Predictors of liver fibrosis in HIV-infected patients with chronic hepatitis C Pablo Barreiro, Luz Mart

1
Predictors of liver fibrosis in HIV-infected
patients with chronic hepatitis CPablo Barreiro,
Luz Martín-Carbonero, Marina Núñez, Pablo Rivas,
Adolfo Morente, Nuria Simarro, Pablo Labarga,
Juan González-Lahoz and Vincent Soriano.
Department of Infectious Diseases, Hospital
Carlos III. Madrid, Spain.
868
Abstract Background Liver fibrosis is
accelerated in HCV/HIV-coinfected patients. The
reasons for this faster liver disease progression
are unclear, although higher plasma HCV-RNA
levels and distinct HCV genotype distribution in
this population compared to HCV-monoinfected
subjects could play a role. Patients and Methods
Liver fibrosis was assessed using elastometry in
all consecutive HIV-infected patients with
chronic hepatitis C attended at our institution
during the last 12 months. Hepatic stiffness was
measured in KiloPascal (KPa) units and
interpreted according to the Metavir score no or
mild fibrosis (F-F1) when 7.1 KPa and fibrosis
with septa or cirrhosis (F2-F4) when gt7.1
KPa. Results A total of 283 patients (71 males
mean age 42 years-old 94 iv drug users 94 on
antiretrovirals mean CD4 count 554 cells/?l, and
72 with plasma HIV-RNA lt50 copies/mL) were
analysed. Mean ALT was 68 IU/L and mean plasma
HCV-RNA was 5.9 log IU/mL. HCV genotype
distribution was as follows 1 (60), 2 (2), 3
(26) and 4 (12). Overall, 164 (58) patients
scored with advanced liver fibrosis (F2-F4) using
elastometry. In the univariate and multivariate
analyses, respectively, significant OR 95 CI
for F2-F4 stages was found for HCV genotype 3
versus others (1.9 1.1-3.4 and 4.3 1.4-13.3),
older age (1.1 1.03-1.17 and 1.1 1.01-1.25),
and elevated ALT levels (1.02 1.01-1.03 and
1.03 1.01-1.04). Although patients with HCV
genotype 1 had higher mean serum HCV-RNA levels
than those with HCV genotype 3 (6.1 vs 5.7 log
IU/mL p0.01), F2-F4 tended to be more frequent
in patients with HCV genotype 3 than in those
with HCV genotype 1 (69 vs 58
pns). Conclusions HCV genotype 3, older age and
elevated ALT levels are independent predictors of
advanced liver fibrosis in HCV/HIV-coinfected
patients.
Patients and Methods Study population All
patients with HIV/HCV coinfection (reactive HCV
serology and detectable HCV-RNA in plasma)
attending at our institution in a 12-month period
were identified and invited to undergo transient
elastometry. Laboratory parameters (ALT, CD4
count, HIV-RNA and HCV load) were recorded at the
last control there were no lags greater than 4
months between elastometry and laboratory
testing. HCV load was measured using a commercial
real-time PCR assay (Cobas Taqman Roche
Diagnostic Systems, Pleasanton, CA). HCV
genotypes were assessed using a commercial
reverse hybridisation method (InnoLiPA HCV II
Innogenetics, Ghent, Belgium). Assessment of
liver fibrosis using elastometry Liver stiffness
was determined using transient elastometry
(figure 2). Briefly, the hepatic region of the
patients was explored with an ultrasound
transducer that was placed in the right
intercostal spaces. When the echography window
showed the characteristic image of liver tissue,
an elastic shear wave was emitted by the
vibration of the ultrasound probe. The speed of
propagation of this vibration through the liver
parenchyma was calculated according to ultrasound
scanning. On the basis of physical principles,
the stiffer the liver the faster the vibration
would pass through the organ. It has been shown
that the PPV of advanced liver fibrosis (fibrosis
with septa to overt cirrhosis (F2-F4 in the
Metavir score) was 95 when elastometry rendered
liver stiffness values gt7.1 KPa. Statistical
analyses Descriptive values are expressed as
percentages, mean (SD) or median (range).
