Maternal TPO-Ab status modulates associations between gestational thyroid hormones and neonatal TSH in 10,154 Chinese mother

Abstract

Background:

Evidence on the maternal–fetal thyroid hormone (TH) axis is limited, especially regarding how maternal thyroid peroxidase antibody (TPO-Ab) status affects the link between gestational thyroid function and neonatal thyroid-stimulating hormone (TSH).

Objectives:

This study explored how maternal TH in late pregnancy influences neonatal TSH and assessed whether TPO-Ab modifies this relationship.

Design:

This retrospective cohort study was conducted at a tertiary specialized hospital.

Methods:

A total of 10,154 eligible mother–newborn pairs were analyzed. Maternal TSH, free thyroxine (FT4), free triiodothyronine (FT3), and TPO-Ab were measured before birth; neonatal TSH was measured 3–7 days postbirth. Mothers were classified as TPO-Ab positive (>34 IU/mL) or negative.

Results:

Maternal TPO-Ab positivity prevalence was 5.5%, while congenital hypothyroidism (CH) screening positivity (TSH ⩾10 mIU/L) occurred in 0.87% of neonates. Among TPO-Ab-negative mothers, maternal TSH positively correlated with elevated neonatal TSH (adjusted β: 0.14, 95% confidence interval (CI): 0.10–0.17, p < 0.001) and CH risk (adjusted odds ratio: 1.21, 95% CI: 1.09–1.33, p < 0.001). A nonlinear dose–response relationship revealed a threshold at 4.85 mIU/L: below this value, neonatal TSH increased sharply with maternal TSH (β: 0.21, 95% CI: 0.16–0.26, p < 0.001). No significant associations existed in TPO-Ab-positive mothers. TPO-Ab status reversed the neonatal TSH-maternal FT3/FT4 ratio relationship: negative correlation in TPO-Ab-negative mothers (β: −1.78, 95% CI: −2.70 to −0.87, p < 0.001) versus positive correlation in TPO-Ab-positive mothers (β: 11.31, 95% CI: 2.03–20.58, p = 0.017).

Conclusion:

Maternal TSH correlates with neonatal TSH in TPO-Ab-negative pregnancies with a threshold at 4.85 mIU/L. TPO-Ab abolishes this association and reverses the FT3/FT4 ratio-neonatal TSH relationship, establishing TPO-Ab status as a critical modifier of maternal–fetal thyroid interactions. These findings support integrating TPO-Ab screening into prenatal thyroid management to optimize neonatal outcomes. However, the observational design limits causal inference, and further prospective studies are warranted.

Keywords

congenital hypothyroidism dose–response relationship maternal–fetal thyroid axis neonatal thyroid-stimulating hormone prenatal thyroid screening thyroid peroxidase antibody

Introduction

Thyroid hormones (TH) are critical regulators of embryonic neural development and fetal growth. Maternal disruptions in TH homeostasis significantly impact pregnancy outcomes such as preterm birth (PTB) and small/large for gestational age (SGA/LGA) deliveries.^1–3 Additionally, maternal thyroid function directly influences the long-term cognitive function and metabolic health of newborns.⁴ The establishment of fetal thyroid function progresses through distinct phases. Initially, the fetus is entirely dependent on maternal TH transferred across the placenta. Autonomous secretion by the fetal thyroid commences at approximately 12–14 weeks of gestation, with production increasing substantially thereafter.⁵ Therefore, maternal thyroid dysfunction during pregnancy constitutes a significant risk factor for neonatal thyroid disorders.^6,7

Neonatal thyroid-stimulating hormone (TSH) levels serve as a critical biomarker in congenital hypothyroidism (CH) screening, with elevated concentrations typically reflecting compensatory thyroid hyperactivity or primary glandular dysfunction.⁸ These levels are modulated by both intrinsic thyroid ontogeny and extrinsic maternal thyroid status.⁹ Although prior investigations have examined maternal–neonatal TH axis interactions, substantial inconsistencies persist across studies.^6,7,10–16 For instance, two cohort studies identified an inverse correlation between maternal free thyroxine (FT4) and neonatal TSH levels, whereas a longitudinal analysis demonstrated a paradoxical positive association.^10,13,14 This marked heterogeneity suggests the involvement of unmeasured effect modifiers within the maternal–fetal thyroid regulatory system.

Among potential unmeasured modifiers, thyroid peroxidase antibodies (TPO-Ab), which are markers of autoimmune thyroid diseases such as Hashimoto’s thyroiditis, are present in approximately 2%—17% of pregnant women.¹⁷ Notably, as TPO-Ab belongs to the immunoglobulin G class, it can cross the placental barrier, potentially exerting direct effects on the fetal thyroid gland or interfering with transplacental TH transport.^18,19 TPO-Ab positivity indicates maternal thyroid autoimmunity, which is associated with an increased risk of adverse birth outcomes.²⁰ Nonetheless, it remains unclear whether maternal TPO-Ab status plays an important modulatory role in the maternal–fetal TH axis.

Addressing the heterogeneity in maternal–neonatal TH axis interactions and the unclear role of TPO-Ab, this retrospective study analyzed data from 10,154 mother–newborn pairs to (1) quantify the association between maternal TH levels (free triiodothyronine (FT3), FT4, and TSH) in late pregnancy and neonatal TSH levels at birth, and (2) evaluate whether maternal TPO-Ab positivity modifies this association.

Materials and methods

Study design and population

This study was a retrospective cohort analysis involving pregnant women who gave birth at the Changzhou Maternal and Child Health Care Hospital from April 2016 to March 2017. The study protocol was approved by the hospital’s ethics committee (Approval No. ZD201803). Given that anonymous data were used, informed consent was not required. Initially, 13,275 pregnant women were enrolled, from which 2202 cases with missing maternal/neonatal thyroid function data were excluded, followed by 488 women with pre-pregnancy comorbidities (chronic hypertension, cardiovascular/hepatic/renal diseases, immune or thyroid disorders, type 1/2 diabetes), 335 multiple pregnancies, 24 miscarriages, 68 fetal malformations, and 4 intrauterine deaths. After exclusions, 10,154 singleton pregnancies were retained for final analysis.

