BLEED-Myocardial Infarction Score: Predicting mid-term post-discharge bleeding events.

Area of Science:

• Cardiology
• Clinical Prediction Models
• Hemorrhagic Risk Assessment

Background:

• Post-myocardial infarction (MI) patients face significant risks of bleeding events after discharge.
• Accurate prediction of mid-term bleeding is crucial for optimizing patient management and preventing adverse outcomes.

Purpose of the Study:

• To derive and validate the BLEED-MI score for predicting mid-term bleeding events in patients post-MI.
• To assess the score's calibration, accuracy, and discrimination in independent cohorts.

Main Methods:

• A risk model (BLEED-MI) was developed using data from 1050 MI patients.
• The model incorporated clinical variables, laboratory values, and treatment data at admission and discharge.
• Validation was performed in an independent cohort of 852 MI patients.

Main Results:

• The BLEED-MI score demonstrated good calibration (P > 0.37) and accuracy (Brier score < 0.07) in both derivation and validation cohorts.
• Discriminative performance for bleeding events was high (c-statistic 0.718–0.753).
• A BLEED-MI score below 2 indicated a 98.7% negative predictive value for significant bleeding in the validation sample.

Conclusions:

• The BLEED-MI score is a novel, validated tool for predicting mid-term bleeding risk in post-MI patients.
• This score offers a bedside solution for identifying patients requiring closer monitoring or tailored therapeutic strategies.
• External validation in larger cohorts is recommended before widespread clinical implementation.