Antiviral Therapy 2017; 22: (doi: /IMP3117)

Transcription

1 Antiviral Therapy 2017; 22: (doi: /IMP3117) Original article Real-world effectiveness and predictors of sustained virological response with all-oral therapy in 21,242 hepatitis C genotype-1 patients Lisa I Backus 1 *, Pamela S Belperio 1, Troy A Shahoumian 1, Timothy P Loomis 1, Larry A Mole 1 1 Department of Veterans Affairs, Population Health Services, Palo Alto Health Care System, Palo Alto, CA, USA *Corresponding author Lisa.Backus@va.gov Background: Predictors of sustained virological response (SVR) to all-oral HCV regimens can inform nuanced treatment decisions. We evaluated effectiveness and identified predictors of SVR for ledipasvir/sofosbuvir ± ribavirin (LDV/SOF ±RBV) and ombitasvir/paritaprevir/ ritonavir + dasabuvir (OPrD) ±RBV in patients treated in routine practice. Methods: Observational, intent-to-treat cohort of 21,142 genotype-1 patients initiating 8 or 12 weeks of LDV/SOF ±RBV or 12 weeks of OPrD ±RBV at any Veterans Affairs facility. Multivariate logistic regression models were constructed to model SVR and identify predictors. Results: SVR was 91.2% (9,781/10,720) for LDV/ SOF, 89.6% (3,266/3,646) for LDV/SOF+RBV, 91.7% (1,197/1,306) for OPrD and 87.8% (3,365/3,832) for OPrD+RBV. For LDV/SOF ±RBV, reduced odds of SVR occurred in African-Americans (0.80, 95% CI 0.70, 0.92, P<0.001), body mass index (BMI)<25 (0.77, 95% CI 0.66, 0.90, P<0.001), BMI 30 (0.77, 95% CI 0.67, 0.89, P<0.001), proton pump inhibitors (PPIs; 0.81, 95% CI 0.71, 0.92, P<0.001), decompensated liver disease (0.58, 95% CI 0.45, 0.74, P<0.001) and FIB4>3.25 (0.60, 95% CI 0.53, 0.69, P<0.001). For OPrD ±RBV, FIB-4>3.25 negatively predicted SVR (0.72, 95% CI 0.59, 0.88, P<0.001). Detectable 4-week on-treatment HCV RNA 15 IU/ml reduced SVR odds for both regimens (LDV/SOF ±RBV OR 0.49, 95% CI 0.41, 0.58, P<0.001; OPrD ±RBV OR 0.38, 95% CI 0.29, 0.50, P<0.001). Receipt of OPrD+RBV compared to LDV/SOF reduced odds of SVR (OR 0.70, 95% CI 0.62, 0.80, P<0.001). Mental health diagnosis did not impact likelihood of SVR. Conclusions: The diversity and size of this cohort allowed for extensive examination of regimen-specific predictors of SVR. FIB-4>3.25 and detectable 4-week on-treatment HCV RNA had the greatest negative impact. African- American race, low or high BMI, and PPIs negatively impacted odds of SVR for LDV/SOF ±RBV. Mental health diagnoses did not. Introduction Clinical trials of the current all-oral HCV regimens for genotype-1 infection have generally reported similarly high sustained virological response (SVR) rates, regardless of the characteristics of the patients at baseline, with few exceptions [1 5]. In contrast, the response to earlier protease inhibitor- and interferon-based regimens for HCV infection varied considerably depending on patient characteristics (for example, baseline viral load, race, HCV genotype, IL28B genotype and extent of fibrosis) and the presence or absence of an early ontreatment response [6]. The carefully selected patients and limited sample sizes of the all-oral clinical trials may not have allowed for adequate determination of differences that may occur in the real-world and that would identify predictors of SVR in this all-oral era. Knowledge of predictors specific to a particular regimen may assist with nuanced decisions for individualizing patient regimens. Understanding the effectiveness of HCV antiviral regimens is a priority for the Department of Veterans Affairs (VA), the largest US provider of health care to HCV-infected individuals. HCV disproportionately affects the veteran population and nearly 5% of all individuals in the US with HCV infection are cared for by VA [7,8]. Given the rapid uptake of all-oral HCV regimens within VA and the diverse HCV-infected veteran population receiving these regimens, we evaluated predictors of response to ledipasvir/sofosbuvir with or without ribavirin (LDV/SOF ±RBV) and ombitasvir/ paritaprevir/ritonavir plus dasabuvir with or without 2017 International Medical Press (print) (online) 481

2 LI Backus et al. ribavirin (OPrD ±RBV) in genotype-1 HCV-infected veterans treated in routine medical practice. Methods This observational intent-to-treat cohort analysis used data from the VA s Clinical Case Registry for HCV, an extract of the VA electronic medical record for all HCVinfected veterans seen at all VA medical facilities [9]. Eligible subjects included all genotype-1 HCVinfected veterans from any VA facility who initiated 8 or 12 weeks of VA-prescribed LDV/SOF ±RBV or 12 weeks of OPrD ±RBV by 30 September 2015 with an end of treatment (EOT) by 15 January For patients who received multiple courses of therapy, only the first course of therapy was included. Regimen choice and timing of follow-up laboratory testing was at the discretion of the provider as is customary in routine practice. Both LDV/SOF and OPrD were on the VA national formulary and available to providers. The present cohort includes 6,961 patients treated with LDV/SOF ±RBV or OPRD ±RBV reported on previously [10]. Patients were excluded for baseline HCV RNA 1,000 IU/ml (n=329), liver transplant (n=240), changing regimens without a treatment interruption (n=98) and receipt of OPrD without RBV for genotype- 1a (n=39). Treatment outcome SVR was defined as HCV RNA results below the limit of quantification after the EOT including at least one test 10 weeks or more after the EOT. 10 weeks was selected to account for the realities of variability in timing of laboratory testing in clinical practice. Patients were categorized as not achieving SVR if they had an HCV RNA above the limit of quantification after the EOT and no subsequent test 10 weeks after EOT, had no HCV RNA testing after the EOT and their last ontreatment HCV RNA was above the limit of quantification, or died while on treatment or within 10 weeks of the EOT. Patients with HCV RNA below the limit of quantification on their last HCV viral load test, either on treatment or after the EOT, but no test 10 weeks of more after the EOT were excluded from the SVR analysis. The