Adverse Effects of Proton Pump Inhibitors: Fact or Fake News?

Mitchell L. Schubert


Curr Opin Gastroenterol. 2018;34(6):451-457. 

In This Article

Potential Adverse Effects of Proton Pump Inhibitors

It should be noted that most, if not all, studies maligning PPIs are retrospective observational studies that, by design, can generate hypotheses but cannot ascribe cause and effect.[1,9] These large database studies have inherent limitations, as the studies were not designed to answer a specific question, information was not collected with a specific hypothesis in mind, PPIs were not assigned at random, experimental and control groups are not comparable, level of exposure (dose, duration and date of initiation) to PPI is not known and unaccounted biases and confounding variables persist, even after multivariable regression and propensity weighting. PPIs users tend to be older, sicker, hospitalized more frequently and taking more medications. Although the greater the magnitude of the association, the more likely that the relationship may be causal, when effect sizes are small [odds ratio (OR) or hazard ratio < 3], as is the case in all the studies, it is not possible to determine whether the association is valid or due to residual bias and confounding variables. Expressing risk in terms of relative risk (OR or hazard ratio), instead of absolute risk or number needed to harm, is misleading and overestimates the risk to an individual, particularly when adverse events are uncommon.

There is little or no biological plausibility linking PPIs to the potential adverse effects to be reviewed (except for alterations in gut microbiome, enteric infection, micronutrient deficiency, fundic gland polyps, and possibly carcinogenesis and interference with drug absorption), as PPIs are in the bloodstream for only a relatively short time period (elimination half-life ~1 h) and peak plasma concentrations are relatively low. The reason PPIs cause prolonged inhibition of acid secretion has little to do with blood concentrations but rather is due to the fact that PPIs become concentrated and trapped within the secretory canaliculus of the parietal cell wherein they covalently bind to cysteine residues on the luminally exposed α-subunit of the H+K+-ATPase to inactivate the acid pump. Recovery of acid secretion depends upon de-novo synthesis of pump protein, a process that takes 54 h in rat.

Gut Microbiome

It is conceivable that PPIs, by producing hypochlorhydria, may allow survival of certain ingested microbes and, thus, alter the composition of the gastrointestinal microbiome. Prior studies have reported an overexpression of oral bacteria in the faeces of individuals taking PPIs as well as an increased ratio of Firmicutes to Bacteroidetes at the phylum level and an increase in Holdemania filiformis and decrease in Psuedoflavonifractor capillosus at the species level.[10] Ten healthy adult volunteers on no medications were enrolled in a study in which salivary, periodontal pocket and faecal samples were analysed for bacterial composition using 16S rRNA before and after 4 weeks of once daily 20 mg esomeprazole.[11] The genus Streptococcus was increased after PPI administration in all three samples but reached statistical significance only in stool. The degree of change in the gut microbiome associated with PPIs is comparable to that induced by antibiotics. The clinical significance of these and other changes in the microbiome induced by PPIs remains to be elucidated.


As gastric acid kills ingested microorganisms, it is biologically plausible that PPIs, by reducing gastric acid secretion, may increase enteric infection, particularly of acid-sensitive organisms such as vibrio cholera, salmonella, campylobacter and, perhaps, norovirus.[12] Other mechanisms may include increased intestinal permeability and altered microbiome. In mice, a PPI (lansoprazole) increased host susceptibility to colitis after oral gavage with Citrobacter rodentium, an enteropathogen used experimentally as a model of human enterohemorrhagic Escherichia coli.[13] The ileal microbiota were altered as a result of the increased gastric pH. The decrease in Clostridiales may have compromised colonization resistance to C. rodentium. It should be noted that rodents practice coprophagy and reduced killing of orally ingested faecal bacteria could have been responsible for the altered microbiome. Studies in humans indicate that PPIs alter the faecal microbiome with increased abundance of oral bacteria.[14]

Numerous observational studies, all with OR less than 3, report an association between PPIs and C. difficile infection (CDI).[14,15] It should be noted that vegetative cells of C. difficile are susceptible to gastric acid, but spores are resistant. Thus, if PPIs are eventually proven to cause CDI, the effect is more likely related to dysbiosis than to an acid antisecretory effect per se.

