Management of fetal hyperthyroidism caused by persistent autoimmune antibodies in a case of previously treated maternal Graves’ disease

Abstract

Maternal Graves’ autoantibodies are well known to cause fetal and neonatal thyroid disturbances. Despite radioiodine therapy, Graves’ autoantibodies are known to persist, which can cross the placenta and cause hyperthyroidism in the fetus. We present the case of a 26-year-old woman in her first pregnancy, clinically and biochemically euthyroid with history of treated Graves’ disease, where the fetus showed signs of hyperthyroidism on antenatal scans. This was confirmed by amniotic fluid testing as fetal blood sampling was not feasible and successfully treated with maternal carbimazole whilst continuing thyroxine for the mother (block-replacement). We discuss the challenges in the diagnosis of fetal hyperthyroidism and treatment whilst maternal thyroid status is maintained on thyroxine.

Maternal hyperthyroidism is documented in about 0.2% of all pregnancies, with Graves’ disease (GD) being the most prevalent cause. ¹ About 2% of neonates born to mothers with known GD demonstrate features of transient autoimmune hyperthyroidism due to the placental crossover of maternal thyroid-stimulating hormone receptor antibodies (TRAb). ² Early diagnosis and treatment of fetal hyperthyroidism is crucial to a successful outcome, but can be challenging as highlighted here.

Case

A 26-year-old woman in her first pregnancy was referred to our high-risk antenatal clinic for a review of fetal tachycardia at 28 weeks of gestation. She had a history of GD treated with radioactive iodine (I131) 2 years prior to presentation, with subsequent thyroxine replacement (100 µg/day). Her TRAb level 3 months post I131 was elevated at 29 IU/L (normal range <1.8) and she had missed subsequent specialist follow-up. She was clinically and biochemically euthyroid (table 1) with slight right eyelid retraction, but no active thyroid eye disease and had elevated TRAb level of 59 IU/l. There was no known family history of thyroid disease or thyroid-stimulating hormone (TSH) receptor-activating mutations.

Table 1.

Maternal and neonatal TSH trends through pregnancy.

Weeks of gestation	TSH (0.1−5 mU/l)	fT4 (4−14 pmol/l)	fT3 (2.5−6 pmol/l)	TRAb (<1 IU/l)	Management
28	0.06	10		59	Drop thyroxine to 75 µg from 100 µg
30	0.05	11	5.2	56	Commenced on a block and replacement regime: PTU 50 mg BD + thyroxine 100 µg
31	0.10	10.8	4.2		Nil changes made
35	1	9.1	3.6	61	Continue PTU 50 mg BD, thyroxine 100 µg from Monday to Friday, 150 µg on Saturday and Sunday
36 + 5 (day of delivery)	4.2	8.9		52	PTU ceased. Maternal thyroxine dose changed to 100 µg daily (pre-pregnancy) post delivery.
Neonatal results
Day 1	<0.05	13	5.1	92
Day 3	<0.05	>77		67	Baby commenced on carbimazole and propranolol
Day 10	<0.05	21		49

TSH, thyroid-stimulating hormone; fT4, free tetra-iodothyronine; fT3, free tri-iodothyronine; TRAb, TSH receptor antibody; PTU, propylthiouracil; BD, twice daily.

Ultrasound scan (USS) at 28 weeks identified a fetal goitre with increased vascularity (Figure 1A to C) (thyroid circumference = 70 mm) and fetal tachycardia (170−190 b/min) with functional tricuspid regurgitation, which are known features but not sufficient for diagnosing fetal hyperthyroidism, and hence invasive testing was considered. Fetal blood sampling was not possible due to velamentous cord insertion. Therefore, an amniocentesis was performed at 30 weeks which showed a free tetra-iodothyronine (fT4) of 8.4 (5−9) pmol/l and free tri-iodothyronine (fT3) of 1.3 pmol/l (no reference range available) with low TSH <0.05 (0.1−0.5) mU/l and elevated TRAb consistent with fetal thyrotoxicosis. TSH, fT3 and the fT4 were measured on the two-step immunoenzymatic quantitative assays on the Beckman Coulter Synchron Clinical Systems. TRAb was measured using a competitive two-step immunoassay on the B.R.A.H.M.S. TRAK. A subsequent USS post amniocentesis revealed premature ossification of the distal femoral epiphysis and mild lateral ventriculomegaly (9−10 mm), again consistent with fetal hyperthyroidism.

Figure 1.

(A) Coronal view of the fetal neck at 30 weeks of gestation – Power Doppler image showing increased vascularity of thyroid 1. (B) Coronal view of the fetal neck at 30 weeks of gestation – calipers demonstrate enlarged thyroid on ultrasound 1. (C) Transverse view of the fetal neck at 30 weeks gestation – calipers demonstrating enlarged thyroid on ultrasound.