EOT was calculated as the last day covered by prescriptions of LDV/SOF or OPrD using all the dates the medication was dispensed and the quantity. HCV RNA was categorized as above or below the lower limit of quantification of 15 IU/ml. Patients were followed from initiation of LDV/SOF ±RBV or OPrD ±RBV through 31 May Control variables Demographic and other baseline variables were determined at the time of treatment initiation and included age, sex, race/ethnicity, body mass index (BMI), diabetes (using International Classification of Diseases [ICD]-9 codes), HIV coinfection, FIB-4 score, history of decompensated liver disease (defined by ICD-9 codes for oesophageal variceal haemorrhage, hepatic coma, hepatorenal syndrome or spontaneous bacterial peritonitis), history of mental health diagnoses (defined by ICD-9 codes for anxiety, bipolar, depression, post-traumatic stress disorder [PTSD] and schizophrenia) at any time prior to treatment initiation and in 1 year prior to treatment initiation, alcohol abuse and hard drug (amphetamine, cocaine, methamphetamine and opioid) use in 1 year prior to treatment initiation (defined by ICD-9 codes), prescribed proton pump inhibitor (PPI; once or twice daily), prescribed histamine 2 receptor antagonists (H2RA), prior HCV antiviral treatment experience and HCV genotype-1 subtype. Subtype-1a included patients with results of 1a, mixed 1a/1b or 1 with subtype unspecified. Prior virological response was based on the most recent VA course of HCV antiviral treatment prior to LDV/SOF ±RBV or OPrD ±RBV. Baseline values for height and weight used to calculate BMI and the laboratory tests for alanine aminotransferase, aspartate aminotransferase, platelets and baseline HCV RNA were defined as the value within 1 year before and closest to the treatment start date. An FIB- 4>3.25 at the start of treatment using baseline laboratory values was used as a marker of advanced liver disease [11]. Patients with FIB were considered to be non-cirrhotic. In VA, HCV antiviral prescriptions are often filled for quantities less than 28 days. Patients were considered to have completed 8 weeks of LDV/SOF if they had received between days worth of medication and 12 weeks LDV/SOF ±RBV or OPrD ±RBV if they received between days worth of medication. Statistical analysis Univariate comparisons used the Pearson c 2 test with Yates continuity correction for categorical variables. Multivariate logistic regression models were constructed to identify predictors of SVR. Models included age, sex, race/ethnicity, BMI, diabetes, history of decompensated liver disease, mental health diagnoses, alcohol abuse diagnosis, hard drug use diagnosis, PPI receipt, treatment experience, FIB-4, genotype subtype and regimen. Additional models included H2RA receipt, PPI dosing of once or twice daily and 4-week on-treatment HCV RNA. Models with the above baseline variables were constructed with all patients, separately for patients who received LDV/SOF ±RBV and OPrD ±RBV, and with only patients who completed full courses of treatment defined as 8 or 12 weeks for LDV/SOF and 12 weeks for all other regimens International Medical Press

3 All-oral HCV treatment in 21,242 patients For all comparisons, a P-value <0.01 was considered statistically significant. All analyses were performed using R version 3.1 (R Foundation for Statistical Computing, Vienna, Austria). The protocol was approved by the Stanford University Institutional Review Board and the VA Palo Alto Health Care System Research and Development Committee. Results There were 21,242 patients with HCV genotype-1 who initiated LDV/SOF (n=11,804), LDV/SOF+RBV (n=3,869), OPrD (n=1,423) or OPRD+RBV (n=4,146) at 128 VA facilities. The mean age for the cohort was 61.7 years, 39.0% were African-American, 32.4% had diabetes, 33.7% had a BMI 30 kg/m 2, 30.4% had an FIB-4>3.25 and 18.9% were treatment-experienced (Table 1). Mental health diagnoses were common with 34.7% of patients having a recent diagnosis of depression and 71.1% having ever had a mental health diagnosis. A quarter of the cohort (26.1%) had concomitant PPI prescriptions, of which 34.9% were dosed twice daily. Only 4.3% of the cohort had prescriptions for H2RAs. Patients with high BMI, diabetes, markers of advanced liver disease such as a history of decompensated liver disease and FIB-4>3.25, prior treatment experience and those receiving PPIs were more likely to have received LDV/SOF+RBV than the other three regimens. Among patients who received LDV/SOF, 5.1% (600/11,804) discontinued treatment early; 3.6% (n=427) discontinued treatment before 8 weeks and 1.5% (n=173) discontinued treatment between 8 and 12 weeks. Early discontinuation occurred in 9.8% (381/3,869) of patients who received LDV/SOF+RBV. Among patients who received OPrD or OPrD+RBV, early discontinuation prior to 12 weeks occurred in 10.5% (150/1,423) and 11.5% (477/4,146) of patients, respectively. 4-week on-treatment HCV RNA testing was performed in 89.6% of the cohort. Over three-quarters (76.6%) were undetectable with 8.2% detectable with HCV RNA<15 IU/ml and 15.3% detectable with HCV RNA 15 IU/ml. These three groups did not differ with regard to baseline characteristics with the exception of race/ethnicity and baseline laboratory data related to their HCV infection. Among those undetectable, detectable <15 IU/ml and detectable 15 IU/ml the percentage of African-Americans increased from 37.1% to 39.9% to 44.6%, respectively. Mean ±sd alanine aminotransferase increased from 71.3 ±53.9 to 73.2 ±53.8 to 77.3 ±59.9, respectively (P<0.001). Mean platelets decreased from ±70.3 to ±67.9 to ±67.2, respectively (P<0.001). The percentage with FIB-4 score >3.25 increased from 29.0% to 31.1% to 38.1%, respectively (P<0.001). The percentage with genotype subtype-1a increased from 71.3% to 73.3% to 74.8%, respectively (P<0.001). SVR results were available for 91.8% (n=19,504) of patients in the cohort, including 110 patients who died while on treatment or shortly after who were categorized as no SVR; 1,738 patients who did not have HCV RNA testing performed