Fifty consecutive patients with fever and cirrhosis were prospectively studied to assess if cirrhotic fever was a true clinical entity and to determine its characteristics and outcome. In 20% (10) of the 50 patients, an identifiable source of fever or infection, was not documented (these patients were defined as having cirrhotic fever). The patients with cirrhotic fever were significantly less toxic, as indicated by lower temperature (P = .0001), tachycardia (P = .0005), and tachypnea (P = .05), but had fever for a longer duration (P = .009) than did patients with infectious fever. Patients with cirrhotic fever were significantly less likely to have focal signs or symptoms (P < .0001) or a portal of infection confirmed by culture (P = .0001), as compared with patients with infectious fever. Outcome (at 30-days or long-term) was not different for patients with cirrhotic fever vs.-patients with infectious fever or matched controls who did not have fever. Eight (80%) of the 10 patients with cirrhotic fever underwent transplantation; fever did not recur after transplantation in any of these patients. Thus, fever in up to 20% of the febrile patients with cirrhosis may be attributable to cirrhosis itself; such patients may be spared the ongoing diagnostic maneuvers and unnecessary trials of antibiotics.