Glycemic control and atrial fibrillation: baseline HbA1c predicts risk, early HbA1c change does not
Y Elhayani, S Shai Yaari, D Ostrovsky, Y Konstantino, A Star, D Shamia, M Haim, H Alnsasra, G TsabanAbstract
Background
Diabetes is associated with increased risk of atrial fibrillation (AF). Whether glycemic control modifies this risk over the long term remains unclear. We examined the association between hemoglobin A1c (HbA1c) and incident AF, and evaluated whether early changes in HbA1c predict subsequent AF risk.
Methods
We analyzed nationwide longitudinal data from a large health maintenance organization insuring more than 4.8 million members. Adults aged 35 years or older in 2005 with at least 2 prior insurance years and at least two documented HbA1c measurements were eligible; patients with previously documented AF were excluded. We first assessed baseline HbA1c as the exposure using restricted cubic splines to characterize the dose response relationship and identify a risk threshold. We then restricted the analysis to patients with baseline HbA1c at or above the threshold and evaluated whether HbA1c trajectories during the first 5 years were associated with incident AF occurring between 5 and 20 years from baseline. Percent change in HbA1c was defined as (HbA1c at year 5 minus HbA1c at year 0) divided by HbA1c at year 0, multiplied by 100. Stable HbA1c was defined as within ±5% of baseline and served as the reference group. Cause-specific hazard models were used with death censored, incorporating time-varying comorbidities that entered the model at first diagnosis. Models adjusted for sex, age, baseline smoking, time-varying estimated glomerular filtration rate, and time-varying clinical covariates including mitral and aortic valve disease, obstructive sleep apnea, thyroid dysfunction, hypertrophic cardiomyopathy, heart failure, myocardial infarction, cerebrovascular disease, peripheral vascular disease, chronic pulmonary disease, and malignancy. Baseline HbA1c was included in trajectory models.
Results
Baseline spline analyses demonstrated increased AF risk beginning at HbA1c 6.5% (Figure). Among patients with baseline HbA1c at or above 6.5%, higher baseline HbA1c remained associated with increased AF risk (hazard ratio 1.03 per 1% higher HbA1c, 95% CI 1.01 to 1.04; p < 0.001). In contrast, early HbA1c change over 5 years was not associated with long-term AF risk during years 5 to 20. Relative to stable HbA1c (±5%), hazard ratios were 1.02 (95% CI 0.96 to 1.08; p = 0.583) for at least 10% reduction, 1.05 (95% CI 1.00 to 1.11; p = 0.061) for 5% to 10% reduction, 1.02 (95% CI 0.97 to 1.07; p = 0.402) for 5% to 10% increase, and 1.01 (95% CI 0.96 to 1.06; p = 0.679) for at least 10% increase.
Conclusions
In this nationwide cohort, AF risk increased above an HbA1c threshold of 6.5%, and higher baseline HbA1c was associated with greater AF risk among patients above this threshold. However, changes in HbA1c over the first 5 years were not associated with long-term AF risk over the subsequent 5 to 20 years, after comprehensive adjustment and accounting for competing mortality.Figure