Although PPIs have been associated with an increased risk of infection, a retrospective cohort study of 24 774 ICU patients failed to show an association between PPIs, used for stress ulcer prophylaxis, and bloodstream infections.[16]


Gastrin is not only a secretagogue but also a growth hormone capable of stimulating proliferation, cell migration and angiogenesis as well as inhibiting apoptosis and activating autophagy.[17] Gastrin receptors (termed gastrin/CCK2, CCK2 or CCK-B receptors) have been identified in various human cancers including adenocarcinomas of the stomach and oesophagus.[18] Gastric hypoacidity, induced by PPIs, interferes with the feedback pathway whereby luminal acid stimulates somatostatin secretion, which, in turn, inhibits gastrin secretion. Patients on high-dose PPIs manifest hypergastrinaemia and hypergastrinaemia has been associated with an increased risk of gastric adenocarcinoma, neuroendocrine tumours (i.e. enterochromaffin-like cell carcinoids) and possibly oesophageal adenocarcinoma.[19]

A total of 1 563 860 individuals in the Danish Prescription Drug Registry taking acid-suppressive medications were matched to unexposed population-based controls.[20] Those with five or more prescriptions for PPIs were six-fold and 10-fold more likely to develop proximal and distal gastric cancer, respectively. In a Hong Kong health database, PPI use, after eradication of H. pylori, was associated with an increased gastric cancer risk (hazard ratio, 2.4).[21] In mice, omeprazole promotes carcinogenesis of the fore-stomach following treatment with the nitrosamine, N-methyl-N'-nitro-N-nitrosoguanidine.[22] Dose–response effects were reported in both the epidemiological and experimental studies.

A study from China reported significant microbiome dysbiosis in gastric biopsies from patients with progressive stages of gastric carcinogenesis starting from superficial gastritis and progressing through atrophy, intestinal metaplasia and cancer.[23] Gastric cancer patients harboured increased numbers of oral bacteria, including P. micra, P. stomatis and F. nucleatum. It is not known what percentage of patients were taking PPIs.

The impact of PPIs on development and progression of oesophageal adenocarcinoma is unclear. However, the gastrin/CCK2 receptor is upregulated in Barrett's oesophagus and L2-IL-1β transgenic mice manifest increased proliferation and expansion of Barrett's-like oesophagus when rendered hypergastrinemic by treatment with PPIs.[24] Furthermore, in cultures of gastric cardia, gastrin stimulates organoid growth while CCK2R inhibition prevents Barrett's-like oesophagus and dysplasia. The data suggest a progression of Barrett's oesophagus to adenocarcinoma in which CCK2R+ progenitor cells, stimulated by hypergastrinaemia, proliferate to give rise to dysplasia.[24] In a Swedish population-based cohort study of 796 492 adults exposed to maintenance therapy with PPIs, an increased risk for development of oesophageal cancer, in particular adenocarcinoma, was reported among individuals using maintenance PPI therapy overall (hazard ratio, 3.9) as well as among those who used PPIs for indications not associated with an increased risk of oesophageal adenocarcinoma (hazard ratio, 2.0–2.7).[25] These findings herald a warning regarding the use of high-dose PPIs, especially in patients with asymptomatic Barrett's oesophagus.

Chronic Kidney Disease

Data purporting an association between PPIs and chronic kidney disease (CKD) have been weak. Two additional studies, one of which is a meta-analysis, were published this past year; both show a weak association and cannot exclude residual confounding.[26,27] A Stockholm administrative database of creatinine measurements of all the region's citizens from 2007 to 2010 was analysed to examine the association between new users of antisecretory medications and progression of CKD, defined as a doubling of creatinine levels or a decline in estimated glomerular filtration rate of at least 30%.[26] Patients using PPIs were more likely to experience progression when compared with those using histamine H2-receptor antagonists (H2As), but the association was relatively weak (hazard ratio, 1.26). Those taking PPIs, were older, had more comorbidities and took more medications, including NSAIDs.

A systematic review and meta-analysis identified five studies with 536 902 participants that investigated that association between antisecretory medications and CKD.[27] When compared with non-PPI users, the pooled risk ratio of CKD with PPI use was only 1.22. When compared with the use of H2RAs, the pooled relative risk of CKD in patients using PPIs was 1.29. Unaccounted biases and confounding variables were most likely responsible for the weak associations reported by these observational studies. For example, PPI users had increased comorbidities, including diabetes, hypertension and ischemic heart disease and dose and frequency of PPI use is not documented.

Cognitive Decline

Inconsistent and conflicting results have been reported regarding the association of PPIs with cognitive decline. Several studies found no association and when an association was present, it was weak (hazard ratio, ~1.4). This past year, four additional studies were published, each of which showed no association between PPIs and cognitive decline.[28–31] A large study of 13 864 participants in the Nurses' Health Study II, who had data on medication use prospectively collected and had completed a self-administered computerized battery of neuropsychological tests, found no convincing association between PPI use and cognitive dysfunction.[28] Data from two large population-based studies of middle-aged and older twins in Denmark, who underwent cognitive assessments over a 2 to 10-year period, also found no association between PPI and cognitive decline.[29] A United States population-based cohort study of individuals aged 65 years and older without dementia at study entry (N = 3484) who were screened for dementia every 2 years for a mean of 7.5 years, also found no association between PPI use and dementia, even for people with high cumulative exposure.[30] Finally, a very large Finnish nested case–control study (70 718 cases of Alzheimer's disease and 282 862 controls) reported no significant association between PPI use and Alzheimer's disease.[31] Thus, patients and providers should be reassured that accumulative data indicate a lack of a causal relationship between PPI use and changes in cognitive function.

