Low incidence of liver-related deaths in HIV positive people without HBV and/or HCV co-infection

Muirgen Stack, HIV i-Base

This anlaysis from the D:A:D study highlights that the risk of liver failure in HIV positive patients is almost entirely driven by coinfection with HBV and/or HCV rather than ART toxicity in HIV monoinfection.

Kovari and colleagues from the Data Collection on Adverse Events of Anti-HIV Drugs (D:A:D) study reported a low incidence ART-related toxicity and that mortality was rare in this large prospective cohort study, published as an advance access paper in the 22 October 2012 edition of CID. [1]

Previously, the group have reported that approximately 16% of deaths in D:A:D are related to liver failure. [2] The effect from ART without HBV/HCV co-infection or alcohol use is less well characterised. However, some ARVs have specific liver concerns including didanosine with non-cirrhotic portal hypertension and nevirapine with liver-relarted hypersensitivity reactions. [3]

All D:A:D study participants with negative HCV and HBV status were included in this analysis. Chronic and previous HCV infection and chronic, active and previous HBV infections were exclusion criteria, as well as unknown HBV or HCV status. Severe alcohol use was defined according to the WHO definition as alcohol consumption of >60 grams/day for men and >40 grams/day for women. When any form of liver failure was cited as the cause of death and HCV or HBV status was negative, clinical charts were obtained and retrospectively reviewed by local investigators. This provided more information on the aetiology of the disease along with laboratory and histology samples if they were available.

Of 49,737 participants followed between 1 December 1999 and 1 February 2010, approxmiately 40% (n=19,618) were HCV or HBV positive either at baseline, or during follow-up and 2506 (5%) had unknown HCV or HBV status. About 9% (n= 4703) participants were excluded because they belonged to cohorts not responding to the requested information for the study. Thus, 22,910 (46%) HCV and HBV negative participants were followed for 114,478 patient-years and included in the analyses.

In the final study group, median age was 38 (IQR 32-46 years), and 73.1% were male. Ethnicity was: white 47%, black 7%, other 2% and unknown 43%. Median year of first HIV diagnosis was 1999 (IQR 1994-2004), median duration of cohort follow-up was 4.9 (IQR 2.2-8.3 years) and cumulative ART exposure was 0.9 (IQR 0-3.5 years). Median CD4 cell count was 410 cells/mm3 (IQR 250-595), percentage with previous AIDS diagnosis was 23% and treatment status was; naïve 38%, on ART 57% and undergoing interruption 5%. BMI, smoking status and diabetes mellitus diagnosis was also recorded.

From a total of 1059 (4.6%) of deaths, only 12 (0.05%) were liver-related. Thus the incidence of liver-related deaths in people without HCV or HBV coinfection was 0.1 per 1000 person-years (95% CI: 0.05 to 0.18). Seven of the 12 patients that died had severe alcohol use of the cause of death, along with one with an additional diagnosis of haemochromatosis. Five of the 12 patients died due to ART toxicity, meaning the rate of ART-related death in treatment-experienced individuals was 0.04 (95% CI: 0.01 to 0.1) with 5 events over 1000 person-years. Two of the five patients experienced acute liver failure with lactic acidosis on regimens including didanosine and stavudine. Two others died of non-cirrhotic portal hypertension – both had been exposed to didanosine. The other patient died from fatal liver failure from a hypersensitivity reaction to nevirapine.

1047 patients died from other causes, with 376 (35.9%) from AIDS, 116 (11.1%) from CVD, 149 (14.2%) from non-AIDS malignancies, 315 (30.1%) from other causes and 91 (8.7%) from unknown causes. The only significant difference between patients with liver-related death and patients who died of other causes was median longer exposure to ART at baseline: 5.5 (IQR 4.1-6.6 years) vs 2.8 (IQR 0.1-5.2 years), p = 0.008. The authors speculated that the longer exposure to ART may have included older drugs such as didanosine and stavudine.


Although the small number of endpoints in the study made multivariate analyses unfeasible, e.g. the role of patient characteristics on risk of liver-related and other cause mortality, it is still important to clarify incidence rates.

In this case, it should reassure HIV positive people that, apart from didanosine and stavudine, now both rarely used in richer countries, or the hypersensitivity reaction to nevirapine, ART is not associated with liver-related mortality when HCV and HBV negative.


  1. Kovari H et al. Antiretroviral drug-related liver mortality among HIV-positive persons in the absence of HBV or HCV co-infection. The Data Collection on Adverse Events of Anti-HIV Drugs (D:A:D) Study. Clin Infect Dis (22 October 2012 Advance access)
  2. Weber R et al. Liver-related deaths in persons infected with the human immunodeficiency virus: the D:A:D study. Arch Intern Med (2006) 166:1632-1641.
  3. Young J et al. Noncirrhotic Portal Hypertension and Didanosine: A Re-Analysis. Clin Infect Dis (2011) 52 (1): 154-155.

Links to other websites are current at date of posting but not maintained.