Ondansetron and the EMA
The European Medicine’s Agency (EMA) released a Pharmacovigilance Risk Assessment Committee (PRAC) update to the manufacturers of ondansetron in July which makes 3 dramatic, incorrect and misleading statements:
- Ondansetron is suspected to cause orofacial malformations
- Ondansetron should not be used in the first trimester
- Women of childbearing potential who are taking ondansetron should consider using contraception
These were based, they claim, on two recent papers and a systematic review. The first paper, Huybrechts et al 2018, is a very well conducted and robust paper looking at 1.8 million pregnancy’s with 88,467 exposures to ondansetron. They found that there were no associations with cardiac (heart) malformations or other congenital malformations, however, they found an association (which is not the same as a cause) with increased incidence of oral cleft. In the general population approximately 11 in 10,000 babies will be born with an oral cleft while in mothers exposed to ondansetron in the first trimester there were 14 in 10,000 babies with oral cleft. So an additional 3 babies in 10,000.
The second paper they reference, Zambelli-Weiner et al 2018, suggested a slight increase risk of both cardiac malformations and oral clefts. However, the quality of the research is poor and not robust and more importantly the funding of the paper is riddled with controversy as the researchers were funded in excess of $200k for their paper by the plaintiffs in a mass tort case against Zofran which is ongoing in the USA. They authors did not admit to this major conflict of interest and others until subpoenaed by the defendants. In effect this paper is discredited and should not hold weight in the eyes of the EMA or anyone else assessing the safety of ondansetron.
The third paper, a systematic review by Kaplan et al 2019 was not referenced by the EMA as it only came out in June 2019, however it crops up later in this story during the teleconference we had with the EMA on the 14th October. The conclusion of the review reads: “These results may be reassuring for pregnant women in whom ondansetron use is clinically indicated since the absolute risks of possible concerns appear to be low.”
From the above papers it is hard to see how the EMA reached their dramatic and scary conclusions that women on ondansetron should use contraception (the sort of statement reserved for the likes of thalidomide!) or that ondansetron is “suspected to cause”. An association is not the same as a cause. It could well be that the women requiring ondansetron in the first trimester are experiencing high levels of malnutrition and are generally not taking folic acid and are micronutrient deficient due to hyperemesis gravidarum (HG) and it is those things that are causing the additional clefts.
So, a group of international experts, clinicians, researchers and patient organisations wrote to the EMA to highlight the real world impact of their statements and ask that they be amended. You can read that letter here.
We also contacted the UK medicine’s regulator and asked them to hold off putting out a Drug Safety Update, which they had been about to do using the same wording as the EMA’s statement (because no one thought to question the legitimacy of the interpretation!). Thankfully they have and I believe Ireland has also waited to consult with stakeholders but other countries have ploughed ahead and released national statements and “Dear Doctor, stop prescribing ondansetron in pregnancy” letters.
A couple of weeks later we got a reply, which you can read here. Basically, their position is that they are only referring to the licensed uses of ondansetron and therefore their statement is not relevant to people with HG. So, lets look at the licensed uses:
1) Chemotherapy nausea and vomiting – DOES NOT OCCUR IN THE FIRST TRIMESTER OF PREGNANCY
2) Post-surgery nausea and vomiting – VERY RARE (GENERALLY EMERGENCY) IN FIRST TRIMESTER
The number of incidences globally would be so few as to be unable to detect an effect from ondansetron exposure in these licensed use cases that the statements seems null and void in its purpose. Furthermore, the data on ondansetron use in pregnancy is collected from off-license indications; almost exclusively hyperemesis gravidarum.
We wrote back, outlining this and accepting their offer of a teleconference. You can read that letter here.
Then we attended a teleconference with the chair of the PRAC, Dr Sabine Strus, the rapporteur, Gabriela Jazbec, various public engagement department employees and other PRAC members such as Hedvig Nordeng and Patrick Batty. From our position we had myself and PSS staff, Clare murphy from bpas, Sally Stephens from the UK Teratology Information Service (in lew of their head, Dr Kenneth Hodson), Prof Cathy Nelson-Piercy, Dr Rebecca Painter, Dr Brian Cleary and Dr Krista Huybrechts (author of the main paper).
The meeting seemed to go well and we outlines our concerns about the interpretation of the data, the patient position and real world impact of HG and how this statement will affect treatment “on the ground”. We highlighted issues around publication bias and drugs with more data being accidentally penalised by effect size (I’ll write on these issues in another post).
Then at the end a very strange thing happened… Dr Straus “summarised” the meeting by announcing that it was wonderful we were all in agreement and the matter had been sorted out! I clarified that while we agree on the Huybrechts et al data being robust we did not agree with their interpretation of it and their statements based on it and asked if they were willing to at least amend their statement. Catagorically they were NOT. At least until new data has come out. Furthermore they made it absolutely clear that conducting a risk-benefit analysis is NOT their job and that it is up to guideline writers, doctors and patients to do that.
This last point is critical because the “risk” of taking ondansetron must be weighed against the risk of not taking ondansetron for a patient who is seriously ill, malnourish and dehydrated. That is likely to require an entirely separate post and I’ve explained it online in a video also but the bottom line is that the equally robust and solid evidence of the effects of unmanaged HG and malnutrition in the first trimester (including cardiometabolic disturbances, small for gestational weight babies, increase risk of autism spectrum disorders, developmental delays and fetal death (the greatest risk being termination of wanted pregnancy)) are far greater than the risk of ondansetron.
Obviously I haven’t left it there and followed up with an email saying that we were disappointed that we felt our queries and concerns remain unanswered, which in turn invited a response to the affect of what questions remain unanswered.
So as of last night the email sitting in their ball court asks the following specific questions:
1) How were the papers critically appraised? Can we see the critical appraisals? Were validated tools used or just experience of the members? I ask because we have concerns over the weight given to the Zambelli-Weiner paper.
2) Why are your conclusions so completely opposite to the conclusions of the authors of both the Huybrechts paper and the Kaplan systematic review? Your recommendations are virtually opposite to the conclusions of the papers. While we agree the data is robust we do not agree with the EMAs conclusions or communication of the data.
3) Why have you used the phrase “suspected to cause” when that is factually incorrect?
4) Why, in light of the vast harm caused by inappropriate phrasing of your recommendation, are you unwilling to amend it?
5) Why are you commenting on the use of a drugs licensed indications in pregnancy when it is not used for those indications in pregnancy and when the data you are using is exclusively collected from a non-licensed indication?
6) How, based on the published data, have you concluded that women of childbearing potential should consider using contraception?
7) Can you clarify how you will communicate your lack of risk/benefit analysis to the various countries authorities so that they understand that responsibility is for them to undertake?
8) You mentioned revising when new data is published. Does this mean that when the forthcoming new paper is published you will be willing to rewrite your statements, and work with us to achieve a balanced and appropriate statement?
I am now off on holiday until the end of the month without internet access but I wanted to get the position out there for people affected by this as I know it is already having a devastating affect on lives. We plan to campaign through November (and until we have had a positive effect) and we will need help with writing letters, articles, giving interviews and so on. So be ready… we are not taking this lying in our sick beds