Nature of the Evidence Base
The current evidence base from clinical studies regarding the use of individual DMDs in pregnancy is summarised below. For convenience we have divided the DMDs arbitrarily into the commonly used categories of “first-line” or “platform” DMDs (which usually include dimethyl fumarate, glatiramer acetate, interferon β, and teriflunomide) and “high-efficacy” DMDs (alemtuzumab, cladribine tablets, fingolimod, natalizumab, ocrelizumab) [
17]. We have sought to cite papers from peer-reviewed journals here. However, the application of DMD-based therapy in the setting of pregnancy is an evolving science and many of the reports considered below are recent presentations at major congresses. These are included to provide the best snapshot possible of the current state of the science.
Glatiramer acetate This treatment has no teratogenic effects, according to data from national registries. In the Italian Multiple Sclerosis Register, analysis of data from 427 pregnancies in mothers with MS from 21 centres found no additional risk of spontaneous abortion or other adverse maternal or foetal outcomes [
28,
29]. A total of 151 women with MS in Germany had been taking glatiramer acetate before the pregnancy, of whom 148 discontinued treatment in the first trimester and 3 discontinued treatment in the second trimester; 95 pregnancies unexposed to DMDs served as a control group [
30]. There was no difference between groups for the proportions of live births or the risk of spontaneous abortion, any congenital anomaly, major congenital anomaly, preterm birth or need for caesarean section. In another study, evaluation of 5042 pregnancies exposed to glatiramer acetate demonstrated low and comparable rates of adverse pregnancy outcomes compared with data from two control databases of birth outcomes that together include > 1.7 million births each year [
31]. Glatiramer acetate therefore appears to be safe with regard to use in pregnancy, at least during the first trimester.
Dimethyl fumarate Analysis of 63 pregnancies in women enrolled in clinical trials and of 135 pregnancies arising from post-marketing reports revealed no adverse effects on pregnancy outcomes [
32]. An international registry is tracking pregnancies in women exposed to dimethyl fumarate; a recent report from this database (194 pregnancies with known outcome) showed that the rate of premature loss of the foetus was 9%, with live births occurring in the remainder, and a rate of birth defects of 4% [
33]. To date, therefore, dimethyl fumarate has not been associated with adverse pregnancy outcomes.
Interferon β Pregnancy outcomes with interferon β have been collected in major registries, namely the Italian Multiple Sclerosis Register (88 exposed and 308 unexposed pregnancies) [
28,
34], the German Multiple Sclerosis and Pregnancy Registry (251 exposed and 194 unexposed pregnancies) [
35], the Merck KGaA Global Drug Safety Database (1022 exposed pregnancies) [
36], and a Nordic Pregnancy Registry (875 exposed pregnancies, 1831 unexposed pregnancies) [
37]. Together, these studies showed that there was no excess risk to the foetus resulting from exposure in utero to interferon β, with regard to rates of live births, spontaneous abortions, or congenital abnormalities; the frequency of these outcomes was comparable to those observed in the general population. Mean birth weight and birth length were also consistent between neonates exposed or not exposed to interferon β in utero [
35,
37].
Teriflunomide and leflunomide An analysis of the global pharmacovigilance database for this agent found a rate of spontaneous abortion of 19% among 70 pregnancies with known exposure to teriflunomide, which was described as being within the range of rates expected for the general population (40% of these women underwent elective terminations of the pregnancy) [
38]. There were no congenital abnormalities in 26 live births. Most of the women who carried the pregnancy to term underwent the rapid elimination procedure for teriflunomide (23/26, 88%). A more recent (up to December 2017) survey of 437 teriflunomide-exposed pregnancies (220 with known outcomes) found a rate of spontaneous abortion of 21% [
39]. There were four birth defects (one considered major). These outcomes were again considered consistent with those expected from the general population, without demonstration of a teratogenic signal for teriflunomide.
The analysis from the global pharmacovigilance database for teriflunomide also documented no adverse birth outcomes from 22 pregnancies of partners of male patients taking teriflunomide [
38]. Similar results were found from a recent analysis of 232 pregnancies exposed to teriflunomide or leflunomide [
40]. The ongoing International Teriflunomide Pregnancy Exposure Registry is aiming to recruit 196 women with pregnancies exposed to teriflunomide (with at least 104 live births) from 17 countries [
41]. This population will provide 80% power to detect a 3.95-fold increase in the risk of birth defects associated with teriflunomide exposure compared with rates from the EUROCAT database.
Leflunomide, which acts as a prodrug for teriflunomide, has been in clinical use for arthritis for about 2 decades. Of 587 leflunomide-exposed pregnancies with known outcome, the rate of birth defects was 7%, with the majority being minor birth defects [
42]. According to the authors, these data suggested a lack of teratogenic potential for leflunomide, consistent with experience with teriflunomide, discussed above.