Cardiovascular safety of proton pump inhibitors in post-percutaneous coronary intervention patients receiving dual antiplatelet therapy: A meta-analysis

Abstract

BACKGROUND

Proton pump inhibitors (PPI) are often prescribed alongside dual antiplatelet therapy (DAPT) to mitigate risk of gastrointestinal (GI) bleeding. Despite GI protection, there is emerging concerns about potential interference with antiplatelet efficacy and increased cardiovascular risk.

AIM

To determine the cardiovascular impact of PPI on DAPT efficacy.

METHODS

Online databases were searched from inception until April 2025, and a total of three randomized controlled trials (RCTs) and five observational studies were identified. Data were pooled using a random-effects model, and results were generated using Review Manager in form of risk ratios (RRs) with 95% confidence intervals (CIs).

RESULTS

The RCTs showed no association between PPI use and risk of ACS (RR = 0.84; 95%CI: 0.70-1.00; P = 0.005), cardiac death (RR = 1.25; 95%CI: 0.35-4.38; P = 0.73) and all-cause mortality (RR = 0.77; 95%CI: 0.370-1.60; P = 0.49). Similarly, observational studies confirmed no significant effect of PPI on risk of ACS (RR = 1.22; 95%CI: 0.89-1.66; P = 0.21), cardiac death (RR = 1.30; 95%CI: 0.66-2.58; P = 0.45), all-cause mortality (RR = 1.21; 95%CI: 0.96-1.53; P = 0.10), stent thrombosis (RR = 1.63; 95%CI: 0.83-3.21; P = 0.16), revascularization (RR = 1.08; 95%CI: 0.59-1.97; P = 0.81) and ischemic stroke (RR = 0.82; 95%CI: 0.34-2.02; P = 0.67), yet and increased risk of major adverse cardiovascular event (RR = 1.29; 95%CI: 1.11-1.50; P = 0.001).

CONCLUSION

Even though RCTs confirmed that PPI use is generally safe in post-percutaneous coronary intervention patients receiving DAPT, observational studies raised suspicion of increased risk warranting a need for further studies.

Key Words: Proton pump inhibitors; Dual antiplatelet therapy; Percutaneous coronary intervention; Major cardiovascular adverse events; Cardiac revascularization

Core Tip: This meta-analysis provides a robust, clinically relevant evaluation of proton pump inhibitor use in patients receiving dual antiplatelet therapy after percutaneous coronary intervention by integrating evidence from both randomized trials and real-world observational studies. Importantly, it demonstrates that concerns regarding impaired antiplatelet efficacy do not translate into increased risks of acute coronary syndrome, mortality, stent thrombosis, or stroke in randomized data, suggesting that adverse signals seen in observational studies likely reflect residual confounding rather than true harm. These findings shift the focus from theoretical drug-drug interactions to patient-centered clinical outcomes and support the safe, guideline-concordant use of PPIs for gastrointestinal protection in appropriately selected post-percutaneous coronary intervention patients.

INTRODUCTION

Dual antiplatelet therapy (DAPT), consisting of aspirin and a P2Y12 receptor inhibitor such as clopidogrel, is a cornerstone of secondary prevention following percutaneous coronary intervention (PCI)[1]. By inhibiting platelet activation and aggregation, DAPT significantly reduces the risk of stent thrombosis and recurrent ischemic events, especially in the early post-PCI period[1]. However, the ischemic benefit of DAPT comes at the cost of increased bleeding risk, most notably, gastrointestinal (GI) bleeding, which has led to the widespread adoption of proton pump inhibitor (PPI) co-prescription in clinical practice[2]. Current guidelines recommend PPIs for patients on DAPT who are at heightened risk for GI bleeding, including those of older age or with a history of peptic ulcer disease, concomitant anticoagulant use, or corticosteroid therapy[3].

While PPIs are effective in reducing GI complications, clinical data on the cardiovascular safety of concomitant PPI and clopidogrel use remain conflicting[4]. The randomized COGENT trial found that PPI use significantly reduced GI bleeding without increasing the incidence of cardiovascular events in patients randomized to omeprazole vs placebo[5]. Conversely, multiple observational studies have reported increased risks of major adverse cardiovascular events (MACEs), myocardial infarction, and all-cause mortality among patients concurrently treated with clopidogrel and a PPI[6,7]. These discordant findings have created uncertainty regarding the safety of this common co-prescription, particularly in patients at high cardiovascular risk following PCI.

