Age Modifies Intracranial and Gastrointestinal Bleeding Risk from P2Y12 Inhibitors in Patients Receiving Dialysis

Abstract

Background: Individuals aged ≥75 years are the fastest-growing population starting dialysis for end-stage kidney disease (ESKD) due to living longer with coronary artery disease. ESKD alone can increase bleeding risk, but P2Y₁₂ inhibitor (P2Y12-I) antiplatelet medications prescribed for cardiovascular treatment can exacerbate this risk in patients with ESKD. The age-specific rates of bleeding complications in dialysis patients with ESKD on P2Y12-I remain unclear, as does how age modifies the bleeding risk from P2Y12-I use in these patients.

Methods: In a retrospective cohort study, we collected data on 40,972 patients receiving maintenance hemo- or peritoneal dialysis who were newly prescribed P2Y12-I therapy between 2011 and 2015 from the USRDS registry. We analyzed the effect of age on the time to first bleed and the interactions between age and P2Y12-I type on modifying the effects of a bleed.

Results: Twenty percent of the cohort were aged ≥75 years. There were 3096 (8%) gastrointestinal (GI) and 1298 (3%) intracranial (IC) bleeding events during a median follow-up of 1 year. Annual incidence rates for IC bleeds were 2% in those aged <55 years and 3% in those aged ≥75 years. Rates for GI bleeds were 4% in those aged <55 years and 9% in those aged ≥75 years. On clopidogrel, prasugrel, and ticagrelor, for every decade increase in age of the cohort members, the risk of IC bleed increased by 9%, 55%, and 59%, and the risk of GI bleed increased by 21%, 28%, and 39%, respectively. At age ≥75 years, prasugrel was associated with a greater risk of IC bleed than clopidogrel. At age ≥60 years, ticagrelor was associated with a greater risk of GI bleed than clopidogrel.

Conclusions: More potent P2Y12-Is (prasugrel and ticagrelor) were associated with a disproportionately higher risk of IC bleed with increasing age compared with that of clopidogrel-prasugrel was much worse than clopidogrel at age ≥75 years. All three drugs were associated with only modest increase in the risk of GI bleed with every decade increase in age-ticagrelor was much worse than clopidogrel at ≥60 years of age. These results highlight the need for head-to-head clinical trials for the use of P2Y12-Is in patients with ESKD to determine age cutoffs where the risk of bleeding outweighs the benefits of thrombosis prevention.

Keywords: P2Y12 inhibitors; age; clopidogrel; dialysis; gastrointestinal bleeding; intracranial bleeding; prasugrel; ticagrelor.

Graphical abstract

Figure 1.

Derivation of the cohort after applying inclusion and exclusion criteria and the remaining cohort at the end of follow-up after death and censoring. P2Y12-I, P2Y₁₂ inhibitor.

Figure 2.

Rise in rates of bleeding events with increasing age of ESKD patients on P2Y12 inhibitors. (A) Annual incidence rate and (B) cumulative risk of gastrointestinal and intracranial bleeds in the study cohort of patients on chronic dialysis with new prescriptions for P2Y12-I.

Figure 3.

Predictors of gastrointestinal bleed in ESKD patients on P2Y12 inhibitors. Forest plot representing Cox proportional hazard ratios of selected risk factors associated with time to first gastrointestinal bleed in (A) the entire study cohort and (B) the subgroup with a cardiovascular event (CV) claim over the 6 months before the index dates. Only select covariates from the Cox model are shown in the forest plot.

Figure 4.

Predictors of intracranial bleed in ESKD patients on P2Y12 inhibitors. Forest plot representing Cox proportional hazard ratios of selected risk factors associated with time to first intracranial bleed in (A) the entire study cohort and (B) the subgroup with a CV claim over the 6 months before the index dates. Only select covariates from the Cox model are shown in the forest plot.

Figure 5.

Hazard ratio (HR; with 90% confidence interval) of the bleeding event as a function of age (with a focus on the 50- to 80-year age range). To create this plot, estimates of the regression coefficients of the interaction term (age×P2Y12-I type) and P2Y12-I type were used to evaluate the relative performance of each pair of P2Y12-I type—prasugrel versus clopidogrel (P versus C), ticagrelor versus clopidogrel (T versus C), and prasugrel versus ticagrelor (P versus T).