240 mg a day given [3]. The esomeprazole NSAID trials used 20 or 40 mg once daily for 26 weeks [7]. The over-the-counter omeprazole label lists one 20 mg tablet a day for 14 days, repeated no more often than every 4 months unless a doctor directs [18].

The intensive care trials used 40 mg of intravenous pantoprazole daily [10][11]. The 3-year trial of long-term harms used pantoprazole 40 mg daily [19].

Risks

Reported risks9

ModerateVery common

Headache, diarrhea and other common side effects

Shown in label trials to raise headache, diarrhea and dizziness against placebo, with headache in 12.2% on pantoprazole against 8.5% on placebo.

The numbers and sources

How common: Very common: more than 1 in 10 people.
Effect: In GERD trials, pantoprazole and placebo rates were headache 12.2% and 8.5%, diarrhea 8.8% and 4.9%, vomiting 4.3% and 2.4%, dizziness 3.0% and 1.2%, and joint pain 2.8% and 1.2%. Nausea (7.0% and 9.8%) and abdominal pain (6.2% and 6.1%) were not higher on the drug [3]. In esomeprazole healing trials, headache occurred in 5.5% on 20 mg, 5% on 40 mg and 3.8% on omeprazole 20 mg [7].
Population: 1,473 adults on pantoprazole and 82 on placebo in nine US trials [3]; 6,682 adults in four esomeprazole trials [7].
Sources: [3], [7]
Caveats: The placebo group in the pantoprazole trials was small (82 people) [3]. The esomeprazole trials had no placebo group [7].

ModerateVery common

Low magnesium

Linked to low blood magnesium in observational studies, in 19.4% of users against 13.5% of non-users. The label warns of rare serious cases, mostly after a year of use.

The numbers and sources

How common: Very common: more than 1 in 10 people.
Effect: Across 16 observational studies, low magnesium was found in 19.4% (95% CI 13.8 to 26.5) of proton pump inhibitor users and 13.5% of non-users (adjusted OR 1.71, 95% CI 1.33 to 2.19), with higher odds on high doses (OR 2.13) [29]. In 1,076 heart patients with normal magnesium at the start, the change in magnesium over about 9.5 months did not differ between users and non-users, and low magnesium occurred in 0.4% of each group [30]. The label states low magnesium has been reported rarely after at least three months of use, most often after a year, with tetany, arrhythmias and seizures among the serious events [3].
Population: 13 cross-sectional, 2 case-control and 1 cohort study with 131,507 patients [29]; 1,076 patients after a coronary procedure [30]; postmarketing reports [3].
Sources: [29], [30], [3]
Caveats: Most studies in the meta-analysis measured magnesium once, so they cannot show it fell after starting the drug [29]. The cohort that measured change found none, except a small effect with multiple diuretics [30]. The band above is the share of users with low magnesium in the meta-analysis.

ModerateVery common

Small intestinal bacterial overgrowth

Linked to about twice the odds of small intestinal bacterial overgrowth, found in 36.8% of users studied against 19.9% of controls, rising with longer use.

The numbers and sources

How common: Very common: more than 1 in 10 people.
Effect: Pooled prevalence of bacterial overgrowth was 36.8% (95% CI 29.7 to 44.3) in proton pump inhibitor users and 19.9% in controls (OR 2.14, 95% CI 1.45 to 3.18), with high heterogeneity (I² 77.6%). Each extra month of treatment was associated with about a 4.3 percentage point higher prevalence [34].
Population: 29 studies with 3,682 users and 2,907 controls [34].
Sources: [34]
Caveats: Observational studies. The band above is the prevalence in the people studied, not a rate in all users [34].

LimitedVery common

Acid symptoms after stopping (rebound)

One trial found 44% of healthy volunteers had acid-related symptoms after stopping 8 weeks of a proton pump inhibitor, against 15% after placebo.

