From January 2017 to December 2019, 516 patients with CCS and AF who had undergone PCI and were indicated to use OAC based on CHA2DS2-VASc score ≥ 1, excluding sex, were included in the final analysis. The median follow-up time was 36 months (interquartile range:22–45).

Baseline characteristics, organized by the year of enrollment, were presented in Table 1. The average age was 66 ± 9 and 95 (18.4%) were female. Those enrolled in 2018 exhibited a slightly higher prevalence of hyperlipidemia and in 2019 a lower creatine clearance (CrCl) was observed. However, other demographic information and the history of comorbidities showed no difference among groups.

Table 2 presented the status of antithrombotic therapy according to the years of enrollment. The utilization of clopidogrel and ticagrelor remained stable across the years without any significant differences. The prescription of aspirin showed a slight decrease. There was a notable increase in the usage of OAC (P < 0.001), primarily attributed to the rising trend of NOAC usage. In contrast, the usage of warfarin showed no significant difference among the groups. DAT showed an increasing trend (P = 0.022). Similarly, there was a significant increase in the group received TAT (P < 0.001).

Figure 2 illustrated a decreasing trend in MACE events, stroke/TIA, MI and SE, while TIMI bleeding events showed an increasing pattern.

Tendency of event rate according to enrolled years. MACE events included all-cause death, myocardial infarction, stroke/transient ischemic attack (TIA), systemic embolism or ischemia-driven revascularization. TIMI bleeding events included minor, minimal and major TIMI bleeding events based on TIMI criteria. TIA, Transient ischemic attack

The characteristics of patients according to antithrombotic therapy were presented in Table 3. Lower creatine clearance (CrCl, P = 0.023) was observed in TAT group. Those with lower left ventricular ejection fraction (LVEF) tended to be prescribed with TAT (P = 0.015). Besides, non-TAT was more likely to be utilized in patients with hyperlipidemia (P < 0.001). On the other hand, patients with diabetes mellitus (DM, P = 0.031) or heart failure (HF, P = 0.034) were more likely to be prescribed with TAT. Therefore, higher HAb1c (P = 0.007) was also observed in the TAT groups. In the comparison between the TAT and non-TAT groups, a higher BMI was associated with TAT therapy (P = 0.037).

Details of the proportion of OAC usage in patients prescribed with TAT were shown in Fig. 3, with rivaroxaban divided into five parts according to the daily prescription dose. Dabigatran was administered at 220 mg daily, and the dose of VKA was tailored for individual patients. Overall, only 46.7% used the standard dose for stroke prevention.

Figure 4 depicted the percentage of events according to antithrombotic therapy during the follow-up period. MACE events showed differences between TAT and antiplatelet therapy. TIMI bleeding events presented differences between combined therapy (antiplatelet therapy plus OAC) and antiplatelet therapy. Similar results were detected in ischemia-driven revascularization.

As demonstrated in the supplementary Table 3, OAC therapy was more likely to be administered to patients with a history of DM (P = 0.010), a diagnosis with HF (P = 0.010), and less likely in those diagnosed with hyperlipidemia (P < 0.001) and previous PCI (P = 0.049). In addition, patients with PAF (P < 0.001) were favored of non-OAC treatment, while patients with peAF (P < 0.001) exhibited the opposite preference. Furthermore, peak cardiac troponin I (cTnI, P < 0.001) and N-terminal pro-B-type natriuretic peptide (NT-proBNP, P < 0.001) also presented an association with OAC usage.

Subsequently, the multivariate logistic regression analysis was performed to assess independent factors capable of predicting the prescription of OAC therapy (Fig. 5). After adjustment for multivariate factors, peak cTnI, peak NT-proBNP and previous PCI showed no influence on OAC choice. However, patients with a history of DM (OR = 1.826; 95%CI, 1.216–2.744) and HF (OR = 1.899; 95%CI, 1.119–3.222) were favored of OAC therapy. Notably, HF emerged as the strongest predictor of OAC prescription. Conversely, hyperlipidemia (OR = 0.335; 95%CI, 0.210–0.533) and PAF (OR = 0.297; 95%CI, 0.199–0.443) were associated with a preference for non-OAC therapy.

Proportion of oral anticoagulation (OAC) usage among patients prescribed with triple antithrombotic therapy (a) and combine therapy (b) based on OAC type and daily dose

Event rate based on antithrombotic therapy. DAT, double antithrombotic therapy TAT, triple antithrombotic therapy; TIA, Transient ischemic attack; MI, myocardial infraction; SE: systemic embolism; MACE events included all-cause death, myocardial infarction, stroke, systemic embolism or ischemia-driven revascularization. TIMI bleeding events included minor, minimal and major TIMI bleeding events based on TIMI criteria. *P < 0.05, P value was based on χ2 test

Predictors of the choice of oral anticoagulant (OAC) therapy. OR, odds ratio; CI, confidence interval; PAF, paroxysmal atrial fibrillation; DM, diabetes mellitus; HF, heart failure. Multiple logistic regression analysis adjusted history of heart failure, diabetes mellitus, hyperlipidemia, paroxysmal atrial fibrillation, PeAF which includes persist atrial fibrillation and permanent atrial fibrillation, creatinine clearance, left ventricular ejection fraction, glycosylated hemoglobin, peak cTnI, peak NT-proBNP

Regarding the prognosis, we conducted separate analyses for MACE events and TIMI bleeding events, respectively. After adjusting for potentially confounding variables through multivariate Cox regression, a history of HF (HR, 1.744; 95%CI, 1.011–3.038) and TAT (HR, 2.708; 95%CI, 1.653–4.436) showed an independent association with MACE events, as detailed in Table 4.

In terms of bleeding events, there was a decrease in the risk of bleeding with an increase in CrCl (HR, 0.986; 95%CI, 0.974–0.997). On the other hand, OAC therapy (HR, 10.378; 95%CI, 6.136–17.555) emerged as a risk factor for TIMI bleeding events. (Table 5)