Prospective Validation of a Metabolic-Comorbidity Risk Score for Early Deterioration in Adult Dengue

Abstract

Background: Early identification of adults with dengue who are likely to deteriorate could improve monitoring and triage, especially in settings with high prevalence of diabetes and hypertension.

Objective: To prospectively validate a parsimonious risk score combining metabolic comorbidity with routine admission variables.

Methods: Adults with laboratory-confirmed dengue were recruited at three hospitals and followed through discharge. A previously specified five-item score included diabetes, hypertension, age ≥50 years, platelet count <100×10^9/L and haematocrit above the sex-specific reference range. The primary outcome was clinical deterioration within 72 hours, defined as severe dengue, shock, major bleeding, organ dysfunction or unplanned high-dependency transfer.

Results: Among 912 participants, 118 (12.9%) met the primary outcome. Event rates increased from 3.1% in the low-risk category to 11.8% in the intermediate category and 31.6% in the high-risk category. The score achieved an AUC of 0.78 (95% CI 0.74–0.82) with good calibration across deciles of predicted risk. Removing the two comorbidity variables reduced the AUC to 0.72. At a threshold selected to prioritise sensitivity, the model identified 86% of deteriorating patients while classifying 41% of the cohort as requiring enhanced monitoring. Performance was similar across hospitals and between men and women.

Conclusion: A simple score combining diabetes, hypertension and routine laboratory findings showed useful prospective discrimination for early dengue deterioration. External validation in different transmission settings and implementation studies are needed before clinical adoption.

Publication format: research abstract. No full-text manuscript accompanies this record.