Emerging trends in post-COVID immune dysregulation: a narrative review

Abstract

Background:

Long coronavirus disease (COVID), or post-acute sequelae of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection, is a multi-system condition associated with persistent immune dysregulation, autoimmunity, and vascular perturbations. Understanding these immunological mechanisms is essential for guiding clinical management and therapeutic strategies.

Methods:

We conducted a narrative review of peer-reviewed human studies published between January 2020 and August 2025, using PubMed/Medline, Scopus, and Web of Science, supplemented by manual searches. Studies reporting immunological assessments in long COVID patients, recovered individuals, or healthy controls were included. Findings were extracted and synthesized thematically across six domains: humoral and cellular immunity, autoimmune signatures, proteomic/metabolic and vascular dysregulation, viral persistence, complement/coagulation/thromboinflammation, and pediatric long COVID.

Results:

Long COVID is characterized by persistent low-grade inflammation, T- and B-cell dysregulation, and prolonged adaptive immune activation. Humoral responses remain elevated, while T-cell exhaustion and loss of coordination between immune compartments are common. Autoimmune phenomena, including latent and polyautoimmunity targeting cytokines, thyroid antigens, and interferons, are frequently observed. Proteomic, metabolic, and vascular perturbations, complement activation, and thromboinflammatory processes contribute to ongoing symptoms. Viral persistence and early immune biomarkers predict long COVID development. These immune alterations correlate with fatigue, cognitive impairment, respiratory dysfunction, and reduced quality of life in both adults and children.

Conclusion:

SARS-CoV-2 infection leaves a lasting immunological footprint marked by chronic inflammation, adaptive immune dysregulation, autoimmunity, and vascular perturbations. Integrating longitudinal immune assessments, biomarker profiling, and early predictive markers is critical for identifying high-risk individuals and informing interventions to reduce long COVID morbidity.

Introduction

The global spread of coronavirus disease (COVID-19), caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), has challenged healthcare systems worldwide and stimulated intensive research into acute immunopathology and vaccine-mediated protection[1]. During the acute phase of infection, SARS-CoV-2 typically triggers a robust innate immune response characterized by elevated pro-inflammatory cytokines, activation of monocytes, macrophages, and neutrophils, and recruitment of T and B lymphocytes^[2,3^]. However, emerging data indicate substantial inter-individual variability in acute immune responses, with some individuals mounting more controlled or less pronounced inflammatory responses while still achieving effective viral clearance[4]. Nevertheless, emerging evidence indicates that a subset of individuals experience persistent or evolving immune alterations even after apparent recovery, which may manifest as longer-term health sequelae^[5,6^].

HIGHLIGHTS

• Long COVID is marked by persistent low-grade inflammation and dysregulated T- and B-cell responses.

• Autoimmune signatures, including multiple autoantibodies, are frequently observed in affected individuals.

• Proteomic, metabolic, vascular, and thrombo-inflammatory changes contribute to symptom persistence.

• Evidence suggests a role for viral persistence and early immune markers in predicting long COVID.

• Immune abnormalities correlate with fatigue, cognitive issues, respiratory symptoms, and reduced quality of life.

Most individuals eventually clear the virus and develop an adaptive immune response that provides protection against reinfection. Globally, long COVID affects a substantial proportion of those recovering from SARS-CoV-2 infection, with pooled prevalence estimates around 36%, and major symptom clusters including respiratory difficulties, general fatigue, and neurological and psychological manifestations[7]. These persistent symptoms contribute to increased healthcare utilization and prolonged functional impairment, highlighting the global public-health burden of post-COVID conditions.

Persistent immune perturbation is captured in the concept of long COVID, or post-acute sequelae of SARS-CoV-2 infection (PASC), which is characterized not only by lingering symptoms but also by immunological signatures that deviate from the expected resolution of inflammation^[8,9^]. Studies have documented altered T-cell subset distributions, elevated markers of immune activation or exhaustion, dysregulated cytokine profiles, and evidence of autoimmunity. For example, patients with PASC 8 months after infection exhibited systemic inflammation, exhausted antigen-specific CD8⁺ T cells, and a loss of coordination between T- and B-cell responses[5]. In another cohort, while pro-inflammatory biomarkers declined over the first 6 months, individuals with long COVID symptoms displayed a shift toward immune suppression or exhaustion rather than full homeostatic recovery[10].

