Prediction of new-onset atrial fibrillation for general population in Asia: A comparison of C2HEST and HATCH scores

https://doi.org/10.1016/j.ijcard.2020.03.036Get rights and content

Highlights

  • This is a big data analysis.

  • This is a nationwide study.

Abstract

Purpose

To explore the use of the C2HEST and HATCH scores to predict the incidence of atrial fibrillation (AF) in Asians.

Method

The predictive capability of AF of C2HEST and HATCH scores was estimated by area under the receiver operating characteristic curve (AUROC). DeLong test was used to compared the difference of AUROC between the 2 scores.

Result

A total of 692,691 subjects were investigated. The risk of AF increased with increasing C2HEST and HATCH scores (p for trend<0.001). The AUROC for C2HEST and HATCHs in predicting AF occurrence was 0.7895 and 0.7711, respectively. C2HEST score had a significant better capability for AF stratification than HATCH score (DeLong test <0.001).

Conclusion

Higher C2HEST and HATCH scores were more strongly associated with the incidence of AF. The C2HEST score appeared to be more predictive of AF than the HATCH score.

Introduction

Risk stratification schemes for atrial fibrillation (AF) in general population have been widely investigated recently [[1], [2], [3], [4], [5], [6], [7]]; mostly because of the emerging huge burden of the associated morbidities and mortality [[8], [9], [10]]. Several AF prediction modalities have been studied; to our knowledge, most are designed for western population [[1], [2], [3], [4], [5]]. Additionally, these scores are relatively complex and not easily approached [[1], [2], [3], [4]].

The C2HEST score - predicting AF in patients without structural heart disease: C2: CAD/COPD (1 point each); H: hypertension (1 point); E: elderly (age ≥ 75 years, 2 points); S: systolic HF (2 points); and T: thyroid disease (hyperthyroidism, 1 point) [11]. The HATCH score - prediction of progression of paroxysmal AF to sustained forms of AF: (hypertension (1 point), age > 75 years (1 point), cerebrovascular accident (CVA) or transient ischemic attack (TIA) (2 points), chronic obstructive pulmonary disease (1 point), and heart failure (2 points) [6,12].

The predictive value of the C2HEST and HATCH scales have been evaluated in Asians [6,[11], [12]]; to our knowledge, comparison of the use of scales for risk stratification in general population in Asian remained unknown. To add data to what is known in this field, we aimed to explore and to compare the predictive capacity of C2HEST and HATCH scores in predicting incident AF in general population, analyzing Taiwanese administrative data to describe the predictive value of each scoring system and to evaluate the discriminative power of both scores using receiver operating characteristics analyses.

Section snippets

Data sources

Taiwan government built a single-payer health insurance database, National Health Insurance Research Database (NHIRD), which included nearly 99% of Taiwan population and the comprehensive health claim data. We conducted this study by using Longitudinal Health Insurance Database 2000 (LHID 2000), which was a subset of the NHIRD and randomly selected 1 million study subjects for research, with data available from 1996 to 2013. The privacy of all subjects was protected by identification encryption

Results

Table 1 shows the baseline characteristics of study subjects. Most individuals were aged 64 years or younger (89.1%), and the mean age was 41.3 years. The distribution of female and male was approximately each half; 4120 (0.59%) had heart failure, 43,301 (6.25%) had coronary artery disease, 25,379 (3.66%) had COPD, 99794 (14.4%) had hypertension, 3206 (0.46%) had hyperthyroidism, 7254 (1.05%) had CVA or TIA. There were 60,935 (8.80%) patients with hyperlipidemia, 21,357 (3.08%) with sleep

Discussion

We studied 692,691 subjects in the Longitudinal Health Insurance Database 2000 in Taiwan. After adjustment for the potential confounding factors other than the components of these scores, incidence and risk of new-onset AF were found to be increased progressively in line with increasing C2HEST and HATCH scores, but the c-statistic is greater for C2HEST score.

The research is clinically relevant for tackling this important clinical comorbidity in general population. So far, the 2 scores had not

Conclusion

Our results showed that these two scores are relatively optimal and might support them as a making-decisions score when predicting AF incidence in Asians, especially for C2HEST score.

CRediT authorship contribution statement

Wei-Syun Hu: Conceptualization, Data curation, Formal analysis, Funding acquisition, Investigation, Methodology, Project administration, Resources, Software, Supervision, Validation, Visualization, Writing - original draft, Writing - review & editing. Cheng-Li Lin: Formal analysis.

Declaration of competing interest

None.

Acknowledgements

This study was supported in part by Taiwan Ministry of Health and Welfare Clinical Trial and Research Center of Excellence (MOHW109-TDU-B-212-114004).

References (16)

There are more references available in the full text version of this article.

Cited by (0)

View full text