Associations between body mass index and lung function using Z-scores: a nonlinear relationship and machine learning...

Area of Science:

• Pulmonary Medicine
• Epidemiology
• Biostatistics

Background:

• Body Mass Index (BMI) is a common metric for assessing body weight.
• Lung function impairment encompasses obstructive, restrictive, and mixed ventilatory defects.
• Understanding the relationship between BMI and lung function is crucial for public health and clinical management.

Purpose of the Study:

• To systematically investigate the association between BMI and different lung function phenotypes.
• To explore the differential impact of BMI on obstructive versus restrictive ventilatory defects.
• To apply Z-scores and machine learning for a novel approach to lung function management.

Main Methods:

• Utilized data from the National Health and Nutrition Examination Survey (NHANES, 2007-2012).
• Calculated lung function Z-scores and analyzed associations using regression models and restricted cubic splines.
• Employed machine learning algorithms and SMOTE for feature importance and class imbalance.

Main Results:

• Higher BMI was inversely associated with obstructive ventilatory defect risk (OR=0.532) and positively associated with restrictive ventilatory defect risk (OR=2.900).
• A U-shaped relationship was observed between BMI and restrictive ventilatory defect risk, with a threshold at 26.39 kg/m².
• BMI-related features were identified as the most important predictors in machine learning models, accounting for over 32% of total feature importance.

Conclusions:

• BMI exhibits differential and opposing associations with obstructive and restrictive lung function impairments.
• Higher BMI is protective against obstructive defects but increases the risk of restrictive defects.
• Findings underscore the complex interplay between BMI and lung health, validated by advanced statistical methods.