TY - JOUR T1 - Disease severity using %predicted and the standardized residual (SR) - A conflict in classification? JF - European Respiratory Journal JO - Eur Respir J VL - 42 IS - Suppl 57 SP - P1274 AU - A.H. Kendrick Y1 - 2013/09/01 UR - http://erj.ersjournals.com/content/42/Suppl_57/P1274.abstract N2 - The SR provides a statistical analysis of normality based on physiological data, whereas the %predicted has little physiological or statistical validity. Commonly used guidelines currently use arbitrary %predicted values as a guide to disease severity. Aim: To assesses the limitations of using %predicted compared to SR. Method: 3474 consecutive datasets (M:1692) aged 20-97yrs (mean 52.9±20.2) for FEV1 were analysed. SR and %predicted values were calculated using the ECCS regression equations, and subdivided based on %predicted. Results: The relationship of %predicted to SR was %pred = 97.6+13.9SR±0.56, r2=0.91.The relationship did not appear to be influenced by sex or age of the subjects. The distribution of SR values according to %predicted is shown in table 1, and indicates that up to 28% of patients would be misclassified in terms of disease severity.View this table:Table 1. Distribution of SR's based on %predicted rangesRecalculating the approximate %predicted values that equate to the standard SR values on -1.64, -2.5 and -3.5, would give %predicted of 75%, 63% and 49%. Conclusion: The conflict between SR and %predicted remains with current guidelines misclassifying disease severity by %predicted. This may be due to choice of reference equations, which needs to be further investigated. ER -