A new simplified predictive model for mortality in methicillin-resistant Staphylococcus aureus bacteremia

Area of Science:

• Infectious Diseases
• Clinical Epidemiology
• Medical Informatics

Background:

• Observational studies of methicillin-resistant Staphylococcus aureus bacteremia (MRSAB) require careful adjustment for confounding factors.
• Patient disease severity and baseline health status in MRSAB vary widely, complicating mortality prediction.

Purpose of the Study:

• To develop a simplified, MRSAB-specific scoring model for estimating 30-day all-cause mortality risk.
• To compare the performance of the new MRSAB score against established models like APACHE II and Pitt Bacteremia scores.

Main Methods:

• Retrospective, single-center cohort study of 455 adult MRSAB patients (2008-2018).
• Multivariable logistic regression identified independent mortality predictors.
• A regression coefficient-based scoring method derived the MRSAB score.
• Discriminatory ability assessed using the c statistic.

Main Results:

• The MRSAB score incorporates six variables: age, respiratory rate, Glasgow Coma Scale, renal failure, hospital-acquired MRSAB, and infection source (infective endocarditis or lower respiratory tract).
• The MRSAB score demonstrated strong discrimination (c statistic = 0.8662), outperforming APACHE II (P=0.043) and Pitt Bacteremia (P<0.001) scores.
• Thirty-day mortality in the cohort was 16.3%.

Conclusions:

• A weighted combination of six routinely measured variables in MRSAB patients effectively predicts 30-day mortality with high discrimination.
• The developed MRSAB score offers a superior alternative to existing models for risk stratification.
• External validation is recommended prior to widespread clinical implementation.