Comparison of severe maternal morbidity between pregnancy with subclinical hypothyroidism and overt hypothyroidism.

Conceptualization: Laurel S. Aberle, Caroline T. Nguyen. Data curation: Rachel S. Mandelbaum. Formal analysis: Koji Matsuo. Funding acquisition: Koji Matsuo. Investigation: All authors. Methodology: Laurel S. Aberle, Koji Matsuo. Project administration: Koji Matsuo. Resources: Shinya Matsuzaki, Tatsuya Miyake, Caroline T. Nguyen, Joseph G. Ouzounian. Software: Koji Matsuo. Supervision: Caroline T. Nguyen, Joseph G. Ouounian, Koji Matsuo. Validation: Koji Matsuo. Visualization: Koji Matsuo. Writing – original draft: Laurel S. Aberle, Koji Matsuo. Writing – review & editing: All authors.

University of Southern California Institutional Review Board (HS‐16‐00481, granted on 7/12/2016).

Ensign Endowment for Gynecologic Cancer Research (Koji Matsuo). The funder had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

A total of 11 620 hospital deliveries with a diagnosis code of subclinical hypothyroidism were compared with 697 320 hospital deliveries with a diagnosis code of overt hypothyroidism (Figure  1 ). Study selection schema. *Including congenital iodine deficiency syndrome, and iodine deficiency goiter. § Cases with both subclinical and overt hypothyroidism. † Cases that did not have subclinical and overt hypothyroidism. Results are given in Table  S2 . Patients in the subclinical hypothyroidism group were more likely to be younger than 25 years, Asian, Black, or Hispanic individuals, reside in the highest quartile census‐level area, located in the Northeast region, and have a diagnosis of obesity disorder, anxiety disorder, polycystic ovary syndrome, endometriosis, and uterine fibroids compared with those in the overt hypothyroidism group. On the contrary, pregnant individuals with subclinical hypothyroidism were less likely to be self‐pay, deliver at urban non‐teaching hospitals, and have a diagnosis of depressive disorder compared with those in the overt hypothyroidism group. Results are given in Table  S3 . Pregnancy with subclinical hypothyroidism was associated with gestational hypertension, placenta previa, multi‐fetal gestations, fetal anomaly, oligohydramnios, and chorioamnionitis. Subclinical hypothyroidism was less likely to be associated with large for gestational age compared with overt hypothyroidism. Results are given in Table  S4 . The rate of delivery in 28–33 and >40 weeks gestation was increased for pregnancy with the subclinical hypothyroidism group compared with pregnancy with overt hypothyroidism. The rates of cesarean delivery and operative delivery were lower for pregnancy with subclinical hypothyroidism compared with overt hypothyroidism. Pregnancy with subclinical hypothyroidism was associated with higher rates of obstructive labor, abnormal fetal heart rate during labor, and fourth degree laceration including anal sphincter laceration compared with overt hypothyroidism. Results are given in Table  1 . The incidence rates of any of the 20 measured severe maternal morbidity at delivery were 18.1 and 11.1 per 1000 deliveries for the subclinical hypothyroidism group and the overt hypothyroidism group, respectively. After controlling for clinico^obstetric factors, subclinical hypothyroidism was associated with a 54% higher rate of severe maternal morbidity at delivery compared with overt hypothyroidism. This association remained when excluding hysterectomy (17.2 vs. 9.7 per 1000 deliveries, aIR 1.65, 95% CI: 1.43–1.90) or restricted to any cardiopulmonary morbidity (6 indicators: 5.2 vs. 2.6 per 1000 deliveries, aIR 1.83, 95% CI: 1.41–2.37). Severe maternal morbidity at delivery. Note : The incidence rates are shown per 1000 deliveries per the exposure group. Abbreviations: aIR, adjusted‐incidence rate ratio; CI, confidence interval. Generalized