The numbers and sources

How common: Very common: more than 1 in 10 people.
Effect: Volunteers took 8 weeks of esomeprazole 40 mg a day and then 4 weeks of placebo, or 12 weeks of placebo. In weeks 9 to 12, 44% (26 of 59) of the drug group and 15% (9 of 59) of the placebo group reported at least one relevant acid-related symptom. Heartburn, dyspepsia or regurgitation was reported by 22% against 5% at week 11 [35].
Population: 120 healthy volunteers, randomized [35].
Sources: [35]
Caveats: One small trial in people without reflux disease. Rebound may be harder to separate from returning symptoms in people with reflux.

ModerateCommon

Gut infections, including C. difficile

Linked to about 1.7 times the odds of C. difficile infection in observational studies. In a 3-year trial, enteric infections were 1.4% on pantoprazole against 1.0% on placebo.

The numbers and sources

How common: Common: between 1 in 100 and 1 in 10 people.
Effect: Across 39 observational studies, proton pump inhibitor use was associated with C. difficile infection (OR 1.74, 95% CI as reported 1.47 to 2.85) and with recurrence (OR 2.51). The odds rose further with antibiotics (OR 1.96 above a proton pump inhibitor alone) [20]. In a randomized trial, enteric infections occurred in 1.4% on pantoprazole and 1.0% on placebo (OR 1.33, 95% CI 1.01 to 1.75). C. difficile infection was about twice as common on pantoprazole, from only 13 events, not statistically significant [19]. The label states observational studies suggest a higher risk of C. difficile diarrhea, especially in hospitalized patients [3].
Population: 42 observational studies with about 313,000 participants [20]; 17,598 adults with stable cardiovascular disease followed a median 3 years [19].
Sources: [20], [19], [3]
Caveats: Heterogeneity was high in the observational analysis (I² 85%) [20]. The trial rate is for all enteric infections, not C. difficile alone [19].

LimitedHow often is not known

Acute interstitial nephritis

The label warns that acute tubulointerstitial nephritis, a kidney inflammation, has occurred with proton pump inhibitors at any point during use. It gives no rate.

The numbers and sources

How common: How often is not known: the sources give no rate.
Effect: The label describes acute tubulointerstitial nephritis from reported case series, presenting with allergic-type symptoms or with nonspecific signs of reduced kidney function such as malaise and nausea [3]. A meta-analysis of observational studies found an association with acute interstitial nephritis (RR 3.61, 95% CI 2.37 to 5.51), rated insufficient strength of evidence [24].
Population: Case series summarized in labeling [3]; observational studies [24].
Sources: [3], [24]
Caveats: No denominator in the case series. The pooled estimate rests on observational data rated insufficient [24].

ModerateHow often is not known

Fundic gland polyps and stomach cancer

Linked to more fundic gland polyps with use of a year or longer, and to about 1.4 times the risk of stomach cancer in a few observational studies.

The numbers and sources

How common: How often is not known: the sources give no rate.
Effect: Pooled odds ratios for fundic gland polyps were 1.43 (95% CI 1.24 to 1.64) in a fixed-effects model and 2.45 (95% CI 1.24 to 4.83) in a random-effects model. The pooled risk ratio for stomach cancer was 1.43 (95% CI 1.23 to 1.66) [36]. The label states most people with these polyps had no symptoms and the polyps were found by chance on endoscopy [3].
Population: 12 studies with more than 87,324 patients, 4 of them on stomach cancer [36].
Sources: [36], [3]
Caveats: The authors state the cancer association could be biased by the few studies and possible confounding [36].

LimitedHow often is not known

Interactions with warfarin, methotrexate and HIV drugs

The labels warn that proton pump inhibitors can raise warfarin's effect and methotrexate levels and change the levels of some HIV drugs. They give no rate.

