Cholestasis response was defined as alkaline phosphatase (ALP) < 1.67 x upper limit of normal (ULN) and total bilirubin (TB) <= ULN and ALP decrease from baseline >= 15% at Week 52 and based on the composite strategy imputing non-response for participants who experienced intercurrent events (ICEs) (study treatment discontinuation or use of rescue therapy for PBC) prior to Week 52.

The ITT population included all randomized participants.

Response to treatment based on normalization of ALP at Week 52 was defined as percentage of participants with ALP =<1.0× ULN and based on the composite strategy imputing non-response for participants who experienced ICEs (study treatment discontinuation or use of rescue therapy for PBC) prior to Week 52.

The ITT population included all randomized participants.

The PBC Worst Itch NRS is a simple, self-administered patient reported outcome (PRO) questionnaire that measures itch intensity. Participants rate intensity of their worst Itch on an 11-point scale ranging from 0 (no itch) to 10 (worst itch imaginable). Total score ranged from 0 (no itch) to 10 (worst itch imaginable). Higher scores indicated worse outcomes. Baseline was defined as the average of the daily PBC Worst itch scores on available data within 14 days prior the day of randomization. Post-baseline data was defined as the average of the daily PBC Worst itch scores on available data in 4-week period intervals (available data every 28 days after the day of first study drug intake).

The Pruritus ITT population included all participants from the ITT analysis set with baseline PBC Worst Itch NRS score ≥4. Only those participants with data collected at Week 52 are reported.

The PBC Worst Itch NRS is a simple, self-administered PRO questionnaire that measures itch intensity. Participants rate intensity of their worst Itch on an 11-point scale ranging from 0 (no itch) to 10 (worst itch imaginable). Total score ranged from 0 (no itch) to 10 (worst itch imaginable). Higher scores indicated worse outcomes. Baseline was defined as the average of the daily PBC Worst itch scores on available data within 14 days prior the day of randomization. Post-baseline data was defined as the average of the daily PBC Worst itch scores on available data in 4-week period intervals (available data every 28 days after the day of first study drug intake).

The Pruritus ITT population included all participants from the ITT analysis set with baseline PBC Worst Itch NRS score ≥4. Only those participants with data collected at Week 24 are reported.

Blood samples were collected at specified timepoints for assessment of ALP. Baseline was defined as the last non-missing central value on or before the first dose of the randomized study medication.

The ITT population included all randomized participants. Only those participants with data collected at specified timepoints are reported.

ALP response was defined as ALP decrease from baseline >= 10%; ALP decrease from baseline >= 20% and ALP decrease from baseline >= 40% at Week 52 and based on the composite strategy imputing non-response for participants who experienced ICEs (study treatment discontinuation or use of rescue therapy for PBC) prior to Week 52. Baseline was defined as the last non-missing central value on or before the first dose of the randomized study medication.

The ITT population included all randomized participants. Only those participants with data collected at Week 52 are reported.

Response to treatment was according to ALP < 1.5x ULN, ALP decrease from baseline >= 40% and TB =\

The ITT population included all randomized participants. Only those participants with data collected at Week 52 are reported.

Response to treatment was according to ALP < 3x ULN, AST < 2x ULN and TB =< 1 mg/dL (Paris I) at Week 52. It was based on the composite strategy imputing non-response for participants who experienced ICEs (study treatment discontinuation or use of rescue therapy for PBC) prior to Week 52.

The ITT population included all randomized participants. Only those participants with data collected at Week 52 are reported.

Response to treatment was according to ALP =< 1.5x ULN, AST =< 1.5x ULN and TB =< 1 mg/dL (Paris II) at Week 52. It was based on the composite strategy imputing non-response for participants who experienced ICEs (study treatment discontinuation or use of rescue therapy for PBC) prior to Week 52.

The ITT population included all randomized participants. Only those participants with data collected at Week 52 are reported.

Response to treatment was according to TB decrease of 15% change from baseline at Week 52. It was based on the composite strategy imputing non-response for participants who experienced ICEs (study treatment discontinuation or use of rescue therapy for PBC) prior to Week 52. Baseline was defined as the last non-missing central value on or before the first dose of the randomized study medication.

The ITT population included all randomized participants. Only those participants with data collected at Week 52 are reported.

