Immune-Metabolic Programs Drive Disease Trajectories in Paediatric Long COVID

Immune-Metabolic Programs Drive Disease Trajectories in Paediatric Long COVID | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Biological Sciences - Article Immune-Metabolic Programs Drive Disease Trajectories in Paediatric Long COVID Monika Brunner-Weinzierl, Katrin Vogel, Irina Han, Pauline Jakobs, and 13 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-7083240/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 04 May, 2026 Read the published version in Nature Communications → Version 1 posted You are reading this latest preprint version Abstract While most children and adolescents recover uneventfully from SARS-CoV-2 infection, some develop persistent symptoms known as paediatric long COVID (LC). Paediatric LC presents with substantial, multisystem health impairment lasting months to years after SARS-CoV-2 infection 1 , 2 . Despite its clinical burden, underpinnings of symptom persistence, heterogeneity, and recovery remain elusive 3 , 4 . Here, we demonstrate that severe symptoms in paediatric LC remained stable over two-to-three years, despite unremarkable cardiopulmonary and routine assessments, and were underpinned by temporally shifting immune-metabolic responses. The first year of LC was marked by viral-associated and Th2-like cytokine responses, transitioning into Th17-like and innate responses over time. Neurofilament light chain, an indicator of neuro-axonal injury, rose with LC-severity, but common autoantibodies remained unchanged. Epstein-Barr virus (EBV) exposure emerged as a key modifier linked to broader immune dysfunction, whereas anti-DFS70 autoantibodies correlated with milder haematological alterations. In EBV-naïve LC cases, symptoms became more severe with altered blood viscosity, but less severe with higher IL-12p40, vitamin B1, and basophils, implicating them as protective. The identified LC subgroups displayed metabolically distinct signatures, supporting the existence of biologically coherent endotypes. These findings uncover immune-metabolic axes linked to resilience and persistence in paediatric LC and may provide a basis for biomarker-informed diagnosis and precision intervention. Health sciences/Biomarkers/Predictive markers Health sciences/Diseases/Immunological disorders Health sciences/Medical research/Paediatric research Figures Figure 1 Figure 2 Figure 3 Figure 4 Introduction The clinical course of SARS-CoV-2 infection in children is typically benign, with many remaining asymptomatic, reflecting robust counter-regulatory immune mechanisms 5 , 6 . Nevertheless, approximately 1–3% of paediatric cases develop long COVID (LC), or post-COVID condition, a debilitating syndrome persisting for at least three months post-infection without an alternative explanation 2 , 7 . Symptoms span over 200 domains, including fatigue, dyspnoea, pain, cognitive impairment, and post-exertional malaise 1 , 3 , 8 , 9 . In only 80% of affected children, these persistent and debilitating symptoms resolve within one year 10 . Autoimmune diseases were found to occur with greater frequency 10 . Risk factors for LC were female sex, >12 years, being infected with original variants, and comorbidities. So far, paediatric LC has no pathognomonic findings or diagnostic tests available. Given that immune differences between children and adults become more similar after early childhood and cumulative exposures in adults complicate interpretation, paediatric LC offers a unique insight into core disease mechanisms 11 – 15 . So far, adult LC has been associated with diverse pathologies, including neuroinflammation 16 , endotheliopathy 17 , complement activation 18 , 19 , reactivation of EBV 19 , metabolic dysfunction 20 , 21 , persistent inflammation and autoimmunity 22 , 23 , whilst an association with differential SARS-CoV-2 antibody responses is unlikely 24 . These diverse manifestations suggest the existence of distinct LC endotypes rather than a single pathway. Whilst the underlying pathophysiological processes driving LC in children remain less well understood 4 , 5 , the initial evidence appears to suggest possible abnormalities in immune responses, cellular metabolism and endothelial function 2 . Endotype-specific understanding is essential to enable precision interventions and improve clinical trial outcomes 25 . Here, we hypothesize that paediatric LC comprises distinct immuno-metabolic endotypes that emerge from dysregulated immune responses and disrupted homeostasis. We aim to identify protective and pathogenic mechanisms that drive disease persistence and define biologically distinct subgroups that can inform targeted clinical trials. Results The study analysed 74 children with active long COVID (LC) symptoms following SARS-CoV-2 infection, up to 166 weeks post primary SARS-CoV-2 infection, alongside 27 controls (Control) (Fig. 1; Extended Data Table 1a,b). The control group (n = 27) included both healthy (n = 14) and well-managed children with cystic fibrosis (CF)(n = 13), as they are frequently exposed to respiratory infections despite stable treatment and similarly affected by SARS-CoV-2 infection 26 . This ensures that the immune changes observed in long COVID are not driven by pre-existing inflammation or infection frequency, but rather reflect LC disease-specific mechanisms. As no significant proinflammatory cytokines IL-6 (Mann-Whitney U-test p = 0.662; mean 20.6/21.4; SD 5.3/5.4) or TNFα (p = 0.923; mean 97/144; SD 61/57) differences were found between both control subgroups, they are considered a single control group. A follow-up assessment (LC visit 2) of LC was performed 3-6 months later. Symptom severity was assessed at each appointment using the following questionnaires: SF-12 Physical Component Summary (PCS), SF-12 Mental Component Summary (MCS), the acute Bell Disability Scale, the KIDSCREEN questionnaire, the Generalized Anxiety Disorder 7-item scale (GAD-7), Patient Health Questionnaire 9-item (PHQ-9), Short Fatigue Severity Scale (FSS), Parent-Reported Symptoms (PRS), Patient-Reported Symptoms (PtRS) and Children’s Sleep Habits Questionnaire (CSHQ); strength was assessed via a hand-grip-test and stamina was assessed via a sit-to-stand test for children (Fig. 1, Extended Data Fig.1) 27 . As indicators of viral response and inflammation, serum concentrations of IFNα were measured, revealing elevated IFNα levels in LC patients during the first year of disease compared to controls, while complement system components C3, C4, and CH50 remained within the normal range (Fig.1d, data not shown). This result contrasts with a study of adult LC 19 , and may be explained by the fact that our study looks at later time points of the disease and that paediatric patients tend to have less pre-exposure pathogens and antigens, resulting in less antigenic sins and innate memory 13,14,28 . Patients within 1 yr active LC showed no clinical improvement in both physical and mental symptoms (Fig. 1c,e, Details of LMM analysis: Extended Data Fig. 1a,b; Extended Data Table 1b). In addition, no improvement at the group level of LC was detected up to 3 years with individual cases showing slight clinical improvement or decline (Fig. 1e, red line, Extended Data Fig. 1a,b; Extended Data Table 1b). This indicates that different pathological mechanisms contribute to disease severity and progression. Since LC patients often experience fatigue and shortness of breath, cardiac and pulmonary involvement were assessed (Fig.1c, Fig.2a,b) 9,29 . To address cardiac involvement, all paediatric LC participants (n = 74) underwent a 12-lead electrocardiogram (ECG) and transthoracic echocardiography upon presentation. Electrocardiographic analysis revealed supraventricular extrasystoles in one case. Echocardiography identified abnormalities in three patients (4.1%): one case of mild mitral regurgitation; one case of a bicuspid aortic valve with a coronary artery fistula; and one case of mild diastolic dysfunction, which resolved at follow-up. Paediatric cardiologists judged these to be incidental. To address pulmonary involvement, we initially clustered these individuals based on the mean of their FEV1 Z-scores across multiple visits to capture differences in lung function (Fig. 2b). FEV1 Z-scores were selected because they provide a standardized measure of lung function relative to population norms, enabling the detection of impairments in respiratory capacity. We used k-means clustering to identify three distinct groups of paediatric LC patients with significantly different levels of lung function capacity (Fig. 2b). However, clusters were not associated with anti-SARS-CoV-2 IgG levels or clinical disease scores, suggesting that heterogeneous immunological mechanisms may underlie active paediatric LC. To explore potential immunological drivers of differences in lung function, we selected biomarkers (IL-13, IL-33) based on their known roles in lung-related immune responses and their critical involvement in SARS-CoV-2 immunity 30 (Fig. 2c). We also included IL-6, a key pro-inflammatory cytokine, due to its role in both acute and chronic inflammatory processes. While IL-6 levels were not elevated systemically during the first year of LC, significant increases of IL-13 and IL-33 were observed (Bonferroni-Holm corrected p < 0.05). We then applied a linear mixed-effects model (LMM), incorporating IL-6, IL-13, IL-33, time since infection, and percentile BMI to assess their combined influence on lung function (Fig. 2d; Extended Data Table 2.1a). Due to collinearity between IL-13 and IL-33 (r > 0.7), only IL-13 was included in the final model. The final model fit yielded R² Marg = 0.162, indicating that the fixed effects (IL-6, IL-13) explained approximately 16% of the variance in FEV1 Z-scores. The R² Cond of 0.72 reflected the additional variance accounted for when random effects were included, highlighting the substantial influence of individual variability. Neither time since infection, percentile BMI nor sex improved the model fit (Fig. 2d; Extended Data Table 2.1b). Notably, our data reveal that better lung function (as reflected by higher FEV1 Z-scores) was associated with elevated systemic IL-13, suggesting a protective tissue repair response operating in parallel with pro-inflammatory mechanisms 30 . This may reflect IL-13's involvement in epithelial repair and immune responses following viral infections 31 . These findings underscore the importance of considering subclinical immunological alterations in paediatric LC, even when lung function remains within normal limits, as these changes may have long-term implications for respiratory health 32 . To investigate the systemic, immune-mediated nature of LC, circulating serum cytokines were assessed within the first year post SARS-CoV-2 infection (Fig.2e-h, Extended Data Fig.2, Extended Data Table 2.2 for multiple testing). The choice of a one-year observation period is based on evidence that long COVID shows spontaneous remission in the majority of cases within 52 weeks; this timeframe is commonly used by others, ensuring comparability, and aligns with the waning of SARS-CoV-2 immunity, such as antibody levels 24,29 . Further analysis of Th2-related cytokines, including IL-4, IL-5, IL-10, and IL-9, revealed no increase compared to controls (Extended Data Fig. 2a). To identify distinct paediatric LC subtypes based on cytokine profiles, we further examined Th1-like (Extended Data Fig. 2b), Th17/22-like (Extended Data Fig. 2d), innate-like and regulatory-like cytokines (Fig. 2d, Fig. 2f,g; Extended Data Fig. 2b, Extended Data Table 2.2). While Th1-related cytokines remained low in LC, this resulted in a significantly elevated Th2/Th1 ratio (Fig. 2e). Among the Th17/22-associated cytokines, IL-12p40 was significantly (corrected p = 0.01) increased, but IL-23 and IL-12p70 were not, whilst the innate-like cytokines IL-1α and IL-1β were also significantly elevated, suggesting ongoing innate-driven immune activation (Extended Data Fig. 2b). These findings highlight a complex immunological dysregulation in LC, characterised by distinct cytokine patterns within the first year of the disease. Autoantibodies (aAb) and Organ Injury in Children with Active Long COVID Under the assumption that long COVID manifests as distinct autoimmune-driven endotypes, we analysed disease-associated autoantibodies (aAb) and, separately, serum neurofilament light chain (NfL) as surrogate marker for central nervous system (CNS) involvement and neuro-axonal injury (Fig. 2i,j, Extended Data Fig. 3.2a). The aAb were surrogate markers for vasculitis, antiphospholipid syndrome, autoimmune connective tissue and coeliac disease. NfL concentrations were converted to age-adjusted z-scores with the Basel paediatric reference dataset and reported as percentiles (P) relative to healthy children 33 . A Wilcoxon signed-rank test against the reference median P 50 showed a significant upward shift (W= 1744, n= 73, p= 0.031; effect size r= 0.25). Clinically, only one child (1.4 %) exceeded the upper reference boundary (P 97.5) and 6/73 (8.2 %) crossed P 90. A Bell score of 40—denoted as severe disability with no more than 3-4 h of light activity per day and a > 50% loss of normal functional capacity—was used as the severe long COVID stratum (Fig. 2i) 33 . Children at or above this cut-off displayed higher NfL percentiles, clustering near the upper tail of the normative distribution (p= 0.003). Pearson´s (and Spearman) correlation showed a significant invers correlation r= -0.33, p= 0.012 (Extended Data Fig. 3.2a). Of note, NfL percentiles were unrelated to the interval since acute SARS-CoV-2 infection (p= 0.48; Extended Data Fig. 3.2a). Collectively, the data indicate ongoing neuro-axonal injury in affected children. In parallel, we assessed functional aAbs against G-protein-coupled receptors (GPCR-aAb) targeting β1- and β2-adrenergic as well as M3- and M4-muscarinic acetylcholine receptors, to capture potential neuro-autonomic involvement. These aAbs did not correlate with time after SARS-CoV-2 infection, systemic cytokine levels, or disease severity (Extended Data Table 2.1c). Furthermore, there was no increase in LC patients with a disease duration of less than 1 year compared to those with a disease duration of more than 1 year (Extended Data Fig. 3.2b). Next, we analysed aAb as surrogate markers for vasculitis (anti-proteinase 3 (PR3) and anti-myeloperoxidase (MPO) aAb), antiphospholipid syndrome (anti-cardiolipin aAb, anti-β2-glycoprotein I aAb (anti-β2GPI), autoimmune connective tissue disease (anti-cyclic citrullinated peptide (CCP) aAb) and coeliac disease (anti-transglutaminase (TransG) Fig.2j). None of them were elevated compared to controls and remained within the normal range during the first year. To evaluate whether enhanced autoimmunity is generally present in patients with active long COVID, we assessed the prevalence of aAb positivity across individuals (Fig.2k). Compared to the control group, no significant increase in aAb positivity was observed, indicating that active paediatric long COVID is not driven by aAb-mediated organ injury. To explore whether isolated anti-DFS70 reactivity might serve as a benign autoreactivity signature, we assayed anti-DFS70 antibodies despite the absence of a positive ANA screen (Fig. 2l). Surprisingly, none of the children in the control group tested positive. In contrast, 11% of patients with active LC tested positive, and those who were anti-DFS70-positive remained so at a second follow-up visit months later. Next, we grouped patients according to their anti-DFS70 status (Fig.2 m,n, Extended Data Fig.3.2 c,d, Extended Data Table 2.3 B-D for multiple testing,). Anti-DFS70-negative LC patients displayed a higher frequency and activity of vWF, and elevated factor VIII levels. Other factors, including antithrombin III, fibrinogen, D-dimer, aPTT, protein C, free protein S, and complement components C3 and C4 remained unchanged; notably, EBV-related antibodies were only observed in anti-DFS70-negative LC patients (Fig. 2n, Extended Data Table 2.3C, Extended Data Table 2.3C). Thus, we provide clear evidence of alterations to clotting factors in a subgroup of paediatric LC patients. Although anti-DFS70 antibodies are classically linked to healthy individuals and are not considered markers of systemic autoimmunity, their presence in this cohort may signify a unique, non-pathogenic immunological response 34 . Notably, Epstein-Barr virus seropositivity was confined to the anti-DFS70-negative subgroup, suggesting the existence of immunologically distinct Long-COVID phenotypes that merit further investigation. Active Paediatric Long COVID beyond the First Year Subsequently, we monitored systemic parameters after one year of LC to delineate the evolving immunopathological phases (Fig. 3a-e, Extended Data Fig. 3.1, Extended Data Table 3.1 for multiple comparisons). Of note, the previously elevated IFNα levels declined, indicating a waning antiviral response (Fig. 3a). Re-evaluation of the earlier Th2-skewed state (Fig. 3b, Extended Data Fig. 3.1a) revealed a marked reduction in IL-4, IL-13, and the type 2-alarmin IL-33, accompanied by a two-fold rise of IL-9, pointing towards a shift towards a Th9-like immunophenotype. Innate-like cytokines GM-CSF (p = 0.037) trended further upward, with a notable and significant increase in IL-1β (Fig. 3c). Within the Th17 axis (Fig. 3d, Extended Data Fig. 3.1d), IL-23 and IL-12p40 increased, while the Th1 axis remained unaltered (Extended Data Fig. 3.1d). IL-12p70 and