Dengue Literature Review

Autoimmunity in Dengue

14 min read

Autoimmunity in Dengue

Overview

Dengue virus infection can trigger transient or prolonged autoimmune responses. Proposed mechanisms include molecular mimicry (cross-reactive antibodies targeting host proteins such as endothelial cells or plasminogen), immune complex (IC) deposition driving tissue inflammation, and persistent Fc receptor-mediated cytokine release. Reported autoimmune sequelae include Guillain-Barré syndrome, retinal vasculitis, IC deposition, and — as a broader syndrome — persistent musculoskeletal and neurological symptoms following acute infection (see Post-Dengue Syndrome). The broader mechanistic framework — including bystander activation and epitope spreading — applies to dengue as it does to other viral infections (see Infection-Triggered Autoimmunity).

Key Points from Literature

ANA in Acute Dengue — Direct HEp-2 Measurement

Chatterjee2024 - ANA Detection in Dengue Kolkata provides the first ANA measurement during acute dengue using the HEp-2 gold standard IIFA (Kolkata, India; n=135 dengue-confirmed, 126 controls):

  • 54.8% of dengue-positive patients ANA-positive by IIFA vs. 10.3% controls (p < 0.001)

  • 18.5% by LIA (18 specific autoantibodies) vs. 7.1% controls (p = 0.009)

  • The ~3:1 IIFA:LIA ratio reveals that the majority of dengue-associated ANAs are non-specific and do not correspond to established autoimmune disease specificities

  • Of 7 autoimmune disease categories tested (SLE, Sjögren’s, CREST, MCTD, PBC, myositis, non-rheumatic diseases), only MCTD (multivariate p = 0.041) and autoimmune myositis (multivariate p = 0.018) were significantly elevated in dengue patients — important caveat: wide confidence intervals (OR 14.01, 95% CI 2.197–89.215 and OR 18.37, 95% CI 2.746–122.944 respectively) and a small sample (n=135) make these hypothesis-generating rather than confirmed findings

  • Among symptomatic patients followed 2 years post-infection, 76.9% (20/26) showed ≥1 autoimmune marker alteration, including elevated IC and CRP in 42.3% each and ANA positivity in 23.1% (see Garcia2009 - Long-term Clinical Symptoms Post-Dengue).

  • Elevated IC correlated with higher anti-dengue IgG titers (p = 0.042), consistent with ongoing antibody production driving IC formation.

  • Elevated CRP suggests immune-mediated (non-degenerative) tissue damage; proposed mediators include IL-6, IL-1, IFN-α/β/γ, and TNF, released via persistent FcγRIIa activation by IC in neutrophils, macrophages, and dendritic cells.

  • The FcγRIIa-HH genotype links impaired IC clearance to autoimmune symptoms: failure to clear IC may cause accumulation in tissues or circulation, triggering inflammatory pain and systemic symptoms (see FcγRIIa Receptor).

ANA Positivity in the Context of Viral Infections

Berlin2007 - Autoantibodies in Nonautoimmune Individuals during Infections provides a direct comparator: ANA positivity was 21.7% during acute viral infections (HAV, HBV, HCV; ELISA ANA 8 Pro at 1:100; P<0.013 vs. controls) vs. 3.8% in healthy blood donor controls. This is strikingly similar to Garcia2009’s 23.1% post-dengue finding. Key interpretive distinction: Berlin2007 measured ANA during the acute infection phase, whereas Garcia2009 measured 2 years post-infection — persistence at 2 years is more clinically significant than transient acute-phase elevations.

Evidence from the same literature suggests infection-triggered ANA can be transient: a cited study found ANA in 20.5% of acute viral hepatitis patients, falling to 6.4% in convalescence. Whether dengue-associated ANA similarly resolves post-infection, or persists due to mechanisms such as ongoing immune complex formation (elevated IC in Garcia2009 correlating with anti-dengue IgG titers), is a key open question.

The three mechanisms by which infections trigger autoimmunity — molecular mimicry, bystander activation, and epitope spreading — are reviewed in Johnson2022 - Infectious Diseases Autoantibodies and Autoimmunity. For dengue specifically, molecular mimicry (dengue NS1 cross-reactivity with endothelial and platelet proteins) and FcγRIIa-mediated IC clearance failure (see FcγRIIa Receptor) are the most directly evidenced mechanisms.

