Cancer immunotherapy based on immune checkpoint inhibitors (ICIs) has dramatically revolutionized the current therapeutic strategies of advanced solid tumors. Generally, three representative ICIs have been commonly applied in clinical practice, including programmed cell death 1 (PD-1), programmed cell death ligand 1 (PD-L1), and cytotoxic drug T lymphocyte-associated antigen 4 (CTLA-4) inhibitors (1). However, despite encouraging results from clinical trials and the real world, a fairly large proportion of patients fail to benefit from ICI therapy due to various inherent and environmental factors, such as microsatellite instability, gut microbiota, and concomitant medications (2–4). Previously, our team has demonstrated that antibiotic administration shortly before or after ICI treatment is associated with poor clinical outcomes in patients with advanced solid cancers (5). Moreover, our team has also identified corticosteroid administration for cancer-related indications as an unfavorable prognostic factor in ICI-treated patients (6). Therefore, a further investigation into the correlation between concomitant medications and ICI efficacy will contribute to more individualized management in patients treated with ICI drugs, therefore benefiting their overall prognosis.

Proton pump inhibitors (PPIs) are the most frequently used drugs for the treatment of gastroesophageal reflux disease and peptic ulcers. Functionally, PPIs bind to the H+/K+-ATPases in gastric oxyntic cells to reduce the secretion of gastric acid (7). However, the gut microbiome-based analysis revealed that PPI use may increase the abundance of some potential pathogenic bacteria such as Escherichia coli, resulting in an increased risk of enteric infections (8). In cancer epidemiology, recent evidence has indicated that PPI use may be correlated with an increased risk of gastrointestinal malignancies (9, 10). With regard to cancer therapy, on the contrary, PPIs not only function as direct anticancer agents through modifying acidic tumor microenvironment but also enhance the chemosensitivity of cancer cells (11). Therefore, the role of PPIs in the cancer field remains inconclusive, and additional investigations are of great clinical significance.

Recently, emerging studies have found that PPIs may diminish the clinical efficacy of ICIs in patients with advanced solid cancers. For instance, a multicenter retrospective study has demonstrated PPI use was significantly correlated with worse overall survival (OS), progression-free survival (PFS), and lower disease control rate (DCR) and objective response rate (ORR) in patients with locally advanced or metastatic urothelial carcinoma (12). Consistently, another retrospective study based on nationwide data has shown that concomitant PPI use was correlated with reduced OS in non-small cell lung cancer (NSCLC) patients treated with ICIs (13). However, a recent study has found that PPI use was not correlated with ORR, PFS, or OS in advanced renal cell carcinoma (RCC) patients treated with anti-PD-1 therapy alone or plus anti-CTLA-4 therapy (14). Furthermore, researchers have even found that PPI use was significantly correlated with prolonged median PFS in metastatic melanoma patients treated with ICI alone or combined ICIs (15). Considering these inconsistent results, in this study, we performed an up-to-date systematic literature review and meta-analysis to evaluate the prognostic significance of PPI use in ICI-treated patients with advanced solid cancers. Our study will not only contribute to more precise PPI management in clinical practice but also further emphasize the non-negligible correlation of concomitant medications with ICI efficacy.

