Abstract
Graves’ disease in pregnancy may be associated with maternal, fetal and neonatal complications, which are proportionate to the severity of hyperthyroidism. Optimal management is detailed preconception counselling, achievement of an euthyroid state prior to conception, and close monitoring of thyroid function and thyroid-stimulating antibodies together with judicious use of anti-thyroid medications during pregnancy. A case of Graves’ disease in pregnancy, complicated by pancytopenia, with a deterioration in thyroid function following cessation of thionamide therapy is described here. Therapeutic plasma exchange was subsequently used to achieve rapid control prior to thyroidectomy. Therapeutic plasma exchange is an effective treatment for hyperthyroidism where thionamides are ineffective or contraindicated, as a bridge to definitive management.
Introduction
Graves’ disease (GD) complicates approximately 0.2% of pregnancies and may be associated with maternal, fetal and neonatal complications. 1 As the majority of women are diagnosed prior to pregnancy, emphasis should be placed on detailed preconception counselling, and in general pregnancy should be postponed until an euthyroid state is achieved. 2 The mainstay of management of GD in pregnancy is close monitoring of mother and fetus with judicious use of thionamide therapy. 2 Maternal risks of poorly controlled GD include pre-eclampsia, cardiac failure, arrhythmias, pulmonary hypertension, periodic paralysis and rarely thyroid storm.3,4 Maternal complications of thionamide therapy include rash, cytopenias, hepatotoxicity and vasculitis; 5 their development may require cessation and definitive management. Teratogenicity may occur due to thionamide exposure between three and eight weeks of gestation. Poorly controlled hyperthyroidism in pregnancy is associated with increased risks of miscarriage, stillbirth, pre-term delivery and low birth weight.3,4,6 Transplacental passage of thyroid-stimulating hormone receptor antibodies may cause fetal and neonatal hyperthyroidism. 7
Case
A 33-year-old Korean woman in her first pregnancy presented to the antenatal clinic at eight weeks of gestation. GD was diagnosed six years earlier. Attempts to wean and cease thionamide therapy prior to pregnancy resulted in relapses of hyperthyroidism, and the woman declined surgery or radioactive iodine. The woman had been euthyroid for more than 12 months while taking propylthiouracil 50 mg twice daily when she conceived (Figure 1). The dose of propylthiouracil was increased to 100 mg twice daily at four weeks of gestation when she became hyperthyroid with free thyroxine (FT4) of 26 pmol/L (first trimester reference range 9–25 pmol/L; Mater Hospital Pathology), free triiodothyronine (FT3) of 9.2 pmol/L (2.6–6.0 pmol/L) and thyroid-stimulating immunoglobulin of 1.7 IU/L (<0.1 IU/L). The dose was increased further to 100 mg three times daily at eight weeks of gestation as she remained hyperthyroid with FT4 24 pmol/L and FT3 8.6 pmol/L. At 15 weeks of gestation she was biochemically euthyroid (FT4 15.2 pmol/L, FT3 5.7 pmol/L) and propylthiouracil was changed to carbimazole 10 mg twice daily to reduce the risk of hepatotoxicity associated with propylthiouracil exposure. At 16 weeks of gestation cytopenias were noted with white cell count (WBC) 3.3 × 109/L, neutrophil count (N) 2.1 × 109/L and haemoglobin (Hb) 9.7 g/dL. The woman’s full blood count had been normal for at least two years prior to this. Carbimazole was ceased at 20 weeks of gestation due to progressive pancytopenia with WBC 2.8 × 109/L, N 1.65 × 109/L Hb 8.6 g/dL and platelet count (Plt) 92 × 109/L.
Trend of FT4, FT3, TSH receptor antibodies, FBC, WBC and Plt.
