Received: 11 November 2019 | Revised: 26 February 2020 | Accepted: 28 February 2020 DOI: 10.1111/liv.14429 ORIGINAL ARTICLE Reduction and stabilization of bilirubin with obeticholic acid treatment in patients with primary biliary cholangitis Albert Parés1 | Mitchell Shiffman2 | Victor Vargas3 | Pietro Invernizzi4 | Elizabeth S. Malecha | Alexander Liberman | Leigh MacConell | Gideon Hirschfield6 5 5 5 1 Hospital Clinic, University of Barcelona, IDIBAPS, CIBERehd, Barcelona, Spain Abstract 2 Liver Institute of Virginia, Bon Secours Background & Aims: Total bilirubin is a predictor of survival in primary biliary chol- Mercy Health, Newport News, VA, USA angitis, with the main elevated component being direct bilirubin. The purpose of this 3 Liver Unit, Hospital Vall d'Hebron, Universitat Autònoma de Barcelona, post hoc analysis was to assess the efficacy and safety of obeticholic acid across CIBERehd, Barcelona, Spain quartiles of varying baseline levels of direct bilirubin in the phase 3, randomized, 4 Division of Gastroenterology and Center placebo-controlled Primary Biliary Cholangitis Obeticholic Acid International Study for Autoimmune Liver Diseases, San Gerardo Hospital, Department of Medicine of Efficacy. and Surgery, University of Milano-Bicocca, Methods: This analysis assessed patients on the basis of their baseline direct bilirubin Monza, Italy 5 level (divided by quartile). Biochemistry and safety outcomes were evaluated within Intercept Pharmaceuticals, San Diego, CA, USA each quartile over time. 6 Toronto Centre for Liver Disease, Toronto Results: In the quartile with the highest baseline direct bilirubin (>5.47 µmol/L), there General Hospital and Department of Medicine, University of Toronto, Toronto, was a significant reduction in both direct and total bilirubin at Month 12 compared Canada with placebo. Least squares mean (standard error) change from baseline in direct bili- Correspondence rubin at Month 12 was 4.17 (1.42) µmol/L for placebo, −3.48 (1.63) µmol/L for obet- Albert Parés, Liver Unit, Hospital Clinic, icholic acid 5-10 mg and −3.66 (1.51) µmol/L for obeticholic acid 10 mg (P < .0001, University of Barcelona, Spain, C/. Villarroel 170, 08036-Barcelona. obeticholic acid vs placebo); the corresponding values for total bilirubin at Month Email: pares@ub.edu 12 were 4.38 (1.55) µmol/L for placebo, −4.53 (1.83) µmol/L for obeticholic acid Funding information 5-10 mg and −5.06 (1.64) µmol/L for obeticholic acid 10 mg (P < .0001, obeticholic The POISE study was sponsored by acid vs placebo). Intercept Pharmaceuticals. Professional writing and editorial services were Conclusions: Obeticholic acid treatment was associated with significant reductions provided by Carolyn Green, PhD (MedLogix in total and direct bilirubin, particularly in patients with high baseline direct bilirubin. Communications, LLC), and sponsored by Intercept Pharmaceuticals. Because raised direct bilirubin levels, even within the normal range, are predictive of survival in primary biliary cholangitis, these results suggest substantial benefits of Handling editor: Ana Lleo obeticholic acid in at-risk patients. KEYWORDS biliary, bilirubin, cholestasis, cirrhosis Abbreviations: ALP, alkaline phosphatase; ALT, alanine aminotransferase; AST, aspartate aminotransferase; GGT, gamma-glutamyl transferase; LS, least squares; OCA, obeticholic acid; OLE, open-label extension; PBC, primary biliary cholangitis; POISE, PBC OCA International Study of Efficacy; SE, standard error; UDCA, ursodeoxycholic acid; ULN, upper limit of normal; VAS, visual analog scale. This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited. © 2020 The Authors. Liver International published by John Wiley & Sons Ltd Liver International. 2020;40:1121–1129.  wileyonlinelibrary.com/journal/liv | 1121 1122 | PARÉS et al. 1 | I NTRO D U C TI O N Key points Primary biliary cholangitis (PBC) is a chronic autoimmune liver dis- Treatment with obeticholic acid for 12 months improved ease characterized by inflammation and the progressive destruction or stabilized biochemical markers associated with negative of bile ducts, cholestasis, eventual cirrhosis and death.1,2 The prog- long-term outcomes in patients with primary biliary chol- nostic significance of several liver chemistries has been evaluated angitis, especially in patients who were considered to be and found to be associated with disease progression, including al- at risk of disease progression. kaline phosphatase (ALP), aspartate aminotransferase (AST) and al- bumin, with bilirubin being identified as one of the key predictors of survival and prognosis in patients with PBC.3-5 Rising total bilirubin levels are a late-stage feature of PBC and are inadequate response to UDCA or as a monotherapy in those unable associated with negative long-term outcomes in patients with the to tolerate UDCA. 24-26 While the approval of OCA is based largely on disease.6,7 In an analysis of liver chemistries associated with long- reduction in ALP, beneficial effects in bilirubin have also been noted. term outcomes in patients with PBC, a total bilirubin level >1x the A phase 2, international, randomized, double-blind, placebo-con- upper limit of normal (ULN) was a key inverse predictor of trans- trolled study of OCA as monotherapy showed a decrease from base- plant-free survival (hazard ratio: 5.06); 86% of patients with a total line in direct bilirubin levels in patients treated for 3 months with bilirubin level ≤1xULN survived for 10 years after enrolment com- OCA 10 mg and OCA 50 mg compared with placebo. 