Utilization of triple antithrombotic therapy in patients with atrial fibrillation undergoing percutaneous coronary intervention

Triple antithrombotic therapy (TAT), a combination of an oral anticoagulant and dual antiplatelet therapy (DAPT), is a key treatment for prevention of ischemic events in patients with atrial fibrillation (AF) undergoing percutaneous coronary intervention (PCI). However, TAT is not extensively used because of the risk of bleeding. This study aimed to determine the utilization and influencing factors of TAT using real-world data in the non-vitamin K antagonist oral anticoagulants (NOACs) era. We analyzed National Inpatient Sample data compiled by the Health Insurance Review & Assessment Service (HIRA-NIS) from 2011 to 2020. Patients with AF who underwent PCI with stent implantation and with an increased stroke risk were selected as candidates for TAT therapy. Demographic and clinical factors associated with TAT use were investigated using the chi-squared test and the Student t-test, and influencing factors were identified using multiple logistic regression. The TAT utilization rate steadily increased from 30.3% in 2011 to 65.4% in 2020 (Cochran-Armitage trend test: p < 0.001) with an average of 45.9%. Positive influencing factors for TAT use were identified as congestive heart failure, history of previous stroke/transient ischemic attack/thromboembolism, valvular heart disease, and year. Negative influencing factors included insurance type (medical aid or Patriots & Veterans Insurance), type of medical institution (general hospitals or primary medical institutions), and comorbidities such as renal disease, liver disease, and history of the previous hemorrhage. The utilization of TAT following PCI among high-stroke risk AF patients steadily increased from 2011 to 2020, reaching 65.4% by the end of the study period. However, in 2020, a significant proportion of 29.4% of patients still received DAPT, indicating that many AF patients undergoing PCI did not receive adequate antithrombotic therapy.

AF is also associated with a four-to fivefold increased risk of ischemic stroke [14]. For AF patients who are at a moderate-to-severe risk of stroke, chronic oral anticoagulants (OACs) are required for stroke prevention [3,4,15]. Therefore, combination therapy with OAC and antiplatelet agents is recommended to prevent both ischemic and thrombotic events in AF patients with acute coronary syndrome (ACS) or undergoing PCI [3,4,15].
According to literature and European Society of Cardiology (ESC) guidelines [3,4,15], patients may benefit from a short-term (1 day to 1 week) triple antithrombotic therapy (TAT) with an OAC and DAPT followed by a 6-to 12-month period of dual antithrombotic therapy (DAT) with OAC and an antiplatelet agent (preferably clopidogrel). OAC monotherapy should be continued thereafter unless recurrent ischemic events occur [3,4,15].
Despite these recommendations, observational studies have reported that AF patients are not adequately treated after PCI. According to Park et al. [16] and Boivin-Proulx et al. [17], DAPT (excluding OAC) was the most frequent treatment instead. These observations may reflect the physician's preference for DAPT, the standard antiplatelet therapy for patients without AF after coronary stent implantation over TAT or DAT, because of the risk of bleeding from the addition of OAC [18,19].
However, DAPT is inferior to OAC for prophylaxis of thromboembolic events in patients with AF and accordingly should not be used as a mainstay therapy [3,4,15]. Even in East Asians, who are known to be more susceptible to antithrombotic therapy-related bleeding events than the Western population [20], DAPT has been reported to have no statistically significant advantages over TAT in terms of ischemia and bleeding outcomes [21].
Underutilization of OAC after PCI was also observed in the Western population [22,23], indicating that low usage of TAT is a global problem. However, to the best of our knowledge, there are not enough studies that analyzed the use of TAT since non-vitamin K antagonist oral anticoagulants (NOACs) replaced warfarin. After the introduction of NOACs, which have a lower bleeding risk compared to that with warfarin, OAC utilization in AF patients increased [24]. Furthermore, it was presumed that TAT utilization has also increased. Therefore, this study aimed to identify the trend of antithrombotic therapy for patients with AF who underwent PCI from 2011 to 2020 and to analyze which factors affect TAT utilization.

