The Impact of Thyroid Disorders on Clinical Pregnancy Outcomes in a Real-World Study Setting

Abstract

Background:

Subclinical hypothyroidism (SCH) and thyroid autoimmunity (TAI) have been associated with poor clinical pregnancy outcomes. However, these outcomes also depend on a number of demographic and obstetric variables. Therefore, the aim of this study was to investigate the impact of thyroid disorders on these outcomes, after adjustment for associated demographic and obstetrical parameters.

Methods:

This is cross-sectional study including 1521 pregnant women who underwent work-up and follow-up in the Centre Hospitalier Universitaire (CHU) Saint-Pierre, Brussels, and had ongoing pregnancies. Thyroid function (thyrotropin [TSH], free thyroxine [fT4]) and TAI (thyroid peroxidase antibodies) was determined at median (Q1-Q3) 13 (11–17) weeks. Baseline parameters and the prevalence of poor clinical pregnancy outcomes were compared between controls (no TAI and TSH <2.51 mIU/L) and three study groups (isolated TAI [TSH <2.51 mIU/L], SCH1 [TSH 2.51–3.7 mIU/L], SCH2 [TSH >3.7 mIU/L]). The impact of the different thyroid groups and demographic/obstetric independent variables on six poor clinical pregnancy outcomes (preeclampsia, intrauterine growth restriction [IUGR], preterm birth, neonatal intensive care unit [NICU] admission, low birth weight, and macrosomia) was investigated in a logistic regression model. Treatment with thyroid hormone before and during pregnancy and assisted and multiple pregnancies were exclusion criteria.

Results:

In total, 79 preeclampsias (5.2%), 40 IUGRs (2.6%), 79 preterm births (5.2%), 10 admissions to NICU (0.66%), 74 low birth weights (4.9%), and 94 babies with macrosomia (6.2%) were documented. TAI was independently associated with NICU admission (adjusted odds ratio [aOR] 16.92 confidence interval [CI 3.36–85.29]; p < 0.001) and TSH, as a continuous variable in the whole range, with preeclampsia (aOR 1.97 [CI 1.18–3.31]; p = 0.010). Trends were present for an association between SCH2 and preeclampsia (aOR 16.73 [CI 1.43–196.42]; p = 0.025), and for SCH1with NICU admission and low birth weight (aOR 19.36 [CI 1.18–316.97]; p = 0.038 and 21.38 [CI 1.29–353.39]; p = 0.032, respectively).

Conclusions:

Pregnant women with TAI had a significantly higher risk of an admission of the baby to the NICU, and SCH tended to be associated with a higher risk of preeclampsia and low birth weight. Other poor clinical pregnancy outcomes were not associated with thyroid disorders, but with demographic and obstetric parameters.

Introduction

Thyroid autoimmunity (TAI) and subclinical hypothyroidism (SCH) have been associated with an increased risk of first trimester miscarriage and preterm birth (1 –3). However, the impact of TAI/SCH on other poor clinical pregnancy outcomes such as preeclampsia, intrauterine growth restriction (IUGR), low birth weight, and admission to neonatal intensive care unit (NICU) remains controversial. Heterogeneity in these results might be due to the absence of (data on) TAI and differences in the study populations (1 –3). Furthermore, and besides demographic data (age, smoking, and obesity), study results are seldom adjusted for covariates such as miscarriage history, a pre-existing hypertensive disorder, gestational diabetes, and others, although these are also risk factors for poor clinical pregnancy outcomes (4 –14). Another issue in studies associating thyroid disorders with pregnancy outcomes is the heterogeneity in the definition of SCH and associated levothyroxine (LT4) treatment during early pregnancy. For many years, the definition of SCH during the first trimester of pregnancy was based on the 2012 Endocrine Society and 2011 American Thyroid Association guidelines (ATA-GL), suggesting an upper cutoff of serum thyrotropin (TSH) of 2.5 mIU/L (15,16). Therefore, women received LT4 whenever they had TSH levels >2.5 mIU/L, independent of the presence of TAI, and this type of treatment might have increased the incidence of preterm delivery and preeclampsia, in women with serum TSH levels <4.0 mIU/L (17,18). As a consequence, in the recent ATA-GL, it has been proposed to determine and use an institutional cutoff for SCH, or when not available, to decrease the upper limit for nonpregnant women by 0.5 mIU/L (19). In the literature, few real-world studies are available investigating the impact of thyroid disorders (SCH/TAI) on pregnancy outcomes in women not treated with LT4, and adjusting the results for demographic and obstetric confounders.

Therefore, the aims of this observational study were to document the prevalence of six poor clinical pregnancy outcomes (preeclampsia, IUGR, preterm delivery, NICU admission, low birth weight, and macrosomia) in pregnant women with different TSH levels, moderate to severe SCH and isolated TAI, and to investigate the impact of these thyroid disorders on those pregnancy outcomes in a logistic regression model, adjusted for demographic and obstetric parameters and gestational ages.

Materials and Methods

Overall study design/definitions

The obstetric clinic is part of a downtown public university hospital (CHU St-Pierre) in Brussels, Belgium. In our center, the first antenatal consultation is systematically completed with a biological analysis including TSH, free thyroxine (fT4), and thyroid peroxidase antibodies (TPO-abs). Furthermore, demographic and obstetrical data (history and follow-up) are noted in a specific obstetric database. Early pregnancy miscarriages are not well coded since it is a heterogeneous outcome (including biochemical and clinical miscarriage) and, therefore, easily underestimated.

Our observational study reports on data of a cross-sectional analysis in pregnant women during the period January 2, 2013, to December 31, 2014, that was nested within the ongoing prospective collection of women's obstetrical parameters and biological data. After the exclusion of women pregnant after assisted reproductive technology (n = 51), multiple pregnancies (n = 77), treated with LT4 before (n = 41) and during pregnancy (n = 139), and finally those with overt hyperthyroidism (n = 3), 1521 women were included in the study.

Figure 1 illustrates the study selection process in a flowchart.

FIG. 1.

Flowchart of the study selection process. fT4, free thyroxine; TPO-abs, thyroid peroxidase antibodies; TSH, thyrotropin.

According to the ATA-GL recommendations, we determined previously first trimester institutional-specific limits for serum TSH and fT4. Our upper limit of serum TSH is 3.74 mIU/L and the lower limit of fT4 is 10.29 pmol/L (19,20). Three study groups were established based on TSH levels, one between 2.51 and 3.74 mIU/L (SCH grade 1 or SCH1), a second with TSH levels >3.74 mIU/L (SCH grade 2 or SCH2) and finally, one including women with TAI only (and TSH levels <2.51 mIU/L). Controls consisted of women without TAI and with TSH levels <2.51 mIU/L.

