Кислотная терапия и риск излечения в случае воспламенительной болезни лопатки: сравнительное исследование в целях информирования о личностном лечении в мире.

1. Introduction

Proton pump inhibitors (PPIs) are among the most frequently prescribed medications worldwide and are commonly used in patients with inflammatory bowel disease (IBD) for gastroesophageal reflux disease, dyspepsia, stress-ulcer prophylaxis, and medication-related upper gastrointestinal symptoms [ 1 , 2 ]. Consequently, a substantial proportion of patients with IBD are exposed to acid-suppressive therapy during their disease. Despite their widespread use, increasing attention has focused on the potential impact of PPIs on intestinal inflammation, infection risk, and IBD course [ 3 , 4 , 5 ].

PPIs produce profound and sustained gastric acid suppression through covalent inhibition of the H + /K + -ATPase, resulting in prolonged suppression of gastric acid until new proton pumps are synthesized [ 6 ]. Previous studies have demonstrated that PPI use is associated with significant changes in gut microbiota composition, reduced microbial diversity, and impaired colonization resistance [ 7 ]. These microbiome effects have been linked to increased susceptibility to enteric infections, particularly Clostridioides difficile ( C. difficile ), a complication of particular concern in patients with IBD who frequently receive antibiotics and immunosuppressive therapies [ 4 , 8 , 9 ]. Given the central role of microbial dysbiosis and immune–microbial interactions in IBD pathogenesis [ 10 , 11 ], these biologic effects raise concern that PPI exposure may influence disease activity and downstream treatment needs.

Observational studies examining the relationship between PPIs and IBD outcomes have yielded inconsistent results. Several population-based analyses and cohort studies have reported associations between PPI use and increased risk of incident IBD or worse clinical outcomes, whereas others have suggested attenuation of these associations after accounting for latency and protopathic bias [ 3 , 5 , 12 ]. Prior studies have varied in their choice of comparator groups and in how exposure timing and duration were defined, which may influence observed associations. Finally, many studies have focused on incident disease or infection risk, with fewer evaluating downstream clinical outcomes such as steroid exposure or escalation to biologic therapy, which may more directly reflect disease activity and treatment burden in patients with established IBD.

Histamine-2 receptor antagonists (H2RAs), another class of acid-suppressive therapy, provide a clinically relevant active comparator, as they are prescribed for similar indications but exert less profound acid suppression and have more modest effects on the gut microbiome [ 13 ]. Comparing PPIs with H2RAs may therefore help disentangle the effects of acid suppression from underlying disease severity and healthcare utilization patterns. In addition, dose–response analyses and a prespecified lag-time (landmark) analysis at 3 months were used to mitigate time-related bias and reduce the influence of early symptom-driven prescribing.

Accordingly, we conducted a large, propensity score-matched cohort study comparing patients with IBD treated with PPIs versus H2RAs. We examined the association between PPI use and subsequent steroid exposure, biologic therapy initiation, C. difficile infection, and emergency room (ER)/inpatient visit across multiple follow-up windows. In the context of personalized medicine, understanding whether commonly prescribed acid-suppressive therapies are associated with differential risks of treatment escalation or infection is directly relevant to individualized treatment decisions in IBD.

2. Methods and Materials

2.1. Data Source

This retrospective cohort study was conducted using the TriNetX [ 14 ] analytics platform (TriNetX, 100 Cambridge Park Drive, Suite 501, Cambridge, MA 02140, USA), a de-identified multicenter, global electronic health record (EHR) database from 172 healthcare organizations (HCOs) across 19 countries, comprising more than 169 million patients. TriNetX enables investigators to design cohort queries, perform propensity-score matching, and compare longitudinal outcomes while ensuring compliance with the Health Insurance Portability and Accountability Act (HIPAA) Privacy Rule (§164.514(a)). To protect patient confidentiality, events involving 10 or fewer patients are suppressed and reported as “≤10”. Because this study used only aggregated, de-identified data, it was exempt from Institutional Review Board (IRB) review. This study adheres to the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guideline for cohort studies [ 15 ] ( Appendix A Table A1 ).