Response to treatment was according to TB (TB =< ULN) and/or albumin (ALB >= LLN) (Rotterdam) at week 52. Participant was considered as responder at Week 52 only if total bilirubin and albumin were normal. The endpoint was analyzed in participants with abnormal total bilirubin (TB > ULN) or albumin (ALB < LLN) at baseline. It was based on the composite strategy imputing non-response for participants who experienced ICEs (study treatment discontinuation or use of rescue therapy for PBC) prior to Week 52.

The ITT population included all randomized participants. Only those participants with data collected at Week 52 are reported.

Response to treatment was according to TB ≤0.6 x ULN at Week 52. It was based on the composite strategy imputing non-response for participants who experienced ICEs (study treatment discontinuation or use of rescue therapy for PBC) prior to Week 52.

The ITT population included all randomized participants. Only those participants with data collected at Week 52 are reported.

Response to treatment was according to ALP ≤1.67 x ULN and TB ≤1 mg/dL (Momah/ Lindor) at Week 52. It was based on the composite strategy imputing non-response for participants who experienced ICEs (study treatment discontinuation or use of rescue therapy for PBC) prior to Week 52.

The ITT population included all randomized participants. Only those participants with data collected at Week 52 are reported.

Response to treatment was according to no worsening of TB, which was defined as level of TB \

The ITT population included all randomized participants. Only those participants with data collected at Week 52 are reported.

Response to treatment according to complete biochemical response which was defined as <= ULN values of ALP, TB, AST, ALT and ALB, and international normalized ratio (INR) at Week 52. It was based on the composite strategy imputing non-response for participants who experienced ICEs (study treatment discontinuation or use of rescue therapy for PBC) prior to Week 52.

The ITT population included all randomized participants. Only those participants with data collected at Week 52 are reported.

The UK-PBC risk score was defined by the mean percentage risk that a PBC participant treated with UDCA would develop liver failure requiring liver transplantation in 5, 10, and 15 years from diagnosis. It is calculated from the following equation at week 52 as follows: UK-PBC score = 0.0287854*(alpEPxULN-1.722136304) - 0.0422873*([{altastEP x ULN/10}^-1] - 8.675729006) + 1.4199 * (LN(bilEPxULN /10)+2.709607778) - 1.960303*(alb x LLN -1.17673001)-0.4161954*(plt x LLN -1.873564875). The survival S(t) for any given participants was then calculated by S(t) = S0(t) exp(UK-PBC score). Here, alpEP x ULN=ALP /Upper Level Normal ALP at the timepoint, altastEPxULN=(ALT, AST or TA) /upper level normal of the value at the timepoint, bilEP x ULN=bilirubin /upper level normal bilirubin at the timepoint, alb x LLN=alb /alb lower level normal at baseline and plt x LLN=plt/plt lower level normal at baseline. Lower UK-PBC score predicts lower risk and better prognosis.

The ITT population included all randomized participants. Only those participants with data collected at Week 52 are reported.

The GLOBE score was a validated risk assessment tool providing an estimate of transplant-free survival for participants with PBC. It was developed by the Global PBC Study Group using Cox regression model on over 4,000 participants with PBC. It is calculated from the following equation at week 52 as follows: GLOBE score = 0.044378 * age at start of ursodeoxycholic acid (UDCA) therapy + 0.93982 * LN (TB x ULN) + 0.335648 * LN (ALP x ULN) - 2.266708 * ALB x LLN - 0.002581 * platelet count per 10^9/L) + 1.216865; where TB x ULN = bilirubin/ULN at the timepoint, ALP x ULN = ALP/ULN at the timepoint, ALB x LLN = ALB/LLN at the timepoint, platelet count per 10^9/L = platelet count per 10^9/L at the timepoint. The Lower GLOBE PBC score predicts lower risk and better prognosis.

The ITT population included all randomized participants. Only those participants with data collected at Week 52 are reported

Hepatobiliary injury and liver function were assessed as measured by AST, ALT, GGT, 5'-NT, and fractionated ALP (hepatic) (H1 and H2). Baseline was defined as the last non-missing central value on or before the first dose of the randomized study medication.

The ITT population included all randomized participants. Only those participants with data collected at Week 52 for each parameter are reported.

Hepatobiliary injury and liver function were assessed as measured by total and conjugated bilirubin. Baseline was defined as the last non-missing central value on or before the first dose of the randomized study medication.