linear model. The exposure–outcome association was adjusted for maternal age, race/ethnicity, obesity, hypertensive disorder, placental pathology, gestational age at delivery, and cesarean delivery. The overt hypothyroidism group served as the reference group. Any one of 20 morbidity indicators per the Centers for Disease Control and Prevention definition. Any one of the following cardiac and pulmonary morbidity indicators: acute myocardial infarction, acute respiratory distress syndrome, cardiac arrest/ventricular fibrillation, conversion of cardiac rhythm, heart failure/arrest during surgery or procedure, and pulmonary edema/acute heart failure. Listed in descending order for aIR for the subclinical hypothyroidism group compared with the overt hypothyroidism group. Small number morbidities were not listed per the program guidelines. Including blood transfusion. Among the individual morbidity indicators (Table  1 ), incidence rate of eclampsia was nearly three times higher among pregnancies with subclinical hypothyroidism compared with pregnancies with overt hypothyroidism (2.2 vs. 0.7 per 1000 deliveries, aIR 2.73, 95% CI: 1.83–4.09), followed by pulmonary edema (2.6 vs. 1.3 per 1000 deliveries, aIR 1.76, 95% CI: 1.22–2.54), acute respiratory distress syndrome (2.6 vs. 1.6 per 1000 deliveries, aIR 1.62, 95% CI: 1.12–2.33), and sepsis (2.2 versus 1.2 per 1000, aIR 1.57, 95% CI: 1.05–2.34). Pregnancy with subclinical hypothyroidism was also associated with higher rates of postpartum hemorrhage with or without blood transfusion (74.1 vs. 54.3 per 1000 deliveries, aIR 1.28, 95% CI: 1.20–1.37) and length of hospital admission of 7 days or longer (35.7 vs. 26.6 per 1000 deliveries, aIR 1.16, 95% CI: 1.05–1.28) compared with pregnancy with overt hypothyroidism. In an exploratory evaluation according to patient demographic (Table  2 ), maternal age younger than 25 years (31.9 vs. 9.4 per 1000 deliveries, aIR 3.62, 95% CI: 2.62–5.01), Black individuals (55.2 vs. 24.7 per 1000 deliveries, aIR 2.21, 95% CI: 1.60–3.06), and pregestational hypertension (70.2 vs. 27.2 per 1000 deliveries, aIR 2.20, 95% CI: 1.60–3.03) were associated with twice or higher rate of severe maternal morbidity for subclinical hypothyroidism compared with overt hypothyroidism, followed by obesity disorder (35.2 vs. 16.7 per 1000 deliveries, aIR 1.87, 95% CI: 1.48–2.35). Sensitivity analysis per patient demographics. Note : The incidence rates for severe maternal morbidity (20 indicators) are shown per 1000 deliveries per the exposure group. Abbreviations: aIR, adjusted‐incidence rate ratio; CI, confidence interval. Generalized linear model. The adjusting model followed the base case model as presented in Table  1 . The overt hypothyroidism group served as the reference group. Post hoc assessment per the combination patterns of two patient demographics exhibiting large effect size. Combination patterns with small number events were not shown per the HCUP guidelines. When assessed based on the combination patterns of any two of these four patient demographics (maternal age younger than 25 years, Black individual, pregestational hypertension, and obesity disorder) (Table  2 ), the rate of any of 20 measured severe maternal morbidity indicators was particularly higher for subclinical hypothyroidism compared with overt hypothyroidism among pregnant patients younger than 25 years with obesity disorder (79.4 vs. 10.1 per 1000 deliveries, aIR 7.89, 95% CI: 4.78–13.03). This was followed by Black individuals with pregestational hypertension (157.9 vs. 40.2 per 1000 deliveries, aIR 3.15, 95% CI: 1.77–5.61), Black individuals with obesity disorder (102.0 vs. 33.9 per 1000 deliveries, aIR 2.81, 95% CI: 1.83–4.32), and pregestational hypertension with obesity disorder (76.9 vs. 26.9 per 1000 deliveries, aIR 1.93, 95% CI: 1.21–3.06).

