It started with a question from a patient.
Nav Persaud, a family doctor in Toronto, was taking care of a pregnant woman who wanted to know about a certain morning sickness drug called Diclectin, or Diclegis in the US.
Dr. Persaud assured her it was the first-line treatment. Diclectin, a combination of pyridoxine (vitamin B6) and the antihistamine doxylamine, is widely considered to be safe and effective. Every year, 300,000 prescriptions are dispensed in Canada, which accounts for more than half of live births. In the US, it is the only morning sickness drug approved by the FDA.
But something made Dr. Persaud curious. He checked the clinical practice guidelines for the drug and found they did not cite any clinical studies. Instead they only cited the Product Monograph, a document produced by the drug company.
“There isn’t really a valid reason for doing a study like this but not publishing it.”
That started a search for information that lasted more than five years. Dr. Persaud and his team found flaws in a 1997 meta-analysis of 24 studies on Diclectin, concluding that the research failed to show that the drug combination was more effective than taking vitamin B6 on its own. He also appealed to health agencies in the Canadian and American governments to release the documentation that led to the drug being approved in both countries. He was especially interested in a 40-year-old study that was conducted for a similar compound called Bendectin.
The Bendectin study included more than 2,300 participants who were divided into eight groups to test different related compounds. This trial was the main source for Bendectin first being approved in Canada and the US, Dr. Persaud said, and it was frequently cited in other studies as evidence for its successor Diclectin’s effectiveness. However, the results were never published — and he wanted to know why.
Diclegis, Diclectin, Bendectin: Under all its names, the drug has been controversial, as are many drugs prescribed to pregnant women. The horrorshow caused by thalidomide, a drug widely prescribed to pregnant women in the 1960s that caused major birth defects, has made us hyper-sensitive to the effects of prenatal medications. In some cases, this caution can harm mothers who are encouraged to endure mental and physical discomfort rather than take probably harmless medications.
Bendectin came under such scrutiny. It was approved by the FDA to treat morning sickness (which doesn’t always occur in the morning, by the way — doctors simply call it NVP, for nausea and vomiting during pregnancy) in 1956. In 1977, the lawsuits started. Mothers believed Bendectin caused birth defects, and lawyers discovered the paucity of testing. Some lawsuits were successful, although the medical question of whether it caused birth defects was not definitively answered. In 1983, the manufacturer took it off the market.
After that, the drug was only available in Canada as a generic produced by Quebec-based Laboratoires Duchesnay Inc. In 1989, fears about birth defects started circulating again in Canada, prompting a public awareness campaign by Duchesnay. Fears calmed, and the drug was newly approved in 2013 by the FDA, which gave it its highest rating for safety. In 2015, it was approved for use in Israel.
Through all this controversy, the focus was on whether the drug is safe, not whether it is effective. The reason may be that most women with NVP get better regardless of whether or not they take medication. NVP tends to go away around 14 weeks of pregnancy, Dr. Persaud said, which some women and doctors may misattribute to the medication.
For Dr. Persaud, all these questions go back to the 40-year-old study that first established the drug in North America. What was in it? And why wasn’t it ever published?
Unfortunately, the original authors were either unreachable or deceased. However, he did get a copy of the study, or most of it, out of the FDA’s microfiche archive. And he’s publishing it.
The study, “8-way randomized controlled trial of doxylamine, pyridoxine, and dicyclomine for nausea and vomiting during pregnancy,” which took place at 14 clinics in the US, has obvious shortcomings.
The trial enrolled 2,308 patients in the first trimester of pregnancy who had complaints of nausea or vomiting. They were directed to take two tablets at bedtime and an additional tablet in the afternoon or morning if needed for one week, then report their frequency of vomiting and hours of nausea.
Right away, things started to go wrong. Three of the 32 investigators never started the study. Data for 30 patients recruited by another investigator was thrown out because it was apparently fabricated. Fifty-one patients were excluded before the trial started for “incomplete data,” and another 709 patients were excluded from the study after it was discovered that they failed to meet the criteria. Finally, another 132 participants did not complete the study. That left 1,599 participants out of the initial 2,359, an attrition rate of more than 30 percent — oddly high for a trial that only took seven days.
These issues — plus the facts that the final results of the study have been lost, the supporting data is unavailable for 37 percent of those in the placebo cohort, and the method of scoring used by the investigators was unclear — led Dr. Persaud to conclude that the study has “some big problems.” Whether due to sloppiness or something more pernicious — the trial was funded by Merrell-National Laboratories, the maker of Bendectin at the time — the study was not deemed publication-worthy. However, it was still submitted to regulators as evidence that the drug should be approved for use. When Bendectin’s descendent Diclegis was submitted to the FDA, drug maker Duchesnay successfully argued for an accelerated path to approval due to the fact that the FDA had already approved Bendectin, which was basically the same.
Diclegis, Diclectin, Bendectin: Under all its names, the drug has been controversial.
Dr. Persaud finds that disconnect disturbing. If a study is used in determining the safety of a medication, it should be made publicly available, he said.
“I don’t know how often FDA approves unpublished studies,” he said. “I would hope the answer is not frequently. There isn’t really a valid reason for doing a study like this but not publishing it.”
The FDA says it is “not unusual” to review unpublished studies when evaluating a drug. “Drug companies are required to provide the FDA with all pertinent studies about a drug in order to gain FDA approval, regardless of whether the studies are published and regardless of whether the study results are positive or negative,” a press officer for the agency said in an email. “The FDA does not have the authority to control what a company chooses to publish or release to the public.”
Duchesnay also submitted its own new two-week, double-blind study of 280 women, which found that the drug was effective, although the FDA noted that “the treatment effect is small." A 2015 review of 41 studies on interventions for nausea and vomiting in pregnancy found insufficient evidence that any treatments including Diclegis are effective.
In an email, the company said, “we have complete confidence in the safety and efficacy of Diclectin® and are very proud to provide it as a safe and effective treatment option for women suffering from nausea and vomiting of pregnancy.” But Dr. Persaud is no longer convinced that Diclegis, or Diclectin as it’s called in his clinic, is worth prescribing to patients.
“I think that both Health Canada and the FDA should revisit their decision to approve this medication and the bodies in Canada and the United States that recommend this medication should revisit their recommendations,” he said. “I have completely stopped prescribing this medication because I’m no longer convinced that it’s been proven to be effective.”
Diclegis impacts a lot of people, and its safety and effectiveness are important. Dr. Persaud’s campaign is part of a bigger story, however.
The 40-year-old Bendectin study was published as part of a global effort to independently publish research that has been buried. The effort is called the “Restoring invisible and abandoned trials” initiative, or RIAT, named after a call to action published in the medical journal The BMJ. Unpublished and misreported studies make it harder to accurately assess treatments, the authors wrote, calling for anyone with access to an abandoned study to publish or republish it.
“We’re talking about a study done in the 1970s,” Dr. Persaud said. “It’s possible that other medications were approved in the 1970s, similar to other studies not being published. It’s possible that there are other medications that haven’t been proven to be effective that are being sold in Canada or the US.”
“All the information should be made publicly available.”
