state

Wiki State

Persistent operational context for the dengue literature review. Read this at the start of every session. Update it after every session.

---

Current Focus

• PRIMARY research thread (set 2026-06-03): is there a correlation between ANA and chronic fatigue in dengue? New direction — supersedes the prior open-ended “autoimmunity / ANA dynamics” framing, which never had a defined clinical endpoint (“why ANA in dengue” was undefined; the literature was reviewed generally). The question bridges two clusters that grew up separately in this wiki: the post-dengue fatigue cluster (Post-Dengue Syndrome ← Seet2007, Garcia2009, Guzman2016, Shih2023) and the ANA / autoimmunity cluster (Antinuclear Antibodies, Autoimmunity in Dengue, ANA and Dengue - Review V2.0). Curator is collecting papers to ingest into this thread (see Queue). First ingest (2026-06-03): Hertanti2024 - Fatigue and Post-Infectious Fatigue in Dengue — the largest fatigue/PIF meta-analysis (40 studies, 38,406 patients). It defines the gap rather than filling it (autoimmunity hypothesized as 1 of 4 PIF mechanisms; zero ANA measured) and recalibrates the shared female-sex effect downward (see below).
  • State of the bridge: Garcia2009 - Long-term Clinical Symptoms Post-Dengue is the only wiki source that co-measures ANA and chronic post-dengue symptoms in one cohort (Cuba, DENV-4, 2-yr follow-up): asthenia 23.0% of symptomatic patients; ANA-positive 23.1% of a 26-patient autoimmune-marker subset. But it does not test the ANA↔fatigue correlation: (1) no individual-level link — the matching 23.0%/23.1% figures sit on different denominators and are coincidental, not correlated; (2) ANA measured only in symptomatic patients, so no control / non-fatigued comparison exists; (3) rat-liver IIF substrate (less sensitive, not HEp-2-comparable). The definitive fatigue paper Seet2007 - Post-Infectious Fatigue Syndrome in Dengue (24.4% fatigue at 2 mo, n=127) measured no autoimmune markers. → The direct ANA↔fatigue correlation is effectively untested in the current wiki — that is the gap this lane targets.
  • Shared signal = shared confounder (female sex) — recalibrated 2026-06-03: Both post-dengue fatigue (cross-study pooled female→PIF OR 1.65, 95% CI 1.27–2.14, I²=0.00 — Hertanti2024 - Fatigue and Post-Infectious Fatigue in Dengue; Garcia2009 65.7% women vs 36.7% men) and ANA positivity (2–3× higher in females — Tan1997 / Satoh2012 / Dinse2022) are female-predominant. Correction: the earlier framing leaned on Seet2007’s female OR 9.687, but that is Seet’s own adjusted single-study estimate (CI 1.5–60.7, a sparse-data artifact); the trustworthy cross-study quantity is the modest, zero-heterogeneity 1.65. Sex remains simultaneously the mechanistic hook and the principal confounder — it could independently produce both with no ANA→fatigue link — but it is a smaller lever than the wiki previously implied. Controlling for sex is still mandatory — the same FATAL-FLAW lesson from the abstract-methodology work (see supporting bullet below).
  • Headwinds already in the wiki (we argue uphill — flag honestly): (1) Shih2023 - Autoimmune Disease Risk After Dengue — no elevated long-term autoimmune disease incidence post-dengue (only ADEM, confined to month 1); ANA positivity largely does not convert to clinical disease. (2) Polyreactive-IgM hypothesis (Zhou2007 - Polyreactive Antibodies Natural Antibody Function, Chatterjee2024 - ANA Detection in Dengue Kolkata) — ~66% of dengue ANA is IIFA-positive / confirmatory-negative, possibly non-specific natural IgM. (3) Existing fatigue mechanisms in the wiki (CD4/CD8 inversion, C3a/C5a, HPA axis — Seet2007) are not ANA-mediated. Sharper framing: does ANA mark a subset of post-dengue fatigue, or is it a sex-linked bystander?
  • Timepoint-mismatch problem: No two ANA/fatigue datapoints share a follow-up window — Seet2007 = 2 mo, Gawali2021 = 6 mo, Garcia2009 = 2 yr. Acute vs convalescent vs persistent ANA cannot currently be aligned to a fatigue trajectory.
  • Open sub-questions: (1) Is ANA-positivity associated with persistent fatigue at the individual level, controlling for sex + age? (2) Which ANA — acute, convalescent, or persistent — tracks fatigue? (3) Does confirmed ANA (HEp-2 + line-immunoassay) behave differently from non-specific IIFA-only ANA against fatigue? (4) What do non-dengue post-infectious fatigue / ME-CFS / post-COVID autoantibody literatures offer as comparators?
• Supporting infrastructure (demoted from primary 2026-06-03) — matched case-control methodology, now in service of the ANA↔fatigue study (above). The curator is preparing an abstract reporting an age-matched 1:1 case-control study of ANA patterns (IIFA, ICAP nomenclature, ≥1:80 screening dilution, HEp-20 substrate) in dengue cases (DF/DHF) vs. healthy controls — headline result 40.0% vs. 17.5% ANA positivity (McNemar exact p=0.049 / Fisher exact p=0.047; matched OR ~3.1–3.25). The wiki’s job on this thread is to build the methodological evidence base needed to scrutinise and strengthen that design before publication. Two analysis pages anchor it: Methodology Critique - ANA IIF Abstract Draft (11 issues — 1 fatal flaw, 5 major) and Dengue vs Healthy Controls - Analysis Methods Workflow (operational checklist derived from the critique). Supporting study-design axis: Age-Matched Case-Control Analysis, drawing on Pearce2016 - Analysis of Matched Case-Control Studies and Iwagami2022 - Introduction to Matching in Case-Control and Cohort Studies.
  • Headline critique findings: (1) FATAL — sex confounding: case sex is unrecorded and controls are predominantly female; ANA prevalence is 2–3× higher in females (Tan1997 / Satoh2012 / Dinse2022), so the direction and magnitude of the primary result are unknown until sex is controlled. (2) Analytical strategy: Fisher’s exact is invalid on matched data; McNemar’s is valid but suboptimal; conditional logistic regression is the only method both Pearce2016 and Iwagami2022 endorse and should be the sole primary analysis. (3) 1:1 ratio wasted power — only 40 of 90 available controls used; 1:2–1:3 is feasible and would have been more robust in the sparse 18–22yr stratum. (4) 39.4% case dropout (26/66, concentrated at ages 18–22) — matched cohort skews older; matched-vs-unmatched ANA comparison is a mandatory disclosure. (5) “DF drives the signal” is overclaimed — every DF/DHF comparison is non-significant (DF vs controls p=0.055). (6) No febrile-control arm to separate dengue-specific ANA from the ~20% generic acute-viral baseline (Berlin2007 / Codes2002).
  • Immediate next actions: recover case sex data → re-match on age+sex; switch the primary analysis to conditional logistic regression; report matched-vs-unmatched ANA positivity; reframe the DF/DHF comparison as exploratory. See the 2026-05-25 Watch Items below for the live issue tracker.
  • Unresolved methodological dispute feeding this thread: Pearce2016 vs. Iwagami2022 disagree on whether unconditional logistic regression is valid for matched case-control data (documented in Age-Matched Case-Control Analysis Contradictions & Debates). Conditional logistic regression sidesteps it. A third methodology paper or a simulation study would resolve it.
• Secondary thread — autoimmunity in dengue / ANA dynamics (the scientific substrate the ANA↔fatigue question and the abstract both draw on): ANA and Dengue - Review V2.0 reviewed and corrected through §3.2 (2026-04-19). Corrections applied: (1) Vo2020 cohort description fixed (n=32 DENV-infected, not n=40; 21 hospitalized: 13 DF, 8 DHF; 11 asymptomatic; 6 primary / 26 secondary); (2) 128 vs 123 antigen array discrepancy resolved (128 total, 123 analysed); (3) PABs reframed as amplified natural polyreactive IgM — not autoantibodies per Zhou2007; (4) Vo2020 hospitalized IgG statement integrated into §3.4 consumption model; (5) Chatterjee2024 dilution limitation rewritten as formal analytical prose; (6) all 46 curator highlights resolved and marks stripped. Review is analytically clean through §3.2; curator has not yet reviewed §3.3 onward. Remaining open ANA gaps: (1) acute-to-chronic HEp-2 IIFA trajectory, (2) NS1 homology with nuclear antigens, (3) FcγRIIa-ANA correlation, (4) longitudinal anti-platelet/anti-endothelial autoantibody titres. The macrophage-driven autoimmunity sub-thread (MAS, ANA-negative) is mechanistically distinguished from the B-cell autoantibody axis.
• Thrombocytopenia mechanism sub-thread (updated 2026-04-17, Saito2004 ingest): Saito2004 extends the bifurcation model: secondary infection has BOTH PAIgG (anti-dengue IgG IC; Oishi2003) and PAIgM (anti-dengue IgM IC; Saito2004) — neither is anti-self. Primary infection has IgM anti-platelet autoantibody (NS1 mimicry; Lin2001). The IgM species are mechanistically opposite: Lin2001’s primary-infection IgM is anti-self; Saito2004’s secondary-infection PAIgM is anti-viral. PAIgM is completely FcγR-independent (IgM pentamer) and predicts DHF with 92.1% specificity. The FcγRIIa paradox is now even sharper: both secondary-infection pathways bypass FcγR at platelet docking; PAIgM bypasses it entirely. Garcia2010’s OR 10.56 must operate through a completely non-platelet pathway.
• Tertiary thread: asymptomatic vs. symptomatic dengue — host genetics (FcγRIIa), immune signatures, and what distinguishes silent from clinical infection. Only 2 sources (Garcia2010, Sungnak2025); under-developed relative to the autoimmunity thread.

Queue

Papers waiting to be ingested (add new entries at the top):

Hertanti2024 ingested 2026-06-03 — first paper in the ANA↔fatigue thread; defines the gap (no ANA measured). Ingest was interrupted by a Bun crash mid-write and completed on resume.

Pearce2016 ingested 2026-05-24 — removed from queue.