Data collection

General characteristics such as height, weight, body mass index, age, parity, gravidity, gestational age, pregnancy complications, and neonatal information (gender, birth length, and birth weight) were obtained from the hospital’s electronic medical record system. Maternal thyroid TH values (FT3, FT4, TSH, and TPO-Ab) were retrieved from the Laboratory Information System. Maternal thyroid function was assessed using blood samples collected upon hospital admission for delivery. These samples were analyzed immediately as part of standard clinical care; consequently, all measurements were obtained during the third trimester (gestational weeks 28–41). Maternal serum TH levels were quantified using electrochemiluminescence immunoassays with a Cobas Elecsys 601 analyzer (Roche Diagnostics, Basel, Switzerland).²¹ Neonatal TSH values were retrieved from the Newborn Screening (NBS) information system. According to NBS protocols, heel prick blood samples were collected onto filter paper between 3 and 7 days after birth. Samples were analyzed by time-resolved fluorescence immunoassay using an Auto DELFIA 1235 analyzer (PerkinElmer Inc., Waltham, MA, USA) to quantify TSH concentrations.²²

Classifications and definitions

Participants were classified into a TPO-Ab negative group (TPO-Ab ⩽34.0 IU/mL) and a TPO-Ab positive group (TPO-Ab >34.0 IU/mL), according to the manufacturer’s guideline that defines the normal reference range for TPO-Ab as ⩽34.0 IU/mL. PTB was defined as delivery occurring before 37 weeks of gestation.²³ According to the Chinese birth weight reference curve, neonates were categorized as follows: SGA if their birth weight was below the 10th percentile for the corresponding gestational age; Appropriate for Gestational Age (AGA) if between the 10th and 90th percentiles; and LGA if above the 90th percentile.²⁴ Pregnancy complications included gestational diabetes mellitus, intrahepatic cholestasis of pregnancy, preeclampsia (PE), and pregnancy-induced hypertension.²⁵ The positive criterion for CH in NBS was defined as TSH ⩾10 mIU/L.²⁶

Statistical analysis

For baseline characteristics, continuous variables with a normal or skewed distribution were expressed as mean ± standard deviation or median (interquartile range), respectively, while categorical variables were expressed as frequencies (percentages). Comparisons between the TPO-Ab negative and positive groups were performed using independent samples T-Test for normally distributed continuous variables, Mann–Whitney U test for non-normally distributed variables, and Chi-square or Fisher’s exact tests for categorical variables. The threshold for statistical significance was set at a two-sided p < 0.05. To account for multiple comparisons, p-values were further adjusted using the Benjamini–Hochberg false discovery rate (FDR) method, with an FDR-corrected q < 0.05 considered significant. Spearman’s correlation coefficient was used to assess the relationship between neonatal TSH levels and maternal demographic characteristics, thyroid function indicators, and neonatal birth parameters. General linear models were employed to evaluate the association between thyroid function indicators and neonatal TSH. The models were adjusted for several covariates, including maternal age, height, weight, parity, gestational age, systolic and diastolic blood pressure at delivery, pregnancy complications, neonatal weight, and length. Smooth curve fitting analysis was conducted to explore potential nonlinear relationships between neonatal TSH and maternal TH, identifying inflection points to develop segmented linear regression models. All data analyses were performed using EmpowerStats software (X & Y Solutions Inc., Boston, MA, USA) and R (version 3.4.2, http://www.R-project.org). A p-value < 0.05 was deemed statistically significant.

Results

Characteristics of the eligible mother–neonate pairs

Baseline characteristics of mothers and neonates are summarized in Table 1. Compared to the TPO-Ab negative group, the TPO-Ab positive group showed numerical differences in several characteristics, including higher maternal age (29.19 vs 28.58 years), higher rates of assisted reproductive technology use (3.57% vs 2.20%), PE (4.10% vs 2.43%), PTB (3.57% vs 1.87%), and positive CH screening results (1.78% vs 0.81%), as well as a slightly lower median gestational age (38.83 vs 38.96 weeks). However, after correction for multiple comparisons using the FDR method, none of these differences remained statistically significant (all q > 0.05). Maternal thyroid function parameters (TSH, FT3, FT4 levels, and FT3/FT4 ratio) and key neonatal anthropometric measures (gender, birth length, and weight) did not differ meaningfully between the two groups.

Table 1.

Demographic characteristics of the participating mother–neonate pairs.

