Purpose
In aging populations, the Global Initiative for Obstructive Lung Disease (GOLD) spirometry threshold may misclassify normal spirometry as airflow limitation. The Global Lung Initiative (GLI) method provides age-adjusted criteria. We investigated how the use of GOLD or GLI thresholds in an algorithm affects the classification of elderly smokers into COPD risk phenotypes.
Methods
Using a modified COPDGene algorithm, including exposure, symptoms, and abnormal spirometry, 200 smokers aged 60 years and older were classified into 4 mutually exclusive phenotypes: Phenotype A (no symptoms, normal spirometry; reference), Phenotype B (symptoms, normal spirometry; possible COPD), Phenotype C (no symptoms, abnormal spirometry; possible COPD), and Phenotype D (symptoms, abnormal spirometry; probable COPD). Abnormal spirometry was defined according to the GOLD or GLI criteria. A comparison was made between the GOLD- and GLI-defined phenotypes.
Results
Using GLI criteria/cut-offs, 18.5% (n = 37) had phenotype A (no COPD), 42% (n = 84) had phenotype B (possible COPD), 7.5% (n = 15) had phenotype C (possible COPD), and 32% (n = 64) had phenotype D (probable COPD). Using GOLD criteria cut-offs, 14.5% (n-29) had phenotype A (no COPD); 31% (n = 62) had phenotype B, 11.5% (n = 23) had phenotype C (probable COPD), and 43% (n = 86) had phenotype D (probable COPD). Eight smokers with GOLD phenotype C were reclassified as GLI phenotype A, while 22 with GOLD phenotype D were reclassified as GLI phenotype B. Smokers identified as ‟probable COPD” by GOLD alone (potential false positives) had better spirometry results than those identified as ‟probable COPD” by both GOLD and GLI.
Conclusion
The use of the GOLD threshold in an algorithm resulted in older smokers being classified into more severe COPD risk phenotypes compared to the GLI threshold. This suggests that GOLD may misclassify smokers with less affected phenotypes as having respiratory impairment, potentially leading to unnecessary and harmful treatments.
