Impact of Proton Pump Inhibitor Use on Progression-Free and Overall Survival in Cancer Patients Undergoing Immune Checkpoint Inhibitor Therapy: A Systematic Review and Meta-Analysis of Recent Studies

Abstract

Background: The introduction of immunotherapy has significantly improved survival outcomes in many solid tumors. However, a subset of patients exhibits limited responsiveness to immune checkpoint inhibitors (ICIs). Emerging evidence indicates that the gut microbiota plays a critical role in modulating the effectiveness of immunotherapy. Consequently, the concurrent use of certain medications that disrupt microbial diversity may contribute to reduced treatment efficacy. Among the agents implicated in altering the gut microbiota are antibiotics and proton pump inhibitors (PPIs). Methods: A systematic literature search was conducted in PubMed, Scopus, and EMBASE. Eligible studies assessed the association between PPI use and progression-free survival (PFS) and/or overall survival (OS) in patients with solid tumors receiving ICIs. They reported hazard ratios (HRs) with 95% confidence intervals (CIs). The analysis focused on studies published between November 2022 and January 2025, in continuity with prior comprehensive meta-analyses that included studies up to November 2022. This contiguity-based approach enabled a focused evaluation of recent evidence, minimizing redundancy while allowing for the detection of evolving trends in clinical practice and methodology. Data were synthesized using both fixed-effects and random-effects models and visualized via Forest plots. Study quality was assessed using the Methodological Index for Non-Randomized Studies (MINORS) and the Newcastle-Ottawa Scale (NOS). Between-study heterogeneity and publication bias were evaluated using I² statistics and funnel plots. Results: From a pool of over 400 screened articles between November 2022 and January 2025, seven studies met the inclusion criteria. The PFS analysis incorporated data from 1367 participants, while the OS analysis included 10,420 individuals. Use of PPIs was linked to a 12% higher risk of disease progression (HR = 1.12; 95% CI: 0.90-1.34) and an 18% increased mortality risk (HR = 1.18; 95% CI: 1.11-1.25). Conclusions: The observed association between PPIs exposure and reduced efficacy of ICIs, as reflected in worsened PFS and OS outcomes, highlights a potential clinical concern that merits further investigation in prospective studies.

Keywords: PPIs; immune checkpoint inhibitors; immunotherapy; microbiota; proton-pump inhibitors; solid tumors.

Figure 1

Flowchart of the study selection process.

Figure 2

(a) The funnel plot displays the standard error on the Y-axis and the HR for OS on the X-axis. The test for heterogeneity yielded a non-significant Cochran’s Q statistic (Q = 5.05, degrees of freedom = 7, p = 0.6545), indicating no substantial between-study variability. The I² statistic, which quantifies the proportion of variance attributable to heterogeneity rather than chance, was 0.0%, with a 95% CI of 0.00% to 55.49%, suggesting negligible inconsistency among the included studies. Evaluation of publication bias showed no significant asymmetry in the funnel plot. Egger’s test yielded an intercept of 0.1371 (95% CI: –1.0092 to 1.2835, p = 0.7796), and Begg’s test based on Kendall’s tau also indicated no evidence of bias (τ = 0.03637, p = 0.8997). Overall, the results suggest a low risk of publication bias and a high degree of consistency across studies reporting HRs for OS. (b) The forest plot illustrates the pooled estimates of the HR for PFS, stratified by PPI use [29,30,32,34,35]. The graphs present these pooled estimates using fixed- and/or random-effects models. To aid interpretation, an arrow on the x-axis indicates the trend of PPI use and its association with varying risk, relative to the unit. Weights, which are based on sample size and estimate precision, reflect the relative contribution of each study to the meta-analysis and are displayed to the right of the graphs. The study by Kawachi H et al. is split into two (D, E) because the authors present separate HRs (with corresponding CIs) for two treatment groups (ICIs, ICIs + chemotherapy). CI—Confidence Interval; HR—Hazard Ratio; ICI—Immune Checkpoint Inhibitor; PFS—Progression-Free Survival; PPIs—Proton Pump Inhibitors.

Figure 3

(a) The funnel plot displays the standard error on the Y-axis and the HR for OS on the X-axis. The test for heterogeneity yielded a non-significant Cochran’s Q statistic (Q = 5.05, degrees of freedom = 7, p = 0.6545), indicating no substantial between-study variability. The I² statistic, which quantifies the proportion of variance attributable to heterogeneity rather than chance, was 0.0%, with a 95% CI of 0.00% to 55.49%, suggesting negligible inconsistency among the included studies. Evaluation of publication bias showed no significant asymmetry in the funnel plot. Egger’s test yielded an intercept of 0.1371 (95% CI: –1.0092 to 1.2835, p = 0.7796), and Begg’s test based on Kendall’s tau also indicated no evidence of bias (τ = 0.03637, p = 0.8997). Overall, the results suggest a low risk of publication bias and a high degree of consistency across studies reporting HRs for OS. (b) The forest plot displays the aggregated HR estimates for OS, with stratification by the use of PPIs [29,30,31,32,33,35]. These estimates are derived using both fixed- and random-effects models. An arrow on the x-axis highlights the direction of PPI usage and its corresponding impact on risk relative to the baseline unit. The weights, shown to the right of the graphs, represent the contribution of each individual study to the overall meta-analysis, adjusted for sample size and the precision of the estimates. The study by Hong S et al. is divided into three sections (A, B, C) as the authors present separate HRs with corresponding CIs for each condition (NSCLC, urothelial carcinoma, malignant melanoma). CI—Confidence Interval; HR—Hazard Ratio; PFS—Progression-Free Survival; PPIs—Proton Pump Inhibitors.