The present study aimed to i) assess the prognostic value of monoethylglycinexylidide (MEGX) formation kinetics in predicting liver dysfunction in patients with severe sepsis, and ii) to compare it with other oxygen indices and with gastric mucosal pH (pHi). Twenty-seven patients meeting the criteria for severe sepsis were prospectively evaluated. Patients with haemodynamic instability, elevated liver enzymes (>2 × normal) and bilirubin levels (>3 mg/dl), coagulation disorders and gastro-intestinal bleeding were not included. A gastric tonometer (TRIP, Tonometrics, Worcester, MA, USA) was introduced in all patients. Gastric pHi was calculated using the Henderson-Hasselbalch equation. Oxygen delivery (DO2) was calculated by the simplified formula DO2 =CI×CaO2, where CI is cardiac index and CaO2 is arterial oxygen content. Oxygen consumption (VO2) was determined by indirect calculation of the product of CI and arteriovenous oxygen content difference. Plasma aspartate aminotransferase, alanine aminotransferase, bilirubin and arterial lactate levels were determined by an enzymatic method. The serum MEGX concentrations were determined using an automated fluorescence polarisation immunoassay (TDX, Abbott, Chicago, IL, USA) before, 15 and 30 minutes after an intravenous injection of 1 mg/kg lignocaine. After recording the admission parameters, all patients were treated according to a supportive therapy protocol for severe sepsis and observed throughout their stay in the intensive care unit. Predetermined liver dysfunction indices (bil >3 mg/dl, liver enzymes >2 × normal) were recorded until discharge or death. The ability to predict liver dysfunction was compared by calculating the sensitivity and specificity for each variable (lactate, pHi, MEGX).To evaluate the accuracy with which these variables were able to discriminate between those patients who developed clinically demonstrable liver dysfunction (LD+) and those who did not (LD-), receiver-operating characteristic (ROC) curves were constructed. Values of the measurements given as medians were compared by the Mann-Whitney U-test. MEGX was the only parameter that showed a significant difference when comparing the two groups (LD+, LD-). In ROC curves constructed to compare the ability of the variables to discriminate between the two patient groups, MEGX displayed the greatest predictability of liver failure development compared to lactate or pHi. Relating these variables to survival, no significant values were found in any of the parameters. No correlation was found between MEGX or pHi and oxygen-derived variables (DO2, VO2). There was also no significant difference in haemodynamics and oxygen-derived variables either between LD+ and LD- groups, or between survivors and non-survivors. Our data demonstrate that MEGX formation kinetics is an early bedside procedure to assess hepatic function dynamically. This test might be an early sign of hepatic insult in sepsis despite global circulatory and splanchnic circulation indices remaining within normal ranges. More research is essential to see whether this parameter can guide therapeutic strategies for preventing liver hypoxia or metabolic impairment in the early phase of severe sepsis.