Management of Thyroid Disorders in Pregnancy

Thyroid disease is a common endocrine disorder in women of childbearing age. There is variation in clinical practice and approach to thyroid diseases globally, in part influenced by differences in population iodine status. There remains controversy regarding testing for and management of thyroid disorders before conception, during pregnancy and postpartum. This guideline presents the available evidence for best practice and, where evidence is lacking, consensus opinion by a multidisciplinary, cross-specialty team of authors is presented.

Preconception testing for thyroid dysfunction in the subfertile and recurrent miscarriage populations is not within the scope of this guideline and is addressed in a separate RCOG Scientific Impact Paper [1].

This guideline is for healthcare professionals who care for women, non-binary and trans people with thyroid disorders in pregnancy.

Within this document we use the terms woman and women's health. However, it is important to acknowledge that it is not only women for whom it is necessary to access women's health and reproductive services in order to maintain their gynaecological health and reproductive wellbeing. Gynaecological and obstetric services and delivery of care must therefore be appropriate, inclusive and sensitive to the needs of those individuals whose gender identity does not align with the sex they were assigned at birth.

Dynamic changes occur in thyroid function through the course of pregnancy, to provide adequate concentrations of thyroid hormone to the woman and fetus [2-4]. Overall, demands on maternal thyroid hormone production increase by approximately 50% during pregnancy; this requires both an adequate supply of iodine for the biosynthesis of thyroid hormones and the absence of significant thyroid disease.

Increased oestrogen in pregnancy raises thyroxine-binding globulin concentrations, starting very early in pregnancy, and plateauing by approximately 18–20 weeks of gestation. To maintain adequate free thyroid hormone concentrations, thyroxine (T4) and tri-iodothyronine (T3) production by the thyroid gland increases during the first half of pregnancy, a new steady-state is reached by mid-gestation and the synthesis of thyroid hormones returns to pre-pregnancy rates. First trimester increases in human chorionic gonadotrophin (hCG), which has weak thyroid-stimulating activity, transiently increases free thyroxine (fT4) and free tri-iodothyronine (fT3) and decrease thyroid stimulating hormone (TSH) [5]. From mid-gestation, as hCG declines, serum fT4 and fT3 concentrations decline gradually, while serum TSH concentrations rise slightly.

Iodine requirement increases considerably during pregnancy as there is increased consumption of iodine for thyroid hormone synthesis and increased renal iodine clearance [6]. The placenta may also be an organ of storage for iodine [7]. The fetal thyroid begins to take up iodine from 10 to 12 weeks of gestation and begins to produce and release appreciable amounts of thyroid hormone from 18 to 22 weeks of gestation. Breast milk production from the second half of gestation adds further to maternal iodine demand [8].

Maternal thyroid hormones are essential for the maintenance of pregnancy and may influence placental development [9]. Transplacental passage of maternal T4 is essential for normal fetal development, especially neurodevelopment during the first half of gestation [10-12]. The fetus is completely dependent upon maternal T4 prior to commencing production of its own thyroid hormone but remains reliant on maternal supply of iodine [2] and continues to receive maternal T4 until birth [13].

Worldwide, iodine deficiency is the leading cause of preventable neurodevelopmental defects [14]. Among populations of severe iodine deficiency (defined by a median urinary iodine concentration in a population of below 20μg/L) there is increased risks of endemic goitre, hypothyroidism, neurological and developmental impairment, subfertility, miscarriage, infant mortality, trophoblastic or embryonic/fetal disorders, profound intellectual impairment, deaf-mutism and motor rigidity in children, and childhood and adult learning difficulties [15]. In these areas of severe iodine deficiency, thyroid nodules have been reported in up to 30% of pregnant women [16]. In regions of mild to moderate iodine deficiency, pregnant women are also at increased risk for the development of goitre [17] and thyroid disorders [18], with one observational study reporting an association with small-for-gestational-age (SGA) fetuses and preterm birth [19]; others showed no association with any adverse obstetric outcome [20]. However, associations between lower maternal urinary iodine concentrations and altered executive function [21], attention deficit and hyperactivity disorders in children [22] and impaired cognitive outcomes [21, 23, 24] have been reported. Meanwhile, in areas with adequate dietary iodine intake, variations in maternal urinary iodine concentrations have been shown to have limited influence on neonatal and infant developmental outcomes.