Key results of the current investigations are the following. First, pregnancy with subclinical hypothyroidism was associated with increased severe maternal morbidity at delivery compared with overt hypothyroidism, especially for eclampsia. Second, severe maternal morbidity rates were increased for pregnancy with subclinical hypothyroidism among patients with a younger maternal age, Black race, pregestational hypertension, and obesity disorder. Subclinical hypothyroidism has been shown to be associated with an increased risk of preterm delivery, miscarriage, impaired glucose tolerance, and gestational diabetes, and pregnancy‐induced hypertensive disorders. 1 , 2 , 3 , 4 , 5 , 6 The findings from this current study further support the association between subclinical hypothyroidism and adverse maternal outcomes in comparison to overt hypothyroidism. In this study, subclinical hypothyroidism was markedly associated with eclampsia. While the exact etiology of this is unknown, the relationship between subclinical hypothyroidism and pregnancy‐induced hypertensive disorders has been studied previously. It is hypothesized that preeclampsia may lead to hypothyroidism due to an increase in anti‐angiogenic factors which leads to decreased nitric oxide production and ultimately reduced capillary flow causing hypothyroidism. 33 Another hypothesis is that hypothyroidism leads to increased peripheral vascular resistance due to abnormal effects on smooth muscle contraction in the renal and systemic arteries, thus predisposing to pregnancy‐induced hypertension. 33 Further research is needed to identify the pathophysiology underlying the association between subclinical hypothyroidism and adverse maternal outcomes. Guidelines on managing subclinical hypothyroidism in pregnancy vary by association. The American College of Obstetricians and Gynecologists 2020 guidelines do not recommend further testing, monitoring, or treatment. 23 This is similar to the 2024 American Society of Reproductive Medicine guidelines. 34 These guidelines are different than the American Thyroid Association 2017 guidelines, 12 which recommend treatment based on thyroid peroxidase antibody status and thyroid stimulating hormone level. These guidelines also recommend ongoing monitoring for those at risk, which include patients with thyroid autoimmunity, history or hemithyroidectomy, or history of treatment with radioactive iodine. 12 Treatment of subclinical hypothyroidism in pregnancy varies depending on association guidelines as mentioned previously. 12 , 23 , 34 This may be due to the limited data supporting a benefit of treatment with levothyroxine. Two randomized clinical trials have reported on the effects of levothyroxine treatment on pregnancy outcomes in patients with subclinical hypothyroidism. 35 , 36 A 2012 trial found no difference in preterm birth rate, delivery gestational age, or birthweight in those with untreated as compared with treated subclinical hypothyroidism. 35 A 2017 trial similarly showed no difference in preterm delivery, preterm birth rate, or birth weight in those with untreated as compared with treated subclinical hypothyroidism. 36 Following this, a 2024 meta‐analysis showed no difference in live birth rate with treatment. 37 In contrast to this, two different randomized clinical trials showed a reduction in preterm birth rate with treatment but only in patients with a thyroid stimulating hormone level >4.0 mU/L. 38 , 39 Additionally, a large‐scale retrospective study found that treatment was associated with a reduced risk of miscarriage in women with a thyroid stimulating hormone level >4.0 mU/L. 7 Another retrospective study showed a reduced risk of preeclampsia and gestational diabetes in patients who received treatment for subclinical hypothyroidism compared with those who did not. 40 Overall, the data remain inconsistent and treatment for subclinical hypothyroidism is not standardized. Screening for thyroid disease in pregnancy is currently recommended only for women with a personal or family history of thyroid disease, type I diabetes mellitus, or clinical suspicion of thyroid disease according to the American College of Obstetricians and Gynecologists. 23 While universal screening has been considered, 12 the benefit of it remains unclear. While some report that universal screening has not been shown to result in reduced adverse outcomes and additionally has not been shown to provide cost benefit, 12 , 14 , 23 others report that universal screening is cost‐effective as it allows for the detection and treatment of cases of subclinical hypothyroidism, further decreasing the risk of complications. 15 , 16 , 41 Consensus on treatment guidelines would allow for an improved ability to decipher the benefit of universal screening. The current study did not aim to evaluate the effect of screening, but multiple risk factors for severe maternal morbidity associated with subclinical hypothyroidism have been identified (age <25, Black individuals, pregestational hypertension, and obesity disorder). Selective screening based on these risk factors may be considered in order to increase the probability of identifying those with subclinical hypothyroidism and most at risk of adverse maternal outcomes. Inclusion of large sample size, pair‐wise comparison of subclinical hypothyroidism and overt hypothyroidism, and several sensitivity analyses enhanced the interpretation of research findings. Key limitations included unmeasured confounding due to the lack of information on clinical details of the subclinical and overt hypothyroidism, including the timing of diagnosis (prior to or during the index pregnancy), severity, and treatment. It is possible that a higher rate of severe maternal morbidity for subclinical hypothyroidism compared with overt hypothyroidism may be due to the treatment effect in the overt hypothyroidism group. Whether practice guidelines have influenced the thyroid screening during pregnancy was not assessed due to lack of information for thyroid function test in the database. The interpretation limits to association only, and the causality for subclinical hypothyroidism and severe maternal morbidity was not evaluated. Identification of exposure, outcomes, and the measured confounders was solely based on the administrative codes, and the accuracy of these data was not assessable with possible misclassification. The generalizability of the study findings in other populations was not examined.