at least 10 weeks post-eot were excluded from the SVR analysis. These excluded patients were slightly younger (mean age ± standard deviation [sd] 60.5 ±7.8 years versus 61.8 ±6.2 years, P<0.001) slightly thinner (mean BMI ±sd, 28.0 ±5.2 versus 28.6 ±5.3, P=0.001), less likely to have diabetes (28.4% versus 32.8%, P<0.001), had slightly higher mean platelet counts (190.0 ±68.2 versus ±70.3, P<0.001) and slightly lower mean FIB-4 scores (3.1 ±3.1 versus 3.3 ±3.1, P=0.009) than patients included in the SVR analysis. There were 972 patients with an undetectable HCV RNA obtained 10 weeks but <12 weeks post-eot included in the SVR analysis. SVR was achieved in 91.2% of 10,720 patients receiving LDV/SOF, 89.6% of 3,646 patients receiving LDV/SOF+RBV, 91.7% of 1,306 patients receiving OPrD and in 87.8% of 3,832 patients receiving OPrD+RBV (Table 2). For patients who received LDV/ SOF, but not for other regimens, SVR rates were significantly lower in African-Americans compared to Caucasians (P=0.001), those with lower or higher BMI compared to those with BMI between kg/m 2 (P=0.001 and P<0.001, respectively) and those with diabetes compared to those without diabetes (P=0.002). Among patients receiving LDV/SOF or LDV/SOF+RBV, SVR rates were significantly lower in those with a history of decompensated liver disease compared to those without (P<0.001), those prescribed PPIs compared to those without PPI prescriptions (P<0.001 for LDV/SOF, P=0.008 for LDV/SOF+RBV) and in those with FIB-4 >3.25 compared to 3.25 (P<0.001). No statistically significant differences in SVR were observed according to baseline patient characteristics among patients receiving OPrD. Despite the contraindication of OPrD use in moderate to severe hepatic impairment, 71 patients with a history of decompensated liver disease received OPRD+RBV and SVR rates were significantly lower in those patients compared to those without such a history (P=0.004). For all four regimens, SVR rates were lower in patients with 4-week on-treatment HCV RNA 15 IU/ml compared to patients with undetectable levels (P<0.001 for LDV/SOF, LDV/SOF+RBV and OPrD+RBV, P=0.014 for OPrD). SVR rates according to baseline characteristics for treatment-naive and treatment-experienced patients separately are available in Additional files 1 and 2. SVR rates in patients completing 8 weeks of LDV/ SOF or 12 weeks of LDV/SOF ±RBV or OPRD ±RBV Antiviral Therapy

4 LI Backus et al. Table 1. Baseline characteristics and 4-week on-treatment response of genotype-1 patients receiving LDV/SOF ±RBV or OPrD ±RBV with treatment durations of 12 weeks or less All patients (n=21,242) LDV/SOF (n=11,804) LDV/SOF+RBV (n=3,869) OPrD (n=1,423) OPrD+RBV (n=4,146) Age 61.7 ±6.4 ( ) 61.4 ±6.8 ( ) 62.2 ±5.3 ( ) 62.7 ±6.0 ( ) 61.8 ±6.1 ( ) <55 years 2,040 (9.6) 1,322 (11.2) 256 (6.6) 89 (6.3) 373 (9.0) years 12,841 (60.5) 7,055 (59.8) 2,436 (63.0) 839 (59.0) 2,511 (60.6) 65 years 6,361 (29.9) 3,427 (29.0) 1,177 (30.4) 495 (34.8) 1,262 (30.4) Male sex 20,529 (96.6) 11,341 (96.1) 3,767 (97.4) 1,377 (96.8) 4,044 (97.5) Race/ethnicity African-American 8,276 (39.0) 4,871 (41.3) 1,220 (31.5) 727 (51.1) 1,458 (35.2) Caucasian 10,447 (49.2) 5,704 (48.3) 2,081 (53.8) 557 (39.1) 2,105 (50.8) Hispanic 1,143 (5.4) 530 (4.5) 279 (7.2) 63 (4.4) 271 (6.5) Other/multiple 1,376 (6.5) 699 (5.9) 289 (7.5) 76 (5.3) 312 (7.5) Body mass index 28.5 ±5.3 ( ) 28.3 ±5.3 ( ) 29.4 ±5.4 ( ) 28.1 ±5.0 ( ) 28.4 ±5.3 ( ) <25 kg/m 2 5,472 (25.8) 3,194 (27.1) 794 (20.5) 386 (27.1) 1,098 (26.5) kg/m 2 8,602 (40.5) 4,750 (40.2) 1,534 (39.6) 618 (43.4) 1,700 (41.0) 30 kg/m 2 7,168 (33.7) 3,860 (32.7) 1,541 (39.8) 419 (29.4) 1,348 (32.5) Decompensated liver disease 677 (3.2) 229 (1.9) 364 (9.4) 11 (0.8) 73 (1.8) Diabetes 6,883 (32.4) 3,675 (31.1) 1,510 (39.0) 466 (32.7) 1,232 (29.7) HIV-coinfected 1,092 (5.1) 811 (6.9) 159 (4.1) 32 (2.2) 90 (2.2) History of mental health diagnosis 15,102 (71.1) 8,606 (72.9) 2,791 (72.1) 919 (64.6) 2,786 (67.2) Anxiety 9,529 (44.9) 5,506 (46.6) 1,760 (45.5) 563 (39.6) 1,700 (41.0) Bipolar 2,708 (12.7) 1,633 (13.8) 488 (12.6) 127 (8.9) 460 (11.1) Depression 13,338 (62.8) 7,630 (64.6) 2,499 (64.6) 799 (56.1) 2,410 (58.1) PTSD 6,740 (31.7) 3,893 (33.0) 1,233 (31.9) 410 (28.8) 1,204 (29.0) Schizophrenia 1,968 (9.3) 1,207 (10.2) 320 (8.3) 117 (8.2) 324 (7.8) Mental health diagnosis past year Anxiety 3,632 (17.1) 2,130 (18.0) 656 (17.0) 208 (14.6) 638 (15.4) Bipolar 1,115 (5.2) 718 (6.1) 197 (5.1) 39 (2.7) 161 (3.9) Depression 7,379 (34.7) 4,228 (35.8) 1,434 (37.1) 404 (28.4) 1,313 (31.7) PTSD 4,351 (20.5) 2,546 (21.6) 785 (20.3) 258 (18.1) 762 (18.4) Schizophrenia 838 (3.9) 536 (4.5) 134 (3.5) 51 (3.6) 117 (2.8) Alcohol abuse diagnosis past year 4,902 (23.1) 2,848 (24.1) 875 (22.6) 266 (18.7) 913 (22.0) Hard drug diagnosis past year 1,937 (9.1) 1,225 (10.4) 297 (7.7) 92 (6.5) 323 (7.8) Histamine-2 receptor antagonist 905 (4.3) 509 (4.3) 184 (4.8) 50 (3.5) 162 (3.9) Proton pump inhibitor 5,543 (26.1) 3,011 (25.5) 1,236 (31.9) 328 (23.0) 968 (23.3) Proton pump inhibitor daily doses n=5,543 n=3,011 n=1,236 n=328 n= ,611 (65.1) 1,982 (65.8) 811 (65.6) 199 (60.7) 619 (63.9) 2 1,932 (34.9) 1,029 (34.2) 425 (34.4) 129 (39.3) 349 (36.1) Continuous variables reported as mean ±standard deviation (range). Categorical variables reported as n (%). ALT, alanine aminotransferase; AST, aspartate aminotransferase; DAA, direct-acting antiviral; LDV/SOF, ledipasvir/sofosbuvir; OPrD, ombitasvir/paritaprevir/ritonavir+dasabuvir; PTSD, post-traumatic stress disorder; RBV, ribavirin International Medical Press