Understanding post-COVID immune dysregulation has critical implications. Persistent immune activation or exhaustion may influence susceptibility to reinfection, vaccine responsiveness, and the development of other infectious or autoimmune conditions^[11,12^]. Immune footprints may also underpin long-term symptoms such as fatigue, cognitive impairment, and cardiovascular sequelae, bridging immunological findings with clinical and public-health perspectives. Given the global burden of SARS-CoV-2 infections across multiple waves, variants, and vaccination statuses, these immune alterations have important public-health consequences, particularly in low- and middle-income countries where healthcare resources and surveillance are limited.

In this narrative review, we synthesize current evidence on post-COVID immune dysregulation, highlighting key domains, including innate immune alterations, adaptive T- and B-cell derangements, and autoimmunity.

In this review, we conceptualize long COVID not as a single nosological entity but as a complex spectrum of clinical phenotypes underpinned by distinct post-infectious endotypes. While the clinical presentation is diverse, emerging evidence suggests that these manifestations can be operationalized into biological subgroups, such as those driven by viral persistence, latent viral reactivation (e.g., EBV), or chronic autoimmune triggers. By framing the evidence through these mechanistic endotypes, we aim to clarify how the diverse immunological findings synthesized in this review relate to specific patient trajectories.

To operationalize this framework, we organized our synthesis across six thematic domains, reflecting distinct pathophysiological mechanisms and clinical subgroups (detailed justification provided in Methods).

There is growing interest in therapeutic strategies to manage post-COVID immune dysregulation and persistent symptoms, with several recent efforts aiming to move beyond descriptive immunopathology toward actionable interventions. Chatterjee and Maparu (2025) provide a comprehensive overview of potential therapies for long COVID, highlighting pharmacological agents, immunomodulatory approaches (e.g., targeting persistent inflammation and T-cell exhaustion), and supportive care strategies that may mitigate chronic inflammation and multisystem sequelae[13]. In parallel, a living systematic review of long COVID interventions suggests that physical and psychological rehabilitation programs probably improve symptoms and quality of life, while evidence for specific immunomodulatory or targeted drug therapies is still emerging and underscores the need for well-designed clinical trials[14].

We further discuss clinical and public-health implications, including long COVID, reinfection risk, and vaccine response, while identifying major knowledge gaps and priorities for future research. This integrative overview provides a timely resource for researchers, clinicians, and policymakers seeking to understand and address the long-term immunological impact of COVID-19 and to inform the development of future therapeutic strategies.

Methods

This narrative review was conducted in accordance with the Scale for the Assessment of Narrative Review Articles guidelines for high-quality narrative reviews[15]. It was undertaken to synthesize current evidence on immune dysregulation following SARS-CoV-2 infection, with a focus on humoral and cellular immunity, autoimmune signatures, proteomic and metabolic pathways, thromboinflammatory processes, viral persistence, and pediatric Long COVID. The review aimed to integrate findings from both early and long-term post-infection studies, highlighting immunological mechanisms, biomarkers, and their clinical correlates.

Eligibility criteria

Studies were included if they met the following criteria: (i) original research on human participants with confirmed SARS-CoV-2 infection; (ii) assessment of immune responses, autoantibodies, proteomic/metabolic pathways, viral persistence, or thromboinflammatory processes in relation to persistent symptoms or long COVID; (iii) reported quantitative or qualitative immune data; and (iv) follow-up of at least 3 weeks post-infection to capture post-acute sequelae. Where comparator groups were included, these comprised fully recovered post-COVID individuals without persistent symptoms, uninfected healthy controls, or vaccinated individuals without prior documented infection, as defined by the original studies. Differences in comparator selection were retained and considered during the interpretation of reported immune alterations.

Exclusion criteria included: non-human studies, case reports without immune data, and studies lacking post-infection follow-up or immunological assessments.

Operational definition

For the purposes of this review, long COVID [also referred to as post-acute sequelae of SARS-CoV-2 infection [PASC] or post-COVID condition] was conceptualized according to the WHO definition of post-COVID-19 condition (symptoms persisting ≥12 weeks after acute infection), while also acknowledging the broader NIH PASC framework (symptoms ≥4 weeks)^[16,17^]. Variation in terminology and case definitions across the included studies (e.g., “Long COVID,” “post-COVID condition,” and PASC) was noted; however, all studies were retained if they described persistent symptoms beyond the acute phase (≥4 weeks as a minimum threshold). Individual study definitions and follow-up durations are summarized in Supplemental Digital Content File S1, available at: http://links.lww.com/MS9/B256.