The ITT population included all randomized participants. Only those participants with data collected at Week 52 for each parameter are reported.

Hepatobiliary injury and liver function were assessed as measured by albumin. Baseline was defined as the last non-missing central value on or before the first dose of the randomized study medication.

The ITT population included all randomized participants. Only those participants with data collected at Week 52 are reported.

Hepatobiliary injury and liver function were assessed as measured by INR. Baseline was defined as the last non-missing central value on or before the first dose of the randomized study medication.

The ITT population included all randomized participants. Only those participants with data collected at Week 52 are reported.

Biomarkers of inflammation were assessed as measured by hsCRP. Baseline was defined as the last non-missing central value on or before the first dose of the randomized study medication.

The ITT population included all randomized participants. Only those participants with data collected at Week 52 are reported.

Biomarkers of inflammation were assessed as measured by fibrinogen and haptoglobin. Baseline was defined as the last non-missing central value on or before the first dose of the randomized study medication.

The ITT population included all randomized participants. Only those participants with data collected at Week 52 for each parameter are reported.

Biomarkers of inflammation were assessed as measured by TNF-alpha. Baseline was defined as the last non-missing central value on or before the first dose of the randomized study medication.

The ITT population included all randomized participants. Only those participants with data collected at Week 52 are reported.

Immune response was assessed as measured by IgG and IgM. Baseline was defined as the last non-missing central value on or before the first dose of the randomized study medication.

The ITT population included all randomized participants. Only those participants with data collected at Week 52 for each parameter are reported.

Biomarkers and non-invasive measures of hepatic fibrosis were assessed as measured by ELF: hyaluronic acid [HA], type 2 procollagen peptide [PIINP], tissue inhibitor of metalloproteinases-1[TIMP-1]) and PAI-1. Baseline was defined as the last non-missing central value on or before the first dose of the randomized study medication.

The ITT population included all randomized participants. Only those participants with data collected at Week 52 for each parameter are reported.

Biomarkers and non-invasive measures of hepatic fibrosis were assessed as measured by CK-18 (M30 and M65). Baseline was defined as the last non-missing central value on or before the first dose of the randomized study medication.

The ITT population included all randomized participants. Only those participants with data collected at Week 52 for each parameter are reported.

Biomarkers and non-invasive measures of hepatic fibrosis were assessed as measured by TGF-beta. Baseline was defined as the last non-missing central value on or before the first dose of the randomized study medication.

The ITT population included all randomized participants. Only those participants with data collected at Week 52 are reported.

FibroScan® TE device (Echosens, Paris, France) is a non-invasive technique used to measure liver stiffness, which correlated with fibrosis. Baseline was defined as the last non-missing value on or before the first dose of the randomized study medication.

The ITT population included all randomized participants. Only those participants with data collected at Week 52 are reported.

Blood samples were collected at specified timepoints for assessment of total cholesterol, low-density lipoprotein (LDL) cholesterol, high-density lipoprotein (HDL) cholesterol, very low-density lipoprotein (VLDL) cholesterol and triglycerides (TG). Baseline was defined as the last non-missing central value on or before the first dose of the randomized study medication.

The ITT population included all randomized participants. Only those participants with data collected at Week 52 for each parameter are reported.

Blood samples were collected at specified timepoints for assessment of FPG. Baseline was defined as the last non-missing central value on or before the first dose of the randomized study medication.

The ITT population included all randomized participants. Only those participants with data collected at Week 52 are reported.

Blood samples were collected at specified timepoints for assessment bile acids and biomarkers of bile acid synthesis. Baseline was defined as the last non-missing central value on or before the first dose of the randomized study medication.

The ITT population included all randomized participants. Only those participants with data collected at Week 52 for each parameter are reported.

Blood samples were collected at specified timepoints for assessment of bile acids and biomarkers of bile acid synthesis as measured by C4. Baseline was defined as the last non-missing central value on or before the first dose of the randomized study medication.

The ITT population included all randomized participants. Only those participants with data collected at Week 52 are reported.

Blood samples were collected at specified timepoints for assessment bile acids and biomarkers of bile acid synthesis as measured by Fibroblast Growth Factor 19. Baseline was defined as the last non-missing central value on or before the first dose of the randomized study medication.

The ITT population included all randomized participants. Only those participants with data collected at Week 52 are reported.