2.2. Study Population, Patient Selection, and Exclusions

We identified all non-deceased patients with active records on the TriNetX platform (1 January 2000–1 March 2025) aged ≥18 years ( Figure 1 ). Patients were required to have a documented diagnosis of IBD, including ulcerative colitis (UC, ICD-10 code K51) or Crohn’s disease (CD, ICD-10 code K50), and subsequent exposure to either PPI (omeprazole, pantoprazole, esomeprazole, lansoprazole, or rabeprazole) or H2RAs (cimetidine, famotidine, nizatidine, or ranitidine) during the study period. Query search criteria were based on standard diagnostic, procedural, or medication codes ( Appendix A Table A2 ).

Patients were excluded if they met any of the following criteria: (1) age <18 years at cohort entry; (2) concomitant use of both PPIs and H2RAs at or before the index date; (3) prior major gastrointestinal surgery, including colectomy or other colon surgery, or receipt of colon-related endoscopic procedures before cohort entry; (4) conditions likely to confound immune status or gastrointestinal outcomes, including immunodeficiency disorders, stem cell transplant, organ and tissue transplant, HIV infection, receipt of antineoplastic agents, celiac disease, rheumatoid disease, diverticular disease, intestinal vascular disease, or infectious gastroenteritis; (5) a documented history of the study outcomes prior to the index date, to ensure assessment of incident events. These exclusion criteria were applied uniformly across all analytic cohorts ( Appendix A Table A2 ).

2.3. Cohort Construction

For the cohort and study design, see Figure 1 . Patients were classified into different exposure cohorts based on medication use at the index date.

(1) Primary-exposure cohort: Patients were classified into (i) a PPI cohort, defined by initiation of any PPI after the first IBD encounter, and (ii) an H2RA cohort, defined by initiation of any H2RA after the first IBD encounter. (2) Dose–response cohort (≥2 prescriptions): To evaluate the impact of sustained exposure, analyses were repeated among patients receiving two or more than two prescriptions of the index medication. (3) Subgroup analysis: Prespecified subgroup analyses were conducted according to IBD subtype (UC or CD). (4) The index date was defined as the first prescription of the index acid-suppressive therapy occurring after the first recorded IBD encounter (The first recorded diagnostic code for UC or CD in the electronic health record). To minimize exposure overlap, patients in the PPI cohort were required to have no H2RA prescriptions after the first IBD encounter, and patients in the H2RA cohort were required to have no PPI prescriptions after the first IBD encounter.

2.4. Outcomes

The primary outcomes of interest were (1) systemic steroid exposure, defined as initiation of oral prednisone (>30 mg daily), intravenous (IV) methylprednisolone, oral or IV hydrocortisone, oral budesonide; (2) initiation of biologic therapy, defined as initiation of monoclonal antibody therapies approved for moderate to severe IBD, including anti TNF agents (infliximab, adalimumab, certolizumab, and golimumab); anti-IL-12/23 or anti-IL-23 agents (ustekinumab and risankizumab); anti-integrin therapy (vedolizumab); (3) C. difficile infection; and (4) ER/Inpatient encounter. Outcomes were identified using diagnosis and prescription codes recorded in the EHR ( Appendix A Table A2 ). Outcomes were assessed during two prespecified follow-up windows: 0–12 months following the index date, and 3–12 months.

2.5. Covariate Selection

Baseline covariates were selected based on clinical relevance and the prior literature. These included demographic characteristics (age, sex, race, and ethnicity); cardiometabolic conditions (hypertension, hyperlipidemia, atherosclerotic cardiovascular disease, heart failure, and peripheral vascular disease); chronic kidney disease; type 1 or 2 diabetes; obesity; gastro-esophageal reflux disease; peptic ulcer disease with/without hemorrhage; medication use (including aspirin, anticoagulants, and high-risk antibiotics); laboratory markers (erythrocyte sedimentation rate, C-reactive protein, and fecal calprotectin); alcohol-related disorder and nicotine-dependent; steroid and immunosuppressant use; and inpatient or ER encounters (all covariates are reported in full in the baseline characteristic Table 1 ) (diagnostic and prescription codes are listed in Appendix A Table A2 ).

