Воздействие корректирующих болезни методов лечения на результаты беременности при множественном склерозе: перспективное когортное исследование из немецкого реестра множественных склерозов и беременности.

Research in context Evidence before this study We searched PubMed until August, 2023, with the terms (“multiple sclerosis”, or “ms”) AND (“pregnancy”, “pregnancy outcome”, “first trimester”, “second trimester”, “third trimester” or “neonatal outcome”) AND (“interferon”, “glatiramer acetate”, “teriflunomide”, ”dimethyl fumarate”, “diroximel fumarate”, ”fumarates”, “S1P-modulators”, “fingolimod”, “ponesimod”, “alemtuzumab”, “natalizumab”, “ocrelizumab”, “ofatumumab”, “rituximab”, “CD20-antibodies” or “cladribine”). We carried out additional searches using Google Scholar. We identified several case studies, some clinical development programmes, and many cohort studies (ranging from 11 to over 2000 enrolled pregnancies) as well as systematic reviews. One systematic review included ten prospective studies and compared pregnancy and neonatal outcomes between disease modifying therapy (DMT) exposed and unexposed multiple sclerosis (MS) patients in almost 5000 pregnancies. Most cohort studies focus on the comparison of one specific DMT with an unexposed control group—usually using descriptive analysis without considering confounders. Despite publications with large sample sizes on pregnancy outcomes in MS patients in general, there is a lack of promising information on pregnancy and neonatal outcomes stratified by different DMT exposure and compared prospectively within a single registry with an unexposed control group. Added value of this study To the best of our knowledge, this cohort study is the largest prospective cohort study of pregnancy and neonatal outcomes in MS patients worldwide to date. It is unique in its stratification, categorizing participants into 12 exposure groups for descriptive analysis and nine exposure groups for multivariate analysis, including recently approved DMTs. This extensive stratification allowed us to directly compare the outcomes between different treatment regimens, providing a detailed and nuanced understanding of their impact on pregnancy and neonatal outcomes. In addition, the comprehensive data collection facilitated by the registry allowed for the inclusion of numerous potential confounders in our analyses, increasing the robustness and reliability of our findings. Implications of all the available evidence Interferon-β-formulations, Glatiramer acetate and Fumarates seem to be safe treatment options during early pregnancy for patients with a mild/moderate active disease course. In patients with highly active MS, Natalizumab and anti-CD20 antibodies have emerged as promising therapy options, but Sphingosin-1-Phosphate-modulators shall be avoided due to an increased growth retardation and potentially increased congenital anomaly risk. Further data from extended follow-up studies on the risk of maternal/infant infection, fetal growth retardation and timing of weight normalization are needed to improve pregnancy recommendations for highly-active patients. For contraindicated DMTs as Teriflunomid, Alemtuzumab, and Cladribine more data are needed in the near future and data from different registries should be merged to achieve sufficient cohort sizes. Taken together, our findings are critical to improving recommendations for the use of DMTs in the context of family planning in women with MS seeking pregnancy by combining established and emerging evidence and increasing the number of reported pregnancies.

Research in context

Evidence before this study

We searched PubMed until August, 2023, with the terms (“multiple sclerosis”, or “ms”) AND (“pregnancy”, “pregnancy outcome”, “first trimester”, “second trimester”, “third trimester” or “neonatal outcome”) AND (“interferon”, “glatiramer acetate”, “teriflunomide”, ”dimethyl fumarate”, “diroximel fumarate”, ”fumarates”, “S1P-modulators”, “fingolimod”, “ponesimod”, “alemtuzumab”, “natalizumab”, “ocrelizumab”, “ofatumumab”, “rituximab”, “CD20-antibodies” or “cladribine”). We carried out additional searches using Google Scholar. We identified several case studies, some clinical development programmes, and many cohort studies (ranging from 11 to over 2000 enrolled pregnancies) as well as systematic reviews. One systematic review included ten prospective studies and compared pregnancy and neonatal outcomes between disease modifying therapy (DMT) exposed and unexposed multiple sclerosis (MS) patients in almost 5000 pregnancies. Most cohort studies focus on the comparison of one specific DMT with an unexposed control group—usually using descriptive analysis without considering confounders. Despite publications with large sample sizes on pregnancy outcomes in MS patients in general, there is a lack of promising information on pregnancy and neonatal outcomes stratified by different DMT exposure and compared prospectively within a single registry with an unexposed control group.

Added value of this study

To the best of our knowledge, this cohort study is the largest prospective cohort study of pregnancy and neonatal outcomes in MS patients worldwide to date. It is unique in its stratification, categorizing participants into 12 exposure groups for descriptive analysis and nine exposure groups for multivariate analysis, including recently approved DMTs. This extensive stratification allowed us to directly compare the outcomes between different treatment regimens, providing a detailed and nuanced understanding of their impact on pregnancy and neonatal outcomes. In addition, the comprehensive data collection facilitated by the registry allowed for the inclusion of numerous potential confounders in our analyses, increasing the robustness and reliability of our findings.

Implications of all the available evidence

Interferon-β-formulations, Glatiramer acetate and Fumarates seem to be safe treatment options during early pregnancy for patients with a mild/moderate active disease course. In patients with highly active MS, Natalizumab and anti-CD20 antibodies have emerged as promising therapy options, but Sphingosin-1-Phosphate-modulators shall be avoided due to an increased growth retardation and potentially increased congenital anomaly risk. Further data from extended follow-up studies on the risk of maternal/infant infection, fetal growth retardation and timing of weight normalization are needed to improve pregnancy recommendations for highly-active patients.

For contraindicated DMTs as Teriflunomid, Alemtuzumab, and Cladribine more data are needed in the near future and data from different registries should be merged to achieve sufficient cohort sizes.

Taken together, our findings are critical to improving recommendations for the use of DMTs in the context of family planning in women with MS seeking pregnancy by combining established and emerging evidence and increasing the number of reported pregnancies.

Introduction

Over the past few decades, many new disease-modifying therapies (DMTs) for multiple sclerosis (MS) have been approved with limited safety data on exposure during pregnancy, even though MS mainly affects women of childbearing age. 1

The European Medicines Agency (EMA) mandates outcomes from at least 1000 pregnancies with first-trimester exposure to assess reproductive safety, 2 a requirement only available for the long established DMTs interferon-β-formulations (IFN-β) and glatiramer acetate (GLAT), 1 , 3 whereas Fumarates (dimethyl fumarate (DMF), diroximel fumarate (DRF)) also most likely do not cause fetal harm. 4 Others like Sphingosin-1-phosphat (S1P)-receptor-modulators, 5 Teriflunomid (TFL), 6 and Cladribine (CLAD) 7 are contraindicated in pregnancy due to a potential, but partly unproven, teratogenic risk or should be avoided during pregnancy, like Alemtuzumab (ALZ). 1 , 8

Although the relapse risk physiologically decreases during pregnancy, 9 pregnancy does not appear to protect women from rebound relapses, which is why women with high disease activity require effective treatment strategies during pregnancy. Natalizumab (NTZ) and anti-CD20 antibodies (Ocrelizumab (OCR), Rituximab (RTX)—off-label for MS, Ofatumumab (OFA)) have become practical therapeutic options with different dosing approaches for use before or throughout pregnancy. 1 While the EMA did not rate an increased teratogenic risk after NTZ first trimester exposure, 10 data on pregnancy outcomes after anti-CD20 antibodies exposure are insufficient. 1 , 11 Fetal and neonatal safety data are crucial factors for MS patients in choosing an appropriate DMT for family planning decisions. Therefore, a comprehensive analysis of pregnancy outcomes stratified by various DMT exposures including non-DMT exposed pregnancies, with the ability to adjust for important confounders is highly needed and the aim of the current study.

Methods

Study design and participants

Pregnancy data from MS-patients, voluntarily enrolled in the German Multiple Sclerosis and Pregnancy Registry (DMSKW) 12 between November 2006 and June 2023, with the following inclusion criteria were analyzed: available dates of last menstrual period, enrollment, and DMT start and stop. Pregnancies with unknown outcome (lost to follow up or ongoing), retrospective enrollment, multiple DMT exposure or exposure to Mitoxantrone, Azathioprine, Immunoglobulins or Daclizumab, were excluded ( Fig. 1 ). Data were primarily collected through standardized telephone interviews in each trimester and at 2, 6, and 12 months postpartum, and, if available, from medical records, maternity logs, or the children's medical check-up booklet. Treating physicians were contacted to verify data on hospitalizations and major congenital anomalies (MCAs). Fig. 1 Flowchart of enrolled pregnancies and stratification system. ALZ, alemtuzumab; CLAD, cladribine; d, days; DMF, dimethyl fumarate; DMT, disease-modifying therapy; DRF, diroximel fumarate; FTY, fingolimod; GLAT, glatiramer acetate; IFN-β, Interferone-β-formulation; LMP, last menstrual period; MS, multiple sclerosis; mth, months; NTZ, natalizumab; OCR, ocrelizumab; OFA, ofatumumab; OZM, ozanimod; PNE, ponesimod; RTX, rituximab; S1P, sphingosin-1-phosphat; TFL, teriflunomide; unexpo, unexposed. ∗retrospective enrollment: end of pregnancy before first contact. ∗∗last NTZ ≤ 1. trimester (until gestational day 83). ∗∗∗last NTZ during 2. trimester (gestational day 84–195). ∗∗∗∗last NTZ ≥ 3. trimester (gestational day 196 to end of pregnancy).