Literature sources and search strategy

We searched three major electronic bibliographic databases: PubMed/Medline, Scopus, and Web of Science for peer-reviewed articles published between 1 January 2020 and 31 August 2025. The search was supplemented by manual screening of reference lists from key articles, systematic reviews, and preprints with peer-reviewed confirmation. Search terms combined keywords and Medical Subject Headings relating to (“post-COVID” OR “post-acute COVID” OR “long COVID” OR “post-SARS-CoV-2”) AND (“immune dysregulation” OR “immune activation” OR “immune exhaustion” OR “autoimmunity” OR “T cell” OR “B cell” OR “cytokine” OR “inflammation” OR “proteomics” OR “metabolomics” OR “viral persistence” OR “complement” OR “thromboinflammation”). Boolean operators (AND, OR) were applied to combine concepts. The final search was conducted on 1 September 2025 (Supplemental Digital Content File S2, available at: http://links.lww.com/MS9/B257).

Study selection and data extraction

Titles and abstracts were screened independently by two reviewers (E.G.A. and S.A.H.) for relevance. Full texts of potentially eligible studies were retrieved and assessed against the inclusion criteria. Key data extracted included study design, country, sample size, population characteristics, long COVID definitions, follow-up duration, immune markers measured, key immune findings, and associated clinical outcomes. Discrepancies were resolved through discussion and consensus; if no agreement was reached, the third reviewer (E.H.G.) made the final decision (Supplemental Digital Content File S1, available at: http://links.lww.com/MS9/B256).

Synthesis of evidence

The six-domain classification was derived from an integrated conceptual framework combining: (i) established pathophysiological mechanisms of post-viral syndromes (autoimmunity, viral persistence, and thromboinflammation); (ii) emerging multi-omics findings in long COVID literature; and (iii) author-defined categories based on the evidence reviewed. These domains span different levels of analysis, from molecular mechanisms to physiological systems to clinical subgroups, reflecting the multidimensional nature of post-COVID immune pathophysiology. Findings were organized thematically into six major domains: humoral and cellular immunity, autoimmune signatures, proteomic/metabolic and vascular dysregulation, viral persistence and biomarkers, complement/coagulation/thromboinflammatory processes, and pediatric long COVID. Narrative synthesis was employed due to heterogeneity in study designs, populations, immune assays, and follow-up durations, which precluded formal meta-analysis. For evidence weighting, priority was given to: (i) longitudinal designs with repeated measures that track immune trajectories over time; (ii) studies with independent validation cohorts confirming findings; and (iii) consistency of effect direction across multiple studies, regardless of sample size. Consistency across findings was assessed by comparing the direction and magnitude of immune alterations reported in at least two independent studies. Discrepancies were resolved by examining differences in study populations, case definitions (as detailed in Supplemental Digital Content File S1, available at: http://links.lww.com/MS9/B256), or assay methods, with conflicts transparently reported in the synthesis. To minimize bias, we focused on peer-reviewed studies, integrated both longitudinal and cross-sectional evidence, and transparently reported heterogeneity and limitations of included studies. Emphasis was placed on integrating longitudinal and cross-sectional evidence to provide a comprehensive overview of immunological patterns and their clinical implications.

Results

Out of 684 retrieved studies, 25 were eligible and included in the final analysis (Fig. 1).

Figure 1.

PRISMA flow diagram of search and study selection process.

Across the included studies, long COVID clinical phenotypes were heterogeneous, with fatigue, cognitive impairment, respiratory dysfunction, and multiorgan/systemic symptoms emerging as predominant clusters. Symptom definitions varied across studies, with most requiring persistence for at least 12 weeks post-infection (aligned with WHO criteria), while some used a minimum of 4 weeks (NIH PASC framework). The co-occurrence of multiple symptom domains was common, with few patients presenting with isolated single-domain symptoms. Fatigue and cognitive symptoms were frequently linked to underlying immune dysregulation, while cardiopulmonary manifestations correlated with elevated inflammatory cytokines, and persistent respiratory symptoms were associated with vascular and angiogenic pathway alterations. This clinical heterogeneity supports framing long COVID as a spectrum of partially overlapping syndromes rather than a single unified construct.

Study characteristics

This analysis included 25 studies comprising a total of 4986 participants, investigating immune alterations and biomarkers associated with long COVID or PASC across multiple regions, including Australia, North America, Europe, Asia, Africa, and South America. The studies used various designs, including prospective longitudinal cohorts, cross-sectional analyses, and case-control studies, with sample sizes ranging from fewer than 21 participants to 845. Study populations included individuals with persistent post-COVID symptoms, fully recovered patients, and healthy controls (Table 1).

Table 1.

Summary of included studies on post-COVID immune dysregulation (n = 23 studies).