5 All-oral HCV treatment in 21,242 patients Table 1. Continued All patients (n=21,242) LDV/SOF (n=11,804) LDV/SOF+RBV (n=3,869) OPrD (n=1,423) OPrD+RBV (n=4,146) Treatment-experienced 4,025 (18.9) 1,463 (12.4) 1,654 (42.8) 232 (16.3) 676 (16.3) DAA-experienced 1,343/4,025 (33.4) 425/1,463 (29.0) 851/1,654 (51.5) 11/232 (4.7) 56/676 (8.3) Prior direct-acting antiviral Simeprevir + sofosbuvir 151 (0.7) 28 (0.2) 120 (3.1) 0 (0.0) 3 (0.1) Sofosbuvir 398 (1.9) 72 (0.6) 304 (7.9) 4 (0.3) 18 (0.4) Boceprevir 886 (4.2) 331 (2.8) 511 (13.2) 7 (0.5) 37 (0.9) Telaprevir 166 (0.8) 45 (0.4) 118 (3.0) 0 (0.0) 3 (0.1) Prior treatment response n=4,025 n=1,463 n=1,654 n=232 n=676 Relapse 1,012 (25.1) 294 (20.1) 573 (34.6) 28 (12.1) 117 (17.3) Partial 422 (10.5) 131 (9.0) 158 (9.6) 41 (17.7) 92 (13.6) Null 533 (13.2) 163 (11.1) 225 (13.6) 42 (18.1) 103 (15.2) Unknown 2,058 (51.1) 875 (59.8) 698 (42.2) 121 (52.2) 364 (53.8) ALT, U/l 71.8 ±54.9 (5 1,062) 68.4 ±53.7 (8 1,062) 79.1 ±54 (10 524) 61.8 ±47.3 (8 552) 78.2 ±60 (5 644) AST, U/l 63.5 ±44.5 (6 614) 59.1 ±41.2 (6-614) 76.6 ±48.9 (10 503) 51.7 ±34.1 (13 334) 67.9 ±48.8 (11 529) Platelets, K/ml ±70.1 (15 857) ±69 (17 707) ±69.3 (15 857) ±60.8 (46 569) ±64.7 (16 588) FIB ±3.1 ( ) 2.8 ±2.6 ( ) 4.9 ±4.3 ( ) 2.3 ±1.5 ( ) 3.3 ±2.8 ( ) >3.25 6,428 (30.4) 2,731 (23.2) 2,079 (53.9) 194 (13.7) 1,424 (34.5) HCV RNA, log IU/ml 6.2 ±0.7 ( ) 6.2 ±0.7 ( ) 6.2 ±0.7 ( ) 6.2 ±0.7 ( ) 6.3 ±0.7 ( ) <6,000,000 IU/ml 17,455 (82.2) 9,905 (83.9) 3,202 (82.8) 1,150 (80.8) 3,198 (77.2) HCV subtype-1b 5,942 (28.0) 2,761 (23.4) 816 (21.1) 1,423 (100.0) 942 (22.7) IL28B polymorphism n=2,629 n=1,434 n=515 n=142 n=538 CC 555 (21.1) 313 (21.8) 81 (15.7) 23 (16.2) 138 (25.7) CT 1,394 (53.0) 748 (52.2) 296 (57.5) 74 (52.1) 276 (51.3) TT 680 (25.9) 373 (26.0) 138 (26.8) 45 (31.7) 124 (23.0) 4-week HCV RNA n=19,041 n=10,402 n=3,617 n=1,264 n=3,758 Undetectable 14,579 (76.6) 8,097 (77.8) 2,694 (74.5) 986 (78.0) 2,802 (74.6) Detectable <15 IU/ml 1,554 (8.2) 861 (8.3) 343 (9.5) 100 (7.9) 250 (6.7) Detectable 15 IU/ml 2,908 (15.3) 1,444 (13.9) 580 (16.0) 178 (14.1) 706 (18.8) Antiviral Therapy

6 LI Backus et al. Table 2. SVR rates by regimen for genotype-1 patients receiving LDV/SOF ±RBV or OPrD ±RBV with treatment durations of 12 weeks or less All patients LDV/SOF LDV/SOF+RBV OPrD OPrD+RBV (n=19,504), % P-value (n=10,720), % P-value (n=3,646), % P-value (n=1,306), % P-value (n=3,832), % P-value Overall SVR 90.3 (17,609/19,504) 91.2 (9,781/10,720) 89.6 (3,266/3,646) 91.7 (1,197/1,306) 87.8 (3,365/3,832) Age <55 years 90.8 (1,634/1,800) 92.5 (1,055/1,140) 90.4 (217/240) 91.1 (72/79) 85.0 (290/341) years 90.1 (10,620/11,782) 91.0 (5,837/6,412) 89.5 (2,041/2,280) 91.1 (697/765) 88.0 (2,045/2,325) 65 years 90.4 (5,355/5,922) 91.2 (2,889/3,168) 89.5 (1,008/1,126) 92.6 (428/462) 88.3 (1,030/1,166) Sex # 0.33 Male 90.2 (16,996/18,848) 91.1 (9,388/10,301) 89.5 (3,176/3,548) 91.4 (1,154/1,262) 87.7 (3,278/3,737) Female 93.4 (613/656) 93.8 (393/419) 91.8 (90/98) 97.7 (43/44) 91.6 (87/95) Race/ethnicity # 0.45 African-American 89.6 (6,789/7,575) 90.3 (3,990/4,417) 88.7 (1,017/1,146) 90.2 (601/666) 87.7 (1,181/1,346) Caucasian 91.0 (8,730/9,595) 92.2 (4,774/5,176) 89.5 (1,753/1,959) 93.9 (480/511) 88.4 (1,723/1,949) Hispanic 89.7 (9,54/1,063) 91.7 (454/495) 89.8 (239/266) 87.3 (48/55) 86.2 (213/247) Other/multiple 89.4 (1,136/1,271) 89.1 (563/632) 93.5 (257/275) 91.9 (68/74) 85.5 (248/290) Body mass index <0.001 < <25 kg/m (4,436/4,962) 90.4 (2,592/2,867) 88.9 (659/741) 90.1 (317/352) 86.6 (868/1,002) kg/m (7,233/7,915) 92.6 (4,011/4,333) 90.8 (1,315/1,449) 92.7 (524/565) 88.2 (1,383/1,568) 30 kg/m (5,940/6,627) 90.3 (3,178/3,520) 88.7 (1,292/1,456) 91.5 (356/389) 88.3 (1,114/1,262) Decompensated liver disease <0.001 <0.001 <0.001 # No 90.6 (17,094/18,873) 91.5 (9,612/10,510) 90.3 (2,985/3,307) 91.6 (1,186/1,295) 88.0 (3,311/3,761) Yes 81.6 (515/631) 80.5 (169/210) 82.9 (281/339) (11/11) 76.1 (54/71) Diabetes < No 90.8 (11,911/13,115) 91.8 (6,739/7,340) 90.2 (2,003/2,220) 92.3 (799/866) 88.1 (2,370/2,689) Yes 89.2 (5,698/6,389) 90.0 (3,042/3,380) 88.6 (1,263/1,426) 90.5 (398/440) 87.1 (995/1,143) HIV coinfection # 0.61 No 90.3 (16,701/18,495) 91.2 (9,099/9,975) 89.7 (3,136/3,495) 91.8 (1,171/1,275) 87.9 (3,295/3,750) Yes 90.0 (908/1,009) 91.5 (682/745) 86.1 (130/151) 83.9 (26/31) 85.4 (70/82) History of mental health diagnosis < No 91.6 (5,201/5,679) 92.3 (2,691/2,917) 90.6 (933/1,030) 94.7 (429/463) 90.5 (1,148/1,269) Yes 89.8 (12,408/13,825) 90.9 (7,090/7,803) 89.2 (2,333/2,616) 91.1 (768/843) 86.5 (423/479) History of mental health diagnosis Anxiety 89.6 (7,787/8,695) (4,508/4,974) (1,470/1,649) (470/511) (1,339/1,561) Bipolar 89.6 (2,182/2,434) (1,323/1,462) (405/447) (96/106) (358/419) 0.14 Depression 89.7 (10,959/12,216) < (6,286/6,923) (2,091/2,346) (662/729) (1,920/2,218) PTSD 89.7 (5,522/6,158) (3,183/3,502) (1,033/1,167) (353/384) (953/1,105) 0.07 Schizophrenia 89.7 (1,606/1,791) (988/1,091) (264/295) (93/104) (261/301) 0.61 DAA, direct-acting antiviral; LDV/SOF, ledipasvir/sofosbuvir; OPrD, ombitasvir/paritaprevir/ritonavir+dasabuvir; PTSD, post-traumatic stress disorder; RBV, ribavirin; SVR, sustained virological response; #, P-values not reported when the minimum expected value in any cell is <5. a 1a includes 1a, mixed 1a/1b and 1 with subtype unspecified International Medical Press

7 All-oral HCV treatment in 21,242 patients Table 2. Continued All patients LDV/SOF LDV/SOF+RBV OPrD OPrD+RBV (n=19,504), % P-value (n=10,720), % P-value (n=3,646), % P-value (n=1,306), % P-value (n=3,832), % P-value Mental health diagnosis past year Anxiety 90.3 (2,983/3,305) (1,731/1,906) (564/616) (184/196) (504/587) 0.13 Bipolar 90.3 (908/1,006) (591/647) (158/177) (29/32) # 86.7 (130/150) 0.76 Depression 89.8 (6,046/6,734) (3,466/3,812) (1,204/1,349) (337/368) (1,039/1,205) 0.05 PTSD 89.3 (3,558/3,985) (2,079/2,287) (658/751) (216/238) (605/709) 0.03 Schizophrenia 89.7 (689/768) (440/483) (111/124) (41/47) # 85.1 (97/114) 0.45 Alcohol abuse diagnosis past year < No 90.8 (13,690/15,085) 91.5 (7,493/8,186) 90.3 (2,564/2,840) 92.0 (979/1,064) 88.6 (2,654/2,995) Yes 88.7% (3,919/4,419) 90.3 (2,288/2,534) 87.1% (702/806) 90.1 (218/242) 84.9 (711/837) Hard drug diagnosis past year < No 90.5 (16,117/17,800) 91.4 (8,826/9,652) 89.9 (3,039/3,381) 91.8 (1,125/1,225) 88.3 (3,127/3,542) Yes 87.6 (1,492/1,704) 89.4 (955/1,068) 85.7 (227/265) 88.9 (72/81) 82.1 (238/290) Histamine-2 receptor antagonist # 0.33 No 90.3 (16,869/18,677) 91.3 (93,71/10,266) 89.6 (3,110/3,472) 91.5 (1,153/1,260) 87.9 (3,235/3,679) Yes 89.5 (740/827) 90.3 (410/454) 89.7 (156/174) 95.7 (44/46) 