Assessing the Quality of Care of Pregnant Patients with Thyrotoxicosis at an Urban Safety Net Hospital

Abstract

Background:

Thyrotoxicosis can adversely affect pregnancy. The quality of care (QoC) for thyrotoxicosis in pregnancy at a tertiary care safety net hospital was evaluated based on current guidelines.

Methods:

Pregnant patients with thyrotoxicosis or a history of Graves’ disease who delivered in 2015–2021 were divided into three groups: low TSH, active Graves’ disease, and past Graves’ disease. The QoC was assessed using thyroid hormone and thyroid stimulating immunoglobulin (TSI) levels, fetal ultrasound, and endocrine referrals. We assessed potential impacts of race/ethnicity and socioeconomic status (SES).

Results:

We included 147 subjects (mean age 31.5yr, 76% Black, 86% non-Hispanic). Of patients with low TSH (n=95), 75% had repeat TSH measurements and 33% had TSI measured. Hispanic patients were more likely to have TSI and repeat TSH measured than non-Hispanics (58% vs 29%; p=0.04, and 100% vs 71%; p=0.03, respectively). In patients with active Graves’ disease (n=23, 70% treated with thionamides), 35% had FT4 levels at goal and 90% had endocrine care or referral. In patients with past Graves’ disease (n=27), 56% had TSI measured, 78% had first-trimester TSH measurements, and 58% had TSH at goal. Black patients were less likely to have TSH checked in the first trimester than other races (85% vs 100%, p=0.048).

Conclusion:

The QoC of thyrotoxicosis in pregnancy at this tertiary care center can be improved. A larger study is needed to assess the potential impacts of race and SES on the care of pregnant patients with thyrotoxicosis.

Introduction

Overt hyperthyroidism, which affects 0.1 to 0.4% of all pregnancies, may have adverse effects on pregnancy and neonatal outcomes.¹ The most common causes of hyperthyroidism during pregnancy include Graves’ disease, an autoimmune disease that may wax and wane, and gestational transient thyrotoxicosis (GTT), which resolves spontaneously. Pregnancy presents unique challenges for hyperthyroidism diagnosis and management since there are physiologic gestational changes in thyroid function and possible adverse fetal effects of both treatment and uncontrolled disease.¹ The American Thyroid Association (ATA) has outlined guidelines for the diagnosis and treatment of hyperthyroidism in patients who are pregnant. These guidelines discuss differentiating Graves’ disease from GTT, monitoring thyroid stimulating hormone (TSH), free thyroid hormone (FT4), and thyroid receptor antibody (TRAb) levels in those with active hyperthyroidism or history of hyperthyroidism, and choosing and appropriately dosing antithyroid drugs (ATD) if required.²

While there have been many published reviews discussing the ATA’s guidelines, no studies have explored how the recommendations from the current ATA guidelines related to hyperthyroidism in pregnancy are followed in real-world practice, or if there are disparities in the quality of care. Previous studies have assessed how disparities may affect the treatment or management of other types of endocrine diseases, such as diabetes and thyroid cancer.³ Some studies have investigated the effect of disparities on management of hyperthyroidism (in particular, Graves’ disease) in the general population. Studies reported that patients with Graves’ disease who were uninsured, had lower income or lower socioeconomic status (SES), or were Black were more likely to undergo thyroidectomy.^4,5 Another study reported that patients who were Black, using Medicaid, or were uninsured were more likely to undergo surgery rather than radioactive iodine (RAI) ablation, although this association was no longer significant after controlling for confounding.⁶ This study concluded that while clinical characteristics such as age, gender, presence of Graves’ ophthalmopathy, and duration of medical treatment determined the choice of treatment, the clinical characteristics themselves were associated with the social determinants of health. Other studies have investigated the effect of indigenous identity on outcomes in Graves’ disease, with lower rates of optimal thyroid hormone levels seen in indigenous populations compared to European counterparts.^7–9 While these studies identified important effects of race and socioeconomic status on the management of Graves’ disease, they did not specifically focus on disparities in the treatment or management of Graves’ disease or other types of hyperthyroidism during pregnancy. However, racial and ethnic disparities in health outcomes in pregnancy are well-described.^10–12 A 2019 study found that skin-tone-based microaggressions may negatively impact African-American patients’ engagement with prenatal care.¹⁰ Another focus group-based study found that Black and Latina patients in particular experienced a perceived lack of attentiveness to their needs during prenatal care.¹¹ Black women are also twice as likely to experience severe maternal morbidity compared to non-Hispanic white women.¹¹ These experiences and disparities may carry over into the management of hyperthyroidism in pregnant patients.