Newly discovered raw PDFs (scope unknown — prioritise after review):

• raw/Ghorai2024.pdf ingested 2026-04-19
• raw/chaturvedi2001.pdf ingested 2026-04-18
• raw/santosa2012.pdf ingested 2026-04-18
• raw/Farias2024.pdf ingested 2026-04-18
• raw/Hung2008.pdf ingested 2026-04-19
• raw/Velazqueza2017.pdf ingested 2026-04-17
• raw/cheng2015.pdf ingested 2026-04-18
• raw/rajadhyaksha2012.pdf ingested 2026-04-17
• raw/jardim2012.pdf ingested 2026-04-18
• raw/codes2002.pdf ingested 2026-04-18

Gawali2021 ingested 2026-04-17 — removed from queue.

Bhatt2020 ingested 2026-04-17 — removed from queue. Dejnirattisai2010 ingested 2026-04-17 — removed from pending. Vo2020 ingested 2026-04-17 — removed from queue. Saito2004 ingested 2026-04-17 — removed from queue. Oishi2003 ingested 2026-04-15 — removed from queue. Lin2001 ingested 2026-04-15 — removed from queue. Morel2014 ingested 2026-04-15 — removed from queue. Palacios2016 ingested 2026-04-15 — removed from queue. Bos2025 ingested 2026-04-14 — removed from queue. Li2018 ingested 2026-04-13 — removed from queue. Seet2007 ingested 2026-04-13 — removed from queue.

Decisions

Structural and workflow decisions with rationale. Append-only.

[2026-04-14] Lint run completed; Bos2025 ingest had 5 metadata/connectivity gaps

Decision: Lint identified and fixed 5 issues from the Bos2025 ingest: (1) index section header counts wrong (Sources 19→20, Geography 5→6); (2) ADE page frontmatter not updated (sources 6→7, updated date stale); (3) Seet2007 misclassified under Related Pages in Autoimmunity in Dengue rather than Sources; (4) Secondary Dengue Infection not updated with Bos2025 kinetics data; (5) IgM-IgG Serology ELISA page missing Bos2025 as source. Three issues flagged but not fixed: ELISA Inhibition Method attribution uncertain (needs PDF verification), Reading Plan orphan, Jhonson2022 filename typo (immutable raw/). Why: Pattern: during Bos2025 ingest, the log correctly recorded what should happen (ADE 6→7, etc.) but some frontmatter edits and secondary-concept-page updates were not executed. How to apply: After any ingest, cross-check the log’s “Updated” list against every page listed in the source’s “Concepts Addressed” and “Entities Mentioned” — both content AND frontmatter (sources count, updated date) must be consistent.

[2026-04-14] “update web” requires clearing public/ when new geography folders are added

Decision: When a new geography page is added (or any new subdirectory), the Quartz build may fail with ENOTEMPTY: directory not empty, rmdir public/geography. Fix: rm -rf public/ before running sync-and-build.ps1. The script itself does not clear the public directory; the Quartz incremental clean can fail on new subdirectories. Rationale: Discovered during Bos2025 ingest when Nicaragua.md created the first new geography in a session with an existing public/ directory. How to apply: If update web fails with an ENOTEMPTY error, run rm -rf webforshare/public/ and retry.

[2026-04-14] Bos2025 preprint ingested with explicit preprint warnings

Decision: Ingested Bos2025 (medRxiv preprint, doi: 10.1101/2025.08.11.25333449) with prominent PREPRINT warnings in the source page, all updated concept/entity pages, and the Notable Findings entry. Citation counts are 0/0 as expected for an unreviewed preprint. Rationale: The paper contains the most detailed longitudinal antibody kinetics data in the wiki and fills important gaps (XR E-IgG trajectories, NS1-IgG waning kinetics, Nicaragua geography). However, preprint findings require explicit caveating because they have not undergone peer review. Mechanistic claims derived from these kinetics (especially ADE risk reframing) are treated as hypotheses, not established findings.

[2026-04-17] Bhatt2020 ingested; sfRNA/TRIM25/RIG-I IFN suppression + Katzelnick ADE window + OAS mechanism

Decision: Ingested Bhatt2020 (Current Microbiology 2021; 78:17–32; DOI 10.1007/s00284-020-02284-w; 300/294 citations; Kasturba Medical College, Manipal, India). Key structural outcomes: (1) sfRNA→TRIM25→RIG-I K172 ubiquitination block mechanism for IFN suppression — first molecular account of DENV innate immune evasion in this wiki; (2) Katzelnick 2017 quantitative ADE window (1:21–1:80 peak enhancement, n=6684 Nicaraguan children) — ADE model now has a bounded enhancement zone, not just a qualitative sub-threshold risk; (3) OAS T cell mechanism detail — low-avidity CD8+ expansion, cytolytic loss, TNF-α/IL-6 excess, delayed clearance; (4) NS1 replication complex cofactor role; (5) glycocalyx degradation dual mechanism (heparanase + cathepsin L); (6) MIF-autophagy viral amplification. New notable finding flagged: Katzelnick ADE window. India: sources 1→2 (Manipal added). 16 pages updated. Why notable: Bhatt2020 is a review but contributes three distinct mechanistic gaps: the IFN suppression mechanism (sfRNA) was completely absent from the wiki; the quantitative ADE window converts a qualitative model into a bounded testable prediction; the OAS mechanism detail provides the cellular basis for what was previously described only conceptually. How to apply: When ingesting future papers on ADE or vaccine safety, check whether antibody titres are reported in absolute numbers that can be benchmarked against the 1:21–1:80 window. When ingesting IFN/innate immunity papers, sfRNA TRIM25/RIG-I is now the established wiki baseline for DENV IFN evasion.

[2026-04-17] Dejnirattisai2010 ingested; anti-prM as dominant ADE-promoting antibody class

Decision: Ingested Dejnirattisai2010 (Science 2010; 328:745-748; DOI 10.1126/science.1185181; 891/820 citations). Key structural outcome: anti-prM antibodies are the dominant structural antibody class (~60%) in DENV-infected humans, fully cross-reactive across all four serotypes, incapable of complete neutralisation (structural ceiling from incomplete prM cleavage), and mediate up to 10^5-fold ADE in primary monocytes/DCs. New entity page: prM Protein. 16 pages touched total. Why notable: This qualitatively changes the ADE model: the dominant non-neutralising cross-reactive antibody pool is anti-prM, not sub-threshold anti-E. Anti-prM cannot neutralise regardless of titre — it is not a sub-threshold problem but a structural one. All current vaccines use native prM and will prime this response. The wiki’s prior ADE discussion was entirely focused on anti-E; this is now corrected. How to apply: When ingesting future ADE or vaccine immunogenicity papers, check whether anti-prM titres are measured separately from anti-E; many studies only report anti-E neutralisation. Papers reporting only anti-E titres should be flagged as incomplete immunogenicity readouts.

[2026-04-17] Vo2020 ingested; primary>secondary IgG autoantibody inversion documented; Cambodia and Autoantigen Microarray pages created

Decision: Vo2020 (autoantibody profiling in Cambodian pediatric dengue, n=40, protein microarray) ingested. Key structural outcome: (1) primary > secondary IgG autoantibody inversion added to Secondary Dengue Infection and Autoimmunity in Dengue as a flagged finding. (2) Nuclear antigen IgG-platelet count positive correlation in DHF (consumption model) documented in Autoimmunity in Dengue and Dengue Pathophysiology. (3) New method page (Autoantigen Microarray) created. (4) New geography page (Cambodia) created. 18 pages touched total. Why: The primary>secondary IgG inversion is directly counter to severity-centric ADE predictions and represents a genuinely new pattern in this wiki, warranting a Notable Finding entry. The nuclear antigen correlation with platelets provides a mechanistic bridge between the ANA thread and the thrombocytopenia thread not previously documented. How to apply: When ingesting future autoantibody profiling studies, check whether the primary>secondary IgG breadth inversion replicates or fails to replicate; the Vo2020 cohort is too small to treat this as established.

[2026-04-17] Saito2004 ingested; bifurcation model fully specified; new method page created

Decision: Saito2004 (PAIgG and PAIgM in Secondary Dengue) ingested. Key structural outcome: the thrombocytopenia bifurcation model is now precisely characterised — primary infection = IgM anti-platelet autoantibody (NS1 mimicry; Lin2001), secondary infection = PAIgG + PAIgM anti-dengue immune complexes (Oishi2003 + Saito2004), both FcγRII-independent at platelet docking, PAIgM additionally FcγR-independent at clearance. A new method page (Platelet-Associated Immunoglobulin ELISA) was created to cover both Oishi2003 and Saito2004 — this method was missing from the wiki despite Oishi2003 having been ingested. Oishi2003 source page updated retroactively to link to it. 15 pages touched total. Why: Saito2004 was the direct follow-on to Oishi2003 by the same group; together they form a complete picture. The IgM distinction (autoantibody vs. immune complex in primary vs. secondary) is a critical nuance that was not visible until Saito2004 was ingested. How to apply: Any future paper reporting PAIgM elevation in dengue must be evaluated for whether eluate specificity was tested; without eluate data, PAIgM elevation is ambiguous between the two IgM mechanisms.

[2026-04-18] ANA Review V2.0 created; all 37 sources incorporated; web deployed

Decision: Created ANA and Dengue - Review V2.0 as a ground-up rewrite superseding the incrementally revised V1.x series. Key structural improvements over V1.x: (1) §IV.2 on anti-prM as a second molecular mimicry arm (Dejnirattisai2010 — entirely absent from V1.x); (2) §V standalone thrombocytopenia bifurcation model with therapeutic implications; (3) §X vaccine design implications (NS1 P311–330 exclusion, anti-prM in all current platforms, Katzelnick ADE window); (4) §XI acute-phase timing comparative table; (5) Li2018, Pang2017, Bhatt2020, Sungnak2025 properly integrated (referenced but not listed in V1.x); (6) Q numbering corrected 1–16 (V1.x had Q14 before Q13); (7) clean document, no revision notes. 37 sources = all wiki sources. V1.x preserved at ANA and Dengue - A Literature Review. Web deployed: bb64e2d. Why: V1.x had grown through 5 patch-on revision passes and had accumulated structural debt (revision notes, transposed Q numbering, missing Li2018 in sources). A V2.0 from scratch integrates everything coherently and will serve as the stable synthesis base for future ingests. How to apply: Future ingests that touch the ANA/autoimmunity thread should update ANA and Dengue - Review V2.0, not V1.x. The V1.x page is an archived historical document.

