Eye shades/sunglasses help alleviate photophobia and shield the eyes from dust and wind. Saline artificial tears can alleviate foreign body sensation, dry eyes, and the gritty sensation. Raising the head of the bed at night may be of some help to reduce periorbital edema in addition to diuretics. Elimination of controllable risk factors for TED progression such as smoking is critical, not only for the benefit of TED but also for the pregnancy. Nocturnal taping of the eyes can be helpful when lagophthalmos is present and prisms can be used to improve cases of mild diplopia. Understanding the natural history of the disease and providing reassurance is one of the most important therapeutic measures since the chance of progression of TED to more severe forms is very low [40]. All the above supportive measures can be safely implemented during pregnancy. Diuretic use to decrease swelling is debatable for routine treatment of TED.

Both hyper- and hypothyroidism can influence the course of TED. Patients with uncontrolled thyroid hormone levels are more likely to exhibit severe TED than euthyroid patients. Anti-thyroid drugs (anti-TDs) do not appear to have a negative influence on the course of TED. Medical treatment is usually followed by a fall in serum antibody concentrations, suggesting a waning of autoimmunity [43]. Since obstetrical and medical complications are directly related to thyroid status during pregnancy, patients should be euthyroid before conceiving. Anti-TDs are the mainstay of treatment during pregnancy. The American Thyroid Association recommends propylthiouracil (PTU) for the first trimester, due to the risk of methimazole (MMI) embryopathy [44]. Following the first trimester, patients on PTU should be switched to MMI in order to decrease the incidence of hepatotoxicity related to PTU. Thyroid function tests should be monitored monthly during pregnancy [15, 21].

Glucocorticoid therapy has been shown to reduce interferon gamma-induced HLA-DR expression in vitro [43–46]. Current recommendations for the treatment of acute and moderate-to-severe TED in adults is a course of 0.5 g of methylprednisone IV once weekly for 6 weeks, followed by 0.25 g/week for 6 weeks. If the patient fails to respond, treatment may be stopped after 6 weeks of 0.5 g/week dosing [47]. Currently there are no specific dosing recommendations for pregnancy. Glucocorticoids, however, have been widely used during pregnancy for the treatment of a variety of medical conditions. They are categorized as a class C pregnancy risk and should be given only if the potential benefits justify the potential risk to the fetus. Various animal studies showed that high doses of glucocorticoid can lead to an increased incidence of cleft palate in exposed offspring [48, 49]. However, several subsequent reports in which pregnant women were treated with glucocorticoids throughout the first trimester for various medical conditions did not reveal a consistent pattern of embryopathy [50]. Glucocorticoids do not appear to have major teratogenic potential [51]. Glucocorticoids during pregnancy, however, may increase the risk of premature rupture of the membranes and intrauterine growth restriction [52, 53]. Antenatal exposure to a relatively large dose of synthetic glucocorticoids is associated with suppression of the hypothalamic-pituitary-adrenal axis [54]. Other maternal adverse effects include pregnancy-induced hypertension, gestational diabetes, osteoporosis, and infection [51]. Therefore, when systemic glucocorticoids are needed in pregnancy, it is generally recommended to use the lowest effective dose for the shortest duration of time, avoiding high doses during the first trimester [55]. Glucocorticoids are metabolized in the placenta by a crucial enzyme, 11-beta-hydroxysteroid dehydrogenase-2. Prednisolone-related drugs (methylprednisone, prednisolone, and prednisone) are mostly metabolized to inactive forms by this enzyme and thus only approximately 10 % of the total dose compared to 33 % of betamethasone and 50 % of dexamethasone will enter the fetal circulation [56].

Prednisone and prednisolone have been measured in breast milk. Levels of prednisolone, the active form of prednisone measured in the milk, are typically less than 0.1 % of the total prednisone dose, even for a dose up to 80 mg. This corresponds to less than 10 % of an infant’s endogenous cortisol production [57] and therefore does not pose a clinically significant risk to a nursing infant. Nonetheless, exposure may be minimized if nursing is performed 3–4 h after a dose [57]. The American Academy of Pediatrics considers steroid therapy compatible with breastfeeding [57]. Guidelines from the National Heart, Lung and Blood Institute for the management of asthma also note that maternal use of systemic glucocorticoids is not a contraindication to breastfeeding [58].

