Clinical Predictors of Confirmed Venous Thromboembolism in Patients with a Suspected Recurrent Venous Thromboembolism: A Retrospective Study of the Reverse I Cohort

Many clinical decision rules (CDR), such as the Wells models for deep vein thrombosis of the legs (DVT) and pulmonary embolism (PE), have been established for patients with suspected first venous thromboembolism (VTE). However these rules may have limitations for patients with suspected recurrent VTE. First, patients with suspected PE and a history of VTE have their diagnosis more likely confirmed than patients with a suspected PE without a history of VTE (40.3 vs. 20.6%) (Le Gal G et al. Arch Intern Med, 2006), suggesting a different pre-test probability. Furthermore, patients who had a previous pulmonary embolism often have complaints of chronic dyspnea (Klok FA et al. Eur J Intern Med, 2008), which could complicate the risk estimation of recurrent PE. A similar scenario arises for patients who suffered a DVT, as 30–50% of patients will show signs and symptoms of post-thrombotic syndrome (PTS) following a deep vein thrombosis (Kahn SR. Curr Opin Pulm Med, 2006). In addition, the value of D-dimer testing is still uncertain in patients with suspected recurrent VTE. Finally, imaging has limitations, as residual thrombosis may be present and may be misdiagnosed as an acute recurrent event. Hence, because pre-test probability, laboratory testing and diagnostic imaging performances differ in patients with a suspected recurrent VTE, a different approach may be needed for effective management of these patients.

Our objective was to study clinical predictors for the diagnosis of recurrent VTE in patients with a history of VTE.

The REVERSE I study enrolled patients with a first unprovoked, objectively proven VTE (Rodger M et al. CMAJ, 2008). Each patient in the REVERSE I cohort that had a suspected recurrent VTE during follow-up was screened for eligibility. Only first adjudicated suspected recurrent events were studied. All these events were blindly adjudicated by an independent committee. Potential clinical predictors of recurrent VTE consisted of clinical predictors collected at the baseline visit (5–7 months after the unprovoked event), and of information collected in physicians' clinical notes, laboratory or imaging results at the time of the suspected recurrent VTE.

The predictive value of each predictor was determined by the Chi-square test for nominal data and the unpaired 2-tailed T-test for continuous data.

In the REVERSE I cohort, out of the 646 patients who were followed, 402 patients had a suspected recurrent VTE within a mean of 20.2 months (range: 0 – 97 months) of follow-up. After screening, 376 patients were enrolled in our study: 52.7 % of patients were males, and the mean age was 53.1 years (± 17.5). Among all suspected recurrent VTE events, male gender and a positive d-dimer result at the time of suspected recurrent VTE (p < 0.01), as well as symptoms occurring for 10 days or less at the time of presentation (p < 0.05) were in favor of a recurrent VTE diagnosis. In addition, mean age was higher in patients with a confirmed recurrent VTE (p < 0.05).

In patients with suspected DVT, the presence of leg swelling was in favor of a confirmed diagnosis (p < 0.01), while leg pain was not associated with a higher rate of confirmed DVT. A suspected DVT in the same leg as the previous event seemed less likely to have a confirmed diagnosis of recurrent DVT. However this trend was not statistically significant.

In patients with suspected PE, mean oxygen saturation was lower among patients with a new PE diagnosis (p < 0.05). Chest pain and shortness of breath were not important predictors.

Finally, in the case of a suspected DVT and PE event, shortness of breath was a significant predictor (p < 0.05) for the diagnosis of recurrent VTE.

This is the first study to show that predictors for the diagnosis of a recurrent VTE may be different than predictors for the diagnosis of a first VTE. For instance, our results show that male gender is not only a risk factor for recurrent VTE (McRae S et al. Lancet, 2006), but is also an important predictor for confirmed VTE among patients with suspected recurrent VTE, while none of the existing CDRs take this in consideration. A CDR specific to patients with a history of VTE may be needed for better management of these patients.

Le Gal:Bayer, bioMérieux, GSK, Leo Pharma, Sanofi Aventis: Honoraria.