Prognosis and treatment of atrial fibrillation in Asian cities: 1-year review of the Asia-Pacific Heart Rhythm Society Atrial Fibrillation Registry

Abstract

Aims

The aim of this study is to describe the implementation of the current guidance for stroke prevention and treatment option in atrial fibrillation (AF) and to evaluate mortality and morbidity in relation to therapeutic decisions, including persistence with treatment at 1 year in Asia-Pacific regions.

Methods and results

We recruited 4664 patients consecutive in- and outpatients with AF who presented to cardiologists in five countries under the Asia-Pacific Heart Rhythm Society (APHRS) in whom 1-year follow-up was completed for 4003 (65.5% male; mean age 68.5 years). Oral anticoagulant (OAC) use remained high, 77% at follow-up, including 17% prescribed a vitamin K antagonist (VKA) and 60% a non-VKA oral anticoagulant (NOAC). At 1-year follow-up, 93% and 88% remained on a VKA or NOAC, respectively. With good adherence to OAC therapy, 1-year mortality was only 2.7%. Most deaths were non-cardiovascular (72.3%) and the 1-year incidence of stroke/transient ischaemic events (TIA) was low (<1%). Hospital readmissions were common for non-cardiovascular cases and atrial tachyarrhythmias. On multivariate analysis, independent baseline predictors of mortality and/or stroke/TIA/peripheral embolism were age, previous heart failure for >12 months, and malignancy. Independent predictors of mortality were age, chronic obstructive pulmonary disease, malignancy, and diuretic use. AF as a primary presentation was predictive of lower mortality and/or stroke/TIA/peripheral embolism as well as mortality.

Conclusion

In this 1-year analysis of the APHRS-AF registry, overall OAC use and persistence were high and were associated with low 1-year cardiovascular mortality and morbidity, but mortality and morbidity related to non-cardiovascular causes were high in AF patients, particularly from malignancy and pneumonia.

Graphical Abstract

Open in new tabDownload slide

• In this 1-year follow-up analysis of the APHRS-AF registry, we provide data on the first contemporary registry focused on management practices for atrial fibrillation (AF) among Asian cardiologists.

• Overall OAC use and persistence were high in patients with AF, especially with the introduction of NOAC into Asia, and was associated with a low 1-year cardiovascular mortality and morbidity.

• The 1-year mortality and morbidity related to non-cardiovascular causes were high in AF patients, particularly from malignancy and pneumonia.

Introduction

Despite a lower overall prevalence reported by observational studies of atrial fibrillation (AF) in Asians compared with other Western countries, Asia has a much higher overall disease burden due to its proportionally larger aged population.^1–3 Stroke is a disabling complication of AF that is of particular cause for concern in patients from Asia-Pacific regions. Implementing consensus expert recommendations about the use oral anticoagulants (OACs) in AF patients can considerably reduce stroke rates.⁴ Nonetheless caution should be exercised when aligning management of AF patients from the Asia-Pacific region to that of other Western countries. Current international guidelines and risk stratification tools for AF management are based predominantly on findings in populations from the Western countries, and may have limited relevance to Asian patients.^5,6 In addition, a vitamin K antagonist (VKA) or non-VKA oral anticoagulants (NOACs) is recommended for reducing the risk of stroke and thromboembolism in high-risk patients with AF.⁵ There are very limited prospective data on the management and treatment outcomes of AF in an Asian population. There is thus a need for a systematic prospective collection of contemporary data regarding the management and treatment of AF in Asia-Pacific countries.

In collaboration with the European Society of Cardiology (ESC), the Asia-Pacific Heart Rhythm Society (APHRS) created the first prospective registry for five Asian metropolitan cities (Hong Kong, South Korea, Japan, Singapore, and Taiwan) to systematically collect contemporary data regarding the management and treatment of AF. In the current analysis, we present the baseline and 1-year data from the APHRS-AF Registry, specifically focusing on symptoms, use of antithrombotic therapy, and rate vs. rhythm strategies, as well as determinants of mortality and stroke.