In nonpregnant women, some studies have found that selenium may lower anti-TPOAb titers. The European Group on Graves’ Orbitopathy (EUGOGO) consortium conducted a trial to compare selenium use with placebo or pentoxifylline. Several ocular parameters (eye involvement (P = 0.01), progression of TED (P = 0.01) as well as quality of life outcome measures (P < 0.001)) improved in the selenium-treated group compared to placebo without significant side effects (see Chap. 6 for a full discussion) [59]. The selenium levels of the study population were marginally low and this may have biased in favor of a positive selenium treatment response [59]. Although selenium levels can be low in full-term pregnant women compared with nonpregnant women, the risk to benefit comparison does not support routine selenium supplementation during pregnancy [21]. Currently, there is an ongoing study on selenium supplementation in pregnancy where secondary outcomes will examine maternal and fetal risks.

Four large double-blind, placebo-controlled studies of somatostatin analogs have shown no significant effects on TED compared with placebo [60–63]. Somatostatin analogs are listed as a category B pregnancy drug. Since somatostatin analogs cross the placenta, potential adverse fetal outcomes must be considered. Several papers reported the safe and effective use of these analogs in pregnancy for various therapeutic indications [64–66].

Rituximab: Rituximab has been assigned a pregnancy category C by the FDA. Because human IgG is known to traverse the placenta, rituximab may potentially cause fetal B-cell depletion. Post-marketing data indicate that B-cell lymphocytopenia generally lasts less than 6 months in infants exposed to the medication in utero. It is only recommended for use in pregnant women when the drug is clearly necessary. Rituximab use during breastfeeding is also not recommended since the drug is excreted in breast milk.

Azathioprine: Azathioprine is a pregnancy category D drug. Azathioprine should not be given during pregnancy and breastfeeding.

In summary, rituximab and azathioprine are contraindicated in pregnancy for the treatment of TED.

Orbital irradiation has been shown to be moderately useful as monotherapy [40]. The main benefit is an improvement in motility [34]. Nonetheless, orbital radiotherapy is contraindicated in pregnancy.

Orbital decompression can be safely performed during pregnancy for severe CON unresponsive to a short course of systemic glucocorticoids or if sight is threatened by corneal exposure. If ophthalmopathy is severe but inactive, orbital decompression may also be performed. Reducing proptosis and decompressing the optic nerve can be achieved by a variety of decompression techniques (see Chaps. 11 and 12). One of the most feared complications of orbital decompression is postoperative diplopia. Extraocular muscle or eyelid surgery is often needed for full rehabilitation but can be delayed until after delivery.

Although orbital decompression is sometimes performed as first-line treatment for CON, most patients will still require glucocorticoid therapy; whereas less than half of patients treated with glucocorticoids initially will require orbital decompression [67]. Although it is preferable to operate on a pregnant patient during the second trimester [68, 69], emergencies require intervention regardless of gestational age.

Whether or not thyroidectomy affects the course of TED remains unsettled. In many studies [41, 70, 71], different thyroid surgeries (subtotal versus total thyroidectomy) seem to carry very low risk for progression of TED [40]. Thyroidectomy is not recommended as first-line therapy for hyperthyroid GD, regardless of pregnancy status, but it should be considered in the event of contraindications to anti-TD therapy [21]. If thyroidectomy is indicated during pregnancy, the optimal time to perform surgery is in the second trimester [21].

With regard to the need for oculoplastic referral, it is the authors’ opinion that because of the rare occurrence of TED during pregnancy and the limited therapeutic options, there should be a very low threshold for referral to oculoplastic specialists.

