Combining DAPT With a PPI Faces the Acid Test of Real-world Use

William A.E. Parker; Robert F. Storey


Eur Heart J. 2019;40(24):1971-1974. 

Dual antiplatelet therapy (DAPT) with aspirin and a P2Y12 inhibitor significantly reduces the risk of major adverse cardiovascular events (MACE) after myocardial infarction (MI), a benefit that persists with long-term use in high-risk patients.[1] The main drawback of DAPT remains an increased risk of bleeding events that lead to distress, discontinuation of therapy, and, most crucially, increased mortality.[2] One significant source of bleeding is the upper gastrointestinal (GI) tract. Whilst spontaneous upper GI haemorrhage can of course occur in drug-free individuals, its incidence is increased in those receiving DAPT.[1] As well as the cumulative anti-haemostatic effects of aspirin and P2Y12 inhibition, aspirin has the direct effect of upper GI irritation and commonly ulceration, mediated through cyclo-oxygenase 1 inhibition that reduces prostaglandin E2 synthesis, responsible for stimulating protective alkaline mucus secretion and reducing acid production.[3] By lowering the acidity in the gastric and duodenal lumen, proton pump inhibitors (PPIs) reduce the risk of peptic ulceration and may also lessen the severity of bleeding in the upper GI tract, should it occur, by enhancing clot stability (Figure 1).[4,5] Combining DAPT with a PPI therefore presents a feasible and biologically plausible strategy to reduce bleeding originating from the upper GI tract, thus allowing continued reduction in ischaemic risk. This hypothesis was supported by a previous randomized controlled trial, albeit with limited power to detect differences in MACE.[6]

Figure 1.

Potential interactions between aspirin, P2Y12 inhibitors, and proton pump inhibitors in patients with myocardial infarction. ADP, adenosine diphosphate; ATPase, adenosine triphosphatase (proton pump); CAM, clopidogrel active metabolite; Cl, chloride ion; COX1, cyclo-oxygenase 1; DAPT, dual antiplatelet therapy; H+, hydrogen ion (proton); HCO 3, bicarbonate ion; K+, potassium ion; Mg2+, magnesium ion; PAM, prasugrel active metabolite; P2Y12, platelet ADP receptor; PGE2, prostaglandin E2; TAM, ticagrelor active metabolite; TPα, thromboxane receptor α; TXA2, thromboxane A2.

Recent years have seen an evolution in guidelines towards more widespread recommendation of PPI prescription in patients receiving DAPT for MI, with the 2017 ESC update giving a class I recommendation for unrestricted PPI use with DAPT.[7]

Until now, there have been few 'real-world' data to support the prescription of a PPI in combination with DAPT after MI, but, in this issue of the European Heart Journal, Sehested and colleagues report a retrospective analysis of >46 000 patients receiving DAPT after MI.[8] They ask, and elegantly answer, two main questions.

First, has the co-prescription of PPIs and DAPT to MI patients become more prevalent over time in line with the evolution of guidelines on the subject, assessed from 2003 to 2014, in a European country? Whilst the time period covered pre-dates the latest position from the ESC, it is clear from this analysis that PPIs have indeed become more commonly prescribed alongside DAPT for MI, but this has lagged behind the pace of changes in the recommendations such that overall rates of use remain relatively low when considering the guidelines. At the end of the analysed period in 2014, only 42% of high-risk patients were prescribed PPIs, when contemporary ESC guidelines would have recommended blanket use in this group,[9] although this did increase from 23% in 2003. There was a similar trend in the low-risk group but, as might be expected, a lower starting and finishing point (8% to 20%).

Secondly, they explored whether PPI prescription improved bleeding outcomes in patients at ESC-defined high and low risk of bleeding. They show a clear benefit in terms of clinical bleeding outcomes in those receiving a PPI vs. those not, with an overall absolute risk ratio of 0.62 [95% confidence interval (CI) 0.48–0.77] and absolute risk difference of 0.44% (95% CI 0.39–0.48). Moreover, the benefit was maintained throughout the ESC-defined high- and low-risk subgroups; in fact, the absolute risk ratio was more favourable in low-risk patients (0.49, 95% CI 0.31–0.73) vs. (0.76, 95% 0.53–1.06) in the high-risk group, with an almost identical absolute risk difference of 0.46% (95% CI 0.40–0.52) if high risk vs. 0.47% (95% 0.43–0.51) if low risk. On the one hand, as the authors identify, this may mean we need better scores to identify those at the highest risk of bleeding but, on the other hand, it is of less relevance if advocating universal prescription of PPIs with DAPT.

The time period included in the analysis was fortunate to capture the transition from clopidogrel to ticagrelor or prasugrel as first-line agents: there was no significant increase in risk of upper GI bleeding with either agent compared with clopidogrel, which provides reassurance that more potent P2Y12 inhibitors can be used preferentially even in those at high risk of upper GI bleeding.