The results of this cross‐sectional study suggest that pregnancy with subclinical hypothyroidism may be associated with severe maternal morbidity at delivery compared with pregnancy with overt hypothyroidism. These data set forth three important findings: (i) recognition of pregnancy with subclinical hypothyroidism as a high‐risk pregnancy group is reasonable; (ii) whether treatment of subclinical hypothyroidism may or may not be beneficial during pregnancy; and (iii) there appears to be a possible role of selective thyroid function screening during pregnancy for “at‐risk” populations.

Subclinical hypothyroidism is diagnosed when the thyroid stimulating hormone is elevated but thyroxine or free‐thyroxine levels are normal. While the prevalence of this condition varies greatly depending on the thyroid stimulating hormone level used, generally, it affects approximately 1.5–2.5% of all pregnancies. 1 , 2 Subclinical hypothyroidism has been associated with adverse maternal and neonatal outcomes. Subclinical hypothyroidism has been shown to be associated with an increased risk of preterm delivery, pregnancy loss, impaired glucose tolerance and gestational diabetes, and pregnancy‐induced hypertensive disorders. 1 , 2 , 3 , 4 , 5 , 6 The mechanism of how or why subclinical hypothyroidism is associated with these maternal adverse effects remains unclear. This is especially unclear due to the limited benefit seen when subclinical hypothyroidism is treated and euthyroidism achieved. 1 , 3 , 7 , 8 While an association between subclinical hypothyroidism and adverse maternal outcomes has been shown, little research has examined the relationship between subclinical hypothyroidism and severe maternal morbidity. Severe maternal morbidity refers to unexpected, serious, and negative outcomes during childbirth. 9 , 10 Severe maternal morbidity is increasing in the United States, and the causes for this are likely multifactorial and not well understood. 10 It is important to further elucidate conditions which are associated with and contribute to severe maternal morbidity to minimize complications and improve counseling and management. 10 Additionally, guidelines regarding monitoring and treatment of subclinical hypothyroidism vary across professional societies. 11 , 12 , 13 Further research regarding the relationship between subclinical hypothyroidism and severe maternal morbidity can help unify treatment and monitoring guidelines. While universal screening is not currently recommended, 11 , 14 it has been shown to identify cases of thyroid disease which otherwise would not have been detected via risk factor‐based testing. 15 Thus, further evaluation of the relationship between subclinical hypothyroidism and severe maternal morbidity may allow for improved screening guidelines. Moreover, prior investigations mainly compared subclinical hypothyroidism to euthyroidism, and there is a scarcity of data regarding whether severe maternal morbidity among patients with subclinical hypothyroidism is different from that in those with overt hypothyroidism. 3 Thus, the objective of the current study was to compare severe maternal morbidity at delivery between pregnancy with subclinical hypothyroidism and overt hypothyroidism.

The manuscript's corresponding author (Koji Matsuo) affirms that the manuscript is an honest, accurate, and transparent account of the study being reported; that no important aspects of the study have been omitted; and that any discrepancies from the study as planned (and, if relevant, registered) have been explained. The National Inpatient Sample is developed for the Healthcare Cost and Utilization Project that is sponsored by the Agency for Healthcare Research and Quality, and the program is the source of the de‐identified data used; race/ethnicity was grouped by the program; and the program has not verified and is not responsible for the statistical validity of the data analysis or the conclusions derived by the study team.

The authors declare no conflicts of interest.