For clinical pregnancy outcomes, data were recorded in women with ongoing pregnancies starting at the first ultrasound at a median gestational age of 13 weeks. Gestational age was based on ultrasound findings and expressed in full weeks. IUGR was defined as a birthweight <10th percentile for gestational age and sex. Gestational diabetes mellitus (GDM) was present when fasting glycemia was ≥92 mg/dL or when women had a positive 75 g oral glucose tolerance test (OGTT) during the second trimester (from the 26th week of pregnancy on), that is, 1 hour postprandial glycemia (≥180 mg/dL) or at 2 hours ≥153 mg/dL. When women were not able to perform the OGTT (vomiting, other reasons), then a standardized breakfast tolerance test was performed and a two hours postmeal glycemia was measured. During all prenatal consultations, a nonquantitative measurement of proteinuria was done and if after 20 weeks of pregnancy women developed hypertension and/or a clinical suspicion of preeclampsia, then a quantitative proteinuria was determined. Preeclampsia was defined as a systolic blood pressure ≥140 mmHg and/or a diastolic blood pressure ≥90 mmHg, associated with a proteinuria >0.3 g/24 hours after 20 weeks of amenorrhea. Preterm birth was defined as birth <37 weeks of gestation, low birth weight when babies weighed <2.5 kg at term, and macrosomia in case of a birth weight >4.0 kg at term. Smoking was stratified as yes/no (yes meant a minimum of five cigarettes daily and women who stopped smoking during pregnancy were also considered as smokers). Details on the definitions of these outcome measures were provided in a previous study of our group (21).

In all study groups and in the control group, parameters were expressed as continuous values for gestational age at blood sampling (or first prenatal visit), serum TSH, fT4, TPO-abs, maternal age, body mass index (BMI), parity, gestational age at birth, and birth weight. Categorical data were expressed: gestational age at blood sampling (<8 or >12 weeks), TAI (TPO-abs ≥60 kIU/L), maternal age (≥30 years), obesity (BMI ≥30 kg/m²), pre-existing hypertensive disorder, high parity (≥3 children), history of first trimester miscarriages (≥2), tobacco use, preeclampsia, GDM, IUGR, preterm birth (<37 weeks), NICU admission, emergency C-section, low birth weight (<2.5 kg), macrosomia (>4.0 kg), and sex.

The study was approved by the institutional review board (AK/15-11-114/4568).

Serum assay

All provisions were performed by the endocrine laboratory at our institution. Serum TSH, fT4, and TPO-abs levels were measured using the Chemiluminescence Centaur XP Siemens immunoanalyzer. The total imprecision coefficients of variation were 6.9%, 4.2%, and 7.6% for TSH, fT4, and TPO-abs, respectively. For conversion of fT4, 1 ng/dL = 12.9 pmol/L.

In accordance with the ATA-GL, we used our institutional cutoff level for TSH in the first trimester, which is 3.74 mIU/L. The lower limit for fT4 was determined at 10.29 pmol/L (19,20). Normal levels of TPO-abs were <60 kIU/L (in nonpregnant women).

Statistical analysis

Data were stored in a Microsoft Excel database and for the pregnancy outcomes in a specific database. Statistical analyses were performed using Stata 11.2 software (Lakeway drive, TX). Continuous data are expressed as median (Q1-Q3) when not normally distributed and as mean ± SD for normally distributed data. Categorical data are presented as number (percentage) of cases. Differences between study groups were analyzed by Fisher's exact tests for categorical data and by a t-test or Mann–Whitney U test for continuous data whether or not they were distributed normally.

In a univariable logistic regression analysis, the impact of independent variables (high maternal age, obesity, pre-existing hypertensive disorders, tobacco use, high parity, history of miscarriage, gestational age at blood sampling, gestational diabetes, pre-existing hypertensive disorders, and relevant pregnancy/obstetric parameters [preeclampsia, IUGR, emergency C-section, low birth weight, and macrosomia]) was investigated on poor clinical pregnancy outcomes (preeclampsia, IUGR, preterm delivery, NICU admission, low birth weight, and macrosomia). As such, we defined additional confounders beyond those commonly used in the literature (e.g., gestational diabetes as risk factor for macrosomia and preeclampsia for IUGR).

The impact of the independent thyroid variables/groups (all serum TSH levels, SCH [TSH >2.5 mIU/L], SCH1 [TSH 2.51–3.74 mIU/L], SCH2 [TSH >3.74 mIU/L], and isolated TAI [and TSH <2.51 mIU/L]) on the dependent poor clinical outcomes (preeclampsia, IUGR, preterm delivery, NICU admission, low birth weight, and macrosomia) was explored in a logistic regression analysis. The results were systematically adjusted for gestational age at blood sampling and for confounders obtained in the univariable analysis (Table 3). For all TSH groups, a correction for TAI was made as well, and controls were always women without TAI and a TSH of <2.51 mIU/L. The regression analysis was repeated for each group, since the inclusion/exclusion criteria changed according to the TSH group. In the analysis of the SCH1 group (TSH 2.51–3.74 mIU/L), women with TSH levels >3.74 mIU/L were excluded and in that of the SCH2 group women with TSH levels 2.51–3.74 mIU/L.

In Tables 1 and 2, p levels <0.025 (0.05/2) were considered as significant according to a Bonferroni correction for 3 thyroid study groups (TAI+, SCH1, and SCH2), and for the regression analyses (Tables 3 and 4), p levels <0.010 (0.05/5) according to a Bonferroni correction for the 6 poor clinical pregnancy outcomes (preeclampsia, IUGR, preterm delivery, NICU admission, low birth weight, and macrosomia).

Table 1.

Thyroid Parameters and Gestational Age at Blood Sampling in Controls and Different Study Groups