Variables	Total (N = 10,154 )	TPO-Ab negative (N = 9593 )	TPO-Ab positive (N = 561)	p-Value	FDR-corrected q-value
Maternal characteristics
Age (years)	28.61 ± 4.38	28.58 ± 4.35	29.19 ± 4.72	0.013	0.163
Height (cm)	161.67 ± 4.58	161.67 ± 4.58	161.71 ± 4.67	0.724	0.887
Weight (kg)	71.48 ± 9.48	71.43 ± 9.45	72.40 ± 9.92	0.056	0.467
BMI (kg/m²)	27.33 ± 3.32	27.31 ± 3.31	27.67 ± 3.53	0.050	0.467
Systolic BP (mmHg)	120.59 ± 11.49	120.53 ± 11.38	121.58 ± 13.15	0.404	0.786
Diastolic BP (mmHg)	74.26 ± 7.94	74.23 ± 7.87	74.77 ± 9.02	0.637	0.848
Gestational age (week)	38.96 ± 1.13	38.96 ± 1.13	38.83 ± 1.17	0.028	0.280
Assisted reproduction (%)	231 (2.27%)	211 (2.20%)	20 (3.57%)	0.035	0.350
Primipara (%)	6089 (59.97%)	5747 (59.91%)	342 (60.96%)	0.620	0.847
Preterm birth (%)	199 (1.96%)	179 (1.87%)	20 (3.57%)	0.005	0.063
Cesarean section (%)	4277 (42.12%)	4019 (41.90%)	258 (45.99%)	0.056	0.466
Pregnancy complication
ICP	596 (5.87%)	565 (5.89%)	31 (5.53%)	0.722	0.887
GDM	828 (8.15%)	780 (8.13%)	48 (8.56%)	0.721	0.887
PE	256 (2.52%)	233 (2.43%)	23 (4.10%)	0.014	0.163
PIH	216 (2.13%)	202 (2.11%)	14 (2.50%)	0.534	0.847
Thyroid hormones
FT3 (pmol/L)	4.05 ± 0.57	4.05 ± 0.57	4.05 ± 0.67	0.181	0.786
FT4 (pmol/L)	12.75 ± 1.89	12.74 ± 1.87	12.92 ± 2.22	0.245	0.786
TSH (mIU/L)	3.03 ± 1.74	3.02 ± 1.71	3.20 ± 2.21	0.532	0.847
TPO-Ab (IU/mL)	11.9 (7.9–17.3)	11.4 (7.7–16.1)	64.3 (42.8–111.1)	<0.001
FT3/FT4 ratio	0.32 ± 0.06	0.32 ± 0.06	0.32 ± 0.07	0.059	0.467
Neonatal characteristics
Sex
Female	4834 (47.61%)	4569 (47.63%)	265 (47.24%)	0.857	0.887
Male	5320 (52.39%)	5024 (52.37%)	296 (52.76%)
Birth length (cm)	50.04 ± 0.61	50.04 ± 0.60	50.08 ± 0.78	0.118	0.725
Birth weight (gram)	3404.78 ± 413.87	3405.47 ± 413.36	3392.95 ± 422.76	0.346	0.786
Weight for gestational age
SGA	790 (7.78%)	745 (7.77%)	45 (8.02%)	0.121	0.725
AGA	7767 (76.49%)	7356 (76.68%)	411 (73.26%)
LGA	1597 (15.73%)	1492 (15.55%)	105 (18.72%)
TSH (mIU/L)	2.88 (1.91–4.21)	2.88 (1.91–4.20)	2.97 (1.87–4.36)	0.008	0.100
TSH ⩾10 mIU/L (%)	88 (0.87%)	78 (0.81%)	10 (1.78%)	0.016	0.160

p Values were derived from comparisons among the characteristics of TPO-Ab negative and TPO-Ab positive groups. Data were presented as mean ± SD, median (IQR), and N (%).

AGA, Appropriate for Gestational Age; BMI, body mass index; BP, blood pressure; FDR, false discovery rate; FT3, free triiodothyronine; FT4, free thyroxine; GDM, gestational diabetes mellitus; ICP, intrahepatic cholestasis of pregnancy; IQR, interquartile range; LGA, large for gestational age; PE, preeclampsia; PIH, pregnancy-induced hypertension; SD, standard deviation; SGA, Small for Gestational Age; TPO-Ab, thyroid peroxidase antibody; TSH, thyroid-stimulating hormone.

Correlations between neonatal TSH and maternal/neonatal parameters

Table 2 presents the correlation coefficients between neonatal TSH levels and maternal characteristics, thyroid function indicators, and neonatal birth parameters. In the maternal TPO-Ab negative group, neonatal TSH levels were negatively correlated with maternal age (r = −0.029; p = 0.005), weight (r = −0.022; p = 0.036), FT3 levels (r = −0.037; p = 0.001), and FT3/FT4 ratio (r = −0.060; p < 0.001), and positively correlated with systolic BP (r = 0.035; p = 0.001), parity (r = 0.022; p = 0.031), gestational age (r = 0.039; p = 0.001), FT4 (r = 0.044; p < 0.001), and TPO-Ab (r = 0.055; p < 0.001). These trends were similar to those observed in the total population. None of these indicators showed significant associations in the TPO-Ab positive group. Maternal TSH levels were positively correlated with neonatal TSH levels in the total population (r = 0.117; p < 0.001), as well as in both the TPO-Ab negative group (r = 0.119; p < 0.001) and the TPO-Ab positive group (r = 0.089; p = 0.035).

Table 2.

Correlation of neonatal TSH with maternal characteristics, neonatal birth characteristics, and laboratory results.

Parameters	Total		TPO-Ab negative		TPO-Ab positive
Parameters	r	p-Value	r	p-Value	r	p-Value
Age (years)	−0.029	0.004	−0.029	0.005	−0.028	0.509
Height (cm)	−0.016	0.101	−0.017	0.105	−0.010	0.813
Weight (kg)	−0.023	0.020	−0.022	0.036	−0.053	0.214
BMI (kg/m²)	−0.018	0.077	−0.015	0.158	−0.070	0.098
Systolic BP (mmHg)	0.029	0.004	0.035	0.001	−0.066	0.121
Diastolic BP (mmHg)	0.016	0.100	0.021	0.039	−0.058	0.173
Parity	0.020	0.047	0.022	0.031	−0.015	0.715
Gestational age (week)	0.040	0.001	0.039	0.001	0.064	0.132
FT3 (pmol/L)	−0.035	0.001	−0.037	0.001	−0.009	0.834
FT4 (pmol/L)	0.040	0.001	0.044	<0.001	−0.024	0.574
TSH (mIU/L)	0.117	<0.001	0.119	<0.001	0.089	0.035
FT3/FT4 ratio	−0.055	<0.001	−0.060	<0.001	0.021	0.615
Birth length (cm)	0.008	0.437	0.009	0.371	−0.018	0.677
Birth weight (gram)	0.018	0.074	0.016	0.119	0.045	0.290

BMI, body mass index; BP, blood pressure; FT3, free triiodothyronine; FT4, free thyroxine; TPO-Ab, thyroid peroxidase antibody; TSH, thyroid-stimulating hormone.