This serial cross‐sectional study utilized the National Inpatient Sample. 16 The database is the Healthcare Cost and Utilization Project's national‐level inpatient database. The Healthcare Cost and Utilization Project is the United States health service data platform that is supported by the Agency for Healthcare Research and Quality, one of the 12 federal agencies within the United States Department of Health and Human Service. The National Inpatient Sample approximates a stratified sample of 20% of discharges in each center from all the participating hospitals across 48 States and the District of Columbia. 16 , 17 Every year the dataset captures nearly seven million inpatient admissions. In 2022, more than 4500 hospitals participated in the program. When weighted for national survey estimates, the National Inpatient Sample represents more than 97% of the U.S. population. The program captures a maximum of 40 diagnoses and 25 procedures for the index admission in each encounter. This study used the National Inpatient Sample due to these robust data capturing mechanisms which enable us to conduct a nationwide overview of pregnancy with subclinical hypothyroidism in the United States. This study was deemed exempt and waived for approval by the University of Southern California Institutional Review Board due to the use of publicly available, de‐identified secondary data (registration number, HS‐16‐00481; patient consent: not required). The study population included hospital deliveries for pregnant patients aged between 15 and 54 years who had a diagnosis code of subclinical hypothyroidism or overt hypothyroidism from 1/2016‐6/2021. Hospital deliveries refer to either vaginal or cesarean deliveries, identified by the Diagnosis‐Related Group codes or the World Health Organization's International Classification of Disease Tenth Revision Clinical Modification or Procedure Classification Schema codes (Table  S1 ). 18 , 19 , 20 , 21 The age range was used to maximize the catchment of pregnancy codes per the classification schema. 18 , 22 The study starting point was chosen due to the introduction of revised administrative coding schema in the National Inpatient Sample and the study endpoint was due to the availability of exposure information. Exclusion criteria included hyperthyroidism, congenital iodine deficiency syndrome, and iodine deficiency goiter (Table  S1 ). Cases that did not have these codes for subclinical hypothyroidism and overt hypothyroidism were also excluded in this study. These exclusions were to eliminate the possible misclassification due to the lack of information on thyroid function results in the database. The exposure was hypothyroidism status, classified as the following groups: subclinical hypothyroidism or overt hypothyroidism. Identifications of subclinical hypothyroidism and overt hypothyroidism were based on the International Classification of Disease Tenth Revision Clinical Modification codes of E02 and E03, respectively (Table  S1 ). Cases that had both codes for subclinical hypothyroidism and overt hypothyroidism were excluded. Overt hypothyroidism was set as the reference group for proof‐of‐principal assessment in this study. 23 The database does not have information for the chronological timing of hypothyroidism diagnosis. The primary outcome measure was set as severe maternal morbidity at delivery. This was selected given the importance of obstetric endpoints. 24 Identifications of these outcome measures followed prior investigations (Table  S1 ). This study followed the Centers for Disease Control and Prevention definition and coding schema to identify severe maternal morbidity (a total of 20 indicators) 25 : acute myocardial infarction, aneurysm, acute renal failure, acute respiratory distress syndrome, amniotic fluid embolism, cardiac arrest / ventricular fibrillation, cardiac rhythm conversion, disseminated intravascular coagulation, eclampsia, heart failure/arrest during surgery or procedure, puerperal cerebrovascular disorders, pulmonary edema/acute heart failure, severe anesthesia complications, sepsis, shock, sickle cell disease with crisis, air and thrombotic embolism, hysterectomy, temporary tracheostomy, and ventilation. The primary outcome measure was set as the composite endpoint of any one of these 20 morbidity indicators. Secondary outcome measures included postpartum hemorrhage including blood product transfusion and prolonged hospitalization defined as the length of stay for hospital delivery of 7 days or longer. Patient baseline demographics, obstetric factors, and delivery characteristics were preselected as study covariates. Identification of these study covariates other than the program defined data followed prior investigations (Table  S1 ). 