The frequency of thyrotoxicosis ranges from 0.1 in 100,000 in young children (i.e., prior to adolescence) to 3.0 in 100,000 in adolescents [7]. Although GD represents the most common cause of thyrotoxicosis, it only affects 0.02 % of all children [78]. GD comprises 10–15 % of thyroid disorders in patients under 18 years of age, is rare under the age of 5, has a peak incidence at 10–15 years of age, and more commonly affects female pediatric patients [75]. Mild signs and symptoms of TED, including lid retraction and lid lag, are common (25–60 %) in pediatric thyrotoxicosis [78]. However, severe TED is rarely seen in pediatric patients. Bartley et al. studied subjects in Olmsted County, Minnesota, and found that only 6/120 cases of TED occurred in patients under the age of 20 years [79]. The reported incidence rate of TED in male children is 0 per 100,000 for ages 5–9 and 15–19 years, and 1.7 per 100,000 for ages 10–14 years. The reported incidence rate of TED in female children is 3.5 per 100,000 for ages 5–9 years, 1.8 per 100,000 for ages 10–14 years, and 3.3 per 100,000 for ages 15–19 years [74, 79].

Pediatric GD may present with many of the same signs and symptoms as in adults, including heat intolerance (61 %), tremor (49 %), and ophthalmopathy (43 %) [80]. Additionally, children with GD may often present with decreased academic performance (50 %) [80]. Favorable predictors of long-term remission include age less than 12 years old, free thyroxine (FT4) level <35–50 pmol/L at diagnosis, initial response to therapy, TSH receptor antibody levels ≤4 times the upper limit of normal, Caucasian ethnicity, and having other autoimmune condition(s) [81, 82].

Thyrotoxicosis has deleterious physiologic effects on children, including low bone mineral density and increased bone turnover, advancement of bone age leading to a lower than predicted final height, diminished left ventricular reserve, thyrotoxic ophthalmologic signs, and cognitive/neuropsychiatric deficits (hyperactivity, irritability, anxiety, and impaired school attention). Correction of the thyrotoxicosis in a timely manner usually reverses the adverse physiologic effects of thyrotoxicosis. Thus, in most children there do not seem to be long-term sequelae as long as the hyperthyroidism is corrected without delay [83].

In adult GD, the standard of care is a 12–18-month trial of anti-TDs followed by definitive therapy (i.e., RAI or thyroidectomy) if remission is not achieved. In contrast, pediatric GD is commonly treated for years with anti-TDs – typically methimazole – as long as the patient tolerates the therapy and euthyroidism can be achieved. Propylthiouracil (PTU) is avoided in the treatment of hyperthyroidism in children due to reports of PTU-induced liver failure (also discussed in chapter 1) [54]. The preference for medical management arises from the theoretical fear of radioiodine-induced thyroid malignancy, as well as the fear of potential morbidity and mortality associated with thyroidectomy [75, 80, 84]. After discontinuation of anti-TD therapy, over 50 % of pediatric patients will remain in remission for at least 2 years; however, relapse rates remain high and on the order of 36–47 % [75, 85–88]. Sustained disease remission for over 2 years post anti-TD therapy is less common in children (15–30 %) compared to adults (40–60 %), and may take up to 10 years to achieve [75, 83, 84, 88]. Gruniero-Papendieck et al. reported that in children who did not receive definitive treatment, the rate of persistent GD at 10 years was 31 % [77]. Up to 28 % of pediatric patients on anti-TD therapy experience adverse effects, and severe complications such as agranulocytosis, aplastic anemia, hepatotoxicity, and mortality have been reported, for PTU in particular [83].