85.0 (130/153) Proton pump inhibitor <0.001 < No 90.8 (13,051/14,369) 91.8 (7,317/7,969) 90.5 (2,234/2,468) 91.7 (921/1,004) 88.1 (2,579/2,928) Yes 88.8 (4,558/5,135) 89.6 (2,464/2,751) 87.6 (1,032/1,178) 91.4 (276/302) 86.9 (786/904) Proton pump inhibitor daily doses (2,977/3,347) 89.4 (1,621/1,813) 89.5 (691/772) 90.7 (166/183) 86.2 (499/579) (1,581/1,788) 89.9 (843/938) 84.0 (341/406) 92.4 (110/119) 88.3 (287/325) Treatment-experienced No 90.2 (14,160/15,694) 91.1 (8,514/9,341) 89.5 (1,851/2,069) 91.7 (996/1,086) 87.5 (2,799/3,198) Yes 90.5 (3,449/3,810) 91.9 (1,267/1,379) 89.7 (1,415/1,577) 91.4 (201/220) 89.3 (566/634) DAA-experienced versus other # 0.99 treatment-experienced No 90.7 (2,293/2,529) 91.3 (891/976) 90.7 (690/761) 91.4 (191/209) 89.4 (521/583) Yes 90.2 (1,156/1,281) 93.3 (376/403) 88.8 (725/816) 90.9 (10/11) 88.2 (45/51) Prior DAA Simeprevir + sofosbuvir 76.0 (98/129) < (8/13) # 77.0 (87/113) < (3/3) # Sofosbuvir 81.8 (292/357) < (41/51) # 81.2 (233/287) < (4/4) # 93.3 (14/15) # Boceprevir 93.1 (787/845) (300/315) (450/489) (6/7) # 91.2 (31/34) # Telaprevir 89.4 (143/160) (41/43) # 88.6 (101/114) (1/3) # Prior treatment response # 0.34 Relapse 91.2 (883/968) 94.7 (268/283) 89.3 (490/549) 88.9 (24/27) 92.7 (101/109) Partial 90.1 (364/404) 92.2 (118/128) 88.1 (133/151) 87.2 (34/39) 91.9 (79/86) Null 90.3 (463/513) 90.8 (138/152) 88.6 (195/220) 97.5 (39/40) 90.1 (91/101) Unknown 90.3 (1,739/1,925) 91.1 (743/816) 90.9 (597/657) 91.2 (104/114) 87.3 (295/338) Antiviral Therapy

8 LI Backus et al. Table 2. Continued All patients LDV/SOF LDV/SOF+RBV OPrD OPrD+RBV (n=19,504), % P-value (n=10,720), % P-value (n=3,646), % P-value (n=1,306), % P-value (n=3,832), % P-value FIB-4 <0.001 <0.001 < (12,372/13,488) 92.3 (7,542/8,169) 92.9 (1,566/1,685) 92.1 (1,030/1,118) 88.8 (2,234/2,516) > (5,187/5,952) 87.9 (2,216/2,522) 86.7 (1,688/1,947) 88.9 (160/180) 86.2 (1,123/1,303) HCV RNA, log IU/ml <6,000,000 IU/ml 90.4 (14,498/16,032) 91.4 (8,221/8,995) 89.4 (2,693/3,013) 91.9 (970/1,055) 88.0 (2,614/2,969) 6,000,000 IU/ml 89.6 (3,109/3,470) 90.4 (1,560/1,725) 90.5 (572/632) 90.4 (227/251) 87.0 (750/862) HCV subtype < < a a 89.6 (12,589/14,047) 90.8 (7,449/8,208) 89.3 (2,570/2,877) 86.8 (2,570/2,962) 1b 92.0 (5,020/5,457) 92.8 (2,332/2,512) 90.5 (696/769) 91.7 (1,197/1,306) 91.4 (795/870) IL28B polymorphism n=2, n=1, n= n=130 # n= CC 91.2 (474/520) 91.1 (267/293) 96.2 (77/80) 90.9 (20/22) 88.0 (110/125) CT 91.7 (1,192/1,300) 92.6 (636/687) 92.3 (262/284) 94.4 (67/71) 88.0 (227/258) TT 89.9 (558/621) 91.2 (311/341) 89.1 (115/129) 86.5 (32/37) 87.7 (100/114) 4-week HCV RNA n=17,549 <0.001 n=9,500 <0.001 n=3,419 <0.001 n=1,160 # n=3,470 <0.001 Undetectable 94.8 (12,717/13,412) 95.2 (7,013/7,370) 93.6 (2,389/2,552) 97.0 (873/900) 94.3 (2,442/2,590) Detectable <15 IU/ml 93.3 (1,337/1,433) 94.8 (740/781) 89.6 (292/326) 96.8 (91/94) 92.2 (214/232) Detectable 15 IU/ml 88.5 (2,393/2,704) 89.1 (1,202/1,349) 88.5 (479/541) 92.8 (154/166) 86.1 (558/648) International Medical Press

9 All-oral HCV treatment in 21,242 patients are available in Additional file 3. Overall, 5,538 patients potentially qualified for 8 weeks of LDV/ SOF using criteria of treatment-naive, baseline HCV RNA<6 million IU/ml and FIB Among these patients, 3,059 (55.2%) received 8 weeks and 2,479 (44.8%) received 12 weeks. SVR rates among African- Americans who received 8 weeks (92.4%, 1,065/1,153) were significantly lower than Caucasians receiving 8 weeks (95.3%, 1,504/1,578, P=0.002), however SVR rates were similar among African-Americans (95.5%, 1,106/1,158) and Caucasians (95.1%, 1,018/1,071) who qualified for 8 weeks but received 12 weeks (P=0.68). Predictors of SVR from multivariate analysis are shown in Table 3. In the overall cohort comparing all regimens to LDV/SOF, receipt of OPrD+RBV was associated with a decreased odds of SVR (OR 0.70, 95% CI 0.62, 0.80, P<0.001). Female sex was associated with increased odds of SVR (OR 1.56, 95% CI 1.14, 2.19, P=0.007). Genotype subtype-1b was also associated with increased odds of SVR (OR 1.40, 95% CI 1.23, 1.60, P<0.001). Mental health diagnoses in the past or in the year prior to treatment initiation (data not shown) and alcohol abuse diagnosis and hard drug use diagnosis in the past year did not predict SVR. Further modelling of the entire cohort using FIB-4 as a continuous variable indicated that each one unit increase in FIB-4 score was associated with a 0.93 (95% CI 0.92, 0.95) reduction in the odds of achieving SVR (P<0.001). In models limited to patients receiving LDV/SOF±RBV, significant independent predictors of decreased odds of SVR were African-American race (OR 0.80, 95% CI 0.70, 0.92, P<0.001), BMI<25 kg/m 2 (OR 0.77, 95% CI 0.66, 0.90, P<0.001), BMI 30 kg/m 2 (OR 0.77, 95% CI 0.67, 0.89, P<0.001), FIB-4>3.25 (OR 0.60, 95% CI 0.53, 0.69, P<0.001), decompensated liver disease (OR 0.58, 95% CI 0.45, 0.74, P<0.001) and PPIs (OR 0.81, 95% CI 0.71, 0.92, P<0.001). In additional models of LDV/SOF±RBV patients, the odds of achieving SVR were reduced more with twice-daily PPI dosing (OR 0.75, 95% CI 0.62, 0.91, P=0.002) than once-daily PPI dosing (OR 0.84, 95% CI 0.72, 0.98, P=0.02). H2RAs did not impact the odds of SVR among patients receiving LDV/SOF±RBV (OR 0.97, 95% CI 0.74, 1.30, P=0.85). Among patients receiving OPrD ±RBV, FIB-4>3.25 was the only significant negative predictor of SVR (OR 0.72, 95% CI 0.59, 0.88, P<0.001). Genotype subtype- 1b was an independent predictor of increased odds of SVR (OR 1.74, 95% CI 1.33, 2.30, P<0.001). In additional sensitivity analysis, HCV RNA 15 IU/ml compared to undetectable at week 4 on-treatment was associated with similarly reduced odds of achieving SVR in models of the entire cohort (OR 0.45, 95% CI 0.39, 0.53, P<0.001) and in models limited to LDV/SOF ±RBV (OR 0.49, 95% CI 0.41, 0.58, P<0.001) and OPrD ±RBV (OR 0.38, 95% CI 0.29, 0.50, P<0.001). In models limited to patients completing 8 weeks of LDV/SOF or 12 weeks of LDV/SOF ±RBV, BMI 30 kg/m 2 (OR 0.72, 95% CI 0.61, 0.85, P<0.001), FIB 4>3.25 (OR 0.54, 95% CI 0.47, 0.64, P<0.001), decompensated liver