A narrow therapeutic window of platelet P2Y₁₂ reactivity in high-risk Chinese percutaneous coronary intervention patients

Study population and design

The Institutional Review Board of The Third Xiangya Hospital and Haikou Affiliated Hospital of Central South University Xiangya School of Medicine approved the protocol and data collection (Approval No. R15006). All patients provided written consent after being fully informed.

This study was a prospective observational study. We evaluated patients with ACS, stable coronary artery disease (CAD), and a clinical need for PCI. These patients underwent successful percutaneous coronary intervention with implantation of at least one drug-eluting stent (DES). Among them, patients with stable coronary disease received elective PCI, whereas those presenting with acute coronary syndromes underwent urgent PCI, all without significant procedural complications. These patients had undergone successful elective implantation of at least one drug-eluting stent without significant complications during the procedure. All patients who underwent PCI at two clinical centers in China from September 10, 2015, to September 9, 2019, were included in the study. Patients 18 years or older who received PCI in 1 native coronary artery and were discharged on DAPT were eligible for enrollment. The study’s exclusion criteria included individuals who are allergic or intolerant to any one of the dual antiplatelet treatment drugs (aspirin, clopidogrel/ticagrelor), those who have undergone or are scheduled to undergo surgery shortly, those with a platelet count lower than 70*10⁹/L, those at high bleeding risk (active gastrointestinal bleeding, severe anemia, or ischemic stroke in the past three months), those with other end-stage diseases, and those with a shorter expected survival period. High-risk was defined as having ≥2 of the following factors for ischemic and/or bleeding events: old age (≥75 years), female gender, renal dysfunction[estimated Glomerular Filtration Ratio (eGFR) <60 mL/min] or anemia at the time of PCI (hemoglobin level <13.2 g/dL for men, <11.8 g/dL for women), low body weight (<60 kg), hypertension (as previously diagnosed), diabetes mellitus, previous stroke, use of non-steroidal anti-inflammatory drugs or selective serotonin reuptake inhibitors, triple anti-hemostatic therapy, previous in-stent thrombosis, and high-risk stenting (multivessel PCI or primary coronary artery stenting) (Vries et al., 2017).

In this exploratory clinical study, we measured platelet reactivity in patients on clopidogrel or ticagrelor after elective or emergency PCI at the Third Xiangya Hospital or Haikou Affiliated Hospital of Central South University Xiangya School of Medicine. The patient’s compliance, co-medication, comorbidities, and previous medical history were documented during the visit.

Antithrombotic therapy

All patients received drug-eluting stents, and dual antiplatelet therapy (aspirin associated with clopidogrel or ticagrelor) was recommended for 12 months. All patients received dual antiplatelet therapy consisting of aspirin and either clopidogrel or ticagrelor. For stable coronary artery disease undergoing elective PCI, clopidogrel was administered as a 300-mg loading dose at least 6 h before the procedure or as maintenance therapy of 75 mg/day for at least 5 days prior to PCI; ticagrelor was administered as a 180-mg loading dose at least 6 h before PCI or as maintenance therapy of 90 mg twice daily for at least 5 days. Conversely, patients presenting with ACS received urgent PCI after an immediate pre-procedural loading dose of clopidogrel (300 mg) or ticagrelor (180 mg). Those already receiving long-term antiplatelet therapy were not administered an additional loading dose. Technicalities of the PCI procedure, including the use of the radial approach and glycoprotein IIb/IIIa inhibitors, were all left to the professional operator’s discretion.

Unfractionated heparin was given during PCI to maintain an activated clotting time between 250 and 300 s.

Platelet function tests

Specialized equipment and trained laboratory personnel are established for platelet function tests. The VASP-platelet reactivity index (PRI) has been utilized to evaluate the effectiveness of platelet reactivity inhibition. Blood samples were drawn 12–24 h after the PCI procedure. The first 3–5 mL of blood was discarded to prevent platelet activation. Blood was then collected into a Vacutainer containing 3.8% trisodium citrate. The vacutainer was gently inverted 3–5 times and sent immediately to the lab. Within 24 h of blood collection, an experienced investigator conducted the standardized flow cytometric analysis (Platelet VASP assay; Diagnostica Stago, Biocytex, Asnières, France) to determine the degree of VASP phosphorylation. The platelet reactivity index was derived from the median fluorescence intensity (MFI) values obtained after incubation with prostaglandin E1 (PGE1) alone or in combination with adenosine diphosphate (ADP), using the following equation: ${\displaystyle PRI\left(\text{\%}\right)=\frac{MFI\left(PGE1\right)-MFI\left(PGE1+ADP\right)}{MFI\left(PGE1\right)}\times 100.}$