Flowchart of enrolled pregnancies and stratification system. ALZ, alemtuzumab; CLAD, cladribine; d, days; DMF, dimethyl fumarate; DMT, disease-modifying therapy; DRF, diroximel fumarate; FTY, fingolimod; GLAT, glatiramer acetate; IFN-β, Interferone-β-formulation; LMP, last menstrual period; MS, multiple sclerosis; mth, months; NTZ, natalizumab; OCR, ocrelizumab; OFA, ofatumumab; OZM, ozanimod; PNE, ponesimod; RTX, rituximab; S1P, sphingosin-1-phosphat; TFL, teriflunomide; unexpo, unexposed. ∗retrospective enrollment: end of pregnancy before first contact. ∗∗last NTZ ≤ 1. trimester (until gestational day 83). ∗∗∗last NTZ during 2. trimester (gestational day 84–195). ∗∗∗∗last NTZ ≥ 3. trimester (gestational day 196 to end of pregnancy).

The study was approved by the institutional review board of Ruhr-University Bochum (18-6474-BR). All women participated voluntarily and gave informed consent.

Outcomes

We defined pregnancy outcomes as spontaneous abortions (SAs; fetal loss before the 22nd completed gestational week (GW)), extrauterine pregnancies (EPs), elective terminations (ETs) with social or medical indication, stillbirths (SBs; fetal loss after the 22nd completed GW), occurrence of severe infections (Grade 3 or higher), 13 systemic antibiotic treatment during pregnancy, and preterm births (PBs; before 37th completed GW).

Neonatal outcomes include neonatal deaths (within 28 days postpartum), small for gestational age (SGA; birthweight ≤10th national age- and sex-specific percentile), 14 birthweight, birth length, head circumference, and MCAs reported up to 1 year postpartum or chromosomal anomalies (classified following the European network of population-based registries for the epidemiological surveillance of congenital anomalies (EUROCAT) guidelines). 15 As the classification is inconsistent, cases with hemangiomas requiring treatment are excluded from MCAs and reported separately.

Statistical analysis

For the analysis of pregnancy and neonatal outcomes, the cohort was stratified by DMT pregnancy exposure and each exposure-group was compared with a DMT unexposed control group (unexpo-group) ( Fig. 1 ). DMT exposure was determined (as previously defined 12 and summarized in Fig. 1 ) by product information from the EMA and FDA, and the drug's pharmacokinetic and pharmacodynamic characteristics. Mostly, this duration is equivalent to about five half-lives of the drug. The NTZ-group was further stratified by latest NTZ administration before/during the first trimester (1st-trimester NTZ-subgroup), during the second trimester (2nd-trimester NTZ-subgroup) or during/after the 3rd trimester (3rd-trimester NTZ-subgroup) to compare systemic antibiotic treatment during pregnancy, PBs, and neonatal outcomes. To evaluate the effect of the underlying disease on SGA and birthweight, we performed a sensitivity analysis comparing the unexpo-group, moderately-effective DMT-group (IFN-β, GLAT, TFL, Fumarates), and highly-effective DMT-group (S1P, ALZ, NTZ, anti-CD20, CLAD) with the general German population. 14 , 16

We analyzed data at a significance level of α = 0.05 using R software (Version 4.2.1) and Rstudio (Version 2023.12.1 + 402).

Descriptive statistics are reported as mean (SD) for continuous normally distributed data, as median (IQR) for non-parametric continuous data and number of pregnancies with outcome/number of pregnancies with available data (percent) (n/N [%]) for categorical data. Each exposure-group was compared to the unexpo-group using the t-test to compare continuous normally distributed outcomes, χ 2 -test for binary outcomes with more or equal to five events, and fisher's exact test for binary outcomes with less than five events.

We used multivariable logistic/linear regression models to estimate odds ratios (OR)/beta-coefficients (β) with 95% confidence intervals (CI) to compare exposure groups with sufficient sample size to the unexpo-group. Due to sample size concerns, the TFL-group (n = 35), ALZ-group (n = 22), and CLAD-group (n = 21) were only analyzed descriptively. Binary logistic regression models were used to estimate ORs for binary outcomes as SAs, antibiotic treatment during pregnancy, PBs, SGAs, and MCAs and linear regression models were used to estimate beta-coefficients for normally distributed continuous variables as birthweight, birth length, and head circumference. We adjusted all models for the factors age, body mass index (BMI), and tobacco use during pregnancy, and for the following factors as applicable: GW at study entry (for SAs, antibiotic treatments during pregnancy, and PBs) and GW at end of pregnancy (for antibiotic treatments during pregnancy, birthweights, birth lengths, head circumferences, and MCAs), high-dose glucocorticosteroids during pregnancy (for all outcomes except of SAs), gynecological complications (for all outcomes except of antibiotic treatments during pregnancy, and SAs), antibiotic use during pregnancy (for all outcomes except of antibiotic treatments during pregnancy, and SAs), newborn's sex (for birthweights, birth lengths, head circumferences, and MCAs), and plurality (for PBs, birthweights, birth lengths, head circumferences, and MCAs), with covariate correlations assessed. In addition, we adjusted for ‘year of enrollment’ for the comparison of MCA rates, to account for adaptions in our standardized questionnaire, and for improvements in diagnosis over the years. We assessed multicollinearity for each model among the predictor variables using the Variance Inflation Factor and we assessed model fit using the Hosmer–Lemeshow Goodness-of-Fit Test for our adjusted binary logistic models, with all outcomes yielding p-values >0.05. The adjusted R-squared values for all linear models exceeded 0.4, indicating satisfactory model performance. Missing data in outcomes and confounders led to a reduction in group sizes for the adjusted logistic/linear regression models. As the resulting reduction in group sizes were only minor, no imputation of missing values was used. The number of events and sample size (n/N [%]) for the crude and adjusted models, as well as a detailed description of all models including covariables is stated in Supplemental Tables a and b .

Role of the funding source

The DMSKW is partly supported by Almirall, Biogen, Hexal, Merck, Novartis, Roche, Sanofi Genzyme, Teva Pharma and Viatris.

The funders did not contribute to the study design, data collection, analysis, interpretation, report writing, or decision to publish.

Results

Baseline characteristics

We analyzed outcomes from 3722 pregnancies in 3247 women, with 1833 pregnancies in the moderately-effective DMT-group, 1052 pregnancies in the highly-effective DMT-group, and 837 in the unexpo-group ( Fig. 1 ).

