Construction of a diagnostic model for chronic mountain sickness among young male migrants to high-altitude areas

To analyze the risk factors for chronic mountain sickness (CMS) in young male migrants living in high-altitude areas and to construct a diagnostic model and evaluate its diagnostic efficacy.

From June 10 to December 29, 2023, a cross-sectional study was conducted on young male migrants subjected with convenience sampling who had been living in high-altitude areas (4500~5000 m) for 6 months or longer. Their demographic data were collected and blood samples were collected for laboratory test. According to the Qinghai Score for Chronic Mountain Sickness, they were divided into CMS group and non-CMS group. Then the participants were randomly divided into a training set and a test set in a ratio of 8:2. Independent risk factors for CMS occurrence were screened out, through random forest variable importance ranking, univariate and multivariable logistic regression analysis, and a diagnostic model was constructed based on these factors. Receiver operating characteristic (ROC) curve analysis, calibration curve analysis, clinical decision curve analysis, and influence curve analysis were used to comprehensively evaluate the diagnostic performance of the model.

According to the inclusion and exclusion criteria, 308 out of 376 participants were finally subjected, and 17.53% of them were diagnosed with CMS. The major clinical symptoms of the CMS patients were dyspnea or palpitations (79.63%) and sleep disorders (85.19%). Further analysis revealed that creatine kinase-MB/creatine kinase (CK-MB/CK, OR=2.17, 95%CI:1.43~3.28), high-altitude residence time (OR=2.44, 95%CI:1.08~5.54), and body mass index (BMI, OR=1.62, 95%CI:1.05~2.50) were 3 major independent risk factors for CMS. The area under the curve (AUC) value of the CMS diagnostic model in the training set and test set was 0.821(95%CI:0.756~0.886) and 0.821(95%CI:0.700~0.944), the specificity was 66.30% and 73.90%, the sensitivity was 89.50% and 81.20%, respectively, indicating good discrimination ability. Hosmer-Lemeshow goodness-of-fit test showed consistency between predicted results and actual observations (χ²=10.029, P=0.263; χ²=4.477, P=0.812). Clinical decision curve analysis demonstrated that within the threshold probability range from 0.1 to 0.7, the net benefit of the model exceeded both full intervention and no intervention strategies. The influence curve analysis showed high consistency between the model predictions and actual incidence when the threshold probability exceeded 0.4. These two analyses together confirmed the clinical application value of the model.

CK-MB/CK, high-altitude residence time and BMI are independent risk factors for CMS, and their diagnostic model helps identify potential individuals at risk for CMS. Early intervention can prevent the harm of CMS to the health of young men migrating to high-altitude areas.