27 A phase 2, 7 pared with 41% of patients with levels >1xULN. In addition, it has randomized, double-blind, placebo-controlled trial of patients with been found that total bilirubin values, even within the normal range, PBC who had an inadequate response to UDCA also showed de- may have prognostic significance; for example, Trivedi et al showed creases in direct bilirubin levels in response to OCA 25 mg and OCA that the optimum cut-off value for total bilirubin in predicting liver 50 mg compared with placebo. 28 5 transplantation or death was 0.70xULN. The Global PBC Study In the pivotal, phase 3 PBC OCA International Study of Efficacy Group similarly established that rising total bilirubin levels, even (POISE) trial, daily doses of OCA 5-10 mg resulted in a signifi- within the normal range, are predictive of the 5-year risk of liver cantly greater number of patients achieving the primary endpoint transplantation or death.8 In fact, multiple guidelines, response cri- (ALP<1.67xULN with a ≥15% reduction in ALP from baseline and a teria and risk score calculators used for PBC agree that the stabili- total bilirubin level ≤ULN at Month 12) compared with placebo. 25 zation or reduction in total bilirubin levels corresponds to reduced Compared with a progressive increase in total bilirubin in the pla- risk for disease progression, highlighting the need for PBC treatment cebo group at Month 12 (2 ± 0.70 μmol/L), patients in the OCA 5 mg options that can stabilize or reduce bilirubin levels in patients with titrated to 10 mg (OCA 5-10 mg) and OCA 10 mg groups experienced PBC.4,6,9-17 a significant reduction in total bilirubin (least squares [LS] mean Most studies and prognostic scores report on only total bilirubin, [standard error, SE]: −0.30 [0.70] and −0.90 [0.70] μmol/L respec- although a key disease-related component of total bilirubin is direct tively). 25 OCA treatment therefore prevents increases in bilirubin or conjugated bilirubin.6-7,18-20 Tahtaci et al evaluated various hae- levels in patients with PBC. matological and biochemical parameters in patients with advanced Although total bilirubin is predictive of outcomes, direct biliru- or early-stage PBC and found that, of total and direct bilirubin, only bin is the main elevated component in PBC. As such, this post hoc direct bilirubin level was significantly higher in patients with late- analysis assessed the efficacy and safety of OCA across quartiles of stage disease. 21 Although fractionation of total bilirubin is not a varying baseline levels of direct bilirubin in POISE. routine clinical laboratory assessment at this time, the American College of Gastroenterology's clinical guidelines for the evaluation of abnormal liver chemistries recommends that total serum biliru- 2 | M ATE R I A L S A N D M E TH O DS bin levels be separated into direct and indirect fractions. 22 These guidelines state that direct bilirubin elevations, and not indirect ele- 2.1 | Study design vations, are often present in cholestatic disorders with impairment in bile flow. 22 In PBC, the levels of bilirubin are elevated mainly by its Detailed methods for POISE, including a CONSORT diagram, were direct fraction.20,23 Taken together, these data and recommenda- previously published (ClinicalTrials.gov, NCT01473524). 25 Briefly, tions suggest that direct bilirubin may be a more sensitive marker for patients with PBC and an inadequate response to or intolerance PBC severity than total bilirubin. However, given that total bilirubin of UDCA were randomly assigned to placebo, OCA 5-10 mg, or is more commonly evaluated, it has emerged as the dominant prog- OCA 10 mg in the 12-month, double-blind, placebo-controlled, nostic marker in various multivariate analyses assessing treatment phase 3 study. 25 The primary composite endpoint in POISE was response and natural history of PBC. ALP<1.67xULN with a ≥15% reduction in ALP from baseline and a Obeticholic acid (OCA) is a potent and selective farnesoid X re- total bilirubin level ≤ULN at Month 12. 25 The ULN for total bilirubin ceptor agonist that is indicated for PBC as second-line therapy in was defined as 19.32 μmol/L for women and 25.48 μmol/L for men, combination with ursodeoxycholic acid (UDCA) in adults with an and the ULN for direct bilirubin was defined as 3.42 μmol/L for all PARÉS et al. | 1123 patients. After the double-blind phase, patients could enrol in the 3 | R E S U LT S open-label extension (OLE) phase, in which all patients were treated with OCA. 3.1 | Baseline characteristics In this post hoc analysis, patients in the POISE intent-to-treat population (N = 216) were pooled across treatment groups and Baseline values for most demographic characteristics, UDCA use/ equally divided into quartiles by baseline direct bilirubin levels (as dose and albumin were generally similar across direct bilirubin this is the main elevated component in PBC) to investigate whether quartiles (Table 2). Of the patients who had direct bilirubin levels there was a relationship between baseline direct bilirubin level above the ULN (quartiles 3 and 4), most (83%) had total bilirubin lev- and laboratory measures throughout the course of the study. 