Investigations revealed normal active B12, red blood cell folate, immunoglobulins and inflammatory markers, absence of haemolysis, negative serology for parvovirus, cytomegalovirus, Epstein-Barr virus and human immunodeficiency virus. Ultrasound evaluation revealed a heterogenous and hypervascular thyroid. Prednisolone 1 mg/kg/day, propranolol and cholestyramine were commenced aiming to prevent recurrence of hyperthyroidism with thionamide cessation. Over the following two weeks FT4 increased from 17.0 to 44.9 pmol/L and FT3 from 7.8 to 28.1 pmol/L. In order to rapidly achieve euthyroidism and minimise risk of thyroid storm prior to planned thyroidectomy, in the setting of thionamide therapy being a potential contributor to cytopenias, she underwent therapeutic plasma exchange (TPE) prior to the operation. One plasma volume (plus an addition 40% to account for pregnancy-related increased circulatory volume) was exchanged with fresh frozen plasma via peripheral intravenous cannulae daily for four consecutive days before surgery. Red cell and platelet transfusions were performed at the end of procedure with a target haemoglobin of at least 90 g/L and platelet count of at least 50 × 109/L. This resulted in a 45.7% fall in fT4 and a 64.8% fall in FT3. Her thyroid hormones and antibodies pre- and post-plasma exchange (measured on commencement and at the immediate conclusion of each exchange session) are summarised in Table 1.
Effect of four sessions of TPE on thyroid hormones and antibodies.
TPE: therapeutic plasma exchange; FT4: free thyroxine; FT3: free triiodothyronine; TSH; thyroid-stimulating hormone.
aTSH Receptor Abs: Negative < 0.10 IU/L; equivocal 0.10–0.55 IU/L; positive > 0.55 IU/L.
Total thyroidectomy was performed at a gestational age of 22 weeks and 2 days under general anaesthesia with no complications. Histopathology of the operative specimen confirmed GD. Cytopenias initially improved; however, a subsequent deterioration in haematological parameters at 26 weeks (WBC 1.9 × 109/L, Hb 8.1 g/dL and Plt 89 × 109/L) prompted bone marrow examination revealing myelodysplasia with excess blasts (MDS-EB-1). Within three weeks, progressive worsening of cytopenias and circulating blasts were noted, prompted repeat bone marrow biopsy which demonstrated MDS-EB-2, indicating rapid progression and transformation to acute myeloid leukaemia (AML) within weeks was anticipated. Lower segment caesarean section was performed at 31 weeks and 2 days of gestation and a live female weighing 1730 g was delivered in good condition. There were no maternal or neonatal complications in the early postpartum period. Shortly after discharge, she commenced chemotherapy for AML.
Discussion
Single lineage haematological cytopenias are not uncommonly observed in individuals with untreated GD; anaemia, leucopenia or thrombocytopenia occur in approximately 33%, 5.8% and 3% of women, respectively.8–12 Pancytopenia with untreated GD is rare; only 30 cases have been described in the literature, all in the setting of severe hyperthyroidism with markedly elevated FT4 and FT3 levels. 13 Bone marrow examination revealed variable findings with hypercellular, hypocellular and normal examinations described. Thionamides were used to manage hyperthyroidism in 80% of cases. Almost all women had resolution of pancytopenia on restoration of euthyroid state. American Thyroid Association guidelines however recommend against using antithyroid drugs if the baseline neutrophil count is less than 1 × 109/L at diagnosis of GD (weak recommendation). 14
Thionamide therapy can lead to agranulocytosis in 0.1–0.5% and pancytopenia in 0.01% of individuals treated. 15 Overlapping autoimmune disease including pernicious anaemia, autoimmune thrombocytopenic purpura and autoimmune haemolytic anaemia should also be considered as potential cause of cytopenias in individuals with GD. Four cases of AML have been described in the setting of long-term propylthiouracil therapy, with no causal evidence established. 16 In the case presented the development of pancytopenia following a recent increase in dose of thionamides, with a resulting euthyroid state, in the setting of normal haematological parameters while on lower dose thionamide in the past, compelled cessation of medication while alternative causes of pancytopenia were excluded. In retrospect it is regrettable that a bone marrow examination was not performed earlier in the woman’s course.