The numbers and sources

How common: How often is not known: the sources give no rate.
Effect: The esomeprazole label reports raised INR and prothrombin time with warfarin, which may lead to abnormal bleeding, lower exposure to rilpivirine, atazanavir and nelfinavir, and higher exposure to saquinavir. Use with rilpivirine products is contraindicated [7]. Both labels state proton pump inhibitors may raise and prolong methotrexate levels, mainly at high doses [7][3].
Population: Drug interaction studies and reports summarized in labeling [7][3].
Sources: [7], [3]
Caveats: Labeling statements, with no rate of harm given.

LimitedHow often is not known

Severe skin reactions and lupus

The label warns of severe skin reactions and of cutaneous and systemic lupus reported with proton pump inhibitors, mostly cutaneous lupus. It gives no rate.

The numbers and sources

How common: How often is not known: the sources give no rate.
Effect: The label lists Stevens-Johnson syndrome, toxic epidermal necrolysis, DRESS and other severe skin reactions as reported. Cutaneous lupus has been reported more often than systemic lupus, as new onset or a flare, within weeks to years of continuous use [3].
Population: Postmarketing reports summarized in labeling [3].
Sources: [3]
Caveats: Reports with no denominator.

Mixed results7

ModerateHow often is not known

Pneumonia

Mixed results: observational studies linked proton pump inhibitors to about 27% higher odds of pneumonia, while a 3-year randomized trial found no difference from placebo.

The numbers and sources

How common: How often is not known: the sources give no rate.
Effect: Across eight observational studies, pneumonia was more common in proton pump inhibitor users (adjusted OR 1.27, 95% CI 1.11 to 1.46), with very high heterogeneity (I² 90.5%) [21]. In a placebo-controlled trial of pantoprazole for a median of 3 years, pneumonia did not differ between groups [19].
Population: 5 case-control and 3 cohort studies [21]; 17,598 adults with stable cardiovascular disease [19].
Sources: [21], [19]
Caveats: Observational studies cannot rule out that sicker people are prescribed acid reducers [21]. The trial abstract does not give pneumonia counts [19].

StrongHow often is not known

Fractures

Linked to about 1.3 times the hip fracture risk in observational studies. A 3-year randomized trial found no difference in fractures. The label warns of fracture risk with long-term use.

The numbers and sources

How common: How often is not known: the sources give no rate.
Effect: One meta-analysis of observational studies found higher risk of hip (RR 1.30, 95% CI 1.19 to 1.43), spine (RR 1.56) and any-site fractures (RR 1.16), and no significant rise with H2 blockers (RR 1.12, 95% CI 0.97 to 1.30) [22]. A second found hip fracture RR 1.26 (95% CI 1.16 to 1.36) [23]. In a randomized trial, fractures did not differ from placebo over a median 3 years [19]. The label states the risk was higher with multiple daily doses and use for a year or longer [3].
Population: 11 observational studies, mostly of older adults [22]; 18 studies with 244,109 fracture cases [23]; 17,598 adults in a randomized trial [19].
Sources: [22], [23], [19], [3]
Caveats: Residual confounding cannot be excluded in observational studies [22], and heterogeneity was high (I² 71.9%) [23]. Three years may be too short for a trial to detect a change in fracture risk.

ModerateHow often is not known

Chronic kidney disease

Linked to about 1.36 times the risk of chronic kidney disease in observational studies, in low-quality evidence. A 3-year randomized trial found no difference from placebo.

The numbers and sources

How common: How often is not known: the sources give no rate.
Effect: Against non-users, proton pump inhibitor users had more acute kidney injury (RR 1.44, 95% CI 1.08 to 1.91), chronic kidney disease (RR 1.36, 95% CI 1.07 to 1.72) and end-stage kidney disease (RR 1.42) [24]. In one community cohort, use was associated with incident chronic kidney disease (adjusted HR 1.50, 95% CI 1.14 to 1.96), including against H2 blocker users (HR 1.39) [25]. In a randomized trial, chronic kidney disease did not differ between pantoprazole and placebo over a median 3 years [19].
Population: 4 cohort and 5 case-control studies with about 2.6 million patients [24]; 10,482 adults in the ARIC cohort with a replication cohort [25]; 17,598 adults in a randomized trial [19].
Sources: [24], [25], [19]
Caveats: The meta-analysis rated the strength of evidence low or insufficient [24]. Observational designs cannot separate the drug from the reasons it was prescribed.

