Epidemiological characteristics of congenital hypothyroidism and risk factors for permanent congenital hypothyroidism in Gansu, China

Abstract

Background

Congenital hypothyroidism (CH) is a major endocrine disorder affecting newborn development, and early differentiation between transient CH (T-CH) and permanent CH (P–CH) is essential for individualized management. This study aimed to describe epidemiological patterns of CH in Gansu Province, China from 2019 to 2021, compare clinical characteristics between T-CH and P–CH, and identify risk factors associated with P–CH to support improved early diagnosis and long-term management in primary care.

Methods

This retrospective cohort study was conducted using data from the Newborn Screening Center of Gansu Provincial Maternity and Child-care Hospital, a provincial referral center in Northwest China. A total of 757,157 newborns screened between January 2019 and December 2021 were included. Newborn heel-prick thyroid-stimulating hormone (TSH) was measured using time-resolved immunofluorescence assay, with TSH ≥ 9 mIU/L as the recall threshold. Recalled infants underwent confirmatory serum testing for TSH, free thyroxine (FT4), and total thyroxine (T4) using chemiluminescent immunoassays. Thyroid autoantibodies (TPOAb, TgAb) were assessed via electrochemiluminescence. Based on treatment response and thyroid function recovery after levothyroxine (L-T4) withdrawal, 275, a total of 332 infants were confirmed to have CH. Among them, 26 cases were excluded due to incomplete clinical data, infants diagnosed with CH were classified into T-CH (n = 87) and P–CH (n = 188) groups.

Results

The incidence of CH in Gansu was 0.044% (332/757,157, 1:2281), with a gradual increase from 2019 to 2021. Compared with T-CH infants, those with P–CH had significantly higher initial screening TSH (42.70 vs 22.45 mIU/L; P < 0.001), lower pretreatment FT4 and T4 levels (both with P < 0.001), markedly lower thyroid autoantibody positivity (1.1% vs 6.9%; P = 0.007), and higher birth weight (3520 vs 2850 g; P < 0.001). P–CH infants also required consistently higher age-specific maintenance doses of L-T4 throughout follow-up (all with P < 0.001). Multivariable logistic regression identified high birth weight (> 2500 g) as an independent risk factor for P–CH (OR = 7.788; 95% CI: 1.917—31.636), while thyroid autoantibody positivity was strongly associated with T-CH (OR = 0.084; 95% CI: 0.016–0.454).

Conclusion

The incidence of CH in Gansu was slightly lower than the national average but demonstrated an upward trend. In our cohort of infants with positive newborn screening results, birth weight > 2500 g was associated with an increased likelihood of permanent CH and may serve as an early clinical indicator for distinguishing P–CH from T-CH. Thyroid autoantibody positivity was more commonly observed in transient CH and may suggest a more favorable prognosis. These findings highlight the importance of long-term follow-up and individualized L-T4 dose adjustment in screening-positive infants.

Introduction

Congenital hypothyroidism (CH) is one of the most common endocrine disorders in neonates and a major preventable cause of neurocognitive impairment and growth retardation. Globally, the incidence of CH has been increasing, and in China recent data indicate a rise in prevalence [1–3]. Clinically, CH often presents with nonspecific features, such as abdominal distension, constipation, jaundice, hypotonia, wide fontanelle, coarse skin, and delayed neuromotor development, making early diagnosis by clinical signs alone difficult. Early detection thus relies critically on newborn screening (NBS). Historically, NBS for CH was introduced in China in the early 1980 s, and since then the detection rate has improved significantly [4].

Thyroid hormone production begins in the fetus by mid-gestation, with T4 levels rising toward term, but the hypothalamic-pituitary-thyroid (HPT) axis does not mature fully until term or the early neonatal period [5, 6]. Without timely treatment, infants with CH may develop severe growth failure (“cretinism”) and profound neurodevelopmental impairment. Lifelong levothyroxine (L-T4) replacement is required in many patients, those classified as having permanent CH (P–CH). However, studies have shown that a substantial proportion of children recover thyroid function and are subsequently reclassified as transient CH [7, 8].

