5
ORIGINAL ARTICLE The development of hepatoportal sclerosis and portal hypertension due to didanosine use in HIV Thomas D. Schiano & Alison Uriel & Douglas T. Dieterich & M. Isabel Fiel Received: 6 August 2010 / Revised: 16 October 2010 / Accepted: 19 October 2010 / Published online: 6 November 2010 # Springer-Verlag 2010 Abstract Hepatoportal sclerosis (HPS) is one of several entities known to cause noncirrhotic portal hypertension. To date, its etiology is unknown. There have been increasing reports of HPS occurring in patients with human immuno- deficiency virus (HIV), and the US Food and Drug Administration (FDA) recently issued an advisory regard- ing the development of noncirrhotic portal hypertension in association with didanosine (ddI) use. We report on a patient with HIV who had taken ddI for 4 years and who developed portal hypertension. Histopathological review of paired liver biopsies showed an initial drug hepatotoxicity, microvascular liver injury, and the presence of HPS. Despite cessation of ddI, the latter biopsy showed resolu- tion of the drug-induced injury, but it also showed progression of the HPS. The patients portal hypertension also progressed suggestive of an unremitting vascular injury. This case demonstrates the development of HPS resulting from a drug-induced microvascular injury. The paired biopsies demonstrate that the initial vascular injury may disappear but that the portal hypertension and HPS progress. Keywords HIV . Noncirrhotic portal hypertension . Hepatoportal sclerosis . Nodular regenerative hyperplasia . Didanosine . Antiretroviral therapy Introduction Hepatoportal sclerosis (HPS) and nodular regenerative hyperplasia (NRH) are two of several entities that are known to cause portal hypertension in the absence of cirrhosis [1, 2]. Unlike NRH which has several well- documented causes (medications such as azathioprine, collagen vascular, and myeloproliferative diseases, etc.) [3], HPS, to date, has no known etiology. In HPS, the primary hepatic lesions are found in the portal tracts which show varying degrees of fibrosis, sclerosis of portal vein radicles, and abnormally formed portal veins. There are increasing reports of both HPS and NRH as well as noncirrhotic portal hypertension occurring in patients having human immunodeficiency virus (HIV) [46]. Several investigators have suggested that the portal hypertension may be related to use of certain highly active anti retroviral therapy (HAART) medications [710]. After reviewing 42 such cases, on January 29, 2010, the FDA issued an advisory regarding the development of portal hypertension occurring in patients with HIV taking didanosine (ddI; http://www.fda.gov/Safety/MedWatch/Safety Information/SafetyAlertsforHumanMedicalProducts/ ucm199343.htm). NRH has been shown to develop after the use of 6- thioguanine in patients being treated for inflammatory bowel disease and acute lymphoblastic leukemia [11, 12]. 6-Thioguanine is an antimetabolite akin in structure to nucleoside reverse transcriptase inhibitors such as zidovu- dine (AZT), stavudine (d4T), and ddI. Such drugs may affect the hepatic microvasculature as other antimetabolite medications are known to do, such as azathioprine causing both veno-occlusive disease and NRH [13]. Herein, we report a case of HPS and NRH developing in a patient who was using ddI with the initial histologic findings suggestive T. D. Schiano : A. Uriel : D. T. Dieterich Division of Liver Diseases, Department of Medicine, The Mount Sinai Medical Center, New York, NY, USA M. I. Fiel (*) Department of Pathology, The Mount Sinai Medical Center, Box 1194, One Gustave Levy Place, New York, NY, USA e-mail: [email protected] Virchows Arch (2011) 458:231235 DOI 10.1007/s00428-010-1004-7

The development of hepatoportal sclerosis and portal hypertension due to didanosine use in HIV

Embed Size (px)

Citation preview

Page 1: The development of hepatoportal sclerosis and portal hypertension due to didanosine use in HIV

ORIGINAL ARTICLE

The development of hepatoportal sclerosis and portalhypertension due to didanosine use in HIV

Thomas D. Schiano & Alison Uriel &Douglas T. Dieterich & M. Isabel Fiel

Received: 6 August 2010 /Revised: 16 October 2010 /Accepted: 19 October 2010 /Published online: 6 November 2010# Springer-Verlag 2010