[2026-04-18] Cheng2015 ingested; P311–330 PDI epitope mapped; ELISA method page created

Decision: Ingested Cheng2015 (Am J Trop Med Hyg 2015; doi:10.4269/ajtmh.14-0162; 22/19 citations). Key structural outcome: (1) PDI-specific NS1 cross-reactive epitope narrowed to P311–330 within the previously established C-terminal region aa 311–352; HSP60 confirmed to use a distinct unidentified NS1 epitope; anti-vimentin elevation anomalous (no NS1 correlation). (2) Primary/secondary independence of anti-endothelial autoantibodies confirmed (convergent with Lin2001 and Saito2004 — Notable Finding flagged). (3) New method page: ELISA.md (first general ELISA page in wiki). (4) Vietnam first appearance as patient source (n=15 DHF + 2 DF, Ho Chi Minh City) — no standalone page per <2 source rule. 14 pages updated; 2 new pages. Why notable: The P311–330 epitope resolution has direct vaccine design implications: NS1-based vaccine constructs containing P311–330 will generate anti-PDI cross-reactive antibodies; constructs with P311–330 deleted or mutated may avoid this. The infection-order independence finding fully decouples NS1 mimicry autoantibodies from the secondary-infection ADE escalation, reframing them as a constitutive primary-infection mechanism rather than a severity modifier. How to apply: When ingesting future NS1 vaccine or NS1 molecular mimicry papers, check whether P311–330 is specifically excluded/mutated — this is now the definitive PDI-cross-reactivity risk locus. When ingesting papers on DHF severity mechanisms, NS1 mimicry autoantibodies should be treated as infection-order independent; ADE/immune complex pathways are the secondary-infection differentiators.

[2026-04-18] Chaturvedi2001 ingested; hCF anti-cytokine autoantibody concept introduced

Decision: Ingested Chaturvedi2001 (FEMS Immunol Med Microbiol 2001; 30:181–186; DOI 10.1016/S0928-8244(00)00251-0; null/1 citations). Key structural outcome: (1) new concept page Cytotoxic Factor in Dengue created — represents the Chaturvedi group’s dengue-specific CD4+ T cell cytokine, with prominent validation caveat throughout; (2) anti-hCF autoantibodies introduced as the first protective autoantibody in the wiki (96% DF → 8% DHF grade IV positivity, P ≤ 0.001); (3) India page expanded to 5 sources with Lucknow/AIIMS 1996 epidemic as the fifth Indian setting; (4) notable finding flagged. 8 pages updated. Why: The anti-protective-autoantibody angle is genuinely novel in this wiki and distinct from all other dengue autoantibody mechanisms. The epistemic challenge (group-specific unvalidated concept) required prominent caveating throughout rather than the standard sourced-fact treatment. How to apply: Any future paper citing hCF or the Chaturvedi cytotoxic factor work should be flagged against this entry — either as independent replication (upgrade epistemic status) or contradictory evidence (document in Contradictions & Debates on the concept page).

[2026-04-19] Deep lint — Hung2008 + Ghorai2024 propagation sweep

Decision: 11 issues fixed: 6 propagation gaps (DENV-1, DENV-3, Aedes aegypti, RT-PCR, ELISA, NS1 Antigen Detection all missing citations from the 2026-04-19 ingests), 1 stale wikilink (Ghorai2024 source page [[NS1 Antigen ELISA]] → [[NS1 Antigen Detection]]), 2 index header count errors (Concepts 23→24, Methods 16→17), Paraguay.md folded into Latin America (Paraguay was the last thin geography page that should have been folded when Latin America was created 2026-04-18 but was overlooked), ANA V2.0 updated to 41 sources (Ghorai2024 added). Why: The Ghorai2024 ingest logged 13 page updates but omitted DENV-1, DENV-3, RT-PCR, ELISA, and NS1 Antigen Detection — the latter because the source page had a stale [[NS1 Antigen ELISA]] link that pointed to a non-existent page. Paraguay.md was created 2026-04-15 under the pre-rule regime; the ≥2-source rule was formalised 2026-04-17 and the Latin America fold 2026-04-18 cleaned Cuba/Nicaragua/Brazil but missed Paraguay. How to apply: When ingesting future papers that list all 4 DENV serotypes as entities, explicitly check all four serotype pages. When a Methods Used section has a wikilink to a non-existent page, the stale link silently breaks the propagation chain.

[2026-04-19] Hung2008 ingested; Southeast Asia regional page created; Vietnam standalone created; Cambodia folded

Decision: Ingested Hung2008 (Am J Infect Dis 2008; 4(1):41–49; DOI 10.3844/ajidsp.2008.41.49; 3/1 citations). Key structural outcome: (1) anti-platelet IgM elevated in both primary (infants) and secondary (children) dengue — extending Lin2001’s primary-infection finding to a secondary-infection context, suggesting anti-platelet IgM is infection-order independent; (2) anti-EC isotype shifts from IgM-only in infants (primary) to IgM+IgG in children (predominantly secondary) — the IgG addition tracks infection order as predicted by memory B-cell class switching; (3) thrombomodulin elevated in both groups as in vivo endothelial structural damage marker independent of autoantibody levels; (4) no autoantibody-severity correlation in either group — anti-EC autoantibodies at these levels do not drive clinical severity directly. Geography cascade: Vietnam = 2nd source → Vietnam.md created; Vietnam + Singapore + Philippines = 3 SEA country pages → Southeast Asia.md created (Cambodia § folded in); Cambodia.md deleted. New method page: Flow Cytometry. 19 pages updated total. Why notable: The isotype shift finding demonstrates that infection-order immunological memory operates at the autoantibody level — IgG emerges in secondary infection context without requiring disease severity as a cofactor. The absence of severity correlation for anti-EC Abs in either group stands in tension with Wan2012’s DHF > DF finding (different methods/antigens) and warrants reconciliation. How to apply: When ingesting future anti-endothelial autoantibody papers, track the measurement method (flow cytometry % reactive cells vs. ELISA OD) separately — Hung2008 and Wan2012 use different readouts and may be measuring different antibody populations. The anti-EC isotype shift (IgM→IgG in secondary infection) is now a documented feature of dengue autoimmunity with confirmed in-vitro endothelial structural damage (TM elevation) but no confirmed severity effect at the cohort level.

[2026-04-18] Curator Highlights workflow added

Decision: Added wiki/analyses/Curator Highlights.md as a light-maintenance page and a new Curator Highlights workflow to CLAUDE.md. The page aggregates all ==highlights== and “ placed by the curator in Obsidian across wiki pages. Regenerated mechanically (full overwrite) during lint and on “update highlights” command. Resolve a highlight by removing its ==...== markers in the source page — it disappears from the aggregated page on next refresh. Why: Curator uses Obsidian’s native highlighting during reading; this gives a single place to review all active annotations without hunting through individual pages. How to apply: During lint, always run step 5 (Curator Highlights refresh). On “update highlights”, grep wiki/ for == patterns, group by page, overwrite Curator Highlights.md.

[2026-04-17] CLAUDE.md schema update — 8 workflow and convention improvements

Decision: Applied 8 structural improvements to CLAUDE.md derived from CLAUDE_UPDATE.md baseline analysis. Key changes: (1) ingest --fast flag skips discussion step for bulk/clear-scope ingests; (2) git snapshot added as ingest step 1; (3) “Questions Raised” now propagates to state.md Watch Items at ingest step 13; (4) lint now runs in folder batches above 80 pages; (5) new retract/correct workflow with warning banners and inline citation flags; (6) analyses page template added to conventions; (7) evidence weighting convention added (study type + n inline on citations); (8) geography hierarchy rule: country pages ≥2 sources, regional pages trigger on next relevant ingest when ≥3 country pages in region — applied retroactively, flagging Cambodia and Nicaragua as thin pending fold. Why: Token efficiency (batched lint, fast-track ingest, template removes re-derivation cost) and inter-session memory quality (questions propagated to Watch Items, retraction tracking, geography rules prevent repeated cleanup). How to apply: ingest --fast is now a valid command. Lint is always batched. Any paper retraction uses retract [AuthorYear]. New country pages require ≥2 sources before standalone creation. Cambodia.md and Nicaragua.md fold into regional pages when Southeast Asia and Americas pages are respectively created.

[2026-04-17] ANA literature review expanded to 22 sources; web redeployed

Decision: Expanded ANA and Dengue - A Literature Review from 12→22 sources to incorporate the 5 recent ingests (Bos2025, Lin2001, Oishi2003, Morel2014, Palacios2016) plus Garcia2010 and Bruhns2009 (referenced but not previously listed). Major additions: §5.1 Lin2001 attribution correction; new §5.5 macrophage axis (ANA-negative MAS via Morel2014 + Palacios2016/Lai2012); new §6.3 NS1-IgG t½ ≈ 2.1 y kinetics + counter-trajectory rising XR E-IgG (Bos2025); §8 Morel/Talib case contrast; §10 expanded with 5 new open questions (Q7–Q11); §11 epistemic taxonomy revised with two new “Established” claims (ANA-negative MAS axis; bifurcated thrombocytopenia mechanisms) and two new “Hypothesis-generating” claims (Talib SLE de novo vs. flare; rising XR E-IgG opposing-vector autoreactivity). Committed as b791fb6 with full lint+ingest sweep; web deployed via sync-and-build.ps1 (push 988cb04 to v4 branch). Why: The 5 ingests since 2026-04-13 each touched the autoimmunity thread but the synthesis page had not been updated since the last comprehensive rewrite. Letting it lag would lose the cross-paper synthesis (especially the macrophage axis vs. B-cell axis distinction, which is only visible when Morel2014/Palacios2016 are read against Lin2001/Lin2006/Wan2012/Chatterjee2024 together). How to apply: When ≥3 sources have been ingested into a thread that has an existing analysis page, audit the analysis page for whether the new findings shift the established/probable/hypothesis-generating taxonomy. Most ingests will only add a citation; some will require new sections.