ModerateHow often is not known

Dementia

Mixed results: one 2026 meta-analysis linked use to about 18% higher dementia risk in older adults, two others found no significant link, and a 3-year trial found no difference.

The numbers and sources

How common: How often is not known: the sources give no rate.
Effect: Across 18 observational studies, the pooled dementia risk ratio was 1.14 (95% CI 0.98 to 1.33, I² 99%), higher in people 65 or older (RR 1.21) [26]. Across 23 longitudinal studies in older adults, the pooled hazard ratio was 1.18 (95% CI 1.02 to 1.37), and risk did not differ from H2 blocker users (HR 1.04) [27]. A 2020 meta-analysis of 12 studies found no association (RR 1.05, 95% CI 0.96 to 1.15) [28]. In a randomized trial, dementia did not differ from placebo over a median 3 years [19].
Population: More than 6.3 million participants in 18 studies [26]; older adults without dementia in 23 studies [27]; 8 cohort and 4 case-control studies [28]; 17,598 adults in a randomized trial [19].
Sources: [26], [27], [28], [19]
Caveats: The observational studies were prone to confounding, exposure misclassification and immortal time bias [26]. Dementia develops over years, longer than the trial ran.

ModerateHow often is not known

Low vitamin B12

Linked to higher odds of a vitamin B12 deficiency diagnosis with 2 or more years of use, while measured B12 levels mostly did not differ from non-users.

The numbers and sources

How common: How often is not known: the sources give no rate.
Effect: In one health system, 2 or more years of proton pump inhibitors was associated with a B12 deficiency diagnosis (OR 1.65, 95% CI 1.58 to 1.73), higher at more than 1.5 pills a day (OR 1.95). For H2 blockers the OR was 1.25 [31]. Across 25 studies, the pooled odds of B12 deficiency were higher in users (OR 1.42, 95% CI 1.16 to 1.73), while most studies found no difference in serum B12 levels [32]. A second meta-analysis found no difference in serum B12 (SMD 0.01, 95% CI -0.14 to 0.16) or homocysteine between long-term users and controls [33]. The label states daily use for longer than 3 years may lead to B12 malabsorption, with rare reports of deficiency [3].
Population: 25,956 people with B12 deficiency and 184,199 without, in a case-control study [31]; 25 studies [32]; 6 studies with 1,587 users and 2,272 controls [33].
Sources: [31], [32], [33], [3]
Caveats: Observational studies. A diagnosis count reflects who was tested as well as who was low [31]. Heterogeneity was significant [32].

LimitedHow often is not known

Death from any cause

Mixed results: a veterans cohort linked proton pump inhibitor use to about 25% higher risk of death than H2 blocker use, while a 3-year randomized trial found no difference.

The numbers and sources

How common: How often is not known: the sources give no rate.
Effect: Over a median 5.7 years, proton pump inhibitor users had a higher risk of death than H2 blocker users (HR 1.25, 95% CI 1.23 to 1.28) and than non-users (HR 1.15). The association held in people without gut conditions (HR 1.24) [37]. In a randomized trial, all-cause mortality did not differ between pantoprazole and placebo over a median 3 years [19].
Population: US veterans in a longitudinal cohort [37]; 17,598 adults in a randomized trial [19].
Sources: [37], [19]
Caveats: The cohort is observational and cannot rule out that sicker people were prescribed proton pump inhibitors [37].

LimitedHow often is not known

Interaction with clopidogrel

The esomeprazole label warns against use with clopidogrel because it reduces clopidogrel's antiplatelet activity. One trial found no rise in heart events with omeprazole.