IFNγ remained low, altogether suggesting a transition towards a Th17-driven state. Taken together (Fig. 3e), these findings outline a dynamic immunological trajectory in paediatric LC, marked by an early anti-viral and Th2-skewed response within the first year. Over time, this shifts towards a Th9/Th17-dominated, innate-oriented endotype, reflecting the progressive immunopathological remodelling that underlies disease chronicity. Next, we employed a linear mixed-effects model (random intercept Patient ID) to analyse factors associated with long COVID, using the Bell score as the dependent variable (Fig. 3f, Extended Data Fig. 3.2a). Covariates were selected based on their biological relevance to long COVID pathophysiology, including markers of viral (re)activation (anti-EBV EBNA IgG), haematological abnormalities in mean corpuscular haemoglobin concentration (MCHC) 35 , and IL-12p40, as it is the most frequently expressed cytokine and has the greatest increase in value since infection (Fig. 2h, 3d), along with basophil granulocyte counts. R² Marg (0.212) indicated that fixed effects explained 21% of the variance. All covariates showed significant associations with the Bell score. Notably, anti-EBV EBNA IgG and MCHC levels decreased as patients improved, while basophil granulocyte counts and IL-12p40 levels increased. Analysis including time post infection did not improve the model (Extended Data Table 3.2b). Molecular Landscape of EBV-Associated Paediatric Long COVID As EBV exposure was found to be significantly different and linked to haematopoietic alterations, we assessed its potential contribution of EBV exposure to paediatric LC endotypes. We applied linear mixed-effects models (LMMs), including EBV EBNA IgG status (cut-off 50 U ml -1 ), BMI percentile, age, sex and time since SARS-CoV-2 infection as covariates. Repeated measures were clustered by patient ID (Fig. 4, left, Table 1 for multiple comparison, Extended Data Table 4). Initially, complement components and anti-SARS-CoV-2 spike protein antibodies were applied, but did not show any dependence on EBV exposure (Table 1a, Extended Data Table 3a). At the molecular level, EBV exposure was associated with a distinct innate immune signature: IL-1α and IL-15 levels were significantly elevated (Fig. 4, Table 1b, Extended Data Table 4). We hypothesised the existence of an additional Th17/Th22 polarised endotype, which was evident in EBV-experienced patients and characterised by significant upregulation of IL-12p40 and IL-22 (Table 1c; Extended Data Table 4). This was paralleled by increased levels of the regulatory cytokines IL-10, IL-11 and IL-27 (Table 1d; Extended Data Table 4), suggesting a concomitant immunoregulatory counter-response. Importantly, EBV exposure was also associated with aAb titres against Prothrombin (Table 1e; Extended Data Table 4), supporting the presence of an autoimmune component within this subgroup. Of note, EBV-Ab response may even cross-react with other aAb 36 . Innate activation was further corroborated by significantly increased neutrophil counts (Table 1f; Extended Data Table 4), indicating sustained myeloid involvement. As chronic EBV responses are known to increase the risk of depression in female adolescents 37 , and depression and anxiety negatively impact recovery in children with chronic fatigue syndrome 38 , the model was applied to mental health scores (Table 1g; Extended Data Table 4). EBV-seropositive patients did not exhibit significantly worse clinical mental health scores than EBV-naïve ones. To summarize the immune and clinical landscape associated with EBV exposure in paediatric LC, we assembled a composite visualization (Fig. 4, left) integrating all significant associations in the above described model (Table 1; Extended Data Table 4), which demonstrates a pronounced inflammatory landscape in the subgroup of EBV-experienced paediatric LC patients. Exploitation of EBV EBNA IgG-negative or low status observed in about half of our paediatric LC cohort, we applied the model from Fig. 3f to this subgroup, excluding EBV status as a covariate (Extended Data Table 5a). The model fit for the Bell score remained highly significant (p < 0.001) with an improved R²Marg of 0.286. IL-12p40 levels and basophil counts were the strongest positive predictors of Bell Score, while lower mean corpuscular haemoglobin concentration (MCHC) correlated with higher Bell scores, indicating better functional status. To probe potential metabolic determinants, we included vitamin B1 (thiamine), based on increasing evidence for mitochondrial dysfunction in LC and on its essential role in mitochondrial energy metabolism 20,21 . This addition significantly improved the model fit for the Bell score further (R² Marg = 0.314)(Fig. 4, right; Extended Data Table 5b), supporting an immunometabolic contribution to clinical severity. This result delineates a distinct EBV-negative LC endotype explainable within an immunometabolic framework. Besides the absence of prior EBV exposure, our findings highlight several protective factors in LC, including IL-13, IL-12p40, basophilic granulocytes, vitamin B1, and autoantigen DFS70-specific aAb. Verification of Subtypes of Paediatric Long COVID We identified three features by which LC subtypes can be categorised: The first relates to the duration post-infection, where LC in the first year is characterized by Th2- and viral-associated cytokines, and LC persisting for 2–3 years is characterized by Th17-like and innate-like systemic cytokines. A second relates to autoimmunity, as anti-DFS70-positivity with an anti-DFS70-positive pattern exhibited markedly fewer coagulation abnormalities. The third involves EBV serostatus which was linked to a pro-inflammatory cytokine profile and granulocytic dysregulation. To further explore these separately immuno-clinical subgroups in paediatric long COVID, we systematically analysed a panel of 43 haematopoietic, coagulation, electrolyte, and vitamin-related biomarkers (Table 2, Extended Data Table 6a,b). Using LMMs for each parameter, we tested for associations with subgroup assignment, EBV serostatus, DFS70 aAb status, and disease duration (first year vs. later), accounting for repeated measures through random intercepts. The 43 parameters assessed included thyroid-stimulating hormone (TSH), activated partial thromboplastin time (aPTT), lipoprotein(a) [LP(a)], electrolytes (Na⁺, K⁺, Cl⁻, Ca²⁺, bicarbonate), vitamins (B1, B6, B12, D, folic acid), metabolic markers (glucose, lactate, creatinine, eGFR), and inflammation/coagulation markers (CRP, D-dimer, fibrinogen, Factor VIII, protein S). After correcting for multiple testing, three metabolic markers were found to be significantly associated with the defined subgroups (each p < 0.001, corrected p = 0.043): TSH, aPTT and LP(a)(Table 2). LP(a) and TSH levels increased over time, whereas aPTT decreased (Extended Data Table 6b). Variation in TSH across subgroups suggests endocrine dysregulation linked to specific immunological profiles. APTT´s association to all subgroups points toward distinct coagulation patterns of each subgroup and LP(a) levels were associated with both disease duration and DFS70 aAb status, likely reflecting inflammation and indicating evolving vascular risk 39,40 . These biomarkers exhibited non-overlapping association patterns, supporting the concept of pathophysiological divergence. Together, these findings provide strong evidence for the existence of biologically meaningful endotypes in paediatric long COVID. Discussion Our cohort study reveals persistent and temporal immune and metabolic dysregulation as a hallmark of paediatric long COVID (LC). Although paediatric and adult LC share symptomatic similarities, our findings demonstrate that distinct molecular and cellular consequences of SARS-CoV-2 defence appear in paediatric LC. While paediatric LC exhibits many symptoms similar to that of adult LC, our data show that it does not demonstrate abnormalities relating to heart or endothelial 17 system function; vitamin or trace element values; gastrointestinal tract parameters; systemic autoimmunity 22 , 23 ; chronic complement activation 18 , 19 ; or reactivation of EBV IgM 19 . This may reflect the greater regenerative capacity of children, the lower cumulative burden of lifestyle-related factors, and a less established and rigid immunological memory compared to adults. However, similar findings include persistent activity of innate immunity and signs of chronic inflammation 22 , 23 . In adult myalgic encephalomyelitis or chronic fatigue syndrome (ME/CFS), elevated cytokine levels have been reported as well as autoantibodies 41 , but at least their cytokine signatures closely resemble those observed in EBV-positive cohorts 42 . Along the same lines, cytokines and metabolic features potentially associated with reduced fatigue as we demonstrate here, were not identified, presumably due to the near-universal prevalence of latent EBV infection in adults, which limits differential analysis. Cytokines known to be involved in the antiviral and anti–SARS-CoV-2 response, such as IL-13, IL-33 and IFNα, were markedly elevated during the first year 23 , 30 , 43 and subsequently decreased during the second and third years. This “early” IFNα rise is consistent with adult long COVID 43 . Comparing the responses of adult and paediatric LC, innate cytokines tend to be more prominent in paediatric SARS-CoV-2 responses 5 . Elevated innate cytokines in LC such as those reported here, may indeed not only drive inflammation, but likely also fuel LC symptoms associated with depressive moods 44 . Although both adult and paediatric LC show persistent deregulation of cytokines, adult LC shows alternative cytokine involvement compared to what we have reported here for paediatric LC 3 . This may be due to differences in the status of the immune system, as well as differences in life-style, antigenic provocation, autoimmunity and exhausted responses. Certain features of paediatric LC—such as Epstein-Barr virus (EBV) seropositivity and innate immune activation—are shared with multisystem inflammatory syndrome in children (MIS-C), a distinct post-infectious complication of SARS-CoV-2 45 . However, the immunometabolic profile observed in EBV-seropositive LC patients suggests a protracted and mechanistically distinct disease process, in contrast to the acute, hyperinflammatory phenotype of MIS-C, which typically manifests 2 to 6 weeks after infection 37 . In postural orthostatic tachycardia syndrome (POTS), which can also occur post SARS-CoV-2 infection, GGPR autoantibodies have been similarly elevated as reported here, and levels did not differ from those in healthy controls and no correlation with disease severity 46 . Critically, we identified several immune markers and clinical features that distinguish children with a more favourable disease course from those with a more protracted trajectory. Linking a decrease in MCHC to an improvement in the Bell score suggests a link to microcirculatory dynamics. Although MCHC cannot be considered a standalone indicator of blood viscosity, its reduction may well reflect improved erythrocyte hydration or membrane stability, which could enhance tissue oxygen delivery 35 ,4748 . As reduced deformability hampers adult LC erythrocytes, restoring paediatric red-cell rheology may likewise accelerate recovery 35 . IL-12p40, mainly known to be a subunit shared by IL-12 and IL-23, plays a key role in T-cell responses, but can also bind monomerically to the IL12 receptor as an antagonist 49 . As its serum concentration is more than 100-fold higher than IL-12 or IL-23 (see Fig. 2 h; 3 d; Extended Fig. 2 b; 3.1b), it may therefore contribute to balancing immune activity rather than fuelling persistent inflammation. The presence of these factors in LC patients associated with milder disease could indicate a shift in those patients towards resolution of inflammation rather than its persistence. In addition, absence of neuronal involvement might be identifiable as only children with a Bell scores of > 40, consistently exhibit NfL levels in the lower normative range, which could provide an additional surrogate marker for a favourable disease course. Severe cases show NfL exclusively in the upper tail of the normal distribution. As this is not the case in MIS-C, it likely reflects a distinct molecular aethiopathology 50 . This reinforces the idea that NfL is a sensitive marker of neuronal involvement in paediatric LC, however, it remains to be determined whether such elevations predict clinical progression or long-term neurological outcomes. The immune dynamics are interwoven with perturbations in other systems, reflected in parallel changes in haematological indices—including evidence of coagulation imbalance—and in metabolic and endocrine markers over time. This highlights the multisystemic nature of paediatric LC and suggests that these non-overlapping metabolic profiles may represent either causally distinct endotypes or downstream consequences of broader immune dysregulation. Regardless of their origin, these features contribute to disease heterogeneity and underscore the need for tailored therapeutic approaches, potentially explaining the failure of previous experimental interventions that targeted LC as a single disease entity. Taken together, these findings demonstrate that the identified endotypes are distinct not only clinically and immunologically, but also in terms of their metabolic and vascular characteristics. This layered heterogeneity reinforces the need for a stratified approach to long COVID diagnostics and management, and supports the utility of biomarker-guided sub-phenotyping in future precision medicine frameworks. Declarations Author Contribution MCBW, KV, IH, PJ, SW, DR, AR, JM, ML, LN, PH, and EU performed or supervised experiments and generated and analysed data. EU, DV, MP, CA, HP, LN evaluated and recruited patients and/or controls. SW and MCBW verified the use and visualization of appropriate statistical methods for modelling. MCBW wrote the original draft with the help of all other co-authors and is the lead corresponding author. MCBW, KV, IH, PJ, JK, DV performed computational analysis of data. DV and MCBW conceptualized the project. DV directed the project and was responsible for funding acquisition. Funding This study was conducted with resources provided by the Multicentre long COVID registry (MLC-R), by the BMBF (to DV: LongCOCID 01EP2101, subproject funding to MCBW 01EP2101C), DFG Br1860/18 (MCBW), and the Ministry of Science, Technology and Environment of Saxony-Anhalt (SarsImmunGender I-196 to MCBW). Open access funding was provided by DFG and Otto-von-Guericke University Magdeburg, Germany. Disclosure Statement Authors declare no competing interests. Data availability Data will be made available upon request. References Sørensen, A. I. V. et al. A nationwide questionnaire study of post-acute symptoms and health problems after SARS-CoV-2 infection in Denmark. Nature communications 13, 4213; 10.1038/s41467-022-31897-x (2022). Morello, R., Martino, L. & Buonsenso, D. Diagnosis and management of post-COVID (Long COVID) in children: a moving target. Current opinion in pediatrics 35, 184–192; 10.1097/MOP.0000000000001221 (2023). Davis, H. E., McCorkell, L., Vogel, J. M. & Topol, E. J. Long COVID: major findings, mechanisms and recommendations. Nature reviews. 