Predictive model for CRT risk in cancer patients with central venous access devices: a systematic review and meta-analysis

  • 0Department of Nephrology, Zigong First People's Hospital, Zigong, Sichuan, China.

|

|

Summary

This summary is machine-generated.

Cancer patients with central venous access devices (CVADs) face a high risk of catheter-related thrombosis (CRT). This review evaluated existing risk prediction models for CVAD CRT, finding most have moderate-to-good performance despite high bias risk.

Area Of Science

  • Oncology
  • Hematology
  • Medical Informatics

Background

  • Central venous access device (CVAD) use is common in cancer patients.
  • Catheter-related thrombosis (CRT) is a significant complication with high incidence in this population.
  • Early detection and risk assessment are crucial for effective thromboprophylaxis.

Purpose Of The Study

  • To systematically review and evaluate risk prediction models for CVAD-related thrombosis in cancer patients.
  • To assess the performance and identify key predictors within these models.

Main Methods

  • Comprehensive literature search across multiple databases (PubMed, Embase, Web of Science, etc.) up to May 2024.
  • Independent screening, data extraction, and quality assessment using the Predictive Model Risk of Bias Assessment Tool (PROBAST).
  • Meta-analysis of Area Under the Curve (AUC) values for model validation.

Main Results

  • Nineteen papers identified 29 predictive models; reported AUC values ranged from 0.470 to 1.000.
  • Common predictors include D-dimer levels, BMI, and diabetes.
  • Combined AUC for six validated models was 0.81, indicating good discrimination, though all studies had a high risk of bias.

Conclusions

  • Available CRT prediction models demonstrate moderate-to-good predictive performance.
  • Significant risk of bias was noted across all included studies, primarily due to reporting limitations.
  • Future research requires methodologically rigorous, multi-center validation and potential development of new models.