This platelet assay has several advantages: it is standardized, reproducible, highly specific to the P2Y₁₂-ADP receptor pathway, and correlates with clopidogrel. To assess the repeatability of the VASP assay in our laboratory, duplicate measurements were performed in a subset of 20 randomly selected samples. The intra-assay coefficient of variation (CV) was 6.9%, while the inter-assay CV, calculated from repeated analyses on different days, was 8.1%. These findings are consistent with previously published validation studies, which reported intra-assay CV < 5% and inter-assay CV < 8% (Bonello et al., 2007). Moreover, reproducibility has been confirmed under modified assay conditions with strong correlation to standard measurements, and repeated testing of the same sample over 24 h demonstrated stable results (Siller-Matula, Panzer & Jilma, 2008; Siller-Matula et al., 2013). Together, these data support the high reproducibility and reliability of the VASP-PRI measurement applied in our study.

Study endpoints

The primary efficacy outcome was a correlation of VASP values with bleeding and ischemic components. The secondary endpoint was a composite of cardiovascular death, myocardial infarction (MI), definite stent thrombosis, urgent revascularization, and stroke at 12 months after PCI. The ischemic events assessed under the primary endpoint were the same components as those predefined for the secondary composite endpoint, namely cardiovascular death, myocardial infarction, definite stent thrombosis, urgent revascularization, and stroke within 12 months after PCI. All clinical outcomes are defined in accordance with the Academic Research Consortium (ARC) recommendations. All deaths were considered to be due to cardiovascular causes unless there was clear evidence of a non-cardiovascular cause. According to the diagnostic criteria recommended by the American College of Cardiology, myocardial infarction was determined by an elevation in serum troponin I or an increase in creatine kinase–myocardial band isoenzyme levels, accompanied by at least one clinical or electrocardiographic indicator of ischemia. These included the sudden onset of typical ischemic chest pain lasting ≥20 min, ST-segment elevation ≥1 mm in two or more contiguous ECG leads, ST-segment depression ≥0.5 mm in at least two contiguous leads, or T-wave inversion >1 mm in leads with dominant R waves. Stent thrombosis was classified as definite or probable based on the definitions proposed by the Academic Research Consortium. Bleeding events were assessed according to the Bleeding Academic Research Consortium (BARC) classification, encompassing types 2, 3, and 5 (Mehran et al., 2011). Major bleeding is specifically referred to as BARC type 3 or 5 events. Each endpoint was independently evaluated and confirmed by the Clinical Events Adjudication Committee (CEAC).

Follow up

Trained research coordinators at each center conducted follow-ups within 12 months via in-person or telephone consultations. For any reported adverse events—such as ischemic episodes, bleeding complications, or discontinuation of dual antiplatelet therapy—the coordinators obtained and verified relevant source documentation from the patients.

Statistical analysis

Continuous data were presented as mean ± SD, and non-normally distributed data were presented as median ± interquartile range. Dichotomous variables were presented as percentages, and the distribution equality between subgroups was assessed using the chi-squared test. Group differences were compared using the nonparametric Kruskal–Wallis test for nominal variables, while dichotomous variables were compared using x² tests. We used ROC curve analysis to determine the optimal threshold values for HPR and LPR by testing the relationship between VASP-PRI and events (ischemia and bleeding).

The final model was developed using forward and backward logistic regression with an entry criterion of P < 0.2. A Cox proportional hazards model was applied to explore predictors of clinical outcomes, incorporating potential covariates such as cardiovascular medications (including statin use), severe left ventricular dysfunction, demographic characteristics (age and sex), acute coronary syndromes, diabetes mellitus, smoking status, serum creatinine concentration, the extent of coronary artery disease, stent type, and residual platelet aggregation categorized into tertiles. Variable selection was performed using a bidirectional stepwise approach, with inclusion criteria set at a P-value <0.05. Statistical analysis was performed using R (version 4.3.1; R Core Team, 2023).

Sample size calculations were based on our pre-experiment in the early stage; the estimated net adverse clinical events (NACE) were 30% in the HPR group, 35% in the LPR group, and 20% in the NPR group. To detect the expected treatment effect within the PRI target range, a sample size of at least 630 high-risk participants was calculated to yield 80% power with a two-sided significance level of 0.05. The percentage of missing values was less than 1% for all variables in the study. Missing values of categorical variables were attributed to their most common value, and continuous variables to the median of the non-missing values.