The majority of women with DMT-exposed pregnancies (2260/2885: 78.3%) discontinued treatment either median 7.4 weeks (IQR: 14.4, 3.0) pre-pregnancy or within the first trimester (median GW: 4.4, IQR: 3.6–5.4), while 11.3% (327/2885) continued DMTs throughout pregnancy ( Table 1 ), mostly IFN-β (81/807: 10.0%), GLAT (88/596: 14.8%), and NTZ (152/674: 22.6%). In pregnancies exposed to contraindicated DMTs, such as S1P-modulators or TFL, the majority were unintended (S1P-group: 54.4%, TFL-group: 61.3%). Two participants received CLAD after the LMP, with the latest dose administered 5.6 weeks post-LMP, and therefore most likely before a positive pregnancy test. Table 1 Characteristics of the study sample. Unexpo-group (n = 837) reference category IFN-β-group (n = 807) GLAT-group (n = 596) TFL-group (n = 35) Fumarates-group (n = 395) S1P-group (n = 167) ALZ-group (n = 22) 1st-trimester NTZ-subgroup (n = 371) 2nd - trimester NTZ-subgroup (n = 105) 3rd-trimester NTZ-subgroup (n = 198) anti-CD20-group (n = 168) CLAD-group (n = 21) Therapy status  Therapy discontinuation prior to pregnancy NA 9/807 (1.1%) NA 16/35 (45.7%) NA 28/167 (16.8%) 21/22 (95.5%) 55/371 (14.8%) NA NA 140/168 (83.3%) 19/21 (90.5%)  Therapy discontinuation in first trimester NA 671/807 (83.1%) 454/596 (76.2%) 16/35 (45.7%) 371/395 (93.9%) 127/167 (76.0%) 0/22 309/371 (83.3%) NA NA 24/168 (14.3%) 0/21  Therapy discontinuation after first trimester NA 31/807 (3.8%) 21/596 (3.5%) 1/35 (2.9%) 8/395 (2.0%) 6/167 (3.6%) 0/22 NA 75/105 (71.4%) 46/198 (23.2%) 0/168 0/21  Therapy during entire pregnancy NA 81/807 (10.0%) 88/596 (14.8%) 0/35 6/395 (1.5%) 0/167 0/22 NA NA 152/198 (76.8%) 0/168 0/21  DMT start initially during pregnancy NA 8/807 (1.0%) 14/596 (2.3%) 2/35 (5.7%) 10/395 (2.5%) 6/167 (3.6%) 1/22 (4.5%) 4/371 (1.1%) 7/105 (6.7%) 0/198 4/168 (2.4%) 2/21 (9.5%)  DMT restart during pregnancy a NA 7/807 (0.9%) 19/596 (3.2%) 0/35 0/395 0/167 0/22 3/371 (0.8%) 23/105 (21.9%) 0/198 0/168 0/21  Exposure duration in weeks prior to/during pregnancy b NA 4.7 (3.9–7.0) 5.0 (4.1–7.5) 1.6 (−12.1 to 4.7) 4.4 (3.9–5.1) 4.6 (2.9–5.9) −11.0 (−13.9 to −6.0) 2.3 (0.6–3.9) 25.4 (18.7–26.8) 32.6 (30.7–39.7) −9.3 (−15.7 to −3.5) −17.3 (−21.6 to −8.1) Demographic information  Age at LMP 32.4 (4.2) 32.4 (4.0) 32.8 (4.0) 31.4 (4.6) 32.4 (4.2) 32.5 (5.2) 29.8 (3.6) 32.0 (4.0) 31.4 (5.1) 32.2 (4.2) 32.7 (3.9) 29.2 (3.7)  GW at study entry 12.7 (7.7–22.4) 11.6 (6.9–20.8) 10.2 (6.1–19.7) 9.9 (6.6–17.6) 8.6 (6.0–14.6) 8.3 (5.7–16.3) 8.7 (6.7–17.0) 9.4 (6.3–16.2) 10.7 (6.7–20.7) 8.1 (5.8–13.7) 7.4 (5.3–11.6) 7.4 (5.7–10.6)  BMI at enrollment in kg/m 2 24.3 (5.1) 24.6 (5.2) 24.4 (4.9) 26.4 (6.5) 23.8 (5.0) 25.1 (6.6) 24.8 (4.8) 24.3 (5.3) 23.7 (5.4) 23.9 (4.5) 24.3 (5.3) 26.1 (5.4)  Smoking during pregnancy 35/827 (4.2%) 27/802 (3.4%) 8/595 (1.3%) 6/35 (17.1%) 16/395 (4.1%) 15/165 (9.1%) 3/22 (13.6%) 26/341 (7.6%) 9/99 (9.1%) 6/196 (3.1%) 4/168 (2.4%) 2/21 (9.5%)  Drugs/alcohol during pregnancy 4/827 (0.5%) 2/806 (0.2%) 2/596 (0.3%) 0/35 3/395 (0.8%) 1/164 (0.6%) 0/22 2/337 (0.6%) 1/99 (1.0%) 3/196 (1.5%) 2/166 (1.2%) 0/21  High-dose glucocorticosteroids during pregnancy 98/837 (11.7%) 48/807 (5.9%) 44/596 (7.4%) 3/35 (8.6%) 18/395 (4.6%) 30/167 (18.0%) 1/22 (4.5%) 60/371 (16.2%) 23/105 (21.9%) c 10/198 (5.1%) 6/168 (3.6%) 0/21  Disease duration at conception, years 5.2 (2.4–8.8) 5.0 (2.3–8.6) 3.4 (1.6–6.9) 4.1 (2.4–7.1) 4.1 (2.0–7.7) 7.5 (4.4–11.4) 4.9 (2.5–8.0) 7.0 (4.2–10.7) 8.1 (4.6–11.4) 8.2 (4.4–11.4) 7.3 (4.3–11.3) 5.3 (4.2–7.7)  MS-Type: RRMS 772/837 (92.2%) 735/807 (91.1%) 563/596 (94.5%) 33/35 (94.3%) 386/395 (97.7%) 161/167 (96.4%) 22/22 (100%) 368/371 (99.2%) 104/105 (99.0%) 197/198 (99.5%) 162/168 (96.4%) 21/21 (100%) Data are n/N (%) or mean (SD) or median (IQR). ALZ, alemtuzumab; BMI, body mass index; CLAD, cladribine; DMT, disease modifying therapy; EDSS, expanded disability status scale; GLAT, glatiramer acetate; GW, gestational week; IFN-β, interferon-β-formulations; IQR, interquartile range; LMP, last menstrual period; MS, multiple sclerosis; n, number of pregnancies with outcome; N, number of pregnancies with available data; NA, not applicable; NTZ, natalizumab; RRMS, relapsing remitting multiple sclerosis; S1P-modulators, sphingosine-1-receptor-modulators; SD, standard deviation; TFL, teriflunomide; unexpo, unexposed. a Dosage interval: IFN-β 1a: >8 days, IFN-β 1b: >8 days, Peg-IFN-β: >21 days, GLAT: >8 days, NTZ >90 days. b Negative values indicate that last application occurred prior to LMP, and positive values indicate that the last application occurred after LMP. c 4/23 (17.4%) started therapy during pregnancy; 16/23 (69.6%) stopped therapy after pregnancy onset and restarted after a pregnancy relapse.

Table 1 Characteristics of the study sample. Unexpo-group (n = 837) reference category IFN-β-group (n = 807) GLAT-group (n = 596) TFL-group (n = 35) Fumarates-group (n = 395) S1P-group (n = 167) ALZ-group (n = 22) 1st-trimester NTZ-subgroup (n = 371) 2nd - trimester NTZ-subgroup (n = 105) 3rd-trimester NTZ-subgroup (n = 198) anti-CD20-group (n = 168) CLAD-group (n = 21) Therapy status  Therapy discontinuation prior to pregnancy NA 9/807 (1.1%) NA 16/35 (45.7%) NA 28/167 (16.8%) 21/22 (95.5%) 55/371 (14.8%) NA NA 140/168 (83.3%) 19/21 (90.5%)  Therapy discontinuation in first trimester NA 671/807 (83.1%) 454/596 (76.2%) 16/35 (45.7%) 371/395 (93.9%) 127/167 (76.0%) 0/22 309/371 (83.3%) NA NA 24/168 (14.3%) 0/21  Therapy discontinuation after first trimester NA 31/807 (3.8%) 21/596 (3.5%) 1/35 (2.9%) 8/395 (2.0%) 6/167 (3.6%) 0/22 NA 75/105 (71.4%) 46/198 (23.2%) 0/168 0/21  Therapy during entire pregnancy NA 81/807 (10.0%) 88/596 (14.8%) 0/35 6/395 (1.5%) 0/167 0/22 NA NA 152/198 (76.8%) 0/168 0/21  DMT start initially during pregnancy NA 8/807 (1.0%) 14/596 (2.3%) 2/35 (5.7%) 10/395 (2.5%) 6/167 (3.6%) 1/22 (4.5%) 4/371 (1.1%) 7/105 (6.7%) 0/198 4/168 (2.4%) 2/21 (9.5%)  DMT restart during pregnancy a NA 7/807 (0.9%) 19/596 (3.2%) 0/35 0/395 0/167 0/22 3/371 (0.8%) 23/105 (21.9%) 0/198 0/168 0/21  Exposure duration in weeks prior to/during pregnancy b NA 4.7 (3.9–7.0) 5.0 (4.1–7.5) 1.6 (−12.1 to 4.7) 4.4 (3.9–5.1) 4.6 (2.9–5.9) −11.0 (−13.9 to −6.0) 2.3 (0.6–3.9) 25.4 (18.7–26.8) 32.6 (30.7–39.7) −9.3 (−15.7 to −3.5) −17.3 (−21.6 to −8.1) Demographic information  Age at LMP 32.4 (4.2) 32.4 (4.0) 32.8 (4.0) 31.4 (4.6) 32.4 (4.2) 32.5 (5.2) 29.8 (3.6) 32.0 (4.0) 31.4 (5.1) 32.2 (4.2) 32.7 (3.9) 29.2 (3.7)  GW at study entry 12.7 (7.7–22.4) 11.6 (6.9–20.8) 10.2 (6.1–19.7) 9.9 (6.6–17.6) 8.6 (6.0–14.6) 8.3 (5.7–16.3) 8.7 (6.7–17.0) 9.4 (6.3–16.2) 10.7 (6.7–20.7) 8.1 (5.8–13.7) 7.4 (5.3–11.6) 7.4 (5.7–10.6)  BMI at enrollment in kg/m 2 24.3 (5.1) 24.6 (5.2) 24.4 (4.9) 26.4 (6.5) 23.8 (5.0) 25.1 (6.6) 24.8 (4.8) 24.3 (5.3) 23.7 (5.4) 23.9 (4.5) 24.3 (5.3) 26.1 (5.4)  Smoking during pregnancy 35/827 (4.2%) 27/802 (3.4%) 8/595 (1.3%) 6/35 (17.1%) 16/395 (4.1%) 15/165 (9.1%) 3/22 (13.6%) 26/341 (7.6%) 9/99 (9.1%) 6/196 (3.1%) 4/168 (2.4%) 2/21 (9.5%)  Drugs/alcohol during pregnancy 4/827 (0.5%) 2/806 (0.2%) 2/596 (0.3%) 0/35 3/395 (0.8%) 1/164 (0.6%) 0/22 2/337 (0.6%) 1/99 (1.0%) 3/196 (1.5%) 2/166 (1.2%) 0/21  High-dose glucocorticosteroids during pregnancy 98/837 (11.7%) 48/807 (5.9%) 44/596 (7.4%) 3/35 (8.6%) 18/395 (4.6%) 30/167 (18.0%) 1/22 (4.5%) 60/371 (16.2%) 23/105 (21.9%) c 10/198 (5.1%) 6/168 (3.6%) 0/21  Disease duration at conception, years 5.2 (2.4–8.8) 5.0 (2.3–8.6) 3.4 (1.6–6.9) 4.1 (2.4–7.1) 4.1 (2.0–7.7) 7.5 (4.4–11.4) 4.9 (2.5–8.0) 7.0 (4.2–10.7) 8.1 (4.6–11.4) 8.2 (4.4–11.4) 7.3 (4.3–11.3) 5.3 (4.2–7.7)  MS-Type: RRMS 772/837 (92.2%) 735/807 (91.1%) 563/596 (94.5%) 33/35 (94.3%) 386/395 (97.7%) 161/167 (96.4%) 22/22 (100%) 368/371 (99.2%) 104/105 (99.0%) 197/198 (99.5%) 162/168 (96.4%) 21/21 (100%) Data are n/N (%) or mean (SD) or median (IQR). ALZ, alemtuzumab; BMI, body mass index; CLAD, cladribine; DMT, disease modifying therapy; EDSS, expanded disability status scale; GLAT, glatiramer acetate; GW, gestational week; IFN-β, interferon-β-formulations; IQR, interquartile range; LMP, last menstrual period; MS, multiple sclerosis; n, number of pregnancies with outcome; N, number of pregnancies with available data; NA, not applicable; NTZ, natalizumab; RRMS, relapsing remitting multiple sclerosis; S1P-modulators, sphingosine-1-receptor-modulators; SD, standard deviation; TFL, teriflunomide; unexpo, unexposed. a Dosage interval: IFN-β 1a: >8 days, IFN-β 1b: >8 days, Peg-IFN-β: >21 days, GLAT: >8 days, NTZ >90 days. b Negative values indicate that last application occurred prior to LMP, and positive values indicate that the last application occurred after LMP. c 4/23 (17.4%) started therapy during pregnancy; 16/23 (69.6%) stopped therapy after pregnancy onset and restarted after a pregnancy relapse.