Data sources
This cross-sectional study utilized National Inpatient Sample data provided by Health Insurance Review & Assessment Service (HIRA-NIS) covering the period from 2011 to 2020. The data represents approximately 10-13% of the Korean population who were admitted to medical institutions and updated annually with a different set of patients selected each year [25].
The International Classification of Diseases, Tenth Revision (ICD-10) code was used to extract diagnostic information. The procedure codes were used to screen PCI with stent implantation. The ingredient codes were used to identify antithrombotic agents: warfarin, four NOACs (dabigatran, rivaroxaban, apixaban, and edoxaban), three P2Y 12 inhibitors (clopidogrel, ticagrelor, prasugrel), and aspirin (Supplementary Table 1).
Social health insurance in Korea consists of three types: National Health Insurance (NHI), Medical Aid (MedAid), and Patriots & Veterans Insurance (PVI). The NHI covers approximately 97% of the population and taxes their income, while MedAid provides coverage for low-income individuals, and PVI is for national veterans [26].

Study population
We identified patients who were diagnosed with AF by using the ICD code I48. We screened patients who underwent PCI with stent implantation by using procedure codes: M6561-6567. We included patients who received PCI between February and November each year. By utilizing yearly claims data, we excluded patients who underwent PCI in January and December from the analysis to ensure a complete examination of comorbidities and drug utilization before and after hospitalization. Among them, those who had a CHA 2 DS 2 -VASc score of ≥ 2 for males or ≥ 3 for females were considered appropriate candidates for TAT in this study [3,15]. We investigated the records of antithrombotic agents prescribed within 14 days after PCI [21].
Ages were classified as ≤ 65, 65-74, and ≥ 75 years. Ischemic heart disease was grouped into ACS (including MI and unstable angina) and non-ACS (including stable angina and chronic ischemic heart disease). The medical institution types were divided into tertiary hospitals and non-tertiary hospitals (including general hospitals and primary medical institutions). Insurance types were divided into NHI and MedAid/PVI. The region was categorized into the capital city, six metropolitan cities, and others.

Antithrombotic therapy
First antithrombotic therapy was defined based on drugs prescribed within 14 days after PCI. Drug therapy was considered continuous if the prescription gap was less than 14 days. The duration of the first antithrombotic therapy was defined as the period in which all first antithrombotic agents were used together.
Owing to the study design of this research, only prescriptions claimed within 30 days of PCI were identified. Therefore, the second antithrombotic therapy was determined only in patients who finished the first antithrombotic therapy within 30 days after PCI.

Influencing factors
The following factors influencing the use of TAT were investigated: age, sex, insurance type, presence of comorbid diseases, non-steroidal anti-inflammatory drug (NSAID) use, region, medical institution type, year, CHA 2 DS 2 -VASc, and HAS-BLED. Comorbid diseases included ischemic heart disease, congestive heart failure, hypertension, diabetes mellitus, prior stroke/TIA/TE, renal/liver disease, prior hemorrhage, valvular heart disease, dyslipidemia, anemia, and any malignancy (excluding non-melanoma skin cancer).

Statistical analysis
The demographic characteristics of patients are presented using frequency analysis. Data for continuous variables are shown as mean and standard deviation (SD) and categorical variables are presented as counts and percentages. For continuous variables, Student's t-test for normally distributed data and the Wilcoxon rank-sum test for non-normally distributed data were applied. For categorical variables, the chi-squared test was used. To investigate the factors

Characteristics of study population
From 2011 to 2020, a total of 3532 patients with AF who underwent PCI with stenting were identified (Fig. 1). Among them, 2684 patients who had a CHA 2 DS 2 -VASc score of ≥ 2 (males) or ≥ 3 (females) were selected as the study population. The demographics and clinical characteristics of the study population are summarized in Table 1. Patients aged over 75 years accounted for the largest proportion of the population (46.2%) followed by those aged 65-74 years (34.2%) and under 65 years old (19.6%). The proportion of male patients (64.3%) was higher than that of female patients (35.7%).