Associations between maternal TH and neonatal TSH

Table 3 presents the associations between maternal TH and neonatal TSH in both unadjusted and multivariable-adjusted regression models. In the unadjusted model for the maternal TPO-Ab negative group, a 1-unit increase in maternal FT3/FT4 ratio was associated with a 1.59 mIU/L decrease in neonatal TSH levels (β: −1.59, 95% confidence interval (CI): −2.45, −0.72). After adjusting for confounding factors, the association remained significant (β: −1.78, 95% CI: −2.70, −0.87). Conversely, in the TPO-Ab positive group, the FT3/FT4 ratio showed a positive association with neonatal TSH levels before adjustment (β: 9.36, 95% CI: 0.88, 17.83), which also persisted after adjustment (β: 11.31, 95% CI: 2.03, 20.58). In the maternal TPO-Ab negative group, maternal TSH levels demonstrated positive associations with both elevated neonatal TSH (β: 0.13, 95% CI: 0.10, 0.17) and CH screening positivity (odds ratio (OR): 1.18, 95% CI: 1.07, 1.30), while maternal FT3 showed an inverse association with neonatal TSH (β: −0.15, 95% CI: −0.25, −0.06); after confounder adjustment, these associations strengthened (TSH–neonatal TSH: β: 0.14, 95% CI: 0.10–0.17; TSH-CH screening positivity: OR: 1.21, 95% CI: 1.09, 1.33; FT3-neonatal TSH: β: −0.18, 95% CI: −0.28, −0.07). Conversely, in TPO-Ab positive mothers, neither maternal TSH–neonatal TSH association (β: 0.04, 95% CI: −0.23, 0.30) nor TSH-CH screening positivity association (OR: 1.14, 95% CI: 0.87, 1.51) reached significance.

Table 3.

Relationship between maternal thyroid hormone and neonatal TSH and CH screening positivity.

Characteristics	Unadjusted^a		Adjusted^a		Unadjusted^b		Adjusted^b
Characteristics	β (95% CI)	p-Value	β (95% CI)	p-Value	OR (95% CI)	p-Value	OR (95% CI)	p-Value
All women
FT3 (pmol/L)	−0.09 (−0.20, 0.02)	0.103	−0.10 (−0.21, 0.01)	0.067	1.11 (0.79, 1.57)	0.531	1.02 (0.70, 1.50)	0.902
FT4 (pmol/L)	0.02 (−0.01, 0.05)	0.216	0.02 (−0.01, 0.05)	0.264	1.02 (0.92, 1.14)	0.704	1.03 (0.91, 1.15)	0.664
TSH (mIU/L)	0.13 (0.09, 0.16)	<0.001	0.13 (0.09, 0.16)	<0.001	1.17 (1.07, 1.27)	<0.001	1.19 (1.08, 1.30)	<0.001
FT3/FT4 ratio	−0.96 (−1.91, −0.02)	0.046	−1.06 (−2.06, −0.05)	0.039	1.09 (0.04, 28.03)	0.957	0.56 (0.02, 19.52)	0.747
Women with TPO ⩽34 IU/mL
FT3 (pmol/L)	−0.15 (−0.25, −0.06)	0.002	−0.18 (−0.28, −0.07)	<0.001	1.09 (0.74, 1.59)	0.668	0.96 (0.63, 1.47)	0.866
FT4 (pmol/L)	0.03 (0.00, 0.06)	0.039	0.03 (0.00, 0.06)	0.049	1.06 (0.95, 1.20)	0.290	1.08 (0.95, 1.22)	0.227
TSH (mIU/L)	0.13 (0.10, 0.17)	<0.001	0.14 (0.10, 0.17)	<0.001	1.18 (1.07, 1.30)	<0.001	1.21 (1.09, 1.33)	<0.001
FT3/FT4 ratio	−1.59 (−2.45, −0.72)	<0.001	−1.78 (−2.70, −0.87)	<0.001	0.35 (0.01, 12.00)	0.558	0.11 (0.00, 5.39)	0.268
Women with TPO >34 IU/mL
FT3 (pmol/L)	0.70 (−0.13, 1.53)	0.098	0.93 (0.00, 1.85)	0.051	1.24 (0.60, 2.56)	0.554	2.27 (0.78, 6.60)	0.133
FT4 (pmol/L)	−0.11 (−0.37, 0.14)	0.377	−0.13 (−0.41, 0.15)	0.368	0.75 (0.53, 1.07)	0.111	0.58 (0.35, 0.95)	0.032
TSH (mIU/L)	0.05 (−0.21, 0.30)	0.712	0.04 (−0.23, 0.30)	0.781	1.11 (0.90, 1.36)	0.334	1.14 (0.87, 1.51)	0.342
FT3/FT4 ratio	9.36 (0.88, 17.83)	0.031	11.31 (2.03, 20.58)	0.017	1279.02 (0.46, 3,551,261.42)	0.077	464,300.93 (20.13, inf.)	0.011

Adjusted for maternal age, height, weight, pregnancy complication, parity, gestational age, systolic and diastolic BP at delivery, rates of assisted reproduction, delivery mode, neonatal sex, height, and weight.

Maternal thyroid hormone-neonatal TSH link.

Maternal thyroid hormone-neonatal CH screening positivity link.

CH, congenital hypothyroidism; CI, confidence interval; FT3, free triiodothyronine; FT4, free thyroxine; OR, odds ratio; TPO-Ab, thyroid peroxidase antibody; TSH, thyroid-stimulating hormone.

