Assessing the Updated Vienna Prediction Model for Unprovoked VTE: Does It Hold Up?

The Vienna Prediction Model was established to identify patients with unprovoked venous thromboembolism (VTE) who have a low risk of recurrence, with the aim of avoiding unnecessary anticoagulant therapy. In validation tests, though, the model showed a “less than optimal calibration, underestimating the observed cumulative recurrence rates at 12 months,” according to Tobias Tritschler, MD, from the Department of General Internal Medicine at Bern University Hospital in Switzerland.

“Given that one-third of the patients with unprovoked VTE have a low recurrence risk and may stop anticoagulation therapy after three months, a risk stratification tool to identify these patients is important,” Dr. Tritschler told ASH Clinical News.

Recently, the original developers of the model published an updated version to address these concerns. Unlike the original version, the updated model allows for prediction of recurrence at several different time points after stopping anticoagulation (3 weeks and 3, 9, 15, and 24 months). However, in a study recently published in Blood, Dr. Tritschler and co-authors discovered that the updated model still leaves something to be desired.

The original Vienna Prediction Model was devised through a prospective cohort study of 929 patients with a first unprovoked VTE, and assigned risk based on the following three parameters:

  • Sex
  • Location of VTE (pulmonary embolism [PE]/proximal deep vein thrombosis [DVT] vs. distal DVT)
  • Quantitative D-dimer level determined by ELISA between three weeks and 15 months after discontinuation of anticoagulation

The updated version was developed in a prospective cohort of 553 patients with unprovoked VTE and uses the same clinical parameters as the original model to estimate recurrent VTE risk at up to 60 months of follow-up using nomograms or a Web-based calculator.

The current multicenter study attempts to validate the updated model included 156 patients ≥65 years old with acute symptomatic unprovoked VTE from nine Swiss hospitals. Patients were excluded from the study if they were immobile, had major surgery, were taking oral estrogen therapy, or had active cancer during the previous three months. All patients had completed a three- to 12-month course of anticoagulation.

Study participants had a median age of 74 years, 41 percent were female and 88 percent had PE or proximal DVT as the initial VTE event (TABLE).

To validate the updated model, Dr. Tritschler and colleagues determined the proportion of patients classified as low-risk according to the updated Vienna Prediction Model and compared the proportion of VTE recurrence between low- and higher-risk patients at 12 and 24 months.

In the updated model, they noted, patients with a predicted 12-month risk <6.2 percent are classified as “low risk.”

Overall, the proportion of VTE recurrence was 11 percent (17 of 156 patients) after 12 months and 17 percent (26 of 156) after 24 months. At 12 months:

  • 5% of men had recurrent VTE compared with 19% of women (p=0.02)
  • 11% of patients with both PE/proximal DVT and distal DVT had recurrent VTE (p=1.0)
  • 13% of patients with a D-dimer ≥1022 ng/ml had recurrent VTE, compared with 9% of those with a D-dimer <1022 ng/mL (p=0.61)

At 24 months:

  • 12% of men and 23% of women had recurrent VTE (p=0.08)
  • 17% of patients with PE/proximal DVT and 16% of patients with distal DVT had recurrent VTE (p=1.0)
  • 18% of patients with a D-dimer ≥1022 ng/mL had recurrent VTE, compared with 15% of patients with a D-dimer <1022 ng/mL (p=0.83)

Ultimately, the updated Vienna Prediction Model failed to correctly identify patients as low- or high-risk for recurrent VTE. “The proportion of recurrent VTE did not differ between low- versus higher-risk patients at 12 months (13% vs. 10%, respectively; p=0.77) and 24 months (15% vs. 17%, respectively; p=1.0),” the authors reported.

Limitations of the study include the relatively small sample size, and that the study may have been underpowered to detect a difference in VTE recurrence. The study also focused on older patients with VTE, whereas the updated Vienna Prediction Model was developed in patients with a broader age distribution. Ultimately, it is still too soon to make any treatment recommendations based on this risk prediction tool, the authors added.

“The updated Vienna Prediction Model does not allow the identification of elderly patients at low risk of VTE recurrence who may stop anticoagulation therapy after three months,” Dr. Tritschler said. “A risk stratification tool that accurately identifies elderly patients with unprovoked VTE who are at low risk of VTE recurrence is still needed.”


Tritschler T, Mean M, Limacher A, et al. Predicting recurrence after unprovoked venous thromboembolism: prospective validation of the updated Vienna Prediction Model. Blood. 2015 September 4. [Epub ahead of print]

TABLE. Patient Baseline Characteristics
All (N=156) Low-risk patients (n=39) Higher-risk patients (n=117) p Value
n (%) or median (interquartile range)
Age (years) 74.0 (69.0; 79.8) 73.0 (68.0; 77.0) 75.0 (69.0; 80.0) 0.27
Male sex 92 (59) 4 (10) 88 (75) <0.01
PE and/or proximal DVT 137 (88) 28 (72) 109 (93) <0.01
D-dimer (ng/ml) 1022 (607; 1755) 717 (410; 1016) 1,161 (694; 1,913) <0.01
BMI (kg/m2) 27.5 (24.8; 30.5) 27.7 (24.9; 31.9) 27.5 (24.8; 30.4) 0.44
Prior VTE 23 (15) 8 (21) 15 (13) 0.30
Time since stopping anticoagulation (months) 5.7 (4.9; 6.9) 5.9 (5.0; 7.6) 5.6 (4.7; 6.6) 0.20
Duration of prior anticoagulation (months) 6.3 (5.3; 7.1) 6.0 (4.1; 6.9) 6.4 (5.5; 7.2) 0.20