Body mass index, weight change, and long-term risk of incident atrial fibrillation in a nationwide cohort
S Shai Yaari, Y Elhayani, D Ostrovsky, M Haim, Y Konstantino, A Star, D Shamia, H Alnsasra, G TsabanAbstract
Background
Body mass index (BMI) is associated with incident atrial fibrillation (AF), but the long-term risk pattern across the BMI spectrum and the prognostic value of early BMI change in routine care remain incompletely defined.
Methods
We analyzed nationwide longitudinal data from a large health maintenance organization covering more than 4.8 million members. Adults aged ≥35 years in 2005 were eligible if they had at least two documented BMI measurements. We excluded individuals with fewer than two years of continuous prior insurance coverage and those with previously documented AF. The association between baseline BMI and incident AF was modeled using restricted cubic splines. Multivariable survival models adjusted for age, sex, renal function, smoking status, socioeconomic status, and both baseline and time-varying comorbidities, including aortic or mitral valve disease, obstructive sleep apnea, thyroid disease, hypertrophic cardiomyopathy, heart failure, myocardial infarction, prior stroke, peripheral vascular disease, pulmonary disease, and cancer. To account for the competing risk of all-cause mortality, we used generalized additive models. We then restricted the analysis to individuals with baseline BMI above the spline-identified risk threshold and examined whether changes in BMI during the first five years were associated with incident AF occurring between years 5 and 20. A BMI change within ±2.5% of baseline was defined as stable. Trajectory models were additionally adjusted for baseline BMI.
Results
The analytic cohort included 431,967 patients; 62,912 developed AF during follow-up (cumulative incidence 14.6%). The baseline BMI-AF association was nonlinear, with progressively higher AF risk above BMI 28 (Figure 1). The association between 5-year changes in BMI and long-term AF risk among patients with baseline BMI above 28 is shown in Figure 2. BMI reduction was associated with lower subsequent AF risk, including more than 10% reduction (adjusted hazard ratio 0.73, 95% CI 0.69 to 0.77). BMI increases were associated with graded higher risk: +2.5% to 5% (adjusted hazard ratio 1.04, 95% CI 1.01 to 1.08), +5% to 10% (1.18, 95% CI 1.12 to 1.23), and at least 10% (1.28, 95% CI 1.14 to 1.43).
Conclusions
In this nationwide cohort, AF risk increased above a BMI threshold of 28. Among patients with elevated BMI, early BMI reduction was associated with lower long-term AF risk, while weight gain conferred incremental excess risk.Figure 1Figure 2