Medical researchers have confirmed what some desperate parents have been claiming for years—that a nonpsychoactive component of marijuana known as cannabidiol (CBD) can reduce epileptic seizures in some children.

Published in The New England Journal of Medicine, the findings stem from a double-blind, placebo-controlled study—the most scientifically rigorous type of investigation possible. “This study clearly establishes cannabidiol as an effective anti-seizure drug for this disorder and this age group," says principal investigator Orrin Devinsky, director of the Comprehensive Epilepsy Center at New York University Langone Medical Center. “It certainly deserves to be studied in other types of epilepsy.”

A total of 120 children and teenagers with Dravet syndrome—a rare disorder marked by drug-resistant seizures that can be nearly continuous in some cases—were part of the study. They were divided into an experimental group, which received the test drug, and a placebo group, which was given a medically inactive compound. Over the course of 14 weeks the youngsters receiving CBD experienced a median number of 5.9 convulsive seizures per month (down from 12.4) compared with 14.1 convulsions per month (down from 14.9) for the placebo group. The new findings are consistent with previous, less-stringent studies of the same drug, a compound called Epidiolex, made by U.K.–based GW Pharmaceuticals. (GW funded the new study.)

As the current paper points out, “interest in cannabidiol for the treatment of epilepsy was generated by media reports of efficacy in children with Dravet syndrome.” The star of many of those reports was Charlotte Figi, now 10, of Colorado, who was having hundreds of seizures a month by age three when her parents decided to treat her with cannabidiol. Unlike the better-known marijuana component tetrahydrocannabinol (THC), cannabidiol does not make users high. Twenty-nine states and the District of Columbia have legalized marijuana for medical use, and the conditions approved for treatment can vary from state to state. Sixteen states have laws that specifically allow the use of CBD to treat seizures.

The Figis treated their daughter with a specially prepared CBD-containing oil, now known as Charlotte’s Web, which is derived from hemp, a type of cannabis containing less than  0.3 percent THC by weight. They reported dramatic improvement. As word spread, more families tried CBD and the positive anecdotes piled up. “That’s certainly one of the motivations for this research,” Devinsky says. He notes there is also plenty of animal model data as well as anecdotes from the late 1800s about Indian hemp, another type of cannabis, that was used to treat epilepsy. And there are even mentions of the approach in Sumerian tablets going back 3,800 years. “The parents added fuel to the fire,” he says, “but that anecdotal evidence was there for millennia.”

All of the patients in the new trial fit the criteria for severe, drug-resistant epilepsy and were taking other seizure medications. Whereas there was a statistically and clinically significant median reduction in convulsive seizures of 39 percent in the treated group, only three of the 52 patients receiving cannabidiol became completely seizure-free. And 93 percent of those patients reported troublesome side effects—including sleepiness (the most common symptom), vomiting, fatigue, decreased appetite, diarrhea and elevated levels of liver enzymes. Eight of the cannabidiol patients stopped participating in the study as a result of the more severe side effects. The study authors indicate some of those issues could have been caused by interactions with other epilepsy drugs. Also, there was no significant reduction in nonconvulsive seizures, which are essentially brief staring spells in which a person is unaware of his or her surroundings for several seconds. The study notes this could be because cannabidiol only affects convulsive seizures or because nonconvulsive seizures “cannot be reliably counted by parents in developmentally delayed children.”

Nevertheless, 62 percent of caregivers in the cannabidiol group said their child’s overall condition improved during the trial, compared with 34 percent in the placebo group. After the trial caregivers of participants in both the placebo and experimental groups were given the option of continuing treatment with cannabidiol in what is called an open-label extension of the study. More than 100 families from both groups took researchers up on the offer.

Sam Riggio, director of operations for Realm of Caring, a cannabis research, education and advocacy group, understands how even treatments with limited benefits can have a big impact on Dravet patients. Riggio and his wife Tara moved to Colorado from Illinois in 2013 when they heard about cannabidiol in a Dravet group on Facebook. (At the time, the hemp supply in the U.S. was spotty. Charlotte’s Web had a reputation as safe and effective for children, but it was only available to Colorado residents.) Their daughter Francesca, who will be seven next month, was having seizures that never lasted less than 15 minutes and sometimes went on for long as 70 minutes. She had her first seizure at six months and at 18 months suffered a severe reaction to an epilepsy drug that left her with profound cognitive deficits. Then her parents started her on cannabidiol. “It wasn’t the answer for her like it was for Charlotte and some of the other kiddos,” Riggio says, “but it provided her enough cushion that she could come off the three pharmaceuticals she was on.” In the years since, while Francesca remains nonverbal with the cognition of an 18-month old, most of her seizures now last less than two minutes and resolve on their own. Her parents have not had to call 911 nor have they had to rush her to the emergency room. “That’s a huge improvement for her quality of life,” says Riggio, who credits the changes to cannabidiol.

Devinsky hopes these latest findings will persuade the U.S. Drug Enforcement Administration to change its classification of marijuana as a Schedule I substance under the Controlled Substances Act, a ruling that hampers medical research and treatment. (Drugs on Schedule I, such as heroin, are deemed by the DEA to have “no currently accepted medical use.”) “To put CBD as a Schedule I drug violates scientific data and common sense,” Devinsky says. He points out that some opiates, which kill thousands of Americans a year, are Schedule III drugs, a less restricted category. “Cannabis is approved in 20 countries for spasms in multiple sclerosis,” he says, “so to say there’s no evidence of efficacy is simply untrue. Those studies came out a decade ago, this study is coming out today. They have to de-schedule this drug. It’s just not fair to the research and clinical communities, or to the patients. It’s medieval.”

Going forward, GW Pharmaceuticals plans to seek U.S. Food and Drug Administration approval of Epidiolex in the next few months as a treatment for Dravet syndrome and Lennox–Gastaut syndrome, another seizure disorder. No one really knows yet how cannabidiol reduces seizures in Dravet, but Devinsky says the leading theory is that the cannabidiol binds to a receptor in nerve cells called GPR55, modulating the calcium activity and excitability of the cells. He also says cannabidiol has many other effects on nerve cells and that there is much to learn.

For Riggio, a gold-standard, double-blind, placebo-controlled clinical trial will do just fine for now. “It’s exciting because it’s a new option for people who feel like they are out of options,” he says. “And it confirms what we’ve seen anecdotally—and the more data we have to back that up, the more confidence we have to use this sooner and hopefully avoid a lot of side effects that my daughter had to endure.”