2.6. Primary and Sensitivity Statistical Analysis

Primary analyses: To compare 0–12 months’ outcomes between PPI and H2RA cohorts. Analyses were conducted in the matched cohorts. Follow-up time was calculated from the index date until outcome occurrence, loss to follow-up, or the end of the observation period. To minimize time-related bias, all outcome events occurring prior to the index date were excluded to ensure assessment of incident events.

Sensitivity analyses: Prespecified lag-time analyses were performed, evaluating outcomes during 3–12 months of follow-up, thereby reducing the influence of early symptom-driven prescribing and reverse causation. Dose–response analyses restricted to patients receiving ≥2 prescriptions of the index medication were also conducted. Finally, prespecified subgroup analyses were performed according to IBD subtype, including ulcerative colitis (UC) and Crohn’s disease (CD).

2.7. Statistical Analysis

All statistical analyses were conducted on the TriNetX platform or R. Continuous variables are reported as means with standard deviations (SD) and categorical variables as counts and proportions as appropriate. Propensity scores were estimated using logistic regression incorporating all prespecified covariates. Patients were matched 1:1 using nearest-neighbor matching with a fixed caliper. Covariate balance before and after matching was assessed using standardized mean differences (SMDs), with values < 0.1 indicating adequate balance [ 16 , 17 ]. Time-to-event analyses were performed on propensity-matched groups using Kaplan–Meier methods implemented on the TriNetX platform, consistent with the prior literature [ 18 ]. Patients were followed up with until the outcome was reached, or were censored after their last data point on record, or when deceased. Log rank testing was performed on the TriNetX platform to assess statistical differences in time to event for each cohort. Hazard ratios (HRs) were estimated using univariate Cox proportional hazards models implemented on the TriNetX platform. TriNetX validated the use of this model with a scaled Schoenfeld residual, wherein a non-random association between these and time is evidence of a violation of the proportional hazard assumption [ 19 ]. Additionally, logistic regression models carried out analyses to estimate odds ratios (ORs), which represent the relative odds of experiencing an outcome during the specified follow-up period. Statistical significance was inferred when the 95% CI did not cross unity, corresponding to a two-sided p -value < 0.05. Figures were generated in Microsoft Excel and R Studio V.2023.3.0.386 (Boston, MA, USA)

3. Results

3.1. Baseline Characteristics

A total of 169,612,700 patients were identified within the TriNetX research network. After applying inclusion and exclusion criteria, 364,537 patients with IBD were eligible for analysis. Among these, 46,209 patients initiated PPIs and 13,044 initiated H2RAs. After 1:1 propensity score matching, 12,808 patients were included in each exposure group. After the match, the mean age of the cohort was 48.0 ± 19.0 years; 54% were female, and 45% were male. Baseline demographic characteristics, comorbidities, markers of inflammatory activity, and medications were well balanced between the PPI and H2RA cohorts, with standardized mean differences (SMD) < 0.1 for all variables ( Table 1 ).

3.2. Association of PPI Use with Clinical Outcomes

During 0–12 months of follow-up, PPI use was associated with higher ORs and HRs of steroid exposure (OR 1.56, 95% CI 1.35–1.79; HR 1.50, 95% CI 1.31–1.72), initiation of biologic therapy (OR 1.99, 95% CI 1.72–2.29; HR 1.91, 95% CI 1.66–2.19), and C. difficile infection (OR 1.42, 95% CI 1.18–1.70; HR 1.38, 95% CI 1.15–1.64). PPI use was also associated with a higher risk of ER/inpatient visits during this period (OR 1.18, 95% CI 1.03–1.36; HR 1.16, 95% CI 1.01–1.32) ( Figure 2 ). Kaplan–Meier analyses demonstrated higher cumulative incidences of systemic corticosteroid exposure, biologic therapy initiation, and C. difficile infection among PPI users compared with H2RAs users, with statistically significant early separation and sustained divergence of the curves ( Figure 3 ).

