麻豆果冻传媒影音

A panel of FDA's outside experts said obeticholic acid (Ocaliva) does not have a favorable risk-benefit profile for treating nonalcoholic steatohepatitis (NASH) with pre-cirrhotic liver fibrosis, and recommended against accelerated approval of the drug.

On Friday, the voted 12-2 (with two abstentions) that the improvements in liver scarring seen in the pivotal with a 25-mg dose of obeticholic acid do not outweigh the various risks, including a rate of drug-induced liver injury (DILI) that surpasses thresholds the FDA considers a threat to drug approvability.

And by a vote of 15-1, panelists recommended against accelerated approval, saying the agency should await additional data on clinical outcomes in the ongoing phase III trial.

"At this point, I do not believe that the benefits outweigh the risks, while keeping in mind that this is a surrogate endpoint among people who are asymptomatic at baseline," said committee chair Benjamin Lebwohl, MD, of Columbia University College of Physicians and Surgeons in New York City. "This is a serious disease. However, the bar needs to be quite high when considering that fact."

He noted that while the phase III study will continue to monitor for clinical outcomes, including death, liver decompensation, and need for transplant, "right now we're seeing numerically more deaths" in the obeticholic acid 25-mg arm than in the placebo arm.

Lebwohl added that the degree to which that surrogate endpoint will ultimately yield benefits is also "marred in uncertainty, particularly regarding the concerns related to DILI."

"We have clear evidence of safety risks," agreed James Floyd, MD, of the University of Washington in Seattle. "It's impossible in my mind to ensure a good risk-benefit profile based on this surrogate endpoint data -- we need to see the full clinical outcomes."

No drug is approved for treating NASH, a severe form of nonalcoholic fatty liver disease (NAFLD), and NASH patients with clinically significant liver fibrosis (stages F2 and F3) are at high risk for progression to cirrhosis, which places them at greater risk for hepatocellular carcinoma, liver decompensation, need for transplant, and death.

REGENERATE met one of its two primary endpoints. At 18 months, a significantly higher proportion of patients on the 25-mg dose of the farnesoid X receptor antagonist experienced at least a 1-stage improvement in fibrosis with no worsening of NASH, a surrogate outcome the FDA agrees is likely to predict clinical benefit for this patient population.

Depending on the analysis, the difference ranged from an absolute difference over placebo of 8.6% (95% CI 4.2-13.0) to 12.8% (95% CI 7.0-18.5). And fewer patients on obeticholic acid had worsening of fibrosis, as well.

No significant difference, however, was seen for the study's second primary endpoint -- NASH resolution with no worsening of fibrosis.

Given that obeticholic acid -- a drug currently approved for treating primary biliary cholangitis -- would require lifelong treatment in NASH, a sticking point was whether risk mitigation for DILI would be feasible in clinical practice.

In a presentation on Friday, FDA's Paul Hayashi, MD, MPH, said predicted DILI fatalities with the proposed 25-mg dose of obeticholic acid was far higher than the agency's typical threshold for concern, put in place after three drugs were pulled from the market in the late 1990s for DILI deaths. Since then, no drug has been withdrawn in the U.S. over fatal DILI.

"So the track record for this change has been good," said Hayashi.

Obeticholic acid's "predicted fatality rate -- which was set by the subjects who required transplant -- is 15- to 30-fold higher than the threshold, and 6- to 13-fold higher than the three drugs removed from the market," he said.

Panelist Jacquelyn Maher, MD, of the University of California San Francisco, who also voted against the drug, noted that the high prevalence of biliary disease in this patient population puts them at even higher risk for DILI, "which can be both sudden and severe."

During the sponsor's presentation, Intercept noted that a trial amendment to monitor and manage hepatic events dropped the incidence of moderate to severe hepatic events possibly related to the drug from 1.5% to 0.13%.

But panelists were not completely sold that the level of monitoring would be feasible in clinical practice, or sufficient.

"The monitoring program that was set up does appear to mitigate, in part, risk of DILI and has been associated with a reduction of events, but does not entirely eliminate the concern," said Lebwohl, summarizing the panelists' views on DILI.

"With regard to frequency of monitoring, there was concern that what's suggested by the sponsor might not be adequate," he continued, "particularly in light of the fact that DILI may occur a long way out from drug initiation and cholestatic liver injury may occur pretty rapidly -- even after one normal spot-check of liver enzymes and bilirubin."

Along with DILI, panelists had other safety issues to consider, including that treatment appeared to increase the risk of diabetes or worsen diabetes in patients with the condition. Other excess risks versus placebo were seen for cholecystitis and bile duct stones or sludge, dyslipidemia requiring new treatment with statins or dose intensification, acute kidney injury, and severe pruritus.

Obeticholic acid in NASH with compensated cirrhosis has , so pre-cirrhotic patients who progress to this level of disease would require treatment discontinuation.

To address the various safety issues, sponsor Intercept said that prescribing would be restricted to gastroenterologists and hepatologists, and that the drug label would exclude patients with complete biliary obstruction and cirrhosis.

But even these plans didn't necessarily allay members' worries.

"I remain concerned that a drug such as this will be able to be restricted to prescription by only experts who are willing to take the necessary steps that are required to mitigate risk," said Maher.

While the FDA is not required to follow the advice of its advisory committees, it typically does.

Comment