The Cardio-Respiratory Polygraphy (CRP) level 3 is a simplified alternative for the diagnosis of Obstructive Sleep Apnea (OSA). CRP devices are provided with an autoanalysis software not yet validated.Outcome: To evaluate the level of concordance between CRP autoanalysis and results obtained after two expertise scorers review and correction.
30 consecutives patients older than 18 years, with suspected OSA studied with CRP level 3 (Stardust II, Respironics). Autoanalysis obtained by soft was analyzed. Amount of different respiratory events with their respective index were considered. Two expertise scorers simultaneously reviewed and corrected every study, epoch by epoch, using standard AASM criteria. These data were compared to those from autoanalysis.OSA severity was established by the Respiratory Disturbance Index (RDI), normal < 5 ev/h; mild 5–14.9; moderate 15–29.9 and severe > 30. The level of concordance in the definition of every respiratory event and in the final diagnostic result for every study between soft and scorers was evaluated with the percentage of concordance and the kappa coefficient, a kappa value > 0.7 was considered acceptable.
21/30 male; age 52.3 ± 13.4 y/o; BMI 29.8 ± 6.1. A total of 11.720 hours of CRP were reviewed in 60 seconds epochs. 6.720 respiratory events were analyzed. The mean RDI obtained by soft was 32.16 ± 20.59 ev/h; and 28.89 ± 24.01 by scorers. The level of concordance in the definition and classification of every event was 31% and the kappa coefficient 0.006. When we compared diagnosis and severity level for every patient according to RDI, the % of concordance resulted 73% and kappa coeff 0.62.
The level of concordance in the detection and classification of respiratory events between autoanalysis software and the expertise scorers correction is low. Diagnostic categorization by RDI has a level of correlation near to the value considered as acceptable.
Autoanalysis results would allow a quick first approach. Nevertheless, it does not seem to be proper to avoid detailed manual review of every study.
Facundo Nogueira, No Financial Disclosure Information; No Product/Research Disclosure Information