Gestational age and thyroid function/TAI data	All women, n = 1521	Controls	Study group 1	Study group 2	Study group 3	p-level^b
Gestational age and thyroid function/TAI data		TAI− and TSH <2.51 mIU/L, n = 1281	TAI+ and TSH <2.51 mIU/L, n = 81	SCH1 TSH 2.51–3.74 mIU/L, n = 140	SCH2 TSH >3.74 mIU/L, n = 19
Continuous data ^a Categorical data, n (%)		TAI− and TSH <2.51 mIU/L, n = 1281	TAI+ and TSH <2.51 mIU/L, n = 81	SCH1 TSH 2.51–3.74 mIU/L, n = 140	SCH2 TSH >3.74 mIU/L, n = 19
Gestational age at blood sampling (weeks)	13 (11–17)	13 (11–16)	13 (11–17)	15 (12–21)	13 (12–22)	<0.001 SCH1 vs. Ctrl
Gestational age at blood sampling (<8 or >12 weeks) TSH (mIU/L)	903 (59.4%)	731 (57.1%)	54 (66.7%)	103 (73.6%)	15 (79.0%)	<0.001 SCH1 vs. Ctrl
	1.35 (0.85–1.89)	1.18 (0.78–1.67)	1.50 (1.09–1.95)	2.85 (2.66–3.20)	4.15 (3.88–4.39)	<0.001 between all groups
TSH 2.51–3.74 mIU/L	140 (9.2%)	N/A	N/A	140 (100%)	N/A	N/A
TSH >3.74 mIU/L	19 (1.3%)	N/A	N/A	N/A	19 (100%)	N/A
fT4 (pmol/L)	14.2 (12.9–15.4)	14.2 (12.9–15.4)	14.2 (12.9–15.4)	12.9 (11.6–14.2)	12.9 (12.9–14.2)	SCH1 vs. Ctrl <0.001 and vs. TAI +0.005
TPO-abs (kIU/L)	28 (28–37)	28 (28–35)	271 (90–1014)	29 (28–35)	32 (28–42)	p < 0.001 TAI+ vs. other groups
TAI (TPO-abs ≥60 kIU/L)	94 (6.2%)	0 (0%)	81 (100%)	12 (8.6%)	1 (5.3%)	N/A

Continuous data are expressed as mean ± SD or median (Q1-Q3).

p-Values <0.025 were significant according to the Bonferroni correction for three study groups (TAI+, SCH1, and SCH2).

fT4, free thyroxine; N/A, not applicable; SCH, subclinical hypothyroidism; TAI, thyroid autoimmunity; TPO-abs, thyroid peroxidase autoantibodies; TSH, thyrotropin.

Table 2.

Demographic, Obstetric, and Pregnancy Outcome Data in Controls and Different Study Groups

Pregnancy and obstetrical data	All women, n = 1521	Controls	Study group 1	Study group 2	Study group 3	p ^b
Pregnancy and obstetrical data		TAI− and TSH <2.51 mIU/L, n = 1281	TAI+ and TSH <2.51 mIU/L, n = 81	SCH1 TSH 2.51–3.74 mIU/L, n = 140	SCH2 TSH >3.74 mIU/L, n = 19
Continuous data ^a Categorical data, n (%)		TAI− and TSH <2.51 mIU/L, n = 1281	TAI+ and TSH <2.51 mIU/L, n = 81	SCH1 TSH 2.51–3.74 mIU/L, n = 140	SCH2 TSH >3.74 mIU/L, n = 19
Maternal age (years)	30.0 ± 5.8	30.0 ± 5.8	31.7 ± 6.5	28.9 ± 5.7	28.2 ± 4.9	0.009 TAI+ vs. SCH1
Maternal age ≥30 years	786 (51.7%)	670 (52.3%)	48 (59.3%)	61 (43.6%)	7 (36.8%)	ns
BMI (kg/m²)	25.7 ± 4.8	25.7 ± 4.9	24.5 ± 3.7	25.5 ± 4.4	25.8 ± 3.7	ns
Obesity (BMI ≥30 kg/m²)	241 (15.9%)	214 (16.7%)	7 (8.6%)	17 (12.1%)	3 (15.8%)	ns
Pre-existing hypertension	14 (0.9%)	12 (0.9%)	2 (2.5%)	0 (0.0%)	0 (0.0%)	ns
Parity (n)	1 (0–2)	1 (0–2)	1 (0–2)	1 (0–1)	0 (0–1)	0.007 SCH1 vs. Ctrl
Multiparity (≥3)	189 (12.4%)	170 (13.3%)	8 (9.9%)	10 (7.1%)	1 (5.3%)	ns
History of ≥2 first trimester MC	100 (6.6%)	88 (6.9%)	6 (7.4%)	5 (3.6%)	1 (5.3%)	ns
Smoking during pregnancy	228 (15%)	184 (14.4%)	18 (22.2%)	23 (16.4%)	3 (15.8%)	ns
GDM	290 (19.1%)	242 (18.9%)	18 (22.2%)	25 (17.9%)	5 (26.3%)	ns
Emergency C-section	97 (6.4%)	86 (6.7%)	5 (6.2%)	4 (2.9%)	2 (10.5%)	ns
Sex (female/male)	51–49%	53–47%	41–59%	39–61%	42–58%	0.002 SCH1 vs. Ctrl
Preeclampsia	79 (5.2%)	67 (5.2%)	4 (4.9%)	5 (3.6%)	3 (15.8%)	ns
IUGR	40 (2.6%)	36 (2.8%)	1 (1.2%)	2 (1.4%)	1 (5.3%)	ns
Gestational age at birth (weeks)	39.6 (38.5–40.4)	39.6 (38.5–40.4)	40.0 (38.8–40.5)	39.6 (39.0–40.5)	39.5 (38.5–40.2)	ns
Preterm birth (<37 weeks)	79 (5.2%)	64 (5.0%)	6 (7.4%)	9 (6.4%)	0 (0%)	ns
Admission to neonatal ICU	10 (0.7%)	5 (0.4%)	3 (3.7%)	2 (1.4%)	0 (0%)	0.009 TAI+ vs. Ctrl
Birth weight (kg)	3.3 ± 0.5	3.3 ± 0.5	3.3 ± 0.5	3.3 ± 0.6	3.3 ± 0.5	ns
Low birth weight (<2.5 kg)	74 (4.9%)	60 (4.7%)	6 (7.4%)	8 (5.7%)	0 (0%)	ns
Macrosomia (>4.0 kg)	94 (6.2%)	74 (5.8%)	5 (6.2%)	12 (8.6%)	3 (15.8%)	ns

Continuous data are expressed as mean ± SD or median (Q1-Q3).

p-Values <0.025 were significant according to the Bonferroni correction for three study groups (TAI+, SCH1, and SCH2).

BMI, body mass index; GDM, gestational diabetes mellitus; ICU, intensive care unit; IUGR, intrauterine growth restriction; MC, miscarriage.

Table 3.