18 , 19 , 20 , 21 , 26 , 27 , 28 , 29 , 30 , 31 Cases with unknown status were grouped as one category in each study covariate. At cohort level, 95.1% of study population had complete information without missing data. Patient baseline demographics included maternal age (<25, 25–29, 30–34, 35–39, and ≥40 years), study period (trisected), race and ethnicity (Asian, Black, Hispanic, Native American, Other, and White) determined and grouped per the Healthcare Cost and Utilization Project, primary payer (Medicaid, private insurance including Health Maintenance Organization, self‐pay, and other), census‐level median household income (every quarter), medical comorbidity (pregestational hypertension, pregestational diabetes mellitus, obesity disorder, and asthma), substance use disorder (tobacco and alcohol), mental health condition (depressive disorder and anxiety disorder), and relevant reproductive factors (prior uterine scar, pregnancy loss, uterine fibroid, endometriosis, and polycystic ovary syndrome). Race and ethnicity were examined as this factor is associated with pregnancy characteristics and maternal outcomes. 19 Hospital parameters where the index delivery occurred were also evaluated, including relative bed capacity (small, mid, and large), location and teaching setting (rural, urban non‐teaching, and urban teaching), and United States census region (Northeast including New England and Mid‐Atlantic divisions, Midwest including East North Central and West North Central divisions, South including South Atlantic, East South Central, and West South Central divisions, and West including Mountain and Pacific divisions). These hospital data were determined by the Healthcare Cost and Utilization Project. Obstetric factors included maternal factors (gestational hypertension, preeclampsia, gestational diabetes mellitus), placental factors (placenta previa and placental abruption), fetal factors (multi‐fetal gestations, fetal growth restriction, fetal anomaly, large for gestational age, fetal breech presentation, and fetal demise), and membranous factors (oligohydramnios, polyhydramnios, preterm premature rupture of membrane, and chorioamnionitis). Delivery characteristics included gestational age at delivery (>40, 39–40, 37–38, 34–36, 28–33, and <28 weeks gestation), cesarean delivery, operative delivery either vacuum or forceps, labor obstruction, abnormal fetal heart rate in labor, and diagnosis of fourth degree laceration including anal sphincter. The analysis in this study was primarily simple descriptive statistics for hypothesis generation. This approach was utilized given the current investigation is an exploration of a database with retrospectively captured information that may include unmeasured confounders. A multivariable regression model was created to assess patient baseline demographics, obstetric factors, and delivery characteristics associated with hypothyroidism status. All the preselected study covariates were entered in the modeling per each stratum. The adjusted‐rate ratio (aRR) and a corresponding 95% confidence interval (CI) for subclinical hypothyroidism comparing to overt hypothyroidism were estimated. The incidence rate of outcome measures was expressed per 1000 deliveries in each hypothyroidism group. The association of exposure and outcome measure was assessed in the multivariable generalized linear model, adjusting for priori clinico‐obstetric factors associated with severe maternal morbidity. These were preselected and included maternal age, race and ethnicity, hypertensive disorder, obesity disorder, placental pathology, gestational age at delivery, and cesarean delivery. 24 Magnitude of statistical strength was expressed with adjusted‐incidence rate ratio (aIR) and a corresponding 95% CI. A number of sensitivity analyses were performed to assess the robustness of study findings. First, severe maternal morbidity was assessed for the individual indicator. This analysis was conducted in a hypothesis‐generating, exploratory fashion. The results were thus intended to raise future study questions rather than guide changes in current clinical practice. Second, composite measures of severe maternal morbidity were assessed using several approaches, removing hysterectomy, adding blood product transfusion, multiple morbidity indicators of two or more, and restricting to six cardiopulmonary morbidity indicators. Blood product transfusion was once included as a severe maternal morbidity indicator per the Centers for Disease Control and Prevention. Third, the exposure–outcome association was evaluated according to the baseline characteristics, assuming that hypothyroidism was a prepregnancy condition or diagnosed in early pregnancy. The characteristics assessed were preselected and included the following four areas: maternal age (<25, 25–29, 30–34, 35–39, and ≥40 years), race and ethnicity (four most frequent groups: Asian, Black, Hispanic, and White), comorbidity (pregestational hypertension, pregestational diabetes mellitus, and obesity disorder), and reproductive factor (prior uterine scar). The adjusting model followed the base analysis. In post hoc, the combination patterns of these assessed baseline characteristics were examined for severe maternal morbidity. Last, the analysis was performed by excluding the unknown cases. Statistical interpretation was based on the 95% confidence interval. The weighted values for national estimates provided by the program were used for the analysis. Statistical Package for Social Sciences (IBM SPSS, version 30.0, Armonk, NY, USA) was used for the analysis. The Strengthening the Reporting of Observational Studies in Epidemiology guidelines were consulted for the performance of this study. 32

Table S1. Coding schema. Table S2. Baseline characteristics. Table S3. Obstetric characteristics. Table S4. Delivery characteristics.