Exercise challenge test is often used to diagnose asthma. Decrease in forced expiratory volume in one second (FEV1) of 10% or 15% is considered as the gold standard to diagnose asthma, but a decrease of 15% in peak expiratory flow (PEF) is also recommended as diagnostic. Our aim was to assess the accuracy of different PEF cut-off points in comparison to FEV1.

We retrospectively studied 326 free running exercise challenge tests with spirometry in children 6 to 16 years old. FEV1 and PEF were measured before and 2, 5, 10 and 15 minutes after exercise. ROC analysis, sensitivity, specificity, positive and negative predictive values and kappa-coefficient were used to analyze how decrease in PEF predicts decrease of 10% or 15% in FEV1.

In the ROC analysis area under the curve was 0.921 (p<0.001) for PEF decrease to predict a 15% decrease in FEV1 and 0.851 (p<0.001) to predict a 10% decrease in FEV1. The agreement between changes in PEF and FEV1 varied from slight to substantial (kappa values of 0.199 ? 0.680) depending on the cut-points. Lower cut-off for decrease in PEF had higher sensitivity and negative predictive value, while higher cut-off values had better specificity and positive predictive value. Decrease of 20% in PEF seemed to be the best cut-off for detecting a 10% decrease in FEV1, and 25% seemed to be the best to detect a 15% decrease in FEV1. Still, about a fifth of the positive findings were false positives.

Change in PEF is not a precise predictor of change in FEV1 in exercise test. The currently recommended cut-point of 15% decrease in PEF seems to be too low and leads to high false positive rate.