Comparison of race-specific and race-neutral GLI spirometric reference equations with an Indian reference equation

Abstract

Background: Despite the increasing popularity and use of Global Lung Function Initiative (GLI) spirometric reference equations, the appropriateness of the race-specific and race-neutral GLI spirometric reference models among the Indian population has not been systematically investigated.

Methods: In this cross-sectional analysis, we used spirometric measurements of 1123 healthy Indian adults (≥18 years of age). We computed reference values and z-scores for forced vital capacity (FVC), forced expiratory volume in 1 s (FEV₁), and FEV₁/FVC from race-specific and race-neutral GLI reference equations as well as from a widely used Indian reference equation. We studied heterogeneity between GLI equations and the Indian equations using Bland-Altman analysis, and the differences between the reference and observed values were compared using the Friedman test.

Results: In Bland-Altman analysis, significant heterogeneity in FVC and FEV₁ between race-specific and Indian equations was observed (bias: 10.4 % and 14.1 %, respectively), with less bias for FEV₁/FVC (3.76 %). The race-neutral equations showed almost similar bias (9.8 %, 13.8 %, and 3.8 % for FVC, FEV₁, and FEV₁/FVC, respectively). Median differences in race-specific reference values from observed values for FVC and FEV₁ were 0.49L and 0.44L, respectively, decreasing slightly with race-neutral equations (0.46L and 0.43L) whereas Indian models showed minimal differences (FVC: 0.10L, FEV₁: 0.05L). Z-scores for FVC and FEV₁ were significantly different between race-specific and race-neutral GLI equations, and both differed from Indian equations.

Conclusion: Both race-specific and race-neutral GLI reference equations are significantly different from the Indian equations, which underscores the importance of determining the suitability of global reference models before being used indiscriminately.

Keywords: FEV(1); FVC; Lung function; Reference values; Z-score.