21, 133–146; 10.1038/s41579-022-00846-2 (2023). Greenhalgh, T., Sivan, M., Perlowski, A. & Nikolich, J. Ž. Long COVID: a clinical update. Lancet 404, 707–724; 10.1016/S0140-6736(24)01136-X (2024). Chou, J., Thomas, P. G. & Randolph, A. G. Immunology of SARS-CoV-2 infection in children. Nature immunology 23, 177–185; 10.1038/s41590-021-01123-9 (2022). Götzinger, F. et al. COVID-19 in children and adolescents in Europe: a multinational, multicentre cohort study. The Lancet. Child & adolescent health 4, 653–661; 10.1016/S2352-4642(20)30177-2 (2020). Stephenson, T. et al. Long COVID (post-COVID-19 condition) in children: a modified Delphi process. Archives of disease in childhood 107, 674–680; 10.1136/archdischild-2021-323624 (2022). Peluso, M. J. & Deeks, S. G. Mechanisms of long COVID and the path toward therapeutics. Cell 187, 5500–5529; 10.1016/j.cell.2024.07.054 (2024). Rao, S. et al. Postacute Sequelae of SARS-CoV-2 in Children. Pediatrics 153 ; 10.1542/peds.2023-062570 (2024). Camporesi, A. et al. Characteristics and predictors of Long Covid in children: a 3-year prospective cohort study. EClinicalMedicine 76, 102815; 10.1016/j.eclinm.2024.102815 (2024). Knolle, J. et al. Children From the Age of Three Show a Developmental Switch in T-Cell Differentiation. Frontiers in immunology 11, 1640; 10.3389/fimmu.2020.01640 (2020). Vogel, K. et al. Bifidobacteria shape antimicrobial T-helper cell responses during infancy and adulthood. Nature communications 14, 5943; 10.1038/s41467-023-41630-x (2023). Shrock, E. et al. Viral epitope profiling of COVID-19 patients reveals cross-reactivity and correlates of severity. Science 370 ; 10.1126/science.abd4250 (2020). Fazekas, d. S. G. & Webster, R. G. Disquisitions of Original Antigenic Sin. I. Evidence in man. The Journal of experimental medicine 124, 331–345; 10.1084/jem.124.3.331 (1966). Zajac, A. J. et al. Viral immune evasion due to persistence of activated T cells without effector function. The Journal of experimental medicine 188, 2205–2213; 10.1084/jem.188.12.2205 (1998). Gutman, E. G. et al. Long COVID: plasma levels of neurofilament light chain in mild COVID-19 patients with neurocognitive symptoms. Molecular psychiatry 29, 3106–3116; 10.1038/s41380-024-02554-0 (2024). Vassiliou, A. G. et al. Endotheliopathy in Acute COVID-19 and Long COVID. International journal of molecular sciences 24 ; 10.3390/ijms24098237 (2023). Cervia-Hasler, C. et al. Persistent complement dysregulation with signs of thromboinflammation in active Long Covid. Science 383, eadg7942; 10.1126/science.adg7942 (2024). Klein, J. et al. Distinguishing features of long COVID identified through immune profiling. Nature 623, 139–148; 10.1038/s41586-023-06651-y (2023). Berezhnoy, G. et al. Maintained imbalance of triglycerides, apolipoproteins, energy metabolites and cytokines in long-term COVID-19 syndrome patients. Frontiers in immunology 14, 1144224; 10.3389/fimmu.2023.1144224 (2023). Molnar, T. et al. Mitochondrial dysfunction in long COVID: mechanisms, consequences, and potential therapeutic approaches. GeroScience 46, 5267–5286; 10.1007/s11357-024-01165-5 (2024). Rojas, M. et al. Autoimmunity is a hallmark of post-COVID syndrome. Journal of translational medicine 20, 129; 10.1186/s12967-022-03328-4 (2022). Lingel, H. et al. Unique autoantibody prevalence in long-term recovered SARS-CoV-2-infected individuals. Journal of autoimmunity 122, 102682; 10.1016/j.jaut.2021.102682 (2021). Altmann, D. M. et al. Persistent symptoms after COVID-19 are not associated with differential SARS-CoV-2 antibody or T cell immunity. Nature communications 14, 5139; 10.1038/s41467-023-40460-1 (2023). Scheibenbogen, C. et al. Fighting Post-COVID and ME/CFS - development of curative therapies. Frontiers in medicine 10, 1194754; 10.3389/fmed.2023.1194754 (2023). Bain, R. et al. Clinical characteristics of SARS-CoV-2 infection in children with cystic fibrosis: An international observational study. Journal of cystic fibrosis 20, 25–30; 10.1016/j.jcf.2020.11.021 (2021). Zeren, M., Gurses, H. N., Denizoglu Kulli, H., Ucgun, H. & Cakir, E. Sit-to-stand test in children with bronchiectasis: Does it measure functional exercise capacity? Heart & 49, 796–802; 10.1016/j.hrtlng.2020.09.017 (2020). Lercher, A. et al. Antiviral innate immune memory in alveolar macrophages following SARS-CoV-2 infection ameliorates secondary influenza A virus disease. Immunity 57, 2530-2546.e13; 10.1016/j.immuni.2024.08.018 (2024). Atchison, C. J. et al. Long-term health impacts of COVID-19 among 242,712 adults in England. Nature communications 14, 6588; 10.1038/s41467-023-41879-2 (2023). Meltendorf, S. et al. IL-13 determines specific IgE responses and SARS-CoV-2 immunity after mild COVID-19 and novel mRNA vaccination. European journal of immunology 52, 1972–1979; 10.1002/eji.202249951 (2022). Gieseck, R. L., Wilson, M. S. & Wynn, T. A. Type 2 immunity in tissue repair and fibrosis. Nature reviews. Immunology 18, 62–76; 10.1038/nri.2017.90 (2018). Zar, H. J. et al. Early-life respiratory syncytial virus disease and long-term respiratory health. The Lancet. Respiratory medicine 12, 810–821; 10.1016/S2213-2600(24)00246-7 (2024). Abdelhak, A. et al. Serum neurofilament light chain reference database for individual application in paediatric care: a retrospective modelling and validation study. The Lancet. Neurology 22, 826–833; 10.1016/S1474-4422(23)00210-7 (2023). Carbone, T. et al. Prevalence and serological profile of anti-DFS70 positive subjects from a routine ANA cohort. Scientific reports 9, 2177; 10.1038/s41598-019-38686-5 (2019). Lechuga, G. C., Morel, C. M. & De-Simone, S. G. Hematological alterations associated with long COVID-19. Frontiers in physiology 14, 1203472; 10.3389/fphys.2023.1203472 (2023). Maguire, C. et al. Molecular mimicry as a mechanism of viral immune evasion and autoimmunity. Nature communications 15, 9403; 10.1038/s41467-024-53658-8 (2024). Ford, J. L. & Stowe, R. P. Depressive symptoms are associated with salivary shedding of Epstein-Barr virus in female adolescents: The role of sex differences. Psychoneuroendocrinology 86, 128–133; 10.1016/j.psyneuen.2017.09.009 (2017). Rowe, K. S. Long Term Follow up of Young People With Chronic Fatigue Syndrome Attending a Pediatric Outpatient Service. Frontiers in pediatrics 7, 21; 10.3389/fped.2019.00021 (2019). Bae, G. et al. Longitudinal lipoprotein and inflammatory mediators analysis uncover persisting inflammation and hyperlipidemia following SARS-CoV-2 infection in long COVID-19. Metabolomics : Official journal of the Metabolomic Society 21, 65; 10.1007/s11306-025-02262-y (2025). Stahl, D. et al. High increase in levels of lipoprotein(a) in plasma of patients with rheumatoid arthritis after COVID-19. Arthritis & rheumatology (Hoboken, N.J.) 75, 329–331; 10.1002/art.42374 (2023). Loebel, M. et al. Antibodies to β adrenergic and muscarinic cholinergic receptors in patients with Chronic Fatigue Syndrome. Brain, behavior, and immunity 52, 32–39; 10.1016/j.bbi.2015.09.013 (2016). Montoya, J. G. et al. Cytokine signature associated with disease severity in chronic fatigue syndrome patients. Proceedings of the National Academy of Sciences of the United States of America 114, E7150-E7158; 10.1073/pnas.1710519114 (2017). Queiroz, M. A. F. et al. Severe COVID-19 and long COVID are associated with high expression of STING, cGAS and IFN-α. Scientific reports 14, 4974; 10.1038/s41598-024-55696-0 (2024). Dantzer, R., O’Connor, J. C., Freund, G. G., Johnson, R. W. & Kelley, K. W. From inflammation to sickness and depression: when the immune system subjugates the brain. Nature reviews. Neuroscience 9, 46–56; 10.1038/nrn2297 (2008). Goetzke, C. C. et al. TGFβ links EBV to multisystem inflammatory syndrome in children. Nature 640, 762–771; 10.1038/s41586-025-08697-6 (2025). Hall, J. et al. Detection of G Protein-Coupled Receptor Autoantibodies in Postural Orthostatic Tachycardia Syndrome Using Standard Methodology. Circulation 146, 613–622; 10.1161/CIRCULATIONAHA.122.059971 (2022). Renoux, C. et al. Impact of surface-area-to-volume ratio, internal viscosity and membrane viscoelasticity on red blood cell deformability measured in isotonic condition. Scientific reports 9, 6771; 10.1038/s41598-019-43200-y (2019). Geisness, A. C. et al. Ionophore-mediated swelling of erythrocytes as a therapeutic mechanism in sickle cell disease. Haematologica 107, 1438–1447; 10.3324/haematol.2021.278666 (2022). Ling, P. et al. Human IL-12 p40 homodimer binds to the IL-12 receptor but does not mediate biologic activity. The Journal of Immunology 154, 116–127; 10.4049/jimmunol.154.1.116 (1995). Geis, T. et al. Serum neurofilament light chain (sNfL) values in a large cross-sectional population of children with asymptomatic to moderate COVID-19. Journal of neurology 268, 3969–3974; 10.1007/s00415-021-10554-1 (2021). Methods Human participants Written informed consent was obtained from all participants and/or their legal representatives in accordance with the Declaration of Helsinki. The study was approved by the ethics committees of the University Hospital Jena (2022-2614_1-BO) and the Otto-von-Guericke University Magdeburg (164-18). Although the study was observational and non-interventional in design, it is registered in the German Clinical Trials Register (DRKS00028523). Extended Data Table 1 and Fig. 1 report the sex, age, percentile BMI values (calculated according to the KiGGS study 51 ) and clinical manifestation details of participants during acute long COVID. The cohort included paediatric long COVID patients infected from October 2020 onwards, the majority of whom were infected during the third wave of the SARS-CoV-2 pandemic in Germany (see Extended Data Table 1). The control group comprised healthy children and adolescents, as well as paediatric cystic fibrosis (CF) patients in a stable condition. Full details of the cohort are provided in Extended Data Table 1a. Only patients who were diagnosed with long COVID at their initial presentation were included in the study. Of the 78 paediatric long COVID patients enrolled, four were excluded: three due to alternative plausible causes of their symptoms (Borrelia infection, psychosomatic symptoms, or symptom onset predating SARS-CoV-2 infection), and one due to withdrawal of consent. Patient Assessments LC patient assessments encompassed both validated questionnaires and objective physical performance tests Fig.1a. Most questionnaires for participants were completed by proxy. Psychometric and functional assessments included the Generalized Anxiety Disorder 7-item scale (GAD-7) for anxiety symptoms, the Patient Health Questionnaire 9-item (PHQ-9) for depressive symptoms, and the Short Form-12 Physical and Mental Component Summary scores (SF-12 PCS and SF-12 MCS) to evaluate physical and mental health-related quality of life. Vital parameters and a wide range of symptoms were assessed through these instruments, including the Munich Long COVID Symptom Questionnaire (MLCSQ), which assesses the frequency of 96 potential symptoms of LC, divided into 13 Systems (Fig.1a,c; Extended Data Fig.1). Symptom burden was captured through both parent-reported symptoms (PRS) and patient-reported symptoms (PtRS). Health-related quality of life in children was assessed using the KIDSCREEN questionnaire, and sleep disturbances were evaluated with the Children’s Sleep Habits Questionnaire (CSHQ). Functional impairment was assessed using the Bell Disability Scale (Bell Score), while fatigue was evaluated with the Fatigue Severity Scale (FSS). Physical performance was evaluated using the Sit-to-Stand Test (STS) and Hand Grip Strength Test 52,53 . Isolation of PBMCs and serum Peripheral blood was collected into EDTA and SST tubes. PBMCs were isolated at room temperature by Ficoll-Paque density centrifugation. Specimens from paediatric LC patients were archived at the Integrated Biobank Jena (University Hospital Jena) and controls at the Medical Faculty, Otto-von-Guericke University Magdeburg. Serum was analysed immediately or stored at −20 °C for batched aAb and cytokine assays with matched controls. Cells were cryopreserved in heat-inactivated FCS supplemented with 10 % (v/v) DMSO. Cytokine analysis in serum Cytokine levels in serum samples were quantified using a flow cytometric multiplex bead-based assay (Supplementary Fig.1) 12,23 . The concentrations of GM-CSF, IL-1α, IL-1β, IL-2, IL-3, IL-4, IL-5, IL-6, IL-9, IL-10, IL-11, IL-12p40, IL-12p70, IL-13, IL-15, IL-18, IL-22, IL-23, IL-27, IL-33, TNFα, TSLP, IFNα, and IFNγ were measured in serum samples from paediatric LC patients and controls using the LegendPlex Human Th Cytokine Panel and LegendPlex Human Cytokine Panel 2 (BioLegend), following the protocols provided by the manufacturer and as described previously 12,23 . Data acquisition was performed on a FACSFortessa X-20 (Becton Dickinson), and results were analysed using the LegendPlex Data Analysis Software Suite (Qognit). Routine Laboratory Testing Laboratory analyses included metabolic markers, blood gases, differential blood counts, general clinical chemistry, coagulation parameters, urinalysis, immunoglobulins, hormone levels, allergy diagnostics, autoantibodies, antiviral antibodies (EBV- and CMV-specific), and anti-SARS-CoV-2 serology. Pulmonary function parameters, vascular function (static and dynamic analysis), and vital signs were also assessed. Arterial/venous blood gases were analysed on ABL 90 Flex Plus instruments. Complete and differential blood counts were obtained on an XN-1000 (Sysmex) using fluorescence flow cytometry, SLS-Hb and impedance channels. Anti-SARS-CoV-2 spike S1 IgG (quantitative; ELISA Genie, CBK4154) and anti-nucleocapsid IgG (qualitative; Roche Elecsys) were performed per manufacturer’s instructions. Values exceeding analytical limits of detection were recorded as maximum + 1 or minimum / 2, respectively. Autoantibody Profiling and Biomarker Quantification Antinuclear antibodies (ANA) were detected by indirect immunofluorescence using the ANA HEp-2 plus kit (#8101, Generic Assays, Dahlewitz, Germany) on HEp-2 cells following serial serum dilutions (1:80–1:2560; positivity defined as ≥1:80), with endpoint titres evaluated using a fluorescence microscope (Olympus, Japan). Presence of anti-DFS70 antibodies was confirmed by immunoblot using the EUROLINE Anti-DFS70 (IgG) kit (#DL 159z-1601 G, Euroimmun, Lübeck, Germany). Anti-CCP IgG (#3665, cut off <30 U/ml), anti-TG IgA (#4033, cut off <20 U/ml; both Generic Assays, Dahlewitz, Germany) as well as anti-Cardiolipin IgG (#ORG 515S, cut off <10 U/ml), anti-β2-Glycoprotein IgG (#ORG 521S, cut off <10 U/ml), anti-Prothrombin IgG (#ORG 541S, cut off <20 U/ml), anti-PR3 IgG (#ORG 618, cut off <10 U/ml) and anti-MPO IgG (#ORG 519, cut off <5 U/ml; all Orgentec, Mainz, Germany) were analysed by ELISA. Serum samples were diluted 1:100 in sample buffer and transferred in duplicate into the respective microtiter plate together with standards and controls. Final analysis was performed using a Tecan Sunrise Microplate reader (Tecan, Männedorf, Switzerland) and calculated according to linearity scale. Values below the diagnostic thresholds were retained as described previously to illustrate the assay’s linear range 23,30 , acknowledging the primarily diagnostic role of established cut-offs. Serum levels of IgG autoantibodies directed against β1-adrenergic, β2-adrenergic, M3-muscarinic, and M4-muscarinic acetylcholine receptors were quantified using a standardized ELISA (CellTrend GmbH, Luckenwalde, Germany) using MAGELLAN TM microplate reader. Each assay run included manufacturer-provided positive and negative controls to ensure assay validity. Each assay included internal calibrators and manufacturer-provided positive and negative controls to ensure linearity and validity. Internal quality control was performed at three concentration levels to assess intra-assay precision. Reference ranges for physiological levels of these functional aAb were defined as <15 U/mL for β1-adrenergic, <8 U/mL for β2-adrenergic, <6 U/mL for M3-muscarinic, and 40 U/ml. Analysis of soluble NfL (NfL) levels was conducted using the Quanterix Simoa NF-Light assay Advantage Kit (Lexington, MA), according to the protocol provided. All samples were analyzed in duplicate. Testing was performed blinded to patient clinical/paraclinical data and outcome measures. To adjust for the influence of age, NfL concentrations were converted to Z scores and (interchangeable) percentiles https://shiny.dkfbasel.ch/baselnflreference-for-kids 33 . The z-score represents the number of standard deviations a particular value deviates from the mean of healthy, age-matched individuals 33 . The z-score represents the number of standard deviations from the mean of healthy age-matched controls. Statistics Linear mixed-effects models (LMMs) were used throughout to assess longitudinal changes in clinical, functional, metabolic and immunological parameters while accounting for repeated measures via a random intercept for patient ID 54 . Fixed effects varied by model and included time since SARS-CoV-2 infection (≤1 vs. >1 year or continuous), EBV EBNA IgG status (cut-off 50 U ml -1 ), DFS70 status (positive/negative), age, sex, BMI percentile, and relevant interactions. Specific LMMs are detailed in Fig. 1e, Fig. 2d, Fig. 3f, Fig. 4, and Extended Data Tables 1–6. Model assumptions were assessed via visual inspection of model residuals; no relevant heteroscedasticity or multicollinearity (VIF < 2) was observed. P-values were calculated using the Satterthwaite approximation and validated via bootstrapping (1,000 iterations); adjusted values were corrected using the Holm–Bonferroni method. Only corrected p-values < 0.05 were considered significant. Marginal R² values indicate variance explained by fixed effects. Statistical analyses were performed using GraphPad Prism, Jamovi (v3.0), and the GamLj module (v3), including linear and linear mixed-effects models (LMMs) with donor ID as a random intercept. Normality was assessed using Shapiro–Wilk and Kolmogorov–Smirnov tests, and visually via Q–Q plots. Outliers were identified using Grubbs’ test. For group comparisons, Mann–Whitney U-tests were used for non-parametric data; the Friedman test with Dunn’s post hoc correction was applied for paired samples. Pearson’s correlation was used for normally distributed variables; Spearman’s rank correlation for ordinal or non-normally distributed data. Diagnostic values below detection limits were imputed as half the limit; values above range were set to maximum + 1, as specified in Extended Data Table 2. References 51. Hölling, H. et al. Die KiGGS-Studie. Bundesweit repräsentative Längs- und Querschnittstudie zur Gesundheit von Kindern und Jugendlichen im Rahmen des Gesundheitsmonitorings am Robert Koch-Institut. Bundesgesundheitsblatt, Gesundheitsforschung, Gesundheitsschutz 55, 836–842; 10.1007/s00103-012-1486-3 (2012). 52. Haile, S. R. et al. Reference values and validation of the 1-minute sit-to-stand test in healthy 5-16-year-old youth: a cross-sectional study. BMJ open 11, e049143; 10.1136/bmjopen-2021-049143 (2021). 