Systematic Review

The Effects of Proton Pump Inhibitors on the Microbiome of the Digestive Tract

Evidence From Next-generation Sequencing Studies

Lukas Macke; Christian Schulz; Leandra Koletzko; Peter Malfertheiner


Aliment Pharmacol Ther. 2020;51(5):505-526. 

In This Article

Abstract and Introduction


Background: Proton pump inhibitors (PPI) are widely used to treat acid-related disorders of the upper gastrointestinal tract. However, large observational studies have raised concerns about PPI-associated adverse events. In recent years, data from next-generation sequencing studies suggested that PPIs affect the composition of the intestinal microbiota, while a balanced gut microbiome is essential for maintaining health.

Aim: To review the available evidence from next-generation sequencing studies on the effect of PPIs on the intestinal microbiome and to discuss possible implications of PPI-induced dysbiosis in health and disease.

Methods: A systematic review was conducted following the recommendations of the Preferred Reporting Items for Systematic reviews and Meta-Analyses statement. A PubMed query yielded 197 results. 19 publications met the prespecified eligibility criteria.

Results: Twelve observational study cohorts with 708 PPI users and 11 interventional cohorts with 180 PPI users were included in the review. In most studies, PPI treatment did not affect microbiological richness and diversity, but was associated with distinct taxonomic alterations: In the upper gastrointestinal tract, PPI users showed overgrowth of orally derived bacteria, mostly Streptococcaceae (findings based on six independent cohorts with 126 PPI users). In faecal samples, PPIs increased multiple taxa from the orders Bacillales (eg, Staphylococcaceae), Lactobacillales (eg, Enterococcaceae, Lactobacillaceae, Streptococcaceae) and Actinomycetales (eg, Actinomycetaceae, Micrococcaceae), the families Pasteurellaceae and Enterobacteriaceae and the genus Veillonella. Taxa decreased by PPIs include Bifidobacteriaceae, Ruminococcaceae, Lachnospiraceae and Mollicutes (findings in faecal samples based on 19 independent cohorts with 790 PPI users).

Conclusion: PPI use is associated with moderate alterations to upper and distal gut microbiota. The available data suggest that PPI-induced hypochlorhydria facilitates colonization of more distal parts of the digestive tract by upper gastrointestinal microbiota.


Introduction: Gastric Acid Inhibition And Adverse Effects of PPIS

Gastric acid eliminates microorganisms, promotes the digestive process and facilitates the absorption of iron, calcium and vitamin B12. However, gastric acid also contributes to the development of mild and severe gastroduodenal pathologies. Proton pump inhibitor (PPIs) are the most effective therapy of acid-related diseases and are among the most commonly used drugs.[1]

While generally considered a safe therapy, PPIs have been associated with numerous adverse effects in population-based observational studies conducted in recent years. These include cardiovascular and kidney disease, micronutrient deficiencies and osteoporosis, cognitive impairment and death.[2]

Most studies are retrospective and not hypothesis-driven in design, might be subject to significant confound and thus they cannot establish causality.[2,3] This is also the case for a recent study on the link between obesity and PPI use in early childhood.[4]

Associations of PPIs with ischemic stroke,[5] myocardial infarction[6] and cognitive impairment were not confirmed by more recent studies.[7]

PPIs and Enteric Infections

The association of PPI use and Clostridium difficile infection appears evident.[8–12] However, due to limitations in study designs, low odds ratios (OR) and some uncertainties concerning biological plausibility, the available studies cannot ultimately provide sufficient evidence of causality of this association.[13,14] Then again, PPI use and community-acquired bacterial enteric infection show a strong association with a pooled OR of 4.28 (95% CI 3.01–6.08) for any infection, 4.84 (95% CI 2.75–8.54) for Salmonella infection and 5.09 (95% CI 3–8.64) for Campylobacter jejuni infection, as shown in a recent meta-analysis of nine observational studies[15] (Table 1, Figure 2).

PPIs and Gastric Cancer

Concern has been raised in population-based studies regarding an increased gastric cancer risk among long-term PPI users. In a large Swedish cohort, PPI therapy for at least 180 days was associated with a standardized incidence ratio of 3.38 (95% CI 3.25–3.53) for any-location gastric cancer during a mean follow-up of 4.9 years.[16] Continuation of PPI therapy after Helicobacter pylori eradication in the population of Hong Kong was associated with an increased risk for non-cardia gastric cancer during a median follow-up of 7.6 years (HR 2.44, 95% CI 1.42–4.20).[17] Two population-based case–control studies from Taiwan demonstrated increased odds for any-site gastric cancer in patients who were treated at least two times for gastro-oesophageal reflux disease between 1996 and 2011 (OR 2.48, 95% CI 1.92–3.20) and in subjects with a cumulative duration of PPI treatment of >6 months for unknown indications between 2000 and 2013 (OR 2.00, 95% CI 1.36–2.95), each compared with those who never used PPIs.[18,19] A recent meta-analysis including >900 000 individuals confirmed this association with a pooled OR of 2.10 (95% CI 1.10–3.09) and the previously reported duration–response relationship [20] (Table 1, Figure 2). To date, the possible biological mechanisms underlying this association are unclear.

