Transcript Paper # H-2320 Diagnostic Accuracy of Serum Hyaluronic Acid for Advanced Fibrosis/Cirrhosis in Patients Coinfected with HIV and HCV S.

Paper # H-2320
Diagnostic Accuracy of Serum Hyaluronic Acid for Advanced Fibrosis/Cirrhosis in Patients Coinfected with HIV and HCV
S. Resino,*1 P. Miralles,2 D. Micheloud,1 J. M. Bellón,2 A. Vargas,2 P. Catalán,2 E. Álvarez,2 J. Cosín,2 R. Lorente,2 M. Sánchez-Conde,2 M. A. Muñoz-Fernández,2 And J. Berenguer
2
Instituto de Salud Carlos III, Majadahonda, Madrid;
2 Hosp.General Universitario Gregorio Marañón, Madrid, Spain
1
Hyaluronic acid (HA), an essential component of the extracellular matrix of the liver is produced by hepatic stellate cells and
degraded by sinusoidal endothelial cells. HA levels increase with the development of liver fibrosis in patients with HCV. We
evaluated serum HA to predict advanced fibrosis (F≥3) and cirrhosis (F4) in HIV+/HCV+ patients.
PATIENS AND METHODS
Patients
The patients for this study came from the Hospital Gregorio Marañón in Madrid (Spain). They all had documented HIV-HCV
coinfection and underwent liver biopsy prior to therapy with interferon and ribavirin.
Laboratory data
A fasting serum sample was immediately stored and frozen (at a temperature of -70ºC) for further assays, after the patient
gave a written consent. Hyaluronic acid (HA) was tested in serum samples mentioned before by a commercially available
quantitative ELISA (HA-ELISA; Echelon Biosciences Inc., Salt Lake City, UT, USA). Concentrations were assayed in
duplicate.
Liver fibrosis was estimated following the criteria established by the METAVIR Cooperative Study Group. Fibrosis was
scored as follows: F0, no fibrosis; F1, portal fibrosis; F2 periportal fibrosis or rare portal-portal septa; F3, fibrous septa with
architectural distortion; no obvious cirrhosis (bridging fibrosis); and F4, definite cirrhosis.
Statistics
Results are presented by median and percentiles, and as frequencies and percentages for categorical data. We calculated
the diagnostic values of HA, Forn’s, APRI and FIB-4 indexes and assessed their diagnostic accuracy by using the area under
the receiver operating characteristic curves (AUROCs). Next, we calculated the sensitivity (Se) and specificity (Sp) of
several cut-off points to evaluate the diagnostic accuracy for advanced fibrosis. Thus, we chose the one with the 95% Se
(low cut-off) and the one with the 95% Sp (high cut-off). We also calculated the positive predictive value (PPV) and
negative predictive value (NPV) for these two cut-off points. Comparison between HA and fibrosis stage were analysed
using Mann-Witney U-test. All tests were two-tailed with P values ≤0.05 considered significant. Statistical analysis was
performed by SPSS 14.0 software (SPSS INC, Chicago, IL, USA) and STATA 9.1.
1
Table 2. Diagnostic accuracy of hyaluronic acid (HA) for significant fibrosis (F≥2), advanced fibrosis (F≥3) and cirrhosis (F4).
19 (9.5)
13 (15.5)
6 (5.1)
>430 (ng/mL)
High cutoff point
<1800 (ng/mL)
182 (90.5)
71 (84.5)
111 (94.9)
161 (80.1)
80 (95.2)
81 (69.2)
>1800 (ng/mL)
40 (19.9)
4 (4.8)
36 (30.8)
AH
Low cutoff point
<687 (ng/mL)
Actual advanced fibrosis (F≥ 3)*
Total
F<3
F≥3
201
137 (68.2)
64 (31.8)
52 (25.9)
49 (35.8)
3 (4.8)
>687 (ng/mL)
High cutoff point
<2290 (ng/mL)
149 (74.1)
88 (64.2)
61 (95.3)
176 (87.6)
130 (94.9)
46 (71.9)
>2290 (ng/mL)
25 (12.6)
7 (5.1)
18 (28.1)
Total
201
Cirrhosis (F4)*
F<4
178 (88.6)
F4
23 (11.4)
<1182 (ng/mL)
101 (50.2)
100 (56.2)
1 (4.3)
>1182 (ng/mL)
High cutoff point
<2400 (ng/mL)
100 (49.8)
78 (43.8)
22 (95.7)
181 (90.0)
169 (94.9)
12 (52.2)
>2400 (ng/mL)
20 (10.0)
9 (5.1)
11 (47.8)
AH
Low cutoff point
94.9
(90.4 – 99.3)
Sp (%,CI95)
15.5
(7.1 – 23.8)
PPV (%,CI95)
61.0
(53.6 – 68.7)
NPV (%,CI95)
Interpretation
68.4
(44.9 – 91.9)
Presence of F<2
(68.4% certainty)
30.8
(21.9 – 39.6)
95.2
(90.1 – 100)
90.0
(79.5 –100 )
49.7
(41.7 – 57.7)
Presence of F≥2
(90.0% certainty)
Se (%, CI95)
Sp (%,CI95)
PPV (%,CI95)
NPV (%,CI95)
Interpretation
Figure 1. Box plots illustrating the distribution of the values of HA
against metavir fibrosis score. Horizontal line inside each box
represents the median, and the lower and upper borders of the box
encompass the interquartile range. The vertical lines from the ends
of each box encompass the extreme data points.
