Physical activity is an important part of life, and hence exercise-induced bronchoconstriction (EIB) can reduce the quality of life. A standardized test is needed to diagnose EIB. The American Thoracic Society (ATS) guidelines recommend an exercise challenge in combination with dry air. We investigated the feasibility of a new, ATS guidelines conform exercise challenge in a cold chamber (ECC) to detect EIB. The aim of this study was to investigate the surrogate marker reaction to methacholine, ECC and exercise challenge in ambient temperature for the prediction of a positive reaction and to re-evaluate the reproducibility of the response to an ECC.

Seventy-eight subjects aged 6 to 40 years with suspected EIB were recruited for the study. The subjects performed one methacholine challenge, two ECCs, and one exercise challenge at an ambient temperature. To define the sensitivity and specificity of the predictor, a receiver-operating characteristic curve was plotted. The repeatability was evaluated using the method described by Bland and Altman (95% Limits of agreement).

The following cut-off values showed the best combination of sensitivity and specificity: the provocation dose causing a 20% decrease in the forced expiratory volume in 1 s (PD₂₀FEV₁) of methacholine: 1.36 m

g (AUC 0.69, p < 0.05), the maximal decrease in FEV₁ during the ECC: 8.5% (AUC 0.78, p < 0.001) and exercise challenges at ambient temperatures: FEV₁ 5.2% (AUC 0.64, p = 0.13). The median decline in FEV₁ was 14.5% (0.0–64.2) during the first ECC and 10.7% (0.0–52.5) during the second ECC. In the comparison of both ECCs, the Spearman rank correlation of the FEV₁ decrease was r = 0.58 (p < 0.001). The 95% limits of agreement (95% LOAs) for the FEV₁ decrease were − 17.7 to 26.4%.

The surrogate markers PD₂₀FEV₁ of methacholine and maxima

l decrease in FEV₁ during ECC can predict a positive reaction in another ECC, whereas the maximal FEV₁ decrease in an exercise challenge at an ambient temperature was not predictive. Compared with previous studies, we can achieve a similar reproducibility with an ECC.