Abstract
Pregnant women with Graves’ disease (GD) who have undergone thyroidectomy or radioactive iodine therapy can have high levels of thyroid-stimulating hormone (TSH) receptor antibodies, which are transferred to the fetus via the placenta, posing a risk for fetal GD. This retrospective observational study, conducted at two high-level perinatal medical centers in Tokyo and Osaka, Japan, aimed to identify predictors of fetal GD in pregnant women with GD who had undergone thyroidectomy or radioactive iodine therapy. In total, 65 women were included, and 79 singleton pregnancies and fetuses were analyzed. Fetal GD occurred in 17.7% of the 79 fetuses. Women in the fetal GD group had higher levels of TSH receptor antibodies and a higher prevalence of ophthalmopathies than did women in the non-fetal GD group. The receiver operating characteristic curve cutoff values of maternal TSH-binding inhibitory immunoglobulin (hereafter referred to as TRAb [TSH receptor antibody from a narrow perspective]) and thyroid-stimulating antibody (TSAb) levels predictive of fetal GD development were as follows: TRAb, 12.8 and 10.2 IU/L at 10 and 20 gestational weeks (GW), respectively; TSAb, 975.4% and 1,259.0% at 10 and 20 GW, respectively. Ophthalmopathy was a predictor of fetal GD; nonetheless, combining the ophthalmopathy and TRAb cutoff values did not improve predictive accuracy. A cutoff value of TRAb ≥10.2 IU/L at 20 GW (highest diagnostic accuracy found) could be a predictor of fetal GD risk for pregnant women with GD who undergo thyroidectomy or radioactive iodine therapy; thus, appropriate fetal monitoring should begin at around 20 GW.
Introduction
Graves’ disease (GD) is common in women of childbearing age, occurring in approximately 0.3% (0.1%–0.4%) of pregnant women [1]. The thyroid-stimulating hormone (TSH) receptor antibodies responsible for GD can cross the placenta; therefore, if maternal TSH receptor antibody levels remain elevated past 18–22 gestational weeks (GW), then when the fetal pituitary-thyroid system has fully developed, these maternal antibodies may stimulate the fetal thyroid, leading to fetal GD [2, 3].
If pregnant women with GD have no history of definitive therapy (thyroidectomy or radioactive iodine therapy), antithyroid drugs or inorganic iodine use can serve as thyroid therapy for the fetus by transfer via the placenta. Conversely, mothers with GD with a history of definitive therapy can have high TSH receptor antibody levels even if maternal thyroid function is normal or decreased [4-8]. In such cases, fetal GD can develop after 18 GW [9], regardless of maternal thyroid function, due to TSH receptor antibodies migrating through the placenta and stimulating the fetal thyroid gland. If not properly treated, fetal GD can lead to serious complications, including heart failure, restricted intrauterine growth, low birth-weight, premature birth, intrauterine death, accelerated bone maturation, and poor long-term mental development [10-12].
TSH-binding inhibitory immunoglobulin (TBII) and thyroid-stimulating antibody (TSAb) have been measured as TSH receptor antibodies (from a broad perspective) for the diagnosis of GD. The former assesses the inhibition of TSH receptor binding, and it has been referred to as TRAb (TSH receptor antibody from a narrow perspective) in this study. The latter evaluates the thyroid-stimulating activity. The 2017 American Thyroid Association [2] and 2018 European Thyroid Association guidelines [13] recommended fetal thyroid function monitoring for all mothers with GD and a TRAb level >5 IU/L or TSAb more than thrice the baseline level, either during early pregnancy stages or at 18–22 GW, irrespective of definitive therapy history. However, TSH receptor antibody thresholds for assessing the fetal GD risk in these guidelines are based on studies on neonatal GD and second or third trimester TSH receptor antibody levels in patients with GD, regardless of whether fetal GD occurred after definitive therapy [14, 15]. Therefore, these thresholds are not based on predictive outcomes for fetal GD onset in women with GD and a history of definitive therapy.
While several reports on the neonatal GD risk in cases of pregnancies with GD and a history of definitive therapy exist [4, 16], no studies have predicted the onset of fetal GD, which requires intrauterine treatment. Given that the availability of perinatal facilities capable of appropriately monitoring fetal GD is limited, appropriately screening high-risk cases, monitoring the fetus throughout pregnancy, and referring patients to facilities capable of multidisciplinary management and treatment are essential. Therefore, this study aimed to determine indicators to identify pregnant women with GD and a history of definitive therapy who are at high risk of fetal GD development.
Materials and Methods
We retrospectively reviewed the medical information, including clinical examination data, of pregnant women with GD and their fetuses who were consecutively treated between March 2002 and August 2021 at the National Center for Child Health and Development, Tokyo, Japan, and between November 2015 and April 2019 at Osaka Women’s and Children’s Hospital, Osaka, Japan. We enrolled pregnant women with GD and a history of definitive therapies (specifically, thyroidectomy before or during pregnancy or radioactive iodine therapy before pregnancy).