Our current study aimed to 1) assess the quality of care of pregnant patients with history of hyperthyroidism at a tertiary care center and safety net hospital, using the 2017 ATA guidelines as metrics, and 2) investigate any disparities in the quality of care of pregnant patients with regards to hyperthyroidism management, related to race, ethnicity, or SES.

Materials and Methods

Study population

This study was approved by the Institutional Review Board. All adult pregnant persons aged ≥18 years with thyrotoxicosis or Graves’ disease who gave birth at an urban safety net hospital from June 2015 to March 2021 were identified by the Clinical Data Warehouse using International Classification of Disease-9 and 10 (ICD-9 and 10) codes corresponding to thyrotoxicosis, Graves’ disease, hyperthyroidism, subclinical hyperthyroidism, thyroid dysfunction in pregnancy, toxic goiter, and gestational transient thyrotoxicosis. Patients were excluded from the study if they 1) were pregnant with a twin or higher order gestation, 2) took lithium and/or amiodarone during pregnancy, 3) had no available records for prenatal care, or 4) did not give birth at this urban safety net hospital. Records of infants born to these patients were also obtained. Medical records of included patients were reviewed by the authors to verify extracted information.

Data collection

Demographic data including age, race, ethnicity, insurance type (public versus private as a surrogate marker for socioeconomic status), and primary language, were collected, maternal information regarding a history of definitive treatment for thyrotoxicosis, serum TSH and FT4 measurements, gestational age at the time of thyroid function test measurement, thyroid stimulating immunoglobulin (TSI) measurements (the preferred TRAb assay used at the study institution), use of thyroid medications including methimazole, propylthiouracil, levothyroxine, and triiodothyronine, date of start of thyroid medications, gestational age at delivery, maternal age at start of pregnancy, endocrine referral (or established endocrine care), as well as information about ultrasounds performed at 23–25 weeks and 32–35 weeks’ gestation were obtained from medical records. Information regarding pregnancy and fetal/neonatal complications were collected, but not analyzed due to small sample size.

Subjects were divided into the following groups based on review of medical records: low TSH regardless of etiology, active Graves’ disease in pregnancy, and past Graves’ disease without current hyperthyroidism during pregnancy (including those who had definitive treatment with surgery or RAI ablation preconception). For each of these three groups, different metrics were used to assess the quality of care, based on the 2017 ATA guidelines for treatment of hyperthyroidism. For patients with low TSH regardless of etiology, the metrics used to assess the quality of care were repeat TSH measurements and TSI measurement. For patients with active Graves’ disease, the metrics used were whether patients were treated with ATDs, if the free thyroxine (FT4) was at goal (defined as at upper one-third of normal range or slightly above the upper limit of normal, which are 1.8ng/dL for FT4 and 4 for FT4 index in assay used), TSI measurement in the first trimester, and engagement with endocrine care. For patients with past Graves’ disease, the metrics used were TSI (if appropriate – in those with history of thyroidectomy or radioactive iodine ablation for Graves’ disease) and TSH measurements in the first trimester.

Statistical analysis

Descriptive data are presented as mean ± standard deviation (SD) or median (range) as appropriate. Correlations were assessed between each metric of quality of care and race, ethnicity, language, insurance type, and patient’s age, using linear or logistic regression models as appropriate. Chi-square tests were used for multivariable analyses to assess potential impacts of race, ethnicity, insurance type, language, and patient’s age on each of the care quality metrics described above. Potential confounders were selected a priori. A two-tailed p-value <0.05 was considered statistically significant. Statistical analysis was performed using SAS version 9.3 (SAS institute, Inc, Cary, NC, USA).

Results

A total of 147 subjects (mean age 32 years old, 76% Black, 86% non-Hispanic) were included (Table 1). Table 2 presents quality of care for patients in each group according to the metrics described above. Of the patients with low TSH (n=95), 75% had a repeat TSH measured and 33% had TSI measured. Hispanic patients were more likely to have TSI measured than non-Hispanic patients (58% vs. 29%, p=0.04). Hispanic patients were also more likely to have a repeat TSH measurement compared to non-Hispanic patients (100% vs 71%, p=0.03). There were no significant differences in the rates of TSI screening, repeat TSH measurement by race, language, or insurance type (Table 2).

Table 1.

Demographic data for pregnant patients with thyrotoxicosis.