Given the substantial heterogeneity and inconsistent findings across existing studies, there remains an unmet need for a comprehensive synthesis of the available evidence. Beyond simply pooling data, it is crucial to critically examine whether study design influences the observed association between PPI use and cardiovascular outcomes in patients receiving DAPT following PCI. Therefore, we conducted a systematic review and meta-analysis that uniquely stratified and compared outcomes derived from randomized controlled trials (RCTs) and observational studies to delineate potential differences in effect estimates attributable to methodological variation. The primary outcomes were MACE, all-cause mortality, and stent thrombosis. Secondary outcomes included cardiac death, angina, arrhythmia, heart failure, stroke, recurrent myocardial infarction, and the need for repeat revascularization.

MATERIALS AND METHODS

Study design and objectives

This systematic review and meta-analysis were conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines (Figure 1)[8]. The primary objective was to evaluate the association between PPI use and the risk of cardiovascular outcomes in patients with acute coronary syndrome (ACS) receiving DAPT following PCI. Both RCTs and observational studies comparing clinical outcomes in PPI user’s vs non-users were included. The protocol was registered with International Prospective Register of Systematic Reviews (No. CRD420251138676).

Figure 1  Preferred Reporting Items for Systematic Reviews and Meta-Analyses Diagram Representing Breakdown of Included and Excluded Studies.

Literature search and data sources

A comprehensive literature search was conducted using PubMed, EMBASE, Cochrane Central Register of Controlled Trials, and ClinicalTrials.gov, covering all available records up to April 2025. The search strategy combined Medical Subject Headings and relevant keywords: (“dual antiplatelet therapy” OR “DAPT” OR “clopidogrel”) AND (“proton pump inhibitor” OR “PPI”) AND (“percutaneous coronary intervention” OR “drug eluting stent” OR “PCI”) AND (“MACE” OR “cardiac complications” OR “myocardial infarction” OR “MI” OR “ACS” OR “acute coronary syndrome”).

Eligibility criteria and study selection

Studies were included if they: (1) Enrolled adult patients (≥ 18 years) with ACS who underwent PCI and were treated with DAPT; (2) On aspirin and clopidogrel; (3) Reported at least one of the following outcomes: All-cause mortality, MACE or stent thrombosis; and (4) Were either RCTs or observational (prospective or retrospective) cohort studies. Studies were excluded if they: (1) Lacked a comparator group (PPI vs no PPI); (2) Not published in English; and (3) Were non-human studies. Two independent reviewers conducted title and abstract screening and full-text review. Discrepancies were resolved by discussion or consultation with a third reviewer.

Data extraction and quality assessment

Data were extracted using a predesigned standardized form, including: (1) Study design and setting; (2) Participant demographics and sample size; (3) DAPT regimen; (4) Duration of follow-up; and (5) Clinical endpoints reported. Results were tabulated in Table 1 (study characteristics)[9-17]. The risk of bias was evaluated using the Cochrane risk of bias 2.0 tool for RCTs and the risk of bias in non-randomized studies of interventions scale for observational studies (Figure 2)[18,19]. Discrepancies in bias assessment were discussed among the authors until consensus was achieved.

Figure 2 Risk of bias assessment of included studies. A: Risk of bias in non-randomized studies of interventions scale graph for included observational studies; B: Traffic plot for included observational studies; C: Risk of bias 2.0 tool graph for included observational studies; D: Traffic plot for included observational studies.

Table 1 Study design of included studies.

Ref.	Year	Country	Type of study	Total patients (n)	Study group (n)	Control group (n)	PPI used	Duration of study
Wei et al[9]	2016	China	RCT	207	123	84	Pantoprazole 40 mg	6 months
Vaduganathan et al[10]	2016	Spain and United States	RCT	2676	870	1806	Omeprazole 20 mg	110 days
Ren et al[11]	2011	Chinese	RCT	172	86	86	Omeprazole	30 days
Jensen et al[12]	2017	Denmark	RCT	2009	870	1806	Pantoprazole	1 year
Kim et al[13]	2024	Korea	Retrospective	2266	1133	1133	Non-selective PPI	15 years
Rossini et al[14]	2011	Italy	Retrospective	1328	1158	170	Non-selective PPI	12 months
Yasu et al[15]	2010	Japan	Retrospective	302	103	199	Rabeprazole	395 days
Aihara et al[16]	2012	Japan	Retrospective	1000	500	500	Non-selective PPI	1 year
Chandrasekhar et al[17]	2017	United States and Europe	Retrospective	4635	1062	3573	Non-selective PPI	24 months

RCT: Randomized controlled trial; PPI: Proton pump inhibitor.

