Aminosalicylates

All aminosalicylates (sulfasalazine, mesalamine, balsalazide) are pregnancy category B except olsalazine, which is pregnancy category C. Sulfasalazine is composed of 5-aminosalicylic acid azo-bonded to sulfapyridine. Initial case reports suggested sulfasalazine teratogenicity with evidence of cardiovascular, genitourinary, and neurologic defects. However, a larger series of 181 pregnant women did not note an increase in congenital anomalies. A population-based study using the Hungarian Case Control Surveillance of Congenital Abnormalities database also did not find a significant increase in the prevalence of congenital abnormalities in the children of women treated with sulfasalazine. Given the concern over potential antifolate effects of the drug, it is recommended that women take folic acid 2 mg daily in the prenatal period and throughout pregnancy. Breast-feeding is also considered low risk with sulfasalazine. Unlike other sulfonamides, bilirubin displacement, and therefore kernicterus, does not occur in the infant. This may be due to negligible transfer via breast milk.

Case series of mesalamine use in pregnancy do not suggest an increased risk to the fetus. This has been supported by a prospective controlled trial of 165 women exposed to mesalamine compared with matched controls with no exposure and a population-based cohort study from Denmark. Breastfeeding while on aminosalicylates has been rarely associated with diarrhea in the infant. Women can breastfeed while being treated with 5-aminosalicylates, but infants should be observed for a persistent change in stool frequency.

Antibiotics

Metronidazole is a pregnancy category B drug. Multiple studies have suggested that prenatal use of metronidazole is not associated with birth defects. These studies include two meta-analyses, two retrospective cohort studies, and a prospective controlled study of 228 women exposed to metronidazole during pregnancy. A population-based case-control study found that overall teratogenic risk was low, but infants of women exposed to metronidazole in the second to third months of pregnancy had higher rates of cleft lip with or without cleft palate. This increase was slight and not believed to be clinically significant.

Metronidazole is excreted in breast milk. If a single dose of metronidazole is given, the American Academy of Pediatrics recommends that breastfeeding should be suspended for 12 to 24 hours. Potential toxicity exists for longer-term use of metronidazole, and it is not compatible with breastfeeding.

Quinolones (eg, ciprofloxacin, levofloxacin, norfloxacin) are pregnancy category C drugs. Quinolones have a high affinity for bone tissue and cartilage and may cause arthropathies in children. The manufacturer reports damage to cartilage in weight-bearing joints after quinolone exposure in immature rats and dogs. However, a prospective controlled study of 200 women exposed to quinolones and a population-based cohort study of 57 women exposed to quinolones did not find an increased risk of congenital malformations. Overall, the risk is believed to be minimal, but given safer alternatives, the drug should be avoided in pregnancy. The data in breastfeeding are limited, but quinolones are probably compatible with use.

In general, given the limited evidence of benefit of these agents in IBD and the extended duration of use in the treatment of CD and UC, they should be avoided during pregnancy. Short courses for the treatment of pouchitis can be considered based on the safety data presented previously. An alternative antibiotic for pouchitis is amoxicillin/clavulanic acid, a pregnancy category B drug. A population-based case-control study and a prospective controlled study did not show evidence of increased teratogenic risk, and it is compatible with breastfeeding.

Corticosteroids

Corticosteroids are pregnancy category C drugs. A case-control study of corticosteroid use during the first trimester of pregnancy noted an increased risk of oral clefts in the newborn. This was confirmed by a large case-control study and a meta-analysis that reported a summary OR for case-control studies examining the risk of oral clefts (3.35 [95% CI, 1.97–5.69]). However, the overall risk of major malformations was low (1.45 [95% CI, 0.80–2.60]). A prospective controlled study of 311 women who received glucocorticosteroids during the first trimester did not note an increased rate of major anomalies and no cases of oral cleft were noted. The study was powered to find a 2.5-fold increase in the overall rate of major anomalies. An increased risk of premature rupture of membranes and adrenal insufficiency in the newborn has been reported in the transplant setting and gestational diabetes is a concern as well. Overall, the use of corticosteroids poses a small risk to the developing infant but the mother needs to be informed of both the benefits and the risks of therapy. Prednisone and prednisolone are compatible with breastfeeding.

A case series of eight patients with CD treated with budesonide did not find an increased risk of adverse outcomes. Inhaled or intranasal budesonide is not associated with adverse fetal outcomes based on large clinical series. Safety in lactation is not known.

Immunomodulators

The immunomodulators are the most controversial agents used in the treatment of the pregnant woman with IBD.