Group	Total N=147	Low TSH (regardless of etiology) N=95	Active Graves’ disease N=23	Past Graves’ disease N=27
Age in years – n (%) Mean Minimum Maximum	31.5 18 49	32.4 18 49	29.4 20 38	33.1 20 43
Race – n (%) White Black Asian Native Hawaiian Other Unidentified	9 (6.1) 111 (75.5) 6 (4.1) 1 (0.7) 1 (0.7) 19 (12.9)	4 (4.2) 71 (74.7) 5 (5.3) 1 (1.05) 1 (1.05) 13 (13.7)	1 (4.35) 14 (60.9) 2 (8.7) 1 (4.35) 0 5 (21.7)	2 (7.4) 20 (74.1) 1 (3.7) 0 0 4 (14.8)
Hispanic Identity – n (%) Yes No	20 (13.6) 127 (86.4)	12 (12.6) 83 (87.4)	6 (26.1) 17 (73.9)	5 (18.5) 22 (81.5)
Language – n (%) English Other	109 (74.1) 38 (25.9)	66 (69.5) 29 (30.5)	19 (82.6) 4 (17.4)	21 (77.8) 6 (22.2)
Insurance – n (%) Private Public	33 (22.4) 114 (77.6)	25 (26.3) 70 (73.7)	4 (17.4) 19 (82.6)	2 (7.4) 25 (92.6)

TSH = thyroid stimulating hormone.

Table 2.

Results. Quality of Care metrics assessed according to thyrotoxicosis status.

Group and metric	Yes – n (%)	No – n (%)	p-value
Low TSH regardless of etiology (n=95)  • TSI measured   ○ All race    ◾ Black    ◾ White    ◾ Asian    ◾ Other   ○ Hispanic ethnicity  • Serum TSH repeated   ○ All race    ◾ Black    ◾ White    ◾ Asian    ◾ Other   ○ Hispanic ethnicity	31 (32.6) 19 (26.8) 3 (75) 2 (40) 1 (50) 7 (58.3) 71 (74.7) 51 (71.8) 3 (75) 2 (40) 1(50) 12 (100)	64 (67.4) 52 (73.2) 1 (25) 3 (60) 1 (50) 5 (41.7) 24 (25.3) 20 (28.2) 1 (25) 3 (60) 1 (50) 0 (0)	0.15 0.042 0.32 0.031
Active Graves’ disease (n=23)  • Treated with antithyroid drugs   ○ All race    ◾ Black    ◾ White    ◾ Asian    ◾ Other   ○ Hispanic ethnicity  • Free T4 at goal level**   ○ All race    ◾ Black    ◾ White    ◾ Asian    ◾ Other   ○ Hispanic ethnicity  • Receiving endocrinology care   ○ All race    ◾ Black    ◾ White    ◾ Asian    ◾ Other   ○ Hispanic ethnicity	16 (69.6) 10 (76.9) 1 (100) 2 (100) 1 (100) 6 (100) 8 (34.8) 4 (28.6) 0 (0) 0 (0) 0 (0) 3 (50) 21 (90.3) 12 (85.7) 1 (100) 2 (100) 1 (100) 6 (100)	7 (30.4) 3 (23.1) 0 (0) 0 (0) 0 (0) 0 (0) 13 (56.5) 9 (64.3) 1 (100) 2 (100) 0 (0) 3 (50) 2 (8.7) 2 (14.3) 0 (0) 0 (0) 0 (0) 0 (0)	N/A N/A 0.024 0.52 0.94 0.52
Past Graves’ disease (n=27)  • TSI measured (if prior definitive therapy)   ○ All race    ◾ Black    ◾ White    ◾ Asian    ◾ Other   ○ Hispanic ethnicity  • TSH measured in 1st trimester   ○ All race    ◾ Black    ◾ White    ◾ Asian    ◾ Other   ○ Hispanic ethnicity	15 (55.6) 10 (50) 1 (50) 1 (100) 0 (0) 3 (60) 21 (77.8) 17 (85) 2 (100) 1 (100) 0 (0) 3 (60)	12 (44.4) 10 (50) 1 (50) 0 (0) 0 (0) 2 (40) 6 (22.2) 3 (15) 0 (0) 0 (0) 0 (0) 2 (40)	0.64 0.82 0.05 0.29

TSH = thyroid stimulating hormone; TSI = thyroid stimulating immunoglobulin; T4 = thyroxine

^**2 patients did not have FT4 levels obtained.