Characteristics of the study sample.

Data are n/N (%) or mean (SD) or median (IQR).

ALZ, alemtuzumab; BMI, body mass index; CLAD, cladribine; DMT, disease modifying therapy; EDSS, expanded disability status scale; GLAT, glatiramer acetate; GW, gestational week; IFN-β, interferon-β-formulations; IQR, interquartile range; LMP, last menstrual period; MS, multiple sclerosis; n, number of pregnancies with outcome; N, number of pregnancies with available data; NA, not applicable; NTZ, natalizumab; RRMS, relapsing remitting multiple sclerosis; S1P-modulators, sphingosine-1-receptor-modulators; SD, standard deviation; TFL, teriflunomide; unexpo, unexposed.

Dosage interval: IFN-β 1a: >8 days, IFN-β 1b: >8 days, Peg-IFN-β: >21 days, GLAT: >8 days, NTZ >90 days.

Negative values indicate that last application occurred prior to LMP, and positive values indicate that the last application occurred after LMP.

4/23 (17.4%) started therapy during pregnancy; 16/23 (69.6%) stopped therapy after pregnancy onset and restarted after a pregnancy relapse.

High-dose glucocorticosteroids during pregnancy were highly dependent on pre-pregnancy DMT use and ranged from 0% in the CLAD-group to 21.9% in the 2nd-trimester NTZ-subgroup. In the 2nd-trimester NTZ-subgroup 16/23 (69.6%) participants with high-dose glucocorticosteroids during pregnancy withdrew NTZ at the onset of pregnancy and restarted NTZ after a pregnancy relapse.

Most women (2110/3722: 56.7%) contacted us during the first trimester, the unexpo-group significantly later (GW 12.7, IQR: 7.7, 22.4) than women in the moderately-effective (GW 10.3, IQR 6.4, 18.9, p-value: <0.001) or highly-effective DMT-group (GW 8.6, IQR 6.0, 15.0, p-value: <0.001).

Pregnancy outcomes

EP (5/3722: 0.1%), ETs (21/3722: 0.6%), and SBs (4/3722: 0.1%) were rare ( Table 2 ). ETs due to fetal defects or maternal risk did not differ between groups, whereas socially indicated ETs were significantly more common with DMTs contraindicated in pregnancy. SAs occurred in 8.2% (245/2977) of all women enrolled prior to GW 22 (77/381: 20.2% in women enrolled prior to GW 6), but did not differ significantly between groups ( Table 2 ). Table 2 Pregnancy outcomes descriptive analysis. Unexpo-group (n = 837) reference category IFN-β-group (n = 807) GLAT-group (n = 596) TFL-group (n = 35) Fumarates-group (n = 395) S1P-group (n = 167) ALZ-group (n = 22) 1st-trimester NTZ-subgroup (n = 371) 2nd - trimester NTZ-subgroup (n = 105) 3rd - trimester NTZ-subgroup (n = 198) anti-CD20-group (n = 168) CLAD-group (n = 21) Adverse pregnancy outcomes  Extrauterine pregnancy 1/837 (0.1%) 0/807 0/596 0/35 1/395 (0.3%) 1/167 (0.6%) 0/22 1/674 (0.1%) 1/168 (0.6%) 0/21  Spontaneous abortion 40/837 (4.8%) 49/807 (6.1%) 33/596 (5.5%) 1/35 (2.9%) 28/395 (7.1%) 16/167 (9.6%) 2/22 (9.1%) 62/674 (9.2%) 14/168 (8.3%) 0/21  GW at enrollment ≤ 22 40/622 (6.4%) 49/618 (7.9%) 33/473 (7.0%) 1/29 (3.4%) 28/339 (8.3%) 16/141 (11.3%) 2/20 (10.0%) 62/563 (11.0%) 14/153 (9.2%) 0/19  GW at enrollment < 6 15/54 (27.8%) 12/65 (18.5%) 18/76 (23.7%) 0/2 9/54 (16.7%) 4/23 (17.4%) 0/2 16/70 (22.9%) 3/30 (10.0%) 0/5  Elective termination 4/837 (0.5%) 4/807 (0.5%) 2/596 (0.3%) 1/35 (2.9%) 2/395 (0.5%) 4/167 (2.4%) 0/22 3/674 (0.4%) 0/168 1/21 (4.8%)  Fetal defects 3/837 (0.4%) 2/807 (0.2%) 0/596 0/35 1/395 (0.3%) 0/167 0/22 3/674 (0.4%) 0/168 0/21  Maternal risk 0/837 1/807 (0.1%) 0/596 0/35 0/395 0/167 0/22 0/674 0/168 0/21  Social indication 0/837 0/807 1/596 (0.2%) 1/35 (2.9%) 1/395 (0.3%) 3/167 (1.8%) 0/22 0/674 0/168 1/21 (4.8%)  Unknown indication 1/837 (0.1%) 1/807 (0.1%) 1/596 (0.2%) 0/35 0/395 1/167 (0.6%) 0/22 0/674 0/168 0/21  Still birth 1/837 (0.1%) 1/807 (0.1%) 0/596 0/35 1/395 (0.3%) 1/167 (0.6%) 0/22 0/674 0/168 0/21  With fetal defects 0/837 1/807 (0.1%) 0/596 0/35 1/395 (0.3%) 0/167 0/22 0/674 0/168 0/21  Without fetal defects 1/837 (0.1%) 0/807 0/596 0/35 0/395 1/167 (0.6%) 0/22 0/674 0/168 0/21 Infections during pregnancy  Severe infections during pregnancy 8/837 (1.0%) 9/807 (1.1%) 7/596 (1.2%) 0/35 11/395 (2.8%) 5/167 (3.0%) 2/22 (9.1%) 4/371 (1.1%) 0/105 6/198 (3.0%) 1/168 (0.6%) 1/21 (4.8%)  GW at birth >22 8/791 (1.0%) 9/755 (1.2%) 7/561 (1.2%) 0/33 11/364 (3.0%) 5/146 (3.4%) 2/19 (10.5%) 4/310 (1.3%) 0/102 6/198 (3.0%) 1/153 (0.7%) 1/20 (5.0%)  Antibiotic treatment during pregnancy 101/837 (12.1%) 84/807 (10.4%) 70/596 (11.7%) 7/35 (20.0%) 53/395 (13.4%) 28/167 (16.8%) 4/22 (18.2%) 42/371 (11.3%) 28/105 (26.7%) 41/198 (20.7%) 39/168 (23.2%) 5/21 (23.8%)  GW at birth >22 100/791 (12.6%) 84/755 (11.1%) 70/561 (12.5%) 7/33 (21.2%) 53/364 (14.6%) 27/146 (18.5%) 4/19 (21.1%) 38/310 (12.3%) 28/102 (27.5%) 41/198 (20.7%) 39/153 (25.5%) 5/20 (25.0%) Pregnancy outcomes  Live birth 791/837 (94.5%) 754/807 (93.4%) 561/596 (94.1%) 33/35 (94.3%) 363/395 (91.9%) 145/167 (86.8%) 20/22 (90.9%) 309/371 (83.3%) 101/105 (96.2%) 198/198 (100%) 153/168 (91.1%) 20/21 (95.2%)  GW at birth 39.6 (38.4–40.4) 39.7 (38.4–40.6) 39.6 (38.6–40.4) 39.1 (38.3–40.3) 39.6 (38.6–40.6) 39.4 (38.1–40.4) 39.9 (38.5–40.8) 39.3 (38.1–40.3) 39.1 (38.1–40.1) 39.6 (38.7–40.1) 39.6 (38.6–40.4) 39.6 (38.6–40.4)  Preterm birth 71/791 (9.0%) 66/754 (8.8%) 44/561 (7.8%) 7/33 (21.2%) 30/363 (8.3%) 18/145 (12.4%) 3/20 (15.0%) 29/309 (9.4%) 13/101 (12.9%) 19/198 (9.6%) 14/153 (9.2%) 2/20 (10.0%)  GW at enrollement < 37 71/763 (9.3%) 66/732 (9.0%) 44/548 (8.0%) 7/32 (21.9%) 30/358 (8.4%) 18/144 (12.5%) 3/20 (15.0%) 29/305 (9.5%) 13/99 (13.1%) 19/195 (9.7%) 14/152 (9.2%) 2/19 (10.5%)  Plurality 11/791 (1.4%) 11/754 (1.5%) 7/561 (1.2%) 1/33 (3.0%) 3/363 (0.8%) 2/145 (1.4%) 1/20 (5.0%) 5/309 (1.6%) 1/101 (1.0%) 2/198 (1.0%) 5/153 (3.3%) 0/20  Twins 11 10 7 1 3 2 1 5 1 2 5 0  Triplets 0 1 0 0 0 0 0 0 0 0 0 0  Number of live born children 802 766 568 34 366 147 21 314 102 200 158 20 Data are n/N (%) or median (IQR). ALZ, alemtuzumab; CLAD, cladribine; GLAT, glatiramer acetate; IFN-β, interferon-β-formulations; IQR, interquartile range; GW, gestational week; n, number of pregnancies with outcome; N, number of pregnancies with available data; NTZ, natalizumab; S1P, sphingosin-1-phosphat; TFL, teriflunomide; unexpo, unexposed.

