# Examining Sleep-Disordered Breathing Events Using Latent Profile Analysis

## TL;DR

This study identifies four distinct subtypes of sleep-disordered breathing based on sleep patterns and clinical features in a large patient group.

## Contribution

The study introduces a novel classification of sleep-disordered breathing subtypes using latent profile analysis.

## Key findings

- Four distinct subtypes of sleep-disordered breathing were identified in the study cohort.
- Class IV showed strong associations with older age, high BMI, and multiple clinical risk factors.
- The subtypes may help improve clinical management by guiding treatment based on disease subtype.

## Abstract

The clinical utility of the ratio of the apnea–hypopnea index (AHI) occurring during rapid eye movement (REM) and non-REM (NREM) sleep (AHIREM/AHINREM ratio) has been debated. We investigated the heterogeneity of REM and NREM sleep behaviors to identify unobserved distinct subtypes of sleep-disordered breathing (SDB) and examine their demographic and clinical features. The present study used a sample of 3626 adult patients who underwent diagnostic polysomnography evaluations at the Sleep Disorders Center of King Abdulaziz Medical City in Riyadh, Saudi Arabia. Latent profile analysis was performed to categorize subjects into distinct profiles of SDB based on AHIREM, AHINREM, and AHIREM/AHINREM ratio. A multinomial logistic model estimated the odds ratio of SDB profiles. Four distinct subtypes of SDB were identified: Class I (low AHIREM; 75.9%) included patients with normal SDB events during REM sleep, serving as the reference group; Class II (REM-OSA, 1.2%) included patients with high AHI during REM sleep but lowest AHI during NREM sleep, resulting in the largest AHIREM/AHINREM ratio; Class III (AHINREM < 30 events per hour, 17.4%); and Class IV (AHINREM ≥ 30 events per hour, 5.5%). Compared to Class I, factors related to Class IV included older age, high BMI, large neck circumference, hypertension, reduced total sleep time, reduced REM sleep, poor sleep efficiency, high desaturation index, low SpO2, high arousal index, and high Epworth Sleepiness Scale. As hypothesized, the study characterized several subtypes of SDB based on the AHIREM, AHINREM, and their ratio (AHIREM/AHINREM) in a large cohort and identified their demographic and clinical features. These subtypes might be clinically useful for defining SDB among adult patients referred to sleep clinics who may have varying responses to treatment depending on their subtype of the disease.