Alcohol-Associated Hepatitis

Larsucosterol (DUR-928) is an investigational product and has not been approved by the FDA for marketing in the U.S. for any indication.

Epigenetic dysregulation in alcohol-associated hepatitis

Development of liver failure in patients with alcohol-associated hepatitis (AH) was reported to be associated with a profound decreased expression of genes that are involved in critical hepatocyte functions. The silencing of these genes was correlated with hypermethylation (addition of methyl groups to DNA), a process facilitated by DNA methyltransferases (DNMTs). Interestingly, the expression of both DNMT1 and DNMT3a was highly increased in patients with AH (see graph below)1.

DURECT is currently conducting a Phase 2b study (AHFIRM) in subjects with severe alcohol-associated hepatitis (also called Alcoholic Hepatitis or AH) to evaluate the safety and efficacy of larsucosterol treatment. For more information visit our Clinical Trials page.

Larsucosterol has already demonstrated a promising safety and efficacy profile in a Phase 2a clinical trial in 19 patients with severe or moderate AH. All patients treated with larsucosterol survived the 28-day follow-up period and there were no drug-related serious adverse events. Patients treated with larsucosterol in this trial had a statistically significant reduction from baseline in bilirubin at day 7 and 28 and MELD at day 28. Lille scores were also statistically significantly lower than those from a well-matched group of patients in a contemporary ongoing trial as well as several published historical controls. 74% of all larsucosterol treated patients and 67% of those with severe AH were discharged from the hospital within four days of receiving a single dose of larsucosterol. For more information about the Phase 2a study design and data, please see the “Data” section below.

Larsucosterol has been shown to inhibit the activity of DNMTs (DNMT-1, 3a and 3b), inhibiting DNA methylation, and upregulating expression of genes involved in signaling pathways associated with cell death, stress response, and lipid biosynthesis, which may ultimately result in increased cell survival, reduced inflammation, and decreased lipotoxicity.

Given larsucosterol's proposed mechanism of action, demonstrated safety profile and promising clinical data observed in a Phase 2a trial, there is strong rationale for investigating larsucosterol for the treatment of AH.

1Argemi et al. 2019. Nature Communications, 10: 3126; Creative Commons Attribution 4.0 International License

About the Phase 2a trial of larsucosterol in Alcohol-Associated Hepatitis

The open-label, dose escalation, multi-center study was designed to determine the safety, pharmacokinetics and pharmacodynamics of larsucosterol in AH patients following treatment. This included assessing liver biochemistry, biomarkers, and prognostic scores such as the Lille score. Final enrollment included 19 patients with moderate and severe AH, who were administered larsucosterol intravenously at three different doses. Eight patients (four moderate and four severe) were dosed at 30 mg, seven patients (three moderate and four severe) were dosed at 90 mg and four patients (all severe) were dosed at 150 mg. After being discharged on day two, one patient did not return for the scheduled day seven and day 28 follow-up visits; therefore Lille, bilirubin and model of end-stage liver disease (MELD) data reported above are based on 18 patients.


Lille scores are used in clinical practice to help determine the prognosis and response of AH patients after seven days of treatment. The lower the Lille score, the better the prognosis. Patients with a Lille score below 0.45 have a six-month survival rate of 85% compared to those with Lille scores above 0.45, with only a 25% six-month survival rate.1The chart below shows the Lille scores for individual AH patients treated with larsucosterol plotted as a function of their baseline MELD scores. In our study, the median Lille score for patients treated with larsucosterol was 0.10. The median Lille score among a cohort of 15 patients treated with standard of care at the University of Louisville (UL) was 0.41 (shown as historical control).

1Louvet A et al. Hepatology 2007; 45: 1348-54.

The chart below shows individual patient Lille scores plotted as a function of their baseline MELD scores.



1)Our advisor, Dr. Craig McClain from the University of Louisville (UL), shared anonymized data from his study, in which 15 AH patients with initial MELD scores ranging from 15-30 received either supportive care alone (n=8 moderate AH patients) or supportive care with corticosteroids (n=7 severe AH patients). Two of the UL patients died by day 28.
2)One patient in the larsucosterol (DUR-928) group did not return for the day 7 visit.
3) Lille scores in the larsucosterol patients were significantly lower than that of the UL patients (p=0.01; Wilcoxon's Rank Sum Test). 

As shown below, 100% of patients in the 30 mg and 90 mg larsucosterol dosing groups were treatment responders based on their Lille scores. 89% of the overall larsucosterol patient population were treatment responders. Patients with severe AH, as defined by Maddrey’s Discriminant Function >32 or MELD 21-30, and baseline serum bilirubin above 8 mg/dL, had similarly high response rates to larsucosterol treatment.

AH Patient Category
Lille Median
All Patients2
30 or 90 mg larsucosterol3
0.10 (0.04, 0.20)
0.05 (0.04, 0.19)
DF>32 (SAH)2
30 or 90 mg larsucosterol3
0.19 (0.05, 0.22)
0.12 (0.05, 0.19)
MELD 21-302
30 or 90 mg larsucosterol3
0.19 (0.11, 0.25)
0.19 (0.10, 0.19)
Baseline bilirubin >8 mg/dl2
30 or 90 mg larsucosterol3
0.10 (0.05, 0.20)
0.10 (0.05, 0.19)
1 One patient did not return for Day 7 visit;
2 Including patients receiving 30, 90 and 150 mg of larsucosterol (DUR-928);
3 Excluding patients receiving 150 mg of larsucosterol.
Maddrey’s Discriminant Function (DF) is calculated using the patient’s prothrombin time and serum bilirubin level.