The eligibility criteria were singleton cases with at least one measurement of TSH receptor antibody levels—TRAb or TSAb—during pregnancy; patients in remission before pregnancy, with insufficient information, or with miscarriages were excluded. Maternal comorbidities, such as diabetes, hypertension, and malignancy, could not be evaluated. We defined cases in remission before pregnancy as negative for TSH receptor antibodies immediately before and during pregnancy, and taking no medications, such as antithyroid drugs or inorganic iodine, before and during pregnancy.
Skilled obstetricians performed fetal monitoring using transabdominal fetal ultrasonography every 2 weeks after 20 GW. We defined fetal GD as clinically evident fetal GD [17], in which the fetus was judged to have fetal goiter [18, 19], fetal tachycardia (fetal heart rate >160 bpm), or signs of heart failure or accelerated bone maturation [20], with intensive maternal treatment with antithyroid drugs or inorganic iodine regardless of indications. Neonatal GD requiring drug treatment within 5 days of birth is also considered fetal GD, even if signs of fetal GD were not present earlier.
The following factors were considered potential predictors of fetal GD development: maternal age; pre-pregnancy body mass index (BMI); smoking during pregnancy; primipara; maternal age at GD onset; time between GD onset and pregnancy; previous maternal definitive therapy; maternal antithyroid drugs or inorganic iodine indication before pregnancy and up to 20 GW; levothyroxine administration in early pregnancy; and the presence or absence of GD ophthalmopathy. Ophthalmopathy was defined as clinically evident ophthalmopathy symptoms, such as eye protrusion, eye movement disorder, eyelid swelling, conjunctival hyperemia, and edema, based on medical records.
We considered maternal TSH receptor antibody levels during pregnancy as the main relevant predictors of fetal GD development. TSH receptor antibody activity was measured by radioreceptor (TBII, TRAb) and biologic stimulation (TSAb) assays. TRAb values were measured mainly with the DYNOtest TRAB Human kit based on a second-generation method (Yamasa, Tokyo, Japan; manufacturer reference limit, <1.0 IU/L) and partly with the ECLusys TRAb kit based on a third-generation method (Roche Diagnostics, Germany; manufacturer reference limit, <2.0 IU/L). Given that TRAb values derived using second- and third-generation methods are considered close approximations, both TRAb values were used. TSAb values were measured mainly with an enzyme immunoassay (EIA)-based TSAb kit (Yamasa, Tokyo, Japan; manufacturer reference limit, <120%). Prior to June 2014, TSAb levels were measured with a radioimmunoassay (RIA) kit; therefore, we converted the RIA TSAb values to EIA TSAb (%) values using the formula according to the manufacturer’s instructions: TSAb (%) = 1.06 × TSAb (%) (RIA) + 43.7. If the test value was outside the assay sensitivity range, it was approximated by an upper or lower limit, e.g., a TRAb value <2.0 IU/L was considered 1.999 IU/L. Maternal TSH receptor antibody levels were categorized by measurement time-point: 10, 20, and 30 GW, when measured from 0 to <16 GW, 16 to <28 GW, and 28 GW to delivery, respectively.
This study was approved by the Ethics Committees of the National Center for Child Health and Development, Tokyo, Japan (2021-166) and Osaka Women’s and Children’s Hospital, Osaka, Japan (899). Consent was obtained via an opt-out method, and details of data use were posted in the hospital and on its website.
Data are presented as median (interquartile range) or n (%) values unless otherwise indicated. JMP version 17 (SAS Institute, Inc., Cary, NC, USA) was used for statistical analysis, and GraphPad Prism (version 9) was used for graphing. Wilcoxon’s rank-sum test and the Chi-square test compared continuous and categorical variables, respectively. Logistic regression and receiver operating characteristic (ROC) curve analyses examined associations between fetal GD and predictors, and determined the optimal cutoff values for predicting fetal GD development. Harrell’s C-statistic assessed whether the discriminative ability was improved by combining the ophthalmopathy and TSH receptor antibody cutoff values (95% confidence intervals [CIs] were calculated by bootstrapping) [21]. The significance level was set at p < 0.05.
Results
We initially enrolled 86 women with GD (112 singleton pregnancies) who had undergone thyroidectomy or radioactive iodine therapy from consecutively enrolled pregnant women with GD. In the final analysis, 65 women (79 pregnancies and fetuses) were included (fetal GD group: n = 14; non-fetal GD group: n = 65; fetal GD incidence: 17.7%) (Fig. 1).
