Two rare liver disease pills got most patients to normal tests

About three in four people with primary biliary cholangitis (PBC) got their main liver blood test back into the normal range after a year on seladelpar, in the phase 3 IDEAL trial. Only 16.1% on placebo did. A second phase 3 trial, ELSPIRE, found the same pattern with elafibranor: 85.3% reached a normal result, against 23.3% on placebo.

Both studies were presented as late-breaking results at The Liver Meeting 2026 in Denver, the annual congress of the American Association for the Study of Liver Diseases (AASLD), on 8 November 2026. IDEAL is in press at the Journal of Hepatology and ELSPIRE was published in Hepatology. Gilead makes seladelpar (sold as Livdelzi, and as Lyvdelzi in some countries). Ipsen makes elafibranor (Iqirvo). Each company funded its own trial.

What makes these trials different is who took part. Everyone had PBC that doctors would usually have called "well enough controlled". Until now, that group has been left out of almost every major trial of add-on treatment.

What is PBC and what does the blood test measure?

PBC is an autoimmune disease in which the immune system slowly damages the small bile ducts inside the liver. Bile backs up, the liver scars over many years, and some people eventually need a transplant. It mostly affects women. Gilead estimates that about 130,000 Americans live with it. Itch and tiredness are the symptoms people talk about most.

The first treatment is a tablet called ursodeoxycholic acid, or UDCA. It works well for many people, but the ELSPIRE investigators say up to 40% respond inadequately.

Doctors judge how well treatment is working mainly with a blood test for alkaline phosphatase, or ALP. It is an enzyme that rises when bile flow is blocked. Higher levels over time are linked with a higher chance of liver failure, transplant and death.

For years the working cut-off has been 1.67 times the upper limit of normal. Below that, a patient on UDCA was usually counted as a responder and stayed on UDCA alone. Above it, a second drug was added. The trials that won approval for seladelpar and elafibranor both enrolled people above 1.67 times normal.

The IDEAL authors point to a growing body of research showing that any ALP above normal still carries extra risk, especially for people diagnosed young or with stiffer livers. Newer guidance from professional societies now names a normal ALP as part of the ideal response. Until IDEAL and ELSPIRE, there was no randomised evidence for treating the people sitting between normal and 1.67 times normal.

How the two trials were set up

The two studies were close copies of each other, which makes them easy to read side by side. Both randomly assigned adults two to one to a daily capsule or a placebo for 52 weeks, on top of their usual UDCA. Neither patients nor doctors knew who got what. Our guide to placebo and blinding explains how blinding protects results like these from bias.

Both drugs act on PPARs, proteins inside cells that control how the liver handles bile acids, fats and inflammation. In IDEAL, 93.8% of participants were women and the average age was 61. ELSPIRE had a median age of 61 and 92.4% women.

One detail differs. IDEAL's main goal was stricter: ALP had to be at or below the upper limit of normal and also at least 15% lower than at the start. ELSPIRE's main goal was simply a normal ALP (at or below 120 U/L), with no rescue treatment, liver event or death. ELSPIRE also reported the stricter version as a secondary result, which is the fairer number to compare: 84.5% on elafibranor against 20.0% on placebo.

What the IDEAL results showed

The response came quickly. After one month, 52.3% of people on seladelpar already had a normal ALP with at least a 15% drop, against 6.5% on placebo. By month 12 it was 75.4% (49 of 65 people) against 16.1% (5 of 31), a difference of 61.3 percentage points. The p value was below 0.0001.

On average, ALP fell by 33.8% on seladelpar and by 5.4% on placebo. Other liver markers moved in the same direction. ALT, a sign of liver cell injury, fell by 22.5% against 5.6%, and GGT, another bile duct enzyme, fell by 29.9% against 4.4%.

The paper also reports what it calls a deep response, meaning normal ALP, bilirubin, ALT and AST together. That was reached by 66.2% on seladelpar and 16.1% on placebo. These secondary and exploratory results were not adjusted for multiple testing, so they are supporting evidence and not formal proof.

Results held up across the subgroups the team looked at, including people diagnosed before age 50 and those with higher bilirubin. Liver stiffness, a scan-based measure of scarring, did not change in either group over the year.

Itch was not a major focus, because few people had much of it. Only 19 participants started with moderate or severe itch. Among them, 9 of 14 on seladelpar had a drop of at least 3 points on a 10-point itch scale, against 1 of 5 on placebo. Those groups are far too small to draw firm conclusions.

Cynthia Levy of the University of Miami, who led the study, noted in Gilead's statement that people in this ALP range had largely been kept out of the main trials of second-line drugs. She said IDEAL "provides randomized, placebo-controlled evidence that seladelpar can improve biochemical response in this population" and added that the findings "can inform individualized treatment discussions," taking into account each person's risk, symptoms and the trade-offs of extra treatment.

What ELSPIRE found with elafibranor

ELSPIRE enrolled between July 2024 and May 2025. At week 52, 85.3% on elafibranor had a normal ALP against 23.3% on placebo, a treatment difference of 62.0 percentage points, according to the Hepatology paper as reported by HCPLive. ALP fell by 36.5% on average with elafibranor and by 4.2% with placebo. Much of that drop had already happened by week 4.

The trial split people by starting ALP. In the lower band, from just above normal to 1.33 times normal, 80.4% on elafibranor reached normal against 26.1% on placebo. In the higher band, nearer 1.67 times normal, it was 87.5% against 14.3%.

A complete biochemical response, meaning normal ALP, bilirubin, liver enzymes, albumin and blood clotting, was seen in 64.3% on elafibranor and 16.7% on placebo. Bilirubin fell more on elafibranor too. As in IDEAL, only the main result was formally tested for significance.

