Bleeding risk in randomized controlled trials comparing warfarin and aspirin: a systematic review and meta-analysis

Page created by Ralph Mclaughlin
 
CONTINUE READING
Journal of Thrombosis and Haemostasis, 10: 512–520                                                       DOI: 10.1111/j.1538-7836.2012.04635.x

  ORIGINAL ARTICLE

Bleeding risk in randomized controlled trials comparing
warfarin and aspirin: a systematic review and meta-analysis
A. E. WARKENTIN,* M. P. DONADINI,                        F. A. SPENCER,          W. LIM       and M . C R O W T H E R
*Faculty of Medicine, University of Toronto, Toronto, ON; and Department of Medicine, McMaster University, Hamilton, ON, Canada

To cite this article: Warkentin AE, Donadini MP, Spencer FA, Lim W, Crowther M. Bleeding risk in randomized controlled trials comparing warfarin
and aspirin: a systematic review and meta-analysis. J Thromb Haemost 2012; 10: 512–20.

                                                                          Introduction
Summary. Background: Warfarin and aspirin (acetylsalicylic
acid [ASA]) are the most commonly used anticoagulant and                  Anticoagulant and antiplatelet drugs are highly effective for the
antiplatelet drugs in the treatment of cardiovascular dis-                prevention and treatment of thrombotic cardiovascular dis-
ease. Objectives: To provide a pooled estimate of the bleeding            eases. Warfarin and aspirin (acetylsalicylic acid [ASA]) are the
risk from randomized controlled trials (RCTs) comparing                   most commonly used anticoagulant and antiplatelet agents for
warfarin and ASA at the dose ranges recommended in evidence-              long-term prophylaxis in patients with atrial fibrillation,
based guidelines. Patients/Methods: Ovid MEDLINE, Em-                     myocardial infarction, peripheral artery disease, cerebrovascu-
base and the Cochrane Library, up to September 2011, were                 lar disease, heart failure, and heart valve replacement.
searched for RCTs comparing bleeding rates in adult patients              Although they are highly effective, both ASA and warfarin
randomized to warfarin, target International Normalized Ratio             can result in bleeding. Data on the risk of bleeding are
(INR) 2.0–3.5, and ASA, 50–650 mg daily, with at least                    surprisingly limited, perhaps because the reported rates vary as
3 months of follow-up. Pooled odds ratios (ORs) and associ-               a result of study design and population, definition, site of
ated 95% confidence intervals (CIs) were calculated with the               bleeding, and drug dosage. The annual incidence of major
inverse variance method and the random effects mod-                        bleeding in trials and cohort studies has been reported to be
el. Results: Four thousand four hundred and forty-two                     between 1.1% and 2.3% [1–5] in patients treated with warfarin
abstracts were screened, resulting in eight included studies for          to achieve an International Normalized Ratio (INR) of 2.0–3.0
final analysis. A pooled estimate derived from the 2904 patients           and between 1.1% and 1.5% in ASA-treated patients [2,3,6,7].
enrolled indicated a trend towards an increase in major bleeding          These rates are derived from carefully selected patients; rates in
risk in those randomized to warfarin (OR 1.27; 95% CI 0.83–               ÔunselectedÕ outpatients are likely to be higher, as demonstrated
1.94). The pooled OR for intracranial hemorrhage in patients              in a recent nationwide registry, where the risks of bleeding were
treated with warfarin vs. ASA was 1.64 (95% CI 0.71–3.78),                found to be 4.3% and 2.6% in patients receiving warfarin and
and that for extracranial major bleeding was 1.03                         ASA, respectively [8]. Taken together, these studies suggest that
(95% CI 0.61–1.75). Minor bleeding, from a 1748-patient                   bleeding rates with warfarin are higher than those seen with
sample, was more common in warfarin patients (OR 1.50;                    ASA, but the difference between these bleeding rates is less than
95% CI 1.13–2.00). Conclusions: This meta-analysis failed to              expected, given that ASA is a less potent anticoagulant than
find a statistically significant difference in major bleeding                warfarin.
between warfarin, target INR 2.0–3.5, and ASA, 50–650 mg                     In clinical practice, the risk of bleeding related to warfarin is
daily. The trend towards increased bleeding with warfarin                 usually considered to be significantly higher than the risk
appears to be explained by an excess of intracranial bleeding in          associated with ASA. However, a pooled estimate of the
warfarin patients.                                                        magnitude of this difference from direct comparison of
                                                                          bleeding risk in trials that have randomized patients to
Keywords: aspirin, bleeding, meta-analysis, systematic review,            warfarin vs. ASA within current therapeutic ranges has not
warfarin.                                                                 been published, to our knowledge. To address this gap, we
                                                                          performed a systematic review and meta-analysis of random-
                                                                          ized controlled trials (RCTs) that enrolled adult patients with
                                                                          any indication for long-term antithrombotic therapy and
Correspondence: Mark Crowther, Rm L301, St JosephÕs Hospital,             randomized them to warfarin or ASA. Our objective was to
50 Charlton Ave East, Hamilton, L8N 4A6, ON, Canada.
                                                                          compare the bleeding rates between these two groups of
Tel.: +1 905 521 6024; fax: +1 905 521 6090.
                                                                          patients. Many studies have used ASA doses or warfarin INR
E-mail: crowthrm@mcmaster.ca
                                                                          target ranges that are not used in modern clinical practice, so
Received 4 June 2011, accepted 4 January 2012                             we sought to assess bleeding risks by use of the dose ranges