No significant between-group differences existed in maternal age, pre-pregnancy BMI, smoking history, age at GD onset, or duration between GD onset and pregnancy (Table 1). The primipara and ophthalmopathy rates were significantly higher in the fetal GD group than in the non-fetal GD group (primipara: 85.7% vs. 46.2%; p < 0.01; ophthalmopathy: 64.3% vs. 20.0%, p < 0.01).
Table 1 Background maternal characteristics and pregnancy outcomes in the fetal Graves’ disease and no fetal Graves’ disease groups
| Variable |
All
n = 79 |
Fetal Graves’ disease group
n = 14 |
No fetal Graves’ disease group
n = 65 |
p value |
| Age (years), median |
35.0 (33.0–38.0) |
35.0 (33.0–37.0) |
35.0 (33.0–38.0) |
0.512 |
| Primipara, n (%) |
42 (53.2) |
12 (85.7) |
30 (46.2) |
<0.01 |
| BMI before pregnancy (kg/m2), median |
20.9 (18.8–22.9) (n = 78) |
19.5 (18.3–23.3) |
21.0 (19.1–23.0) (n = 64) |
0.266 |
| Pre-pregnancy smokinga, n (%) |
12 (15.2) |
3 (21.4) |
9 (13.4) |
0.437 |
| Smoking, n (%) |
4 (5.1) |
1 (7.1) |
3 (4.6) |
0.549 |
| Age of onset of Graves’ disease (years), median |
26.0 (21.0–31.0) |
29.0 (20.3–32.3) |
26.0 (20.5–31.0) |
0.662 |
| Duration of disease until pregnancy (years), median |
6.0 (3.0–12.0) |
3.5 (3.0–14.5) |
7.0 (4.0–11.5) |
0.247 |
| Duration from previous definitive treatments to pregnancy (years), median |
3.4 (1.2–6.7) |
2.2 (1.1–6.3) |
4.0 (1.3–6.8) |
0.624 |
| Ophthalmopathyb, n (%) |
22 (27.8) |
9 (64.3) |
13 (20.0) |
<0.01 |
| Previous definitive treatments (including during pregnancy) |
44 (55.7) |
11 (78.6) |
33 (50.8) |
0.133 |
| Thyroidectomy only, n (%) |
20 |
8c |
12 |
|
| Total thyroidectomy, n (1 case during pregnancy) |
24 |
3 |
21c |
|
| Subtotal thyroidectomy, n (1 case during pregnancy) |
34 (43.0) |
3 (21.4) |
31 (47.7) |
|
| Radioactive iodine therapy only, n (%) |
1 (1.3) |
0 (0.0) |
1 (1.5) |
|
| Subtotal thyroidectomy and radioactive iodine therapy, n (%) |
|
|
|
|
| Anti-thyroid drugs or inorganic iodine before pregnancy, n (%) |
17 (21.5) |
4 (28.6) |
13 (20.0) |
0.486 |
| Methimazole, n |
7 |
2 |
5 |
|
| Propylthiouracil, n |
5 |
1 |
4 |
|
| Inorganic iodine, n |
4 |
0 |
4 |
|
| Methimazole and inorganic iodine, n |
1 |
1 |
0 |
|
| Levothyroxine therapy before pregnancy, n (%) |
50 (63.3) |
10 (71.4) |
40 (61.5) |
0.486 |
| Levothyroxine, n |
47 |
8 |
39 |
|
| Levothyroxine and methimazole, n |
3 |
2 |
1 |
|
| Levothyroxine dosage before pregnancy (μg/day), median |
100.0 (75.0–125.0) (n = 49) |
100.0 (71.9–125.0) (n = 10) |
100.0 (75.0–125.0) (n = 39) |
0.829 |
| Anti-thyroidal drugs or inorganic iodine use (pre-pregnancy to 20 weeks) by maternal indication, n (%) |
19 (24.0) |
5 (35.7) |
14 (21.5) |
0.306 |
| Weeks of labor (weeks), median |
39.3 (38.3–40.1) |
37.9 (37.3–39.5) |
39.6 (38.6–40.1) |
<0.05 |
| Delivery method, n (%) |
|
|
|
0.918 |
| Spontaneous vaginal delivery |
45 (57.0) |
9 (64.3) |
36 (55.4) |
|
| Caesarean section |
20 (25.3) |
3 (21.4) |
17 (26.1) |
|
| Forceps delivery, vacuum extraction |
12 (15.2) |
2 (14.3) |
10 (15.4) |
|
| Unknown |
2 (2.5) |
0 (0.0) |
2 (3.1) |
|
| Female sex, n (%) |
44 (57.1) (n = 77) |
8 (57.1) |
36 (57.1) (n = 63) |
1.000 |
| Birthweight (g), median |
2988.0 (2660.0–3310.0) |
2801.0 (2394.5–3425.5) |
3046.0 (2709.0–3301.0) |
0.190 |
Data are presented as median (interquartile range) values or n (%) unless otherwise indicated. Abbreviation: BMI, body mass index
aPre-pregnancy smoking was defined as “no smoking” if the woman had never smoked or had stopped before pregnancy and as “smoking” if the woman had stopped smoking after finding out she was pregnant or was still smoking at the time of enrollment.
bOphthalmopathy was defined as clinically evident signs of ophthalmopathy, such as eye protrusion, eye movement disorder, eyelid swelling, conjunctival hyperemia, or edema. Patients in whom ophthalmopathy was not noted in the medical records were classified as having no ophthalmopathy.
cOne patient underwent thyroidectomy during pregnancy.
