GW's Epidiolex Epilepsy Data Paves Way For Cannabinoids

Safety and efficacy in study published in New England Journal of Medicine bode well for cannabinoid therapy Epidiolex, which is headed to FDA mid-year.

The positive reception of Phase III data for GW Pharmaceuticals PLC's cannabinoid therapy Epidiolex for Dravet syndrome, a severe type of pediatric epilepsy, in the New England Journal of Medicine is an important step for pharmaceutical-grade cannabinoids as a therapeutic option for neurological disorders as well as other diseases.

Derived from the marijuana plant, Epidiolex includes cannabidiol (CBD) but not the psychoactive tetrahydrocannabinol (THC). The drug is one of over 50 cannabinoids in any stage of development, according to Pharmaprojects. As of late 2016, candidates were in development for about 40 indications, including epilepsy, multiple sclerosis disorders, pain and schizophrenia (see sidebar). Epidiolex is the most advanced in development, followed by Insys Therapeutics Inc.'s synthetic cannabidiol therapy, which is in Phase III for similar indications.

Full results from the Phase III GWPCARE1 study, which showed a significantly lower rate of convulsive seizures for Epidiolex versus placebo, were reported by Orrin Devinksy of the New York University Langone Comprehensive Epilepsy Center and colleagues in the NEJM on May 25, following a top-line release last year. (Also see "GW Succeeding With Cannabinoid Drug This Time Around" - Pink Sheet, 14 Mar, 2016.)

Dravet syndrome is a tough-to-treat and potentially life-threatening form of pediatric epilepsy with a genetic cause. Because many patients still have seizures despite taking epilepsy medicine, there is high unmet need and some patients have resorted to giving marijuana to their children. However, data has been lacking and the products available have not been subjected to rigorous testing and review.

In an NEJM editorial accompanying the data, Samuel Bercovic of the University of Melbourne said the "trial represents the beginning of solid evidence for the use of cannabinoids in epilepsy."

The results now require replication; future trials may answer further questions about the applicability of cannabinoids to the many other syndromes of childhood epilepsy and to treatment in adults, Bercovic said.

GW is poised to file Epidolex with the US FDA in the middle of this year for Dravet syndrome and Lennox-Gastaut syndrome, a different form of severe, child-onset epilepsy. The filing will include the one pivotal Dravet study and data from two positive Lennox-Gastaut trials (Also see "Dravet Syndrome Indication Will Set GW Pharma Apart On US Epilepsy Market" - Scrip, 27 Sep, 2016.). A second Phase III Dravet study is recruiting.

CEO Justin Gover commented to Scrip that he hopes the data bodes well for the approval and for the "acceptance of the product by medical community if and when it is approved by the FDA."

The Epidiolex application has fast track and orphan drug status and the company is hoping for priority review with FDA. GW Pharma has already been building a commercial team in the US to launch the product.

"We will be ready to launch this drug in the US ourselves if and when FDA approves it," Gover said.

GW plans to file in Europe later this year and commercialize the drug in that region by itself as well.

Cowen has projected sales for the drug of $250m worldwide in 2018 and $1.2bn in 2021. Leerink Swann projects sales of $88.6m globally in 2018 and $1.1bn in 2021.

Push From NEJM Write-Up

Publication of full results for Epidiolex in Dravet syndrome in the NEJM – and with an encouraging accompanying editorial – should go a long way toward helping the drug gain some publicity and momentum ahead of submission and review.

The study of 120 children and young adults with Dravet syndrome and drug-resistant seizures tested Epidiolex vs. placebo, both on top of standard epileptic treatment, using the primary endpoint of change in convulsive seizure frequency over a 14-week treatment period. On average, patients were taking three anti-epileptic drugs, including valproate and clobazam, as background therapy, having tried and failed an average of more than four therapies.

