While ARV can be toxic and indeed be associated with mortality, the
risk is not as high as NOT treating. Indeed, as the study found,
before COMBINATIONS of ARV were available, liver related deaths (LRD)
were much higher: "A total of 184 (1.7%) died from LRD during 52 236
person-years of follow-up (PYFU). The death rate from LRD declined
from 6.9 per 1000 PYFU before 1995 [95% confidence interval (CI),
3.9-9.9] to 2.6 at/after 2004 (95% CI, 1.6-4.0)."

Still, methods and means to protect the liver and offset damage are
needed. Agents like NAC, alpha lipoic acid, whey proteins, milk
thistle and others may help to further reduce the risks of ARV-related
hepatotoxicity.

        George M. Carter

***
NATAP http://natap.org/
_______________________________________________

European HIV study finds rise in liver-related deaths

The study authors concluded: “….we found a siignificant increase
over time across Europe in the death rates from LRD in patients with
similar CD4 cell counts. Combination antiretroviral therapy reduced
the death rate from LRD, apparently by increasing the CD4 cell count.
In patients who had started cART and with similar CD4 cell counts,
longer exposure was associated with an increased death rate from LRD.
This may be due to progression of liver disease due to hepatitis B or
C over time as patients survive longer, worsening of chronic hepatitis
due to direct liver toxicity of antiretrovirals, or some other factor.
cART continues to have a major benefit for reducing the risk of AIDS
or HIV-related deaths, and in reducing liver-related deaths, and the
increased death rate from LRD with longer exposure to combination
therapy should be balanced against its many benefits.�

By Anne Harding

NEW YORK (Reuters Health) - Deaths from liver-related disease (LRD)
among patients with HIV and similar CD4 cell counts have increased
over time, but the reasons behind the rise remain unclear, according
to a new report.

"Hopeful this study will stimulate additional reports so we can start
to get a picture together on what's actually going on," Dr. Jens D.
Lundgren of the Copenhagen HIV Program in Hvidovre, Denmark told
Reuters Health.

Possibilities include exposure to combination antiretroviral therapy
(cART), the fact that patients are living longer with hepatitis B or
hepatitis C infections, or even increased use of alcohol, he noted.

Dr. Lundgren and his colleagues in the EuroSIDA Study Group conducted
the current study to investigate reports of increases in liver-related
disease among patients with HIV, and to determine if length of
exposure to cART had any effect on liver-related disease.

The researchers looked at 10,937 patients participating in EuroSIDA,
1.7% of whom died from liver-related disease during follow-up. Their
initial analysis found that liver-related disease declined from 6.9
per 1,000 person-years of follow-up (PYFU) before 1995 to 2.6 per
1,000 PYFU at or after 2004.

The rate of liver-related disease was strongly associated with CD4
cell count, with patients with current counts below 50 cells per
microliter having a liver-related disease death rate of 14.1 per 1,000
PYFU, compared to 7.0 per 1,000 PYFU for patients with counts above
50.

Once the data was adjusted for CD4 cell count, the researchers found
death rates from liver-related disease actually increased by 13% per
year. Among patients on cART, deaths from liver-related causes
increased by 12% for every year of exposure to the drugs. Patients
positive for HbsAg had nearly triple the risk of death from
liver-related disease, while anti-HCV increased liver-related disease
risk more than five-fold.

Based on the findings, Dr. Lundgren said, "we need to understand that
liver-related death is a problem in HIV and we need to target the
modifiable risk factors that influence the risk of liver related
death. They include ongoing hepatitis B infection, ongoing hepatitis C
infection, and alcohol."

Consulting with patients on alcohol use is particularly important for
HIV patients with hepatitis C infection, he added. "We know that, for
example, in hepatitis C, if you drink, the alcohol actually works
synergistically with the infection to accelerate fibrosis of the
liver."

He and his colleagues conclude: "cART continues to have a major
benefit for reducing the risk of AIDS or HIV-related deaths, and in
reducing liver-related deaths, and the increased death rate from
liver-related disease with longer exposure to combination therapy
should be balanced against its many benefits."