NS1 Molecular Mimicry — Mechanism and Targets

The NCKU group (Lin, Lei et al.) has provided the most detailed experimental evidence for dengue-specific molecular mimicry. Anti-NS1 antibodies cross-react with platelet and endothelial cell surface proteins, causing immune-mediated platelet destruction (thrombocytopenia) and vascular leakage (haemorrhagic syndrome). Key established points:

  • Anti-platelet and anti-endothelial autoantibody levels are higher in DHF/DSS than DF; absorption with NS1 antigen confirms anti-NS1 accounts for cross-reactivity (see Lin2006 - Autoimmune Pathogenesis in Dengue Virus Infection)
  • Specific host proteins cross-targeted: PDI (protein disulfide isomerase), vimentin, ATP synthase β-chain, and HSP60; the responsible NS1 domain is the C-terminal region (aa 311–352) — deletion abolishes cross-reactivity (see Lin2011 - Molecular Mimicry Virus Host Dengue Pathogenesis)
  • Anti-NS1 causes endothelial apoptosis (via NO/p53/Bax/caspase-3 pathway) and inflammatory activation (NF-κB → IL-6, IL-8, MCP-1; ICAM-1↑) — two distinct endothelial pathology mechanisms (see Lin2006 - Autoimmune Pathogenesis in Dengue Virus Infection)
  • The dengue E protein contains a WGNGCG motif (aa 101–106) homologous to coagulation factors XI, X, IX, VII, thrombin, plasminogen, and tPA; anti-E Abs inhibit plasmin activity; this motif is conserved in haemorrhagic flaviviruses (JEV, WNV, YFV) but absent in HCV — consistent with flavivirus haemorrhagic phenotype (see Lin2011 - Molecular Mimicry Virus Host Dengue Pathogenesis)
  • Dengue autoimmunity is distinctive in occurring during the acute phase of infection (unlike EBV/SLE, C. jejuni/GBS, which manifest post-infectious) — see NS1 Molecular Mimicry in Dengue for full synthesis

This NS1 molecular mimicry mechanism provides a dengue-specific explanation for why dengue-associated ANA may differ from generic viral-infection-triggered ANA: the direct NS1/host-protein cross-reactivity produces demonstrable functional damage (platelet lysis, vascular leakage) rather than merely epiphenomenal autoantibody production.

Autoantibody Kinetics and Temporal Profile

Wan2012 - Autoimmunity in Dengue Pathogenesis — the capstone review from the NCKU group — provides the most explicit characterisation of dengue autoantibody dynamics: titres peak in the acute phase, decline during convalescence, and persist for several months. This is described as “different from chronic virus infection-associated autoimmune disease.” The paper also adds new mimicry targets beyond Lin2006/Lin2011: LYRIC protein (NS1 aa 116–119), RGD structural mimicry in NS1, and capsid (C) protein as a fourth cross-reactive viral protein.

The “intrinsic ADE” hypothesis in Wan2012 links ADE to autoimmunity mechanistically: FcγR-mediated DENV entry suppresses type I IFN and promotes IL-10/Th2 skewing, creating conditions for enhanced autoantibody production alongside high viral loads.

The temporal profile — resolution within months — makes Garcia2009’s 2-year ANA persistence a significant outlier. Possible explanations: (1) FcγRIIa-HH genotype impairs IC clearance, prolonging immune stimulation; (2) ongoing anti-dengue IgG production (correlated with elevated IC in Garcia2009) sustains the autoantibody response beyond the expected window; (3) the Garcia2009 ANA may reflect a population of autoantibodies distinct from anti-NS1 cross-reactive Abs.

Case reports cited in Wan2012 document individual patients progressing from dengue to frank SLE/lupus nephritis (Rajadhyaksha 2012) and multiple autoimmune features (Jardim 2012) — though Shih2023 shows this is not detectable as a population-level risk.

Contextualising ANA Positivity Against Healthy-Population Baselines

The 23.1% ANA positivity reported in Garcia2009 must be evaluated against contemporary healthy-population reference data:

All comparisons support elevated ANA positivity in post-dengue symptomatic patients, though the exact testing dilution in Garcia2009 is not explicitly stated and the populations differ. Crucially, no contemporaneous healthy control group was tested for ANA in the Garcia2009 study.

The 2019 EULAR/ACR SLE classification criteria (see Aringer2019 - 2019 EULAR ACR SLE Classification Criteria) establish that ANA ≥1:80 at least once is a mandatory entry criterion for SLE classification; however, the low specificity of ANA means that most ANA-positive post-dengue patients do not have SLE — ANA here likely reflects broader non-specific or infection-triggered autoimmune activation.