[2026-04-16] Deep lint propagation sweep — recurring miss across 5 recent ingests

Decision: Second-pass lint identified that ~18 pages were named in source pages’ Entities/Concepts/Methods sections of the last 5 ingests (Bos2025, Lin2001, Oishi2003, Morel2014, Palacios2016) but had not been visited to add the source to their own Sources lists. Fixed in this session: NS1 Protein 7→9, Autoimmunity 20→21, Cross-Reactive Abs 6→7, Secondary Dengue 7→9, Dengue Pathophysiology 2→5, ADE 7→8, Cytokine Storm 6→7, Vaccine Candidates 3→4, DENV-1 2→4, DENV-2 3→5, DENV-3 4→7, DENV-4 3→5, Aedes albopictus 1→2, IgM-IgG ELISA 3→7, NS1 Antigen 3→6, RT-PCR 2→3, Singapore 1→2. Also: Nicaragua orphan resolved (added to DENV-1, DENV-3, Cross-Reactive Antibodies Related Pages); Morel2014 erroneous DENV-1 entity link removed (paper does not specify serotype). Index per-page counts and frontmatter updated: dates synced. Why: The previous lint (2026-04-15) verified single-page metadata consistency but did not perform the inverse check — for each source’s outbound link list, has every target page added the source to its Sources? This pattern was already in feedback memory (“ingest secondary sources require full propagation”) but the lint workflow did not enforce it systematically. How to apply: Lint should run a per-source forward propagation check: for each new source page, programmatically enumerate all wikilinks under “Entities Mentioned,” “Concepts Addressed,” and “Methods Used”; then for each target page confirm the source appears in its Sources section AND that the frontmatter sources count was incremented. Missing entries are propagation gaps to fix.

[2026-04-15] Lint run completed; Palacios2016 → Dengue Neurological Complications connection was the only missed update

Decision: Lint found 1 content gap (Dengue Neurological Complications missing Palacios2016 retinal vasculitis case) and 1 log typo (NS1 Protein “5→7” should be “6→7”). Fixed both. All other frontmatter counts, link targets, and orphan checks clean. The “stale” log.md links ([[IgM/IgG Serology ELISA]], [[Lin2006]]) are inside backtick code spans — not live wikilinks in Obsidian; no fix needed. Why: Palacios2016 was ingested with minimal updates (only index.md), but its source page listed Dengue Neurological Complications as a concept addressed. The retinal vasculitis case (Chang 2007, Singapore; immune complex deposition) was never propagated to that concept page. How to apply: When ingesting a secondary source (letter, review with no original data), still audit every concept listed in “Concepts Addressed” and confirm the concept page was updated — not just the index.

[2026-04-15] Lin2001 attribution correction propagated across wiki

Decision: Lin2001 (2001) is the original source for IgM anti-platelet autoantibody findings in dengue — not Lin2006 (2006), which confirms and mechanistically extends the 2001 finding. NS1 Protein and NS1 Molecular Mimicry in Dengue pages were corrected to credit Lin2001 as the origin, with Lin2006 explicitly repositioned as the downstream confirmation paper. All pages that referenced this finding now carry both citations with the correct priority. Why: The attribution error was propagated from Lin2006/Lin2011/Wan2012 back-citing each other without surfacing the founding paper. Lin2001 precedes these by 5 years and contains the original flow-cytometry MFI data; Lin2006 adds complement-mediated lysis mechanism. How to apply: When ingesting new papers that cite anti-platelet IgM in dengue, credit Lin2001 first, then Lin2006 for mechanism detail.

[2026-04-15] Site deployed live; update web now includes git push

Decision: Site is live at https://dengue-wiki-web.pages.dev (Cloudflare Pages, GitHub repo OsandaC/dengue-wiki-web, branch v4). sync-and-build.ps1 now ends with git add -A, a timestamped commit, and git push — update web is a single end-to-end deploy command. Port numbers removed from commands.md; Cloudflare Tunnel section removed (superseded). baseUrl set to dengue-wiki-web.pages.dev. Rationale: Site is now permanently hosted; no tunnel needed. Single-command deploy reduces friction. How to apply: When curator says “update web”, run sync-and-build.ps1 — it syncs, builds, commits, and pushes. Cloudflare redeploys automatically in ~1 minute. Note: After Dependabot PRs are merged on GitHub, always run npm install in webforshare/ and push the updated package-lock.json before the next deploy, or Cloudflare’s npm ci will fail with a lock file sync error.

[2026-04-14] “update web” workflow added; commands.md created

Decision: Added update web as a curator command in CLAUDE.md — triggers sync-and-build.ps1 via Bash. Created wiki/commands.md as a permanent reference for all curator and terminal commands. Rationale: Curator needed a single-command web update workflow and a durable reference for all operational commands. (Superseded by [2026-04-15] decision above for hosting details.)

[2026-04-14] Citation counts added to index sources table

Decision: The index Sources table now includes a Citations (SS / CR) column, and a ranked-by-citation subsection is appended below it. Rationale: Quick signal for paper influence and field weight without opening individual source pages.

[2026-04-13] Quartz static site set up in webforshare/

Decision: webforshare/ folder is at Literature Review Dengue/webforshare/ (sibling of dengue-wiki/, not inside it). Contains a Quartz v4 static site generator. Wiki files are synced from dengue-wiki/wiki/ into webforshare/content/ via sync-and-build.ps1. Hosting provider not yet decided — baseUrl will be updated to the permanent Cloudflare domain once chosen. Rationale: Colleagues need to browse the wiki without Obsidian. Quartz natively supports wikilinks, frontmatter, search, backlinks, and graph view — no changes to wiki source files required.

[2026-04-13] ANA literature review completed; Notable Findings and Gap Analysis updated

Decision: Completed the ANA and Dengue literature review as a synthesis analysis page drawing on all 16 sources. Updated Notable Findings with one new synthetic entry (NS1 mimicry targets are not nuclear antigens → epitope spreading implied). Created Gap Analysis 2026-04-13 superseding the 2026-04-12 version with a full status audit. Rationale: The ANA thread had enough source coverage to support a comprehensive review. The new Notable Finding (epitope spreading implication) was only visible from full-wiki synthesis — it does not appear in any individual source page — making it a genuine addition to knowledge rather than a restatement.

[2026-04-13] Added Epistemic Honesty Principle to CLAUDE.md

Decision: Added a top-level section to CLAUDE.md codifying the “caring honesty” framework — suppress sycophancy signals, preserve warmth signals, apply the “respected colleague” heuristic for edge cases. Rationale: The wiki serves the curator’s understanding, not comfort. Methodological criticism must not be softened because the curator seems invested; praise must be specific, not generic. This governs all interactions: ingest discussions, query answers, lint reports.

[2026-04-13] Added proactive query compounding to Query workflow

Decision: The Query workflow now proactively suggests saving non-trivial synthesis answers as wiki pages, rather than waiting for the curator to ask. Rationale: Inspired by Karpathy’s LLM Wiki pattern — valuable query synthesis should compound into the wiki just like ingested sources do. Prevents good analysis from disappearing into chat history.

[2026-04-13] Added wiki/state.md as persistent session context

Decision: Created this file to track operational state (focus, queue, decisions, watch items) across sessions. Rationale: CLAUDE.md documents the schema (stable), log.md tracks what happened (history), index.md catalogs content. None of them captured current priorities, queued work, or the reasoning behind structural choices. This file fills that gap so every new session starts with full context.

[2026-04-12] Citation count fields added to source frontmatter

Decision: Source pages include citations_semantic_scholar, citations_crossref, and citations_retrieved fields fetched from APIs during ingest. Rationale: Provides a rough proxy for paper influence. Two sources (Semantic Scholar + CrossRef) because neither is complete alone.

[2026-04-12] Notable Findings as a single running page

Decision: wiki/analyses/Notable Findings.md is one append-only page, not a folder of individual findings. Rationale: Keeps striking observations in one scannable list. The bar is deliberately high — surprising given existing wiki content, crosses multiple concepts, or challenges an existing claim. Most ingests produce zero or one entry.

[2026-05-24] Pearce2016 ingested; new methods dimension — Age-Matched Case-Control Analysis

Decision: Ingested Pearce2016 (BMJ 2016;352:i969; DOI 10.1136/bmj.i969; 684/575 citations). Not a dengue paper — a highly cited epidemiological methods reference on matched case-control study analysis. New methods page Age-Matched Case-Control Analysis.md created. This is the first non-laboratory methods page in the wiki (all prior methods pages cover diagnostic/experimental assays). Motivated by scrutiny of the curator’s age-matched case-control analysis in the ANA IIF abstract draft. Why: The curator’s ANA research methodology (age-matched case-control) was criticized as biased and non-conventional. Pearce2016 provides the authoritative BMJ methodological framework for evaluating whether the criticism is valid — specifically addressing the two most common misconceptions (matching eliminates confounding; matched design requires matched analysis). This establishes a methodological evidence base within the wiki. How to apply: When evaluating any case-control dengue study in this wiki (e.g., Garcia2010), assess whether matching factors were analytically controlled and whether the conditional vs. unconditional analysis choice was appropriate. If more epidemiological methodology papers are ingested (RCT design, propensity scoring, cohort methods), consider creating a dedicated study-design/ axis — for now, methods/ is adequate.

[2026-05-24] Iwagami2022 ingested; study-design/ axis created; Age-Matched moved from methods/

Decision: Ingested Iwagami2022 (Annals of Clinical Epidemiology 2022;4(2):33–40; DOI 10.37737/ace.22005; 115/122 citations). Not a dengue paper — a methodological tutorial on matching in case-control and cohort studies. Created wiki/study-design/ as a new axis for epidemiological/statistical study design pages. Moved Age-Matched Case-Control Analysis.md from methods/ to study-design/. CLAUDE.md updated: Architecture tree, frontmatter type enum, page naming conventions, Domain Context. Age-Matched page updated with Iwagami2022 content including the Pearce2016/Iwagami2022 contradiction on unconditional logistic regression, matching ratios, over-matching, cohort vs. case-control matching distinction. Why: Two methodology papers (Pearce2016, Iwagami2022) plus the curator’s active ANA IIF research methodology provide enough signal to justify a dedicated axis. Study-design pages are categorically different from laboratory methods (RT-PCR, ELISA) — they cover research design and statistical methodology, not assay protocols. How to apply: Future epidemiological methodology papers (RCT design, propensity scoring, bias analysis) go to study-design/. Laboratory methods remain in methods/. The Age-Matched page now documents a genuine methodological disagreement (Pearce vs. Iwagami on unconditional logistic regression) — this should be referenced when evaluating any matched case-control design in dengue papers.