Given the cross-sectional nature of the study,
the association of multiple variables with the
two distinct groups, patients with minimal
(F0-F1) and advanced liver fibrosis (F2-F4), were
analysed in uni- and multivariate analyses.
Comparisons were made using the Chi square test
for proportions, and parametric or non-parametric
tests, as required, for continuous variables.
Finally, all variables included in the univariate
analysis that had p values 0.5 were considered
for a logistic regression analysis. All data were
recorded and analysed using SPSS (version 11.01)
software package (SPSS Inc, Chicago, IL, USA).
Comparisons between HCV genotypes No significant
differences between patients with HCV genotype 3
versus other genotypes were found when comparing
demographics or main laboratory parameters,
except for a lower proportion of men (57 versus
76, p0.004) and higher mean ALT values (7578
IU/L versus 6546 IU/L plt0.001) (Table 2).
Likewise, no significant differences in plasma
HIV-RNA or CD4 counts were noticed when comparing
patients with distinct HCV genotypes. Although
subjects with HCV genotype 3 tended to have been
exposed more frequently to anti-HCV therapy in
the past than the rest, the difference did not
reach statistical significance (56.3 versus
45.9 pns). However, as expected, sustained
viral clearance had been obtained by a greater
proportion of them compared to others (19.6
versus 7.7 p0.01). The success rate of
elastometry was similar in patients with HCV
genotype 3 compared to the rest (0.880.17 vs
0.870.14, respectively). The proportion of
patients with significant liver stiffness (gt7.1
KPa, Metavir F2-F4) was significantly higher in
the group infected by HCV genotype 3 as compared
with other genotypes (69 versus 54, p0.02).
Overall, 119 (42) patients had median
elastometric values compatible with no or mild
liver fibrosis (lt7.1 KPa) the remaining 164
patients (58) had advanced liver fibrosis (gt7.1
KPa). The accuracy of elastometry in terms of
mean number of valid measures and success rates
was comparable when comparing fibrotic and
non-fibrotic patients. The distribution of
patients according to estimated Metavir scores
for liver fibrosis (mean elastometric values) was
as follows F0-F1 (5.40.95 KPa) in 69 patients
(42), F2 (8.20.6 KPa) in 36 (22 ), F3
(10.60.7 KPa) in 16 (10), and F4 (23.913.3
KPa) in 43 (26).
Mean plasma HIV-RNA and CD4 counts, as well as
the proportion of patients on HAART and time on
antiretroviral therapy did not differ
significantly when comparing patients with F2-F4
and F0-F1 liver fibrosis stages. Likewise, plasma
HCV-RNA and prior experience to IFN-based
therapies did not different either. As previously
mentioned, the proportion of patients with
estimates of advanced liver fibrosis was higher
in subjects with HCV genotype 3 (50 out of 72
69) as compared with other genotypes (114 out of
211 54) p0.02. While no differences in liver
fibrosis were observed with respect to alcohol
consumption, all 6 patients with positive HBsAg
were in the group with advanced liver
fibrosis. The multivariate analysis was performed
including in the model all variables with a p
value 0.5 in the univariate analysis (Table 3).
Then, infection with HCV genotype 3 becomes the
stronger independent predictor of advanced liver
fibrosis (OR 4.3 95 CI, 1.4-13.3), being older
age (OR 1.12 95 CI, 1.01-1.25 and elevated ALT
(OR 1.03 95 CI, 1.01-1.04) the other two
variables independently associated with it.
