Spirometric reference equations for First Nations children and adolescents living in rural Saskatchewan
Background: The spirometric reference values are of great importance for diagnosis and treatment of lung diseases. At present, there are no spirometric reference values for First Nations children and adolescents living in Canada. Objectives: The objectives of the present study were (1) to identify the flexible and efficient statistical method to derive lung function reference equations that can be used to obtain the predicted values and Lower Limit of Normal (LLN) for lung function in children and adolescents, and (2) to obtain prediction equations for FVC, FEV1 and FEV1=FVC for First Nations children and adolescents living in rural Saskatchewan, Canada. Methods: Spirometric results from a prospective cohort study, "First Nations Lung Health Project" were used to identify 130 healthy non-smoking children and adolescents. The predicted values and LLN of spirometric indices [Forced Vital Capacity (FVC), Forced Expiratory Volume at one second (FEV1) and FEV1 and FVC ratio (FEV1=FVC)] were calculated for school-going children and adolescents ages 6-17 years. The subjects participating in the study were from two Cree First Nations on-reserve communities located in rural Saskatchewan, Canada. All lung function values were reviewed by a respirologist for acceptability of the test. Following an extensive literature review, the Generalized Additive Models for Location, Scale and Shape (GAMLSS) was identified as a flexible statistical tool to model the lung function variables. The lung function indices were assumed to follow a Box-Cox-Cole-Green (BCCG) distribution with median, , coe ffcient of variation and skewness . Akaike Information Criteria (AIC) approach was used to obtain the reference models. The LLN was calculated by taking the fifth percentile of the prediction equations of the lung function variables. The above approach is recommended for the prediction of lung function of multi-ethnic people aged 3-95 years from different ethnic groups by the Global Lung Function Initiative (GLI). Results: Significant differences were observed in lung function (FVC, FEV1 and FEV1=FVC) and anthropometric measurements between both boys and girls. Therefore, fitting separate equations for both sexes are justified. In GLI, polynomial bases of order 6-7 were used for modeling the meadian, coefficient of variation and skewness . In this study, lower order polynomial bases (up to order 4) were enough to obtain the reference models. In GLI, the polynomial bases were divided by 100 to let it lie within 0 to 1. In this study, the polynomials were divided by 20 to lie these between 0 and 1. The predicted values of FVC was higher than the values for FEV1 in both boys and girls. Therefore the values of FEV1=FVC ratios is less than 100% in this population. In girls, the difference between the curves of FVC and FEV1 was smaller compared to boys. Thus, the total volume of air for girls during exhalation are close to the volume of air exhaled at the first second. The estimated curves showed that the models fitted the lung function data reasonably well. Conclusions: The results in this study showed that the optimum model for the prediction of lung function were almost similar to the ones used by GLI for the prediction of lung function of all-age multi-ethnic populations.The predicted values and LLN values of the lung function variables reported in this study can be recommended to health-care providers for the use in diagnosis respiratory diseases in First Nations children and adolescents in rural Saskatchewan. Small sample (n < 150) was a limitation of this study. This study limitation can be overcome by including more individuals from the follow-up study, which will be conducted in 2016.
DegreeMaster of Science (M.Sc.)
DepartmentSchool of Public Health
SupervisorKhan, Shahedul; Rennie, Donna C.
CommitteePahwa, Punam; Szafron, Michael
Copyright DateFebruary 2016
Reference Equations, Predicted Equations, Children and Adolescents, Cree First Nations, Lower Limit of Normal