How Long Is Long Enough? Extending OAC After Unprovoked PE

Written By

Glenn D.O. Pebanco, Pharm.D., BCPS
Jaclyn Hernandez, Pharm.D.

Reviewed By

Henry I. Bussey, Pharm.D.
Mary Amato, Pharm.D., MPH, BCPS
Adrian Wong, Pharm.D., BCPS


Couturaud F, Sanchez O, Pernod G, et al. Six months vs extended oral anticoagulation after a first episode of pulmonary embolism – the PADIS-PE randomized clinical trial. JAMA 2015; 314: 31-40.

The recommended treatment duration for a first episode of unprovoked venous thromboembolism (VTE) is, at a minimum, 3 months with extended anticoagulation favored for those who are not at high risk for bleeding.  However, the optimal duration of anticoagulation therapy remains unknown.1  Earlier studies have shown that an extended course of oral anticoagulation (OAC) with warfarin reduced VTE recurrence but at the expense of increased bleeding.2,3  The Prolonged Anticoagulation Treatment for a First Episode of Idiopathic Pulmonary Embolism (PADIS-PE) study revisits this question in patients with an initial, unprovoked pulmonary embolism (PE) but, most importantly, provides insights about patient outcomes after anticoagulation treatment is discontinued.4


The PADIS-PE study was a multicenter, randomized, double-blinded, parallel-group trial conducted in France and enrolled 371 adult patients who experienced a first episode, unprovoked, symptomatic PE.  All participants were treated with an oral vitamin K antagonist (VKA) to a goal INR between 2 and 3 for 6 months before being randomized to either continued VKA or placebo for an additional 18 months. After completing the assigned treatment, subjects were followed up for an additional 24 months (post-treatment period).  The mean percentage of time in therapeutic range (TTR) was nearly 70% prior to randomization. Major exclusion criteria were recurrent VTE or bleeding while on treatment during the first 6 months of anticoagulation; known major thrombophilia; platelet count < 100 x 103/uL; major surgery planned within 18 months; and life expectancy less than 18 months. The primary outcome was a composite of symptomatic recurrent VTE and major bleeding. Symptomatic VTE was diagnosed if the clinical suspicion was objectively confirmed, whereas bleeding was deemed major if it was fatal, involved a critical organ, or was overt and associated with a decrease in hemoglobin of 2g/dL or more, or required 2 or more units of packed red blood cells. A lung ventilation-perfusion scan was performed and D-dimer concentration was measured prior to randomization. Baseline characteristics were similar in both groups in which the ‘average’ patient was approximately 58 years old with a body mass index of 27.  There was a nearly equal balance of male and female participants.  The mean D-dimer was elevated (greater than 250 ng/mL) in both treatment groups and more than 30% of patients had evidence of a residual perfusion defect at baseline.  Approximately half of the patients were at ‘low or moderate’ risk for bleeding and the other half were deemed to be at high risk for bleeding based on the presence of 2 or more risk factors.1,4


During the 18-month treatment period, the primary composite outcome (VTE or major bleeding) occurred in 6 patients (3.3%) in the VKA group and 25 patients (13.5%) in the placebo group (HR 0.22; 95% CI 0.09-0.55; p <0.001).  The TTR was 70% in the VKA treatment group and the mean interval between INR measurements was approximately 20 days. Patients assigned to the placebo group had sham INR measurements performed approximately every 23 days and were reported to be in target INR range 75% of the time.  It should be noted that the 3 recurrent VTE events in the VKA-treated group occurred after the treatment had been stopped and all 3 were unprovoked. In the placebo-treated group, 23 VTE events were unprovoked and only 1 occurred after placebo treatment had been stopped. These results equate to an absolute risk reduction of 9.5 VTE events per 100 patients years with a number-needed-to-treat of 11. Major bleeding was higher in the warfarin-treated group (2.2% vs. 0.5%, p = 0.22; NNH = 59).


