A physiologically based model for spirometric reference equations in adults

A spirometric reference equation consists of a mathematical model with constants and coefficients optimized to fit a specific data set from healthy individuals. Commonly applied models are selected on statistical rather than physiological considerations. A predetermined model with constants and coefficients optimized to various populations would enable interpretable and interesting comparisons between populations. Lubinski and Golczewski recently presented a piecewise linear model with constants and coefficients claimed to be physiologically interpretable (Lubinski model). Three questions were addressed: Is the Lubinski model as useful clinically as other models: multiple linear, piecewise polynomial and exponential with splines? Will reference equations based on the Lubinski model and optimized to a Swedish and to a Polish population allow for interpretable comparisons? Are three well-known reference equations clinically useful in the Swedish adult population? A recent Swedish random population sample with high-quality spirometric measurements enabled the present analyses. When optimized to fit the Swedish population sample, the Lubinski model and two other models provided accurate predictive normal values. Interesting differences were demonstrated between the Polish and Swedish populations. The proportion of subjects below lower limit normal was adequate for the piecewise polynomial equations but too low and not clinically useful for the advocated exponential equations with splines. It is concluded that the Lubinski model is clinically as useful as other models, and it adds important value and is recommended for future spirometric reference equations for adults. The advocated exponential equations with splines are not recommended for Swedish adults because of too wide normal limits.