Gansu Province, located in northwestern China, remains underrepresented in epidemiological studies of CH, particularly with respect to long-term outcomes and risk factors for P–CH. In this context, our study aims to characterize the incidence, clinical phenotypes, and predictors of permanent versus transient CH in Gansu. By elucidating these regional features, we hope to support optimized screening strategies and individualized management of CH in resource limited settings. Although previous studies and consensus guidelines have described clinical differences between transient CH (T-CH) and permanent CH (P–CH), most evidence has been derived from limited regional cohorts, and data from northwestern China remain scarce. In addition, early clinical indicators that may help distinguish P–CH from T-CH among infants with positive newborn screening results remain insufficiently characterized. Gansu Province, located in northwestern China, is underrepresented in epidemiological studies of CH, particularly regarding long-term outcomes and factors associated with persistent disease. Therefore, this study aimed to characterize the incidence and clinical features of CH in Gansu and to explore potential early indicators associated with P–CH among screening-positive infants. By providing region-specific epidemiological data and evaluating readily available clinical parameters, this study may contribute to improved risk stratification and more individualized management of CH.

Methods

Study design and ethical approval

This retrospective cohort study was conducted using NBS data collected in Gansu Province, China, between January 2019 and December 2021. The study included a total of 757,157 newborns, comprising 389,824 males and 367,332 females. The study protocol was reviewed and approved by the Ethics Committee of Gansu Provincial Maternal and Child Health Hospital (File No.: 2023GSFY28), and all procedures conformed to the principles of the Declaration of Helsinki for medical research involving human subjects.

Newborn screening and laboratory measurements

All neonates underwent routine NBS after birth. Heel-prick blood samples were analyzed for thyroid-stimulating hormone (TSH) using an enzyme-linked immunofluorescence assay. A TSH level ≥ 8 mIU/L was considered a positive primary screen. Newborns with positive screening results were recalled for confirmatory testing of serum TSH, free thyroxine (FT4), and total thyroxine (T4) levels. Reference ranges for serum TSH, FT4, and total T4 were 0.31—5.30 mIU/L, 9.8—22.0 pmol/L, and 75.45—148.18 nmol/L, respectively.

Diagnostic criteria for CH

Diagnosis of congenital hypothyroidism followed the 2020—2021 European Society for Paediatric Endocrinology consensus guidelines. For neonates aged 1—4 weeks, treatment with oral levothyroxine (L-T4) was initiated if TSH > 20 mIU/L or FT4 < 5 pmol/L. For infants older than 21 days, L-T4 treatment could be started when serum TSH was 6—20 mIU/L, and CH was confirmed if TSH remained elevated while FT4 decreased after treatment initiation. In this cohort, neonates with heel-prick TSH ≥ 8 mIU/L and confirmatory serum TSH ≥ 20 mIU/L plus FT4 < 12 pmol/L were classified as CH. A total of 332 CH cases were confirmed, corresponding to an incidence of 0.044%.

Classification of transient and permanent CH

CH patients were categorized into transient (T-CH) and permanent (P–CH) subtypes according to diagnostic consensus. T-CH criteria included: 1) normal thyroid morphology on ultrasonography; 2) no family history of CH in first-degree relatives; 3) L-T4 dose sufficient to maintain TSH within low-normal and FT4 within or above normal range; 4) gradual tapering or cessation of L-T4 after 6 months to 1 year, with complete discontinuation by age 2—3; and 5) stable TSH and FT4 within reference range at follow-up at 1, 3, 6 months, and 1 year after cessation. Patients not meeting these criteria were classified as P–CH. The final study population included 87 T-CH and 188 P–CH patients.

Exclusion criteria

Fifty-seven neonates were excluded due to the following reasons: 8 cases lacked parental consent, 3 infants died within 2 years of birth, 20 cases had complex neonatal comorbidities or preterm birth with normal follow-up TSH values, and 26 cases had incomplete follow-up data that could compromise analysis accuracy.