                                                                                  2012 International Society on Thrombosis and Haemostasis
Comparative bleeding risk: warfarin vs. aspirin 513

recommended in evidence-based guidelines: ASA 50–                    abstract selection phase, the article was discussed by the two
650 mg day)1 and INR range 2.0–3.5 [9–11]. We were also              reviewers. If consensus could not be reached, a third reviewer
interested in, as a secondary question, what the bleeding risk       (M.C.) could be consulted. Consultation with a third reviewer
was in trials that randomized patients to warfarin vs. warfarin      could also be utilized for disagreements during the full-text
plus ASA (at equivalent warfarin targets in both treatment           review.
arms, and within the same therapeutic ranges specified                   We included all studies meeting our inclusion criteria,
already). Prespecified secondary analyses of this study included      irrespective of the efficacy outcome of the individual studies
an investigation of bleeding risk according to age (< 70 and         and the indication for antithrombotic therapy, as it has been
‡ 70 years old) and clinical indication for therapy.                 previously reported that the therapeutic indication for anti-
                                                                     thrombotic therapy does not significantly impact on bleeding
                                                                     rates with long-term treatment [12].
Methods
A protocol for assessment of study eligibility was drafted, and
                                                                     Data extraction and study bias assessment
is available upon request from the authors. No study review
protocol is available for data extraction and data analysis.         Two reviewers (A.E.W. and M.P.D.) independently per-
                                                                     formed the data extraction, using an extraction sheet drafted
                                                                     for the current study, and completed a bias assessment of
Data sources and searches
                                                                     included articles utilizing the Cochrane Handbook for
We searched the Ovid MEDLINE (1948 to August week 4                  Systematic Reviews of Interventions [13]. Data pertaining to
2011) and Ovid MEDLINE In-Process and Other Non-                     study inclusion criteria, such as target INR, ASA dosage, and
Indexed Citations (2 September 2011), Embase (1980–2011              duration of follow-up, were extracted, as were data pertaining
week 35) and the Cochrane Central Register of Controlled             to factors such as number of patients enrolled in the study,
Trials (2011, issue 3) databases. The search strategies for these    patient characteristics, clinical indication for antithrombotic
three databases were formulated by A.E.W., following consul-         medication, bleeding definitions, bleeding outcomes in all
tation with a professional librarian. For a complete description     patients and in specific subgroups, and number of patients
of search terms that were applied to each of these three             assigned to the treatment arms of interest to the current
databases, see Appendix S1. The electronic search strategy was       review (for a complete description, data extraction forms are
complemented by manually reviewing the reference lists of            available from the authors upon request). Owing to hetero-
articles that were identified for full-text review, as well as        geneity in across-study definitions for bleeding, major
through contact with content experts.                                bleeding events were counted as they were defined in each
                                                                     study.
                                                                        Discrepancies in the extracted data were resolved by
Study selection
                                                                     consensus between the two reviewers.
The included studies, for the purposes of full-text review,             The bias assessment of studies was conducted in accor-
satisfied the following criteria: (i) the study considered            dance with that specified in the Cochrane Handbook for
patients prescribed warfarin with a target INR of 2.0–3.5 as         Systematic Reviews of Interventions [13]. Studies were
compared with those precribed ASA at a dose of 50–                   assessed for selection bias, performance bias, detection bias,
650 mg day)1, or warfarin vs. warfarin plus ASA (with the            attrition bias, and reporting bias. Selection bias assessment
same warfarin INR ranges and ASA dosages already                     involved analysis of random sequence generation and allo-
specified); (ii) the study was an RCT; (iii) adult patients were      cation concealment. Performance bias assessment involved
enrolled; (iv) patient follow-up was carried out for at least        the blinding of study participants and personnel. Detection
3 months; and (v) bleeding was an outcome ascertained in the         bias assessment involved the blinding of study outcome
study. The study titles and abstracts were screened for study        assessors. Attrition bias assessment involved an analysis of
inclusion, and if any of the five criteria were not satisfied, the     the completeness of study outcome data. Reporting bias
study was excluded. If there was ambiguity as to whether each        assessment involved an analysis of selective reporting of
criterion was satisfied, the abstract under consideration was         study outcomes [13]. For each component of the analysis,
noted for full-text review. In addition, we excluded articles        studies could receive a label of: low risk of bias; high risk of
that were not published in English, if the full-text article could   bias; or unclear risk of bias. In an effort to avoid a possible
not be retrieved (online, from the library, or via interlibrary      selection bias, all studies that were retrieved and subsequently
loan), and if other oral anticoagulants were used in addition        analyzed with this bias assessment were ultimately included
to warfarin. If data from the same patients were published in        in our final analysis.
multiple articles (i.e. duplicate data), data from the most
recent publication were used, unless the data were not
                                                                     Data synthesis and analysis
extractable. The title and abstract review, and full-text review,
were performed in duplicate by two reviewers (A.E.W. and             Pooled odds ratios (ORs) and associated 95% confidence
M.P.D.). If there was disagreement in the initial title and          intervals (CIs) were calculated with the inverse variance

 2012 International Society on Thrombosis and Haemostasis
514 A. E. Warkentin et al

method for both major and minor bleeding in patients receiving
                                                                              Results
warfarin vs. those receiving ASA. The analysis was performed
with Review Manager statistical software (REVMAN ver-
                                                                              Study selection
sion 5.0; Copenhagen: The Nordic Cochrane Centre, The
Cochrane Collaboration, 2008).                                                We identified 4442 studies with the electronic search strategy,
   To explore between-study variability and the appropriate-                  and 47 of these were ultimately identified for full-text evaluation
ness of pooling the results from individual studies, the I2-test              (Fig. 1). Of these 47 full-text articles, 40 studies were excluded
for heterogeneity was used. The I2-value expresses the                        for one or more of the following reasons: arms of treatment
percentage of between-study variability that is attributable to               and/or INR range and/or ASA dosage were not appropriate
heterogeneity rather than chance. An I2-value of ‡ 60%                        (36 studies); duplicate data (two studies); and study was not an
suggests significant heterogeneity, indicating that a formal                   RCT (two studies). Therefore, seven studies remained for data
meta-analysis would be of limited value, given the amount of                  extraction and bias assessment. One additional article was
heterogeneity. To further reduce the impact of heterogeneity,                 identified from a manual review of study bibliographies. There
we chose to use a random effects model that would allow for                   were no additional eligible studies identified after communica-
heterogeneity and incorporate its impact in the meta-analysis.                tion with four content experts. Therefore, eight studies were
Funnel plots of effect size against standard error were used to               included in the systematic review and meta-analysis [14–21].
display small-study effects that may arise from reporting bias,                  For the 47 full-text articles screened, there was perfect
differences in methodological quality, or true heterogeneity                  agreement between reviewers regarding inclusion and exclu-
among studies. Prespecified subgroup analyses were performed                   sion, except for one article that was deemed to be ambiguous
to evaluate the bleeding risk: (i) according to age (< 70 and                 by one of the reviewers. For this study, a third party (M.C.) was
‡70 years); and (ii) according to the clinical indication for                 consulted to resolve the ambiguity. This article was ultimately
antithrombotic therapy.                                                       excluded. [22]
   A secondary analysis, specified a priori, was a comparison of
bleeding risk in studies that randomized patients to warfarin
                                                                              Study characteristics
plus ASA vs. warfarin alone (with the same target INR in both
treatment arms, and the same prespecified INR range and ASA                    The main characteristics of the included studies are described in
dosages specified previously).                                                 Table 1. The patient population included atrial fibrillation in