Similar associations were observed during the 3–12-month follow-up window, with persistently increased risks for steroid exposure and biologic therapy initiation. In contrast, the associations with C. difficile infection and ER/inpatient visits were attenuated and did not reach statistical significance in time-to-event analyses ( Figure 2 ). The Kaplan–Meier curves are shown in Figure 3 .

Also, crude absolute event proportions were calculated among patients at risk during follow-up. During 0–12 months, systemic steroid initiation occurred in 5.50% of PPI users compared with 3.60% of H2RA users (absolute risk difference 1.90%). Biologic therapy initiation occurred in 5.30% versus 2.74%, respectively (absolute risk difference 2.56%). In the 3–12-month lag-time window, biologic initiation occurred 1.90% of PPI users versus 0.97% of H2RA users (absolute risk difference 0.93%), and systemic steroid exposure occurred in 2.02% versus 1.40% (absolute risk difference biologic initiation occurred 1.90% PPI users versus 0.97% H2RA users 0.62%).

3.3. Dose–Response Analysis (≥2 Prescriptions)

In analyses restricted to patients receiving ≥ 2 prescriptions, 6644 patients were matched in each group. PPI use for 0–12 months was not significantly associated with steroid exposure or ER/inpatient visits. However, the risk of biologic therapy initiation remained increased (OR 1.56, 95% CI 1.31–1.86; HR 1.54, 95% CI 1.30–1.82). The association with C. difficile infection was modest but statistically significant in both odds-based and time-to-event models (OR 1.26, 95% CI 1.02–1.57; HR 1.26, 95% CI 1.02–1.56) ( Table 2 ).

During the 3–12-month period, point estimates for C. difficile infection, steroid use, and ER/inpatient visits remained above unity but were not statistically significant. In contrast, the association with biologic (OR 1.48, 95% CI 1.12–1.97; HR 1.45, 95% CI 1.09–1.93) therapy initiation persisted significantly ( Table 2 ).

3.4. Subgroup Analysis by IBD Subtype

In UC, 5982 patients were matched in each of the PPI and H2RA groups. PPI use was associated with significantly higher risks of steroid exposure, biologic therapy initiation, C. difficile infection, and ER/inpatient visits during both the 0–12-month and 3–12-month follow-up windows ( Figure 4 ).

In CD, 6471 patients were matched in each exposure group. During 0–12 months of follow-up, PPI use was associated with increased risks of steroid exposure, biologic therapy initiation, and C. difficile infection, whereas no significant association was observed for ER/inpatient visits. During the 3–12-month period, associations with steroid exposure and biologic therapy initiation persisted, while associations with C. difficile infection and ER/inpatient visits were attenuated and did not reach statistical significance ( Figure 4 ).

4. Discussion

To our knowledge, this is among the first large, real-world, active-comparator studies to examine the association between PPI use and downstream treatment escalation outcomes, C. difficile infection in patients with established IBD. We found that PPI use was associated with higher risks of treatment escalation and C. difficile infection compared with H2RA use. Across multiple analytic approaches, the most consistent associations were observed for systemic steroid exposure and initiation of biologic therapy outcomes that reflect clinically meaningful disease activity and treatment burden in IBD. These findings extend prior observational work by incorporating active comparator design, incident-event assessment, prespecified follow-up windows, and dose–response analyses to address key sources of bias.

Associations between PPI use and adverse clinical outcomes were strongest during early follow-up (0–12 months). Similar associations were observed during the 3–12-month follow-up window, with persistently increased risks for steroid exposure and biologic therapy initiation. The use of a 3–12 month lag-time window was specifically intended to reduce the influence of early symptom-driven prescribing and reverse causation, which are well-recognized sources of protopathic bias in pharmacoepidemiologic studies [ 20 , 21 ]. However, PPIs are often initiated during periods of symptom exacerbation or early flare-related care. Even with the lag-time analyses, these findings should be interpreted cautiously, and the observed association may partially reflect underlying disease severity rather than direct pharmacologic effect. In contrast, the associations with C. difficile infection and ER/inpatient visits were attenuated in the 3–12 months window. Healthcare utilization outcomes also demonstrated greater heterogeneity across analyses and variability by IBD subtype, which is not unexpected given that such outcomes reflect not only disease activity but also symptom perception, access to care, and health-seeking behavior. In addition, ER/inpatient encounters in TriNetX represent all-cause utilization and cannot be specifically attributed to IBD-related activity, which may also contribute to greater heterogeneity in these outcomes and represent an inherent limitation of EHR-based analyses.

