REVIEW
SA JOURNAL OF DIABETES & VASCULAR DISEASE
58
VOLUME 7 NUMBER 2 • JUNE 2010
Treatment of underlying conditions
It is important to establish there is no underlying thyroid disease,
as AF responds poorly to treatment until euthyroid status has
been re-established. Uncontrolled hypertension should be treated
(if severe,
before
institution of antithrombotic therapy, given the
risk of haemorrhagic stroke),
15
and the choice of agents may be
influenced by the need for some AF rate-controlling properties.
Patients in heart failure and AF, with progressive symptoms, should
be considered early for specialist referral.
Assessment and optimisation of thromboembolic risk
AF is a major cause of thromboembolic stroke, conferring a
five-fold increase in annual risk. Assessment and reduction of
thromboembolic risk should be the priority in every patient with
newly diagnosed AF. Stasis in the non-contracting atria encourages
thrombus formation, most commonly in the left atrial appendage.
12
Hypercoagulability and endothelial dysfunction also play a role.
2
Warfarin, adjusted to a target INR of 2.5 (range 2–3), is known
to provide the greatest risk-reduction for thromboembolic stroke.
Aspirin (75–325mg once daily) also reduces stroke risk, but has been
shown to be less effective than warfarin in most trials, including the
elderly
13
and even when given in combination with clopidogrel.
14
Who needs warfarin?
All patients with
mitral stenosis or rheumatic mitral valve disease
and AF should be anticoagulated unless there is a very significant
contraindication because there is an extremely high risk of
thromboembolism.
2
Even those in sinus rhythm have a significant
risk of thromboembolism, particularly in the presence of significant
atrial enlargement (
>
50 mm) or (in mitral regurgitation) congestive
cardiac failure. Anticoagulation is not recommended purely for
aortic or tricuspid disease with AF, unless concomitant with
congestive cardiac failure or other risk factors.
16
For
non-valvular
AF (the majority), risk stratification scoring
systems can help to decide on which patients to anticoagulate.
NICE has published a stroke risk-stratification algorithm as part of
its AF management guideline. Internationally, it is common practice
to use the CHADS2 score: congestive cardiac failure, hypertension,
age
>
75, and diabetes mellitus all score 1 point, stroke/TIA 2 points
(see Table 1). Warfarin is recommended for CHADS2
≥
2, warfarin
or aspirin at 1, and aspirin (unless contraindications) for a score of
0.
1,2,17,18
At low CHADS2 score (0–1), bleeding risk may outweigh
the benefits of anticoagulation, and this remains an area of
continued debate. Assessment of bleeding risk is complex,
19
as is
the quantitative comparison of the clinical impact of intracranial
haemorrhage or other major bleeds versus that of ischaemic stroke
or thromboembolism.
There is evidence from the ACTIVE-W study that, even at a
CHADS2 score of 1, warfarin confers a modest but significant
advantage over antiplatelet therapy, with a low bleeding risk.
20
The
recently published ATRIA study of 13 559 patients found that the
net clinical benefit of warfarin was negligible in CHADS2 score 0
and 1, increasing significantly, as expected, at higher scores. For
example, the risk (event rate) of thromboembolism with CHADS2
score 4–6 was 6.34 off warfarin compared to 3.25 on warfarin
per 100 patient years, while the risk of intracranial haemorrhage
was 0.51 versus 1.08 per 100 patient years, respectively. Using
an impact factor of 1.5 for intracranial haemorrhage, reflecting
its greater average morbid impact than ischaemic stroke, the net
clinical benefit was 0.97 (CHADS2
=
2), 2.07 (3) and 2.2 (4–6) events
per 100 patient years. After prior history of stroke (net benefit
=
2.48), the group with next-greatest net benefit of warfarin was
those aged over 85 (
=
2.34).
21
To add to the complexity, a recent study in patients with
implanted pacemakers has shown that a combination of CHADS2
score with burden (frequency/duration) of AF episodes may
enhance risk stratification and optimise choice of antithrombotic
therapy. Although the CHADS2 score remains the basis for choice
of antithrombotic therapy, a high AF burden may need to be
an additional marker or risk that would push one towards full
anticoagulation with warfarin even if the CHADS2 score is relatively
low (0–1).
Conversely, patients with brief AF paroxysms may be at lower
risk than that purely predicted by the CHADS2 score alone.
22
This issue requires further clarification from larger studies but,
given that most patients do not have such constant heart rhythm
monitoring, and that many episodes of AF are asymptomatic,
suitable thromboprophylaxis should currently be offered to all
patients regardless of whether their AF is paroxysmal or persistent,
predominantly based upon the CHADS2 score (or equivalent).
Anticoagulation in the elderly
The elderly have a higher stroke risk and potentially the most to
gain from anticoagulation.
23
Although bleeding risk also increases
with age, the true risk of significant haemorrhage – particularly
in patients deemed at risk from falls – may be overestimated by
physicians, leading to underprescribing of warfarin, and the
alternative use of aspirin.
24,25
One recent study showed a high rate of major haemorrhage
in the over-80s, particularly in the first three months of warfarin
therapy.
26
However, the risk of bleeding was greatest in those with
the highest stroke risk (CHADS2 score
≥
3). More recently, a study
of 783 elderly patients (median age 75 years, with 180 patients
over 80 years at study onset) showed a lower bleeding risk (major
bleed rate 1.4 per 100 patient years), closer to earlier data, with the
notable differences from the previous study being low concomitant
use of antiplatelets, and a computer-assisted warfarin dosing
regimen.
27
Importantly, warfarin is far superior to aspirin in over-
75s for the prevention of stroke and with no significant difference
in bleeding risk,
13
and the net clinical benefit of warfarin in elderly
patients is also supported by the ATRIA study.
21
The decision to anticoagulate elderly patients remains a balanced
one, but should take into account the ability to achieve a stable
INR, control of hypertension, and (quality-adjusted) life expectancy
more than a perceived bleeding risk from falls or age per se.
Table 1.
CHADS2 stroke risk score for patients with non-valvular atrial
fibrillation
18
Risk factor
Score
Congestive heart failure
1
Hypertension
1
Age
≥
75 years
1
Diabetes mellitus
1
Stroke or TIA
2
