Impact of proton pump inhibitors on efficacy of antiplatelet strategies with ticagrelor or aspirin after percutaneous coronary intervention: Insights from the GLOBAL LEADERS trial

Abstract Background Several studies have suggested that proton pump inhibitors (PPIs) may reduce the antiplatelet effects of clopidogrel and/or aspirin, possibly leading to cardiovascular events. Aims We aimed to investigate the association between PPI and clinical outcomes in patients treated with ticagrelor monotherapy or conventional antiplatelet therapy after percutaneous coronary intervention (PCI). Methods This is a subanalysis of the randomized GLOBAL LEADERS trial, comparing the experimental antiplatelet arm (23‐month ticagrelor monotherapy following 1‐month dual antiplatelet therapy [DAPT]) with the reference arm (12‐month aspirin monotherapy following 12‐month DAPT) after PCI. Patient‐oriented composite endpoints (POCEs: all‐cause mortality, myocardial infarction, stroke, or repeat revascularization) and its components were assessed stratified by PPI use as a time‐dependent covariate in patients with the experiment or reference antiplatelet arm. Results Among 15,839 patients, 2115 patients (13.5%) experienced POCE at 2 years. In the reference arm, the use of PPIs was independently associated with POCE (hazard ratio [HR]: 1.27; 95% confidence interval [CI]: 1.12–1.44) and its individual components, whereas it was not in the experimental arm (HR: 1.04; 95% CI: 0.92–1.19; p interaction = 0.035). During the second‐year follow‐up, patients taking aspirin with PPIs had a significantly higher risk of POCE compared to those on aspirin without PPIs (HR: 1.57; 95% CI: 1.27–1.94), whereas the risk did not differ significantly irrespective of PPI in ticagrelor monotherapy group (HR: 1.03; 95% CI: 0.83–1.28; p interaction = 0.008). Conclusions In contrast to conventional antiplatelet strategy, there were no evidence suggesting the interaction between ticagrelor monotherapy and PPIs on increased cardiovascular events, which should be confirmed in further studies. Clinical Trial Registration URL: https://clinicaltrials.gov


| INTRODUCTION
Antiplatelet therapy has been established as a standard of care in patients with coronary artery disease (CAD), especially after percutaneous coronary intervention (PCI), to reduce the risk of adverse ischemic events, although potent antiplatelet therapies augment the risk of bleeding. 1 Numerous studies have reported conflicting results about the drug interaction between those antiplatelet agents and PPIs. [5][6][7][8][9] Thus far, a few randomized controlled trials comparing antiplatelet therapy with concomitant PPI in patients with cardiovascular disease have questioned whether any significant association exists between PPIs and adverse clinical outcomes when used together with clopidogrel or aspirin after PCI; however, there are scarce data on the interaction between PPIs and more potent P2Y 12 inhibitors, such as ticagrelor. 7 Guideline recommendations on the use of PPIs in patients on dual antiplatelet therapy (DAPT) differ on both sides of the Atlantic.
The European guidelines on DAPT recommend the routine use of PPIs when treated with DAPT (Class I), however they did not endorse extended use after discontinuation of DAPT. 10 The 2016 ACC/AHA guidelines on DAPT only gave a Class I recommendation for PPI use in patients treated with DAPT who had a history of gastrointestinal bleeding; the routine use of PPIs was not recommended (Class III). 2 Similarly, the 2019 ESC guidelines on chronic coronary syndrome (CCS) only recommended PPI use in those with high risk of GI bleeding (Class I). 11 Despite these recommendations, vast numbers of practitioners are still routinely prescribing PPIs without taking into account the potential risks and benefits. 12,13 In the present subgroup analysis of the GLOBAL LEADERS trial, we aimed to investigate the effect of PPIs on clinical outcomes under different antiplatelet regimens, including DAPT, aspirin monotherapy, and ticagrelor monotherapy following PCI in the largest contemporary PCI cohort of an all-comers randomized controlled trial.

| Study design and patient population
The present study is a post hoc subgroup analysis of the GLOBAL LEADERS study (NCT01813435). 14 The details of the trial have been previously reported elsewhere. 14,15 In brief, the GLOBAL LEADERS study was an investigator-initiated, prospective randomized, multicenter, multicontinental, open-label trial designed to evaluate two antiplatelet strategies after PCI using uniformly bivalirudin and a biolimus A9 eluting stent (BioMatrix; Biosensors) in an all-comers population with no restriction regarding clinical presentation (CCS or acute coronary syndrome [ACS] 13 ), complexity of the lesions, or number of stents used. 14 Patients who required oral anticoagulation therapy after PCI, had known overt major bleeding, were planned for surgery within 12 months of PCI, or had severe hepatic impairment were not eligible for the study. In the experimental antiplatelet strategy arm, patients received aspirin 75-100 mg once daily in combination with ticagrelor 90 mg twice daily for 1 month after the index PCI; followed by ticagrelor 90 mg twice daily monotherapy for 23 months (from 1 to 24 months, regardless of the clinical presentation). In the conventional antiplatelet strategy arm (the reference arm), patients received aspirin 75-100 mg daily in combination with either clopidogrel 75 mg once daily in patients with CCS or ticagrelor 90 mg twice daily in patients with ACS for 1 year after the index PCI; followed by aspirin 75-100 mg once daily monotherapy for the following 12 months (from 12 to 24 months).
The institutional review board at each participating institution approved the GLOBAL LEADERS study. All patients provided informed consent. The study complied with the Declaration of Helsinki and Good Clinical Practice. An independent data and safety monitoring committee oversaw the safety of all patients.