There was no significant difference between the two groups in the percentage of mothers who underwent thyroidectomy or radioactive iodine therapy. Two mothers underwent thyroidectomy during pregnancy—one underwent total thyroidectomy at 21 GW (fetal GD group), and the other underwent subtotal thyroidectomy at 22 GW (non-fetal GD group). There was no significant difference between mothers who were taking antithyroid drugs, inorganic iodine, or levothyroxine before pregnancy or by maternal indication up to 20 GW.
Supplementary Table 1 shows the pregnancy and postpartum courses of the 14 mothers in the fetal GD group. They were diagnosed with GD between 12–34 years of age and gave birth between 30–39 years of age. Fetal GD was diagnosed in 11 cases during the fetal period (19.6–36.4 GW), after which the fetus was carefully monitored or treated with antithyroid drugs or inorganic iodine administered to the mother. Among these cases, three neonates did not require medication with antithyroid drugs or inorganic iodine (Cases 4, 7, and 10). No mother with GD showed signs of fetal edema or heart failure during pregnancy. Neonatal GD necessitating medications developed in the neonates of three mothers (Cases 8, 11, and 13) without clinically evident signs of fetal GD after total or subtotal thyroidectomy, as well as the neonate of one mother (Case 12) who did not require treatment during pregnancy for fetal GD after total thyroidectomy. The neonates of cases 8,12, and 13 showed TSH suppression or/and elevated free thyroxine in umbilical cord blood or/and peripheral blood on day 0 after birth, while the neonate of case 11 showed accelerated bone maturation findings on X-rays taken on day 0 after birth. One mother after radioactive iodine therapy had markedly elevated TRAb levels during both pregnancies (Cases 9 and 14).
Table 2 shows the maternal TSH receptor antibody levels at 10, 20, and 30 GW. Both maternal TRAb and TSAb levels were significantly higher in the fetal GD group than in the non-fetal GD group at all time points.
Table 2 TRAb and TSAb values at 10, 20, and 30 gestational weeks in the fetal Graves’ disease and no fetal Graves’ disease groups
| Variable |
Fetal Graves’ disease group (n = 14) |
No fetal Graves’ disease group (n = 65) |
p value |
| TRAb (IU/L) |
| 10 gestational weeks |
41.5 (12.9–76.0) (n = 10) |
4.3 (1.5–11.0) (n = 47) |
<0.01 |
| 20 gestational weeks |
39.6 (11.4–66.5) (n = 14) |
3.0 (1.0–7.1) (n = 40) |
<0.0001 |
| 30 gestational weeks |
12.5 (9.6–62.4) (n = 14) |
2.0 (1.0–4.8) (n = 53) |
<0.0001 |
| TSAb (%) |
| 10 gestational weeks |
3480.4 (1133.0–4126.4) (n = 10) |
355.2 (184.9–1256.6) (n = 36) |
<0.001 |
| 20 gestational weeks |
2681.7 (1567.1–3292.5) (n = 12) |
198.5 (144.5–544.3) (n = 41) |
<0.0001 |
| 30 gestational weeks |
2120.5 (905.1–3459.7) (n = 14) |
206.4 (174.2–363.8) (n = 34) |
<0.0001 |
Data are presented as median (interquartile range) values.
Abbreviations: TRAb, TSH-binding inhibitory immunoglobulin (TBII); TSAb, thyroid-stimulating antibody
The optimal maternal TSH receptor antibody level cutoff values and corresponding ROC areas under the curve (AUCs) for predicting fetal GD development were evaluated (Fig. 2). These cutoff values were: TRAb 12.8 IU/L (AUC, 0.860) and TSAb 975.4% (AUC, 0.892) at 10 GW, and TRAb 10.2 IU/L (AUC, 0.954) and TSAb 1,259.0% (AUC, 0.945) at 20 GW, indicating both TRAb and TSAb cutoff values were extremely useful in predicting fetal GD development. Maternal TRAb and TSAb values at 20 GW had higher AUCs than those at 10 GW. Levels at 30 GW were not analyzed because fetal GD onset had already occurred.