Methotrexate

Methotrexate, a pregnancy category X drug, is clearly teratogenic and should not be used in women considering conception. Methotrexate is a folic acid antagonist, and use during the critical period of organogenesis (6–8 weeks postconception) is associated with multiple congenital anomalies collectively called methotrexate embryopathy or the fetal aminopterin-methotrexate syndrome. The syndrome is characterized by intrauterine growth retardation; decreased ossification of the calvarium; hypoplastic supraorbital ridges; small, low-set ears; micrognathia; limb abnormalities; and sometimes mental retardation. Exposure in the second and third trimesters may be associated with fetal toxicity and mortality. Methotrexate may persist in tissues for long periods, and it is suggested that patients wait at least 6 months from the discontinuation of the drug before attempting conception.

Methotrexate is excreted in breast milk and may accumulate in neonatal tissues. The AAP classifies methotrexate as a cytotoxic drug with the potential to interfere with cellular metabolism. It is contraindicated in breastfeeding.

Azathioprine/6-mercaptopurine

6-mercaptopurine (6-MP) and its prodrug azathioprine (AZA) are pregnancy category D drugs. Animal studies have demonstrated teratogenicity with increased frequencies of cleft palate, open-eye, and skeletal anomalies seen in mice exposed to AZA and cleft palate, skeletal anomalies, and urogenital anomalies seen in rats. Transplacental and transamniotic transmission of AZA and its metabolites from the mother to the fetus can occur. The oral bioavailability of AZA (47%) and 6-MP (16%) is low, and the early fetal liver lacks the enzyme inosinate pyrophosphorylase needed to convert AZA to 6-MP. Both features may protect the fetus from toxic drug exposure during the crucial period of organogenesis.

The largest evidence on safety comes from transplantation studies where rates of anomalies ranged from 0.0% to 11.8% and no evidence of recurrent patterns of congenital anomalies emerged. A population-based cohort study from Denmark compared 11 women exposed to AZA or 6-MP with the general population. The adjusted OR for congenital malformations was 6.7 (95% CI, 1.4–32.4). However, when a single severely ill patient with autoimmune hepatitis and multiple other medications was removed from the cohort, the OR was 3.4 (95% CI, 0.4–27.3).

In IBD, multiple case series have not noted an increase in congenital anomalies, although one study did report a higher incidence of fetal loss in women with IBD with prior treatment on 6-MP compared with those who never had 6-MP exposure. However, a Danish nationwide cohort study found that women with CD exposed to corticosteroids and AZA/6-MP were more likely to have preterm birth (12.3% and 25.0% respectively) compared with non-IBD controls (6.5%). Congenital anomalies were also more prevalent among AZA/6-MP exposed cases compared with the reference group (15.4% vs 5.7%) with an odds ratio of 2.9 (95% CI 0.9–8.9). However, only 26 women were exposed to AZA/6-MP during conception versus 628 patients in the reference group, and the authors controlled for “disease activity,” which they defined as more than or fewer than two admissions for disease exacerbation, accounting for only the most severe patients. Finally, the largest single study to date studied 189 women who called a teratogen information service after exposure to AZA during pregnancy and compared them with 230 women who did not take any teratogenic medications during pregnancy. The rate of major malformations did not differ between groups with six neonates in each; for AZA, the rate was 3.5% and for the control group rate it was 3.0% ( P = .775; OR 1.17; CI 0.37, 3.69).

Breastfeeding, initially discouraged, may now be compatible with AZA use. Small studies in IBD suggest that the overall exposure to the infant is low. Moretti and colleagues reported four women breastfeeding on AZA. In two women, multiple breast milk samples did not have detectable levels of drug by high-performance liquid chromatography, and none of the four infants had any complications. Three other studies measured metabolite levels in the breastfeeding infant. Gardiner reported four infants with undetectable metabolite levels despite mothers whose levels were in the therapeutic range. Sau collected 31 samples from 10 breastfeeding women on AZA/6-MP. Only two samples had low levels of 6-MP in breast milk (1.2 and 7.6 ng/mL in one patient vs a serum level of 50 ng/mL). There were no detectable 6-thioguanine nucleotide or 6-methyl mercaptopurine levels in the 10 infants, nor were there signs of hematologic or clinical immunosuppression. Finally, an elegant study of eight lactating women on AZA obtained milk and plasma samples at 30 and 60 minutes after drug administration and hourly for the following 5 hours. The variation in the bioavailability of the drug was reflected in a wide range of peak plasma values of 6-MP within the first 3 hours. A similar curve, but with an hour’s delay and at significantly lower concentrations varying from 2-50 μg/L, was seen in maternal milk. Most 6-MP in breast milk was excreted within the first 4 hours after drug intake. On the basis of maximum concentration measured, the infant ingested 6-MP of less than 0.008 mg/kg bodyweight/24 h. The risks and benefits of breastfeeding must be considered carefully; however, at this time there does not appear to be an absolute contraindication to breastfeeding while on AZA/6-MP and mothers should be advised to wait 4 hours after dosing to feed.

