The introduction of direct oral anticoagulants (DOAC) has revolutionised the treatment of atrial fibrillation (AF) and venous thromboembolism (VTE). Previous clinical trials have demonstrated that DOACs are non-inferior to warfarin for treatment and prevention of VTE and at least as safe in terms of bleeding risk. However, there is limited data on real-world outcomes. We aim to evaluate the local real-world DOAC use with particular focus on safety and to specifically compare the safety outcomes of the warfarin era vs the DOAC era in the management of VTE.


Retrospective evaluation was performed with a total of 1696 patients reviewed including demographics and safety outcomes. Patients commenced or continuing on a DOAC for the treatment of AF and VTE at the Northern Hospital, a tertiary hospital in Victoria, Australia (September 2013 to September 2016, n=1079) were identified.

A sub analysis was then performed comparing patients on DOAC for VTE indications to a separate retrospective local audit of VTE patients during the warfarin era (July 2011 to December 2012, n=617). For the purpose of the sub analysis, patients with active malignancy and thromboses other than deep vein thrombosis (DVT) and pulmonary embolus (PE) were excluded to maintain consistency with the warfarin era data.


A total of 1079 patients on DOAC with median age 70 years (range 17-96) were identified. The indications for DOAC were AF 61.4% (n=663), VTE treatment 30.5% (n=329) and VTE maintenance/prophylaxis 8.1% (n=87). The most commonly prescribed DOAC was rivaroxaban 60.8% (n=656) followed by apixaban 28.7% (n=310) and dabigatran 10.5% (n=113). Of these patients 29.7% (n=320) were on low dose anticoagulation. Forty episodes of clinically significant bleeding (ISTH-SCC score 3-4) (3.7%) were captured [Table 1]. The average HASBLED score of these patients was 2 and significant risk factors for bleeding included low dose anticoagulation, AF patients, concurrent antiplatelet use, prior bleeding and high falls risk.

In terms of thrombotic complications, there were 12 episodes of thrombotic stroke (1.8%) despite DOAC use with risk factors including low dose anticoagulation (p=0.03), prior stroke (p=0.04) and high falls risk (p=0.03). Nine patients on DOACs for VTE (2.2%) reported recurrence including a patient with active malignancy.

374 patients on DOAC for VTE management were included in the sub-analysis comparison to the warfarin audit (n=617). Of the DOAC patients, 88.2% (n=330) were on rivaroxaban and 11.8% (n=44) were on apixaban - 31.8% (n=315) were on DOAC for acute VTE treatment and 6.0% (n=59) for maintenance or prophylaxis at time of data collection. There were more unprovoked VTE in the DOAC group (66.0% (n=208) vs 42.0% (n=259), p<0.001). The patient demographics are outlined in Table 2. Overall, there were 28 clinically significant bleeding (ISTH-SCC score 3-4) events in the warfarin group (4.5%) and 8 in the DOAC group (2.1%; p=0.05). The sites of bleeding were similar in both groups with gastrointestinal tract bleeding being the most common source. Forty-eight recurrent VTE events were captured - 39 occurred on warfarin (6.3%) and 9 while on DOAC (2.4%; p=0.005).


This retrospective audit shows that our local safety outcomes are comparable to clinical trials. Low dose anticoagulation and high falls risk (a surrogate marker of frailty) were significant risk factors for both clinically significant bleeding and thrombotic stroke in the DOAC population. These patients are likely frailer with greater co-morbidities and have shared risk factors for bleeding and stroke, suggesting that for these high risk patients, low dose anticoagulation does not negate their risk of complications and careful prescribing and close monitoring remain essential.

The sub-comparison between VTE patients on DOAC and warfarin found that patients on warfarin had a higher rate of VTE recurrence in comparison to patients on DOAC. This may reflect patients maintaining more time on therapeutic anticoagulation with the more predictable DOACs compared to warfarin patients who may have more time below therapeutic range. There may also be a selection bias as-high risk patients with co-morbidities and those not meeting criteria for DOACs continue to be prescribed warfarin in the DOAC era. This data is similar to recent real-life DOAC experience in the atrial fibrillation population.


No relevant conflicts of interest to declare.

Author notes


Asterisk with author names denotes non-ASH members.