Despite the recognition for ICI-based cancer immunotherapies by the 2018 Nobel Prize, in clinical practice, the majority of patients fail to benefit from these therapies, even in combined therapies with two ICI drugs and/or chemoradiotherapy (32). A recent review has proposed that the anticancer efficacy of ICI drugs may be affected by some concomitant medications such as antibiotics, corticosteroids, PPIs, and metformin (3). For instance, antibiotic administration was proved to diminish the clinical efficacy of ICI drugs by influencing the gut microbiome (33). Similar to antibiotics, PPIs were also found to significantly affect the composition of the gut microbiota and are negatively associated with ICI efficacy (23, 34, 35). However, there are some studies suggesting that PPI use was unrelated to ICI efficacy (14, 28). Therefore, in this study, we performed an up-to-date meta-analysis to clarify the correlation between PPI use and ICI efficacy in advanced solid cancers. During the preparation of our manuscript, it is worth mentioning that three similar studies have been published so far (36–38). Compared with them, on the one hand, our present work is an up-to-date analysis including more studies (n = 17 vs. n = 7 for Li et al., n = 5 for Li et al., and n = 7 for Qin et al.) and more patients (n = 9,978 vs. n = 1,482 for Li et al., n = 1,167 for Li et al., and n = 3,647 for Qin et al.), especially including 8 studies published in 2021, which have not been included in these three studies. On the other hand, in addition to cancer types, our subgroup analysis was performed based on more factors including region, age, sample size, therapeutic strategy, and PPI exposure timeframes, contributing to further understanding of the actual role of PPIs during ICI therapy. Moreover, the results varied wildly among these three meta-analyses. Therefore, our up-to-date investigation is of great necessity and hoped to provide novel insights into the precise management of PPIs in clinical practice.

In our study, we firstly found that PPI use was significantly correlated with worse OS in patients treated with ICI therapy, while no significant correlation was found between it and PFS. The two meta-analyses published in 2020 have collectively demonstrated that PPI use was unrelated to OS nor PFS of patients treated with ICI drugs (36, 37). Both the studies have attributed their results to the sophisticated role of PPIs in cancer therapy. For instance, PPIs may disturb the gut microbiome that was related to ICI efficacy and was suggested to be discontinued or switched to a histamine H2-receptor antagonist (39). However, accumulating evidence demonstrated PPIs not only inhibit tumor growth and enhance chemosensitivity through modulating an acid milieu but also promote immune reaction and prevent tumor immune escape (40). Furthermore, compared with antibiotics, the impact of PPIs on the gut microbiome is relatively limited, and there is a lack of sequencing studies to identify dominant bacteria affected by PPIs in ICI non-responders. On the contrary, an updated investigation by Qin et al. in 2021 demonstrated a significant negative correlation of PPI use with OS and PFS in advanced cancer patients treated with ICIs (38). Despite the lack of supporting mechanism investigations, Qin et al. hypothesized that PPIs may decrease ICI efficacy by altering the abundance of some bacteria via increasing pH levels. Our present result revealed that PPI use was unrelated to ICI efficacy, which is consistent with 5 of 11 included studies (21, 25, 28, 30, 31). Despite the fact that PFS is superior to OS in evaluating therapy response, however, we were unable to assert that PPI use was unrelated to PFS. In the future, multicenter validations based on large samples are still essential, and cautious use of PPIs for ICI-treated patients is advocated.

To further investigate the correlation of PPI with ICI efficacy, the subgroup was stratified according to region, cancer type, sample size, the timeframe of PPI exposure, and immunotherapy drug. In region subgroups, PPI use was significantly correlated with worse OS and PFS of patients from worldwide multicenter studies, but the number of included studies is limited (n = 2–3), and more cross-continent multicenter cooperation is encouraged. In cancer-type subgroups, the negative correlation of PPI use with PFS was significant in patients with NSCLC and urothelial carcinoma, while it was the opposite in patients with melanoma. Metagenomic evidence has proved Ruminococcaceae was enriched in the stools of NSCLC patients benefiting from ICIs, while PPIs were known to be negative with Ruminococcaceae abundance (4, 41). Therefore, PPIs may diminish the clinical efficacy of ICIs partly through modulating gut microbiota in NSCLC. A previous study has found esomeprazole as a PPI that inhibited the malignant progression of melanoma through modifying the acidic microenvironment (42). A similar study also found that esomeprazole inhibited the growth of melanoma through downregulating VEGF-C expression via inactivating NF-κB (43). Both the studies supported the hypothesis that PPIs may directly exert their anticancer role rather than affecting ICI efficacy in melanoma, partly explaining the opposite results in the subgroup analysis. In age subgroups, PPI use was significantly correlated with worse OS and PFS only in patients over 65 years old. Considering the fact that the elderly received PPI therapy more frequently than other age groups, PPIs should be especially cautiously used or switched to histamine H2-receptor antagonists in these patents during immunotherapy. In terms of sample size, we found that the negative correlation of PPIs with OS/PFS was significant in studies enrolling more than 200 patients, although the included studies were limited. This finding highlights the importance of sample size in determining the prognostic role of PPIs and encourages large-scale clinical validations in the future. Since the timeframes of PPI exposure varied among studies, the subgroup analysis was subsequently performed to clarify whether the unfavorable prognostic impact of PPIs was dependent on their exposure timeframes (18, 19, 22). As result, we found that the negative correlation of PPIs with OS and PFS was significant in patients receiving PPIs 30 days before and/or after ICI initiation. This finding suggests that PPIs should be discontinued shortly before or after starting ICI therapy in order to avoid their potential unfavorable impact. Finally, we noted the unfavorable prognostic impact of PPIs was significant in patients receiving anti-PD-L1 therapies. However, considering limited included studies in each subgroup, more retrospective investigations are needed, which may help further clarify the correlation between PPIs and the efficacy of various types of ICIs.

