Telomere Length and Oxidative Damage in Children and Adolescents with Autism Spectrum Disorder: A Systematic Review and Meta-Analysis

Abstract

Background:

Autism spectrum disorder (ASD) has been reported to confer an increased risk of natural premature death. Telomere erosion caused by oxidative stress is a common consequence in age-related diseases. However, whether telomere length (TL) and oxidative indicators are significantly changed in ASD patients compared with controls remains controversial. The aim of this study was to determine the associations of ASD with TL and oxidative indicators by performing a meta-analysis of all published evidence.

Methods:

The PubMed and Embase databases were searched for articles published up to April, 2024. The effect size was expressed as standardized mean difference (SMD) and 95% confidence interval (CI) via Stata 15.0 software.

Results:

Thirty-nine studies were included. Pooled results showed that compared with controls, children and adolescents with ASD were associated with significantly shorter TL (SMD = –0.48; 95% CI = –0.66– –0.29; p < 0.001; particularly in males), lower total antioxidant capacity (TAC: SMD = –1.15; 95% CI = –2.01– –0.30; p = 0.008), and higher oxidative DNA (8-hydroxy-2^′-deoxyguanosine, 8-OHdG: SMD = 0.63; 95% CI = 0.03–1.23; p = 0.039), lipid (hexanolyl-lysine, HEL: SMD = 0.37; 95% CI = 0.13–0.62; p = 0.003), and protein (3-nitrotyrosine, 3-NT: SMD = 0.86; 95% CI = 0.21–1.51; p = 0.01; dityrosine, DT: SMD = 0.66; 95% CI = 0.521–0.80; p < 0.01) damage. There were no significant differences between ASD and controls in 8-isoprostane and oxidative stress index after publication bias correction, and in N-formylkynurenine during overall meta-analysis.

Conclusions:

TL, 8-OHdG, TAC, HEL, 3-NT, and DT represent potential biomarkers for prediction of ASD in children and adolescents.

Graphical abstract

Keywords

autism / telomere length / oxidative stress / biomarker / meta-analysis

1. Introduction

Autism spectrum disorder (ASD) characterized by impairments in social communication and interaction as well as the presence of limited interests or repetitive, stereotypic behaviors, is a common neurodevelopmental disorder in children and adolescents, with a global prevalence of 0.6% [1]. Children with ASD have been reported at a significantly increased risk for premature mortality compared with controls [2]. Causes of death examination show that the premature mortality of ASD subjects mainly results from natural death, unintentional injury (drowning or traffic accident due to wandering or elopement) and intentional self-harm [3, 4]. Subsequently adjusted analysis confirms deaths from natural causes predominate [3]. These findings indicate that ASD may be associated with accelerated biological aging.

Telomeres are the repetitive DNA repeat sequences of 5^′-TTAGGG-3^′ located at the ends of linear eukaryotic chromosomes to prevent chromosome degradation and maintain genomic stability [5, 6]. Telomere length (TL) has been found to be progressively shortened with biological aging, which consequently limits cell proliferation and induces senescence or apoptosis in somatic cells, ultimately promoting the development of aging-related diseases and premature death [7, 8, 9]. Theoretically, TL should be shorter in patients with ASD than that in those without. This hypothesis was demonstrated by Zhang et al. [10] who detected the absolute quantitative TL was 5042.068 \pm 1950.595 kb in the ASD group and 6199.267 \pm 2931.947 kb in the typical development (TD) control group (p = 0.002). Li et al. [11] observed that the relative TL was 0.88 \pm 0.28 in patients with ASD and 1.01 \pm 0.43 in control (p = 0.006). Multivariate analysis also verified a significant correlation between TL shortening and the occurrence of ASD [10, 11]. However, Panahi et al. [12] and Lewis et al. [13] found no significant difference in average TL between ASD and controls (TD or healthy siblings). These inconsistent results point the necessity of re-examining the associations between TL and ASD.

Accumulating evidence has implied that activation of oxidative stress may play a central role in the pathogenesis of ASD [14]. Recently published meta-analyses have identified significant changes in levels of oxidative stress-related biomarkers for ASD patients: increased pro-oxidant nitric oxide, malondialdehyde and oxidized glutathione; decreased anti-oxidant glutathione and glutathione peroxidase [15, 16]. Studies on both WI-38 fibroblasts and HL-60 cells have found that relative to non-telomere sequence, telomere DNA sequence is five times more susceptible to oxidative stress induced by ultraviolet [17, 18]. A continuous, exponential correlation between reactive oxygen species (ROS) levels and telomere shortening rates is observed in fibroblasts from human and sheep [19]. Supplementation of antioxidant vitamin E and selenium is reported to slow telomere shortening of free-living white stork chicks [20]. The interaction of ROS with guanines in telomere sequences can result in the formation of 8-hydroxy-2^′-deoxyguanosine (8-OHdG) [17]. Therefore, oxidative stress-induced up-regulation of 8-OHdG may represent the mechanisms of telomere attrition in ASD patients. However, contradictory results had been revealed by different studies investigating the levels of 8-OHdG in ASD patients compared with controls: either a significant increase [10, 21] or without a statistical difference [22, 23].

Here, a comprehensive, systematic review and meta-analysis is designed to analyze the relationships between the risk of ASD and TL, 8-OHdG as well as other oxidative stress-related biomarkers that have not been investigated in published meta-analyses [15, 16]. The results may provide potential predictive biomarkers for the early detection and intervention of children and adolescents with ASD to prevent poor prognosis and guide their life trajectory to a more positive path [24].

2. Materials and Methods

2.1 Search Strategy

This systematic review and meta-analysis was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) 2020 guidelines [25]. The PRISMA_2020_checklist is included in the Supplementary Materials-PRISMA_2020_checklist. Two reviewers independently searched PubMed and Embase databases to identify articles published until April 2024 using the following keywords (“telomere” OR “8-hydroxy-2^′-deoxyguanosine” OR “8-OHdG” OR “8-oxo-7,8-dihydro-2^′-deoxyguanosine” OR “8-oxoguanine” OR “8-isoprostane” OR “8-iso-PGF2\alpha” OR “total antioxidant capacity” OR “total antioxidant power” OR “total antioxidant status” OR “thiol” OR “disulfide” OR “nitrotyrosine” OR “hexanolyl-lysine” OR “N-formylkynurenine” OR “oxidative stress markers” OR “oxidative stress index”) AND (“children” or “childhood” or “adolescents”) AND (“autism” OR “autistic disorder”). Searches were not restricted by language. Additionally, the reference lists of selected studies and reviews were manually checked to retrieve potentially eligible literatures.

