Morra2018 - Defining Warning Signs and Severe Dengue

Full citation: Morra, M. E., Altibi, A. M. A., Iqtadar, S., Minh, L. H. N., Elawady, S. S., Hallab, A., Elshafay, A., Omer, O. A., Iraqi, A., Adhikari, P., Labib, J. H., Elhusseiny, K. M., Elgebaly, A., Yacoub, S., Huong, L. T. M., Hirayama, K., & Huy, N. T. (2018). Definitions for warning signs and signs of severe dengue according to the WHO 2009 classification: Systematic review of literature. Reviews in Medical Virology, 28(3), e1979. https://doi.org/10.1002/rmv.1979

Raw file: [[raw/morra2018.pdf]]

Summary

This systematic review asks a deceptively simple operational question: when published studies use the WHO-2009 dengue classification, do they actually agree on what each “warning sign” and each “sign of severe dengue” means? The WHO-2009 scheme names its categories (Dengue without Warning Signs / Dengue with Warning Signs / Severe Dengue) but leaves most of the individual signs without precise, measurable defining criteria. The authors set out to document, sign by sign, how the literature has operationalized those criteria — frequency thresholds for vomiting, intensity descriptors for abdominal pain, hematocrit-rise cutoffs, platelet thresholds, blood-pressure combinations for shock, and so on.

Methodologically this is a PRISMA-compliant review (protocol registered PROSPERO CRD42015024105, July 2015) built on a Scopus-only electronic search — Scopus chosen for its peer-review coverage — restricted to English-language original articles published from 2010 onward. Three independent reviewers screened, with Endnote X7 deduplication and full texts retrieved through the Nagasaki University library. Of 490 records retrieved, 446 were excluded (389 provided no additional definitions, 30 case reports, 14 reviews, 3 posters, 10 non-English), leaving 44 included studies spanning roughly 15 countries, mixed designs, and both pediatric and adult cohorts, 2010–2015. (The abstract’s “491” is an internal typo; the Results text and PRISMA Figure 2 give 490.)

The headline conclusion is stark: of 16 warning/severe signs examined, only 2 had a consensus definition across studies — “liver enlargement” (warning sign) and “liver involvement” (severe sign) — and both are precisely the signs that WHO-2009 itself already pre-defines. The remaining 14 signs varied widely in their cutoffs and criteria. The authors conclude that this definitional heterogeneity undermines comparability across studies and likely contributes to the WHO-2009 scheme’s lower specificity, and they call for standardized, consensus-built definitions (via the Delphi technique under the WHO umbrella) validated against large multicentre datasets.

Study Design

• Type: Systematic review (PRISMA; PROSPERO CRD42015024105)
• Sample size: 44 included studies (from 490 screened; 446 excluded)
• Setting: Scopus search, English-language original articles 2010+; included studies span ~15 countries (Singapore, Brazil, Indonesia, Sri Lanka, Vietnam, Thailand, India, Malaysia, Nicaragua, Colombia, Philippines, Mexico, Paraguay, Germany, Netherlands), pediatric + adult, published 2010–2015
• Population: Patients in original retrospective/prospective/cross-sectional studies that classified dengue by WHO-2009 and provided additional sign definitions beyond WHO-2009 itself