Dose-response analyses of patients receiving two or more than two prescriptions further contextualized these findings. Attenuation of some associations in this subgroup likely reflects exclusion of short-term or situational PPI use initiated during periods of acute illness or disease flare, thereby reducing the influence of early symptom-driving prescribing and protopathic bias. In this more stable exposure group, biologic therapy initiation, an indicator of sustained disease severity, remained consistently associated with PPI use. The persistence of associations in the ≥2-prescription analyses raises the possibility of a more sustained association beyond early symptom-driven prescribing. However, residual confounding by underlying disease trajectory cannot be excluded, and these findings should be interpreted cautiously.

Subgroup analyses further demonstrated phenotype-specific patterns. Among patients with UC, PPI use was associated with higher risks of treatment escalation, C. difficile infection, and healthcare utilization across both 0–12-month and 3–12-month follow-up periods. In contrast, among patients with CD, associations with steroid exposure and biologic initiation persisted, whereas associations with C. difficile infection and ER/inpatient visits were attenuated over time. These differences may reflect underlying phenotypic and pathophysiological distinctions between UC and CD, including differences in disease distribution, mucosal involvement, microbial ecology, and clinical management strategies [ 22 , 23 ]. UC, characterized by continuous colonic involvement, may be particularly susceptible to perturbations in colonic microbiota and mucosal immune responses associated with acid suppression. However, confidence intervals overlapped across subgroups, and formal interaction testing was not performed; therefore, these findings should be interpreted cautiously and not as definitive evidence of differential effects between ulcerative colitis and Crohn’s disease.

Several biologic mechanisms may plausibly link PPI exposure to adverse outcomes in IBD. PPIs induce profound gastric acid suppression, leading to alterations in gut microbiota composition, reduced microbial diversity, and impaired colonization resistance [ 7 , 24 , 25 ]. Recent meta-analysis indicates an increase in the risk of C. difficile infection with increased dose and duration of PPI [ 26 ]. In patients with IBD who already exhibit baseline dysbiosis [ 27 , 28 ], impaired mucosal barrier function, and frequent exposure to antibiotics and immunosuppressive therapies, these microbiome perturbations may be particularly consequential, consistent with our observation of increased C. difficile infection risk in both UC and CD.

Beyond infection risk, multiple pathways may link PPI exposure to treatment escalation. Dysbiosis and C. difficile infection can precipitate mucosal inflammation and disease flares, thereby increasing the need for systemic corticosteroids and subsequent biologic therapy. In addition, PPIs have been shown to impair innate immune function, including neutrophil activity and lysosomal acidification [ 29 ], and to alter antigen processing and T-cell activation. Acid suppression has also been associated with increased intestinal permeability [ 30 ], potentially facilitating luminal antigen translocation and amplifying mucosal immune responses. Together, these mechanisms may lower the threshold for clinically apparent disease activity and accelerate escalation to immunosuppressive therapies.

This study has several strengths, including a large, well-characterized IBD cohort; use of an active comparator to minimize confounding by indication; comprehensive adjustment for demographic, medication, and laboratory covariates; and multiple prespecified sensitivity analyses addressing dose, timing, and protopathic bias. Nonetheless, limitations should be acknowledged. Metabolic comorbidities and broader patient health factors may influence inflammatory bowel disease progression and treatment decisions [ 31 ]. Although several comorbid conditions were included in the propensity matching process, unmeasured or incompletely captured comorbidities may still contribute to residual confounding in real-world observational analyses. Inflammatory biomarkers were incompletely captured, and PPIs initiation may have occurred during early flare-related care. Therefore, observed associations should be interpreted as reflecting treatment patterns within real-world clinical contexts rather than direct causal effects. In addition, treatment escalation endpoints may be associated with provider behavior, regional practice variation, and healthcare access within the TriNetX network. These outcomes, therefore, represent real-world clinical management patterns rather than direct measures of inflammatory severity. The attenuation observed in lag-time and sustained-exposure analyses further supports cautious interpretation of the findings and suggests that early prescribing dynamics and residual confounding may contribute to the observed associations. Medication exposure was based on prescription records and may not fully capture adherence or over-the-counter PPI use. Outcome ascertainment relied on EHR coding, and detailed measures of disease activity, endoscopic findings, and indication for PPI use were not available. Utilization-based outcomes may be influenced by non-IBD-related factors, reflecting a limitation of EHR data. As an observational study, these findings demonstrate associations rather than causation, and residual confounding cannot be excluded.

