Poster, Podium & Video Sessions
Presentation Authors: Ross Krasnow*, Mark Preston, Boston, MA, Benjamin Chung, Stanford, CA, Adam Kibel, Steven Chang, Boston, MA
Introduction: The Caprini Risk assessment model is widely used to risk stratify patients for the occurrence venous thromboembolism (VTE), and has been validated for non-urologic surgery. We sought to validate the Caprini risk assessment model in a contemporary cohort of patients undergoing major urologic cancer surgery.
Methods: A population-based cohort study comprised of a weighted sample of 1,099,093 patients from 490 United States hospitals undergoing radical prostatectomy, radical nephrectomy, partial nephrectomy, or radical cystectomy for malignancy from 2003 to 2013. The primary outcome was 90-day symptomatic VTE (pulmonary embolism or deep vein thrombosis). Patients were scored according to the Caprini risk assessment model. The association of risk factors with VTE was determined with logistic regression. The performance of the Caprini score, as a predictor of VTE, was quantified using receiver operating characteristic (ROC) curves.
Results: There were no patients in the low risk category, 0.9% in the moderate risk category, and 99.1% were high risk (38.1% high risk, 48.5% higher risk, and 12.5% highest risk). The incidence of postoperative VTE was 1.2% (0.6%, radical prostatectomy; 1.9%, radical nephrectomy; 1.0%, partial nephrectomy; 5.4%, radical cystectomy). Old age, obesity, central venous access, acute MI, abnormal pulmonary function, hypercoagulable states, past history of VTE, recent surgery, and immobilization were independent risk factors for VTE. While the Caprini score was generally associated with increased risk of VTE (odds ratio [OR] 1.21, 95% confidence interval [CI] 1.17-1.25, p<0.001), the Caprini score demonstrated poor discrimination in the prediction of VTE no matter if they received VTE chemoprophylaxis (ROC area 0.53, 95% CI 0.50-0.56) or did not receive VTE chemoprophylaxis (ROC area 0.58, 95% CI 0.56-0.59) (Figure).
Conclusions: While the Caprini risk assessment model has been validated in other surgical specialties, it is not a good predictor of venous thromboembolism in patients undergoing major urologic cancer surgery. It should not be used to risk stratify patients undergoing major urologic cancer surgery.
Source Of Funding: None