Statistical analysis

Meta-analyses were conducted using a random-effects model to account for between-study variability. Pooled effect sizes were reported as risk ratios (RRs) with corresponding 95% confidence intervals (CIs). Heterogeneity was quantified using the I² statistic, with I² > 50% considered substantial[20]. A two-sided P < 0.05 was considered statistically significant.

Certainty of evidence

The certainty of the evidence for each outcome was assessed using the GRADE (Grading of Recommendations, Assessment, Development, and Evaluation) framework[21]. Each outcome was rated as high, moderate, low, or very low based on study design, risk of bias, inconsistency, indirectness, and imprecision. Results are presented in Table 2 and Supplementary Table 1.

Table 2 Baseline characteristics of included population, mean ± SD/n (%).

Ref.	Age		Male gender		Hypertension		Diabetes mellitus
	Study group (years)	Control group (years)	Study group	Control group	Study group	Control group	Study group	Control group
Vaduganathan et al[10], 2016	NA	NA	NA	NA	NA	NA	NA	NA
Wei et al[9], 2016	59.32 ± 9.14	58.47 ± 10	69 (56.1)	48 (57.1)			502 (50.4)	496 (49.0)
Yasu et al[15], 2010	69.9 ± 9.6	67.4 ± 10.1	69 (67.0)	144 (72.4)	66 (64.1)	129 (64.8)	70 (68.0)	115 (57.8)
Kim et al[13], 2024	65.1 ± 11.2	65.6 ± 11.2	754 (66.5)	745 (65.8)	677 (59.8)	705 (62.2)	762 (65.8)	128 (72.5)
Jensen et al[12], 2017	64.7 ± 10.2	64.8 ± 10.6	729 (73.1)	758 (74.9)	535 (53.7)	546 (54.0)	113 (11.3)	103 (10.2)
Rossini et al[14], 2011	64 ± 11	63 ± 11	875 (75.6)	138 (81.2)	736 (63.6)	111 (65.2)	14338 (40.3)	13975 (39.3)
Ren et al[11], 2011	62.08 ± 10.62	61.84 ± 11.21	62	63			415 (83)	419 (83.8)
Aihara et al[16], 2012	68 ± 11	69 ± 10	363 (72.6)	379 (75.8)	356 (71.2)	345 (69.0)	205 (61.9)	579 (61.7)
Chandrasekhar et al[17], 2017	65.4 ± 11.1	64.1 ± 11.4	2686 (75.2)	738 (69.5)	847 (79.8)	2873 (80.4)	349 (32.9)	1171 (32.8)

NA: Not available.

RESULTS

This meta-analysis included 9 studies with a total of 12586 patients, comprising 5035 patients receiving PPI therapy (PPI group) and 7551 patients on DAPT without PPIs (non-PPI group). The included studies consisted of 4 RCTs and 5 observational retrospective cohorts.

Primary outcomes

ACS: Pooled analysis of RCTs only showed no significant difference in the risk of ACS between the PPI and non-PPI groups (RR = 0.84 ; 95%CI: 0.70-1.00; P = 0.05; I² = 0%; Figure 3), similarly analysis of observational studies confirmed no impact of PPI on risk of ACS (RR = 1.22; 95%CI: 0.89-1.66; P = 0.21; I² = 0%; Figure 3).

Figure 3 Forest plot assessing the risk of acute coronary syndrome associated with proton pump inhibitors in comparison with placebo. A: Randomized controlled trials only; B: Observational studies only. PPI: Proton pump inhibitor; M-H: Mantel-Haenszel; CI: Confidence interval.

