Influence of obesity and overweight on the association between sleep-disordered breathing and atrial fibrillation: the DASAP-HF study

Background

The association between sleep apnea (SA) and atrial fibrillation (AF) has been well described. However, it remains unclear whether the association is causative or primarily dependent on shared comorbidities such as obesity. The Respiratory Disturbance Index (RDI) computed by an implantable cardioverter defibrillator (ICD) algorithm accurately identifies severe SA, whether central or obstructive in origin.

Purpose

In the present analysis we studied in patients with heart failure the contribution of obesity in the relationship between SA, measured by RDI, and AF.

Methods

Patients with ejection fraction ≤35% implanted with an ICD endowed with an algorithm (ApneaScan) that calculates the RDI each night, were enrolled and followed-up for 24 months. The weekly mean RDI value was considered, as calculated during the entire follow-up period. The endpoint was daily AF burden of ≥6 hours.

Results

164 patients (age 67±10 years, 75% male, ejection fraction 29±5%) had usable RDI values during the entire follow-up period. Body mass index (BMI) was <25 kg/m2 in 62 patients (normal), 25.0–29.9 kg/m2 in 66 patients (overweight), ≥ 30 kg/m2 in 36 patients (obese). When compared with normal patients (31±11 episodes/h), the average RDI value calculated during the entire follow-up period did not differ in overweight patients (35±13 episodes/h, p=0.114), but was significantly higher in obese patients (39±12 episodes/h, p=0.002). During follow-up, AF burden ≥6 hours/day was documented in 48 (29%) patients (BMI ≥ versus <25 kg/m2; HR: 1.47, 95% CI: 0.83–2.60, p=0.197; BMI ≥ versus <30 kg/m2; HR: 0.98, 95% CI: 0.46–2.09, p=0.963). Based on the ROC curve analysis, average RDI ≥37 episodes/h maximized sensitivity and specificity for the prediction of AF (Area under the curve: 0.63, 95% CI: 0.55–0.70, p=0.011). Device-detected RDI ≥37 episodes/h was associated with the occurrence of AF on univariate analysis (HR: 3.88, 95% CI: 2.02–7.44, p<0.001), as well as after correction for either BMI ≥25 kg/m2 (HR: 3.76, 95% CI: 1.94–7.26, p<0.001), or BMI ≥30 kg/m2 (HR: 4.15, 95% CI: 2.15–8.04, p<0.001).

Conclusions

In heart failure patients, we confirmed the association between ICD-detected SA and AF, an association that persisted independent of patient body habitus.

Funding Acknowledgement

Type of funding sources: Private grant(s) and/or Sponsorship. Main funding source(s): The study is supported by a research grant from Boston Scientific