Diagnosis
Abdominal pain and persistent fever in a patient with pancreatitis.
Epidemiologic Risks
Severe pancreatitis often goes on to become complicated by infection of the phlegmon and abscess formation.
Microbiology
Some variation of the theme of bowel flora (Escherichia coli, streptococci, enterococci, anaerobes and not infrequently Candida), you really need to get a specific organism so tap the pus for cultures, usually treat like intra-abdominal infections.
Like so many infections, when more sensitive modalities are used, bacteremia with a wide number of organisms is common: "Bacterial DNA was detected in peripheral blood from 68.8% of patients with acute pancreatitis, and more than half (60.4%) of the patients encountered polymicrobial flora. Translocated bacteria in patients with acute pancreatitis were primarily constituted of opportunistic pathogens derived from the gut, including Escherichia coli, Shigella flexneri, Enterobacteriaceae bacterium, Acinetobacter lwoffii, Bacillus coagulans, and Enterococcus faecium. The species of circulating bacteria shifted remarkably among the patients with different severity. The presence of the bacteremia correlated positively with the Acute Physiology and Chronic Health Evaluation-II scores of patients with acute pancreatitis (r = 0.7918, p < 0.0001)(PubMed)".
Empiric Therapy
I prefer a third-generation cephalosporin PLUS metronidazole. Alternatives include carbapenem (one trial suggests imipenem is best and may prevent the need for surgery) OR penicillin/beta-lactamase inhibitors OR a quinolone PLUS metronidazole.
A meta-analysis suggests benefit from a carbapenem (PubMed). It is common to give prophylactic antibiotics IV 10 to 14 days (often a carbapenem) as it (maybe) decreases super-infection and (perhaps) mortality in patients with severe acute pancreatitis with CT findings of pancreatic necrosis, use the same antibiotics as for empiric therapy. But the data is not convincing of benefit: "Antibiotic prophylaxis does not significantly reduce the incidence of infected pancreatic necrosis but may affect allcause mortality in acute necrotizing pancreatitis (PubMed)."
Fluconazole for the prevention of Candida also supported in some studies. Data to support this practice is marginal and contradictory. I bet we need more studies.
Pearls
The problem is one cannot tell the difference between infected pancreatitis and non-infected pancreatitis, although in my role as a consultant I rely on my infallibility to be able to make just that differentiation. They both have fevers, increased WBC and abdominal pain. So they end up receiving lots of antibiotics with negative cultures, in the meantime breeding resistant organisms. Given the 16S studies, perhaps a short course of antibiotics should be considered (PubMed).
The data and increasing consensus opinion are that antibiotics for necrotizing pancreatitis is a waste of time. It is better to sample the fluid and treat what you grow. But the whole topic is as clear as a pancreatic phlegmon aspirate.
It is probably better when you suspect an abscess to drain it and not send the patient to OR, the conservative approach has fewer complications and maybe less death. If the NEJM says it, it must be so.
Rants
Curious Cases
Relevant links to my Medscape blog
Last Update: 12/11/18.