Type of atrial fibrillation and outcomes in patients without oral anticoagulants

Abstract Background The effect of type of atrial fibrillation (AF) on adverse outcomes in Chinese patients without oral anticoagulants (OAC) was controversial. Hypothesis The type of AF associated with adverse outcomes in Chinese patients without OAC. Methods A total of 1358 AF patients without OAC from a multicenter, prospective, observational study was included for analysis. Univariable and multivariable Cox regression models were utilized. Net reclassification improvement analysis was performed for the assessment of risk prediction models. Results There were 896(66%) patients enrolled with non‐paroxysmal AF (NPAF) and 462(34%) with paroxysmal AF (PAF). The median age was 70.9 ± 12.6 years, and 682 patients (50.2%) were female. During 1 year of follow‐up, 215(16.4%) patients died, and 107 (8.1%) patients experienced thromboembolic events. Compared with the PAF group, NPAF group had a notably higher incidence of all‐cause mortality (20.2% vs. 9.4%, p < .001), thromboembolism (10.5% vs. 3.8%, p < .001). After multivariable adjustment, NPAF was a strong predictor of thromboembolism (HR 2.594, 95%CI 1.534–4.386; p < .001), all‐cause death (HR 1.648, 95%CI 1.153–2.355; p = .006). Net reclassification improvement analysis indicated that the addition of NPAF to the CHA2DS2‐VASc score allowed an improvement of 0.37 in risk prediction for thromboembolic events (95% CI 0.21–0.53; p < .001). Conclusions In Chinese AF patients who were not on OAC, NPAF was an independent predictor of thromboembolism and mortality. The addition of NPAF to the CHA2DS2‐VASc score allowed an improvement in the accuracy of the prediction of thromboembolic events.


| INTRODUCTION
In 2020 ESC Guidelines for the diagnosis and management of atrial fibrillation, the "4S-AF" scheme 1 was proposed to be considered in managing atrial fibrillation (AF) patients. The burden of AF has been listed as one component, and studies demonstrated that the duration or pattern of AF correlates with the extent of the atrial substrate, remodeling, and AF-related outcomes. 1 The AF pattern is now the simplest and quick way to assess patients' AF burden without much examination. Though the effects of AF pattern on outcomes had been investigated for nearly 20 years, the results were controversial. Most of these papers enrolled patients with oral anticoagulants (OACs), and anticoagulant state may also be a confounder. As so far, only two studies 2,3 focused the non-anticoagulated patients, and the results were not consistent. More information is needed to assess the prognosis of AF type on adverse events in patients without OAC. Though China has a heavy burden of AF that the annual risk of thromboembolic events in Chinese AF patients ranged from 3.7% to 9.2%, [4][5][6] there was little data about the relationship between AF pattern and outcomes. The purpose of this article was to explore the association between AF type and adverse outcomes in Chinese patients who were not on anticoagulation and further evaluated the value of AF pattern in decision-making for stroke prevention.

