To investigate differences in the classification of subjects, into Chronic Obstructive Pulmonary Disease (COPD) patients or not, using global commonly used equations - European Coal and Steal Community (ECSC)equations - and equations derived from local healthy subjects.
796 patients underwent spirometry during the last 4 years in our laboratory in Athens-Greece. 177 of them, had symptoms and signs denoting COPD. 238 of them, (120 women and 118 men, age 14–88 years old, characterized as “healthy”) were never-smokers, with no reported diagnosis of asthma or respiratory symptoms. Spirometry tests with missing data and tests from non-Caucasian subjects were excluded (55 tests). Using linear regression analysis, equations predicting forced expiratory volume in one second (FEV1) & forced vital capacity (FVC) in the healthy subjects, with age and height as independent variables, were estimated. Applying the global initiative for chronic obstructive lung disease (GOLD) spirometric criteria and the two sets of equations (ECSC & the ones estimated) subjects were diagnosed, or not, with COPD.Since our sample derives from a hospital population (inpatients and screening control of outpatients), the prediction equations may not be representative of Athens’ population.
None of the healthy subjects had had abnormal spirometry results. From the 177 patients with a clinical suspicion of COPD, 122 fulfilled the GOLD criteria, with both prediction equations. Applying GOLD criteria in the total of 741 patients (as a screening method), 149 were classified as patients with COPD. 122 of them were correctly classified. 27 had either asthma or COPD with overlapping bronchiectasis.Using the estimated local equations, 1 extra patient (without clinical suspicion of COPD) was classified as COPD.
The classification of patients with COPD, using GOLD criteria, did not significantly alter, when local equations were applied instead of the ECSC equations, despite the differences in the predicted normal values.
Diagnosis of COPD, using GOLD criteria, may not be sensitive to differences in predicted values and thus application of locally derived equations for that purpose, may not be necessary.
Nikolaos Tatsis, No Financial Disclosure Information; No Product/Research Disclosure Information