Abstract
Aims: Obstructive sleep aponea (OSA) is associated with metabolic changes. OSA can effectively be treated with continuous positive airway pressure (CPAP) which is presumed to reverse its pathophysiological consequences. Exhaled breath contains biochemical information on the pathophysiological state. We aimed to answer the questions whether a disease specific breath profile can be detected by real-time breath analysis using mass spectrometry in OSA.
Methods: 28 patients with OSA effectively treated with CPAP were randomly assigned to either subtherapeutic or to continue therapeutic CPAP for two weeks. The composition of exhaled breath (primary outcome) was analysed by real-time mass spectrometry and further processed by 1-way ANOVA (Tukey´s multiple comparison test) and random forest algorithm for disease prediction. The discriminating exhaled compounds were identified by breath condensate.
Results: Compared to therapeutic CPAP, CPAP withdrawal led to a recurrence of OSA (oxygen-desaturation index +30.3/h; 95%CI:19.8/h,40.7/h, p<0.001) accompanied by significant changes in 10 exhaled compounds (example in ). Based on breathprints, untreated OSA could be predicted in a leave-one-out-cross-validation with a sensitivity of 92.9 % and specificity of 84.6%.
Conclusions: Analysis of exhaled breath by mass spectrometry allows rapid non-invasive identification of untreated OSA with high accuracy, and identification of target molecules.
- Copyright ©ERS 2015