[2026-05-14] Council review workflow added to CLAUDE.md

Decision: Ported “Summon the Council” multi-agent critical review workflow from efb-dengue-wiki into dengue-wiki CLAUDE.md. Added Claude-council/ output folder to Architecture; added full 7-step workflow to §Workflows (after Remove/Merge Axis, before Domain Context). Council member role descriptions adapted from EFB-specific (flow cytometry, gating strategies) to dengue-broad (diagnostic assay validity, sample selection). Council is read-only — produces reports in Claude-council/, does not modify wiki pages. Why: Curator wants consistent paper review capability across both wikis. The council provides structured multi-perspective critique that complements the standard ingest workflow. How to apply: “summon the council [paper]” or “council review [paper]” triggers the workflow. Custom roles can be added/swapped. Reports land in Claude-council/ — wiki changes require explicit curator direction.

Watch Items

Issues, thin areas, or things to revisit. Remove items as they’re resolved.

• [NEW 2026-06-03] PIF risk-factor evidence is geographically narrow (Hertanti2024): The significant PIF risk-factor pools (female sex OR 1.65, DHF 1.80, comorbidities 2.14) rest on only 2–3 studies each, dominated by Sri Lankan cohorts (Abeysena2019, Sigera2021, Perera2023, Umakanth2018). Whether these associations replicate outside South Asia is untested. A non-South-Asian PIF cohort reporting risk-factor ORs would test generalisability.

• [NEW 2026-06-03] Fatigue-instrument harmonisation (Hertanti2024): 87.5% of pooled studies assessed fatigue by clinical-symptom checklist vs only 12.5% by validated questionnaire; within the PIF pool, checklists detected 12.6% vs 29.5% by questionnaire. The wiki’s fatigue datapoints (Seet2007 FQ, Gawali2021, Garcia2009 asthenia) are therefore not directly comparable without harmonising the measure — a prerequisite for aligning them onto one fatigue trajectory. Decide which instrument the wiki privileges before pooling rates.

• [NEW 2026-06-03] DHF→PIF vs DHF→acute-fatigue tension (Hertanti2024): DHF→PIF is borderline-positive (pooled OR 1.80, p=0.042, 2 studies — Perera2023 sig, Sigera2021 individually NS), but DHF→acute fatigue is null and internally contradictory (Ferreira2018 2.31 harmful vs Recker2024 0.75 protective, pooled NS, I²=93.5%). Whether severity acts specifically on the persistent phase, or the 1.80 is a 2-study artifact, is unresolved — hold the tension open, do not read it as overturning the wiki’s severity-independence finding. A single cohort co-measuring acute severity and PIF trajectory would resolve it.

• [NEW 2026-05-25] ANA IIF Abstract — sex confounding (FATAL FLAW): The abstract cannot rule out that sex imbalance between cases (sex unknown) and controls (predominantly female) drives the primary ANA result. Recovering sex data for cases is essential before publication. If unavailable, a sensitivity analysis estimating the sex distribution required to nullify the result is the minimum acceptable disclosure.

• [UPDATED 2026-05-25] ANA IIF Abstract — analytical strategy + matching design: Fisher’s exact invalid on matched data; McNemar’s valid but suboptimal. Conditional logistic regression is the recommendation both Pearce2016 and Iwagami2022 endorse. Additionally: 1:1 matching ratio discarded ~50 available controls unnecessarily (1:2 or 1:3 feasible per Iwagami2022); 39.4% case dropout from 18–22 age range compromises representativeness and compounds with sex confounding. Matched vs. unmatched ANA rate comparison is a mandatory disclosure.

• [UPDATED 2026-05-24] Age-Matched Case-Control Analysis (2 sources): Now in study-design/ (moved from methods/). Two methodology references (Pearce2016, Iwagami2022) with a documented contradiction on unconditional logistic regression validity. A third methodology paper or a simulation study comparing unconditional vs. conditional analysis would resolve the dispute.

• Unconditional vs. conditional analysis in dengue case-control studies (Iwagami2022/Pearce2016 Q): The two methodology references disagree on whether unconditional logistic regression is valid for matched case-control data. A survey of which analytical method is used in dengue case-control studies would clarify practical significance.

• Over-matching risk in dengue case-control designs (Iwagami2022 Q): Dengue studies sometimes match on age + sex + hospital admission date + neighborhood. Whether this level of matching produces over-matching (too many matching variables reducing informative discordant pairs) has not been assessed for any dengue study in this wiki.

• Garcia2010 analytical method audit (Pearce2016 Q): Garcia2010 used a population-based case-control design (symptomatic vs. asymptomatic dengue, FcγRIIa genotyping). Per Pearce2016, if age-matching was used, the matching factor must be controlled in the analysis. Whether Garcia2010’s reported ORs (e.g., FcγRIIa-HH OR 10.56 for DHF) correctly account for matching-induced confounding has not been verified against the raw paper.

• Anti-prM vaccine design gap (Dejnirattisai2010): All current vaccines (CYD-TDV, TAK-003, TV003) use native prM — all will prime ADE-potent anti-prM responses. Have any post-2010 vaccine candidates adopted heterologous or modified prM sequences? No source in this wiki addresses this. A review of current DENV vaccine design literature for prM modification strategies would directly address this.

• Anti-prM kinetics across inter-infection interval gap: Bos2025 reports rising XR EDI/II E-IgG but does not distinguish anti-prM from total XR structural antibody. Whether anti-prM titres follow a similar rising or waning trajectory during the inter-infection window is unknown and directly relevant to ADE risk quantification.

• Anti-prM clinical correlate gap: No cohort study in this wiki has measured anti-prM titres as a predictor of DHF risk in secondary infection. This is the most clinically direct implication of Dejnirattisai2010 and remains completely unaddressed.

• prM Protein entity thin (1 source): New page, 1 source. Any additional structural virology or immunology paper that characterises prM biology, cleavage efficiency, or mixed virion populations would strengthen it.

• Thin entity pages (post-2026-04-16 propagation sweep): CYD-TDV (1 source), Wolbachia (1 source). Aedes albopictus now 2 (Oishi2003 added — methodological role only); DENV-1 now 4; DENV-3 now 7; DENV-2 now 5; DENV-4 now 5 — no longer thin.

• Thin method pages: Several methods (Single-Cell RNA Sequencing, V(D)J Sequencing, qRT-PCR, ELISA Inhibition Method) have only 1 source

• Geography: 10 geography pages now (111 total wiki pages). Southeast Asia regional page created 2026-04-19; Americas/Latin America page created 2026-04-18. Vietnam standalone created 2026-04-19 (2 sources). Paraguay folded into Latin America 2026-04-19 (lint). Remaining countries without standalone pages: Mexico (1 source, will fold into Latin America). Countries with thin pages: Thailand (2 sources, both same DENFREE cohort). India (6 sources) and Taiwan (5 sources) are well-covered. Philippines (2 sources, both same Manila group) has narrow thematic coverage.

• Missing concept page: Herd immunity — mentioned in CLAUDE.md domain context but no dedicated page yet; no sources directly address it. Cross-Reactive Antibodies and Cytokine Storm pages created 2026-04-13

• Thin concept pages (newly flagged 2026-04-17 lint): NK Cell Responses in Dengue (1 source) — alongside CYD-TDV and Wolbachia. [RESOLVED 2026-04-17] Original Antigenic Sin: 1→2 sources (Bhatt2020 added OAS T cell mechanism).

• Thin method pages (newly flagged 2026-04-17 lint): Line Immunoassay ANA (1 source), Surface Plasmon Resonance (1 source) — add to the existing thin method list (Single-Cell RNA Sequencing, V(D)J Sequencing, qRT-PCR, ELISA Inhibition Method, Autoantigen Microarray)

• [RESOLVED 2026-04-19] ANA Review V2.0 updated: Hung2008 (anti-EC isotype shift, TM, infection-order independence extended to 4 sources), Santosa2012 (false-positive serology), and Farias2024 (dengue→autoimmune misclassification, bidirectional differential) all integrated. Sources 37→40. New §7.4 added. Infection-order independence established claim updated to 4 sources.

• OROV/Mayaro co-infection severity gap (Farias2024 Q): Brazil co-circulates DENV, CHIKV, ZIKV, OROV (Oropouche), and MAYV (Mayaro). Farias2024 raises but does not address whether DENV co-infection with OROV or MAYV worsens clinical outcomes compared to mono-infection. No source in this wiki addresses this question. A search for dengue + OROV co-infection + severity in the Brazilian literature would be productive.

• GBWT reliability validation gap (Farias2024 Q): The gallbladder wall thickness (GBWT) >3mm criterion for DHF prediction (90.5% sensitivity, 69.6% specificity per Farias2024) comes from three Brazilian studies in a specialist emergency context. Applicability in primary care or non-hospital settings with variable ultrasound quality is unknown. A multi-centre evaluation across different resource settings would clarify clinical utility.

• Still’s disease as dengue trigger — single-source gap (Farias2024 Q): Farias2024 mentions dengue potentially triggering adult-onset Still’s disease (AOSD), but does not cite primary data — only contextual discussion. No source in this wiki addresses dengue-triggered AOSD with primary evidence. A targeted literature search for dengue + Still’s disease case reports would establish whether this is a documented complication or a theoretical concern.

• [RESOLVED 2026-04-17] ANA analysis page updated with Vo2020/Saito2004. Sources 22→24. Key additions: §4.4 nuclear antigen IgG consumption model; §5.4 extended with Saito2004 PAIgM (anti-viral, FcγR-independent, 92.1% DHF specificity); §7 Vo2020 primary>secondary IgG inversion; §9 infection-order host factor; §10 Q12–Q13; §11 three-pathway thrombocytopenia established claim and two new hypothesis-generating entries.

• ANA thread substantially complete: The literature review, gap analysis, and notable findings are all current. The remaining ANA gaps are specific and well-defined (see Wiki State and Gap Analysis 2026-04-13). No further source ingestion is needed to consolidate the current thread — but a longitudinal HEp-2 cohort paper would be the single highest-value addition.