53. Reychler, G., Audag, N., Mestre, N. M. & Caty, G. Assessment of Validity and Reliability of the 1-Minute Sit-to-Stand Test to Measure the Heart Rate Response to Exercise in Healthy Children. JAMA pediatrics 173, 692–693; 10.1001/jamapediatrics.2019.1084 (2019) 54. Arra, A. et al. PD1+ innate lymphoid cells 3 predict JAK-dependent inflammation in rheumatoid arthritis. Journal of autoimmunity 154, 103424; 10.1016/j.jaut.2025.103424 (2025). Tables Tables 1 and 2 are available in the Supplementary Files section. Additional Declarations There is NO Competing Interest. Supplementary Files Tables.docx SFig12LegendPlexV2.docx Supplementary Fig. 1 ExtendedDataFigureandTables.docx Cite Share Download PDF Status: Published Journal Publication published 04 May, 2026 Read the published version in Nature Communications → Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-7083240","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Biological Sciences - Article","associatedPublications":[],"authors":[{"id":485063360,"identity":"5a391b8c-af52-4923-9dd2-9975f2fab626","order_by":0,"name":"Monika Brunner-Weinzierl","email":"data:image/png;base64,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","orcid":"https://orcid.org/0000-0002-6838-3568","institution":"University Hospital, Otto-von-Guericke-University Magdeburg","correspondingAuthor":true,"prefix":"","firstName":"Monika","middleName":"","lastName":"Brunner-Weinzierl","suffix":""},{"id":485063361,"identity":"4450e4f6-7405-4768-8f91-caeaef336409","order_by":1,"name":"Katrin Vogel","email":"","orcid":"","institution":"Department of Experimental Pediatrics, University Hospital, Otto-von-Guericke-University Magdeburg","correspondingAuthor":false,"prefix":"","firstName":"Katrin","middleName":"","lastName":"Vogel","suffix":""},{"id":485063362,"identity":"ad9cf5ae-367b-4eca-8b5f-5a29ebc6d1e2","order_by":2,"name":"Irina Han","email":"","orcid":"","institution":"Department of Experimental Paediatrics, University Hospital, Otto-von-Guericke-University Magdeburg","correspondingAuthor":false,"prefix":"","firstName":"Irina","middleName":"","lastName":"Han","suffix":""},{"id":485063363,"identity":"19c031f9-7b0c-4149-9d4b-9ece1d046236","order_by":3,"name":"Pauline Jakobs","email":"","orcid":"","institution":"Department of Experimental Pediatrics, University Hospital, Otto-von-Guericke-University Magdeburg","correspondingAuthor":false,"prefix":"","firstName":"Pauline","middleName":"","lastName":"Jakobs","suffix":""},{"id":485063364,"identity":"c1234dd9-c787-4b50-8695-625b09e0f78a","order_by":4,"name":"Michael Lorenz","email":"","orcid":"https://orcid.org/0000-0001-6153-2068","institution":"Department of Paediatric Pulmonology, Clinic of Paediatrics, Jena University Hospital, Friedrich Schiller University","correspondingAuthor":false,"prefix":"","firstName":"Michael","middleName":"","lastName":"Lorenz","suffix":""},{"id":485063365,"identity":"10648246-0fb7-46ee-9983-fb9a0205a5d3","order_by":5,"name":"Lars Newman","email":"","orcid":"","institution":"Post-COVID Outpatient Clinic, Department of Paediatrics and Adolescent Medicine, Jena University Hospital, Friedrich Schiller University","correspondingAuthor":false,"prefix":"","firstName":"Lars","middleName":"","lastName":"Newman","suffix":""},{"id":485063366,"identity":"6ab92226-62ce-4df3-a072-4b7c6a5c8e4c","order_by":6,"name":"Annegret Reinhold","email":"","orcid":"","institution":"Institute of Molecular and Clinical Immunology, Otto-von-Guericke-University Magdeburg","correspondingAuthor":false,"prefix":"","firstName":"Annegret","middleName":"","lastName":"Reinhold","suffix":""},{"id":485063367,"identity":"0cd1c98a-7e3e-49b8-8576-2f437b96ec3f","order_by":7,"name":"Juliane Mohr","email":"","orcid":"https://orcid.org/0000-0002-0612-2952","institution":"Institute of Molecular and Clinical Immunology, Otto-von-Guericke-University Magdeburg","correspondingAuthor":false,"prefix":"","firstName":"Juliane","middleName":"","lastName":"Mohr","suffix":""},{"id":485063368,"identity":"f29ae271-5a76-4407-9cf4-df1a92abf86e","order_by":8,"name":"Clara Aign","email":"","orcid":"","institution":"Department of Experimental Paediatrics, University Hospital, Otto-von-Guericke-University Magdeburg","correspondingAuthor":false,"prefix":"","firstName":"Clara","middleName":"","lastName":"Aign","suffix":""},{"id":485063369,"identity":"2fd0b1bd-be22-4aff-9cd6-9976d6b111c9","order_by":9,"name":"Michelle Paszkier","email":"","orcid":"","institution":"Department of Experimental Paediatrics, University Hospital, Otto-von-Guericke-University Magdeburg","correspondingAuthor":false,"prefix":"","firstName":"Michelle","middleName":"","lastName":"Paszkier","suffix":""},{"id":485063370,"identity":"0c5d1a6c-c413-449c-add8-0c0ffe813b2f","order_by":10,"name":"Kuhle Jens","email":"","orcid":"","institution":"Department of Neurology, University Hospital, University of Basel","correspondingAuthor":false,"prefix":"","firstName":"Kuhle","middleName":"","lastName":"Jens","suffix":""},{"id":485063371,"identity":"b12bc641-2ea3-449f-8838-daae502d9334","order_by":11,"name":"Peter Huppke","email":"","orcid":"","institution":"Department of Paediatric Neurology, Jena University Hospital, Friedrich Schiller University","correspondingAuthor":false,"prefix":"","firstName":"Peter","middleName":"","lastName":"Huppke","suffix":""},{"id":485063372,"identity":"a5452c9f-7b35-498b-ac31-d4fee9b7904b","order_by":12,"name":"Stefan Weinzierl","email":"","orcid":"","institution":"Audio Communication Group, Technical University Berlin","correspondingAuthor":false,"prefix":"","firstName":"Stefan","middleName":"","lastName":"Weinzierl","suffix":""},{"id":485063373,"identity":"da84cb6f-a1fe-49e1-a6cf-3839f7cff92a","order_by":13,"name":"Dirk Reinhold","email":"","orcid":"https://orcid.org/0000-0002-6441-9631","institution":"Otto-von-Guerike-University Magdeburg","correspondingAuthor":false,"prefix":"","firstName":"Dirk","middleName":"","lastName":"Reinhold","suffix":""},{"id":485063374,"identity":"0a65871d-19b9-40b5-88d8-fe9b324480a7","order_by":14,"name":"Elisabeth Ullmann","email":"","orcid":"","institution":"Post-COVID Outpatient Clinic, Department of Paediatrics and Adolescent Medicine, Jena University Hospital, Friedrich Schiller University","correspondingAuthor":false,"prefix":"","firstName":"Elisabeth","middleName":"","lastName":"Ullmann","suffix":""},{"id":485063375,"identity":"ece111d8-e4e6-4457-9e63-f2f42cb037db","order_by":15,"name":"Hans Proquitté","email":"","orcid":"","institution":"Section of Neonatology and Paediatric Intensive Care Medicine, Department of Paediatrics and Adolescent Medicine, Jena University Hospital, Friedrich Schiller University","correspondingAuthor":false,"prefix":"","firstName":"Hans","middleName":"","lastName":"Proquitté","suffix":""},{"id":485063376,"identity":"e7e2db0e-3168-434d-9e81-9c24303cc544","order_by":16,"name":"Daniel Vilser","email":"","orcid":"","institution":"Post-COVID Outpatient Clinic, Department of Paediatrics and Adolescent Medicine, Jena University Hospital, Friedrich Schiller University","correspondingAuthor":false,"prefix":"","firstName":"Daniel","middleName":"","lastName":"Vilser","suffix":""}],"badges":[],"createdAt":"2025-07-09 11:21:40","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-7083240/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-7083240/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1038/s41467-026-72224-y","type":"published","date":"2026-05-04T04:00:00+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":87386261,"identity":"41a79d26-5812-4872-8e84-d03bcc976615","added_by":"auto","created_at":"2025-07-23 08:55:56","extension":"jpg","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":158052,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eDemographic and clinical characterisation of paediatric participants\u003c/strong\u003e\u003cbr\u003e\n\u003cstrong\u003ea\u003c/strong\u003e A schematic overview illustrates participant classification and associated analyses. Control participants (n= 27) had no clinical diagnosis of long COVID (LC). LC patients were predominantly enrolled within the first year after SARS-CoV-2 infection (Visit 1, n = 74) and re-evaluated 3–6 months later (Visit 2, n = 74). The outer ring highlights the range of analyses (n \u0026gt; 300) performed to confirm the presence and persistence of long COVID–associated symptoms. \u003cstrong\u003eb\u003c/strong\u003e Demographics of the control (upper row) and long COVID Visit 1 (lower row) groups include age (mean displayed centrally), sex distribution, and COVID-19 infection rates. Statistical analysis was performed using the Chi-square test, and p-values are indicated. \u003cstrong\u003ec\u003c/strong\u003e Symptom prevalence is shown as a histogram from least to most prevalent symptoms, categorized by system: constitutional (Const., lilac), ear, nose, throat (ENT, burgundy), cardiovascular (Cardio., green), gastrointestinal (GI, yellow), neurological (Neuro., dark blue), psychological (Psych., light blue), musculoskeletal (MSK, red), and immunological (Immun., orange). \u003cstrong\u003ed\u003c/strong\u003e A boxplot compares plasma IFNα levels between long COVID patients within the first year after infection (LC \u0026lt;1\u0026nbsp;yr) and controls. For individuals with multiple visits, a single representative value was used. The median is indicated by the horizontal line. Statistical significance was assessed using the Kruskal–Wallis test with Dunn’s post hoc test for pairwise comparisons and adjusted for multiple testing using the Mann–Whitney U-test. \u003cstrong\u003ee\u003c/strong\u003e Correlation plots show Physical Component Score (PCS), Bell Score, Fatigue Severity Scale (FSS), and Sit-to-Stand Score in relation to time post SARS-CoV-2 infection. R²\u003csub\u003eMarg\u003c/sub\u003e values indicate model fit of linear mixed-effects models (LMMs) applied either to the first year post-infection (LC \u0026lt;1 yr) or the full observation period (see Extended Data Fig. 1a,b). Donor ID was included as a random effect and time as a fixed covariate. Each dot represents an individual donor at a specific visit.\u003c/p\u003e","description":"","filename":"1.jpg","url":"https://assets-eu.researchsquare.com/files/rs-7083240/v1/8c5b24e6244a829dce588ab3.jpg"},{"id":87386256,"identity":"cd181889-a933-4c31-b79e-7f38edec84d8","added_by":"auto","created_at":"2025-07-23 08:55:56","extension":"jpg","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":188188,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eImmune, pulmonary and endothelial features of paediatric long COVID.\u003c/strong\u003e \u003cstrong\u003ea\u003c/strong\u003e Pie chart illustrating the prevalence of cardiac findings in long COVID (LC) patients, including abnormal electrocardiogram or echocardiography. \u003cstrong\u003eb\u003c/strong\u003e Boxplot illustrating the distribution of Forced Expiratory Volume in 1 second (FEV1) Z-scores across three clusters, derived from the mean FEV1 Z-score per donor to account for repeated measurements. The pathological threshold for FEV1 is indicated in red at –1.64. A Mann–Whitney U-test was used to compare cluster 0 and cluster 1. \u003cstrong\u003ec, f–h\u003c/strong\u003e Systemic cytokine levels were compared between controls (n = 27) and paediatric LC patients (\u0026lt;1 yr after infection) as indicated. For LC patients with two assessments, the mean was used (n = 47). Multiple comparisons and all cytokines tested are presented in Extended Data Fig. 2 and Extended Data Table 2.2 for multiple comparison. \u003cstrong\u003ed\u003c/strong\u003eModel fit of linear mixed-effects regression using FEV1 Z-score as the dependent variable and patient ID as the clustering variable (Extended Data Table 2). \u003cstrong\u003ee\u003c/strong\u003e Ratio of IL-4 to IFNγ for the Th2/Th1 ratio. \u003cstrong\u003ei\u003c/strong\u003eSerum levels of NfL in LC patients grouped according to low or intermediate/high Bell scores. \u003cstrong\u003ej \u003c/strong\u003eSerum levels of aAbs in controls (grey) and LC \u0026lt;1 yr (light orange). \u003cstrong\u003ek\u003c/strong\u003eAutoantibody reactivity in controls (grey) and LC \u0026lt;1 yr as well as 1–3 yr was quantified. ANOVA with post hoc testing. \u003cstrong\u003el\u003c/strong\u003e Proportion of anti-DFS70-positive LC patients among all patients based on dilution series and anti-DFS70 progression in the same patients between 2 visits. \u003cstrong\u003em\u003c/strong\u003eBoxplots showing coagulation in anti-DFS70-positive (orange) versus anti-DFS70-negative patients (beige). \u003cstrong\u003en\u003c/strong\u003eAnti-EBNA IgG levels to EBV in anti-DFS70-positive LC (orange) compared to anti-DFS70-negative LC (beige). Outliers were detected using the Grubbs test, normality was assessed via the Kolmogorov–Smirnov test, and significance evaluated using the Mann–Whitney U-test (\u003cstrong\u003ec, e–h, m, n\u003c/strong\u003e) or Welch test (\u003cstrong\u003ei\u003c/strong\u003e). Outliers are shown with open circles. Data points represent individual donors; boxplots and violin plots indicate medians. Significance levels: ****p \u0026lt; 0.001; ***p \u0026lt; 0.005; **p \u0026lt; 0.01; *p \u0026lt; 0.05.\u003c/p\u003e","description":"","filename":"2.jpg","url":"https://assets-eu.researchsquare.com/files/rs-7083240/v1/6f35b04f505ee87c2f75e1f1.jpg"},{"id":87386259,"identity":"b61bff02-7b19-4a31-aaba-eccf6f7e6185","added_by":"auto","created_at":"2025-07-23 08:55:56","extension":"jpg","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":142172,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eLong-term changes in paediatric long COVID\u003c/strong\u003e \u003cstrong\u003ea–d\u003c/strong\u003e Systemic cytokine levels were compared between controls and paediatric long COVID (LC) patients (1–3 yr post-infection) as indicated. Multiple comparisons and full cytokine panel shown in Extended Data Fig. 3.1 and Extended Data Table 3.1. Each dot represents an individual donor or the mean value within the respective time frame (n=47/n=45). Statistical analysis was performed using the Mann-Whitney U-test, as in Fig. 2. \u003cstrong\u003ee\u003c/strong\u003eNet plot summarizes significant cytokine differences for LC \u0026lt;1 yr and LC 1-3 yr. Differences relative to controls (grey circle) are visualized using ordinal coding, indicating no significant difference, significant increase, or further increase compared to LC \u0026lt; 1 yr. \u003cstrong\u003ef\u003c/strong\u003e Model fit of a linear mixed-effects regression using the Bell Score of LC patients as the dependent variable and patient ID as the clustering variable (see Extended Data Table 3.2). Significance levels: ****p \u0026lt; 0.001; ***p \u0026lt; 0.005; **p \u0026lt; 0.01; *p \u0026lt; 0.05\u003c/p\u003e","description":"","filename":"3.jpg","url":"https://assets-eu.researchsquare.com/files/rs-7083240/v1/1e54d58fce028492dfabcfbb.jpg"},{"id":87386255,"identity":"5d669ac6-f3bf-4c92-af9b-67d09a73b4be","added_by":"auto","created_at":"2025-07-23 08:55:56","extension":"jpg","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":57350,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eBiological EBV-linked subtypes of paediatric long COVID\u003c/strong\u003e. \u003cstrong\u003eLeft panel:\u003c/strong\u003e Summary of significant dependent variables identified through linear mixed-effects models (LMMs) (see Table 1 and Extended Data Table 4). Only model fits surviving Bonferroni–Holm correction for multiple testing are shown. Overlapping circles indicate shared associations with key covariates: \u003cem\u003etime since SARS-CoV-2 infection\u003c/em\u003e, \u003cem\u003eage\u003c/em\u003e, and \u003cem\u003esex\u003c/em\u003e (BMI showed no significant association). Repeated measures were accounted for using Patient ID as a random intercept.\u003cstrong\u003e Right panel:\u003c/strong\u003e\u0026nbsp; \u003cstrong\u003eEBV-naïve subgroup\u003c/strong\u003e (defined as IgG EBV EBNA \u0026lt; 50\u0026nbsp;U\u0026nbsp;ml\u003csup\u003e-1\u003c/sup\u003e) is shown with LMM model fit examining associations between the dependent variable Bell score and the covariates basophilic granulocytes, MCHC, IL-12p40, vitamin B1, with Patient ID included as a random intercept (Extended Data Table 5b). Significant associations are marked (*p\u0026lt;0.05; **p\u0026lt;0.01; ****p\u0026lt;0.001). Model fit statistics is indicated.\u003c/p\u003e","description":"","filename":"4.jpg","url":"https://assets-eu.researchsquare.com/files/rs-7083240/v1/6bd0510cd69e02cf139abb85.jpg"},{"id":108476809,"identity":"2a4f6d69-1e49-4aae-80e1-a9b4a0a992fa","added_by":"auto","created_at":"2026-05-05 07:05:38","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":934138,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7083240/v1/f7cb1066-b4ee-44da-b047-7f73484c0c86.pdf"},{"id":87386254,"identity":"97f6952f-228c-40f6-88c9-a04e55ac37c1","added_by":"auto","created_at":"2025-07-23 08:55:55","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":22225,"visible":true,"origin":"","legend":"","description":"","filename":"Tables.docx","url":"https://assets-eu.researchsquare.com/files/rs-7083240/v1/59b1e39448d80b6a489674c7.docx"},{"id":87386258,"identity":"40cc994c-5af0-4d90-ada2-7c78f17901b1","added_by":"auto","created_at":"2025-07-23 08:55:56","extension":"docx","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":998491,"visible":true,"origin":"","legend":"Supplementary Fig. 1","description":"","filename":"SFig12LegendPlexV2.docx","url":"https://assets-eu.researchsquare.com/files/rs-7083240/v1/bf2a434140635460e247937c.docx"},{"id":87386954,"identity":"284b5979-9799-4bd9-9b41-af87ba4733a4","added_by":"auto","created_at":"2025-07-23 09:03:56","extension":"docx","order_by":3,"title":"","display":"","copyAsset":false,"role":"supplement","size":1812788,"visible":true,"origin":"","legend":"","description":"","filename":"ExtendedDataFigureandTables.docx","url":"https://assets-eu.researchsquare.com/files/rs-7083240/v1/5a9156a7df0a4ba92340b97a.docx"}],"financialInterests":"There is \u003cb\u003eNO\u003c/b\u003e Competing Interest.","formattedTitle":"Immune-Metabolic Programs Drive Disease Trajectories in Paediatric Long COVID","fulltext":[{"header":"Introduction","content":"\u003cp\u003eThe clinical course of SARS-CoV-2 infection in children is typically benign, with many remaining asymptomatic, reflecting robust counter-regulatory immune mechanisms\u003csup\u003e\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e,\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e\u003c/sup\u003e. Nevertheless, approximately 1\u0026ndash;3% of paediatric cases develop long COVID (LC), or post-COVID condition, a debilitating syndrome persisting for at least three months post-infection without an alternative explanation\u003csup\u003e\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e,\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e\u003c/sup\u003e. Symptoms span over 200 domains, including fatigue, dyspnoea, pain, cognitive impairment, and post-exertional malaise\u003csup\u003e\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e,\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e,\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e,\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e\u003c/sup\u003e. In only 80% of affected children, these persistent and debilitating symptoms resolve within one year \u003csup\u003e\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e\u003c/sup\u003e. Autoimmune diseases were found to occur with greater frequency\u003csup\u003e\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e\u003c/sup\u003e. Risk factors for LC were female sex, \u0026gt;12 years, being infected with original variants, and comorbidities. So far, paediatric LC has no pathognomonic findings or diagnostic tests available.