Data collection

Clinical data collected included: 1) demographic information, gestational age (weeks + days), birth weight (grams), sex, mode of delivery (vaginal or cesarean); 2) neonatal health status, postnatal disease history, including timing, treatment, and prognosis of conditions such as asphyxia, infection, or jaundice; 3) thyroid-related parameters, primary TSH screening values, serum TSH, FT4, total T4, and thyroid antibody status (TPOAb, TgAb). Follow-up schedule consisted of biweekly visits in the neonatal period, monthly visits from 1—6 months, quarterly visits from 6 months to 2 years, and biannual visits thereafter. Follow-up was primarily conducted in outpatient clinics, supplemented by remote telephone or video consultations. Adherence was promoted via individualized health records, reminder messages, and free thyroid function testing for families attending scheduled follow-ups.

Statistical analysis

Statistical analyses were performed using SPSS version 27.0 (IBM, Armonk, NY). Normality of continuous variables was assessed using the Shapiro–Wilk test. Normally distributed data were presented as mean ± standard deviation (SD) and compared using independent-samples t-tests. Non-normally distributed continuous variables were expressed as median [interquartile range, IQR] and compared using the Mann–Whitney U test. Categorical variables were reported as counts and percentages and analyzed using chi-square or rank-sum tests. Variables showing significance in univariate analysis were entered into a binary logistic regression model to identify independent risk factors for P–CH. A two-tailed P value < 0.05 was considered statistically significant.

Results

From 2019 to 2021, a total of 757,157 newborns underwent screening in Gansu Province. In 2019, 294,389 infants were screened, of whom 121 were diagnosed with congenital hypothyroidism, corresponding to an incidence of 0.0411%. In 2020, 248,051 newborns were screened, with 113 confirmed cases (incidence: 0.0456%). In 2021, 214,717 infants were screened, and 98 were diagnosed, yielding an incidence of 0.0456%. Overall, the incidence of congenital hypothyroidism in Gansu showed a slight upward trend over the three-year period (Table 1).

Table 1.

Incidence of congenital hypothyroidism in Gansu Province, 2019—2021

Year	Number Screened	Number Diagnosed	Incidence (%)
2019	294,389	121	0.04110
2020	248,051	113	0.04556
2021	214,717	98	0.04564

Among patients with T-CH, 56.3% were male and 43.7% were female, whereas in the P–CH group, males accounted for 51.6% and females 48.4%. Chi-square analysis indicated no significant difference in sex distribution between the two groups (χ² = 0.533, P = 0.465). Regarding family history, 79.3% of T-CH patients had no family history of CH and 20.7% had a positive family history; in the P–CH group, 85.6% had no family history and 14.4% had a positive history. This difference was not statistically significant (χ² = 1.740, P = 0.187). For maternal thyroid disease during pregnancy, 21.8% of T-CH and 14.9% of P–CH cases were affected, with no significant difference observed (χ² = 2.025, P = 0.155). Similarly, medication history did not differ significantly between T-CH (21.8%) and P–CH (14.9%) groups (χ² = 2.025, P = 0.155) (Table 2).

Table 2.

Sex and family history distribution in transient and permanent congenital hypothyroidism [n, %]

	T-CH(N = 87)	P–CH(N = 188)	χ2	P
Variable			0.533	0.465
Male	49 (56.3)	97 (51.6)
Female	38 (43.7)	91 (48.4)
			1.740	0.187
Family history negative	69 (79.3)	161 (85.6)
Family history positive	18 (20.7)	27 (14.4)
Maternal thyroid disease			2.025	0.155
Negative	68 (78.2)	160 (85.1)
Positive	19 (21.8)	28 (14.9)
Medication history			2.025	0.155
No	68 (78.2)	160 (85.1)
Yes	19 (21.8)	28 (14.9)

The median initialprimary screening TSH value in the T-CH group was 22.45 mIU/L (interquartile range [IQR]: 14.56—52.54), whereas in the P–CH group it was 42.70 mIU/L (IQR: 18.99—116.45). The mean TSH values were 8.5 mIU/L in the T-CH group and 12.3 mIU/L in the P–CH group. Z-test analysis revealed a statistically significant difference between the two groups (Z = − 3.707, P < 0.001) (Table 3). These results indicate that infants with permanent congenital hypothyroidism had significantly higher primary screening TSH levels compared with those with transient CH (Fig. 1).

