Prediction of Catheter-Associated Thrombosis in Critically Ill Children

Area of Science:

• Pediatric critical care medicine
• Vascular access complications
• Thrombosis risk prediction

Background:

• Untunneled central venous catheters are essential in critically ill children but carry a risk of catheter-associated deep venous thrombosis (CA-DVT).
• Early identification of children at high risk for CA-DVT is crucial for timely intervention and prevention of complications.

Purpose of the Study:

• To determine if the risk of CA-DVT in critically ill children with untunneled central venous catheters can be predicted within 24 hours of insertion.
• To develop and validate risk prediction models for CA-DVT in this pediatric population.

Main Methods:

• Secondary analysis of two multicenter prospective cohort studies involving 175 critically ill children.
• Development of risk prediction models using logistic regression and recursive partitioning, incorporating predictors present within 24 hours of catheter insertion.
• Active surveillance with ultrasonography was used to detect CA-DVT, with 30.3% of children developing thrombosis.

Main Results:

• Logistic regression and recursive partitioning models were developed to predict CA-DVT risk.
• Models incorporating predictors within 24 hours of catheter insertion (including age, recent surgery, catheter location, blood product transfusion, and severity of illness) showed good predictive performance (Area Under Curve ~0.80).
• Model 4, using recursive partitioning with severity of illness, effectively identified groups of children at low, intermediate, and high risk for CA-DVT.

Conclusions:

• A predictive model can identify critically ill children at high risk for catheter-associated thrombosis within 24 hours of untunneled central venous catheter insertion.
• This early risk stratification allows for targeted monitoring and potential preventative strategies in pediatric intensive care units.