Methods

We adopted the study methods and protocol of the EURObservational Research Programme on Atrial Fibrillation (EORP-AF) Long-Term General Registry, as reported previously.^7,8 The registry was established in late 2015 and enrolment finished in early 2017. In brief, the registry population comprises consecutive in- and outpatients who presented with AF to a cardiologist, derived from 52 centres in five countries with a broad mix of tertiary and general hospitals. Consecutive AF patients seen by a cardiologist in those centres were screened at the time of presentation or clinic visit; and eligible patients approached for written informed consent according to local regulations. All patients enrolled had an ECG, including Holter, event monitor or implantable loop recorder-documented AF within the 12 months prior to enrolment.

After baseline assessment at enrolment, 1-year follow-up was performed by the local cardiologist investigator. The major study endpoints of interest were mortality, occurrence of stroke/thromboembolism, cardiovascular comorbidities, and hospital readmission within 1 year. As reported in the EORP-AF general registry,^7,8 stroke risk was categorized using the CHA₂DS₂-VASc score and further sub-categorized as ‘low risk’ [CHA₂DS₂-VASc score= 0 (male) or 1 (female)], ‘moderate risk’ [CHA₂DS₂-VASc score = 1 (males)] or ‘high risk’(CHA₂DS₂-VASc score ≥2). Similarly, bleeding risk was categorized using the HAS-BLED score.^7,8

The study protocol was approved by local ethics committee and trial was ClinicalTrials.gov (NCT04807049). Informed consent was obtained from all patients.

Statistical analyses

Univariate analysis was applied to both continuous and categorical variables. Continuous variables are reported as mean ± standard deviation and/or as median and interquartile range. Among-group comparisons were made using a non-parametric test (Kruskal–Wallis test). Categorical variables are reported as percentages. Among-group comparisons were made using a χ² test or Fisher’s exact test as appropriate.

Plots of Kaplan–Meier curves for time to all-cause death in relation to AF subtype were performed. The survival distribution between the types of AF was compared using the log-rank test. All variables at entry that were statistically significant at univariate analysis and variables considered of relevant clinical interest were included in the multivariable model (logistic regression) to identify the independent predictors of all-cause death and/or stroke/transient ischaemic attack (TIA)/peripheral embolism during the 1-year follow-up.

A two-sided P-value <0.05 was considered statistically significant. All analyses were performed using the SAS statistical software version 9.4 (SAS Institute, Inc., Cary, NC, USA).

Results

Table 1 summarizes the patient characteristics in relation to clinical subtype of AF. The patient disposition is shown in Figure 1, where 4003 (85.8%) of the 4664 patients who were enrolled and alive at baseline had at least one visit/contact during follow-up, while 118 (2.5%) died and 543 (11.6%) had unknown status or were lost to follow-up. The mean follow-up for the whole cohort was 364 days.

As expected, patients with long-standing persistent AF and permanent AF were older. Although there was no statistically significant difference in gender ratio between different AF subtypes, the highest prevalence of male gender was observed in patients with persistent AF. Of the patients with permanent AF, there were more who were at high risk of stroke according to the CHA₂DS₂-VASc score, as well as bleeding as reflected by HAS-BLED strata. On the contrary, more low-risk patients for stroke and bleeding presented with paroxysmal AF (Table 1).

Symptoms at follow-up

At 1-year follow-up, only 15.6% of patients were symptomatic. As shown in Supplementary material online, Table S1, all the AF symptoms, especially palpitation were significantly correlated with the EHRA score. Those with paroxysmal AF (19.6%) were the most frequent to report symptoms. The most common symptoms were palpitations (56.2%) and shortness of breath (30.1%). Patients with paroxysmal and persistent AF were more likely to have palpitations; those with long-standing persistent AF and permanent AF were more likely to report shortness of breath.

