Abstract
BACKGROUND:
The BOLD method (questionnaires and standardized post-bronchodilator spirometry) estimates and compares the prevalence of COPD in different countries. Its diagnostic accuracy has not been fully evaluated.
AIM:
To assess the diagnostic accuracy of the BOLD methodology.
METHODS:
We performed a BOLD assessment, additional lung function and clinical assessment in subjects 5 years after their COPD diagnosis in the Cape Town BOLD study in 2005.
RESULTS:
The response rate was 56% (106 of 196 eligible subjects). Sixteen subjects (15%) had a post-bronchodilator FEV1:FVC≥0.7 (i.e. non-obstructed), a further 11 subjects (10%) were non-obstructed using lower limit of normal (7 of whom were >70yr). Using 2 sets of clinical criteria of different stringency to define asthma, 10 or 18% of the cohort were considered to have asthma: 6 or13% without current airflow limitation (AFL); 11 or19% with current AFL (n=90); and 10 or26% with both AFL and reversibility (n=31). In a same-day comparison between the BOLD handheld (ndd EasyOne) and office spirometry (nSpire), the FEV1 differed by >150ml in 23%, resulting in reclassification of COPD status in 17% of subjects [Bland-Altman: mean diff=41ml;sd 146ml; 95%CI: -250:+333ml]. Between visit variability of the ndd EasyOne (performed<1 month apart in a subgroup) resulted in reclassification of 11 of 56 subjects (20%), with 16 (29%) having FEV1 values that differed by >150ml between visits. [mean diff=18ml;sd 220ml; 95%CI: -0.422:+0.458ml].
CONCLUSIONS:
The BOLD methods overestimate the burden of COPD owing to use of the fixed FEV1/FVC definition, lack of discrimination from asthma, and test variability of the handheld spirometer.
- © 2014 ERS