| METHODS
The present study was based on a multicenter, prospective, observational study 7  representative hospitals around China (including rural and urban, academic and community, general and specialized, public and private hospitals) had participated. A total of 1358 patients with non-valvular AF and no OAC (both discharged without OAC and no initiation during the follow-up) from the multicenter study were included for analysis in the present article. The process of patient selection was shown in Figure S1.
The low percentage of anticoagulation therapy in our study was a national medical status, and the proportion was only 2.7% in 2004 8 and increased to 18.6%~31.7% in the early period of the 2010s. 7,9 The study was approved by the ethics committee of each center and obeyed the Declaration of Helsinki. All patients have provided written consent to participate in the study.
The demographic information, admission vital signs, medical histories, and treatments were collected at baseline by interviewing the participants, reviewing their medical records, and contacting their treating physicians. The type of AF was defined according to the 2006 ACC/AHA/ESC guidelines for the management of patients with AF. 10 Briefly, if the arrhythmia terminated spontaneously, AF was designated paroxysmal; when sustained beyond 7 days, it was termed persistent. Termination with pharmacological therapy or direct-current cardioversion does not alter the designation. Permanent AF was defined that AF did not terminate either spontaneously or with electrical or chemical cardioversion, or cardioversion had not been attempted. Both persistent AF and permanent AF were divided into non-paroxysmal AF (NPAF) group in the following analysis. There were 896 (66%) patients enrolled with NPAF and 462 (34%) with paroxysmal AF (PAF). The classification of the subtype of AF relied on the attending physician's interpretation. Patients were also divided into AF/flutter group or other rhythm group by the electrocardiogram rhythm at discharge for sensitivity analysis. The CHA 2 DS 2 -VASc score by giving 2 points to each patient of age ≥75 years and a history of prior stroke or TIA and 1 point to each patient of age 65-74 years, history of hypertension, diabetes mellitus, congestive heart failure, vascular diseases, and female sex.
The registry was designed to have a 1-year follow-up. The followup was completed in November 2012 by trained research personnel via clinic visit, telephone or delivery of medical records. The status of OAC during the follow-up period was collected again at the visit. In this study, the primary outcome was thromboembolic events (TE events, including stroke and non-central nervous system embolism), and secondary outcomes were defined as all-cause death, cardiovascular death and stroke. Cardiovascular death included sudden cardiac death and death caused by heart failure, stroke, myocardial infarction, pulmonary embolus, peripheral embolus, aortic dissection.
Continuous variables were expressed as means with SDs or medians with quartiles; categorical variables were expressed as frequencies and percentages. Differences in continuous variables between groups according to the type of AF were analyzed using unpaired t-test or the Mann-Whitney U test; comparison of categorical variables was performed using χ 2 test or Fisher's exact test. Kaplan-Meier curves and logrank tests were performed to illustrate the discrepancies among the AF patterns. Univariable and multivariable Cox regression analysis was utilized to evaluate the effects of AF type on the TE events, stroke, all-cause death and cardiovascular death. The following covariables were adjusted in the multivariable model: sex, age ≥75-years-old, body mass index, admission systolic blood pressure, admission diastolic blood pressure, admission heart rate, tobacco use, previous stroke or transient ischemic attack (TIA), coronary artery disease (CAD), prior myocardial infarction, hypertension, HF, significant valvular heart disease, diabetes mellitus, emphysema/chronic obstructive pulmonary disease, hyperthyroidism, sleep apnea, previous major bleeding, dementia or cognitive defects, antiplatelet drug, β-blocker, ACEI/ARB, calcium channel blocker, diuretics, digoxin, statin, antiarrhythmic drug. Further, we conducted subgroup analyses to assess whether the difference between types on the risk of TE events and all-cause death existed among the following specific subsets of patients, including sex, age, presence of CAD, hypertension, HF, CHA 2 DS 2 -VASc score and antiplatelet drugs at discharge. Hazard ratio was estimated using Cox proportional hazards models fitted separately in subgroups of patients. The association between the rhythm at discharge (AF/flutter group vs other rhythm group) and adverse outcomes were also explored as the sensitivity analysis. The incremental contribution of AF type in predicting the risk of TE events based on CHA 2 DS 2 -VASc score was presented as net reclassification improvement and integrated discrimination improvement using the PredicABEL, an R package for the assessment of risk prediction models. The prognostic utility of CHA 2 DS 2 -VASc score and after adding NPAF as a risk factor (1 point) into CHA 2 DS 2 -VASc score was assessed by C-statistic estimates. The comparison between the two scores was also made.
Hazard ratio and 95% confidence intervals (CI) were calculated.
The software package SPSS version 25.0 (IBM Corporation, New York, NY), the R software version 4.0.2 (R Foundation for Statistical Computing) and the MedCalc version 19.0.7 were used for statistical analysis. GraphPad Prism version 6.01 was utilized for figures. All statistical tests were two-tailed, and a p-value <.05 were considered significant.

| RESULTS
Baseline characteristics were given in Table 1 and had higher systolic blood pressure, CHA 2 DS 2 -VASc score, but lower body mass index, diastolic blood pressure, and heart rate. They suffered from comorbidities more frequently. Patients with NPAF were more likely to take antiplatelet drugs, diuretics and digoxin, while they had less proportion of antiarrhythmic agents. Baseline characteristics of patients classified by the rhythm at discharge were given in Table S1.
The adverse outcomes during the 1314 person-year of follow-up and the relationship with AF type were given in Table 2 Table S2). Other baseline risk factors of TE events in this population were age ≥75-years-old, female and prior stroke or TIA (Table S3).  19 Besides, EORP-AF General Pilot Registry 11 reported comparable risk between NPAF and PAF under the high rates of anticoagulation use. Two reasons might explain the controversial results. On the one hand, the AF type might change during the follow-up. Furthermore, AF progression significantly increased the risk of adverse events, [20][21][22] especially during the progression period from paroxysmal AF to sustain AF. 23 The Chinese atrial fibrillation registry had a follow-up period of up to 4 years. Previous study 23 reported that the cumulative rate of progression to sustained type was only 2.6% at 1 year, 11.4% at 3 years, and 28.3% at 5 years, which indicated that the progression of AF type was time-dependent. Our study only had a 1-year follow up, which relatively reduced the confounding effects due to AF progression. On the other hand, anticoagulated therapy might diminish the difference in thromboembolic risk resulting from the AF pattern. According to the Chinese atrial fibrillation registry, 18 Stockholm cohort trial 13  We also evaluated the potential of the AF pattern in decisionmaking for stroke prevention. Though increasing evidence showed that patterns or duration of AF associated with adverse events, 26

| CONCLUSIONS
In Chinese non-anticoagulated AF patients, NPAF was an independent predictor of thromboembolism and mortality. The addition of NPAF to the CHA 2 DS 2 -VASc score allowed an improvement in the accuracy of the prediction of thromboembolic events. Further investigations are needed to confirm the results and assess its utility in therapy strategy.