In patients with active Graves’ disease (n=23), 70% were treated with antithyroid drugs. Eight (35%) of patients with active Graves’ disease had FT4 levels at goal, and 21 (90%) patients had endocrine care or referral. Most patients had fetal ultrasound performed at appropriate time points: 74% in the second trimester and 83% in the third trimester. While there was a statistically significant difference in the proportion of FT4 levels at goal by race (p=0.02), the sample size was too small to compare each race separately. There was no significant difference in the rates of FT4 levels at goal or in rates of endocrinology referral or care by ethnicity, language, or socioeconomic status (Table 2).

In patients with past Graves’ disease (n=27), 56% had TSI measured (5/13 patients with prior thyroidectomy and 4/5 patients with prior RAI ablation), 78% had TSH measured in the first trimester. TSH was at goal (defined as < 2.5mIU/L if on levothyroxine or < 4mIU/L if not on levothyroxine) in 58% of all patients with past history of Graves’ disease and in 44% (7/16) of patients with history of definitive therapy for Graves’ disease. Black patients were less likely to have TSH checked in the first trimester than other races, although the total sample size was small. Eighty-five percent of Black patients had their TSH checked in the first trimester compared to 100% of white and Asian patients and 25% of patients with unidentified race (p=0.048). There was no significant association between having TSH measured in the first trimester and language, Hispanic identity, or socioeconomic status (Table 2). There was no significant association between having TSI measured and race, ethnicity, language, or socioeconomic status (Table 2).

Twenty-five patients were on levothyroxine treatment during pregnancy, of whom 22 had a prior history of Graves’ disease. No patients who had active or past Graves’ disease were on a block and replacement regimen (levothyroxine with an ATD). One patient with a prior history of Graves’ disease had a fetal demise in the second trimester during her second pregnancy.

Discussion

Our study of 147 pregnant people at a tertiary safety-net hospital demonstrated that the quality of care in pregnant patients who have current or prior hyperthyroidism can be improved. TSH abnormalities during the first trimester of pregnancy are common. Low TSH in the first trimester of pregnancy is most frequently due to gestational transient thyrotoxicosis, where maternal hCG directly stimulates TSH receptor on the thyroid gland, increasing thyroid hormone production with subsequent reduced serum TSH levels through negative feedback. Graves’ disease is the second most common cause of hyperthyroidism in pregnancy. The current ATA guidelines recommend that any subnormal serum TSH in the first trimester should be evaluated by measuring serum total T4 or FT4, and T3 levels, as well as using a careful history and physical examination to determine the etiology. If overt hyperthyroidism is present, measurement of TRAb titers can be helpful in clarifying the diagnosis or underlying cause.² Based on these recommendations, we elected to consider our metrics for the assessment of quality of care in patients with low TSH during their pregnancy as repeat TSH and TSI measurements. In patients who were found to have low TSH during pregnancy, regardless of etiology, our analysis showed that 75% had a repeat TSH measurement, but only 33% had their TSI measured. TSI measurement may not always have been indicated for each of these patients, if they had a presentation that was strongly suggestive of GTT and lacked findings consistent with Graves’ disease. Notably, our analysis showed that Hispanic patients were more likely to have TSI and repeat TSH measured than non-Hispanic patients. The study’s setting at an urban safety net hospital with a diverse, multicultural patient population makes this finding particularly important, as this hospital specifically aims to optimize the care of patients from diverse backgrounds.

In patients with active Graves’ disease during pregnancy, the ATA recommends treatment as needed to minimize the potential adverse effects of overt hyperthyroidism on pregnancy and child outcomes.² The main treatment option during pregnancy is ATDs. Therefore, we selected use of ATDs as one of our metrics for assessing the quality of care in patients with active Graves’ disease. In our cohort, 70% of patients with active Graves’ disease were taking ATDs. However, this result does not fully reflect the more nuanced conversation that may have taken place about side effects of ATDs and consideration of duration of therapy, magnitude of thyroid hormone abnormalities (as subclinical hyperthyroidism may not warrant treatment), and severity of Graves’ disease. Over 90% of patients with active Graves’ disease were under the care of endocrinologists. However, the proportion of patients with active Graves’ disease with FT4 levels at goal was low, at 35%. Per the ATA guidelines, target FT4 levels in ATD-treated patients are at or just above the upper limit of normal in patients to avoid fetal overtreatment, since the fetal thyroid is more sensitive than the maternal thyroid to ATD effects.^2,13 In our study, all patients with suboptimal FT4 levels had FT4 below recommended targets by the ATA guidelines. The majority (92%) of the patients with lower FT4 than appropriate were treated with ATDs, which suggests that ATD doses were too high. However, it should be noted that we are unable to assess whether appropriate treatment decision to change ATD dose were made in response to FT4 levels not at goal. Two patients (9%) did not have any FT4 level measured during pregnancy.