The numbers and sources

How common: How often is not known: the sources give no rate.
Effect: The label states that esomeprazole inhibits CYP2C19, the enzyme that converts clopidogrel to its active form, that esomeprazole 40 mg reduces clopidogrel's pharmacological activity, and that the label's instruction is to avoid using the two together [7]. In a randomized trial of people on aspirin and clopidogrel, cardiovascular events were 4.9% on omeprazole and 5.7% on placebo (HR 0.99, 95% CI 0.68 to 1.44) [8].
Population: Drug interaction studies summarized in labeling [7]; 3,761 patients on dual antiplatelet therapy [8].
Sources: [7], [8]
Caveats: The trial ended early and was not powered to rule out a small rise in heart events [8].

Checked, no increase found1

ModerateStudied in pregnancy

Pregnancy

Not linked to major birth defects or miscarriage in observational studies summarized on the labels, including a meta-analysis of first-trimester use.

The numbers and sources

How common: How often is not known: the sources give no rate.
Effect: The pantoprazole label cites a meta-analysis of 1,530 women exposed in at least the first trimester and 133,410 unexposed, with no significant rise in major malformations (OR 1.12, 95% CI 0.86 to 1.45) or miscarriage (OR 1.29, 95% CI 0.84 to 1.97). It states observational limits cannot establish or exclude a risk, and that rat studies showed bone changes in exposed pups [3]. The esomeprazole label states epidemiologic data on first-trimester omeprazole show no increase in major malformations [7].
Population: Pregnancy cohorts and a meta-analysis summarized in labeling [3][7].
Sources: [3], [7]
Caveats: Observational data only [3].

Claims tested and not found

That long-term proton pump inhibitor use causes pneumonia, fractures, kidney disease, dementia or earlier death. In 17,598 people randomized to pantoprazole 40 mg or placebo for a median 3 years, none of these differed between groups. Enteric infections were the one outcome that rose (1.4% against 1.0%) [19].
That proton pump inhibitors reduce crying and irritability in infants. A systematic review of randomized trials found no effect on infant crying or irritability [38].
That proton pump inhibitors resolve chronic cough. In adult trials lasting 2 to 3 months, total cough resolution did not differ from placebo (OR 0.46, 95% CI 0.19 to 1.15) [16].