disease (OR 0.60, 95% CI 0.45, 0.80, P<0.001) and PPIs (OR 0.75, 95% CI 0.64, 0.87, P<0.001) remained significant independent negative predictors of SVR (Additional file 4). Genotype subtype-1b remained a positive predictor of SVR (OR 1.34, 95% CI 1.13, 1.61, P<0.001) and receipt of LDV/ SOF+RBV compared to LDV/SOF was also associated with a higher likelihood of achieving SVR (OR 1.27, 95% CI 1.06, 1.52, P=0.008). In patients completing 12 weeks of OPrD ±RBV, FIB-4>3.25 remained a significant negative predictor (OR 0.61, 95% CI 0.45, 0.84, P=0.002) and genotype subtype-1b remained a positive predictor (OR 3.95, 95% CI 2.29, 7.42, P<0.001). Discussion In this large diverse population of genotype-1 HCV infected veterans treated in routine medical practice, SVR rates of 88 92% were observed with LDV/ SOF ±RBV and OPrD ±RBV regimens. Though the veteran population has historically been considered difficult to treat because of complex comorbidities, underlying mental health and substance use diagnoses and other social determinants, these data demonstrate that LDV/SOF- and OPrD-based regimens are highly effective in this population notably even in those with underlying mental health diagnoses. This study represents the largest evaluation of LDV/SOF ±RBV and OPrD ±RBV in HCV genotype-1 including over 14,000 and 5,000 patients, respectively. The diversity and size of this cohort provides the opportunity to examine extensively regimen-specific predictors of SVR. Several negative predictors were identified for both LDV/SOF ±RBV and OPrD ±RBV including a history of decompensated liver disease and FIB-4>3.25, which reduced the odds of SVR by 40% among those receiving LDV/ SOF ±RBV and 28% among those receiving OPrD ±RBV. Detectable 4-week on-treatment HCV RNA had the greatest negative impact on achieving SVR for both regimens and was associated with a 51% and 62% reduced odds of achieving SVR in those receiving LDV/SOF ±RBV or OPrD ±RBV regimens, respectively. Compared to treatment with LDV/SOF, treatment with OPrD+RBV was 30% less likely to result in SVR. Historically, HCV RNA levels had been an on-therapy predictor of treatment outcome but their value has been less clear with all-oral genotype-1 regimens. More recently, small studies evaluating the utility of 4-week on-treatment responses with SOF-based regimens have Antiviral Therapy

10 LI Backus et al. Table 3. Significant predictors of SVR in intent-to-treat multivariable models for genotype-1 patients treated with LDV/SOF ±RBV or OPrD ±RBV Overall OR LDV/SOF ±RBV OR OPrD ±RBV OR (95% CI), n=19,440 P-value (95% CI), n=14,323 P-value (95% CI), n=5,117 P-value Females (ref male) 1.56 (1.14, 2.19) (1.02, 2.12) (1.05, 4.27) 0.05 Age <55 (ref 55 64) 0.98 (0.82, 1.17) (0.87, 1.34) (0.57, 1.07) 0.11 Age 65 (ref 55 64) 1.03 (0.92, 1.16) (0.89, 1.17) (0.86, 1.31) 0.58 African-Americans (ref Caucasians) 0.82 (0.73, 0.91) < (0.70, 0.92) < (0.70, 1.05) 0.14 Hispanic (ref Caucasians) 0.93 (0.75, 1.16) (0.77, 1.33) (0.55, 1.14) 0.19 Other/multiple (ref Caucasians) 0.82 (0.68, 1.01) (0.68, 1.12) (0.52, 1.02) 0.06 BMI <25 (ref BMI 25 29) 0.79 (0.69, 0.89) < (0.66, 0.90) < (0.66, 1.04) 0.09 BMI 30 (ref BMI 25 29) 0.83 (0.74, 0.93) (0.67, 0.89) < (0.80, 1.24) 0.98 Decompensated liver disease (ref no) 0.57 (0.46, 0.72) < (0.45, 0.74) < (0.33, 1.04) 0.05 Diabetes (ref no) 0.88 (0.79, 0.98) (0.77, 1.00) (0.71, 1.06) 0.15 Proton pump inhibitor (ref no) 0.84 (0.75, 0.94) < (0.71, 0.92) < (0.77, 1.17) 0.58 Treatment-experienced (ref naive) 1.04 (0.91, 1.19) (0.87, 1.19) (0.87, 1.43) 0.40 FIB-4 >3.25 (ref FIB ) 0.64 (0.57, 0.72) < (0.53, 0.69) < (0.59, 0.88) <0.001 Subtype 1b (ref 1a a ) 1.40 (1.23, 1.60) < (1.12, 1.50) < (1.33, 2.30) <0.001 LDV/SOF+RBV (ref LDV/SOF) 0.99 (0.86, 1.14) (0.88, 1.18) 0.79 OPrD (ref LDV/SOF) 0.75 (0.59, 0.95) 0.02 OPrD+RBV (ref LDV/SOF) 0.70 (0.62, 0.80) < (0.81, 1.56; ref OPrD) 0.49 Anxiety diagnosis (ref no) 0.92 (0.81, 1.03) (0.80, 1.06) (0.72, 1.11) 0.31 Bipolar diagnosis (ref no) 1.01 (0.86, 1.18) (0.86, 1.24) (0.71, 1.31) 0.81 Depression diagnosis (ref no) 0.92 (0.81, 1.04) (0.81, 1.09) (0.68, 1.07) 0.16 PTSD diagnosis (ref no) 0.98 (0.87, 1.11) (0.85, 1.13) (0.79, 1.23) 0.92 Schizophrenia diagnosis (ref no) 1.05 (0.88, 1.26) (0.86, 1.30) (0.75, 1.48) 0.80 Alcohol abuse diagnosis past year 0.91 (0.79, 1.05) (0.79, 1.11) (0.67, 1.12) 0.25 Hard drug diagnosis past year 0.85 (0.69, 1.03) (0.67, 1.08) (0.55, 1.15) 0.21 Results in bold are statistically significant. BMI, body mass index; LDV/SOF, ledipasvir/sofosbuvir; OPrD, ombitasvir/paritaprevir/ritonavir+dasabuvir; OR, odds ratio; PTSD, post-traumatic stress disorder; RBV, ribavirin; SVR, sustained virological response. a 1a includes 1a, mixed 1a/1b and 1 with subtype unspecified. presented conflicting results [12 16]. Because national VA guidance recommends week 4 HCV RNA testing and nearly 90% of patients had such testing, we were able to analyse comprehensively the predictive value of 4-week on-treatment HCV RNA for SVR. 4-week ontreatment HCV RNA levels 15 IU/ml, which constituted above the lower limit of quantification (LLOQ) independently predicted a significantly reduced likelihood of achieving SVR which suggests a potential role for viral load monitoring. Patients with detectable early on-treatment HCV RNA may benefit from review of potential drug interactions, extension of therapy or modification of therapy such as addition of ribavirin, to improve the chance of SVR. For example, in cirrhotic patients treated with LDV/SOF who had quantifiable HCV RNA early in therapy, higher SVR rates were observed with 24 weeks versus 12 weeks of therapy [17]. Thus, clinicians should be aware that monitoring of 4-week on-treatment HCV RNA levels may have clinical utility in predicting treatment outcome and guiding decisions regarding duration or augmentation with ribavirin. More consensus exists regarding the severity of liver disease as a negative predictor of SVR. Our data further indicates that the impact correlates directly with FIB-4 score the more advanced the degree of liver disease the less likely SVR will