Abstract
The Royal College of Obstetricians (RCOG) have recently published their Green-top Guideline on the Management of Thyroid Disorders in Pregnancy. 1 These guidelines are welcomed considering that thyroid disorders are common in pregnancy and inappropriate or inadequate management can lead to preventable maternal and fetal morbidity. Their publication provides the opportunity to ensure that the advice women receive from their primary care clinicians, the maternity team and their endocrinologist is consistent. While the guideline provides clarity on multiple aspects of care, we have concerns that some of the recommendations (Table 1) lack an evidence-base and feasibility in implementation.
Relevant Green-Top guideline 1 recommendations.
We agree that treatment of subclinical hypothyroidism (SCH) and overt hypothyroidism needs to be adequate and ideally optimised pre-pregnancy. We support and welcome the recommendations for adequate iodine supplementation and the use of trimester- and manufacturer-specific reference intervals in pregnancy. However, we are concerned that the recommended antenatal target concentration of thyroid stimulating hormone (TSH) of < 2.5 mU/L is not evidence-based and risks overtreatment.
A recommended TSH target of <2.5 mU/L is lower than that recommended in other international expert guidance. 2 While overt hypothyroidism has been associated with adverse pregnancy outcomes, there is no consistent evidence that TSH concentrations in the range 2.5–4.0 mU/L with normal thyroxine (T4) concentration lead to maternal or fetal harm. A cohort analysis by Li et al. found that, provided women have a TSH within the local pregnancy-specific reference interval, there is no increased risk of miscarriage even when antenatal TSH is >2.5mU/L. 3 Similarly, a large cohort study demonstrated that although a TSH >4 mU/L was associated with increased risks of preterm birth and neonatal respiratory distress syndrome, this was not the case with TSH concentrations 2.5–4.0 mU/L. 4 In a large US study of SCH, thyroxine treatment was associated with a reduced risk of pregnancy loss among those with pre-treatment TSH 4.1–10 mU/L but not in those with pre-treatment TSH 2.5–4.0 mU/L. 5 The Controlled Antenatal Thyroid Screening (CATS) trial 6 failed to demonstrate that treating women to a target TSH in the low end of the pregnancy reference interval leads to improved childhood anthropometric, bone, and cardiometabolic measurements 7 or short- and long-term cognitive outcomes.8,9
The harms of overtreating SCH in pregnancy are increasingly recognised. Both low and high maternal free thyroxine concentrations in pregnancy are associated with lower child IQ, grey matter and cortex volumes, 10 and the children of ‘overtreated’ mothers in the CATS II trial displayed measurably more ADHD symptoms and behavioural difficulties at age 9–10 years old. 8 We are concerned that RCOG recommendations to both empirically increase thyroxine doses on confirmation of pregnancy and the use of a target TSH concentration in the lower end of the pregnancy reference interval increase the risk of overtreatment in pregnancy. The recommendation that all women undergo a standardised increase in thyroxine dose is based on historical data from 19 women, including six needing thyroxine replacement after treatment for thyroid cancer which has specific and different treatment targets compared to the majority who have autoimmune thyroid disease. 11 An empirical increase in thyroxine dose can also lead to overtreatment, with TSH suppression seen in 32–65% of women depending on the recommended empirical increase. 12
It is perhaps the case that the risks of overtreatment are mitigated by the RCOG recommendation to increase thyroid function testing in pregnancy. RCOG guidance now recommends that pregnant women treated with levothyroxine have TSH and free T4 concentrations checked every 4–6 weeks until 20 weeks of gestation. This is arguably the only way to reliably achieve a target TSH in the lower end of the pregnancy reference interval, which typically requires 3–4 dose changes in pregnancy in empirically increased cohorts. 13 In practical terms, this represents up to five blood tests in the first half of pregnancy if pregnancy is diagnosed early, for up to 2% of the pregnant population, leading to an estimated 30,000 extra clinical attendances in England compared to previous frequency of thyroid function monitoring in pregnancy. The logistics, cost, clinical responsibility, and patient burden for this are not clear, but likely to be significant, without clear clinical benefit for those with adequate thyroxine replacement according to pregnancy-specific reference intervals.
We suggest that aetiology of hypothyroidism, and preconception TSH concentration and levothyroxine doses are taken into account when considering thyroxine dose increases in pregnancy. We also suggest that, in the absence of evidence of harm, a target TSH using an assay- and pregnancy-specific reference interval or <4.0 mU/L should be used in pregnancy. This has the advantage of requiring less frequent monitoring and reduces the risks of overtreatment and over medicalisation of pregnancy.
Finally, we welcome the recommendations regarding the management of Graves’ disease in pregnancy. The teratogenic risk with carbimazole (CBZ) exceeds that of propylthiouracil (PTU), and the spectrum of malformations with CBZ is more severe than for PTU. There is therefore consensus that a switch from CBZ to PTU should be made preconception or before 10 weeks' gestation. The RCOG guidance recommends that switching from CBZ to PTU after 10 weeks is not advisable as organogenesis is complete at this stage and a medication switch offers no benefit. This, in addition to supporting CBZ use beyond 20 weeks' gestation to reduce the risk of hepatotoxicity, is worthy of highlight as neither are common current practice.