Thyroidectomy and radioiodine represent definitive treatment options. Sherman et al. retrospectively investigated 78 patients under 18 years of age who underwent thyroidectomy for GD. The surgical indication for more than half the patients was failed medical therapy (60 %). Other indications included intolerance to anti-TDs (9 %), severe thyrotoxicosis, severe TED (concern for worsening with RAI), significant thyromegaly (gland >30 g), suspicious/malignant thyroid nodule, and unwillingness to receive RAI. Postoperative morbidity was transient and consisted of hypocalcemia (96 %) and unilateral recurrent laryngeal nerve palsy (1 %). It should be noted that the surgery was performed by experienced thyroid surgeons in a high-volume center [80]. Additional indications leading to thyroidectomy over RAI included age less than 10 years (due to theoretical higher risk of thyroid malignancy with RAI in this age group) and consideration of pregnancy in late adolescent females (TRAb levels can transiently increase and persist for at least 5 years after ablation, thus increasing risk of fetal/neonatal GD in the offspring) [75, 89–91].

Although radiation-induced thyroid malignancy as evidenced by the Chernobyl nuclear accident is a concern, retrospective studies suggest a very low risk of malignancy associated with RAI therapy. Patients with GD who are over 10 years old can be treated with RAI with the consultation of a pediatric ophthalmologist prior to ablation [75, 91]. Published retrospective studies with up to 36 years of follow-up have not supported the concerns of increased thyroid malignancy post RAI in the pediatric population [90–92]. In a retrospective review of the medical records of 78 pediatric patients who underwent thyroidectomy at Mayo Clinic, 4 patients (5 %) were incidentally found to have thyroid malignancies on histologic examination [80]. Thus, pediatric patients may harbor occult thyroid malignancy irrespective of RAI therapy.

The definition of TED with respect to adults has been reviewed elsewhere in this textbook. In children, Antoniazzi et al. defined TED as clinical signs of ocular involvement in patients with GD (lid retraction, lid lag, or stare), who also have exophthalmos or signs of enlarged extraocular muscles on orbital MRI [93].

In a 2004 review by Wiersinga, 23 % of pediatric patients with GD were found to have TED, which is very similar to that seen in adult patients with GD (18 %) [74]. However, TED prevalence rates up to 39 % have been reported in childhood GD [94]. A few small studies have attempted to further characterize the epidemiology of pediatric TED and to identify risk factors for the development of clinically significant TED. Durairaj et al. found that 10.0 % (2/20) of pediatric patients with TED had a family history of TED, and 60.0 % (12/20) had a family history of thyroid dysfunction [95]. Young, however, did not find any specific HLA Ag to be prevalent in patients with TED versus patients without TED [96]. Thus, currently there are no reliable genetic markers for identifying which pediatric patients with GD are at increased risk of developing TED.

According to Gogakos and coauthors, although severe TED is less common in children than adults, the epidemiology overall is similar [102]. Furthermore, TED has a similar female-to-male predominance in both children and adults (87 % and 83 %, respectively), and is associated with more severe thyrotoxicosis at the time of diagnosis in all ages. There is often a 2-year lag between the onset of thyroid dysfunction and TED, and a notable family history of thyroid disease in both children and adults. In late adolescence, the clinical features of TED resemble those of the adults [74, 97]. A 2005 European survey of 23 pediatric and 44 adult endocrinologists suggested that smoking was a potential risk factor for the development of TED, as countries with higher smoking prevalence had a higher prevalence of TED among children and adolescents with GD [84]. Although it is not possible to draw conclusions about causality from a survey study, in adults, smoking is an established risk factor for TED (see Chap. 5). Wiersinga postulates that the lower frequency of smoking in children may at least in part explain the tendency for milder clinical manifestations in pediatric TED [74].

Two studies have examined the severity of TED in a total of 163 prepubertal versus postpubertal children and adolescents with GD. Holt et al. did not find evidence of restrictive strabismus in prepubertal subjects. Severe TED seemed to occur less frequently in prepubertal subjects, but the number of subjects in that group was too small to achieve statistical significance (total N = 41; subjects with TED N = 6) [98]. In their prospective study of pediatric patients with GD followed by orbital MRI, Antoniazzi et al. reported a better prognosis in subjects who were prepubertal at the time of diagnosis. The authors hypothesized that increasing orbital volume with advancing age and puberty serves as a physiologic decompression [93].