Anomalous Differences Between the Global Lung Function Initiative 2023 and 2012 Spirometry Reference Values

Area of Science:

• Pulmonary Medicine
• Biostatistics
• Population Health

Background:

• The Global Lung Function Initiative (GLI) established spirometry reference values in 2012 based on international data from healthy individuals, stratified by sex, age, and height across four ancestral groups.
• In 2023, a unified set of predicted spirometry values was developed to be ancestry-independent for the entire GLI database.

Purpose of the Study:

• To investigate the impact of varying proportions of ancestral groups within age cohorts on predicted spirometry values derived from the 2012 and 2023 GLI equations.
• To assess the consequences of using inverse probability weighting in the 2023 GLI equations across the full age range (3.5 to 95 years).

Main Methods:

• Utilized the Global Lung Function Initiative (GLI) spirometry database, noting the overrepresentation of European ancestry (77.4%).
• Applied inverse probability weighting in the 2023 GLI equations to approximate equal contribution from the four ancestral groups.
• Analyzed the effect of this weighting strategy on predicted spirometry values across different age cohorts.

Main Results:

• Participant proportions from each ancestry group varied significantly across age cohorts.
• The European ancestry group, with the highest predicted lung volumes, had a disproportionately high contribution in younger children and older adults.
• GLI 2023 predicted lung volumes were higher in older adults and lower in children/adolescents compared to GLI 2012 predictions for individual ancestral groups.

Conclusions:

• The inverse weighting strategy in the 2023 GLI equations introduced systematic bias, skewing predicted spirometry values.
• This bias leads to progressively overestimated predicted lung volumes with increasing age in adults.
• Conversely, it results in progressively underestimated predicted lung volumes with increasing age in children and adolescents.