Article Text
PDF
Abstract
Introduction Positional Obstructive Sleep Apnoea (P-OSA) – where apnoeas or hypopnoeas occur predominantly or exclusively in the supine sleep position – is common. 5% of patients attending our service for home sleep studies meet current diagnostic criteria. Sleep positional trainer (SPT) devices provide vibro-positional feedback to minimise supine sleep time. As discomfort difficulties which limited previous positional therapy approaches are potentially avoided, SPT devices may be a worthwhile alternative to CPAP therapy for P-OSA patients. Accuracy of SPT’s positional analysis has however not been benchmarked. We compared positional analysis from 2 SPT devices with that from polygraphy equipment and sleep video recordings.
Methods 21 patients attending for home (Somnoscreen, S-Med) or in-hospital (S-Med +video) sleep study wore additional collar (Nightshift, NS) and/or belt (Nightbalance, NB) SPT device, with vibration positional feedback disabled.%sleep time supine for each device was determined. Concordance analysis evaluating sleep position registered by each device for 15 min epochs of sleep was performed from the NS, S-Med and video sleep positional analysis reports (this data is not reported by the Nightbalance device software).
Results Despite standard instructions, device setup failure occurred with NS device in 6 patients and NB device in 9 patients. Bland-Altman plots (figure 1) demonstrate reasonable but incomplete agreement between SPT devices, S-Med and video for%sleep time supine. Kappa analysis demonstrated poor overall sleep positional concordance for NS vs S-Med (k=-0.002), NS vs video (k=0.196) and S-Med vs video (k=0.092). Individual patient data analysis demonstrated complete sleep time concordance between NS vs S-MED or video, and S-Med vs video in some patients but consistent concordance for all 3 modalities in any individual was not seen, and there was complete discordance between modalities in other patients. Mismatch between head and torso sleep position accounted for some of these differences.
Bland-Altman plots with comparison of% night lying supine determined by the four different modalities (Nightshift collar and Nightbalance belt SPT devices, Somnoscreen polygraphy device and sleep video analysis).
Conclusion In a real world setting we identified difficulties with SPT device setup and incomplete accuracy for sleep positional analysis by both SPT and home polygraphy equipment. Further consideration of cost-effective diagnostic precision, patient selection for SPT therapy, SPT device choice and setup and clinically relevant monitoring is required before SPT therapy is adopted.