Table 3.

Primary screening TSH values in transient and permanent congenital hypothyroidism

	T-CH(N = 87)	P–CH(N = 188)	Z	P
Primary Screening TSH	22.45 (14.56—52.54)	42.70 (18.99—116.45)	−3.707	0.000

Fig. 1.

Primary screening TSH values in infants with transient and permanent congenital hypothyroidism

Prior to treatment, the median serum TSH level in the T-CH group was 27.11 mIU/L (interquartile range [IQR]: 16.50—100.00), compared with 98.45 mIU/L (IQR: 30.63—100.00) in the P–CH group, representing a statistically significant difference (Z = −4.457, P < 0.001). Two weeks after initiation of levothyroxine therapy, TSH levels decreased to 2.74 mIU/L (IQR: 1.18—5.02) in the T-CH group and 2.97 mIU/L (IQR: 1.03—5.98) in the P–CH group, with no significant difference observed between the groups (Z = −0.776, P = 0.438).

Before treatment, the median total T4 level was 76.40 nmol/L (IQR: 46.70—104.90) in T-CH patients and 56.60 nmol/L (IQR: 26.73—89.05) in P–CH patients, showing a significant difference (Z = −3.580, P < 0.001). After two weeks of therapy, total T4 increased to 127.80 nmol/L (IQR: 106.60—142.30) in T-CH and 129.80 nmol/L (IQR: 116.60–147.53) in P–CH, with no significant difference (Z = −1.352, P = 0.176).

Prior to treatment, median FT4 levels were 11.60 pmol/L (IQR: 6.40—16.50) in T-CH and 7.60 pmol/L (IQR: 3.23—12.90) in P–CH, a statistically significant difference (Z = −4.084, P < 0.001). Mean Total T4 in the T-CH group was 76.40 pmol/L, higher than the P–CH group (mean 56.60 pmol/L), while mean serum TSH was lower in T-CH (27.11 mIU/L) than in P–CH (98.45 mIU/L). Two weeks after treatment, FT4 levels were 20.80 pmol/L (IQR: 18.50—23.80) in T-CH and 21.40 pmol/L (IQR: 18.45—24.65) in P–CH, with no significant difference observed (Z = −0.323, P = 0.747) (Table 4).

Table 4.

Comparison of thyroid function parameters and treatment response between transient and permanent congenital hypothyroidism

	T-CH(N = 87)	P–CH(N = 188)	Z/χ2	P
TSH
Before treatment	27.11(16.50—100.00)	98.45(30.63—100.00)	−4.457	0.000
Two weeks after treatment	2.74(1.18—5.02)	2.97(1.03—5.98)	−0.776	0.438
Total T4
Before treatment	76.40(46.70—104.90)	56.60(26.73—89.05)	−3.580	0.000
Two weeks after	127.80(106.60—142.30)	129.80(116.60—147.53)	−1.352	0.176
Free T4
Before treatment	11.60(6.40—16.50)	7.60(3.23—12.90)	−4.084	0.000
Two weeks after	20.80(18.50—23.80)	21.40(18.45—24.65)	−0.323	0.747
Thyroid autoantibodies			7.164	0.007
Absent	81(93.1)	186(98.9)
Present	6(6.9)	2(1.1)

Based on these comparisons, we observed that the initial screening TSH levels were significantly higher in the P–CH group than in the T-CH group. After two weeks of oral L-T4 therapy, TSH levels markedly decreased in both groups (Fig. 2). In addition, the baseline total T4 and FT4 concentrations were substantially lower in the P–CH group compared with the T-CH group, whereas both indices increased markedly following two weeks of treatment (Figs. 3 and 4).

Fig. 2.

Comparison of TSH levels between transient and permanent congenital hypothyroidism before treatment and after 2 weeks of therapy

Fig. 3.