In iodine-replete and mildly iodine deficient populations, autoimmunity is the commonest aetiology for thyroid disorders (Table 1). Untreated and inadequately-treated overt hypothyroidism (OH) is associated with an increased risk of spontaneous miscarriage, perinatal death, pre-eclampsia, pregnancy-induced hypertension, preterm birth, low birth weight and postpartum haemorrhage [25-27]. Untreated overt hyperthyroidism, commonly due to Graves' disease, is also associated with increased risks, in particular pre-eclampsia, preterm birth, fetal growth restriction and maternal heart failure [28-30].

Gestational transient thyrotoxicosis is common, affecting 1%–5% of pregnancies in Europe [28], and is usually benign and self-limiting. Hyperthyroidism in pregnancy is rarer and usually caused by Graves' disease; prevalence in pregnancy: 0.5%–1.3% pre-existing Graves', 0.05% new onset Graves', 0.1% toxic nodular disease [4, 29]. Autoimmune Graves' disease is mediated by the presence of stimulating TSH-receptor antibodies (TRAb), and commonly improves with advancing gestation. Gestational transient thyrotoxicosis usually does not require treatment, whereas hyperthyroidism (caused by Graves' or toxic nodular disease) does. Subclinical hyperthyroidism is defined as below normal serum TSH concentrations with normal concentrations of circulating thyroid hormones. Current evidence indicates that it is not associated with adverse fetal or pregnancy outcomes and therefore, does not require treatment (see Appendix F) [31, 32].

Thyroid autoimmunity is the presence of circulating antithyroid autoantibodies that are targeted against the thyroid, with or without thyroid dysfunction. Thyroid peroxidase antibodies (TPOAb) are the most common antithyroid autoantibodies. The prevalence of thyroid autoantibodies varies between 5% and 31% across studies and populations [33]. In a large cohort of 19 556 women with a history of miscarriage or subfertility, thyroid autoantibodies were present in approximately 10% [34]. Euthyroidism with thyroid autoimmunity has been associated with increased risk of miscarriage, preterm birth and postpartum thyroiditis (PPT) [33, 35]. When looking at thyroid autoimmunity in conjunction with thyroid dysfunction, there may additionally be an association with increased risks of pre-eclampsia and gestational diabetes [36-38].

The incidence of clinically apparent nodules or goitre presenting in pregnancy in iodine-replete and mildly iodine-deficient areas is low [39]. Ultrasound-detected nodules are more common with increasing parity and age [40, 41]. When a new thyroid nodule or goitre is diagnosed in pregnancy, local symptoms such as tracheal compression should be assessed, and malignancy and hyperthyroidism excluded.

There is controversy concerning the need and cost-effectiveness of routinely testing for thyroid disease (Table 1) and for thyroid autoimmunity in the first trimester of pregnancy or in women who are planning for pregnancy. Whether levothyroxine treatment improves pregnancy and offspring outcomes in subclinical hypothyroidism (SCH) and isolated hypothyroxinaemia (IH) remains debatable. Controversies in the care of these conditions for the general obstetric population will be discussed in this guideline. Uncertainties in the management of thyroid function abnormalities in the care of subfertility and recurrent miscarriage is addressed specifically in a separate RCOG Scientific Impact Paper [1].

Both inadequate and excessive treatment of thyroid disorders, the choice of treatment, as well as delayed commencement and adjustment of treatment, can also result in detrimental effects on the pregnancy and fetus. Therefore, care should be optimised prior to conception, during pregnancy and after birth, and should be provided by clinicians with appropriate experience (see Appendix C for suggested designated roles and responsibilities of healthcare professionals, and care pathways). There should be a clear designated primary clinician and this will depend upon local expertise. This is important to ensure continuity of care over the course of pregnancy, minimise confusion with regards to treatment adjustments and to improve overall experience. Where possible, women and people with thyroid disorders should be seen in joint multidisciplinary clinics, both when planning pregnancy and during pregnancy, comprising clinicians with obstetric and endocrine expertise.

