A composite sleep and pulmonary phenotype predicting hypertension.

Area of Science:

• Sleep Medicine
• Cardiovascular Health
• Pulmonary Function

Background:

• Sleep Disordered Breathing (SDB) and sleep disturbances are linked to hypertension.
• The standard Apnea Hypopnea Index (AHI) is insufficient for identifying patients who will benefit from SDB treatment regarding blood pressure.
• Novel predictive markers are needed for hypertension development in SDB patients.

Purpose of the Study:

• To utilize machine learning for identifying sleep and pulmonary measures that jointly predict hypertension development.
• To develop a more accurate predictive model for incident hypertension compared to the AHI.

Main Methods:

• Applied feature screening and Elastic Net penalized regression to polysomnography and lung function data from the Sleep Heart Health Study (SHHS).
• Utilized a wide array of polysomnography measures and lung function data.
• Validated findings in the Multi-Ethnic Study of Atherosclerosis (MESA) cohort.

Main Results:

• A composite of pulmonary function and 18 sleep phenotypes predicted incident hypertension (OR=1.43, 95% CI [1.14, 1.80] per SD).
• The AHI showed weak association with incident hypertension (OR=1.13, 95% CI [0.97, 1.33] per SD).
• The new phenotype was cross-sectionally associated with hypertension in the MESA cohort (OR=1.26, 95% CI [1.10, 1.45]).

Conclusions:

• A unique combination of sleep and pulmonary function measures offers superior prediction of hypertension compared to the AHI.
• The composite measure includes indices of apnea and hypopnea event durations.
• Shorter sleep event durations are associated with an increased risk of hypertension.