Randomised controlled trials (RCTs) of prophylactic antibiotic treatment (beta-lactams, quinolones or cephalosporins) versus placebo/no treatment (control) in patients with acute necrotising pancreatitis with computer tomography-proven necrosis were eligible for inclusion. Trials had to provide outcome data on infected pancreatic necrosis, mortality, non-pancreatic infection or need for surgical intervention. It appeared that trials that studied oral antibiotics were excluded.
The included trials studied prophylaxis with imipenem, cefuroxime, ofloxacin, metronidazole, imipenem-cilastatin or meropenem (various doses and regimens) versus placebo or no treatment. Four trials were multicentre and five trials were single centre. Trials were published between 1993 and 2009. Acute necrotising pancreatitis was mainly due to gallstones and alcohol.
The authors did not state how many reviewers undertook study selection.