Abstract Hepatoportal sclerosis (HPS) is one of severalentities known to cause noncirrhotic portal hypertension. Todate, its etiology is unknown. There have been increasingreports of HPS occurring in patients with human immuno-deficiency virus (HIV), and the US Food and DrugAdministration (FDA) recently issued an advisory regard-ing the development of noncirrhotic portal hypertension inassociation with didanosine (ddI) use. We report on apatient with HIV who had taken ddI for 4 years and whodeveloped portal hypertension. Histopathological review ofpaired liver biopsies showed an initial drug hepatotoxicity,microvascular liver injury, and the presence of HPS.Despite cessation of ddI, the latter biopsy showed resolu-tion of the drug-induced injury, but it also showedprogression of the HPS. The patient’s portal hypertensionalso progressed suggestive of an unremitting vascularinjury. This case demonstrates the development of HPSresulting from a drug-induced microvascular injury. Thepaired biopsies demonstrate that the initial vascular injurymay disappear but that the portal hypertension and HPSprogress.

Keywords HIV. Noncirrhotic portal hypertension .

Hepatoportal sclerosis . Nodular regenerative hyperplasia .

Didanosine . Antiretroviral therapy

Introduction

Hepatoportal sclerosis (HPS) and nodular regenerativehyperplasia (NRH) are two of several entities that areknown to cause portal hypertension in the absence ofcirrhosis [1, 2]. Unlike NRH which has several well-documented causes (medications such as azathioprine,collagen vascular, and myeloproliferative diseases, etc.)[3], HPS, to date, has no known etiology. In HPS, theprimary hepatic lesions are found in the portal tracts whichshow varying degrees of fibrosis, sclerosis of portal veinradicles, and abnormally formed portal veins. There areincreasing reports of both HPS and NRH as well asnoncirrhotic portal hypertension occurring in patientshaving human immunodeficiency virus (HIV) [4–6].Several investigators have suggested that the portalhypertension may be related to use of certain highly activeanti retroviral therapy (HAART) medications [7–10]. Afterreviewing 42 such cases, on January 29, 2010, the FDAissued an advisory regarding the development of portalhypertension occurring in patients with HIV takingdidanosine (ddI; http://www.fda.gov/Safety/MedWatch/SafetyInformation/SafetyAlertsforHumanMedicalProducts/ucm199343.htm).

NRH has been shown to develop after the use of 6-thioguanine in patients being treated for inflammatorybowel disease and acute lymphoblastic leukemia [11, 12].6-Thioguanine is an antimetabolite akin in structure tonucleoside reverse transcriptase inhibitors such as zidovu-dine (AZT), stavudine (d4T), and ddI. Such drugs mayaffect the hepatic microvasculature as other antimetabolitemedications are known to do, such as azathioprine causingboth veno-occlusive disease and NRH [13]. Herein, wereport a case of HPS and NRH developing in a patient whowas using ddI with the initial histologic findings suggestive

T. D. Schiano :A. Uriel :D. T. DieterichDivision of Liver Diseases, Department of Medicine,The Mount Sinai Medical Center,New York, NY, USA

M. I. Fiel (*)Department of Pathology, The Mount Sinai Medical Center,Box 1194, One Gustave Levy Place,New York, NY, USAe-mail: [email protected]

Virchows Arch (2011) 458:231–235DOI 10.1007/s00428-010-1004-7

Page 2: The development of hepatoportal sclerosis and portal hypertension due to didanosine use in HIV

of drug hepatotoxicity and injury to the hepatic microvas-culature. This thus suggests drug hepatotoxicity (possiblyfrom ddI) as an etiology for HPS. Also of concern in thiscase is the progressive development of portal hypertensiondespite cessation of ddI, which suggests an unremittinghepatic vascular injury.

