Exhaled breath analysis is a potential non-invasive tool for diagnosing and monitoring airway diseases. Gas chromatography-mass spectrometry and electrochemical sensor arrays are the main techniques to detect volatile organic compounds (VOC) in exhaled breath. We developed a broadband quantum cascade laser spectroscopy technique for VOC detection and identification. The objective of this study was to assess the repeatability of exhaled breath profiling with broadband quantum cascade laser-based spectroscopy and to explore the clinical applicability by comparing exhaled breath samples from healthy children with those from children with asthma or cystic fibrosis (CF). Healthy children and children with stable asthma or stable CF, aged 6-18 years, were included. Two to four exhaled breath samples were collected in Tedlar bags and analyzed by quantum cascade laser spectroscopy to detect VOCs with an absorption profile in the wavenumber region between 832 and 1262.55 cm^-1. We included 35 healthy children, 39 children with asthma and 15 with CF. Exhaled breath VOC profiles showed poor repeatability (Spearman's rho = 0.36 to 0.46) and agreement of the complete profiles. However, we were able to discriminate healthy children from children with stable asthma or stable CF and identified VOCs that were responsible for this discrimination. Broadband quantum cascade laser-based spectroscopy detected differences in VOC profiles in exhaled breath samples between healthy children and children with asthma or CF. The combination of a relatively easy and fast method and the possibility of molecule identification makes broadband quantum cascade laser-based spectroscopy attractive to investigate the diagnostic and prognostic potential of volatiles in exhaled breath.