Antithrombotic therapy

Table 2 summarizes the prescription of anti-thrombotic agents before and after the 1-year follow-up visit. Overall, 77% of patients were prescribed an OAC at follow-up of whom 17% were prescribed a VKA and 60% a NOAC. Similar proportions of VKA and NOAC prescription were observed before and after 1-year follow-up (Table 2a). Among different subgroups of AF patients, OAC use was highest among those with persistent and permanent AF (86–90%), with significantly more VKA use in permanent AF and more NOACs used among patients with paroxysmal or persistent AF (Table 2a). Antiplatelet therapy was used in only 12% at follow-up and with no difference among different AF subgroups (Table 2a).

The use of antithrombotic therapy by stroke risk strata (based on the CHA₂DS₂-VASc score) is shown in Table 2b. An increasing trend to prescribe OACs including VKA and NOACs was observed from low to moderate and high-risk patients. Antiplatelet therapy (commonly aspirin) was more commonly prescribed to patients at moderate risk.

Changes in antithrombotic therapy use at 1-year follow-up in relation to the baseline entry visit are shown in Figure 2A. Of those prescribed a VKA, 62% remained on a VKA during follow-up; and 10.3% and 2.1% changed to NOACs or antiplatelet therapy alone, respectively. Of those prescribed an NOAC at baseline, 76% remained on an NOAC, and 0.5% and 1.1% had changed to a VKA or antiplatelet therapy alone, respectively. Of those on antiplatelet therapy, 28% were prescribed an OAC on follow-up, including 6% in combination with an antiplatelet agent.

The continued use of antithrombotic therapy at 1-year follow-up in relation to the baseline visit is shown in Figure 2B. Of those on a VKA or NOAC, 93% and 88% remained on their VKA or NOAC, respectively. Of those prescribed antiplatelet therapy, 84% were continuing treatment and only 8.5% had changed to an OAC.

Overall, reduced or adjusted dose NOACs were prescribed in 44% of patients at baseline, and remained similar at 44% and 45%, respectively, at pre-follow-up and after follow-up consultation (Supplementary material online, Table S2).

Rate and rhythm control drugs

Table 2c summarizes the rhythm/rate control drugs used at follow-up. Beta-blockers (50%) were the most common drugs used, especially among first diagnosed AF patients. Digitalis was used in 24% of patients with permanent AF. Overall, the use of Class Ic (11%) and III (9%) drugs was low but common in paroxysmal and persistent AF.

Interventions performed by 1-year follow-up were shown in Supplementary material online, Table S3. Electrical and pharmacological cardioversion had been performed in 3.9% and 4.1% of patients, respectively, especially among those with persistent or long-standing persistent AF. Catheter ablation had been performed in 8.9%, particularly among those with paroxysmal or persistent AF.

As shown in Supplementary material online, Table S4, the utilization of non-pharmacological interventions, especially catheter ablation was significantly correlated with AF symptoms as determined by European Heart Rhythm Association (EHRA) score.

Mortality and morbidity

After 1 year, 2.7% (120/4382) of the study patients had died. The highest mortality was in those with first diagnosed AF (5.7%) and permanent AF (4.7%). Kaplan–Meier curves for mortality between known AF subgroups are shown in Figure 3.

Table 3a summarizes the detailed causes of death for different AF subgroups. Overall, the reported aetiologies of death included cardiac (23.4%; 22/94), vascular (4.3%; 4/94), and non-cardiovascular (72.3%; 68/94), with no significant differences between AF subgroups. Cardiac causes were more common in patients with long-standing persistent AF. Vascular causes such as ischaemic or haemorrhagic stroke were rare in different AF subgroups. Among the non-cardiovascular causes of death, pneumonia (33.8%) and malignancy (22.1%) were the most common (Table 3a).