Table 2 Pregnancy outcomes descriptive analysis. Unexpo-group (n = 837) reference category IFN-β-group (n = 807) GLAT-group (n = 596) TFL-group (n = 35) Fumarates-group (n = 395) S1P-group (n = 167) ALZ-group (n = 22) 1st-trimester NTZ-subgroup (n = 371) 2nd - trimester NTZ-subgroup (n = 105) 3rd - trimester NTZ-subgroup (n = 198) anti-CD20-group (n = 168) CLAD-group (n = 21) Adverse pregnancy outcomes  Extrauterine pregnancy 1/837 (0.1%) 0/807 0/596 0/35 1/395 (0.3%) 1/167 (0.6%) 0/22 1/674 (0.1%) 1/168 (0.6%) 0/21  Spontaneous abortion 40/837 (4.8%) 49/807 (6.1%) 33/596 (5.5%) 1/35 (2.9%) 28/395 (7.1%) 16/167 (9.6%) 2/22 (9.1%) 62/674 (9.2%) 14/168 (8.3%) 0/21  GW at enrollment ≤ 22 40/622 (6.4%) 49/618 (7.9%) 33/473 (7.0%) 1/29 (3.4%) 28/339 (8.3%) 16/141 (11.3%) 2/20 (10.0%) 62/563 (11.0%) 14/153 (9.2%) 0/19  GW at enrollment < 6 15/54 (27.8%) 12/65 (18.5%) 18/76 (23.7%) 0/2 9/54 (16.7%) 4/23 (17.4%) 0/2 16/70 (22.9%) 3/30 (10.0%) 0/5  Elective termination 4/837 (0.5%) 4/807 (0.5%) 2/596 (0.3%) 1/35 (2.9%) 2/395 (0.5%) 4/167 (2.4%) 0/22 3/674 (0.4%) 0/168 1/21 (4.8%)  Fetal defects 3/837 (0.4%) 2/807 (0.2%) 0/596 0/35 1/395 (0.3%) 0/167 0/22 3/674 (0.4%) 0/168 0/21  Maternal risk 0/837 1/807 (0.1%) 0/596 0/35 0/395 0/167 0/22 0/674 0/168 0/21  Social indication 0/837 0/807 1/596 (0.2%) 1/35 (2.9%) 1/395 (0.3%) 3/167 (1.8%) 0/22 0/674 0/168 1/21 (4.8%)  Unknown indication 1/837 (0.1%) 1/807 (0.1%) 1/596 (0.2%) 0/35 0/395 1/167 (0.6%) 0/22 0/674 0/168 0/21  Still birth 1/837 (0.1%) 1/807 (0.1%) 0/596 0/35 1/395 (0.3%) 1/167 (0.6%) 0/22 0/674 0/168 0/21  With fetal defects 0/837 1/807 (0.1%) 0/596 0/35 1/395 (0.3%) 0/167 0/22 0/674 0/168 0/21  Without fetal defects 1/837 (0.1%) 0/807 0/596 0/35 0/395 1/167 (0.6%) 0/22 0/674 0/168 0/21 Infections during pregnancy  Severe infections during pregnancy 8/837 (1.0%) 9/807 (1.1%) 7/596 (1.2%) 0/35 11/395 (2.8%) 5/167 (3.0%) 2/22 (9.1%) 4/371 (1.1%) 0/105 6/198 (3.0%) 1/168 (0.6%) 1/21 (4.8%)  GW at birth >22 8/791 (1.0%) 9/755 (1.2%) 7/561 (1.2%) 0/33 11/364 (3.0%) 5/146 (3.4%) 2/19 (10.5%) 4/310 (1.3%) 0/102 6/198 (3.0%) 1/153 (0.7%) 1/20 (5.0%)  Antibiotic treatment during pregnancy 101/837 (12.1%) 84/807 (10.4%) 70/596 (11.7%) 7/35 (20.0%) 53/395 (13.4%) 28/167 (16.8%) 4/22 (18.2%) 42/371 (11.3%) 28/105 (26.7%) 41/198 (20.7%) 39/168 (23.2%) 5/21 (23.8%)  GW at birth >22 100/791 (12.6%) 84/755 (11.1%) 70/561 (12.5%) 7/33 (21.2%) 53/364 (14.6%) 27/146 (18.5%) 4/19 (21.1%) 38/310 (12.3%) 28/102 (27.5%) 41/198 (20.7%) 39/153 (25.5%) 5/20 (25.0%) Pregnancy outcomes  Live birth 791/837 (94.5%) 754/807 (93.4%) 561/596 (94.1%) 33/35 (94.3%) 363/395 (91.9%) 145/167 (86.8%) 20/22 (90.9%) 309/371 (83.3%) 101/105 (96.2%) 198/198 (100%) 153/168 (91.1%) 20/21 (95.2%)  GW at birth 39.6 (38.4–40.4) 39.7 (38.4–40.6) 39.6 (38.6–40.4) 39.1 (38.3–40.3) 39.6 (38.6–40.6) 39.4 (38.1–40.4) 39.9 (38.5–40.8) 39.3 (38.1–40.3) 39.1 (38.1–40.1) 39.6 (38.7–40.1) 39.6 (38.6–40.4) 39.6 (38.6–40.4)  Preterm birth 71/791 (9.0%) 66/754 (8.8%) 44/561 (7.8%) 7/33 (21.2%) 30/363 (8.3%) 18/145 (12.4%) 3/20 (15.0%) 29/309 (9.4%) 13/101 (12.9%) 19/198 (9.6%) 14/153 (9.2%) 2/20 (10.0%)  GW at enrollement < 37 71/763 (9.3%) 66/732 (9.0%) 44/548 (8.0%) 7/32 (21.9%) 30/358 (8.4%) 18/144 (12.5%) 3/20 (15.0%) 29/305 (9.5%) 13/99 (13.1%) 19/195 (9.7%) 14/152 (9.2%) 2/19 (10.5%)  Plurality 11/791 (1.4%) 11/754 (1.5%) 7/561 (1.2%) 1/33 (3.0%) 3/363 (0.8%) 2/145 (1.4%) 1/20 (5.0%) 5/309 (1.6%) 1/101 (1.0%) 2/198 (1.0%) 5/153 (3.3%) 0/20  Twins 11 10 7 1 3 2 1 5 1 2 5 0  Triplets 0 1 0 0 0 0 0 0 0 0 0 0  Number of live born children 802 766 568 34 366 147 21 314 102 200 158 20 Data are n/N (%) or median (IQR). ALZ, alemtuzumab; CLAD, cladribine; GLAT, glatiramer acetate; IFN-β, interferon-β-formulations; IQR, interquartile range; GW, gestational week; n, number of pregnancies with outcome; N, number of pregnancies with available data; NTZ, natalizumab; S1P, sphingosin-1-phosphat; TFL, teriflunomide; unexpo, unexposed.

Pregnancy outcomes descriptive analysis.

Data are n/N (%) or median (IQR).

ALZ, alemtuzumab; CLAD, cladribine; GLAT, glatiramer acetate; IFN-β, interferon-β-formulations; IQR, interquartile range; GW, gestational week; n, number of pregnancies with outcome; N, number of pregnancies with available data; NTZ, natalizumab; S1P, sphingosin-1-phosphat; TFL, teriflunomide; unexpo, unexposed.

Severe infections during pregnancy were rare (54/3452: 1.6%–in pregnancies lasting >22 GW) and occurred without a specific pattern ( Supplemental Table c ), but were observed significantly more often in Fumarates (11/395: 2.8% vs. 8/837 unexposed-group: 1.0%, p-value: 0.03) and ALZ-group (2/22: 9.1% vs. 8/837: 1.0% unexpo-group, p-value: 0.02) compared to the unexpo-group.

On the other hand, systemic antibiotic treatment during pregnancy was significantly higher in the 2nd-trimester NTZ-subgroup (adjusted OR: 2.47, CI: 1.47, 4.05, p-value: <0.001), 3rd-trimester NTZ-subgroup (adjusted OR: 1.75, CI: 1.15, 2.63, p-value: 0.01) and anti-CD20-group (adjusted OR: 2.16, CI: 1.41, 3.29, p-value: <0.001) ( Fig. 2 and Supplemental Table a ). Fig. 2 Impact of DMTs on pregnancy and neonatal outcomes. Adjusted linear/logistic regression models: Odds ratio (logistic) or β (linear). 1 Odds ratio—binary logistic regression model, 2 β—linear regression model. DMT, disease modifying therapy; GLAT, glatiramer acetate; IFN-β, interferon-β-formulations; MCA, major congenital anomaly; NTZ, natalizumab; S1P, sphingosine-1-phosphat; SGA, small for gestational age; unexpo, unexposed.

