OnabotulinumtoxinA in chronic migraine has been approved by the NICE in the UK based on evidence from the PREEMPT studies . However, none of the patients in the original trials were reported to be pregnant and so the efficacy and safety of onabotulinumtoxinA in this population remains unknown. A US survey in 2006 of physicians using onabotulinumtoxinA for all indications found that only 12/396 (3%) had any experience of use in pregnant patients . Of the 12 physicians, 1 declared they were ‘very comfortable’, 5 ‘somewhat comfortable’ and the rest presumably ‘uncomfortable’ with the procedure. Currently in the UK, the advice from the Medicine and Healthcare product Regulatory Agency (MHRA) is that onabotulinumtoxinA should not be used in pregnancy unless clearly necessary’ . Similarly, in the US, the Food and Drug Administration (FDA) lists onabotulinumtoxinA as category C meaning risk is not ruled out but patient benefits may warrant use of the drug in pregnancy . Based on the information and lack of evidence on its harm or safety in the existing literature, we counselled and appraised the patient in detail on the risk-benefit evaluation and choice of other treatment options. The patients were also informed that the typical course of migraine during pregnancy is towards improvement, particularly during the second and third trimester. The discussion was to help them decide if they wished to continue the treatment. Our patients were highly refractory migraineurs with most failing at least three and some up to 6 prior preventive treatments.
In the present study, the 32 patients who remained on treatment with onabotulinumtoxinA through pregnancy continued to show a good response whilst 11/13 who withdrew from treatment showed a relapse in their condition. One patient had a miscarriage (1/32) but the rest had normal vaginal or planned forceps or caesarean deliveries. It is hard to determine the true miscarriage rate in the general population due to reporting bias but the NHS estimated 1 in 8 known pregnancies will miscarry  and a 17% miscarriage rate was estimated in one US study . Of the 32 live births, there were no fetal malformations identified, compared to a UK average prevalence rate of 2% . Patients who continued to have onabotulinumtoxinA would have received injections in the second or third trimesters.
Patients in our study were refractory migraineurs with the vast majority tried and failed at least three oral preventive agents with many having tried more than 5–6 class of drugs. Hence, they were keen to continue their treatment with onabotulinumtoxinA in spite of uncertain impact of the drug on pregnancy. None of the patients were overusing painkillers as those with previous excessive analgesic consumption were asked to reduce the intake of painkillers prior to commencing onabotulinumtoxinA and were not on any other concomitant medications.
To our knowledge, there is only 1 case report in the literature of onabotulinumtoxinA treatment for chronic migraine during pregnancy . In this US study, a 26-year-old migraineur was treated with onabotulinumtoxinA but later chose to discontinue this because of the unknown risks in pregnancy. Her migraines worsened but then treatment resumed from 18 weeks gestation with good effect. Fetal movements were noted to be normal throughout and there was no interuterine growth restriction. The delivery was a planned caesarean section at 39 weeks because of breech presentation and the infant had an uneventful birth. The authors reviewed the child at 6.5 years and noted no complications to development.
Safety considerations are of paramount importance for all treatments in pregnancy. OnabotulinumtoxinA exerts its effects by inhibiting nerve terminal exocytosis and by doing so causes neuromuscular blockade. Although there are reports that diffusion of onabotulinumtoxinA after local injection does occur , in rabbits, injections of radioactive onabotulinumtoxinA into the eyelids did not lead to spread to distant structures in the body, including the eye . Moreover, onabotulinumtoxinA is a relatively large molecule of 150 kDa and is therefore unlikely to cross the placenta passively, although active transport cannot be excluded . Therefore, at least in theory, one might speculate that onabotulinumtoxinA injections should not have a significant effect on the uterus nor the fetus during pregnancy.
The challenge in conducting trials in pregnancy means that evidence is commonly derived from animal studies or observational studies at best. Animal studies have demonstrated that onabotulinumtoxinA injections are associated with reduced fetal body weight and skeletal ossification as well as abortions, early deliveries and maternal deaths . Such findings have not been confirmed in humans. A 24-year retrospective study of the Allergan global safety database reviewed 574 onabotulinumtoxinA treatments in pregnancy . Overall the prevalence of fetal defects was comparable in both the toxin treated and general populations. It is important to note that the toxin treated group was a heterogeneous population, with a large variety of indications and therefore differences in the sites and doses of toxins injected. This relatively large study did include 22 patients treated for migraine, however the specific outcomes of these pregnancies were not stated and overall results came from pooled data across the different indications . It is also worthy to note that 96% of all cases exposed to onabotulinumtoxinA in this study occurred either prior to conception or during the first trimester and so the effects of injections in the second and third trimester remain largely unknown. Brin et al. in the Allergan study also noted some cases of maternal botulism in the literature which did not report adverse effects on the pregnancy nor the fetus despite significant clinical neuromuscular weakness in the mother [19,20,21].