Author(s)	Year	Country	Study design	Sample size
Bodansky et al[15]	2023	USA	Cross-sectional	242
Buonsenso et al[8]	2024	Italy	Cross-sectional	21
Cervia-Hasler et al[16]	2024	Switzerland	Case-control	152
Farré et al[7]	2025	Spain	Cohort	171
Gabernet et al[11]	2023	USA	Cohort	513
Huqin et al[17]	2025	China	Cohort	105
Jiang et al[18]	2024	China	Cohort	66
Kallaste et al[9]	2025	Estonia	Cohort	105
Klein et al[19]	2023	USA	Case-control	275
Phetsouphanh et al[3]	2024	Australia	Cohort	62
Pretorius et al[20]	2021	South Africa	Case-control	47
Pretorius et al[21]	2022	South Africa	Cross-sectional	845
Rojas et al[10]	2022	Colombia	Cross-sectional	130
Simón-Rueda et al[22]	2024	Spain	Cohort	67
Su et al[5]	2022	USA	Cohort	309
Swank et al[23]	2023	USA	Cohort	63
Thompson et al[24]	2023	USA	Cohort	567
Vazquez-Alejo et al[2]	2023	Spain	Cohort	64
Wei et al[25]	2024	China	Case-control	50
Wolday et al[26]	2025	Ethiopia	Cohort	78
Wynberg et al[27]	2024	South Africa	Cohort	223
Yin et al[4]	2024	USA	Cohort	43
Zhang et al[28]	2024	China	Cohort	362
Kwissa et al[29]	2025	USA	Cohort	Not specified
Bae et al[30]	2025	Germany	Cohort	426

Humoral and cellular immune responses

Long COVID patients exhibited persistent alterations in both humoral and cellular immunity. Longitudinal data allow differentiation between early post-acute alterations (up to 6 months) and long-term persistent changes (beyond 6 months), highlighting which immune perturbations resolve versus those that remain chronically elevated. In the ADAPT cohort[3], it was observed that at 3 and 8 months post-infection, participants with long COVID had approximately three-fold higher anti-nucleocapsid IgG, elevated neutralization capacity, and increased frequencies of spike- and nucleocapsid-specific CD4⁺ and CD8⁺ T-cells compared to matched controls. Exhaustion markers PD-1 and TIM-3 on T-cells remained elevated alongside enhanced innate activation of monocytes and dendritic cells. By 24 months, many differences had largely resolved, although subtle elevations in IFN-β, IFN-γ, and CD8⁺ T-cells persisted. Similarly, Yin et al[5] reported higher frequencies of exhausted CD8⁺ T-cells and increased tissue-migratory CD4⁺ cells in long COVID individuals, accompanied by a loss of coordination between T- and B-cell responses. Kwissa et al (2025) additionally reported that long COVID patients maintain persistently elevated IgG titers against SARS-CoV-2 Envelope and Nucleocapsid proteins, exhibit altered antibody class switching, and show increased circulating T-follicular helper and MAIT cells alongside elevated cytokines, including LIF, IL-11, Eotaxin-3, and HMGB-1, collectively supporting a state of prolonged immune activation and dysregulation distinct from convalescent individuals[18].

Persistent innate immune alterations, including heightened monocyte activation, altered natural killer cell function, and dysregulated interferon signaling, appear to interact with adaptive T- and B-cell responses, potentially sustaining chronic inflammation and contributing to the heterogeneity of long COVID manifestations[19], and evidence for trained monocyte/macrophage activation has been described in signal transduction studies of post-COVID innate immune imprinting[20].

In cohorts of patients recovering from severe acute COVID-19, Vazquez-Alejo et al[2] found that T-cell exhaustion markers, including TIM-3 and PD-1, remained elevated at 6 months, while SARS-CoV-2-specific memory T-cell subsets persisted, indicating prolonged immune activation and incomplete resolution. Thompson et al (2023)[21] demonstrated that acute-phase gene-expression signatures in whole blood, including immunoglobulin and immune cell markers, were associated with the later development of post-acute sequelae, highlighting how early immune responses may shape long COVID risk. Wynberg et al[22] found that elevated plasma cytokines, such as IL-6, IL-10, IL-17, TNFα, and IP10, were associated with persistent symptoms at 12–24 weeks, with early IL-1β levels predictive of subsequent impairment in diffusion capacity. In addition, Klein et al[23] reported differences in lymphoid and myeloid populations, exaggerated humoral responses, and higher antibody titers against SARS-CoV-2, EBV, and VZV, which correlated with long COVID status and cognitive/fatigue symptom clusters. While these observations highlight persistent immune perturbations in long COVID, they are primarily associative, and causal mechanisms linking specific immune signatures to clinical phenotypes remain to be established (Supplemental Digital Content Table S1, available at: http://links.lww.com/MS9/B251).