The mean frequency of convulsive seizures dropped from 12.4 to 5.9 for patients on Epidiolex vs. 14.9 to 14.1 for placebo, a statistically significant difference that satisfied the primary endpoint.

In the NEJM editorial, Bercovic commented that "anecdote has been confirmed by data, and one might ask whether a controlled trial was really necessary." But, he added, "the answer is absolutely yes. Perhaps counterintuitively, the rate of response to placebo in clinical trials is higher among children than among adults," he said. Furthermore, "parents who go to enormous efforts to get cannabis for their children report a higher response rate than those who can easily obtain it."

Side Effects Were Manageable

Side effects associated with Epidiolex included diarrhea, vomiting, fatigue, and elevated liver enzymes. Overall, the rate of adverse events at any grade was 93% for Epidiolex vs. 75% for placebo, but most were mild-to-moderate. The rate of severe events was 16% for Epidiolex vs. 5% for placebo. The drop-out rate due to side effects was 14% for Epidiolex vs. 5% for placebo (see table below).

"With additional experience, perhaps these effects can be modified with dose adjustment and other strategies," Bercovic suggested.

Leerink Swann analyst Paul Matteis commented in a May 24 note that overall "the disclosure of the data in a very high quality journal adds further corroboration for Epidiolex's efficacy profile in Dravet and offers incremental support for a high odds of FDA approval."

However, the data were not perfect. As Mattteis noted, there was no statistically significant difference for several secondary endpoints including the quality of life in childhood epilepsy score nor on certain types of seizures, e.g. tonic-clonic and non-convulsive.

"The lack of a significant reduction in nonconvulsive seizure frequency suggests that the antiseizure effect of cannabidiol may be specific to convulsive seizures in the Dravet syndrome or that the frequency of nonconvulsive seizures (e.g., brief staring spells) cannot be reliably counted by parents in developmentally delayed children," the study notes.

Furthermore, some investors question whether concomitant use of the anticonvulsant clobazam is having a synergistic effect and influencing the efficacy results in the test drug arm. Clobazam was used in 66% of the patients on Epidiolex and 64% on placebo. Both are metabolized in the cytochrome P450 pathway.

Matteis noted that safety was acceptable and consistent with previously released results – elevation of liver enzymes, which was likely related to the concomitant use of valproate, was an issue (12 cases for Epidiolex vs. one for placebo) but there were no cases of Hy's law.

"Cannabidiol may potentiate a valproic acid-induced change in hepatic aminotransferase levels. The observation that the increases in hepatic aminotransferase levels mostly resolved while the patients continued taking the drug suggests that a transient metabolic stress on the liver may be responsible," Devinsky and colleagues reported.

Matteis noted that in the nine cases where an Epidiolex-treated patient who suffered from elevated liver enzymes remained in the trial, all returned to normal levels, "suggesting that the issue is usually manageable."

Investors are concerned that the liver results "could in part hinder broad use," depending on labeling language, but valproate is widely used in refractory epilepsy despite "more serious liver enzyme issues" and a boxed warning for liver toxicity, the analyst pointed out.

Gover commented that the rate of elevated liver enzymes was similar to what has been reported in trials of other epilepsy drugs and that most resolved while on treatment. Furthermore, he said, the result shows the importance of characterizing a therapy in a rigorous, scientific way.

"We continue to believe that the clinical profile of Epidiolex is differentiated from other anti-epileptics," Leerink's Matteis concluded, pointing out that "105 of the 108 patients who completed the study elected to transition to the open label study."

Phase III Adverse Events Results In Dravet Syndrome
Adverse Event Measure	Epidiolex (n=61)	Placebo (n=59)
Percent with adverse events	93%	75%
Of those with AEs, % mild-to-moderate	84%	95%
Severe adverse events	16%	5%
Somnolence	36%	10%
Diarrhea	31%	10%
Fatigue	20%	3%
Vomiting	15%	5%
Drop-out rate	14%	5%
Source: Devinsky et. al., NEJM