1. health and fitness
2. disease
3. heart disease

European Heart Journal (1996) 17, 674-681
Evidence-based Cardiology
Should all patients with atrial fibrillation receive warfarin?
Evidence from randomized clinical trials
J. G. F. Cleland*, P. J. Cowburn* and R. H. Falkf
* Medical Research Council Clinical Research Initiative in Heart Failure, University of Glasgow, Glasgow,
Scotland, U.K.; ^Boston University Medical Centre, Massachusetts, U.S.A.
Introduction
Key Words: Atrial fibrillation, heart failure, aspirin,
warfarin.
Correspondence: Dr John G. F. Cleland. British Heart Foundation
Senior Fellow. Medical Research Council Clinical Research
Initiative in Heart Failure, West Medical Building. University of
Glasgow. Glasgow G12 8QQ. Scotland. U.K
0195-668X/96/050674 + 08 S18.00/0
Population statistics (Table 1)
Atrial fibrillation is a common condition affecting about
0-5% of the adult population, the prevalence rising
progressively with age to affect 5-12% of those over 74
years'1"12"15'. Population surveys with serial assessments
indicate that stroke may be preceded by chronic atrial
fibrillation in 1 -4%—14-5% of cases, although the total
number of patients with atrial fibrillation in these
surveys is comparatively small'1'16171. The surveys
underestimate the total number of strokes due to atrial
fibrillation, as patients presenting with stroke and new
onset atrial fibrillation will not be identified prospectively. However, only atrial fibrillation identified prior to
an embolic event allows an opportunity for prevention.
Comparing patients in sinus rhythm to those with atrial
fibrillation, the age-adjusted relative risk of stroke
ranges from 2-3 to 6-9 with a relative risk of about 4 in
the report with the largest number of events'1181. Some
retrospective studies'191 have suggested a much lower
risk of stroke in 'lone atrial fibrillation' (i.e. atrial
fibrillation in the absence of any cardiac abnormality). It
is clear from the randomized trials that this term applies
to only a minority of patients with atrial fibrillation.
Data on the absolute annual risk of stroke in
patients with atrial fibrillation from serial follow-up in
community studies suggest rates of stroke between 0-3
and 4 1 % per annum' 11617 '. The latter figure'171 is very
similar to the rates described in the control groups of the
randomized controlled trials. Thus, despite exclusion of
many patients, the randomized studies do seem to reflect
the problem in the community. The risk of stroke over
5 years for the general population with atrial fibrillation
may be as high as 25%. In addition, the prevalence
1996 The European Society of Cardiology
Downloaded from by guest on October 21, 2014
Stroke is usually devastating and often fatal. It consumes about 1% of all health care spending in industrialized countries. Up to 15% of strokes may be preceded
by chronic atrial fibrillation1'1. Seven randomized controlled trials of aspirin or anticoagulants for the prophylaxis of stroke and other thrombo-embolic events have
been published since 1989[2"91. Despite the evidence that
warfarin can reduce the risk of stroke, audit of the use of
anticoagulants even in hospital practice suggests that
only a third of patients without contraindications to
warfarin are so treated'10'. It is likely that prescription of
anticoagulants for patients with atrial fibrillation is even
lower in the community setting.
The reluctance of doctors to prescribe anticoagulants and for patients to take them stem partly from the
perceived dangers of such therapy but mainly from the
inconvenience and expense, both to the patient and
doctor, of repeated monitoring. However, among volunteers for randomized studies warfarin has not been
associated with an adverse impact on the quality of life
unless bleeding, minor or major, has occurred'11].
Three strategies, not mutually exclusive, may be
pursued to overcome these problems. Firstly, if the
excess risk of atrial fibrillation is small then it may be
best to treat only high risk groups. Most doctors would
be reluctant not to anticoagulate a patient with mitral
stenosis or a mechanical mitral prosthesis and atrial
fibrillation; many doctors would be reluctant to anticoagulate a young patient with an otherwise normal heart.
Secondly, lower, safer doses of anticoagulants, that may
require less frequent monitoring, could be used. Finally,
a less expensive more efficient infrastructure for
anticoagulation could be inaugurated.
Evidence-based cardiology
675
Table 1 Incidence of stroke in community studies in patients with and without non-rheumatic atrial fibrillation
(NRAF) (only long-term studies with repeated assessments to detect new onset atrial fibrillation have been chosen)
Study
n=
duration of
follow-up
Flegel (a) 1987
19018
~ 15 years
Whitehall Study1551
7727
Flegel (b) 1987
British Regional Heart Survey155'
~ 8 years
Wolf; Framingham 1978[1?1
5184
~24 years ~
Wolf: Framingham 19911'1
5070
~34 years!
Age range Total no. of XT
,
,.
°
. .. ,
. i j .
No. ofr strokes of patients
at initial
strokest
assoc. with with NRAF
developing
(% of whole
screen
NRAF
NRAF
developing
population)
(years)
stroke
INO.