                                                         Recods identified throgh
                                                           database screening
                                                               (n = 4442)

                                                      Duplicates removed (n = 767)

                                                      Records screened (n = 3675)                   Excluded (n = 3628)

                                                                                               Excluded (n = 40)
                                                                                                 - Treatment arms and/or
                                                      Full-text articles assessed for              INR range and/or ASA
                                                             eligibillity (n = 47)                 dosage not suitable (n =
                                                                                                   36)
                                                                                                 - Duplicate data (n = 2)
                                                                                                 - Not an RCT (n = 2)
                        Articles identified through
                          full-text reference list
                               review (n = 1)

                                                        Studies included in meta-
                                                             analysis (n = 8)

Fig. 1. Study selection. ASA, aspirin (acetylsalicylic acid); INR, International Normalized Ratio; RCT, randomized controlled trial.

                                                                                      2012 International Society on Thrombosis and Haemostasis
Comparative bleeding risk: warfarin vs. aspirin 515

Table 1. Study and patient characteristics
                                             Total no.
                                             of patients   Mean age       Female/
                                             (no. of       (years)        male (%)                        Mean         Patients
First author,          Indication for        patients in   (patients in   (patients in           ASA      follow-up    lost to       Withdrawal/
year (study            antithrombotic        arms of       arms of        arms of        INR     dose     duration     follow-up,    dropout*,
acronym)               therapy               interest)     interest)      interest)      range   (mg)     (months)     no. (%)       no. (%)

Huynh, 2001 [14]       UA/NSTEMI             135           67             20/80          2–2.5   80       13           3 (2)         12 (9)
                        and prior
                        CABG
Colli, 2007 [15]       Aortic valve          69            70             14/86          2–3     100      3            NA            6 (8)
 (WoA Epic)             replacement
Gullov, 1999 [16]      Atrial                677 (339)     73.2           38.9/61.1      2–3     300      NAà          0             170 (25.1)
 (AFASAK 2)             fibrillation
Rash, 2007 [17]        Atrial                75            83§            53/47          2–3     300      12           NA            11 (15)
 (WASPO)                fibrillation
Mant, 2007 [18]        Atrial                973           81.5           45.4/54.6      2–3     75       32           8 (0.8)       279 (28.7)
 (BAFTA)                fibrillation
Cleland, 2004 [19]     Chronic heart         279 (180)     64             25/75          2–3     300      27           NA            NA
 (WASH)                 failure
Cokkinos, 2006         Chronic heart         197 (114)     62             11/89          2–3     325      19           NA            3 (2)–
 [20] (HELAS)           failure
Massie, 2009 [21]      Chronic heart         1587(1063)    63             15.2/84.8      2–3.5   162      23           76 (4.8)**    312 (19.7)
 (WATCH)                failure

ASA, acetylsalicylic acid; CABG, coronary artery bypass graft; INR, international normalized ratio; NA, not available; NSTEMI, non-ST-
elevation myocardial infarction; UA, unstable angina. *For withdrawal/dropout, all patients who stopped using the study drug or switched group
were counted. Seventy-five patients were initially enrolled, six of whom were withdrawn from final statistical analysis in the primary article. àThe
study was stopped prematurely, after 42 months. The mean duration of follow-up is not provided in the article describing the study. §This value
represents the mean of the reported median age in the primary article. –Only one of three patients belongs to one of the arms of interest. **Fifty
patients (4.7%) were lost to follow-up in the arms of interest. Two hundred and ten patients (19.8%) were withdrawn from study medication in
the arms of interest.

three studies [16–18], chronic heart failure in three studies [19–            < 90 mmHg or a fall in Hb level below 9.7 g dL)1 in one
21], acute coronary syndrome in one study [14], and heart valve               study [16].
replacement in one study [15]. The study sample sizes varied                     Five studies reported the frequency of minor bleeding
from 69 to 1587 patients. A total of 2948 patients were enrolled              outcomes [14,16,17,19,21]. A formal minor bleeding definition
in the therapeutic arms relevant to the systematic review. ASA                was, however, present in only two studies [14,16]. One study
dosage varied from 75 to 325 mg day)1; target INR ranges                      defined minor bleeding as all non-major bleeding events
were 2.0–2.5 in one study, 2.0–3.0 in six studies, and 2.0–3.5 in             reported by the physician or patient [14], and one study
one study. The mean age varied from 62 to 83 years. Duration                  defined it as all non-major and non-threatening bleeds [16].
of follow-up varied among the included studies.                                  For the purpose of the meta-analysis, we made an attempt to
   Only one study included an additional arm assigned to                      differentiate intracranial from non-intracranial major bleeding
warfarin plus ASA at the same prespecified INR range and                       events. However, data on these events were reported only in
ASA dosage [14]. In four studies, other drugs or no treatment                 four studies [16,18,19,21]. After the authors of the other four
were used in additional arms not considered for the purpose of                studies had been contacted, the relevant information was made
this meta-analysis: placebo or no treatment in two studies                    available for only one study [17].
[19,20]; clopidogrel in one study [21]; and fixed-dose warfarin
and fixed-dose warfarin plus ASA in one study [16].
                                                                              Bias assessment
   Major bleeding was defined in six studies, and differed across
the studies (Table 2). Among the criteria that defined major                   Only one of the eight included studies [20] was identified as low
bleeding, a need for transfusion was present in five studies                   risk of bias among all categories, as shown in Table 3. One
[14,17–19,21]; a fall in hemoglobin (Hb) level of ‡ 2 g dL)1 in               additional study was identified as low risk in all but the random
three studies [14,17,21]; a fatal outcome in three studies                    sequence generation and allocation concealment categories, in
[16,18,21]; intracranial bleeding in two studies [17,18]; a need              which insufficient information was provided to classify these
for surgical intervention in three studies [16,18,21]; an outcome             items as low risk [14]. The remaining six studies all contained
of disability in one study [21]; an outcome of cardiopulmonary                elements of high risk of bias with respect to blinding of
arrest or irreversible damage in one study [16]; and the presence             participants/personnel [15–19,21]. This was because of the
of at least two characteristics including a request for more than             unblinded nature of warfarin delivery in these trials. All studies
three red blood cell units, a systolic blood pressure of                      were identified as low risk in terms of selective reporting, and all