Introduction Liver fibrosis leads to most
clinical complications in patients with chronic
HCV infection. Several factors, such as male
gender, older age, longer duration of HCV
infection and/or high alcohol consumption, have
classically been associated to more severe liver
damage in patients with chronic hepatitis C. More
recently, HIV coinfection has been demonstrated
to be an strong independent predictor of
accelerated HCV-related liver fibrosis. HCV
genotypes were considered to cause liver fibrosis
with similar frequency however, patients
infected with HCV genotype 1 seem to show higher
HCV-RNA levels both plasma and the liver, with
more rapid progression to end-stage liver disease
than patients carrying other HCV genotypes. On
the other hand, patients infected with HCV
genotype 3 show frequently liver steatosis, which
ultimately accelerates liver fibrosis. Liver
biopsy has been for many years the most reliable
procedure to assess hepatic fibrosis. However,
this tool is not free of complications and may
not always be accurate. Transient elastometry
(FibroScan) is a new non-invasive procedure to
measure the stiffness of the liver tissue. The
accuracy of the method for measuring liver
fibrosis has been successfully compared with that
of liver biopsy (figure 1). The factors
associated with more advanced liver fibrosis
according to elastometry were examined in a
cohort of HIV/HCV coinfected patients.
Discussion We found an association between HCV
genotype 3 and advanced liver fibrosis stages.
This observation has significant clinical
implications, since patients with HCV genotype 3
are amongst the best responders to current
anti-HCV therapy, and therefore reinforces that
treatment should particularly be pursued in them.
HCV genotype 3 infection has been associated with
higher rates of liver steatosis. More recently a
link between hepatic steatosis and inflammation
has been reported, which may explain a faster
progression to liver fibrosis in patients with
chronic hepatitis C due to genotype 3. In the
setting of HIV infection, other factors might as
well contribute to explain the greater fibrogenic
effect of HCV genotype 3. Firstly, the risk of
hepatotoxicity following initiation of
antiretroviral drugs is greater in patients with
HCV genotype 3 than with other genotypes.
Secondly, the use of some antiretrovirals could
favour steatosis of the liver. In agreement with
prior studies, we found that patients carrying
HCV genotype 1 presented higher plasma HCV-RNA
levels than those infected with HCV genotype 3.
In HCV/HIV-coinfected hemophiliacs, this finding
has been associated with more rapid CD4 declines
and faster HIV disease progression . However, in
our knowledge no studies have found a significant
correlation between HCV-RNA levels and the extent
of liver fibrosis. Our results using elastometry
confirm that there is no significant correlation
between HCV load and the extent of liver
fibrosis. ALT elevations are not a good marker of
liver fibrosis. This is particularly true in
HIV-infected patients in whom other factors
besides chronic hepatitis C, such as
antiretroviral drugs, may also contribute to ALT
abnormalities 22. However, the probability of
having F2-F4 estimates using elastometry in our
patients with elevated ALT was as high as 87 for
HCV genotype 3 and 68 for other genotypes.
Conversely, it was below 40 in patients with
normal ALT. The simplicity and indulgence of
transient elastometry as compared with liver
biopsy may probably facilitate in the near future
a better understanding of the role of other
factors involved in the progression of liver
fibrosis in the HCV/HIV-coinfected population.
Results Main characteristics of the study
population A total of 285 HCV/HIV-coinfected
patients were identified. In two cases invalid
elastometric measures were obtained due to severe
obesity. Analyses were carried out on the
remaining 283 patients (Table 1). Mean laboratory
parameters were ALT 6856 IU/L, CD4 count 554287
cells/?l, and viral load 20.9 HIV-RNA log
copies/ml. Most patients (94) were under HAART
and 72 had less than 50 HIV-RNA copies/ml. The
distribution of patients according to the type of
antiretroviral regimen being received at the time
of examination was fairly heterogeneous. Mean HCV
viremia was 5.90.9 HCV-RNA log IU/ml, and the
distribution of genotypes was as follows
genotype 1 in 60, genotype 2 in 2, genotype 3
in 25, genotype 4 in 12 and mixed genotypes (2
and 3) in two patients. Nearly half of the study
population (49) had been exposed to anti-HCV
therapy in the past (37 patients to IFN
monotherapy, 32 to IFNRBV, and 65 to pegylated
IFNRBV). However, only 17 of them had attained
sustained virological response. Given that no
definitive information exists regarding a
possible reversion of liver fibrosis in patients
who have cleared HCV infection with anti-HCV
therapy, we decided to keep this small group of
patients in our analysis.