Strokes with
AF as % of
total strokes
63
40-69
251(1-3%)
10(RR6-9)*
16%
4-0%
48
40-59
71 (0-9%)
1 (RR2-3)*
2%
1-4%
78
30-62
345 (6-7%)
20 (RR 5-6)
26%
5-8%
311
311
30-62
572(11-3%)
83t(RR ~4)
27%
50-59
60-69
70-79
80-89
14-5%
Age
AttrR 1-5%
AttrR 2-8%
AttrR 9-9%
AttrR 23-5%
*Age adjusted relative risk.
fRecalculated (includes strokes and transient ischaemic attack).
JMedian duration of atrial fibrillation to stroke >3 years; about 50% of patients also had heart failure.
AttrR = percentage of strokes that could be attributed to atrial fibrillation after correction for other concomitant risk factors.
The randomized controlled trials
The seven randomized controlled trials included patients
with both paroxysmal and chronic stable atrial fibrillation (Tables 2 and 3). About 35% of patients were
women, with a mean age of 70 years; about 25% had
heart failure and with the exception of EAFT less than
10% had had a prior stroke. The mean follow-up was
around 2 years and the target range for anticoagulation
varied from what many would consider subtherapeutic
to dangerous (International Normalized Ratio (INR)
1-4-4-5). Warfarin was administered double-blind only
in the CAFA and SP1NAF studies and aspirin in all but
the SPAF-II study. The overall risk of stroke was much
greater in the EAFT study, reflecting the high risk of a
recurrent event in patients with prior stroke and atrial
fibrillation.
Overall, about 10% (about 5% per annum) of the
control groups in these studies developed a stroke. The
risk of stroke did not seem to vary with time, suggesting that the annual risk of stroke in survivors would
continue indefinitely at around 5% per annum.
Only one trial (SPAF-I) showed aspirin to be
effective in reducing stroke in atrial fibrillation. The
inability of clinical trials to show superiority of aspirin
over warfarin should not be accepted as evidence for the
efficacy of aspirin. The combined results from the trials
suggest, at best, that treating 200 patients with aspirin
for one year will prevent one stroke, no deaths but with
little evidence of excess bleeding. Five of six trials
showed that warfarin could reduce stroke compared to
control. One trial was stopped early, in view of the
positive results in the other trials, with a trend in favour
of warfarin. The combined results from these trials
suggesting that treating 200 patients for one year with
warfarin will prevent about six strokes and one to two
deaths, but will be associated with one to two major
bleeding episodes.
Studies comparing aspirin and warfarin have
incorporated fewer patients and had fewer events. The
AFASAK study suggested that warfarin was superior to
aspirin. The SPAF studies also suggest that warfarin is
superior to aspirin in the prevention of thrombo-embolic
stroke but that, especially in the elderly, this is offset by
an increase in intra-cranial haemorrhage.
Eur Heart J, Vol. 17, May 1996
Downloaded from by guest on October 21, 2014
of 'silent' cerebral infarction detected by computed
tomography scans may be as high as 15%'20'. Whether
this is truly silent or leads to important cognitive impairment is unclear'211. Therefore the risk of stroke is high
enough to warrant anticoagulating most patients with
atrial fibrillation.
The Framingham data suggest that atrial fibrillation increases the relative risk of stroke more in older
patients and that stroke may be attributed to atrial
fibrillation in almost 25% of events in patients over the
age of 80 years1'1. Older patients are also at greatest risk
of complications from anticoagulants, although perhaps due to the presence of concomitant complicating
conditions such as peptic ulcer rather than due to age
itself*22'23'. Thus, at least one group with much to gain
also have an increased risk from therapy.
The prevalence of both stroke and atrial fibrillation increase with age, raising the possibility that atrial
fibrillation is a risk marker rather than a risk factor124^.
However, patients without atrial fibrillation have a
lower risk of stroke than those with atrial fibrillation,
even after matching for the presence of carotid and
cardiac disease, indicating that atrial fibrillation is
indeed a risk factor'1'.
ON
s
n
o
Table 2 Randomized controlled trials comparing warfarin andlor aspirin andlor control
Study
Design
AFASAK121
1007 CAF
SPAF-I'4-51
1330 CAF or PAF
BAATAF
131
CAFA' 6 '
171
SPINAF
8
EAFT' -*"
420 CAF or PAF
Double-blind aspirin
Open label warfarin
Double-blind aspirin
Open-label warfarin
Open label warfarin
% Female
, A g e ; year , S . . % with HF
(mean or median)
Downloaded from by guest on October 21, 2014
8
(V
5"
% with prior CVA
Follow-up
(months)
Aspirin dose
(mg)
24
75 mg
7
16
325
21 years
26
—
46
74
52
4
29
67
19
28
68
26
7
(CVA or TIA)
3
378 CAF or PAF
Double-blind warfarin
25
68
22
4
15
—
571 CAF
Double-blind warfarin
0
67
30
8
20
—
44
73
10
100
(TIA or minor stroke)
28
300
24
64
17
??
37
325
41
80
26
77
24
325
1007 CAF or PAF
SPAF-ll <75 years''
715
SPAF-II >75 years1'"
385
Double-blind aspirin
Open-label warfarin
Open-label aspirin
Open-label warfarin
Open-label aspirin
Open-label warfarin
CAF=chronic atrial fibrillation. PAF=paroxysmal fibrillation. CVA = cerebrovascular accident. TIA = transient ischaemic attack. HF = heart failure
AFASAK = Atrial Fibrillation Aspirin Anticoagulation Kobenhavn Trial.