AIDS 2005;19:2117-2125.

**********
Is there evidence for an increase in the death rate from liver-related
disease in patients with HIV?

AIDS:  Volume 19(18)  2 December 2005  p 2117-2125

Mocroft, Amandaa; Soriano, Vincentb; Rockstroh, Jurgenc; Reiss,
Peterd; Kirk, Olee; de Wit, Stephanef; Gatell, Joseg; Clotet,
Bonaventurah; Phillips, Andrew Na; Lundgren, Jens De; for the EuroSIDA
Study Group

From the aRoyal Free Centre for HIV Medicine and Dept Primary Care and
Population Sciences, Royal Free and University College Medical School,
London, UK

Abstract
Background: Increases in deaths due to liver-related disease (LRD)
among HIV-infected individuals have been reported although the
influence of combination antiretroviral therapy (cART) on LRD is
controversial.

Aims: To determine changes over time in the death rate from LRD and if
longer exposure to cART was associated with an increased death rate
from LRD in 10 937 patients from EuroSIDA, an observational
longitudinal cohort study.

Results: A total of 184 (1.7%) died from LRD during 52 236
person-years of follow-up (PYFU). The death rate from LRD declined
from 6.9 per 1000 PYFU before 1995 [95% confidence interval (CI),
3.9-9.9] to 2.6 at/after 2004 (95% CI, 1.6-4.0). When the current CD4
cell count and other factors were taken into account, there was a 13%
increase in the death rate from LRD per year (95% CI, 5-20%, P =
0.0008). In patients who had started cART, there was a 12% increase in
the death rate from LRD per additional year exposure to cART (95% CI,
4-20%, P = 0.022) after adjustment for current CD4 cell count and
other factors.

Conclusions: Death rates from LRD appeared to decrease across Europe.
However after adjustment for the current CD4 cell count, and therefore
increases in CD4 cell counts in patients taking cART, there was a
significant increase over time in death rates from LRD. In patients
with similar CD4 cell counts, longer exposure to cART was associated
with an increased death rate from LRD. This may be due to direct liver
toxicity of antiretrovirals, progression of liver disease due to
hepatitis B virus or hepatitis C virus over time as patients survive
longer, or some other factor.

Introduction
Combination antiretroviral therapy [cART; also called highly active
antiretroviral therapy (HAART)] has significantly reduced the impact
of AIDS and death in patients with HIV [1-3]. HIV has become a
long-term chronic disease and the occurrence of fewer deaths from
HIV/AIDS has led to an increase in deaths due to other causes [4-9].
As the impact of HIV disease is reduced, co-infection with other
viruses, such as hepatitis C virus (HCV) and hepatitis B virus (HBV)
have emerged as more important causes of morbidity and mortality, with
several studies reporting an increase in the proportion of deaths
attributable to liver-related disease (LRD) [10-13]. However, this
increase in the proportion of deaths due to LRD does not necessarily
mean that the death rate (or incidence) from LRD has similarly
increased, or that patients with HIV are more likely to die from LRD
since the introduction of cART. There have been few studies looking at
death rates due to LRD [8,14], with conflicting results. These studies
focused on all causes of death and deaths from LRD were included in a
sub-analysis.

Death rates due to LRD among HIV-infected individuals might increase
with longer exposure to cART for several reasons. Progression to liver
cirrhosis in patients co-infected with HBV or HCV takes on average 20
years [15,16]. The prevention of deaths from opportunistic illnesses
in patients treated with cART increases the duration of exposure to
HBV or HCV infection. In addition, cART can lead to hepatotoxicity,
causing treatment interruption, liver decompensation and death
[17-20]. The risk of liver-related toxicities is considerably higher
in patients co-infected with HBV and/or HCV [18,21]. There are
currently no data considering whether longer exposure to cART might
increase the death rate from LRD.