Cyclosporine and tacrolimus

Cyclosporine is a pregnancy category C drug. A meta-analysis of 15 studies of pregnancy outcomes after cyclosporine therapy reported a total of 410 patients with data on major malformations. The calculated OR of 3.83 for malformations did not achieve statistical significance (95% CI, 0.75–19.6). The rate of malformations was 4.1%, which is not different from the general population. The conclusion of the study was that cyclosporine did not appear to be a major human teratogen. In a study published in the obstetric literature, a retrospective review of 38 pregnancies in 29 women between 1992 and 2002 was conducted. There were four spontaneous abortions and 10 first-trimester terminations for worsening liver function. The mean gestational age was 36.4 weeks, and there were no intrauterine or neonatal deaths. Five minor congenital anomalies were noted. The investigators concluded that planned pregnancy at least 2 years after liver transplantation with stable allograft function and continued immunosuppression had an excellent maternal and neonatal outcome.

There are several case reports of successful cyclosporine use during pregnancy to control UC and complete the pregnancy. In the setting of severe, corticosteroid-refractory UC, cyclosporine may be a better option than colectomy given the substantial risk to the mother and fetus of surgery during this time.

Cyclosporine is excreted into breast milk in high concentrations. Therefore, the AAP considers cyclosporine contraindicated during breastfeeding because of the potential for immune suppression and neutropenia.

Tacrolimus is also a pregnancy category C drug. The earliest experience with this medication was in 1997, with a report of 27 pregnancies with exposure to tacrolimus. Two infants died at weeks 23 and 24, but the mean gestational period was 36.6 weeks. There was a 36% incidence of transient perinatal hyperkalemia. One newborn had unilateral polycystic renal disease. Another study from Germany reported on 100 pregnancies in transplant recipients followed from 1992 to 1998. There were a 68% live birth rate, 12% spontaneous abortion rate, and 3% stillbirth rate. Fifty-nine percent of the infants were premature. Malformations occurred in four neonates with no consistent defects. In a later single-center experience, 49 pregnancies in 37 women over 13 years were followed prospectively. Thirty-six women survived the pregnancy, and two premature babies were seen. One infant died of Alagille syndrome; the rest survived, and 78% were of normal birth weight. No other congenital abnormalities were noted. A single case report of a patient with UC who had a successful pregnancy on maintenance tacrolimus was recently published. No other data on IBD are published at this time. Tacrolimus is contraindicated in breastfeeding because of the high concentrations found in breast milk.

Thalidomide

Thalidomide, a pregnancy category X drug, has some anti–tumor necrosis factor (anti-TNF) effects and has been used successfully for the treatment of CD. However, its teratogenicity has been extensively documented and includes limb defects, central nervous system effects, and abnormalities of the respiratory, cardiovascular, gastrointestinal, and genitourinary systems. Thalidomide is contraindicated during pregnancy and in women of childbearing age who are not using two reliable methods of contraception for 1 month before starting therapy, during therapy, and for 1 month after stopping therapy. There are no human data on breastfeeding, but it is not advised given the potential toxicity.

Infliximab

Infliximab (INF), a pregnancy category B drug, is used for the management of CD and UC. INF is an IgG1 antibody, which does not cross the placenta in the first trimester, but very efficiently crosses in the second and third trimester. Although this protects the infant from exposure during the crucial period of organogenesis, INF levels cross efficiently in the third trimester and are present in the infant for several months from birth.