Nonlinear relationship between neonatal TSH and maternal TH

Smooth curve fitting and linear regression analysis revealed a nonlinear correlation between neonatal TSH levels and maternal TSH levels (Figure 1), with the model results presented in Table 4. Standard linear regression models showed that, in the TPO-Ab negative group, neonatal TSH increased with rising maternal TSH levels (β: 0.14, 95% CI: 0.10, 0.17). A similar trend was observed in the total population group (β: 0.13, 95% CI: 0.09, 0.16), with p < 0.001 for both groups, indicating significant statistical significance.

Figure 1.

Smooth curve fitting analysis (fully adjusted model) of the association between maternal TSH and neonatal TSH. (a) Stratified analysis: Separate curves were shown for the TPO-Ab positive (blue) and TPO-Ab negative (red) groups. (b) Total population analysis: A single curve represented the association in the combined cohort. The model was adjusted for maternal age, height, weight, pregnancy complication, parity, gestational age, systolic and diastolic blood pressure at delivery, rates of assisted reproduction, neonatal sex, and neonatal height and weight.

Table 4.

The results of two-piece wise linear regression model about neonatal TSH and maternal TSH.

Outcome	Total		TPO-Ab negative		TPO-Ab positive
Outcome	β (95% CI)	p-Value	β (95% CI)	p-Value	β (95% CI)	p-Value
Fitting model by standard linear regression	0.13 (0.09, 0.16)	<0.001	0.14 (0.10, 0.17)	<0.001	0.04 (−0.22, 0.30)	0.781
Fitting model by two-piece wise linear regression
Inflection point of the maternal TSH	3.4		4.85		2.74
⩽Inflection point	0.26 (0.18, 0.34)	<0.001	0.21 (0.16, 0.26)	<0.001	0.69 (−0.28, 1.65)	0.163
>Inflection point	0.05 (−0.01, 0.11)	0.09	0.01 (−0.07, 0.09)	0.752	−0.12 (−0.46, 0.23)	0.509
p for log-likelihood ratio test		<0.001		<0.001		0.163

CI, confidence interval; TPO-Ab, thyroid peroxidase antibody; TSH, thyroid-stimulating hormone.

Further analysis using a two-piece wise linear regression model revealed an inflection point of 4.85 mIU/L in the TPO-Ab negative group. Below this threshold, neonatal TSH increased by 0.21 mIU/L for every 1 mIU/L increase in maternal TSH (β: 0.21, 95% CI: 0.16, 0.26). Above the threshold, no statistically significant association was observed (p = 0.752). In the total population, the inflection point was 3.4 mIU/L. Below this threshold, neonatal TSH increased with maternal TSH levels (β: 0.26, 95% CI: 0.18, 0.34), while no significant association was found above the threshold (p = 0.090). A log-likelihood ratio test comparing the two models demonstrated that the two-piece wise linear regression model provided a significantly better fit to the data in both the TPO-Ab negative group and the total population group (p < 0.001).

Additionally, this study examined the relationship between neonatal TSH and maternal FT3/FT4 ratio using smooth curve fitting. The results revealed a nonlinear trend between the two variables (Figure 2). Subsequently, two-piece wise linear regression models were employed to assess the statistical significance of potential inflection points. The log-likelihood ratio test demonstrated that the two-piece wise model did not significantly improve the fit compared to the standard linear model (p = 0.168, Table 5), indicating that the proposed inflection points were not statistically significant.

Figure 2.

Smooth curve fitting analysis (fully adjusted model) of the association between maternal FT3/FT4 ratio and neonatal TSH. (a) Stratified analysis: Separate curves were shown for the TPO-Ab positive (blue) and TPO-Ab negative (red) groups. (b) Total population analysis: A single curve represented the association in the combined cohort. The model was adjusted for maternal age, height, weight, pregnancy complication, parity, gestational age, systolic and diastolic blood pressure at delivery, rates of assisted reproduction, neonatal sex, and neonatal height and weight.

Table 5.

The results of a two-piecewise linear regression model about maternal FT3/FT4 ratio and neonatal TSH.

Outcome	Total		TPO-Ab negative		TPO-Ab positive
Outcome	β (95% CI)	p-Value	β (95% CI)	p-Value	β (95% CI)	p-Value
Fitting model by standard linear regression	−1.06 (−2.05, −0.05)	0.039	−1.78 (−2.70, −0.87)	<0.001	11.31 (2.03, 20.58)	0.017
Fitting model by two-piecewise linear regression
Inflection point of the maternal FT3/FT4 ratio	0.38		0.38		0.34
⩽Inflection point	−1.97 (−3.36, −0.57)	0.006	−2.40 (−3.68, −1.13)	<0.001	0.93 (−15.82, 17.67)	0.914
>Inflection point	1.17 (−1.40, 3.74)	0.372	−0.27 (−2.61, 2.06)	0.818	22.47 (4.84, 40.10)	0.013
p for log-likelihood ratio test		0.065		0.168		0.138

CI, confidence interval; FT3, free triiodothyronine; FT4, free thyroxine; TPO-Ab, thyroid peroxidase antibody; TSH, thyroid-stimulating hormone.

Discussion

Main findings

This hospital-based cohort study on a Chinese population yielded several key findings. Firstly, in the TPO-Ab negative group, maternal TSH levels were positively associated with both neonatal TSH levels (adjusted β: 0.14, 95% CI: 0.10, 0.17) and CH screening positivity (adjusted OR: 1.21, 95% CI: 1.09, 1.33), while no significant association was observed in the positive group. Secondly, through smooth curve fitting analysis, neonatal TSH levels exhibit a nonlinear relationship and saturation effect with maternal TSH levels in the TPO-Ab negative group, with evidence of a threshold effect (4.85 mIU/L). On the left side of the inflection points, neonatal TSH levels increase with rising maternal TSH levels (β: 0.21, 95% CI: 0.16, 0.26). However, on the right side of the inflection point, this association is not statistically significant (p = 0.752). Thirdly, this study is the first to indicate that TPO-Ab status significantly alters the direction of the association between maternal FT3/FT4 ratio and neonatal TSH. In the TPO-Ab negative group, a higher maternal FT3/FT4 ratio was negatively correlated with neonatal TSH (adjusted β: −1.78, 95% CI: −2.70, −0.87), whereas this trend was reversed in the TPO-Ab positive group (adjusted β: 11.31, 95% CI: 2.03, 20.58).