2.2 Selection Criteria

Studies that met the following criteria were considered eligible: (1) Evaluated the associations between TL, 8-OHdG, oxidative stress biomarkers of interest (8-isoprostane, 8-iso-PGF2\alpha; hexanolyl-lysine, HEL; total antioxidant capacity, TAC; 3-nitrotyrosine, 3-NT; dityrosine, DT; N-formylkynurenine, NFK; oxidative stress index, OSI; thiol/disulfide) and ASD in children and adolescents (defined as 22 years based on the degree of individual cognition and socialization [24, 26]); (2) Designed as a case (ASD)-control comparative study; (3) Provided sufficient data to estimate effect sizes.

Exclusion criteria included: (1) Duplicate publications; (2) Non-original articles (case reports, reviews, conference abstract or comments); (3) Non-human experimental studies (in vitro and in vivo); (4) Lack of controls; (5) Biomarkers of interest not measured, measured in less than three studies or data unavailable; (6) Age of partial or all ASD patients >22 years; (7) Other irrelevant topics.

2.3 Data Extraction

Two reviewers independently collected the following information from each study: the first author, publication year, country, sample size, average age, gender ratio, diagnostic criteria for ASD, control type, sampling source and data of outcome measures (mean \pm standard deviation). Data were directly obtained from tables/texts or indirectly estimated from figures using the GetData Graph digitizer software (version 2.26; https://getdata-graph-digitizer.software.informer.com/). With regard to any discrepancy in the extracted data, a consensus was reached by thorough discussion with a third reviewer.

2.4 Quality Assessment

The quality of each included article was judged by two authors using the Newcastle–Ottawa Scale (NOS) [27] that consisted of three domains: patient selection (0–4 points), comparability (0–2 points) and outcome (0–3 points). NOS scores ranged from 0 to 9 stars and higher quality studies were defined as NOS scores >6.

2.5 Statistical Analysis

The meta-analysis was conducted using Stata software (version 15.0; Stata Corporation, College Station, TX, USA). Pooled effect size was presented as standardized mean difference (SMD) with 95% confidence interval (CI) and determined by Z-test, with a p-value 0.05 considered statistically significant. The heterogeneity between studies was evaluated using the Cochrane’s Q test and I² statistic. A value of p 0.1 and I² > 50% indicated substantial heterogeneity between studies and thus a random-effect model was selected to calculate composite results; a fixed-effect model was used when there was no significant heterogeneity (p > 0.1 and I² 50%). Subgroup analysis stratified by control type (TD or unaffected sibling), assay unit of TL [absolute TL in kbp and relative TL based on telomere (T) to single-copy gene (S) sequence (T/S) ratio], sample size (100 or \ge100), race (Asian or non-Asian), sample source (blood, urine, saliva or brain) and detection method for each variable was performed for variables with at least five analyzed data items to explore the potential cause of heterogeneity. Moreover, to further identify potential factors for heterogeneity, a multivariable meta-regression analysis was performed to test the impact of various subgroup covariates on study effect sizes (for variables with at least ten analyzed data items). Egger’s linear regression test was utilized to analyze publication bias. If bias was present (p 0.05), the trim-and-fill method was used to correct pooled results. A sensitivity analysis was carried out with the leave-one-out method to explore the effect of each study on the overall result.

3. Results

3.1 Literature Search

Electronic search located 1403 studies, 969 of which were discarded due to duplicate publications. Reading titles and abstracts excluded 390 records as they were case reports (n = 4), reviews/meta-analysis (n = 40), conference abstract or comments (n = 11), animal experiments (n = 104), cell experiments (n = 1), lack of controls (n = 5), adult ASD (n = 50) and irrelevant topic (n = 175). The full-text of the remaining 44 studies was downloaded and six were removed due to: age of partial ASD patients >22 years (n = 3), data unavailable (n = 1) and biomarkers detected for mothers of ASD children (n = 1). Ultimately, 39 eligible studies were included in this meta-analysis [10, 11, 12, 13, 21, 22, 23, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59] (Fig. 1).

3.2 Study Characteristics

Details of all included studies are summarized in Table 1 (Ref. [10, 11, 12, 13, 21, 22, 23, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59]) and Supplementary Table 1. All 39 studies were published in English from 2005 to 2024. Eight studies were undertaken in the USA, four in Egypt, Japan, Saudi Arabia, two in China, India, Iran, Slovenia, Turkey, UK and one in the Belgium, Italy, Nigeria, Qatar, Romania, Slovakia, Spain, respectively. Although the gender ratio of ASD patients varied in different studies, the number of males was typically greater than females in most of the studies. Age was 12 years old (defined as children, which is a stage for the acquisition of brain plasticity and learning capacity) for most participants and 12 to 22 years old in sporadic cases (defined as adolescents, which is the transition period between childhood and adulthood to have the capacity for planning and self-regulation) [22]. ASD patients were diagnosed according to the criteria detailed in the diagnostic and statistical manual of mental disorders (fourth or fifth version), autism diagnostic interview-revised, the autism diagnostic observation schedule, the social communication questionnaire, the childhood autism rating scale, the autism behavior checklist, the developmental, dimensional and diagnostic interview and the international classification of diseases. All studies were high in methodological quality, achieving eight or nine stars.

3.3 Meta-Analysis of the Association of TL with ASD in Children and Adolescents

Six studies [10, 11, 12, 13, 28, 29] with 11 datasets detected TL in ASD patients and controls by polymerase chain reaction assays (Supplementary Table 1). Meta-analysis results showed that TL tended to be significantly shortened in patients with ASD compared with controls (SMD = –0.48; 95% CI = –0.66– –0.29; p 0.001) (Table 2; Fig. 2). Subgroup analysis indicated that TL was only significantly shorter in ASD when compared to TD (SMD = –0.49; 95% CI = –0.67– –0.31; p 0.001), but not when compared to unaffected siblings (p = 0.200). Subgroup analyses based on TL assay unit, sample size, race and sample source also demonstrated telomere shortening in ASD subjects (Supplementary Table 2). Also, the heterogeneity was removed when the sample size was larger than 100. This heterogeneity factor was confirmed in meta-regression analysis (Supplementary Table 3): effect size tended to increase with increasing sample size.