Sources

Sigterman KE, van Pinxteren B, Bonis PA, et al. Short-term treatment with proton pump inhibitors, H2-receptor antagonists and prokinetics for gastro-oesophageal reflux disease-like symptoms and endoscopy negative reflux disease. Cochrane Database of Systematic Reviews. 2013. https://doi.org/10.1002/14651858.CD002095.pub5
Khan M, Santana J, Donnellan C, et al. Medical treatments in the short term management of reflux oesophagitis. Cochrane Database of Systematic Reviews. 2007. https://doi.org/10.1002/14651858.CD003244.pub2
U.S. Food and Drug Administration. PROTONIX (pantoprazole sodium) delayed-release tablets and oral suspension, prescribing information. 2026. DailyMed. https://dailymed.nlm.nih.gov/dailymed/drugInfo.cfm?setid=08098cb2-c048-4640-f387-6beec4a38936
Gralnek IM, Dulai GS, Fennerty MB, et al. Esomeprazole versus other proton pump inhibitors in erosive esophagitis: a meta-analysis of randomized clinical trials. Clinical Gastroenterology and Hepatology. 2006. https://doi.org/10.1016/j.cgh.2006.09.013
Leontiadis GI, Sharma VK, Howden CW. Proton pump inhibitor therapy for peptic ulcer bleeding: Cochrane collaboration meta-analysis of randomized controlled trials. Mayo Clinic Proceedings. 2007. https://doi.org/10.4065/82.3.286
Rostom A, Dube C, Wells G, et al. Prevention of NSAID-induced gastroduodenal ulcers. Cochrane Database of Systematic Reviews. 2002. https://doi.org/10.1002/14651858.CD002296
U.S. Food and Drug Administration. NEXIUM (esomeprazole magnesium) delayed-release capsules and oral suspension, prescribing information. 2025. DailyMed. https://dailymed.nlm.nih.gov/dailymed/drugInfo.cfm?setid=f4853677-1622-4037-688b-fdf533a11d96
Bhatt DL, Cryer BL, Contant CF, et al. Clopidogrel with or without omeprazole in coronary artery disease. New England Journal of Medicine. 2010. https://doi.org/10.1056/NEJMoa1007964
Moayyedi P, Eikelboom JW, Bosch J, et al. Pantoprazole to Prevent Gastroduodenal Events in Patients Receiving Rivaroxaban and/or Aspirin in a Randomized, Double-Blind, Placebo-Controlled Trial. Gastroenterology. 2019. https://doi.org/10.1053/j.gastro.2019.04.041
Cook D, Deane A, Lauzier F, et al. Stress Ulcer Prophylaxis during Invasive Mechanical Ventilation. New England Journal of Medicine. 2024. https://doi.org/10.1056/NEJMoa2404245
Krag M, Marker S, Perner A, et al. Pantoprazole in Patients at Risk for Gastrointestinal Bleeding in the ICU. New England Journal of Medicine. 2018. https://doi.org/10.1056/NEJMoa1714919
Ford AC, Gurusamy KS, Delaney B, et al. Eradication therapy for peptic ulcer disease in Helicobacter pylori-positive people. Cochrane Database of Systematic Reviews. 2016. https://doi.org/10.1002/14651858.CD003840.pub5
Pinto-Sanchez MI, Yuan Y, Hassan A, et al. Proton pump inhibitors for functional dyspepsia. Cochrane Database of Systematic Reviews. 2017. https://doi.org/10.1002/14651858.CD011194.pub3
Guo H, Ma H, Wang J. Proton Pump Inhibitor Therapy for the Treatment of Laryngopharyngeal Reflux: A Meta-Analysis of Randomized Controlled Trials. Journal of Clinical Gastroenterology. 2016. https://doi.org/10.1097/MCG.0000000000000324
Wei C. A meta-analysis for the role of proton pump inhibitor therapy in patients with laryngopharyngeal reflux. European Archives of Oto-Rhino-Laryngology. 2016. https://doi.org/10.1007/s00405-016-4142-y
Chang AB, Lasserson TJ, Gaffney J, et al. Gastro-oesophageal reflux treatment for prolonged non-specific cough in children and adults. Cochrane Database of Systematic Reviews. 2011. https://doi.org/10.1002/14651858.CD004823.pub4
Gibson PG, Henry RL, Coughlan JL. Gastro-oesophageal reflux treatment for asthma in adults and children. Cochrane Database of Systematic Reviews. 2003. https://doi.org/10.1002/14651858.CD001496
U.S. Food and Drug Administration. PRILOSEC OTC (omeprazole magnesium) delayed-release tablets, drug facts label. 2026. DailyMed. https://dailymed.nlm.nih.gov/dailymed/drugInfo.cfm?setid=37291cab-e350-d8e3-e063-6294a90a9cb1
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. https://doi.org/10.1053/j.gastro.2019.05.056