PPIs and Upper Gastrointestinal Bacterial Overgrowth

Prior to the introduction of the first PPI to the market in 1988,[21] a 2-week course of omeprazole was shown to increase bacterial counts in gastric juice of healthy volunteers.[22] Subsequently, controlled trials confirmed an increased incidence of gastric and duodenal bacterial overgrowth during short-term omeprazole therapy. The bacteria were primarily those resident in the oral cavity, but included also faecal type and anaerobic bacteria in cultures from gastric and duodenal aspirates.[23,24] Considering that the gastric mucosa represents a distinct niche, protected from the stomach lumen by the gastric mucous layer, Sanduleanu et al observed luminal and mucosal growth of non-H pylori bacteria following long-term omeprazole therapy. Luminal bacteria were related to gastric pH, and mucosal bacterial growth increased with duration of acid inhibition[25] (Table 1, Figure 2).

Small intestinal bacterial overgrowth (SIBO) is defined by >105/mL colony forming units of bacteria in jejunal aspirates.[26] This condition can be associated with non-specific mucosal inflammation, malabsorption and abdominal symptoms such as bloating and diarrhoea.[27] In patients with liver cirrhosis, SIBO is also associated with ascites, minimal hepatic encephalopathy and spontaneous bacterial peritonitis.[28] Likewise, PPI use has been associated with hepatic encephalopathy and spontaneous bacterial peritonitis [29–32] and therefore, these conditions are believed to be mediated by PPI-induced SIBO. While the risk of PPIs for the development of SIBO remained equivocal and inconclusive in many individual studies, two recent meta-analyses including 11 and 19 studies with 3134 and 7055 subjects both revealed a statistically significant association of PPI use with an increased risk of SIBO, with pooled ORs of 2.282 (95% CI 1.238–4.205) and 1.71 (95% CI 1.20–2.43)[33,34] (Table 1, Figure 2).

Next-generation Sequencing and Gut Microbiota Dysbiosis

An estimated 80% of microorganisms present in the gut cannot be detected by traditional culturing methods. This limitation was overcome by the development of parallelized high-throughput sequencing technologies, commonly referred to as next-generation sequencing, which demonstrated the enormous variety and diversity of non-culturable microorganisms in the digestive tract.[35–37] Targeted sequencing of amplicons of the variable regions of the 16S rRNA is the most widely used approach for the identification of bacterial taxa and their relative abundance. However, a sufficient sequence coverage is required for obtaining resolutions down to species level, and there are concerns regarding reproducibility between the established sequencing platforms and computational tools.[38–40]

The application of next-generation sequencing indicated numerous associations between gastrointestinal conditions and altered composition of the microbiome, for example in inflammatory bowel disease[41] and C difficile infection,[42] and a wide range of non-intestinal diseases such as type 1 and 2 diabetes,[43,44] liver and cardiovascular disease,[45,46] rheumatic diseases,[47] multiple sclerosis[48] and autism-spectrum disorders.[49] Microbiome compositions that differ from a balanced, healthy state are often referred to as dysbiosis, although there is no consensus on the composition of a eubiotic, healthy microbiome, nor is there a widely agreed definition of the term dysbiosis.[50,51] One definition of dysbiosis was proposed by Levy et al featuring the following characteristics: the bloom of pathobionts, the loss of commensals and the loss of diversity.[52]

Currently, much effort is concentrated on exploring causal relationships between dysbiotic microbiota and disease. The hypothesis has been proposed, that a combination of genetic susceptibility and environmental exposure may trigger mucosal injury, inflammation, epithelial barrier dysfunction and dysbiosis of the gut.[53] This homeostatic imbalance may lead to altered microbial metabolite production,[54] dysregulation of host immunity[52,55] and reduced colonization resistance against pathogens,[56,57] ultimately promoting the pathogenesis of chronic systemic disorders.[53]

By abrogation of the stomach acid barrier, PPI use might permit survival and colonization of the stomach with oropharyngeal and environmental microbiota and their transition further down to small and large intestine. This pathway is supposed to lead to intestinal dysbiosis (Figure 2). Furthermore, it has been hypothesized that PPIs might directly affect microbes including H pylori, Streptococci, Lactobacilli, Candida albicans and Saccharomyces, which express P-type ATPases with high homology to the H+/K+ATPase of parietal cells.[58–65] Beyond conclusion by analogy, the mechanism of action of PPIs on microbial P-type ATPases has not been explored experimentally.