\u003c/p\u003e\u003cp\u003eGiven that immune differences between children and adults become more similar after early childhood and cumulative exposures in adults complicate interpretation, paediatric LC offers a unique insight into core disease mechanisms\u003csup\u003e\u003cspan additionalcitationids=\"CR12 CR13 CR14\" citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e\u003c/sup\u003e. So far, adult LC has been associated with diverse pathologies, including neuroinflammation\u003csup\u003e\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e\u003c/sup\u003e, endotheliopathy\u003csup\u003e\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e\u003c/sup\u003e, complement activation\u003csup\u003e\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e,\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e\u003c/sup\u003e, reactivation of EBV\u003csup\u003e\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e\u003c/sup\u003e, metabolic dysfunction\u003csup\u003e\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e,\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e\u003c/sup\u003e, persistent inflammation and autoimmunity\u003csup\u003e\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e,\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e\u003c/sup\u003e, whilst an association with differential SARS-CoV-2 antibody responses is unlikely\u003csup\u003e\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e\u003c/sup\u003e. These diverse manifestations suggest the existence of distinct LC endotypes rather than a single pathway. Whilst the underlying pathophysiological processes driving LC in children remain less well understood \u003csup\u003e\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e,\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e\u003c/sup\u003e, the initial evidence appears to suggest possible abnormalities in immune responses, cellular metabolism and endothelial function\u003csup\u003e\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e\u003c/sup\u003e. Endotype-specific understanding is essential to enable precision interventions and improve clinical trial outcomes\u003csup\u003e\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e\u003cp\u003eHere, we hypothesize that paediatric LC comprises distinct immuno-metabolic endotypes that emerge from dysregulated immune responses and disrupted homeostasis. We aim to identify protective and pathogenic mechanisms that drive disease persistence and define biologically distinct subgroups that can inform targeted clinical trials.\u003c/p\u003e"},{"header":"Results","content":"\u003cp\u003eThe study analysed 74 children with active long COVID (LC) symptoms following SARS-CoV-2 infection, up to 166 weeks post primary SARS-CoV-2 infection, alongside 27 controls (Control) (Fig. 1; Extended Data Table 1a,b). The control group (n = 27) included both healthy (n = 14) and well-managed children with cystic fibrosis (CF)(n = 13), as they are frequently\u0026nbsp;exposed to respiratory infections\u0026nbsp;despite stable treatment and similarly affected by SARS-CoV-2 infection\u003csup\u003e26\u003c/sup\u003e. This ensures that the immune changes observed in long COVID are not driven by pre-existing inflammation or infection frequency, but rather reflect LC disease-specific mechanisms. \u0026nbsp;As no significant proinflammatory cytokines IL-6 (Mann-Whitney U-test p = 0.662; mean 20.6/21.4; SD 5.3/5.4) or TNF\u0026alpha; (p = 0.923; mean 97/144; SD 61/57) differences were found between both control subgroups, they are considered a single control group. A follow-up assessment (LC visit 2) of LC was performed 3-6 months later. Symptom severity was assessed at each appointment using the following questionnaires: SF-12 Physical Component Summary (PCS), SF-12 Mental Component Summary (MCS), the acute Bell Disability Scale, the KIDSCREEN questionnaire, the Generalized Anxiety Disorder 7-item scale (GAD-7), Patient Health Questionnaire 9-item (PHQ-9), Short Fatigue Severity Scale (FSS), Parent-Reported Symptoms (PRS), Patient-Reported Symptoms (PtRS) and Children\u0026rsquo;s Sleep Habits Questionnaire (CSHQ); strength was assessed via a hand-grip-test and stamina was assessed via a sit-to-stand test for children (Fig. 1, Extended Data Fig.1)\u003csup\u003e27\u003c/sup\u003e. As indicators of viral response and inflammation, serum concentrations of IFN\u0026alpha; were measured, revealing elevated IFN\u0026alpha; levels in LC patients during the first year of disease compared to controls, while complement system components C3, C4, and CH50 remained within the normal range (Fig.1d, data not shown). This result contrasts with a study of adult LC\u003csup\u003e19\u003c/sup\u003e, and may be explained by the fact that our study looks at later time points of the disease and that paediatric patients tend to have less pre-exposure pathogens and antigens, resulting in less antigenic sins and innate memory\u003csup\u003e13,14,28\u003c/sup\u003e.\u003c/p\u003e\n\u003cp\u003ePatients within 1 yr active LC showed no clinical improvement in both physical and mental symptoms (Fig. 1c,e, Details of LMM analysis: Extended Data Fig. 1a,b; Extended Data Table 1b). In addition, no improvement at the group level of LC was detected up to 3 years with individual cases showing slight clinical improvement or decline (Fig. 1e, red line, Extended Data Fig. 1a,b; Extended Data Table 1b). This indicates that different pathological mechanisms contribute to disease severity and progression.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eSince LC patients often experience fatigue and shortness of breath, cardiac and pulmonary involvement were assessed (Fig.1c, Fig.2a,b)\u003csup\u003e9,29\u003c/sup\u003e. To address cardiac involvement, all paediatric LC participants (n = 74) underwent a 12-lead electrocardiogram (ECG) and transthoracic echocardiography upon presentation. Electrocardiographic analysis revealed supraventricular extrasystoles in one case. Echocardiography identified abnormalities in three patients (4.1%): one case of mild mitral regurgitation; one case of a bicuspid aortic valve with a coronary artery fistula; and one case of mild diastolic dysfunction, which resolved at follow-up. Paediatric cardiologists judged these to be incidental. To address pulmonary involvement, we initially clustered these individuals based on the mean of their FEV1 Z-scores across multiple visits to capture differences in lung function (Fig. 2b). FEV1 Z-scores were selected because they provide a standardized measure of lung function relative to population norms, enabling the detection of impairments in respiratory capacity. We used k-means clustering to identify three distinct groups of paediatric LC patients with significantly different levels of lung function capacity (Fig. 2b). However, clusters were not associated with anti-SARS-CoV-2 IgG levels or clinical disease scores, suggesting that heterogeneous immunological mechanisms may underlie active paediatric LC. To explore potential immunological drivers of differences in lung function, we selected biomarkers (IL-13, IL-33) based on their known roles in lung-related immune responses and their critical involvement in SARS-CoV-2 immunity \u003csup\u003e30\u003c/sup\u003e (Fig. 2c). We also included IL-6, a key pro-inflammatory cytokine, due to its role in both acute and chronic inflammatory processes. While IL-6 levels were not elevated systemically during the first year of LC, significant increases of IL-13 and IL-33 were observed (Bonferroni-Holm corrected p \u0026lt; 0.05). We then applied a linear mixed-effects model (LMM), incorporating IL-6, IL-13, IL-33, time since infection, and percentile BMI to assess their combined influence on lung function (Fig. 2d; Extended Data Table 2.1a). Due to collinearity between IL-13 and IL-33 (r \u0026gt; 0.7), only IL-13 was included in the final model. The final model fit yielded R\u0026sup2;\u003csub\u003eMarg\u003c/sub\u003e = 0.162, indicating that the fixed effects (IL-6, IL-13) explained approximately 16% of the variance in FEV1 Z-scores. The R\u0026sup2;\u003csub\u003eCond\u003c/sub\u003e of 0.72 reflected the additional variance accounted for when random effects were included, highlighting the substantial influence of individual variability. Neither time since infection, percentile BMI nor sex improved the model fit (Fig. 2d; Extended Data Table 2.1b). Notably, our data reveal that better lung function (as reflected by higher FEV1 Z-scores) was associated with elevated systemic IL-13, suggesting a protective tissue repair response operating in parallel with pro-inflammatory mechanisms\u003csup\u003e30\u003c/sup\u003e. This may reflect IL-13\u0026apos;s involvement in epithelial repair and immune responses following viral infections\u003csup\u003e31\u003c/sup\u003e. These findings underscore the importance of considering subclinical immunological alterations in paediatric LC, even when lung function remains within normal limits, as these changes may have long-term implications for respiratory health\u003csup\u003e32\u003c/sup\u003e.\u003c/p\u003e\n\u003cp\u003eTo investigate the systemic, immune-mediated nature of LC, circulating serum cytokines were assessed within the first year post SARS-CoV-2 infection (Fig.2e-h, Extended Data Fig.2, Extended Data Table 2.2 for multiple testing). The choice of a one-year observation period is based on evidence that long COVID shows spontaneous remission in the majority of cases within 52 weeks; this timeframe is commonly used by others, ensuring comparability, and aligns with the waning of SARS-CoV-2 immunity, such as antibody levels\u003csup\u003e24,29\u003c/sup\u003e. Further analysis of Th2-related cytokines, including IL-4, IL-5, IL-10, and IL-9, revealed no increase compared to controls (Extended Data Fig. 2a). To identify distinct paediatric LC subtypes based on cytokine profiles, we further examined Th1-like (Extended Data Fig. 2b), Th17/22-like (Extended Data Fig. 2d), innate-like and regulatory-like cytokines (Fig. 2d, Fig. 2f,g; Extended Data Fig. 2b, Extended Data Table 2.2). While Th1-related cytokines remained low in LC, this resulted in a significantly elevated Th2/Th1 ratio (Fig. 2e). Among the Th17/22-associated cytokines, IL-12p40 was significantly (corrected p = 0.01) increased, but IL-23 and IL-12p70 were not, whilst the innate-like cytokines IL-1\u0026alpha; and IL-1\u0026beta; were also significantly elevated, suggesting ongoing innate-driven immune activation (Extended Data Fig. 2b). These findings highlight a complex immunological dysregulation in LC, characterised by distinct cytokine patterns within the first year of the disease.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAutoantibodies (aAb) and Organ Injury in Children with Active Long COVID\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eUnder the assumption that long COVID manifests as distinct autoimmune-driven endotypes, we analysed disease-associated autoantibodies (aAb) and, separately, serum neurofilament light chain (NfL) as surrogate marker for central nervous system (CNS) involvement and neuro-axonal injury (Fig. 2i,j, Extended Data \u0026nbsp;Fig. 3.2a). The aAb were surrogate markers for vasculitis, antiphospholipid syndrome, autoimmune connective tissue and coeliac disease. NfL concentrations were converted to age-adjusted z-scores with the Basel paediatric reference dataset and reported as percentiles (P) relative to healthy children\u003csup\u003e33\u003c/sup\u003e. A Wilcoxon signed-rank test against the reference median P\u0026nbsp;50 showed a significant upward shift (W= 1744, n= 73, p=\u0026nbsp;0.031; effect size r=\u0026nbsp;0.25). Clinically, only one child (1.4\u0026nbsp;%) exceeded the upper reference boundary (P\u0026nbsp;97.5) and 6/73 (8.2\u0026nbsp;%) crossed P\u0026nbsp;90. A Bell score of 40\u0026mdash;denoted as severe disability with no more than 3-4 h of light activity per day and a \u0026gt; 50% loss of normal functional capacity\u0026mdash;was used as the severe long COVID stratum (Fig.\u0026nbsp;2i)\u003csup\u003e33\u003c/sup\u003e. Children at or above this cut-off displayed higher NfL percentiles, clustering near the upper tail of the normative distribution (p= 0.003). Pearson\u0026acute;s (and Spearman) correlation showed a significant invers correlation r=\u0026nbsp;-0.33, p=\u0026nbsp;0.012 (Extended Data Fig.\u0026nbsp;3.2a). Of note, NfL percentiles were unrelated to the interval since acute SARS-CoV-2 infection (p=\u0026nbsp;0.48; Extended Data Fig.\u0026nbsp;3.2a). Collectively, the data indicate ongoing neuro-axonal injury in affected children.\u003c/p\u003e\n\u003cp\u003eIn parallel, we assessed functional aAbs against G-protein-coupled receptors (GPCR-aAb) targeting \u0026beta;1- and \u0026beta;2-adrenergic as well as M3- and M4-muscarinic acetylcholine receptors, to capture potential neuro-autonomic involvement. These aAbs did not correlate with time after SARS-CoV-2 infection, systemic cytokine levels, or disease severity (Extended Data Table 2.1c). Furthermore, there was no increase in LC patients with a disease duration of less than 1 year compared to those with a disease duration of more than 1 year (Extended Data Fig. 3.2b). Next, we analysed aAb as surrogate markers for vasculitis (anti-proteinase 3 (PR3) and anti-myeloperoxidase (MPO) aAb), antiphospholipid syndrome (anti-cardiolipin aAb, anti-\u0026beta;2-glycoprotein I aAb (anti-\u0026beta;2GPI), autoimmune connective tissue disease (anti-cyclic citrullinated peptide (CCP) aAb) and coeliac disease (anti-transglutaminase (TransG) Fig.2j). None of them were elevated compared to controls and remained within the normal range during the first year. To evaluate whether enhanced autoimmunity is generally present in patients with active long COVID, we assessed the prevalence of aAb positivity across individuals (Fig.2k). Compared to the control group, no significant increase in aAb positivity was observed, indicating that active paediatric long COVID is not driven by aAb-mediated organ injury.\u003c/p\u003e\n\u003cp\u003eTo explore whether isolated anti-DFS70 reactivity might serve as a benign autoreactivity signature, we assayed anti-DFS70 antibodies despite the absence of a positive ANA screen (Fig. 2l).\u0026nbsp;Surprisingly, none of the children in the control group tested positive. In contrast, 11% of patients with active LC tested positive, and those who were anti-DFS70-positive remained so at a second follow-up visit months later. Next, we grouped patients according to their anti-DFS70 status (Fig.2 m,n, Extended Data Fig.3.2 c,d, Extended Data Table 2.3 B-D for multiple testing,). Anti-DFS70-negative LC patients displayed a higher frequency and activity of vWF, and elevated factor VIII levels. Other factors, including antithrombin III, fibrinogen, D-dimer, aPTT, protein C, free protein S, and complement components C3 and C4 remained unchanged; notably, EBV-related antibodies were only observed in anti-DFS70-negative LC patients (Fig. 2n, Extended Data Table 2.3C, Extended Data Table 2.3C). Thus, we provide clear evidence of alterations to clotting factors in a subgroup of paediatric LC patients. Although anti-DFS70 antibodies are classically linked to healthy individuals and are not considered markers of systemic autoimmunity, their presence in this cohort may signify a unique, non-pathogenic immunological response\u003csup\u003e34\u003c/sup\u003e. Notably, Epstein-Barr virus seropositivity was confined to the anti-DFS70-negative subgroup, suggesting the existence of immunologically distinct Long-COVID phenotypes that merit further investigation.