Abstract:
Objective To evaluate the effectiveness of the Caprini risk assessment model in predicting deep venous thrombosis in hospitalized patients with malignant tumors.
Methods Deep venous thrombosis screening was performed in 504 patients with malignant tumors who were hospitalized in Beijing Shijitan Hospital between January 2015 and January 2017. Their Caprini thrombosis risk model scores and risk classifications were analyzed and compared with those of the Khorana risk model.
Results The median Caprini score of patients with deep venous thrombosis was 6 (range 4-8), which was higher than the score of 5 (range 4-7) in the group without deep venous thrombosis (Z=10.033, P=0.004). Statistically significant differences in the incidence of deep venous thrombosis were found among the low-medium, high-, and extremely high-risk groups (Z=-1.933, P=0.053). The area under the receiver-operating characteristic curve (AUC) of the Caprini scores was 0.61195% confidence interval (CI):0.54-0.69, P=0.004, and the cutoff value was 6 points, with the largest Youden index. The AUC of the Khorana model was 0.65 (95% CI:0.57-0.72, P < 0.001), and the difference between the Khorana and Caprini models was not statistically significant (Z=0.674, P=0.500). The AUC of the Caprini model was 0.85 (95% CI:0.66-0.96, P < 0.01) and that of the Khorana model was 0.68 (95% CI:0.47-0.84, P=0.18) in the patients who underwent malignant tumor surgery. The AUC of the Khorana model was 0.72 (95% CI:0.61-0.82, P=0.01) and that of the Caprini model was 0.55 (95% CI:0.44-0.67, P=0.54) in the non-operative patients who received chemotherapy.
Conclusions The Caprini and Khorana risk assessment models have certain predictive values, but the discrimination is not good. The Caprini model is providing better predictability in patients with malignant tumors treated with surgery. The Khorana model is suitable for non-operative patients who received chemotherapy. Further studies on the application of the Caprini risk assessment model in patients with malignant tumors are needed.