Univariable Logistic Regression Analysis with Demographic Obstetric/Pregnancy Parameters as Categorical Independent Variables, and Major Poor Pregnancy Outcomes as Dependent Variables

Dependent outcomes	Preeclampsia	IUGR	Preterm birth	NICU admission	Low birth weight	Macrosomia
Independent variables	OR	OR	OR	OR	OR	OR
	[95% CI]	[95% CI]	[95% CI]	[95% CI]	[95% CI]	[95% CI]
	p	p	p	p	p	p
High maternal age (≥30 years)	1.06	1.27	1.25	3.76	1.10	1.11
	[0.67–1.67]	[0.68–2.37]	[0.79–1.97]	[0.74–26.51]	[0.69–1.76]	[0.73–1.69]
	0.786	0.455	0.334	0.110	0.675	0.606
Obesity (BMI ≥30 kg/m²)	3.36	1.13	1.04	1.33	0.82	1.91
	[2.08–5.42]	[0.50–2.53]	[0.57–1.91]	[0.13–6.72]	[0.42–1.60]	[1.18–3.09]
	<0.001	0.771	0.879	1.000	0.573	0.008
Pre-existing hypertensive disorder	280	6.42	7.61	7.89	8.18	0.76
	[41.00–12068]	[0.67–30.41]	[1.70–27.17]	[0.00–52.98]	[1.82–29.26]	[0.00–4.61]
	<0.001	0.050	0.005	1.000	0.004	0.621
Tobacco use	0.91	1.20	1.23	2.44	1.74	0.73
	[0.48–1.73]	[0.53–2.71]	[0.69–2.22]	[0.40–10.81]	[1.01–3.04]	[0.39–1.39]
	0.785	0.652	0.485	0.179	0.049^*	0.357
High parity rate (≥3)	1.28	0.56	1.14	1.77	0.97	1.73
	[0.68–2.39]	[0.18–1.74]	[0.60–2.19]	[0.18–8.95]	[0.48–1.96]	[1.02–2.96]
	0.444	0.339	0.679	0.623	0.944	0.041^*
History of miscarriages (≥2)	1.90	0.74	1.90	1.58	1.26	1.15
	[0.93–3.89]	[0.09–2.94]	[0.93–3.89]	[0.03–11.63]	[0.43–3.01]	[0.52–2.51]
	0.076	0.770	0.076	1.000	0.627	0.725
Gestational age at blood sampling (<8 or >12 weeks)	1.26	0.92	1.12	0.68	0.94	0.96
	[0.78–2.01]	[0.49–1.72]	[0.70–1.78]	[0.15–2.98]	[0.59–1.51]	[0.63–1.46]
	0.335	0.807	0.623	0.749	0.821	0.861
Gestational diabetes	0.99	0.46	1.08	1.82	0.70	1.49
	[0.56–1.75]	[0.17–1.26]	[0.61–1.89]	[0.51–6.53]	[0.38–1.39]	[0.92–2.41]
	0.985	0.139	0.783	0.378	0.346	0.099
Preeclampsia	—	19.04	7.29	4.64	12.14	0.12
		[9.79–37.06]	[4.11–12.92]	[0.47–23.68]	[6.99–21.11]	[0.00–0.70]
		<0.001	<0.001	0.092	<0.001	0.026^*
IUGR	—	—	10.32	4.18	42.93
			[5.15–10.71]	[0.09–31.52]	[21.64–85.12]
			<0.001	0.235	<0.001
Preterm birth (<37 weeks)	—	—	—	4.64	33.71	—
				[0.47–23.83]	[22.75–49.96]
				0.092	<0.001
Emergency C-section	—	—	—	3.72	—	—
				[0.37–19.00]
				0.130
Low birth weight (<2.5 kg)	—	—	101.00	8.66	—	—
			[55.00–186.00]	[1.41–38.92]
			<0.001	0.010
Macrosomia (>4.0 kg)	—	—	—	3.84	—	—
				[0.39–19.67]
				0.123

Bold indicates p values are significant.

p-Values <0.010 were significant according to the Bonferroni correction for six poor pregnancy outcomes (preeclampsia, IUGR, preterm birth, NICU, low birth weight, and macrosomia).

CI, confidence interval; NICU, neonatal intensive care unit; OR, odds ratio.

Table 4.

Logistic Regression Analysis with Different Thyroid Groups as Independent Variables and Poor Pregnancy Outcomes as Categorical Dependent Variables

Dependent outcomes	Preeclampsia	IUGR	Preterm birth	NICU admission	Low birth weight	Macrosomia
Independent thyroid variables	aOR ^a [95% CI] p	aOR ^b [95% CI] p	aOR ^c [95% CI] p	aOR ^d [95% CI] p	aOR ^e [95% CI] p	aOR ^f [95% CI] p
Isolated TAI+ (TSH <2.50 mIU/L)	0.80[0.20–3.15]0.752	0.80[0.12–5.46]0.816	N/A	16.92 [3.36–85.29] <0.001	3.51[0.40–30.47]0.255	0.81[0.25–2.65]0.726
TAI− and TSH all range	1.97 [1.18–3.31] 0.010	0.59[0.24–1.47]0.259	0.80[0.30–2.10]0.647	1.49[0.59–3.80]0.402	2.65[0.82–8.61]0.105	1.45[0.83–2.53]0.195
TAI− and TSH >2.50 mIU/L ( = SCH)	1.16[0.15–9.13]0.891	N/A	N/A	16.86[1.03–275.2]0.047^*	19.44[1.18–320.3]0.038^*	1.85[0.40–8.58]0.432
TAI− and TSH 2.51–3.74 mIU/L ( = SCH1)	N/A	N/A	N/A	19.36[1.18–316.9]0.038^*	21.38[1.29–353.4]0.032^*	1.93[0.42–8.99]0.401
TAI− and TSH >3.74 mIU/L ( = SCH2)	16.73[1.43–196.42]0.025^*	N/A	N/A	N/A	N/A	N/A

Bold indicates p values are significant.

p-Values <0.010 were significant according to the Bonferroni correction for six poor pregnancy outcomes (preeclampsia, IUGR, preterm birth, NICU, low birth weight and macrosomia).

All outcomes were corrected for gestational age at blood sampling, thyroid autoimmunity (besides the “TAI” group), and furthermore, for each outcome group separately with:

Obesity and pre-existing hypertensive disorder.

Preeclampsia.

Preeclampsia, pre-existing hypertensive disorder, IUGR, low birth weight.

Low birth weight.

Preeclampsia, pre-existing hypertensive disorder, IUGR, tobacco use, preterm birth.

Preeclampsia, high parity, obesity, gestational diabetes.

aOR, adjusted odds ratio; N/A, not applicable (too low number of events); TAI, thyroid autoimmunity (TPO-abs ≥60 kIU/L).

Otherwise, all statistical tests were considered significant when p < 0.05.

Results

Table 1 gives thyroid parameters and gestational ages at blood sampling in controls and different study groups.

The median (Q1-Q3) time of blood sampling was higher in women with SCH1 than controls (15 [12–21] weeks vs. 13 [11–17] weeks; p < 0.001). This was also the case when expressed as the prevalence of women who underwent their blood sampling earlier than 8 or later than 12 weeks of gestation (73.6% vs. 57.1%; p < 0.001). By definition, serum TSH levels were higher in the SCH1 and SCH2 groups than controls (p < 0.001). fT4 levels were significantly lower in the SCH1 group than those in controls and women with only TAI (12.9 [11.6–14.2] pmol/L vs. 14.2 [12.9–15.4] pmol/L in both groups; p < 0.001). In the SCH1 group, there was no significant difference in serum TSH levels between women screened in or outside the period of pregnancy weeks 8–12 (data not shown).