Of the 3978 patients who completed the 1-year follow-up visit, 325 (8.2%) had been readmitted for AF/atrial flutter, 222 (5.6%) for other cardiovascular events and 546 (13.7%) for non-cardiovascular events (including 43, 1.1% hospital admissions for major extracranial bleeding, Table 3b). Readmission for AF/atrial flutter was more common in first diagnosed AF patients, and readmission for heart failure or non-cardiovascular events was less common in patients with paroxysmal AF (Table 3b).

Multivariate analysis

On a stepwise model, multivariate predictors of stroke/TIA/peripheral embolism and/or mortality are shown in Supplementary material online, Table S5. For stroke/TIA/peripheral embolism and/or mortality, independent predictors were age [odds ratio (OR): 1.09, P < 0.001], previous heart failure for >12 months (OR: 2.21, P ≪ 0.001), and malignancy (OR: 2.12, P = 0.001). AF as a primary presentation was predictive of lower stroke/TIA/peripheral embolism and/or mortality (OR: 0.37, P < 0.001).

For mortality, independent predictors were age (OR: 1.09, P < 0.001), chronic obstructive pulmonary disease (OR: 1.70, P < 0.001), malignancy (OR: 2.278, P = 0.001), and diuretic use (OR: 2.75, P < 0.001). AF as a primary presentation was predictive of lower mortality (OR: 0.38, P < 0.001).

Discussion

This survey represents the first contemporary prospective registry focused on management practices among Asian cardiologists, with associated follow-up data. There are several major findings of this 1-year follow-up analysis of the APHRS-AF registry. First, most patients were relatively asymptomatic and only 1/8 of those with AF were symptomatic with palpitations and/or shortness of breath. Second, a relatively high proportion of patients (77%) were treated with an OAC: 60% prescribed an NOAC and 17% a VKA at 1-year follow-up. Third, rate control was the mainstay of treatment for most AF patients (72%), and rhythm control was infrequent, with electrical cardioversion and catheter ablation being performed in only 4% and 9%, respectively, and anti-arrhythmic agents using Class Ic and III agents were prescribed in only 11% and 9%, respectively. Fourth, overall 1-year mortality was only 2.7% in our AF patients and the major causes of death were non-cardiovascular including pneumonia (33.8%) and malignancy (22.1%). Fifth, 1-year incidence of TIA/stroke was low (<1%) but hospital readmissions were common for non-cardiovascular events (13.7%) and AF/atrial flutter (5.6%). Finally, although new-onset AF was associated with increased risk of stroke/TIA/peripheral embolism and/or mortality, AF as a primary presentation predicted lower stroke/TIA/peripheral embolism and/or mortality.

Prior studies have shown that patients in Asia were less likely to be prescribed OAC for stroke prevention in AF compared with other regions.^9–11 One of the potential reasons for this under-utilization is concern about the bleeding risk, especially intracranial bleeding due to VKA use in the Asian population.^1,12 Compared with VKAs, NOACs are associated with a lower risk of intracranial bleeding, especially in the Asian population.¹³ Therefore, the introduction of NOACs in Asian countries should have improved the utilization of OACs in Asian patients with AF.^14–16 OAC use (∼77%) in different Asian countries was much higher in the current APHRS-AF registry than that reported in other international registries (∼55%).^9–11 Moreover, among those patients treated with OACs, most were prescribed NOACs (78%) and only 22% were prescribed a VKA at 1-year follow-up.

As expected, an increasing trend to prescribe OACs including VKA and NOACs was observed from low-to-moderate and high-risk patients. Furthermore, our 1-year follow-up data show that continued use of an OAC was higher for both patients treated with a VKA (93%) and NOACs (88%) at baseline. These findings are of clinical importance since adherence and compliance with guideline recommendations for OACs in Asian AF patients have been shown to reduce stroke rates.⁴ Indeed, the high utilization and continued use of OACs observed in this study was associated with a very low incidence of stroke/TIA (<1%) at 1-year follow-up. Interestingly, this very low stroke/TIA rate was observed in this study despite the fact that up to 44% of patients on NOACS were prescribed with low reduced or adjusted dose.

