Infective endocarditis: Do we have an effective risk score model? A systematic review.

Area of Science:

• Cardiology
• Infectious Diseases
• Medical Informatics

Background:

• Infective endocarditis (IE) is a severe condition with high mortality and surgical rates.
• Predicting patient outcomes and guiding interventions for IE remains challenging.
• Existing IE risk scores lack consistent validation and transportability.

Approach:

• A systematic review following PRISMA guidelines evaluated available IE risk scores.
• 32 papers analyzing 20 distinct risk scores were included, with a focus on performance metrics like AUC/ROC.
• Risk-of-bias assessment was conducted using PROBAST guidelines.

Key Points:

• 14 scores were specific to IE, with variable numbers of predictors (3-14).
• Many scores performed well in derivation cohorts but poorly in external validation cohorts.
• Few scores incorporated microbiological variables (50%) or biomarkers (15%).

Conclusions:

• Current IE risk scores are limited by small sample sizes, retrospective data, and inadequate external validation.
• The transportability of existing scores is compromised, hindering widespread clinical application.
• Development of robust IE risk prediction models requires large-scale population studies and comprehensive registries.