Comparison of total T4 levels between transient and permanent congenital hypothyroidism before treatment and after 2 weeks of therapy

Fig. 4.

Comparison of free T4 levels between transient and permanent congenital hypothyroidism before treatment and after 2 weeks of therapy

All patients in the T-CH group had both left and right thyroid lobes present. In the P–CH group, 3.7% of patients exhibited absence of either the left or right thyroid lobe. The difference approached statistical significance based on the χ² test (χ² = 3.324, P = 0.068).

Thyroid autoantibodies were detected in 6.9% of patients with T-CH, compared with 1.1% in the P–CH group, and this difference was statistically significant (χ² = 7.164, P = 0.007). Thus, the proportion of thyroid antibody-positive individuals was significantly higher in the T-CH group. The positivity rate in the T-CH group (32.1%) exceeded that in the P–CH group (8.7%).

The median birth weight was significantly higher in the P–CH group (3,200 g) compared with the T-CH group (2,850 g) (Z = −5.123, P < 0.001). A significant difference in the distribution of birth-weight categories was also observed between groups (χ² = 17.337, P < 0.001). Specifically, infants with birth weight > 2,500 g accounted for 94.1% (177/188) of the P–CH group, markedly higher than the 86.2% (75/87) in the T-CH group. Conversely, low birth weight (< 2,500 g) was more prevalent in the T-CH group (13.8%, 12/87) than in the P–CH group (5.9%, 11/188).

It should be noted that the “higher birth-weight group” in this study was defined as > 2,500 g. This threshold was selected based on the baseline characteristics of our cohort: the overall median birth weight among 275 infants with CH was 3,020 g, with medians of 2,850 g in the T-CH group and 3,200 g in the P–CH group. The value of 2,500 g represented a meaningful inflection point in the distribution between groups. Additionally, this categorization demonstrated significant discriminatory value in univariable analysis and remained an independent risk factor in multivariable logistic regression. Therefore, this study did not adopt the clinical definition of macrosomia (> 4,000 g) (Table 5).

Table 5.

Comparison of birth-weight categories between infants with transient and permanent congenital hypothyroidism

	T-CH(N = 87)	P–CH(N = 188)	χ2	P
< 1500 g	9(10.3)	3(1.6)	17.337	0.000
1500—2500 g	11(12.6)	8(4.3)
> 2500 g	75(86.2)	177(94.1)

Comparison of gestational age distributions revealed no significant differences between infants with T-CH and those with P–CH. Preterm birth (< 37 weeks) occurred in 10.3% of the T-CH group and 6.4% of the P–CH group, whereas term birth (≥ 37 weeks) accounted for 89.7% and 93.6% of each group, respectively (χ² = 1.324, P = 0.250) (Table 6).

Table 6.

Distribution of gestational age in infants with transient versus permanent congenital hypothyroidism

	T-CH(N = 87)	P–CH(N = 188)	χ2	P
Preterm	9(10.3)	12(6.4)	1.324	0.250
Term	78(89.7)	176(93.6)

Infants with transient congenital hypothyroidism consistently required lower levothyroxine doses than those with permanent CH at all measured time points, including initiation of therapy, 6 months, 1 year, and 2 years of age. The differences were highly significant at each interval (baseline: Z = −4.113, P < 0.001; 6 months: Z = −7.337, P < 0.001; 1 year: Z = −8.576, P < 0.001; 2 years: Z = −7.035, P < 0.001). Levothyroxine therapy in all infants was initiated promptly after diagnosis in accordance with clinical guidelines, and the subsequent time points represent routine follow-up assessments during treatment.

At treatment initiation, the mean L-T4 dose in the T-CH group was 10 μg/kg/day, notably lower than the 15 μg/kg/day required for the P–CH group. These findings indicate that permanent CH generally necessitates higher and more sustained L-T4 supplementation, reflecting the persistent nature of thyroid hormone deficiency in this subgroup (Table 7).

Table 7.