Antithrombotic therapy
Out of the 2684 study patients, 1233 (45.9%) received TAT within 14 days after PCI, the remaining patients received either DAPT (50.9%) or DAT (1.5%). Over time warfarinbased TAT therapy declined from 30.3% in 2011 to 5.2% in 2020, whereas the use of NOAC-based TAT therapy emerged in 2013 and steadily increased to surpass 60% (Supplementary Table 2). Table 1 also presents the results of the chi-squared test and Student's t-test. TAT use increased with increasing age and was lower in MedAid/PVI than in NHI (Table 1). TAT use was higher in patients with congestive heart failure, history of previous stroke/TIA/TE, and valvular heart disease, whereas it was lower in patients with ACS, diabetes mellitus, liver disease, and history of previous hemorrhage. TAT utilization rate was highest in the capital city (51.9%), followed by the six metropolitan cities (46.6%), and other regions (42.5%). It was higher in tertiary hospitals (52.8%) than in non-tertiary hospitals (40.2%), including general hospitals and primary medical institutions.
The mean CHA 2 DS 2 -VASc score was significantly higher in TAT users than in non-TAT users (4.20 ± 1.54 versus 4.06 ± 1.48). However, there was no statistically significant difference between the two groups in HAS-BLED scores (2.33 ± 0.94 versus 2.29 ± 0.96) ( Table 1).
The average duration for TAT was 20 days. Post-TAT antithrombotic therapy (PTA) use was further investigated in patients with TAT as the first antithrombotic therapy, which was terminated within 30 days after PCI, and identified a second antithrombotic therapy. Of 597 PTA users, DAT and DAPT were prescribed 40.2% and 52.4% of the time, respectively (Supplementary Table 3); however, since 2019, the utilization of DAT was more compared to that of DAPT.  In addition, non-tertiary hospitals, including general hospitals and primary medical institutions, were associated with approximately 0.572 times lower odds of TAT use than tertiary hospitals (p < 0.001).

