Some observations on the variation in height coefficients in prediction equations for forced vital capacity

Intercept and height coefficients in prediction equations for forced vital capacity (FVC) are highly correlated and highly variable, due in part to the small coefficient of variation of height (4%). This degree of imprecision and relationship was observed among the coefficients of a set of 26 unrelated published prediction equations, and appeared to dominate any systematic differences between them in methodology and technique. The findings indicate that care needs to be taken when using equations for predicting the FVC of people at the extremes of height. They also demonstrate the need for careful planning in the design of studies intended for the compilation of prediction equations, including provision for sufficient numbers of observations. Support is also given to the use of a composite formula based on the averaging of many individual prediction equations. The results are applicable to other similar and commonly used lung function equations (e.g. that for forced expiratory volume in 1 second, FEV1) that employ height as a predictor variable.