Summary
Predictive tools for major bleeding (MB) using machine learning (ML) might be advantageous over traditional methods. We used data from the Registro Informatizado de Enfermedad TromboEmbólica ...(RIETE) to develop ML algorithms to identify patients with venous thromboembolism (VTE) at increased risk of MB during the first 3 months of anticoagulation. A total of 55 baseline variables were used as predictors. New data prospectively collected from the RIETE were used for further validation. The RIETE and VTE‐BLEED scores were used for comparisons. External validation was performed with the COMMAND‐VTE database. Learning was carried out with data from 49 587 patients, of whom 873 (1.8%) had MB. The best performing ML method was XGBoost. In the prospective validation cohort the sensitivity, specificity, positive predictive value and F1 score were: 33.2%, 93%, 10%, and 15.4% respectively. F1 value for the RIETE and VTE‐BLEED scores were 8.6% and 6.4% respectively. In the external validation cohort the metrics were 10.3%, 87.6%, 3.5% and 5.2% respectively. In that cohort, the F1 value for the RIETE score was 17.3% and for the VTE‐BLEED score 9.75%. The performance of the XGBoost algorithm was better than that from the RIETE and VTE‐BLEED scores only in the prospective validation cohort, but not in the external validation cohort.
The duration of anticoagulation for a first episode of unprovoked venous thromboembolism (VTE) should balance the likelihood of VTE recurrence against the risk of major bleeding.
Analyze rates and ...case-fatality rates (CFRs) of recurrent VTE and major bleeding after discontinuing anticoagulation in patients with a first unprovoked VTE after at least 3 months of anticoagulation.
We compared the rates and CFRs in patients of the Registro Informatizado Enfermedad Trombo Embólica (RIETE) and Contemporary management and outcomes in patients with venous thromboembolism registries. We used logistic regression models to identify predictors for recurrent pulmonary embolism (PE) and major bleeding.
Of 8261 patients with unprovoked VTE in RIETE registry, 4012 (48.6%) had isolated deep vein thrombosis (DVT) and 4250 had PE. Follow-up (median, 318 days) showed 543 recurrent DVTs, 540 recurrent PEs, 71 major bleeding episodes, and 447 deaths. The Contemporary management and outcomes in patients with venous thromboembolism registry yielded similar results. Corresponding CFRs of recurrent DVT, PE, and major bleeding were 0.4%, 4.6%, and 24%, respectively. On multivariable analyses, initial PE presentation (hazard ratio HR, 3.03; 95% CI, 2.49-3.69), dementia (HR, 1.47; 95% CI, 1.01-2.13), and anemia (HR, 0.72; 95% CI, 0.57-0.91) predicted recurrent PE, whereas older age (HR, 2.11; 95% CI, 1.15-3.87), inflammatory bowel disease (HR, 4.39; 95% CI, 1.00-19.3), and anemia (HR, 2.24; 95% CI, 1.35-3.73) predicted major bleeding. Prognostic scores were formulated, with C statistics of 0.63 for recurrent PE and 0.69 for major bleeding.
Recurrent DVT and PE were frequent but had low CFRs (0.4% and 4.6%, respectively) after discontinuing anticoagulation. On the contrary, major bleeding was rare but had high CFR (24%). A few clinical factors may predict these outcomes.