aSt Mary's Hospital, London, UK
bImperial College London, UK.
Received 8 February, 2007
Accepted 5 March, 2007
The recent report by Mallet et al.  confirmed nodular regenerative hyperplasia on liver biopsies in seven out of eight HIV patients with unexplained portal hypertension. At this London centre we have also observed a series of patients with HIV infection who have non-cirrhotic portal hypertension (NCPH).
Between 2004 and 2006, six patients with chronic HIV infection were found to have unexplained portal hypertension. The reasons for investigating the portal system included deranged liver function tests, haematemesis and bilateral leg swelling.
All patients had splenomegaly and patent portal veins on ultrasound or computed tomography examination of the abdomen. Five out of six cases had endoscopically confirmed oesophageal varices, and one case had portal hypertensive gastropathy. Screening for causes of cirrhosis (liver antibodies, autoantibodies, hepatitis B and C, ferritin, alpha-feta protein) showed evidence of past hepatitis B infection in five out of six patients and chronic hepatitis C infection (genotype 3a) in one case although histology did not reveal cirrhosis.
In addition we noted abnormal thrombophilic tendency in four out of six cases (protein C deficiency, protein S deficiency and lupus anticoagulant). Unlike Mallet et al. , no cases of portal vein thrombosis were identified.
These six patients were aged between 26 and 48 years. Their CD4 cell counts ranged between 200 and 600 (median 445) at the time of diagnosing NCPH. Five out six had been diagnosed with HIV infection over 10 years previously, and also had triple-class antiretroviral experience. Four out of six patients were from Uganda. Mallet et al.  observed a history of didanosine therapy in their cohort. Similarly, five out of six of our cases had taken didanosine therapy for at least 2 years (range 24–90 months, median 58 months).
Transabdominal liver biopsies were performed in all cases. Histology confirmed nodular regenerative hyperplasia in two cases, features of sinusoidal dilatation and venous outflow obstruction in three cases, and normal liver appearances in one case (see Fig. 1).
Three patients developed severe variceal bleeding despite beta-blocker therapy and one patient died from this cause.
In summary, we reached similar conclusions to Mallet et al. . NCPH is emerging as an important cause of liver disease and morbidity in patients with HIV infection, and there may be many unrecognized less severe cases. Diagnosis and management can be challenging, especially for patients with varices that require anticoagulation for prothrombotic tendencies, which means that early detection and diagnosis is paramount . The mechanism of portal damage remains unclear and may be multifactorial. Possible causes include vaso-occlusive damage, mitochondrial damage from antiretroviral therapies , or HIV infection itself. Our small series suggests that prothrombotic tendency may be an important factor. Larger studies will be required to investigate this subject further.
1. Mallet V, Blanchard P, Verkarre V, Vallet-Pichard A, Fontaine H, Lascoux-Combe C, Pol S. Nodular regenerative hyperplasia is a new cause of chronic liver disease in HIV infected patients. AIDS 2007; 21:187–192.
2. Okuda K. Noncirrhotic portal hypertension versus idiopathic portal hypertension. J Gastroenterol Hepatol 2002; 17(Suppl 3):S204–S213.
3. Abrescia N, D'Abbraccio M, Figoni M, Busto A, Maddaloni A, De Marco M. Hepatotoxicity of antiretroviral drugs. Curr Pharm Des 2005; 11:3697–3710.