BAATAF = Boston Area Anticoagulation Trial in Atrial Fibrillation Trial
CAFA = Canadian Atrial Fibrillation Anticoagulation Trial.
SPAF=Stroke Prevention in Non-rheumatic Atrial Fibrillation Trial.
SPINAF=Stroke Prevention in Non-rheumatic Atrial Fibrillation Trial.
EAFT= European Atrial Fibrillation Trial.
Median duration
of atrial fibrillation
32%
<1 year
19%
<1 year
~ 8 years
(mean)
46%
< 1, year
Patients
from
SPAF-I
Evidence-based cardiology
677
Table 3 Outcome in randomized controlled trials comparing warfarin andlor aspirin andlor control
Study
AFASAK
SPAF-I (a)
SPAF I (b)
BAATAF
CAFA
SPINAF
EAFT (a)
EAFT (b)
Stroke
(controls)
Stroke
(aspirin)
Stroke
(warfarin)
Mortality
(controls)
Mortality
(aspirin)
Mortality
(warfarin)
Haemorr
(controls)
Haemorr
(aspirin)
n= 16/336
2-4% p.a.
n= 17/211
6 0 % p.a.
n=42/568
6-3% p.a.
n= 13/208
2-9% p.a.
n = 9/191
4-6% p.a.
n = 23/290
4-8% p.a.
n = 50/214
100% p.a.
n = 90/3 78
10 2% p.a.
n= 15/336
2-2% p.a.
—
n = 5/335
0-7% p.a.
n=7/210
2-5% p.a.
—
n = 17*
2-5% p.a.
n=8
3 1 % p.a.
n = 50
8-8% p.a.
n = 26
5-8% p.a.
n=8
3-3% p.a.
n = 26
5-9% p.a.
n=44
8-8% p.a.
n = 99
11 -2% p.a.
n=15*
2-2% p.a.
—
n = 3*
0-4% p.a.
n=6
2-2% p.a.
—
n=0
n=0
SPAF-I I <75 years
SPAF-II >75 years
Control vs aspirin
Control vs warfarin
—
n = 88/404
9-3% p.a.
n = 21/357
1-9% p.a.
n = 21/188
5-6% p.a.
n = 2/212
0-4% p.a.
n = 8/187
3-4% p a.
n = 7/281
1-5% p.a.
n=20/225
3-8% p.a.
n= 18/358
1-6% p.a.
n = 20/197
5-1% p.a.
127/1292
(9-8%) ns
57/881
(6-5%)
49/1450
(3-4%)
/><00001
43/890
(4-8%) ns
166
(12.9%)
129
(8-9%)
n = 39
7 1% p.a.
—
—
n=102
10-8% p.a.
n=41
3-7% p.a.
n = 24
6-1% p.a.
n= 11
2-4% p.a.
n=10
4 3% p.a.
n = 20
4-4% p.a.
n=41
7-8% p.a.
n = 36
3 2% p.a.
n = 26
6-4% p.a.
156
(12-1%) ns
91
(6-3%)
80
(9-1%)
p<om
65
(7-3%) ns
18
(1 4%)
24
(1-7%)
n=2
0-80% p.a.
n=4
1 -5% p.a.
—
n=8
1-8% p a .
n=5
2-5% p.a.
n=7
1-5% p.a.
n=13
2-8% p.a.
n~19
1-7% p.a.
n-16
4-2% p.a.
16
(1-2%) ns
16
(1 8%)
39
(2-7%)
/><005
37
(4-2%)
P<00l
Figures in bold represent significant differences from comparative groups. Numbers of strokes excludes transient ischaemic attacks and
includes intracerebral haemorrhage. Rates of major haemorrhage only are quoted and have been taken from ref1261 where possible due to
variable reporting. Major haemorrhage included fatal or requiring hospitalization with or without blood transfusion. About 20% of
patients will have minor bleeding (nose bleeds, bleeding piles etc) on warfarin although it should be appreciated there is a significant
hazard of minor bleeding in the control groups in these studies * = unknown or vascular death Total mortality in the AFASAK study
was 71.
Annual rates of stroke are re-calculated rather than those given in the paper due to inclusion of variable end-points in quoted rates.
For abreviations see Table 2.
Although the selection process for patients randomized in these studies does appear to have produced a
control group with an event rate representative of the
community studies, there must be concerns that patients
with even a moderate increased risk of bleeding complications have been excluded'22'251. This, combined with
the more careful measurement of INR and better patient
compliance with treatment than in clinical practice
means that the trials may overestimate the safety of
warfarin. However, as the trial results are presented as
'intention-to-treat' and many strokes in patients randomized to warfarin occurred among patients who had
stopped treatment, the trials also underestimate the real
benefits on stroke reduction among compliant patients.
Moreover, stroke is usually a more devastating event
with longer lasting consequences than haemorrhage'26'.
The most common reason for stopping warfarin was
patient refusal to continue with monitoring.
