Neurodevelopmental disorders (NDDs) represent a group of complex and heterogeneous conditions that manifest as impairments in cognition, communication, behavior and motor skills due to abnormal neural development [1]. Common NDDs usually have their initial-onset during childhood and adolescence, with an estimated global prevalence rate of 2–5.7% for attention deficit hyperactivity disorder (ADHD), 0.4–4.4% for autism spectrum disorder (ASD) and 0.3–0.8% for tic disorder (TD) [2, 3]. Children and adolescents with these disorders are at increased risk for poor long-term outcomes, including lower levels of educational attainment, employment status and high rates of substance abuse, delinquency and psychiatric comorbidity (e.g., anxiety and depression), which not only significantly affect their quality of life, but also impose a tremendous stress on families and society [4, 5]. This highlights the urgent need for early diagnosis and interventions, in children and adolescents with NDDs, to prevent serious consequences.

Although the etiology of NDDs is considered to be multifactorial, existing evidence suggests thyroid dysfunction as a probable trigger [6]. There have been meta-analyses that demonstrate a significant association exists between maternal thyroid dysfunction and NDDs in children [6, 7]. Pooling of 29 articles by Ge et al. [6] showed the risk of ADHD in offspring was increased in mothers with thyroid hormones outside the normal range during pregnancy, including hyperthyroidism [odds ratio (OR) = 1.18; 95% confidence interval (CI) = 1.04 to 1.34] or hypothyroidism (OR = 1.14; 95% CI = 1.03 to 1.26). Maternal hypothyroidism (OR = 1.41; 95% CI = 1.05 to 1.90) [6] or autoimmune thyroid diseases (OR = 1.28; 95% CI = 1.14 to 1.44) [7] was also found to elevate the prevalence ratio of ASD in children through integration of 29 [6] or 6 [7] studies, respectively. Additionally, mice overexpressing thyroid hormone-responsive protein or lacking thyroid hormone receptors (THRB or THRA) typically manifested with ADHD [8, 9, 10]. An animal model of hypothyroidism by treatment with propylthiouracil exhibited developmental cerebellar defects and autism-like behaviors [11, 12]. All these observations imply thyroid function status-related parameters in children and adolescents [such as free triiodothyronine (FT3), free thyroxine (FT4), total triiodothyronine (TT3), total thyroxine (TT4), thyroid stimulating hormone (TSH), thyroid peroxidase antibody (TPO-Ab) or thyroglobulin antibody (TG-Ab)] may represent potential biomarkers for predicting the development of NDDs. This hypothesis has been supported by some studies: Meng et al. [13] found FT3 was significantly increased, while FT4 was significantly decreased in children with ADHD, ASD and TD compared to healthy controls (HCs). Furthermore, both TT3 and TT4 were significantly lower in children with ADHD/TD than those in HCs, while TG-Ab was significantly reduced in children with TD/ASD relative to HCs. A receiver operating characteristic (ROC) curve analysis demonstrated FT4 was of high clinical value in the diagnosis of children with NDDs, with an area under the curve (AUC) of 0.76 [13]. Singh et al. [14] found the TSH level was 30% lower in male children with ASD than that in typically developing boys. The diagnostic accuracy and AUC for ASD based upon TSH levels was 76% and 0.674, respectively [14]. However, controversial results were also reported, including significantly higher levels of TSH and lower levels of FT4 in children with ASD [15, 16] and no statistical differences in TSH and TT4 between ADHD/ASD cases and HCs [17, 18]. Therefore, the relationship between thyroid hormones and NDDs in children and adolescents still requires further evaluation.

It furthermore has been reported that overweight and obesity are more prevalent among children and adolescents with NDDs [19]. Patients with hypothyroidism are often obese [20], which on one hand may be a consequence of lowered energy expenditure due to a decrease in the basal metabolic rate [21] or on the other hand associated with uncontrolled food intake due to disruptions in growth [insulin, insulin-like growth factor-1 (IGF-1) or IGF-binding protein 3 (IGFBP-3)] [22] and appetite (leptin, ghrelin or adiponectin) [23] hormone signaling pathways, ultimately facilitating weight gain. Thus, growth and appetite hormone-related parameters may serve as underlying diagnostic biomarkers for NDDs. Unfortunately, there were also conflicting findings in these indicators among different studies, including reduced [17, 24, 25], increased [26, 27] and non-significance [28, 29] of the various possible.