Key Findings

• Only 2 of 16 signs had a consensus definition across studies — “liver enlargement” (warning sign) and “liver involvement” (severe sign) — and both are the signs WHO-2009 already pre-defines. The other 14 signs showed wide definitional variation (systematic review of 44 studies). This is the paper’s load-bearing result: agreement exists only where WHO supplied the definition; wherever the literature had to fill a gap, it diverged.
• Shock is the single starkest example of heterogeneity. Across 17 studies, Table 7 catalogs 23 distinct parameter-combinations used to define shock, built from narrow pulse pressure <20 mmHg (15 studies, 88.2%), hypotension <90 mmHg (12, 70.6%), tachycardia >100/min (11, 64.7%), and poor capillary perfusion (7, 41.2%). Most authors (n=14) accepted narrow pulse pressure alone as sufficient; Macedo et al notably required ≥2 hypoperfusion signs and did NOT accept narrow pulse pressure alone.
• Increase in hematocrit (22 studies): 19 used “>20% rise from baseline”; Thai et al used “>15%”; 4 used a gender-adjusted absolute cutoff (e.g. men >46–50% / women >40–44%); Rodrigues used a non-sex-adjusted >48%.
• Rapid decrease in platelet count (described 19× across 20 studies): thresholds ranged across <20,000 (2), <50,000 (10), <100,000 (8), <150,000 (1), plus one “drop ≥10,000/mm³ in 24h.”
• Persistent vomiting (6 studies): defined along three parameters — frequency, duration, impact — e.g. ≥6 episodes/24h or CTC grade ≥3 (Mercado); ≥3 episodes/12h preventing oral hydration (Malavige 2011); ≥2 consecutive days (Carrasco/Gan/Leo); vomiting with dehydration signs (Aung).
• Abdominal pain (5 studies): none specified an exact duration or intensity score — only qualitative descriptors (tenderness/continuous/diffuse; intense+continuous; continuous-not-intermittent; increasing/intense).
• Lethargy/restlessness (5 studies): Glasgow <15 ± Blantyre <5; alteration of consciousness; drowsiness/irritability.
• Clinical fluid accumulation (17 studies): all use pleural effusion + ascites but split on detection method — 6 require radiographic documentation (ultrasound for ascites, CXR for effusion), 2 allow clinical OR imaging; additional criteria included gallbladder wall thickening (5), edema (1), free fluid around bladder (1).
• Mucosal bleeding (15 studies): site frequencies — nose 15 (100%), gingival 13 (86.7%), GI 10 (66.7%), vaginal 11 (73.3%), respiratory/hemoptysis 6 (40%), urinary 5 (33.3%), skin 4 (26.7%), eye 4 (26.7%), ear 1 (6.7%).
• Liver enlargement [CONSENSUS SIGN #1] (14 of 16 studies): used the WHO-2009 definition (hepatomegaly >2 cm); 2 added “painful hepatomegaly.”
• Liver involvement [CONSENSUS SIGN #2] (19 studies): AST or ALT >1000 IU/L in 18 of 19 (94.7%) — the near-universal definition; 1 used an acute-liver-failure definition. This is the closest the severe-sign side comes to consensus, and again it tracks a WHO-pre-defined sign.
• Other severe signs varied: severe plasma leakage (7 studies, variably defined); respiratory distress (4 studies, all using respiratory rate but with incongruent cutoffs ≥24 / ≥30 / ≥40 / ≥60 breaths/min, some adding PaO2:FiO2, O2 saturation, or mechanical-ventilation criteria); cardiac involvement (12 studies — myocarditis 8, heart failure 2, both by echo 1, cardiomyopathy 1; most gave no diagnostic criteria); CNS involvement (10 studies — encephalitis 6 (60%), encephalopathy 5 (50%), convulsions 3 (30%), coma 1 (10%)); renal impairment (8 studies — creatinine ≥2× upper limit of normal 6 (75%), ≥2× baseline 5 (62.5%), >1.2 mg/dL 1, gender-specific 1); severe bleeding (13 studies — GI 12 (92%), need for transfusion 9 (69%) = WHO bleeding-scale grade 3, vaginal 7 (54%), plus idiosyncratic single-study criteria).
• Specificity comparison is borrowed, not pooled: the figures “specificity 73.0% (WHO-2009) vs 93.4% (WHO-1997)” come from Macedo et al (a Rio de Janeiro pediatric cohort, cited in this review) — they are NOT Morra’s own data, and this review performs no diagnostic-accuracy meta-analysis of its own. The authors attribute WHO-2009’s lower specificity partly to the very lack of clear defining criteria that the review documents.
• Recommendation: build consensus definitions via the Delphi technique at national/international forums under the WHO umbrella, validated against large multicentre data (the review points to the IDAMS consortium study, >8000 patients, identifying early <72h-fever progression markers). Stated value of standardization: preventing unnecessary hospitalization during outbreaks and harmonizing the comparability of dengue epidemiological studies and vaccine trials.