The high prevalence of heart failure among
patients with atrial fibrillation is a further reason to be
cautious about the use of aspirin in this setting as aspirin
may negate the benefits of treatment with an ACE
inhibitor'271.
In summary, there is little evidence to support a
role for aspirin specifically for the management of atrial
fibrillation (but see below). Warfarin is clearly effective.
However, the risk of stroke in many patients with
atrial fibrillation is low. Even if the relative benefits of
warfarin are great in these patients the absolute risk may
be sufficiently small to warrant withholding warfarin.
Likewise the risk of haemorrhage in some patients is
high and this needs to be taken into account.
Risk stratification in atrial fibrillation
The duration of atrial fibrillation and whether it is
continuous or paroxysmal does not affect the risk of
stroke'281. Age is an important risk factor'281. Patients
over 65 years of age have a greater than 3-5% annual risk
of stroke even if other risk factors are absent'281. Other
important risk factors for stroke in atrial fibrillation on
Eur Heart J, Vol. 17, May 1996
Downloaded from by guest on October 21, 2014
Aspirin vs warfarin
148/1282
(11-5%)
128/1450
(8-8%)
n = 24/552
3-8% p.a.
—
n=4
—
1-6% p.a.
n=14
n=10
1-9% p.a. 1-4% p.a.
n=8
—
1-8% p.a.
n=l
—
0-5% p.a.
n=4
0-9% p.a
n=3
0-7% p.a.
n=4
n=6
0-7% p.a 0-9% p.a.
n~10
0 9% p.a
n~6
1-6% p.a.
Haemorr
(warfarin)
678 J. C. F. Cleland et al.
Table 4 Incidence of embolic events in landmark studies of heart failure[56]
Trial
V-HeFT-I
NYHA
I I/I 11
CONSENSUS
III/IV
V-HeFT-II
II/III
SOLVD-T + P
PROMISE
I—111
III/IV
n =
follow-up
Anti-platelet
agents
Anticoagulants
AF
(%)
642
2-3 years
253
0-5 years
804
2 6 years
6796
3 3 years
1088
0-5 years
13%*
13%
(31% if AF)
33-5%
15
77
27%*
34%/54%
21%
(41% if A F)
16%/12%
CVA
%
PTE
%
Other
Total
%
%
40
0-8
0-8
*r>
7?
5-6
(2-5)t
77
0-7
0-2
0-7
11
r>
7?
l-8t
50
13
2-3*
4-6f
4-7
1 8t
6
38
l-2t
4-0
3-5J
5-6
(2-2)t
5-6
(1 7)t
•77
Rows in bold identify the two studies of patients with the most severe heart failure; they also have the highest annualized stroke
incidence.
*Only fatal strokes reported. This may only be one third of the total stroke event rate.
•(•Calculated annualized stroke risk. J = reported annualized stroke rate.
PTE = Previous thrombolic embolism.
V-HeFT-I = Vasodilator-Heart Failure Trial I.
CONSENSUS = Comparative North Scandinavian Enalapril Survival Study.
SOLVD = Studies of Left Ventricular Dysfunction.
PROMISE = Prospective Randomised Milnnone Survival Evaluation trial.
For other abbreviations, see Table 2.
Eur Heart J. Vol 17. May 1996
patients with heart failure have atrial fibrillation a high
risk of embolic stroke might be anticipated but this may
be offset by lower risk of thrombotic and haemorrhagic
stroke due to lower blood pressures (due to ventricular
dysfunction and its treatment). Thus, heart failure may
change the balance of the aetiology of stroke in heart
failure rather than the overall rate. Aspirin and warfarin
treatment in heart failure trials has not been randomized
and no conclusion of their efficacy, or lack of it, can be
made from these data.
In summary, there is a powerful argument for
treating all patients with atrial fibrillation with warfarin
if they are over 65 years of age. However, the SPAF-I I
study shows that the risk of intracranial haemorrhage
increases in patients over 75 years, especially when the
INR rises above 30. Interestingly, the EAFT study also
suggested that warfarin was less effective in reducing the
overall incidence of stroke in patients >75 years of
age'381, perhaps for the same reasons. Accordingly the
target INR should be lower in older patients. An INR of
20-2-5 may be ideal, but lower target ranges should be
explored.
In younger patients without risk factors an
echocardiogram should be done. If this shows normal
left ventricular function and a left atrium <40 mm the
risk is probably < 10% and no treatment is necessary. In
those with risk factors or an abnormal echocardiogram
warfarin should be instituted.
Are any benefits of warfarin or aspirin
on stroke confined to patients with
atrial fibrillation (Table 5)
Older people and those who have already had a stroke
are more likely to have an event whether or not the
Downloaded from by guest on October 21, 2014
a multivariate analysis are previous transient ischaemic
attack or stroke, hypertension and diabetes1281 and, less
consistently, evidence of impaired left ventricular
function129"311.
The EAFT study identified a population with an
annual risk of recurrent stroke of about 12% and
demonstrated that initiation of warfarin after the index
event (43% within 2 weeks) was highly effective in
preventing further events. Although data from surveys
suggested that the risk of recurrent stroke was greatest in
the months after the initial event, this was not true in
EAFT, the excess risk being similar early and late after
the initial event.