This guideline was developed using standard methodology for developing RCOG Green-top Guidelines (GTGs) [42]. The Cochrane Library (including the Cochrane Database of Systematic Reviews, the Database of Abstracts of Reviews of Effects [DARE] and the Cochrane Central Register of Controlled Trials [CENTRAL]), EMBASE, MEDLINE and Trip were searched for relevant papers. The search was inclusive of all relevant articles published until July 2023. The databases were searched using the relevant Medical Subject Headings (MeSH) terms, including all subheadings and synonyms, and this was combined with a keyword search. Search terms included ‘*thyroid diseases’, ‘*euthyroid sick syndromes’, ‘*goitre’, ‘*hyperthyroidism’, ‘*hyperthyroxinemia’, ‘*thyroid dysgenesis’, ‘*thyroid nodule' and ‘*thyroiditis'. The search was limited to studies on humans and papers in the English language. Relevant guidelines were also searched for using the same criteria in the National Guideline Clearinghouse and the National Institute for Health and Care Excellence (NICE) Evidence Search.

Where possible, recommendations are based on available evidence. Areas lacking evidence are highlighted and graded accordingly. Further information about the assessment of evidence and the grading of recommendations may be found in Appendix A.

Retrospective observational studies of presumed treated OH have shown no difference in pregnancy outcome compared with women without OH. However, the adequacy of OH treatment or the absence of hypothyroidism in the control group was not specifically ascertained [107]. Treatment needs to be adequate, and ideally optimised pre-conception, with appropriate advice given before pregnancy on instituting an empirical dose increase upon conception to prevent hypothyroidism in early pregnancy. Adverse pregnancy outcomes including premature birth, low birth weight, pregnancy loss, and impaired neurological development in babies are more common and severe in OH than in SCH. A retrospective study of more than 1000 pregnant women on levothyroxine replacement therapy, demonstrated that the risk of pregnancy loss increased proportionally to the degree of TSH elevation above 4.5 mU/L in the first trimester [108]. Similarly, the incidence of neurodevelopmental defects and lowering of children's IQ at 7–9 years of age demonstrated a graded response with higher maternal TSH concentrations during pregnancy associated with higher risk to children [109]. [Evidence level 2+]

The goal of levothyroxine treatment is to normalise and maintain maternal serum TSH values within the trimester-specific pregnancy reference range, and to mimic physiological changes and prospectively prevent abnormalities in thyroid function. Hence, dose titration using the lower half of the TSH reference range as a guide is commonly adopted during pregnancy (i.e., 0.1–2.5 mU/L) [110]. [Evidence level 2–]

In pregnancy, SCH may be defined as an increased TSH above the upper limit of the trimester-specific pregnancy reference range, with severe SCH defined as cases with TSH greater than 10 mU/L, accompanied by normal concentrations of thyroid hormones. The annual rate of progression of SCH to OH in the non-pregnant population ranges from 2% to 6% [111]. Risk factors for progression include increased TPOAb and an initial TSH above 10 mU/L [112]. In pregnant women who are TPOAb positive, post-hemithyroidectomy, or treated with radioactive iodine, progression to OH is more likely due to the inability of the thyroid to augment production when needed during pregnancy. [Evidence level 2+]

Observational studies have linked SCH with adverse pregnancy outcomes. Meta-analyses of individual participant data (n = 47 045, 19 cohorts) showed that SCH was associated with a higher risk of pre-eclampsia (odds ratio [OR] 1.53; 95% confidence interval [CI] 1.09–2.15) [113], preterm birth (OR 1.29; CI 1.01–1.64) [114], and SGA at birth (OR 1.24; CI 1.04–1.48) [115]. [Evidence level 2+]

Other meta-analyses reported increased risks of pregnancy loss (OR 1.93; CI 1.40–2.64); hypertensive disorders of pregnancy (OR 1.54; CI 1.21–1.96) [116]; placental abruption (OR 2.16; CI 1.15–4.06) [117]; breech presentation at term (OR 2.3; CI 1.50–3.51); and an increased incidence of neurodevelopmental defects (correlated with degree of TSH elevation) [109]. [Evidence level 2–]

