Potential determinants of nocturnal hypoxemic burden and its association with atrial fibrillation recurrence after catheter ablation

Abstract

Abstract Background Sleep-disordered breathing (SDB) is present in approximately half of patients with atrial fibrillation (AF) scheduled for catheter ablation (CA) and coincides with an elevated nocturnal hypoxemia. Purpose To evaluate determinants of nocturnal hypoxemic burden metrics and its association with AF recurrence after CA in patients undergoing a systematic SDB management and treatment pathway. Methods Consecutive AF patients scheduled for CA without history of SDB and/or SDB screening were included. Patients were referred to a virtual SDB management pathway including mobile health supported overnight home sleep test. An apnoea-hypopnoea index (AHI) of 5 to <15 and ≥ 15 were interpreted as mild and moderate-to-severe SDB, respectively. Oximetry data from home sleep tests were analyzed by a computer algorithm to evaluate metrics of nocturnal hypoxemic burden: time spent below 90% oxygen desaturation (T90) based on acute desaturations (T90 desat) or non-specific drifts in oxygen desaturation (T90 non-spec). Patients were monitored for AF recurrences through standard 12-lead ECG or Holter ECG and/or smartphone-based photophletysmography (PPG) performed at 3- and 12-month follow up. Multiple regression analysis was employed to (1) evaluate determinants of nocturnal hypoxemic burden metrics and (2) identify predictors of AF recurrence during 12 months after CA (excluding 3 months blanking period). Results The analysis included 210 patients (42% female, mean age 64± 9 years). Median AHI was 14.7 [6.7-24.7], median T90desat was 0.25 [0.03-1.42] minutes, T90non-spec 0.26 [0.08-0.78] minutes. Overall, 101 (48%) patients were diagnosed with moderate-to-severe SDB and 68 (32%) had mild SDB. Of patients with moderate-to- severe SDB and mild SDB, 87% and 46%, respectively, received a SDB therapy recommendation. Multiple linear regression analysis identified diabetes mellitus (B 8.611 95%CI 1.350-15.871, p=0.020) and AHI ≥ 15 (B 5.886, 95%CI 1.793-9.979, p= 0.005) as predictors of T90non-spec. Thyroid dysfunction (B 1.849, 95%CI 0.623-3.075, p=0.003) and AHI ≥ 15 (B 2.161, 95%CI 1.320-3.001, p<0.001) were significant predictors of T90desat. During 12 months after CA, 65 (31%) patients had experienced at least one AF recurrence after the blanking period. After multiple Cox-regression, age (HR 1.032, 95%CI 1.002-1.063, p= 0.035) and BMI ≥ 25 (HR 2.029, 95%CI 1.032-3.989, p=0.040) remained predictors for AF recurrences. Conclusions In AF patients scheduled for CA, SDB (AHI ≥ 15) and thyroid dysfunction were predictors for nocturnal hypoxemic burden with repetitive desaturations (T90desat). SDB and diabetes mellitus were predictors for nocturnal hypoxemic burden without repetitive desaturations (T90non-spec). In patients with AF who followed a virtual SDB management and treatment pathway, SDB presence and nocturnal hypoxemic burden metrics did not emerge as predictors, whereas age and BMI (≥25) remained significant predictors for AF recurrence.