The PBC Worst Itch NRS is a simple, self-administered PRO questionnaire that measures itch intensity. Participants rate intensity of their worst Itch on an 11-point scale ranging from 0 (no itch) to 10 (worst itch imaginable). Total score ranged from 0 (no itch) to 10 (worst itch imaginable). Higher scores indicated worse outcomes. Response to PBC Worst Itch NRS was according to clinically meaningful change at least 30% reduction; one point, two points or three points decrease in score from baseline in participants with a baseline NRS score ≥4. Baseline was defined as the average of the daily PBC Worst itch scores on available data within 14 days prior the day of randomization. Post-baseline data was defined as the average of the daily PBC Worst itch scores on available data in 4-week period intervals (available data every 28 days after the day of first study drug intake).

The Pruritus ITT population included all participants from the ITT analysis set with baseline PBC Worst Itch NRS score ≥4. Only those participants with data collected at Weeks 24 and 52 are reported.

The PBC Worst Itch NRS is a simple, self-administered PRO questionnaire that measures itch intensity. Participants rate intensity of their worst Itch on an 11-point scale ranging from 0 (no itch) to 10 (worst itch imaginable). Total score ranged from 0 (no itch) to 10 (worst itch imaginable). Higher scores indicated worse outcomes. Baseline was defined as the average of the daily PBC Worst itch scores on available data within 14 days prior the day of randomization. Post-baseline data was defined as the average of the daily PBC Worst itch scores on available data in 4-week period intervals (available data every 28 days after the day of first study drug intake). A worsening of Pruritus was defined by a positive change from baseline in PBC Worst Itch NRS score greater than 2 at week 24 and 52.

The Pruritus ITT population included all participants from the ITT analysis set with baseline PBC Worst Itch NRS score ≥4.. Only those participants with data collected at Week 24 and Week 54 are reported.

The 5-D Itch scale is a questionnaire that has been validated in several different diseases. It assesses symptoms in terms of 5 domains: duration: score range 1 (less than 6 hours) to 5 (all day); degree: score range 1 (not present) to 5 (unbearable); duration: score range 1 (less than 6 hours) and 5 (all day); direction: score range 1(completely resolved) and 5 (getting worse); disability: score range 1 (never affects or not applicable) and 5 (delays falling asleep and frequently wakes up at night or always affects this activity) and distribution: where the question was to mark where itching is present in body parts over the last two weeks. Participants rated their symptoms over preceding 2-week period on a 1 to 5 scale, with 5 being the most affected. The 5-D Itch total score ranged between 5 (no pruritus) and 25 (most severe pruritus). Higher scores indicated worse outcomes. Baseline= last non-missing value on or before the first dose of the randomized study medication.

The ITT population included all randomized participants. Only those participants with data collected at Week 52 are reported.

The PROMIS Fatigue Short Form 7a consists of seven items that measure both the experience of fatigue and the interference of fatigue on daily activities over the past week. Participants completed the PROMIS Fatigue Short Form 7a and answered following questions on a scale from 1 (never) to 5 (always) : "How often did you feel tired"; "How often did you experience extreme exhaustion"; "How often did you run out of energy"; "How often did your fatigue limit you at work"; "How often were you too tired to think clearly"; "How often were you too tired to take a bath or shower" and 'How often did you have enough energy to exercise strenuously". T-score values include mean of 50 and standard deviation of 10. The PROMIS Fatigue T-score ranged from 29.4 (best) to 83.2 (worst). Higher scores indicated worse outcomes. Baseline was defined as the last non-missing value on or before the first dose of the randomized study medication.

The ITT population included all randomized participants. Only those participants with data collected at Week 52 are reported.

The ESS is a short, self-administered questionnaire that consists of 8 questions asking to rate how likely it is to fall asleep in everyday situations (each question can be scored from 0 to 3 points; '0' indicates no chance of dozing and '3' indicates high chance of dozing). It provides a total score which has been shown to relate to the participant's level of daytime sleepiness (total score range 0 [no sleepiness] to 24 points [significant sleepiness]). The total score is the sum of all item scores and ranged from 0 (best) to 24 (worst). Higher scores indicated more chances of sleepiness. Participants were asked to complete the ESS with regard to the level of sleepiness they experienced over approximately the past 7 days. Baseline was defined as the last non-missing value on or before the first dose of the randomized study medication.

The ITT population included all randomized participants. Only those participants with data collected at Week 52 are reported.