• Potential page split: Autoimmunity in Dengue now has 15 direct sources; mechanistic (NS1 mimicry, endothelial, platelet) vs. epidemiological (Li2018, Shih2023, Garcia2009) content may warrant splitting into acute mechanisms and population-level risk pages.

• Li2018 ingest complete: The adrenocortical insufficiency TLR mechanism and ADEM cross-design convergence are now documented. The adrenocortical finding remains unvalidated in any lab-confirmed study — a targeted search for dengue + adrenal + cortisol literature would clarify whether this signal is real.

• Bruhns2009 ingest complete: FcγRIIa Receptor Overview corrected — old “HH lower IgG1 affinity” claim refuted; H131/R131 difference is IgG2-specific (4.5×). FcγRIIB inhibitory deficit quantified. The mechanism linking HH genotype to worse dengue outcomes is now an open question — the IgG2 angle is the leading candidate. A search for anti-dengue IgG2 subclass distribution in secondary infection would directly test this.

• FcγRIIa mechanism gap (new): No study in this wiki directly measures anti-dengue IgG2 titres in primary vs. secondary infection with genotype stratification. This is the most direct experiment needed to determine whether the Bruhns IgG2-asymmetry is clinically relevant in dengue.

• Post-Dengue Syndrome thread strengthened: Now 5 sources (Hertanti2024 meta-analysis added 2026-06-03); Seet2007 provides first non-Cuba cohort. Cross-cohort severity-independence finding is now the strongest structural claim on this page. Short follow-up gap (Seet2007 2 months vs Garcia2009 2 years) still unresolved — a meta-analysis of mixed windows does not bridge it; a single prospective cohort spanning 2 mo → 2 yr would.

• NS1 mimicry / nuclear ANA gap: The new Notable Finding flags that NS1 mimicry cannot explain nuclear IIFA positivity; epitope spreading is the implied mechanism. No paper in this wiki directly tests this — a targeted literature search for dengue + epitope spreading + nuclear antigen would be productive.

• Bos2025 preprint status: Ingested 2026-04-14 as preprint. Track for peer-reviewed publication; revisit mechanistic claims (rising XR EDI/II IgG as ADE driver; IgM persistence at 18M) once peer-reviewed.

• Rising XR E-IgG mechanism gap (new): Bos2025 shows XR EDI/II IgG rises 6–18M post-primary, challenging the classical waning ADE model. Whether these rising antibodies are functionally non-neutralising (ADE-capable) vs. cross-neutralising is unknown. A functional ADE assay study stratifying by EDI/II vs EDIII antibody titres would directly test this.

• NS1-IgG waning → autoantibody decline (new): If NS1-IgG wanes with t½≈2.1 years (Bos2025), the NS1-mimicry component of dengue ANA should decline on a similar curve. No study in this wiki has measured anti-platelet / anti-endothelial autoantibody titres longitudinally (acute to 18M+) to test this prediction.

• Nicaragua geography thin — pending fold (retroactive rule): Only 1 source (Bos2025 preprint). ⚠ Thin country page: Nicaragua.md has <2 sources and will be folded into a Latin America / Americas regional page when that page is created. Until then remains as a thin Watch Item. Would benefit from a peer-reviewed Nicaraguan cohort study to reach the ≥2 threshold independently.

• Geography and non-ANA content remain thin: Now 8 countries (Cuba, Thailand, Taiwan, India, Singapore, Nicaragua, Paraguay, Philippines); no Southeast Asia or Americas regional pages; vaccine and vector biology still largely derived from Guzman2016. Singapore now 2 sources (Seet2007 + Palacios2016/Chang 2007 retinal vasculitis). Broadening into these areas would balance the wiki’s coverage.

• Mitotic spindle ANA pattern in dengue — literature search warranted (Jardim2012 Q): The mitotic spindle ANA pattern in Jardim2012 is the first documentation of this pattern class in dengue in this wiki. It targets centromere/spindle apparatus proteins rather than nuclear DNA/histones — a different autoantigen class from NS1 mimicry targets. Whether other dengue papers document this pattern specifically is unknown. A targeted PubMed search for “dengue AND antinuclear antibody AND mitotic spindle” or “dengue AND centromere antibody” would determine whether this is an isolated observation or a recurring feature.

• C3-selective depression vs. classical pathway in dengue — differential value (Jardim2012 Q): Jardim2012 shows selective C3 depression (0.39 g/L) with C4 normal — contrasting with Rajadhyaksha2012 (both C3 22 mg/dL, C4 5 mg/dL severely depressed). If this C4-sparing pattern is consistent across dengue complement consumption (vs. SLE-type classical pathway depletion of both), it could serve as a rapid differential distinguishing point in endemic settings. No systematic study of complement pathway involvement patterns in dengue serositis exists in this wiki.

• Cryoglobulin type in dengue — single source gap (Jardim2012 Q): Jardim2012 is the only cryoglobulinemia case in this wiki. Whether dengue-associated cryoglobulins are type I (monoclonal IgM), type II (mixed IgM monoclonal + IgG polyclonal), or type III (polyclonal both) affects the mechanistic interpretation. A literature search for dengue + cryoglobulinemia would clarify whether this is a rare individual case or an under-documented dengue complication.

• [RESOLVED 2026-04-18] Americas regional page created: Latin America.md now consolidates Cuba (3 sources), Nicaragua (1), Brazil (2). Cuba.md, Nicaragua.md, Brazil.md deleted. 10 wikilinks updated. Index Geography table revised. Mexico (1 source, no page) and Paraguay (1 source, no page) remain as thin watch items that will fold into Latin America when they reach ≥2 sources. Brazil content is thin on dengue-specific epidemiology — a peer-reviewed Brazilian dengue cohort would substantially improve the § Brazil section.

• False-positive dengue IgM cross-kit variability (Santosa2012 Q): The 15% false-positive rate in RF-positive patients uses only the Panbio immunochromatography kit (unpublished local data, n=20 RF-positive). Whether this generalises to MAC-ELISA platforms is unknown; the Hunsperger 2009 multi-kit evaluation found wide variability across 10 kits. A systematic comparison of false-positive rates across major dengue IgM kit designs in RF-positive, ANA-positive, and dengue-phase-IgM-positive patients would clarify the scope of the problem.

• Dengue-phase IgM ANA cross-reactivity with dengue serology kits (Santosa2012 Q): The ANA-positive group in Santosa2012 showed 0/10 false-positive dengue IgM — but these were SLE patients with predominantly IgG ANA. Dengue patients with acute-phase non-specific IIFA-positive ANA (predominantly IgM; see Chatterjee2024) present a different question: could their dengue-phase polyreactive IgM amplification produce false-positive readings on other immunological assays? This has not been tested.

• ANA threshold comparability gap (Codes2002 Q): The 20.5% Codes2002 acute-phase rate uses ≥1:40 as the cutoff. At ≥1:80 only 4.5% (7/156) were positive. No dengue study in this wiki reports acute-phase ANA using ≥1:80 as threshold on a comparable IIF platform. Whether dengue-specific ANA positivity at ≥1:80 exceeds the hepatitis baseline (4.5%) is unknown.

• ASMA in dengue — complete evidence gap (Codes2002 Q): Anti-smooth muscle antibody was detected in 14.8% of viral hepatitis patients acutely (Codes2002). No dengue study in this wiki has measured ASMA. Whether dengue similarly elevates ASMA is unknown — and would be relevant to distinguishing dengue-triggered autoimmune hepatitis from primary autoimmune hepatitis in endemic settings.

• IIF substrate comparison gap (Codes2002 Q): Codes2002 does not specify the IIF substrate (HEp-2 vs. rat/mouse liver). For the 20.5%/6.4% figures to serve as a valid benchmark for dengue ANA studies using HEp-2 (Chatterjee2024, Gawali2021), substrate equivalence must be assumed — not established. A study measuring viral-infection ANA on matched HEp-2 substrate with dengue comparator would provide a proper benchmark.

• HSP60 NS1 epitope identification gap (Cheng2015 Q): Cheng2015 confirms HSP60 cross-reactivity is driven by a distinct, unidentified NS1 epitope (not P311–330). Identification of this epitope would clarify whether anti-HSP60 independently contributes to endothelial damage or merely co-tracks with the anti-PDI response at lower magnitude. A synthetic peptide panel covering the NS1 C-terminal domain excluding P311–330 would be the experimental approach.

• Anti-vimentin mechanism gap (Cheng2015 Q): Anti-vimentin IgM is elevated in DHF vs. controls but does not correlate with anti-NS1 IgM or anti-EC IgM — suggesting a non-NS1, non-P311–330 mechanism. Whether another DENV protein (capsid, prM cross-reactivity, structural protein beyond NS1) drives anti-vimentin elevation is unknown. A study testing anti-vimentin levels in viruses lacking NS1 (chimeric or NS1-deleted DENV) would directly test NS1 dependency.

• Anti-PDI antiviral potential gap (Cheng2015 Q): PDI facilitates DENV entry on endothelial cells via integrin activation (Wan2012). Anti-PDI antibodies could theoretically block PDI-mediated viral entry — a self-limiting antiviral mechanism. No study has tested whether anti-PDI autoantibodies generated via NS1 mimicry reduce DENV infectivity in endothelial cell lines. This is testable by comparing viral entry rates in HMEC-1 cells pre-treated with anti-PDI serum vs. controls.

• Primary/secondary DHF equivalence for anti-endothelial autoantibodies — confirmation needed (Cheng2015 Q): Cheng2015’s primary DHF group is n=2 — insufficient for reliable primary/secondary equivalence claim. A prospective study enrolling ≥20 primary DHF patients with confirmed seronegative baseline would directly test whether the infection-order independence of anti-PDI/HSP60/vimentin/P311–330 autoantibodies generalises.

• hCF independent validation gap (Chaturvedi2001): The cytotoxic factor (hCF) concept is the Chaturvedi group’s own proprietary construct — no sequence deposited, no independent replication. A literature search for any independent citation of hCF, or any proteomics paper attempting to identify the HPLC-purified dengue cytokine fraction described by this group, would determine whether this is a real uncharacterised cytokine or an artefact. If hCF maps to a known cytokine or dengue structural protein, the Chaturvedi2001 anti-hCF autoantibody finding would gain greatly in interpretive power.