Some studies additionally analyzed male or female cases independently or only enrolled males. Therefore, meta-analysis for males and females was separated. Meta-analysis of four studies with seven datasets [11, 12, 13, 28] (Supplementary Table 1) showed that male ASD patients possessed a highly significant decrease in TL in comparison to controls (SMD = –0.25; 95% CI = –0.42– –0.09; p = 0.002; particularly for subgroups with TD controls, p = 0.014; Asian population, p = 0.009; saliva samples, p = 0.006) (Table 2). There was no significant difference in TL between female ASD patients and controls regardless of overall (SMD = –0.20; 95% CI = –1.11– 0.72; p = 0.670) [12, 13, 28] or subgroup meta-analyses (Supplementary Table 2).

3.4 Meta-Analysis of the Association of Oxidative Stress Biomarkers with ASD in Children and Adolescents

3.4.1 8-OHdG

Eleven studies [10, 21, 22, 23, 30, 31, 32, 33, 34, 35, 36] investigated the difference in levels of oxidative DNA damage biomarker 8-OHdG between ASD and controls (Supplementary Table 1). Pooled analysis of the data from these studies found that the concentration of 8-OhdG was significantly increased in patients with ASD compared with controls (SMD = 0.63; 95% CI = 0.03–1.23; p = 0.039) (Table 2, Fig. 3), particularly for those with mild-moderate severity [21, 23] (SMD = 1.97; 95% CI = 0.32–3.62; p = 0.019) (Table 2).

3.4.2 8-iso-PGF2\alpha

Seven studies [31, 32, 33, 37, 38, 39, 40] compared 8-iso-PGF2\alpha levels in ASD patients with that in healthy controls (Supplementary Table 1). Meta-analysis observed a significant increase in levels of 8-iso-PGF2\alpha in the ASD group compared with controls (SMD = 4.61; 95% CI = 2.61–6.60; p 0.001) (Table 2). Subgroup analysis based on sample size, race, sample source and detection method all further demonstrated this significant association (Supplementary Table 2). Furthermore, two studies [37, 38] with three datasets specifically measured 8-iso-PGF2\alpha levels in ASD patients with mild-moderate or severe impairments. Pooled analysis was independently performed for them. A consistent conclusion was achieved with this analysis and a significantly larger effect size was found (mild-moderate: SMD = 12.11; 95% CI = 7.93–16.30; p 0.001; severe: SMD = 8.48; 95% CI = 4.93–12.03; p 0.001) (Table 2).

3.4.3 TAC

TAC in ASD patients and controls was assessed in 13 studies [10, 22, 33, 34, 35, 41, 42, 43, 44, 45, 46, 47, 48] with 18 datasets (Supplementary Table 1). Random-effects meta-analysis revealed that TAC levels were significantly decreased in ASD patients compared with controls (SMD = –1.15; 95% CI = –2.01– –0.30; p = 0.008) (Table 2). Subgroup analysis showed that significantly lower TAC levels were only detected in urine samples of ASD patients compared with TD controls (Supplementary Table 2).

3.4.4 HEL

Five studies [22, 31, 33, 34, 35] provided the HEL levels in ASD patients and controls (Supplementary Table 1). Results from meta-analyses indicated that HEL levels were significantly increased in ASD patients compared with controls (SMD = 0.37; 95% CI = 0.13–0.62; p = 0.003) (Table 2). Subgroup analysis showed a significant increase in studies with sample size 100 (p = 0.001) and non-Asian population (p = 0.006) (Supplementary Table 2).

3.4.5 OSI

Four studies [41, 42, 43, 47] with six datasets evaluated changes in OSI between ASD patients and controls (Supplementary Table 1). Pooled results observed that OSI in ASD patients was significantly higher than that in controls (SMD = 6.76; 95% CI = 4.20–9.31; p 0.001) (Table 2). Subgroup analysis demonstrated a significant increase in studies with sample size 100 (p = 0.005) and Asian population (p = 0.002) (Supplementary Table 2). Meta-analysis of data from ASD patients with mild-moderate (SMD = 14.00) or severe (SMD = 5.76) impairments also indicated a remarkable increase in OSI (Table 2).

3.4.6 Total Thiol

Four independent publications [45, 49, 50, 51] measured total thiol levels in ASD patients and controls (Supplementary Table 1). Meta-analysis found decreased levels of total thiol in ASD patients compared with controls, but the p-value was approximately 0.05 (SMD = –1.65; 95% CI = –3.24– –0.06; p = 0.042) (Table 2), indicating its association with ASD needs further investigation.

3.4.7 3-NT

Eight studies [52, 53, 54, 55, 56, 57, 58, 59] involving 29 datasets were included to evaluate the association of 3-NT with ASD (Supplementary Table 1). Pooled analysis of these data found that 3-NT was significantly increased in ASD patients compared with controls (SMD = 0.86; 95% CI = 0.21–1.51; p = 0.01) (Table 2). Subgroup analysis showed that 3-NT was significantly increased in the blood samples of ASD patients (p = 0.005), but not significantly changed in brain tissues (p = 0.117) (Supplementary Table 2).

3.4.8 DT

Four studies [31, 53, 58, 59] involving six datasets explored the association of DT with ASD (Supplementary Table 1). Combined analysis found that DT was significantly increased in ASD patients compared with controls (SMD = 0.66; 95% CI = 0.52–0.80; p 0.001) (Table 2). Subgroup analysis showed that DT residue contents of proteins in blood and urine were both significantly increased (p 0.01). Sample size, race and detection method factors did not change this significant result (Supplementary Table 2).