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eActive Paediatric Long COVID beyond the First Year\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eSubsequently, we monitored systemic parameters after one year of LC to delineate the evolving immunopathological phases (Fig. 3a-e, Extended Data Fig. 3.1, Extended Data Table 3.1 for multiple comparisons). Of note, the previously elevated IFN\u0026alpha; levels declined, indicating a waning antiviral response (Fig. 3a). Re-evaluation of the earlier Th2-skewed state (Fig. 3b, Extended Data Fig. 3.1a) revealed a marked reduction in IL-4, IL-13, and the type 2-alarmin IL-33, accompanied by a two-fold rise of IL-9, pointing towards a shift towards a Th9-like immunophenotype. Innate-like cytokines GM-CSF (p = 0.037) trended further upward, with a notable and significant increase in IL-1\u0026beta; (Fig. 3c). Within the Th17 axis (Fig. 3d, Extended Data Fig. 3.1d), IL-23 and IL-12p40 increased, while the Th1 axis remained unaltered (Extended Data Fig. 3.1d). IL-12p70 and IFN\u0026gamma; remained low, altogether suggesting a transition towards a Th17-driven state. Taken together (Fig. 3e), these findings outline a dynamic immunological trajectory in paediatric LC, marked by an early anti-viral and Th2-skewed response within the first year. Over time, this shifts towards a Th9/Th17-dominated, innate-oriented endotype, reflecting the progressive immunopathological remodelling that underlies disease chronicity.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eNext, we employed a linear mixed-effects model (random intercept Patient ID) to analyse factors associated with long COVID, using the Bell score as the dependent variable (Fig. 3f, Extended Data Fig. 3.2a). Covariates were selected based on their biological relevance to long COVID pathophysiology, including markers of viral (re)activation (anti-EBV EBNA IgG), haematological abnormalities in mean corpuscular haemoglobin concentration (MCHC) \u003csup\u003e35\u003c/sup\u003e, and IL-12p40, as it is the most frequently expressed cytokine and has the greatest increase in value since infection (Fig.\u0026nbsp;2h, 3d), along with basophil granulocyte counts. R\u0026sup2;\u003csub\u003eMarg\u003c/sub\u003e (0.212) indicated that fixed effects explained 21% of the variance. All covariates showed significant associations with the Bell score. Notably, anti-EBV EBNA IgG and MCHC levels decreased as patients improved, while basophil granulocyte counts and IL-12p40 levels increased. Analysis including time post infection did not improve the model (Extended Data Table 3.2b).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eMolecular Landscape of EBV-Associated Paediatric Long COVID\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAs EBV exposure was found to be significantly different and linked to haematopoietic alterations, we assessed its potential contribution of EBV exposure to paediatric LC endotypes. We applied linear mixed-effects models (LMMs), including EBV EBNA IgG status (cut-off 50\u0026nbsp;U\u0026nbsp;ml\u003csup\u003e-1\u003c/sup\u003e), BMI percentile, age, sex and time since SARS-CoV-2 infection as covariates. Repeated measures were clustered by patient ID (Fig. 4, left, Table 1 for multiple comparison, Extended Data Table 4). Initially, complement components and anti-SARS-CoV-2 spike protein antibodies were applied, but did not show any \u0026nbsp;dependence on EBV exposure (Table 1a, Extended Data Table 3a). \u0026nbsp;At the molecular level, EBV exposure was associated with a distinct innate immune signature: IL-1\u0026alpha; and IL-15 levels were significantly elevated (Fig. 4, Table 1b, Extended Data Table 4). We hypothesised the existence of an additional Th17/Th22 polarised endotype, which was evident in EBV-experienced patients and characterised by significant upregulation of IL-12p40 and IL-22 (Table 1c; Extended Data Table 4). This was paralleled by increased levels of the regulatory cytokines IL-10, IL-11 and IL-27 (Table 1d; Extended Data Table 4), suggesting a concomitant immunoregulatory counter-response. Importantly, EBV exposure was also associated with aAb titres against Prothrombin (Table 1e; Extended Data Table 4), supporting the presence of an autoimmune component within this subgroup. Of note, EBV-Ab response may even cross-react with other aAb\u003csup\u003e36\u003c/sup\u003e. Innate activation was further corroborated by significantly increased neutrophil counts (Table 1f; Extended Data Table 4), indicating sustained myeloid involvement. As chronic EBV responses are known to increase the risk of depression in female adolescents\u003csup\u003e37\u003c/sup\u003e, and depression and anxiety negatively impact recovery in children with chronic fatigue syndrome\u003csup\u003e38\u003c/sup\u003e, the model was applied to mental health scores (Table 1g; Extended Data Table 4). EBV-seropositive patients did not exhibit significantly worse clinical mental health scores than EBV-na\u0026iuml;ve ones. To summarize the immune and clinical landscape associated with EBV exposure in paediatric LC, we assembled a composite visualization (Fig. 4, left) integrating all significant associations in the above described model (Table 1; Extended Data Table 4), which demonstrates a pronounced inflammatory landscape in the subgroup of EBV-experienced paediatric LC patients.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eExploitation of EBV EBNA IgG-negative or low status observed in about half of our paediatric LC cohort, we applied the model from Fig. 3f to this subgroup, excluding EBV status as a covariate (Extended Data Table 5a). The model fit for the Bell score remained highly significant (p\u0026nbsp;\u0026lt;\u0026nbsp;0.001) with an improved R\u0026sup2;Marg of 0.286. IL-12p40 levels and basophil counts were the strongest positive predictors of Bell Score, while lower mean corpuscular haemoglobin concentration (MCHC) correlated with higher Bell scores, indicating better functional status. To probe potential metabolic determinants, we included vitamin B1 (thiamine), based on increasing evidence for mitochondrial dysfunction in LC and on its essential role in mitochondrial energy metabolism\u003csup\u003e20,21\u003c/sup\u003e. This addition significantly improved the model fit for the Bell score further (R\u0026sup2;\u003csub\u003eMarg \u003c/sub\u003e= 0.314)(Fig. 4, right; Extended Data Table 5b), supporting an immunometabolic contribution to clinical severity. This result delineates a distinct EBV-negative LC endotype explainable within an immunometabolic framework. Besides the absence of prior EBV exposure, our findings highlight several protective factors in LC, including IL-13, IL-12p40, basophilic granulocytes, vitamin B1, and autoantigen DFS70-specific aAb.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eVerification of Subtypes of Paediatric Long COVID\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe identified three features by which LC subtypes can be categorised: The first relates to the duration post-infection, where LC in the first year is characterized by Th2- and viral-associated cytokines, and LC persisting for 2\u0026ndash;3 years is characterized by Th17-like and innate-like systemic cytokines. A second relates to autoimmunity, as anti-DFS70-positivity with an anti-DFS70-positive pattern exhibited markedly fewer coagulation abnormalities. The third involves EBV serostatus which was linked to a pro-inflammatory cytokine profile and granulocytic dysregulation. To further explore these separately immuno-clinical subgroups in paediatric long COVID, we systematically analysed a panel of 43 haematopoietic, coagulation, electrolyte, and vitamin-related biomarkers (Table 2, Extended Data Table 6a,b). Using LMMs for each parameter, we tested for associations with subgroup assignment, EBV serostatus, DFS70 aAb status, and disease duration (first year vs. later), accounting for repeated measures through random intercepts. The 43 parameters assessed included thyroid-stimulating hormone (TSH), activated partial thromboplastin time (aPTT), lipoprotein(a) [LP(a)], electrolytes (Na⁺, K⁺, Cl⁻, Ca\u0026sup2;⁺, bicarbonate), vitamins (B1, B6, B12, D, folic acid), metabolic markers (glucose, lactate, creatinine, eGFR), and inflammation/coagulation markers (CRP, D-dimer, fibrinogen, Factor VIII, protein S). After correcting for multiple testing, three metabolic markers were found to be significantly associated with the defined subgroups (each p \u0026lt; 0.001, corrected p = 0.043): TSH, aPTT and LP(a)(Table 2). LP(a) and TSH levels increased over time, whereas aPTT decreased (Extended Data Table 6b). Variation in TSH across subgroups suggests endocrine dysregulation linked to specific immunological profiles. APTT\u0026acute;s association to all subgroups points toward distinct coagulation patterns of each subgroup and LP(a) levels were associated with both disease duration and DFS70 aAb status, likely reflecting inflammation and indicating evolving vascular risk\u003csup\u003e39,40\u003c/sup\u003e. These biomarkers exhibited non-overlapping association patterns, supporting the concept of pathophysiological divergence. Together, these findings provide strong evidence for the existence of biologically meaningful endotypes in paediatric long COVID.\u003c/p\u003e"},{"header":"Discussion","content":"\u003cp\u003eOur cohort study reveals persistent and temporal immune and metabolic dysregulation as a hallmark of paediatric long COVID (LC). Although paediatric and adult LC share symptomatic similarities, our findings demonstrate that distinct molecular and cellular consequences of SARS-CoV-2 defence appear in paediatric LC. While paediatric LC exhibits many symptoms similar to that of adult LC, our data show that it does not demonstrate abnormalities relating to heart or endothelial\u003csup\u003e\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e\u003c/sup\u003e system function; vitamin or trace element values; gastrointestinal tract parameters; systemic autoimmunity\u003csup\u003e\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e,\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e\u003c/sup\u003e; chronic complement activation\u003csup\u003e\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e,\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e\u003c/sup\u003e; or reactivation of EBV IgM\u003csup\u003e\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e\u003c/sup\u003e. This may reflect the greater regenerative capacity of children, the lower cumulative burden of lifestyle-related factors, and a less established and rigid immunological memory compared to adults. However, similar findings include persistent activity of innate immunity and signs of chronic inflammation\u003csup\u003e\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e,\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e\u003c/sup\u003e. In adult myalgic encephalomyelitis or chronic fatigue syndrome (ME/CFS), elevated cytokine levels have been reported as well as autoantibodies\u003csup\u003e\u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e41\u003c/span\u003e\u003c/sup\u003e, but at least their cytokine signatures closely resemble those observed in EBV-positive cohorts\u003csup\u003e\u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e42\u003c/span\u003e\u003c/sup\u003e. Along the same lines, cytokines and metabolic features potentially associated with reduced fatigue as we demonstrate here, were not identified, presumably due to the near-universal prevalence of latent EBV infection in adults, which limits differential analysis.\u003c/p\u003e\u003cp\u003eCytokines known to be involved in the antiviral and anti\u0026ndash;SARS-CoV-2 response, such as IL-13, IL-33 and IFNα, were markedly elevated during the first year \u003csup\u003e\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e,\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e,\u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e43\u003c/span\u003e\u003c/sup\u003e and subsequently decreased during the second and third years. This \u0026ldquo;early\u0026rdquo; IFNα rise is consistent with adult long COVID\u003csup\u003e\u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e43\u003c/span\u003e\u003c/sup\u003e. Comparing the responses of adult and paediatric LC, innate cytokines tend to be more prominent in paediatric SARS-CoV-2 responses\u003csup\u003e\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e\u003c/sup\u003e. Elevated innate cytokines in LC such as those reported here, may indeed not only drive inflammation, but likely also fuel LC symptoms associated with depressive moods\u003csup\u003e\u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e44\u003c/span\u003e\u003c/sup\u003e. Although both adult and paediatric LC show persistent deregulation of cytokines, adult LC shows alternative cytokine involvement compared to what we have reported here for paediatric LC\u003csup\u003e\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e\u003c/sup\u003e. This may be due to differences in the status of the immune system, as well as differences in life-style, antigenic provocation, autoimmunity and exhausted responses.\u003c/p\u003e\u003cp\u003eCertain features of paediatric LC\u0026mdash;such as Epstein-Barr virus (EBV) seropositivity and innate immune activation\u0026mdash;are shared with multisystem inflammatory syndrome in children (MIS-C), a distinct post-infectious complication of SARS-CoV-2\u003csup\u003e45\u003c/sup\u003e. However, the immunometabolic profile observed in EBV-seropositive LC patients suggests a protracted and mechanistically distinct disease process, in contrast to the acute, hyperinflammatory phenotype of MIS-C, which typically manifests 2 to 6 weeks after infection\u003csup\u003e\u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e\u003c/sup\u003e. In postural orthostatic tachycardia syndrome (POTS), which can also occur post SARS-CoV-2 infection, GGPR autoantibodies have been similarly elevated as reported here, and levels did not differ from those in healthy controls and no correlation with disease severity \u003csup\u003e\u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e46\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e\u003cp\u003eCritically, we identified several immune markers and clinical features that distinguish children with a more favourable disease course from those with a more protracted trajectory. Linking a decrease in MCHC to an improvement in the Bell score suggests a link to microcirculatory dynamics. Although MCHC cannot be considered a standalone indicator of blood viscosity, its reduction may well reflect improved erythrocyte hydration or membrane stability, which could enhance tissue oxygen delivery\u003csup\u003e\u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e,4748\u003c/sup\u003e. As reduced deformability hampers adult LC erythrocytes, restoring paediatric red-cell rheology may likewise accelerate recovery\u003csup\u003e\u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e\u003c/sup\u003e. IL-12p40, mainly known to be a subunit shared by IL-12 and IL-23, plays a key role in T-cell responses, but can also bind monomerically to the IL12 receptor as an antagonist\u003csup\u003e\u003cspan citationid=\"CR49\" class=\"CitationRef\"\u003e49\u003c/span\u003e\u003c/sup\u003e. As its serum concentration is more than 100-fold higher than IL-12 or IL-23 (see Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e2\u003c/span\u003eh; \u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e3\u003c/span\u003ed; Extended Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e2\u003c/span\u003eb; 3.1b), it may therefore contribute to balancing immune activity rather than fuelling persistent inflammation. The presence of these factors in LC patients associated with milder disease could indicate a shift in those patients towards resolution of inflammation rather than its persistence. In addition, absence of neuronal involvement might be identifiable as only children with a Bell scores of \u0026gt;\u0026thinsp;40, consistently exhibit NfL levels in the lower normative range, which could provide an additional surrogate marker for a favourable disease course. Severe cases show NfL exclusively in the upper tail of the normal distribution. As this is not the case in MIS-C, it likely reflects a distinct molecular aethiopathology\u003csup\u003e\u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e50\u003c/span\u003e\u003c/sup\u003e. This reinforces the idea that NfL is a sensitive marker of neuronal involvement in paediatric LC, however, it remains to be determined whether such elevations predict clinical progression or long-term neurological outcomes.\u003c/p\u003e\u003cp\u003eThe immune dynamics are interwoven with perturbations in other systems, reflected in parallel changes in haematological indices\u0026mdash;including evidence of coagulation imbalance\u0026mdash;and in metabolic and endocrine markers over time. This highlights the multisystemic nature of paediatric LC and suggests that these non-overlapping metabolic profiles may represent either causally distinct endotypes or downstream consequences of broader immune dysregulation. Regardless of their origin, these features contribute to disease heterogeneity and underscore the need for tailored therapeutic approaches, potentially explaining the failure of previous experimental interventions that targeted LC as a single disease entity.\u003c/p\u003e\u003cp\u003eTaken together, these findings demonstrate that the identified endotypes are distinct not only clinically and immunologically, but also in terms of their metabolic and vascular characteristics. This layered heterogeneity reinforces the need for a stratified approach to long COVID diagnostics and management, and supports the utility of biomarker-guided sub-phenotyping in future precision medicine frameworks.