Kris Kowdley of Liver Institute Northwest, the lead author, said in Ipsen's announcement that "patients below this threshold can still have active disease," and that the data support "consideration of normal ALP as an achievable treatment goal" for people who respond only partly to UDCA.

A worked example with real numbers

It helps to put the percentages into the units that appear on a blood test report. IDEAL used an upper limit of normal of 116 U/L for both women and men.

1.67 times 116 is about 194 U/L. Under the old rule, someone on UDCA with an ALP anywhere from 117 to 193 U/L would usually have been told their response was good enough.

The average person entering IDEAL had an ALP of 153 U/L, which is 1.3 times normal. Apply the average 33.8% fall seen on seladelpar: 153 multiplied by 0.662 gives about 101 U/L, comfortably inside the normal range. Apply the 5.4% fall seen on placebo: 153 multiplied by 0.946 gives about 145 U/L, still above normal.

This is an illustration using group averages, not a prediction for any one person. Individual results varied, and a quarter of people on seladelpar did not reach the main goal. It does show why a modest-sounding percentage drop can be enough to cross the line when someone starts only a little above it.

In absolute terms, IDEAL's 61.3-point gap means that for every 10 people treated for a year, about six more reached normal ALP on seladelpar than would have on placebo. Our guide to reading absolute and relative risk walks through this kind of sum.

Rounded to whole numbers, IDEAL's main result was 75% against 16%, and ELSPIRE's was 85% against 23%.

What about side effects?

In IDEAL, side effects were about as common on seladelpar (78.5%) as on placebo (83.9%). Serious adverse events were less frequent on the drug, at 3.1% against 12.9%, and investigators judged none to be caused by it. There were no deaths and no signs of liver, muscle or kidney harm. Joint pain was reported by 13.8% on seladelpar and 6.5% on placebo, and tiredness by 10.8% against 3.2%.

In ELSPIRE, side effects were reported by 79.0% on elafibranor and 70.0% on placebo, mostly mild or moderate. Serious events were 4.8% against 13.3%, with none judged treatment related. Two findings are worth knowing about. Weight gain of more than 5% affected 22.6% of people on elafibranor against 6.7% on placebo, though the average gain was 1.4 kg. Bone fractures occurred in 8.1% against 3.3%. All were linked to injuries such as falls, and investigators did not attribute them to the drug.

Neither trial was big enough to pick up rare side effects, and neither lasted more than a year.

The case for caution

An editorial published with ELSPIRE in Hepatology raised several points that apply to both trials. Nearly a quarter of people on placebo in ELSPIRE reached a normal ALP on UDCA alone. Three-quarters of participants started within 1.33 times normal, close enough that small swings can tip someone over the line.

The bigger question is whether a normal ALP on these drugs actually protects the liver. That link comes from studies of people whose ALP fell on its own or on UDCA. The editorial notes it is unknown whether the same benefit follows when the fall is caused by a PPAR drug. In ELSPIRE, two established scarring measures, liver stiffness and the ELF blood score, did not improve compared with placebo. Itch and fatigue scores were flat too. In IDEAL, liver stiffness was also stable in both groups.

ELSPIRE reported projected 15-year transplant-free survival from two risk calculators. Those projections improved with elafibranor (86.2% against 78.7% on the GLOBE score), but they are model outputs from the same blood tests, not observed outcomes. The ELSPIRE authors themselves cautioned that the calculators were built for use after UDCA.

Both trials enrolled very few people with cirrhosis. IDEAL was almost entirely American, with 86.5% of participants White. ELSPIRE had no participants who were not White or Asian.

Longer outcome trials are under way. Gilead's AFFIRM study (NCT06051617) is testing seladelpar against placebo on clinical outcomes in people with PBC and compensated cirrhosis. Ipsen's placebo-controlled ELFIDENCE study (NCT06016842) is looking at long-term clinical outcomes with elafibranor. Both drugs received accelerated approval in the US, a route that requires companies to confirm a clinical benefit in later studies. The FDA cleared elafibranor in June 2024 and seladelpar in August 2024.

Can I get these drugs in the UK?

Both are already licensed and recommended in the UK for people whose ALP stays above the 1.67 threshold, or who cannot take UDCA. The MHRA approved seladelpar on 16 January 2025. NICE recommended elafibranor in November 2024 and seladelpar in June 2026, within their marketing authorisations.

What the new trials could change is who counts as eligible. Gilead says it is "engaging with global regulatory authorities" about IDEAL, and Ipsen says the ELSPIRE data are being shared with regulators. Neither company has announced a timetable. Whether NHS prescribing follows will depend on licence wording and on how UK guidelines treat a near-normal ALP. Our explainer on how new medicines get approved by the FDA, MHRA and NICE covers those steps.

Other rare liver conditions also featured at this year's meeting. We have covered maralixibat for itch in biliary atresia, and the late-breaking programme included the hepatitis B drug bepirovirsen, which the FDA approved in October.

What to ask at your next liver appointment

If you have PBC and take UDCA, these questions can help you make sense of the new results with your hepatologist:

• What was my most recent ALP, and what is the upper limit of normal at my lab?
• Has my ALP been above normal for a long time, even if it is under 1.67 times normal?
• Do I have features that raise my risk, such as a diagnosis before 50, a high liver stiffness reading or raised bilirubin?
• Would adding seladelpar or elafibranor be an option for me now, or only if my ALP rises?
• How would we watch for weight gain, joint pain or bone problems if I started one?

You can look up IDEAL and ELSPIRE, and search for other PBC trials, using our guide to finding and joining a clinical trial.

This article reports trial results for general information. It is not medical advice, and treatment decisions should be made with your own liver specialist.