The SPAF-I study showed that a history of
hypertension, heart failure, previous thromboembolism, left ventricular dysfunction or left atrial dilatation on echocardiography were all related to an
increased risk of thrombo-embolism and that absence of
these conditions could identify 26% of the population
with a thrombo-embolic risk of 10% per year, although
the confidence intervals around this estimate could not
exclude a risk as high as 4% per year130'3'1. Mitral
annular calcification also appears to be a risk factor132'.
Women also appear to be at higher risk of stroke
and get greater benefit from warfarin, the latter perhaps
reflecting better compliance'28'. Measurements of haemostatic factors and spontaneous echocardiographic
contrast may be alternative methods for assessing the
risk but more experience is required before they can be
recommended in clinical practice133'341.
Interestingly, studies of patients with moderately
severe heart failure do not suggest that atrial fibrillation
is associated with a higher overall risk of stroke and
there is a dispute about whether patients with heart
failure are at higher or lower risk of stroke than the
general population (Table 4)[3O-3l-35-37i. As up to 50% of
Evidence-based cardiology 679
Table 5 Effects of aspirin and warfarin on stroke, mortality and haemorrhage in the two largest long-term post
myocardial infarction trials with each agent
Stroke
(controls)
Stroke
(aspirin)
AMIS' 4 '
45/2257f
0-6%
27/2267
0-4% p.a.
PARIS II[401
33/1565
1 1 % p.a.
20/1563
0-6% p.a.
WARIS
(INR 2-8-4-8)1391
44/607
2-4% p.a.
20/607
1-1% p.a.
123
6 6% p a.
ASPECT
(INR 2-8^-8)' 42!
42/1704
0-8% p.a.
24/1700
0-4% p.a.
P<005
n=189
3-6% p.a.
Study
Stroke
(warfarin)
Mortality
(controls)
Mortality
(aspirin)
Mortality
(warfarin)
Haemorr
(controls)
Haemorr
(aspirin)
219
3 0% p.a.
246
3-4% p.a. ns
—
41*
0-6% p.a.
73**
11% p.a.
114
3-7% p.a.
111
3-6% p.a. ns
—
50**
1-6% p.a.
53**
1 -7% p.a. ns
94
5 0 p.a.
/><003
n=170
3-2% p.a. ns
0
Haemorr
(warfarin)
P<005
n=19
0-4% p.a.
8
0-4% p.a.
P<001
n=73
1-4% p.a.
P<0-0l
fNon-fatal only
*Defined as a fall in haematocrit associated with a gastro-intestinal problem
**Haematemesis or melaena only.
AMIS = Aspirin Myocardial Infarction Study
PARIS II = Sersantine-Aspirin Reinfarction Study (Part II).
WARIS = Warfarin Reinfarction Study.
ASPECT=Anticoagulants in the Secondary Prevention of Events in Coronary Thrombosis
Table 6 Adequacy of anti-coagulant control in studies of atrial fibrillation
AFASAK
SPAF-I
BAATAF
CAFA
SPINAF
EAFT
SPAF-I I
<75 years/>75 years
Mode of
monitoring
Anticoagulation
below range
Estimated
INR range
Estimated
average INR
Anticoagulation
above range
INR
PTR
PTR
INR
PTR
INR
PTR/INR
58%
23%
8%
40%
29%
32%
20%/22%
2-8-4-2
2-0-4-5
1-5-2-7
2-0-3-0
1 4-2-8
2-5-4-0
2 0-4-5
Not given
0-6%
5%
9%
17%
15%
9%
5%/6%
~2-5
Not given
2-4
Not given
2-9
2-7/2-6
Downloaded from by guest on October 21, 2014
Study
I NR = International normalized ratio.
PTR = Prothrombin time ratio.
For other abbreviations see Table 2.
patient has atrial fibrillation. Review of postinfarction
studies with aspirin and warfarin show that both
agents can reduce stroke, the benefits of warfarin being
strikingly greater than those of aspirin in the long-term
trials'39"^2'. Patients with atrial fibrillation are not immune to stroke due to carotid disease, although it does
not appear to confer a greater risk'431, and it is possible
that aspirin reduces the risk-marker component of
stroke in atrial fibrillation rather than stroke caused by
embolism from the heart itself. It is likely that warfarin
reduces stroke of both vascular and cardiac origin.
What is the lowest effective INR?
(Table 6)
The risk of haemorrhage with warfarin is generally
related to the intensity of anticoagulation, the age of the
patient and any underlying diseases. It is not clear that
the efficacy of warfarin is related to the intensity of the
INR when the INR is maintained above 1-4. The
difficulties in translating prothrombin times into INR
values adds further uncertainty about what the optimal
degree of anticoagulation is. The two trials with the
greatest reduction in stroke also had the lowest therapeutic range'3"71. In the other trials a third or more of
INRs were below the target range (Table 6). These data
suggest that much lower doses of warfarin than currently recommended may be effective in atrial fibrillation. The EAFT study suggested a target INR of 30
with a lower limit of 20 [44) . Unfortunately the EAFT
investigators did not analyse for a category between 1-5
and 20. An INR of 20-2-4 appeared to be associated
with as much benefit as more intense anticoagulation
with a lower risk of adverse events.