Phenotypic heterogeneity of long COVID

Long COVID presents as a spectrum of clinical phenotypes with corresponding immune signatures. Fatigue-dominant symptoms and cognitive impairment are associated with broad immune activation and dysregulated antibody responses[23]. Cardiopulmonary manifestations, including persistent dyspnea and reduced diffusion capacity, correlate with elevated systemic cytokines such as IL-6, IL-10, IL-17, TNFα, and IP10[22], and endothelial dysfunction[24]. Immune exhaustion and dysregulated adaptive responses, reflecting T- and B-cell imbalance, are observed across multiple phenotypes^[5,25^]. This evidence emphasizes that long COVID is immunologically heterogeneous, supporting the need for phenotype-specific investigation and tailored clinical strategies.

Autoimmune signatures

Long COVID has been associated with persistent autoantibody production and dysregulated immune self-recognition. Bodansky et al (2023)[26] identified shared autoantibody signatures in individuals with long COVID and fully recovered patients, suggesting that some autoimmunity may persist independently of ongoing symptoms. In a Colombian cohort, Rojas et al[11] found that 83% of Long COVID patients exhibited latent autoimmunity, while 62% displayed polyautoimmunity, including IgG antibodies against IL-2, CD8B, and thyroglobulin. Notably, anti-interferon antibodies were present in 5–10% of patients, and higher anti-SARS-CoV-2 IgG titers correlated with the presence of autoantibodies, age, and body mass index. The presence of anti-IFN-λ IgG was specifically associated with persistent respiratory symptoms, highlighting a mechanistic link between autoimmunity and long COVID manifestations. Similarly, Su et al[25] reported that early detection of autoantibodies, together with EBV viremia and SARS-CoV-2 RNAemia, predicted the development of post-acute sequelae, revealing distinct immune endotypes that may underlie heterogeneous long COVID phenotypes. These associations do not prove causality, and further longitudinal studies are required to determine whether autoantibody production contributes directly to clinical manifestations or represents an epiphenomenon (Table 2).

Table 2.

Stratification of post-COVID autoantibody findings into transient auto-reactivity versus clinically significant autoimmune disease.

Study (year)	Autoantibodies/findings	Classification
Bodansky et al (2023)[15]	Low-level autoantibodies detected in both groups, without a clear association with symptoms	Transient auto-reactivity
Rojas et al (2022)[10]	Anti-IL-2, anti-CD8B, anti-thyroglobulin; 62% polyautoimmunity; anti-IFN-λ associated with persistent respiratory symptoms	Clinically significant autoimmune disease
Su Y et al (2022)[5]	Early autoantibodies + EBV/SARS-CoV-2 RNAemia predicted post-acute sequelae	Clinically significant autoimmune disease/predictive endotype

Proteomic, metabolic, and vascular dysregulation

Long COVID is characterized by persistent alterations in vascular, proteomic, and metabolic pathways, reflecting a complex interplay between immune dysregulation and tissue repair mechanisms. Where reported, these studies distinguish early post-infection molecular changes from long-term persistent dysregulation, allowing assessment of which pathways may contribute to chronic symptoms versus transient post-acute alterations. Farré et al [8] reported that VEGFA was overexpressed in long COVID patients, with sex-specific differences observed in postmenopausal women, alongside activation of vascular, immune, and angiogenesis-related pathways. Kallaste et al[10] found evidence of early low-grade inflammation in long COVID patients that largely subsided by 6 months, suggesting that transient immune activation may contribute to long-term symptom persistence. In a Chinese cohort, Huqin et al[27] demonstrated persistent molecular alterations in proteomic, metabolomic, and transcriptomic profiles up to 3 years post-infection, with these changes correlating with impaired pulmonary function.

Wolday et al[28] identified that acute-phase elevations in SLAMF1, IL15RA, IL18, and CXCL9/10 predicted long COVID development, while TRANCE expression appeared protective. Jiang et al[29] highlighted enhanced neutrophil activity in post-Omicron long COVID patients, accompanied by distinct proteomic and metabolomic signatures, suggesting a unique immune-metabolic endotype. Bae et al (2025) found a persistent hyperlipidemic phenotype alongside elevated proinflammatory mediators, including IFN-α2, IFN-γ, TNF-α, CXCL8/IL-8, IL-12p70, IL-17A, and IL-23, in long COVID patients up to 20 months post-acute infection, indicating enduring metabolic and inflammatory dysregulation consistent with chronic immune activation[30]. Finally, Simón-Rueda et al[24] reported ongoing immune dysregulation, endothelial dysfunction, and a pro-thrombotic profile in long COVID patients even 2 years post-infection, emphasizing the long-lasting impact on vascular and immune homeostasis (Supplemental Digital Content Table S2, available at: http://links.lww.com/MS9/B252).

