Saturday Poster Sessions, October 28, 2006
HCV: Diagnosis and Natural History
J. H. Ishida; C. Jin; P.
Bacchetti; V. Tan; M. G. Peters; N. A. Terrault
Background:
Complications of hepatitis C virus (HCV) infection are
primarily related to the development of advanced fibrosis. Risk factors for
fibrosis include male gender, older age at HCV infection and heavy alcohol use,
but the role of cannabis remains controversial.
Objective:
To study the association between cannabis use and fibrosis
severity in persons with chronic HCV infection.
Methods:
Between 2001 and 2004, 328 HCV-infected subjects enrolled
from university and community sources underwent in-person interviews (to assess
demographics, risk factors for HCV, and use of cannabis and alcohol), virologic
testing and liver biopsy. Biopsies were scored for fibrosis using the Ishak
method (scale F0 to F6) by a single pathologist blinded to clinical data.
Biopsy adequacy was assessed by length and number of portal tracts. 204 completed all the baseline
requirements. Included and excluded
subjects were not significantly different except that the former used cannabis
more frequently than the later group.
Results:
The median age of the cohort was 46.8 years, 69% were male,
49% were Caucasian, 61% earned ≤$15,000 per year, and the presumed route
of infection was intravenous drug use in 70%.
Ninety-seven percent had at least one drink witin year prior to
enrolment. Current (within 12 months of
enrollment) cannabis daily use frequency was daily (14%), used cannabis within
the year prior to enrolment (20%). Fibrosis stage median was 1 (scale 0-6),
necoinflammation median was 5.
Daily cannabis use was associated with moderate to severe
fibrosis compared to mild fibrosis in univariate analysis. Gender, race, enrollment age, HCV duration
and infection source, HCV genotype, HIV, daily tobacco use, and lifetime
alcohol use were not significantly associated with mild fibrosis or moderate to
severe fibrosis stage in univariate and multivariate analysis.
Conclusion:
·
Current
daily cannabis use independently increased the odds of moderate to severe
fibrosis by nearly 7-fold in persons with chronic HCV infection.
o
The
duration of lifetime moderate to heavy alcohol use was also independent
predictor of moderate to severe fibrosis
·
There
was no association between current daily cannabis use and mild fibrosis.
o
HCV
viral load was an independent predictor of mild fibrosis.
·
The
number of portal tracts was associated with both mild and moderate to severe
fibrosis, highlighting the importance of controlling for biopsy adequacy.
·
This
cross-sectional design limits the ability to establish a temporal relationship
between cannabis use and fibrosis stage.
·
Quantity,
duration or method of cannabis use were not assessed in detail.
·
Our
results indicate that HCV-infected individuals should be counseled to reduce or
abstain from cannabis use.
Future Directions:
·
Confirm
differential effect of cannabis use by severity of disease.
·
Prospective cohort study of cannabis
versus non-cannabis users with chronic HCV using more detailed cannabis
assessments.
P. Halfon; Y. Bacq; A. De
Muret; G. Pénaranda; M. Bourlière; D. Ouzan; A. Tran; D. Botta-Fridlund; C.
Renou; M. Brechot; C. Degott; V. Paradis
Introduction:
Several blood scores of liver fibrosis have been published.
The most validated scores are the Fibrotest (FT) and APRI. However, possibly
more accurate tests have been recently published: Fibrometer (FM) and Hepascore
(HS). The aim of this study was to compare the diagnostic accuracy of those 4
tests in independent conditions and populations.
Methods:
356 patients with chronic viral hepatitis C from 2 different
French areas with a liver specimen length ≥ 15 mm (22 ± 7 mm) were
included. Liver specimens were evaluated by 2 independent pathologists
according to the Metavir staging before reaching a consensus. Blood variables
were determined on fresh or frozen blood samples.
Results:
The distribution of Metavir stages according to liver
specimens was: F0: 4%, F1: 55%, F2: 26%, F3: 11%, F4: 4%. The AUROC are listed
as a function of blood score and fibrosis cut-off in the Table 1.
The AUROC of those 4 blood scores were not significantly
different in that population characterized by a high proportion of F1 and a low
proportion of F4. On the other hand, those scores had different patterns.
Therefore, we evaluated the rate of misclassified patients (%) as a function of
F stages or cut-off, FT vs FM respectively in the following Table 2.
The comparison of misclassification (blood vs liver) rates
indicated significantly different patterns, FT vs FM respectively: F0+1: 18 vs
28 % (p=0.002), ≥F2: 43 vs 31 % (p=0.005), ≥ F3: 28 vs 12 %
(p=0.065), F4: 15 vs 0 % (p: NA).