P<0.001
P<0.001
95.3
(89.3 – 100)
35.8
(27.4 – 44.2)
40.9
(32.7 – 49.2)
A
94.2
(86.9 – 100)
94.9
(90.8 – 98.9)
72.0
(52.4 – 91.6)
73.9
(67.1 – ,8
80.6)
Se (%, CI95)
Sp (%,CI95)
PPV (%,CI95)
NPV (%,CI95)
95.7
(85.1 – 100)
56.2
(48.6 – 63.7)
22.0
(13.4 – 30.6)
P=0.004
55.0
(30.7 – 79.3)
3000
Interpretation
2500
* Number and (percentage). Abbreviations: Se, sensitivity; Sp, specificity; PPV, positive predictive value; NPV, negative predictiveFibrosis
value.
Stage
Presence of F4
(55.0%53certainty)40
F0-F1
F2
F3
0. 90
1. 00
HA
0. 863
Forn's
F4
0. 817
APRI
0 .801
FIB -4
0. 819
HA
0 . 772
Forn's
F≥3
0 .761
APRI
0. 762
FIB -4
F≥2
0 .740
APRI
0. 765
FIB -4
0. 50
0. 60
0. 70
0. 80
0. 90
1. 00
2000
1500
Presence of F<4
(99.0% certainty)
82
0. 80
0 .731
Presence of F≥3
(73.9% certainty)
,6
93.40
(89.5 – 97.3)
N=
0. 70
0 . 676
P=0.001
CONCLUSION
1000
500
94.9
(91.4 – 98.4)
0. 60
Forn's
,2
47.8
(25.2 – 70.4)
0. 50
Figure 2. Summary of area under the receiver
operating characteristic curves (AUROCs) of
hyaluronic acid (HA), FIB-4, APRI and Forn’s
indexes for significant fibrosis (F≥2),
advanced fibrosis (F≥3) and cirrhosis (F4).
HA
P<0.001
3500
99.0,4
(96.6 – 100)
AUROCs
0. 753
P<0.001
B
P<0.001 of F<3
Presence
(94.2%P<0.001
certainty) P<0.001
1,0
28.1
(16.3 – 39.9)
Our study cohort included 201 patients whose characteristics at the time of liver biopsy are shown in Table 1. Overall, HA
increased significantly with the advanced stage of hepatic fibrosis (Figure 1). Moreover, the differences were higher between
F0-F1 and F4 in AH levels distribution. For purposes of comparison, we also evaluated three reported simple models consisting of
routine parameters to predict liver fibrosis: a) Forn’s index; b) the APRI index; c) FIB-4. The AUROC values of the HA for
significant fibrosis (F≥2), advanced fibrosis (F≥3) and cirrhosis (F4) were similar to FIB-4, APRI, and Forn´s indexes (Figure 2).
The 95% confidence interval of AUROC of HA for significant fibrosis (0.676; CI95%: 0.603; 0.750) and cirrhosis (0.863; CI95%:
0.795; 0.931) are not overlapped and there is difference statistically significant (p<0.05).
Next, we calculated the Se, Sp, PPV and NPV of several cutoff points to evaluate the diagnostic accuracy of HA for significant
fibrosis (F≥2), advanced fibrosis (F≥3) and cirrhosis (F4) (Table 2).