20,25 els within the normal range. The number of male patients and ALP Efficacy and safety analyses were evaluated by direct baseline and total bilirubin levels increased progressively by direct bilirubin bilirubin quartiles and treatment groups (Table 1). For comparison quartile, whereas platelet count progressively decreased within in- with results from direct bilirubin quartiles, efficacy was also eval- creasing direct bilirubin quartiles. Baseline values for AST, ALT and uated by total bilirubin quartiles. For the assessment of efficacy GGT did not show clear patterns across quartiles, although the mean response rate, discontinuations were considered non-responders value of each of these markers was highest in quartile 4. as a conservative form of imputation. Analyses were performed for the double-blind phase (initial 12 months); long-term evaluation was performed for the OLE of POISE for up to 36 months of OCA 3.2 | Efficacy treatment and is presented in the supplemental material. For the OLE analyses, patients were evaluated from the time they initiated In a comparison between quartiles, direct bilirubin levels remained OCA treatment; for patients randomly assigned to placebo during stable in quartiles 1 and 2 across all treatment groups (Figure 1A,B). the double-blind phase, the baseline value was the visit immediately In contrast, both dosages of OCA resulted in a reduction in direct prior to OCA initiation. bilirubin levels in quartiles 3 and 4 after 12 months of randomized The POISE protocol was approved by local and national ethics treatment, whereas direct bilirubin levels increased in the placebo and regulatory agencies and was conducted in accordance with Good group (Figure 1C,D). Comparison between treatment groups within Clinical Practice guidelines and the Declaration of Helsinki (Seoul, quartiles 3 and 4 showed a statistically significant difference in LS 27 South Korea, October 2008 amendment). Written informed con- mean change from baseline in direct bilirubin levels at Month 12 rela- sent was obtained from each patient included in the study. tive to placebo (quartile 3: OCA 5-10 mg, −0.26 µmol/L; OCA 10 mg, −0.28 µmol/L; placebo, 1.12 µmol/L, both OCA groups, P < .05; quartile 4: OCA 5-10 mg, −3.48 µmol/L; OCA 10 mg, −3.66 µmol/L; 2.2 | Statistical analyses placebo, 4.17 µmol/L, both OCA groups, P < .0001). Long-term as- sessment throughout 36 months of the OLE also showed the most For the evaluation of the primary endpoint of POISE in each direct pronounced decrease in direct bilirubin levels in direct bilirubin bilirubin quartile, P-values for comparing treatments were obtained quartiles 3 and 4 (Figure S1). using the Cochran-Mantel-Haenszel general association test strati- When total bilirubin was compared across direct bilirubin quar- fied by the randomization strata factor. P-values for comparing ac- tiles, total bilirubin levels remained relatively stable in quartiles 1 and tive treatments vs placebo for direct and total bilirubin, ALP, AST, 2 (Figure 2A,B). In contrast, both dosages of OCA resulted in reduction alanine aminotransferase (ALT) and gamma-glutamyl transferase of total bilirubin levels in quartiles 3 and 4 after 12 months of treat- (GGT) levels, as well as for the pruritus visual analog scale (VAS) ment, whereas total bilirubin levels steadily increased from Months 3 scores, were obtained using an analysis of covariance model with to 12 in the placebo group (Figure 2C,D). The LS mean change in total baseline value as a covariate and fixed effects for treatment and the bilirubin levels from baseline at Month 12 was statistically significant randomization strata factor. for both dosage groups of OCA in direct bilirubin quartile 4 relative TA B L E 1 Baseline direct bilirubin level for each quartile in POISE Direct Bilirubin Cut-off Quartile (µmol/L) (mg/dL) Placebo (n) OCA 5-10 mg (n) OCA 10 mg (n) 1 ≤2.05 ≤0.12 18 20 17 2 2.05 to 3.33 0.12 to 0.20 20 17 18 3 3.33 to 5.47 0.20 to 0.32 17 16 19 4 >5.47 >0.32 18 17 19 Note: Bilirubin values in µmol/L were divided by 17.10 to convert to mg/dL. Abbreviations: OCA, obeticholic acid; PBC, primary biliary cholangitis; POISE, PBC OCA International Study of Efficacy. 1124 | PARÉS et al. TA B L E 2 Baseline characteristics for direct bilirubin quartiles in POISE Quartile 1 (n = 55) Quartile 2 (n = 55) Quartile 3 (n = 52) Quartile 4 (n = 54) Age, years 57.30 (8.35) 55.40 (10.97) 54.90 (12.03) 55.70 (10.47) Female, n (%) 54 (98) 52 (95) 47 (90) 43 (80) Caucasian, n (%) 48 (87) 53 (96) 51 (98) 51 (94) Duration of PBC, years 7.80 (5.67) 8.80 (6.17) 9.40 (5.58) 8.50 (6.70) History of PBC-related 34 (62) 28 (51) 35 (67) 40 (74) pruritus, n (%) UDCA use, n (%) 54 (98) 49 (89) 47 (90) 50 (93) UDCA daily dose, mg/kg 16 (4.61) 15.20 (4.25) 16.70 (6.04) 16 (4.47) ALP, U/L 293.57 (95.03) 312.11 (114.21) 317.66 (88.88) 369.98 (130.11) AST, U/L 41.07 (13.16) 45.04 (25) 43.83 (15.30) 72.03 (34.08) ALT, U/L 45.95 (18.89) 53.28 (33.39) 50.52 (24.33) 81.88 (50.25) GGT, U/L 223.61 (154.46) 210.71 (117.49) 223.35 (164.63) 441.73 (518.85) Total bilirubin, μmol/L 6.08 (2.58) 7.79 (1.69) 11.55 (3.41) 19.11 (6.97) >ULN, n (%) 0 0 2 (4) 16 (30) Direct bilirubin, μmol/L 1.61 (0.15) 2.73 (0.38) 4.23 (0.61) 11.19 (7.05) >ULN, n (%) 0 0 50 (96) 54 (100) Platelets, ×10 9/L 255.19 (83.76) 247.28 (64.75) 234.84 (87.53) 170.81 (75.65) Albumin, g/dL 4.29 (0.25) 4.40 (0.24) 4.39 (0.28) 4.20 (0.37) Note: Data are mean (SD) unless otherwise specified. Quartiles were defined by baseline direct bilirubin levels as follows: quartile 1, ≤2.05 µmol/L; quartile 2, 2.05 to 3.33 µmol/L; quartile 3, 3.33 to 5.47 µmol/L; quartile 4, >5.47 µmol/L. The ULN for total bilirubin was as follows: female: 19.32 μmol/L, male: 25.48 μmol/L; the ULN for direct bilirubin was 3.42 μmol/L for all patients. Abbreviations: ALP, alkaline phosphatase; ALT, alanine aminotransferase; AST, aspartate aminotransferase; GGT, gamma-glutamyl transferase; POISE, Primary Biliary Cholangitis OCA International Study of Efficacy; UDCA, ursodeoxycholic acid; ULN, upper limit of normal. to placebo (OCA 5-10 mg, −4.53 µmol/L; OCA 10 mg, −5.06 µmol/L; of OCA in quartiles 2 and 3 and OCA 5-10 mg in quartile 4, and placebo, 4.38 µmol/L, both OCA groups, P < .0001). There was a this was generally consistent with or greater than the effect in the non-significant reduction in total bilirubin from baseline at Month 12 overall cohort (Figure S6). In a long-term assessment of open-label in quartile 3 (OCA 5-10 mg, −0.84 µmol/L; OCA 10 mg, −0.48 µmol/L; treatment, patients in direct bilirubin quartile 4 who received either placebo, 0.81 µmol/L). Long-term assessment throughout 36 months dose of OCA in the double-blind portion of the study had higher of the OLE showed sustained reductions in total bilirubin levels for responder rates at 12, 24 and 36 months than patients who had direct bilirubin quartiles 3 and 4 (Figure S2). received placebo in the double-blind phase; the other quartiles did Additional analyses of patients in the double-blind phase showed not show clear differences or patterns across groups (Figure S7A-C). results similar to those of direct bilirubin when assessed by baseline In a comparison of all baseline direct bilirubin quartiles, ALT, total bilirubin quartiles (Figure S3) or total bilirubin >0.70xULN, a GGT and AST levels were also reduced throughout 12 months of cut-off value for bilirubin in predicting liver transplantation or death treatment in both OCA dosage groups vs placebo, and within each (Figure S4).5,7 quartile, this effect was generally significant throughout each time Over 12 months of randomized treatment, both dosages of OCA point for each marker (Figure S8-10). In addition, 12 months of treat- resulted in statistically significant reductions in ALP levels across ment with OCA reduced the median estimated risk of death or liver each direct bilirubin quartile relative to placebo (Figure 3). The LS transplantation at years 5, 10 and 15 per the GLOBE score and UK- mean change in ALP from baseline at Month 12 for OCA groups PBC risk score (Figure S11-12). The largest reduction in estimated relative to placebo was statistically significant for each quartile (all risk was observed in quartile 4, where comparison between OCA quartiles, P < .05 for both OCA groups vs placebo). The magnitude and placebo achieved statistical significance (P < .01) for both OCA of the LS mean change in ALP was generally similar for each OCA dosage groups across all time points. dose across quartiles. Long-term assessment throughout 36 months of the OLE also showed reductions in ALP levels for each direct bili- rubin quartile (Figure S5). 3.3 | Safety and tolerability Compared with patients randomly assigned to placebo, more patients randomly assigned to OCA achieved the primary endpoint In patients receiving OCA 5-10 mg during the double-blind phase, of POISE for each baseline direct bilirubin quartile at Month 12, an similar incidences of pruritus were observed in each baseline direct effect that was significant for OCA 10 mg in quartile 1, both doses bilirubin quartile, and they remained generally similar to placebo PARÉS et al. | 1125 (A) Quartile 1: BL DBILI ≤2.05 µmol/L (B) Quartile 2: BL DBILI 2.05 to 3.33 µmol/L 7 7 6 6 5 LS mean (SE) change from baseline 5 LS mean (SE) change from baseline 4 4 Direct Bilirubin (µmol/L) 3 Direct Bilirubin (µmol/L) 3 2 2 1 1 0 * 0 * –1 –1 * –2 –2 –3 –3 –4 –4 Placebo ± UDCA, n = 18 –5 –5 Placebo ± UDCA, n = 20 OCA 5-10 mg ± UDCA, n = 20 –6 OCA 5-10 mg ± UDCA, n = 17 OCA 10 mg ± UDCA, n = 17 –6 OCA 10 mg ± UDCA, n = 18 –7 –7 0 3 6 9 12 0 3 6 9 12 Months Months (C) Quartile 3: BL DBILI 3.33 to 5.47 µmol/L (D) Quartile 4: BL DBILI >5.47 µmol/L 7 7 6 6 5 5 LS mean (SE) change from baseline 4 LS mean (SE) change from baseline 4 Direct Bilirubin (µmol/L) 3 Direct Bilirubin (µmol/L) 3 2 2 1 1 0 * 0 –1 ** ** * * –1 ** * * –2 *** –2 ** –3 –3 –4 –4 *** Placebo ± UDCA, n = 17 ** *** –5 –5 Placebo ± UDCA, n = 18 *** OCA 5-10 mg ± UDCA, n = 16 OCA 5-10 mg ± UDCA, n = 17 –6 –6 OCA 10 mg ± UDCA, n = 19 OCA 10 mg ± UDCA, n = 19 –7 –7 0 3 6 9 12 0 3 6 9 12 Months Months F I G U R E 1 Changes in direct bilirubin in POISE across 12 months of treatment in each direct bilirubin quartile. (A-B) Quartiles 1 and 2 represent patients with normal baseline direct bilirubin levels. (C-D) Quartiles 3 and 4 represent patients with baseline direct bilirubin levels generally above the ULN (defined as 3.42 μmol/L). Levels of significance: *P < .05, **P < .01, ***P < .0001; P-values were obtained using an analysis of covariance model. BL, baseline; DBILI, direct