Comparison of levothyroxine dosing (μg/kg/day) between infants with transient and permanent congenital hypothyroidism

	T-CH	P–CH	Z	P
Before treament	5.71(4.50—6.82)	6.81(5.23—7.89)	−4.113	0.000
6 months	2.22(1.88—2.89)	3.05(2.67—3.76)	−7.337	0.000
1 years old	1.67(1.16—2.03)	2.51(2.09—3.21)	−8.576	0.000
2 years	1.04(0.78—1.64)	2.01(1.65—2.51)	−7.035	0.000

Variables that demonstrated significant differences in univariable analyses were included in a multivariable logistic regression model to identify independent predictors of permanent congenital hypothyroidism (P–CH). The analysis revealed that birth weight was a significant risk factor for P–CH. Infants with birth weight > 2,500 g had 7.788-fold higher odds of developing P–CH compared with those with birth weight < 1,500 g (95% CI: 1.917—31.636).

In contrast, the presence of thyroid autoantibodies was associated with a higher likelihood of transient CH. Infants who tested positive for thyroid antibodies had only 0.084-fold the odds of having P–CH compared with those who were antibody-negative (95% CI: 0.016—0.454), suggesting that antibody positivity predicts a transient course of hypothyroidism.(Table 8).

Table 8.

Multivariable logistic regression analysis of risk factors for permanent congenital hypothyroidism

Factors	B	SE	Wald	P	OR(95%CI)
Follow-up TSH	0.010	0.008	1.480	0.224	1.010(0.994—1.025)
Total T4	0.010	0.010	0.923	0.337	1.010(0.99.−1.029)
Free T4	−0.070	0.072	0.933	0.334	0.933(0.809—1.074)
Birth weight			11.950	0.003
< 1500
1501—2500	0.979	0.842	1.351	0.245	2.661(0.511—13.870)
> 2500	2.053	0.715	8.238	0.004	7.788(1.917—31.636)
Usage	0.155	0.113	1.901	0.168	1.168(1.917—1.456)
Thyroid autoantibodie	−2.473	0.858	8.300	0.004	0.084(0.016—0.454)

Discussion

In our large, population‑based cohort from Gansu Province, we not only documented a stable, but slightly increasing, incidence of CH over three years, but also characterized key clinical and biochemical differences between T‑CH and P‑CH forms. This is particularly significant because relatively few population-based studies have described CH subtypes in western regions of China, and our data provide regional evidence that may help inform screening strategies and long-term management in this population [1].

Recent international evidence further highlights that the proportion of transient CH may be higher than previously recognized, particularly in population-based screening programs, emphasizing the importance of distinguishing T-CH from P–CH early in clinical practice [9, 10]. In addition, long-term analyses spanning several decades have demonstrated that advances in newborn screening have progressively shifted our understanding of CH etiology, with increasing recognition of transient forms and milder phenotypes detected through early screening [11].

One of the most important findings of our study is that, among infants with positive newborn screening results for CH, birth weight > 2,500 g was associated with a higher likelihood of P–CH. After adjusting for other variables, infants above this threshold showed approximately 7.8-fold higher odds of permanent disease compared with those with very low birth weight (< 1,500 g) within this screening-positive cohort. This finding suggests that birth weight, a routinely collected clinical variable, may serve as a practical early clinical indicator to help differentiate P–CH from T-CH among infants diagnosed through newborn screening. While classical risk stratification often focuses on overt factors such as thyroid dysgenesis or family history, our data indicate that anthropometric characteristics at birth may also reflect underlying pathophysiology [12]. This aligns with emerging international work: for example, a study in Iran developed a risk prediction model for CH subtypes and identified confirmatory TSH, total T4, and certain clinical features as discriminators [13].

This observation is consistent with emerging studies that aim to identify early predictive markers of CH subtypes. For example, recent retrospective cohort analyses have demonstrated that early-life clinical parameters, combined with biochemical indices, can significantly improve prediction of transient versus permanent CH [14]. Moreover, predictive models incorporating hormonal levels and clinical features have been proposed to enhance early differentiation between subtypes [15].