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eAuthor Contribution\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eMCBW, KV, IH, PJ, SW, DR, AR, JM, ML, LN, PH, and EU performed or supervised experiments and generated and analysed data. EU, DV, MP, CA, HP, LN evaluated and recruited patients and/or controls. SW and MCBW verified the use and visualization of appropriate statistical methods for modelling. MCBW wrote the original draft with the help of all other co-authors and is the lead corresponding author. MCBW, KV, IH, PJ, JK, DV performed computational analysis of data. DV and MCBW conceptualized the project. DV directed the project and was responsible for funding acquisition.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis study was conducted with resources provided by the Multicentre long COVID registry (MLC-R), by the BMBF (to DV: LongCOCID 01EP2101, subproject funding to MCBW 01EP2101C), DFG Br1860/18 (MCBW), and the Ministry of Science, Technology and Environment of Saxony-Anhalt (SarsImmunGender I-196 to MCBW). Open access funding was provided by DFG and Otto-von-Guericke University Magdeburg, Germany.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eDisclosure Statement\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAuthors declare no competing interests.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eData availability\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eData will be made available upon request.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eS\u0026oslash;rensen, A. I. V.\u003cem\u003e et al. \u003c/em\u003eA nationwide questionnaire study of post-acute symptoms and health problems after SARS-CoV-2 infection in Denmark. \u003cem\u003eNature communications \u003c/em\u003e\u003cstrong\u003e13, \u003c/strong\u003e4213; 10.1038/s41467-022-31897-x (2022).\u003c/li\u003e\n\u003cli\u003eMorello, R., Martino, L. \u0026amp; Buonsenso, D. Diagnosis and management of post-COVID (Long COVID) in children: a moving target. \u003cem\u003eCurrent opinion in pediatrics \u003c/em\u003e\u003cstrong\u003e35, \u003c/strong\u003e184\u0026ndash;192; 10.1097/MOP.0000000000001221 (2023).\u003c/li\u003e\n\u003cli\u003eDavis, H. E., McCorkell, L., Vogel, J. M. \u0026amp; Topol, E. J. Long COVID: major findings, mechanisms and recommendations. \u003cem\u003eNature reviews. \u003c/em\u003e\u003cstrong\u003e21, \u003c/strong\u003e133\u0026ndash;146; 10.1038/s41579-022-00846-2 (2023).\u003c/li\u003e\n\u003cli\u003eGreenhalgh, T., Sivan, M., Perlowski, A. \u0026amp; Nikolich, J. Ž. Long COVID: a clinical update. \u003cem\u003eLancet \u003c/em\u003e\u003cstrong\u003e404, \u003c/strong\u003e707\u0026ndash;724; 10.1016/S0140-6736(24)01136-X (2024).\u003c/li\u003e\n\u003cli\u003eChou, J., Thomas, P. G. \u0026amp; Randolph, A. G. Immunology of SARS-CoV-2 infection in children. \u003cem\u003eNature immunology \u003c/em\u003e\u003cstrong\u003e23, \u003c/strong\u003e177\u0026ndash;185; 10.1038/s41590-021-01123-9 (2022).\u003c/li\u003e\n\u003cli\u003eG\u0026ouml;tzinger, F.\u003cem\u003e et al. \u003c/em\u003eCOVID-19 in children and adolescents in Europe: a multinational, multicentre cohort study. \u003cem\u003eThe Lancet. Child \u0026amp; adolescent health \u003c/em\u003e\u003cstrong\u003e4, \u003c/strong\u003e653\u0026ndash;661; 10.1016/S2352-4642(20)30177-2 (2020).\u003c/li\u003e\n\u003cli\u003eStephenson, T.\u003cem\u003e et al. \u003c/em\u003eLong COVID (post-COVID-19 condition) in children: a modified Delphi process. \u003cem\u003eArchives of disease in childhood \u003c/em\u003e\u003cstrong\u003e107, \u003c/strong\u003e674\u0026ndash;680; 10.1136/archdischild-2021-323624 (2022).\u003c/li\u003e\n\u003cli\u003ePeluso, M. J. \u0026amp; Deeks, S. G. Mechanisms of long COVID and the path toward therapeutics. \u003cem\u003eCell \u003c/em\u003e\u003cstrong\u003e187, \u003c/strong\u003e5500\u0026ndash;5529; 10.1016/j.cell.2024.07.054 (2024).\u003c/li\u003e\n\u003cli\u003eRao, S.\u003cem\u003e et al. \u003c/em\u003ePostacute Sequelae of SARS-CoV-2 in Children. \u003cem\u003ePediatrics \u003c/em\u003e\u003cstrong\u003e153\u003c/strong\u003e; 10.1542/peds.2023-062570 (2024).\u003c/li\u003e\n\u003cli\u003eCamporesi, A.\u003cem\u003e et al. \u003c/em\u003eCharacteristics and predictors of Long Covid in children: a 3-year prospective cohort study. \u003cem\u003eEClinicalMedicine \u003c/em\u003e\u003cstrong\u003e76, \u003c/strong\u003e102815; 10.1016/j.eclinm.2024.102815 (2024).\u003c/li\u003e\n\u003cli\u003eKnolle, J.\u003cem\u003e et al. \u003c/em\u003eChildren From the Age of Three Show a Developmental Switch in T-Cell Differentiation. \u003cem\u003eFrontiers in immunology \u003c/em\u003e\u003cstrong\u003e11, \u003c/strong\u003e1640; 10.3389/fimmu.2020.01640 (2020).\u003c/li\u003e\n\u003cli\u003eVogel, K.\u003cem\u003e et al. \u003c/em\u003eBifidobacteria shape antimicrobial T-helper cell responses during infancy and adulthood. \u003cem\u003eNature communications \u003c/em\u003e\u003cstrong\u003e14, \u003c/strong\u003e5943; 10.1038/s41467-023-41630-x (2023).\u003c/li\u003e\n\u003cli\u003eShrock, E.\u003cem\u003e et al. \u003c/em\u003eViral epitope profiling of COVID-19 patients reveals cross-reactivity and correlates of severity. \u003cem\u003eScience \u003c/em\u003e\u003cstrong\u003e370\u003c/strong\u003e; 10.1126/science.abd4250 (2020).\u003c/li\u003e\n\u003cli\u003eFazekas, d. S. G. \u0026amp; Webster, R. G. Disquisitions of Original Antigenic Sin. I. Evidence in man. \u003cem\u003eThe Journal of experimental medicine \u003c/em\u003e\u003cstrong\u003e124, \u003c/strong\u003e331\u0026ndash;345; 10.1084/jem.124.3.331 (1966).\u003c/li\u003e\n\u003cli\u003eZajac, A. J.\u003cem\u003e et al. \u003c/em\u003eViral immune evasion due to persistence of activated T cells without effector function. \u003cem\u003eThe Journal of experimental medicine \u003c/em\u003e\u003cstrong\u003e188, \u003c/strong\u003e2205\u0026ndash;2213; 10.1084/jem.188.12.2205 (1998).\u003c/li\u003e\n\u003cli\u003eGutman, E. G.\u003cem\u003e et al. \u003c/em\u003eLong COVID: plasma levels of neurofilament light chain in mild COVID-19 patients with neurocognitive symptoms. \u003cem\u003eMolecular psychiatry \u003c/em\u003e\u003cstrong\u003e29, \u003c/strong\u003e3106\u0026ndash;3116; 10.1038/s41380-024-02554-0 (2024).\u003c/li\u003e\n\u003cli\u003eVassiliou, A. G.\u003cem\u003e et al. \u003c/em\u003eEndotheliopathy in Acute COVID-19 and Long COVID. \u003cem\u003eInternational journal of molecular sciences \u003c/em\u003e\u003cstrong\u003e24\u003c/strong\u003e; 10.3390/ijms24098237 (2023).\u003c/li\u003e\n\u003cli\u003eCervia-Hasler, C.\u003cem\u003e et al. \u003c/em\u003ePersistent complement dysregulation with signs of thromboinflammation in active Long Covid. \u003cem\u003eScience \u003c/em\u003e\u003cstrong\u003e383, \u003c/strong\u003eeadg7942; 10.1126/science.adg7942 (2024).\u003c/li\u003e\n\u003cli\u003eKlein, J.\u003cem\u003e et al. \u003c/em\u003eDistinguishing features of long COVID identified through immune profiling. \u003cem\u003eNature \u003c/em\u003e\u003cstrong\u003e623, \u003c/strong\u003e139\u0026ndash;148; 10.1038/s41586-023-06651-y (2023).\u003c/li\u003e\n\u003cli\u003eBerezhnoy, G.\u003cem\u003e et al. \u003c/em\u003eMaintained imbalance of triglycerides, apolipoproteins, energy metabolites and cytokines in long-term COVID-19 syndrome patients. \u003cem\u003eFrontiers in immunology \u003c/em\u003e\u003cstrong\u003e14, \u003c/strong\u003e1144224; 10.3389/fimmu.2023.1144224 (2023).\u003c/li\u003e\n\u003cli\u003eMolnar, T.\u003cem\u003e et al. \u003c/em\u003eMitochondrial dysfunction in long COVID: mechanisms, consequences, and potential therapeutic approaches. \u003cem\u003eGeroScience \u003c/em\u003e\u003cstrong\u003e46, \u003c/strong\u003e5267\u0026ndash;5286; 10.1007/s11357-024-01165-5 (2024).\u003c/li\u003e\n\u003cli\u003eRojas, M.\u003cem\u003e et al. \u003c/em\u003eAutoimmunity is a hallmark of post-COVID syndrome. \u003cem\u003eJournal of translational medicine \u003c/em\u003e\u003cstrong\u003e20, \u003c/strong\u003e129; 10.1186/s12967-022-03328-4 (2022).\u003c/li\u003e\n\u003cli\u003eLingel, H.\u003cem\u003e et al. \u003c/em\u003eUnique autoantibody prevalence in long-term recovered SARS-CoV-2-infected individuals. \u003cem\u003eJournal of autoimmunity \u003c/em\u003e\u003cstrong\u003e122, \u003c/strong\u003e102682; 10.1016/j.jaut.2021.102682 (2021).\u003c/li\u003e\n\u003cli\u003eAltmann, D. M.\u003cem\u003e et al. \u003c/em\u003ePersistent symptoms after COVID-19 are not associated with differential SARS-CoV-2 antibody or T cell immunity. \u003cem\u003eNature communications \u003c/em\u003e\u003cstrong\u003e14, \u003c/strong\u003e5139; 10.1038/s41467-023-40460-1 (2023).\u003c/li\u003e\n\u003cli\u003eScheibenbogen, C.\u003cem\u003e et al. \u003c/em\u003eFighting Post-COVID and ME/CFS - development of curative therapies. \u003cem\u003eFrontiers in medicine \u003c/em\u003e\u003cstrong\u003e10, \u003c/strong\u003e1194754; 10.3389/fmed.2023.1194754 (2023).\u003c/li\u003e\n\u003cli\u003eBain, R.\u003cem\u003e et al. \u003c/em\u003eClinical characteristics of SARS-CoV-2 infection in children with cystic fibrosis: An international observational study. \u003cem\u003eJournal of cystic fibrosis \u003c/em\u003e\u003cstrong\u003e20, \u003c/strong\u003e25\u0026ndash;30; 10.1016/j.jcf.2020.11.021 (2021).\u003c/li\u003e\n\u003cli\u003eZeren, M., Gurses, H. N., Denizoglu Kulli, H., Ucgun, H. \u0026amp; Cakir, E. Sit-to-stand test in children with bronchiectasis: Does it measure functional exercise capacity? \u003cem\u003eHeart \u0026amp; \u003c/em\u003e\u003cstrong\u003e49, \u003c/strong\u003e796\u0026ndash;802; 10.1016/j.hrtlng.2020.09.017 (2020).\u003c/li\u003e\n\u003cli\u003eLercher, A.\u003cem\u003e et al. \u003c/em\u003eAntiviral innate immune memory in alveolar macrophages following SARS-CoV-2 infection ameliorates secondary influenza A virus disease. \u003cem\u003eImmunity \u003c/em\u003e\u003cstrong\u003e57, \u003c/strong\u003e2530-2546.e13; 10.1016/j.immuni.2024.08.018 (2024).\u003c/li\u003e\n\u003cli\u003eAtchison, C. J.\u003cem\u003e et al. \u003c/em\u003eLong-term health impacts of COVID-19 among 242,712 adults in England. \u003cem\u003eNature communications \u003c/em\u003e\u003cstrong\u003e14, \u003c/strong\u003e6588; 10.1038/s41467-023-41879-2 (2023).\u003c/li\u003e\n\u003cli\u003eMeltendorf, S.\u003cem\u003e et al. \u003c/em\u003eIL-13 determines specific IgE responses and SARS-CoV-2 immunity after mild COVID-19 and novel mRNA vaccination. \u003cem\u003eEuropean journal of immunology \u003c/em\u003e\u003cstrong\u003e52, \u003c/strong\u003e1972\u0026ndash;1979; 10.1002/eji.202249951 (2022).\u003c/li\u003e\n\u003cli\u003eGieseck, R. L., Wilson, M. S. \u0026amp; Wynn, T. A. Type 2 immunity in tissue repair and fibrosis. \u003cem\u003eNature reviews. Immunology \u003c/em\u003e\u003cstrong\u003e18, \u003c/strong\u003e62\u0026ndash;76; 10.1038/nri.2017.90 (2018).\u003c/li\u003e\n\u003cli\u003eZar, H. J.\u003cem\u003e et al. \u003c/em\u003eEarly-life respiratory syncytial virus disease and long-term respiratory health. \u003cem\u003eThe Lancet. Respiratory medicine \u003c/em\u003e\u003cstrong\u003e12, \u003c/strong\u003e810\u0026ndash;821; 10.1016/S2213-2600(24)00246-7 (2024).\u003c/li\u003e\n\u003cli\u003eAbdelhak, A.\u003cem\u003e et al. \u003c/em\u003eSerum neurofilament light chain reference database for individual application in paediatric care: a retrospective modelling and validation study. \u003cem\u003eThe Lancet. Neurology \u003c/em\u003e\u003cstrong\u003e22, \u003c/strong\u003e826\u0026ndash;833; 10.1016/S1474-4422(23)00210-7 (2023).\u003c/li\u003e\n\u003cli\u003eCarbone, T.\u003cem\u003e et al. \u003c/em\u003ePrevalence and serological profile of anti-DFS70 positive subjects from a routine ANA cohort. \u003cem\u003eScientific reports \u003c/em\u003e\u003cstrong\u003e9, \u003c/strong\u003e2177; 10.1038/s41598-019-38686-5 (2019).\u003c/li\u003e\n\u003cli\u003eLechuga, G. C., Morel, C. M. \u0026amp; De-Simone, S. G. Hematological alterations associated with long COVID-19. \u003cem\u003eFrontiers in physiology \u003c/em\u003e\u003cstrong\u003e14, \u003c/strong\u003e1203472; 10.3389/fphys.2023.1203472 (2023).\u003c/li\u003e\n\u003cli\u003eMaguire, C.\u003cem\u003e et al. \u003c/em\u003eMolecular mimicry as a mechanism of viral immune evasion and autoimmunity. \u003cem\u003eNature communications \u003c/em\u003e\u003cstrong\u003e15, \u003c/strong\u003e9403; 10.1038/s41467-024-53658-8 (2024).\u003c/li\u003e\n\u003cli\u003eFord, J. L. \u0026amp; Stowe, R. P. Depressive symptoms are associated with salivary shedding of Epstein-Barr virus in female adolescents: The role of sex differences. \u003cem\u003ePsychoneuroendocrinology \u003c/em\u003e\u003cstrong\u003e86, \u003c/strong\u003e128\u0026ndash;133; 10.1016/j.psyneuen.2017.09.009 (2017).\u003c/li\u003e\n\u003cli\u003eRowe, K. S. Long Term Follow up of Young People With Chronic Fatigue Syndrome Attending a Pediatric Outpatient Service. \u003cem\u003eFrontiers in pediatrics \u003c/em\u003e\u003cstrong\u003e7, \u003c/strong\u003e21; 10.3389/fped.2019.00021 (2019).\u003c/li\u003e\n\u003cli\u003eBae, G.\u003cem\u003e et al. \u003c/em\u003eLongitudinal lipoprotein and inflammatory mediators analysis uncover persisting inflammation and hyperlipidemia following SARS-CoV-2 infection in long COVID-19. \u003cem\u003eMetabolomics : Official journal of the Metabolomic Society \u003c/em\u003e\u003cstrong\u003e21, \u003c/strong\u003e65; 10.1007/s11306-025-02262-y (2025).\u003c/li\u003e\n\u003cli\u003eStahl, D.\u003cem\u003e et al. \u003c/em\u003eHigh increase in levels of lipoprotein(a) in plasma of patients with rheumatoid arthritis after COVID-19. \u003cem\u003eArthritis \u0026amp; rheumatology (Hoboken, N.J.) \u003c/em\u003e\u003cstrong\u003e75, \u003c/strong\u003e329\u0026ndash;331; 10.1002/art.42374 (2023).\u003c/li\u003e\n\u003cli\u003eLoebel, M.\u003cem\u003e et al. \u003c/em\u003eAntibodies to \u0026beta; adrenergic and muscarinic cholinergic receptors in patients with Chronic Fatigue Syndrome. \u003cem\u003eBrain, behavior, and immunity \u003c/em\u003e\u003cstrong\u003e52, \u003c/strong\u003e32\u0026ndash;39; 10.1016/j.bbi.2015.09.013 (2016).\u003c/li\u003e\n\u003cli\u003eMontoya, J. G.\u003cem\u003e et al. \u003c/em\u003eCytokine signature associated with disease severity in chronic fatigue syndrome patients. \u003cem\u003eProceedings of the National Academy of Sciences of the United States of America \u003c/em\u003e\u003cstrong\u003e114, \u003c/strong\u003eE7150-E7158; 10.1073/pnas.1710519114 (2017).\u003c/li\u003e\n\u003cli\u003eQueiroz, M. A. F.\u003cem\u003e et al. \u003c/em\u003eSevere COVID-19 and long COVID are associated with high expression of STING, cGAS and IFN-\u0026alpha;. \u003cem\u003eScientific reports \u003c/em\u003e\u003cstrong\u003e14, \u003c/strong\u003e4974; 10.1038/s41598-024-55696-0 (2024).\u003c/li\u003e\n\u003cli\u003eDantzer, R., O\u0026rsquo;Connor, J. C., Freund, G. G., Johnson, R. W. \u0026amp; Kelley, K. W. From inflammation to sickness and depression: when the immune system subjugates the brain. \u003cem\u003eNature reviews. Neuroscience \u003c/em\u003e\u003cstrong\u003e9, \u003c/strong\u003e46\u0026ndash;56; 10.1038/nrn2297 (2008).\u003c/li\u003e\n\u003cli\u003eGoetzke, C. C.\u003cem\u003e et al. \u003c/em\u003eTGF\u0026beta; links EBV to multisystem inflammatory syndrome in children. \u003cem\u003eNature \u003c/em\u003e\u003cstrong\u003e640, \u003c/strong\u003e762\u0026ndash;771; 10.1038/s41586-025-08697-6 (2025).\u003c/li\u003e\n\u003cli\u003eHall, J.\u003cem\u003e et al. \u003c/em\u003eDetection of G Protein-Coupled Receptor Autoantibodies in Postural Orthostatic Tachycardia Syndrome Using Standard Methodology. \u003cem\u003eCirculation \u003c/em\u003e\u003cstrong\u003e146, \u003c/strong\u003e613\u0026ndash;622; 10.1161/CIRCULATIONAHA.122.059971 (2022).\u003c/li\u003e\n\u003cli\u003eRenoux, C.\u003cem\u003e et al. \u003c/em\u003eImpact of surface-area-to-volume ratio, internal viscosity and membrane viscoelasticity on red blood cell deformability measured in isotonic condition. \u003cem\u003eScientific reports \u003c/em\u003e\u003cstrong\u003e9, \u003c/strong\u003e6771; 10.1038/s41598-019-43200-y (2019).\u003c/li\u003e\n\u003cli\u003eGeisness, A. C.\u003cem\u003e et al. \u003c/em\u003eIonophore-mediated swelling of erythrocytes as a therapeutic mechanism in sickle cell disease. \u003cem\u003eHaematologica \u003c/em\u003e\u003cstrong\u003e107, \u003c/strong\u003e1438\u0026ndash;1447; 10.3324/haematol.2021.278666 (2022).\u003c/li\u003e\n\u003cli\u003eLing, P.\u003cem\u003e et al. \u003c/em\u003eHuman IL-12 p40 homodimer binds to the IL-12 receptor but does not mediate biologic activity. \u003cem\u003eThe Journal of Immunology \u003c/em\u003e\u003cstrong\u003e154, \u003c/strong\u003e116\u0026ndash;127; 10.4049/jimmunol.154.1.116 (1995).\u003c/li\u003e\n\u003cli\u003eGeis, T.\u003cem\u003e et al. \u003c/em\u003eSerum neurofilament light chain (sNfL) values in a large cross-sectional population of children with asymptomatic to moderate COVID-19. \u003cem\u003eJournal of neurology \u003c/em\u003e\u003cstrong\u003e268, \u003c/strong\u003e3969\u0026ndash;3974; 10.1007/s00415-021-10554-1 (2021).