Predictability of the Caprini Risk Assessment Model for Postoperative Venous Thromboembolism may be Increased by Global Test of Haemostasis

Abstract

Introduction - Postoperative venous thromboembolism (VTE) is still a challenge. Standard preventive protocol, that includes elastic compression stockings and low-molecular-weight heparin (LMWH) injection, may be not equally effective in different patients at high risk. Previously it has been shown that patients with 11 and more Caprini scores have a dramatically increased risk of VTE at the background of elastic compression and LMWH injections. The way to improve such situation is to understand, why standard prophylaxis does not work properly in some individuals. To achieve this goal may be used modern global tests of hemostasis. The aim of this study was to compare predictability for postoperative deep vein thrombosis (DVT) of the classic Caprini 2005 RAM and its modified version, considering the results of TD test in surgical patients at high risk. Methods - This was a prospective observational clinical study involving 80 patients (33 men and 47 women, mean age - 73,9±7,2 years), undergoing major (79 cases) or minor (1 case) surgery for colorectal cancer. Patients were at high risk of postoperative venous thromboembolism: 5-15 Caprini scores (mean - 9,9±2,0) and received combined prophylaxis: anti-embolic above-knee compression stockings and LMWH injections (enoxaparin 40 mg once daily) until discharge. LMWH was administered in a fixed hour according to the time of blood shambling for TD. Before and after the surgery (0 day, 5-7 day and 1 month) duplex ultrasound was performed to detect postoperative DVT. Results - DVT was found in 21 of 80 patients (26.3%; 95% CI: 17,9-36,8%) at 5-7 day after surgery. No new venous thromboembolism events were detected at 1 month. Regression analysis and ROC-curve showed a significant predictability of Caprini scores for DVT (p<0,0001, S=0,839±0,045). Analysis of ROC-curve coordinates showed a cut-off point of 11 scores, which with the sensitivity of 76.2% and a specificity of 74.6% predicted DVT. The results of TD test showed significant hypercoagulation at the background of LMWH administration in patients with DVT. Regression analysis and ROC-curves demonstrated that TD parameters as stationary velocity of clot growth, measured at 2 hours after surgery (Vst), and initial velocity of clot growth, measured 24 hours after the LMWH injection (Vin), had the greatest predictability for postoperative DVT: S=0,764 ± 0,109, p=0,049 and S=0,946±0,047, p=0,001, respectively. The cut-off points for DVT prediction with high sensitivity/specificity appeared to be Vst>32,5 (normal range 20-29) μm/min and Vin>64,5 (normal range 36-56) μm/min. Identified thresholds for TD parameters have been integrated into Caprini RAM under the item "other congenital or acquired thrombophilia" (score of 3). Total Caprini scores were recalculated in patients who had one or both TD parameters exceeded the cut-off, followed by re-analysis of the ROC-curves. It was found that the Caprini RAM with integrated Vin>64,5 μm/min characterized by the highest predictability for DVT (S=0,876±0,038, p<0,0001) and increased cut-off up to 12 scores with sensitivity of 85,7% and specificity of 72.9%. Conclusion - The integration of TD parameters into Caprini RAM allows increasing predictability for postoperative DVT.