She was managed with 100 mg/day of propylthiouracil (PTU). Thyroxine dose was increased aiming for a placental transfer of this medication. The dose was titrated every 2 weeks based on fetal heart rate (FHR) and an USS at 36 weeks revealed normal FHR (145−160 b/min) with significant improvement in the goitre (Table 1).

She underwent an elective caesarean section at 36 weeks and 5 days of gestation and the baby weighed 2.5 kg (Apgar 9 at 1 and 5 min). Neonatal heart rate was within limits (130−160 /min); however, a physical examination showed exophthalmos but no palpable goitre. Clinically, the baby showed no evidence of fevers, sweats, irritability or any signs suggestive of hypermetabolism. Liver functions and platelet counts were within normal limits. Cord blood analysis revealed mild neonatal hyperthyroidism: TSH <0.05 (0.8−9.8) mU/l, fT4 13 (9.8−23 pmol/l), fT3 5.1 (3.5−6 pmol/l). This worsened markedly over 48 h in keeping with washout of maternal PTU levels. The neonate had an elevated TRAb titre of 92 IU/l and was commenced on 0.5 mg/kg/day of carbimazole and 2 mg/kg/day of propranolol. Subsequent, tests showed improvement in TSH and TRAb. Carbimazole was titrated and ceased during neonatology follow-up with no long-term sequelae.

Discussion

Maternal TRAbs are known to cross the placental barrier and stimulate fetal TSH receptors. ³ Although fetal thyroid development begins at 7 weeks of gestation, the thyroid hormone axis is functionally mature only later in gestation. ⁴ Also, transplacental passage of maternal immunoglobulin G antibodies increase during the second half of pregnancy. ⁴ Hence, placental cross-over of TRAb manifests in the third trimester as tachycardia, fetal growth restriction, abnormal amniotic fluid volume, advanced skeletal maturation and craniosynostosis on fetal ultrasound.^3,4 Fetal tachycardia, if untreated can lead to heart failure, which is associated with a perinatal mortality rate of about 12% to 45%.^3,5

It is well known that despite receiving ablative therapy for GD, approximately 40% of patients are TRAb positive post 5 years of therapy,^6,7 hence TRAb levels must be checked during early pregnancy. ¹ This was not organised in our patient as she was based rurally and had missed early specialist review. TRAb levels over three times of normal are increasingly associated with hyperthyroidism in fetus.^1,8

Fetal USS is reliable at detecting thyroid dysfunction, ruling out tracheal and oesophageal compression and monitoring response to anti-thyroid drugs. Studies demonstrate sensitivity and specificity of 92% and 100%, respectively, in expert hands for tertiary fetal scans. ⁹ Cordocentesis is being increasingly performed and remains the gold standard ¹⁰ but was not possible in our case as a result of the velamentous cord insertion; hence, we had to revert to amniotic fluid sampling. Amniotic fluid does not have established reference intervals for thyroid function tests and hence determining the normality of the result was based on reports from the published literature, albeit from other analyser platforms, and also in different trimesters. ¹¹ TRAbs, if present in amniotic fluid, are almost always of maternal origin and therefore likely pathologic and contributing to fetal thyrotoxicosis.

Treatment includes using methimazole or PTU, both known to cross the placenta.^4,12,13 We used PTU due to its known action of interfering with peripheral deiodination of T4 to T3, thereby acting faster to prevent further cardiac complications in fetus and premature delivery. Given short treatment duration, the risk of hepatotoxicity was deemed as low. The rate of fetal anomalies with both the drugs is similar at around 3% ¹² and risk is minimal in the third trimester. FHR was used as a guide to titrate PTU levels to ensure a balance between fetal hyperthyroidism and hypothyroidism. Maternal thyroxine was increased to prevent any reduction in maternal thyroid levels and allow adequate transfer of thyroxine to the fetus if required by iatrogenic fetal thyroid axis over suppression.

Improvement in the fetal goitre and FHR seemed to be reliable markers of fetal thyroid status as the neonate was born with relatively normal fT4 levels and suppressed TSH in cord blood at birth with no goitre. The neonate, however, had significant exophthalmos consistent with the exposure to maternal TRAbs. Neonatal blood on day 2 after birth showed marked hyperthyroidism explained by the decline in PTU levels after birth and persisting maternal TRAbs which was managed by carbimazole and propranolol. Treatment is often down-titrated until maternal TRAbs are eliminated from neonatal circulation ranging from 1 to 3 months. ²

Conclusion

Our case highlights the importance of checking TRAbs in pregnant women with past GD, especially if elevated more than 3 times the normal limit, even if well-controlled post-definitive therapy. It also demonstrates the challenges of diagnosing and managing fetal thyroid status using amniotic fluid assay and tertiary ultrasounds if cord sampling is not possible. Appropriate multidisciplinary team management involving an experienced endocrinologist, obstetrician or maternal–fetal medicine specialist is crucial.