\u003c/li\u003e\n\u003c/ol\u003e"},{"header":"Methods","content":"\u003cp\u003e\u003cstrong\u003eHuman participants\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWritten informed consent was obtained from all participants and/or their legal representatives in accordance with the Declaration of Helsinki. The study was approved by the ethics committees of the University Hospital Jena (2022-2614_1-BO) and the Otto-von-Guericke University Magdeburg (164-18). Although the study was observational and non-interventional in design, it is registered in the German Clinical Trials Register (DRKS00028523). Extended Data Table 1 and Fig. 1 report the sex, age, percentile BMI values (calculated according to the KiGGS study \u003csup\u003e51\u003c/sup\u003e) and clinical manifestation details of participants during acute long COVID. The cohort included paediatric long COVID patients infected from October 2020 onwards, the majority of whom were infected during the third wave of the SARS-CoV-2 pandemic in Germany (see Extended Data Table 1). The control group comprised healthy children and adolescents, as well as paediatric cystic fibrosis (CF) patients in a stable condition. Full details of the cohort are provided in Extended Data Table 1a. Only patients who were diagnosed with long COVID at their initial presentation were included in the study. Of the 78 paediatric long COVID patients enrolled, four were excluded: three due to alternative plausible causes of their symptoms (Borrelia infection, psychosomatic symptoms, or symptom onset predating SARS-CoV-2 infection), and one due to withdrawal of consent.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ePatient Assessments\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eLC patient assessments encompassed both validated questionnaires and objective physical performance tests Fig.1a. Most questionnaires for participants were completed by proxy.\u0026nbsp;Psychometric and functional assessments included the Generalized Anxiety Disorder 7-item scale (GAD-7) for anxiety symptoms, the Patient Health Questionnaire 9-item (PHQ-9) for depressive symptoms, and the Short Form-12 Physical and Mental Component Summary scores (SF-12 PCS and SF-12 MCS) to evaluate physical and mental health-related quality of life. Vital parameters and a wide range of symptoms were assessed through these instruments, including the Munich Long COVID Symptom Questionnaire (MLCSQ), which assesses the frequency of 96 potential symptoms of LC, divided into 13 Systems (Fig.1a,c; Extended Data Fig.1). Symptom burden was captured through both parent-reported symptoms (PRS) and patient-reported symptoms (PtRS). Health-related quality of life in children was assessed using the KIDSCREEN questionnaire, and sleep disturbances were evaluated with the Children\u0026rsquo;s Sleep Habits Questionnaire (CSHQ). Functional impairment was assessed using the Bell Disability Scale (Bell Score), while fatigue was evaluated with the Fatigue Severity Scale (FSS). Physical performance was evaluated using the Sit-to-Stand Test (STS) and Hand Grip Strength Test \u003csup\u003e52,53\u003c/sup\u003e.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eIsolation of PBMCs and serum\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003ePeripheral blood was collected into EDTA and SST tubes. PBMCs were isolated at room temperature by Ficoll-Paque density centrifugation. Specimens from paediatric LC patients were archived at the Integrated Biobank Jena (University Hospital Jena) and controls at the Medical Faculty, Otto-von-Guericke University Magdeburg. Serum was analysed immediately or stored at \u0026minus;20\u0026nbsp;\u0026deg;C for batched aAb and cytokine assays with matched controls. Cells were cryopreserved in heat-inactivated FCS supplemented with 10\u0026nbsp;% (v/v) DMSO.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCytokine analysis in\u003c/strong\u003e\u003cstrong\u003e\u0026nbsp;serum\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eCytokine levels in serum samples were quantified using a flow cytometric multiplex bead-based assay (Supplementary Fig.1) \u003csup\u003e12,23\u003c/sup\u003e. The concentrations of GM-CSF, IL-1\u0026alpha;, IL-1\u0026beta;, IL-2, IL-3, IL-4, IL-5, IL-6, IL-9, IL-10, IL-11, IL-12p40, IL-12p70, IL-13, IL-15, IL-18, IL-22, IL-23, IL-27, IL-33, TNF\u0026alpha;, TSLP, IFN\u0026alpha;, and IFN\u0026gamma; were measured in serum samples from paediatric LC patients and controls using the LegendPlex Human Th Cytokine Panel and LegendPlex Human Cytokine Panel 2 (BioLegend), following the protocols provided by the manufacturer and as described previously\u003csup\u003e12,23\u003c/sup\u003e. Data acquisition was performed on a FACSFortessa X-20 (Becton Dickinson), and results were analysed using the LegendPlex Data Analysis Software Suite (Qognit).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eRoutine Laboratory Testing\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eLaboratory analyses included metabolic markers, blood gases, differential blood counts, general clinical chemistry, coagulation parameters, urinalysis, immunoglobulins, hormone levels, allergy diagnostics, autoantibodies, antiviral antibodies (EBV- and CMV-specific), and anti-SARS-CoV-2 serology. Pulmonary function parameters, vascular function (static and dynamic analysis), and vital signs were also assessed. Arterial/venous blood gases were analysed on ABL 90 Flex Plus instruments. Complete and differential blood counts were obtained on an XN-1000 (Sysmex) using fluorescence flow cytometry, SLS-Hb and impedance channels. Anti-SARS-CoV-2 spike S1 IgG (quantitative; ELISA Genie, CBK4154) and anti-nucleocapsid IgG (qualitative; Roche Elecsys) were performed per manufacturer\u0026rsquo;s instructions. Values exceeding analytical limits of detection were recorded as maximum + 1 or minimum / 2, respectively.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAutoantibody Profiling and Biomarker Quantification\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAntinuclear antibodies (ANA) were detected by indirect immunofluorescence using the ANA HEp-2 plus kit (#8101, Generic Assays, Dahlewitz, Germany) on HEp-2 cells following serial serum dilutions (1:80\u0026ndash;1:2560; positivity defined as \u0026ge;1:80), with endpoint titres evaluated using a fluorescence microscope (Olympus, Japan). Presence of anti-DFS70 antibodies was confirmed by immunoblot using the EUROLINE Anti-DFS70 (IgG) kit (#DL 159z-1601 G, Euroimmun, L\u0026uuml;beck, Germany).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eAnti-CCP IgG (#3665, cut off \u0026lt;30 U/ml), anti-TG IgA (#4033, cut off \u0026lt;20 U/ml; both Generic Assays, Dahlewitz, Germany) as well as anti-Cardiolipin IgG (#ORG 515S, \u0026nbsp;cut off \u0026lt;10 U/ml), anti-\u0026beta;2-Glycoprotein IgG (#ORG 521S, cut off \u0026lt;10 U/ml), anti-Prothrombin IgG (#ORG 541S, cut off \u0026lt;20 U/ml), anti-PR3 IgG (#ORG 618, cut off \u0026lt;10 U/ml) and anti-MPO IgG (#ORG 519, cut off \u0026lt;5 U/ml; all Orgentec, Mainz, Germany) were analysed by ELISA. Serum samples were diluted 1:100 in sample buffer and transferred in duplicate into the respective microtiter plate together with standards and controls. Final analysis was performed using a Tecan Sunrise Microplate reader (Tecan, M\u0026auml;nnedorf, Switzerland) and calculated according to linearity scale. Values below the diagnostic thresholds were retained as described previously to illustrate the assay\u0026rsquo;s linear range\u003csup\u003e23,30\u003c/sup\u003e, acknowledging the primarily diagnostic role of established cut-offs.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eSerum levels of IgG autoantibodies directed against \u0026beta;1-adrenergic, \u0026beta;2-adrenergic, M3-muscarinic, and M4-muscarinic acetylcholine receptors were quantified using a standardized ELISA (CellTrend GmbH, Luckenwalde, Germany) using MAGELLAN\u003csup\u003eTM\u003c/sup\u003e microplate reader. Each assay run included manufacturer-provided positive and negative controls to ensure assay validity. Each assay included internal calibrators and manufacturer-provided positive and negative controls to ensure linearity and validity. Internal quality control was performed at three concentration levels to assess intra-assay precision. Reference ranges for physiological levels of these functional aAb were defined as \u0026lt;15 U/mL for \u0026beta;1-adrenergic, \u0026lt;8 U/mL for \u0026beta;2-adrenergic, \u0026lt;6 U/mL for M3-muscarinic, and \u0026lt;10.7 U/mL for M4-muscarinic receptor aAb; upper assay limits were \u0026gt;40 U/ml.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eAnalysis of soluble NfL (NfL) levels was conducted using the Quanterix Simoa NF-Light assay Advantage Kit (Lexington, MA), according to the protocol provided. All samples were analyzed in duplicate. Testing was performed blinded to patient clinical/paraclinical data and outcome measures. To adjust for the influence of age, NfL concentrations were converted to Z scores and (interchangeable) percentiles https://shiny.dkfbasel.ch/baselnflreference-for-kids\u003csup\u003e33\u003c/sup\u003e. The z-score represents the number of standard deviations a particular value deviates from the mean of healthy, age-matched individuals\u003csup\u003e33\u003c/sup\u003e. The z-score represents the number of standard deviations from the mean of healthy age-matched controls.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eStatistics\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eLinear mixed-effects models (LMMs) were used throughout to assess longitudinal changes in clinical, functional, metabolic and immunological parameters while accounting for repeated measures via a random intercept for patient ID\u003csup\u003e54\u003c/sup\u003e. Fixed effects varied by model and included time since SARS-CoV-2 infection (\u0026le;1 vs. \u0026gt;1 year or continuous), EBV EBNA IgG status (cut-off 50 U\u0026nbsp;ml\u003csup\u003e-1\u003c/sup\u003e), DFS70 status (positive/negative), age, sex, BMI percentile, and relevant interactions. Specific LMMs are detailed in Fig. 1e, Fig.\u0026nbsp;2d, Fig. 3f, Fig. 4, and Extended Data Tables 1\u0026ndash;6. Model assumptions were assessed via visual inspection of model residuals; no relevant heteroscedasticity or multicollinearity (VIF\u0026nbsp;\u0026lt;\u0026nbsp;2) was observed. P-values were calculated using the Satterthwaite approximation and validated via bootstrapping (1,000 iterations); adjusted values were corrected using the Holm\u0026ndash;Bonferroni method. Only corrected p-values \u0026lt; 0.05 were considered significant. Marginal R\u0026sup2; values indicate variance explained by fixed effects.\u003c/p\u003e\n\u003cp\u003eStatistical analyses were performed using GraphPad Prism, Jamovi (v3.0), and the GamLj module (v3), including linear and linear mixed-effects models (LMMs) with donor ID as a random intercept. Normality was assessed using Shapiro\u0026ndash;Wilk and Kolmogorov\u0026ndash;Smirnov tests, and visually via Q\u0026ndash;Q plots. Outliers were identified using Grubbs\u0026rsquo; test. For group comparisons, Mann\u0026ndash;Whitney U-tests were used for non-parametric data; the Friedman test with Dunn\u0026rsquo;s post hoc correction was applied for paired samples. Pearson\u0026rsquo;s correlation was used for normally distributed variables; Spearman\u0026rsquo;s rank correlation for ordinal or non-normally distributed data. Diagnostic values below detection limits were imputed as half the limit; values above range were set to maximum + 1, as specified in Extended Data Table 2.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eReferences\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e51. H\u0026ouml;lling, H.\u003cem\u003e\u0026nbsp;et al.\u0026nbsp;\u003c/em\u003eDie KiGGS-Studie. Bundesweit repr\u0026auml;sentative L\u0026auml;ngs- und Querschnittstudie zur Gesundheit von Kindern und Jugendlichen im Rahmen des Gesundheitsmonitorings am Robert Koch-Institut.\u0026nbsp;\u003cem\u003eBundesgesundheitsblatt, Gesundheitsforschung, Gesundheitsschutz\u0026nbsp;\u003c/em\u003e\u003cstrong\u003e55,\u0026nbsp;\u003c/strong\u003e836\u0026ndash;842; 10.1007/s00103-012-1486-3 (2012).\u003c/p\u003e\n\u003cp\u003e52.\u0026nbsp; \u0026nbsp;Haile, S. R.\u003cem\u003e\u0026nbsp;et al.\u0026nbsp;\u003c/em\u003eReference values and validation of the 1-minute sit-to-stand test in healthy 5-16-year-old youth: a cross-sectional study. \u003cem\u003eBMJ open\u0026nbsp;\u003c/em\u003e\u003cstrong\u003e11,\u0026nbsp;\u003c/strong\u003ee049143; 10.1136/bmjopen-2021-049143 (2021).\u003c/p\u003e\n\u003cp\u003e53.\u0026nbsp; \u0026nbsp;Reychler, G., Audag, N., Mestre, N. M. \u0026amp; Caty, G. Assessment of Validity and Reliability of the 1-Minute Sit-to-Stand Test to Measure the Heart Rate Response to Exercise in Healthy Children. \u003cem\u003eJAMA pediatrics\u0026nbsp;\u003c/em\u003e\u003cstrong\u003e173,\u0026nbsp;\u003c/strong\u003e692\u0026ndash;693; 10.1001/jamapediatrics.2019.1084 (2019)\u003c/p\u003e\n\u003cp\u003e54.\u0026nbsp; \u0026nbsp;Arra, A.\u003cem\u003e\u0026nbsp;et al.\u0026nbsp;\u003c/em\u003ePD1+ innate lymphoid cells 3 predict JAK-dependent inflammation in rheumatoid arthritis. \u003cem\u003eJournal of autoimmunity\u0026nbsp;\u003c/em\u003e\u003cstrong\u003e154,\u0026nbsp;\u003c/strong\u003e103424; 10.1016/j.jaut.2025.103424 (2025).\u003c/p\u003e"},{"header":"Tables","content":"\u003cp\u003eTables 1 and 2 are available in the Supplementary Files section.\u003c/p\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":true,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"nature-portfolio","isNatureJournal":true,"hasQc":false,"allowDirectSubmit":false,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"","title":"Nature Portfolio","twitterHandle":"","acdcEnabled":false,"dfaEnabled":false,"editorialSystem":"ejp","reportingPortfolio":"","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"","lastPublishedDoi":"10.21203/rs.3.rs-7083240/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-7083240/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eWhile most children and adolescents recover uneventfully from SARS-CoV-2 infection, some develop persistent symptoms known as paediatric long COVID (LC). Paediatric LC presents with substantial, multisystem health impairment lasting months to years after SARS-CoV-2 infection\u003csup\u003e\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e,\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e\u003c/sup\u003e. Despite its clinical burden, underpinnings of symptom persistence, heterogeneity, and recovery remain elusive\u003csup\u003e\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e,\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e\u003c/sup\u003e. Here, we demonstrate that severe symptoms in paediatric LC remained stable over two-to-three years, despite unremarkable cardiopulmonary and routine assessments, and were underpinned by temporally shifting immune-metabolic responses. The first year of LC was marked by viral-associated and Th2-like cytokine responses, transitioning into Th17-like and innate responses over time. Neurofilament light chain, an indicator of neuro-axonal injury, rose with LC-severity, but common autoantibodies remained unchanged. Epstein-Barr virus (EBV) exposure emerged as a key modifier linked to broader immune dysfunction, whereas anti-DFS70 autoantibodies correlated with milder haematological alterations. In EBV-na\u0026iuml;ve LC cases, symptoms became more severe with altered blood viscosity, but less severe with higher IL-12p40, vitamin B1, and basophils, implicating them as protective. The identified LC subgroups displayed metabolically distinct signatures, supporting the existence of biologically coherent endotypes. These findings uncover immune-metabolic axes linked to resilience and persistence in paediatric LC and may provide a basis for biomarker-informed diagnosis and precision intervention.\u003c/p\u003e","manuscriptTitle":"Immune-Metabolic Programs Drive Disease Trajectories in Paediatric Long COVID","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-07-23 08:55:51","doi":"10.21203/rs.3.rs-7083240/v1","editorialEvents":[],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"nature-communications","isNatureJournal":true,"hasQc":false,"allowDirectSubmit":false,"externalIdentity":"NCOMMS","sideBox":"Learn more about [Nature Communications](http://www.nature.com/ncomms/)","snPcode":"","submissionUrl":"https://mts-ncomms.nature.com/","title":"Nature Communications","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"ejp","reportingPortfolio":"Nature Communications","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"2ff00461-cfe3-4f0f-b41a-294b8768a9ee","owner":[],"postedDate":"July 23rd, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[{"id":51489733,"name":"Health sciences/Biomarkers/Predictive markers"},{"id":51489734,"name":"Health sciences/Diseases/Immunological disorders"},{"id":51489735,"name":"Health sciences/Medical research/Paediatric research"}],"tags":[],"updatedAt":"2026-05-05T07:05:34+00:00","versionOfRecord":{"articleIdentity":"rs-7083240","link":"https://doi.org/10.1038/s41467-026-72224-y","journal":{"identity":"nature-communications","isVorOnly":false,"title":"Nature Communications"},"publishedOn":"2026-05-04 04:00:00","publishedOnDateReadable":"May 4th, 2026"},"versionCreatedAt":"2025-07-23 08:55:51","video":"","vorDoi":"10.1038/s41467-026-72224-y","vorDoiUrl":"https://doi.org/10.1038/s41467-026-72224-y","workflowStages":[]},"version":"v1","identity":"rs-7083240","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-7083240","identity":"rs-7083240","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}