Material and methods

Clinical history

A 45-year-old man with HIV presented in 2003 withabnormal liver chemistry tests. His risk factor for HIVacquisition was men having sex with men. The HIV wasinitially diagnosed in 1992, and, to date, he has not had anAIDS-defining illness. The HIV RNAwas 98 copies/ml andthe CD4 count was 376/mm3. There was no history ofalcohol use or previous history of liver disease. Uponpresentation, AST was 70 U/l, ALT 87 U/l, alkalinephosphatase 286 U/l, and GGTP 584 U/l with normalserum bilirubin, albumin, platelets, and INR. Liver tests1 year previously were AST 59 U/l, ALT 46 U/l, alkalinephosphatase 151 U/l, and GGTP 300 U/l. Fasting glucose/insulin and lipid panel were normal with no clinicalevidence of metabolic syndrome. Extensive serologicwork-up for etiology of the abnormal liver chemistry testsas well as abdominal imaging was unremarkable except formild hepatomegaly. The patient had been taking antiretro-viral medications since 1994, starting with AZT andlamivudine which were subsequently changed to d4T, ddI,and saquinavir in 1998. In 1999, nelfinavir was substitutedfor saquinavir. Two liver biopsies were performed, andpaired specimens were histopathologically analyzed.

Results

The first liver biopsy performed consisted of a 3.8-cmlength of liver tissue with bland hepatocellular cholestasisand formation of rosettes as well as mild featherydegeneration of hepatocytes. There was no portal or lobularinflammation. Portal areas were densely fibrotic and portalvein radicles were either totally obliterated or had markedlydiminished luminal caliber (phlebosclerosis). Hepaticarterioles had thickened walls and a decreased lumen(Fig. 1a). Patchy chronic inflammatory cells were presentaround the obliterated portal veins; foci of central venulitiswere also present (Fig. 1b). Some densely fibrotic portaltracts were in close proximity to each other indicative ofan abnormality of the hepatic architecture (Fig. 1c).Furthermore, although there was no cirrhosis, a distinctlynodular architecture was apparent and, when highlighted

by a reticulin stain, consisted of hyperplastic hepatocytesbordered by condensed reticulin fibers with atrophichepatocytes, consistent with NRH (Fig. 1d). There wasno microvesicular or macrovesicular steatosis present inthe biopsy. The constellation of findings was consistentwith a drug-induced liver injury and concomitant HPSwith NRH.

Based on the liver biopsy findings and concern for drughepatotoxicity, the patient’s antiretroviral regimen waschanged to Trizivir (AZT/abacavir/lamivudine) and tenofovir(Viread). He subsequently followed up on a biannual basisand was asymptomatic until 2008, when a computerizedtomography (CT) scan was performed as part of a work-up foranemia and thrombocytopenia. The platelet count haddecreased from 179,000 mm3 in July 2005 to 149,000 mm3

in May 2006 to 104,000 mm3 in November 2008. Otherlaboratory testing at this time included hemoglobin 11.2 g/dl,GGTP 333 U/l, albumin 4.0 g/dl, bilirubin 1.0 mg/dl,alkaline phosphatase 166 U/l, AST 44 U/l, ALT 43 U/l,total protein 7.5 g/dl, rapid plasma reagin (RPR) negative,testosterone 318 ng/dl, and CD4 count 370/mm3. CT scanshowed hepatosplenomegaly and varices; the portal vein waspatent. Further extensive serologic work-up was negative forviral hepatitis and other etiologies of liver disease such asautoimmune hepatitis, Wilson disease or hemochromatosis,and alpha-1-anti-trypsin deficiency. He was not taking anyother prescription or herbal medications.

The second needle liver biopsy that was performedconsisted of a 2.1-cm length of liver tissue that showedsmall portal tracts with no fibrosis, markedly dilated anddystrophic portal vein branches, some of which herniatedinto the periportal zones consistent with HPS (Fig. 2). Bileducts appeared normal; however, the accompanying hepaticarterioles were difficult to identify. Portal tracts were distantfrom each other with thickening of hepatocyte platesindicative of regeneration; mild condensation of reticulinfibers and subtle nodularity were also noted. Compared tothe initial biopsy, the current liver histology showed asignificant regression in portal fibrosis with dramaticchanges in portal vein configuration (consistent withHPS), the cholestasis had resolved, and the hepatic venulitishad disappeared; the NRH was less defined.

