Statistics from Altmetric.com
Thiopurines, including azathioprine and mercaptopurine, have a good safety record in pregnancy.1–3 Meta-analysis shows that maternal thiopurine use is not associated with either congenital abnormality or low birth weight.1 A positive association with preterm delivery has been attributed to the severity of the underlying condition for which thiopurine use acts as a surrogate marker.1 However, Jharap et al suggest that all infants exposed to thiopurines in pregnancy should be routinely tested for anaemia. This conclusion is based on a subgroup analysis of only 16 infants, a median ‘anaemic’ haemoglobin of 14.9 g/dL, which lies within the 95% reference range, and a statistically insignificant correlation between thiopurine metabolite concentration and anaemia.4
We challenged this recommendation by undertaking a retrospective cohort study of pregnant women receiving thiopurine therapy referred to two UK tertiary referral centres over a 3-year period (2011–2013). This identified 58 women who had received thiopurine therapy in pregnancy and who delivered after 24 weeks gestation (azathioprine: n=54; mercaptopurine: n=4). The majority of patients had IBD (n=25; 43%), but there were also women with vasculitides including systemic lupus erythematosus (n=20; 34%), renal transplants (n=3; 5%) and interstitial lung disease (n=3; 5%). The majority of women were receiving a daily dose of 100 mg of azathioprine (range 50–150 mg daily, 1.0–1.5 mg/kg/day). Concomitant medication use included prednisolone (n=18; 31%), hydroxychloroquine (n=12; 21%) and mesalazine (n=3; 5%). The rate of adverse pregnancy outcome reflected the high-risk nature of the cohort, with 17% (10/59) of babies delivering preterm, 19% (11/59) being small for gestational age and 14% (8/59) being admitted to the neonatal unit. One baby had Down's syndrome.
A full blood count was performed in 11 neonates out of 59 (18.6%) for clinical reasons including extreme prematurity and suspected maternal and neonatal sepsis. Neonatal haemoglobin concentrations in this group ranged from 14.1 to 19.6 g/dL, and were all within the normal range corrected for gestation (see figure 1). Neonatal white cell concentration ranged from 5.4×109/L to 30.8×109/L. A single infant was identified to have neutropenia, attributable to subclinical chorioamnionitis. There was no clinical suspicion of anaemia in the routine neonatal examination of the remaining 48 (81.4%) infants in the cohort.
Our data support the evidence-based consensus of the European Crohn's and Colitis Organisation that thiopurines should be considered low risk during both pregnancy and lactation.5 In addition, long-term follow-up of babies exposed to azathioprine in pregnancy and lactation shows no difference in either the risk of infection or hospitalisation under 5 years of age.6
Our study highlights the inadequacy of the US Food and Drug Administration (FDA) classification of drugs in pregnancy. Thiopurines have an FDA category D status based on skeletal, visceral and haematological abnormalities identified in high-dose animal studies (5 mg/kg) and in three isolated case reports in humans.7 In June 2015, the FDA classification was replaced by the Pregnancy and Lactation Labelling Rule (or final rule). We hope that this will improve prescription drug labelling and enable healthcare professionals to assess the true risk of medication use in pregnancy, thereby avoiding harm due to inappropriate discontinuation of medication including thiopurines.
There remains a need for prospective studies of the use of thiopurines in pregnancy with myelotoxicity as a primary outcome, including a comprehensive evaluation of confounding factors including maternal disease activity and peripartum events. Until such studies are completed, there remains insufficient evidence to recommend routine testing for anaemia following intrauterine exposure to thiopurines.
Contributors CN-P conceived and designed the study. MC, SJ and KW acquired the data and undertook the analysis and interpretation. All authors jointly contributed to writing and editing the manuscript for submission.
Competing interests None declared.
Provenance and peer review Not commissioned; internally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.