- Email: [email protected]
450 A NOVEL PREDICTION MODEL FOR LIVER FIBROSIS IN PATIENTS WITH CHRONIC HEPATITIS C VIRUS USING FIBROSCAN AND ROUTINE LABORATORY DATA
POSTERS immune genes, activation and recruitment of NK cells and T cells to the liver. Methods: We performed longitudinal quantification of the levels of IL28B by ELISA on plasma samples collected from a cohort of injection drug users (n = 30) during acute HCV infection with different outcomes. In addition, we used qRT-PCR to monitor expression of seven genes previously associated with response to interferon following HCV infection in in vitro models (IFI6, IFIT1, Mx1, USP18, IP-10) and NK cell activity (NCR3, KLRD1). Results: We observed that patients homozygous for the favourable IL28B rs12979860 C allele (CC) expressed higher levels of IL28B during early acute infection as compared to patients bearing the non-favourable T allele (*/T) (p = 0.0205). This observation became more significant when only acute infected patients with chronic evolution were analysed (p = 0.0047). IL28B plasma levels did not correlate with induction of genes examined in PBMCs except for IP-10, KLRD1 and Mx1. Higher IL28B plasma levels correlated with decreased expression of IP-10 (p = 0.0016) and KLRD1 (CD94) (p = 0.0002) and increased expression of Mx1 (p = 0.0283) only in patients with chronic evolution despite having the favourable genotype. Conclusions: Our results suggest that IL28B polymorphism may modulate the recruitment of immune cells to the liver and the NK cell activity by affecting the expression of the interferonstimulated chemokine IP-10 and antiviral protein Mx1, and the NK cell inhibitory receptor KLRD1. 450 A NOVEL PREDICTION MODEL FOR LIVER FIBROSIS IN PATIENTS WITH CHRONIC HEPATITIS C VIRUS USING FIBROSCAN AND ROUTINE LABORATORY DATA M. El Raziky1 , A.B. Awad2 , M. Youssef1 , A. Awad2 , A. Elsharkawi1 , G. Esmat1 , M. Mustafa1 , M. Mabrouk1 . 1 Faculty of Medicine Cairo University, 2 Computer Science, Faculty of Computers and Information, Cairo University, Cairo, Egypt E-mail: [email protected] Background: Data mining analysis explores data to discover hidden patterns, trends and enables the development of model to assess liver fibrosis utilizing liver stiffness measured by fibro-scan and simple laboratory data. Aim: To develop a novel model to predict the stage of fibrosis in chronic HCV patients using fibro-scan and routine clinical and laboratory data. Methods: Decision tree learning algorithm was applied to routine laboratory data of 296 chronic HCV Egyptian patients for model building using 20 attributes. Internal cross-validation was performed with 10-folds, and confidence (0.01). Liver histopathology (Metavir scoring) was used to assess accuracy for the model. Transient elastography measurement was used. Results: Fibrosis was classified into three groups F0–F1 (minimal), F2–F3 (moderate), and F4 (severe). The correctly classified instances were 218/296 (74%). Decision tree was able to diagnose F0–F1, F2–F3, and F4 with sensitivity 59% and specificity 36% using only routine data, and with sensitivity 70% and specificity 68.5% using fibroscan. At stiffness 7.25 fibroscan was able to diagnose liver fibrosis F0–F1with sensitivity 84% and specificity 85% and F2 F3 F4 with sensitivity 78% and specificity 76%. Out of 20 attributes the decision tree models showed that liver stiffness was selected as the variable of initial split (most decisive), with optimal cut-off value of <7.1 the possibility of being F0–F1 167/18 was 89.2%. At liver stiffness from 7.1–13.6 the possibility of being F2–F3 76/33 was 56.7%. Patients with liver stiffness >13.6, hepatic texture was the second important splitting attribute, other attributes as albumin, and AFP have less decisive role for prediction of fibrosis. as shown in fig (1). These results were confirmed statistically using univariate logistic regression analysis with P value <0.01. The reproducibility S184
of the model was confirmed by external validation set on 249 at cut off value <7.1 patients with correctly classified instances 90/15 (83.3%). Liver stiffness from 7.1–13.6 correctly classified instances is 76/20 (73.6). Conclusion: The model tree using fibro-scan and routine clinical and laboratory data can predict degree of hepatic fibrosis in chronic HCV patients with high accuracy and reproducibility. 451 EVALUATION OF PHARMACOKINETIC DRUG–DRUG INTERACTION (DDI) BETWEEN BMS-791325, AN NS5B NON-NUCLEOTIDE POLYMERASE INHIBITOR, DACLATASVIR AND ASUNAPREVIR IN TRIPLE COMBINATION IN HCV GENOTYPE 1-INFECTED PATIENTS X. Wang, W. Li, S.