How safe are anti-rheumatic drugs during pregnancy?

Highlights

• The risk for congenital malformations from low-dose weekly methotrexate is low.

• Mycophenolate mofetil increases both miscarriage and birth defects significantly.

• Leflunomide does not appear to be a human teratogen when a washout is performed.

• Azathioprine is not gonadotoxic in men.

• Certolizumab does not increase adverse pregnancy outcomes.

Rheumatic diseases may be active during pregnancy necessitating drug treatment in order to control maternal disease activity and ensure a successful pregnancy outcome. The present literature survey of the last 2 years does not profoundly change the recommendations given in recent reviews: the teratogenic drugs cyclophosphamide, methotrexate, mycophenolate mofetil, and biologics without or with few pregnancy data must be withdrawn before a planned pregnancy. Leflunomide has up to date not shown to be a human teratogen. Drugs that can be used throughout pregnancy include corticosteroids, sulfasalazine, antimalarials, cyclosporine, tacrolimus and azathioprine. Among biological drugs extended pregnancy experience exists only for TNF-inhibitors. The effect of immunosuppressive drugs and biologics on male reproductive function is only partly known.

Introduction

Planning pregnancy in women with rheumatic diseases requires adjustment of therapy to drugs, which keep the underlying maternal disease quiescent and are compatible with embryonic and fetal development. Withdrawal of all drugs administered before conception resulting in a disease flare may be equally disastrous for pregnancy outcome as continuation of agents that are able to harm the developing child.

Information on usage of drugs during pregnancy is limited, since pregnant women are never included in drug trials and knowledge usually accumulates over the years from inadvertent drug exposure during unplanned pregnancies. Regular updates of information are therefore timely. This survey is based on a literature search on immunosuppressive, antirheumatic drugs used in pregnancy published during 2010–2012. Several excellent reviews have given a complete overview of published experience of antirheumatic drugs in pregnancy, and these should be consulted for details [1•, 2, 3].

Risks of pregnancy in the mother with rheumatic disease depend on disease activity shortly before conception and throughout pregnancy, extent of organ involvement and presence of certain types of autoantibodies (Table 1). Risks for the fetus are related to maternal disease activity, presence of autoantibodies and maternal therapy (Table 1). As shown in Table 2, drug therapy of rheumatic diseases differs and will as a rule be tailored according to type of organ involvement, severity of the disease and individual response to treatment. Adjustment of therapy in a patient planning a pregnancy or consulting during early pregnancy will aim at optimal disease control in the mother and healthy development of the child (Figure 1).

Methotrexate (MTX) and cyclophosphamide (CYC) inhibit cell division and are teratogenic in animals and humans. No new reports on CYC in autoimmune disease pregnancies have been located. A teratogen update has analyzed all the published cases with MTX exposures to doses between 7.5 mg/week for rheumatic diseases and up to 150 mg/week for induction of abortion for the proportion of all malformations [4^••] and compared with the same proportion from the Metropolitan Atlanta Congenital Defects Program [5]. The data have confirmed that the sensitive period for the production of malformations is 6–8 weeks after conception for doses greater than 10 mg of MTX/week. Higher doses increase the risk for MTX embryopathy including cardiac defects, pulmonary atresia, craniosynostosis, and limb deficiencies [4^••].

Two new studies have addressed pregnancy outcomes in men treated with MTX. A prospective study of men treated with weekly doses 7.5 mg and 30 mg of MTX three months before or until conception found no child with birth defects in 42 pregnancies [6]. A population based study of 50 pregnancies with fathers exposed to MTX has reported two children with orofacial malformations [7].

Mycophenolate (MMF) is used successfully in the treatment of lupus nephritis. First trimester exposure increases the risk for multiple congenital malformations with a special phenotype including microtia, auditory canal atresia, cleft lip and palate, micrognathia, hypertelorism, ocular coloboma, short fingers and hypoplastic nails [8]. A new prospective study has investigated 57 pregnancies exposed to MMF in the first trimester [9]. 45% of the pregnancies ended in spontaneous abortion. Six out of 29 live born children showed congenital malformations, which in four infants had a clinical phenotype consistent with mycophenolate embryopathy [9]. The malformation rate for all the pregnancies after MMF exposure was 26%. Patients who start treatment must be informed that discontinuation of MMF is necessary at least 6 weeks before trying conception, and be advised to use effective birth control during therapy.

