A New Derivation of the Reference Equations of Spirometric Values in the Thai Population Using the GAMLSS Method

Objective: The Global Lung Initiative (GLI)-2012 equation has now been used as the reference for the interpretation of the spirometry worldwide. The equations have been derived using modern statistical methods called Generalized Additive Models for Location, Scale and Shape (GAMLSS) and Spline. A part of the spirometric data from which the equations were derived is the data collected from the Thai population. The present study aimed to create the reference equations using only the data from the Thai population retrieved from the previous cross-sectional study and compared them to the GLI-2012 fs. Materials and Methods: The previous data of each spirometric parameters including FVC, FEV., FEV./FVC, and FEF25-75% in 1,616 male and 2,280 female subjects were retrieved for the statistical modeling by GAMLSS and Spline packages in R Statistics. Lambda-Mu-Sigma (LMS) models were used to establish the best fitted models defined by the smallest Schwarz fs Bayesian Criterion (SBC). The authors also used a set of spirometric data of healthy elderly subjects to validate the equations and compare with the GLI-2012. Results: The GAMLSS-Thai equations were then derived to establish the predicted (Mu) and the lower limit of normal (LLN) values for each parameter. For 140 male and 651 female subjects aged 60 to 80 years, the predicted performance of GAMLSS-Thai is better than GLI-2012 equations as shown by various statistical tools including z-score, mean absolute percentage error (MAPE), and root mean square deviation (RMSD) in all spirometric parameters. Conclusion: The GAMLSS-Thai equation can be alternatively used as reference spirometric values in subjects aged 10 to 80 years in Thai population. It is more appropriate for subjects aged 60 to 80 years than the GLI-2012 equation.