While bleeding is generally considered adverse,
data from the AFASAK study reminds us that bleeding
can be an early warning sign of underlying malignant or
inflammatory disease'2221.
Eur Heart J, Vol. 17, May 1996
680 J. C. F. Cleland et al.
What can be done to improve
anticoagulant services?
The major barrier to the use of effective anticoagulation
is the need for frequent monitoring. The best solution
would be to reduce the frequency of monitoring but this
can probably only safely be done by proving that a
lower target range of INR is effective. Alternative solutions would be to develop home (or work place) monitoring by a nurse/technician based system. Kits are
already available although wider experience of their
strengths and limitations is required before they can be
recommended. For the moment regional laboratory
services with a team dedicated to good anti-coagulant
control would seem ideal. Efficient monitoring of
anticoagulation is probably highly cost-effective'451.
for patients with atrial fibrillation at a low risk of stroke
but should not be used as an excuse to withhold
anticoagulants in patients at greater risk.
Several larger studies investigating the effects of
different intensities of anticoagulation and the use of
aspirin-warfarin combinations are underway. Indeed
SPAF-III, comparing a combination of low dose
warfarin and aspirin with formal anticoagulation has
been stopped and reported in March 1996. A summary
of the results will appear in the July issue. Identification
of the minimum effective dose of warfarin and effective
monitoring systems remain a priority.
Dr J. G. F. Cleland is supported by the British Heart Foundation. The authors would like to thank Dr Gordon Murray
(Robertson Centre for Biostatistics, University of Glasgow) for
statistical advice.
References
Patients undergoing cardioversion are at increased risk
of thrombo-embolic events because of pre-formed
thrombus, and the risk of developing thrombus after
cardioversion due to relapse of atrial fibrillation and
the delay in recovery of atrial function'461. In patients
receiving less than 2 days anticoagulation, the risk of
an embolic incident associated with cardioversion is
between 3 and 7%t47~*9]. Early hopes that transoesophageal echocardiography could identify patients without
pre-formed thrombus'501 who could be cardioverted
without anticoagulation have been dashed'51521. However, transoesophageal echocardiography to exclude
pre-formed thrombus is justified when rapid cardioversion is required, and could be a cost-effective way
of dealing with acute atrial fibrillation'53'541. Such
patients should be given heparin, cardioverted and
subsequently anticoagulated with warfarin for 1-2
months'53541. In the absence of a transoesophageal
echocardiogram excluding atrial thrombus, anticoagulation for 4-6 weeks is required whether cardioversion is
attempted electrically or by drugs. Patients with acute
atrial fibrillation are at risk of stroke and should be
heparinized immediately.
[1] Wolf PA, Abbott RD, Kannel WB. Atrial fibrillation as an
independent risk factor for stroke. The Framingham Study.
Stroke 1991; 22. 983-8.
[2] Petersen P, Godtfredsen J, Andersen B, Boysen G, Andersen
ED. Placebo-controlled, randomised trial of warfarin and
aspirin for prevention of thromboembolic complications in
chronic atrial fibrillation. The Copenhagen AFASAK. study.
Lancet 1989; 1: 175-9.
[3] Singer DE, Hughes RA, Gress DR et al. The effect of
low-dose warfarin on the risk of stroke in patients with
nonrheumatic atrial fibrillation. N Engl J Med 1990; 323:
1505-11.
[4] McBnde R, Anderson DC, Asinger RW et al. Preliminary
report of the stroke prevention in atrial fibrillation study.
N Engl J Med 1990; 322: 863-8.
[5] McBridge R. Stroke prevention in atrial fibrillation study
Final results. Circulation 1991; 84: 527-39.
[6] Connolly SJ, Laupacis A, Gent M et al. Canadian Atrial
Fibrillation Anticoagulation (CAFA) Study. J Am Coll
Cardiol 1991; 18: 349-55.
[7] Ezekowitz MD, Bridgers SL, James KE et al. Warfarin in the
prevention of stroke associated with nonrheumatic atrial
fibrillation. N Engl J Med 1992; 327: 1406-12.
[8] Van Latum JC, Vermeulen PC, Den Ouden A et al. Secondary
prevention in non-rheumatic atrial fibrillation after transient
ischaemic attack or minor stroke. Lancet 1993; 342: 1255-62.
[9] Chesebro JH, Wiebers DO, Holland AE et al. Warfarin versus
aspirin for prevention of thromboembolism in atrial fibrillation: Stroke prevention in atrial fibrillation II study. Lancet
1994; 343: 687-91.
[10] Lip GYH, Nyen Tean K, Dunn FG. Treatment of atrial
fibrillation in a district general hospital. Br Heart J 1994; 71:
92-5.
[11] Lancaster TR, Singer DE, Sheehan MA et al. The impact of
long-term warfarin therapy on quality of life: Evidence from a
randomized trial. Arch Intern Med 1991; 151: 1944-9.
[12] Boysen G, Nyboe J, Appleyard M et al. Stroke incidence and
risk factors for stroke in Copenhagen, Denmark. Stroke 1988;
19: 1345-53.