Table 2 shows the demographic, obstetric, and pregnancy outcome data in controls and different study groups.

The mean maternal age was higher in women with TAI than in women in the SCH1 group (31.7 ± 6.5 years vs. 28.9 ± 5.7 years; p = 0.009). The rate of NICU admissions was significantly higher in women with TAI (n = 3) than that in controls (n = 5); 3.7% versus 0.4% (p = 0.009). Three women (of whom none with TAI) had a placental abruption; one of the babies was admitted to the NICU (data not shown).

Table 3 gives the results (odds ratio [OR] confidence interval [CI]) of the univariable logistic regression analysis with demographic obstetric/pregnancy parameters as categorical independent variables, and major poor pregnancy outcomes as dependent variables.

The strongest associations with preeclampsia were (a) pre-existing hypertensive disorder (OR 280 [CI 41–12068.00] and obesity [OR 3.36 CI 2.08–5.42]; both p < 0.001); (b) with NICU admission low birth weight (OR 8.66 [CI 1.41–38.92]; p = 0.010); (c) with low birth weight pre-existing hypertensive disorder (OR 8.18 [CI 1.82–29.26]; p = 0.004), preeclampsia (OR 12.14 [CI 6.99–21.11]; p < 0.001), IUGR (OR 42.93 [CI 21.64–85.12]; p < 0.001), and preterm birth (OR 33.71 [CI 22.75–49.96]; p < 0.001). Further details are listed in Table 3.

Table 4 gives the results (adjusted odds ratio [aOR] [CI]) of the logistic regression analysis with different thyroid groups as independent variables and poor pregnancy outcomes as categorical dependent variables. The presence of TAI was independently associated with NICU admission (aOR 16.92 [CI 3.36–85.29]; p < 0.001). TSH as a continuous variable was associated with preeclampsia (aOR 1.97 [CI 1.18–3.31]; p = 0.010).

SCH (all TSH >2.50 mIU/L) tended to be associated with NICU admission and low birth weight (aOR 16.86 [CI 1.03–275.17]; p = 0.047 and aOR 19.44 [CI 1.18–320.28]; p = 0.038, respectively).

SCH1 (TSH 2.51–3.74 mIU/L) tended to be associated with NICU admission and low birth weight (aOR 19.36 [CI 1.18–316.97]; p = 0.038 and aOR 21.38 [CI 1.29–353.39]; p = 0.032, respectively).

SCH2 (TSH >3.74 mIU/L) tended to be associated with preeclampsia (aOR 16.73 [CI 1.43–196.42]; p = 0.025).

Discussion

The main finding in the study presented here is the independent association between TAI and admission to the NICU (NICU admission), and furthermore, a trend between (high) serum TSH/SCH, preeclampsia, and low birth weight.

The prevalence of NICU admission varies between 5% and 18% in the OB/GYN literature, and many maternal and fetal conditions are associated, such as labor induction with prostaglandin analogues, length of first stage of labor, prolonged rupture of membranes, preeclampsia, the use of IV antihypertensives, maternal fever, and low birth weight (13,14). We also noted a significant association between low birth weight and NICU admission in our study. However, in none of the previous cited studies on NICU admission, thyroid (dys)function or TAI was taken into account as a potential risk factor. In the endocrine literature, the prevalence of NICU admission varies between 0.6% and 20.7% depending on the presence of thyroid dysfunction (22 –28). The prevalence in our study was within that range with 3.7%. Hyperthyroidism, overt hypothyroidism, and SCH have been associated with an increased risk of NICU admission in some studies (22 –26), but not in others (1,27,28). Furthermore, in studies investigating overt and SCH in relation to NICU admission, it was not always clear to what extend the presence of TAI played a role, sometimes because it was not investigated, the number of women with TAI was too low, or because it was an exclusion criterion (2,23,25,26). Also, associations between thyroid disorders and NICU admission were not corrected for covariates known to be associated with it (e.g., low birth weight). In our study, TAI was an independent risk factor for NICU admission even in the presence of a normal thyroid function (TSH <2.51 mIU/L and absent hyperthyroidism) and after correction for low birth weight. The underlying reasons for the association between TAI and NICU admission remain speculative. However, explanations could be provided by the fact that TAI is associated with higher serum TSH levels on one hand and an immune imbalance on the other hand, conditions that can be accentuated and co-occur during pregnancy (1,3). Indeed, SCH and SCH1 also tended to be associated with an increased prevalence of NICU admission. Recent data have also shown that TAI (increased TPO-abs) is associated with an attenuated thyroidal response to human chorionic gonadotropin (hCG) stimulation and that it may exert effects on pregnancy outcomes independently of thyroid function through a disturbed humoral and innate immunity, cross reactivity between TPO-abs and extrathyroidal sites, and, among others, the presence of anticardiolipin antibodies (29 –31). Pregnancy affects the maternal immune system to maintain tolerance of the fetus, with a switch to a predominantly T-helper-2-type cytokine profile, and an important role of the regulatory T cells (32). These alterations might be impaired in women with TAI and lead to pregnancy complications (32). Placental abruption is a rare complication that has been associated with isolated TAI and might contribute to the decision to admit the neonate to the NICU (27,33). In our study, three women had a placental abruption and one baby of which was admitted to the NICU. However, none of the women had TAI. In a recent study in women with hypothyroidism and TAI, it was shown that LT4 treatment could reduce the prevalence of NICU admissions in the group of women with serum TSH levels >4 mIU/L, a decrease that was obtained through a lower number of preterm deliveries (24). The absence of positive effects of LT4 treatment in women with TAI and TSH levels <4 mIU/L and together with the results presented here are arguments in favor of a non-TSH–dependent effect of TAI. Finally, NICU admissions might not be an optimal pregnancy outcome measurement, since the decision to admit a newborn to the NICU is not straightforward and subject to clinical judgment. In daily practice, pregnant women with known TAI should aim to have a serum TSH <2.5 mIU/L according to the recent ATA-GL (19). In women with TAI and a normal serum TSH (<2.5 mIU/L), other preventive measures could be considered such as shortening the duration of the first stage of labor, as well as other measures as suggested by Burges et al. (14). The results of our study might add arguments in favor of systematically measuring TPO-abs in pregnant women, although screening for thyroid dysfunction itself is a topic of ongoing debate, with some proposing systematic screening and others only targeted screening of women at high risk (19,34). In the ATA-GL, it is mentioned that there is insufficient evidence to recommend for or against universal screening for abnormal TSH concentrations in early pregnancy and that universal screening for TPO-abs in early pregnancy (or preconception) may also prove to be an attractive alternative; however, there is no data to support the latter approach at present and further investigation is needed (19). In the gynecological literature, the 2015 recommendation of the American College of Obstetricians and Gynecologists is firmly against universal screening (35).