Levothyroxine treatment is recommended for pregnant women with severe SCH (TSH above 10 mU/L) where there is considerable risk of further progression to OH. Levothyroxine treatment should also be considered for those with SCH where the TSH is between the upper limit of the pregnancy reference range and 10 mU/L, particularly if TPOAb positive, in order to reduce the risk of developing OH and associated complications in pregnancy [51, 118]. Even though the two largest trials of treatment of SCH [91, 92] (discussed in Section 6) reported no difference in maternal and child outcomes with levothyroxine treatment, they were still limited statistically by both sample size and late commencement of treatment, usually from the second trimester of pregnancy. When these trials were considered together with others in a systematic review and meta-analysis totalling seven RCTs and six observational studies (N = 7342) [119] it was concluded that levothyroxine treatment of SCH may reduce the risk of pregnancy loss (relative risk [RR] 0.79; CI 0.67–0.93) and neonatal death (RR 0.35; CI 0.17–0.72). [Evidence level 2–]

Another meta-analysis included only studies which defined SCH in pregnancy using a TSH threshold of greater than 4 mU/L; this study which included three RCTs and three observational studies (N = 7955) [120] reported a reduction in pregnancy loss (OR 0.55; CI 0.43–0.7), as well as preterm birth (OR 0.63; CI 0.41–0.98) and gestational hypertension (OR 0.78; CI 0.63–0.97) with levothyroxine treatment. [Evidence level 2–]

IH is considered to be a distinct biochemical entity, usually defined as a fT4 concentration below the 2.5th percentile, with a TSH within the reference range. However, there remains some variability of IH definition, and an absence of established regional reference ranges (that account for iodine status) in TPOAb negative populations could lead to misclassification of IH status [121]. The most common cause of IH is iodine deficiency [84]. Other proposed causes include obesity [122], iron deficiency [123], and exposure to organochlorine pesticides [124].

Some studies have shown an association between hypothyroxinaemia and poorer cognitive development of the children [125-127]. Results from observational studies of IH on pregnancy outcomes are conflicting and a meta-analysis identified placental abruption alone to be increased in women with IH (pooled OR 2.3; CI 1.1–4.8) [117, 128] but a causal link has not been established. [Evidence level 2–]

Existing randomised trials of women with SCH and IH diagnosed in pregnancy, with levothyroxine treatment mostly commenced in the second trimester of pregnancy, did not show improvement in child neurodevelopmental outcomes [91, 92], however, they were underpowered to assess efficacy within the subgroup of women who commenced levothyroxine in the first trimester of pregnancy (less than 12 weeks of gestation). The potential consequences of levothyroxine overtreatment should also be considered (see below).

Transplacental delivery of specifically maternal T4 is essential for the developing fetal brain from early first trimester of pregnancy [12, 126]. The recommended treatment of maternal hypothyroidism is oral levothyroxine. It is strongly recommended that other thyroid preparations that contain non-T4 forms of thyroid hormone, such as desiccated thyroid or combined levothyroxine and liothyronine therapy are not used in pregnancy, as these may result in insufficient transfer of maternal T4 to the fetal brain. Women already on such treatments should be strongly advised to switch to a levothyroxine only preparation, but a graded switch can be considered. [Evidence level 3]

Untreated or poorly controlled hyperthyroidism is associated with a number of adverse outcomes, but it remains unclear whether these consequences relate to maternal hyperthyroidism, to fetal hyperthyroidism (caused by transplacental transfer of thyroxine or stimulating TSH-receptor antibodies [TRAb]) or to antithyroid drug treatment, which may cause fetal hypothyroidism as well as direct toxicity [147]. Large record linkage studies [148-151] have confirmed increased risks of pre-eclampsia, stillbirth, maternal admission to intensive care unit, lower birth weight and higher rates of attention deficit hyperactivity disorder and autism in children when comparing women with hyperthyroidism and control subjects [152, 153]. Observational studies reported increased risks of intrauterine growth restriction, pre-eclampsia, preterm birth and maternal heart failure [154-157]. [Evidence level 2++]

The risk is directly related to control of maternal hyperthyroidism, both in terms of severity of hyperthyroidism and how soon in pregnancy euthyroidism is achieved [147, 158, 159]. [Evidence level 2+]

Maintenance of euthyroidism with optimal treatment throughout pregnancy coupled with adequate antenatal care, has not been associated with increased obstetric risks except for a possible residual risk of placental abruption [154, 156] [Evidence level 2+] and antithyroid drug associated teratogenicity [Evidence level 2++], where applicable (see Section 9.2).