Mechanistically, several plausible explanations could underlie this birth‑weight association. Larger infants likely require greater thyroid hormone production to meet metabolic demands, and if their synthetic capacity is intrinsically limited (perhaps due to subtle developmental anomalies in the thyroid), they may be more vulnerable to permanent dysfunction. Genetic variation may also play a role; developmental genes such as TSHR, PAX8, or NKX2—1, when altered, could reduce compensatory capacity [16–18]. Interestingly, we observed a slightly higher, though not statistically significant, rate of thyroid lobe absence in P‑CH infants, which mayindirectly hint at structural vulnerability [19–21]. However, in line with previous studies, the distinction between absence of the left or right lobe itself is unlikely to carry substantial clinical relevance. What appears to be more informative is whether the thyroid gland is present in situ or absent/ectopic, as transient CH is generally associated with a higher probability of an anatomically normal thyroid gland, whereas permanent CH more frequently reflects dysgenesis or structural abnormalities reported in previous epidemiological studies [22, 23]. Therefore, in the present analysis we focused primarily on the broader classification of thyroid morphology rather than detailed lobe-specific subdivision.

Recent imaging-based studies further support the prognostic value of thyroid morphology. For instance, thyroid volume and structural characteristics assessed by ultrasound have been shown to provide additional discriminatory power between transient and permanent CH [24]. Similarly, studies focusing on children with eutopic thyroid glands or hemiagenesis have identified structural and functional parameters as important predictors of long-term disease outcome [25, 26].

Our biochemical data reinforce the distinction between P‑CH and T‑CH: at diagnosis, P‑CH infants exhibited significantly higher TSH and lower total and free T4, but after two weeks of levothyroxine (L‑T4) therapy, these parameters converged. This pattern suggests that early thyroid hormone replacement is effective in normalizing thyroid function, but the initial imbalance may reflect a more entrenched glandular dysfunction in P‑CH. Comparable studies have reported that diagnostic TSH and treatment dosing differ between subtypes, making these metrics valuable for early prognostication [22]. Several cohort studies from Europe, Asia, and North America have similarly demonstrated that higher initial TSH concentrations, increased levothyroxine requirements, and structural thyroid abnormalities are strongly associated with permanent CH, whereas infants with milder biochemical abnormalities and an in situ thyroid gland are more likely to have transient disease [27, 28]. These findings support the growing consensus that early biochemical severity and thyroid morphology provide useful clues for differentiating CH subtypes during follow-up.

Consistent with these findings, recent studies have further emphasized the role of levothyroxine dosing as a practical predictor. Higher L-T4 requirements during treatment and at the time of treatment interruption have been associated with permanent CH, while lower dose requirements are more suggestive of transient disease [10, 29]. These observations reinforce the value of integrating treatment response into subtype prediction models.

Another compelling result from our cohort is the association of thyroid autoantibodies with T‑CH. Although the presence of thyroid autoantibodies in transient CH has been reported previously, our findings provide additional evidence in a regional cohort that antibody positivity is more frequently observed in T-CH than in P–CH. We found that antibody positivity greatly reduced the odds of P‑CH, indicating that autoimmune mechanisms may contribute more to the transient form. While the maternal–fetal immune interplay in CH remains underexplored, our observations align with prior case‑control data showing that family history or parental thyroid disease can influence CH outcomes [30]. This is in line with broader etiological evaluations suggesting that transient CH may arise from a combination of environmental, maternal, and immunological factors, rather than fixed structural abnormalities [11, 15].

However, our study has limitations. First, we lacked detailed maternal data on iodine intake, thyroid function during pregnancy, and autoantibody status, all of which might confound the relationship between birth weight and CH subtype. Without such information, the potential influence of maternal iodine status on fetal thyroid development could not be evaluated in the present study. Second, as a retrospective study, there is potential selection bias. We initially identified 332 CH cases, but only 275 were included in our final analysis; cases lost to follow-up or with incomplete data may have led to underestimation of P‑CH prevalence or biased associations. Third, the small number of thyroid autoantibody-positive infants limits the precision of our findings regarding immunological predictors. Fourth, genetic testing was not routinely performed in this cohort. Given that pathogenic variants in genes such as TSHR, PAX8, NKX2—1, DUOX2, and TG are increasingly recognized as important contributors to congenital hypothyroidism, the absence of molecular data limits our ability to fully explore etiological differences between permanent and transient CH. In addition, recent studies have emphasized the importance of re-evaluation protocols and long-term follow-up in accurately distinguishing transient from permanent CH, suggesting that misclassification may occur without standardized reassessment strategies [14, 26].