bilirubin; OCA, obeticholic acid; PBC, primary biliary cholangitis; POISE, PBC OCA International Study of Efficacy; UDCA, ursodeoxycholic acid; ULN, upper limit of normal in quartiles 1 and 2; however, the incidence of pruritus gener- to placebo during the first 3-6 months of OCA treatment in direct bil- ally increased with increasing quartiles in the OCA 10 mg group irubin quartiles 3 and 4 (Figure S13). The median change from base- compared with placebo (Table S1). One death unrelated to OCA line in pruritus VAS scores for all treatment groups was ≤1 at all time occurred in a patient receiving OCA 5-10 mg; the patient was in points in the double-blind phase in quartiles 1 and 2 (data not shown). direct bilirubin quartile 4. During the OLE, the exposure-adjusted After 36 months of the OLE, a general increase in pruritus VAS scores treatment-emergent adverse event rate for the total OCA group was observed in direct bilirubin quartiles 3 and 4 (Figure S14). generally increased with increasing direct bilirubin quartile for overall adverse events, pruritus, serious adverse events and hepa- tobiliary disorders (Table S2). Hyperbilirubinemia was reported in 4 | D I S CU S S I O N 1 patient during the OLE in quartile 4 (data not shown). One death occurred in the OLE that was deemed unrelated to OCA.29 4.1 | Summary of the results and their implications In direct bilirubin quartiles 1 and 2, there were no statistical dif- ferences between treatment groups in pruritus VAS scores; pruritus These results suggest that high baseline direct as well as total biliru- VAS scores were statistically higher in the OCA 10 mg group relative bin levels identify the risk for worsening of biochemistry in patients 1126 | PARÉS et al. Quartile 2: BL DBILI 2.05 to 3.33 µmol/L 7 (A) Quartile 1: BL DBILI ≤2.05 µmol/L (B) 7 6 6 5 LS mean (SE) change from baseline 5 4 LS mean (SE) change from baseline 4 3 Total Bilirubin (µmol/L) 3 Total Bilirubin (µmol/L) 2 2 1 1 0 0 –1 –1 * –2 –2 –3 –3 –4 –4 Placebo ± UDCA, n = 20 –5 –5 Placebo ± UDCA, n = 18 OCA 5-10 mg ± UDCA, n = 17 OCA 5–10 mg ± UDCA, n = 20 –6 –6 OCA 10 mg ± UDCA, n = 18 OCA 10 mg ± UDCA, n = 17 –7 –7 0 3 6 9 12 0 3 6 9 12 Months Months Quartile 3: BL DBILI 3.33 to 5.47 µmol/L Quartile 4: BL DBILI >5.47 (C) 7 (D) 7 6 6 5 5 LS mean (SE) change from baseline LS mean (SE) change from baseline 4 4 3 Total Bilirubin (µmol/L) 3 Total Bilirubin (µmol/L) 2 2 1 1 0 0 –1 –1 –2 * –2 * –3 * –3 –4 –4 ** –5 Placebo ± UDCA, n = 17 –5 Placebo ± UDCA, n = 18 *** –6 OCA 5-10 mg ± UDCA, n = 16 –6 OCA 5-10 mg ± UDCA, n = 17 ** *** OCA 10 mg ± UDCA, n = 19 OCA 10 mg ± UDCA, n = 19 –7 –7 0 3 6 9 12 0 3 6 9 12 Months Months F I G U R E 2 Changes in total bilirubin in POISE across 12 months of treatment in each direct bilirubin quartile. (A-B) Quartiles 1 and 2 represent patients with normal baseline direct bilirubin levels. (C-D) Quartiles 3 and 4 represent patients with baseline direct bilirubin levels generally above the ULN (defined as 3.42 μmol/L). Levels of significance: *P < .05, **P < .01, ***P < .0001; P-values were obtained using an analysis of covariance model. BL, baseline; DBILI, direct bilirubin; OCA, obeticholic acid; POISE, Primary Biliary Cholangitis OCA International Study of Efficacy; UDCA, ursodeoxycholic acid; ULN, upper limit of normal with PBC. Patients receiving placebo with baseline direct bilirubin Although total bilirubin is commonly reported as a prognostic >ULN (quartiles 3 and 4) showed evidence of disease progression measure for PBC, data from this study support greater exploration over the 12-month course of the POISE placebo-controlled study, of direct bilirubin levels as a measure of disease progression. 20 The as indicated by rising direct 12-month bilirubin levels. Treatment majority of patients in quartiles 3 and 4 had total bilirubin levels with OCA for 12 months led to significant improvements in direct within the normal range (70% of patients in quartile 4), but their di- and total bilirubin levels in quartile 4 compared with placebo and rect bilirubin levels were above the ULN (mean direct bilirubin level prevented a rise in ALP, AST, ALT and GGT liver chemistries across in quartile 4 was 3.27xULN). These patients showed a decreasing each baseline direct bilirubin quartile. Significant reductions in the trend in platelet count and an increasing trend in ALP level at base- estimated risk of negative outcomes per the GLOBE score and UK- line, corresponding with increasing total and direct bilirubin levels PBC risk score at 5-, 10- and 15-year time points were also observed between quartiles. Analyses of quartiles defined by baseline total in quartile 4 compared with placebo. The results of this study further bilirubin levels showed results similar to those of direct bilirubin support the efficacy of OCA, with observed general reductions from quartiles, and a clear increasing trend was observed in the placebo baseline in each of these parameters regardless of disease severity. group for direct bilirubin more so than total bilirubin in quartile 3, PARÉS et al. | 1127 (A) (B) Quartile 2: BL DBILI 2.05 to 3.33 µmol/L Quartile 1: BL DBILI ≤2.05 µmol/L 100 100 LS mean (SE) change from baseline LS mean (SE) change from baseline 50 50 0 0 ALP (U/L) –50 ALP (U/L) –50 ** –100 –100 ** * ** * ** ** ** ** –150 *** ** ** –150 *** ** ** Placebo ± UDCA, n = 18 ** Placebo ± UDCA, n = 20 –200 –200 OCA 5-10 mg ± UDCA, n = 17 OCA 5-10 mg ± UDCA, n = 20 OCA 10 mg ± UDCA, n = 18 OCA 10 mg ± UDCA, n = 17 –250 –250 0 3 6 9 12 0 3 6 2 Months Months (C) Quartile 3: BL DBILI 3.33 to 5.47 µmol/L (D) Quartile 4: BL DBILI >5.47 µmol/L 100 100 LS mean (SE) change from baseline LS mean (SE) change from baseline 50 50 0 0 * ** ALP (U/L) ALP (U/L) –50 –50 ** ** ** ** ** –100 ** *** –100 ** ** *** *** *** ** *** –150 ** *** –150 *** Placebo ± UDCA, n = 17 Placebo ± UDCA, n = 18 *** –200 –200 OCA 5-10 mg ± UDCA, n = 17 OCA 5-10 mg ± UDCA, n = 16 OCA 10 mg ± UDCA, n = 19 OCA 10 mg ± UDCA, n = 19 –250 –250 0 3 6 2 0 3 6 9 12 Months Months F I G U R E 3 Changes in ALP in POISE across 12 months of treatment in each direct bilirubin quartile. (A-B) Quartiles 1 and 2 represent patients with normal baseline direct bilirubin levels. (C-D) Quartiles 3 and 4 represent patients with baseline direct bilirubin levels generally above the ULN (defined as 3.42 μmol/L). Levels of significance: *P < .05, **P < .01, ***P < .0001; P-values were obtained using an analysis of covariance model. ALP, alkaline phosphatase; BL, baseline; DBILI, direct bilirubin; OCA, obeticholic acid; POISE, Primary Biliary Cholangitis OCA International Study of Efficacy; UDCA, ursodeoxycholic acid; ULN, upper limit of normal further supporting the possible role of direct bilirubin as a more moderate to severe hepatic impairment (Child-Pugh B/C cirrhosis), sensitive measure of disease severity. These data combined suggest and it is important to follow recommended dosing and administra- that direct bilirubin measurements should be considered in a clinical tion provisions with OCA. 24 setting when a patient presents with increasing total bilirubin levels, Evaluation of treatment with OCA in patients with advanced even if the increase in total bilirubin is considered to be within the PBC continues in two ongoing phase 4 clinical studies. The Study normal range. of Obeticholic Acid Evaluating Clinical Outcomes in Patients With Regarding safety, the incidence of treatment-emergent pruri- Primary Biliary Cholangitis (COBALT, study 747-302; EudraCT, tus in patients receiving OCA 5-10 mg was similar to that of pla- 2014-005012-42; ClinicalTrials.gov, NCT02308111) is assessing the cebo in the lower baseline direct bilirubin quartiles (quartiles 1 and long-term clinical outcomes of OCA, and study 747-401 (EudraCT, 2), whereas patients in quartile 4 (with the highest baseline direct 2017-001762-13; ClinicalTrials.gov, NCT03633227) is evaluating bilirubin levels) showed an increased incidence of pruritus-related the pharmacokinetics and safety of OCA in patients with PBC and discontinuations and the largest increase in pruritus severity relative moderate to severe hepatic impairment (Child-Pugh B and Child- to other quartiles, as assessed by the VAS. This group of patients Pugh C cirrhosis).30,31 also had the greatest benefit from OCA treatment per biochemical evaluations. It should be noted that all patients participating in the POISE study had a high risk of disease progression because of an 4.2 | Literature context inadequate response to or intolerance of UDCA; however, the base- line disease severity for all patients was generally mild to moderate, In patients with PBC, the development of hyperbilirubinemia is as- with few patients having compensated cirrhosis (Child-Pugh A cir- sociated with disease progression and impaired bilirubin excretion, rhosis). 25 Incorrect doses of OCA can pose a risk to patients with and the elevation of direct bilirubin may result from obstructed 1128 | PARÉS et al. bile flow, altered bile ductular integrity or reduced production of AC K N OW L E D G E M E N T S 20,32 bile because of defective activity of bile efflux transporters. Technical and editorial support for this manuscript was provided by Results from this study support previous data showing that OCA Carolyn Green, PhD, of MedLogix Communications, LLC, and funded may improve total and direct bilirubin levels25; however, the nov- by Intercept Pharmaceuticals. The authors thank Janet Owens- elty of the analyses performed in this study resides in the in-depth Grillo, PhD, and Juan Carlos Lopez-Talavera, MD, PhD, for their con- evaluation of the effect of OCA across various baseline direct and tributions to this work. The POISE study was sponsored by Intercept total bilirubin levels. The demonstrated benefits of OCA in the Pharmaceuticals. analyses shown here are expected to be clinically significant, as UDCA-treated patients with PBC who had higher total bilirubin C O N FL I C T S O F I N T E R E S T levels, even within the normal range, were shown to have an in- AP has received grant funding, personal fees and advisory board creased risk of liver transplantation or death compared with pa- fees from Intercept Pharmaceuticals; advisory board fees and tients with lower total bilirubin levels. 