Conclusions:
These results show that the accuracy of those tests is
sensitive to the fibrosis stage prevalence which depends on the population
studied. For the diagnosis of clinically significant fibrosis (F≥2),
those 4 blood scores had a similar performance in this population that lies in
the lower range of published results. However, those tests have performance profiles
significantly different as a function of fibrosis stages or cut-off. This has
to be taken into account during the interpretation process. Moreover, the
diagnostic target (fibrosis cut-off) has to be clearly defined.
|
Table 1 |
|||||
|
AUROC |
FT |
FM |
APRI |
HS |
P |
|
≥ F2 |
0.79 |
0.78 |
0.76 |
0.76 |
NS |
|
≥ F3 |
0.81 |
0.85 |
0.81 |
0.81 |
NS |
|
F4 |
0.86 |
0.94 |
0.92 |
0.89 |
NS |
|
Table 2 |
|||||
|
% |
F0 |
F1 |
F2 |
F3 |
F4 |
|
FT |
0 |
19 |
52 |
29 |
15 |
|
FM |
0 |
30 |
41 |
16 |
0 |
S. Chevaliez; M. Bouvier-Alias; G. Dameron; F. Darthuy; J. Remire; J. Pawlotsky
Introduction:
Accurate, sensitive and specific HCV RNA quantification is
mandated to monitor the virological responses and individually tailor pegylated
interferon-ribavirin combination therapy. Real-time PCR quantification is
likely to become standard technology for HCV RNA quantification. The Cobas
AmpliPrep-Cobas TaqMan48 (CAP-CTM48) real-time PCR platform (Roche Molecular
Systems) provides almost fully automated real-time PCR quantification of HCV
RNA.
Objective:
The objective of this study was to assess the intrinsic
performance of the CAP-CTM48 real-time PCR platform for HCV RNA detection and
quantification.
Methods:
Analytical sensitivity, precision and reproducibility were
assessed by testing serial dilutions of a standard containing 5x106 HCV RNA
IU/ml three times in different experiments. The lower limit of detection was
studied by testing serial dilutions of a standard containing 50 HCV RNA IU/ml.
Specificity was assessed by testing 200 anti-HCV antibody-negative samples.
Quantification of the different HCV genotypes was studied by testing 123
clinical samples (genotype 1, n=29; 2, n=27; 3, n=28; 4, n=30; 5, n=9; and 6,
n=2) neat and after serial dilutions in comparison with third-generation bDNA
(Versant HCV RNA 3.0, Bayer HealthCare).
Results:
(i) Testing serial dilutions of a standard containing 5x106
HCV RNA IU/ml showed linear quantification over the full range of HCV RNA
levels and a strong correlation with the expected values (slope: 0.9671,
R=0.9981). (ii) The coefficient of variations of triplicate testing ranged from
0.2% to 7.2% (1.9% on average). (iii) Specificity was 100%. (iv) HCV RNA was
detected in 8 repeat samples from 50 IU/ml down to 22 IU/ml, in 6/8 samples at
the 15 IU/ml concentration, 4/8 samples at the 10 IU/ml concentration and 5/8
samples at the 7 IU/ml concentration. (v) Serial dilutions of genotype 1 to 6
clinical samples showed HCV RNA levels in IU/ml generally higher in CAP-CTM48
than in bDNA (difference of the order of 0.7 log on average) when neat samples
were tested, but the difference was reduced after the first dilution. (vi)
Quantification of clinical samples was linear over the full range of tested
values whatever the HCV genotype. (vii) A few genotype 2 and 4 samples were
underquantified in the CAP-CTM48 assay relative to bDNA. Full-length sequence
analysis of the 5’UTR showed potential mismatch positions that could explain
these differences.
Conclusion:
CAP-CTM48 real-time PCR has good intrinsic performance, but two issues remain: a global overestimation of HCV RNA levels in IU/ml relative to bDNA in neat (but not diluted) samples, and occasional underquantification of genotype 4 and, sometimes, genotype 2 samples.
M. Ziol; P. Marcellin; C.
Douvin; V. de Ledinghen; R. Poupon; M. Beaugrand
Background:
Liver stiffness measurement (LSM) has been proposed as a
noninvasive tool to assess liver fibrosis in patients with chronic hepatitis C.
Several publications(1-3) have presented diagnosis accuracies and cut-off
values for the diagnosis of significant fibrosis and cirrhosis but with
different cut-off values. The aim of this study was to validate the diagnosis
accuracy and cut-off values of LSM in an independent HCV population.
Method:
We prospectively enrolled 639 HCV patients in five French
centers (Clichy: 178; Bondy: 163; Creteil: 129; Pessac: 101; Paris: 68). Within
six month each of them had a liver biopsy (LB) and LSM. Fibrosis stage and
activity grade were assessed using the METAVIR scoring system. Steatosis was
scored as follows: S0: no; S1:1-10%; S2:11-30% and S3 more than 30%. ROC
analyses were performed and cut-off values chosen to maximize the sum of
sensitivity and specificity.