3.4.9 NFK

Three studies [53, 58, 59] involving five datasets explored the association of NFK, a marker of oxidation in the tryptophan residue of proteins, with ASD (Supplementary Table 1). Meta-analysis revealed no significant difference in NFK residue contents of proteins between ASD and control groups (p = 0.525). No significant result was obtained in all subgroups with the number of datasets >1 (Supplementary Table 2).

3.5 Publication Bias and Sensitivity Analysis

Egger’s linear regression test did not find evidence of publication bias for analysis of TL, 8-OHdG, HEL, total thiol, 3-NT and NFK, whereas potential publication bias was observed for 8-iso-PGF2\alpha (all cases, p = 0.002; severe cases, p = 0.044), TAC (all cases, p = 0.006; mild-moderate cases, p = 0.014), OSI (all cases, p = 0.025; mild-moderate cases, p = 0.005) and DT (p = 0.004) (Table 2).

The trim-and-fill method was subsequently used to adjust the effect sizes for variables with publication bias. Results showed that although the SMD was slightly changed, the levels of DT (SMD = 0.59; 95% CI = 0.39–0.80; p 0.01) and 8-iso-PGF2\alpha (severe: SMD = 5.39; 95% CI = 1.82–8.96; p = 0.03) were still significantly increased in ASD similar to the pre-correction analysis; the SMD and 95% CI were not significantly changed for TAC; 8-iso-PGF2\alpha (all cases, p = 0.333) and OSI (all cases, p = 0.450; mild-moderate cases, p = 0.074) levels in ASD patients were not significantly different from those in controls.

Sensitivity analyses by excluding one study in turn did not find significant changes in the effective estimates for all biomarkers, indicating the robustness of the meta-analysis outcomes (Fig. 4).

4. Discussion

Although there have meta-analyses exploring the association between TL and neurological disorders, they have mainly focused on adult disease, such as schizophrenia [60, 61], Alzheimer [62] and Parkinson [63]. To our knowledge, this is the first meta-analysis to integrate all published studies to comprehensively confirm TL changes in children and adolescents with ASD. Consistent with other neurological disorders [60, 61, 62, 63], pooled results showed that TL in children and adolescents with ASD was significantly reduced compared with healthy controls (particularly TD). Also, this significant result was not influenced by TL assay unit, sample size, race and sample source. Furthermore, analysis of prevalence studies indicated that the ratio of male to female was about 3:1 among ASD children [64]. First impressions have suggested males with ASD were rated lower than females on social cognition and object relations, emotional investment, and social causality scales [65, 66]. Females decreased more in symptom severity than males across childhood [67]. These findings indicated that TL may be particularly decreased in male ASD patients, which was confirmed by this study. The negative associations of TL with female ASD may be resulted from the small sample size of them (relative to males) and needs to be further confirmed through increased statistical power.

The underlying mechanism of telomere shortening in ASD patients remains not completely understood, but activation of oxidative stress was considered as a major factor because Zhang et al. [10] found that decreased TL was positively correlated with lower activity of catalase (CAT) which is an important antioxidant enzyme for catalyzing H₂O₂ to H₂O and O₂ in maintaining the redox balance. The accumulated H₂O₂ due to lower activity of CAT may specifically attack the 8th carbon atom of the DNA guanine base in the telomere sequence and lead to the production of 8-OHdG [17, 18, 68]. The presence of oxidative lesions in the DNA can promote DNA single- or double-strand breaks at telomeric regions and cause the loss of the distal fragments of telomeric DNA and consequent telomere attrition [69, 70]. To confirm oxidative DNA damage status in children and adolescents with ASD, studies with 8-OHdG in ASD patients were also retrieved. As expected, meta-analysis of 11 studies showed that subjects with ASD (particularly mild-moderate severity) had an increased 8-OHdG concentration, which was concordant with meta-analyses for mental illnesses [71, 72].

Although multivariate analysis by Zhang et al. [10] found that CAT was an independent biomarker for prediction of ASD, meta-analysis of all evidence did not detect their significant association [16], indicating the necessity of identification of more effective oxidative stress biomarkers for young subjects with ASD. In this study, studies providing the data of 8-iso-PGF2\alpha, OSI, TAC, HEL, thiol, 3-NT, DT and NFK in ASD patients and controls were incorporated. Meta-analysis showed that 8-iso-PGF2\alpha, OSI, HEL, 3-NT and DT were significantly increased, whereas TAC was significantly decreased in ASD patients compared with controls. However, publication bias correction analysis excluded the significance of 8-iso-PGF2\alpha and OSI. TAC is an assessment of the cumulative action of all antioxidants (including redox synergistic interactions) and may represent a better biomarker than individual anti-oxidants (such as CAT) [73]. Moreover, subgroup analysis showed that TAC in urine and saliva samples of ASD patients was significantly changed, but not in blood, indicating urine is more valuable biological matrix for TAC assay [35]. With the exception of DNA, oxidative stress also causes lipid and protein damages. HEL is a lipid peroxidation biomarker formed by the reaction of lysine with peroxidized n-6 polyunsaturated fatty acid (PUFA) [74]. n-6 PUFA has been reported to exert protective roles against neuronal degeneration and neuronal loss in experimental rats with valproic acid-induced autism [75]. A decline in levels of n-6 PUFA to cause HEL elevation due to oxidative stress has been found as a risk factor for inducing autism [22, 35, 76]. 3-NT is a marker of tyrosine nitration of proteins, which is generated through a reaction between the tyrosyl radical and nitrogen dioxide [77], whereas DT is a marker of oxidation in the tyrosine residue of proteins, which is derived from a reaction between the tyrosyl radical and reactive oxygen species. An in vivo experiment has demonstrated that compared to free-tyrosine, free-3-NT (nitration of amino acids) treatment induces more nigrostriatal cell body loss and increases net ipsilateral turning behavior (consistent with degeneration of striatal dopaminergic nerve terminals) [78]. An in vitro study has also proved that introduction of 3-NT can damage the cytoskeleton and induce apoptosis of neuronal PC12 cells [79]. Compared with the saline group, DT administration has been observed to impair hippocampus-dependent nonspatial memory in young adulthood mice [80]. Therefore, decreased TAC and increased HEL, 3-NT and DT may be potential biomarkers reflecting the oxidative stress status of ASD.

