CARDIOVASCULAR JOURNAL OF AFRICA • Vol 21, No 6, November/December 2010
342
AFRICA
Rivaroxaban equals warfarin treatment in atrial fibrillation
patients at high risk of stroke
Rivaroxaban has met its primary objective
in the ROCKET trial of non-inferiority to
warfarin in the treatment of atrial fibrilla-
tion (AF) patients at high risk of stroke.
In fact the study included a majority of
patients who had already experienced
either a stroke or transient ischaemic
event, therefore needing warfarin for
secondary prevention.
For the primary-efficacy endpoint,
rivaroxaban was superior to warfarin,
delivering a 21% relative risk reduction
in stroke and non-central nervous system
(CNS) systemic embolism in the pre-
specified on-treatment population (1.70
vs 2.15%,
p
=
0.015).
Could these results of rivaroxaban
in the ROCKET trial have been fore-
seen by Prof John Camm, University
of London, in his commentary a year
ago on the RELY study?
1
He pointed
out then, ‘much more information will
be needed before regulators can decide
on the approvability of the drug [in this
case, he was referring to dabigatran] for
the management of patients with atrial
fibrillation with thrombo-embolic risk
(CHADS score greater than 2%)’.
We now know that the FDA has
approved the higher dose of dabigatran
(150 mg twice daily) to prevent stroke
in patients with atrial fibrillation, and
a 75-mg dose for patients with severe
renal impairment (15–30 ml/mm). The
FDA made no additional stipulations for
dabigatran usage in AF except to say, ‘for
stroke prevention’; while the Canadian
approval says, ‘for AF patients in Canada
in whom anti-coagulation is appropriate’.
The
randomised,
double-blind
ROCKET study could make an argument
for the usage of rivaroxaban in patients
with AF and a high CHADS score. In
ROCKET, the majority of patients were
in the 3–5% CHADS score risk assess-
ment, with a mean of 3.48% (Table 1).
The CHADS score in the RELY study
was 2.1%. The primary efficacy outcome
for the ROCKET study is shown in Fig. 1
and Table 2.
The event rate for stroke and non-CNS
embolism was 1.7 per 100 patient years
for rivaroxaban and 2.15 for warfarin,
based on the on-treatment population.
The intention-to-treat (ITT) event rates
were higher; 2.12 with rivaroxaban and
2.42 for warfarin.
In ROCKET, the time in therapeutic
INR range was 57.8%. There were similar
rates of bleeding and adverse events in
the rivaroxaban and warfarin arms, but
less intra-cranial
haemorrhage and
fatal
bleeding
with
rivaroxa-
ban. Prof Kenneth
W Maheffey of
Duke University
concluded
that
rivaroxaban has now been shown to be a
proven alternative to warfarin for stroke
prevention in moderate- or high-risk
patients with non-valvular AF.
South Africa participated in the
ROCKET study and entered 247 patients
into the study.
The discussant of this study at the
American Heart Association, Dr Elaine
M Hylek, Boston, USA pointed out that
ROCKET had recruited the oldest and
highest-risk AF patient cohort, with
55% having had a stroke and therefore
requiring warfarin for secondary preven-
tion, 91% with hypertension, 62% with
chronic heart failure (which would have
contributed to INR variability), and 39%
with diabetes. ‘These were patients at the
highest risk for intra-cranial haemorrhage
due to age, high blood pressure and prior
stroke’, she pointed out.
In this high-risk population, less than
50% achieved the time-in-therapeutic-
range (TTR) threshold of
>
58–60% need-
ed to realise the benefits of warfarin. She
pointed out that per-protocol or ‘on-treat-
ment analysis’ is important to confirm
non-inferiority of a primary intention-to-
treat analysis. Then, after non-inferiority
is evident, superiority may be assessed,
preferably defined at the outset and with
an intention-to-treat analysis.
2
Dr Hylek’s
conclusions are summarised:
•
Based on both the ITT and per proto-
col (PP) analysis, rivaroxaban is non-
Days from randomisation
Fig. 1. Primary efficacy outcome: stroke and non-CNS
embolism.
No. at risk:
Rivaroxaban 6958 6211 5786 5468 4406 3407 2472 1496 634
Warfarin 7004 6327 5911 5542 4461 3478 2539 1538 655
Event Rates are per 100 patient-years
Based on Protocol Compliant Treatment Population
HR (95% CI): 0.79 (0.66–0.96)
p
-value non-inferiority:
<
0.001
Rivaroxaban
Warfarin
Rivaroxaban Warfarin
Event rate
1.71
2.16
6
5
4
3
2
1
0
0 120 240 360 480 600 720 840 960
Cumulative event rate (%)
TABLE 2. PRIMARY EFFICACY OUTCOME:
STROKEAND NON-CNS EMBOLISM
0
1
2
Rivaroxaban
better
Warfarin
better
On treatment
n
=
14 143
ITT
n
=
14 171
Rivaroxaban
event rate
Warfarin
event rate
HR (95% CI)
p
-value
1.70
2.15
0.79
(0.65–0.95)
0.015
2.12
2.42
0.88
(0.74–1.03)
0.117
Event rates are per 100 patient years
Based on safety on treatment or intention-to-treat through site-notifi-
cation populations
TABLE 1. BASELINE DEMOGRAPHICS
Rivaroxaban
(
n
=
7 081)
Warfarin
(
n
=
7 090)
CHADS
2
score (mean)
2 (%)
3 (%)
4 (%)
5 (%)
6 (%)
3.48
13
43
29
13
2
3.46
13
44
28
12
2
Prior VKA use (%)
62
63
Congestive heart failure
63
62
Hypertension (%)
90
91
Diabetes mellitus (%)
40
39
Prior stroke/TIA/
embolism (%)
55
55
Prior myocardial
infarction (%)
17
18
Based on intention-to-treat population
