Predictability of airflow obstruction using normal spirometry measurements

Airflow Obstruction (AO) is of importance as it is a critical element of chronic respiratory diseases such as chronic obstructive pulmonary disease (COPD) and asthma. Early identification of AO helps to manage and to improve long-term outcomes. The aim is to assess the ability of spirometric markers to predict (the risk of) AO development in the general population. To this end, 5141 participants with at least two consecutive valid measurements were selected from the Austrian LEAD (Lung, hEart, sociAl, boDy) cohort. Participants had normal lung function, with no AO, and no prior diagnosis of asthma or COPD at baseline. Lung function (LF) predictors were used to develop statistical models predicting the incidence of AO at follow-up, approximately four years in advance. Beyond defining cut-offs to identify participants at risk , the analyses showed that a combined LF model provided higher discrimination than single-parameter models based on FEV₁ and FEF 25–75% , while its discrimination was not significantly different from FEV₁/FVC. Incorporating total lung capacity (TLC) as a lung volume (LV) parameter into the LF model resulted in a small but statistically significant increase in discrimination (AUC = 0.901 vs. 0.891; ΔAUC = 0.010, p = 0.017). However, the added clinical value of lung volume parameter requires further evaluation. The positive predictive values (PPVs) of the models ranged from 0.08 to 0.16, indicating an increased probability of developing AO among participants identified as being at risk compared with the overall study population (0.0393); however, PPVs remained relatively low, consistent with the low incidence of AO in the study population.