Rituximab before and during pregnancy: A systematic review, and a case series in MS and NMOSD.

OBJECTIVE: To evaluate the safety of rituximab treatment before and during pregnancy in women with MS and neuromyelitis optica spectrum disorders (NMOSDs) who may be at risk of relapses by performing a systematic literature review combined with a retrospective single-center case series.
METHODS: Studies were systematically identified in the PubMed, Google Scholar, and EMBASE using the key terms "pregnancy" and "rituximab"; 22 articles were included for review (>17,000 screened). Then, patients with MS and NMOSD from 1 center (University of California, San Francisco) exposed to rituximab before conception were identified through medical record review.
RESULTS: Systematic review: We identified 102 pregnancies with rituximab use within 6 months of conception: 78 resulted in live births and 12 in spontaneous abortions. Of 54 live births with reported gestational age, 31 occurred at term (37 weeks+) and 2 before 32 weeks. When checked, B-cell counts were low in 39% of newborns and normalized within 6 months. Case series: we identified 11 pregnancies (1 ongoing) in 10 women (7 MS and 3 NMOSD) treated with rituximab within 6 months of conception. All completed pregnancies resulted in term live births of healthy newborns (1 lost to follow-up at term). No maternal relapses occurred before/during pregnancy; 1 occurred postpartum (NMOSD).
CONCLUSION: No major safety signal was observed with rituximab use within 6 months of conception. Beyond the need for monitoring neonatal B cells, these observations support prospectively monitoring a larger patient cohort to determine whether rituximab may safely protect women with MS and NMOSD who are planning a pregnancy against relapses.

Women are disproportionately affected by MS and neuromyelitis optica spectrum disorders (NMOSDs), and management of disease-modifying therapies (DMTs) before pregnancy represents an ongoing challenge for neurologists. No safety concerns have been identified with platform injectable DMTs,[1-3] but discontinuation before pregnancy is typically recommended for the more potent oral and infusible DMTs. Therefore, many women may face a heightened risk of relapses during the period between DMT discontinuation and the potentially protective (in MS[4,5] but not in NMOSD[6,7]) effects of pregnancy. This risk could be further magnified by recurrence of severe “rebound” MS disease activity after discontinuing natalizumab[8-12] or fingolimod,[13-16] and in fact these two DMTs appear associated with a higher risk of relapse during pregnancy in the new treatment era.

Rituximab, frequently used off-label for the treatment of MS and NMOSD,[17,18] may offer distinct advantages for managing women at the time of conception. First, its biological effects (B-cell depletion) persist long after the drug is effectively eliminated (typically, 5 maximal half-lives[3] each of 19–22 days or approximately 110 days[19]). These data suggest that women could attempt conception approximately 3.5 months after their last infusion without significant risk of fetal exposure to the monoclonal antibody, while conferring protection against MS flares throughout the pregnancy. In addition, should a woman unintentionally conceive before rituximab elimination, the risk of fetal exposure is low, as IgG1 subclass antibodies are not transferred across the placenta during the first trimester. Finally, transition to rituximab from natalizumab may confer protection against the risk of a rebound of disease activity associated with natalizumab withdrawal.[20]

To date, pregnancy and neonatal outcomes in women with MS and NMOSD treated with rituximab are largely unreported.[3,20] To bridge this gap, we performed a systematic review of the medical literature, combined with a retrospective single-center case series.

Methods

Systematic review

To summarize and analyze the existing literature on pregnancy outcomes in women treated with rituximab for any indication within 6 months of conception through delivery, we performed a systematic review.

Data sources

Original research studies were identified from the PubMed/MEDLINE, EMBASE, and Google Scholar databases. The search terms “pregnancy” and “rituximab” were used in combination to include all articles with the key words for all years (last updated July 3, 2017). Further hand searching of reference lists of obtained articles was performed.