This study addresses a common clinical dilemma: whether the choice of acid-suppressive therapy matters in patients with IBD who frequently require these medications. By comparing PPIs with H2RAs using a real-world active-comparator design, we provide evidence that therapy selection may be associated with differences in downstream treatment escalation. Although relative risks were elevated, absolute differences were modest. In the 3–12-month window, biologic initiation occurred in 1.90% of PPI users versus 0.97% of H2RA users (absolute difference 0.93%), and systemic steroid exposure differed by 0.62%. Even small differences in escalation risk may influence individualized treatment decisions when multiple acid-suppressive options are clinically appropriate, though causality cannot be established.

5. Conclusions

In patients with IBD who require acid-suppressive therapy, the choice of agent may be clinically relevant. In this real-world cohort, PPI use was associated with higher risks of treatment escalation and C. difficile infection compared with H2RAs use, with the strongest and most consistent associations observed for systemic corticosteroid exposure and biologic therapy initiation. Although some associations attenuated over time, suggesting a contribution of early symptom-driven prescribing, persistent signals for treatment escalation suggest that the findings are not entirely explained by early symptom-driven prescribing; however, residual confounding by underlying disease severity cannot be excluded. When high-potency acid suppression is not clearly indicated, H2RAs may represent a reasonable alternative, particularly in patients at higher risk for disease exacerbation or infection. Together, these results support judicious selection and periodic reassessment of acid-suppressive therapy as part of a personalized management strategy in IBD.

Abbreviations

The following abbreviations are used in this manuscript

Appendix A

Author Contributions

Conceptualization, G.S. and N.W.; Data Curation, Y.S. and D.V.; Investigation, Y.S.; Software, D.V.; Validation, Y.S. and D.V.; Formal Analysis, Y.S. and D.V.; Visualization, Y.S.; Writing—Original Draft Preparation, Y.S.; Writing—Review and Editing, Y.S., G.S. and N.W.; Supervision, G.S. and N.W. All authors have read and agreed to the published version of the manuscript.

Institutional Review Board Statement

The MetroHealth Medical Center Institutional Review Board (IRB) determined this study to be exempt from IRB approval because it uses only de-identified and aggregated data, complying with HIPAA Privacy Rule (§164.514(a)).

Informed Consent Statement

The study was conducted using de-identified retrospective data from the TriNetX global database. As such, informed consent was not required.

Data Availability Statement

All raw output data from TriNetX are available on reasonable requests by contacting the corresponding author.

Conflicts of Interest

The authors have no conflicts of interest to declare.

Funding Statement

This research received no external funding.

Footnotes

Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

References

1. Forgacs I., Loganayagam A. Overprescribing proton pump inhibitors. BMJ. 2008;336:2–3. doi: 10.1136/bmj.39406.449456.BE.

2. Heidelbaugh J.J., Kim A.H., Chang R., Walker P.C. Overutilization of proton-pump inhibitors: What the clinician needs to know. Therap. Adv. Gastroenterol. 2012;5:219–232. doi: 10.1177/1756283X12437358.

3. Bin X., Yang M., Nguyen L.H., He Q., Zhen J., Yu Y., Di M., Qin X., Lu Q., Kuo Z.C., et al. Regular Use of Proton Pump Inhibitor and the Risk of Inflammatory Bowel Disease: Pooled Analysis of 3 Prospective Cohorts. Gastroenterology. 2021;161:1842–1852.e10. doi: 10.1053/j.gastro.2021.08.005.

