Validation of a predictive model for identifying an increased risk for recurrence in adolescents and young adults with a first provoked thromboembolism

Background: To develop and validate a predictive model to determinate patients at increased risk to suffer from recurrence following a first provoked deep vein thrombosis (VTE). Methods: Predictive variables, i.e. male sex [1 point], inherited thrombophilia (IT) status (none [0 points], single [1 point], combined variants [2 points]), blood group non-0, and age at first VTE onset were included into a risk assessment model, which was derived in 511 patients and then validated in 509 independent subjects. Results: VTE recurrence risk score (maximum 4 points, range 0–3) was below two for patients scored as low-risk (LRS) and ≥2 for patients at high-risk (HRS). Within a median time of 3 years after withdrawal of anticoagulation (AC) recurrence rate in LRG (derivation) was 11.8% versus 26.0% in HRS (p < 0.001). In the validation cohort within 2.2 years the recurrence rate was 9.8% in LRS versus 30.1% in HRS (p < 0.001). In multivariable analysis adjusted for age at first VTE and blood group the recurrent risk in HRS was significantly increased compared with the LRS (derivation: hazard/95% confidence interval: 3.7/1.75–7.91; validation: 4.7/2.24–9.81; combined 5.2/1.92–13.9). Model specificity (sensitivity) was 79.0% (52.0%) in the derivation cohort compared with 78.0% (43.0%) in the validation group. In conclusion, in the prediction model presented here the risk of VTE recurrence was associated with male gender and combined ITs. Based on the negative predictive value calculated the model may identify patients with a first provoked VTE not being at risk for recurrence.