 2012 International Society on Thrombosis and Haemostasis
516 A. E. Warkentin et al

Table 2. Major bleeding definitions in the included studies
                     Components of major bleeding definition

                                                                                                 Leading to
                     Fall in                                                                     CP arrest or        Potentially            Definition
                     Hbof           Transfusion                       Surgey       Disability    irreversible        life-                  not
Study                ‡ 2 g dL)1     required           Intracranial   required     outcome       damage*             threatening    Fatal   available

Huynh, 2001 [14]     4              4                  -              -            -             -                   -              -       -
Colli, 2007 [15]     -              -                  -              -            -             -                   -              -       4
 (WoA Epic)
Gullov, 1999 [16]    -              -                  -              4ৠ         -             4à                  4              4       -
 (AFASAK 2)
Rash, 2007 [17]      4              4                  4              -            -             -                   -              -       -
 (WASPO)
Mant, 2007 [18]      -              4                  4              4            -             -                   -              4       -
 (BAFTA)
Cleland, 2004        -              4                  -              -            -             -                   -              -       -
 [19] (WASH)
Cokkinos, 2006       -              -                  -              -            -             -                   -              -       4
 [20] (HELAS)
Massie, 2009         4              4–                 -              4            4             -                   -              4       -
 [21] (WATCH)

CP, cardio-pulmonary; Hb, hemoglobin. *Myocardial infarction, stroke, blindness. Defined in the presence of at least two of the following:
> 3 U of red blood cells required; systolic blood pressure of < 90 mmHg; Hb < 6 mmol L)1 (9.7 g dL)1). àBleeding event defined as Ôlife-
threateningÕ. §Surgery or angiographic intervention required. –At least two units of packed red blood cells or whole blood.

Table 3. Study bias assessment
                                        Random                                   Blinding of          Blinding of          Incomplete
First author, year                      sequence             Allocation          participants/        outcome              outcome              Selective
(study acronym)                         generation           concealment         personnel            assessment           data                 reporting

Huynh, 2001 [14]                        Unclear risk         Unclear risk        Low risk             Low risk             Low risk             Low   risk
Colli, 2007 [15] (WoA Epic)             Unclear risk         Unclear risk        High risk            Unclear risk         Unclear risk         Low   risk
Gullov, 1999 [16] (AFASAK 2)            Low risk             Low risk            High risk            Low risk             Unclear risk         Low   risk
Rash, 2007 [17] (WASPO)                 Low risk             Low risk            High risk            Unclear risk         Unclear risk         Low   risk
Mant, 2007 [18] (BAFTA)                 Low risk             Low risk            High risk            Low risk             Unclear risk         Low   risk
Cleland, 2004 [19] (WASH)               Low risk             Low risk            High risk            Low risk             Unclear risk         Low   risk
Cokkinos, 2006 [20] (HELAS)             Low risk             Low risk            Low risk             Low risk             Low risk             Low   risk
Massie, 2009 [21] (WATCH)               Low risk             Low risk            High risk            Low risk             Unclear risk         Low   risk

but two studies [14,15] were low risk for selection biases. Owing               A distinction between intracranial hemorrhage and non-
to conservative estimation of risk in incomplete outcome data,               intracranial major bleeding events was possible for only five
six of eight studies [15–19,21] were classified as unclear risk.              studies [16–19,21] (2630 patients). The pooled OR for
Fully detailed bias assessments can be made available upon                   intracranial hemorrhage in patients treated with warfarin vs.
request to the authors.                                                      ASA was 1.64 (95% CI 0.71–3.78) (Fig. 4), and that for
                                                                             extracranial major bleeding was 1.03 (95% CI 0.61–1.75)
                                                                             (Fig. 5).
Data synthesis
                                                                                A sensitivity analysis was performed in two groups of trials
Major bleeding occurred in 69 of 1455 (4.7%; 95% CI 3.8–6.0)                 according to ASA dosage, in order to assess the robustness of
patients treated with warfarin and in 54 of 1449 (3.7%;                      the results. Considering only the RCTs that used ASA dosages
95% CI 2.9–4.8) patients treated with ASA (Fig. 2). The                      of < 300 mg [14,15,18,21], the pooled OR for major bleeding
pooled OR for major bleeding showed a non-statistically                      confirmed a non-statistically significant increased risk for
significant increase in major bleeding for patients receiving                 patients receiving warfarin vs. those receiving ASA (OR 1.25;
warfarin vs. those receiving ASA (OR 1.27; 95% CI 0.83–                      95% CI 0.83–1.86). A similar result for the same comparison
1.94).                                                                       was also obtained with consideration of the RCTs that used
   The pooled OR for minor bleeding events (data available for               ASA dosages equal to or greater than 300 mg [16,17,19,20]
five studies [14,16,17,19,21], 1748 patients) showed a signifi-                (pooled OR 1.45; 95% CI 0.31–6.88).
cantly increased risk in patients treated with warfarin vs. ASA                 Funnel plots of effects size against standard error were
(OR 1.50; 95% CI 1.13–2.00) (Fig. 3).                                        created for major and minor bleeding. The funnel plots

