Computed Tomography-Proven Active Bleeding with Negative Emergency Angiography: Early Mortality Predictors and
Louis Russier1, Ilies El Boukili2, Nory Elhadjene3
1Department of Radiology, University Hospital of Saint-Étienne, 42270, Saint-Étienne, Saint-Priest-en-Jarez, France.
Insights
This study identified key predictors of 30-day mortality in patients with CT-proven bleeding and negative angiography. The developed NAB-L score offers a simple bedside tool for risk stratification in these high-risk patients.
Area of Science:
- Radiology and Imaging
- Emergency Medicine
- Critical Care Medicine
Background:
- Computed tomography (CT) is crucial for diagnosing active bleeding.
- Negative emergency angiography in CT-proven bleeding patients presents a management challenge.
- These patients face significant risks of mortality and rebleeding.
Purpose of the Study:
- To identify predictors of 30-day mortality.
- To develop a bedside risk score for patients with CT-proven active bleeding and negative emergency angiography.
- To improve risk stratification and clinical decision-making in this cohort.
Main Methods:
- Retrospective analysis of 118 adult patients with CT-detected active bleeding and negative angiography.
- Univariate and multivariable logistic regression, including LASSO-penalized modeling.
- Development and validation of the Negative Angiography Bleeding-Lethality (NAB-L) score.
- Assessment of discrimination (AUC) and calibration (Hosmer-Lemeshow test).
Main Results:
- 30-day mortality was 14.4% and rebleeding occurred in 33.9% of patients.
- Independent predictors of mortality included low hemoglobin nadir (<8 g/dL), elevated arterial lactate (>2 mmol/L), and age (>70 years).
- The NAB-L score demonstrated good discrimination (AUC 0.854) and adequate calibration, with mortality increasing significantly with higher scores.
Conclusions:
- CT-positive/angiography-negative bleeding is associated with substantial early mortality and rebleeding.
- The NAB-L score provides a simple, bedside tool for early risk stratification.
- External validation is recommended for the NAB-L score pending further studies.
Purpose:
To identify predictors of 30-day mortality and derive a bedside risk score in patients with computed tomography (CT)-proven active bleeding and negative emergency angiography.
Materials And Methods:
We retrospectively included consecutive adults with CT-detected active bleeding and negative emergency angiography in a single-center cohort (2012-2026). Univariate analyses (Fisher exact, Mann-Whitney U) were followed by multivariable logistic regression, with a LASSO-penalized model (fivefold cross-validation) performed as sensitivity analysis. A simplified score, Negative Angiography Bleeding-Lethality (NAB-L), was derived from regression coefficients. Discrimination was assessed by area under the receiver operating characteristic curve (AUC) with bootstrap internal validation, calibration by the Hosmer-Lemeshow test, and survival by Kaplan-Meier methods with log-rank comparison.
Results:
A total of 118 patients were included (median age 67.5 years [IQR 51-78]; 55% male). Thirty-day mortality was 14.4% (17/118) and rebleeding occurred in 33.9% (40/118). Hemoglobin nadir < 8 g/dL (adjusted odds ratio (aOR) 11.44; 95% confidence intervals (CI) 1.45-90.35), arterial lactate > 2 mmol/L (aOR 5.69; 1.27-25.42), and age > 70 years (aOR 3.63; 1.01-13.02) were independently associated with mortality. The NAB-L score (0-5) showed good discrimination (AUC 0.854; 95% CI 0.760-0.941; optimism-corrected 0.816) and adequate calibration (Hosmer-Lemeshow p = 0.946). Mortality rose from 0% (score 0-1) to 56% (score 5). Rebleeding was more frequent in high-risk patients (52% vs. 17%, p = 0.036).
Conclusion:
CT-positive/angiography-negative bleeding carries substantial early mortality and rebleeding risk. The NAB-L score enables simple early bedside risk stratification but should be considered hypothesis-generating pending external validation.