When considering potential benefits of PPIs in this patient group, we should also consider any negative effects that might detract from any positive effects on upper GI bleeding (Figure 1). Notably, hypomagnesaemia, caused by reduced absorption of this essential ion from the alimentary canal, has frequently been associated with PPI therapy,[10] with potential sequelae including increased risk of arrhythmia,[11] which, considering the elevated baseline risk in this population, may be of particular concern. There has also been some concern that hypomagnesaemia may increase the risk of fractures,[12] with the associated risks of traumatic bleeding and interruption of DAPT, potentially leading to ischaemic events. Similarly, an association between PPIs and chronic kidney disease has been reported.[13] Currently, the evidence for an adverse effect of PPIs on renal function is relatively weak, but this area warrants further scrutiny in view of the increased overall cardiovascular risk attributable to chronic kidney disease.

Finally, the hepatic conversion of clopidogrel, but not other P2Y12 inhibitors, to its active metabolite is inhibited by some PPIs, most notably omeprazole and esomeprazole but potentially also other PPIs that inhibit hepatic cytochrome P450 (CYP)2C19. It remains unclear whether this translates directly into an increased risk of MACE due to possible confounders that might affect any observational study examining the issue of the effect of PPI prescription on bleeding outcomes, including the present analysis.[14] Nevertheless, caution should be exercised when co-prescribing PPIs with clopidogrel, and using either a PPI that does not significantly inhibit CYP2C19, such as pantoprazole, or, in patients with MI, a different P2Y12inhibitor seems sensible. In the present analysis, there was no consideration of whether PPIs affected cardiovascular outcomes, but this merits further exploration.

The latest recommendation by the ESC comes about after the time period covered by the present analysis, but appears to be supported by it, in the absence of safety data. However, to contextualize these findings fully, they should be considered alongside the results of a recently presented randomized controlled trial of 17 598 patients with a history of ischaemic heart disease receiving pantoprazole or placebo alongside aspirin, rivaroxaban, or a combination of the two: the Cardiovascular OutcoMes for People using Anticoagulation StrategieS (COMPASS) PPI study. This showed disappointingly little effect of a PPI compared with placebo on a composite primary endpoint made up of various upper GI bleeding-related events [hazard ratio (HR) 0.88; 95% CI 0.67–1.15, P = 0.35], although a secondary endpoint of bleeding gastroduodenal events did show some superiority (HR 0.52; 95% CI = 0.28–0.94). Further post-hoc analyses supported other benefits, particularly in those at the highest risk, but were exploratory, and the fact remains that the study missed its primary endpoint. The full results of COMPASS PPI are yet to be published and details of safety outcomes will be viewed with interest. The stable nature of the participants compared with the acute cases included in the present analysis, combined with the definitions of events used in the COMPASS PPI study, may explain some of the lack of demonstrated efficacy of PPI therapy.[15] Nevertheless, the lack of significant effect in COMPASS PPI is difficult to rationalize, and this may dissuade some from being insistent on PPI prescription in MI patients, who may experience medication fatigue due to polypharmacy, with consequent risk of reduced adherence to cardiovascular medications.

It should also be remembered that upper GI bleeding represents just one of the modalities of haemorrhage that can occur in a patient receiving DAPT and, when considering the most catastrophic endpoint of fatal bleeding, it may not be particularly important, although this may depend on definitions. On the one hand, for example, in the PLATelet inhibition and patient Outcomes (PLATO) trial, out of 18 624 randomized patients, protocol-defined fatal bleeding occurred in 20 patients receiving aspirin and ticagrelor, and in 23 patients receiving aspirin and clopidogrel.[16] Only five of these cases (11.6%) were due to GI haemorrhage and all occurred in the clopidogrel group despite the higher rate of GI haemorrhage in the ticagrelor group. On the other hand, upper GI haemorrhage may be associated with mortality occurring further away from the bleeding event: the present analysis found that it was associated with 11.0% risk of death at 30 days, and 25.4% at 1 year, although causality cannot be assumed. Other measures to reduce bleeding such as maintaining tight blood pressure control should be considered alongside a PPI. Similarly, targeting other non-GI bleeding complications when receiving DAPT, such as intracranial haemorrhage (which accounted for 30.2% of fatal bleeding in PLATO), must also remain a priority. Predictably this analysis confirms that these are not reduced by PPI therapy.

In summary, the data presented by Sehested and colleagues suggest that there remains good rationale for considering PPI prescription in patients receiving DAPT for MI, particularly in individuals at high risk of upper GI bleeding; however, clinicians should be mindful of potential adverse effects of PPI therapy when making this decision. Further work is required to characterize the frequency and severity of these adverse effects more comprehensively. Novel strategies for reducing bleeding from all sites in patients receiving DAPT should also be pursued.