be achieved. For each one unit increase in FIB-4 score the odds of achieving SVR were reduced by 7%. This is similar to findings observed in other cirrhotic populations whereby SVR was inversely related to pretreatment MELD score [18]. Each point increase in baseline MELD was associated with a 37% increase in the risk of relapse and lower pretreatment MELD score was a strong positive predictor of SVR. FIB-4 score represents an inexpensive easily calculated surrogate marker of degree of liver disease which can be quickly assessed to determine likelihood of SVR and thus decisions regarding potential use of ribavirin and or treatment duration. Ledipasvir is known to exhibit ph-dependent solubility with increased solubility at lower ph, thus coadministration of LDV/SOF with PPIs or H2RAs can reduce ledipasvir gastrointestinal absorption resulting in lower peak plasma concentrations and systemic exposure. Receipt of an H2RA with LDV/SOF ±RBV did not have significant impact on SVR, however, PPI use did. SVR rates reported here were significantly lower in the nearly 4,000 patients who received concomitant International Medical Press

11 All-oral HCV treatment in 21,242 patients PPI prescriptions during LDV/SOF ±RBV treatment; these patients were 20% less likely to achieve SVR compared to those that did not receive concomitant PPI therapy in multivariate models controlling for other clinical characteristics. This is in contrast to another real-world study in which neither low nor high dose PPI was associated with lower SVR [19]. In our study, approximately one-third of patients were prescribed twice-daily PPI which resulted in even lower SVR rates among those receiving LDV/SOF+RBV and reduced the likelihood of achieving SVR by 25%. It is unclear why only LDV/SOF+RBV SVR rates were affected more by twice-daily than once-daily PPI dosing except perhaps that it is a marker of more severe liver disease as PPI use was highly correlated with higher FIB-4 scores. Given the widespread use, the availability of PPIs without a prescription, and the potential negative impact on achieving SVR, prescribed and over-the-counter medications should be thoroughly reviewed with patients prior to prescribing LDV/SOF ±RBV. To maximize the likelihood of treatment success, consideration should be given to permanently or temporarily discontinuing PPIs if possible or to use of an alternative HCV regimen if PPIs are deemed medically necessary. Several other factors were found to negatively impact SVR for patients who received LDV/SOF ±RBV. For African-Americans, the odds of achieving SVR were reduced by 20% compared to Caucasians receiving LDV/SOF ±RBV in multivariate models. Similarly reduced odds of SVR were observed for African- Americans compared to Caucasians even when limited to patients who completed a full treatment course. This observation may be a reflection of the lower SVR rates we observed in African-Americans who received 8 weeks of LDV/SOF compared to Caucasians who received 8 weeks of LDV/SOF. In contrast, SVR rates in African-Americans who received 12 weeks of LDV/ SOF were higher and comparable to SVR rates achieved in Caucasians. African-Americans, particularly those with additional negative predictors, may benefit from a 12-week rather than an 8-week course of LDV/SOF in order to maximize the likelihood of SVR. Additional support for such practice may be inferred from recent reports indicating that response to DAA treatment may be influenced by a combination of negative predictors, such that three or more negative factors may be necessary to significantly impact SVR rates [20]. Similarly, even in the presence of baseline resistance-associated variants, virological failure characteristically only occurred if other negative predictive factors were present [21]. Both high and low BMI were associated with a 23% reduced odds of SVR with LDV/SOF ±RBV treatment, but only high BMI still negatively impacted SVR in patients completing a full treatment course. High BMI has been previously associated with a decreased response to pegylated interferon therapy due to a lower bioavailability and thus less drug exposure [22]. Patients with BMI 30 kg/m 2 also have been shown to have more pronounced steatosis and features of metabolic syndrome which have resulted in lower SVR rates [22,23]. It is possible that these syndromes may have been present in a larger proportion of patients in this cohort thus impacting observed SVR rates. In DAA clinical trials, high BMI did not appear to impact treatment response, however, this may reflect the small numbers of patients with extremes in BMI enrolled in such trials. While reduced likelihood of SVR with higher BMI may be related to metabolic syndrome and steatosis, there is little information in the literature as to why lower BMI may be associated with reduced likelihood of SVR. Perhaps low muscle mass and associated immune dysfunction or malnutrition may contribute, similar to the worse outcomes associated with low BMI in HIV infection [24]. Further study is necessary to determine the potential aetiologies including pharmacokinetic or pharmacodynamics mechanisms for the reduced SVR observed with low and high BMI. Until more is known, weight reduction may provide an important adjunct treatment strategy with oral HCV regimens in patients with high BMI. While this study represents one of the largest realworld cohorts of diverse HCV-infected patients, there are limitations. Baseline resistance testing was not performed during this time, thus we were unable to assess the impact of this factor. Likewise, there may be other factors or characteristics that could have influenced response that we did not examine. While we examined alcohol abuse and hard drug use diagnoses in the past year, the existence of a recent diagnosis code is not a marker of on-going abuse at the time of DAA initiation especially given the often episodic nature of substance abuse. Providers are generally unlikely to begin DAAs with on-going problematic substance abuse. We could not assess either counselling in regard to PPI use or overthe-counter PPI or H2RA use, though because of lower cost most veterans likely receive these medications from the VA. Duration of treatment and early treatment discontinuation rates were determined based on the cumulative quantity dispensed which may overestimate the treatment duration as patients may discontinue treatment even with medication in their possession. Specific reasons for early discontinuation and patient adherence could not be determined from the electronic data. While VA treatment guidance recommended 12 weeks of therapy with OPrD+RBV for GT1a patients with cirrhosis, AASLD/IDSA guidance recommended 24 weeks of therapy in this population. Some providers may have elected to treat such patients for 24 weeks and these patients would not be included in this evaluation as Antiviral Therapy

