Recurrence of spontaneous bacterial peritonitis in cirrhotic patients non-prophylactically treated with norfloxacin: serum albumin as an easy but reliable predictive factor

Background Several large studies revealed that selective intestinal decontamination (SID) prevented recurrence of spontaneous bacterial peritonitis (SBP) in cirrhotic patients. Nonetheless, there are no definitive patient selection parameters identifying who would benefit from SID. Aims To investigate long-term outcomes in cirrhosis patients with recurrence of SBP and to identify predictive factors for SBP recurrence. Methods We retrospectively studied 146 cirrhosis patients diagnosed with a first episode of SBP from 2005 to 2006. Of these, 89 patients survived; the survivors were divided into two groups based on recurrence and non-recurrence of SBP, and clinical parameters, survival time and cause of death were analysed. Results The in-hospital mortality was 39% (57/146). The SBP recurrence rate was 42.7% (38/89). The survival rate between patients with recurrent SBP and those without recurrence did not differ (P=0.092). Sepsis was the major cause of death in the recurrent SBP group, but not in the non-recurrent group. Serum albumin level before discharge and beta-blocker use between the two groups differed significantly (P < 0.0001). Using the cut-off point for serum albumin level before discharge of 2.85 g/dl as a predictor for recurrence of SBP, the sensitivity was 70.2% and the specificity was 76.3%. Furthermore, long-term survival of the group with high albumin before discharge was better than that of the corresponding group with low albumin (P=0.007). Conclusion Spontaneous bacterial peritonitis was associated with high sepsis-related mortality in cirrhotic patients. Serum albumin before discharge was a useful single parameter to predict the recurrence of SBP and long-term survival.