To further elucidate the link between the development of NDDs and levels of thyroid, growth and appetite hormones in children and adolescents, a meta-analysis that integrated all relevant evidence was conducted here. The sample size and statistical power can be increased through pooling of data from multiple individual studies, which may thus provide more convincing conclusions.

Although there was one meta-analysis [13] that explored the link of thyroid hormones in children with the development of NDDs, this study only collected data of FT4 and TSH before August 23, 2024. In the present study, an updated meta-analysis that retrieved studies providing all thyroid hormones (e.g., FT3, FT4, TT3, TT4, TSH and TPO-Ab) in both children and adolescents with NDDs and HCs until March 1, 2025, was performed. Consequently, 26 and 32 datasets were respectively obtained for analysis of FT4 and TSH, the sample size of which was clearly larger than those of the study of Meng et al. [13] (FT4, 9; TSH, 14) and thus, the conclusion may be more certain. As expected, the results of the overall meta-analysis were non-significant for both FT4 (SMD = –0.02; 95% CI = –0.19 to 0.15) and TSH (SMD = –0.08; 95% CI = –0.23 to 0.08), which was somewhat different from those of Meng et al. [13] (FT4: MD = –0.29, 95% CI: –0.50 to –0.09, significant; TSH: MD = –0.07, 95% CI: –0.36 to 0.2, non-significant). However, significantly reduced FT4 was observed in the overall analysis for the ADHD-IT subtype, subgroup analyses for total NDDs/ADHD (multicenter studies) and trim-and-fill adjusted analysis for total NDDs, ADHD and ASD. More interestingly, subgroup analysis revealed lower TSH levels were associated with an increased risk for NDDs diagnosis in children, but not in adolescents. Furthermore, this study, for the first time, identified that compared to HCs, FT3 levels were significantly increased in total NDD and ADHD patients; TT4 levels were significantly decreased in total NDD and ADHD-IT patients; TT3 levels were only elevated in each ADHD subtype (ADHD-IT, ADHD-HIT, ADHD-CT); TPO-Ab levels were significantly increased in total NDD, ADHD and ASD patients. This suggests TSH, FT4 and TPO-Ab may represent potential biomarkers for diagnosing and monitoring children and/or adolescents with various NDDs (both ADHD and ASD), while FT3, TT3 and TT4 may be associated with the development of ADHD.

Thyroid hormones are synthesized and secreted by the thyroid gland. T4 is the major thyroid hormone secreted by this gland. However, T4 is a biologically inactive form that must be converted to T3 by iodothyronine deiodinases for biological activity in target cells. T4 and T3 can be free or bound to plasma proteins, among which, only FT4 and FT3 are pivotal for physiological functions [32]. Therefore, theoretically, the importance sequence of thyroid hormones for the development of diseases may be FT3 $>$ FT4 $>$ TT3 $>$ TT4, which was demonstrated here (i.e., overall meta-analysis for NDDs and ADHD achieved significant results for FT3. Also, the number of analyzed data was more than six; while only subgroup or pooled analysis with a small number of datasets confirmed significant findings for other hormones). Existing evidence has demonstrated these thyroid hormones play fundamental roles in brain development. For example, mutations in THRB or THRA were reported to cause syndromes of resistance to thyroid hormone (RTHB or RTHA) in patients who exhibited an ADHD-like phenotype [33]. T3 functioned mainly via binding to THRB or THRA. THRB or THRA mutations may disrupt T3 binding and inactivate downstream signaling pathways, ultimately promoting elevated T3 levels in the blood [34]. Furthermore, the production and release of T3 are stimulated by TSH through the hypothalamic-pituitary thyroid axis. The T3 accumulated in the blood may exert a negative feedback effect on the anterior pituitary and thereby suppress TSH secretion [35]. Higher FT3, TT3 and lower TSH represent the status of hyperthyroidism which has several symptoms in common with ADHD, including hyperactivity and inattention. It had been demonstrated that the prevalence ratio of ADHD in those with hyperthyroidism increased 1.7-fold when compared to the reference population [36]. An in vitro study also revealed that treatment with high-dose T3 reduced the viability of neural precursor cells and induced hyperactivity of cortical neurons as well as similar transcriptomic changes with ASD and ADHD patients [37]. Lower levels of T4 may be associated with increased expression and enzymatic activity of deiodinase 2 [38, 39]. TPO-Ab is a thyroid autoimmune antibody. High levels of TPO-Ab indicate the presence of inflammatory diseases in the thyroid gland. Numerous studies have implied activation of inflammation may be a risk factor for the development of ADHD and ASD symptoms in adolescence [40, 41], including increased number of M1 macrophages and secretion of pro-inflammatory cytokines [e.g., interleukin (IL)-1$\beta$, IL-6, IL-8 and tumor necrosis factor-$\alpha$] when compared to HCs [42]. Maternal hypothyroxinemia was also reported to induce ASD-like phenotypes in the offspring who simultaneously exhibited immune disorders manifested primarily by the significant increase in levels of pro-inflammatory cytokine IL-17A in blood and brain tissues, as well as the enhancement in the ratio of Th17/Treg cells and frequencies of M1-like macrophages in the spleen. Supplementation with T4 prevented maternal hypothyroxinemia-induced ASD phenotypes and immune responses [43]. Such findings may explain results reported here that high levels of TPO-Ab, FT3, TT3 and low levels of FT4, TT4, TSH predicted a high risk of ADHD and/or ASD.