Interpretation

Several epidemiological studies from Denmark, India, Japan, Spain, and China (with sample sizes ranging from 142 to 3339) have demonstrated that maternal TSH levels are positively correlated with neonatal TSH levels.^6,10–15 This correlation remains unaffected by maternal gestational stage (early or late pregnancy) or neonatal sample sources (umbilical cord blood or heel stick blood). Consistent with prior reports, this large cohort study involving over 10,000 mother–infant pairs confirmed the positive correlation. Additionally, smooth curve fitting analysis revealed a nonlinear dose–response relationship between maternal and neonatal TSH concentrations, with a maternal TSH threshold of 4.85 mIU/L identified in TPO-Ab-negative women. Below this threshold, a positive association was observed between maternal and neonatal TSH levels. Above this threshold (which likely indicates maternal hypothyroidism), no significant association was found. Similarly, an observational study involving 90 hypothyroid mothers and their neonates in India found no significant association between maternal and neonatal TSH levels.¹⁷ Furthermore, linear regression analyses demonstrated no statistically significant association between maternal and neonatal TSH levels in the TPO-Ab-positive subgroup. Notably, our study represents the first comprehensive evaluation of the TSH relationship between TPO-Ab-positive mothers and their offspring.

In contrast to the well-documented positive association between maternal and neonatal TSH levels, existing evidence regarding the relationship between other maternal TH (e.g., FT3 and FT4) and neonatal TSH concentrations remains inconsistent. Korevaar et al.,¹⁰ in a prospective cohort study of 3339 Dutch mother–infant pairs, demonstrated a negative correlation between maternal FT4 and neonatal cord-blood TSH levels. Similarly, a cross-sectional study of 203 euthyroid pregnant women and their newborns revealed negative correlations between neonatal cord-blood TSH and both maternal free FT3 and FT4 levels.¹⁴ However, a longitudinal study from China, involving 2991 mother–infant pairs, found a significant positive association of neonatal TSH (heel prick samples) with maternal FT3 and FT4 levels.¹³ Furthermore, our study confirmed through both correlation analyses and regression models that in the TPO-Ab-negative subgroup, maternal FT3 levels were inversely associated with neonatal TSH concentrations (heel prick samples), whereas maternal FT4 levels exhibited a positive correlation (in crude model: p = 0.039; in adjusted model: p = 0.052). In contrast, no meaningful correlations were observed between neonatal TSH and either maternal FT3 or FT4 levels in the TPO-Ab-positive subgroup. The observed heterogeneity across studies likely stems from multifactorial sources: (1) methodological variations in cohort design and statistical power (reflected in sample size differences); (2) population diversity in genetic/epigenetic predispositions (ethnicity) and environmental exposures (iodine status); (3) preanalytical technical confounders spanning biospecimen procurement (blood source), processing timelines, and assay batch effects; and (4) pathophysiological modifiers such as maternal thyroid autoimmunity (TPO-Ab positivity).

The FT3/FT4 ratio, a key indicator reflecting the peripheral metabolic balance of TH, sensitively captures the dynamic conversion efficiency between FT3 and FT4. Its significance extends beyond the isolated assessment of thyroid function provided by individual markers.²⁷ Previous study have shown that this ratio is more advantageous than single indicators in revealing the association between TH and metabolic health, as well as in evaluating fetal TH exposure and the intrauterine developmental environment.²⁸ In the present study, we further explored the relationship between maternal FT3/FT4 ratio and neonatal TSH levels. The diametrically opposed associations of FT3/FT4 ratio with neonatal TSH levels, characterized by an inverse correlation in TPO-Ab-negative mothers and a positive correlation in TPO-Ab-positive mothers, may plausibly reflect autoimmune-mediated disruption of thyroid homeostasis.

The direct impact of maternal TPO-Ab status on maternal–neonatal TH relationships requires further elucidation. In TPO-Ab-negative women, maternal TSH and FT4 levels exhibited positive correlations with neonatal TSH, while maternal FT3 showed an inverse correlation. Notably, these correlations were absent in TPO-Ab-positive mothers, indicating a fundamental disruption of the fetoplacental thyroid axis. Interpreting these findings necessitates consideration of the dynamic gestational context. Maternal thyroid physiology undergoes profound changes across trimesters, characterized by rising thyroxine-binding globulin, hCG-mediated thyroid stimulation in the first trimester, and increasing placental deiodinase activity later in gestation. In TPO-Ab-negative pregnancies, these relationships may reflect a coordinated physiological adaptation. By contrast, in TPO-Ab-positive women, whose thyroid reserve may be compromised, this adaptive capacity appears to be strained, potentially exacerbating relative hormonal insufficiency. In these pregnancies, the breakdown of the expected maternal–neonatal hormone equilibrium appears to be driven by autoimmune mechanisms. We therefore hypothesize that this disconnection involves a dual-pathway mechanism: first, TPO-Ab may downregulate placental TH transporters, limiting maternal thyroxine transfer and weakening the expected link between maternal FT4 and neonatal TSH^29,30; second, transplacental TPO-Ab could bind to fetal thyroid TPO, impairing fetal hormone synthesis and triggering a compensatory fetal TSH rise independent of maternal FT3.^19,31 This fetal TSH response must be interpreted in light of the standard timing of neonatal screening (typically 3–7 days postpartum), which captures the natural postnatal TSH surge. Alterations in this surge pattern in TPO-Ab-exposed neonates could reflect in utero thyroid axis perturbation. Critically, TPO-Ab serves primarily as a marker of broader autoimmunity; more direct fetal thyroid threats stem from co-existing thyrotropin receptor antibodies (e.g., thyroid-stimulating immunoglobulins, or thyrotropin-blocking antibodies), which can actively stimulate or inhibit the fetal thyroid. Furthermore, the overall iodine status of the population is a key modifier. Iodine sufficiency is a prerequisite for normal thyroidal adaptation to pregnancy. In regions of even mild iodine insufficiency, the added burden of autoimmune thyroiditis (marked by TPO-Ab) may precipitate more significant dysfunction, potentially amplifying the observed disruptions in maternal–fetal thyroid correlations. Consequently, the obstetric and neonatal impacts of TPO-Ab positivity may be understood as stemming from combined hormonal and immune dysregulation, with the clinical expression of these effects further modulated by gestational timing, postnatal physiology, and nutritional context.