Study selection

This search yielded over 17,000 results; titles and abstracts were screened for relevance, and relevant manuscripts underwent subsequent review. Studies were excluded if they were not written in English, were reviews with no specific individual- or cohort-level data, or if mothers were exposed to rituximab more than 6 months before conception (list of citations available upon request). Twenty-two publications were included in the current review, with 102 mothers exposed to rituximab in the desired timeline (see Preferred Reporting Items for Systematic Reviews and Meta-Analyses flow diagram,[21] figure).

Figure

Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines for systematic review

Data extraction and analysis

Data were extracted (G.D.) and checked (R.B.) for individual-level information relating to maternal and fetal outcomes. Most articles were case reports, a few were retrospective, and 1 was a meta-analysis of a database; none included control groups.

Retrospective single-center case series

Sample selection

To identify a cohort of women with MS or NMOSD treated with rituximab within 6 months of conception or during delivery at the University of California, San Francisco (UCSF) Multiple Sclerosis and Neuroinflammation Center, we performed search of relevant medical records. Among 323 patients with CNS inflammatory disorders who treated with rituximab between August 2010 and July 2017, we identified 160 women who received rituximab infusions before the age of 50 years. Their medical records were manually reviewed to identify pregnancies occurring within 6 months of exposure to rituximab. We identified 10 women with at least 1 pregnancy occurring within the selected timeframe. These were cross-referenced with participating clinicians' individual caseloads.

Data collection

Medical records were reviewed to record pregnancy outcomes (delivery or fetal loss), maternal MS or NMOSD disease activity during pregnancy and postpartum (relapses, symptoms, and medications), maternal pregnancy-related and other complications (e.g., preeclampsia), and neonatal outcomes (gestational age, delivery mode, and newborn health status).

Standard protocol approvals, registrations, and patient consents

The UCSF Committee of Human Research approved the study protocol for retrospective analysis of electronic medical record–derived MS data with no patient contact (Ref #13-11686).

Results

We identified 102 women who became pregnant within 6 months of exposure to rituximab or who were treated with rituximab while pregnant. Of these 102 patients, 38 were described in case reports or small case series, and 64 were described in a meta-analysis[22] (table 1). Medical indications for rituximab treatment included 2 cases with NMOSD, as well as lymphoma, rheumatoid arthritis, and immune-mediated thrombocytopenic purpura, among others. Many mothers experienced successful medical management for their primary condition with rituximab.

Table 1

Systematic review of 22 articles reporting maternal and fetal outcomes for 102 pregnancies characterized by maternal treatment with rituximab within 6 months of conception or during pregnancy

A total of 74 of the 102 pregnancies reported resulted in live births. In the 41 pregnancies from the case reports and case series, 1 fetal demise was reported at 21 weeks of gestation in a patient with a history of miscarriages[23] and 1 stillbirth was reported at 27 weeks of gestation due to placental insufficiency.[24] There was also 1 therapeutic abortion and 1 spontaneous abortion in a patient taking methotrexate until confirmation of pregnancy.[24] In the larger meta-analysis of the global drug safety database investigating fetal outcomes in patients with rituximab exposure (64 pregnancies with 1 ongoing at the time of publication), 11 spontaneous abortions were reported, as well as 15 medical abortions.[22] These mothers were often taking other medications, including methotrexate, an antifolate drug that is often used to treat ectopic pregnancies.

Newborns were delivered at term (37 weeks of gestational age or after) in 31 of the 54 live births, where gestational age was reported. None of the deliveries were severely preterm (less than 28 weeks of gestation, according to the World Health Organization guidelines), and 2 were born before 32 weeks. One mother on an rituximab, cyclophosphamide, doxorubicin, vincristine and prednisone regimen for lymphoma underwent emergency cesarean section at 34 weeks due to fetal distress. The baby was born with a patent ductus arteriosus that was fixed the day after birth.[25] Besides, the patent ductus arteriosus, the only other newborn physical abnormalities or medical conditions reported involved twins: one was born with a clubfoot and the other with erythema toxicum neonatorum. Both twins developed normally and did not have any complications.[26]

Neonatal B-cell depletion was found in 9 of the 23 pregnancies in which B-cell counts were measured and reported. None of these neonates experienced any infectious complications, and none were noted to experience any adverse reactions to vaccinations. All B-cell levels normalized within 6 months, some sooner.