Discussion

The development of noncirrhotic portal hypertension due toboth NRH and HPS in patients with HIV is beingincreasingly recognized. Several investigators have alsonoted that patients with HIV may develop cryptogenicliver injury and cirrhosis, even in the pediatric population[6, 14]. In January 2010, the FDA issued an advisoryregarding the potential causation of this hepatotoxicity as

232 Virchows Arch (2011) 458:231–235

Page 3: The development of hepatoportal sclerosis and portal hypertension due to didanosine use in HIV

Fig. 2 a Second liver biopsy. A portal tract is shown with adisproportionately enlarged dystrophic portal vein that herniates intothe surrounding parenchyma (arrow). Inset shows another portal tractwith herniated portal veins creating megasinusoids (arrows). (H&E,

original magnification ×100). b A portal tract with fibrosis is shown(arrow). A slightly damaged bile duct with the accompanying arteryare present; however, the portal vein branch is missing

Fig. 1 a First liver biopsy. A densely fibrotic portal tract is shownwith partial obliteration of the portal vein branch (arrowheads). Afocus of inflammatory cells lifting the endothelial lining at the edge ofthe vein indicative of venulitis is also shown (arrow). The hepaticarteriole has a thickened wall. (H&E, original magnification ×400). bA representative centrilobular area showing central venulitis (arrow).The lobules show cholestasis with formation of rosettes (arrowheads).

Note the presence of markedly dilated sinusoids (asterisks). (H&E,original magnification ×100). c Two portal tracts with dense fibrosiswith no apparent portal vein branch identified (phlebosclerosis).(Trichrome stain, original magnification ×100). d Nodular regenera-tive hyperplasia showing condensation of reticulin fibers at the edgeof a poorly defined nodule (arrows). (Reticulin stain, originalmagnification ×100)

Virchows Arch (2011) 458:231–235 233

Page 4: The development of hepatoportal sclerosis and portal hypertension due to didanosine use in HIV

being due to ddI (http://www.fda.gov/Safety/MedWatch/SafetyInformation/SafetyAlertsforHumanMedicalProducts/ucm199343.htm). The FDA’s decision to revise the druglabel for ddI was based on 42 postmarketing reports ofnoncirrhotic portal hypertension occurring in patients usingddI. Twenty-six patients were male, ages ranging from 10 to65 years, with a duration of ddI treatment ranging frommonths to years.

The current case demonstrates the initial development ofhistologic liver damage strongly suggestive of drughepatotoxicity with hepatic microvascular injury in apatient taking ddI for 4 years. The initial changes wereconsistent with a cholestatic-type drug-induced liverinjury and portal tracts showed changes of HPS.Endotheliitis of the portal vein branches and centralvenulitis were present. NRH was also noted at initialpresentation. Subsequent liver biopsy showed the well-established portal venopathy of HPS and concomitantNRH with the resolution of the portal vein endotheliitisand central venulitis. These paired biopsies suggest thatHPS stemmed from an endothelial injury most likely dueto ddI. This is the first documentation that HPS is relatedto a drug-induced vascular injury and is the root causefor the cases of noncirrhotic portal hypertension beingdescribed in HIV.

HPS and NRH are known to be part of the samespectrum of conditions causing noncirrhotic portal hyper-tension that include incomplete septal cirrhosis, portal veinthrombosis, different vasculitides, and schistosomiasis[15–17]. NRH has been known to result from hepaticvascular injury so the current report now confirms asimilar etiology for HPS. Although it has been longpresumed that HPS results from hepatic vascular injury,the current report strongly suggests that direct drughepatotoxicity can now be implicated.

Although it is possible that ddI is not the cause ofhepatotoxicity in the current case, the increasing reportsof liver disease and noncirrhotic portal hypertensionoccurring with ddI use strongly implicate it. What ismost concerning about the progression of the histologicchanges in the patient we describe is the rapid develop-ment of portal hypertension (as suggested by worseningthrombocytopenia) and persistence of HPS and NRHover 5 years despite cessation of ddI. This occurreddespite minimal abnormalities in liver chemistry tests;clearly, a similar development might be otherwisemasked in patients having concurrent viral hepatitis.Patients with HIV are often coinfected with viralhepatitides, such as hepatitis B and C, and can also haveappreciable steatosis [18]. These heterogeneous histologicfeatures may potentially confound the identification ofHPS especially on needle liver biopsy in which it can beeasily missed. Although HPS appears rarely associated

with hepatic synthetic compromise and liver failure [19–21], it is now being increasingly recognized. At our owncenter, retrospective review of the liver explant of a patientwith HIV transplanted in 1999 for presumed cryptogeniccirrhosis also showed HPS. This patient had taken ddI forseveral years.