• Anti-hCF dual biomarker potential (Chaturvedi2001): If anti-hCF autoantibody levels (96% DF → 8% DHF grade IV, Chaturvedi2001) could be combined with PAIgM >20 ng/10⁷ (92.1% DHF specificity, Saito2004), this might produce a two-biomarker prognostic panel distinguishing DF from DHF at presentation. No study has tested this combination. Requires hCF to be independently validated first.

• Propagation discipline (new, post-2026-04-16 deep lint): The recurring miss of forward propagation from source → linked target pages has been documented in feedback memory and a fix workflow added to state.md decisions. Future ingests should be checked with a forward-propagation script before completion. Until automated, every ingest should be followed by a manual audit of every wikilink in the source page’s Entities/Concepts/Methods sections — particularly for secondary sources (letters, reviews) where the propagation overhead is least proportional to the source’s perceived weight.

• [RESOLVED 2026-04-18 deep lint] Two propagation gaps fixed: (1) Infection-Triggered Autoimmunity.md missing Santosa2012 → sources 13→14; (2) IgM-IgG Serology ELISA.md missing Chaturvedi2001 → sources 16→17. Full forward-propagation check of all 6 recent ingests (Santosa2012, Farias2024, Cheng2015, Chaturvedi2001, Jardim2012, Codes2002) × all linked target pages completed and confirmed clean.

• [RESOLVED 2026-04-19 deep lint] Hung2008 + Ghorai2024 propagation sweep completed: 6 propagation gaps fixed (DENV-1, DENV-3, Aedes aegypti, RT-PCR, ELISA, NS1 Antigen Detection), 1 stale wikilink fixed (Ghorai2024: NS1 Antigen ELISA → NS1 Antigen Detection), 2 index header count errors fixed (Concepts 23→24, Methods 16→17), Paraguay.md folded into Latin America (sources 7→8), ANA V2.0 updated to 41 sources. Total pages: 112→111.

• Dengue-MAS mechanism gap (new): No study in this wiki directly measures anti-endothelial cell Abs by flow cytometry (Wan2012 method) in dengue-MAS patients. Morel2014’s MAS cases were ANA-negative, but it is unknown whether the Wan2012-type anti-endothelial Abs were also absent or elevated. A study measuring both standard ANA and flow cytometric anti-endothelial Abs in dengue patients across the full severity spectrum (DF, DHF, MAS) would clarify whether these represent truly separate immune effector mechanisms or a continuum.

• Morel2014 SS citation gap: Morel2014 is not indexed on Semantic Scholar at the English-edition DOI (10.1016/j.reumae.2014.03.008). The Spanish original may be indexed under a different DOI — worth a manual check if citation count is needed. CrossRef confirms 0 citations at the English-edition DOI.

• ELISA Inhibition Method — verify Bos2025 attribution: Bos2025 source page lists ELISA Inhibition Method under “Methods Used.” The wiki’s ELISA Inhibition Method page describes a specific Garcia2009/Vazquez 2003 yellow-fever-adapted competitive assay. Whether Bos2025 used the identical protocol or a different inhibition ELISA for cross-reactivity determination could not be verified from the wiki alone — verify against raw PDF if method accuracy matters.

• Reading Plan orphan: Reading Plan - ANA and Dengue Dynamics is only linked from index.md. No content page links back to it. Minor connectivity issue; acceptable for a meta/reference document.

• Primary-infection DHF/DSS complement mechanism gap (new): Lin2001 establishes that IgM anti-platelet autoantibodies arise in primary DENV-3 infection and that complement-mediated platelet lysis correlates with DHF/DSS severity. No wiki source tests whether complement pathway efficiency (C3/C4 levels, C1q polymorphisms, MBL status) predicts which primary infections progress to DHF/DSS. This is the most targeted experiment implied by the Lin2001 data.

• NS1 Protein page — possible redundant bullet (minor): After the Lin2001 attribution rewrite of the “Autoantigenic properties” section, there may be a legacy summary-level bullet that overlaps with the new detailed block. Review NS1 Protein page for redundancy if editing the section again.

• Thrombocytopenia bifurcation — FcγRIIa paradox deepened (updated Saito2004): Saito2004 adds PAIgM (completely FcγR-independent) as a second secondary-infection immune complex pathway that independently predicts DHF (92.1% specificity). The paradox is now even sharper: both thrombocytopenia mechanisms in secondary infection (PAIgG and PAIgM) bypass FcγR at platelet docking, and PAIgM has no FcγR involvement at any step. The Garcia2010 FcγRIIa-HH DHF risk (OR 10.56) must operate entirely through a non-platelet pathway — ADE in monocytes/macrophages is the leading hypothesis. No study in this wiki directly tests FcγRIIa genotype effects on PAIgG/PAIgM levels in secondary infection.

• Direct viral platelet infection as independent mechanism (Saito2004 Q): 42.8% of platelet samples in Saito2004 were dengue RNA-positive. No study in this wiki tests whether direct viral platelet infection contributes to platelet activation and endothelial adhesion independently of immune complex (PAIgG/PAIgM) deposition. This is a distinct third mechanism hypothesis not currently addressed anywhere in the wiki.

• Autoantibody depletion as plasma leakage predictor (Vo2020 Q): Vo2020 raises whether the temporal depletion of DHF-correlated autoantibodies (IgG against complement/coagulation/nuclear antigens) precedes plasma leakage onset. If depletion precedes leakage, these antibodies are candidate prognostic biomarkers. No longitudinal study in this wiki tracks autoantibody kinetics against plasma leakage timing.

• PAIgM >20 ng/10⁷ DHF predictor — requires independent validation: Saito2004 reports PAIgM >20 ng/10⁷ predicts DHF with 92.1% specificity and 48.6% sensitivity in 78 secondary-infection patients (DHF grades I–II only; no shock). No independent replication exists in this wiki. Validation in a DSS cohort and in a non-Philippines population would establish whether this is a generalizable predictor or cohort-specific.

• PAIgG-DHF correlation gap (now partially resolved by Saito2004): Saito2004 confirms PAIgG is significantly higher in DHF than DF in secondary infection (37.1 vs. 25.0 ng/10⁷, P < 0.01), though PAIgM outperforms it as an independent DHF predictor by logistic regression. The original Oishi2003 limitation (no severity stratification) is now addressed by the follow-on study.

• Philippines geography strengthened (Saito2004 ingest): Now 2 sources (Oishi2003 + Saito2004). Still under-represented relative to the Philippines’ dengue burden; both current sources are from the same Manila research group studying the same PAIgG/PAIgM mechanism — coverage is thematically narrow. A Philippines epidemiology or serotype distribution paper would diversify coverage.

• [RESOLVED 2026-04-19] Cambodia geography: Cambodia.md folded into Southeast Asia.md §Cambodia. Deletion executed; all Cambodia wikilinks updated to Southeast Asia. Cambodia § in Southeast Asia.md captures the single Vo2020 (Kampong Cham DENV-1 2012–2013) source.

• Primary > secondary IgG autoantibody inversion (Vo2020): The finding that primary infection generates broader IgG autoantibody repertoire than secondary inverts severity-centric expectations. Needs replication in a larger cohort with adequate primary/secondary matching (the Vo2020 primary group was n=6, all male, all DENV-1 — heavily confounded). If replicated, this would have major implications for when dengue-induced autoimmunity risk is highest in infection history.

• Nuclear antigen IgG consumption in DHF (Vo2020): IgG autoantibodies against KU, Smith, histone, Sm/RNP, nucleosome (canonical ANA targets) are positively correlated with platelet counts in DHF — lower in severe disease, suggesting consumption via immune complex formation. This provides a mechanistic explanation for why ANA titres might not be higher in severe dengue (compatible with Morel2014’s ANA-negative MAS). No study in this wiki has directly measured ANA titres serially across the DHF severity spectrum while also measuring complement consumption — a targeted experiment that would directly test the consumption model.

• Autoantigen Microarray method — single source: New method page has only 1 source (Vo2020). Technically a well-established platform used primarily in SLE (Zhu et al. 2015), but no other dengue study in this wiki uses it. If other protein array studies in dengue exist, they would strengthen this page.

• sfRNA TRIM25/RIG-I mechanism — in vitro only (Bhatt2020 Q): The Manokaran 2015 sfRNA/TRIM25 mechanism was demonstrated in cell lines (cited in Bhatt2020). Whether it operates in primary human monocytes and DCs — the principal DENV target cells in vivo — has not been confirmed in available wiki sources. A paper using primary human macrophages or monocyte-derived DCs to test sfRNA-TRIM25 interaction would strengthen this claim.

• Katzelnick ADE window — post-vaccination titre waning kinetics gap (Bhatt2020 Q): The 1:21–1:80 ADE window identifies a quantitative danger zone, but no source in this wiki maps post-vaccination Ab waning kinetics for TAK-003 or TV003 against this threshold. A paper tracking longitudinal titres post-vaccination in seropositive and seronegative vaccinees would directly test whether vaccinee titres pass through the enhancement zone.

• MIF antagonists as potential dengue therapy (Bhatt2020 Q): MIF induces autophagy that DENV exploits for replication. MIF antagonists (e.g., ISO-1) have been trialled in other inflammatory diseases. No source in this wiki tests MIF antagonism in dengue — a targeted search for MIF inhibition in flavivirus models would clarify whether this is a viable therapeutic direction.

• New raw PDFs to queue: raw/chaturvedi2001.pdf, raw/santosa2012.pdf, raw/Farias2024.pdf, raw/Gawali2021.pdf, raw/Hung2008.pdf, raw/Velazqueza2017.pdf, raw/cheng2015.pdf — all appear in raw/ directory but are not yet ingested. Scope unknown — prioritise by relevance to current threads when ready.

• Polyreactive IgM hypothesis for dengue ANA non-specific fraction (new, Zhou2007): The ~66% IIFA-positive, LIA-negative dengue ANA fraction is now hypothesised to reflect normal polyreactive IgM amplification rather than antigen-induced autoimmunity. This can be directly tested by BCR V-region sequencing of IIFA-positive B cells from dengue patients — germline (not somatically hypermutated) usage would confirm polyreactive origin. No study in this wiki has done this.

• Polyreactive Antibodies page thin (1 source): New page, 1 source (Zhou2007). Additional immunology papers on polyreactive antibodies in infection contexts (especially virus infection) would strengthen it. The field is primarily murine/in vitro — a human dengue study directly testing polyreactive IgM would be the highest-value addition.