Several limitations should be acknowledged when interpreting results reported here. First, the number of published studies was relatively small (10 for most indicators except of TAC and 8-OHdG) and the sample size in these studies was not large (100 for most articles), which may lead to underestimation or overestimation of the effect size for variables. Additionally, Egger’s linear regression test showed the presence of publication bias for 8-iso-PGF2\alpha and OSI and the trim-and-fill correction analysis removed the significance of them in ASD, indicating “missing” studies obviously influenced the results. Subgroup analysis for TL detected the heterogeneity was removed when the sample size was \ge100. Meta-regression analysis also confirmed the sample size as a moderator variable on effect sizes of TL. Second, Qasem et al. [37] found that age had a significant negative correlation with 8-iso-PGF2\alpha plasma levels in ASD patients. Hirayama et al. [36] reported that 8-OHdG was only significantly increased in ASD children with an age under 6 years compared to that in the age-matched male TD group, not significantly changed in the analysis of all cases. TL of ASD patients was identified to be progressively reduced with an increase in the chronological age [12, 81]. These findings indicate biomarkers may be useful as a screening test for specific age or sex group. However, no raw data for each patient with different behavioral and clinicopathological characteristics (e.g., age, sex) were available and thus subgroup or meta-regression analyses based on these risk factors could not be performed for all indicators. Third, the receiver operating characteristic curve was rarely drawn to confirm the diagnostic efficiency of TL [10] and oxidative stress biomarkers [49]. Detection method for biomarkers in different articles also varied. Therefore, which length or concentration should be designed as cut-off values for diagnosis of ASD remains uncertain. Fourth, this meta-analysis evaluated the associations between ASD and biomarkers based on the univariate analysis results in included articles. Multivariate analysis data were rarely available [10, 11]. Fifth, accumulation of oxidative damage at telomeres and telomere shortening also suggested to be attributed to decreased DNA repair [82, 83]. Recovering DNA repair gene expression can ameliorate autism-like behaviors in mice [84]. However, until now, no human studies detected the expression of DNA repair genes. Sixth, there was a lack of sufficient studies that analyzed the changes of selected biomarkers in adult ASD subjects, which could be interesting for future longitudinal study of ASD. Accordingly, future analyses in larger, more diverse populations with detailed individual information are warranted to powerfully confirm the connection of TL, oxidative stress biomarkers and DNA repair genes with ASD before clinical use.

5. Conclusions

This meta-analysis demonstrates that children and adolescents with ASD are associated with significantly shorter TL, lower TAC and higher oxidative DNA (8-OHdG), lipid (HEL) and protein (3-NT and DT) damages. These indicators may have the potential to be used as biomarkers for early prediction of ASD in children and adolescents. However, more clinical researches in larger, more diverse ASD populations are still needed in order to provide a strong evidence for these biomarkers because the number of articles for most indicators is relatively less.

References

[1] Salari N, Rasoulpoor S, Rasoulpoor S, Shohaimi S, Jafarpour S, Abdoli N, et al. The global prevalence of autism spectrum disorder: a comprehensive systematic review and meta-analysis. Italian Journal of Pediatrics. 2022; 48: 112.

[2] Yoo SM, Kim KN, Kang S, Kim HJ, Yun J, Lee JY. Prevalence and Premature Mortality Statistics of Autism Spectrum Disorder Among Children in Korea: A Nationwide Population-Based Birth Cohort Study. Journal of Korean Medical Science. 2022; 37: e1.

[3] Jokiranta-Olkoniemi E, Gyllenberg D, Sucksdorff D, Suominen A, Kronström K, Chudal R, et al. Risk for Premature Mortality and Intentional Self-harm in Autism Spectrum Disorders. Journal of Autism and Developmental Disorders. 2021; 51: 3098–3108.

[4] Catalá-López F, Hutton B, Page MJ, Driver JA, Ridao M, Alonso-Arroyo A, et al. Mortality in Persons With Autism Spectrum Disorder or Attention-Deficit/Hyperactivity Disorder: A Systematic Review and Meta-analysis. JAMA Pediatrics. 2022; 176: e216401.

[5] Blackburn EH. Structure and function of telomeres. Nature. 1991; 350: 569–573.

[6] Lu W, Zhang Y, Liu D, Songyang Z, Wan M. Telomeres-structure, function, and regulation. Experimental Cell Research. 2013; 319: 133–141.

[7] Ishikawa N, Nakamura KI, Izumiyama-Shimomura N, Aida J, Matsuda Y, Arai T, et al. Changes of telomere status with aging: An update. Geriatrics & Gerontology International. 2016; 16: 30–42.

[8] Gampawar P, Schmidt R, Schmidt H. Telomere length and brain aging: A systematic review and meta-analysis. Ageing Research Reviews. 2022; 80: 101679.

[9] Wang Q, Zhan Y, Pedersen NL, Fang F, Hägg S. Telomere Length and All-Cause Mortality: A Meta-analysis. Ageing Research Reviews. 2018; 48: 11–20.

[10] Zhang T, Sun Y, Wei J, Zhao G, Hao W, Lv Z, et al. Shorter telomere length in children with autism spectrum disorder is associated with oxidative stress. Frontiers in Psychiatry. 2023; 14: 1209638.

[11] Li Z, Tang J, Li H, Chen S, He Y, Liao Y, et al. Shorter telomere length in peripheral blood leukocytes is associated with childhood autism. Scientific Reports. 2014; 4: 7073.

[12] Panahi Y, Salasar Moghaddam F, Babaei K, Eftekhar M, Shervin Badv R, Eskandari MR, et al. Sexual Dimorphism in Telomere Length in Childhood Autism. Journal of Autism and Developmental Disorders. 2023; 53: 2050–2061.

[13] Lewis CR, Taguinod F, Jepsen WM, Cohen J, Agrawal K, Huentelman MJ, et al. Telomere Length and Autism Spectrum Disorder Within the Family: Relationships With Cognition and Sensory Symptoms. Autism Research: Official Journal of the International Society for Autism Research. 2020; 13: 1094–1101.

[14] Liu X, Lin J, Zhang H, Khan NU, Zhang J, Tang X, et al. Oxidative Stress in Autism Spectrum Disorder-Current Progress of Mechanisms and Biomarkers. Frontiers in Psychiatry. 2022; 13: 813304.

