Inability to validate a predictive score for infective endocarditis in intravenous drug users

Area of Science:

• Infectious Diseases
• Cardiology
• Medical Diagnostics

Background:

• Endocarditis is a serious infection affecting heart valves, particularly in intravenous drug use (IVDU) patients.
• Accurate and timely diagnosis of endocarditis in IVDU patients remains a clinical challenge.
• Identifying reliable diagnostic markers is crucial for improving patient outcomes.

Purpose of the Study:

• To identify presenting features that improve the diagnostic accuracy of endocarditis in symptomatic IVDU patients.
• To develop and validate a clinical scoring system for endocarditis in this high-risk population.
• To assess the utility of clinical, radiographic, and laboratory findings in predicting endocarditis.

Main Methods:

• Retrospective analysis of two datasets of admitted IVDU patients from an urban teaching hospital.
• Development and validation of an endocarditis diagnostic score using multivariable analysis and likelihood ratios.
• Comparison of receiver operating characteristic curves to evaluate score performance.

Main Results:

• Multivariable analysis identified factors such as past endocarditis history (negative correlation), white blood cell count, neutrophil/band percentage, chest x-ray infiltrate, and arterial oxygenation (negative correlation) as associated with endocarditis.
• Prospective validation demonstrated poor predictive performance of the developed endocarditis score.
• No single clinical, radiographic, or laboratory feature reliably estimated endocarditis likelihood.

Conclusions:

• Presenting clinical, radiographic, and laboratory features are not sufficiently accurate for estimating endocarditis likelihood in symptomatic IVDU patients.
• The developed endocarditis score showed poor predictive performance upon validation.
• Further research is needed to identify more effective diagnostic strategies for endocarditis in IVDU populations.