12 LI Backus et al. we limited it to patients receiving 12-week treatment courses. Finally, approximately 8.2% of patients lacked definitive SVR data and were excluded from the SVR analysis. Given the slightly lower mean FIB-4 score in these excluded patients compared to included patients and given the substantial impact of FIB-4 score on SVR, it is likely that if we had definitive laboratory data on these excluded patients the observed SVR rates would have been higher with little change in the other SVR predictors since the excluded patients otherwise differed little from the included patients. In conclusion, clinical trials with all-oral regimens have generally identified few predictors of SVR in part because of narrow patient selection and small numbers of treatment failures. In this large, diverse real-world cohort, however, numerous predictors of SVR were identified which can be used to help guide treatment decisions and manage expectations. A better understanding of these factors may lead to clinical decisions which ultimately maximize the chance of virological success. Acknowledgements Guarantor of the article: LIB. Specific author contributions: study concept and design: LIB, PSB, TPL and LAM; analysis and interpretation of data: LIB, PSB, TAS, TPL and LAM; drafting of the manuscript: LIB, PSB, critical revision of the manuscript for important intellectual content: LIB, PSB and LAM; statistical analysis: TAS. This statement acknowledges that all authors approved the final version of the article, including the authorship list. This work was prepared independently without financial support. Disclosure statement All authors have no financial, professional or personal disclosures to report. Additional files Additional file 1: A table of sustained virological response rates by regimen for treatment-naive genotype-1 patients receiving LDV/SOF ±RBV or OPrD ±RBV with treatment durations of 12 weeks or less can be found at documents/3966_backus_addfile1.pdf Additional file 2: A table of sustained virological response rates by regimen for treatment-experienced genotype-1 patients receiving LDV/SOF ±RBV or OPrD ±RBV with treatment durations of 12 weeks or less can be found at documents/3966_backus_addfile2.pdf Additional file 3: A table of sustained virological response rates by regimen for genotype-1 patients receiving LDV/ SOF ±RBV or OPrD ±RBV who completed 8 or 12 weeks of LDV/SOF and 12 weeks of all other regimens can be found at documents/3966_backus_addfile3.pdf Additional file 4: A table of significant predictors of sustained virological response in multivariable models for genotype-1 patients treated with LDV/SOF ±RBV or OPrD ±RBV who completed a treatment course can be found at documents/3966_backus_addfile4.pdf References 1. Afdhal N, Zeuzem S, Kwo P, et al. Ledipasvir and sofosbuvir for untreated HCV genotype 1 infection. N Engl J Med 2014; 370: Afdhal N, Reddy KR, Nelson DR, et al. Ledipasvir and sofosbuvir for previously treated HCV genotype 1 infection. N Engl J Med 2014; 370: Ferenci P, Bernstein D, Lalezari J, et al. ABT-450/r-ombitasvir and dasabuvir with or without ribavirin for HCV. N Engl J Med 2014; 370: Poordad F, Hezode C, Trinh R, et al. ABT-450/r-ombitasvir and dasabuvir with ribavirin for hepatitis C with cirrhosis. N Engl J Med 2014; 370: Zeuzem S, Ghalib R, Reddy KR, et al. Grazoprevir-elbasvir combination therapy for treatment-naive Cirrhotic and noncirrhotic patients with chronic hepatitis C virus genotype 1, 4, or 6 infection: a randomized trial. Ann Intern Med 2015; 163: Ghany MG, Nelson DR, Strader DB, et al. An update on treatment of genotype 1 chronic hepatitis C virus infection: 2011 practice guideline by the American Association for the Study of Liver Diseases. Hepatology 2011; 54: Armstrong GL, Wasley A, Simard EP, et al. The prevalence of hepatitis C virus infection in the United States, 1999 through Ann Intern Med 2006; 144: Backus LI, Belperio PS, Loomis TP, et al. Hepatitis C virus screening and prevalence among US veterans in Department of Veterans Affairs care. JAMA Intern Med 2013; 173: Backus LI, Gavrilov S, Loomis TP, et al. Clinical case registries: simultaneous local and national disease registries for population quality management. J Am Med Inform Assoc 2009; 16: Backus LI, Belperio PS, Shahoumian TA, et al. Comparative effectiveness of ledipasvir/sofosbuvir ± ribavirin vs. ombitasvir/paritaprevir/ritonavir + dasabuvir ± ribavirin in 6961 genotype 1 patients treated in routine medical practice. Aliment Pharmacol Ther 2016; 44: Chou R, Wasson N. Blood tests to diagnose fibrosis or cirrhosis in patients with chronic hepatitis C virus infection: a systematic review. Ann Intern Med 2013; 158: Thornton K, Deming P, Manch RA, et al. Is response guided therapy dead? Low cure rates in patients with detectable hepatitis C virus at week 4 of treatment. Hepatol Int 2016; 10: Maasoumy B, Vermehren J, Welker MW, et al. Clinical value of on-treatment HCV RNA levels during different approved sofosbuvir-based antiviral regimens. J Hepatol 2016; 65: Kowdley KV, Nelson DR, Lalezari JP, et al. On-treatment HCV RNA as a predictor of sustained virologic response in HCV genotype 3-infected patients treated with daclatasvir and sofosbuvir. Liver Int 2016; 36: International Medical Press