Determinants of isolated reduction of diffusing capacity for carbon monoxide

Diffusing capacity for carbon monoxide (D_LCO) is a measure of gas exchange across the alveolar-capillary membrane. It is known to be affected in lung disease, for example COPD; but also by non-respiratory conditions, such as heart failure. Our aim was to study determinants of reduced DLCO in a population-based setting. Swedish CArdioPulmonary bioImage Study (SCAPIS) is a multicenter population-based study of adults aged 50-64 years. We have defined normal values for D_LCO in healthy never-smokers and used 5th percentile as lower limit of normal (LLN). We studied, by means of multivariable analysis, factors associated with D_LCO<LLN in 24,261 subjects without signs of airflow obstruction. Emphysema and interstitial lung abnormalities were assessed by computed tomography, other factors by questionnaires. Having a D_LCO<LLN (n=1472) was independently related to emphysema (odds ratio and 95% confidence interval: 2.6 (2.2, 3.9)), other lung disease than asthma or COPD (1.8 (1.2, 2.8)), chest pain walking uphill (1.6 (1.3, 2.1)), smoking history (per 10 pack-years) (1.6 (1.4, 1.9)), interstitial lung abnormalities (1.4 (1.2, 1.7)), breathing problems affecting daily life (1.4 (1.2, 1.7), male sex (1.3 (1.1, 1.6)) and cough (1.2 (1.1, 1.5)). D_LCO<LLN was less likely if subjects were older, had asthma, higher BMI, higher hemoglobin levels, higher forced vital capacity and higher physical activity levels. In conclusion, we present determinants of having isolated reduction of D_LCO in a large population-based sample. Strongest association was found with emphysema, but also with other lung diseases than asthma and COPD. The association with chest pain walking uphill might signal for need of cardiovascular risk assessment.