Introduction Monitoring of breath volatiles could provide non-invasive rapid assessment of metabolic stress, inflammation and intravenous anaesthetic drug concentrations in anaesthetised patients. This study was designed to validate an off-line single breath end-tidal sampling method with analysis by selected ion flow tube mass spectrometry (SIFT-MS). Exhaled acetone was used as a model compound due to its abundance in breath and ease of measurement.

Methods End-expiratory breath samples from 22 healthy, non-diabetic, elective and semi-elective surgical patients were collected into Tedlar^® bags via a T-piece adjacent to the endotracheal tube. The effects of different breathing systems, the breathing circuit filter, and consequences of altering the inspiratory gas flow rate and adjustable pressure limiting (APL) valve on exhaled acetone concentrations were explored in subgroups of these patients.

Results Median exhaled acetone concentration was 738 ppb (range 257–6594 ppb) for samples collected on the patient side of the circuit filter with the APL valve open (usual position) (n = 22). Median intra-subject coefficient of variation for exhaled acetone concentration using this method was 8.3% (interquartile range 6.9–14.5%). Higher inspiratory but not exhaled acetone concentrations were seen when using the ADU Carestation compared to the Aysis Carestation anaesthesia machines (median inspiratory concentration 276 ppb v 131 ppb, p = 0.0005; median exhaled concentration 630 ppb v 513 ppb, p = 0.95). Altering the inspiratory gas flow rate did not significantly affect exhaled acetone concentration; however APL valve closure resulted in a reduction in exhaled acetone concentration. Higher concentrations of acetone were measured in breath samples collected from the patient side of the circuit filter, since collection of samples after the filter resulted in dilution by deadspace air. Breath acetone concentration was related to plasma acetone (r_s = 0.80, p < 0.0001) and plasma beta-hydroxybutyrate concentrations (r_s = 0.49, p = 0.029).

Conclusions This non-invasive method of end-tidal breath collection in anaesthetised patients is reproducible for the analysis of acetone and is suitable for repeated sampling. With appropriate validation, the same method could be applied to the collection of other volatiles in the breath of intubated and ventilated patients, making it possible to investigate the concentrations of other potential biomarkers in this patient group.