| PPI use
Patients had an outpatient clinic visit at 1, 3, 6, 12, 18, and 24 months after the index procedure. The status of PPI use was collected at discharge and subsequent clinical visits. Individual PPI type was not collected. In 20 patients (0.13%) who were hospitalized longer than 1 month after the index procedure, the PPI status at discharge were regarded as same as the PPI status at 1 month. Forty-four patients (0.28%) who died during the index hospitalization were excluded from the present analysis.

| Study end points
The primary endpoint of the present study was the patient-oriented composite endpoint (POCE) within 2 years of randomization, which was defined as a composite of all-cause mortality, any myocardial infarction (MI) (periprocedural or spontaneous), any stroke (ischemic, hemorrhagic, or uncertain), and any revascularization (re-PCI or coronary artery bypass graft surgery [CABG] in target or nontarget vessel) according to the Academic Research Consortium (ARC) II definition. 16,17 The survival status of patients lost to follow-up was obtained through public civil registries. The composite endpoints were analyzed according to time-to-first event analysis. Other endpoints included individual component of POCE, definite stent thrombosis according to the ARC definition, 18 Bleeding Academic Research Consortium (BARC) type 3 or 5 bleeding, 19 BARC type 2 bleeding, and the net adverse clinical events (NACE: defined as POCE plus BARC type 3 or 5 bleeding). 23 All endpoints were investigatorreported without a clinical adjudication committee (CEC).

| Statistical analysis
Continuous variables are expressed as mean ± standard deviation and are compared using independent t-test. Categorical variables are presented as counts and percentages and are compared using χ 2 test or Fisher's exact test as appropriate. The event rates were calculated by using Kaplan-Meier method.
Active PPI treatment was examined as a time-dependent covariate, where if use of a PPI was documented at any time point (index hospital discharge, 1, 3, 6, 12, or 18 months) the patient was treated as continuing to take the PPI until the next timepoint. Clinical outcomes at 2 years were compared between patients taking and The effects of PPI versus no-PPI use on clinical outcomes were also assessed stratified according to the randomly assigned antiplatelet strategies (the experimental antiplatelet strategy arm and the conventional antiplatelet strategy arm). Unadjusted and inverse probability of treatment weighting (IPTW)-adjusted Kaplan-Meier curves were generated to estimate a cumulative incidence of the primary endpoint up to 2 years of follow-up. 21 In addition, landmark analyses with the prespecified timepoint of 1 year (at the time of the planned cessation of a P2Y 12 inhibitor in the conventional strategy) was performed to evaluate the effects of PPI use between ticagrelor monotherapy versus DAPT (up to 12 months), and ticagrelor monotherapy versus aspirin monotherapy (from 12 to 24 months).
Since in the conventional antiplatelet strategy arm, the antiplatelet regimens differed between CCS (clopidogrel + aspirin) and ACS (ticagrelor + aspirin), we also assessed the impact of PPI on clinical outcomes in the two clinical presentation subgroups (CCS and ACS) of the reference arm.
As a sensitivity analysis, we also conducted a propensity-scorematched analysis to compare PPI with no-PPI in the two antiplatelet arms (Supporting Information: Appendix). We used a greedy algorithm to match 1:1 without replacement between PPI and no-PPI use at discharge by using a caliper width of 0.1 SD of the logit of the propensity score.  The baseline patient characteristics between patients with PPI or no-PPI at discharge stratified by the two antiplatelet strategies are shown in Table 1. Regardless of antiplatelet strategies, patients taking PPI were older, were more frequently female, had greater frequency of ACS, had higher prevalence of hypercholesterolemia, PVD, and COPD, had a more frequent history of previous MI and previous PCI, more frequently underwent PCI for LMCAD and MVD, and had higher PRECISE-DAPT scores compared to those with no-PPI at discharge.  (Figures 2 and 3). In the conventional antiplatelet strategy arm, PPI use was also associated with a significantly higher risk in all the components of POCE, including all-cause mortality, any MI, any stroke, and any revascularization, as well as NACE at 2 years.

| Association of PPI with outcomes in the randomized antiplatelet arms
In the experimental ticagrelor monotherapy arm, PPI use was not associated with the risk of POCE at 2 years (adjusted HR: 1.04; 95% CI: 0.92-1.19; p = 0.503; p interaction = 0.035) (Figures 2 and 3). In addition, there was no significant association between PPI use and any clinical endpoints at 2 years ( Figure 3).