Predictors of advanced liver fibrosis A total of
164 (58) patients were found to show advanced
liver fibrosis (Metavir F2-F4) according to
elastometry (gt7.1 KPa). Older age was the only
demographic variable which could be associated to
a greater liver stiffness (p0.001). No
differences were found with respect to gender,
BMI or risk behaviour (Table 3). Mean ALT values
were higher in patients with advanced liver
fibrosis as compared with those with F0-F1 (8166
versus 4827 IU/L plt0.001). Likewise, the
proportion of patients with ALT levels above the
upper limit of normality (gt55 IU/L) was higher in
subjects with gt7.1 KPa compared to those with
7.1 KPa (56 versus 29, plt0.001). Of note, the
positive predictive value for significant liver
stiffness (Metavir F2-F4) in subjects with
elevated ALT levels was particularly high for HCV
genotype 3 (87 95 CI, 75-99), while it was
lower (68 95 CI, 59-78) for patients with
other HCV genotypes. These differences persisted
when the subset of patients cured from chronic
hepatitis C with anti-HCV therapy were excluded
from the analysis (data not shown).
Conclusions HCV genotype 3, older age and
elevated ALT levels are independent predictors of
advanced liver fibrosis in HCV/HIV-coinfected
patients. Transient elastometry seems to be a
valid method to study liver fibrosis in HIV/HCV
coinfected patients.
Contact pabarre_at_auna.com
Figure 1. Correlation between FibroScan and liver
biopsy in HIV/HCV coinfected patients De
Ledinghen, et al. JAIDS 200641175
2
Sustained Virological Response Following HCV
Therapy is Associated with Regression of Liver
Fibrosis in HCV/HIV-Coinfected Patients P.
Barreiro, P. Labarga, N. Simarro, M. Núñez, L.
Martín-Carbonero, M. Romero, P. Rivas, J.
García-Samaniego, J. González-Lahoz and V.
Soriano Service of Infectious Diseases. Hospital
Carlos III, Madrid, Spain
Abstract Background Chronic hepatitis C leads to
progressive liver fibrosis, which is accelerated
in HCV/HIV-coinfected patients. Therapy with
interferon (IFN) /- ribavirin (RBV) for 6 to 12
months allows reaching sustained virological
response (SVR) in less than half of coinfected
patients. An improvement in liver fibrosis should
be expected in the subset of patients attaining
SVR. However, this benefit has not been proven in
HCV/HIV-coinfected patients. Material and
Methods All HIV/HCV-coinfected patients who had
completed a full course of HCV therapy with IFN
(or pegylated IFN) /- RBV in the past at our
institution and were seen during the last 12
months were identified. All had elevated liver
enzymes before receiving HCV therapy and some
extent of hepatic fibrosis (F1-F4) in the liver
biopsy. Current liver fibrosis was measured in
all using elastometry by FibroScan. Results A
total of 112 HIV/HCV-coinfected patients were
analysed (76 males, mean age 367 years, 67 on
HAART). HCV genotype distribution was 1 (70), 3
(24) and 4 (6). A total of 44 had SVR while the
remaining 68 were non-responders or relapsers.
The main demographic features were comparable
between both groups. Information for other
variables is recorded in the table. F3-F4
estimates were less frequent in SVR than in
non-SVR (OR 2.6 p0.04). Interestingly, in
patients with SVR the mean lag between the end of
HCV therapy and elastometry assessment was longer
in patients showing F0-F1 as compared with those
with F2-F4 (38 vs 22 months p0.06). Moreover,
all 3 patients cured 10 years earlier were
F0-F1. Conclusions SVR after IFN-based
therapies may lead to regression of HCV-related
liver fibrosis in HIV-coinfected patients.
However, long periods of time seem to be required
to show this benefit.
859
The comparison between patients with F0-F2 and
those with F3-F4 estimates at FibroScan is shown
in Table 2. No differences were found regarding
demographic parameters. Patients with higher
grades of fibrosis presented with higher ALT
levels (83 vs 47 IU/L, p0.001) and lower
platelet counts (169 vs 199 per ?L, p0.01).
While HIV related parameters were comparable, in
patients with F3 vs ltF3, the HCV load was
greater (5.1 vs 3.2 log IU/mL, p0.002) and the
proportion of genotypes 1 was higher (76 vs 54,
p0.03). The proportion of responders was higher
in the group of patients with ltF3 vs F3 (48 vs
10, plt0.001). The lag between the end of
anti-HCV therapy and FibroScan assessment was
comparable between the two groups. Only attaining
SVR after anti-HCV therapy was related with lower
liver fibrosis in the univariate analysis. We
also analyzed the influence of different
variables on the degree of liver stiffness in
patients that had attained SVR and in
non-responders. We found no significant
differences in the group of responders (Table 3)
however, patients with lower grades of liver
fibrosis had a longer duration of sustained
anti-HCV treatment response (40 months) than
those with more fibrosis (18 months). In patients
with persistent HCV viremia, if more advanced
liver fibrosis was detected, ALT values were
greater (91 vs 68 IU/L, p0.06) and platelet
counts were lower (169 vs 196 per ?L, p0.05)
(Table 4). Interestingly, the length of
persistent viremia after the end of anti-HCV
therapy was higher as the degree of liver
stiffness was greater (50 months in F3-F4 and 30
months in F0-F2, p0.02). The lag between
anti-HCV therapy and FibroScan was the only
variable associated with the extent of liver
fibrosis in the multivariate analysis. We
performed a survival analysis to compare, between
responders and non-responders, the detection of
F3-F4 along time after end of anti-HCV therapy
(Figure 2). As shown, an accumulation of advanced
fibrosis was observed in patients with persistent
HCV viremia, while the incidence of this adverse
outcome remained were stable in patients that had
attained SVR. The difference between the curves
obtained was statistically significant (p0.01).
Finally, a correlation analysis was done, in
responders and in non-responders, in order to
estimate the influence of time after anti-HCV
therapy on the progression of liver fibrosis
(Figure 3). In patients with persistent viremia a
direct and significant correlation was observed
between time after therapy and the degree of
liver fibrosis (rho 0.25, p0.03). Conversely,
the longer the time after therapy in patients
that had attained SVR to (peg)IFNRBV, the less
intense the degree of liver stiffness detected by
FibroScan (rho -0.39, p0.02). Discussion We
have evaluated the long-term impact on liver
fibrosis of sustained HCV clearance. Patients
attaining SVR to anti-HCV regimens were compared
with those that had viral rebound after therapy.
The most striking finding was that, responders
presented significantly lower liver stiffness
than those with persistent HCV replication. These
results underscore the importance of pegIFN-RBV
therapy in patients with chronic hepatitis C as
the benefits of long-term viral clearance on
liver fibrosis worth the effort. This affirmation
is probably more valid for HIV/HCV coinfected
patients in whom the progression of liver
fibrosis is accelerated with respect to HCV
monoinfected patients. Conclusion Sustained
virological response to anti-HCV therapy seems to
be associated with regression in liver fibrosis.
Table 1. Main characteristics of the HCV/HIV
co-infected population examined with FibroScan
following a prior course of IFN (or pegylated
IFN) RBV therapy.
Table 3. Factors associated with less liver
stiffness (F0-F2 vs F3-F4) by FibroScan in
HCV/HIV co-infected patients who attained
sustained virological response after a prior
course of IFN (or pegylated IFN) RBV therapy.
Introduction Chronic hepatitis C is a relevant
cause of morbidity and mortality in patients with
HIV infection. Treatment with the combination of
(peg)interferon (IFN) plus ribavirin (RBV) for 6
to 12 months may provide sustained clearance of
HCV from the plasmatic compartment in a
significant number of patients, 30-45 across
studies. Some studies have shown that patients
attaining sustained virological response to
anti-HCV therapy present fewer liver
complications than non-responders in the long
term. There are other reports showing a
regression in liver fibrosis years after the
achievement of HCV sustained clearance. Herein,
we have examined the degree of liver fibrosis
according to transient elastometry in a group of
HIV/HCV coinfected patients that received
treatment with (peg)IFNRBV in order to analyze
the impact of HCV clearance on liver histology.
Results A total of 106 consecutive HIV/HCV
coinfected patients that had received a full
course of (peg)IFNRBV were examined with
FibroScan between September 2004 and December
2005. In three instances elastometry was not
available due to severe obesity, so that the
analyses were performed in 103 patients. Among
these, 69 did not respond to therapy and 34 were
sustained virological responders. The main
characteristics of the study population are
depicted in Table 1. As shown, most of them were
males (67) and mean age was 43 years ALT values
were normal in responders and, on average,
elevated in non-responders (24 vs 80 IU/L,
plt0.001). There were no differences between
responders and non responders regarding
HIV-related parameters (viral load (1.87 log
copies/ml), CD4 count (597 cells/?L), proportion
under HAART (83), or HAART modality).
Non-responders had a mean HCV load of 5.68 log
IU/mL, while all responders presented
undetectable values the distribution of HCV
genotypes was different between these two groups,
there were more genotypes 1 in non-responders (75
vs 37, plt0.001) and more genotypes 3 in
responders (56 vs 17, plt0.001). Importantly, the
lag between the end of (peg)IFN RBV therapy and
FibroScan examination was long and comparable
between the two groups (40 months). Mean liver
stiffness resulted significantly lower in
responders as compared with non-responders (6.6
vs 11.7 KPa, plt0.001) the distribution of
patients according to estimated Metavir scores is
shown in the next figure
Table 4. Factors associated with less liver
stiffness (F0-F2 vs F3-F4) by FibroScan in
HCV/HIV co-infected patients who did not attained
sustained virological response after a prior
course of IFN (or pegylated IFN) RBV therapy.
Table 2. Factors associated with less liver
stiffness (F0-F2 vs F3-F4) by FibroScan in
HCV/HIV co-infected patients who had received a
prior course of IFN (or pegylated IFN) RBV
therapy.
Patients and Methods All patients with HIV/HCV
coinfection that had received a full course of
(peg)IFN RBV therapy were identified and invited
to undergo transient elastometry. Main laboratory
parameters (ALT, CD4 count, HIV-RNA, HCV load,
etc.) were recorded at the last control there
were no lags greater than 4 months between
elastometry and laboratory testing. HCV load was
measured using a commercial real-time PCR assay
(Cobas Taqman Roche Diagnostic Systems,
Pleasanton, CA). HCV genotypes were assessed
using a commercial reverse hybridisation method
(InnoLiPA HCV II Innogenetics, Ghent, Belgium).
Liver stiffness was determined using transient
elastometry (figure 1). Briefly, the hepatic
region of the patients was explored with an
ultrasound transducer that was placed in the
right intercostal spaces. When the echography
window showed the characteristic image of liver
tissue, an elastic shear wave was emitted by the
vibration of the ultrasound probe. The speed of
propagation of this vibration through the liver
parenchyma was calculated according to ultrasound
scanning. On the basis of physical principles,
the stiffer the liver the faster the vibration
would pass through the organ. It has been shown
that the AUROC curve for advanced liver fibrosis
(bridging fibrosis to overt cirrhosis (F3-F4 in
the Metavir score) was 97 when elastometry
rendered liver stiffness values gt9.5 KPa. Given
the cross-sectional nature of the study, the
association of multiple variables with the two
distinct groups, patients with minor/moderate
(F0-F2) and advanced liver fibrosis (F3-F4), were
analysed in uni- and multivariate analyses.
BMI, body mass index PI, protease inhibitors
NAN, non-nucleoside analogs NA, nucleoside
analogs ddX dideoxy-nucleosides FS,
FibroScan OR, odds ratio CI, confidence
interval ns, not significant na, not applicable.
Figure 2
Contact pabarre_at_auna.com
Figure 3