Risk prediction model for early outcomes of revascularization for chronic limb-threatening ischaemia

Area of Science:

• Vascular Surgery
• Medical Informatics
• Predictive Analytics

Background:

• Chronic limb-threatening ischemia (CLTI) poses significant risks, necessitating accurate outcome prediction after revascularization.
• Quantifying risks and benefits is crucial for optimizing treatment strategies in CLTI patients.
• The JAPAN Critical Limb Ischemia Database (JCLIMB) provides valuable data for clinical research.

Purpose of the Study:

• To develop and validate a risk prediction model for 30-day post-revascularization outcomes in CLTI patients.
• To identify key predictors of major amputation/death and major adverse limb events.
• To provide a tool for treatment decision-making and quality assessment in CLTI care.

Main Methods:

• Utilized data from 2906 CLTI patients in the JCLIMB database (2013-2016).
• Developed multivariable risk models using least absolute shrinkage and selection operator (LASSO) logistic regression.
• Validated the model's performance using c-statistic and calibration plots.

Main Results:

• The model predicted major amputation and/or death (5.0%) with a c-statistic of 0.82.
• Predictors for amputation/death included abnormal white blood cell count, emergency procedures, heart failure, fever, and hemodialysis.
• The model predicted major adverse limb events (6.4%) with a c-statistic of 0.79.
• Predictors for adverse limb events included abnormal white blood cell count and fever.

Conclusions:

• The developed risk prediction model demonstrates high accuracy for 30-day outcomes in CLTI patients.
• This model can assist clinicians in selecting appropriate treatment strategies for CLTI.
• The model may serve as a benchmark for evaluating the quality of care at medical facilities treating CLTI.