A Cluster Analysis of the Japanese Multicenter Outpatient Registry of Patients With Atrial Fibrillation

Taku Inohara, Jonathan P. Piccini, Kenneth W. Mahaffey, Takehiro Kimura, Yoshinori Katsumata, Kojiro Tanimoto, Kohei Inagawa, Nobuhiro Ikemura, Ikuko Ueda, Keiichi Fukuda, Seiji Takatsuki, Shun Kosaka

Research output: Contribution to journalArticle

Abstract

Recently, cluster analysis was used to identify unique clinically relevant phenotypes of atrial fibrillation (AF) in a cohort from the United States (US) and classified clusters according to the presence of comorbid behavioral disorders, those with conduction disorders, or atherosclerotic comorbidities. Whether these phenotypes are consistent in AF cohorts outside the US remains unknown. Thus, we sought to conduct a cluster analysis in a cohort of Japanese AF patients. We conducted a cluster analysis of phenotypic data (46 variables) in an AF patient cohort recruited from 11 Japanese sites participating in the KiCS-AF Registry. Overall, 2,458 AF patients (median [IQR] age, 68.0 [60.0 to 76.0]; 30.3% female; median [IQR] CHA2DS2-Vasc, 2 [1, 3]) were analyzed. Similar to the US cohort, atherosclerotic comorbidities were identified as distinguishing factors to characterize clusters. Distribution of AF type and left atrial (LA) size substantially varied and was the key feature for cluster formation. CHA2DS2-Vasc score also contributed to cluster formation, although behavioral disorders and/or conduction disorders did not readily characterize clusters. Subsequently, the cohort was classified into 3 clusters: (1) Younger paroxysmal AF (n = 1,190); (2) Persistent/permanent AF with LA enlargement (n = 1,143); and (3) Atherosclerotic comorbid AF in elderly patients (N = 125). In conclusion, conventional classifications, such as atherosclerotic risk factors and CHA2DS2-Vasc score contributed to cluster formation in mutually, whereas in nonatherosclerotic clusters, AF type or LA size rather than the presence or absence of behavior risk factors or sinus node dysfunction (tachy-brady syndrome) seemed to contribute to cluster formation in the Japanese cohort.

Original languageEnglish
JournalAmerican Journal of Cardiology
DOIs
Publication statusPublished - 2019 Jan 1

Fingerprint

Atrial Fibrillation
Cluster Analysis
Registries
Outpatients
Comorbidity
Phenotype
Sick Sinus Syndrome

ASJC Scopus subject areas

  • Cardiology and Cardiovascular Medicine

Cite this

A Cluster Analysis of the Japanese Multicenter Outpatient Registry of Patients With Atrial Fibrillation. / Inohara, Taku; Piccini, Jonathan P.; Mahaffey, Kenneth W.; Kimura, Takehiro; Katsumata, Yoshinori; Tanimoto, Kojiro; Inagawa, Kohei; Ikemura, Nobuhiro; Ueda, Ikuko; Fukuda, Keiichi; Takatsuki, Seiji; Kosaka, Shun.

In: American Journal of Cardiology, 01.01.2019.

Research output: Contribution to journalArticle

Inohara, Taku ; Piccini, Jonathan P. ; Mahaffey, Kenneth W. ; Kimura, Takehiro ; Katsumata, Yoshinori ; Tanimoto, Kojiro ; Inagawa, Kohei ; Ikemura, Nobuhiro ; Ueda, Ikuko ; Fukuda, Keiichi ; Takatsuki, Seiji ; Kosaka, Shun. / A Cluster Analysis of the Japanese Multicenter Outpatient Registry of Patients With Atrial Fibrillation. In: American Journal of Cardiology. 2019.
@article{613c32be70e04f31a6d685532d5a2adc,
title = "A Cluster Analysis of the Japanese Multicenter Outpatient Registry of Patients With Atrial Fibrillation",
abstract = "Recently, cluster analysis was used to identify unique clinically relevant phenotypes of atrial fibrillation (AF) in a cohort from the United States (US) and classified clusters according to the presence of comorbid behavioral disorders, those with conduction disorders, or atherosclerotic comorbidities. Whether these phenotypes are consistent in AF cohorts outside the US remains unknown. Thus, we sought to conduct a cluster analysis in a cohort of Japanese AF patients. We conducted a cluster analysis of phenotypic data (46 variables) in an AF patient cohort recruited from 11 Japanese sites participating in the KiCS-AF Registry. Overall, 2,458 AF patients (median [IQR] age, 68.0 [60.0 to 76.0]; 30.3{\%} female; median [IQR] CHA2DS2-Vasc, 2 [1, 3]) were analyzed. Similar to the US cohort, atherosclerotic comorbidities were identified as distinguishing factors to characterize clusters. Distribution of AF type and left atrial (LA) size substantially varied and was the key feature for cluster formation. CHA2DS2-Vasc score also contributed to cluster formation, although behavioral disorders and/or conduction disorders did not readily characterize clusters. Subsequently, the cohort was classified into 3 clusters: (1) Younger paroxysmal AF (n = 1,190); (2) Persistent/permanent AF with LA enlargement (n = 1,143); and (3) Atherosclerotic comorbid AF in elderly patients (N = 125). In conclusion, conventional classifications, such as atherosclerotic risk factors and CHA2DS2-Vasc score contributed to cluster formation in mutually, whereas in nonatherosclerotic clusters, AF type or LA size rather than the presence or absence of behavior risk factors or sinus node dysfunction (tachy-brady syndrome) seemed to contribute to cluster formation in the Japanese cohort.",
author = "Taku Inohara and Piccini, {Jonathan P.} and Mahaffey, {Kenneth W.} and Takehiro Kimura and Yoshinori Katsumata and Kojiro Tanimoto and Kohei Inagawa and Nobuhiro Ikemura and Ikuko Ueda and Keiichi Fukuda and Seiji Takatsuki and Shun Kosaka",
year = "2019",
month = "1",
day = "1",
doi = "10.1016/j.amjcard.2019.05.071",
language = "English",
journal = "American Journal of Cardiology",
issn = "0002-9149",
publisher = "Elsevier Inc.",