It is well-established that patients with HIV and HCVcoinfection have a more rapid progression in hepaticfibrosis [22]. The use of HAART including ddI hasactually been shown to slow fibrosis progression in suchpatients [23, 24]. Several investigators have shown thedevelopment of HPS occurring in the absence of ddI usewith the suggestion of there being an underlyinghypercoagulable disorder or concurrent portal veinthrombosis [4, 8, 25]. However, it is difficult todefinitively prove that these patients never previouslyused ddI as it was a major part of the early anti-HIVarmamentarium. It appears from the current study andfrom others that even short-term use of ddI can result inhepatotoxicity.

It is clear that in some patients with HIV having used ddIeven for short periods of time, the gradual development ofbiochemical and radiologic evidence of portal hypertensioncan occur in the presence of minimal liver chemistry testabnormalities. The current case strongly suggests that thedevelopment of noncirrhotic portal hypertension from ddIuse is directly related to a drug hepatotoxicity andresultant hepatic vascular injury. Similar precedents existwith patients developing veno-occlusive disease andNRH after the short-term use of azathioprine and 6-thioguanine. In this case, the initial insult resulted inprogressive and more chronic liver injury despite cessa-tion of the inciting agent. The paired liver biopsies inthis case showed resolution of cholestasis, absence ofendotheliitis and central venulitis, consistent with theremoval of a hepatotoxic agent. The fibrosis seen on theinitial liver biopsy was also resolved, suggesting theremoval of an inciting agent. HPS and NRH were seenon both liver biopsies with NRH being less defined onthe latter biopsy. In addition, HPS was manifested bydystrophic veins without surrounding hepatocellulardamage, suggesting that the increasing portal hyperten-sion is ascribable to this vascular abnormality. Whetherintervention at the initial detection of the hepatic vascularinjury would prevent the development of HPS or NRH isnot known.

In conclusion, the current case demonstrates thedevelopment of HPS and noncirrhotic portal hypertensionin a patient initially developing drug hepatotoxicity andhepatic microvascular injury from ddI use. Thus, HPSshould no longer be considered idiopathic, as this casesuggests that it may share similar pathophysiologicorigins as NRH. It appears that drug hepatotoxicity from

234 Virchows Arch (2011) 458:231–235

Page 5: The development of hepatoportal sclerosis and portal hypertension due to didanosine use in HIV

ddI may be the cause of this hepatic microvascular insultand supports the recent concerns regarding its use.

Conflicts of Interest None.

References

1. Bioulac-Sage P, Le Bail B, Bernard PH, Balabaud C (1995)Hepatoportal sclerosis. Semin Liver Dis 15:329–339

2. Sarin SK, Kumar A, Chawla YK et al (2007) Noncirrhotic portalfibrosis/idiopathic portal hypertension: APASL recommendationsfor diagnosis and treatment. Hepatol Int 1:398–413

3. Reshamwala PA, Kleiner DE, Heller T (2006) Nodular regener-ative hyperplasia: not all nodules are created equal. Hepatology44:7–14

4. Schiano TD, Kotler DP, Ferran E, Fiel MI (2007) Hepatoportalsclerosis as a cause of noncirrhotic portal hypertension in patientswith HIV. Am J Gastroenterol 102:2536–2540

5. Mallet V, Blanchard P, Verkarre V et al (2007) Nodularregenerative hyperplasia is a new cause of chronic liver diseasein HIV-infected patients. AIDS 21:187–192

6. Maida I, Nunez M, Rios MJ et al (2006) Severe liver diseaseassociated with prolonged exposure to antiretroviral drugs. JAcquir Immune Defic Syndr 42:177–182