Study population

From September 10, 2015, through September 9, 2019, a total of 1,892 post-PCI patients were screened, 1,451 patients met the entry and discharge criteria of the experimental design scheme, and 1,012 completed the collection of relevant clinical and laboratory data. According to the high-risk criteria, 860 patients who were at high risk for bleeding or ischemic events were finally enrolled. Of these, 535 patients presented with ACS, and 325 had stable coronary disease with additional risk factors. Regarding the choice of the second antiplatelet, 250 patients received ticagrelor, while the remainder received clopidogrel. Follow-up was successfully completed for all high-risk patients (Fig. 1). The baseline clinical and procedural features of the study population stratified by VASP-PRI are listed in Table 1.

Primary and secondary endpoints

ROC curve analysis was performed to evaluate the predictive value of the VASP-PRI for ischemic and bleeding events. ROC curve analysis demonstrated that VASP-PRI could significantly discriminate between high-risk PCI patients with and without ischemic events (area under the curve (AUC):0.77; 95%CI [0.72–0.82]; P < 0.0001). A VASP-PRI ≥ 0.45 was identified as the optimal cutoff point for predicting ischemic events. It had a sensitivity of 86.6% and a specificity of 63.6%, with a negative predictive value (NPV) of 98.2% and a positive predictive value (PPV) of 16.7% (Fig. 2A). Similarly, VASP-PRI effectively distinguished between PCI patients with and without bleeding events (AUC: 0.77, 95%CI [0.73–0.81], P < 0.0001). A VASP-PRI cutoff point of ≤0.24 was optimal for predicting bleeding events, with 72.1% sensitivity, 70.3% specificity, 92.8% NPV, and 32.1% PPV (Fig. 2B).

According to the VASP-PRI, we divided the high-risk patients into LPR, NPR, and HPR. A VASP-PRI ≥ 0.45 was identified as HPR, A VASP-PRI ≤ 0.24 was identified as LPR, and a VASP-PRI between 0.24 and 0.45 was identified as NPR. Females, smokers, peripheral vascular disease (PVD), and previous PCI were more frequent among patients with HPR, while low body mass index (BMI) and low systolic blood pressure (SBP) were more frequent among patients with LPR. Patients in the NPR group had significantly lower rates of using calcium antagonists, glycoprotein IIb/IIIa (GPIIb/IIIa) inhibitors, and proton pump inhibitors than the HPR and LPR groups (Table 1).

The incidence of ischemic events was 1.33% in the LPR group (P < 0.001 vs. HPR group), 3.57% in the NPR group (P < 0.001 vs. HPR group), and 16.42% in the HPR group. Bleeding events occurred in 31.89% of the LPR group, 14.73% of the NPR group (P < 0.001 vs. LPR group), and 3.28% of the HPR group (P < 0.001 vs. LPR group) (Fig. S1).

In high-risk Chinese PCI patients, the average VASP-PRI was 37.12% ± 24.84%, and 39.41% ±  23.41% among those free of adverse outcomes at 1 year. Patients experiencing ischemic events had the highest VASP-PRI value (58.82% ± 18.12%, P < 0.001 vs. patients without ischemic events), whereas patients undergoing bleeding events had the lowest VASP-PRI value (18.77% ± 17.86%, P < 0.001 vs. patients without bleeding events) (Fig. S2).

At univariate analysis, heart failure, transient ischemic attack (TIA), clopidogrel use, morphine use, and VASP-PRI were significantly associated with the occurrence of Major Adverse Cardiovascular Events (MACE). As shown in Table 2, VASP-PRI was independently associated with a decreased incidence of MACE in the multivariate analysis.The Kaplan–Meier cumulative survival curves of high-risk PCI patients showed that patients in the NPR group had a significantly higher 1-year survival rate than those in the LPR group (94.68% vs. 95.96%, P < 0.01). The NPR group also had a significantly higher 1-year survival rate than the HPR group (94.68% vs. 97.61%, P < 0.01) (Fig. 3).

Study limitations

It is important to recognize that these trials have significant limitations. First, this observational study aimed to investigate the relationship between platelet reactivity and major adverse cardiac events. Therefore, a prospective study is necessary to confirm the clinical relevance of the PRI cutoff value. Second, the clinical utility of the VASP assay may differ, but it remains unaffected by the GPIIb/IIIa inhibitor. Third, the frequency of CYP2C19 polymorphisms is higher in Asian populations compared to Western populations, but genetic studies were lacking in our study. Thus, we cannot rule out the possibility that ethnic factors contribute to differences in the PRI therapy window. Fourth, our study had a limitation in that it was an observational design, which means we cannot infer causation from the results. Patients taking anticoagulants were not included in the DAPT study, which excludes 4% to 7% of all PCI patients. Our study emphasizes the need for further research to determine if personalized therapies improve outcomes in high-risk vascular disease patients.