Population-Level Autoimmune Disease Risk After Dengue

Li2018: Broad Autoimmune Risk Claimed (ICD-Coded Cohort, Taiwan NHIRD)

Li2018 - Increased Risk of Autoimmune Diseases in Dengue used the Taiwan NHIRD to identify 12,506 hospitalised dengue patients (ICD-9 coded, 2000–2010) matched 1:9 to controls, with 3-year follow-up. Key findings:

  • Overall aHR 1.88 (95% CI 1.49–2.37; p<0.001) for any autoimmune disease
  • Most frequent outcome by case count: primary adrenocortical insufficiency (n=19; aHR 2.05)
  • Significant associations (without multiple comparison correction): SLE (aHR 3.50; n=13), ADEM (aHR 3.80; n=9), systemic vasculitis (aHR 3.70; n=4), myasthenia gravis (aHR 5.35; n=4)
  • GBS was NOT significant (aHR 0.97) — a null finding that persists in the stronger Shih2023 study
  • Proposed mechanism: NS1 activates TLR2/6 and TLR4; TLRs expressed on adrenocortical cells, endothelial cells, and chondrocytes → tissue-specific autoimmune pathology. This extends the established NS1-TLR4 pathway (see Guzman2016 - Dengue Infection) to a new organ target (adrenal cortex), though the hypothesis is unvalidated in experimental models.
  • Substantial residual confounding: dengue group 6 years younger (45.1 vs. 51.1 yrs), more male (50.5% vs. 44.5%), and urban (SMD 0.30–0.38 after matching)

Shih2023: Broad Claim Refuted (Lab-Confirmed Cohort, Taiwan NHIRD)

Shih2023 - Autoimmune Disease Risk After Dengue provides the most rigorous epidemiological test using the same NHIRD, but restricted to laboratory-confirmed dengue. Using 63,814 lab-confirmed dengue patients and 255,256 matched controls, followed for a mean of 4.57 years, the study found:

  • Overall autoimmune disease risk marginally elevated (aHR 1.16; P = 0.0002), but clinically small
  • Only ADEM survives Bonferroni correction across 14 specific outcomes (aHR 2.72; P < 0.0001). All other diseases — including SLE, RA, Sjögren’s, GBS, post-infectious arthritis, and MG — non-significant after correction
  • The ADEM risk is entirely acute: 16 dengue patients (0.025%) developed ADEM in the first month vs. 0 controls (HR >9999); excess disappeared after month 1
  • Identified three methodological problems in Li2018: (1) only 51.4% of hospitalised ICD-coded “dengue” patients were actually lab-confirmed — ~half the Li2018 cohort may not have had dengue; (2) no correction for multiple comparisons across 20+ diseases; (3) restriction to hospitalised patients introduces selection bias

Cross-Study Convergences

The comparison between Li2018 and Shih2023 reveals two findings robust enough to survive completely different methodological approaches:

  1. ADEM risk is real and consistent — aHR 3.80 (Li2018) and aHR 2.72 (Shih2023), both significant, with independent case counts from different study populations. Of all specific autoimmune diseases tested across both papers, ADEM is the one finding that holds.
  2. GBS risk is not elevated — aHR 0.97 (Li2018, non-significant) and non-significant in Shih2023. The GBS-dengue association seen in case reports does not reach statistical significance in either population-based study.

Key implication for this wiki’s core research thread: The Garcia2009 and Lin group findings (elevated autoimmune markers, NS1 molecular mimicry) reflect genuine acute-phase biological phenomena — but they do not translate into lasting autoimmune disease at the population level. The dengue-driven autoimmunity is predominantly transient. The one exception — ADEM — is consistent: it is an acute demyelinating response, not a chronic autoimmune disease.

Indirect Evidence: Sex and Severity as Cross-Outcome Modifiers

Seet2007 - Post-Infectious Fatigue Syndrome in Dengue does not measure autoimmune markers and is not directly about autoimmunity — but it provides two findings that are relevant here as indirect evidence of a shared host-immunological axis across post-dengue outcomes:

1. Severity-independence replication: Post-dengue fatigue at 2 months was not significantly associated with DHF vs. DF (p = 0.855). This replicates Garcia2009’s finding that post-dengue autoimmune markers at 2 years were also not significantly predicted by acute severity (p = 0.086). Two independent cohorts, different countries (Singapore vs. Cuba), different serotypes (DEN-1 vs. DENV-4), different follow-up durations (2 months vs. 2 years), different outcomes (fatigue vs. ANA/IC/CRP) — both show the same pattern. This strengthens the inference that post-dengue immune sequelae are governed by host-intrinsic factors rather than the degree of viral pathological burden.

