Prediction of Venous Thromboembolism in Patients with Cancer By the Activated Partial Thromboplastin Time: Results from the Vienna Cancer and Thrombosis Study

INTRODUCTION: Venous thromboembolism (VTE) is a frequent and burdensome complication of malignancy. Accurate prediction of cancer-associated VTE may pave the way towards targeted thromboprophylaxis. In this study, we investigated whether the activated Partial Thromboplastin Time (aPTT), a global function test of coagulation reflecting the activity of the intrinsic pathway, can be used to predict the occurrence of VTE in patients with cancer.

PATIENTS & METHODS: In a prospective and observational cohort study, 1869 patients with newly-diagnosed malignancy or progressive disease after remission were prospectively followed for 2 years until the onset of VTE, death or censoring (included tumor sites: brain (n=261), breast (n=275), lung (n=317), gastric (n=64), colorectal (n=187), pancreas (n=131), kidney (n=43), prostate (n=157), haematologic (n=310), others (n=124)). Tumor entities were grouped into 3 groups of low/moderate (n=504), high (n=909), and very high (n=456) VTE risk tumor sites in analogy to the Khorana score with extensions according to our in-house VTE results (Ay et al. Blood. 2010. 24: 5377). Primary endpoint was symptomatic, objectively-confirmed and independently adjudicated fatal or non-fatal VTE. The baseline aPTT was measured using a commercially available assay (STA-PTT A, Diagnostica Stago, Asnieres, France).

RESULTS: During the two-year study period, 151 patients (8.1%) developed VTE, and 733 (39.2%) patients died. The cumulative 6-month, 1-year, and 2-year incidence of VTE (incorporating mortality as a competing risk) was 5.8% (95%CI: 4.8-6.9), 7.4% (95%CI: 6.2-8.6), and 8.7% (95%CI: 7.4-10.0), respectively. The median baseline aPTT was 33.2 seconds (25^th – 75^th percentile: 30.8-36.3). The aPTT was found to be shorter in patients with higher VTE risk tumor sites (Kruskal-Wallis p<0.001), and patients with high-grade malignancy in histological investigations (p<0.001). A negative correlation was observed between the aPTT and factor VIII activity (rho=-0.30, p<0.001). In competing risk analysis, the 12-month cumulative incidence of VTE was almost two times higher in patients with short baseline aPTT (defined as an aPTT below a cut-off at the 25^th percentile (Q1, 30.8 seconds) of its distribution) than in patients with an aPTT ≥ this cut-off (11.4% (95%CI: 8.6-14.6) vs. 6.1% (95%CI: 4.9-7.4), Gray’s test p<0.001, Figure 1). In time-to-VTE regression analysis, we observed a highly significant association between shorter aPTT and a higher risk of VTE (Subdistribution Hazard Ratio (SHR) per 1 second increase in aPTT = 0.90, 95%CI: 0.86-0.94, p<0.0001). This result prevailed after adjusting for factor VIII and low/moderate, high, and very high VTE risk tumor sites (adjusted SHR per 1 second increase in aPTT = 0.93, 95%CI: 0.90-0.97, p=0.001). Further multivariable adjustments for D-Dimer, soluble P-Selectin, peak of thrombin generation, or prothrombin fragment 1+2 did not alter the strength of association between the aPTT and the risk of VTE.

CONCLUSION: We observed an impressive and highly significant association between a shorter aPTT and an increased risk of VTE in a large, prospective, observational cohort of cancer patients including a wide range of tumor entities. This finding was independent of factor VIII activity, high and very high VTE risk tumor sites, and several important biomarkers of hypercoagulability. It could be suggested that a shortened aPTT indicates the activation of the intrinsic pathway in cancer patients, and may reflect the factor VIII-independent convergence of one or more specific prothrombotic pathways of malignancy. Further, this routinely available and low-cost parameter may provide predictive potential in stratifying cancer patients regarding their risk of developing cancer-associated thrombosis.