[Comparative profitability of hepatic biopsy and microbiological tests in patients with HIV infection]

Diagnostic liver biopsy is proposed in HIV-positive patients who present unexplained fever. This invasive procedure is truly useful if it allows establishing a difficult diagnosis or improves survival rate. We conducted a retrospective study to determine the diagnostic and prognostic power of liver biopsy in HIV-positive patients with fever.

One hundred thirty-eight liver biopsies were performed in 129 patients. Utility was defined as demonstration of the pathogen or identification of a tumoral process.

The liver biopsy met the utility criteria in 27 cases showing mycobacterial infections (n = 22) and herpes hepatitis, type 1 herpes simplex virus, cytomegalovirus and cryptococcosis infections (n = 1 each). These last 4 diagnoses were also possible with other tests. Comparing non-contributive liver biopsies (n = 111) with those demonstrating hepatic mycobacterial infection (n = 22) showed that the two groups were not different in terms of demographic data. Splenomegalia was more frequent in the non-contributive group (68% vs 37%, p = 0.007) as was superficial lymph node enlargement (45% vs 12%, p < 10(-3)). Laboratory tests were not discriminating. Mycobacterial infection was diagnosed in 22 patients in the non-contributive group. Bacteriological samples were positive for mycobacterium in 20 of the 22 patients in the contributive group. The mean delay to the first positive test for mycobacterium was 15 +/- 8 days compared with 30 +/- 10 days for liver tissue cultures. Mean survival after liver biopsy was 10 months: patients with a positive Ziehl-Neelson stain on the liver biopsy did not have a longer survival (9.7 +/- 7.6 vs 10.2 +/- 10.4 months).

In most cases, liver biopsy in HIV-positive patients with fever provides a diagnosis which can be obtained with non-invasive techniques without improving prognosis.