There are several limitations to our study. First, this is a meta-analysis based on previously published retrospective studies. During data extraction, much important information such as HR values of OS/PFS, PPI types, PPI exposure timeframes, and ICI therapeutic strategies was unavailable or not detailed. This limitation partly impedes our global and subgroup analyses and is hoped to be improved by following up-to-date investigations based on more high-quality literatures. Second, the included studies were restricted to literatures published in English, which may miss some potential eligible ones published in other languages such as Chinese. In the future, international collaborations are highly encouraged and will benefit from a more comprehensive literature search. Third, various inherent factors of included retrospective studies such as patient selection, drug types/doses, and therapeutic strategies may also impact our result. Fourth, our subgroup analysis mainly focused on NSCLC, melanoma, and urothelial carcinoma due to limited studies. The correlation of PPIs with ICI efficacy in other cancers such as digestive malignancies is worth investigating. Fifth, since our finding is partially inconsistent with other similar studies, one or more independent cohorts from our hospital should be utilized for further validation. Finally, the direct evidence supporting the correlation between PPI use and ICI efficacy is still insufficient. For instance, to our knowledge, there are no studies investigating the changes of the gut microbiome in NSCLC patients receiving PPIs during ICI therapy, nor the experiments validating the beneficial role of PPIs in enhancing the anticancer efficacy of ICI drugs in melanoma. Therefore, mechanism investigations based on multi-omics detection and animal experiment will be emphasized in our following work.

In conclusion, based on updated evidence, our study demonstrated that concomitant PPI was significantly correlated with worse OS instead of PFS in advanced solid cancer patients receiving ICIs. The subgroup analysis revealed the negative correlation of PPI use with PFS was significant in NSCLC and urothelial carcinoma, while it was the opposite for melanoma. Furthermore, the unfavorable prognostic impact of PPIs was observed in patients receiving anti-PD-L1/PD-1 therapies or those receiving PPIs 30 days before and/or after ICI initiation. Therefore, although our global analysis failed to confirm the negative correlation of PPIs with PFS, cautious use of PPIs is still recommended to avoid their potentially detrimental impact on ICI efficacy. In addition, more clinical validations and related up-to-date meta-analyses are necessary, and mechanism investigations in vivo and in vitro will be helpful in elucidating the complicated role of PPIs in cancer immunotherapy.

The original contributions presented in the study are included in the article/supplementary material. Further inquiries can be directed to the corresponding authors.

XY and MY: conceptualization and methodology. CL and HG: data collection and analysis and writing—original draft preparation. HM and JT: data validation and writing—reviewing and editing.

This study was supported by the Traditional Chinese Medicine Science and Technology Development Plan Project of Jiangsu (No. YB2020107) and Yangzhou Science and Technology Bureau (No. YZ2020085).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.