Implementing a pharmacogenomic-driven algorithm to guide antiplatelet therapy among Caribbean Hispanics: a non-randomised clinical trial

Abstract

Objectives: To assess whether genotype-guided selection of oral antiplatelet drugs using a clinical decision support (CDS) algorithm reduces the rate of major adverse cardiovascular and cerebrovascular events (MACCEs) among Caribbean Hispanic patients, after 6 months.

Design: An open-label, multicentre, non-randomised clinical trial.

Setting: Eight secondary and tertiary care hospitals (public and private) in Puerto Rico.

Participants: 300 Caribbean Hispanic patients on clopidogrel, both genders, underwent percutaneous coronary intervention (PCI) for acute coronary syndromes, stable ischaemic heart disease and documented extracardiac vascular diseases.

Interventions: Patients were separated into standard-of-care (SoC) and genotype-guided (pharmacogenetic (PGx)-CDS) groups (150 each) and stratified by risk scores. Risk scores were calculated based on a previously developed CDS risk prediction algorithm designed to make actionable treatment recommendations for each patient. Individual platelet function, genotypes, clinical and demographic data were included. Ticagrelor was recommended for patients with a high-risk score ≥2 in the PGx-CDS group only, the rest were kept or de-escalated to clopidogrel. The intervention took place within 3-5 days after PCI. Adherence medication score was also measured.

Primary and secondary outcomes: The occurrence rate of MACCEs (primary) and bleeding episodes (secondary). Statistical associations between patient time free of events and predictor variables (ie, treatment groups, risk scores) were tested using Kaplan-Meier survival analyses and Cox proportional-hazards regression models.

Results: The genotype-guided group had a clinically lower but not significantly different risk of MACCEs compared with the SoC group (8.7% vs 10.7%, p=0.56; HR=0.56). Among high-risk score patients, genotype-driven guidance of antiplatelet therapy showed superiority over SoC in reducing MACCE incidence 6 months postcoronary stenting (adjusted HR=0.104; p< 0.0001).

Conclusions: The potential benefit of implementing our PGx-CDS algorithm to significantly reduce the incidence rate of MACCEs in post-PCI Caribbean Hispanic patients on clopidogrel was observed exclusively among high-risk patients, with apparently no evident effect in other patient groups.

Trial registration number: NCT03419325.

Keywords: adverse events; cardiovascular disease; clinical decision-making; genetics; risk factors.

Figure 1. Kaplan-Meier plot of time free of both primary and secondary endpoints combined at 6 months of follow-up, grouped by treatment interventions (ie, SoC vs genotype-guided groups). The adjusted HR and the corresponding log-rank test p value are presented. Age, sex, BMI, T2DM, stent length, statins, aspirin and PPIs users were considered as covariates in the analysis. Data were censored for 6 months. BMI, body mass index; SoC, standard-of-care; T2DM, type 2 diabetes mellitus.

Figure 2. The PGx-guided CDS algorithm reduced the incidence rate of MACCEs and major/minor bleedings in Caribbean Hispanic patients. The incidence of MACCEs and major/minor bleeding events in the SoC group was 40% while in the genotype-guided group was 23%. The OR (95%CI) and p value of the association were computed using a one-sided χ² test. Data are presented as frequency of events in patients. CDS, clinical decision support; MACCEs, major adverse cardiovascular and cerebrovascular events; PGx, pharmacogenetic; SoC, standard of care. ** indicates a very significant result (p < 0.01)

Figure 3. Kaplan-Meier plots of time free of primary endpoint (MACCEs) grouped by risk scores (6 months of follow-up). (A) The comparison between SoC and genotype-guided treatment groups for low-risk patients (risk score<2); (B) the comparison between SoC and genotype-guided treatment groups for high-risk patients (risk score ≥2). HRs and corresponding log-rank test p values are shown, and data are censored for 6 months. Age, sex, BMI, T2DM, stent length, statins, aspirin and PPIs users were considered as covariates in the analysis. BMI, body mass index; MACCEs, major adverse cardiovascular and cerebrovascular events; PPIs, proton pump inhibitors; SoC, standard of care; T2DM, type 2 diabetes mellitus.

Figure 4. Forest plot displaying HRs for MACCEs (with 95% CIs), absolute differences in percentages, and corresponding p values obtained from the multivariate Cox proportional hazards regression model adjusted for relevant covariates (sex, age, BMI, T2DM, stent length, statins, aspirin and PPIs users). The results illustrate the comparison between high-risk patients in both treatment groups (ie, PGx CDS vs SoC). ^#Median length of adjoined stents ≥30 mm. BMI, body mass index; CDS, clinical decision support; MACCEs, major adverse cardiovascular and cerebrovascular events; PGx, pharmacogenetic; PPIs, proton pump inhibitors; SoC, standard of care; T2DM, type 2 diabetes mellitus.