These results are consistent with previous trials including the data from extended therapy with direct oral anticoagulants (DOACs).2-3,5-7  Extended use of OAC prevents recurrent VTEs with or without increased risk of bleeding, regardless of how long the treatment is extended (See Table 1)


Table 1:  Extended Treatment for Idiopathic VTE – Continued Treatment vs. Placebo


(OAC used)

Mean duration of extended
treatment (months)

Recurrence Rate for VTE
HR  (95% CI)

Number Needed to Treat

Major Bleeding
(Active vs. Placebo)

Kearon C et al 2a* (Warfarin)


0.05 †
(0.01 – 0.37)


3.8% vs. 0%

WODIT-PE 3a* (Warfarin)


0.99 †
(0.45 – 2.16)



EINSTEIN-Cont 5 (Rivaroxaban)


0.18 †
(0.09 – 0.39)


0.7% vs. O%

AMPLIFY-EXT 6b (Apixaban 2.5mg BID and 5mg BID)


0.19 ¶
(0.11 – 0.33)

0.20 ¶
(0.11 – 0.34)


0.2% vs 0.1% vs. 0.5%

RE-SONATE 7 (Dabigatran)


0.08 †
(0.02 – 0.25)


0.3% vs. 0%

PADIS-PE 4*  (Warfarin)


0.15 †
(0.05 – 0.43)


2.2% vs. 0.5%

NA: Not applicable, NE: Not estimable, NR: not reported

a Trial stopped early

b Relative risk reported for recurrent VTE or VTE-related death

* Trials which included a warfarin/VKA arm

† Hazard ratio is for OAC as compared to placebo

¶ Relative risks are for apixaban as compared to placebo


During the 24-month post-treatment period, the beneficial effects of warfarin were essentially erased. Extending the length of VKA therapy by 18 additional months did not modify the long-term event rate — 25 patients (16.4%) in the warfarin-treated group had a recurrent VTE during the 24 month post-treatment follow-up period.  Two-thirds of these events were unprovoked and 4 were fatal PEs. During the post-treatment period, 14 patients (8.7%) in the placebo-treated group had a recurrent VTE and the majority were unprovoked.  Over the entire 48-month study, the composite event rate was 6.7 per 100 patients years in warfarin treated patients and 8.9 per 100 patients years in the placebo group (HR = 0.75, 955 CI, 0.47-1.18). Only one other trial, RE-SONATE, followed patients after OAC was discontinued.  Twelve months after treatment was stopped,  the cumulative incidence of recurrent or fatal VTE was 6.9% in the dabigatran group compared to 10.7% in the placebo group (HR 0.61; 95% CI 0.42 – 0.88).7  Perhaps if the RESONATE investigators had followed these patients for an additional 12 months, the difference might have evaporated.


Clearly the benefits of extended treatment with warfarin after an unprovoked PE outweigh the potential risks.  Strengths of the PADIS-PE include all the usual study design features we’d like to see - randomization, double blinding, and long-term follow-up. However, the relatively high TTR might actually limit the generalizability of the study results.  A recent systematic review and meta-analysis of 40 studies in VTE treated with vitamin K antagonist therapy, the mean TTR was only about 60% during the first 6 months of treatment.9


Tools to predict VTE recurrence remain elusive.  Some have advocated repeat ultrasonography, serial d-Dimer testing, patient age, gender, and hormone use to determine whether a patient should receive extended treatment following a DVT.10  Whether these or other parameters are useful in assessing the risk of recurrent PE requires further study. The decision to extend OAC therapy for unprovoked VTE depends not only on individual assessment of risk of recurrence and bleeding but also on the burden of treatment.1 The PADIS-PE trial provides additional evidence to support extended use of warfarin in patients with unprovoked PE.  But several questions remain unanswered.  Should we continue OAC therapy indefinitely?  Should we use DOACs preferentially for extended treatment given their improved safety profile and ease of use?  What would you do?