The aims of this study were therefore two-fold. First to describe
changes over time in the death rates from LRD and second, to determine
if there was any change in the death rate from LRD with longer
exposure to cART.

Patients
EuroSIDA is a prospective, European study of 11 229 patients with
HIV-1 infection in 80 centres across Europe (including Israel and
Argentina as non-European representatives; see Appendix). Details of
the study have been published [22]. At recruitment, in addition to
demographic and clinical information, a complete antiretroviral
history was collected, together with the eight most recent CD4 cell
counts and viral load measurements. Patients are seen within their
clinics as required, but at 6-monthly intervals, relevant data is
extracted from patient clinical charts onto follow-up forms. This
analysis includes data up to winter 2004 and includes details on all
CD4 lymphocyte counts and viral loads measured since last data
collection, the date of starting and stopping each antiretroviral drug
and the use of drugs for prophylaxis against opportunistic infections.
Data on HBV and HCV antibody status was recorded in 1997 at the
inception of Cohort III and for patients still under follow-up from
Cohorts I and II, similarly at the recruitment of Cohort IV and
annually thereafter. HCV and HBV serological markers were assessed
using commercial enzyme-linked immunosorbent assays. Patients were
classified as HCV positive on the basis of a reactive antibody test
and HBV positive when serum HBsAg was present. Serum HCV-RNA and/or
HBV-DNA was not routinely performed. Cause of death has been routinely
recorded in EuroSIDA since the outset of the study (see page 8 of the
form at http://www.cphiv.dk/pdf_folder/eurosida_enrol_2004.pdf for
details). For the purposes of this analysis, any death where the cause
was recorded as liver failure (unspecified, hepatitis or non-hepatitis
related) were classified as liver related. cART was defined as a
minimum of three drugs, of which at least one had to be a protease
inhibitor (PI), non-nucleoside reverse transcriptase inhibitor
(NNRTI), abacavir or tenofovir.

Members of the co-ordinating office visited all centres to ensure
correct patient selection and that accurate data was provided by
checking the information provided against case notes for all reported
clinical events and a random sample of 10% of all other patients.

Results
A total of 10 937 patients satisfied the inclusion criteria and are
described in Table 1, of whom 184 (1.7%) died from LRD during 52 236
PYFU. Patients who died from LRD tended to be recruited to EuroSIDA,
on average, 12 months before those without a LRD (P < 0.0001). They
also had lower CD4 cell counts (medians of 162 versus 254 cells/μl, P
< 0.0001) and were less likely to be taking cART at recruitment (22.8
versus 36.7%, P < 0.0001). In addition, patients who died from LRD
were more likely to be either HBsAg-positive (13.6 versus 5.2%, P <
0.0001) or HCV-antibody positive at recruitment (45.7 versus 16.7%, P
< 0.0001). At recruitment to the study, HBV or HCV status was unknown
for 4125 (37.7%) and 5117 (46.8%) patients, respectively.

The death rate from LRD was 3.5 per 1000 PYFU [95% confidence interval
(CI), 3.0-4.0]. Figure 1 shows the change in the death rate from LRD
over calendar time. There was an estimated 7% decline in the death
rate from LRD per year (95% CI, 1-12%, P = 0.015). However, other
factors are related to deaths from LRD which may also have changed
over time and which might explain the apparent decrease seen in Fig.
1. Table 2 illustrates the univariate and multivariate incidence rate
ratios (IRR) of deaths due to LRD. Of note, age was not related to the
death rate from LRD (P = 0.26), nor was treatment for HCV (P = 0.66),
for HBV using either tenofovir (P = 0.92, used by 175 patients before
recruitment) or lamivudine (P = 0.093, used by 4816 patients before
recruitment). There was no evidence for a change over time in the
death rate from LRD (IRR, 0.98; 95% CI, 0.92-1.04; P = 0.49) after
adjusting for the factors shown in Table 2. Of note, patients with HBV
(IRR, 2.83; 95% CI, 1.81-4.41; P < 0.0001) or HCV (IRR, 4.24; 95% CI,
2.38-7.57; P < 0.0001) had an increased death rate from LRD.