Impact of DMTs on pregnancy and neonatal outcomes. Adjusted linear/logistic regression models: Odds ratio (logistic) or β (linear). 1 Odds ratio—binary logistic regression model, 2 β—linear regression model. DMT, disease modifying therapy; GLAT, glatiramer acetate; IFN-β, interferon-β-formulations; MCA, major congenital anomaly; NTZ, natalizumab; S1P, sphingosine-1-phosphat; SGA, small for gestational age; unexpo, unexposed.

The majority of pregnancies (3448/3722, 92.6%) resulted in live births, with 316 PBs (9.4% out of 3367 pregnancies enrolled before GW 37), significantly more often after TFL pregnancy exposure (7/32: 21.9% vs. 71/763: 9.3% unexpo-group, p-value: 0.03).

Neonatal outcomes

3399 singleton, 48 twin and 1 triplet pregnancy resulted in 3498 liveborn neonates, of which 14 (0.4%) ended in neonatal deaths: 12 (85.7%) attributed to PBs, one caused by a chromosomal defect with multiple MCAs (unexpo-group), and one for no apparent reason. Both neonatal deaths within the ALZ-group occurred in one preterm (20th GW) twin pregnancy ( Table 3 ). Table 3 Neonatal Outcomes descriptive analysis. Unexpo-group (n = 802) reference category IFN-β-group (n = 766) GLAT-group (n = 568) TFL-group (n = 34) Fumarates-group (n = 366) S1P-group (n = 147) ALZ-group (n = 21) 1st-trimester NTZ-subgroup (n = 314) 2nd-trimester NTZ-subgroup (n = 102) 3rd-trimester NTZ-subgroup (n = 200) anti-CD20-group (n = 158) CLAD-group (n = 20) Neonatal death 3/802 (0.4%) 4/766 (0.5%) 1/568 (0.2%) 0/34 1/366 (0.3%) 0/147 2/21 (9.5%) 2/314 (0.6%) 1/102 (1.0%) 0/200 0/158 0/20  Thereof preterm births 2/3 (66.7%) 4/4 (100%) 1/1 (100%) NA 1/1 (100%) NA 2/2 (100%) 2/2 (100%) 1/1 (100%) NA NA NA SGA a 140/795 (17.6%) 124/753 (16.5%) 100/561 (17.8%) 4/34 (11.8%) 73/362 (20.2%) 40/146 (27.4%) 1/20 (5.0%) 60/312 (19.2%) 20/100 (20.0%) 49/198 (24.7%) 38/158 (24.1%) 2/20 (10.0%) Birthweight b , in g 3414.2 (443.6) 3415.0 (449.7) 3400.6 (429.1) 3391.1 (332.9) 3392.8 (447.7) 3277.2 (450.2) 3432.9 (381.9) 3355.9 (436.8) 3304.8 (443.9) 3310.0 (427.2) 3423.0 (449.3) 3485.0 (418.0)  Missing 0/716 3/680 (0.4%) 2/513 (0.4%) 0/26 2/330 (0.6%) 1/125 (0.7%) 0/17 0/276 0/88 0/178 0/138 0/18 Birth length b , in cm 51.7 (2.5) 51.6 (2.5) 51.7 (2.3) 51.3 (2.0) 51.5 (2.4) 50.7 (2.4) 51.8 (3.0) 51.4 (2.5) 51.1 (2.5) 51.0 (2.6) 51.8 (2.5) 52.2 (2.6)  Missing 0/716 4/680 (0.6%) 2/513 (0.4%) 0/26 3/330 (0.9%) 1/125 (0.8%) 0/17 2/276 (0.7%) 0/88 0/178 1/138 (0.7%) 0/18 Head circumference b , in cm 35.1 (1.4) 35.0 (1.5) 35.1 (1.4) 35.0 (1.4) 34.9 (1.4) 35.0 (1.5) 35.2 (1.2) 35.2 (1.4) 34.6 (1.6) 35.0 (1.4) 35.0 (1.4) 35.1 (1.4)  Missing 83/716 (11.6%) 77/766 (10.1%) 43/568 (7.6%) 1/26 (3.8%) 14/330 (4.2%) 5/125 (4.0%) 1/17 (5.9%) 47/276 (17.0%) 7/88 (8.0%) 10/178 (5.6%) 3/138 (2.2%) 1/18 (5.6%) MCA in live birth c 24/666 (3.6%) 29/662 (4.4%) 15/497 (3.0%) 3/25 (12.0%) 5/316 (1.6%) 7/133 (5.3%) 2/19 (10.5%) 10/295 (3.4%) 4/92 (4.3%) 9/188 (4.8%) 8/141 (5.7%) 0/14 Hemangioma requiring treatment c 4/666 (0.6%) 2/662 (0.3%) 1/497 (0.2%) 0/25 1/316 (0.3%) 2/133 (1.5%) 0/19 3/295 (1.0%) 0/92 0/188 2/141 (1.4%) 1/14 (7.1%) Chromosomal anomalies in live birth c 2/666 (0.3%) 1/662 (0.2%) 3/497 (0.6%) 0/25 0/316 0/133 0/19 2/295 (0.7%) 0/92 1/188 (0.5%) 0/141 0/14 Data are n/N (%) or mean (SD). ALZ, alemtuzumab; CLAD, cladribine; GLAT, glatiramer acetate; IFN-β, interferon-β-formulations; MCA, major congenital anomaly; n, number of neonates with outcome; N, number of neonates with available data; NA, not applicable; NTZ, natalizumab; SGA, small for gestational age; S1P, sphingosin-1-phosphate; SD, standard deviation; TFL, teriflunomide; unexpo, unexposed. a Missings: triplets (n = 3), missing sex (n = 16), missing birth weight (n = 20). b Preterm births excluded. c Reported up to 1 year postpartum, N includes neonates with 12 months follow up.

Table 3 Neonatal Outcomes descriptive analysis. Unexpo-group (n = 802) reference category IFN-β-group (n = 766) GLAT-group (n = 568) TFL-group (n = 34) Fumarates-group (n = 366) S1P-group (n = 147) ALZ-group (n = 21) 1st-trimester NTZ-subgroup (n = 314) 2nd-trimester NTZ-subgroup (n = 102) 3rd-trimester NTZ-subgroup (n = 200) anti-CD20-group (n = 158) CLAD-group (n = 20) Neonatal death 3/802 (0.4%) 4/766 (0.5%) 1/568 (0.2%) 0/34 1/366 (0.3%) 0/147 2/21 (9.5%) 2/314 (0.6%) 1/102 (1.0%) 0/200 0/158 0/20  Thereof preterm births 2/3 (66.7%) 4/4 (100%) 1/1 (100%) NA 1/1 (100%) NA 2/2 (100%) 2/2 (100%) 1/1 (100%) NA NA NA SGA a 140/795 (17.6%) 124/753 (16.5%) 100/561 (17.8%) 4/34 (11.8%) 73/362 (20.2%) 40/146 (27.4%) 1/20 (5.0%) 60/312 (19.2%) 20/100 (20.0%) 49/198 (24.7%) 38/158 (24.1%) 2/20 (10.0%) Birthweight b , in g 3414.2 (443.6) 3415.0 (449.7) 3400.6 (429.1) 3391.1 (332.9) 3392.8 (447.7) 3277.2 (450.2) 3432.9 (381.9) 3355.9 (436.8) 3304.8 (443.9) 3310.0 (427.2) 3423.0 (449.3) 3485.0 (418.0)  Missing 0/716 3/680 (0.4%) 2/513 (0.4%) 0/26 2/330 (0.6%) 1/125 (0.7%) 0/17 0/276 0/88 0/178 0/138 0/18 Birth length b , in cm 51.7 (2.5) 51.6 (2.5) 51.7 (2.3) 51.3 (2.0) 51.5 (2.4) 50.7 (2.4) 51.8 (3.0) 51.4 (2.5) 51.1 (2.5) 51.0 (2.6) 51.8 (2.5) 52.2 (2.6)  Missing 0/716 4/680 (0.6%) 2/513 (0.4%) 0/26 3/330 (0.9%) 1/125 (0.8%) 0/17 2/276 (0.7%) 0/88 0/178 1/138 (0.7%) 0/18 Head circumference b , in cm 35.1 (1.4) 35.0 (1.5) 35.1 (1.4) 35.0 (1.4) 34.9 (1.4) 35.0 (1.5) 35.2 (1.2) 35.2 (1.4) 34.6 (1.6) 35.0 (1.4) 35.0 (1.4) 35.1 (1.4)  Missing 83/716 (11.6%) 77/766 (10.1%) 43/568 (7.6%) 1/26 (3.8%) 14/330 (4.2%) 5/125 (4.0%) 1/17 (5.9%) 47/276 (17.0%) 7/88 (8.0%) 10/178 (5.6%) 3/138 (2.2%) 1/18 (5.6%) MCA in live birth c 24/666 (3.6%) 29/662 (4.4%) 15/497 (3.0%) 3/25 (12.0%) 5/316 (1.6%) 7/133 (5.3%) 2/19 (10.5%) 10/295 (3.4%) 4/92 (4.3%) 9/188 (4.8%) 8/141 (5.7%) 0/14 Hemangioma requiring treatment c 4/666 (0.6%) 2/662 (0.3%) 1/497 (0.2%) 0/25 1/316 (0.3%) 2/133 (1.5%) 0/19 3/295 (1.0%) 0/92 0/188 2/141 (1.4%) 1/14 (7.1%) Chromosomal anomalies in live birth c 2/666 (0.3%) 1/662 (0.2%) 3/497 (0.6%) 0/25 0/316 0/133 0/19 2/295 (0.7%) 0/92 1/188 (0.5%) 0/141 0/14 Data are n/N (%) or mean (SD). ALZ, alemtuzumab; CLAD, cladribine; GLAT, glatiramer acetate; IFN-β, interferon-β-formulations; MCA, major congenital anomaly; n, number of neonates with outcome; N, number of neonates with available data; NA, not applicable; NTZ, natalizumab; SGA, small for gestational age; S1P, sphingosin-1-phosphate; SD, standard deviation; TFL, teriflunomide; unexpo, unexposed. a Missings: triplets (n = 3), missing sex (n = 16), missing birth weight (n = 20). b Preterm births excluded. c Reported up to 1 year postpartum, N includes neonates with 12 months follow up.