Discussion
In this study using HIRA-NIS, we found that the utilization of TAT for patients with AF after PCI increased from 30.3% in 2011 to 39.8% in 2015. Previous studies using claims data reported that TAT utilization increased from 30.3% in 2011 to 38.2% in 2015, which is consistent with the findings of our study [16]. The temporal trend of TAT use increased significantly between 2011 and 2020. The TAT use increased   [16,17,27,28]. Recent guidelines recommend short-term use of TAT for 1 day to 1 week after PCI [3,4,15]. Despite these recommendations, one-third of AF patients in 2020 did not use TAT after PCI, which could be because of clinicians overestimating the risk of bleeding [16,27]. A previous study has reported that DAPT alone is inferior to OAC in the prevention of AF-related ischemic events (relative risk [RR] = 1.44; p < 0.001) and hemorrhagic events (RR = 1.21; p = 0.001) [29]. Therefore, DAT with OAC is recommended as an alternative regimen when TAT is not selected as the main antithrombotic therapy owing to the risk of bleeding. However, even in 2020, only a small fraction of patients received DAT (2.4%) as a substitute therapy for TAT, and the majority received DAPT (29.4%).
In this study, valvular heart disease was the most influential factor in TAT utilization. According to Başaran et al. the combination therapy of antithrombotic agents was more common in patients with significant valvular disease than in those with non-significant valvular disease [30]. Park et al. reported that congestive heart failure, previous stroke/TIA/ TE, ≥ 75 years of age, and hypertension positively affected TAT use, whereas MI negatively impacted TAT use [16]. In our study, ischemic risk factors, such as congestive heart failure and previous stroke/TIA/TE, positively influenced TAT use; however, age and hypertension did not affect TAT use. Both factors are also risk factors for bleeding; hence, it is presumed that the effect was offset. According to the guidelines [4,13,15], ACS in patients with AF is considered a compelling indication for TAT. However, in our study, ACS was found to be a negative factor for TAT, which aligns with findings from the previous study [16]. This discrepancy could be due to clinicians not recognizing the increased risk of stroke from vascular disease in AF patients [23,31] or the concern about the elevated bleeding risk associated with the concomitant use of antiplatelet agents and OAC in ACS patients [32]. In addition, bleeding risk factors, such as kidney disease, liver disease, and previous hemorrhage, also influenced the underutilization of TAT, which was consistent with the findings of Bocchino et al. [33].
The beneficiaries of MedAid/PVI were identified to have lower utilization of TAT than NHI subscribers. Wang et al. reported that Medicaid beneficiaries most frequently reported having difficulties getting care in the USA (OR = 3.411; p < 0.05) [34]. Although it is known that South Korea has a more generous medical aid system for the underprivileged, our findings suggest that it may not sufficiently guarantee access to necessary medical needs in this circumstance.
At the level of medical institutions, the utilization of TAT was found to be low in non-tertiary medical institutions. Other studies have consistently shown that tertiary care institutions are more aggressive in adapting recommendations of clinical guidelines into clinical practice [24,35].
In this study, the average duration of use of TAT was 20 days, whereas the guideline recommended the use of TAT for up to 7 days. This difference is probably owing to the extension of TAT use by up to 1 month in patients with high ischemic/thrombotic and low bleeding risk [3,4,15]. However, even considering the insufficient information available in estimating the bleeding risk factors of individual patients, the duration of TAT was longer than expected. It is presumably because a 2016 ESC guideline advocated a longer period of TAT (1 month) to prevent ischemic events in most patients without consideration of risk factors than did the current guidelines [3,4,15,36].
This study has some limitations. First, we used claims data that were obtained for reimbursement and not for clinical or research purposes; therefore, providers who want a higher reimbursement rate might up-code the diagnosis information [37]. Moreover, because the data did not include uninsured events, we were unable to gather some information, such as over-the-counter NSAID use. Second, clinical information such as the number, type, and length of stents, alcohol use, laboratory test results, or disease severity, were not included in these data [38]. Therefore, we estimated the clinical status Adj. OR adjusted odds ratio, CHA 2 DS 2 -VASc congestive heart failure, hypertension, age ≥ 75 years (doubled), diabetes mellitus, prior stroke, transient ischemic attack (TIA), or thromboembolism (TE) (doubled), vascular disease, age 65-74 years, and sex, CI confidence interval, HAS-BLED hypertension, abnormal renal/liver function, stroke, bleeding history or predisposition, labile INR, elderly, and drugs/alcohol, MedAid medical aid, NHI National Health Insurance, NSAID non-steroidal anti-inflammatory drug, PVI Patriots & Veterans Insurance, R reference, TE thromboembolism, TIA transient ischemic attack of patients using the provided disease codes (Supplementary  Table 1). Third, socioeconomic factors, including income, education, and health behaviors, were not reflected. Bleeding risk is an important factor in deciding whether to use TAT. However, claims data do not include clinical data such as stent length and number, which made it difficult to adequately identify bleeding risks such as those mentioned by the Academic Research Consortium for High Bleeding Risk (ARC-HBR). Therefore, further studies including clinical information are needed.
Despite these limitations, this study provides insight into several aspects. We reported recent data on utilization rate of TAT compared to other treatments, such as DAPT and DAT, in patients with AF undergoing PCI. In addition, we were able to identify the characteristics of patients who were less subjected to the most recent recommendation.

Conclusions
In AF patients who underwent PCI with stenting and were at an increased stroke risk, the utilization of TAT in 2020 was 65.4%. When TAT was not selected, the most used alternative treatment of TAT was DAPT rather than DAT with OAC. It should be widely recognized that TAT and DAT should be more actively implemented than DAPT.