Abstract
BACKGROUND: Risk of venous thromboembolism (VTE) recurrence remains high in patients with cancer-associated thrombosis (CAT), despite therapeutic anticoagulation. Identifying patients at risk of treatment failure is still a challenge. OBJECTIVES: We aimed to assess the performance of the Ottawa score in predicting VTE recurrence in a large homogeneous population of patients with CAT treated with the same anticoagulant, tinzaparin, for at least 3 months. METHODS: Individual patient data from 3 prospective cohort studies and 1 randomized controlled trial were pooled (PROSPERO: CRD42019119907). Clinical events of interest were adjudicated by independent central adjudication committees in all 4 studies. RESULTS: Among the 1413 patients included, the Ottawa score could be calculated for 1088 of whom 646 (59.4%) were classified at high risk of recurrence (Ottawa score ≥ 1). The 6-month cumulative incidence of recurrent VTE was 5.0% (95% CI, 3.2-7.8) in the Ottawa low-risk group and 8.5% (95% CI, 6.6-10.8) in the high-risk group. The area under the receiver operating characteristic curve was 0.56 (95% CI, 0.51-0.62). The sensitivity of the dichotomized Ottawa score (score ≥ 1) was 72.8% (95% CI, 62.6%-83.0%), the specificity was 41.9% (95% CI, 37.8%-45.9%), the positive predictive value was 8.6% (95% CI, 6.4%-10.8%), and the negative predictive value was 95.3% (95% CI, 93.3%-97.4%). Introducing additional predictive factors failed to significantly improve the score's performance. CONCLUSIONS: Despite the large number of patients and anticoagulant treatment standardization, the Ottawa score failed to accurately predict recurrent VTE in patients with CAT treated with tinzaparin.