Clustering analysis of acute mountain sickness susceptibility among young adults during rapid ascent using low-altitude step test data

Acute mountain sickness (AMS) compromises health and work efficiency after rapid ascent. This study aimed to explore a low-altitude-data-based method for classifying AMS-susceptible young adults. Forty-four low-altitude residents (18-31 years) completed a step test at 50 m. Measurements included maximal oxygen uptake (V̇O₂max), peripheral oxygen saturation at rest (re_SpO₂), exercise (ex_SpO₂), and recovery (rec_SpO₂), heart rate at rest (re_HR), exercise (ex_HR), and recovery (rec_HR), and step index. Participants ascended to 3650 m within 6 h, and AMS was assessed using the 2018 Lake Louise Score (LLS). Associations with LLS were examined before k-means clustering. Neither re_SpO₂ nor re_HR was significantly correlated with LLS, whereas non-resting-state variables (V̇O₂max, ex_SpO₂, rec_SpO₂, ex_HR, rec_HR, and step index) were. Clustering using V̇O₂max, ex_SpO₂, rec_SpO₂, and step index showed strong structure (silhouette coefficient = 0.767) and 93.18% agreement with LLS. Step index and ex_SpO₂ showed the greatest between-cluster separation, followed by rec_SpO₂ and V̇O₂max. Exercise- and recovery-phase variables from a low-altitude step test may classify AMS susceptibility in young adults via unsupervised learning. Evaluated after rapid ascent to 3650 m, this noninvasive, low-cost, and practical approach shows potential for pre-ascent AMS susceptibility screening in healthy young adults.