2. Female sex as an outsized modifier: In Seet2007, female sex predicted post-dengue fatigue with OR 9.687 (95% CI 1.546–60.684). Garcia2009 found female predominance in post-dengue autoimmune markers (65.7% of women vs. 36.7% of men, p = 0.008). The female sex excess in ANA positivity is well-established in healthy populations (see Antinuclear Antibodies). The convergence of the female excess across three separate observations — (a) healthy-population ANA, (b) post-dengue autoimmune markers (Garcia2009), (c) post-dengue fatigue (Seet2007) — suggests sex-linked immunological mechanisms (hormonal regulation of FcγR expression, Th1/Th2 balance, B cell tolerance thresholds) may be the common thread modifying the post-dengue immune response. This does not establish mechanism but it refines the phenotyping: whatever the shared pathophysiology of post-dengue fatigue and post-dengue autoimmunity, it is substantially female-biased.

Contradictions & Debates

  • No asymptomatic control group was tested for autoimmune markers in Garcia2009 - Long-term Clinical Symptoms Post-Dengue, making it uncertain whether these findings are specific to symptomatic sequelae or occur more broadly in dengue-recovered individuals.
  • The appropriate healthy-population comparator for the Garcia2009 ANA finding is uncertain: the dilution used is not clearly stated, the population is Cuban (not US or European), and ANA prevalence has been rising over time (Dinse2022). The most conservative comparison (against the contemporary 16.1% US rate) still shows elevation, but by a smaller margin than earlier estimates suggested.
  • Substrate incompatibility (critical): Garcia2009 used rat liver tissue as the IIF substrate, whereas all healthy-population reference studies in this wiki use HEp-2 cells (see Indirect Immunofluorescence ANA Test). Rat liver is less sensitive — it detects fewer ANA specificities and likely underestimates ANA prevalence relative to HEp-2. This means the 23.1% post-dengue figure is a floor estimate; had HEp-2 been used, the rate would likely be higher. The cross-study comparisons (23.1% vs. 13.8–16.1%) are therefore directionally valid but the gap may be larger than reported.
  • The Berlin2007 viral-infection ANA rate (21.7%) closely matches Garcia2009 (23.1%), raising the question of whether dengue-associated ANA is qualitatively distinct from generic viral-infection-triggered ANA, or a manifestation of the same phenomenon measured at an unusually late (2-year) time point.
  • Johnson2022 documents that bystander polyclonal activation during severe viral infection (COVID-19) does NOT significantly elevate ANA above ICU-level controls, suggesting that cytokine storm alone is insufficient to explain ANA elevation — making dengue-specific mechanisms (molecular mimicry, FcγRIIa-driven IC persistence) the more plausible explanations for the Garcia2009 finding.
  • Discrepant specific autoimmune signals: Shih2023 (Taiwan, n=63,814 lab-confirmed, population-based) found only ADEM significantly elevated after dengue. Chatterjee2024 (India, n=135, hospital-based) found MCTD and myositis significantly elevated. These findings are not necessarily contradictory — different study designs, settings, populations, and outcome definitions — but they cannot be reconciled into a single picture yet. Shih2023 may have lacked statistical power for rare rheumatological outcomes (MCTD and myositis cases would be very few in any cohort). Alternatively, the Chatterjee2024 signal may reflect setting-specific factors (endemic dengue, hospital enrichment of severe/unusual cases).
  • Unresolved tension: Garcia2009 found elevated ANA, IC, and CRP at 2 years post-dengue (23.1% ANA positivity), yet Shih2023 finds no elevated incidence of clinical autoimmune diseases over a mean 4.57-year follow-up. These findings appear contradictory but may be reconciled: (1) ANA positivity is not equivalent to autoimmune disease — most ANA-positive individuals, especially at low titres, never develop clinical disease; (2) the Garcia2009 autoimmune markers were measured in a highly selected symptomatic subset with multiple prior infections, while Shih2023 covers all dengue patients including mild/asymptomatic; (3) elevated IC and CRP may reflect post-infectious inflammation rather than true autoimmunity. The Garcia2009 autoimmune marker elevations may represent a biologically real phenomenon that nonetheless falls below the threshold for clinical disease in the vast majority of patients.

Sources

Graph View

Backlinks