All-cause mortality: RCTs demonstrated no significant association between PPI use and risk of all-cause mortality (RR = 0.77; 95%CI: 0.37-1.60; P = 0.49; I² = 0%). Observational studies depicted a similar impact of PPI use on all-cause mortality (RR = 1.21; 95%CI: 0.96-1.53; P = 0.10; I² = 0%; Figure 4).

Figure 4 Forest plot assessing the risk of all-cause mortality associated with proton pump inhibitors in comparison with placebo. A: Randomized controlled trials only; B: Observational studies only. PPI: Proton pump inhibitor; M-H: Mantel-Haenszel; CI: Confidence interval.

Secondary outcomes

Cardiac death: Both RCTs (RR = 1.25; 95%CI: 0.35-4.38; P = 0.73; I² = 0%) and observational studies (RR = 1.30; 95%CI: 0.66-2.58; P = 0.45; I² = 0%) showed that PPI use was no associated with significant risk of cardiac death (Figure 5).

Figure 5 Forest plot assessing the risk of cardiac mortality associated with proton pump inhibitors in comparison with placebo. A: Randomized controlled trials only; B: Observational studies only. PPI: Proton pump inhibitor; M-H: Mantel-Haenszel; CI: Confidence interval.

Observational studies only analysis: The risk of MACE was significantly increased in post-PCI patients using PPI (RR = 1.29; 95%CI: 1.11-1.50; P = 0.001; I² = 0%). However, pooled analysis of observational studies failed to show any significant impact of PPI use on stent thrombosis (RR = 1.63; 95%CI: 0.83-3.21; P = 0.16; I² = 0%), coronary revascularization (RR = 1.08; 95%CI: 0.59-1.97; P = 0.81; I² = 77%) and ischemic stroke (RR = 0.82; 95%CI: 0.34-2.02; P = 0.67; I² = 38%; Figure 6).

Figure 6 Observational studies only forest plot assessing the impact of proton pump inhibitors. A: Major adverse cardiovascular event; B: Stent thrombosis; C: Coronary revascularization; D: Ischemic stroke. PPI: Proton pump inhibitor; M-H: Mantel-Haenszel; CI: Confidence interval.

DISCUSSION

Summarization of results

Our meta-analysis demonstrated that the use of PPIs in patients receiving DAPT after PCI was associated with a significant increase in MACE (seen only in observational studies). However, we failed to find any notable signal for increased stent thrombosis, all-cause mortality, cardiac mortality, need for revascularization, ACS, or stroke. Collectively, these results suggest that there is no adverse cardiovascular impact of PPIs on DAPT-treated patients.

Pathophysiology behind the impact of PPI on DAPT

The mechanistic concern regarding PPI use in patients on DAPT primarily stems from clopidogrel’s metabolism. As a prodrug, clopidogrel requires biotransformation via the cytochrome P450 (CYP) system, particularly CYP2C19, into its active metabolite (thiol) that irreversibly inhibits platelet P2Y12 receptors. PPIs, most notably omeprazole and esomeprazole, are competitive inhibitors of CYP2C19 and can reduce the efficiency of this activation, leading to attenuated platelet inhibition and increased platelet reactivity[22]. The underlying genetic polymorphisms in CYP2C19 also play a role, as carriers of the loss-of-function allele may be more sensitive to any further inhibition imposed by PPIs[23]. Since stent thrombosis is an event highly dependent on platelet inhibition during the early endothelial healing phase, even a modest reduction in clopidogrel activation can have clinically significant consequences[24]. This effect is likely more pronounced in patients at high thrombotic risk, such as those with ACS, multiple stents, or diabetes, where the margin for platelet reactivity is narrow. This effect failed to be translated in our analysis, most likely due to small number of included participants.

However, the clinical expression of this interaction appears to be selective and variable. Outcomes that rely heavily on uninterrupted platelet inhibition, such as early thrombotic complications, may be more sensitive to this interaction, whereas broader endpoints like myocardial infarction, stroke, or mortality are influenced by multiple overlapping mechanisms, including progression of atherosclerosis, systemic comorbidities, and competing non-cardiovascular risks[25]. In addition, protective therapies frequently prescribed in this population, such as statins, angiotensin-converting enzyme inhibitors, and beta-blockers, may attenuate small differences in platelet reactivity[26]. Variability among PPIs further complicates interpretation: Omeprazole and esomeprazole are stronger CYP2C19 inhibitors, while pantoprazole and rabeprazole exert weaker effects, potentially leading to heterogeneous findings across studies[27].