[13] Lake FR, Cullen KJ, De Klerk NH, McCall MG, Rosman
DL. Atrial fibrillation and mortality in an elderly population.
Aust New Zealand J Med 1989; 19: 321-6.
[14] Hill JD, Mottram EM, Killeen PD. Study of the prevalence of
atrial fibrillation in general practice in patients over 65 years
of age. J R Coll Gen Pract 1987; 37: 521-7.
[15] Campbell A, Caird Fl, Jackson TFM. Prevalence of abnormalities of electrocardiogram in old people. Br Heart J 1974;
36: 1005-11.
Summary
Anticoagulants should be used more widely in patients
with atrial fibrillation. Legitimate concerns exist about
the risk/benefit ratio in younger patients with no risk
factors and in patients over the age of-75 years. Use of
lower doses of anticoagulation (potential target range
INR of 1-5-2-5) than used heretofore is probably the
solution to most of the problems associated with anticoagulation, but conclusive proof of the efficacy of this
strategy is needed. Although aspirin may reduce the risk
of stroke the effect may be no more than among patients
with a similar level of cardiovascular risk factors and
in sinus rhythm. As such, aspirin is a valid alternative
Eur Heart J. Vol. 17, May 1996
Downloaded from by guest on October 21, 2014
Management of cardioversion
Evidence-based cardiology
[39]
[40]
[41]
[42]
[43]
[44]
[45]
[46]
[47]
[48]
[49]
[50]
[51]
[52]
[53]
[54]
[55]
[56]
patients with a transient ischemic attack or minor ischemic
stroke and with nonrheumatic atrial fibrillation. Stroke 1995;
26: 801-6.
Smith P, Arnesen H, Holme I. The effect of warfarin on
mortality and reinfarction after myocardial infarction. N Engl
J Med 1990; 323: 147-52.
Klimt CR, Knatterud GL, Stamler J, Meier P. Persantineaspirin reinfarction study. Part II. Secondary coronary prevention with persantine and aspirin. J Am Coll Cardiol 1986;
7: 251-69.
Anonymous. A randomized, controlled trial of aspirin in
persons recovered from myocardial infarction. J Am Med
Assoc 1980; 243. 661-9.
Van Bergen PFMM, Jonker JJC, Van der Meer FJM et al.
Effect of long-term oral anticoagulant treatment on mortality
and cardiovascular after myocardial infarction. Lancet 1994;
343: 499-503.
Kanter MC, Tegeler CH, Pearce LA et al. Carotid stenosis in
patients with atrial fibrillation. Prevalence, risk factors, and
relationship to stroke in the stroke prevention in Atrial
Fibrillation Study. Arch Intern Med 1994; 154: 1372-7.
The European Atrial Fibrillation Trial Study Group. Optimal
oral anticoagulant therapy in patients with nonrheumatic
atrial fibrillation and recent cerebral ischaemia. N Engl J Med
1995; 333: 5-10.
Laupacis A. Cost effectiveness of stroke prevention in atrial
fibrillation. Ann Intern Med 1993; 118- 92 (Letter).
Fatkin D, Kuchar DL, Thorburn CW, Feneley MP. Transesophageal echocardiography before and during direct current
cardioversion of atrial fibrillation: Evidence for 'atrial
stunning' as a mechanism of thromboembolic complications.
J Am Coll Cardiol 1994; 23: 307-16.
Arnold AZ, Mick MJ, Mazurek RP, Loop FD, Trohman RG.
Role of prophylactic anticoagulation for direct current cardioversion in patients with atrial fibrillation or atrial flutter. J Am
Coll Cardiol 1992; 19: 851-5.
Weinberg DM, Mancini GBJ. Anticoagulation for cardioversion of atrial fibrillation. Am J Cardiol 1989; 63: 745-6
Bjerkelund CJ, Orning OM. The efficacy of anticoagulant
therapy in preventing embolism related to D.C. electrical
conversion of atrial fibrillation. Am J Cardiol 1969; 23:
208-16.
Manning WJ, Silverman DI, Gordon SPF, Krumholz HM,
Douglas PS. Cardioversion from atrial fibrillation without
prolonged anticoagulation with use of transesophageal
echocardiography to exclude the presence of atrial thrombi. N
Engl J Med 1993; 328: 750-5.
Black IW, Fatkin D, Sagar KB et al. Exclusion of atrial
thrombus by transesophageal echocardiography does not preclude embolism after cardioversion of atrial fibrillation: A
multicenter study. Circulation 1994; 89: 2509-13.
Missault L, Jordaens L, Gheeraert P, Adang L, Clement D.
Embolic stroke after unanticoagulated cardioversion despite
prior exclusion of atrial thrombi by transoesophageal echocardiography. Eur Heart J 1994; 15: 1279-80.
Manning WJ, Silverman DI, Keighley CS, Oettgen P, Douglas
PS. Transesophageal echocardiographically facilitated early
cardioversion from atrial fibrillation using short-term anticoagulation: Final results of a prospective 4.5-year study. J
Am Coll Cardiol 1995; 25: 1354-61.