PBC-40 assess symptoms across 6 domains:symptoms,itch,fatigue,cognitive,emotional and social. Participants responded on verbal response scale,depending on section options range from 'never'/'not at all'/'strongly disagree' to 'always'/'very much'/'strongly agree'.5 items (3/3 in itch domain and 2/10 in social domain)also included a 'does not apply' option.Domains:Symptoms(7 questions)score range 7-35,Itch(3 questions)score range 3-15,Fatigue(11 questions)score range 11-55,Cognitive(6 questions)score range 6-30,Emotional(3 questions)score range 3-15,Social(10 questions)score range 10-50.Score for each domain was provided(total score was not calculated),with each verbal response scale correlating to score of 1-5 per item(0-5 on items with a 'does not apply' option);5=most affected.Individual item scores are summed to give total domain score.Higher scores=worse outcomes.PBC-40 had 4-week recall period.Baseline=last non-missing value on or before first dose of randomized study medication.

The ITT population included all randomized participants. Only those participants with data collected at Week 52 are reported.

The EQ-5D-5L is a 6-item, standardized questionnaire that assesses mobility, self-care, usual activities, pain/discomfort, anxiety/depression, and overall health state. The visual analog scale (VAS) of EQ-5D-5L questionnaire is numbered from 0 (worst) to 100 (best). Higher scores indicated better outcomes. Baseline was defined as the last non-missing value on or before the first dose of the randomized study medication.

The ITT population included all randomized participants. Only those participants with data collected at Week 52 are reported.

DEXA scanning (hip and lumbar) was performed in participants to assess bone mineral density (femoral neck, lumbar and total hip). T-scores were calculated based on actual measured bone density value. The T-score compared a participant's bone density against that of a healthy 30-year-old adult. T-scores were used to determine primary osteoporosis, which exists on its own without any other cause and were divided into 3 categories: low risk, medium risk, and high risk (osteoporosis). T-scores measure the number of minerals in a participant's bone. A participant's level of bone loss was compared to that of a typical, healthy 30-year-old adult. A score ≥-1.0 indicates low risk, medium risk is a score between-2.5 and -1.0, and a high risk/osteoporosis is a score ≤-2.5. A positive change in T-score indicates an improvement in bone mineral density. Baseline was defined as the last non-missing central value on or before the first dose of the randomized study medication.

The ITT population included all randomized participants. Only those participants with data collected at Week 52 for each parameter are reported.

Blood samples were collected at specified timepoints for assessment of serum markers of bone turnover, CTX. Baseline was defined as the last non-missing central value on or before the first dose of the randomized study medication.

The ITT population included all randomized participants. Only those participants with data collected at Week 52 are reported.

Blood samples were collected at specified timepoints for assessment of serum markers of bone turnover, P1NP. Baseline was defined as the last non-missing central value on or before the first dose of the randomized study medication.

The ITT population included all randomized participants. Only those participants with data collected at Week 52 are reported.

Onset of clinical outcomes were described as a composite endpoint composed of the following: model for end stage liver disease sodium (MELD-Na) >14 for participants with baseline MELD-Na <12; liver transplant; uncontrolled ascites requiring treatment; hospitalization for new onset or recurrence of any including, variceal bleed, hepatic encephalopathy defined as West-Haven score of 2 or more and spontaneous bacterial peritonitis; and death.

The ITT population included all randomized participants.

An AE was any untoward medical occurrence in a participant or clinical investigation participant who was administered an investigational product, and which does not necessarily have a causal relationship with treatment. A TEAE was defined as any AE with onset during TEAE assessment period, or any event started prior to first dose of treatment and worsened or became serious during the TEAE assessment period. An SAE was any AE that resulted in death, was life-threatening, required inpatient hospitalization or prolongation of existing hospitalization, resulted in persistent or significant disability/incapacity, was a congenital abnormality/birth defect. An AESI was an AE that might not be serious but was of special importance to a particular treatment or class.

Safety analysis set included all participants who were administered at least one dose of DB study drug irrespective of the treatment received.

Blood samples were collected at specified timepoints to evaluate plasma concentration of Elafibranor and its metabolite GFT1007.

Pharmacokinetics Analysis Set included all participants who were administered at least one dose of elafibranor and have at least one post-dose PK sample. Only those participants with data collected at specified timepoints are reported.