Agreement between spirometric quality reviewers in an epidemiological study.

OBJECTIVE

. Spirometry is the gold standard among pulmonary function tests, with acceptability criteria established to ensure reliable results. However, variability in the interpretation of spirometry by trained reviewers remains unexplored. This study assessed agreement among trained reviewers regarding spirometric acceptability and its components, including the number of acceptable maneuvers, predominant defect, and repeatability.

METHODS

. Spirometries from 3,982 adults (≥40 years) participating in an epidemiological study across six cities in Argentina were independently reviewed by junior reviewers (JRs) and adjudicated by senior reviewers (SRs), following the 2005 ATS/ERS guidelines. Agreement was analyzed using Cohen's kappa, McNemar's test, and proportion comparisons.

RESULTS

. A total of 7,964 pre- and post-bronchodilator spirometries were analyzed. Agreement on overall acceptability between JRs was 76% and 84% in two independent reviewer teams (RvTa and RvTb, respectively). Agreement on specific defects was highest for numerical criteria, such as insufficient expiratory time (k=0.61) and excessive back-extrapolated volume (BEV; k=0.52), but lower for subjective criteria, including weak effort (k=0.19) and cough in the first second (k=0.36). No agreement was observed for glottic closure (k=0). SR adjudication increased the proportion of acceptable tests from 54% to 65%.

CONCLUSIONS

Reviewer variability significantly affects spirometric acceptability, with numerical criteria achieving higher agreement than subjective assessments. These discrepancies have implications for large-scale studies, data quality, and automated algorithm development. Cross-checking reviews is recommended to enhance reliability. Further standardization of acceptability criteria is warranted.