There is a growing body of evidence that suggests INF is low risk in pregnancy. The two largest studies are from the TREAT Registry and the INF Safety Database maintained by Centocor (Malvern, PA). The TREAT Registry is a prospective registry of patients with CD. Patients may or may not be treated with INF. Of the more than 6200 patients enrolled, 168 pregnancies were reported, 117 with INF exposure. The rates of miscarriage (10.0% vs 6.7%) and neonatal complications (6.9% vs 10.0%) were not significantly different between INF-treated and INF-naïve patients, respectively. The INF Safety Database is a retrospective data collection instrument. Pregnancy outcome data are available for 96 women with direct exposure to INF. This was primarily exposure in during conception and the first trimester. When patients found out they were pregnant, the treatment was often stopped. The 96 pregnancies resulted in 100 births. The expected versus observed outcomes among women exposed to INF were not different from those of the general population. A series of 10 women with maintenance INF use throughout pregnancy was also reported. All 10 pregnancies ended in live births, with no reported congenital malformations. Another series reported 22 patients with exposure to INF within 3 months of conception, continued until 20 weeks of gestation at which time the drug was stopped to minimize placental transfer. Several of the patients did have a flare of disease in the third trimester. There were three spontaneous abortions, one missed abortion, one stillbirth at 36 weeks (umbilical strangulation), two preterm births, three low birth weight infants, and no congenital anomalies.

INF crosses the placenta and is detectable in the infant for several months after birth. A case report noted higher than detectable INF levels in an infant born to a mother on INF therapy every 4 weeks. The mother breast-fed and continued to receive INF but the infant’s INF level dropped over 6 months, suggesting placental rather than breast milk transfer. The effect of high INF levels on the infant’s developing immune system is not known, although at 7 months the infant had appropriate responses to vaccination. In a case series of eight patients receiving INF during pregnancy, all eight patients delivered a healthy infant. The mothers were receiving INF 5 mg/kg every 8 weeks and the mean time between delivery and the last infusion was 66 days (range 2–120 days). The INF level at birth was always higher in the infant and cord blood than in the mother and it took anywhere from 2 to 7 months for the infant to have undetectable INF levels. These findings support the fact that IgG1 antibodies are very efficiently transported across the placenta in the third trimester, but the infant reticuloendothelial system is too immature to effectively clear the antibody rapidly. INF has not been detected in breast milk.

So far, there has been no reported adverse event associated with elevated INF levels in the newborns. In our experience, infants exposed to INF in utero have an appropriate response to standard early vaccinations. In adults receiving a similar agent, adalimumab, pneumococcal and influenza vaccinations were given safely and effectively. However, live vaccinations, such as varicella and small pox, are contraindicated in immunosuppressed patients, such as those on anti-TNF therapy. Traditionally, the first live virus encountered by an infant was at 1 year of age (varicella, measles-mumps-rubella) when INF levels would be undetectable. However, now, rotavirus live vaccine is given at 2 months of age. Although it is given orally and is significantly attenuated, its safety in this setting is not known and the mother and pediatrician should be cautioned against its use if INF or ADA levels may be present.

Adalimumab

Adalimumab (ADA), a pregnancy category B drug, is FDA approved for induction and maintenance of remission in CD. Three case reports document the successful use of ADA to treat CD during pregnancy, including one in which the patient received weekly dosing throughout pregnancy for a total of 38 doses. OTIS (Organization for Teratology Information Specialists) reports 27 women enrolled in a prospective study of ADA in pregnancy and an additional 47 ADA exposed pregnant women in a registry (Chambers CD Johnson D, Jones KL. Pregnancy outcomes in women exposed to adalimumab: The OTIS autoimmune diseases in pregnancy project, personal communication; July 13, 2007). The rate of spontaneous abortion and stillbirth was similar to the diseased comparison and the general population. The rates of congenital malformation and preterm delivery are also within the expected range.

ADA, an IgG1 antibody, would be expected to cross the placenta in the third trimester as INF does. However, as ADA levels cannot be checked commercially, this has not been confirmed. ADA is considered compatible with breastfeeding although there are no human data.

Certolizumab pegol

Certolizumab pegol (CZP) is a PEGylated Fab’ fragment of a humanized anti-TNFα monoclonal antibody. As it does not have an Fc portion, it should not be actively transported across the placenta as INF and ADA are. A study of pregnant rats receiving a murinized IgG1 antibody of TNFα and a PEGylated FAB’ fragment of this antibody, demonstrated much lower levels of drug in the infant and in breast milk with the Fab’ fragment compared with the full antibody. These findings were confirmed in two patients with CD receiving CZP during pregnancy. Both patients received the drug during the 2 weeks before delivery. The mothers’ levels were high (19.60) but the infants’ level (1.02) and cord blood level (1.65) were low on the day of birth. However, one concern may be that in theory the Fab’ fragment (and the IgG1 antibodies) may cross the placenta passively in low levels in the first trimester during the period of organogenesis. A single case report in a patient with Crohn reported a successful pregnancy on CZP. Further data are clearly needed on all the anti-TNF agents.