Development of a Multivariable Prediction Model to Estimate Probability of Sleep-Disordered-Breathing...

Area of Science:

• Cardiology
• Sleep Medicine
• Medical Diagnostics

Background:

• Sleep-disordered breathing (SDB) is prevalent in atrial fibrillation (AF) patients, worsening outcomes.
• Current methods for identifying AF patients needing SDB testing are insufficient.

Purpose of the Study:

• To develop and validate a predictive tool for identifying AF patients with moderate-to-severe SDB.
• To create a practical scoring system for clinical use.

Main Methods:

• A prediction model was developed using data from 442 AF patients undergoing polysomnography.
• External validation was performed on a separate cohort of 409 AF patients.
• A simplified score (MOODS) was derived from significant predictors: age, sex, BMI, diabetes, and prior stroke/TIA.

Main Results:

• Significant SDB was found in 34% (derivation) and 54% (validation) of AF patients.
• The multivariable model showed good discrimination (C-statistic: 0.75).
• The MOODS score demonstrated good discrimination (C-statistic: 0.73) and high sensitivity/specificity for SDB detection.

Conclusions:

• The MOODS score accurately estimates the probability of significant SDB in AF patients.
• MOODS can assist in clinical decision-making and optimize resource allocation for SDB testing.