Diagnosis of airway obstruction: A comparison between the GLI 2012 reference equations, the NHANES III reference equations and the fixed cut-off for FEV ₁ /FVC

Objectives To assess the agreement between three recommended definitions of airflow obstruction, Methodology A large existing dataset of spirometric examinations from previous studies was used. Considering the lower limits of normal (LLN) according to the Global Lungs Initiative (GLI) 2012 equations as a gold standard for the diagnosis of obstruction, sensitivity, specificity, positive and negative predictive values were calculated for the LLN according to NHANES III equations, and the fixed cut-off FEV 1 /FVC < 70%. In addition, discordant cases, i.e. false positives and false negatives, were analyzed. Results Spirometric data from 11,676 Caucasian subjects aged ≥ 20 yrs were studied (60% men). For the fixed cut-off point relative to the GLI 2012 LLN, sensitivity, specificity, PPV and NPV were 96%, 86%, 81% and 97%, respectively. There was a substantial rate of false positive diagnoses, increasing with age. For the NHANES III LLN compared to the GLI 2012 LLN, sensitivity, specificity, PPV and NPV were 99%, 98%, 97% and 99%, respectively. There were very few discordant cases. Analysis of the discordant cases using the fixed cut-off point showed a mean difference of FEV 1 /FVC with the GLI 2012 LLN of 4,2% in the 1030 false-positive cases. In the 192 false-negative cases, the mean difference was – 1,9%. When using the NHANES III LLN, discordant cases were very close to their LLN. Conclusions This study showed very good agreement between the GLI 2012 and NHANES III reference equations for the diagnosis of airway obstruction. Using a fixed cutoff for FEV 1 /FVC resulted in a substantial overdiagnosis of obstruction, increasing with age.