The Lille scores of patients treated with larsucosterol in this trial were also significantly lower than several selected published historical studies (Hepatology 2007, 45:1348-1354; Gut 2011, 60:255-260), in which patients had similar baseline bilirubin, albumin, creatinine, prothrombin time and DF scores, and were treated with standard of care with or without corticosteroids. Of course, due to the historical nature of these studies, such comparisons should be taken cautiously.

A sub-group analysis was conducted to compare severe AH patients in the 30mg and 90mg dosing groups (n=8) with well-matched severe AH patients (n=13) who received corticosteroids for 28 days in a contemporaneous study at the University of Louisville (UL). Patients shown below in the UL steroid group had a baseline mean MELD of 24.46 and baseline mean Maddrey’s DF score of 62.98. The 8 patients in the larsucosterol group had baseline mean MELD of 24.50 and baseline mean Maddrey’s DF score of 61.25. All patients treated with larsucosterol survived the 28-day follow up period, while 3 patients in the UL steroid group died within the first 28 days.

The steroid group in the above graph includes the 7 severe AH patients treated with steroids from the UL group shown in the MELD vs Lille graph above plus an additional 6 severe AH patients subsequently treated in the UL study; The larsucosterol (DUR-928) group included 8 patients.


Bilirubin is formed by the breakdown of red blood cells in the body. The level of total bilirubin in the blood is an indication of how the liver is functioning. In this trial, patients treated with larsucosterol had a significant early reduction from baseline in bilirubin by day 7. Patients with more elevated bilirubin at baseline (serum bilirubin >8 mg/dL) had a median reduction from baseline of 25% by day 7 and 48% by day 28.

www.durect-1-NEW (1).jpg

*p<0.05 compared to baseline (Wilcoxon's Signed Rank Test)

Model of End-Stage Liver Disease (MELD)

MELD is another common scoring system used to assess the severity and prognosis of AH patients. Patients with MELD scores of 11-20 are classified as having moderate AH and patients with MELD scores of 21-30 are classified as having severe AH. As with Lille scores, the lower the MELD score, the better the prognosis for the AH patient. In this study (shown in the chart below), the median reduction from baseline in MELD among all larsucosterol treated patients was over 2 points and among those with baseline bilirubin levels >8 mg/dL was 5 points by day 28.


*p<0.05 compared to baseline (Wilcoxon's Signed Rank Test)
MELD is calculated based on (a) bilirubin, (b) serum creatinine (sCr), and (c) International Normalized Ratio (INR), which is a measure of prothrombin time.

Safety and Pharmacokinetics

Larsucosterol was well tolerated at all doses tested. There were no drug-related serious adverse events and only three adverse events designated as possibly related to larsucosterol: one occurrence of moderate generalized pruritus, one mild rash and one grade two alkaline phosphatase. There were no discontinuations, early withdrawals or termination of study drug or study participation due to adverse events. All patients treated with larsucosterol survived through the 28-day follow-up period. Drug exposures were dose proportional and were not affected by the severity of the disease.

AH is a life-threatening acute alcohol-associated liver disease (ALD) often caused by chronic heavy alcohol use or a drastic increase in alcohol consumption (e.g., a binge). It is characterized by severe inflammation and destruction of liver tissue (i.e., necrosis). AH may occur suddenly after binge drinking, potentially leading to life-threatening complications, including liver failure, acute renal injury and multi-organ failure.

AH can affect both women and men, and while more than half of the people diagnosed with AH are between 40 and 60 years old, a growing number of young adults have been diagnosed in recent years, as a result of increasingly heavy consumption and binge drinking prevalence.

AH is associated with about 137,000 US hospitalizations per year. An analysis of 77 studies published between 1971 and 2016, which included data from a total of 8,184 patients, showed the overall mortality from AH was 26% at 28 days, 29% at 90 days and 44% at 180 days. A subsequent global study published in December 2021, which included 85 tertiary centers in 11 countries across 3 continents, prospectively enrolled 2,581 AH patients with a median MELD score of 23.5, reported mortality at 28 and 90 days of 20% and 30.9% respectively. There are no FDA approved therapies for AH and stopping alcohol consumption is not sufficient for recovery in many moderate and severe patients. Treatments to reduce liver inflammation, such as corticosteroids, have not demonstrated a benefit in survival at 90 days or 1 year, and have demonstrated an increased risk of infection (STOPAH trial). In addition, only 25 to 45% of AH patients are eligible for corticosteroids. The cost related to treating hospitalized AH patients and their length of hospital stay were reported at AASLD’s The Liver Meeting 2021: 

Each hospitalization episode with AH diagnosis for patients who:Average length of stayTotal healthcare cost per hospitalization episode
Died during the hospitalization 9 days $151,500
Were discharged 6 days


Marlowe, et. Al., AASLD 2021 Poster No. 381#


The rate of AH patients undergoing liver transplants has more than doubled since the first COVID-19 shelter in-place orders in early 2020. The procedure involves a long waiting period, a burdensome selection process and costs more than $875,000 on average.

DURECT currently holds the worldwide development and commercialization rights to larsucosterol and other molecules identified as part of the Epigenetic Regulator Program.