5,8,33 Our finding that OCA personal fees from Novartis; and personal fees from CymaBay reduces both total and direct bilirubin in patients with elevated Therapeutics and Inova Diagnostics. MLS has received grant fund- direct bilirubin (quartile 4) supports the potential for OCA to delay ing and consulting, advisory board and speaker fees from Intercept the occurrence of outcome events, as a bilirubin level of 1.60xULN Pharmaceuticals; grant funding and consulting and speaker fees is associated with an accelerated increase in bilirubin and a median from Gilead Sciences; grant funding and speaker fees from Enanata time of 19 months to an outcome event.34 Even with total biliru- Pharmaceuticals; grant funding from CymaBay Therapeutics, bin levels well within the normal range, the data presented here Genkyotex, Afimmune, Conatus Pharmaceuticals, Exalenz indicate that direct bilirubin levels may be significantly elevated. Bioscience, Genfit, NGM Biopharmaceuticals, Novartis and Shire; Although many guidelines and established risk criteria include total grant funding and personal fees from AbbVie, Dova Pharmaceuticals bilirubin, lower thresholds within the normal range may need to and Valeant Pharmaceuticals; and personal fees from Bayer, Bristol 4,6,9-17 be considered for clinical use. This is further exemplified by Myers-Squibb, Easai, Daiichi Sankyo, Mallinckrodt, Merck, Optum the inclusion of total bilirubin as a continuous variable within the RX and Shionogi. VV has received speaking fees from Intercept UK-PBC risk score and Global PBC risk score, suggesting that lower Pharmaceuticals. PI has received grant funding and advisory total bilirubin levels within any range contribute to a lower risk of board fees from Intercept Pharmaceuticals; grant funding from 9,16 liver-related/all-cause mortality and liver transplantation. These Bruschettini and Gilead Sciences; grant funding and consultant findings support the use of direct bilirubin measurements as a po- fees from Menarini Diagnostics; consultant fees from Genkyotex tentially more distinct marker of disease progression, which may and International Laboratory; and is co-owner of LIVERA. GMH has shape the standards currently used in the field. However, in clinical received personal fees from Intercept Pharmaceuticals, CymaBay practice, direct bilirubin levels are typically measured only when Therapeutics and GSK; and grant funding from Gilead Sciences and total bilirubin level is above the ULN, a fact that may attenuate this Faulk Pharmaceuticals. ESM, AL and LM are employees of and stock proposal. shareholders for Intercept Pharmaceuticals. E T H I C A L A P P R OVA L S TAT E M E N T 4.3 | Limitations of the study The POISE protocol was approved by local and national ethics and regulatory agencies and was conducted in accordance with Good The analyses performed here are post hoc; therefore, this study Clinical Practice guidelines and the Declaration of Helsinki (Seoul, was not specifically powered to detect differences within quar- South Korea, October 2008 amendment). tiles. In addition, data presented from the OLE did not include a placebo group for comparison to account for natural disease pro- PAT I E N T C O N S E N T S TAT E M E N T gression. The study population was not representative of one with Written informed consent was obtained from each patient included a high incidence of jaundice; jaundice occurs at total bilirubin levels in the study. above 51.30 μmol/L, and the highest total bilirubin quartile in POISE showed a mean level of 19.12 μmol/L.35 ORCID Albert Parés https://orcid.org/0000-0002-5413-9687 Pietro Invernizzi https://orcid.org/0000-0003-3262-1998 4.4 | Strengths of the study REFERENCES The analyses presented here support the benefits of OCA for vari- 1. Hirschfield GM, Gershwin ME. The immunobiology and pathophys- ous PBC liver chemistries, including total and direct bilirubin levels, iology of primary biliary cirrhosis. Annu Rev Pathol. 2013;8:303-330. 2. Lindor KD, Bowlus CL, Boyer J, Levy C, Mayo M. Primary biliary even in at-risk patient populations. The data used for these analyses cholangitis: 2018 practice guidance from the American Association were collected prospectively as part of an international, randomized, for the Study of Liver Diseases. Hepatology. 2019;69(1):394-419. placebo-controlled, phase 3 study. PARÉS et al. | 1129 3. Kouroumalis E, Samonakis D, Voumvouraki A. Biomarkers for pri- 21. Tahtaci M, Yurekli OT, Bolat AD, et al. Increased mean platelet vol- mary biliary cholangitis: current perspectives. Hepat Med. 2018;10: ume is related to histologic severity of primary biliary cirrhosis. Eur 43-53. J Gastroenterol Hepatol. 2015;27(12):1382-1385. 4. Lindor KD, Gershwin ME, Poupon R, Kaplan M, Bergasa NV, 22. Kwo PY, Cohen SM, Lim JK. ACG clinical guideline: evaluation of Heathcote EJ. Primary biliary cirrhosis. Hepatology. 2009;50(1): abnormal liver chemistries. Am J Gastroenterol. 2017;112(1):18-35. 291-308. 23. Wang H, Xu H, Wang X, et al. Red blood cell distribution width to 5. Trivedi PJ, Bruns T, Cheung A, et al. Optimising risk stratification platelet ratio is related to histologic severity of primary biliary cir- in primary biliary cirrhosis: AST/platelet ratio index predicts out- rhosis. Medicine (Baltimore). 2016;95(11):e3114. come independent of ursodeoxycholic acid response. J Hepatol. 24. Intercept Pharmaceuticals. OCALIVA [package insert]. New York, 2014;60(6):1249-1258. NY: Intercept Pharmaceuticals; 2018. 6. European Association for the Study of the Liver. EASL Clinical 25. Nevens F, Andreone P, Mazzella G, et al. A placebo-controlled Practice Guidelines: the diagnosis and management of patients with trial of obeticholic acid in primary biliary cholangitis. N Engl J Med. primary biliary cholangitis. J Hepatol. 