We identified a total of 11 pregnancies from 10 patients who became pregnant within 6 months of rituximab exposure (table 2): 6 with MS and 3 with NMOSD, with pregnancy ongoing in 1 patient at the time of data collection. No patients received rituximab infusions while pregnant. No patients experienced relapses between rituximab treatment and conception.

Table 2

Pregnancy and neonatal characteristics in 11 pregnancies characterized by maternal exposure to rituximab for treatment of demyelinating diseases within 6 months of conception: a single-center case series

Medical complications arose in 4 of the 10 completed pregnancies and included pregnancy-related (gestational diabetes, pre-eclampsia with postpartum eclampsia) and disease-related (blurring of vision and lightheadedness at about 5 months of gestation and worsening spasticity; numbness in fingertips 2 weeks before delivery); neither of the disease-related complications was determined by the treating neurologist to represent a clinical relapse, and neither required additional medications or hospitalizations.

Of the 10 completed pregnancies, 9 resulted in term live births, with healthy children (status unknown for 1 as the mother moved out of the area and closer to familial supports for delivery). Most mothers breastfed for at least a few weeks before restarting treatment. One patient of the 8 with data for the full-year postpartum, who carried an NMOSD diagnosis, relapsed postpartum.

Discussion

Rituximab, which seems to reduce relapse frequency in both MS and NMOSD, may offer significant advantages for women with MS and NMOSD who are planning a pregnancy and require ongoing DMT, given that its biological effect extends significantly beyond its pharmacokinetic half-life. Although individual variability exists in terms of the exact half-life of rituximab,[27] or of any administered monoclonal antibody, an estimate of effective elimination would require 5 maximal half-lives[3] each of 19–22 days or approximately 110 days.[19] B-cell repletion occurs in most individuals within 8 months, but in the phase 2 study of rituximab, MS disease activity remained suppressed 1 year after a single course of treatment. However, not enough is known about pregnancy outcomes in women with MS and NMOSD who conceive after rituximab treatment to fully understand the risks and advantages of this approach.

In the current systematic review of pregnancy outcomes in women with a number of medical conditions treated with rituximab (some with severe diseases and using other concomitant medications), we calculated an overall reported rate of spontaneous abortions of 12%, and 41% of reported deliveries occurred before 37 weeks (2 before 32 weeks). Three malformations or medical conditions were reported among the 67 newborns (as a point of reference from the general population in the United States, the rate of major malformations at birth is 3%[28]). The primary adverse effect noted was a low neonatal B-cell count in 39% of the newborns evaluated that normalized within 6 months in all cases. Of interest, in 1 patient with NMOSD whose last treatment with 100 mg rituximab was 7 months before conception, a low B-cell percentage detected in fetal cord blood suggested that maternal rituximab treatment might influence fetal B-cell counts even after rituximab should have been completely eliminated from the maternal circulation.[21]

Limitations of this literature review include the retrospective nature and small size of the cases series and reports, potential under-reporting, and lack on information on potential confounders. For example, the coexisting use of glucocorticoids, antimetabolites, and other chemotherapy agents could potentially confound any causal role attributed to rituximab for these adverse events. Furthermore, there was little information on other possible confounding variables such as obstetrical or clinical disease histories, which could also influence the pregnancy course and outcome. Prospective evaluation of pregnancy outcomes, including case-control studies evaluating the rates of pregnancy loss or prematurity, as well as longer term information about children's immunologic trajectories, is required to overcome the current limitations and biases. Nonetheless, this review did not identify major concerns that would preclude treatment with rituximab within 6 months of conception in women with demyelinating diseases at risk of inflammatory activity when discontinuing other MS medications before conception.