The diagnostic accuracy for predicting fetal GD development using the maternal antibody cutoff values at each gestational age was as follows—sensitivity and specificity, respectively: 90.0% and 76.6% using TRAb 12.8 IU/L at 10 GW; 100.0% and 72.2% using TSAb 975.4% at 10 GW; 100.0% and 90.0% using TRAb 10.2 IU/L at 20 GW; and 91.7% and 87.8% using TSAb 1,259.0% at 20 GW. The best diagnostic accuracy was obtained for TRAb 10.2 IU/L at 20 GW. Scatter plots of maternal TRAb values, and maternal TRAb and TSAb values at 20 GW are shown in Figs. 3, 4 and the Graphical Abstract, respectively. After combining these values, the sensitivity and specificity of the accuracy of predicting fetal GD onset were 91.7% and 97.0%, respectively, indicating decreased sensitivity but increased specificity. Thus, the combination of maternal TRAb and TSAb cutoff values at 20 GW could be a good predictor of fetal GD onset. We obtained maternal TRAb and TSAb cutoff values for predicting the development of fetal GD by excluding two cases in which the mother underwent thyroidectomy at 21 and 22 GW, and the data showed similarities for both 10 and 20 GW (data not shown). The changes in TRAb and TSAb levels in the two mothers who underwent thyroidectomy during pregnancy are shown in Supplementary Table 2. In both these cases, a decrease in maternal TRAb and TSAb levels appeared to be similar to the decrease caused by the general course of pregnancy; however, there could have been a further decrease as a result of the thyroidectomy.
Graphical Abstract
Prediction of the development of fetal Graves’ disease
We examined factors other than maternal TRAb and TSAb levels associated with or without fetal GD, including primipara and ophthalmopathy, which differed significantly between the fetal GD and non-fetal GD groups. Maternal TRAb and TSAb levels were higher in primiparous women than in multiparous women at each stage of pregnancy (data not shown). Additionally, there was a correlation between ophthalmopathy and maternal TSAb levels (data not shown), while that between ophthalmopathy and maternal TRAb levels was not significant. Subsequently, multivariate logistic regression (using the Firth adjusted likelihood method using a generalized linear model) was performed, with ophthalmopathy and maternal TRAb ≥12.8 IU/L at 10 GW, and ophthalmopathy and maternal TRAb ≥10.2 IU/L at 20 GW serving as adjustment variables for fetal GD development. In the former model, the adjusted odds ratio for ophthalmopathy was 13.1 (95% CI, 2.3–112.3), p = 0.0018, and that for maternal TRAb ≥12.8 IU/L at 10 GW was 23.2 (95% CI, 3.8–288.2), p = 0.0001; each was an independent predictor. Whereas in the latter model, ophthalmopathy was no longer a significant explanatory factor for fetal GD development (data not shown). Therefore, we tested the ROC AUCs combining ophthalmopathy and maternal TRAb ≥12.8 IU/L at 10 GW: the AUCs were 0.81 for maternal TRAb ≥12.8 IU/L at 10 GW alone and 0.89 for the combination with ophthalmopathy, with a difference in the AUCs of 0.08 (95% CI, –0.186–0.029). Combining ophthalmopathy with maternal TRAb ≥12.8 IU/L at 10 GW did not improve the diagnostic accuracy of predicting fetal GD development.
Discussion
We found that the TRAb cutoff value of ≥10.2 IU/L at around 20 GW to be the most suitable predictor of fetal GD in pregnancies with GD after definitive therapies, such as thyroidectomy or radioactive iodine therapy (sensitivity: 100.0%, specificity: 90.0%). Furthermore, combining the cutoff values for TRAb ≥10.2 IU/L and TSAb ≥1,259.0% around 20 GW suggested the possibility of increased specificity (97.0%). The presence of ophthalmopathy was an independent predictor of fetal GD, although combining ophthalmopathy with TRAb levels did not improve diagnostic accuracy.
Several studies have reported on the value of maternal TRAb and TSAb levels in pregnancy among women with GD in predicting neonatal GD [14, 15, 22]. In a retrospective study of 47 infants born to mothers with GD in France, no cases of neonatal GD occurred when maternal second trimester TRAb levels were ≤5.6 IU/L. Furthermore, in an analysis of 41 mothers, a TRAb cutoff value of ≥5 IU/L predicted neonatal GD with 100% sensitivity and 43% specificity; in 18 of the mothers, combining TRAb ≥5 IU/L with TSAb ≥400% increased diagnostic accuracy (sensitivity: 100%, specificity: 86%) [14].