7. Maida I, Garcia-Gasco P, Sotgiu G et al (2008) Antiretroviral-associated portal hypertension: a new clinical condition? Prevalence,predictors and outcome. Antivir Ther 13:103–107

8. Saifee S, Joelson D, Braude J et al (2008) Noncirrhotic portalhypertension in patients with human immunodeficiency virus-1infection. Clin Gastroenterol Hepatol 6:1167–1169

9. Chang PE, Miquel R, Blanco JL et al (2009) Idiopathic portalhypertension in patients with HIV infection treated with highlyactive antiretroviral therapy. Am J Gastroenterol 104:1707–1714

10. Kovari H, Ledergerber B, Peter U et al (2009) Association ofnoncirrhotic portal hypertension in HIV-infected persons andantiretroviral therapy with didanosine: a nested case-control study.Clin Infect Dis 49:626–635

11. Geller SA, Dubinsky MC, Poordad FF et al (2004) Early hepaticnodular hyperplasia and submicroscopic fibrosis associated with6-thioguanine therapy in inflammatory bowel disease. Am J SurgPathol 28:1204–1211

12. De Bruyne R, Portmann B, Samyn M et al (2006) Chronic liverdisease related to 6-thioguanine in children with acute lymphoblasticleukaemia. J Hepatol 44:407–410

13. Katzka DA, Saul SH, Jorkasky D, Sigal H, Reynolds JC, SolowayRD (1986) Azathioprine and hepatic venocclusive disease in renaltransplant patients. Gastroenterology 90:446–454

14. Kochin I, Magid M, Arnon R, Glasscock A, Kerkar N, Miloh T(2010) Variceal bleeding in an adolescent with HIV diagnosedwith hepatoportal sclerosis and nodular regenerative hyperplasia. JPediatr Gastroenterol Nutr 50:340–343

15. Ludwig J, Hashimoto E, Obata H, Baldus WP (1993) Idiopathicportal hypertension; a histopathological study of 26 Japanesecases. Histopathology 22:227–234

16. Hillaire S, Bonte E, Denninger MH et al (2002) Idiopathic non-cirrhotic intrahepatic portal hypertension in the West: a re-evaluationin 28 patients. Gut 51:275–280

17. Okudaira M, Ohbu M, Okuda K (2002) Idiopathic portalhypertension and its pathology. Semin Liver Dis 22:59–72

18. Crum-Cianflone N, Dilay A, Collins G et al (2009) Nonalcoholicfatty liver disease among HIV-infected persons. J Acquir ImmuneDefic Syndr 50:464–473

19. Geramizadeh B, Malek-Hosseini SA, Salahi H, Bahador A,Nikeghbalian S (2008) Liver failure and the need for transplantationin three patients with hepatoportal sclerosis. Transplant Proc40:3526–3528

20. Isabel Fiel M, Thung SN, Hytiroglou P, Emre S, Schiano TD(2007) Liver failure and need for liver transplantation in patientswith advanced hepatoportal sclerosis. Am J Surg Pathol 31:607–614

21. Krasinskas AM, Eghtesad B, Kamath PS, Demetris AJ, AbrahamSC (2005) Liver transplantation for severe intrahepatic noncirrhoticportal hypertension. Liver Transpl 11:627–634

22. Gonzalez SA, Talal AH (2003) Hepatitis C virus in humanimmunodeficiency virus-infected individuals: an emerging comor-bidity with significant implications. Semin Liver Dis 23:149–166

23. Marine-Barjoan E, Saint-Paul MC, Pradier C et al (2004) Impactof antiretroviral treatment on progression of hepatic fibrosis inHIV/hepatitis C virus co-infected patients. AIDS 18:2163–2170

24. Brau N, Salvatore M, Rios-Bedoya CF et al (2006) Slowerfibrosis progression in HIV/HCV-coinfected patients withsuccessful HIV suppression using antiretroviral therapy. J Hepatol44:47–55

25. Mallet VO, Varthaman A, Lasne D et al (2009) Acquired protein Sdeficiency leads to obliterative portal venopathy and to compensatorynodular regenerative hyperplasia in HIV-infected patients. AIDS23:1511–1518

Virchows Arch (2011) 458:231–235 235