There was a strong relationship between current CD4 cell count and
death rates from LRD. Patients with a current CD4 cell count < 50
cells/μl had a death rate from LRD of 14.1 per 1000 PYFU (95% CI,
10.3-18.0), compared with 7.0 (95% CI, 5.6-9.0) in patients whose
current CD4 cell count was between 51 and 200 cells/μl and 1.6 (95%
CI, 1.2-2.0) in patients whose current CD4 cell count was > 200
cells/μl. This relationship was highly consistent in each of the
calendar periods 1994-1996, 1997-1999 and 2000 or later. Within the
lower CD4 cell count strata (< 50 cells/μl and ≤ 200 cells/μl),
there was an increase in death rates from LRD over time. For example,
in patients with a current CD4 cell count of < 50 cells/μl, the death
rate from LRD increased from 11.2 per 1000 PYFU (95% CI, 6.4-16.0)
during 1994-1996 to 22.6 per 1000 PYFU (95% CI, 12.7-32.5) at/after
2000.

Clearly the current CD4 cell count was related to the death rate from
LRD and will have changed over time due to the introduction of cART.
This might explain the changes in the death rate from LRD. Patients
with unknown HBV or HCV status at baseline who were subsequently
tested can also be included. Table 3 shows the results from a
multivariate Poisson regression model after adjusting for current CD4
cell count and current HBV and HCV status. AIDS diagnosis, starting
cART, and using lamivudine (used by 9215 patients) were also included
as time-updated variables. There was, on average, a 13% increase in
the death rate from LRD per year (95% CI, 5-20%; P = 0.0008). Patients
with HBsAg had an increased death rate from LRD (IRR, 2.82; 95% CI,
1.88-4.25; P < 0.0001), as did patients with anti-HCV (IRR, 5.12; 95%
CI, 3.01-8.73; P < 0.0001) in comparison with patients without
hepatitis co-infections. The increase in the death rate from LRD was
consistent in patients with a current CD4 cell count of < 200
cells/μl (IRR, 1.15; 95% CI, 1.06-1.24; P = 0.0010) or in patients
with a current CD4 cell count > 200 cells/μl (IRR, 1.15; 95% CI,
1.03-1.29; P = 0.016) in models adjusting for the same variables.
Starting tenofovir was not associated with a significantly reduced
incidence of deaths from LRD (P = 0.63; 871 patients started
tenofovir).

By death, or last follow-up, HBV status remained unknown for 2583
patients (23.6%) and HCV status was unknown for 3063 patients (28.0%).
Treatment with interferon and/or ribavarin was uncommon; 292 patients
started interferon (4.1%), of whom 12 died (4.1%), whereas 181
patients started ribavarin (2.2%), of whom 4 (2.6%) died. Of 184
deaths from LRD, 30 (16.3%) occurred in patients co-infected with HBV
and 101 (54.9%) in patients co-infected with HCV. It is unknown if
patients died from a liver-related death directly attributable to HBV
or HCV infection, as such data is not currently available. Overall, 88
patients (47.8%) died from LRD without ever starting cART and 24
patients (13.0%) were antiretroviral naive at the time of death.
Patients who died from a LRD after starting cART had significantly
higher CD4 cell counts immediately prior to death (medians 134 versus
102 cells/μl; P = 0.039) and lower viral loads (medians 3.05 versus
4.45 log10copies/ml; P = 0.026).

