Hepatic encephalopathy (HE) is a neuropsychiatric syndrome which develops commonly in liver cirrhosis following the build up of toxic substances in the blood that cross the blood-brain barrier and affect normal brain function. The diagnosis of HE is difficult due to only subtle impairments of cognitive function at early stages of the disease and the lack of a gold standard test that specifically and reliably detects the condition. HE is associated with a poor prognosis and effective treatment largely depends on early diagnosis. Thus the aim of this work was to investigate the use of breath analysis as a non-invasive and simpler means of diagnosing HE in cirrhotic patients. This was based on the hypothesis that toxins accumulating in the blood may, if sufficiently volatile, undergo alveolar gas exchange in the lungs to be excreted in the breath. Bespoke breath testing devices were utilised for the collection of breath onto Solid-Phase Micro-Extraction (SPME) fibres and adsorbent packed Automated Thermal Desorption (A TD) tubes from cirrhotic patients with and without HE, patients with early alcohol-related health problems, patients with respiratory disease, and healthy controls. Analysis of the breath samples collected was undertaken using Gas Chromatography Mass Spectrometry. In total, 237 different compounds were identified from all samples collected using the SPME breath analyser system and 385 using the ATD breath sampling device. Multivariate discriminant analysis was used to identify Volatile Organic Compounds (VOCs) that will discriminate patients according to disease status. More compounds were associated with the presence of HE compared to the absence of HE in alcoholic cirrhotic patients. Classification rules based on the presence or absence of volatiles correctly classified the presence of HE in 86% of patients tested with the SPME technique and 88% of patients tested with the A TD technique. Breath tests based on the presence or absence of discriminatory volatiles, correctly predicted the presence of cirrhosis in 93% and 96% of alcoholic patients tested using the SPME and ATD techniques, respectively. The presence or absence of four key volatiles on the breath also helped discriminate patients with early alcohol related health problems (ARHP) from healthy cases, correctly predicting the presence of ARHP in 78% and 91 % of alcoholic patients tested using the SPME and ATD systems, respectively. The use of a targeted sensor-array device found that, on average, HE patients exhaled higher concentrations of hydrogen, alcohol, and total VOCs compared to alcoholic cirrhotic patients without clinical signs of HE, especially in the non- smoking cohort of subjects studied. This finding shows that the use of gas sensor technologies in clinical practice can provide useful diagnostic information for clinical conditions at the bedside of patients in real-time. Not all breath volatiles come from the alveolar-blood interface; many can also be produced in the oral cavity by the action of bacterial or salivary enzymes on a range of substrates. Thus gases were sampled from tongue biofilm models in vitro and this identified 32 compounds commonly detected on breath. This highlights the need for further research to determine the major source of breath volatiles in order that suitable markers of systemic disease or metabolic disorders can be identified. Overall, the results reported in this thesis suggest that breath analysis is a useful tool for the non-invasive diagnosis of a range of conditions associated with the liver such as HE, cirrhosis, and, most importantly the presence of early alcohol related health problems before any significant damage to the liver has occurred.