In a recent article, Mallet and colleagues [1] reported a series of eight cases of nodular regenerative hyperplasia (NRH) of the liver as a new cause of chronic liver disease in HIV-infected patients. NRH is characterized ‘by the presence of diffuse rather small regenerative nodules in the absence of significant fibrosis’. All patients were treated with HAART and all received didanosine included in one of several lines of antiretroviral drugs they had before the diagnosis of NRH. The authors speculate that a prothrombotic state enhanced by both HIV infection and didanosine toxicity may result in intrahepatic thrombosis leading to NRH. The prevalence of such cases could have been underestimated and we think is of interest to report similar cases. We would therefore like briefly to report a case diagnosed in Martinique, where we are presently following 700 HIV-infected patients. A 49-year-old woman was diagnosed with HIV in November 1996. In her records she reported a liver fracture associated with multiple trauma after a motorcycle crash in 1976. The liver was sutured and recovery was good. No analgesics were taken between 1976 and 1996. Oral contraceptives were the only drugs taken during that period (less than 24 months' cumulative exposure). At HIV diagnosis, the viral load was 36 000 copies/ml and the CD4 cell count was 414 cells/μl. In December 1996, a first line of HAART was initiated including didanosine, zidovudine and nevirapine and was stopped in July 2005 after 111 months for possible viramune hepatic toxicity. One year after the initiation of HAART non-icteric cholestasis had developed, with the gamma glutamyl transferase level reaching 10 times the upper limit of normal (ULN) and alkaline phosphatases reaching five times the ULN. Three years later there was a moderate cytolytic hepatitis with aspartate aminotransferase reaching 2.5 × ULN and alanine aminotransferase 1.5 × ULN, whereas the prothrombin time decreased to 60%, factor V to 65%, and thrombocytopenia was approximately 70 000 copies/m3, suggesting mild liver failure. In 2006, 5 years after the initiation of HAART, the patient presented with enlarged liver and spleen, portal hypertension with ascites and grade 1 gastroesophageal varices that improved with treatment by aldactone started in October 2005. She complained of iterative acute splenic pain. There was neither a history of alcohol intake nor toxic exposure. Laboratory work-up excluded shistosomiasis, viral hepatitis, autoimmune processes, haemochromatosis, alpha-1 antitrypsin deficiency and Wilson's disease. In April 2004 a liver biopsy had ruled out a specific parasitosis, mycosis or tumour and showed nodules of regenerating hepatocytes. Nodules were separated from the normal area by a dense net of reticuline fibres, and inside were rows of two to three enlarged hepatocytes (see Fig. 1) and vascular abnormalities with peliosis. There was no fibrosis. Although Bartonella serology was negative, the patient received clarithromycin without improvement.Fig. 1: Liver biopsy of the patient performed in April 2004, showing nodules of regenerating hepatocytes. Nodules are separated from the normal area by a dense net of reticuline fibres, and inside are rows of two to three enlarged hepatocytes and vascular abnormalities with peliosis. There is no fibrosisThe HIV-RNA level had been undetectable (below 200 copies/ml) under the first line,whereas theCD4 cell count slowly dropped to 176 cells/μl. After the first line was stopped, cotrimoxazole prophylaxis was started in June 2005, but no antiretroviral treatment was given during the 8 months after HAART ended. A new line of treatment was started in March 2006 with lopinavir, tenofovir and emtricitabine. Under this treatment, the HIV-RNA level is presently below 40 copies/ml and the CD4 cell count is 220 cells/μl. A retrospective diagnosis of NRH was made based on criteria defined by Mallet et al.[1]: (i) cryptogenic liver disease with abnormal liver function for more than 6 months, signs of portal hypertension and ultrasonographic liver dysmorphia; (ii) micronodulation with an alteration of thickened and compressed liver cells or suspected with the association of sinusoidal dilatation and clinical signs of portal hypertension. Finally, NHR is a classic evolution of peliosis. Our case is very similar to those reported by Mallet et al.[1] and emphasizes the seriousness of this complication: the risk of variceal haemorrhage, indication for some patients of transplantation (three out of eight of the cases reported by Mallet et al. [1]). Only HIV infection and HAART have been identified as potential co-factors. Nevirapine toxicity is well known, has not been specifically incriminated in NRH pathogenesis, but was also taken by our patient as well as by eight patients reported by Mallet et al. [1]. We agree with the authors that special attention should be paid to any hepatic dysfunction, however minor, under HAART that cannot be related to an identifiable cause.
Read full abstract