Patients who started cART had a significantly lower death rate from
LRD (2.8 per 1000 PYFU; 95% CI, 2.2-3.3) in comparison with patients
who had not started cART (6.2 per 1000 PYFU; 95% CI, 4.9-7.6; P <
0.0001). This difference was largely explained by the increases in CD4
cell count after starting cART. Thus after adjustment for the
variables known to be related to deaths from LRD (from Table 3) and
the current CD4 cell count there was no difference in the death rates
from LRD (IRR, 0.69; 95% CI, 0.44-1.09, P = 0.11). It was not possible
to stratify this analysis by the type of cART regimen (e.g., NNRTI
versus PI). Figure 2 illustrates the unadjusted and adjusted (for
exposure group, current HBV and HCV status, current CD4 cell count and
diagnosis of AIDS as time-updated covariates) death rates from LRD,
using never taken cART as the reference group. After adjustment, it
was clear that patients taking cART for < 2 years have a lower death
rate from LRD in comparison with patients who have never taken cART
(IRR, 0.49; 95% CI, 0.29-0.83; P = 0.0082), but thereafter the rate
was increasing. Of note, among patients who had started cART, there
was a 12% increase in the death rate from LRD with each additional
year exposure to cART (95% CI, 1.04-1.20; P = 0.022) after adjustment
for similar variables.

Table 3. Multivariate incidence rate ratios (IRR) of liver related
death.

Table 2. Univariate and multivariate incidence rate ratios (IRR) of
deaths due to liver-related disease.

Discussion
Although death rates from LRD in almost 11 000 patients with HIV
appeared to decrease across Europe, this was explained by increases in
CD4 cell count after starting cART. After taking such increases into
account, there has been a significant increase over time in the death
rate from LRD. This means that of two patients of similar HCV or HBV
status, exposure group and current CD4 cell count, the patient under
follow-up in 2004 had a higher risk of death due to LRD in comparison
with a patient under follow up in 2000. In addition, there was an
increase in the death rate from LRD with longer exposure to cART in
patients who had started cART. This may be due to longer survival in
co-infected patients or to prolonged treatment with potentially
hepatotoxic drugs.

Before adjustment, we found a small decrease in the death rates from
LRD. Such a decrease was explained by differences in patient
characteristics at baseline, and after adjustment for these, there was
no significant change over time in the death rate from LRD. In a
previous report from EuroSIDA, there was a small decrease in the death
rates from LRD, but this analysis was not adjusted [8]. In a smaller
study from Spain, also unadjusted, there was a significant increase in
the death rates from LRD [14]. The risk of death due to LRD decreases
with increasing CD4 cell counts [24-26]. cART-induced changes in the
immune system may reduce the risk of liver fibrosis and cirrhosis, as
HIV may accelerate progression of HCV because of its associated
immunodeficiency [20,27,28]. In addition, improved immunity will also
decrease the risk of opportunistic infections and malignancies which
may involve the liver. After adjustment for these increases in CD4
cell count, we found an increase in the death rates from LRD over time
and the increase was consistent in patients with a current CD4 cell
count above or below 200 cells/μl. There are several explanations for
these findings. The decline in death rates from AIDS-defining
illnesses associated with cART may have led to an increase in other
underlying comorbidities, such as liver disease. It is also possible
that hepatotoxicity has played a role. Four main mechanisms of
drug-related liver toxicity that could account for a deleterious
effect of antiretroviral therapy on liver function in HIV-infected
patients have been recognized in HIV-infected individuals exposed to
antiretroviral therapy: (1) direct drug toxicity; (2) immune
reconstitution following initiation of cART in the presence of HCV
and/or HBV or other opportunistic infections involving the liver
[20,29-31]); (3) hypersensitivity reactions with liver involvement;
and (4) mitochondrial toxicity.

Patients co-infected with either HBV or HCV had an increased death
rate from LRD, as reported previously by others [32-35]. This could be
explained by a worsened natural history of both HBV and HCV in
patients co-infected with HIV and a faster progression to end-stage
liver disease [36-38]. Despite the increased death rate from LRD in
patients with HBV or HCV, it was reassuring that deaths from LRD only
accounted for a small proportion of the total deaths that occur in
EuroSIDA [22]. Surprisingly, the death rate from LRD was not increased
in older patients, and this variable was not included in multivariate
models. Identical results were found in a multivariate model which
included age (data not shown). Bonacini et al. [39] also found no
relationship between age and progression of liver disease, whereas
other studies found age to play a significant role [25,38,40,41]).
Apart from differences in the patient populations studied or the
statistical analyses, the reasons for these conflicting results are
not clear.