Several reports have described maternal TRAb levels as predictive of neonatal GD development in pregnancies with GD after definitive therapies [16, 4]. However, no reports exist on the prediction of fetal GD onset, which essentially requires fetal treatment during pregnancy. Yoshihara et al. evaluated 145 mothers with GD within 2 years after radioactive iodine therapy and reported that a maternal TRAb cutoff level >9.7 IU/L in the third trimester predicted neonatal hyperthyroidism with 100% sensitivity and 88.3% sensitivity [16]. In another study, 44 pregnancies in 35 mothers with GD who underwent radioactive iodine therapy were examined, and neonatal hyperthyroidism occurred in 11.3% of cases; moreover, mothers who delivered infants with neonatal hyperthyroidism had significantly higher TRAb levels at delivery than mothers who delivered euthyroid infants (65.8% ± 22.5% vs. 23.3% ± 16.0%, p < 0.05) [4]. However, these reports did not address fetal GD. Additionally, while the differentiation of fetal thyroid goiter has been reported [17], cases of fetal GD in pregnancies with high maternal TRAb and TSAb levels after surgery or radioactive iodine therapy exist; however, no reports have focused on predicting fetal GD onset.
Using the cutoff value of maternal TRAb ≥10.2 IU/L at approximately 20 GW had the best diagnostic accuracy (sensitivity: 100%, specificity: 90.0%), suggesting the possibility of predicting fetal GD risk. While maternal TRAb and TSAb levels tend to decrease during the course of pregnancy [23], high maternal levels persist throughout pregnancy in some cases [4], and elevated TRAb levels during pregnancy have been reported [24], indicating that screening at around 20 GW rather than in early pregnancy may lead to a more accurate diagnosis of fetal GD. Additionally, combining TRAb ≥10.2 IU/L and TSAb ≥1,259.0% at around 20 GW resulted in an increased specificity of 97.0%. Payrat et al. reported an improvement in specificity from 43% to 86% by combining second trimester TRAb ≥5.6 IU/L and TSAb ≥400% in predicting neonatal GD onset. Furthermore, the TRAb and TSAb cutoff values in our fetal GD prediction were approximately twice that of those reported by Payrat et al. [14] Yoshihara et al. also reported that third trimester TRAb >9.7 IU/L predicted neonatal GD onset in pregnancies after radioactive iodine therapy [16], which is consistent with our 20 GW cutoff value. The reason why the cutoff values of TRAb and TSAb levels for predicting fetal GD are higher in pregnant patients with GD who have undergone definitive therapy, despite often not receiving antithyroid drug therapy during pregnancy, is not entirely clear. In pregnant patients with GD who have undergone definitive therapy, TSH receptor antibodies are still produced even though there is a lower amount of the main antigen, the TSH receptor on thyroid epithelial cells. This indicates the possibility that, compared to GD without definitive therapy, the subtypes of immunoglobulin G produced in the maternal body, as well as the proportions of stimulatory, neutral, and inhibitory immunoglobulin G, may differ. Additionally, based on the mechanism of GD, the lower maternal production of pro-inflammatory cytokines could also be considered as a contributing factor [25].
Herein, the proportion of primiparas was higher in the fetal GD group than in the non-fetal GD group, and primiparous women had higher TRAb and TSAb levels than multiparous women. We hypothesized that the primiparous women had become pregnant without a sufficient decrease in TRAb and TSAb levels after definitive therapy.
Fetal GD was noted as early as approximately 20 GW in the fetal GD group (Supplementary Table 1). As fetal GD development has also been reported at 18 GW [9], we recommend that the fetal GD risk be assessed at 20 GW or earlier in cases of high maternal TSH receptor antibody levels before pregnancy.
Some limitations of this study should be noted. First, fetal GD developed in 14 of 79 cases of mothers with GD, TSH receptor antibodies positive status, and a history of definitive therapy, with a high incidence rate of 17.7%. Consecutive pregnant women with GD who were positive for TSH receptor antibodies after definitive therapy were retrospectively enrolled at two facilities. However, both are perinatal tertiary facilities and likely to have a larger population of women with maternal GD with higher TSH receptor antibody levels. Therefore, the prevalence of fetal GD may be overestimated. Second, fetal GD diagnosis was based on fetal ultrasound by a skilled obstetrician every 2 weeks, starting at around 20 GW, at both centers. However, there were three cases of neonatal GD that did not show clinically evident signs of fetal GD; therefore, fetal GD can be underdiagnosed by ultrasound. Two of these three neonates showed TSH suppression or/and elevated free thyroxine in umbilical cord blood or/and peripheral blood on day 0 after birth, while one showed accelerated bone maturation findings on X-rays taken on day 0 after birth. Thus, we considered these cases to have continuous hyperthyroidism from the fetal to the neonatal period. Third, although an association between ophthalmopathy and fetal GD development has been suggested, the presence of ophthalmopathy was not a useful predictor of fetal GD development in this study. However, ophthalmopathy was diagnosed based only on clinical signs, and the diagnoses were therefore not definitive. We suspected that the failure of ophthalmopathy to improve the predictive accuracy for fetal GD was due, in part, to the notable influence of high maternal TSH receptor antibody levels. Nonetheless, ascertaining whether ophthalmopathy can predict the development of fetal GD warrants further investigation by objectively assessing ophthalmopathy or increasing the study sample size. Fourth, the TSAb measurement method was changed from TSAb-RIA to TSAb-EIA during the study. We converted the TSAb-RIA values to TSAb-EIA values using the conversion formula published by the manufacturer; however, it is unclear whether the formula was applicable during pregnancy. Additionally, many active stimulated antibody assay kits have not been standardized, even for nonpregnant women, making generalizing TSAb values difficult. Considering the upper reference limit for the TSAb level in this study was 120%, the maternal TSAb cutoff at 20 GW for predicting fetal GD was approximately 10 times the upper limit; moreover, it is unclear whether this value can be universalized to other kits. Furthermore, TSAb levels will be measured by bioassay in Japan starting in 2022, and data on the TSAb levels of pregnant women obtained with this method are limited. Fifth, this study included two cases in which thyroidectomy was performed during pregnancy. In both cases, the decrease in maternal TRAb and TSAb levels appeared to be similar to the decrease caused by the general course of pregnancy; however, the possibility of a further decrease as a postoperative effect of thyroidectomy cannot be ruled out.