4. Bavishi C., Dupont H.L. Systematic review: The use of proton pump inhibitors and increased susceptibility to enteric infection. Aliment. Pharmacol. Ther. 2011;34:1269–1281. doi: 10.1111/j.1365-2036.2011.04874.x.

5. Abrahami D., Pradhan R., Yin H., Yanofsky R., McDonald E.M., Bitton A., Azoulay L. Proton pump inhibitors and the risk of inflammatory bowel disease: Population-based cohort study. Gut. 2023;72:1288–1295. doi: 10.1136/gutjnl-2022-328866.

6. Sachs G., Shin J.M., Howden C.W. Review article: The clinical pharmacology of proton pump inhibitors. Aliment. Pharmacol. Ther. 2006;23:2–8. doi: 10.1111/j.1365-2036.2006.02943.x.

7. Imhann F., Bonder M.J., Vila A.V., Fu J., Mujagic Z., Vork L., Tigchelaar E.F., Jankipersadsing S.A., Cenit M.C., Harmsen H.J.M., et al. Proton pump inhibitors affect the gut microbiome. Gut. 2016;65:740–748. doi: 10.1136/gutjnl-2015-310376.

8. Tariq R., Singh S., Gupta A., Pardi D.S., Khanna S. Association of Gastric Acid Suppression with Recurrent Clostridium difficile Infection: A Systematic Review and Meta-analysis. JAMA Intern. Med. 2017;177:784–791. doi: 10.1001/jamainternmed.2017.0212.

9. Deshpande A., Pant C., Pasupuleti V., Rolston D.D.K., Jain A., Deshpande N., Thota P., Sferra T.J., Hernandez A.V. Association between proton pump inhibitor therapy and Clostridium difficile infection in a meta-analysis. Clin. Gastroenterol. Hepatol. 2012;10:225–233. doi: 10.1016/j.cgh.2011.09.030.

10. Foppa C., Rizkala T., Repici A., Hassan C., Spinelli A. Microbiota and IBD: Current knowledge and future perspectives. Dig. Liver Dis. 2024;56:911–922. doi: 10.1016/j.dld.2023.11.015.

11. Stange E.F., Schroeder B.O. Microbiota and mucosal defense in IBD: An update. Expert Rev. Gastroenterol. Hepatol. 2019;13:963–976. doi: 10.1080/17474124.2019.1671822.

12. Choden T., Zhang H., Sakuraba A. Influence of proton pump inhibitor use on clinical outcomes of patients with inflammatory bowel disease. Ann. Med. 2023;55:2198775. doi: 10.1080/07853890.2023.2198775.

13. Lin Y.-T., Lin T.-Y., Hung S.-C., Yiu P.-Y., Wu P.-H., Chuang Y.-S., Hung W.-C., Chiu Y.-W., Kuo M.-C., Wu C.-Y. Anti-Acid Drug Treatment Induces Changes in the Gut Microbiome Composition of Hemodialysis Patients. Microorganisms. 2021;9:286. doi: 10.3390/microorganisms9020286.

14. TriNetX. [(accessed on 10 January 2026)]. Available online: https://trinetx.com .

15. Von Elm E., Altman D.G., Egger M., Pocock S.J., Gøtzsche P.C., Vandenbroucke J.P., STROBE Initiative Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: Guidelines for reporting observational studies. J. Clin. Epidemiol. 2008;61:344–349. doi: 10.1016/j.jclinepi.2007.11.008.

16. Austin P.C. Balance diagnostics for comparing the distribution of baseline covariates between treatment groups in propensity-score matched samples. Stat. Med. 2009;28:3083–3107. doi: 10.1002/sim.3697.

17. Haukoos J.S., Lewis R.J. The Propensity Score. JAMA. 2015;314:1637–1638. doi: 10.1001/jama.2015.13480.

18. Austin P.C. The use of propensity score methods with survival or time-to-event outcomes: Reporting measures of effect similar to those used in randomized experiments. Stat. Med. 2014;33:1242–1258. doi: 10.1002/sim.5984.