Myocardial Infarction and Stroke

The evidence linking PPIs to an increased risk of major adverse cardiac events has been weak. A 2017 meta-analysis of 17 GERD trials reported a hazard ratio of only 1.7 for patients taking PPIs.[32] The quality of evidence, however, was moderate or low, cardiovascular outcomes were not study endpoints and confounding variables such as obesity, smoking, alcohol and family history were not assessed.

Three studies were subsequently published, in 2018, that support the cardiovascular safety of PPIs. A Danish registry study of 214 998 individuals, without prior history of MI or stroke who underwent an elective upper endoscopy between 1997 and 2012, reported only a minimal increase in ischaemic stroke and MI for current PPI users (hazard ratio, 1.13 and 1.31, respectively).[33] Although obesity, smoking and exercise status were addressed, a major limitation of this and all such observational studies is the inability to correct for unmeasured confounding variables and medication adherence.

A United States study, using a large administrative database of 80 million individuals covered by commercial and Medicare supplemental plans, reported no evidence that prescription PPIs increased risk of MI compared with prescription H2RAs.[34] Patients diagnosed with an MI within 1 year of the first prescription claim were excluded.

Another United States study of 68 514 women (mean age, 65 ± 7 years) enrolled in the Nurses' Health Study and 28 989 men (mean age, 69 ± 8 years) enrolled in the Health Professionals Follow-up Study report no increased risk of stroke in regular PPI users.[35] The individuals were followed biennially with detailed questionnaires and follow-up rates exceeded 90%.

Bone Fracture

Although there may be some evidence suggesting an association between PPIs and osteoporosis, the evidence linking PPIs to bone fracture is more tenuous.[36,37] It has been proposed that PPIs may induce bone fracture by reducing calcium absorption and/or increasing falls due to vitamin B12 deficiency. A Finnish nested case–control study using a nationwide database of elderly patients with Alzheimer's disease, half of which used PPIs, found no association between PPIs used for longer than a year and hip fracture.[37]

Drug Interaction

It is conceivable that PPIs, by increasing gastric pH, could adversely affect the bioavailability and absorption of certain medications. Although the package insert for direct-acting antiviral medications used to treat hepatitis C cautions against concomitant use of PPIs, a recent secondary analysis of data from six phase III trials of ombitasvir, ritonavir and dasabuvir with or without ribavirin reports that PPI use was not a predictor of treatment failure.[38] A subsequent meta-analysis, however, comprising nine cohort studies with 32 684 participants, reports that use of PPIs was associated with a lower odds of achieving a sustained virologic response (1.4-fold increased risk of failure).[39] There are some data to suggest that twice-daily PPI, but not once daily, may be associated with a lower OR for treatment success.[40]

PPIs might affect the absorption and therefore the effectiveness of certain chemotherapeutic drugs. Preclinical data indicate that a number of tyrosine-kinase inhibitors rely on pH-dependent solubility to dissolve within the stomach and be absorbed. Capecitabine is an oral 5-fluorouracil prodrug used in the adjuvant treatment of gastrointestinal cancer. A secondary analysis of TRIO-013, a phase III randomized trial, comparing capecitabine and oxaliplatin with or without lapatinib in 545 patients with metastatic gastroesophageal cancer, reports that PPI use, identified by medication records, was associated with a poorer progression-free and overall survival.[41]


Although PPI associations with cardiac events and dementia have largely been debunked and concerns assuaged, a recent study associating PPI use with a higher risk of death has grabbed attention and rekindled unnecessary angst and alarm.[42] Researchers mined administrative data from the U.S. Department of Veterans Affairs to evaluate the risk for all-cause mortality in mainly white male veterans who began taking PPIs (n = 275 933) or H2RAs (n = 73 355) between October 2006 and September 2008. Over a median follow-up of 5.7 years, PPIs were associated with a minimal increased risk for death compared with no PPI (hazard ratio, 1.15) and to H2RAs (hazard ratio, 1.25). The cause of death was not collected and there is no clear mechanism to explain the association. As discussed previously for these types of retroactive analyses of large databases, patients using antisecretory medications, especially PPIs, are different from patients not taking these medications, that is confounding by indication; the patients are sicker, have more severe degrees of disease and have increased exposure to the healthcare system itself, including hospitalizations. Thus, this study provides no convincing evidence that PPI use is the cause of the excess deaths.