Diagnosis
Fever, RUQ pain and abscess in the liver on CT or US. It is always annoying how often patients lack RUQ pain, around a third (PubMed). Lab findings usually not abnormal enough to help.
Epidemiologic Risks
Most occur from an underlying biliary disease of any kind, other causes include downstream from colon pathology (especially appendicitis and diverticulitis), hematogenous, and bad luck.
C. albicans and other fungi can occur in the neutropenic and if aspergillus occurs in a young person for no reason, think of Chronic Granulomatous Disease.
Diabetics are also at risk especially with Klebsiella (PubMed).
PPIs increase the risk of liver abscess and mortality (PubMed).
Microbiology
Pyogenic: often mixed: E. coli + Streptococci (especially milleri group) + anaerobes.
Occasionally S. aureus.
Mono-bacterial infections with Klebsiella pneumoniae (the hyper-viscous kind) are not uncommon, especially in diabetics (PubMed). About 10% of Klebsiella liver abscess will have a metastatic complication, most often eyes and lung (PubMed). There may be an association between Klebsiella liver abscesses and malignancy (PubMed).
Amoebic: E. histolytica.
Candida: Hmmmm. Candida? Yep. Candida, especially C. albicans after a prolonged course of neutropenic fever.
AIDS: Bartonella (aka cat scratch disease), especially in HIV disease.
There is a syndrome of aseptic liver abscesses, an autoimmune disease most often associated with IBD and related to Sweets, pyoderma gangrenosum and Behçet disease (PubMed).
Empiric Therapy
I prefer a third-generation cephalosporin PLUS metronidazole.
Alternatives include a carbapenem OR penicillin/beta-lactamase inhibitors OR a quinolone PLUS metronidazole. Clindamycin can be substituted for the metronidazole.
Drain (if < 2-3 cm, probably can take care of it medically) PLUS antibiotics if bacterial
Antibiotics alone if amoebic.
Candida, probably lipid-based amphotericin B.
Duration? Until it is better. I go until radiographic resolution. Usually 4,6,8 weeks of antibiotics, iv or po depending on what grows and how well it has been drained.
Pearls
Patients with CGD associated liver abscesses do better with steroids (PubMed).
There is no way to tell the difference clinically or radiographically between bacterial and amoebic abscesses(Pubmed).
Culture negative abscesses seem to do as well as those that grow something (Pubmed).
Curious Cases
Relevant links to my Medscape blog
Good time are slowing fading away
Great Ideas Disproved by Reality Yet Again
Probably Not Relevant But Fun to Think About
Last Update: 06/18/19.