Ambulatory polygraphy as a screening tool for obstructive sleep apnea in combination with semi-continuous rhythm monitoring in patients with atrial fibrillation: a pilot study

Abstract Background Obstructive sleep apnea (OSA) is a common, yet underdiagnosed risk factor in atrial fibrillation (AF) patients. Polysomnography (PSG) is the gold standard for OSA diagnosis but is also costly, time-consuming, and labor-intensive. (Cardio)-respiratory polygraphy (PG) offers an ambulatory screening for OSA detection. Studying the impact of OSA diagnosis and treatment on AF recurrence will require scalable tools for semi-continuous heart rhythm monitoring. Purpose The NOXFib-AF pilot study assessed the feasibility of PG as an OSA screening tool in patients with AF. The study monitored AF recurrences using a Fitbit smartwatch for heart rhythm monitoring with the FibriCheck algorithm. This pilot assessed the feasibility and patient compliance. Methods Overnight OSA testing was performed with the Noxturnal T3 (NOX-T3s) PG device used at home by AF patients, followed by four weeks of heart rhythm monitoring. Both semi-continuous monitoring with the Fitbit Versa 2 smartwatch with FibriCheck algorithm, and spot-check measurements on a smartphone (twice daily and in case of symptoms) were used. Patients with OSA suspicion on PG (Apnea-Hypopnea Index (AHI) cut-off ≥ 11.1 based on previous studies) were referred for PSG confirmation. Patient satisfaction with the PG and Fitbit smartwatch was evaluated through questionnaires. Results Thirty symptomatic AF patients underwent an overnight PG-based OSA screening, achieving an overall success rate of 96.7% (29/30). Interim analysis of 19 patients who underwent PSG evaluation upon referral showed that 12 (63.2%) were correctly diagnosed with at least moderate OSA (i.e., AHI ≥ 15) and were eligible for reimbursed CPAP treatment. PG and PSG agreement for OSA classification was fair (Cohen’s kappa = 0.230), with 68.4% (13/19) correctly classified (normal-mild, moderate-severe). There was high patient adherence to wearing the smartwatch for four weeks, with median compliance (i.e., adherence to the expected number of measurements) of 96.9% (IQR: 94.3% – 97.6%). A median of 150 (IQR: 145 – 151) daily automatic measurements were performed. Spot-check measurements via the smartphone also showed high compliance (110%, IQR: 100.0% – 135.5%) and high patient motivation (i.e., consistency of performing measurements; 90.3%, IQR: 67.7% – 96.8%). Patients were positive for both PG screening (Figure 1A) and Fitbit-based FibriCheck monitoring (Figure 1B). Conclusions PG is a feasible approach as a home screening tool for OSA, positively perceived by patients. Also, the acceptance of semi-continuous Fitbit-based FibriCheck monitoring was high. In a follow-up study, both tools will be used to evaluate the impact of a structured OSA screening and treatment trajectory on AF burden.