| Landmark analyses with timepoint of 12 months
Up to 12 months (the experimental arm: ticagrelor monotherapy following 1-month DAPT, the reference arm: aspirin + clopidogrel/ ticagrelor), PPI use was not associated with the incidence of any endpoints including POCE irrespective of antiplatelet strategies ( p interaction = 0.008) as well as any other clinical endpoints between PPI and no-PPI use (Table 2).
Among ACS patients of the reference arm (ticagrelor + aspirin), PPI use was independently associated with increase in any MI or BARC type 3 or 5 bleeding, whereas among CCS patients of the reference arm (clopidogrel + aspirin), PPI use was not independently associated with increased risks of any clinical endpoints (Table 3).
However, there were no evidence of significant treatment-bysubgroup interactions between PPI use and clinical presentation (CCS or ACS) on any clinical endpoints (all p interaction > 0.05, Table 3).

| Sensitivity analysis
In the propensity score matched cohort (N = 9724), the results were consistent; there was a significant treatment-by-subgroup interaction between PPIs and antiplatelet regimens in terms of POCE at 2 years (p interaction = 0.021), which was mainly driven by the second-year results (Supporting Information: Table 4).

| DISCUSSION
Previous studies proposed different potential mechanisms of how PPIs reduced the antiplatelet efficacy of aspirin or clopidogrel. The pH partition hypothesis has been suggested to explain the drug interaction between aspirin and PPIs. 4,22 PPIs increase intragastric pH by inhibiting the H + /K + -exchanging ATPase of the gastric parietal cells, potentially resulting in a reduced lipophilicity of aspirin, and lowering its gastric absorption. 23 PPIs may inhibit the activity of CYP2C19, which may cause insufficient bioactivation of clopidogrel and an impaired platelet inhibitory effect. 5 However, the COGENT trial, 8 which is the only large randomized trial evaluating the association of PPI use on clinical outcomes among CAD patients treated with clopidogrel on top of aspirin, demonstrated that using omeprazole yielded a significant reduction in the incidence of upper gastrointestinal bleeding without increasing cardiovascular events.
Unlike clopidogrel, ticagrelor is a noncompetitive, direct-acting P2Y 12 -receptor antagonist, and does not require hepatic bioactivation. Theoretically therefore, the efficacy of ticagrelor should not be affected by any PPIs. In fact, in the PLATO PLATELET substudy, PPI use was associated with higher platelet reactivity with clopidogrel but not ticagrelor. 24 Another possible mechanism for the adverse effect of PPIs was proposed by Ghebremariam et al. 25 The authors found that PPIs can interfere with the clearance of asymmetric dimethylarginine, which can reduce nitric oxide synthesis and impair endothelium-dependent vasodilatation, possibly resulting in increased cardiovascular events.
According to this hypothesis, the risk of ischemic events would be increased regardless of the antiplatelet regimen. However, in a substudy of the randomized COMPASS study, where patients were randomly assigned to pantoprazole or placebo, there was no significant difference on cardiovascular death, MI, or stroke at F I G U R E 3 Effect of time-dependent PPI use on clinical outcomes at 2 years in patients treated with experimental antiplatelet strategy or conventional antiplatelet strategy. In the conventional antiplatelet strategy arm, PPI was associated with a significantly higher risks of POCE and the components, including all-cause mortality, any MI, any stroke, and any revascularization at 2 years, whereas in the experimental strategy arm there were no significant association between PPI and any endpoints at 2 years. The event rates were calculated by using Kaplan-Meier method.
The PPI use was treated as a time-dependent covariate. The adjusted covariates are listed in Table 2 suggesting that the adverse effect of PPIs was not evident even treated with antiplatelet therapy with aspirin at least when using pantoprazole.
In the current post hoc analysis of the GLOBAL LEADERS trial, continued use of a PPI in the second-year after PCI was associated with a significant increase in cardiovascular events in patients on aspirin monotherapy, which might be affected by unadjusted confounders in line with previous observational studies. In fact, PPI group had numerically higher bleeding risks than non-PPI group both in the experimental and reference arms, albeit there were no statistical significances. In contrast, however, the PPI use was not associated with adverse events during the first year after PCI in the reference arm, in which DAPT was continued for 12 months (CCS: aspirin and clopidogrel; ACS: aspirin and ticagrelor). We also assessed the impacts of PPI use on clinical outcomes up to 12 months between the two antiplatelet regimens of the reference arm (CCS and ACS), however, there were no evidence suggesting that the effects of PPI use differed depending on the different DAPT regimens (Table 3). In addition, among those treated with ticagrelor monotherapy, there were no significant differences in clinical outcomes between PPI-user and non-PPI-user. Our study therefore could support the safety of PPI use in patients with ticagrelor monotherapy up to 2 years or DAPT up to 1 year after PCI. 10 Our analyses highlight that in the contemporary PCI cohort, the prescription rate of PPIs is nearly 50% despite patient's having a relatively low bleeding risk (e.g., mean PRECISE-DAPT score was 16.4 ± 8.8, 26   advisory boards and/or consulting from Amgen, Ava and Fresenius,