Viral persistence and biomarkers of post-acute sequelae

Persistent viral antigens and early biomarker profiles have been strongly associated with long COVID development. Longitudinal follow-up clarifies that certain viral antigens or biomarker elevations are transient in the early post-acute phase, while persistence beyond 6–12 months is associated with ongoing symptomatology. Swank et al[31] detected circulating SARS-CoV-2 spike protein in patients up to 12 months post-infection, with antigen persistence closely linked to the presence of ongoing symptoms. In a large cohort, Gabernet et al[12] identified a low “recovery factor,” reflecting chronic inflammation and erythropoietic stress, which was predictive of persistent long COVID regardless of acute disease severity. Su et al[25] further demonstrated that early SARS-CoV-2 RNAemia, EBV viremia, and the presence of autoantibodies served as key predictors of post-acute sequelae, revealing distinct immune endotypes associated with symptom heterogeneity (Supplemental Digital Content Table S3, available at: http://links.lww.com/MS9/B253).

Complement, coagulation, and thromboinflammatory processes

Long COVID has been associated with persistent alterations in coagulation pathways, complement activation, and thromboinflammatory processes. Available longitudinal evidence suggests that while some coagulation and complement abnormalities are prominent in the early post-acute phase, a subset of patients demonstrates persistence of thromboinflammatory markers beyond 6–12 months, supporting a distinction between transient post-infectious effects and sustained vascular dysregulation. Pretorius et al[32] identified anomalous fibrin(ogen) microclots and increased antiplasmin levels in long COVID patients, with these abnormalities correlating with ongoing symptoms. This observation was further supported by Pretorius et al[33], who found a high prevalence of fibrin amyloid microclots and platelet pathology in long COVID, highlighting persistent vascular dysregulation. In Switzerland, Cervia-Hasler et al[34] reported dysregulation of complement proteins and terminal complement complexes, accompanied by platelet activation, which correlated with tissue injury markers. Similarly, Wei et al[35] observed dysregulation in coagulation, complement, platelet function, and lipid metabolism in long COVID patients, suggesting impaired energy metabolism and vascular repair pathways. Zhang et al[36] further demonstrated that complement and coagulation pathways remained altered up to 2 years post-hospitalization, with specific proteins associated with lung-function recovery (Supplemental Digital Content Table S4, available at: http://links.lww.com/MS9/B254).

Pediatric long COVID

Children with long COVID exhibit distinct immunological signatures and functional impairments compared to adults. Although pediatric data remain more limited, emerging follow-up studies indicate that certain cytokine elevations may attenuate over time, whereas functional impairment can persist, underscoring the importance of distinguishing early immune activation from longer-term sequelae in this population. Buonsenso et al[9] analyzed pediatric cohorts and found elevated cytokines, including CCL23, Flt3L, CD5, uPA, CD40, and TGFα, with CCL23 particularly associated with long COVID after adjusting for age and sex. These immune alterations were linked to functional impairment and suggested enhanced monocyte and T-cell migration (Supplemental Digital Content Table S5, available at: http://links.lww.com/MS9/B255).

Discussion

This narrative review highlights the complex and heterogeneous immunological landscape underlying long COVID across both adults and children, encompassing persistent humoral and cellular immune alterations, autoimmune signatures, viral persistence, endothelial dysfunction, thromboinflammatory pathways, and immune-metabolic dysregulation. Collectively, the included studies suggest that long COVID is not a single pathological entity but rather a spectrum of post-infectious immune endotypes with variable clinical manifestations and biological mechanisms^[2,3,5,11,18,23,25^].

Persistent immune activation appears to be a central feature of long COVID. Multiple studies have demonstrated prolonged elevations in SARS-CoV-2-specific antibodies, sustained T-cell activation or exhaustion, altered memory-cell populations, and dysregulated cytokine responses months after acute infection^[2,3,5,18,22,23^]. These findings support the concept that incomplete immune resolution may contribute to ongoing symptoms, including fatigue, cognitive dysfunction, and cardiopulmonary impairment. Early immune signatures during acute infection were also associated with subsequent post-acute sequelae, suggesting that the nature of the initial immune response may influence long COVID risk and trajectory^[21,22^].

