(178) Towards targeted exhaled breath analysis for young children in CF care to detect bacteria in the lungs

Background: Children with Cystic Fibrosis (CF) are prone to bacterial infections, which can be associated with lung function decline. Consequently, the detection of bacteria plays an important role in CF care. However, culture dependent diagnosis of such infections is a challenge in pediatric patients due to limited sample availability and quality. As an alternative, exhaled breath analysis measuring volatile organic compounds (VOCs) might be a promising non-invasive and culture independent tool to identify bacteria. Several explorative studies have suggested that VOCs can differentiate between bacterial species, however these results are not always consistent. The next step forward for exhaled biomarkers would be to move from discovery studies towards hypothesis driven targeted analysis, an essential step for clinical application. Staphylococcus aureus (SA) is a common bacteria found in young patients with CF. We aimed to test the feasibility of targeted exhaled breath analysis in young children with CF, in a proof-of-concept study, for the detection of SA using previously identified VOCs, and whether this could discriminate between SA positive and negative cultures.

Methods: Children with CF aged 1-6 years were included. Exhaled breath samples were taken and stored in thermal desorption tubes and analysed by gas chromatography – mass spectrometry (GC-MS). Microbiological cultures were collected on the same day. After a literature review to identify SA related VOCs, a targeted peak detection via AMDIS (v.2.68) was performed on the GC-MS dataset. The resulting peak intensities were used for a multivariate modelling by Sparse Partial Least Squares Discriminant Analysis (sPLS-DA), after which the predicted outcomes (SA positive/negative) were compared with the actual microbiological outcomes. Statistical analyses were performed in R-Studio (v.1.3).

Results: Exhaled breath collection was attempted in 24 children with CF. The procedure was successful in 83% of the patients (N= 20, median age = 4 yrs. (IQR: 2-6 yrs.)). 9 patients had a positive culture for SA. 6 were taken by a nasopharyngeal swab, 2 from sputum and 1 from bronchoalveolar lavage fluid. 2 patients had a positive culture for H. Influenzae and 1 patient for Moraxella Cattharalis. 8 patients had a negative culture. We identified 5 VOCs associated with SA from the literature review (≥2 consistent notations). These VOCs were used to predict culture positivity for SA resulting in an accuracy of 70%, specificity of 89% and a sensitivity of 55%. The positive and negative predictive value of the test was 86% and 62%, respectively.

Conclusions: Targeted analysis of MS driven exhaled VOC profiles appears to enable detection of pathogens in exhaled breath. We showed that non-invasive breath collection for targeted exhaled breath analysis is feasible in young children, and that VOC’s associated with SA were detectable. One of the limitations of this study is the small sample size and that most microbiology cultures were taken by a nasopharyngeal swab, which may explain the low sensitivity of the model. To conclude, more research should be done to evaluate this promising methodology in larger cohorts. Preferably, using more sensitive microbiological sampling methods.

Acknowledgements: This work was funded by a Vertex Innovation Award (VIA) 2020.

References: