Integrated proteomic and metabolomic profiling reveals sex‐stratified biomarkers predicting chronicity in paediatric primary immune thrombocytopenia Authors: Yuanyuan Xue, Rongrong Zhang, Dongmei Liang, Yupeng Xu, Zhaofang Tian, Xiaohong Dai, Wei Xu, Qingqing Cao, Haiyan Zhu, Yun Wang, Yufang Yuan Published: 2026 ABSTRACT SummaryThis study aimed to develop prediction models for chronic immune thrombocytopenia (ITP) using a sex‐stratified proteomic and metabolomic approach, providing a framework for individualized prognosis evaluation and timely clinical management. This investigation was designed as a non‐interventional, prospective observational cohort. Plasma samples were collected from 67 children initially diagnosed with ITP along with 40 healthy controls. After a minimum of 1 year of regular follow‐up, participants were classified according to sex and disease progression. The male and female cohorts each comprised individuals with chronic ITP, non‐chronic ITP and healthy controls. From each subgroup, three peripheral blood samples were randomly chosen for proteomic and metabolomic profiling. Integrative omics were analysed for correlations using Pearson's coefficient (threshold: |r| > 0.8, p < 0.05). Predictive models were constructed using sex‐specific biomarkers associated with chronic progression. The analysis identified intercellular adhesion molecule‐1 (ICAM‐1) and biopterin in males and actin, alpha 2, smooth muscle, aorta (ACTA‐2) and N6‐acetyl‐L‐lysine in females as associated factors. Cross‐sex applications of these biomarkers revealed limited predictive value. Furthermore, the receiver operating characteristics curves, calibration curves and clinical decision curve analysis demonstrated good predictive efficacy of these predictive models. The underlying mechanisms of interaction between these biomarkers, sex differences and ITP chronicity progression warrant further investigation.Keywords: children, chronicity, immune thrombocytopenia, metabolomic, plasma biomarkers, proteomic, sex INTRODUCTION Immune thrombocytopenia (ITP) is defined as an acquired autoimmune haemorrhagic disorder, presented as isolated thrombocytopenia after secondary causes are excluded. 1 , 2 Although the majority of paediatric cases demonstrate a benign and self‐limiting trajectory, approximately 10%–20% of patients ultimately develop chronic ITP (CITP). 3 , 4 Children with CITP are subjected to not only a recurrent risk of bleeding but also treatment‐related adverse effects, compromising their psychological well‐being, quality of life and overall development. In addition, the prolonged disease course and ongoing therapeutic demands impose sustained pressure on families and increase healthcare resource utilization. 5 , 6 Consequently, reliable prediction of disease progression at initial diagnosis could support timely, individualized interventions aimed at reducing the adverse consequences for both children and their families. The prognostic role of sex in paediatric ITP remains controversial. While some studies reported female sex as a risk factor for chronicity, 7 , 8 including one that identified an initial diagnosis age >10 years and a platelet count >20 × 109/L as risk factors among girls, 9 others found no association. 10 Of particular relevance, sex‐stratification approaches have been utilized in related research. 11 , 12 A large‐scale retrospective study suggested that female CITP patients exhibit a higher risk of immunopathological manifestations and an increased likelihood of progression to systemic lupus erythematosus, 11 providing preliminary clinical evidence to support sex‐stratified analysis. From an immunopathological perspective, sex‐specific molecular mechanisms in autoimmune diseases are mainly reflected in the effects of sex hormones, X‐chromosome dosage and sex‐biased autosomal gene expression. 13 Sex hormones contribute to disease pathogenesis through regulating lymphocyte activity, giving rise to sexually dimorphic immune responses. Females generally present stronger humoral immunity (e.g. increased plasma cells) and heightened pro‐inflammatory activity (e.g. enhanced lymphocyte proliferation, activation and cytokine secretion), whereas males are characterized by expanded regulatory immune cell subsets. 14 Oestrogen has been shown to induce platelet autoantibody production and complement activation, impairing immune tolerance. 15 Although direct evidence for these mechanisms in paediatric ITP remains limited, they provide a plausible biological hypothesis for conducting sex‐stratified analyses in ITP research. Proteomics relies on data‐independent acquisition technology for the systematic analysis of protein expression, modifications and interaction networks in biological samples, thereby revealing key regulatory protein changes associated with biological functions. 16 Meanwhile, metabolomics employs liquid chromatography–mass spectrometry to comprehensively qualify and quantify small molecule metabolites within organisms. This approach has become a vital tool for identifying disease‐associated metabolites and dysregulated metabolic pathways. 17 Considerable advances have been achieved in the application of omics technologies to ITP research. Proteomic strategies were previously employed by the research team to predict responses to glucocorticoid treatment. 18 Concurrently, plasma metabolomics has shown effectiveness in forecasting glucocorticoid resistance in chronic ITP patients, serving as an auxiliary diagnostic tool and in assessing prognosis. 19 No investigations to date have integrated proteomics with metabolomics for the identification of chronicity‐associated biomarkers in paediatric ITP, a gap that provides a novel framework for ongoing research. In this study, a plasma integrated proteomic and metabolomic analytical strategy incorporating sex stratification was applied to identify early‐warning biomarkers predictive of chronic progression in paediatric ITP and to construct a prognostic prediction model. MATERIALS AND METHODS Study cohort and grouping This investigation was designed as a non‐interventional, prospective observational cohort study. A total of 67 children with an initial diagnosis of ITP who were hospitalized in the Department of Pediatrics at The Affiliated Huaian No.1 People's Hospital of Nanjing Medical University, together with 40 healthy volunteers recruited concurrently from the outpatient clinic with normal peripheral blood counts, were enrolled between 1 June 2023 and 1 June 2024. The inclusion and exclusion criteria and treatment protocol are provided in Supporting Information. CITP was defined as disease duration longer than 12 months, whereas non‐chronic (NC) ITP comprised newly diagnosed cases (<3 months) and persistent cases (3–12 months). 20 After a minimum of 1 year of regular follow‐up, participants were classified according to sex and disease progression. Group 1 was composed of male participants, including CITP (C‐1), NC ITP (NC‐1) and healthy control (HC‐1) subgroups. Group 2 consisted of female participants, also comprising CITP (C‐2), NC ITP (NC‐2) and healthy control (HC‐2) subgroups. The protocol was executed as per the ethical principles of the Declaration of Helsinki. The studies involving human participants were reviewed and approved by the ethics committee of the Affiliated Huaian No.1 People's Hospital of Nanjing Medical University. Written informed consent for participation in this study was provided by the participants' legal guardians/next of kin. Omics profiling and validation Peripheral blood samples were collected and processed for integrated proteomic and metabolomic profiling. Differential proteins and metabolites were analysed using bioinformatics approaches and validated by enzyme‐linked immunosorbent assay (ELISA). Detailed methodologies pertaining to sample preparation, instrumental parameters, chromatographic conditions and analytical procedures are provided in Supporting Information, Methods. Statistical analysis Statistical analyses were executed utilizing GraphPad Prism 9.1 (San Diego, CA, USA), SPSS 26.0 (IBM, Armonk, NY, USA) and R software (version 4.3.1; R Core Team, 2023). Continuous variables were denoted as mean ± standard deviation (x¯ ± s) when assessed with the t‐test or as median (interquartile range) [M (Q 1, Q 3)] when evaluated by the Mann–Whitney U‐test. Comparisons among multiple groups were analysed by two‐way analysis of variance (ANOVA) and the Tukey multiple comparisons tests. Categorical variables were denoted as frequencies with corresponding proportions and compared by chi‐square tests. Independent risk factors were determined through the logistic regression analysis. Receiver operating characteristic (ROC) curves were constructed. Calibration plots and clinical decision curve analysis (DCA) were applied to evaluate model performance. Statistical significance was attributed to p‐value <0.05. RESULTS Patient characteristics This study enrolled 67 ITP patients, comprising 30 chronic ITP cases (16 males, 14 females) and 37 non‐chronic ITP cases (19 males, 18 females). Additionally, 40 healthy controls (20 males, 20 females) were recruited for comparative purposes. Moreover, no statistically significant differences were identified in either sex or age between ITP patients and healthy controls (Table S1, p > 0.05). Furthermore, subgroup analyses demonstrated that chronic and non‐chronic ITP groups did not differ markedly with respect to any assessed parameters, encompassing age, weight, height and platelet count, even following sex‐stratified evaluation (Table 1, p > 0.05). TABLE 1Clinical profiles of immune thrombocytopenia (ITP) patients.VariablesChronic group (n = 30)Non‐chronic group (n = 37) Z/t/χ 2 p Sex (n, %) χ 2 = 0.0260.872Male16 (53.33%)19 (51.35%)Female14 (46.67%)18 (48.65%)Age (years, M (Q 1, Q 3))Total6.00 (4.00, 8.00)3.00 (2.00, 9.00) Z = −1.6510.099Male5.00 (3.00, 8.25)3.00 (2.00, 8.00) Z = −0.8320.406Female6.50 (4.25, 8.00)4.00 (2.00, 8.50) Z = −1.4480.148Weight (kg, M (Q 1, Q 3))Total29.500 (20.250, 38.500)21.000 (11.500, 37.000) Z = −1.6780.093Male31.500 (20.750, 37.750)24.000 (11.250, 43.500) Z = −0.5960.551Female26.500 (17.025, 38.250)18.250 (12.500, 27.875) Z = −1.6350.102Height (cm, x¯±s)Total123.100 ± 25.512111.851 ± 33.218 t = 1.5250.132Male120.938 ± 22.558112.711 ± 37.191 t = 0.8040.427Female125.571 ± 29.196110.944 ± 29.505 t = 1.3980.173Platelet count (×109/L, M (Q 1, Q 3))Total13.50 (10.00, 21.75)18.00 (8.00, 32.00) Z = −0.7190.472Male13.50 (10.75, 18.75)18.00 (10.00, 28.00) Z = −0.7790.436Female13.50 (8.50, 28.00)18.00 (5.75, 35.00) Z = −0.3610.718 Clinical profiles of immune thrombocytopenia (ITP) patients. Ultra deep proteomics profiling results In total, 4298 quantifiable proteins were identified (Figure S1A). Principal component analysis (PCA) demonstrated distinctive clustering patterns among specimens from the C‐1, C‐2 and NC‐1 groups as well as between the C‐2, NC‐1 and NC‐2 groups. Strong intragroup aggregation was observed within each cohort, whereas intergroup separation indicated marked differences across groups (Figure 1A,B). Differentially expressed proteins (DEPs) were screened using thresholds of protein expression fold change (FC) ≥1.5 or ≤0.67 combined with p‐value <0.05. A sum of 458 DEPs emerged in males and 55 in females (Figure 1C). After overlap adjustment, three non‐sex‐specific DEPs and 507 sex‐specific DEPs were linked to ITP chronicity, with the latter serving as the primary focus of screening (Figure 1D). Furthermore, expression differences between chronic and non‐chronic ITP groups within each sex were visualized by heat maps, and volcano plots illustrated protein alterations between the respective groups (Figure 1E–H). Heat map analyses of DEPs demonstrated clear distinctions in protein expression patterns between male and female cohorts with the same disease course, while volcano plots displayed the corresponding differential protein distributions (Figure S1B–E). Similar clustering distributions were also observed in PCA comparisons of ITP patients and healthy controls within each sex group (Figure S2A,B). Notably, substantial differences in protein expression profiles were detected between ITP groups and their corresponding healthy controls (Figure S2C,D). Sex‐stratified analysis further revealed 16 proteins in males and 13 in females that were consistently differentially expressed in both ITP versus healthy control and chronic versus non‐chronic comparisons (Figure S2E,F). Plasma proteomic profiles of the chronic (C) versus non‐chronic (NC) groups within each sex. (A, B) Principal component analysis (PCA): PC1 signifies the first principal component, PC2 signifies the second principal component and PC3 signifies the third principal component; (A) C‐1 versus C‐2 versus NC‐1; (B) C‐2 versus NC‐1 versus NC‐2. (C) Outcomes of differential proteins in NC‐1 versus C‐1 and NC‐2 versus C‐2 groups. (D) Venn diagram depicting the differential protein expression between males and females. (E, F) Clustering heat map, in which red signifies upregulation and blue signifies downregulation, with colour intensity indicating varying degrees of up‐ and downregulation; (E) NC‐1 versus C‐1; (F) NC‐2 versus C‐2. (G, H) Volcano plot of the plasma differential proteins in NC‐1 versus C‐1 (G) and NC‐2 versus C‐2 groups (H), in which the horizontal coordinate signifies the log2(FC) and the vertical coordinate signifies ‐log10(p‐value). The red and green scatter dots signify increased and decreased differential proteins respectively. Metabolomics profiling results PCA demonstrated clear clustering patterns among samples from the C‐1, C‐2 and NC‐1 groups as well as between the C‐2, NC‐1 and NC‐2 groups. Strong intragroup aggregation was observed within each cohort, whereas intergroup separation indicated notable differences across groups (Figure 2A,B). Differentially expressed metabolites (DEMs) were screened under thresholds of FC ≥1.5 or ≤0.67 combined with p‐value <0.05. In ITP patients, 56 DEMs were identified in males and 40 in females (Figure 2C). After overlap adjustment, five non‐sex‐specific DEMs and 86 sex‐specific DEMs were linked to ITP chronicity, with sex‐specific metabolites serving as the main focus of screening (Figure 2D). Furthermore, metabolite expression differences between chronic and non‐chronic ITP groups within each sex were visualized by heat maps, whereas volcano plots depicted expression variations between the respective groups (Figure 2E–H). Comparable clustering distributions were also observed in PCA comparisons of ITP patients and healthy controls within each sex group (Figure S3A,B). Marked differences in metabolite expression profiles were detected between ITP groups and their matched healthy controls (Figure S3C,D). Sex‐stratified analysis further identified four male‐specific and three female‐specific metabolites that were differentially expressed in both ITP versus healthy control and chronic versus non‐chronic comparisons (Figure S3E,F). Heat map analyses of DEMs revealed clear distinctions between male and female groups with the same disease course, while volcano plots illustrated the corresponding metabolite distributions (Figure S4A–D). Radar charts displayed the top 10 metabolites with the most prominent changes across group comparisons (Figure S5A–D). Plasma metabolomic profiles of chronic (C) versus non‐chronic (NC) groups within each sex. (A, B) Principal component analysis (PCA): PC1 signifies the first principal component, PC2 signifies the second principal component and PC3 signifies the third principal component; (A) C‐1 versus C‐2 versus NC‐1; (B) C‐2 versus NC‐1 versus NC‐2. (C) Findings of differential metabolites in NC‐1 versus C‐1 and NC‐2 versus C‐2 groups. (D) Venn diagram depicting the differential metabolites expression between males and females. (E, F) Clustering heat map, with red indicating upregulation and green indicating downregulation, and the colour intensity indicating varying degrees of up‐ and downregulation; (E) NC‐1 versus C‐1; (F) NC‐2 versus C‐2. (G, H) Volcano plot of the plasma differential metabolites in NC‐1 versus C‐1 (G) and NC‐2 versus C‐2 groups (H), with the horizontal axis representing the log2(FC) and the vertical axis representing ‐log10(p‐value). The red and green scatter points signify elevated and diminished differential metabolites respectively. Integrative omics profiling By applying integrative analysis through Venn diagrams, biomarkers exhibiting differential expression in both ITP versus healthy control and chronic versus non‐chronic comparisons within each sex group were identified, while demonstrating a sex‐specific differential expression model. Candidate DEMs and DEPs were subsequently analysed for correlations using Pearson's coefficient (threshold: |r| > 0.8, p < 0.05) (Figure S6). The sex‐stratified biomarkers predictive of ITP chronicity were as follows: in males, intercellular adhesion molecule‐1 (ICAM‐1) (DEP) and biopterin (DEM); in females, actin, alpha 2, smooth muscle, aorta (ACTA‐2) (DEP) and N6‐acetyl‐L‐lysine (DEM). Gene Ontology (GO) classification and the Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis of DEPs and DEMs GO enrichment analysis demonstrated notable similarities between males and females. Within biological processes, DEPs were predominantly enriched in cellular process, metabolic process and biological regulation. For cellular components, DEPs were mainly localized to cellular anatomical entities and protein‐containing complex. The molecular functions primarily included binding, catalytic activity and molecular function regulatory activity (Figure S7A,B). KEGG pathway analysis further revealed sex‐specific enrichment profiles. In males, DEPs were largely associated with motor proteins, whereas in females, significant enrichment was observed in the neurotrophin signalling pathway and vascular smooth muscle contraction (Figure S7C,D). For metabolites, male‐associated DEMs were mainly enriched in porphyrin metabolism and steroid hormone biosynthesis, while female‐associated DEMs were enriched in histidine metabolism and arginine biosynthesis (Figure S7E,F). ELISA results The candidate DEPs and DEMs were verified by ELISA in the validation cohort (Figure 3). Two‐way ANOVA revealed that in males, the C group showed decreased ICAM‐1 concentrations (p < 0.05) but increased biopterin levels (p < 0.01) when contrasted with the NC group; no significant differences were observed for ACTA‐2 and N6‐acetyl‐L‐lysine (all p > 0.05). In contrast, among females, the C group showed elevated ACTA‐2 levels (p < 0.01) and decreased N6‐acetyl‐L‐lysine levels (p < 0.0001) compared to the NC group, while ICAM‐1 and biopterin showed no notable differences (all p > 0.05). Results of enzyme‐linked immunosorbent assay (ELISA) validation for the biomarkers. Plasma levels of intercellular adhesion molecule‐1 (ICAM‐1) (A), biopterin (B), actin, alpha 2, smooth muscle, aorta (ACTA‐2) (C) and N6‐acetyl‐l‐lysine (D) are presented across the C, NC and HC groups in males and females respectively. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. ITP, immune thrombocytopenia; C, chronic ITP group; NC, non‐chronic ITP group; HC, healthy control group. Efficacy of the model Multifactorial logistic regression analyses were conducted for all biomarkers in both male (Table S2) and female cohorts (Table S3). ICAM‐1 and N6‐acetyl‐L‐lysine were identified as statistically significant factors associated with chronicity in males and females respectively (all p < 0.05). The findings indicated that reduced levels of ICAM‐1 and N6‐acetyl‐L‐lysine were linked to an elevated risk of chronicity. Moreover, sex‐specific biomarkers showed no statistically significant predictive value in opposite‐sex groups (all p > 0.05). ROC curve analyses compared the predictive capacity of biomarkers for chronicity progression in males and females (Figure 4A,B). Sex‐stratified ROC analysis demonstrated distinct predictive performance. In males, the combination of ICAM‐1 and biopterin markedly enhanced predictive accuracy, yielding an area under the curve (AUC) of 0.875 (95% confidence interval [CI]: 0.762–0.988) (Table S4). In females, the combination of ACTA‐2 and N6‐acetyl‐L‐lysine produced an excellent predictive performance, with an AUC of 0.909 (95% CI: 0.811–1.000) (Table S5). Efficacy of the prediction model across sex subgroups. (A, B) Receiver operating characteristic (ROC) curves of predictive indicators in male (A) and female (B) cohorts. (C, D) Calibration curves for the prediction model in males (C) and females (D). (E, F) Decision curve analysis (DCA) curves for the prediction model in males (E) and females (F). Calibration and DCA curves were further applied to evaluate sex‐specific prediction models (Figure 4C–F). The Hosmer–Lemeshow goodness‐of‐fit test confirmed adequate calibration for both sexes (males: χ 2 = 11.476, p = 0.119; females: χ 2 = 5.872, p = 0.662; all p > 0.05), indicating no significant lack of fit. Additionally, clinical DCA demonstrated higher net benefits across threshold probabilities, verifying the strong performance of these sex‐stratified prediction models. DISCUSSION ITP is characterized by considerable biological heterogeneity, with pronounced interindividual variability in disease progression and therapeutic responses, promoting a shift towards individualized clinical management. 21 Age has been identified as a key predictor of CITP in children. 11 , 22 , 23 Schmidt et al. developed a biological risk score incorporating haemoglobin, platelet count, FcγRIIc genotype, immunoglobulin G anti‐platelet antibodies and preceding vaccination, which effectively stratified disease courses. 24 Notably, in our cohort, neither age nor platelet count emerged as significant predictors, which may be attributed to differences in sample size or population characteristics. A sex‐stratified analysis based on targeted biomarkers was innovatively applied to construct personalized prognostic models for disease progression. The identified chronicity‐associated biomarkers included ICAM‐1 and biopterin in males as well as ACTA‐2 and N6‐Acetyl‐L‐Lysine in females. ICAM‐1, an immunoglobulin superfamily membrane protein, exists primarily on endothelial cells and leucocytes, functioning in lymphocyte stimulation, and immune synapse development, maintaining immune balance. During inflammatory responses, ICAM‐1 expression exhibits marked elevation through proinflammatory cytokines and demonstrates involvement in the development of autoimmune diseases like multiple sclerosis and inflammatory bowel disease. 25 Notably, consistent with Garabet et al., 26 this study also observed elevated ICAM‐1 levels in ITP patients compared to healthy controls, indicating a consistent finding across both adult and paediatric populations. However, in male paediatric patients, a low baseline ICAM‐1 level at initial diagnosis was associated with an increased risk of chronic ITP (p < 0.05). In contrast, Li et al. 27 reported that, in adult ITP patients, high ICAM‐1 levels were associated with glucocorticoid resistance and poor response to decitabine therapy. This discrepancy suggests that the prognostic value of ICAM‐1 may differ between children and adults. The upregulation of ICAM‐1 is mediated through the nuclear factor kappa B (NF‐κB) and Janus kinase‐signal transducer and activator of transcription. 25 Recent research indicated that NF‐κB activity can be attenuated by androgens via the AR–USP18–TAK1–NF‐κB axis. Specifically, androgen receptor activation induces USP18 expression, which suppresses TAK1 phosphorylation and subsequently diminishes NF‐κB activity in anti‐tumour T cells, thereby exerting sex‐specific anti‐inflammatory effects. 28 Collectively, these findings provide a potential mechanistic framework for understanding the observed sex‐specific disparity in ICAM‐1 levels. However, the precise mechanism by which low ICAM‐1 contributes to chronicity in male children warrants further investigation. Lysine acetylation, a fundamental post‐translational modification, results in the production of N6‐acetyl‐L‐lysine. 29 This modification modulates platelet physiology by covalently altering proteins engaged in cytoskeletal regulation, energy metabolism, endoplasmic reticulum homeostasis and cellular contractility. 30 Lysine deacetylase inhibitors employed in the treatment of haematological malignancies were shown to induce thrombocytopenia through interference with megakaryocyte differentiation, as demonstrated in human CD34+ haematopoietic progenitor models. 31 This study confirmed the clinical relevance of N6‐acetyl‐L‐lysine, where diminished baseline levels were linked to a higher likelihood of chronicity in female ITP patients. This indicator achieved an AUC of 0.861 (95% CI: 0.726–0.996), with excellent discriminatory power observed in females. Nitric oxide, an essential signalling mediator in vascular homeostasis, is generated by endothelial nitric oxide synthase. Tetrahydrobiopterin (BH4) is a critical cofactor required for endothelial nitric oxide synthase activity. 32 Evidence suggested that testosterone‐driven immunomodulation characterized by enhanced tolerogenic CD11c+ dendritic cell function, inhibition of conventional CD4+ T‐cell clonal expansion and promotion of CD4+Foxp3+ regulatory T‐cell populations. 33 Collectively, these mechanisms attenuate pro‐inflammatory T‐cell activity while facilitating anti‐inflammatory cytokine production, thereby potentially contributing to ITP remission. The synthesis of BH4 is influenced by inflammatory mediators, and biopterin, as an oxidized derivative of BH4, is considered a marker reflecting systemic BH4 availability. 34 , 35 This study demonstrated that biopterin did not independently predict chronicity progression in males (p = 0.070). ROC curve assessment revealed an AUC of 0.783 (95% CI: 0.628–0.938), and when combined with ICAM‐1 in a combined model, the AUC reached 0.875 (95% CI: 0.762–0.988), with significant predictive accuracy. The ACTA‐2 encodes α‐smooth muscle actin, which is primarily localized within microfilament bundles of smooth muscle cells, facilitating contractile activity. 36 In chronic kidney disease, an inflammatory disorder, ACTA‐2 has been recently suggested as a diagnostic biomarker. 37 Although ACTA‐2 did not independently predict chronicity in females (p = 0.126), ROC curve analysis yielded AUC values of 0.720 (95% CI: 0.528–0.912), implying that ACTA‐2 may represent a potential female‐specific indicator of chronicity. This investigation carries several limitations that warrant acknowledgement. First, being a single‐centre study with a relatively small cohort, the non‐significant associations observed for biopterin in males and ACTA‐2 in females may have resulted from limited statistical power. Second, the absence of a non‐ITP thrombocytopenia control group and an independent validation cohort constrain the generalizability of these findings. Consequently, it is essential to validate all candidate biomarkers in larger, multi‐ethnic paediatric populations through studies that integrate them with clinical variables to develop more accurate predictive models. Third, it should be highlighted that no differentially screened metabolites appeared in the radar plots across multiple group comparisons, as their fold change values were markedly lower than those of the top‐ranking metabolites. Furthermore, integrative omics‐based correlation analysis identified a direct relationship between ICAM‐1 and Phe‐Ala‐Leu in male patients. However, Phe‐Ala‐Leu was excluded from subsequent validation because of its poor stability. Finally, future investigations should adopt longitudinal study designs to monitor serial changes in biomarkers within the same cohort from diagnosis to the chronic phase, thereby elucidating their potential mechanistic contributions to ITP progression. CONCLUSION This prospective observational cohort study revealed that ICAM‐1 served as an independent predictor of chronicity in males, whereas the combined assessment of ICAM‐1 and biopterin yielded superior predictive performance. In females, N6‐acetyl‐L‐lysine displayed independent prognostic value, which was further strengthened when evaluated in conjunction with ACTA‐2. AUTHOR CONTRIBUTIONS Dongmei Liang: Investigation; methodology; software. Zhaofang Tian: Investigation; methodology; writing – review and editing. Yupeng Xu: Investigation; methodology; software. Rongrong Zhang: Conceptualization; methodology; writing – review and editing. Yun Wang: Investigation; methodology; writing – review and editing. Yuanyuan Xue: Data curation; validation; software; writing – original draft. Yufang Yuan: Conceptualization; funding acquisition; methodology; supervision; writing – review and editing. Wei Xu: Investigation; methodology; writing – review and editing. Haiyan Zhu: Investigation; methodology; writing – review and editing. Xiaohong Dai: Investigation; methodology; writing – review and editing. Qingqing Cao: Investigation; methodology; writing – review and editing. FUNDING INFORMATION This work was supported by The Affiliated Huaian No.1 People's Hospital of Nanjing Medical University ‘Research and Innovation Team’ project (YCT202301). CONFLICT OF INTEREST STATEMENT All authors have no conflict of interest to disclose. ETHICS STATEMENT The protocol was executed as per the ethical principles of the Declaration of Helsinki. The studies involving human participants were reviewed and approved by the ethics committee of the Affiliated Huaian No.1 People's Hospital of Nanjing Medical University (Approval No. KY‐2023‐132‐01). PATIENT CONSENT STATEMENT Written informed consent for participation in this study was provided by the participants' legal guardians/next of kin. Supporting information Figure S1. Figure S2. Figure S3. Figure S4. Figure S5. Figure S6. Figure S7. Tables S1–S5. Appendix S1. ACKNOWLEDGEMENTS This investigation was carried out at the Affiliated Huaian No.1 People's Hospital of Nanjing Medical University. The authors extend their deepest appreciation to all staff members and study participants for their valuable contributions. Xue Y, Zhang R, Liang D, Xu Y, Tian Z, Dai X, et al. Integrated proteomic and metabolomic profiling reveals sex‐stratified biomarkers predicting chronicity in paediatric primary immune thrombocytopenia. Br J Haematol. 2026;208(6):1954–1964. 10.1111/bjh.70521 [Correction added on 15 June 2026, after first online publication: The subcategory has been changed.] DATA AVAILABILITY STATEMENT The original contributions presented in the study are publicly available. The proteomics data have been deposited in the ProteomeXchange database via the PRIDE repository with the accession number PXD067864, and the metabolomics data are available in the MetaboLights database under the accession number MTBLS12920. REFERENCES Associated Data Supplementary Materials Figure S1. Figure S2. Figure S3. Figure S4. Figure S5. Figure S6. Figure S7. Tables S1–S5. Appendix S1. Data Availability Statement The original contributions presented in the study are publicly available. The proteomics data have been deposited in the ProteomeXchange database via the PRIDE repository with the accession number PXD067864, and the metabolomics data are available in the MetaboLights database under the accession number MTBLS12920.