Introduction Hepatic encephalopathy (HE) has a detrimental effect on patients' health-related quality of life and a significant negative impact on survival. Nevertheless, there is no diagnostic gold standard so the condition is often poorly diagnosed and managed.

Aim The aim of this study was to evaluate the performance validity of breath sample analysis for volatile organic compounds (VOCs) in the diagnosis of hepatic encephalopathy (HE) in patients with cirrhosis.

Method The patient population comprised 26 individuals (17 men, 9 women) of mean (range) age 60 (45 to 75) years, with biopsy-proven cirrhosis secondary to alcohol (n=21), or non-alcoholic steatosis, hepatitis C, autoimmune hepatitis, haemochromatosis or cryptogenic (n=1 each); none of the patients was currently misusing alcohol. Patients were classified using clinical, psychometric and electroencephalographic variables as either neuropsychiatrically unimpaired (n=10) or as having minimal (n=6) or overt (n=10) HE. Exhaled breath samples were collected, on the same day as the neuropsychiatric assessment, in a chemically inert 3L Tedlar bag (Adtech, UK). An evacuation device was used to evacuate 2L of gas from the bag into an empty glass tube which was packed with graphitised carbon adsorbents able to trap a wide range of VOCs (Sigma Aldrich, UK). Each tube was processed using a TurboMatrix automated thermal desorption unit (PerkinElmer, UK) attached to a Clarus 500 gas chromatograph mass spectrometer (GC-MS). In total 280 discrete peaks from the chromatographic output, excluding known experimental contaminants, were investigated for potential use as markers of HE.

Results A number of peaks were identified in the patients with cirrhosis which were absent or present in significantly different quantities in ten healthy controls. Discriminant analysis allowed the generation of two classification equations using standardised data from 12 peaks to build a predictive model for HE. This model correctly classified all patients from the original population (Abstract P13 figure 1).

• Download figure
• Open in new tab
• Download powerpoint

Abstract P13 Figure 1

Classification scores for the model used to predict HE.

Conclusion Analysis of VOCs in breath samples identifies patients with HE with a high degree of accuracy. Future work will validate the classification equations in a new group of patients, while identification of the individual compounds involved will provide insights into the pathogenesis of the syndrome and potential new therapeutic targets.