Introduction: While the 2012 Global Lung Initiative spirometry reference equations (GLI₂₀₁₂) are widely recommended, studies investigating their performance in Africa are scarce and have reported conflicting results.

Aim: This study evaluated the fit of GLI₂₀₁₂ spirometry prediction equations in Tanzania.

Methods: In this population-based cross-sectional study, participants were randomly recruited from urban and rural areas in 4 Tanzanian regions. Each participant completed an interviewer-administered questionnaire and performed spirometry according to the ATS/ERS recommendations. Spirometry z-scores were generated using the GLI₂₀₁₂ calculator for ?Black? and ?Other? equations. A mean z-score of more than ±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 poor spirograms. The remaining 586 participants aged 23-73 years had high-quality spirograms. In this study, 51% were females and were on average older than males (median: 43 vs 33 years; p<0.001) with a slightly higher BMI (median: 28 vs 27; p<0.001). Overall, the GLI₂₀₁₂ ?Other? equations performed poorly compared to the ?Black? equations, and both accurately predicted FEV₁/FVC but poorly predicted FVC and FEV₁: ?Other?: mean z-score ± SD of -1.17±1.18 for FVC, -1.14±1.16 for FEV₁, and -0.11±1.20 for FEV₁/FVC; ?Black?: -0.51±1.11 for FVC, -0.52±1.17 for FEV₁, and -0.05±1.14 for FEV₁/FVC.

Conclusion: The GLI₂₀₁₂ spirometry prediction equations do not seem valid for the Tanzanian population.