Leflunomide, an inhibitor of pyrimidine synthesis was shown to be teratogenic in rats, rabbits and mice, and therefore labeled as contraindicated in human pregnancy [10]. A few case reports during the years 2000–2010 were not conclusive until the Organization of Teratology Information Specialists (OTIS) published the results of a prospective study of birth outcomes in women exposed before or during pregnancy to leflunomide [11^••]. Comparing pregnancy outcomes of 64 patients with RA exposed during the first trimester to 108 RA pregnancies not exposed neither showed increase in birth defects nor any recognizable malformation pattern [11^••]. This result was confirmed in additional 45 pregnancies of patients who were exposed to leflunomide either within two years before conception or during the first trimester [12]. These data indicate that leflunomide is not a strong human teratogen though due to the limited number of exposed pregnancies contraception is recommended during use and a washout procedure in case of a planned or unintended pregnancy.

A controversy exists as to an association of first trimester exposure to corticosteroids and a significant increase of oral clefts [13]. A new population based study with 51 973 corticosteroid exposed pregnancies did not find an increase of orofacial clefts regardless of type of corticosteroid administered (oral, inhalation, nasal spray, topical) [14^••]. After completion of placenta development, all corticosteroids are partly inactivated by 11-beta-hydroxylases of the placenta with relative protection of the fetus. Prolonged use of corticosteroids increases the risk of preterm delivery, intrauterine growth restriction, and the frequency of maternal side effects like hypertension, diabetes mellitus and osteoporosis. However, severe flares of rheumatic disease during pregnancy most often need doses of corticosteroids exceeding 20 mg/day for rapid control of symptoms.

Hydroxychloroquine (HCQ) exerts a number of positive effects that are particularly beneficial for SLE patients [15]. These include prevention of lupus flares, lowering of lipid levels, and moderate protections against thrombosis [16], and reduction of development of congenital heart block (CHB) in children of mothers positive for anti-Ro/anti-La antibodies [17]. Even in anti-Ro/anti-La positive mothers with a previous child with CHB the recurrence rate of cardiac-NL was 64.6% lower in pregnancies exposed to HCQ compared with those unexposed [18]. The new data indicate that HCQ should be continued in SLE patients and other patients positive for anti-Ro/anti-La antibodies throughout pregnancy.

A registry based Norwegian study has identified 589 pregnancies with exposure to antirheumatic drugs during the first trimester, thereof 74 exposed to sulfasalazine. No increase in congenital malformations was found in exposed children [7], confirming previous studies.

Cyclosporine (Cs) and tacrolimus are calcineurin inhibitors and inhibit the activation of T and B cells. Two forms of cyclosporine with different bioavailabilities are available: CsA and modified Cs (Neoral) the latter with much improved bioavailability. Most pregnancy experience is present for CsA and stems from registries of pregnant transplant recipients whereas experience from pregnant patients with autoimmune disease is limited to case reports and case series. Thus far, most of the studies of pregnancies exposed to CsA or Neoral have found an increase in premature delivery and low birth weight, but no increase of congenital malformations nor any particular malformation pattern [19, 20]. Whether prematurity and low birth weight actually are a side effect of Cs or related to the underlying disease and comorbidities or multidrug therapy is not clarified. Pregnancy experience for tacrolimus is nearly exclusively related to transplant recipients, but a few case reports on patients with autoimmune diseases exist [20, 21]. As with Cs no increase in the malformation rate or a malformation pattern has emerged. In patients with autoimmune disease refractory to other immunosuppressive treatment Cs or tacrolimus may be administered during pregnancy.

In a systematic review and meta-analysis the effect of thiopurines (6-mercaptopurine and azathioprin) on birth outcome from female and male inflammatory bowel disease (IBD) patients were analyzed [22^••]. The pooled analysis of five studies showed no significant association between thiopurines and congenital defects (OR 1.45; 95% CI 0.99, 2.13) or low birth weight (OR 1.01; 95% CI 0.96, 1.06). However, the analysis disclosed a significant increase for the association between preterm birth and women taking thiopurines (OR 2.28; 95% CI 0.67, 7.73), yet it remains unclear whether this was secondary to the drug or to active IBD [22^••]. The findings confirm results in transplant recipients treated with AZA and support its use during pregnancy when indicated.

The pooled analysis of three studies showed no significant risk for congenital anomalies after paternal exposure at the time of conception to thiopurines [22^••]. Another study evaluated the outcome of 46 pregnancies when fathers were exposed to thiopurines [23]. No significant difference in regard to unsuccessful pregnancies or congenital malformations was found compared with pregnancies fathered by patients who had never been treated with thiopurines or who stopped thiopurines more than 3 months before conception.