CROI 2016 Abstract eBook

Abstract Listing

Poster Abstracts

528

Variation in Liver Fibrosis Staging Using FibroTest, FIB-4, and APRI Scores Aurielle Thomas 1 ; Charisse Ahmed 2 ; Chloe Gross 3 ; Rachel Silk 3 ; Elizabeth Akoth 3 ; Eleanor M.Wilson 4 ; Angie Price 3 ; Shyam Kottilil 5 ; Henry Masur 6 ; Sarah Kattakuzhy 3 1 NIH Clinical Cntr, Bethesda, MD, USA; 2 NIH Clinical Cntr, Bethesda, MD, USA; 3 Inst of Human Virology, Bethesda, MD, USA; 4 Inst of Human Virology, Baltimore, MD, USA; 5 Univ of Maryland Med Cntr, Baltimore, MD, USA; 6 NIH, Bethesda, MD, USA

Background: Patients with Hepatitis C (HCV) infection develop varying degrees of hepatic fibrosis, and fibrosis stage affects multiple aspects of HCV care. Increasingly, non-invasive methods are used to determine the stage of liver fibrosis. The objective of this analysis was to analyze three common staging methods from a large urban cohort and determine their degree of correlation among HCV mono-infected and HIV/HCV co-infected patients. Methods: Analysis was completed by a chart review of 263 patients from the ASCEND study, a phase IV multi-site trial of community-based hepatitis C treatment. FibroTest, FIB-4, and APRI were obtained for each patient using standard equations, and all were adjusted to a four-point scale for comparison. Serum biomarkers used for FIB-4 and APRI scoring were obtained within 30 days from FibroTest. Statistical Analysis was completed using nonparametric Wilcoxon rank sum and column statistics were performed in Graphpad Prism 6.0. Results: The patient cohort was predominantly African American, with a mean age of 59, and included 61 HIV/HCV co-infected and 203 HCV mono-infected patients. Overall, mean FibroTest score was higher than both APRI (mean difference -0.77+/- .09) and FIB-4 (mean difference -0.44+/- .08), The calculated difference between FibroTest and APRI was significantly higher (p<0.0001) when compared to the mean difference between FibroTest and FIB-4. In contrast, the mean difference in between FIB-4 vs APRI was 0.33 (SEM:0.06). In co-infected patients, the calculated mean difference of both FibroTest-FIB-4 and FibroTest-APRI was not significantly different (p=0.5, p=0.57) compared to mono-infected patients. Conclusions: Overall, this analysis demonstrates significant differences in liver fibrosis staging between FibroTest, FIB-4, and APRI scores in an urban cohort representative of the HCV epidemic, including HIV/HCV co-infected patients. There was minimal variation between APRI and FIB-4, and these scores were significantly lower than calculated FibroTest, with no effect of HIV co-infection on these relationships. In the era of non-invasive staging, more research on the correlation of each scoring systemwith long-term outcomes is required to determine optimal staging method.

p=0.57

p=0.5

p<0.0001

Mean:-0.74 SEM:0.1

p=0.0077

4

Mean:-0.41 SEM:0.09

Mean:-0.86 SEM:0.18

Mean:-0.53 SEM:0.16

2

0

-2

Difference in Score

-4

Co Fibrotest-FIB4

Co Fibrotest-APRI

Mono Fibrotest-FIB4 Test Comparisons

Mono Fibrotest-APRI

529

New Predictive Index Based on the Combination of Liver Stiffness and CTP Nicolás Merchante 1 ; Antonio Rivero-Juárez 2 ; FranciscoTéllez 3 ; Dolores Merino 4 ; María José Ríos 5 ; Guillermo Ojeda-Burgos 6 ; Mohamed Omar Mohamed-Balghata 7 ; Patricia Monje- Agudo 1 ; Montserrat Pérez-Pérez 8 ; Juan A. Pineda 1 ; for the HEPAVIR-Cirrhosis Study Group 1 Hosp Universitario de Valme, Sevilla, Spain; 2 IMIBIC, Córdoba, Spain; 3 Hosp La Línea de la Concepción (AGS Campo de Gibraltar), La Línea de la Concepción, Spain; 4 Complejo Hospario de Huelva, Huelva, Spain; 5 Hosp Universitario Virgen de la Macarena, Seville, Spain; 6 Hosp Universitario Virgen de la Victoria, Málaga, Spain; 7 Complejo Hospario de Jaén, Jaén, Spain; 8 Hosp de La Línea de la Concepción, Cádiz, Spain Background: Liver stiffness (LS) predicts clinical outcome in patients with cirrhosis. However, LS is not being routinely used for some clinical decision-making, which still mainly relies on the Child-Turcotte-Pugh (CTP) or the MELD scores. Our objective was to develop a new predictive index, which includes LS, and to evaluate its ability to predict events in HIV/HCV-coinfected patients with compensated cirrhosis. Methods: Prospective study of 446 HIV/HCV-coinfected patients with a new diagnosis of cirrhosis and no previous liver decompensation (LD). The time from diagnosis to LD and to liver-related death (LRD), as well as its predictors, were evaluated. A new score based on the combination of LS and CTP was built. The ability of the new score to predict outcomes was compared to that of other classical scores, by means of the comparisons of the AUROC and the integrated discrimination improvement (IDI) between models. Results: Median (IQR) follow-up of 49 (25-68) months. 80 (17.9%) patients had a LD. Variables independently associated with LD: age, HBV, SVR, AIDS, CTP stage B (vs A) (AHR 4.18, p<0.0001) and baseline LS (comparison group: LS < 21 kPa) (LS 21-39,9 kPa: AHR 2.48, p=0.005; LS ≥ 40 kPa: AHR 3.68, p< 0.0001). LS yielded a better performance than MELD (IDI 3.3%; p=0.01) and a similar performance than CTP (IDI 0.13%; p=0.9) to predict LD. Consequently, LS and CTP were combined in a new predictive score. 3-year probability of LD increased across stages of the new index: stage 1 (CTP A and LS < 21 kPa) 5%, stage 2 (CTP A and LS 21-39,9 kPa) 10%, stage 3 (CTP A and LS ≥ 40 kPa) 17%, stage 4 (CTP B and LS < 21 kPa) 29%, stage 5 (CTP B and LS 21-39,9 kPa) 52% and stage 6 (CTP B and LS ≥ 40 kPa) 73% (p<0.0001). A newmultivariate analysis including the new index demonstrated its independent association with LD (stage 1: comparison group): stage 2 [AHR 2.9; p=0.002], stage 3 [AHR 3.5; p=0.001], stage 4 [AHR 3.7; p=0.2], stage 5 [AHR 10.4; p<0.0001] and stage 6 [AHR 20.7; p<0.0001]. The AUROC of this model was higher than that of the model based on the CTP stage (AUROC 0.612 vs AUROC 0.575; p=0.06). LRD occurred in 37 (8.3%) patients. 3-year probability of LRD increased from stage 1 to 3 (stage 1: 1%, stage 2: 4%, stage 3: 10%) and was significantly higher in stages 4 to 6 (stage 4: 29%, stage 5: 27%, stage 6: 28%). Conclusions: The combination of LS and CTP stage in a new predictive index improves the ability of CTP to predict clinical outcome in HIV/HCV-coinfected patients with compensated cirrhosis.

Poster Abstracts

206

CROI 2016