In conclusion, pregnant women with GD who have undergone definitive therapies, such as thyroidectomy or radioactive iodine therapy, should have maternal TSH receptor antibody levels evaluated at approximately 20 GW, regardless of maternal thyroid function, and be considered at high risk for fetal GD if their TRAb level is ≥10.2 IU/L. Thereafter, they should have the fetal thyroid gland monitored regularly at an advanced perinatal facility and undergo multidisciplinary intervention as needed.
Acknowledgments
We thank all those involved in this study and Editage (www.editage.com) for assistance in preparing the manuscript.
Author Contribution Statement
A.H.: data curation (lead); formal analysis (lead); investigation; visualization; writing – original draft (lead).
N.A.: conceptualization (lead); data curation; funding acquisition (lead); investigation; methodology (lead); project administration (lead); supervision (lead); visualization; writing – review & editing (lead).
N.U., Y.I., A.M., S.W., M.W., N.M.: investigation, writing – review & editing.
S.S., H.K.: data curation, investigation, writing – review & editing.
A.S.: formal analysis (supporting), writing – review & editing.
C.N.: formal analysis (supporting); methodology (supporting), writing – review & editing.
H.S.: writing – review & editing.
Author Disclosure Statement
Naoko Arata received payment for the speaker bureau from ASKA Pharmaceutical Co., Ltd. and research funds from Yamasa Co. Hitoshi Shimano received payment for lecture fees from Kowa Company, Ltd. The other authors declare no conflicts of interest.
Funding Statement
This study was funded by a grant from the Children and Families Agency (Grant No. 23DA0201) received by Naoko Arata. The funders were not involved in this study.
Availability of Data and Materials
The study protocol and datasets for this study are available from the corresponding author upon reasonable request.
Ethics Approval and Consent to Participate
This study was approved by the Ethics Committees of the National Center for Child Health and Development (Approval No. 2021-166) and Osaka Women’s and Children’s Hospital (Approval No. 899). Consent was obtained via an opt-out method, and details of data use were posted in the hospital and on its website.
References
- 1 Krassas GE, Poppe K, Glinoer D (2010) Thyroid function and human reproductive health. Endocr Rev 31: 702–755.
- 2 Alexander EK, Pearce EN, Brent GA, Brown RS, Chen H, et al. (2017) 2017 Guidelines of the American Thyroid Association for the diagnosis and management of thyroid disease during pregnancy and the postpartum. Thyroid 27: 315–389.
- 3 Illouz F, Luton D, Polak M, Besançon A, Bournaud C (2018) Graves’ disease and pregnancy. Ann Endocrinol (Paris) 79: 636–646.
- 4 Hamada N, Momotani N, Ishikawa N, Yoshimura Noh J, Okamoto Y, et al. (2011) Persistent high TRAb values during pregnancy predict increased risk of neonatal hyperthyroidism following radioiodine therapy for refractory hyperthyroidism. Endocr J 58: 55–58.
- 5 Atkinson S, McGregor AM, Kendall-Taylor P, Peterson MM, Smith BR (1982) Effect of radioiodine on stimulatory activity of Graves’ immunoglobulins. Clin Endocrinol (Oxf) 16: 537–543.
- 6 Kim J, Choi MS, Park J, Park H, Jang HW, et al. (2021) Changes in thyrotropin receptor antibody levels following total thyroidectomy or radioiodine therapy in patients with refractory Graves’ disease. Thyroid 31: 1264–1271.