}

TY - JOUR

T1 - A Cluster Analysis of the Japanese Multicenter Outpatient Registry of Patients With Atrial Fibrillation

AU - Inohara, Taku

AU - Piccini, Jonathan P.

AU - Mahaffey, Kenneth W.

AU - Kimura, Takehiro

AU - Katsumata, Yoshinori

AU - Tanimoto, Kojiro

AU - Inagawa, Kohei

AU - Ikemura, Nobuhiro

AU - Ueda, Ikuko

AU - Fukuda, Keiichi

AU - Takatsuki, Seiji

AU - Kosaka, Shun

PY - 2019/1/1

Y1 - 2019/1/1

N2 - Recently, cluster analysis was used to identify unique clinically relevant phenotypes of atrial fibrillation (AF) in a cohort from the United States (US) and classified clusters according to the presence of comorbid behavioral disorders, those with conduction disorders, or atherosclerotic comorbidities. Whether these phenotypes are consistent in AF cohorts outside the US remains unknown. Thus, we sought to conduct a cluster analysis in a cohort of Japanese AF patients. We conducted a cluster analysis of phenotypic data (46 variables) in an AF patient cohort recruited from 11 Japanese sites participating in the KiCS-AF Registry. Overall, 2,458 AF patients (median [IQR] age, 68.0 [60.0 to 76.0]; 30.3% female; median [IQR] CHA2DS2-Vasc, 2 [1, 3]) were analyzed. Similar to the US cohort, atherosclerotic comorbidities were identified as distinguishing factors to characterize clusters. Distribution of AF type and left atrial (LA) size substantially varied and was the key feature for cluster formation. CHA2DS2-Vasc score also contributed to cluster formation, although behavioral disorders and/or conduction disorders did not readily characterize clusters. Subsequently, the cohort was classified into 3 clusters: (1) Younger paroxysmal AF (n = 1,190); (2) Persistent/permanent AF with LA enlargement (n = 1,143); and (3) Atherosclerotic comorbid AF in elderly patients (N = 125). In conclusion, conventional classifications, such as atherosclerotic risk factors and CHA2DS2-Vasc score contributed to cluster formation in mutually, whereas in nonatherosclerotic clusters, AF type or LA size rather than the presence or absence of behavior risk factors or sinus node dysfunction (tachy-brady syndrome) seemed to contribute to cluster formation in the Japanese cohort.

AB - Recently, cluster analysis was used to identify unique clinically relevant phenotypes of atrial fibrillation (AF) in a cohort from the United States (US) and classified clusters according to the presence of comorbid behavioral disorders, those with conduction disorders, or atherosclerotic comorbidities. Whether these phenotypes are consistent in AF cohorts outside the US remains unknown. Thus, we sought to conduct a cluster analysis in a cohort of Japanese AF patients. We conducted a cluster analysis of phenotypic data (46 variables) in an AF patient cohort recruited from 11 Japanese sites participating in the KiCS-AF Registry. Overall, 2,458 AF patients (median [IQR] age, 68.0 [60.0 to 76.0]; 30.3% female; median [IQR] CHA2DS2-Vasc, 2 [1, 3]) were analyzed. Similar to the US cohort, atherosclerotic comorbidities were identified as distinguishing factors to characterize clusters. Distribution of AF type and left atrial (LA) size substantially varied and was the key feature for cluster formation. CHA2DS2-Vasc score also contributed to cluster formation, although behavioral disorders and/or conduction disorders did not readily characterize clusters. Subsequently, the cohort was classified into 3 clusters: (1) Younger paroxysmal AF (n = 1,190); (2) Persistent/permanent AF with LA enlargement (n = 1,143); and (3) Atherosclerotic comorbid AF in elderly patients (N = 125). In conclusion, conventional classifications, such as atherosclerotic risk factors and CHA2DS2-Vasc score contributed to cluster formation in mutually, whereas in nonatherosclerotic clusters, AF type or LA size rather than the presence or absence of behavior risk factors or sinus node dysfunction (tachy-brady syndrome) seemed to contribute to cluster formation in the Japanese cohort.

UR - http://www.scopus.com/inward/record.url?scp=85069639114&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85069639114&partnerID=8YFLogxK

U2 - 10.1016/j.amjcard.2019.05.071

DO - 10.1016/j.amjcard.2019.05.071

M3 - Article

C2 - 31350002

AN - SCOPUS:85069639114

JO - American Journal of Cardiology

JF - American Journal of Cardiology

SN - 0002-9149

ER -