Validation of a Predictive Model for Indentifying An Increased Risk for Thromboembolism in Children with Acute Lymphoblastic Leukemia: Results of a Multicenter Cohort Study

Abstract

Children with acute lymphoblastic leukemia (ALL) are at increased risk for venous thromboembolism (VTE), however, not all children experience a VTE. Developing a predictive model for determining children at increased risk would be beneficial in targeting interventional studies to only high risk groups. A recent meta-analysis of studies in VTE in children with ALL identified four potential risk factors:treatment with Escherichia coli asparaginase (CASP),concomitant use of steroids,presence of central venous lines andthrombophilic genetic abnormalities.As VTE in childhood ALL is well recognized as serious clinical problem and due to the lack of studies on prevention, the standard of practice varies and some centres use enoxaparin prophylaxis for these children. However, the risks and benefits of the intervention are unknown. The aim of the study wasto develop a simple model for predicting ALL-chemotherapy-associated VTE using baseline clinical and laboratory variables, andto evaluate, on an explorative basis, the increasing off-label use of enoxaparin for VTE prophylaxis in ALL children.For development of the risk model the predictive variables were scored as follows: treatment with CASP (5000–10000/m2) in combination with prednisone or dexamethasone, presence of central venous lines, thrombophilic genetic abnormalities, e.g. positive family history for VTE or identification of a single thrombophilic trait (1 point each), or carrier status of combined thrombophilic traits (2 points). A definition of VTE risk by score was low (1–2) and high (□ 3). The risk score was than prospectively validated in an independent cohort of 136 newly recruited patients enrolled into the German database. Seven patients were excluded (lost to follow-up n=2; death n=2, secondary malignancy, VTE before ALL-onset, infant < 12 months of age: each n=1). The cumulative VTE rates at 3.5 months in the validation cohorts were 3.6% (95% CI 1%–9%) in the low-risk group (4 of 112), and 47% (95%CI 23%–72%) in the high-risk category (8 of 17). In multivariate analysis [Cox regression] the high risk group was significantly associated with VTE when compared to the low risk group even after adjusting for age at ALL-onset, duration of CASP administration, steroid administered (prednisone/dexamethasone), and presence or absence of enoxaparin prophylaxis [hazard/95%CI: 4.16/1.13–15.34]. The negative predictive value for VTE was 96.3% [95%CI: 92.9–99.8]. Early enoxaparin prophylaxis reduced the absolute VTE-risk about 60% [95%CI: 23–96]. Therefore, the model can identify ALL-children with an increased risk for symptomatic VTE.