- 7 Laurberg P, Wallin G, Tallstedt L, Abraham-Nordling M, Lundell G, et al. (2008) TSH-receptor autoimmunity in Graves’ disease after therapy with anti-thyroid drugs, surgery, or radioiodine: a 5-year prospective randomized study. Eur J Endocrinol 158: 69–75.
- 8 Teng CS, Yeung RT, Khoo RK, Alagaratnam TT (1980) A prospective study of the changes in thyrotropin binding inhibitory immunoglobulins in Graves’ disease treated by subtotal thyroidectomy or radioactive iodine. J Clin Endocrinol Metab 50: 1005–1010.
- 9 Donnelly MA, Wood C, Casey B, Hobbins J, Barbour LA (2015) Early severe fetal Graves disease in a mother after thyroid ablation and thyroidectomy. Obstet Gynecol 125: 1059–1062.
- 10 Davis LE, Lucas MJ, Hankins GD, Roark ML, Cunningham FG (1989) Thyrotoxicosis complicating pregnancy. Am J Obstet Gynecol 160: 63–70.
- 11 Millar LK, Wing DA, Leung AS, Koonings PP, Montoro MN, et al. (1994) Low birth weight and preeclampsia in pregnancies complicated by hyperthyroidism. Obstet Gynecol 84: 946–949.
- 12 Kriplani A, Buckshee K, Bhargava VL, Takkar D, Ammini AC (1994) Maternal and perinatal outcome in thyrotoxicosis complicating pregnancy. Eur J Obstet Gynecol Reprod Biol 54: 159–163.
- 13 Kahaly GJ, Bartalena L, Hegedüs L, Leenhardt L, Poppe K, et al. (2018) 2018 European Thyroid Association guideline for the management of Graves’ hyperthyroidism. Eur Thyroid J 7: 167–186.
- 14 Abeillon-du Payrat J, Chikh K, Bossard N, Bretones P, Gaucherand P, et al. (2014) Predictive value of maternal second-generation thyroid-binding inhibitory immunoglobulin assay for neonatal autoimmune hyperthyroidism. Eur J Endocrinol 171: 451–460.
- 15 Besançon A, Beltrand J, Le Gac I, Luton D, Polak M (2014) Management of neonates born to women with Graves’ disease: a cohort study. Eur J Endocrinol 170: 855–862.
- 16 Yoshihara A, Iwaku K, Noh JY, Watanabe N, Kunii Y, et al. (2019) Incidence of neonatal hyperthyroidism among newborns of Graves’ disease patients treated with radioiodine therapy. Thyroid 29: 128–134.
- 17 Huel C, Guibourdenche J, Vuillard E, Ouahba J, Piketty M, et al. (2009) Use of ultrasound to distinguish between fetal hyperthyroidism and hypothyroidism on discovery of a goiter. Ultrasound Obstet Gynecol 33: 412–420.
- 18 Ranzini AC, Ananth CV, Smulian JC, Kung M, Limbachia A, et al. (2001) Ultrasonography of the fetal thyroid: nomograms based on biparietal diameter and gestational age. J Ultrasound Med 20: 613–617.
- 19 Funaki S, Umehara N, Mezawa H, Kurakazu M, Matsushima S, et al. (2020) Ultrasonographic assessment of fetal thyroid in Japan: thyroid circumference and distal femoral and proximal tibial ossification. J Med Ultrason (2001) 47: 603–608.
- 20 Goldstein I, Lockwood C, Belanger K, Hobbins J (1988) Ultrasonographic assessment of gestational age with the distal femoral and proximal tibial ossification centers in the third trimester. Am J Obstet Gynecol 158: 127–130.
- 21 Pencina MJ, D’Agostino RB (2004) Overall C as a measure of discrimination in survival analysis: model specific population value and confidence interval estimation. Stat Med 23: 2109–2123.
- 22 Tamaki H, Amino N, Aozasa M, Mori M, Iwatani Y, et al. (1988) Universal predictive criteria for neonatal overt thyrotoxicosis requiring treatment. Am J Perinatol 5: 152–158.
- 23 Zakarija M, McKenzie JM (1983) Pregnancy-associated changes in the thyroid-stimulating antibody of Graves’ disease and the relationship to neonatal hyperthyroidism. J Clin Endocrinol Metab 57: 1036–1040.
- 24 Suzuki N, Yoshihara A, Yoshimura Noh J, Kinoshita K, Ohnishi J, et al. (2020) TRAb elevations occurred even in the third trimester; a case of a mother of a child with neonatal thyroid dysfunction, who received radioactive iodine therapy for Graves’ disease. Endocr J 67: 1019–1022.
- 25 Davies TF, Andersen S, Latif R, Nagayama Y, Barbesino G, et al. (2020) Graves’ disease. Nat Rev Dis Primers 6: 52.
