Validation of the Global Lung Function Initiative Spirometry Reference Equations in Tanzania

Abstract Background Spirometry is the most widely used pulmonary function test. Lung function varies with age, sex, height, and ethnicity, therefore accurate interpretation requires comparison with appropriate reference values from healthy individuals with similar characteristics. The Global Lung Function Initiative introduced multi-ethnic race-specific equations in 2012 (GLI 2012 ) and race-neutral equations in 2022 (GLI 2022 ). However, their validity in the Tanzanian population remains largely untested, and no reference values have been recommended for this group. This study aimed to evaluate the performance of GLI 2012 and GLI 2022 equations in healthy, non-smoking Tanzanian adults. Methods This was a population-based cross-sectional study, where participants were randomly recruited from four Tanzanian regions. The participants were interviewed using a questionnaire and were examined by spirometry according to the ATS/ERS recommendations. Spirometry Z-scores were generated using the GLI calculator for ‘Black’, ‘Other’ and GLI 2022 ‘race-neutral’ equations. A mean Z-score outside the range of ± 0.5 was considered clinically significant. Results Out of 968 participants recruited, 325 (34%) were excluded due to acute respiratory symptoms, chronic cardio-respiratory diseases or past/current smoking. Of 643 participants who performed spirometry, 57 (9%) were excluded due to technically poor spirograms. The remaining 586 participants were 23–73 years. There were 51% females in the study, and they were older than males (median: 43 vs. 33 years; p < 0.001) with a slightly higher BMI (median: 28 vs. 27; p < 0.001). While the GLI 2012 ‘Other’ and GLI 2022 race-neutral equations poorly predicted FEV 1 and FVC, the ‘Black’ equations demonstrated borderline performance: ‘ Other’ : mean Z-score ± SD of -1.17 ± 1.18 for FVC and − 1.14 ± 1.16 for FEV 1 ; ‘ Black’ : -0.51 ± 1.11 for FVC and − 0.52 ± 1.17 for FEV 1 ; ‘ Race-neutral ’: -0.97 ± 0.98 for FVC and − 1.01 ± 1.02 for FEV 1 . However, all three equations accurately predicted FEV 1 /FVC (mean Z-score within ± 0.5). Conclusions All three equations reliably predict the FEV₁/FVC ratio and can be used to diagnose airway obstruction. The GLI 2012 ‘Other’ and GLI 2022 race-neutral equations do not accurately predict FEV₁ and FVC in the Tanzanian population and are therefore not recommended. The borderline performance observed with the GLI 2012 Black equations suggests that they may be used with caution in Tanzania until a specific correction factor or locally-derived reference equations become available.