76 / 210
XXXII Congreso Nacional de la Sociedad Española de Trombosis y Hemostasia
76
studies is much broader and, therefore, it could be anticipated that
higher major bleeding rates would be observed in these studies.
As seen in the phase III clinical trials, overall rates of ICH with
the NOACs were also low
6,14
in real-world studies and, if direct
comparator treatments were also assessed, lower than
9,11,15,16
or
similar to
3,17
those seen with warfarin. Regarding GI bleeding, the
real-world evidence with rivaroxaban seems not to reflect the cli-
nical trial results. In ROCKETAF and the meta-analysis described
above, GI bleeding rates were significantly higher in patients recei-
ving rivaroxaban versus warfarin therapy
2
. However, in real-world
studies, this signal was not reproducible and GI bleeding rates
with rivaroxaban were consistently similar to rates seen with war-
farin
3,18,19
. For dabigatran, a similar increase in GI bleeding rates
compared with warfarin was demonstrated in the RE-LY trial, but
real-world findings are inconsistent. Several studies have shown
similar or even lower GI bleeding rates for dabigatran compared
with warfarin
9,18-21
whereas other studies confirmed higher rates of
GI bleeding with dabigatran
11,16,20,22
.
Another important finding is the low case-fatality rate after
NOAC related major bleeding. For instance, in the Dresden NOAC
Registry, the case-fatality rate for a major bleeding event during
rivaroxaban therapy was 6.3%
4
, which compares favourably with
the approximately 15% case-fatality rate reported previously for
patients 90 days after a VKA-related major bleeding.
Different from case-fatality rates of major bleeding are data on
the incidence of fatal bleeding. A large pharmacovigilance study
demonstrated low rates of fatal bleeding for rivaroxaban of 0.08
per 100 person years
5
, and a Danish registry study demonstrated
fatal bleeding rates of 0.13-0.27 for dabigatran 150 mg, compared
with 0.25-0.46 for warfarin
25
. Although real-world data on fatal
bleeding rates and post-major bleeding mortality rates are limited,
the data that are available are consistent with findings from RCTs.
A meta-analysis by Caldeira et al., of data from ARISTOTLE,
ENGAGE-AF, ROCKET AF, J-ROCKET AF and RE-LY, found
that NOACs decreased the risk of fatal bleeding by 47% and redu-
ced the case-fatality rate by 32 % compared with VKAs
26
.
Effectiveness
In addition to safety information, real-world studies can also
provide supportive information on the effectiveness of NOACs
to prevent stroke. In phase III clinical trials, the NOACs were
shown to be at least as effective as warfarin for the prevention of
stroke and systemic embolism in patients with non-valvular AF;
findings confirmed in a meta-analysis of the four pivotal trials
2
.
This meta-analysis also demonstrated that the NOACs were asso-
ciated with a significant reduction in all-cause-mortality compared
with warfarin
2
.
Several rea-world studies have evaluated the effectiveness of
the NOACs and, as with the bleeding data, it is reassuring to note
that the findings of the real-world studies reflect the outcomes
observed in the clinical trials. Of note, retrospective healthcare
database analyses tend to result in higher event rates compared
with prospectively collected data from registries. The event rates
also vary widely between the studies because of differences in
patient populations and study design; in particular differences in
the effectiveness endpoints used.
If event rates of the equivalent composite outcome of stroke
and systemic embolism are compared between the real-world and
phase III NOAC studies, they are lower in the real-world studies
(0.8
vs.
1.1-1.7 per 100 patient-years, respectively)
6,12
. An excep-
tion is the Laliberté et al. study, in which event rates with riva-
roxaban were 4.6 per 100 patient-years and as high as 5.9 per
100 patient-years with VKA
3
. In the phase III studies, te highest
event rates were associated either with lower dose use (RE-LY
dabigatran 110 mg dose; 1.5 events per 100 patient-years) or a
higher-risk patient population (ROCKETAF mean CHADS
2
= 3.5;
1.7 events per 100 patient-years). The mean CHADS
2
score in the
Laliberté et al. study was 2.0; therefore, this cannot account for
the higher event rates observed; these differences are more likely
to be because of how the events were identified,
i.e.
a retrospective
search strategy using ICD-9-CM codes
3
.
Independent of the applied design, the available real-word
effectiveness data consistently demonstrate that NOACs appear
to be at least as effective as VKAs in reducing the risk of cardiovas-
cular outcomes including stroke, systemic embolism and transient
ischaemic attack in patients with AF.
Furthermore, there is some real-world evidence to support a
reduction in mortality risk with NOAC use as demonstrated in the
Ruff
et al.
meta-analysis
2
; three real-world studies have demonstrated
significant reductions in mortality with NOACs
versus
warfarin
9,11,27
.
Conclusions
The value of real-world evidence to provide important informa-
tion about a drug beyond that which can be established from RCTs is
becoming increasingly recognized. What is less well understood are
the differences between the varying sources of real-world data and how
these can impact on study results and their interpretation. However,
with these caveats in mind, when evaluating the currently available
real-world evidence relating to the NOACs for stroke prevention in
patients withAF, what is most striking is the consistency of the effec-
tiveness and safety findings. Irrespective of the data source, in clinical
practice the NOACs have been shown to be at least as effective as
VKAs for the prevention of stroke and other cardiovascular outcomes,
with similar or lower rates of major bleeding. Even more important-
ly, so far there is no evidence to suggest that major bleeding is more
problematic or associated with inferior outcomes in NOAC recipients;
indeed the first real-world evidence data have indicated the opposite.
Overall, the available real-world evidence provides confirmation that
the benefits of the NOACs for stroke prevention in patients withAF, as
demonstrated in phase III RCTs, have translated into clinical practice.
In conjunctionwith the additional advantages of simplified dosing, low
potential for food and drug interactions and no requirement for routine
coagulationmonitoring, the NOACs are an attractive therapeutic option
for patients with non-valvular AF who are at risk of stroke. It will be
interesting to see if future real-world data on the use of apixaban and
edoxabanwill add to the emerging evidence base demonstrating strong
effectiveness and safety profiles of the NOACs in clinical practice.
References
1. Garrison LP Jr, Neumann PJ, Erickson P, Marshall D, Mullins
CD. Using real-world data for coverage and payment decisions: