There
is a growing interest in developing alternatives to liver biopsy for the assessment
of the degree of fibrosis in patients with chronic hepatitis C virus (HCV) infection.
This is also true for HIV-HCV coinfected subjects. However, the applicability
of these biochemical systems in HIV-HCV coinfected patients has not been determined.
Two posters
at the 57th Annual Meeting of the American Association for the Study of Liver
Diseases (AASLD) held recently in Boston evaluated the role of several fibrosis
scoring systems based on serum markers in this population.
Comparison
of Non-invasive Methods
French
investigators assessed the correlation between the METAVIR fibrosis scores of
liver biopsies and 7 different biochemical scores in 272 HIV-HCV coinfected patients
who had participated in the Ribavic trial. FibroTest (FT), SHASTA, Fib-4, Hepascore,
Fibrometer, AST to platelet ration index (APRI), and the Forns index were included
in the evaluation of diagnostic accuracy and cost-effectiveness.
Alpha-2-macroglobulin,
apolipprotein A1, haptoglobulin, total bilirubin, gamma-glutamyl-transferase (GGT),
alanine aminotransferase (ALT), aspartate aminotransferase (AST), hyaluronic acid,
urea, total cholesterol, albumin, and platelet counts were the biomarkers measured
for these models.
The
researcher used area under receiver operating characteristics curve (AUROC) measurements
to assess the accuracy of the indexes. Overall, higher AUROCs were obtained the
diagnosis of advanced fibrosis (F4) with all scoring systems.
For
the diagnosis of fibrosis F3 or greater, Fibrometer, Hepascore and Fibrotest had
higher AUROCs than the others. For fibrosis of F2 or greater, Fibrometer and Hepascore
had higher AUROCs than the Forns index. The accuracy in the diagnosis of significant
liver fibrosis (F2 or greater) was significantly higher for Fibrometer (71%),
Hepascore (68%), and Fibrotest (62%) than for Fib-4 (43%), APRI (33%), or Forns
index (22%) (P < 0.05).
For
all systems, the positive predictive value (80%-88%) was superior to the negative
predictive value (32%-44%). A cost-performance analysis showed that Hepascore
had the best cost-performance ratio.
The
authors concluded that, "among noninvasive liver fibrosis biomarkers in HIV-HCV
coinfected patients, Hepascore, Fibrometer, and Fibrotest can always be applied
and can accurately classify patients with severe fibrosis". However, the
ability of these tests to reliably diagnose a low degree of fibrosis seems to
be limited.
APRI
In
like manner, APRI has not yet been validated in the HIV-HCV coinfected population.
A group from Seattle evaluated the performance of APRI in 111 patients coinfected
with HCV and HIV who had undergone liver biopsy between 2000 and 2005 period.
As a control group, 111 HCV monoinfected patients matched for age, sex and liver
disease stage were randomly selected.
Fibrosis
was staged using the Batts-Ludwing system (F0-F4). The APRI cut-off values of
0.5 and 1.5 were used to predict the absence or presence of significant fibrosis
(fibrosis scores 3 or 4), respectively. AUROCs were compared between monoinfected
and coinfected patients.
The
characteristics of the coinfected and monoinfected groups were comparable, except
for ALT levels, which surprisingly were significantly lower in the coinfected
group (mean 92 IU/L vs.124 IU/L; P = 0.02). In each group, 24 patients (22%) had
histological fibrosis stages 3 or 4. There were significantly more patients with
advanced fibrosis among the 46 coinfected patients who had CD4 counts below 300
cells/mm3 (F4 = 13.1%) than among the 111 HCV monoinfected patients (F4 = 6.3%)
(P = 0.048).
As
shown in the table below, the AUROC was significantly lower in HIV-HCV coinfected
patients with CD4 counts below 300 cells/mm3 than in the HCV monoinfected patients
(P = 0.048). As a result, only 30% of patients with low CD4 counts were correctly
classified by APRI, compared with 48% of HCV monoinfected patients and coinfected
individuals with higher CD4 counts.
AUROC
(95% CI)
Correctly Classified (%)
HCV monoinfected
.88 (.81-.96)
53/111 (48)
HIV-HCV coinfected
.79 (.70 -.88)
45/111 (41)
CD4 < 300 cells/mm3
.72 (.57-.87)
14/46 (30)
CD4 ≥ 300 cells/mm3
.82 (.70 -.95)
31/65 (48)
The authors explored
the reasons for the lower accuracy of this index in the subset of patients with
lower CD4 counts. They found that AST levels were not associated with fibrosis
stage in these patients. According to the authors, this did not appear to be driven
by other mechanism for increased AST, such as antiretroviral hepatotoxicity.
Contrary
to previous reports, the researchers concluded, this study showed that, "APRI
did not perform as well in the HIV-HCV coinfected patients with low CD4 counts.
However, its accuracy in those with higher CD4 counts approached that in the HCV
[monoinfected] group. While the model was significantly less accurate in coinfected
patients with low CD4 counts, the mechanisms behind this difference remain to
be elucidated."
This
limitation adds to the low sensitivity and specificity of APRI for the diagnosis
of significant fibrosis in patients with HCV and alcoholic liver disease [4].
Given the frequent history of alcohol use in HCV/HIV coinfected patients, APRI
may be of limited usefulness in the diagnosis of fibrosis in many patients.
In
conclusion, although indexes based on serum markers may be of help in the assessment
of liver fibrosis in some HCV/HIV-coinfected patients, they need to be better
refined. Their limitations need to be taken into account and may be better used
in conjunction with other tools such as transient elastography.
11/10/06
References
P
Cacoub, F Carrat, P Bedossa, and others. Independent assessment and cost-performance
analysis of noninvasive liver fibrosis biomarkers in HIV-HCV coinfected patients:
the Fibrovic Study-ANRS HC02. 57th Annual Meeting of the American Association
for the Study of Liver Diseases (AASLD). October 27-31, 2006. Boston, MA. Abstract
215.
A Singal,
L V Thomassen, D R Gretch, and others. The AST to platelet ratio index does not
accurately predict significant fibrosis in HCV/HIV coinfected patients with low
CD4 T cell counts. 57th AASLD. Abstract 226.
Nunes
D, and others. HIV infection does not affect the performance of noninvasive markers
of fibrosis for the diagnosis of hepatitis C virus-related liver disease. Journal
of Acquired Immune Deficiency Syndromes 40: 538-544. 2005.
Lieber
C, and others. Aspartate aminotransferase to platelet ratio index in patients
with alcoholic liver fibrosis. American Journal of Gastroenterology 101:
1500-1508. 2006.
