Background/aims: Patients with liver cirrhosis and ascites have a high risk of spontaneous bacterial peritonitis, but the prognostic impact of spontaneous bacterial peritonitis has not been well examined.
Methodology: Patients with liver cirrhosis and ascites were included at the time of their first paracentesis during hospitalization in the Department of Hepatology, Aarhus University Hospital, Denmark, between September 1992 and September 2000. Cox regression was used to estimate the mortality of patients with spontaneous bacterial peritonitis (ascites leukocyte count > or = 250 per mm3) relative to controls without spontaneous bacterial peritonitis. Furthermore, we used Cox regression to estimate the change in mortality when controls developed spontaneous bacterial peritonitis during follow-up.
Results: Of 286 patients, 76 (27%) had spontaneous bacterial peritonitis at the first paracentesis. The mortality ratio of patients with spontaneous bacterial peritonitis relative to controls was 1.0 (95% confidence interval 0.7-1.5) after adjustment for age, gender, comorbidity, and alcohol abuse. Of the 210 controls, 42 (20%) were found to have spontaneous bacterial peritonitis at a later paracentesis. Their mortality rate more than doubled with the onset of spontaneous bacterial peritonitis.
Conclusions: Spontaneous bacterial peritonitis at the first paracentesis did not affect the prognosis of patients with liver cirrhosis, whereas development of spontaneous bacterial peritonitis during follow-up doubled the mortality risk. This may be due to a longer diagnostic delay in those who developed spontaneous bacterial peritonitis during follow-up.