                                                                                        2012 International Society on Thrombosis and Haemostasis
Comparative bleeding risk: warfarin vs. aspirin 517

                                         Warfarin       ASA                               Odds ratio                      Odds ratio
        Study or subgroup               Events Total Events Total          Weight    IV, Random, 95% Cl              IV, Random, 95% Cl
        Huynh et al, 2001 [14]               1    45      0   46            1.7%        3.13 [0.12–79.00]
        Colli et al, 2007 [15]               3    34      1   35            3.3%        3.29 [0.32–33.31]
        Gullov et al,1999 [16]               4 170        5 169             9.6%         0.79 [0.21–3.00]
        Rash et al, 2007 [17]                0    36      3   39            2.0%         0.14 [0.01–2.87]
        Mant et al, 2007 [18]              25 488       25 485             40.1%         0.99 [0.56–1.76]
        Cleland et al, 2004 [19]             4    89      1   91            3.6%       4.24 [0.46– 38.66]
        Cokkinos et al, 2006 [20]            4    53      0   61            2.1%     11.18 [0.59–212.70]
        Massie et al, 2009 [21]            28 540       19 523             37.6%         1.45 [0.80–2.63]

        Total (95% Cl)                          1455          1449 100.0%                 1.27 [0.83–1.94]
        Total events                        69           54
        Heterogeneity: Tau2 = 0.03, Chi2 = 7.61, d.f. = 7 (P = 0.37); l 2 = 8%
        Test for overall effect: Z = 1.09 (P = 0.28)                                                       0.005      0.1    1     10            200
                                                                                                                         ASA Warfarin

Fig. 2. Risk of major bleeding in patients receiving warfarin or aspirin (acetylsalicylic acid) (ASA). CI, confidence interval; d.f., degrees of freedom; IV,
inverse variance.

                                        Warfarin       ASA                            Odds ratio                         Odds ratio
        Study or subgroup              Events Total Events Total          Weight IV, Random, 95% Cl                 IV, Random, 95% Cl
        Huynh et al, 2001 [14]            10     45      2   46            3.2%     6.29 [1.29–30.57]
        Gullov et al, 1999 [16]           42 170       26   169           22.7%      1.80 [1.05–3.11]
        Rash et al, 2007 [17]               6    36      4   39            4.3%      1.75 [0.45–6.79]
        Cleland et al, 2004 [19]          15     89    12    91           11.0%      1.33 [0.59–3.04]
        Massie et al, 2009 [21]          155 540      123   523           58.9%      1.31 [0.99–1.72]

        Total (95% Cl)                          880            868 100.0%               1.50 [1.13–2.00]
        Total events                     228           167
        Heterogeneity: Tau2 = 0.02; Chi2 = 4.57, d.f. = 4 (P = 0.33); l 2 = 12%
                                                                                                          0.01      0.1       1       10          100
        Test for overall effect: Z = 2.79 (P = 0.005)
                                                                                                                          ASA Warfarin

Fig. 3. Risk of minor bleeding in patients receiving warfarin or aspirin (acetylsalicylic acid) (ASA). CI, confidence interval; d.f., degrees of freedom; IV,
inverse variance.

                                       Warfarin       ASA                             Odds ratio                         Odds ratio
        Study or subgroup             Events Total Events Total           Weight IV, Random, 95% Cl                 IV, Random, 95% Cl
        Gullov et al, 1999 [16]            2 170        1 169             12.0%     2.00 [0.18–22.27]
        Rash et al, 2007 [17]              0    36      0   39                          Not estimable
        Mant et al, 2007 [18]              7 488        5 485             52.1%      1.40 [0.44–4.43]
        Cleland et al, 2004 [19]           0    89      0   91                          Not estimable
        Massie et al, 2009 [21]            6 540        3 523             35.9%      1.95 [0.48–7.83]

        Total (95% Cl)                         1323           1307 100.0%               1.64 [0.71–3.78]
        Total events                       15              9
        Heterogeneity: Tau2 = 0.00; Chi2 = 0.16, d.f. = 2 (P = 0.92); l 2 = 0%
                                                                                                          0.01      0.1      1       10           100
        Test for overall effect: Z = 1.17 (P = 0.24)
                                                                                                                          ASA Warfarin

Fig. 4. Risk of intracranial hemorrhage in patients receiving warfarin or aspirin (acetylsalicylic acid) (ASA). CI, confidence interval; d.f., degrees of
freedom; IV, variance.

                                      Warfarin       ASA                              Odds ratio                         Odds ratio
        Study or subgroup            Events Total Events         Total    Weight IV, Random, 95% Cl                 IV, Random, 95% Cl
        Gullov et al, 1999 [16]           2 170        4          169      8.7%      0.49 [0.09–2.72]
        Rash et al, 2007 [17]             0    36      3           39      3.0%      0.14 [0.01–2.87]
        Mant et al, 2007 [18]           18 488        20          485     41.7%      0.89 [0.47–1.70]
        Cleland et al, 2004 [19]          4    89      1           91      5.4%     4.24 [0.46–38.66]
        Massie et al, 2009 [21]         22 540        16          523     41.2%      1.35 [0.70–2.59]

        Total (95% Cl)                        1323           1307 100.0%                 1.03 [0.61–1.75]
        Total events                      46            44
        Heterogeneity: Tau2 = 0.06; Chi2 = 4.78, d.f. = 4 (P = 0.31); l 2 = 16%
        Test for overall effect: Z = 0.12 (P = 0.91)                                                      0.01      0.1       1      10           100
                                                                                                                           ASA Warfarin

Fig. 5. Risk of extracranial major bleeding in patients receiving warfarin or aspirin (acetylsalicylic acid) (ASA). CI, confidence interval; d.f., degrees of
freedom; IV, inverse variance.

