## Abstract

Abstract 394

Background:

Long-term low molecular weight heparin (LMWH) is the current standard for treatment of venous thromboembolism (VTE) in cancer patients. Whether treatment strategies should vary according to individual risk of VTE recurrence remains unknown. We have derived a clinical prediction rule that stratifies VTE recurrence risk in patients with cancer-associated VTE. The derivation model includes 4 independent predictors (sex, primary tumor site, stage and prior VTE). The score sum ranges between −3 and +3 points. Patients with a score ≤ 0 had low risk (≤4.5%) for recurrence and patients with a score above 1 had a high risk (≥ 19%) for VTE recurrence. Subsequently, we applied and validated the rule in an independent set of 819 patients from 2 randomized controlled trials comparing LMWH to warfarin for VTE treatment in cancer patients. In the current study we aim to externally validate our clinical prediction rule with an independent population of patients with cancer-associated VTE followed at the Thrombosis clinics of two tertiary Canadian centres.

Methods:

We conducted a retrospective cohort study of patients with cancer and VTE diagnosed and/or followed at the Thrombosis Clinic of the Victoria Hospital (London, Canada) from January 2006 to December 2010; and the Thrombosis Unit of the Ottawa Hospital (Ottawa, Canada) from January 2009 to December 2011. We included data from adult patients with active malignancy and objectively diagnosed acute pulmonary embolism (PE) or deep venous thrombosis (DVT) of the lower extremity (above knee), upper extremity and neck veins, or unusual site thrombosis. The primary outcome measure was VTE recurrence during the first six months of anticoagulation.

Results:

353 patients fulfilled our inclusion criteria and were included in the study. There were 149 males, and the overall population had a median age of 64 years (range: 18 – 95). One hundred and twenty-three patients had lower extremity DVT, 93 had PE and 57 had both. The remaining 80 patients had either upper extremity/neck DVT (n = 55) or unusual site thrombosis (n = 25). 77 patients had a prior history of VTE. The most common primary tumour site was gastrointestinal, followed by the lung. Of the 304 patients with solid tumours, 230 (75.7%%) had TNM greater than I.

Two hundred and ninety-three (83.0%) patients were treated with longterm low molecular weight heparin (LMWH) only and 60 (17.0%) with warfarin (VKA). VTE recurrence occurred in 44 of 353 patients (12.4%). When we evaluated VTE recurrence risk per site, there was no significant difference: London 13 of 90 and Ottawa 31 of 263 [RR=1.23 (95%CI= 0.671 – 2.237; p=0. 507)]. In addition, there was no significant benefit with the use of LMWH (37 of 293) over VKA (7 of 60) in the risk of recurrence [RR=0.92 (95%CI= 0.433 – 1.973; p= 0.8379)]. When we applied our clinical prediction rule (Table 1) in the entire study population, recurrent VTE occurred in 12 of 204 (5.8%) patients stratified as low risk probability and in 32 of 149 (21.4%) patients stratified as high risk probability (Table 2).

Conclusions:

Our prediction rule has been adequately validated to now be used in prospective trials of treatment. Future trials evaluating novel treatment strategies for high risk patients are warranted.

Table 1.

The Clinical prediction rule for risk stratification of VTE recurrence risk in patients with cancer- associated venous thrombosis

PredictorPoints
Female
Lung cancer
Breast Cancer −1
TNM$Stage I −2 Previous VTE** Clinical Probability Low risk of *rVTE (< or = 0) −3 to 0 High risk of *rVTE (> or = 1) 1 to 3 PredictorPoints Female Lung cancer Breast Cancer −1 TNM$ Stage I −2
Previous VTE**
Clinical Probability
Low risk of *rVTE (< or = 0) −3 to 0
High risk of *rVTE (> or = 1) 1 to 3
*

rVTE= recurrent venous thromboembolism.

**

VTE= venous thromboembolism.

\$

TNM= tumour- node- metastasis classification of solid tumours.

Table 2.

Results of the clinical prediction rule in the London-Ottawa retrospective study

Sum of PointsTotal Patients (n = 353)*rVTE (n = 44)Frequency of rVTE (%)
-3
-2 45 6.7
-1 56 8.9
102 3.9
97 16 16.5
31 13 42.0
37.5
Clinical Probability
Low Risk of rVTE (< or = 0) 5.8%
High Risk of rVTE (> or = 1) 21.4%
Sum of PointsTotal Patients (n = 353)*rVTE (n = 44)Frequency of rVTE (%)
-3
-2 45 6.7
-1 56 8.9
102 3.9
97 16 16.5
31 13 42.0
37.5
Clinical Probability
Low Risk of rVTE (< or = 0) 5.8%
High Risk of rVTE (> or = 1) 21.4%
*

rVTE= recurrent venous thromboembolism.

Disclosures:

No relevant conflicts of interest to declare.

## Author notes

*

Asterisk with author names denotes non-ASH members.