The heterogeneity of immune findings across cohorts further supports the existence of distinct biological endotypes in long COVID. Persistent antigenic stimulation, altered adaptive immune coordination, interferon dysregulation, and innate immune activation may contribute differently across symptom phenotypes^[5,18–20,25^]. This variability may also partly reflect differences in acute disease severity, infecting viral variants, host immune background, and vaccination status^[2,5,25^]. Recognizing these distinct immune trajectories may help explain the broad spectrum of clinical manifestations observed in post-COVID conditions.

Autoimmune phenomena have emerged as another important component of long COVID pathophysiology. Several studies have identified persistent autoantibodies targeting cytokines, interferons, and self-antigens, together with evidence of latent or polyautoimmunity^[11,25,26^]. Associations between autoantibodies and persistent respiratory or systemic symptoms suggest that immune self-recognition may contribute to ongoing inflammation and tissue dysfunction in at least a subset of patients[11]. However, whether these autoantibodies are directly pathogenic or represent secondary markers of immune dysregulation remains uncertain and requires further longitudinal investigation.

Persistent viral antigens and viral reactivation may also contribute to chronic immune stimulation in long COVID. Detection of circulating spike protein months after infection, together with associations involving SARS-CoV-2 RNAemia and EBV reactivation, supports the hypothesis that incomplete viral clearance may sustain inflammatory responses in some individuals^[12,25,31^]. Mechanistically, SARS-CoV-2-induced immune dysregulation may trigger reactivation of latent EBV, and subsequent viral replication could further amplify inflammatory responses through T-cell cross-reactivity, bystander activation, and sustained cytokine production. Whether EBV reactivation directly drives pathology or merely reflects underlying immune dysfunction remains unclear. Clinically, these findings suggest that EBV serology, alongside autoantibody profiling, could aid in risk stratification, and EBV-positive long COVID subgroups may potentially benefit from antiviral or targeted immunomodulatory approaches, though this requires prospective validation.

At the same time, evidence of immune exhaustion and dysregulated immune memory suggests that maladaptive host immune responses may persist even in the absence of active viral replication^[25,37^]. These observations indicate that long COVID pathogenesis is likely multifactorial and involves interactions between viral persistence, immune dysregulation, and host susceptibility factors.

Complement activation, endothelial dysfunction, platelet activation, and thromboinflammatory pathways were also consistently implicated in long COVID^[24,32–36^]. Persistent fibrin microclots, complement dysregulation, and vascular abnormalities may contribute to impaired tissue perfusion, chronic inflammation, and multisystem symptoms^[24,32–35^]. The overlap between immune dysregulation and vascular injury highlights the interconnected nature of inflammatory and thrombotic pathways in post-COVID conditions. Emerging evidence further suggests that these mechanisms may share similarities with pathways involved in chronic cardiometabolic diseases, including endothelial dysfunction and low-grade systemic inflammation^[38,39^].

Proteomic and metabolomic studies additionally demonstrated persistent molecular alterations extending months to years after infection^[8,10,27–30^]. These findings support the presence of sustained immune-metabolic and vascular dysregulation in long COVID and may help identify biological pathways linked to symptom persistence and functional impairment. Although several candidate biomarkers have been proposed, including cytokine profiles, endothelial markers, autoantibodies, and thromboinflammatory proteins, their clinical applicability remains limited by variability in study design, laboratory methods, and lack of standardized validation across populations[40].

Pediatric long COVID remains comparatively understudied, although emerging evidence suggests that children may also develop persistent immune dysregulation and functional impairment following SARS-CoV-2 infection^[9,41^]. Reported cytokine alterations and immune-cell migration signatures indicate that post-COVID immune perturbations occur across age groups, although developmental immunology may influence both susceptibility and clinical presentation. Longitudinal pediatric studies remain limited and are needed to better define immune trajectories and long-term outcomes in children and adolescents.

The findings of this review have important translational implications. Persistent immune activation, endothelial dysfunction, viral persistence, and autoimmune features provide potential rationale for targeted therapeutic strategies, including immunomodulatory therapies, antiviral approaches, and interventions aimed at vascular stabilization or thromboinflammatory pathways[42]. Rehabilitation and supportive care strategies also remain essential for addressing functional and cognitive impairment in affected individuals. However, robust evidence from large prospective studies and randomized clinical trials is still lacking, and treatment approaches should be interpreted cautiously.

