ABSTRACT
Decreased oxygen saturation (SO2) at high altitude is associated with potentially life-threatening diseases, e.g., high-altitude pulmonary edema. Wearable devices that allow continuous monitoring of peripheral oxygen saturation (SpO2), such as the Garmin Fenix® 5X Plus (GAR), might provide early detection to prevent hypoxia-induced diseases. We therefore aimed to validate GAR-derived SpO2 readings at 4559 m. SpO2 was measured with GAR and the medically certified Covidien Nellcor SpO2 monitor (COV) at six time points in 13 healthy lowlanders after a rapid ascent from 1130 m to 4559 m. Arterial blood gas (ABG) analysis served as the criterion measure and was conducted at four of the six time points with the Radiometer ABL 90 Flex. Validity was assessed by intraclass correlation coefficients (ICCs), mean absolute percentage error (MAPE), and Bland-Altman plots. Mean (±SD) SO2, including all time points at 4559 m, was 85.2 ± 6.2% with GAR, 81.0 ± 9.4% with COV, and 75.0 ± 9.5% with ABG. Validity of GAR was low, as indicated by the ICC (0.549), the MAPE (9.77%), the mean SO2 difference (7.0%), and the wide limits of agreement (-6.5; 20.5%) vs. ABG. Validity of COV was good, as indicated by the ICC (0.883), the MAPE (6.15%), and the mean SO2 difference (0.1%) vs. ABG. The GAR device demonstrated poor validity and cannot be recommended for monitoring SpO2 at high altitude.