Objective: The aim of the study was to validate the automatic and manual analysis of ApneaLink Ox™ (ALOX) in patients with suspected obstructive sleep apnea (OSA).
Methods: All patients with suspected OSA had a polysomnography (PSG) and an ALOX performed in the sleep laboratory. For automatic analysis, hypopnea was defined as a decrease in airflow ≥30 % of baseline for at least 10 s plus oxygen desaturation ≥3 or 4 %. While for the manual analysis, hypopnoea was considered when a reduction of airflow ≥30 % of ≥10 s plus oxygen desaturation ≥3 % or increase in cardiac rate ≥5 beats/min were identified or, when only a reduction of airflow ≥50 % was observed. OSA was defined as a respiratory disturbance index (RDI) ≥5. The apnea/hypopnea automatic index (AHI3-a, AHI4-a) and manual index were estimated. Receiver operating characteristics (ROC) analysis and the agreement between ALOX and PSG were performed.
Results: Fifty-five patients were included (38 men; mean age, 48.2; median, RDI 15.1; median BMI, 30 Kg/m(2)). The automatic analysis of ALOX under-estimated the RDI from PSG, mainly for the criterion of oxygen desaturation ≥4 % (AHI3-a-RDI, -3.6 ± 10.1; AHI4-a-RDI, -6.5 ± 10.9, p < 0.05). The autoscoring from ALOX device showed a better performance when it was set up to identify hypopneas with an oxygen desaturation criterion of ≥3 % than when it was configured with an oxygen desaturation criterion of ≥4 % (area under the receiver operator curves, 0.87 vs. 0.84). Also, the manual analysis was found to be better than the autoscoring set up with an oxygen desaturation of ≥3 % (0.923 vs. 0.87). The manual analysis showed a good interobserver agreement for the classification of patients with or without OSA (k = 0.81).
Conclusion: The AHI obtained automatically from the ApneaLink Ox™ using oxygen desaturation ≥3 % as a criterion of hypopnea had a good performance to diagnose OSA. The manual scoring from ApneaLink Ox™ was better than the automatic scoring to discriminate patients with OSA.