Currently, there were no meta-analyses that investigated the association of IGF-1/IGFBP-3 with NDDs. In this study, six and nine datasets that respectively analyzed ADHD and ASD were, for the first time, integrated. Although the overall meta-analysis did not obtain significant differences between NDDs (ADHD and ASD) and HCs, subgroup analysis indicated significantly reduced IGF-1 levels in CSF samples of ASD when compared to HCs. IGF-1, a polypeptide hormone, was implicated in exerting neuro-protective roles in brain cells, including increasing cell viability and decreasing both pro-apoptotic caspase-3 activity and lactate dehydrogenase release through activation of the PI3K/AKT-YAP/TAZ cascade [44, 45]. In SHANK3-deficient mice that mimicked ASD, daily intraperitoneal injections of IGF-1 or its derivative were observed to reverse deficits in hippocampal $\alpha$-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid signaling, long-term potentiation and motor performance [46]. Clinical trials found that treatment with IGF-1 triggered significant improvements in social impairment, restrictive behaviors and hyperactivity in SHANK3-deficient children [47, 48]. These findings may support low levels of IGF-1 as a pathogenesis of ASD.

Chen et al. [49] used a meta-analysis to explore the changes of appetite hormones (including leptin, n = 11; ghrelin, n = 2; adiponectin, n = 7) in children with ASD, with leptin significantly increased to be identified. However, articles included in the study of Chen et al. [49] contained data from patients undergoing medical treatment (methylphenidate or atomoxetine) which may have influenced levels of those appetite hormones (e.g., increase leptin, ghrelin and adiponectin) [29, 50]. In this study, all data were collected from drug-naïve patients with ADHD and ASD. When compared to Chen et al.’ study [49], more or newly published [51, 52, 53] datasets were included for analysis of ASD (leptin: n = 15; ghrelin: n = 2; adiponectin: n = 7). Moreover, appetite hormones were, for the first time, analyzed for ADHD patients (leptin: n = 9; ghrelin: n = 5; adiponectin: n = 3). Regardless of pooled or subgroup analysis, no remarkably significant differences in levels of adiponectin were identified between total NDDs/ADHD/ASD and HCs. Leptin and ghrelin levels were found to be significantly increased in male NDDs (leptin: ADHD, n = 2; ASD, n = 4; ghrelin: ADHD, n = 1; ASD, n = 1); leptin levels were also higher in independent analyses of all ASD patients and several subgroups. Leptin was originally discovered to be secreted from adipose tissues into the blood and function as a regulator of body-weight by reducing food intake and increasing energy expenditure [54]. Administration of recombinant leptin was found to induce weight loss [55, 56], which was similar to the symptoms of hyperthyroidism (FT3 increase) observed in ASD and ADHD. Additionally, leptin was previously identified to share structural and functional similarities with the IL-6 family of cytokines [57]. In vitro incubation with or in vivo injection of leptin was shown to stimulate the release of pro-inflammatory mediators (e.g., TNF-$\alpha$, IL-1$\beta$ or cyclooxygenase-2) in brain cells [58, 59]. Therefore, high levels of leptin may impact the development of ASD by a pro-inflammatory mechanism, which was consistent with the result of the meta-analysis of high TPO-Ab in ASD.