Therapeutic options for GD in non-pregnant individuals awaiting definitive management in the setting of thionamide-intolerance include potassium iodide (KI), corticosteroids, cholestyramine and perchlorate. KI use in first trimester has demonstrated significantly lower risk of congenital anomalies compared to continuation on methimazole (1.53 vs. 4.14%); however, hyperthyroidism worsened in 9.2% without any clear predictors. 17 American Thyroid Association guidelines state KI treatment for GD in pregnancy cannot be recommended outside of Japan (a region of high iodine intake) until more evidence on safety and efficacy is available. 2 In our case, oral iodide therapy was considered but not used because of concerns regarding lack of efficacy in rapidly progressive thyrotoxicosis together with potential exacerbation without thionamide therapy. Several small studies in non-pregnant subjects demonstrated rapid improvement in thyroid function with corticosteroid monotherapy in GD, including pre-operative preparation for thyroidectomy in the setting of thionamide intolerance.18–22 Cholestyramine binds to iodothyronines reducing their enterohepatic circulation and several studies have demonstrated more rapid decline in hyperthyroidism when added to thionamides compared to thionamides alone. 23 The maternal, fetal and neonatal effects of perchlorate during pregnancy are unknown. 24
TPE treatment has been described in 91 non-pregnant individuals with GD.25–30 Use of TPE for hyperthyroidism in pregnant women31–33 or during molar pregnancies34–36 has only been rarely reported. No controlled trial has been performed. The American Society for Apheresis guidelines recommend TPE as category II (disorder for which apheresis is accepted as second-line therapy, either as a standalone treatment or in conjunction with other modes of treatment) with evidence grade 2 C (weak recommendation, low-quality or very low-quality evidence) for the management of thyroid storm. 37 In the largest retrospective case series of 40 non-pregnant individuals with GD, a median of four TPE treatments for 2.5–3 h each resulted in a 44.8% fall in FT4 and 61.2% fall in FT3. 27 TPE clears protein-bound thyroid hormones, aided by the extensive protein binding which T4 and T3 undergo in the bloodstream. 38 The process results in an effective decrease in total thyroid hormones which is coupled with new binding sites from the colloid exchange product, and after some time there is a newly established equilibrium resulting in a fall in free thyroid hormone levels. It is postulated to also help to clear circulating thyroid-stimulating hormone receptor antibodies, cytokines and deiodinase. 39 As the effect of TPE on hyperthyroidism is transient, treatment should be followed promptly by definitive therapy. TPE has been used in a variety of conditions in pregnancy, particularly thrombotic microangiopathies, antiphospholipid syndrome, hypertriglyceridemia, vasculitides and myasthenia gravis. Overall complication rates from TPE may be as high as 10%, including electrolyte abnormalities such as hypocalcaemia. Severe complications including cardiorespiratory events and anaphylaxis occur in less than 1%. 40
Conclusion
GD in pregnancy requires detailed pre-conception counselling and ideally achievement of an euthyroid state prior to conception. During pregnancy close maternal and fetal monitoring is required, and hyperthyroidism is usually effectively managed with anti-thyroid medications after discussion of risks. Pancytopenia is a rare complication of untreated Grave’s disease with severe hyperthyroidism, and of thionamide therapy. Non-thyroid–related causes should be considered when pancytopenia occurs with GD. In rare situations where thionamides are contraindicated or ineffective, TPE is an effective treatment to reduce thyroid hormone levels but should be followed by definitive management.
Footnotes
Declaration of conflicting interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.
Ethical approval
The Mater Health Human Research Ethics Committee waived approval for this publication.
Informed consent
Written informed consent was obtained from the patient for their anonymized information to be published in this article.
Guarantor
AM is the guarantor of the present work.
Contributorship
ML, AM, MY and JL cared for the patient and wrote the manuscript.
Acknowledgements
None.