Discrepancies between study designs

Through our analysis, we found a noble discrepancy between the two study types, with pooled analysis of observational studies demonstrating a statistically significant association between PPI and higher rates of MACE, an effect not translated in RCTS. This variation is likely in the setting of methodological differences between the two study deigns. Observational studies, being unblinded and non-randomized, are susceptible to biases such as selection bias. Since PPI are more likely to be prescribe to patients who are sicker, have higher baseline cardiovascular risk, increased risk of bleeding and higher comorbidities, factors that can individually increase risk of complications. Furthermore, residual confounders such as frailty, lack of compliance to medication and diet, or severity of cardiac disease may not be well balanced among the study and control groups. On the contrary, RCTs minimize both measured and unmeasured confounders through randomization and blinding, thus providing more accurate results. The absence of increased MACE risk in RCTS likely reflect patient-level risk factors rather than a true deleterious cardiovascular effect of PPI therapy. These findings collectively underscore the importance of interpreting results with caution.

Comparison with existing literature

Our findings were compliant with those of Li et al[28], who concluded that PPIs did not increase the risk of MACE in a similar population. Similarly, our observation of non-significant increased risk of stent thrombosis supports Luo et al[29] who reported no significant differences between PPI and non-PPI groups. For mortality, our results align with Li et al[28] and Luo et al[29], both of whom found no difference in all-cause death, whereas Melloni et al[30] reported increased mortality with PPI use. The neutral effect we observed on ACS is consistent with the meta-analysis by Saeed et al[31], while stroke outcomes mirror those of Khan et al[32], who found no significant increase in ischemic or hemorrhagic stroke. Collectively, these comparisons emphasize the variability in published evidence and highlight the importance of pooled analyses that integrate both RCTs and observational cohorts.

Limitations and strengths

This meta-analysis has a few important limitations. First, a substantial proportion of the included studies were observational in nature, introducing potential for residual confounding, selection bias, and treatment indication bias. For example, patients prescribed PPIs may have had more comorbidities or prior bleeding risk factors, which could confound cardiovascular outcomes. Inclusion of observational studies was important to better understand side effects of PPI. Second, the types and dosages of PPIs varied across studies, with some using omeprazole (a potent CYP2C19 inhibitor) and others using pantoprazole or rabeprazole, which may have different pharmacokinetic profiles. Lack of subgroup results prevented us from performing meta-regression and subgroup analysis. This heterogeneity limits the ability to generalize findings to all PPIs equally or do subgroup analysis. Moreover, several outcomes such as stroke, heart failure, and angina were not reported uniformly across studies, leading to potential underpowering of subgroup analyses. Lastly, long-term outcomes beyond 1-2 years were infrequently reported, restricting assessment of the durability of PPI-related cardiovascular effects over time.

Despite these limitations, this study offers several strengths. It is one of the few comprehensive analyses to comparing RCTs and large observational cohorts examining the cardiovascular effects of PPI use in DAPT-treated patients post-PCI. By including data from diverse geographic populations and various clinical settings, the findings offer broad external validity. Additionally, the inclusion of multiple outcomes, ranging from stent thrombosis to ACS and stroke, provides a more nuanced understanding of the net clinical effect of PPIs. The methodological rigor in study selection, quality assessment, and stratified analysis enhances the reliability of our conclusions.

CONCLUSION

In conclusion, PPI use in patients receiving DAPT after PCI was not associated with increased risk of ACS, stroke, cardiac death, coronary revascularization, stent thrombosis or all-cause mortality. Although, an increased risk of MACE was seen in observational studies, this association not confirmed in RCTs and is likely attributed to residual confounders associated with the study design, limiting it generalizability. While PPIs remain essential for GI protection in patients at elevated bleeding risk, the available evidence suggests that their use does not confer a clinically meaningful increase in adverse cardiovascular outcomes when appropriately prescribed. Collectively, these findings support the cardiovascular safety of PPI co-therapy following PCI and underscore the importance of individualized risk-benefit assessment, with careful PPI selection when indicated. Although further prospective studies may help refine the choice and duration of therapy, and clarify the discrepancy associated with risk of MACE.