In this cohort of Asian patients with AF, most were relatively asymptomatic. As a result, rhythm control therapies including use of antiarrhythmic drugs, cardioversion, and catheter ablation were less frequently used as they are mainly recommended for patients with symptomatic AF.^5,6 Rate control with a beta-blocker and/or digoxin was the most common treatment in different subtypes of AF.

In contrast to the EORP-AF Registry,⁸ we observed an overall lower mortality rate (2.7%) associated with AF in contemporary clinical practice among Asian cardiologists. Moreover, non-cardiovascular rather than cardiac were the most common causes of mortality in our Asian patients compared with European cohorts.⁸ While the reasons for the lower rate of cardiovascular mortality and morbidity in our Asian patients compared with EORP-AF population remain unclear, it is possible attributed to the higher proportion of patients with high risk CHA₂DS₂-VASc score in the EORP-AF cohort (81% vs. 73%).⁸ Among different subtypes of AF, the highest mortality was observed in patients with first-diagnosed AF (5.7%). Nonetheless, AF as the primary presentation predicted lower morbidity and mortality. These findings suggest that AF is a marker for a poor prognosis when associated with underlying medical comorbidities such as pneumonia or malignancy rather than directly contributing to the mortality in those with new-onset AF.^17,18 Indeed, other than age, the presence of obstructive pulmonary disease and malignancy were independent predictors of mortality in our cohort. Moreover, the main reasons for hospital readmission were non-cardiovascular, rather than atrial tachyarrhythmias and heart failure as reported in the EORP-AF Registry.⁸ Interestingly, diuretic use was associated with increased mortality in our study as well as in European cohorts.⁸ Although the reasons remain unclear, it is possible that changes in serum potassium level in AF patients treated with diuretics contributed to an increase in short-term mortality.¹⁹

Limitations

This study has several limitations. First, this was an observational registry study, but consecutive patients were recruited from different centres in contemporary clinical practice reflecting the country-specific patterns. Second, our patients in this registry were followed-up by cardiologists, and this might partly account for the high utilization of OAC. Third, ∼12% of patients were lost to follow-up (15%), although this figure is lower than the ∼15% observed in the EORP-AF Registry.⁸ Fourth, similar to the EORP-AF Registry,⁸ we did not record data on INR control in our patients treated with VKA. Therefore, it remains unclear whether the INR control in patients treated with VKA has any impact on their outcomes. Nonetheless, we did not observe any significant high ischaemia or bleeding events among those treated with VKA (data no show). Fifth, our results are limited by the ‘real-world’ observational design and their non-randomized nature in a broad mix of tertiary and general hospitals, and thus preclude any detailed analyses of the impact of different treatment interventions on the clinical outcomes in this registry. Sixthly, this registry included only five Asian metropolitan cities. Whether similar findings would be observed in other Asian countries remains unclear. Finally, the current analysis precludes any meaningful indirect comparison with EORP-AF Registry as the study centres selection and patients characteristics were not matched between those two cohorts.

Conclusions

We present the first contemporary registry focused on management practices among cardiologists in Asia, with associated follow-up data, conducted since publication of the new international guidelines on AF,^5,6 and the introduction of NOACs. Overall OAC use, especially NOACs in Asia, is increasing and 1-year persistence with therapy is high. Moreover, the 1-year mortality and morbidity related to AF in this study appears to be lower than that of the European cohort.

Supplementary material

Supplementary material is available at Europace online.

Funding

This study was an independent research grant by Pfizer and Bristol Myers Squibb (BMS) to Asia-Pacific Heart Rhythm Society.

Data availability

The data underlying this article will be shared on reasonable request to the corresponding author.

References

Author notes