Strengths

The strengths of this cohort study include the following: (1) our study, through stratified analysis based on maternal TPO-Ab status, has revealed a significant pattern of association between the maternal FT3/FT4 ratio and neonatal TSH levels, highlighting the critical role of autoimmune thyroid status in modulating maternal–fetal thyroid interactions; (2) By employing covariate-adjusted regression models to control for potential confounders and incorporating smooth curve fitting analysis to identify nonlinear relationships and threshold effects, our research overcomes the limitations of traditional linear analyses; (3) As the first cohort study in China to systematically investigate the relationship between the maternal FT3/FT4 ratio and neonatal TSH, this study fills a gap in race-specific data, providing important regional public health reference values; and (4) The observed positive correlation between maternal TSH and neonatal TSH levels and CH screening positivity underscores the potential value of maternal TSH monitoring as a prenatal screening tool for neonatal thyroid dysfunction.

Limitations

Several limitations of this study should be acknowledged. Firstly, the retrospective observational design fundamentally limits causal inference. While statistical adjustments were made for available confounders, unmeasured or residual confounding cannot be ruled out. Further limitations inherent to the design include the hospital-based recruitment, which may have introduced selection bias (e.g., toward higher-risk pregnancies), and the reliance on a single measurement of maternal thyroid parameters, which captures only a cross-sectional snapshot of a dynamic physiological state. Secondly, we were unable to adjust for certain key confounding factors, such as maternal iodine nutrition status and exposure to environmental endocrine-disrupting chemicals. Thirdly, the lack of placental histological evidence limits our mechanistic understanding of how TPO-Ab might affect hormone transport. Furthermore, neonatal TSH, used here as a screening marker for CH, lacks diagnostic specificity for definitive thyroid dysfunction. Finally, regarding statistical power for stratified analysis, the prevalence of TPO-Ab positivity was 5.5% (N = 561). Although this constitutes a substantial subgroup, the imbalance between groups limits the precision of effect estimates within the TPO-Ab positive group, necessitating validation in larger, specifically designed cohorts.

Conclusion

This study demonstrates that maternal TSH levels positively correlate with neonatal TSH in TPO-Ab-negative pregnancies, with a nonlinear threshold at 4.85 mIU/L, whereas TPO-Ab positivity eliminates this association and reverses the maternal FT3/FT4 ratio-neonatal TSH relationship. These findings advocate for TPO-Ab screening to stratify prenatal risk, targeted maternal TSH control below 4.85 mIU/L, and dual monitoring of TSH and TH ratios in TPO-Ab-positive pregnancies to optimize neonatal outcomes. While observational design limits causal interpretation, the results emphasize TPO-Ab status as a pivotal modifier of maternal–fetal thyroid dynamics, urging integration into clinical guidelines and mechanistic research on placental TPO-Ab transfer.

Supplemental Material

sj-docx-1-tae-10.1177_20420188261446409 – Supplemental material for Maternal TPO-Ab status modulates associations between gestational thyroid hormones and neonatal TSH in 10,154 Chinese mother–newborn pairs

Supplemental material, sj-docx-1-tae-10.1177_20420188261446409 for Maternal TPO-Ab status modulates associations between gestational thyroid hormones and neonatal TSH in 10,154 Chinese mother–newborn pairs by Huihui Wang, Sijie Xi, Xusheng Chen, Zhaolong Zhan, Xiaosong Yuan and Bin Zhang in Therapeutic Advances in Endocrinology and Metabolism

Footnotes

Acknowledgements

We thank all participants of this study and the staff of the laboratory and medical record sections from Changzhou Maternal and Child Health Care Hospital for their technical assistance and information service.

Declarations

ORCID iD

Xiaosong Yuan

Supplemental material

Supplemental material for this article is available online.

Clinical trial number

Not applicable.

References

Leung

MTY

Man

KKC

, et al. Maternal thyroid dysfunction during pregnancy and the risk of adverse outcomes in the offspring: a systematic review and meta-analysis. J Clin Endocrinol Metab 2020; 105(12): dgaa555.

Yuan

Wang

Gao

, et al. Impact of maternal thyroid hormone in late pregnancy on adverse birth outcomes: a retrospective cohort study in China. Endocr J 2021; 68(3): 317–328.

Zhang

Zhan

, et al. Maternal obesity and the incidence of large-for-gestational-age newborns in isolated hypothyroxinemia pregnancies: a comparative cohort study. Reprod Biol Endocrinol 2025; 23(1): 60.

Korevaar

Muetzel

Medici

, et al. Association of maternal thyroid function during early pregnancy with offspring IQ and brain morphology in childhood: a population-based prospective cohort study. Lancet Diabetes Endocrinol 2016; 4(1): 35–43.