This review also highlights several important research gaps. Heterogeneity in long COVID definitions, follow-up duration, symptom characterization, and immune-assay methodologies limits comparability across studies and complicates interpretation of findings. Standardization of diagnostic criteria and immune profiling approaches will be essential to improve reproducibility and enable future meta-analyses. In addition, important factors, including sex-based immune differences, genetic susceptibility, aging-related immune changes, and variant-specific immune responses, remain insufficiently understood^[43,44^]. Addressing these gaps through longitudinal, multi-omic, and demographically diverse studies will be critical for refining mechanistic understanding and advancing precision approaches to long COVID management.

Recent advances in computational biology and artificial intelligence may further support the integrative analysis of complex immunological and molecular datasets generated in long COVID research. Emerging approaches using multidimensional immune profiling, proteomics, metabolomics, and machine-learning-based analyses may help identify biologically meaningful subgroups, improve biomarker discovery, and guide personalized therapeutic strategies^[23,45–48^]. Future integration of these approaches with translational and clinical research may enhance the understanding of post-COVID immune dysregulation and support the development of targeted interventions.

Strengths and limitations

This review provides a comprehensive synthesis of current evidence on long COVID immunology and post-acute sequelae, integrating findings from 23 studies across multiple countries and diverse populations, including adults and children with varying acute COVID-19 severity. By organizing the evidence thematically – covering humoral and cellular immunity, autoimmune signatures, proteomic and metabolic dysregulation, viral persistence, thromboinflammatory processes, and pediatric long COVID – the review captures the multidimensional nature of post-COVID immune alterations. The inclusion of both longitudinal and cross-sectional studies allows consideration of early, intermediate, and long-term immune trajectories, while peer-reviewed references ensure data traceability and reliability.

However, several limitations must be acknowledged. As a narrative synthesis, the search was not systematic, so relevant studies may have been missed. The included studies are heterogeneous in sample size, follow-up duration, definitions of long COVID, population demographics, and immune assessment methods, which limits direct comparison and meta-analytic evaluation. This methodological and clinical heterogeneity may influence the consistency of reported immune signatures, affect the reproducibility of findings across cohorts, and necessitate cautious interpretation when drawing overarching mechanistic conclusions. Formal risk-of-bias assessments were not performed, and mechanistic interpretations are largely inferred from observational associations rather than established causal pathways. Furthermore, many studies had small sample sizes or single-center designs, particularly in pediatric cohorts, which may reduce generalizability. These limitations highlight the need for larger, standardized, and mechanistic studies to validate and expand the findings presented here.

Conclusion

Long COVID is a multi-system syndrome driven by a core set of primary immunological mechanisms: persistent immune activation with T-cell exhaustion, autoimmunity (particularly anti-cytokine and anti-interferon antibodies), and thromboinflammation. Secondary phenomena, including complement dysregulation, metabolic alterations, and endothelial dysfunction, likely represent downstream consequences or modifiers of these primary drivers rather than independent causes. Among the candidate biomarkers, autoantibody profiles and cytokine panels show the most promise for clinical translation: autoantibody testing could stratify patients for immunomodulatory therapy, while elevated IL-6, IL-10, TNFα, and IP10 may serve as monitoring tools for disease activity and treatment response. Early viral persistence (spike antigen, SARS-CoV-2 RNAemia, and EBV reactivation) identifies high-risk individuals who might benefit from antiviral strategies.

While heterogeneity in study design and immune assessment limits comparability across studies, the evidence supports integrating longitudinal immunological assessments and biomarker profiling into clinical management frameworks. Future studies should prioritize standardized longitudinal designs with harmonized immune assays to validate these biomarkers, establish clinically meaningful thresholds, and test mechanism-targeted interventions. Such approaches will be critical for reducing morbidity and improving functional recovery in both adults and children affected by long COVID.

Ethical approval

Not applicable.

Consent

Not applicable.

Sources of funding

The authors received no specific funding for this work.

Author contributions

E.G.A. conceived the study, designed the review framework, and contributed to the overall organization of the manuscript. E.G.A. and S.A.H. conducted the literature search and independently screened the identified studies. S.W.M., A.H.A., and K.N.T. participated in study screening, data verification, and contributed to the interpretation of the findings and manuscript revision. Any discrepancies between E.G.A. and S.A.H. during study selection were resolved through discussion, with final adjudication provided by E.H.G. E.H.G. supervised the overall review process, provided critical input during data interpretation, and contributed to manuscript refinement. All authors reviewed and approved the final version of the manuscript and agreed to be accountable for all aspects of the work.

Conflicts of interest disclosure

The authors report no conflicts of interest in this work.

Research registration unique identifying number (UIN)

None.

Guarantor

Eyob Girma Abera.

Provenance and peer review

Not commissioned, externally peer-reviewed.

Data availability statement

All relevant data are within the manuscript.