Looking forward, future research should adopt a prospective, ideally multicenter design to validate our observations. Key priorities include systematic collection of maternal iodine status and thyroid function, genetic and structural assessments (e.g., thyroid ultrasound or molecular testing), and long-term neurodevelopmental follow-up. Furthermore, integrating dynamic predictors such as early hormone levels, treatment response, thyroid imaging, and clinical characteristics into unified predictive models may improve early subtype stratification and guide individualized management [14].

In summary, our study provides strong evidence that in a large, under-studied Chinese population, higher birth weight is a robust independent risk factor for permanent congenital hypothyroidism. Combined with biochemical and immunological data, these findings may support more precise risk stratification, earlier prognostic counseling, and tailored management strategies, ultimately improving outcomes for affected infants.

Abbreviations

CH

Congenital Hypothyroidism

TCH

Transient Congenital Hypothyroidism

PCH

Permanent Congenital Hypothyroidism

NBS

Newborn Screening

TSH

Thyroid-Stimulating Hormone

FT4

Free Thyroxine

T4

Total Thyroxine

TPOAb

Thyroid Peroxidase Antibody

TgAb

Thyroglobulin Antibody

L-T4

Levothyroxine

HPT

Hypothalamic-Pituitary-Thyroid

SD

Standard Deviation

IQR

Interquartile Range

OR

Odds Ratio

CI

Confidence Interval

Authors’ contributions

All authors conducted all data analyses and were the primary contributors to writing the manuscript. All authors reviewed and revised the manuscript.

Funding

This study was supported by the Gansu Provincial Health Industry Research Program (Grant No.GSWSKY2022-33).

Data availability

The datasets used and/or analyzed during the current study are available from the corresponding author (Xiujing Li, E-mail: 13,919,157,945@163.com) upon reasonable request. All data supporting the findings of this study are derived from the newborn screening records and clinical follow-up data of Gansu Provincial Maternity and Child-care Hospital (Gansu Provincial Central Hospital), which are managed in accordance with institutional data governance regulations. Due to ethical and legal restrictions related to the protection of personal health information (in compliance with the Personal Information Protection Law of the People’s Republic of China and the ethical approval requirements), the raw data cannot be made publicly available in an open-access repository. Researchers seeking access to the data must submit a formal request to the corresponding author, including a detailed research plan and a commitment to adhere to data privacy and security protocols. The data will be provided in a de-identified format to ensure the confidentiality of participants’ personal information, and access will be granted following review and approval by the institutional data access committee and the Ethics Committee of Gansu Provincial Maternal and Child Health Hospital. All materials, reagents, and analytical tools used in the study (including time-resolved immunofluorescence assay kits, chemiluminescent immunoassay reagents, and SPSS 27.0 statistical software) are commercially available and do not require special access permissions.

Declarations

Ethics approval and consent to participate

The study protocol was reviewed and approved by the Ethics Committee of Gansu Provincial Maternal and Child Health Hospital (File No.: 2023GSFY28). All procedures involving human participants in this study adhered to the ethical standards of the Declaration of Helsinki and relevant national and institutional guidelines for medical research involving human subjects. This study was a retrospective cohort study based on anonymized newborn screening data and medical records. The requirement for individual written informed consent to participate was waived by the Ethics Committee of Gansu Provincial Maternal and Child Health Hospital, because all personal identifying information was fully de-identified, and the use of existing routine clinical data posed minimal risk to participants. During data collection and follow-up, all personal and medical information of newborns was kept strictly confidential in accordance with the Personal Information Protection Law of the People’s Republic of China and institutional data security regulations. For infants requiring long-term follow-up, verbal consent was obtained from their legal guardians after explanation of the study purpose and procedures.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.