A randomized open label study of 'imipenem vs. cefepime' in spontaneous bacterial peritonitis.

BACKGROUND & AIMS: Spontaneous bacterial peritonitis (SBP), in the presence of bacterial resistance or failure of third generation cephalosporins (3rd GC) has poor outcome. Empirical antibiotic(s) options are limited in these scenarios.

METHODS: Consecutive cirrhotics with SBP because of hospital acquired SBP (>48 h of admission), microbial resistance or non-response (no resolution of SBP at 48 h) were randomized to Cefepime (n = 88) or Imipenem (n = 87) plus standard medical therapy. We assessed for 'response at 48 h' (reduction in ascitic fluid absolute neutrophil count (ANC) by >25% at 48 h), resolution of SBP (<250 cu/mm ANC at day 5) and their clinical outcome.

RESULTS: Of 957 paracentesis in 1200 hospitalized cirrhotics, 253 (26.4%) had SBP and 175 (69.6%) were randomized. Baseline parameters were comparable in two groups. Response at 48 h (58.6% vs. 51.7%; P = 0.4) and resolution of SBP in those with response at 48 h were comparable with no difference in mortality at week 2, month 1 and 3. Patients with 'No response at 48 h' had higher mortality compared with responders (73.8% vs. 25%; P < 0.001). Resolution of SBP was associated with 'response at 48 h' and septic shock, latter being main pre-terminal event. AKI at enrolment [Hazard ratio (HR), 2.6], pneumonia [HR, 2.9], septic shock [HR, 2.2] and response at 48 h [HR, 4.6] predicted poor outcome.

CONCLUSIONS: In hospitalized cirrhotics with SBP and risk factors for treatment failure, cefepime showed comparable efficacy and survival to imipenem. Non-response to therapy at 48 h is a reliable predictor of treatment failure and mortality. Antibiotic combinations and novel options are needed for these patients.