Several limitations should be acknowledged. First, the number of published studies was relatively small for some indicators, such as TPO-Ab, IGFBP3, ghrelin and TG-Ab (only one study; thus, TG-Ab was not analyzed) [13]. Patients in female, adolescent and overweight subgroups were rarely enrolled, which led to the inconclusive ES estimate. Second, the mean and SD were provided for thyroid, growth and appetite hormones in NDDs. Whether the data of these indicators were outside the normal range and the corresponding proportion were very little described [60]. Third, few studies explored the diagnostic efficiency and threshold values of these biomarkers for predicting NDDs by ROC curve analysis and calculation of AUCs [13, 14, 61, 62]. Fourth, few research evaluated clinical factors that affected levels of thyroid, growth and appetite hormones, such as dietary intake, nutrition (albumin) [63], genetics (polymorphism) [14, 24, 62] and environmental factors (perfluoroalkyl substances) [64]. Fifth, there was evidence of significant heterogeneity in the majority of pooled analyses and it could not be eliminated after subgroup analyses. Sixth, considering the possibly same mechanisms for the development of various NDDs, the meta-analysis was firstly performed for the whole NDDs and then each NDD type. However, our research mainly focused on three types (ADHD, ASD and TD) of NDDs because of their high prevalence and relatively numerous studies. No studies reported the association of thyroid, growth and appetite hormones-related indicators with other types of NDDs [e.g., disorder of intellectual development (DID), developmental speech or language disorder (DSD) or developmental learning disorder (DLD)]. Seventh, the diagnosis criterion of NDDs was somewhat different among included studies (different DSM versions or others), which may influence prevalence rates of patients and levels of studied indicators in them. Therefore, more clinical trials with larger sample sizes and detailed information need to be designed to further understand the associations of these thyroid, growth and appetite biomarkers with various NDDs (including ADHD, ASD, TD, DID, DSD and DLD; classified using the same diagnosis criterion), and confirm their diagnostic thresholds and prediction performance before clinical generalization of these results.

The current meta-analysis showed a significant association of the high risk of NDDs in children and adolescents with significantly low levels of FT4, TT4, TSH, IGF-1 and high levels of FT3, TT3, TPO-Ab and leptin in bodily fluids. Thyroid hormones (FT3, FT4, TT3, TT4, TSH and TPO-A) were particularly suited as noninvasive biomarkers for monitoring the development of ADHD, while growth factor IGF-1 and the appetite hormone leptin (in addition to FT4 and TPO-Ab) may serve as more desirable biomarkers for predicting ASD.

NDDs, neurodevelopmental disorders; ADHD, attention deficit hyperactivity disorder; ASD, autism spectrum disorder; TD, tic disorder; OR, odds ratio; CI, confidence interval; THR, thyroid hormone receptor; FT3, free triiodothyronine; FT4, free thyroxine; TT3, total triiodothyronine; TT4, total thyroxine; TSH, thyroid stimulating hormone; TPO-Ab, thyroid peroxidase antibody; TG-Ab, thyroglobulin antibody; ROC, receiver operating characteristic; AUC, an area under the ROC curve; IGF-1, insulin-like growth factor-1; IGFBP-3, IGF-binding protein 3; PRISMA, Preferred Reporting Items for Systematic Reviews and Meta-Analyses; CSF, cerebrospinal fluid; NOS, Newcastle-Ottawa Scale; SMD, standardized mean difference; ES, effect size; H, heterogeneity; HC, healthy controls; DSM, diagnostic and statistical manual of mental disorders; ADI-R, autism diagnostic interview-revised; ADOS, autism diagnostic observation schedule; CARS, childhood autism rating scale; K-SADS-PL, kiddie schedule for affective disorders and schizophrenia of school-age children-present and lifetime version; DBDRS, diagnostic interview and disruptive behavior disorder rating scale; ICD-10, international statistical classification of diseases and related health problems 10th revision; CCMD, Chinese classification of mental disorders; ADH-IT, ADHD-predominantly inattentive type; ADHD-HIT, ADHD-the predominantly hyperactive/impulsive type; ADHD-CT, ADHD-combined type.

The data used for meta-analysis in the study are displayed in Supplementary Table 1.

HW: Data curation, Formal analysis, Writing — original draft. KH: Data curation, Formal analysis, Writing — original draft. LZP: Conceptualization, Methodology, Supervision, Writing — review & editing. XCX: Conceptualization, Supervision, Writing — review & editing. All authors read and approved the final manuscript. All authors have participated sufficiently in the work and agreed to be accountable for all aspects of the work.

Not applicable.

This research received no external funding.

The authors declare no conflict of interest.

Supplementary material associated with this article can be found, in the online version, at https://doi.org/10.31083/JIN39816.