Neonatal Outcomes descriptive analysis.

Data are n/N (%) or mean (SD).

ALZ, alemtuzumab; CLAD, cladribine; GLAT, glatiramer acetate; IFN-β, interferon-β-formulations; MCA, major congenital anomaly; n, number of neonates with outcome; N, number of neonates with available data; NA, not applicable; NTZ, natalizumab; SGA, small for gestational age; S1P, sphingosin-1-phosphate; SD, standard deviation; TFL, teriflunomide; unexpo, unexposed.

Missings: triplets (n = 3), missing sex (n = 16), missing birth weight (n = 20).

Preterm births excluded.

Reported up to 1 year postpartum, N includes neonates with 12 months follow up.

Overall, among 3048 neonates with 12-month follow-up, 116 (3.8%) had MCAs, without a specific pattern ( Supplemental Table d ) or a significant difference between groups neither in the descriptive nor in the regression analysis ( Table 3 , Fig. 2 , Supplemental Table b ). MCAs were numerically higher in the TFL- (3/25: 12.0%, CI: 2.5%, 31.2%) and ALZ-group (2/19: 10.5%, CI: 1.3%, 33.1%).

A higher risk of SGA neonates was associated with exposure to S1P (adjusted OR: 1.65, CI: 1.07, 2.50, p-value: 0.02), and anti-CD20 (adjusted OR: 1.54, CI 1.01, 2.32, p-value: 0.04) compared to the unexpo-group ( Supplemental Table b ).

Additionally, neonates in the S1P-group had significantly lower birthweights (adjusted β: −132 g, CI: −205 g, −60 g, p-value: <0.001) and were smaller (adjusted β: −0.91 cm, CI: −1.3, −0.49, p-value: <0.001) compared to the unexpo-group, as in the 3rd-trimester NTZ-subgroup (birthweight adjusted β: −74 g, CI: −138, −9.4, p-value: 0.02) ( Fig. 2 , Supplemental Table b ).

Birth weight and SGA comparison with the general German population

The number of SGA children in the entire cohort (651/3459: 18.8%) were significantly higher compared to general reference values in Germany (10.0%, p-value: <0.001), 14 as were SGAs in the unexpo-group (140/795: 17.6%), moderately-effective DMT-group (301/1710: 17.6%), and highly-effective DMT-group (210/954: 22.0%) (p-value: <0.001 in each group).

We also observed significant differences in birthweights of our entire cohort at GW 40 (female: 3381.2 g, SD: 402.4, p-value: <0.001, male: 3509.5 g, SD: 395.6, p-value: <0.001) and within the unexpo-group only for female neonates (3380.4 g, SD: 397.7, p-value: 0.02) compared to the German population (female: 3473 g, SD: 422; male: 3624 g, SD: 441). 16

Supplemental Figure a visualizes the median birthweight of males and females in GW 38 to 41, which lies between the 25th and 50th percentile and IQRs mostly lying between the 10th and 75th percentile of the German population.

Discussion

In this large, prospective pregnancy cohort study in women with MS, DMT pregnancy exposure did not adversely affect most pregnancy outcomes (EP, SAs, medically indicated ETs, and SBs). A novel finding is that growth restriction was associated with second line DMTs, especially S1P-modulators, anti-CD20 antibodies, and NTZ during third trimester, but also MS itself. Our study is one of the first evaluating the maternal risk for severe infections which ranged between 0% (TFL) and 9.1% (ALZ) and were significantly higher in the descriptive analysis for Fumarates and Alemtuzumab exposed pregnancies. More women treated with NTZ during the 2nd or 3rd trimester or with anti-CD20 antibodies received systemic antibiotics during pregnancy. MCAs ranged between 0% (CLAD) and 12% (TFL) with the caveat that for some contraindicated DMTs only few exposed pregnancies could be analyzed.

The MCA prevalence in our cohort (3.8%) is slightly above the background risk of 2.8% reported for the general German population by EUROCAT 17 and very similar to an alternative study in a MS cohort (3.7%). 3 Comparing MCAs between populations using different methods is challenging, as demonstrated by the variance in MCAs reported by EUROCAT from two different German Sites in Saxony-Anhalt (2.6%) 17 and Mainz (4.1%). 17 In Saxony-Anhalt, MCAs were reported from multiple hospitals including data up to 1 year postpartum, whereas in Mainz, trained pediatricians examined each neonate within the first week of life. 17 In our registry, patient-reported outcomes, verified by the treating pediatrician, were collected with three follow-ups during the postpartum year, potentially introducing surveillance bias.

Although we found a higher prevalence in the TFL- (12.0%) and ALZ-group (10.5%), teratogenicity was not reported in pregnancies exposed to TFL (1%) 6 or ALZ (3%) 8 before, underlining the need for at least 300 exposed pregnancies as required by EMA, to draw firm conclusions. 2 In contrast Fingolimod (FTY) has been associated with a 2-fold elevated teratogenic risk, 5 with 5.3% MCAs in our cohort (statistically not significant). We predominantly observed congenital heart defects, especially atrial and ventricular septal defects, and renal malformations mirroring the EMA's findings. 5 The discrepancy in MCA prevalences with 4.0% (n = 75) 18 in the Gilenya Pregnancy Exposure registry and 2.7% (n = 219) 18 in the pregnancy outcomes intensive monitoring by Novartis, highlight the challenges in comparing different post-approval safety collection methods. Therefore, the ConcePTION project aims to create an ecosystem for consistent monitoring and communication of medicine safety during pregnancy and breastfeeding, and to compare strategies for pregnancy outcome collection. 19 Our data cannot confirm whether S1P-modulators are teratogenic. However, as small molecules that cross the placental barrier and block S1P-receptor signaling—critical for vascular formation during embryogenesis 20 -teratogenicity, as observed in animal studies, is biologically plausible.

In contrast monoclonal antibodies are large molecules and their active transplacental transport begins in the second trimester, after the sensitive phase of organogenesis. 21 Therefore, teratogenicity assuming exposure shortly before or early in pregnancy is biologically implausible. In our cohort, MCAs occurred in 5.7% of anti-CD20 exposed pregnancies, not statistically significantly higher than internal controls. Other studies did not report on an increased MCA risk (0% n = 38, 22 4.8% n = 125 23 ) after anti-CD20 pregnancy exposure. The conservative European label advising anti-CD20 antibody withdrawal at least 12 months prior to pregnancy, might lead to increased scrutiny and thorough antenatal examinations after pregnancy exposure. It remains unclear if surveillance bias or registry methodology account for this non-statistically significant increase in our cohort.

Other adverse pregnancy outcomes (EPs, SAs, SBs, and PBs) did not demonstrate a higher prevalence, which is in line with findings from other studies and meta-analysis including IFN-β, GLAT, Fumarates, S1P-modulators, anti-CD20 antibodies, and NTZ exposed pregnancies. 1 , 3 , 4 , 11

The relatively low incidence of SAs observed across the entire cohort (8.2%) compared to the general population (12–15%) 24 and prospective or population-based MS studies (10.9% 3 –27.3%) 22 reflects the difficulties in accurately determining the frequency of SAs. In our registry, abortions occur at an early gestational stage (9.6, IQR: 7.7–11.1), contrasting with the GW at enrollment (10.3, IQR: 6.4–18.7). Consequently, sensitivity analysis of pregnancies enrolled in the registry prior to the 5th GW show a higher and probably more realistic frequency of SAs (20.2%). This emphasizes that SAs in other cohorts may be underestimated and can only be compared with cohorts capturing pregnancies very early.