An additional question that arises is whether thyroglobulin antibodies (Tg-abs) should be measured in addition to TPO-abs. In a study performed in Brussels by another research group, it was shown that Tg-abs are increased independently of TPO-abs in 5% of women, and associated with higher serum TSH levels (36). In contrast to the ATA-GL, the guidelines from the European Thyroid Association recommend the measurement of Tg-abs in case of elevated serum TSH levels and negative TPO-abs (19,37). However, in a recent study, TSH reference ranges determined during pregnancy were comparable between women with and without Tg-abs, but no pregnancy outcomes were investigated (38). In a Chinese study, it has been shown that increased Tg-abs and not TPO-abs were associated with miscarriage, but in a Danish study, no impact of Tg-abs was noted on pregnancy outcomes (39,40). Other poor clinical pregnancy outcomes investigated in the study presented here (preeclampsia, preterm birth, IUGR, low birth weight, and macrosomia) were not associated with the presence of TAI. In a meta-analysis in 2011, and in another study published later, isolated TAI was associated with preterm delivery (OR 1.4–1.7) (41,42), although this was not confirmed in a more recent study (43). Differences between studies might be due to other designs and definitions of positivity for thyroid antibodies. Indeed, in a recent study it has been shown that preterm delivery rates were increased at antibody levels that were below the cutoff recommended by the manufacturer (44). It is also important to emphasize that outcomes such as preterm birth and macrosomia were more strongly associated with nonthyroidal parameters such as pre-existing hypertensive disorders and obesity, thereby providing less weight on the importance of thyroid disorders.

Another observation made in our study is a positive linear relationship between all TSH levels and preeclampsia, and a trend with SCH2. Furthermore, SCH and SCH1 were closely related with the appearance of low birth weight. Preeclampsia is an outcome that is dependent on many variables and present in 3–5% of pregnancies (4,5). The prevalence in our study was 5.2% in the whole study group, but it tripled to 15.8% in women with SCH2. Also, after correction for obesity and pre-existing hypertension, SCH2 tended to remain associated with preeclampsia. We did not observe an association with preeclampsia when the TSH cutoff was set >2.5 mIU/L, in analogy with the results in the study by Plowden et al. (43). In a meta-analysis published in 2016, in which most studies defined SCH as a serum TSH >3.5 mIU/L, SCH was significantly associated with preeclampsia (relative risk 1.3) (2). However, it should be mentioned that preeclampsia has also been associated with overt hypo- and hyperthyroidism (OR 1.38 and 1.8, respectively), and that in another study, women with TSH levels in the high-normal range but <4.0 mIU/L before treatment and treated with LT4 had an increased risk of preeclampsia (17,23,25). The link between thyroid dysfunction and preeclampsia has been explained in detail in a study by Wilson et al. (45). In summary, thyroid hormones can have cardiovascular effects (genomic and nongenomic), and lead to endothelial dysfunction. More recent studies also assign a role to the concentration of serum hCG levels in the pathophysiology of preeclampsia, which could explain the association with both hypo- and hyperthyroidism. Women with thyroid dysfunction, but high serum hCG levels (reflecting a normal functioning placenta) seem to have a preeclampsia risk that is comparable with that in euthyroid women (46). Finally, SCH could be a consequence of preeclampsia and not the cause of it. In two studies, placental antiangiogenic factors (such as the soluble FMS-like tyrosine kinase-1) known to be associated with preeclampsia were shown to have an impact on the highly vascularized thyroid gland, and, as a consequence, lead to SCH independent of the presence of TAI (47,48).

Remarkably, in reviews on preeclampsia, thyroid dysfunction is not even mentioned as a risk factor (4,5). This may imply that gynecologists do not consider thyroid dysfunction as an important risk factor, or that it is unknown to them; this may contribute to the views of obstetric societies that there is insufficient evidence to perform systematic screening for thyroid disorders in pregnant women (35). Low birth weight tended to be associated with SCH and SCH1, independently of preeclampsia, preterm delivery, tobacco use, and IUGR. Moreover, in the study by Karakosta et al., SCH was also associated with low birth weight with and without correction for preterm birth (42). However, in two recent meta-analyses, SCH was not associated with low birth weight (1,2). Furthermore, in some studies, low birth weight was associated rather with hyperthyroidism than with SCH (3,23). Mechanisms underlying the association between SCH and low birth weight remain speculative, but it is known that thyroid hormone is essential for fetal growth and maturation of many target tissues through the actions of GH and IGF-1, and they might interfere with cardiovascular homeostasis in utero (3,49).

Strengths of our study are its real-world setting and the classification of thyroid disorders in groups separating TAI and thyroid dysfunction, which allowed us to gain more insight into the recurring question whether TAI or SCH is associated with poor pregnancy outcomes. Furthermore, this is one of the few studies in which pregnancy outcomes in relation to thyroid disorders were adjusted for pregnancy/obstetric-related conditions in addition to the regular corrections for demographic parameters. We defined SCH according to our institutional cutoff, and our data set provides insight into the consequences of SCH in women not treated with LT4, despite the recommendations at that period (15,16,19). These strengths of our study are in contrast also limitations. The reasons why women with SCH were not treated was not always clear (some refused treatment and others with higher serum TSH were TAI negative). Only a low number of women were included in the SCH2 group, which was due to the fact that most of these women were treated with LT4. Women in the SCH1 group were screened rather late during pregnancy (due to the real-world setting), but the outcomes were corrected for gestational age at blood sampling. A number of women with (controlled) chronic disease (HIV, hepatitis B/C) were included in the analysis, without adjusting for it. And finally, categorizing women into groups according to TSH levels that were based on a single thyroid function test determination can also be a limitation, although that is probably of lesser importance in the context of TAI.

Conclusions

Pregnant women with TAI had a significantly higher risk of an NICU admission of their baby, and SCH tended to be associated with a higher risk of preeclampsia and low birth weight. Other poor clinical pregnancy outcomes were not associated with thyroid disorders, but with demographic and obstetric parameters. Prospective studies including a large number of women are requested to confirm our results, to unravel underlying mechanisms, and ultimately leading to focused treatments besides LT4 in case of hypothyroidism.

Footnotes

Author Disclosure Statement

K.P. received lecture fees from the IBSA Institut Biochimique SA (satellite meeting of the European Thyroid Association) in 2016 and the Berlin-Chemie AG company (ETA educational thyroid meeting) in 2017 and 2018.

Funding Information

No funding was received for this article.

References

Chan

, Boelaert

. 2015. Optimal management of hypothyroidism, hypothyroxinaemia and euthyroid TPO antibody positivity preconception and in pregnancy. Clin Endocrinol (Oxf), 82:313–326.