[15] Frustaci A, Neri M, Cesario A, Adams JB, Domenici E, Dalla Bernardina B, et al. Oxidative stress-related biomarkers in autism: systematic review and meta-analyses. Free Radical Biology & Medicine. 2012; 52: 2128–2141.

[16] Chen L, Shi XJ, Liu H, Mao X, Gui LN, Wang H, et al. Oxidative stress marker aberrations in children with autism spectrum disorder: a systematic review and meta-analysis of 87 studies (N = 9109). Translational Psychiatry. 2021; 11: 15.

[17] Kawanishi S, Oikawa S. Mechanism of telomere shortening by oxidative stress. Annals of the New York Academy of Sciences. 2004; 1019: 278–284.

[18] Oikawa S, Tada-Oikawa S, Kawanishi S. Site-specific DNA damage at the GGG sequence by UVA involves acceleration of telomere shortening. Biochemistry. 2001; 40: 4763–4768.

[19] Richter T, von Zglinicki T. A continuous correlation between oxidative stress and telomere shortening in fibroblasts. Experimental Gerontology. 2007; 42: 1039–1042.

[20] Pineda-Pampliega J, Herrera-Dueñas A, Mulder E, Aguirre JI, Höfle U, Verhulst S. Antioxidant supplementation slows telomere shortening in free-living white stork chicks. Proceedings. Biological Sciences. 2020; 287: 20191917.

[21] El-Ansary A, Cannell JJ, Bjørklund G, Bhat RS, Al Dbass AM, Alfawaz HA, et al. In the search for reliable biomarkers for the early diagnosis of autism spectrum disorder: the role of vitamin D. Metabolic Brain Disease. 2018; 33: 917–931.

[22] Imataka G, Yui K, Shiko Y, Kawasaki Y, Sasaki H, Shiroki R, et al. Urinary and Plasma Antioxidants in Behavioral Symptoms of Individuals With Autism Spectrum Disorder. Frontiers in Psychiatry. 2021; 12: 684445.

[23] Osredkar J, Kumer K, Fabjan T, Jekovec Vrhovšek M, Maček J, Zupan M, et al. Determination of modified nucleosides in the urine of children with autism spectrum disorder. Acta Biochimica Polonica. 2023; 70: 335–342.

[24] Pinto TM, Laurence PG, Macedo CR, Macedo EC. Resilience Programs for Children and Adolescents: A Systematic Review and Meta-Analysis. Frontiers in Psychology. 2021; 12: 754115.

[25] Page MJ, McKenzie JE, Bossuyt PM, Boutron I, Hoffmann TC, Mulrow CD, et al. The PRISMA 2020 statement: An updated guideline for reporting systematic reviews. Journal of Clinical Epidemiology. 2021; 134: 178–189.

[26] Huang J, Du C, Liu J, Tan G. Meta-Analysis on Intervention Effects of Physical Activities on Children and Adolescents with Autism. International Journal of Environmental Research and Public Health. 2020; 17: 1950.

[27] Lo CKL, Mertz D, Loeb M. Newcastle-Ottawa Scale: comparing reviewers’ to authors’ assessments. BMC Medical Research Methodology. 2014; 14: 45.

[28] Nelson CA, Varcin KJ, Coman NK, De Vivo I, Tager-Flusberg H. Shortened Telomeres in Families With a Propensity to Autism. Journal of the American Academy of Child and Adolescent Psychiatry. 2015; 54: 588–594.

[29] Salem S, Ashaat E. Association of Relative Telomere Length and LINE-1 Methylation with Autism but not with Severity. Journal of Autism and Developmental Disorders. 2024; 54: 2266–2273.

[30] Sajdel-Sulkowska EM, Xu M, Koibuchi N. Increase in cerebellar neurotrophin-3 and oxidative stress markers in autism. Cerebellum (London, England). 2009; 8: 366–372.

[31] Osredkar J, Gosar D, Maček J, Kumer K, Fabjan T, Finderle P, et al. Urinary Markers of Oxidative Stress in Children with Autism Spectrum Disorder (ASD). Antioxidants (Basel, Switzerland). 2019; 8: 187.

[32] Ming X, Stein TP, Brimacombe M, Johnson WG, Lambert GH, Wagner GC. Increased excretion of a lipid peroxidation biomarker in autism. Prostaglandins, Leukotrienes, and Essential Fatty Acids. 2005; 73: 379–384.

[33] Ghezzo A, Visconti P, Abruzzo PM, Bolotta A, Ferreri C, Gobbi G, et al. Oxidative Stress and Erythrocyte Membrane Alterations in Children with Autism: Correlation with Clinical Features. PloS One. 2013; 8: e66418.

[34] Yui K, Tanuma N, Yamada H, Kawasaki Y. Reduced endogenous urinary total antioxidant power and its relation of plasma antioxidant activity of superoxide dismutase in individuals with autism spectrum disorder. International Journal of Developmental Neuroscience: the Official Journal of the International Society for Developmental Neuroscience. 2017; 60: 70–77.

[35] Yui K, Tanuma N, Yamada H, Kawasaki Y. Decreased total antioxidant capacity has a larger effect size than increased oxidant levels in urine in individuals with autism spectrum disorder. Environmental Science and Pollution Research International. 2017; 24: 9635–9644.

[36] Hirayama A, Wakusawa K, Fujioka T, Iwata K, Usui N, Kurita D, et al. Simultaneous evaluation of antioxidative serum profiles facilitates the diagnostic screening of autism spectrum disorder in under-6-year-old children. Scientific Reports. 2020; 10: 20602.

[37] Qasem H, Al-Ayadhi L, El-Ansary A. Cysteinyl leukotriene correlated with 8-isoprostane levels as predictive biomarkers for sensory dysfunction in autism. Lipids in Health and Disease. 2016; 15: 130.

[38] Mostafa GA, El-Hadidi ES, Hewedi DH, Abdou MM. Oxidative stress in Egyptian children with autism: relation to autoimmunity. Journal of Neuroimmunology. 2010; 219: 114–118.

[39] Pop B, Niculae AȘ Pop TL, Răchișan AL. Individuals with autism have higher 8-Iso-PGF2α levels than controls, but no correlation with quantitative assay of Paraoxonase 1 serum levels. Metabolic Brain Disease. 2017; 32: 1943–1950.

[40] Yao Y, Walsh WJ, McGinnis WR, Praticò D. Altered vascular phenotype in autism: correlation with oxidative stress. Archives of Neurology. 2006; 63: 1161–1164.

[41] Omotosho IO, Akinade AO, Lagunju IA, Yakubu MA. Oxidative stress indices in ASD children in Sub-Sahara Africa. Journal of Neurodevelopmental Disorders. 2021; 13: 50.

[42] Saleem TH, Shehata GA, Toghan R, Sakhr HM, Bakri AH, Desoky T, et al. Assessments of Amino Acids, Ammonia and Oxidative Stress Among Cohort of Egyptian Autistic Children: Correlations with Electroencephalogram and Disease Severity. Neuropsychiatric Disease and Treatment. 2020; 16: 11–24.

[43] Damodaran LPM, Arumugam G. Urinary oxidative stress markers in children with autism. Redox Report: Communications in Free Radical Research. 2011; 16: 216–222.

[44] Rai K, Hegde AM, Jose N. Salivary antioxidants and oral health in children with autism. Archives of Oral Biology. 2012; 57: 1116–1120.

[45] Ranjbar A, Rashedi V, Rezaei M. Comparison of urinary oxidative biomarkers in Iranian children with autism. Research in Developmental Disabilities. 2014; 35: 2751–2755.

[46] Parellada M, Moreno C, Mac-Dowell K, Leza JC, Giraldez M, Bailón C, et al. Plasma antioxidant capacity is reduced in Asperger syndrome. Journal of Psychiatric Research. 2012; 46: 394–401.

[47] Hassan MH, Desoky T, Sakhr HM, Gabra RH, Bakri AH. Possible Metabolic Alterations among Autistic Male Children: Clinical and Biochemical Approaches. Journal of Molecular Neuroscience: MN. 2019; 67: 204–216.

[48] Jasenovec T, Radosinska D, Jansakova K, Kopcikova M, Tomova A, Snurikova D, et al. Alterations in Antioxidant Status and Erythrocyte Properties in Children with Autism Spectrum Disorder. Antioxidants (Basel, Switzerland). 2023; 12: 2054.

[49] Ayaydın H, Kılıçaslan F, Koyuncu İ Çelik H, Çalık M, Güzelçiçek A, et al. Impaired Thiol/Disulfide Homeostasis in Children Diagnosed with Autism: A Case-Control Study. Journal of Molecular Neuroscience: MN. 2021; 71: 1394–1402.

[50] Efe A, Neşelioğlu S, Soykan A. An Investigation of the Dynamic Thiol/Disulfide Homeostasis, As a Novel Oxidative Stress Plasma Biomarker, in Children With Autism Spectrum Disorders. Autism Research: Official Journal of the International Society for Autism Research. 2021; 14: 473–487.

[51] Ramaekers VT, Sequeira JM, Thöny B, Quadros EV. Oxidative Stress, Folate Receptor Autoimmunity, and CSF Findings in Severe Infantile Autism. Autism Research and Treatment. 2020; 2020: 9095284.

[52] Khan A, Harney JW, Zavacki AM, Sajdel-Sulkowska EM. Disrupted brain thyroid hormone homeostasis and altered thyroid hormone-dependent brain gene expression in autism spectrum disorders. Journal of Physiology and Pharmacology: an Official Journal of the Polish Physiological Society. 2014; 65: 257–272.

[53] Anwar A, Marini M, Abruzzo PM, Bolotta A, Ghezzo A, Visconti P, et al. Quantitation of plasma thiamine, related metabolites and plasma protein oxidative damage markers in children with autism spectrum disorder and healthy controls. Free Radical Research. 2016; 50: S85–S90.

[54] Nadeem A, Ahmad SF, Attia SM, Al-Ayadhi LY, Bakheet SA, Al-Harbi NO. Oxidative and inflammatory mediators are upregulated in neutrophils of autistic children: Role of IL-17A receptor signaling. Progress in Neuro-psychopharmacology & Biological Psychiatry. 2019; 90: 204–211.

[55] Sajdel-Sulkowska EM, Xu M, McGinnis W, Koibuchi N. Brain region-specific changes in oxidative stress and neurotrophin levels in autism spectrum disorders (ASD). Cerebellum (London, England). 2011; 10: 43–48.

[56] Frye RE, Delatorre R, Taylor H, Slattery J, Melnyk S, Chowdhury N, et al. Redox metabolism abnormalities in autistic children associated with mitochondrial disease. Translational Psychiatry. 2013; 3: e273.

[57] Al-Bishri WM. Glucose transporter 1 deficiency, AMP-activated protein kinase activation and immune dysregulation in autism spectrum disorder: Novel biomarker sources for clinical diagnosis. Saudi Journal of Biological Sciences. 2023; 30: 103849.

[58] Anwar A, Abruzzo PM, Pasha S, Rajpoot K, Bolotta A, Ghezzo A, et al. Advanced glycation endproducts, dityrosine and arginine transporter dysfunction in autism - a source of biomarkers for clinical diagnosis. Molecular Autism. 2018; 9: 3.

[59] Al-Saei ANJM, Nour-Eldine W, Rajpoot K, Arshad N, Al-Shammari AR, Kamal M, et al. Validation of plasma protein glycation and oxidation biomarkers for the diagnosis of autism. Molecular Psychiatry. 2024; 29: 653–659.

[60] Polho GB, De-Paula VJ, Cardillo G, dos Santos B, Kerr DS. Leukocyte telomere length in patients with schizophrenia: A meta-analysis. Schizophrenia Research. 2015; 165: 195–200.

[61] Rao S, Kota LN, Li Z, Yao Y, Tang J, Mao C, et al. Accelerated leukocyte telomere erosion in schizophrenia: Evidence from the present study and a meta-analysis. Journal of Psychiatric Research. 2016; 79: 50–56.

[62] Forero DA, González-Giraldo Y, López-Quintero C, Castro-Vega LJ, Barreto GE, Perry G. Meta-analysis of Telomere Length in Alzheimer’s Disease. The Journals of Gerontology. Series A, Biological Sciences and Medical Sciences. 2016; 71: 1069–1073.

[63] Forero DA, González-Giraldo Y, López-Quintero C, Castro-Vega LJ, Barreto GE, Perry G. Telomere length in Parkinson’s disease: A meta-analysis. Experimental Gerontology. 2016; 75: 53–55.

[64] Loomes R, Hull L, Mandy WPL. What Is the Male-to-Female Ratio in Autism Spectrum Disorder? A Systematic Review and Meta-Analysis. Journal of the American Academy of Child and Adolescent Psychiatry. 2017; 56: 466–474.

[65] Cola ML, Plate S, Yankowitz L, Petrulla V, Bateman L, Zampella CJ, et al. Sex differences in the first impressions made by girls and boys with autism. Molecular Autism. 2020; 11: 49.

[66] Mattern H, Cola M, Tena KG, Knox A, Russell A, Pelella MR, et al. Sex differences in social and emotional insight in youth with and without autism. Molecular Autism. 2023; 14: 10.

[67] Waizbard-Bartov E, Ferrer E, Heath B, Rogers SJ, Nordahl CW, Solomon M, et al. Identifying autism symptom severity trajectories across childhood. Autism Research: Official Journal of the International Society for Autism Research. 2022; 15: 687–701.

[68] Midorikawa K, Hirakawa K, Kawanishi S. Hydroxylation of deoxyguanosine at 5’ site of GG and GGG sequences in double-stranded DNA induced by carbamoyl radicals. Free Radical Research. 2002; 36: 667–675.

[69] Tarry-Adkins JL, Martin-Gronert MS, Chen JH, Cripps RL, Ozanne SE. Maternal diet influences DNA damage, aortic telomere length, oxidative stress, and antioxidant defense capacity in rats. FASEB Journal: Official Publication of the Federation of American Societies for Experimental Biology. 2008; 22: 2037–2044.

[70] Srikanth P, Chowdhury AR, Low GKM, Saraswathy R, Fujimori A, Banerjee B, et al. Oxidative Damage Induced Telomere Mediated Genomic Instability in Cells from Ataxia Telangiectasia Patients. Genome Integrity. 2022; 13: 2.

[71] Çeli K HEA, Tuna G, Ceylan D, Küçükgöncü S. A comparative meta-analysis of peripheral 8-hydroxy-2’-deoxyguanosine (8-OHdG) or 8-oxo-7,8-dihydro-2’-deoxyguanosine (8-oxo-dG) levels across mood episodes in bipolar disorder. Psychoneuroendocrinology. 2023; 151: 106078.

[72] Goh XX, Tang PY, Tee SF. 8-Hydroxy-2’-Deoxyguanosine and Reactive Oxygen Species as Biomarkers of Oxidative Stress in Mental Illnesses: A Meta-Analysis. Psychiatry Investigation. 2021; 18: 603–618.

[73] Serafini M, Del Rio D. Understanding the association between dietary antioxidants, redox status and disease: is the Total Antioxidant Capacity the right tool? Redox Report: Communications in Free Radical Research. 2004; 9: 145–152.

[74] Kato Y, Osawa T. Detection of a lipid-lysine adduct family with an amide bond as the linkage: novel markers for lipid-derived protein modifications. Methods in Molecular Biology (Clifton, N.J.). 2009; 580: 129–141.

[75] Yadav S, Tiwari V, Singh M, Yadav RK, Roy S, Devi U, et al. Comparative efficacy of alpha-linolenic acid and gamma-linolenic acid to attenuate valproic acid-induced autism-like features. Journal of Physiology and Biochemistry. 2017; 73: 187–198.

[76] Brigandi SA, Shao H, Qian SY, Shen Y, Wu BL, Kang JX. Autistic children exhibit decreased levels of essential Fatty acids in red blood cells. International Journal of Molecular Sciences. 2015; 16: 10061–10076.

[77] Campolo N, Issoglio FM, Estrin DA, Bartesaghi S, Radi R. 3-Nitrotyrosine and related derivatives in proteins: precursors, radical intermediates and impact in function. Essays in Biochemistry. 2020; 64: 111–133.

[78] Mihm MJ, Schanbacher BL, Wallace BL, Wallace LJ, Uretsky NJ, Bauer JA. Free 3-nitrotyrosine causes striatal neurodegeneration in vivo. The Journal of Neuroscience: the Official Journal of the Society for Neuroscience. 2001; 21: RC149.

[79] Blanchard-Fillion B, Prou D, Polydoro M, Spielberg D, Tsika E, Wang Z, et al. Metabolism of 3-nitrotyrosine induces apoptotic death in dopaminergic cells. The Journal of Neuroscience: the Official Journal of the Society for Neuroscience. 2006; 26: 6124–6130.

[80] Ran Y, Yan B, Li Z, Ding Y, Shi Y, Le G. Dityrosine administration induces novel object recognition deficits in young adulthood mice. Physiology & Behavior. 2016; 164: 292–299.

[81] Okazaki S, Kimura R, Otsuka I, Funabiki Y, Murai T, Hishimoto A. Epigenetic clock analysis and increased plasminogen activator inhibitor-1 in high-functioning autism spectrum disorder. PloS One. 2022; 17: e0263478.

[82] Barnes RP, Fouquerel E, Opresko PL. The impact of oxidative DNA damage and stress on telomere homeostasis. Mechanisms of Ageing and Development. 2019; 177: 37–45.

[83] Mueller FS, Amport R, Notter T, Schalbetter SM, Lin HY, Garajova Z, et al. Deficient DNA base-excision repair in the forebrain leads to a sex-specific anxiety-like phenotype in mice. BMC Biology. 2022; 20: 170.

[84] Hussein MH, Alameen AA, Ansari MA, AlSharari SD, Ahmad SF, Attia MSM, et al. Semaglutide ameliorated autism-like behaviors and DNA repair efficiency in male BTBR mice by recovering DNA repair gene expression. Progress in Neuro-psychopharmacology & Biological Psychiatry. 2024; 135: 111091.