2017;67(1):145-172. 2016;375(7):631-643. 7. Lammers WJ, van Buuren HR, Hirschfield GM, et al. Levels of al- 26. Pellicciari R, Fiorucci S, Camaioni E, et al. 6alpha-ethyl-chenodeoxy- kaline phosphatase and bilirubin are surrogate end points of out- cholic acid (6-ECDCA), a potent and selective FXR agonist endowed comes of patients with primary biliary cirrhosis: an international with anticholestatic activity. J Med Chem. 2002;45(17):3569-3572. follow-up study. Gastroenterology. 2014;147(6):1338-1349.e1335; 27. Kowdley KV, Luketic V, Chapman R, et al. A randomized trial of obe- quiz e1315. ticholic acid monotherapy in patients with primary biliary cholangi- 8. Murillo Perez CF, Gulamhusein A, van Buuren HR, et al. Bilirubin tis. Hepatology. 2018;67(5):1890-1902. is predictive of transplant-free survival even within the nor- 28. Hirschfield GM, Mason A, Luketic V, et al. Efficacy of obeticholic mal range in primary biliary cholangitis patients. Hepatology. acid in patients with primary biliary cirrhosis and inadequate re- 2017;66:41A-42A. sponse to ursodeoxycholic acid. Gastroenterology. 2015;148(4):751- 9. Carbone M, Sharp SJ, Flack S, et al. The UK-PBC risk scores: deri- 761.e758. vation and validation of a scoring system for long-term prediction 29. Trauner M, Nevens F, Shiffman ML, et al. Long-term efficacy and of end-stage liver disease in primary biliary cholangitis. Hepatology. safety of obeticholic acid for patients with primary biliary cholan- 2016;63(3):930-950. gitis: 3-year results of an international open-label extension study. 10. Corpechot C, Abenavoli L, Rabahi N, et al. Biochemical response Lancet Gastroenterol Hepatol. 2019;4(6):445-453. to ursodeoxycholic acid and long-term prognosis in primary biliary 30. Clinicaltrials.gov. Phase 4 study of obeticholic acid evaluating clini- cirrhosis. Hepatology. 2008;48(3):871-877. cal outcomes in patients with primary biliary cholangitis (COBALT). 11. Corpechot C, Chazouilleres O, Poupon R. Early primary biliary cir- 2019. https://clini​c altr​ials.gov/ct2/show/NCT02​3 08111. Accessed rhosis: biochemical response to treatment and prediction of long- February 20, 2019 term outcome. J Hepatol. 2011;55(6):1361-1367. 31. Clinicaltrials.gov. Study of OCA evaluating pharmacokinetics and 12. Freeman RB Jr, Wiesner RH, Harper A, et al. The new liver alloca- safety in patients with PBC and hepatic impairment. 2019. https:// tion system: moving toward evidence-based transplantation policy. clini​c altr​ials.gov/ct2/show/NCT03​633227. Accessed February 20, Liver Transpl. 2002;8(9):851-858. 2019. 13. Hirschfield GM, Dyson JK, Alexander GJM, et al. The British Society 32. Levitt DG, Levitt MD. Quantitative assessment of the multiple pro- of Gastroenterology/UK-PBC primary biliary cholangitis treatment cesses responsible for bilirubin homeostasis in health and disease. and management guidelines. Gut. 2018;67(9):1568-1594. Clin Exp Gastroenterol. 2014;7:307-328. 14. Kilmurry MR, Heathcote EJ, Cauch-Dudek K, et al. Is the Mayo 33. Gordon S, Rodriguez C, Romanelli R, Haller I. Serum bilirubin within model for predicting survival useful after the introduction of normal range is associated with an increasing risk of mortality in pa- ursodeoxycholic acid treatment for primary biliary cirrhosis? tients with primary biliary cholangitis regardless of ursodeoxycholic Hepatology. 1996;23(5):1148-1153. acid treatment. Hepatology. 2018;68(1):31A. 15. Kuiper EMM, Hansen BE, de Vries RA, et al. Improved prog- 34. Harms MH, Pares A, Mason AL, et al. Behavioral patterns of total nosis of patients with primary biliary cirrhosis that have a bio- serum bilirubin prior to major clinical endpoints in 3529 patients chemical response to ursodeoxycholic acid. Gastroenterology. with primary biliary cholangitis. J Hepatol. 2016;64(2):S633-S634. 2009;136(4):1281-1287. 35. Fargo MV, Grogan SP, Saguil A. Evaluation of jaundice in adults. Am 16. Lammers WJ, Hirschfield GM, Corpechot C, et al. Development and Fam Physician. 2017;95(3):164-168. validation of a scoring system to predict outcomes of patients with primary biliary cirrhosis receiving ursodeoxycholic acid therapy. Gastroenterology. 2015;149(7):1804-1812.e1804. 17. Mayo MJ, Parkes J, Adams-Huet B, et al. Prediction of clinical out- S U P P O R T I N G I N FO R M AT I O N comes in primary biliary cirrhosis by serum enhanced liver fibrosis Additional supporting information may be found online in the assay. Hepatology. 2008;48(5):1549-1557. Supporting Information section. 18. Bowlus CL. Obeticholic acid for the treatment of primary biliary cholangitis in adult patients: clinical utility and patient selection. Hepat Med. 2016;8:89-95. How to cite this article: Parés A, Shiffman M, Vargas V, et al. 19. Jung HE, Jang JY, Jeong SW, et al. Prognostic indicators in pri- mary biliary cirrhosis: significance of revised IAHG (International Reduction and stabilization of bilirubin with obeticholic acid Autoimmune Hepatitis Group) score. Clin Mol Hepatol. 2012;18(4): treatment in patients with primary biliary cholangitis. Liver 375-382. Int. 2020;40:1121–1129. https://doi.org/10.1111/liv.14429 20. Reshetnyak VI. Primary biliary cirrhosis: clinical and laboratory crite- ria for its diagnosis. World J Gastroenterol. 2015;21(25):7683-7708.