We found that patients starting cART had a lower death rate due to
LRD. As reported by others [25,26,42], this difference remained
significant after adjustment for differences in patient
characteristics at recruitment to EuroSIDA (data not shown). However,
after adjustment for latest CD4 cell count this association was no
longer significant and there was no difference in the rate of deaths
due to LRD for patients taking cART compared to patients not taking
cART, suggesting that cART reduced deaths due to LRD by increasing CD4
cell count. However, the death rate from LRD increased by 13% per year
of additional exposure to cART. For example, in two patients with
similar current CD4 cell count, HBV or HCV status and exposure group,
a patient taking cART for 2 years would have a 13% higher death rate
from LRD in comparison with a patient who had taken cART for 1 year.
This is in contrast to Mehta et al. [41], who showed that cumulative
exposure to cART was not associated with serious histological liver
disease. Further, Tural et al. [43] showed that the time on
antiretroviral therapy decreased the odds ratio of liver fibrosis in a
population of patients who underwent liver biopsy. In a separate
study, Tural et al. showed that liver fibrosis was more severe in
patients co-infected with HCV and HIV, even when taking cART, in
comparison with patients with HCV alone [44]. The statistical
methodology used to analyse the data presented here was different to
previous analyses as it used cumulative exposure to cART and allowed
the estimation of the death rate from LRD according to exposure to
cART. In addition to the larger size of EuroSIDA and the use of
different methodology, it is possible that the duration of cART may
have been too brief in previous analyses to demonstrate this effect.
Other differences in the study populations, such as alcohol use or
active intravenous drug use, may also play a role.

There are several limitations of our study which should be taken into
account when interpreting our results. HBV or HCV status was not
available for all patients in EuroSIDA. Data on HCV or HBV viraemia
was not available to determine if the disease was active. It is
possible that the proportion of patients with active disease has
changed over time due to increased survival. Our finding of an
increased rate of deaths due to LRD over time was however independent
of HCV or HBV status, and it is unlikely information about unknown HBV
or HCV status or HCV viraemia would substantially alter these
findings. We found similar results in analyses limited to those with
known HBV or HCV status, or among IDUs. We did not adjust for HIV
viral load as this data was missing from a number of patients, and
particularly before 1997 when this test was introduced as part of
routine clinical care. We have no data on current illicit drug use,
alcohol use or adherence to cART, all of which are potentially related
to deaths due to LRD. We have routinely collected data on cause of
death since the study started in 1994, and a rigorous data quality
assurance system is in place. However, in recognition of the
increasing complexity involved with determining the cause of death in
patients with HIV, EuroSIDA is an active participant in the CoDe
project (details at www.cphiv.dk/ ) which aims to produce a uniform
coding system for cause of death that can be applied to studies of
individuals with HIV. This may also allow an improved classification
of the actual pathogenesis which underlies any death. For example, it
will allow us to distinguish between deaths from LRD resulting from
treatment toxicity, hepatitis-associated end-stage liver disease or a
HIV-associated opportunistic infection which involves the liver.

To conclude, we found a significant increase over time across Europe
in the death rates from LRD in patients with similar CD4 cell counts.
Combination antiretroviral therapy reduced the death rate from LRD,
apparently by increasing the CD4 cell count. In patients who had
started cART and with similar CD4 cell counts, longer exposure was
associated with an increased death rate from LRD. This may be due to
progression of liver disease due to hepatitis B or C over time as
patients survive longer, worsening of chronic hepatitis due to direct
liver toxicity of antiretrovirals, or some other factor. cART
continues to have a major benefit for reducing the risk of AIDS or
HIV-related deaths, and in reducing liver-related deaths, and the
increased death rate from LRD with longer exposure to combination
therapy should be balanced against its many benefits.