Maraka

, Ospina

, O'Keeffe

, Espinosa De Ycaza

, Gionfriddo

, Erwin

, Coddington

3rd , Stan

, Murad

, Montori

. 2016. Subclinical hypothyroidism in pregnancy: a systematic review and meta-analysis. Thyroid, 26:580–590.

Korevaar

TIM

, Medici

, Visser

, Peeters

. 2017. Thyroid disease in pregnancy: new insights in diagnosis and clinical management. Nat Rev Endocrinol, 13:610–622.

Bartsch

, Medcalf

, Park

, Ray

; High Risk of Pre-eclampsia Identification Group. 2016. Clinical risk factors for pre-eclampsia determined in early pregnancy: systematic review and meta-analysis of large cohort studies. BMJ, 353:i1753.

Mol

BWJ

, Roberts

, Thangaratinam

, Magee

, de Groot

CJM

, Hofmeyr

. 2016. Pre-eclampsia. Lancet, 387:999–1011.

Lausman

, Kingdom

; Maternal Fetal Medicine Committee. 2013. Intrauterine growth restriction: screening, diagnosis, and management. J Obstet Gynaecol Can, 35:741–748.

Figueras

, Gardosi

. 2011. Intrauterine growth restriction: new concepts in antenatal surveillance, diagnosis, and management. Am J Obstet Gynecol, 204:288–300.

Romero

, Dey

, Fisher

. 2014. Preterm labor: one syndrome, many causes. Science, 345:760–765.

Keelan

, Newnham

. 2017. Recent advances in the prevention of preterm birth. F1000Res, 6. DOI: 10.12688/f1000research.11385.1.

10.

Ogawa

, Matsuda

, Kanda

, Konno

, Mitani

, Makino

, Matsui

. 2013. Survival rate of extremely low birth weight infants and its risk factors: case-control study in Japan. ISRN Obstet Gynecol, 2013:873563.

11.

Cutland

, Lackritz

, Mallett-Moore

, Bardaji

, Chandrasekaran

, Lahariya

, Nisar

, Tapia

, Pathirana

, Kochhar

, Munoz

, Brighton Collaboration Low Birth Weight Working G. 2017. Low birth weight: case definition & guidelines for data collection, analysis, and presentation of maternal immunization safety data. Vaccine, 35:6492–6500.

12.

Araujo Junior

, Peixoto

, Zamarian

, Elito Junior

, Tonni

. 2017. Macrosomia. Best Pract Res Clin Obstet Gynaecol, 38:83–96.

13.

Alkiaat

, Hutchinson

, Jacques

, Sharp

, Dickinson

. 2013. Evaluation of the frequency and obstetric risk factors associated with term neonatal admissions to special care units. Aust N Z J Obstet Gynaecol, 53:277–282.

14.

Burgess

, Katz

, Pessolano

, Ponterio

, Moretti

, Lakhi

. 2016. Determination of antepartum and intrapartum risk factors associated with neonatal intensive care unit admission. J Perinat Med, 44:589–596.

15.

Stagnaro-Green

, Abalovich

, Alexander

, Azizi

, Mestman

, Negro

, Nixon

, Pearce

, Soldin

, Sullivan

, Wiersinga

; American Thyroid Association Taskforce on Thyroid Disease During Pregnancy and Postpartum. 2011. Guidelines of the American Thyroid Association for the diagnosis and management of thyroid disease during pregnancy and postpartum. Thyroid, 21:1081–1125.

16.

De Groot

, Abalovich

, Alexander

, Amino

, Barbour

, Cobin

, Eastman

, Lazarus

, Luton

, Mandel

, Mestman

, Rovet

, Sullivan

. 2012. Management of thyroid dysfunction during pregnancy and postpartum: an Endocrine Society clinical practice guideline. J Clin Endocrinol Metab, 97:2543–2565.

17.

Maraka

, Mwangi

, McCoy

, Yao

, Sangaralingham

, Singh Ospina

, O'Keeffe

, De Ycaza

, Rodriguez-Gutierrez

, Coddington

3rd , Stan

, Brito

, Montori

. 2017. Thyroid hormone treatment among pregnant women with subclinical hypothyroidism: US national assessment. BMJ, 356:i6865.

18.

Rodriguez-Gutierrez

, Maraka

, Ospina

, Montori

, Brito

. 2017. Levothyroxine overuse: time for an about face?. Lancet Diabetes Endocrinol, 5:246–248.

19.

Alexander

, Pearce

, Brent

, Brown

, Chen

, Dosiou

, Grobman

, Laurberg

, Lazarus

, Mandel

, Peeters

, Sullivan

. 2017. Guidelines of the American Thyroid Association for the diagnosis and management of thyroid disease during pregnancy and the postpartum. Thyroid, 27:315–389.

20.

Veltri

, Belhomme

, Kleynen

, Grabczan

, Rozenberg

, Pepersack

, Poppe

. 2017. Maternal thyroid parameters in pregnant women with different ethnic backgrounds: do ethnicity-specific reference ranges improve the diagnosis of subclinical hypothyroidism?. Clin Endocrinol (Oxf), 86:830–836.

21.

Veltri

, Kleynen

, Grabczan

, Salajan

, Rozenberg

, Pepersack

, Poppe

. 2017. Pregnancy outcomes are not altered by variation in thyroid function within the normal range in women free of thyroid disease. Eur J Endocrinol, 178:189–197.

22.

van den Boogaard

, Vissenberg

, Land

, van Wely

, van der Post

, Goddijn

, Bisschop

. 2011. Significance of (sub)clinical thyroid dysfunction and thyroid autoimmunity before conception and in early pregnancy: a systematic review. Hum Reprod Update, 17:605–619.

23.

Mannisto

, Mendola

, Grewal

, Xie

, Chen

, Laughon

. 2013. Thyroid diseases and adverse pregnancy outcomes in a contemporary US cohort. J Clin Endocrinol Metab, 98:2725–2733.

24.

Nazarpour

, Ramezani Tehrani

, Simbar

, Tohidi

, Alavi Majd

, Azizi

. 2017. Effects of levothyroxine treatment on pregnancy outcomes in pregnant women with autoimmune thyroid disease. Eur J Endocrinol, 176:253–265.

25.

Turunen

, Vaarasmaki

, Mannisto

, Hartikainen

, Lahesmaa-Korpinen

, Gissler

, Suvanto

. 2019. Pregnancy and perinatal outcome among hypothyroid mothers: a population-based cohort study. Thyroid, 29:135–141.

26.

Casey

, Dashe

, Wells

, McIntire

, Byrd

, Leveno

, Cunningham

. 2005. Subclinical hypothyroidism and pregnancy outcomes. Obstet Gynecol, 105:239–245.