Manning WJ, Silverman DI, Katz SE et al. Impaired left atrial
mechanical function after cardioversion: Relation to the
duration of atrial fibrillation. J Am Coll Cardiol 1994; 23:
1535-40.
Flegel KM, Shipley MJ, Rose G. Risk of stroke in nonrheumatic atrial fibrillation. Lancet 1987; 1: 526-9.
Cleland JGF. The clinical course of heart failure and its
modification by ACE inhibitors: Insights from recent clinical
trials. Eur Heart J 1994; 15: 125-30.
Eur Heart J, Vol. 17, May 1996
Downloaded from by guest on October 21, 2014
[16] Flegel KM, Hanley J Risk factors for stroke and other
embolic events in patients with nonrheumatic atrial fibrillation. Stroke 1989; 20: 1000-4.
[17] Wolf PA, Dawber TR, Thomas EJ, Kannel WB. Epidemiologic assessment of chronic atrial fibrillation and risk of
stroke. The Framingham Study. Neurology 1978; 28: 973-7.
[18] Wolf PA, Abbott RD, Kannel WB. Atrial fibrillation: A
major contributor to stroke in the elderly. The Framingham
study. Arch Intern Med 1987; 147: 1561-4.
[19] Kopecky SL, Gersh BJ, McGoon MD et al. The natural
history of lone atrial fibrillation. A population-based study
over three decades. N Engl J Med 1987; 317: 669-74.
[20] Ezekowitz MD, James KE, Nazarian SM et al. Silent cerebral
infarction in patients with non-rheumatic atrial fibrillation.
Circulation 1995; 92: 2178-82.
[21] O'Connell JE, Gray CS. Atrial fibrillation and cognitive
impairment. Lancet 1992; 340: 1169.
[22] Gitter MJ, Jaeger TM, Petterson TM, Gersh BJ, Silverstein
MD. Bleeding and thromboembolism during anticoagulant
therapy: a population-based study in Rochester, Minnesota.
Mayo Clin Proc 1995; 70: 725-33.
[23] Gurwitz JH, Goldberg RJ. Holden A, Knapic N, Ansell J.
Age-related risk of long-term oral anticoagulant therapy.
Arch Intern Med 1988; 148. 1733-6.
[24] Chesebro
JH,
Fuster
V,
Halperin
JL.
Atrial
fibrillation — Risk marker for stroke. N Engl J Med 1990;
323: 1556-8.
[25] Gottlieb LK, Salem Schatz S. Anticoagulation in atrial fibrillation: Does efficacy in clinical trials translate into effectiveness in practice? Arch Intern Med 1994; 154: 1945-53.
[26] Caro JJ, Groome PA, Flegel KM. Atrial fibrillation and
anticoagulation: From randomised trials to practice. Lancet
1993; 341: 138 l^t.
[27] Cleland JGF, Bulpitt CJ, Falk RH et al. Is aspirin safe for
patients with heart failure? Br Heart J 1995; 74: 215-19.
[28] Laupacis A, Boysen G, Connolly S et al. Risk factors
for stroke and efficacy of antithrombotic therapy in atrial
fibrillation: Analysis of pooled data from five randomized
controlled trials. Arch Intern Med 1994; 154: 1449-57.
[29] Petersen P, Kastrup J, Helweg Larsen S, Boysen G,
Godtfredsen J. Risk factors for thromboembohc complications in chronic atrial fibrillation. The Copenhagen
AFASAK study. Arch Intern Med 1990; 150: 819-21.
[30] Pearce LA. Predictors of thromboembolism in atrial fibrillation: II. Echocardiographic features of patients at risk Ann
Intern Med 1992; 116: 6-12.
[31] Pearce LA. Predictors of thromboembolism in atrial fibrillation: I. Clinical features of patients at risk Ann Intern Med
1992; 116: 1-5.
[32] Benjamin EJ, Plehn JF, D'Agostino RB et al. Mitral annular
calcification and the risk of stroke in an elderly cohort. N Engl
J Med 1992; 327: 374-9.
[33] Lip GYH, Lowe GDO, Rumley A, Dunn FG. Increased
markers of thrombogenesis in chronic atrial fibrillation.
Effects of warfarin treatment. Br Heart J 1995; 73: 527-33.
[34] Leung DYC, Black IW, Cranney GB, Hopkins AP, Walsh
WF. Prognostic implications of left atrial spontaneous echo
contrast in nonvalvular atrial fibrillation. J Am Coll Cardiol
1994; 24: 755-62.
[35] Dunkman WB, Johnson GR, Carson PE, Bhat G, Farrell L,
Cohn JN. Incidence of thromboembolic events in congestive
heart failure. Circulation 1993; 87: VI94-VI101.
[36] Katz SD, Marantz PR, Biasucci L et al. Low incidence of
stroke in ambulatory patients with heart failure: A prospective
study. Am Heart J 1993; 126: 141-6
[37] Middlekauff HR, Stevenson WG, Stevenson LW. Prognostic
significance of atrial fibrillation in advanced heart failure: A
study of 390 patients. Circulation 1991; 84: 40-8.
[38] Van Latum JC, Koudstaal PJ, Venables GS, Van Gijn J,
Kappelle LJ, Algra A. Predictors of major vascular events in
681