Predicted versus absolute values in the application of exhaled nitric oxide measurements

Abstract

Constitutional factors such as age, sex and height, and acquired factors such as atopy and smoking, influence exhaled nitric oxide (F(E)NO) levels. The utility of predicted values based on reference equations which account for these factors has not been evaluated. To compare the performance characteristics of absolute versus % predicted values for F(E)NO as predictors of diagnosed asthma and steroid response. We compared the sensitivities, specificities and likelihood ratios using F(E)NO (% predicted) with absolute values for F(E)NO (ppb) in 52 steroid-naive subjects with non-specific respiratory symptoms. The reference equations of Olin et al. (Chest, 2007) and Dressel et al. (Resp. Med., 2008) were used to derive predicted values. Receiver operating curve analyses were performed and the areas under the curve (AUC) were calculated for two outcomes: diagnosed asthma (yes/no), and steroid response after fluticasone for 4 weeks (defined as ≥ 12% increase in FEV(1); increase in mean morning PEF ≥ 15%; reduction in symptoms ≥ 1 point; increase in PC(20)AMP of ≥ 2 doubling doses). The AUCs for diagnosed asthma were: F(E)NO (absolute) 0.770; F(E)NO (% pred.): 0.758 (Olin) and 0.775 (Dressel) (NS). The AUCs for F(E)NO (abs.) and F(E)NO (% pred.) with respect to the four indices of steroid response were likewise not significantly different. Correcting F(E)NO for combinations of age, sex, height, smoking and atopy using reference equations did not enhance the performance characteristics of F(E)NO as a predictor of either the diagnosis of asthma or steroid responsiveness in patients with chronic airways-related symptoms.