 2012 International Society on Thrombosis and Haemostasis
518 A. E. Warkentin et al

appeared symmetric, suggesting the absence of publication            able to an increased risk of intracranial bleeding. The
bias or other reasons for small-study effects (Figs S1 and           frequency of extracranial major bleeding was equivalent
S2).                                                                 between warfarin and ASA.
   Two prespecified subgroup analyses were performed                     The results of the secondary analysis based on age deserve
according to: (i) the mean age of the study participants;            some consideration. We found a tendency for a higher risk of
and (ii) the clinical indication for antithrombotic therapy.         bleeding related to warfarin than to ASA in patients younger
With regard to the secondary analysis based on patient age,          than 70 years (OR 1.71; 95% CI 0.98–2.98), but not in those
the pooled ORs for major bleeding for patients receiving             older than 70 years (OR 0.96; 95% CI 0.58–1.59). These
warfarin vs. ASA in studies that included patients with a            findings coming from secondary analysis should be considered
mean age of < 70 years (four studies [14,19–21], 1492                with caution. Indeed, they may be attributable to chance, given
patients) or ‡ 70 years (four studies [15–18], 1456 patients)        the small number of studies included in this secondary analysis.
were 1.71 (95% CI 0.98–2.98) and 0.96 (95% CI 0.58–1.59),            Age has been described previously as a risk factor for bleeding
respectively. With regard to the secondary analysis based on         in both patients treated with warfarin and those treated with
the clinical indication for antithrombotic therapy, and the          ASA [3,4,8]. Additional data are needed to confirm or refute
risk of major bleeding related to warfarin vs. ASA in studies        this finding.
that enrolled patients with atrial fibrillation (three studies           This review does have limitations. First, there were
[16–18], 1387 patients) or with congestive heart failure (three      relatively few RCTs comparing ASA, 50–650 mg day)1, and
studies [19–21], 1357 patients), the pooled ORs were 0.91            warfarin within an INR range of 2.0–3.5. Four of these
(95% CI 0.54–1.52) and 2.08 (95% CI 0.81–5.36), respec-              studies were prematurely interrupted, because of slow enroll-
tively.                                                              ment of patients in all three congestive heart failure studies
   Finally, the bleeding risk of warfarin plus ASA vs. warfarin      [19–21], and for ethical reasons in the AFASAK-2 trial [16].
alone was only evaluated in one study that met our inclusion         We excluded many older RCTs because they did not fulfill the
criteria [14]. In this study two of 44 (5%; 95% CI 0–16)             prespecified criteria including the use of an INR target range
patients receiving warfarin and ASA experienced an episode of        of 2.0–3.5. These issues resulted in fewer patients being
major bleeding, as compared with one of 45 (2%; 95% CI 0–            available for the meta-analysis than we would have hoped, but
13) patients receiving warfarin (OR 2.10; 95% CI 0.18–23.98).        we are confident that we identified all eligible studies. We
For the same comparison, minor bleeding events were expe-            excluded observational studies, because we wished to present
rienced by nine of 44 (20%; 95% CI 11–35) patients and 10 of         the most accurate estimate of bleeding risks by using a direct
45 (22%; 95% CI 12–36) patients, respectively (OR 0.90;              comparison of warfarin and ASA. Second, the definition for
95% CI 0.33–2.48).                                                   major bleeding differed between studies. This heterogeneity,
                                                                     which limits our ability to provide reliable estimates of
                                                                     bleeding risk, argues strongly for the use of a standardized
Discussion
                                                                     definition of major bleeding. The ISTH has established a
This systematic review provides a pooled analysis of the risk of     definition for major bleeding in both medical and surgical
bleeding in RCTs that directly compared warfarin with ASA            patients [23,24]; however, all of the studies included in this
using the INR and ASA dose ranges commonly used and                  systematic review were initiated before those definitions were
recommended by recent clinical practice guidelines [9–11].           published. We considered major bleeding and other types of
Interestingly, our meta-analysis showed a non-statistically          bleeding as they were defined in each study and the meta-
significant increased risk of major bleeding in patients treated      analysis compared the ORs calculated within each study, thus
with warfarin as compared with those treated with ASA                allowing for a meaningful pooled estimate. This approach was
(OR 1.27; 95% CI 0.83–1.94). This observation may have               also utilized in a similar meta-analysis [12]. Third, we were
clinical significance, insofar as clinicians will frequently switch   particularly interested in intracranial hemorrhage, because of
patients with a perceived increased risk of bleeding from            its high case-fatality rate [25] and its unique association with
warfarin therapy to ASA; the findings of this meta-analysis,          antithrombotic therapies [26,27]. However, separate data for
along with considerations of the relative efficacy of ASA and         intracranial and non-intracranial hemorrhage were available
warfarin, should be considered by clinicians when balancing          for only five studies. Fourth, the risk of bleeding may vary
the risks and benefits of warfarin or ASA treatment.                  across different vitamin K antagonists, as a possible conse-
   The results of our analysis of intracranial hemorrhage            quence of significant differences in half-life. For this reason,
and non-intracranial major bleeding should be considered             we decided to focus on warfarin only, even though we
with caution, as data were available for only five studies,           acknowledge that the results may be less generalizable.
and only three of these studies [16,18,21] had estimable ORs         However, our choice was made in order to provide more
for intracranial hemorrhage, owing to the lack of reported           accurate and precise results. Finally, the patients enrolled in
bleeding events in the other two studies [17,19]. In addition,       our analysis were selected because they were participants in
the CIs of the pooled ORs are wide. However, the results             RCTs. Such patients probably have a lower risk of bleeding
seem to suggest that the trend towards increased major               or other complications than unselected patients in the
bleeding in patients receiving warfarin was entirely attribut-       community.