For patients who have undergone definitive therapy (RAI or thyroidectomy) for Graves’ disease prior to pregnancy, it is recommended that maternal serum TRAb levels are measured in early pregnancy, as elevated levels of TRAb can increase risks of fetal or neonatal goiter and hyperthyroidism. In our study, only 56% of patients with a history of Graves’ disease with prior definitive therapy had TSI measured. In addition, among patients with a history of Graves’ disease, Black patients were less likely to have TSH checked in the first trimester than other races, although the sample size was small. It is difficult to assess what factors had led to these discrepancies, but a Canadian study by Stoll et al. suggested that access to regular medical care is associated with higher odds of being screened for thyroid disease, either through ultrasound or laboratory testing.¹⁴ In the United States, there have been studies examining the effect of Black race on access to healthcare. One study found that Black patients in the United States had more frequent emergency department visits and inpatient hospital stays compared to White patients, which may be attributable to a lack of outpatient care and appropriate control of ambulatory conditions.^15,16 Another study found that prenatal care was less likely to be established a timely manner in Black women compared to White women¹⁷ despite similar insurance status. In addition, multiple studies have demonstrated that pregnant patients from minoritized group may experience greater health care communication gaps, poorer continuity of care, racial discrimination, and poorer health outcomes, including increased maternal mortality.^11–13

Our study is the first to assess the quality of care of thyrotoxicosis in pregnancy in a diverse population based on the current ATA guidelines. We focused especially on racial diversity and socioeconomic diversity within our study, conducted in urban safety-net tertiary care center, with over 75% of the patients being Black, and over 75% having public rather than private health insurance. Limitations of our study include its small sample size. Among the patients studied, only 23 had active Graves’ disease in pregnancy, and only 27 had past Graves’ disease. The small sample size limited our ability to elucidate the effects of race and socioeconomic status on the quality of care. Additionally, while the measures we examined were important aspects of the care of the pregnant patient with new or past diagnosis of thyrotoxicosis, they are not all encompassing and do not fully reflect the more nuanced conversations that take place between providers and their patients, particularly regarding decisions to continue (or discontinue) ATDs during gestation. Another important limitation to consider is that our patient population in the safety-net hospital may face challenges such as limited resources and competing priorities, which may result in suboptimal frequency of follow up during pregnancy, and thus less than ideal quality of care.

For patients who had past Graves’ disease, we considered obtaining TSH in the first trimester to be a metric of appropriate care. However, due to availability, we used TSH and TSI results, not whether the lab was ordered. Therefore, it is possible that labs may have been ordered but not drawn in some patients, indicating that in addition to provider-level bias, access to care, likely due to systemic factors, may be another barrier in the outpatient setting for patients in our population. Finally, this study was conducted at an urban safety net hospital with a multicultural population, ready access to interpreter services, and substantial institutional efforts to minimize healthcare disparities, which may not be representative of care in other settings.

Conclusion

Our study suggests that the quality of care for pregnant patients with thyrotoxicosis at our urban safety net hospital can be improved. Only 35% of patients with active Graves’ disease had FT4 levels at goal, which may indicate inappropriate titration of ATDs. Care should be taken in ensuring that those with active Graves’ disease have FT4 levels at or just above the upper reference limit to optimize fetal thyroid hormone exposure. Only half of our subjects with past Graves’ disease and history of definitive therapy had recommended TSI monitoring or optimal TSH levels during pregnancy, which should be improved to minimize potential adverse impact on pregnancy and fetal outcomes. A larger study is needed to assess the potential impact of race and SES on care quality of pregnant patients with hyperthyroidism.

Clinical Relevance.

This study enhances the understanding of the care gaps in thyrotoxicosis in a vulnerable adult population. While there have been reviews discussing the ATA guidelines, these articles have not explored how the guidelines are followed in practice. Our study highlights potential disparities and current gap in care in treatment of pregnant patients with Graves’ disease.

Teaching Points.

• The quality of care of pregnant patients at this tertiary care center can be improved based on the American Thyroid Association (ATA) guidelines.

• Only 35% of patients with active Graves’ disease had free T4 levels at goal.

• Half of the patients with prior diagnosis of Graves’ disease who received definitive treatment had the recommended TSI monitoring by the ATA guidelines.

• Of those who had a prior diagnosis of Graves’ disease, Black patients were less likely to have TSH measured in the 1^st trimester than White patients.