LaFranchi

Haddow

Hollowell

JG.

Is thyroid inadequacy during gestation a risk factor for adverse pregnancy and developmental outcomes?

Thyroid 2005; 15(1): 60–71.

Medici

De Rijke

Peeters

, et al. Maternal early pregnancy and newborn thyroid hormone parameters: the generation R study. J Clin Endocrinol Metab 2012; 97(2): 646–652.

Menon

Sreejyothi

Raveendranath

Can thyroid dysfunction in mothers influence neonatal thyroid profile?

Pediatr Rev Int J Pediatr Res 2016; 3(7): 503–508.

Ding

Liu

Zhang

, et al. Permanent vs transient congenital hypothyroidism in Chinese children: physical growth and predictive nomogram. J Clin Endocrinol Metab 2024; 109(8): e1616–e1622.

Tonyushkina

Shen

Ortiz-Toro

, et al. Embryonic exposure to excess thyroid hormone causes thyrotrope cell death. J Clin Invest 2014; 124(1): 321–327.

10.

Korevaar

TIM

Chaker

Jaddoe

VWV

, et al. Maternal and birth characteristics are determinants of offspring thyroid function. J Clin Endocrinol Metab 2016; 101(1): 206–213.

11.

Itoh

Baba

Yuasa

, et al. Association of maternal serum concentration of hydroxylated polychlorinated biphenyls with maternal and neonatal thyroid hormones: the Hokkaido birth cohort study. Environ Res 2018; 167: 583–590.

12.

Sinha

Mandal

Gupta

, et al. Effect of maternal hypothyroidism on thyroid hormone status of fetus: a hospital based study. J Med Sci Clin Res 2019; 7(5): 576–581.

13.

Zhang

S-J

, et al. Higher maternal thyroid resistance indices were associated with increased neonatal thyroid-stimulating hormone—analyses based on the Huizhou mother-infant cohort. Front Endocrinol 2022; 13: 937430.

14.

Chen

Shen

, et al. Association between maternal age and sex-based neonatal free triiodothyronine levels. BMC Endocr Disord 2024; 24(1): 98.

15.

Velasco

Martín

Gallego

, et al. Maternal-fetal thyroid function at the time of birth and its relation with iodine intake. Thyroid 2013; 23(12): 1619–1626.

16.

Chandrasekhar

Barla

Pusapati

, et al. Maternal hypothyroidism and its effects on neonatal thyroid function: a retrospective study. Asian J Pharm Clin Res 2023; 16(9): 70–72.

17.

Alexander

Pearce

Brent

, et al. 2017 Guidelines of the American thyroid association for the diagnosis and management of thyroid disease during pregnancy and the postpartum. Thyroid 2017; 27(3): 315–389.

18.

Lee

Lau

, et al. Increased fetal abortion rate in autoimmune thyroid disease is related to circulating TPO autoantibodies in an autoimmune thyroiditis animal model. Fertil Steril 2009; 91(5 Suppl.): 2104–2109.

19.

Seror

Amand

Guibourdenche

, et al. Anti-TPO antibodies diffusion through the placental barrier during pregnancy. PLoS One 2014; 9(1): e84647.

20.

Dhillon-Smith

Coomarasamy

TPO antibody positivity and adverse pregnancy outcomes. Best Pract Res Clin Endocrinol Metab 2020; 34(4): 101433.

21.

Wang

Huang

, et al. Establishment of self-sequential longitudinal reference intervals of maternal thyroid function during pregnancy. Exp Biol Med (Maywood) 2010; 235(10): 1212–1215.

22.

Long

Guo

Yao

, et al. Genetic and phenotypic characteristics of congenital hypothyroidism in a Chinese cohort. Front Endocrinol (Lausanne) 2021; 12: 705773.

23.

Yuan

Han

Jia

, et al. Low fetal fraction of cell free DNA at non-invasive prenatal screening increases the subsequent risk of preterm birth in uncomplicated singleton pregnancy. Int J Womens Health 2022; 14: 889–897.

24.

Zhang

Zhan

, et al. Impact of serum retinol-binding protein 4 levels in late pregnancy on the incidence of small/large for gestational age infants among 11,854 pregnant women: a retrospective study. J Epidemiol 2025; 35(6): 287–296.

25.

Yuan

Gao

Zhang

, et al. Association of maternal D-dimer level in late pregnancy with birth outcomes in a Chinese cohort. Clin Chim Acta 2020; 501: 258–263.

26.

Nascimento

Dornbusch

, et al. Impact of the reduction in TSH cutoff level to 6 mIU/L in neonatal screening for congenital hypothyroidism in Santa Catarina: final results. Arch Endocrinol Metab 2021; 64(6): 816–823.

27.

Bassols

Prats-Puig

Soriano-Rodríguez

, et al. Lower free thyroxin associates with a less favorable metabolic phenotype in healthy pregnant women. J Clin Endocrinol Metab 2011; 96(12): 3717–3723.

28.

Knight

Shields

Hattersley

, et al. Maternal hypothyroxinaemia in pregnancy is associated with obesity and adverse maternal metabolic parameters. Eur J Endocrinol 2016; 174(1): 51–57.

29.

Patel

Landers

, et al. Delivery of maternal thyroid hormones to the fetus. Trends Endocrinol Metab 2011; 22(5): 164–170.

30.

Fan

Luo

Tang

, et al. Advanced maternal age, mode of delivery, and thyroid hormone levels in Chinese newborns. Front Endocrinol (Lausanne) 2020; 10: 913.

31.

Hoermann

Pekker

Midgley

JEM

, et al. The role of supporting and disruptive mechanisms of FT3 homeostasis in regulating the hypothalamic-pituitary-thyroid axis. Ther Adv Endocrinol Metab 2023; 14: 20420188231158163.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.03 MB