Mh. Collins et al., HEPATIC GRANULOMAS IN CHILDREN - A CLINICOPATHOLOGICAL ANALYSIS OF 23CASES INCLUDING POLYMERASE CHAIN-REACTION FOR HISTOPLASMA, The American journal of surgical pathology, 20(3), 1996, pp. 332-338
In a 15-year period at the Riley Hospital for Children, granulomas wer
e found in 23 (4%) of a total of 521 liver biopsies. An etiology was i
dentified in 87%: Histoplasma was diagnosed in 15 cases (65%) by polym
erase chain reaction (PCR) on paraffin-embedded tissue, serology, and
special stains; sarcoidosis was diagnosed in four cases; and schistoso
miasis was diagnosed in one case. Serial liver biopsies were available
from five patients; granulomas occurred in only one biopsy of the ser
ies from each patient. Extrahepatic tissue from six patients contained
granulomas, and an etiology for the liver granulomas was identified i
n all six patients (four histoplasmosis, two sarcoidosis). The extrahe
patic tissue from two patients with Histoplasma was diagnostic. We mad
e the following conclusions: that PCR is applicable to archival materi
al and greatly increases the yield of specific infectious diagnoses of
liver granulomas compared with conventional diagnostic methods (65 ve
rsus 22%); that the infections causing liver granulomas are those that
are endemic in a community (e.g., Histoplasma in Indiana); that Histo
plasma can coexist with a wide variety of systemic and primary liver d
iseases; that the likelihood of identifying a cause of liver granuloma
s is increased if there are extrahepatic granulomas; and that hepatic
granulomas may have a limited life span. Treatment of liver granulomas
should be determined by the clinical setting and directed at the unde
rlying cause.