19. Zhang Z., Reinikainen J., Adeleke K.A., Pieterse M.E., Groothuis-Oudshoorn C.G.M. Time-varying covariates and coefficients in Cox regression models. Ann. Transl. Med. 2018;6:121. doi: 10.21037/atm.2018.02.12.

20. Tamim H., Monfared A.A., LeLorier J. Application of lag-time into exposure definitions to control for protopathic bias. Pharmacoepidemiol. Drug Saf. 2007;16:250–258. doi: 10.1002/pds.1360.

21. Suissa S., Dell’Aniello S. Time-related biases in pharmacoepidemiology. Pharmacoepidemiol. Drug Saf. 2020;29:1101–1110. doi: 10.1002/pds.5083.

22. Qiu P., Ishimoto T., Fu L., Zhang J., Zhang Z., Liu Y. The Gut Microbiota in Inflammatory Bowel Disease. Front. Cell. Infect. Microbiol. 2022;12:733992. doi: 10.3389/fcimb.2022.733992.

23. Singh N., Bernstein C.N. Environmental risk factors for inflammatory bowel disease. United Eur. Gastroenterol. J. 2022;10:1047–1053. doi: 10.1002/ueg2.12319.

24. Seto C.T., Jeraldo P., Orenstein R., Chia N., DiBaise J.K. Erratum to: Prolonged use of a proton pump inhibitor reduces microbial diversity: Implications for Clostridium difficile susceptibility. Microbiome. 2016;4:10. doi: 10.1186/s40168-016-0158-1.

25. Imhann F., Vila A.V., Bonder M.J., Manosalva A.G.L., Koonen D.P.Y., Fu J., Wijmenga C., Zhernakova A., Weersma R.K. The influence of proton pump inhibitors and other commonly used medication on the gut microbiota. Gut Microbes. 2017;8:351–358. doi: 10.1080/19490976.2017.1284732.

26. Finke M., Boven A., Vlieghe E., Engstrand L., Orsini N., Brusselaers N. Proton pump inhibitors and the risk of Clostridioides difficile infection: A systematic review and dose-response meta-analysis. J. Infect. 2025;90:106488. doi: 10.1016/j.jinf.2025.106488.

27. Vich Vila A., Imhann F., Collij V., Jankipersadsing S.A., Gurry T., Mujagic A., Kurilshikov A., Bonder M.J., Jiang X., Tigchelaar E.F., et al. Gut microbiota composition and functional changes in inflammatory bowel disease and irritable bowel syndrome. Sci. Transl. Med. 2018;10:eaap8914. doi: 10.1126/scitranslmed.aap8914.

28. Lloyd-Price J., Arze C., Ananthakrishnan A.N., Schirmer M., Avila-Pacheco J., Poon T.W., Andrews E., Ajami N.J., Bonham K.S., Brislawn C.J., et al. Multi-omics of the gut microbial ecosystem in inflammatory bowel diseases. Nature. 2019;569:655–662. doi: 10.1038/s41586-019-1237-9.

29. Fowler J.F., Eubank T.A., Garey K.W. Proton pump inhibitor effect on macrophage and neutrophil function: A systematic review. Front. Immunol. 2024;15:1477993. doi: 10.3389/fimmu.2024.1477993.

30. Zhang M., Liu W., Liu Y., Cheng Q., Zhang X., Qian N., Qi J., Chen T., Yu Y. Proton pump inhibitor increases intestinal epithelial paracellular permeability via the p38-MAPK/NF-kappaB signaling pathway. Mol. Biol. Rep. 2025;53:170. doi: 10.1007/s11033-025-11336-y.

31. Massironi S., Solitano V., Furfaro F., D’Amico F., Zilli A., Allocca M., Peyrin-Biroulet L., Jairath V., Danese S. Impact of type 2 diabetes on clinical outcomes and advanced therapy use in patients with Crohn’s disease: A real-world propensity score-matched analysis. J. Crohns Colitis. 2025;19:jjaf157. doi: 10.1093/ecco-jcc/jjaf157.

Associated Data

Data Availability Statement

All raw output data from TriNetX are available on reasonable requests by contacting the corresponding author.