                                                                            2012 International Society on Thrombosis and Haemostasis
Comparative bleeding risk: warfarin vs. aspirin 519

   As noted previously, our results may be useful as a guide       authors. Any queries (other than missing material) should be
for clinical practice. Clinicians frequently switch patients       directed to the corresponding author for the article.
from warfarin to ASA as a result of a perception of
increased risk of bleeding; this switch is made on the
                                                                   References
assumption that ASA is associated with a lower risk of
hemorrhage. Switching from warfarin to ASA has been                 1 Ost D, Tepper J, Mihara H, Lander O, Heinzer R, Fein A. Duration of
shown to be associated with reduced efficacy in many                   anticoagulation following venous thromboembolism: a meta-analysis.
                                                                      JAMA 2005; 294: 706–15.
clinical situations, particularly in those patients with atrial     2 Atrial Fibrillation Investigators. Risk factors for stroke and efficacy of
fibrillation and additional risk factors for ischemic stroke           antithrombotic therapy in atrial fibrillation: analysis of pooled data
[28]. The inability to demonstrate any difference in extra-           from five randomized controlled trials. Arch Intern Med 1994; 154:
cranial major bleeding (OR 1.03; 95% CI 0.61–1.75) and the            1449–57.
failure to identify a reduced risk of bleeding in those             3 The Stroke Prevention in Atrial Fibrillation Investigators. Bleeding
                                                                      during antithrombotic therapy in patients with atrial fibrillation. Arch
patients over 70 years of age (OR 0.96; 95% CI 0.58–1.59)             Intern Med 1996; 156: 409–16.
suggest that the practice of switching from warfarin to ASA         4 Palareti G, Leali N, Coccheri S, Poggi M, Manotti C, DÕAngelo A,
because ASA is less likely to be associated with bleeding             Pengo V, Erba N, Moia M, Ciavarella N, Devoto G, Berrettini M,
should be re-evaluated.                                               Musolesi S. Bleeding complication of anticoagulant treatment: an
   In conclusion, this systematic review and meta-analysis            inception-cohort, prospective collaborative study (ISCOAT). Lancet
                                                                      1996; 348: 423–8.
comparing the bleeding risk associated with warfarin vs. ASA        5 Abdelhafiz AH, Wheeldon NM. Results of an open-label, prospective
in RCTs did not show a significant difference in major bleeding        study of anticoagulant therapy for atrial fibrillation in an outpatient
risk between these two agents. Further study of this observa-         anticoagulation clinic. Clin Ther 2004; 26: 1470–8.
tion is warranted, given this finding.                               6 Gulløv AL, Koefoed BG, Petersen P, Pedersen TS, Andersen ED,
                                                                      Godtfredsen J, Boysen G. Fixed minidose warfarin and aspirin alone
                                                                      and in combination vs adjusted-dose warfarin for stroke prevention
Acknowledgements                                                      in atrial fibrillation: second Copenhagen Atrial Fibrillation,
                                                                      Aspirin, and Anticoagulation Study. Arch Intern Med 1998; 158: 1513–
J. McKinnell acted in a guidance capacity for how to properly         21.
search the databases utilized in this investigation. She is a       7 Ng W, Wong WM, Chen WH, Tse HF, Lee PY, Lai KC, Li SW, Ng
professional librarian affiliated with the McMaster Health              M, Lam KF, Cheng X, Lau CP. Incidence and predictors of upper
                                                                      gastrointestinal bleeding in patients receiving low-dose aspirin for
Science Library (McMaster University).                                secondary prevention of cardiovascular events in patients with coro-
                                                                      nary artery disease. World J Gastroenterol 2006; 12: 2923–7.
                                                                    8 Sørensen R, Hansen ML, Abildstrom SZ, Hvelplund A, Andersson C,
Disclosure of Conflict of Interests                                   Jørgensen C, Madsen JK, Hansen PR, Køber L, Torp-Pedersen C,
M. Crowther has sat on advisory boards for Leo Pharma,                Gislason GH. Risk of bleeding in patients with acute myocardial
                                                                      infarction treated with different combinations of aspirin, clopidogrel,
Pfizer, Bayer, BI, Alexion, and Artisan, has prepared educa-           and vitamin K antagonists in Denmark: a retrospective analysis of
tional materials for Pfizer, Octapharm and CSL Behring, has            nationwide registry data. Lancet 2009; 374: 1947–8.
provided expert testimony for Bayer, and holds a Career             9 Fuster V, Rydén LE, Cannom DS, Crijns HJ, Curtis AB, Ellenbogen
Investigator award from the Heart and Stroke Foundation of            KA, Halperin JL, Le Heuzey JY, Kay GN, Lowe JE, Olsson SB,
Ontario, and the Leo Pharma Chair in Thromboembolism                  Prystowsky EN, Tamargo JL, Wann S, Smith SC Jr, Jacobs AK,
                                                                      Adams CD, Anderson JL, Antman EM, Halperin JL et al. ACC/
Research at McMaster University. M. Crowther and F. A.
                                                                      AHA/ESC 2006 Guidelines for the Management of Patients with
Spencer are Career Investigators of the Heart and Stroke              Atrial Fibrillation: a report of the American College of Cardiology/
Foundation of Ontario. W. Lim is the recipient of the EJ              American Heart Association Task Force on Practice Guidelines and
Moran Campbell Internal Career Award (McMaster Univer-                the European Society of Cardiology Committee for Practice Guide-
sity). F. A. Spencer is affiliated with the Thrombosis and Ath-        lines (Writing Committee to Revise the 2001 Guidelines for the
                                                                      Management of Patients with Atrial Fibrillation): developed in col-
erosclerosis Research Institute (McMaster University). The            laboration with the European Heart Rhythm Association and the
other authors state that they have no conflict of interest.            Heart Rhythm Society. Circulation 2006; 114: e257–354.
                                                                   10 Salem DN, OÕGara PT, Madias C, Pauker SG. Valvular and structural
                                                                      heart disease: American College of Chest Physicians Evidence-Based
Supporting information                                                Clinical Practice Guidelines (8th Edition). Chest 2008; 133: 593S–629S.
                                                                   11 Becker RC, Meade TW, Berger PB, Ezekowitz M, OÕConnor CM,
Additional Supporting Information may be found in the online
                                                                      Vorchheimer DA, Guyatt GH, Mark DB, Harrington RA. The pri-
version of this article:                                              mary and secondary prevention of coronary artery disease: American
Figure S1. Funnel plot for assessment of publication bias for         College of Chest Physicians Evidence-Based Clinical Practice Guide-
major bleeding.                                                       lines (8th Edition). Chest 2008; 133: 776S–814S.
Figure S2. Funnel plot for assessment of publication bias for      12 Dentali F, Douketis JD, Lim W, Crowther M. Combined aspirin–oral
                                                                      anticoagulant therapy compared with oral anticoagulant therapy alone
minor bleeding.
                                                                      among patients at risk for cardiovascular disease: a meta-analysis of
Appendix S1. Supplemental materials.                                  randomized trials. Arch Intern Med 2007; 167: 117–24.
Please note: Wiley-Blackwell are not responsible for the content   13 Higgins JPT, Altman DG, Sterne JAC (eds). Chapter 8: assessing risk
or functionality of any supporting materials supplied by the          of bias in included studies. In: Higgins JPT, Green S (eds). Cochrane