Next, in our preliminary case series of 11 pregnancies in women with demyelinating diseases treated with rituximab within 6 months of conception, none of the patients experienced a relapse before conception or during pregnancy. Those with completed pregnancies did not experience major rebound activity after delivery; only 1 patient (NMOSD) experienced a postpartum relapse (up to 1/3 women with MS have been reported to relapse postpartum). Most of the treated patients breastfed for at least a few weeks before receiving another dose of rituximab. All children were reported to be healthy at birth and to remain healthy at follow-up.

Limitations from the current case series include the small sample size and possible bias if some pregnancies in women not receiving primary or obstetrical care in the same hospital system as our tertiary care center were not captured in the patient's medical record. However, given the complex medical decision making that typically happens at the time of conception and delivery in women with demyelinating diseases, it is likely that we identified most of the pregnancies within our rituximab-treated cohort. Furthermore, to date, regular B-cell monitoring in the mother during and after pregnancy and in the neonate has not been routinely performed in our clinic. As decreased neonatal B-cell counts were noted in the systematic review, this information would provide additional insights into the effect of rituximab intrapartum and postpartum and would allow optimization of the timing of rituximab treatment. Finally, there were no patients in our single-center case series who received rituximab during pregnancy, limiting discussion of possible effects of intrapartum maternal treatment and fetal exposure.

Currently, to treat women with demyelinating diseases during pregnancy, glatiramer acetate can be used before, and even during, pregnancy; but some therapies commonly used in MS are relatively contraindicated.[29] Fingolimod, dimethyl fumarate, and teriflunomide are small molecules that could cross the placenta and potentially cause birth defects. Natalizumab treatment during pregnancy may be associated with neonatal pancytopenia,[30] and natalizumab treatment discontinuation can be associated with recurrence of severe MS disease activity.[31] For example, in a retrospective study evaluating 22 pregnancies after discontinuation of natalizumab, recurrence of disease activity was noted in 95.5% of the cases, despite little to no activity in the year before natalizumab discontinuation.[32] Disease activity seemed more limited when conception occurred shortly after or even before discontinuation of natalizumab, with recurrence of disease activity often occurring 4–6 months after discontinuation. To our knowledge, no data on daclizumab use during pregnancy are available in humans. Like rituximab, alemtuzumab has a pharmacodynamic effect that is far longer than its pharmacokinetic half-life and theoretically could be used in women with MS who are planning pregnancy. However, alemtuzumab-treated patients are at high risk for treatment-related, de novo thyroid and other autoimmune diseases. Because maternal autoantibodies can be transmitted across the placental barrier and thereby can cause disease in the fetus or newborn, treatment with alemtuzumab before pregnancy must be approached with caution. In March 2017, ocrelizumab, a humanized monoclonal anti-CD20 antibody, was approved by the US Food and Drug Administration for both relapsing and progressive MS.[33] Its average half-life is 26 days—possibly shortening the preconception therapeutic window relative to rituximab. In an initial report of 9 women whose embryo/fetus was considered “exposed” to ocrelizumab (i.e., here, infusion within 3 months of conception) during the MS clinical trials, pregnancy outcomes included 1 healthy term baby, 6 elective terminations, and 2 ongoing pregnancies.[34] Therefore, an unmet need for effective treatments for women living with MS and NMOSD who are considering pregnancy remains.

While awaiting prospective pregnancy and postpartum monitoring, the current systematic review and case series provide some preliminary reassurance that rituximab may offer a window of time before conception during which inflammatory activity can be mitigated, without evidence of major adverse effects during pregnancy. Longer term follow-up and a larger sample size are needed to determine the safety of rituximab before and during pregnancy in women with MS and NMOSD, and independent studies will be required to assess the potential benefits and risks of other B-cell depleting agents such as ocrelizumab in this situation.