Global Lung Function Initiative 2012 reference values for spirometry in South Italian children

Rationale Despite the widespread use of the Global Lung Function Initiative (GLI) 2012 reference values, there is still the need of testing their applicability in local areas. Objectives The aims of this study are to evaluate applicability of GLI reference equations in a large population-based sample of normal schoolchildren from Sicily, and to compare GLI and previous prediction equations in terms of spirometry test interpretation. Methods GLI equations were evaluated in 1243 normal schoolchildren, 49% males, aged 7–16 years, height 116–187 cm. Normality assumptions for the GLI z-scores (FEV1, FVC, FEV1/FVC) were tested, and bootstrap confidence intervals for the mean (0 expected) and the variance (1 expected) were derived. GLI and other reference equations were compared in terms of probabilities to fall below the lower limit of normal (LLN). Results The GLI z-score normality assumption held for males but not for females (p < 0.001). According to the mean z-score, predicted values were: slightly underestimated for FEV1 (0.15 in males, 0.07 in females); overestimated for FVC (−0.27 in males, −0.32 in females); highly underestimated for FEV1/FVC (0.75 in males, 0.81 in females). Variability was correctly estimated. The probability of FEV1<LLN correctly approached 0.05 when using GLI, Hankinson and Quanjer equations. Wang equations yielded correct probabilities of abnormal FVC; Pistelli equations yielded correct probabilities of abnormal FEV1/FVC for females. Conclusions GLI 2012 references underestimate FEV1/FVC predicted values in a sample of normal South Italian children. Physicians interpreting spirometry should be aware to test reference values prior to their use in a local area.