Postoperative venous thromboembolism risk stratification in patients with uterine cancer

Background: Uterine cancers are associated with a high risk for venous thromboembolisms. The American Society of Clinical Oncology practice guidelines recommend that all patients undergoing pelvic surgery for cancer should receive extended pharmacologic thromboprophylaxis with the duration being dependent on risk. However, risk stratification for patients with uterine cancer is not clearly defined. The Caprini score is the most widely used risk assessment model but it has been found to have limited use in the gynecologic oncology population. A modified Caprini score has been explored in other populations. The Khorana score is an additional risk assessment model that has not been studied in this context.

Objective: Our objective was to evaluate the ability of a modified Caprini model and the Khorana score to risk stratify patients with uterine cancer for postoperative venous thromboembolisms within 90 days of surgery.

Study design: Following institutional review board approval, a retrospective cohort study was performed, and all patients with uterine cancer who underwent a hysterectomy over a 4-year period were included. The Caprini and Khorana scores were calculated for each patient. The Caprini score cutoff for highest risk was evaluated at ≥7, ≥8, and ≥9 (modified Caprini) and the Khorana score cutoff was evaluated at ≥2 and ≥3. To determine the prognostic use of each score and other clinico-pathologic criteria related to the development of a venous thromboembolism, univariate analyses were performed using independent t tests, chi-square tests, or Fisher's exact tests; a multivariate analysis was performed using logistic regression.

Results: A total of 954 patients were included. The rate of venous thromboembolism development was 1.7% (16/954). A minimally invasive surgical approach was used in 90.5% (863/954) of patients. The mean Caprini score for patients with a venous thromboembolism was 10.3 compared with 8.1 for patients without a venous thromboembolism (95% confidence interval, 1.17-3.33; P<.0001). The mean Khorana score for the venous thromboembolism group was 2.4 vs 1.9 for those without (95% confidence interval, 0.04-0.82; P=.03). Both the Caprini and Khorana scores were found to be associated with venous thromboembolisms, but only a Caprini score with a cutoff of ≥8 or ≥9 was statistically significant (risk ratio, 31.25; 95% confidence interval, 1.88-519.49; risk ratio, 4.59; 95% confidence interval, 1.49-14.13, respectively), with high accuracy based on the area under the curve (0.75 and 0.68, respectively). Of the minimally invasive subgroup, 11.7% (101/863) of patients had same-day discharge with no postoperative thromboprophylaxis; none of these patients developed venous thromboembolisms. Despite extended prophylaxis among the laparotomy patients (30 days), the rate of venous thromboembolisms was more than 3 times that of the minimally invasive group (5.49% vs 1.7%). Advanced tumor stage and leukocytosis were noted to be independent risk factors for venous thromboembolisms.

Conclusion: Our study suggests that using a modified Caprini score could help to identify the highest-risk patients who would benefit from prolonged thromboprophylaxis, could reduce the incidence of postoperative venous thromboembolisms, and could minimize the cost and harm of overtreatment. These findings need to be validated in a prospective manner, and further research is needed to determine the optimal duration of therapy.