Kwok CS, Arthur AK, Anibueze CI, et al. Risk of Clostridium difficile infection with acid suppressing drugs and antibiotics: meta-analysis. American Journal of Gastroenterology. 2012. https://doi.org/10.1038/ajg.2012.108
Eom CS, Jeon CY, Lim JW, et al. Use of acid-suppressive drugs and risk of pneumonia: a systematic review and meta-analysis. CMAJ. 2011. https://doi.org/10.1503/cmaj.092129
Yu EW, Bauer SR, Bain PA, et al. Proton pump inhibitors and risk of fractures: a meta-analysis of 11 international studies. American Journal of Medicine. 2011. https://doi.org/10.1016/j.amjmed.2011.01.007
Zhou B, Huang Y, Li H, et al. Proton-pump inhibitors and risk of fractures: an update meta-analysis. Osteoporosis International. 2016. https://doi.org/10.1007/s00198-015-3365-x
Nochaiwong S, Ruengorn C, Awiphan R, et al. The association between proton pump inhibitor use and the risk of adverse kidney outcomes: a systematic review and meta-analysis. Nephrology Dialysis Transplantation. 2018. https://doi.org/10.1093/ndt/gfw470
Lazarus B, Chen Y, Wilson FP, et al. Proton Pump Inhibitor Use and the Risk of Chronic Kidney Disease. JAMA Internal Medicine. 2016. https://doi.org/10.1001/jamainternmed.2015.7193
Peng TR, Lin HH, Yang LJ, et al. Inconsistency in the Association Between Proton Pump Inhibitor Use and Dementia Risk: An Updated Meta-Analysis. Brain Sciences. 2026. https://doi.org/10.3390/brainsci16020159
Ha H, Lee HW, Park JE, et al. Evaluating the Association Between PPI Use and Dementia Risk Considering Methodological Choices: An Updated Meta-Analysis. Clinical Pharmacology and Therapeutics. 2026. https://doi.org/10.1002/cpt.70534
Hussain S, Singh A, Zameer S, et al. No association between proton pump inhibitor use and risk of dementia: Evidence from a meta-analysis. Journal of Gastroenterology and Hepatology. 2020. https://doi.org/10.1111/jgh.14789
Srinutta T, Chewcharat A, Takkavatakarn K, et al. Proton pump inhibitors and hypomagnesemia: A meta-analysis of observational studies. Medicine. 2019. https://doi.org/10.1097/MD.0000000000017788
Park SH, Lee SH, Lee JS, et al. Changes in serum magnesium concentration after use of a proton pump inhibitor in patients undergoing percutaneous coronary intervention. Kidney Research and Clinical Practice. 2015. https://doi.org/10.1016/j.krcp.2015.03.001
Lam JR, Schneider JL, Zhao W, et al. Proton pump inhibitor and histamine 2 receptor antagonist use and vitamin B12 deficiency. JAMA. 2013. https://doi.org/10.1001/jama.2013.280490
Choudhury A, Jena A, Jearth V, et al. Vitamin B12 deficiency and use of proton pump inhibitors: a systematic review and meta-analysis. Expert Review of Gastroenterology & Hepatology. 2023. https://doi.org/10.1080/17474124.2023.2204229
Parnham O, Patient W. Association Between Long-Term Proton Pump Inhibitor Therapy and Vitamin B12 Status: A Systematic Review and Meta-Analysis. Cureus. 2025. https://doi.org/10.7759/cureus.90038
Khurmatullina AR, Andreev DN, Kucheryavyy YA, et al. The Duration of Proton Pump Inhibitor Therapy and the Risk of Small Intestinal Bacterial Overgrowth: A Systematic Review and Meta-Analysis. Journal of Clinical Medicine. 2025. https://doi.org/10.3390/jcm14134702
Reimer C, Søndergaard B, Hilsted L, et al. Proton-pump inhibitor therapy induces acid-related symptoms in healthy volunteers after withdrawal of therapy. Gastroenterology. 2009. https://doi.org/10.1053/j.gastro.2009.03.058
Tran-Duy A, Spaetgens B, Hoes AW, et al. Use of Proton Pump Inhibitors and Risks of Fundic Gland Polyps and Gastric Cancer: Systematic Review and Meta-analysis. Clinical Gastroenterology and Hepatology. 2016. https://doi.org/10.1016/j.cgh.2016.05.018
Xie Y, Bowe B, Li T, et al. Risk of death among users of Proton Pump Inhibitors: a longitudinal observational cohort study of United States veterans. BMJ Open. 2017. https://doi.org/10.1136/bmjopen-2016-015735
Gieruszczak-Białek D, Konarska Z, Skórka A, et al. No effect of proton pump inhibitors on crying and irritability in infants: systematic review of randomized controlled trials. Journal of Pediatrics. 2015. https://doi.org/10.1016/j.jpeds.2014.11.030

Everything in the library