Derivation and Validation of Risk Prediction Models for Cardiovascular and Kidney Outcomes of...

Area of Science:

• Cardiology
• Nephrology
• Clinical Risk Prediction

Background:

• Patients with chronic kidney disease (CKD) face elevated risks of adverse events post-acute coronary syndrome (ACS).
• Optimized treatment strategies are crucial for this high-risk population.
• Existing risk stratification may not adequately address combined cardiac and renal outcomes in CKD patients post-ACS.

Purpose of the Study:

• To derive and validate risk models for predicting key adverse outcomes in patients with CKD following ACS.
• To assess the models' performance in predicting 1-year mortality, myocardial infarction (MI) readmission, and kidney failure.
• To provide tools for enhanced shared decision-making in managing CKD patients with ACS.

Main Methods:

• Development of risk models using a large cohort of adults with CKD admitted for non-ST-segment-elevation ACS in Alberta, Canada (2004-2017).
• External validation and updating of models in temporally and geographically distinct CKD patient cohorts in Canada.
• Utilized Cox proportional hazard and Fine and Gray competing risk models for analysis.

Main Results:

• The derivation cohort comprised 11,980 patients; validation cohorts included 4,204 (temporal) and 1,787 (geographic) patients.
• Models demonstrated very good discrimination for kidney failure (C-indexes up to 0.93) and modest discrimination for mortality (up to 0.77) and MI readmission (up to 0.65).
• Models were well-calibrated in validation cohorts, especially after updating, showing comparable performance.

Conclusions:

• Validated risk models for cardiac and renal outcomes in CKD patients with ACS have been developed and updated.
• These models can significantly aid clinicians in shared decision-making regarding patient management.
• The tools support informed choices about diagnostic testing, therapeutic interventions, and monitoring strategies.