 2012 International Society on Thrombosis and Haemostasis
520 A. E. Warkentin et al

     Handbook for Systematic Reviews of Interventions, Version 5.1.0 (up-        21 Massie BM, Collins JF, Ammon SE, Armstrong PW, Cleland JGF,
     dated March 2011). The Cochrane Collaboration, 2011. Available                 Ezekowitz M, Jafri SM, Krol WF, OÕConnor CM, Schulman KA, Teo
     from http://www.cochrane-handbook.org.                                         K, Warren SR, for the WATCH Trial Investigators. Randomized trial
14   Huynh T, Théroux P, Bogaty P, Nasmith J, Solymoss S. Aspirin,                 of warfarin, aspirin, and clopidogrel in patients with chronic heart
     warfarin, or the combination for secondary prevention of coronary              failure: the Warfarin and Antiplatelet Therapy in Chronic Heart
     events in patients with acute coronary syndromes and prior coronary            Failure (WATCH) trial. Circulation 2009; 119: 1616–24.
     artery bypass surgery. Circulation 2001; 103: 3069–74.                      22 Powers PJ, Gent M, Jay RM, Julian DH, Turpie AG, Levine M, Hirsh
15   Colli A, Mestres CA, Castella M, Gherli T. Comparing warfarin to               J. A randomized trial of less intense postoperative warfarin or aspirin
     aspirin (WoA) after aortic valve replacement with the St Jude Medical          therapy in the prevention of venous thromboembolism after surgery
     EpicTM heart valve bioprosthesis: results of the WoA Epic Pilot Trial.         for fractured hip. Arch Intern Med 1989; 149: 771–4.
     J Heart Valve Dis 2007; 16: 667–71.                                         23 Schulman S, Kearon C. Definition of major bleeding in clinical
16   Gulløv AL, Koefoed BG, Petersen P. Bleeding during warfarin and                investigations of antihemostatic medicinal products in non-surgical
     aspirin therapy in patients with atrial fibrillation: the AFASAK 2              patients. J Thromb Haemost 2005; 3: 692–4.
     Study. Arch Intern Med 1999; 159: 1322–8.                                   24 Schulman S, Angeras U, Bergqvist D, Eriksson B, Lassen MR, Fisher
17   Rash A, Downes T, Portner R, Yeo WW, Morgan N, Channer KS. A                   W. Definition of major bleeding in clinical investigations of antihe-
     randomised controlled trial of warfarin versus aspirin for stroke pre-         mostatic medicinal products in surgical patients. J Thromb Haemost
     vention in octogenarians with atrial fibrillation (WASPO). Age Ageing           2010; 8: 202–4.
     2007; 36: 151–6.                                                            25 van Asch CJ, Luitse MJ, Rinkel GJ, van der Tweel I, Algra A, Klijn
18   Mant J, Hobbs FD, Fletcher K, Roalfe A, Fitzmaurice D, Lip GYH,                CJ. Incidence, case fatality, and functional outcome of intracerebral
     Murray E, on behalf of the BAFTA investigators and the Midland                 haemorrhage over time, according to age, sex, and ethnic origin: a
     Research Practices Network (MidReC). Warfarin versus aspirin for               systematic review and meta-analysis. Lancet Neurol 2010; 9: 167–76.
     stroke prevention in an elderly community population with atrial            26 Appelboam R, Thomas EO. Warfarin and intracranial haemorrhage.
     fibrillation (the Birmingham Atrial Fibrillation Treatment of the Aged          Blood Rev 2009; 23: 1–9.
     Study, BAFTA): a randomised controlled trial. Lancet 2007; 370: 493–        27 Toyoda K, Yasaka M, Nagata K, Nagao T, Gotoh J, Sakamoto T,
     503.                                                                           Uchiyama S, Minematsu K. Antithrombotic therapy influences loca-
19   Cleland JGF, Findlay I, Jafri S, Sutton G, Falk R, Bulpitt C, Prentice C,      tion, enlargement, and mortality from intracerebral hemorrhage. The
     Ford I, Trainer A, Poole-Wilson PA. The Warfarin/Aspirin Study in              Bleeding with Antithrombotic Therapy (BAT) retrospective study.
     Heart failure (WASH): a randomized trial comparing antithrombotic              Cerebrovasc Dis 2009; 27: 151–9.
     strategies for patients with heart failure. Am Heart J 2004; 148: 157–64.   28 Singer DE, Albers GW, Dalen JE, Fang MC, Go AS, Halperin JL, Lip
20   Cokkinos DV, Haralabopoulos GC, Kostis JB, Toutouzas PK, for the               GY, Manning WJ. Antithrombotic therapy in atrial fibrillation:
     HELAS investigators. Efficacy of antithrombotic therapy in chronic               American College of Chest Physicians Evidence-Based Clinical Prac-
     heart failure: the HELAS study. Eur J Heart Fail 2006; 8: 428–32.              tice Guidelines (8th Edition). Chest 2008; 133: 546S–92S.

                                                                                         2012 International Society on Thrombosis and Haemostasis
You can also read