Comparison of residual apnea-hypopnea index given by positive airway pressure software and polysomnography in patients with sleep apnea syndrome

Purpose: the aim of this study was to compare residual apnea-hypopnea index (AHI) provided by positive airway pressure (PAP) device software to the index from a simultaneous polysomnogram in patients with obstructive sleep apnea syndrome (OSAS). Method: patients with OSAS diagnosis treated by PAP and referred for a polysomnography on PAP were eligible. Data recorded by the PAP software were downloaded after polysomnography. Correlations between manually-scored data and PAP-retrieved data softwares were compared (R software). Results: 65 patients were prospectively included. Apnea hypopnea index (AHI) and obstructive apnea index (OAI) assessed during polysmonographic recording significantly correlate with AHI and OAI provided by ventilator software. Mean bias between AHI assessed during polysomnographic recording (AHIPSG) and AHI provided by ventilator software (AHIVS) was 6.9 with a limit of agreement of - 14 to 28 (± 2 SD). Mean bias between OAI assessed during polysomnographic recording (OAIPSG) and OAI provided by ventilator software (OAIVS) was - 0.4 with a limit of agreement of - 9.6 to 8.8 (± 2 SD). An AHIVS higher than 7.75 had a sensibility of 0.83 and a specificity of 1 to detect an AHIPSG higher than 15/h. An OAIVS higher than 1.9 has a sensibility of 0.75 and a specificity of 0.79 to detect OAIPSG higher than 5/h. Conclusion: data provided by home PAP software does not alleviate the need of a polysomnographic record to assess PAP efficiency. However, we suggest AHI and OAI thresholds, which may be used to detect patients with a poorly controlled OSAS.