-P. Huang, B. He, E. Chung, A. Griffies, E. Cooney, E. Hughes, H. Kandoussi, K. Sims, D. Gardiner, R. Bertz, T. Eley. Bristol-Myers Squibb, Princeton, NJ, USA E-mail: [email protected] Background and Aims: Interferon and ribavirin free treatments for HCV may be achieved by combinations of direct-acting antivirals. Daclatasvir (DCV), an NS5A replication complex inhibitor, and asunaprevir (ASV), an NS3 protease inhibitor were combined with BMS-791325, a potent, selective non-nucleoside inhibitor of the NS5B polymerase, to treat HCV genotype (GT) 1-infected patients. All three drugs are P-glycoprotein substrates, CYP3A4 substrates, OATP1B1 inhibitors and P-glycoprotein inhibitors in vitro and/or in vivo. ASV induces CYP3A4, and is an OATP1B1 substrate. BMS791325 appears to induce CYP3A4. BMS-794712 is the active metabolite of BMS-791325. No clinically meaningful DDI occurred previously between DCV and ASV. In this study, potential DDIs of the triple combination were assessed in a subset of patients. Methods: Study AI443014 is a phase 2a open-label, multipledose study combining DCV (60 mg QD), ASV (200 mg BID, tablet), and BMS-791325 at two doses (75 mg BID or 150 mg BID) in 32 treatment-naive, HCV GT 1-infected, non-cirrhotic patients for 12 or 24 weeks. Non-compartmental pharmacokinetic parameters were derived on Day 14. The pharmacokinetics of all analytes in each regimen (N = 12 for 75 mg BID and N = 18 for 150 mg BID), including BMS-794712, were explored versus historical data graphically and using descriptive statistics. Results: See Table. Day 14 exposures: DCV and BMS-791325 were comparable to historical data. ASV exposures appeared to be reduced by ~30%; variability was high. Metabolic ratio for BMS794712 was ~25%. Conclusion: No clinically meaningful interaction was observed by addition of BMS-791325 to DCV and ASV. Additional study of both doses of BMS-791325 in this triple combination is warranted to confirm the most appropriate dose in broader patient populations. Table: AUC(TAU) (ng*h/mL), Geometric Mean (CV%) on Day 14
Daclatasvir Asunaprevir BMS-791325 BMS-794712
This study
Historical values
11248 (36)* 1065 (78)* 9554 (65)‡ 2364 (48)‡
10700 (30.7) (N = 11)† 1528 (106) (N = 12)† 9170 (34)§ 2150 (35)§
*DCV 60 mg QD+ASV 200 mg BID+BMS-791325 150 mg BID; † DCV 60 mg QD+ASV 200 mg BID in HCV-1 patients; ‡ DCV 60 mg QD+ASV 200 mg BID+BMS-791325 75 mg BID; § BMS-791325 75 mg BID+peginterferon alfa2a+ribavirin in HCV-1 patients (N = 12).
Journal of Hepatology 2013 vol. 58 | S63–S227
of the model was confirmed by external validation set on 249 at cut off value <7.1 patients with correctly classified instances 90/15 (83.3%). Liver stiffness from 7.1–13.6 correctly classified instances is 76/20 (73.6). Conclusion: The model tree using fibro-scan and routine clinical and laboratory data can predict degree of hepatic fibrosis in chronic HCV patients with high accuracy and reproducibility. 451 EVALUATION OF PHARMACOKINETIC DRUG–DRUG INTERACTION (DDI) BETWEEN BMS-791325, AN NS5B NON-NUCLEOTIDE POLYMERASE INHIBITOR, DACLATASVIR AND ASUNAPREVIR IN TRIPLE COMBINATION IN HCV GENOTYPE 1-INFECTED PATIENTS X. Wang, W. Li, S.-P. Huang, B. He, E. Chung, A. Griffies, E. Cooney, E. Hughes, H. Kandoussi, K. Sims, D. Gardiner, R. Bertz, T. Eley. Bristol-Myers Squibb, Princeton, NJ, USA E-mail: [email protected] Background and Aims: Interferon and ribavirin free treatments for HCV may be achieved by combinations of direct-acting antivirals. Daclatasvir (DCV), an NS5A replication complex inhibitor, and asunaprevir (ASV), an NS3 protease inhibitor were combined with BMS-791325, a potent, selective non-nucleoside inhibitor of the NS5B polymerase, to treat HCV genotype (GT) 1-infected patients. All three drugs are P-glycoprotein substrates, CYP3A4 substrates, OATP1B1 inhibitors and P-glycoprotein inhibitors in vitro and/or in vivo. ASV induces CYP3A4, and is an OATP1B1 substrate. BMS791325 appears to induce CYP3A4. BMS-794712 is the active metabolite of BMS-791325. No clinically meaningful DDI occurred previously between DCV and ASV. In this study, potential DDIs of the triple combination were assessed in a subset of patients. Methods: Study AI443014 is a phase 2a open-label, multipledose study combining DCV (60 mg QD), ASV (200 mg BID, tablet), and BMS-791325 at two doses (75 mg BID or 150 mg BID) in 32 treatment-naive, HCV GT 1-infected, non-cirrhotic patients for 12 or 24 weeks. Non-compartmental pharmacokinetic parameters were derived on Day 14. The pharmacokinetics of all analytes in each regimen (N = 12 for 75 mg BID and N = 18 for 150 mg BID), including BMS-794712, were explored versus historical data graphically and using descriptive statistics. Results: See Table. Day 14 exposures: DCV and BMS-791325 were comparable to historical data. ASV exposures appeared to be reduced by ~30%; variability was high. Metabolic ratio for BMS794712 was ~25%. Conclusion: No clinically meaningful interaction was observed by addition of BMS-791325 to DCV and ASV. Additional study of both doses of BMS-791325 in this triple combination is warranted to confirm the most appropriate dose in broader patient populations. Table: AUC(TAU) (ng*h/mL), Geometric Mean (CV%) on Day 14
Daclatasvir Asunaprevir BMS-791325 BMS-794712
This study
Historical values
11248 (36)* 1065 (78)* 9554 (65)‡ 2364 (48)‡
10700 (30.7) (N = 11)† 1528 (106) (N = 12)† 9170 (34)§ 2150 (35)§
*DCV 60 mg QD+ASV 200 mg BID+BMS-791325 150 mg BID; † DCV 60 mg QD+ASV 200 mg BID in HCV-1 patients; ‡ DCV 60 mg QD+ASV 200 mg BID+BMS-791325 75 mg BID; § BMS-791325 75 mg BID+peginterferon alfa2a+ribavirin in HCV-1 patients (N = 12).
Journal of Hepatology 2013 vol. 58 | S63–S227