External validation of a fingertip wearable device for obstructive sleep apnea diagnosis and split-night tracking of...

To externally validate a novel Type IV fingertip wearable device, TipTraQ, against polysomnography (PSG) for diagnosing obstructive sleep apnea (OSA) in the western-country population with both AASM AHI1A(3%) and AHI1B(4%) labeling rules. Furthermore, evaluate its ability to dynamically track positive airway pressure (PAP) treatment effectiveness using split-night titration as surrogate with short-term apnea-hypopnea index (AHI) and TipTraQ-derived Respiratory Events Index (TQ-REI) sequences.

In a prospective study at the Duke Sleep Disorders Center, 198 participants were enrolled, including a diagnostic validation cohort (n = 147) and a split-night titration cohort (n = 29). Short-term AHI and TQ-REI sequences were calculated using a 15-min moving window updated every 5 min to capture respiratory event dynamics. Agreement between TipTraQ and PSG was evaluated using standard performance metrics.

In the diagnostic cohort, TipTraQ correlated strongly with PSG for total sleep time (TST; PCC = 0.786) and REM duration (PCC = 0.655). Bland-Altman analysis revealed mean biases of -9.50 min for TST and -9.46 min for REM. Epoch-by-epoch sleep staging (Wake, REM, Non-REM) yielded a Cohen's Kappa of 0.567, with highest sensitivity in non-REM sleep (0.824). TQ-REI correlated highly with AHI under AASM scoring rules (PCC = 0.879 for 1A; 0.910 for 1B). In the titration cohort, TipTraQ maintained consistent performance. Analysis of short-term AHI and TQ-REI sequences demonstrated the device's ability to track respiratory event dynamics, yielding a median PCC of 0.882 (1A) and 0.872 (1B).

TipTraQ shows strong agreement with PSG for respiratory metrics and moderate agreement with PSG for sleep metrics, and captures temporal dynamics of sleep-disordered breathing events.