Abstract P167: Associations of Sleep-Disordered Breathing in Relation to Incident Atrial Fibrillation and Change in Serum NT-proBNP Levels Among Japanese: The Circulatory Risk in Communities Study (CIRCS)

Abstract

Introduction: Sleep-disordered breathing (SDB) is considered as a risk factor of cardiometabolic diseases. However, the evidence on prospective association between SDB and risk of heart failure is limited. Therefore, we aimed to examine whether SDB is associated with the incidence of atrial fibrillation (AF) which is linking to heart failure, and high serum NT-proBNP levels during follow-up in general population. Hypothesis: Since SDB is a known risk factor of hypertension, we hypothesized that SDB was associated with higher risk of atrial fibrillation and high serum NT-proBNP levels. Methods: Among 4,971 Japanese men and women, we measured the frequency of 3% or more oxygen desaturation during night (3%ODI) using Pulse-oximeter (Pulsox-3Si; Minolta Inc.) from 2000 through 2005 in three areas across Japan. SDB was defined as 5 or more events/hour of 3%ODI. Incidence of AF was documented by ECG in annual cardiovascular health check-up until the end of March, 2017. Among 2,588 participants who participated in both of baseline survey (2000-2005) and follow-up survey (2009-2011), we measured serum NT-proBNP levels by electrochemiluminescence immunoassay method (ECLusys NT-proBNP II; Roche Diagnostics K.K.) at both periods. High NT-proBNP levels were defined as 400 pg/mL or more. HRs and 95%CIs of incident AF were calculated by using the Cox regression. Difference in change of serum NT-proBNP levels were tested between the presence and absence of SDB using ANCOVA, and ORs and 95%CIs were estimated by using the logistic regression. Analyses were stratified jointly by sex and fifths of age, and adjusted for drinking status, and current smoking. Statistical analyses were performed using the Statistical Analysis System 9.4 (SAS Institute Inc.). All P values were two-sided, and statistical and borderline significances were defined as P<0.05 and <0.10, respectively. Results: During 42,437 person-years of follow-up, 79 individuals developed AF. Multivariable-adjusted HR of incident AF in relation to SDB was 1.58 (95%CI: 1.00-2.48). However, after further adjustment for BMI, the association was attenuated and became non-significant. Mean change of serum NT-proBNP levels during follow-up was 21.9 (95%CI: 16.5-27.4) pg/mL for non-SDB and 35.7 (27.0-44.5) pg/mL for SDB. Further adjustment for BMI made that association weak and of borderline significance. In those who was free of high serum NT-proBNP levels at baseline, multivariable-adjusted OR of high serum NT-proBNP levels for SDB was 1.64 (95%CI: 0.91-2.97) with borderline significance (P=0.10). Further adjusting for BMI weakened this association up to non-significant level. Conclusions: SDB is associated with a higher AF risk and increasing serum NT-proBNP levels independently of age, sex, and heavy drinking habit, but not BMI.