Real-time electronic nose based pathogen detection for respiratory intensive care patients

Abstract An acoustic wave based electronic nose was used to monitor the exhaled breath of patients in an intensive care unit. The system could be used for detecting and identifying bacterial infections of the lungs and airways in real-time. The patients all had ventilator assisted breathing and were diagnosed with respiratory failure due to severe pneumonia and other extrapulmonary diseases by two chest physicians. The electronic nose was based on piezoelectric quartz crystal microbalance sensors. The system used an array of 24 individual transducers each coated with a different peptide sequence ranging from 5 to 10 amino acids in length. The overall pattern response of the electronic nose to the patients’ breath was subjected to multiple discriminant analysis (MDA). The results of this were compared to data collected by conventional swab and sputum cultures taken from the same patients. Six different bacterial pathogens were identified and grouped into clusters by the MDA with 98% accuracy these were Pseudomonas aeruginosa, Acinetobacter baumannii, Klebsiella pneumoniae, Staphylococcus aureus and Acinetobacter lwoffii.