• [RESOLVED 2026-04-17] ANA analysis page Zhou2007 integration: New §4.5 added to the literature review covering polyreactive IgM as the mechanistic basis for the non-specific IIFA fraction. Sources 24→28 (added Zhou2007, Gawali2021, Velazqueza2017, Rajadhyaksha2012 in third-pass revision).

• Gawali2021 control-group gap (new): The 18.33% ANA rate at 6 months post-dengue (Gawali2021) cannot be interpreted without a contemporaneous dengue-negative control from the same Central Indian population. India’s background ANA rate at 1:100 is unknown — Li2019 (Chinese health-checkup at >1:100: 14.01%) is the best available proxy, leaving only a ~4 percentage point excess. A follow-up study from Central India or another endemic Indian setting with a matched control group would resolve whether 6-month post-dengue ANA elevation is real or artefactual.

• 6-month ANA trajectory gap (new, Gawali2021): The three dengue ANA data points now in the wiki (acute 54.8%, 6 months ~18%, 2 years 23%) are from three different countries (India/Kolkata, India/Gwalior, Cuba), different serotypes, different substrates, and no shared denominator. Whether this represents a true decline-and-stabilisation pattern or simply population heterogeneity cannot be determined. A single prospective cohort with ANA measured at baseline, 1M, 3M, 6M, 12M, and 24M post-dengue on a consistent HEp-2 platform with a control arm would directly test the trajectory.

• Mexico geography thin (Velazqueza2017): Mexico now has 1 source (Velazqueza2017) — no standalone page per ≥2-source rule. Will fold into a Latin America / Americas regional page when that page is created. Note: Cuba and Nicaragua are the existing Americas standalone pages; Mexico is a third Americas-region datapoint but without a page. Americas regional page requires ≥3 country pages in the region — currently Cuba + Nicaragua (2 pages). Mexico reaching ≥2 sources OR another Americas country reaching ≥2 sources would add a third Americas page, triggering the regional page.

• SLE-dengue temporal sequence gap (Velazqueza2017 Q): Velazqueza2017 Case 2 has the most temporally suggestive dengue→SLE triggering sequence in the wiki (SLE diagnosed 2 months post-dengue), but the authors acknowledge chronological ambiguity. No prospective study in this wiki has tracked the appearance of anti-dsDNA and antinucleosome antibodies in dengue patients before and after acute infection to determine whether these arise de novo from dengue or represent pre-existing subclinical SLE revealed by the workup.

• Antinucleosome in dengue+SLE vs. DHF consumption (Velazqueza2017/Vo2020 tension): Vo2020 shows nuclear antigen IgGs (incl. nucleosome) are lower in severe dengue DHF, implying consumption. Velazqueza2017 shows very high antinucleosome (up to 258 IU/ml) in dengue+active SLE cases. Whether the SLE-driven antinucleosome response simply outpaces consumption, or whether these represent immunologically distinct antibody populations, cannot be resolved from either study. A study measuring antinucleosome kinetics in dengue patients stratified by pre-existing SLE vs. dengue-only would directly test this.

• IL-10 → T cell apoptosis therapeutic paradox (Pang2017 Q): IL-10 blockade reduces T cell apoptosis in dengue (Mathew & Rothman 2008), but IL-10 is also anti-inflammatory. Targeting IL-10 therapeutically in dengue could impair immune regulation while improving cellular clearance — the net effect is unknown. A study measuring T cell apoptosis rates, IL-10 levels, and outcome severity simultaneously in a dengue cohort would clarify whether the apoptotic arm drives pathology independently of IL-10’s immunosuppressive roles.

• C5b-C9/NLRP3 link — dengue-specific evidence gap (Pang2017 Q): The C5b-C9 → NLRP3 inflammasome pathway invoked in Pang2017 (citing Suresh 2016) is based on general complement biology; direct evidence that DENV-activated C5b-C9 activates NLRP3 in dengue patient samples has not been published in this wiki’s sources. A paper measuring NLRP3 activation markers (IL-1β, IL-18, caspase-1) alongside NS1 and C5b-C9 levels in dengue patients would directly test this step.

• Anti-NS1 GPI-signalling → viral replication (Pang2017 Q): This mechanism (Jacobs 2000 via Pang2017) would operate in secondary infection after anamnestic anti-NS1 IgG responses develop. No source in this wiki has directly tested whether pre-formed anti-NS1 antibodies augment viral replication in infected cells in a secondary infection model. A mechanistic study comparing viral titre in anti-NS1-opsonised vs. non-opsonised monocytes would test this.

• Anti-cardiolipin persistence in dengue-triggered SLE (Rajadhyaksha2012 Q): Combined IgM 44 + IgG 12 MPLU/mL anti-cardiolipin persisting to 4 months in a dengue→SLE case. Whether this represents dengue-induced transient antiphospholipid antibodies (seen transiently in Morel2014 Case 1 and other viral infections) vs. SLE-driven nascent antiphospholipid syndrome with thrombotic risk is unknown. No source in this wiki longitudinally tracks antiphospholipid antibodies from dengue through post-dengue follow-up.

• Renal biopsy immunofluorescence gap (Rajadhyaksha2012 Q): The paper reports Class IV GN grade but not biopsy immunofluorescence findings (IgG/IgM/C3/C4 deposits; viral antigen staining). Whether dengue viral IC deposits vs. anti-nuclear IC deposits drove the GN cannot be determined. A study with both viral antigen staining and standard lupus nephritis immunofluorescence in dengue-associated GN would distinguish the two IC mechanisms.

• [RESOLVED 2026-04-17] ANA analysis page Rajadhyaksha2012 integration: Fully integrated as a primary source in the third-pass revision. §8 case contrast table expanded; Q14 on anti-cardiolipin added to open questions; §11 hypothesis-generating claim updated to reflect multiple independent case sources.

• Anti-EC severity dissociation — method heterogeneity gap (Hung2008 Q): Hung2008 (flow cytometry, % reactive cells) finds no anti-EC severity correlation; Wan2012 (unspecified ELISA method) finds anti-EC levels higher in DHF/DSS than DF. These may be measuring different antibody populations (IgM vs. IgG, different endothelial antigens). A head-to-head study measuring anti-EC by both flow cytometry (Hung2008 protocol) and ELISA in the same dengue cohort would determine whether the discrepancy is methodological or biological.

• Anti-EC IgG long-term autoimmune risk gap (Hung2008 Q): The isotype shift to anti-EC IgG in secondary dengue children (Hung2008) does not correlate with acute severity, but IgG autoantibodies have longer half-lives and complement-activating properties. Whether anti-EC IgG generated during secondary dengue contributes to the elevated autoimmune disease risk documented at 1–2 years post-dengue (Shih2023 aHR 1.88 for all autoimmune diseases) is unknown. A study measuring anti-EC IgG at the time of infection and again at 6M and 12M post-dengue would test whether acute-phase anti-EC IgG predicts later autoimmune diagnosis.

• TM non-correlation with anti-EC severity (Hung2008 Q): Thrombomodulin is elevated in both infants (6.1 vs 2.7 ng/mL) and children (8.8 vs 2.4 ng/mL), documenting in vivo endothelial structural damage in both groups. Yet TM does not correlate with anti-EC levels in either group, suggesting structural and immunological endothelial damage are dissociated. Whether TM correlates with vascular leakage markers (e.g., hematocrit rise, albumin drop) in the same cohort would test whether structural endothelial damage is clinically relevant even in the absence of anti-EC autoantibody severity effect.

• Anti-platelet IgM infection-order independence — confirmation gap (Hung2008 Q): Hung2008 shows anti-platelet IgM in both infants (primary) and children (secondary), consistent with Lin2001’s primary-infection finding. However, the children’s cohort in Hung2008 is predominantly but not exclusively secondary infection — anti-platelet IgM in confirmed primary vs. secondary children within the same cohort is not directly compared. A study directly comparing anti-platelet IgM in primary vs. secondary dengue children using identical methods would confirm infection-order independence.

• [RESOLVED 2026-04-19] ANA Review V2.0 Hung2008 integration complete: See resolved Farias2024 item above.

• Profibrinolysis/hyperfibrinolysis cohort validation gap (Ghorai2024 via Chuang 2013/2014): The two coagulation mimicry mechanisms (profibrinolysis via anti-plasminogen → fibrinogen cleavage; hyperfibrinolysis via anti-plasminogen complex → excess FDP) are based on Chuang group’s in vitro scFv experiments and reach the wiki only through Ghorai2024 (secondary review). No polyclonal dengue patient serum study has measured plasma FDP, anti-plasminogen titres, and fibrinogen levels simultaneously to test whether the bifurcation maps onto DF vs. DHF phenotypes in vivo. A prospective coagulation panel study would directly test this.

• Anti-DSG/pemphigus incidence in dengue — evidence level very low (Ghorai2024 Q): The anti-DSG pemphigus mechanism is presented in Ghorai2024 as a narrative review claim; the underlying evidence appears to be individual case reports rather than cohort data. No dengue study in this wiki has measured anti-DSG-1 or anti-DSG-3 serology prospectively. Whether dengue-triggered pemphigus is a reproducible clinical entity or an isolated case report coincidence cannot be determined. A study measuring anti-DSG ELISA in a dengue cohort would establish whether seroconversion occurs at detectable frequency.

• Lin 2008 hepatic model — human validation gap (Ghorai2024 Q): Anti-NS1 antibody-targeting of central and portal hepatic vein endothelium → fatty liver, necrotic body, liver fibrosis, mononuclear infiltration is a murine model finding (Lin 2008, cited in Ghorai2024). No human histopathological study in this wiki directly tests whether anti-NS1 AECA mediates hepatic vein endothelial apoptosis vs. other dengue hepatitis mechanisms (direct viral hepatocyte infection; NS1 TLR4 activation). A liver biopsy study correlating anti-NS1 AECA titres with histopathological endotheliitis patterns would directly test this in humans.

• Pemphigus and Acantholysis in Dengue page — thin (1 source): New concept page has only 1 source (Ghorai2024, itself a secondary review). Clinical case reports of dengue-triggered pemphigus or a primary study measuring anti-DSG serology in dengue would strengthen this page.