For significant fibrosis (F≥2) and with the low AH cut-off (<430 ng/mL), 13 of 84 (15.5%) patients without significant fibrosis
were correctly identified, and only 6 of 117 (5.1%) patients with significant fibrosis were misclassified (61.0% PPV and 68.4%
NPV). When we applied the high AH cut-off (> 1800 ng/mL), 36 of 117 (30.8%) patients with significant fibrosis were correctly
identified, and only 4 of 84 (4.8%) of the patients without significant fibrosis were misclassified (90% PPV and 49.7% NPV)
(Table 2).
For advanced fibrosis (F≥3) and with the low AH cut-off (<687 ng/mL), 49 of 137 (35.8%) patients without advanced fibrosis
were correctly identified, and only 3 of 64 (4.8%) patients with advanced fibrosis were misclassified (40.9% PPV and 94.2%
NPV). When we applied the high AH cut-off (> 2290 ng/mL), 18 of 64 (28.1%) patients with advanced fibrosis were correctly
identified, and only 7 of 137 (5.1%) of the patients without advanced fibrosis were misclassified (72% PPV and 73.9% NPV)
(Table 2).
For cirrhosis (F4) and with the low AH cut-off (<1182 ng/mL), 100 of 178 (56.2%) patients without cirrhosis (F4) were correctly
identified, and only 1 of 23 (4.3%) patients with cirrhosis (F4) were misclassified (22.2% PPV and 99% NPV). When we applied
the high AH cut-off (> 2400 ng/mL), 11 of 23 (47.8%) patients with cirrhosis (F4) were correctly identified, and only 9 of 178
(5.1%) of the patients without cirrhosis (F4) were misclassified (55% PPV and 93.4% NPV) (Table 2).
P<0.001
HA
<430 (ng/mL)
Se (%, CI95)
New index (PAHH)
AH
Low cutoff point
Actual significant fibrosis (F≥ 2)*
Total
F<2
F≥2
201
84 (41.8)
117 (58.2)
RESULTS
Indexes
Table 1. Characteristics of 201 HIV/HCV coinfected patients, who
underwent a liver biopsy.
Characteristic
No. HIV-1 patients *
201
Male sex *
152 (75.6)
Age (years) #
39.4 (36.8; 43.3)
HIV acquired by IVDU *
180 (89.6)
Prior AIDS *
62 (30.8)
Years since HCV infection #
21.3 (17.7; 24.3)
High alcohol intake *
28 (14)
Antiretroviral therapy
Non treated
10 (5)
PI-based *
47 (23.4)
Months on HAART (n=190) #
50.2 (34.9; 65.7)
Stage of liver fibrosis *
F0
16 (8)
F1
68 (33.8)
F2
53 (26.4)
F3
41 (20.4)
F4
23 (11.4)
Fibrosis progression index #
0.08 (0.05; 0.15)
HIV markers
Nadir CD4+ T-cells #
210 (103; 324)
Baseline CD4+ T-cells/μL #
490 (373; 660)
HIV-RNA < 50 cp/mL *
156 (77.6)
Log10 VL copies/mL (n=45)
3.23 (2.71; 3.98)
HCV markers *
HCV genotype
1-4
153 (77.3)
3
45 (22.7)
HCV-RNA >850,000 cp/ml
125 (75.8)
Biochemical parameters #
ALP (UI/dL)
124 (81; 196)
AST (UI/dL)
57 (37.5; 85)
GGT (UI/dL)
113 (58; 208)
ALT (UI/dL)
77 (49; 117)
AST/ALT
0.75 (0.6; 0.97)
* Absolute number (percentage). # Median (percentile 25; percentile 75).
Abbreviations: HCV: Hepatitis C virus. HIV-1: Human immunodeficiency
virus type 1. HAART: highly active antiretroviral therapy. PI: protease
inhibitor. HIV-RNA: HIV plasma viral load. HCV-RNA: HCV plasma viral
load. INR: international normalized ratio. ALP: alkaline phosphatase.
AST: aspartate aminotransferase. GGT: gamma glutamyl transpeptidase.
ALT: alanine aminotransferase.
OBJETIVE
0
23
F4
N=
Fibrosis
Stage
84
53
41
23
F0-F1
F2
F3
F4
HA appears to be and accurate non-invasive method for detection of of advanced fibrosis and cirrhosis in HIV/HCV-coinfected
patients.
Corresponding author: Salvador Resino, Unidad de Investigación, Instituto de Salud Carlos III (Campus Majadahonda), Carretera Majadahonda- Pozuelo, Km 2.2; 28220 Majadahonda (Madrid); Telf.: +34 918 223 266; Fax: +34 915 097 946; e-mail: [email protected]