Can Oximetry-Derived Parameters Predict 3-Year Cardiovascular Mortality?

Area of Science:

• Cardiology
• Sleep Medicine
• Medical Informatics

Background:

• Obstructive sleep apnoea (OSA) is linked to cardiovascular disease (CVD).
• Current diagnostic tools like the apnoea-hypopnoea index (AHI) are insufficient for predicting CVD outcomes.
• Oximetry-derived parameters are explored as potential CVD predictors in OSA.

Purpose of the Study:

• To evaluate the predictive performance of oximetry-derived features, alone and combined with demographic and lifestyle data, for 3-year CVD mortality.
• To investigate the combined predictive power of various factors for CVD outcomes in OSA patients.

Main Methods:

• Utilized oximetry-derived parameters, demographic information, and lifestyle habits.
• Developed predictive models for 3-year CVD mortality.
• Assessed model performance using F1 scores on training and testing datasets.

Main Results:

• Oximetry-derived parameters alone showed limited predictive ability (F1 scores: 58.1% training, 53.65% testing).
• Combining oximetry data with demographic and lifestyle features significantly enhanced prediction (F1 scores: 79% training, 77.47% testing).
• The combined model achieved statistically significant results (p-values <0.05).

Conclusions:

• Oximetry-derived parameters are insufficient predictors of CVD mortality when used in isolation.
• Integrating demographic and lifestyle data with oximetry significantly improves the prediction of future CVD events.
• Future models for clinical use should incorporate multiple covariates for accurate CVD risk assessment in OSA patients.