Admission hemoglobin-to-red cell distribution width ratio predicts short- and long-term mortality in critically ill patients with spontaneous intracerebral hemorrhage: a retrospective cohort study using the MIMIC-IV database

Technical Summary: Admission HRR as a Prognostic Biomarker in Spontaneous Intracerebral Hemorrhage

Problem Statement
Spontaneous intracerebral hemorrhage (ICH) remains a devastating subtype of stroke with high mortality and disability rates. While current prognostic assessment relies on clinical severity scores (e.g., GCS, ICH score, SOFA), these tools often require multiple variables or radiological data not immediately available in emergency settings. There is a need for inexpensive, rapidly obtainable, and reproducible laboratory biomarkers to aid in early risk stratification. Although the hemoglobin-to-red blood cell distribution width ratio (HRR) has shown prognostic value in malignancies, sepsis, and cardiovascular diseases, its specific utility in predicting short- and long-term mortality in critically ill patients with spontaneous ICH remains unclear. Previous studies have focused primarily on in-hospital outcomes or dynamic changes in HRR, leaving the relationship between baseline admission HRR and extended mortality periods (30, 90, and 365 days) insufficiently explored.

Methodology
This study utilized a retrospective cohort design based on data from the Medical Information Mart for Intensive Care IV (MIMIC-IV) database (2008–2019).

• Study Population: The analysis included 2,069 adult patients (≥18 years) diagnosed with nontraumatic intracerebral hemorrhage.
• Exclusion Criteria: Patients who died within 24 hours of ICU admission, those with multiple ICU admissions (only the first was included), patients lacking key laboratory data within the first 24 hours, and those with severe hepatic/renal disease or malignant tumors were excluded.
• Exposure Variable: The primary exposure was the admission Hemoglobin-to-Red Blood Cell Distribution Width Ratio (HRR), calculated as Hemoglobin (g/dL) ÷ RDW (%), measured within 24 hours of ICU admission.
• Outcomes: The primary outcome was 365-day all-cause mortality. Secondary outcomes included 30-day and 90-day all-cause mortality.
• Statistical Analysis:
  • Multivariable logistic regression models were constructed with progressive adjustment: Model 1 (unadjusted), Model 2 (adjusted for age and sex), and Model 3 (fully adjusted for age, sex, hypertension, diabetes, myocardial infarction, and dementia).
  • Generalized Estimating Equation (GEE) models were employed to account for within-subject correlations across repeated mortality outcomes (30, 90, and 365 days).
  • Restricted cubic spline regression was used to assess dose-response relationships, and Kaplan-Meier survival analysis was performed to evaluate survival curves stratified by HRR quartiles.

Key Results

• Baseline Characteristics: Patients who died during follow-up were generally older, had a higher prevalence of dementia, lower hemoglobin, higher RDW, and consequently lower admission HRR compared to survivors. They also presented with higher illness severity scores (SOFA, SAPS II, APS III).
• Association with Mortality: Lower admission HRR was significantly associated with higher mortality at all time points.
  • In the fully adjusted logistic regression models, each one-unit increase in HRR was associated with significantly lower odds of mortality:
    • 30-day mortality: OR 0.38 (95% CI 0.20–0.70)
    • 90-day mortality: OR 0.33 (95% CI 0.18–0.60)
    • 365-day mortality: OR 0.24 (95% CI 0.13–0.42)
• Longitudinal Analysis: GEE models confirmed that admission HRR is an independent predictor of mortality across repeated follow-up periods (OR 0.25, 95% CI 0.14–0.44).
• Dose-Response and Survival: Restricted cubic spline analysis indicated an approximately linear inverse association between HRR and 365-day mortality. Kaplan-Meier analysis showed a clear separation of survival curves, with 365-day survival probability increasing from 22.0% in the lowest HRR quartile (Q1) to 56.0% in the highest quartile (Q4).

Significance and Claims
The authors claim that admission HRR is a simple, readily available, and cost-effective biomarker that is independently associated with short-, intermediate-, and long-term mortality in critically ill patients with spontaneous ICH.

• Clinical Utility: Because HRR is derived from routine complete blood counts, it requires no additional testing or expenditure. It can be calculated immediately upon admission, potentially facilitating early identification of high-risk patients before complex prognostic models are available.
• Mechanistic Plausibility: The study posits that HRR captures the combined effects of anemia (reflecting impaired oxygen delivery) and systemic inflammation (reflected by RDW), both of which are central to secondary brain injury following ICH.
• Study Strengths: The research highlights the use of a large, high-quality database with comprehensive follow-up, the consistency of findings across multiple statistical models (logistic regression and GEE), and the stability of the association across different time horizons.
• Limitations: The authors acknowledge the retrospective nature of the study (precluding causal inference), the reliance on a single admission measurement without dynamic tracking, potential residual confounding, and the need for external validation in independent multicenter cohorts before widespread clinical application.

The paper concludes that while HRR should not replace existing severity scores, it may serve as a useful adjunctive tool to improve early risk stratification and prognostic assessment in this population.