Ferritin-to-Platelet Ratio Predicts Dengue Shock and Ferritin/ALT Ratio Discriminates Severe Hepatitis: Novel Composite Indices from a Prospective Serial Biomarker Study 

Technical Summary: Novel Ferritin Composite Indices as Dengue Triage Biomarkers

Problem Statement
Dengue fever represents a critical public health challenge in India, particularly in the state of Gujarat, where seasonal outbreaks are recurrent. A primary clinical difficulty lies in the early prediction of severe complications, specifically Dengue Shock Syndrome (DSS) and severe hepatitis, during the critical phase of the disease (Days 3–7). Current triage protocols in Indian tertiary hospitals rely heavily on serial platelet counts; however, thrombocytopenia is often a delayed and non-specific marker that fails to reliably precede hemodynamic shock or identify impending organ injury early enough to alter management. While serum ferritin is known to be an acute-phase reactant elevated by macrophage activation in dengue, its serial kinetic profile across the febrile-to-critical phase transition and the utility of novel ferritin-derived composite indices have not been systematically evaluated in adult Indian populations.

Methodology
This prospective observational study was conducted at SMIMER Medical College and Hospital in Surat, India, between June 2020 and November 2021. The cohort consisted of 100 adult patients (≥18 years) with WHO 2009-confirmed dengue fever. Patients with conditions independently altering ferritin (e.g., chronic inflammation, iron overload, severe anemia) were excluded.

• Data Collection: Serum ferritin was measured via sandwich ELISA on Day 1 (admission), Day 3, and Day 5. Clinical severity was classified as non-severe (n=72) or severe (n=28) based on WHO 2009 criteria.
• Endpoints: To avoid circularity (where the biomarker is part of the severity definition), the study utilized four non-circular independent endpoints for diagnostic accuracy: shock/DSS, severe thrombocytopenia (<20,000/μL), severe hepatitis (AST or ALT ≥1,000 IU/L), and severe bleeding.
• Novel Indices: Two pre-specified composite indices were derived:
  1. Ferritin-to-Platelet Ratio (FPR): Ferritin (ng/ml) ÷ Platelet count (×10³/μL).
  2. Ferritin/ALT Ratio (FLR): Ferritin (ng/ml) ÷ ALT (IU/L).
  3. (A Ferritin/AST Ratio was also calculated but less emphasized in the primary findings).
• Statistical Analysis: Non-parametric methods were used due to non-normal data distribution. Diagnostic accuracy was assessed via ROC curve analysis with bootstrap confidence intervals. Incremental predictive value was evaluated using Net Reclassification Index (NRI), Integrated Discrimination Improvement (IDI), and Decision Curve Analysis (DCA). Advanced analyses included trajectory clustering (k-means, Gaussian mixture models) and Principal Component Analysis (PCA).

Key Results

• Ferritin Kinetics: Serum ferritin followed a distinct "rise-peak-fall" trajectory, peaking on Day 3 (mean 955±177 ng/ml) before declining by Day 5. Severe patients exhibited significantly higher ferritin levels at all time points compared to non-severe patients.
• Diagnostic Accuracy of Raw Ferritin:
  • Against composite WHO severity, Day 1 ferritin showed an AUC of 0.997, but this was identified as circular due to quasi-complete separation in logistic regression.
  • Against non-circular outcomes, Day 1 ferritin predicted shock/DSS with an AUC of 0.789 and severe thrombocytopenia with an AUC of 0.727.
  • Raw ferritin showed no discriminatory value for severe hepatitis (AUC 0.539).
  • Ferritin did not provide incremental predictive benefit over platelet count alone (NRI=0.143, IDI=0.000).
• Performance of Composite Indices:
  • Ferritin-to-Platelet Ratio (FPR): This index demonstrated superior performance, achieving an AUC of 0.913 (95% CI 0.858–0.969) for predicting composite severity (specifically shock/DSS) on Day 1. Severe patients had a median FPR eight-fold higher than non-severe patients (56.0 vs. 7.3).
  • Ferritin/ALT Ratio (FLR): This index discriminated severe hepatitis with near-perfect accuracy (AUC 0.995; 95% CI 0.988–1.000). The ratio collapsed in severe hepatitis cases (median 0.30) compared to non-hepatitis cases (median 8.43) due to the disproportionate surge in ALT relative to ferritin.
• Correlations: A robust inverse correlation was observed between ferritin and platelet counts (partial Spearman rho = −0.392), which remained independent of age and sex.
• Kinetic Velocity: The rate of change (velocity) of ferritin between days performed poorly as a standalone predictor compared to single-point measurements, suggesting that a single admission-day measurement is sufficient for triage.

Significance and Claims
The authors claim that the study establishes two novel, cost-free composite indices that address specific gaps in dengue triage:

1. FPR as a Shock Predictor: The Ferritin-to-Platelet Ratio is presented as a highly accurate bedside tool for predicting dengue shock and severe thrombocytopenia, outperforming raw ferritin or platelet counts alone in terms of composite discrimination.
2. FLR as a Hepatitis Discriminator: The Ferritin/ALT Ratio is identified as a novel marker capable of distinguishing severe hepatitis with near-perfect accuracy, a capability raw ferritin lacks.
3. Kinetic Insight: The study characterizes the ferritin trajectory in Indian adults, confirming a Day 3 peak and noting that severe patients present with higher baseline ferritin but a slower rate of rise (ceiling effect) compared to non-severe patients.

The paper concludes that while ferritin alone does not add significant classification value over platelet counts, the derived composite indices (FPR and FLR) warrant prospective multi-center validation as practical triage tools in dengue-endemic settings. The authors explicitly state that these indices were derived and evaluated on the same dataset and require independent validation before clinical implementation.