Sepsis-associated liver injury: Incidence, classification and the clinical significance

Although it is a common complication of sepsis, sepsis-associated liver injury has not been substantially recognized, because its diagnostic criteria and clinical implications are unclear. We aimed to elucidate the incidence, manifestation, disease type classification and prognosis of sepsis-associated liver injury.

The subjects were 588 patients admitted to our hospital for sepsis between 2001 and 2010. They were classified into "normal liver function", "sepsis-associated liver injury" and "sepsis-not-associated liver injury" groups. Sepsis-associated liver injury was classified as either "cholestatic", "hepatocellular" or "shock liver." Each of these three subgroups was further classified into "with jaundice" or "without jaundice". The primary end-point was the "poor prognosis ratio", defined as the proportion of patients whose prognosis was "unchanged", "worsened" or "died".

Among the 449 subjects except for sepsis-not-associated liver injury (n = 139), the incidence of sepsis-associated liver injury was 34.7% (156/449), including 75 cholestatic (48.1%), 34 hepatocellular (21.8%) and 47 shock liver (30.1%) cases. Jaundice was a complication in 25 (33%), six (17.6%) and four (8.5%) patients in each group, respectively. The poor prognosis ratio was higher in males (37.5%) and in the elderly (47.7%); it was 48.0%, 38.2% and 62.8% in the cholestatic, hepatocellular and shock liver groups, respectively, and higher than the normal liver function (18.4%) group (P < 0.0001). It was also higher in patients with jaundice (68.6%) than in those without (45.5%) (P < 0.0001).

Sepsis-associated liver injury, especially with jaundice, is a significant predictive sign of poor prognosis in patients with sepsis.