27.

Abbassi-Ghanavati

, Casey

, Spong

, McIntire

, Halvorson

, Cunningham

. 2010. Pregnancy outcomes in women with thyroid peroxidase antibodies. Obstet Gynecol, 116:381–386.

28.

Negro

, Schwartz

, Gismondi

, Tinelli

, Mangieri

, Stagnaro-Green

. 2011. Thyroid antibody positivity in the first trimester of pregnancy is associated with negative pregnancy outcomes. J Clin Endocrinol Metab, 96:E920–E924.

29.

Korevaar

, Steegers

, Pop

, Broeren

, Chaker

, de Rijke

, Jaddoe

, Medici

, Visser

, Tiemeier

, Peeters

. 2017. Thyroid autoimmunity impairs the thyroidal response to human chorionic gonadotropin: two population-based prospective cohort studies. J Clin Endocrinol Metab, 102:69–77.

30.

Kim

, Cho

, Kim

, Yang

, Ahn

, Thornton

, Park

, Beaman

, Gilman-Sachs

, Kwak-Kim

. 2011. Thyroid autoimmunity and its association with cellular and humoral immunity in women with reproductive failures. Am J Reprod Immunol, 65:78–87.

31.

Twig

, Shina

, Amital

, Shoenfeld

. 2012. Pathogenesis of infertility and recurrent pregnancy loss in thyroid autoimmunity. J Autoimmun, 38:J275–J281.

32.

Turhan Iyidir

, Konca Degertekin

, Sonmez

, Atak Yucel

, Erdem

, Akturk

, Ayvaz

. 2015. The effect of thyroid autoimmunity on T-cell responses in early pregnancy. J Reprod Immunol, 110:61–66.

33.

Nazarpour

, Ramezani Tehrani

, Simbar

, Azizi

. 2016. Thyroid autoantibodies and the effect on pregnancy outcomes. J Obstet Gynaecol, 36:3–9.

34.

Sitoris

, Veltri

, Kleynen

, Belhomme

, Rozenberg

, Poppe

. 2019. Screening for thyroid dysfunction in pregnancy with targeted high-risk case finding: can it be improved?. J Clin Endocrinol Metab, 104:2346–2354.

35.

American College of Obstetricians and Gynecologists. 2015. Practice Bulletin No. 148: thyroid disease in pregnancy. Obstet Gynecol, 125:996–1005.

36.

Unuane

, Velkeniers

, Anckaert

, Schiettecatte

, Tournaye

, Haentjens

, Poppe

. 2013. Thyroglobulin autoantibodies: is there any added value in the detection of thyroid autoimmunity in women consulting for fertility treatment?. Thyroid, 23:1022–1028.

37.

Lazarus

, Brown

, Daumerie

, Hubalewska-Dydejczyk

, Negro

, Vaidya

. 2014. 2014 European thyroid association guidelines for the management of subclinical hypothyroidism in pregnancy and in children. Eur Thyroid J, 3:76–94.

38.

Andersen

, Andersen

, Carle

, Christensen

, Handberg

, Karmisholt

, Knosgaard

, Kristensen

, Bulow Pedersen

, Vestergaard

. 2019. Pregnancy week-specific reference ranges for thyrotropin and free thyroxine in the North Denmark Region Pregnancy Cohort. Thyroid, 29:430–438.

39.

Liu

, Shan

, Li

, Mao

, Xie

, Wang

, Fan

, Wang

, Zhang

, Han

, Wang

, Liu

, Fan

, Bao

, Teng

. 2014. Maternal subclinical hypothyroidism, thyroid autoimmunity, and the risk of miscarriage: a prospective cohort study. Thyroid, 24:1642–1649.

40.

Bliddal

, Boas

, Hilsted

, Friis-Hansen

, Tabor

, Feldt-Rasmussen

. 2015. Thyroid function and autoimmunity in Danish pregnant women after an iodine fortification program and associations with obstetric outcomes. Eur J Endocrinol, 173:709–718.

41.

, Wang

, He

, Xu

, Wang

. 2012. Thyroid antibodies and risk of preterm delivery: a meta-analysis of prospective cohort studies. Eur J Endocrinol, 167:455–464.

42.

Karakosta

, Alegakis

, Georgiou

, Roumeliotaki

, Fthenou

, Vassilaki

, Boumpas

, Castanas

, Kogevinas

, Chatzi

. 2012. Thyroid dysfunction and autoantibodies in early pregnancy are associated with increased risk of gestational diabetes and adverse birth outcomes. J Clin Endocrinol Metab, 97:4464–4472.

43.

Plowden

, Schisterman

, Sjaarda

, Perkins

, Silver

, Radin

, Kim

, Galai

, DeCherney

, Mumford

. 2017. Thyroid-stimulating hormone, anti-thyroid antibodies, and pregnancy outcomes. Am J Obstet Gynecol, 217:697e691.e1–e697.e7.

44.

Korevaar

TIM

, Pop

, Chaker

, Goddijn

, de Rijke

, Bisschop

, Broeren

, Jaddoe

VWV

, Medici

, Visser

, Steegers

EAP

, Vrijkotte

, Peeters

. 2018. Dose dependency and a functional cutoff for TPO-antibody positivity during pregnancy. J Clin Endocrinol Metab, 103:778–789.

45.

Wilson

, Casey

, McIntire

, Halvorson

, Cunningham

. 2012. Subclinical thyroid disease and the incidence of hypertension in pregnancy. Obstet Gynecol, 119:315–320.

46.

Korevaar

, Steegers

, Chaker

, Medici

, Jaddoe

, Visser

, de Rijke

, Peeters

. 2016. The risk of preeclampsia according to high thyroid function in pregnancy differs by hCG concentration. J Clin Endocrinol Metab, 101:5037–5043.

47.

Venkatesha

, Toporsian

, Lam

, Hanai

, Mammoto

, Kim

, Bdolah

, Lim

, Yuan

, Libermann

, Stillman

, Roberts

, D'Amore

, Epstein

, Sellke

, Romero

, Sukhatme

, Letarte

, Karumanchi

. 2006. Soluble endoglin contributes to the pathogenesis of preeclampsia. Nat Med, 12:642–649.

48.

Levine

, Vatten

, Horowitz

, Qian

, Romundstad

, Yu

, Hollenberg

, Hellevik

, Asvold

, Karumanchi

. 2009. Pre-eclampsia, soluble fms-like tyrosine kinase 1, and the risk of reduced thyroid function: nested case-control and population based study. BMJ, 17:339:b4336.

49.

Moreh-Waterman

, Miller-Lotan

, Tamir

, Hochberg

. 2003. Maternal hypothyroidism may affect fetal growth and neonatal thyroid function. Obstet Gynecol, 102:232–241.