Results:
Eighty-six (13%) LB were considered unsuitable for fibrosis
assessment (less than 1cm and/or 10 portal tracts if no obvious cirrhosis) and
59 (9%) of LSM unreliable (less than 8 valid measurements or
IQR/median>50%). Analysis was conducted in 494 patients (319 men; age: 49 ±
12 years; BMI 25 ± 4 kg/m2). Patients' distribution was: F0=30; F1=193; F2=155;
F3=47; F4=69. LSM was significantly (p<0.001) correlated to fibrosis stage
(Spearman: 0.70), activity (0.45) and steatosis (0.35) but in multivariate
analysis LSM was correlated to fibrosis only (partial correlation coefficient:
0.57).AUROC (95%CI) were 0.84 (0.80-0.87), 0.93 (0.90-0.95) and 0.96
(0.94-0.98) for F≥2, F≥3 and F=4 respectively. Optimal cut-off
values for this population and those previously published applied to this same
population are presented in the Table.
Conclusion:
These results obtained in an independent HCV population
confirmed those previously published. Optimal cut-off values are slightly
different but their performances are very close. This emphasizes the fact that
cut-off values depend on the chosen balance between sensitivity and
specificity. Finally, LSM appears as an accurate noninvasive method to assess
liver fibrosis in HCV patients.
1. Ziol Hepatology 2005:48-54.
2. Castera Gastroenterology 2005: 343-50.
3. Coco AASLD 2005.
|
|
Cut-off
values (kPa) |
Sensitivity |
Specificity |
VPP |
VPN |
Diagnosis
accuracy |
|
F01 / F2,3,4 |
7.5 |
67 |
87 |
86 |
68 |
76 |
|
F01 / F2,3,4 |
8.7 (1) |
59 |
92 |
90 |
65 |
74 |
|
F01 / F2,3,4 |
7.1 (2) |
71 |
81 |
82 |
70 |
76 |
|
F01 / F2,3,4 |
8.3 (3) |
61 |
89 |
87 |
65 |
74 |
|
F01,2,3 / F4 |
10.2 |
99 |
85 |
51 |
100 |
87 |
|
F01,2,3 / F4 |
14.5 (1) |
80 |
94 |
69 |
97 |
92 |
|
F01,2,3 / F4 |
12.5 (2) |
84 |
93 |
65 |
97 |
92 |
|
F01,2,3 / F4 |
14.0 (3) |
81 |
94 |
68 |
97 |
92 |
P. Cacoub; F. Carrat; P. Bedossa; J. Lambert; G. Pénaranda; C. Perronne; S. Pol; P. Halfon
Introduction:
Many non invasive liver fibrosis scores have been proposed as
alternatives to liver biopsy (LB) in HCV infected patients, i.e. Fibrotest
(FT), APRI, Forns, Hepascore (HS) and Fibrometer (FM). FT, SHASTA, and Fib-4
scores have been tested in small cohorts of HIV-HCV co-infected patients.
Aim:
To compare diagnostic accuracies of FT, APRI, Forns, HS, FM,
SHASTA, and Fib-4 in a large cohort of HIV-HCV co-infected patients.
Patients & Methods:
274 HIV-HCV coinfected patients, naïve of HCV treatment, had a LB before inclusion in the prospective ANRS HC02 Ribavic trial: 198 (72%) men, age 39.9 years, LB 18.5 mm, fibrosis stage (Metavir) of F1 in 68 (25%) patients, F2 in 110 (41%), F3 in 67 (23%), and F4 in 29 (11%). Diagnostic accuracies of each score were determined by area under the roc curve (AUROCs), and the accuracy (rate of well classified patient) compared to LB. AUROCs were compared with Hanley-McNeil test, and accuracies with McNemar or Chi2 test. Note that low values of AUROCs were expected, since our sample did not include patients with F0.
Results:
FT, HS, and FM were more accurate than other scores with
AUROCs of 0.64 [95% CI 0.58 ;0.70], 0.69 [0.63 ;0.74], and 0.70 [0.61 ;0.73]
for the diagnosis of significant fibrosis (F2F3F4), respectively. Values for
the diagnosis of F3F4 were 0.72 [0.66 ;0.77], 0.76 [0.71 ;0.81], and 0.78 [0.73
;0.83], respectively. For the diagnosis of cirrhosis (F4), AUROCs were 0.81
[0.76 ;0.85], 0.83 [0.78 ;0.88], and 0.84 [0.78 ;0.88]. Comparison of these 3
tests (Table) showed that for global analysis, accuracy FM > FT but not
different from HS. In a stage by stage fibrosis analysis, accuracy FM > HS
> FT except for F1 where FT > FM. Search for a better test by using combination
of FT, HS and FM did not increase significantly the diagnostic performances of
each test.
Conclusion:
In HIV-HCV co-infected patients, use of non invasive liver
fibrosis biomarkers can correctly classify only patients with severe fibrosis (≥
F3, accuracies 74%-97%). For global analysis, accuracy of FM is higher than FT
but similar to HS. In a stage by stage fibrosis analysis, FM is better for the
diagnosis of ≥F2 and ≥F3 stages whereas FT is better for the
diagnosis of F1.
|
|
Accuracy
= rate of well classified patients (%) |
|||||
|
Liver biopsy
fibrosis (Metavir) |
Global Analysis |
F1 |
F2 |
≥F2 |
F3 |
≥F3 |
|
Number of patients |
274 |
68 |
110 |
206 |
67 |
96 |
|
FT* |
61 |
60 |
||||