Comparison of FEV reference equations for evaluating a cystic fibrosis therapeutic intervention.

Objectives: The Global Lung Function Initiative (GLI, 2012) developed reference equations for forced expiratory volume in 1 s (FEV ). Previous equations were developed by groups led by Knudson (1983), Wang (1993), Hankinson (1999), and Stanojevic (2008). We assessed how different prediction equations affect the conclusions from a therapeutic intervention study that evaluated the rate of percent predicted FEV (ppFEV ) decline.

Methodology: Using data from the Epidemiologic Study of cystic fibrosis (CF), we re-analyzed our previous study evaluating the relationship of dornase alfa (DA) use with ppFEV using the Knudson, Wang & Hankinson, Stanojevic, and GLI equations. The change in intercept and change in slope of ppFEV from a 2-year pre-index period and 2-year post-index period were compared between the treated (N = 2483) and comparator groups (N = 6992, from 4110 unique patients).

Results: Change in intercept for the comparator group was similar across equations except that Wang & Hankinson values were more negative. The difference in change in intercept between the DA and comparator groups ranged from 3.38 to 4.02% predicted. The change in slope for the comparator group ranged from -0.58 to +0.30 ppFEV /year, but the difference in change in slope between the DA and comparator groups was in a narrower range from +0.53 to +0.89 ppFEV /year.

Conclusions: Although individual patient results are impacted by the choice of reference equations, the study conclusions from this evaluation of a therapeutic intervention were minimally affected. GLI equations are recommended for future studies, but prior results based on other equations should be accepted as reliable.