The predictive value of risk assessment models for venous thromboembolism on gynaecological cancer patients

Abstract

Background: Venous thromboembolism (VTE) is a multifactorial disease. There are two main clinical entities that are associated with morbidity and mortality, deep vein thrombosis and pulmonary embolism. Our study aimed to compare the three risk assessment models (RAMs), Khorana, Caprini and Padua in terms of predicting VTE in gynaecologic oncology patients.

Methods: Patients were retrospectively scored according to Caprini, Padua and Khorana scoring models to assess the risk for VTE. Accuracy analysis of risk assessment models was performed using sensitivity, specificity, positive and negative predictive values as well as the area under the curve of each model per patient.

Results: The Caprini score has good sensitivity (80.0), a poor specificity (24.3), low positive predictive value (7.2) and good negative predictive value (94.3) (95% CI). The Khorana score has a poor sensitivity (30.0), a fair specificity (62.5), low positive predictive value (5.6) and good negative predictive value (92.4) (95% CI). The Padua score has an average sensitivity (60.0), a poor specificity (42.6), low positive predictive value (7.1) and good negative predictive value (93.5) (95% CI). The Caprini score had the overall best performance.

Conclusion: Caprini score performed better and proved to be the best score. It has the potential to reduce mortality associated with VTE in gynaecological cancer patients. However, the Caprini score needs to be tested in the same population in a prospective study in a multicentre.

Contribution: The results of this study prove to us that Caprini score is the best to be used in a South African setting.