The prevalence of thyroid nodules in pregnancy based on ultrasound studies in areas with mild-to-moderate iodine deficiency varies between 15% and 21% [40, 41], but the incidence of clinically apparent nodules presenting in pregnancy in non-iodine-deficient areas is likely to be under 1% [39]. [Evidence level 2+]

Ultrasound-detected nodules are more common with increasing parity and age [40, 41], and may increase in size during pregnancy [40, 189]. Thyroid cancer in association with pregnancy is very rare, with a prevalence of 14 in 100 000 [190] and is more likely to be diagnosed postpartum than at other times in pregnancy [39, 190]. [Evidence level 2+]

PPT is defined as the development of thyroid dysfunction, excluding other thyroid diseases, within the first 12 months following a pregnancy in a previously euthyroid woman [203]. This is an autoimmune disorder associated with antibodies to TPO and thyroglobulin [204], caused by a reactivation of the immune system following the relative immune suppression during pregnancy [205].

PPT occurs in 5%–10% of unselected pregnancies [206]. Women with other autoimmune disorders are at increased risk, in particular, those with type 1 diabetes mellitus [207], systemic lupus erythematosus [102] and a previous history of Graves' disease [208]. PPT may also occur in those with Hashimoto's thyroiditis [209] or with a personal or family history of thyroid disease [205]. Overall, 30%–50% of women with positive TPOAb develop PPT with higher risk in those with higher TPO antibody concentrations [210]. [Evidence level 3]

The classical form of PPT is triphasic with an initial thyrotoxic phase followed by a transient hypothyroid phase and then a return to euthyroidism. The clinical course is variable with 20%–40% of women exhibiting the classical form, 20%–30% developing only thyrotoxicosis and 40%–50% presenting with isolated hypothyroidism [205, 210]. The thyrotoxic phase usually occurs between 2 and 6 months postpartum but may present up to 12 months following birth. The hypothyroid phase typically presents between 3 and 12 months postpartum and results in permanent hypothyroidism in up to 50% [51, 211, 212]. Risk factors for permanent hypothyroidism include multiparity, higher concentrations of TPOAb, greater maternal age, more severe hypothyroidism, thyroid hypoechogenicity on ultrasound scanning and a history of pregnancy loss [51, 205, 213]. The risk of relapse of PPT with subsequent pregnancies is as high as 70%, especially in TPOAb positive women [205]. Some studies have indicated a link between TPOAb positivity [214-216], PPT [217] and postpartum depression, but an RCT of levothyroxine prophylaxis in TPOAb positive women did not lower rates of postpartum depression [218]. [Evidence level 1+]

British Thyroid Association (https://www.british-thyroid-association.org/).

British Thyroid Foundation (https://www.btf-thyroid.org/).

The UK Iodine group (https://www.ukiodine.org/).

SC is part of an academic consortium that has received grants from Société Des Produits Nestlé S.A. and Bayer, and is a co-inventor on patent filings by Nestlé S.A. unrelated to the published work. SC has received reimbursement from the Expert Group on Inositol in Basic and Clinical Research (EGOI; a not-for-profit academic organisation) and honoraria from Nestlé Nutrition Institute for speaking at conferences. JG is a co-inventor of patent filings by King’s College London unrelated to the published work. KB has received consulting fees from Eli Lilly, Immunovant, SERB, and Egetis Pharmaceuticals, unrelated to the published work; she has received speaker honoraria from the Gulf Association for Endocrinology and Diabetes (GAED) and from SERB; she is Associate Editor for the Journal of the Endocrine Society for which she receives honoraria. RS received reimbursement to attend scientific conferences from IBSA Pharma and Theramex, and speaker honoraria from Ferring. All other authors declare no Conflicts of Interest.