Our study includes comprehensive data on growth parameters. Low birthweight and SGA are important risk factors for both fetal and neonatal death, and many diseases later in life, including type 2 diabetes mellitus or cardiovascular diseases. 25 When comparing the entire cohort with the German population, more neonates were smaller than expected, with a birthweight difference of 92 g (female) and 115 g (male) in the 40th GW, respectively. 17.6% of neonates unexposed to any DMT were SGA, more in the highly-effective DMT-group (22.0%), compared to 10% in Germany. 14 Our data supports the long-lasting debate on fetal growth retardation attributed to MS itself, 26 , 27 as previously suggested in both exposed and DMT-unexposed MS-patients (25.3%) 26 and lower birthweights around 110 g compared to healthy control groups (10%, 26 3517.4 g 27 ). When comparing only women with high-dose glucocorticosteroids during pregnancy, indicating an active disease course, there was no significant difference in SGAs between the unexposed- (25.3%), the moderately-effective (25.5%), and the highly-effective DMT-group (27.6%). Therefore, it is possible, that not only specific DMTs and MS itself, but also disease activity may play a crucial role regarding fetal growth retardation. Mechanisms that could lead to fetal growth retardation due to MS itself may include neuronal dysfunction affecting pelvic organs, resulting in suboptimal intrauterine conditions 28 and urinary tract infections, other subclinical infections, 29 maternal vitamin D deficiencies, 30 , 31 or other comorbidities not accounted for in this analysis. In addition, steroid use during pregnancy due to increased disease activity in some patients may also contribute to increased rates of SGA. 32

Beside MS itself, we observed additional effects on weight reduction caused by different DMTs with the strongest influence being associated with S1P-exposure (SGAs: 27.4%, birth weight (β): −132 g). Preclinical data for FTY indicated growth retardation in animals 20 but human data was lacking. An effect on fetal development is plausible not only due to placental transit, but also from a pharmacodynamic point of view, as S1P is a signaling molecule that plays a crucial role in various cellular processes and influences placental development, function, and blood flow, thus could affect nutrient and oxygen delivery to the fetus. 20 Anti-CD20 pregnancy exposure also seemed to increase SGA neonates (24.1%), although the mean birthweight (3423.0 g, SD 449.3) is in line with the German population (mean 3456 g) 14 and a small cohort study (n = 14, 3364.5 g), 33 suggesting that our sample size may be insufficient to draw final conclusions. Analysis with stratification of the anti-CD20 group into exposure duration and timing might reveal if this effect is observed throughout the whole group or mainly in neonates exposed later in pregnancy.

The human fetus experiences the most significant growth in absolute terms during the middle and late stages of gestation. 34 Therefore, it seems plausible that we only observed a birthweight reduction of 74 g in the 3rd-trimester NTZ-subgroup compared to the control group. We also observed a tendency towards reduced birthweights following prolonged exposure to NTZ during pregnancy, in our likely underpowered recent study with 200 pregnancies less. 35 Outcomes of one of the largest recently published cohorts on NTZ exposure reported by Portaccio et al. 36 revealed a significant birthweight reduction of 169 g for NTZ first trimester exposure, but included higher numbers of smokers (14.5% vs. 7.6% in our 1st-trimester NTZ-subgroup) and a higher rate of high-dose glucocorticosteroids during pregnancy (36.5% vs. 16.2% in our 1st-trimester NTZ-subgroup). 36 Return of disease activity is a recognized risk after withdrawal of NTZ 37 and FTY 38 and might lead to high-dose glucocorticosteroid treatments during pregnancy. Glucocorticosteroids are known for their increased PB risk 39 and birthweight reduction (12–332 g). 32 In our cohort we observed higher numbers of PBs (14.6%), SGAs (26.2%) and birth weight reduction (β: 87 g) linked to high-dose glucocortiocsteroids during pregnancy.

Severe infections are a relatively novel but important outcome, as pregnancy infections may cause complications including SAs, SBs, PBs and lower birthweights. 29 They occurred significantly more often following exposure to fumarates (11/395: 2.8% vs. 8/837 unexposed-group: 1.0%, p-value: 0.03), and ALZ (2/22: 9.1% vs. 8/837: 1.0% unexpo-group, p-value: 0.02). The observed increase in infections among the fumarates- and ALZ-group remains difficult to conclusively explain due to the rarity of severe infections in our cohort (54/3722: 1.5%). However, it is plausible that drug effects such as the reduction of circulating B and T lymphocytes (ALZ) and drug side effects such as lymphopenia (fumarates) could play a role in increasing severe infections as lymphopenia was recently connected with cases of progressive multifocal leukoencephalopathy (PML) and other opportunistic infections like Varicella-Zoster-Virus, or Herpes simplex virus. 40

Exposure to 2nd- (OR: 2.47, CI: 1.47, 4.05, p-value: <0.001) and 3rd-trimester NTZ (OR: 1.75, CI: 1.15, 2.63, p-value: 0.01), or anti-CD20-antibodies (OR: 2.16, CI: 1.41, 3.29, p-value: <0.001) was connected to a higher risk of antibiotic use during pregnancy. This is in line with a Swedish cohort study—published as an abstract so far–reporting a higher risk for any infections especially for RTX (24%) but also for NTZ exposed pregnancies (9.7%) compared to unexposed pregnancies (7.7%). 23 Potential mechanisms may include lower IgG levels (for anti-CD20 antibodies) or impaired immune surveillance (for NTZ). 40 , 41 The longer exposure to NTZ during the third trimester may have contributed to a higher risk of infections with antibiotic treatments, especially as the immune system shifts to a more immune-tolerant state during the second and third trimester, increasing susceptibility to infections. 40 The increased rates of antibiotic exposure in the 2nd-trimester NTZ subgroup may have also been influenced by the increased steroid use during therapy breaks and the resulting higher frequency of hospital visits, which may have led to over-prescribing in some cases due to heightened awareness or more frequent monitoring. Further research is warranted to confirm the potential impact on antibiotic therapy and severe infections, taking into account that the administration of newer therapeutic agents, such as anti-CD20 antibodies, occurred primarily during the COVID-19 pandemic, which may have led to the confounding effect of generally lower hospitalizations during this period. 42 It will be crucial to collect more long-term data on these children after birth on growth parameters, child development and infection risks.

The strength of the study is the large cohort size, which allowed stratification of different DMT groups, and close prospective follow up, which enhanced the analytical robustness and generalizability. Detailed examination of pregnancy outcomes, adjustment for various confounders, comparisons with unexposed, and assessment of DMT-associated risks, contribute to scientific rigor. However, an insufficient sample size for regression analysis in groups exposed to newer or contraindicated substances (TFL, ALZ, CLAD) and in outcomes with rare cases (e.g. severe infections) limits the analysis to descriptive approaches and requires cautious interpretation.

Voluntary participation, particularly of individuals receiving potentially problematic medications or with high disease activity, may introduce selection or surveillance bias, affecting the representativeness of the sample compared with population-based studies. Despite efforts to control for variables, unaddressed confounders may influence observed associations.

Conclusion

Overall, the therapeutic options for MS patients have improved tremendously in recent years and with this study we have made an important contribution to assessing the safety of DMTs during pregnancy. Our study provides confirmative findings, that IFN-β, GLAT seem to be safe treatment options during early pregnancy. 1 , 3 Fumarates also seem to be safe treatment options during early pregnancy 1 , 4 despite the unexpectedly increased number of severe infections compared to unexposed pregnancies, which are numerically small, but should be further investigated. S1P-modulators should be avoided due to the increased SGA-, potentially increased MCA- and relapse associated disability risk. 5 , 38

NTZ, if maintained during pregnancy due to its high rebound risk, 37 has emerged as a promising therapy option for women with highly active disease courses seeking pregnancy, as have anti-CD20 antibodies since early exposure does not seem to significantly impact B-cell development in neonates. 43 Further data are needed from extended follow-up studies, especially after long-term exposure, on the risk of maternal/infant infections, fetal growth retardation and timing of weight normalization. To improve pregnancy recommendations, further research should determine whether disease activity or DMT pregnancy exposure impacts fetal growths retardation.

For smaller groups such as TFL, ALZ, and CLAD and rare outcomes such as severe infections, more data are needed in the near future and data from different registries should be merged. However, because of its cyclic dosing regimen and long-lasting effect, CLAD may be a reasonable treatment option if pregnancy is planned at the earliest after a 6-month washout period, with effective disease control during pregnancy and postpartum in a small cohort. 44

Taken together, our findings are critical to improve future pregnancy recommendations in women with MS seeking pregnancy by combining established and emerging evidence and increasing the number of reported pregnancies.

Contributors

S.T. and K.H. designed the study. N.B., K.D.K., and S.T. checked the plausibility of the data. N.B., S.T., and S.H. analyzed the data. N.B., S.T., K.H., N.F., L.W., and R.G. interpreted the data. N.B. and S.T. wrote the manuscript and figures with input from all authors. All authors critically reviewed the manuscript, provided critical feedback, and approved the final version. No medical writer or editor was involved in the creation of the manuscript.

Data sharing statement

Upon reasonable request and in compliance with data protection policies, anonymized data supporting the findings of this study will be made available.

Declaration of interests

The DMSKW is partly supported by Almirall, Biogen, Hexal, Merck, Novartis, Roche, Sanofi Genzyme, Teva Pharma and Viatris. NB received payment for manuscript writing from Thieme Verlag and had a paid internship at Merck in 2021. KDK, SH and TO have nothing to disclose. NF received speakers honoraria and a sponsorship for congress participation from Biogen GmbH. LW received travel grants from Novartis. RG has received speaker honoraria and research support from Bayer-Schering Healthcare, Biogen-Idec Germany, Chugai, Eisai, Merck Serono, Nikkiso Pharma, Novartis, Roche, Sanofi-Genzyme and TEVA, has received consulting honoraria from CSL Behring, Baxter, Janssen and Talecris and has stock options in Bayer, Merck and Roche. S.T. received speakers’ honoraria from Bayer Healthcare and Biogen GmbH, payment for manuscript writing from HEXAL AG as well as sponsorship for congress participation from Biogen GmbH. K.H. has received speaker honoraria and research support from Bayer, Biogen, Merck, Novartis, Sanofi-Genzyme, Roche and Teva, has received support for congress participation from Bayer, Biogen, Merck, Roche, Sanofi Genzyme and Teva, and has served on scientific advisory boards for Bayer, Biogen, Sanofi, Teva, Roche, Novartis and Merck.