Methods Used

Plain-text — this is a clinical-classification systematic review, outside the wiki’s flow-cytometry / BCR method scope, so no method wikilinks are warranted:

• PRISMA-compliant systematic review (protocol PROSPERO CRD42015024105).
• Single-database electronic search (Scopus, chosen for peer-review coverage); keywords “warning signs”, “severe dengue”, “classification”; English-language original articles from 2010.
• Three independent reviewers; Endnote X7 deduplication; full-text retrieval via Nagasaki University library.
• Qualitative synthesis of how each warning/severe sign is operationally defined (no quantitative pooling, no diagnostic-accuracy meta-analysis).

Entities Mentioned

This is a clinical-classification paper with no B-cell, surface-marker, or flow-cytometry content — no B-cell entity links are warranted. The clinical signs it catalogs (vomiting, abdominal pain, lethargy, fluid accumulation, hematocrit rise, thrombocytopenia, hepatomegaly, mucosal/severe bleeding, shock, respiratory distress, cardiac/CNS/renal/liver involvement) are recorded here only as plain text.

Concepts Addressed

Dengue Severity Classification

Relevance & Notes

This is the second source on the wiki’s severity axis, which was opened on 2026-06-29 by Narvaez2011 - Evaluating WHO Dengue Severity Classifications. The two sources stack into a compounding lesson. Narvaez documents between-scheme heterogeneity — WHO-1997 and WHO-2009 agree only fairly on who is “severe” (κ=0.25), and a DENV-2→severe association present under 1997 vanishes under 2009. Morra adds the within-scheme layer: two studies that both label themselves “WHO-2009” may still operationalize “warning signs” and “severe dengue” differently, because WHO-2009 leaves 14 of 16 signs without precise criteria. The compounded takeaway for this wiki: a severity-stratified claim must carry not just which scheme produced it but, ideally, which operational definitions — the shared “WHO-2009” label can conceal real differences in who got counted as severe.

This matters concretely because the wiki’s own dengue sources do not share definitions: the GodoyLozano2016 (Lower IgG SHM Rates in Acute Dengue) and Ansari2025 (Peripheral T Helper Subset Drives B Cell Response in Dengue) source pages stratify by WHO-2009, while the GarciaBates2013 (Plasmablast Response and Dengue Severity) source page uses Brazil’s national DF/DFC criteria. Even among the WHO-2009 studies, Morra implies their operational definitions of “severe” may not align. (These sibling source pages are named in plain text here to keep this page’s only wikilinks the two below.)

Foregrounded cross-link: Morra is partly a meta-review that contains the wiki’s existing severity source. Narvaez et al 2011 is reference #40 in Morra and appears in Morra’s Table 1 (row: “Narvaez et al 2011, cross-sectional, Nicaragua, 2005–2010, 544 patients, median age 8.5”), and Morra cites it across several sign-definition tables — abdominal pain, fluid accumulation, the platelet <100,000 group, and mucosal bleeding. So Narvaez2011 - Evaluating WHO Dengue Severity Classifications is one of the 44 studies whose definitional choices Morra is cataloging — a directly checkable nesting of the two sources.

Honest scoping of Morra’s contribution: it pools no diagnostic accuracy of its own and runs no quantitative meta-analysis. Its value is descriptive — it documents definitional practice across the literature, and the one accuracy comparison it reports (73.0% vs 93.4% specificity) is borrowed from Macedo et al, not generated here. Read this paper as evidence about how the field defines its terms, not as evidence about which scheme performs better. Limitations the authors themselves note, balanced: the review captured only papers that set out to define warning signs (it may miss studies where a definition appeared incidentally), it was restricted to English-language articles, and it searched Scopus only — each of these bounds the 44-study sample, though the heterogeneity finding is so wide that a larger sample would likely widen, not narrow, it.

Questions Raised

• Which of the wiki’s severity-stratified B-cell studies used divergent operational definitions of “severe” beneath a shared “WHO-2009” label, such that their findings are less comparable than the label implies — and can that be checked against the actual cutoffs each reported?
• Would harmonized (Delphi-consensus) definitions change which cohorts count as “severe,” and therefore which B-cell/plasmablast/DN-cell signals get attributed to severe disease?
• Given that shock alone was defined 23 different ways, are any wiki sources that invoke “shock” or “DSS” using mutually incompatible criteria — and does that affect how their cellular findings should be compared?
• Could a future B-cell pilot in this program pre-specify its severity endpoint at the level of operational definitions (not just scheme name) to avoid the comparability problem this review documents?