Abstract

Do the safety and efficacy outcomes reported in the clinical trials of direct oral anticoagulants (DOAC) translate to the 'real-world'?

Background: A number of DOACs are now available for clinicians to prescribe in clinical practice. Whilst the results from large clinical trials demonstrate that these agents are as effective as vitamin K antagonists, there is some concern that the patients studied in the trials were not representative of patients, clinicians encounter in everyday practice. The aim of our study was to compare the real-world clinic population commenced on a DOAC to that from the clinical trials for these agents, in order to assess potential differences in safety and efficacy.

Patients and methods: A retrospective observational cohort study was undertaken. Patients who were initiated on a DOAC (apixaban, dabigatran and rivaroxaban) at a large teaching hospital in South East London between 1st August 2012 and 31st July 2014 were identified through pharmacy issue data with those followed-up for a minimum of 6 months included. Baseline demographic data, rates of stroke/VTE and rates of major/non-major clinically relevant (NMCR) (ISTH definition) bleeding were assessed and compared to pooled data reported from the corresponding Phase III trials. Differences between groups were compared using t-tests or chi-squared tests.

Results: During the review period, 748 patients were initiated on a DOAC, 365 for atrial fibrillation (AF) and 383 for venous thromboembolism (VTE). In terms of demographic differences, the real-world AF population comprised more females, were significantly older, had poorer renal function and a lower body weight. In contrast, the real-world VTE population typically had a higher body weight and poorer renal function, compared to the trial population, (table 1). Efficacy of DOACs was found to be similar across both the VTE and AF populations. With respect to safety, the real-world AF population experienced similar rates of major bleeding and a significantly lower rate of NMCR bleeding compared to the trial populations. In contrast, the real-world VTE population experienced a significantly higher rate of major bleeding, particularly gastrointestinal bleeding. Although the rate of NMCR bleeding was similar, there was a significantly higher rate of urogenital bleeding in the real-world VTE population, specifically heavy menstrual bleeding in women.

Conclusions: The efficacy outcomes of DOAC use in a real-world AF and VTE population are consistent with the Phase III trials, despite some significant differences in baseline characteristics. However, a significantly increased rate of major bleeding was observed in the real-world VTE population, which requires further investigation.

Table 1.

Baseline demographic characteristics, efficacy and safety outcomes in the real-world population versus the trial population

 Atrial Fibrillation Venous Thromboembolism 
 Trial population+
N=28,342 
Real-world population Trial population++ Real-world population 
 N=365 N=8,716 N=383 
Baseline Demographics, mean (SD) unless otherwise specified         
Age, years 72 (9.6) 76.8 * (12.1) 56.9 (14.2) 55.6 (18.7) 
Female (%) 10451 (36.9) 215 * (58.9) 3753 (43.1) 184 (48.0) 
Weight, kg 82.7 (19.5) 77.3 * (22.6) 84.9 (19.6) 88.2 * (23.0) 
Creatinine clearance, mL/min 69 (26.7) 58.1 * (26.9) 105.8 (40.7) 91.1 * (37.6) 
Concomitant aspirin therapy 10341 (36.5) 49 * (13.4) (0) 
Previous VKA use (%) 15711 (55.4) 193 (52.9) 85 (22.2) 
         
Efficacy (%)         
All-cause mortality 1695 (6.0) 37 * (9.1) 160 (1.8) 10 (2.5) 
Stroke 676 (2.4) (2.0) (0.3) 
VTE 39 (0.1) (0.2) 192 (2.2) (1.8) 
         
Safety (%)         
Major Bleeding 1419 (5.0) 17 (4.2) 79 (0.9) 15 * (3.8) 
Intracranial 170 (0.6) (0.2) (0.1) 2 * (0.5) 
Gastrointestinal 644 (2.3) (2.0) (0.1) 8 * (2.0) 
Non-major Clinically relevant (NMCR) bleeding 4824 (17.0) 30 * (7.4) 540 (6.2) 26 (6.6) 
Gastrointestinal (2.2) 53 (4.2) 10 (2.5) 
Urogenital 296 (4.2) 16 (3.9) 100 (2.5) 38 * (9.6) 
 Atrial Fibrillation Venous Thromboembolism 
 Trial population+
N=28,342 
Real-world population Trial population++ Real-world population 
 N=365 N=8,716 N=383 
Baseline Demographics, mean (SD) unless otherwise specified         
Age, years 72 (9.6) 76.8 * (12.1) 56.9 (14.2) 55.6 (18.7) 
Female (%) 10451 (36.9) 215 * (58.9) 3753 (43.1) 184 (48.0) 
Weight, kg 82.7 (19.5) 77.3 * (22.6) 84.9 (19.6) 88.2 * (23.0) 
Creatinine clearance, mL/min 69 (26.7) 58.1 * (26.9) 105.8 (40.7) 91.1 * (37.6) 
Concomitant aspirin therapy 10341 (36.5) 49 * (13.4) (0) 
Previous VKA use (%) 15711 (55.4) 193 (52.9) 85 (22.2) 
         
Efficacy (%)         
All-cause mortality 1695 (6.0) 37 * (9.1) 160 (1.8) 10 (2.5) 
Stroke 676 (2.4) (2.0) (0.3) 
VTE 39 (0.1) (0.2) 192 (2.2) (1.8) 
         
Safety (%)         
Major Bleeding 1419 (5.0) 17 (4.2) 79 (0.9) 15 * (3.8) 
Intracranial 170 (0.6) (0.2) (0.1) 2 * (0.5) 
Gastrointestinal 644 (2.3) (2.0) (0.1) 8 * (2.0) 
Non-major Clinically relevant (NMCR) bleeding 4824 (17.0) 30 * (7.4) 540 (6.2) 26 (6.6) 
Gastrointestinal (2.2) 53 (4.2) 10 (2.5) 
Urogenital 296 (4.2) 16 (3.9) 100 (2.5) 38 * (9.6) 

+Pooled data from ARISTOTLE, RE-LY and ROCKET-AF trials

++Pooled data from AMPLIFY, RE-COVER and EINSTEIN-PE/DVT trials

*p<0.05

Disclosures

Patel:Bayer plc: Research Funding. Auyeung:Bayer PLC: Research Funding. Arya:Bayer plc: Research Funding.

Author notes

*

Asterisk with author names denotes non-ASH members.