European Journal of General Practice, 2014; Early Online: 1–5
Background Paper
Eur J Gen Pract Downloaded from informahealthcare.com by Biblioteka Uniwersytetu Warszawskiego on 01/23/15 For personal use only.
New oral anticoagulants for nonvalvular atrial fibrillation in the elderly: Limited applicability in primary care
Wim Opstelten, Maureen van den Donk, Ton Kuijpers & Jako Burgers Dutch College of General Practitioners, Utrecht, the Netherlands
KEY MESSAGE: • New oral anticoagulants (NOACs) are alternatives of vitamin K antagonists (VKAs) because of their effectiveness, lower risk of harms, and ease of use • The effectiveness and safety of NOACs has been assessed in randomized controlled trials, but the experience with these drugs is still limited outside research context. • As long as additional data about the benefits and harms of NOACs in primary care setting are lacking, general practitioners should be cautious in prescribing these drugs, especially in elderly patients with multiple comorbidity.
ABSTRACT Background: Based on the results from randomized controlled trials (RCTs), new oral anticoagulants (NOACs) seem attractive alternatives to vitamin K antagonists (VKAs) because of their effectiveness, safety, and ease of use. However, the use of NOACs in unselected elderly patients with atrial fibrillation (AF) in primary care is arguable. Objectives: To assess the evidence for the effectiveness and safety of NOACs compared with VKAs in elderly patients with nonvalvular AF in primary care. Methods: Starting from the meta-analysis of Ruff et al. (Lancet 2014;383:955–62), we used the GRADE-approach to make a transparent and explicit judgement of the quality of evidence. Results: The meta-analysis reviewed four non-inferiority RCTs, including 58 634 AF patients with an average age of 70–73 years. Inconsistency of results, indirectness of evidence, and imprecision of risk reductions resulted in downgrading of the quality of evidence available from these studies. The quality of evidence for a decrease in all-cause stroke and systemic embolism (RR: 0.81; 95%CI: 0.73–0.91) for elderly patients using NOACs compared to VKAs in routine primary care was low. The quality of evidence for a lower risk for haemorrhagic stroke (RR: 0.49; 95% CI: 0.38–0.64) and for a lower risk of intracranial bleeding (RR: 0.48; 95% CI: 0.39–0.59) was moderate. Conclusion: There is uncertainty about effectiveness and safety of NOACs in unselected elderly patients with AF in primary care. Therefore, the balance between benefit and harm is still unclear. For this reason, routine use of NOACs is not recommended in elderly patients in primary care.
INTRODUCTION Atrial fibrillation (AF) is a common disorder with an increasing prevalence in our ageing society. The prevalence in patients older than 65 years is over 8% and the lifetime risk to develop AF at the age of 55 years amounts 24% in men and 22% in women (1). AF increases the risk of ischaemic stroke, which can be determined using the CHA2DS2-VASc score (2). Almost all AF patients older than 65 years are eligible for treatment with oral anticoagulants. Until recently, vitamin K
antagonists (VKAs) were available as potent anticoagulant drugs, lowering the risk of an ischemic stroke by approximately 60% (3). The major side effect of these drugs, however, is the increased risk of bleeding. In addition, VKAs have a narrow therapeutic window and food and drugs influence their anticoagulant effect, which necessitates frequent coagulation monitoring using the international normalized ratio (INR) and dose adjustment. Moreover, many patients perceive the repeated venepuncture as a burden. Recently, new oral anticoagulants (NOACs) became available for treatment
Correspondence: W. Opstelten, Dutch College of General Practitioners, PO Box 3231, 3502 GE Utrecht, the Netherlands. Tel: ⫹ 31 30 282 35 00. Fax: ⫹ 31 30 282 35 01. Email: [email protected] (Received 28 February 2014; accepted 15 November 2014) ISSN 1381-4788 print/ISSN 1751-1402 online © 2014 Informa Healthcare DOI: 10.3109/13814788.2014.989986
Eur J Gen Pract Downloaded from informahealthcare.com by Biblioteka Uniwersytetu Warszawskiego on 01/23/15 For personal use only.
2
W. Opstelten et al.
of nonvalvular AF. These drugs seem promising, as clinical trials showed similar efficacy as VKAs (i.e., warfarin), fewer side effects so far, and regular INR monitoring is not needed. However, whether or not use of NOACs is justified for the broad spectrum of unselected elderly patients in primary care is arguable. During the recent revision of the AF guideline of the Dutch College of General Practitioners (DCGP), we evaluated the evidence for the efficacy and safety of NOACs in primary care according to the Grading of recommendations assessment, development and evaluation (GRADE) methodology, and, subsequently, formulated a recommendation for the use of these drugs (4,16).
MATERIAL AND METHODS The meta-analysis of Ruff et al. was used as a basis for determining the impact of NOACs on elderly patients with AF (5). This meta-analysis was not funded by industry, although individual authors reported potential conflicts of interest. It included four large phase III randomized controlled trials (RCTs): the randomized evaluation of long-term anticoagulation therapy (RE-LY; dabigatran), the rivaroxaban once daily oral direct factor Xa inhibition compared with vitamin K antagonism for prevention of stroke and embolism trial in atrial fibrillation (ROCKET AF), the apixaban for reduction of stroke and other thromboembolic events in atrial fibrillation (ARISTOTLE) trial, and the effective anticoagulation with factor Xa next generation in atrial fibrillation—thrombolysis in myocardial infarction 48 (ENGAGE AF-TIMI 48) trial (edoxaban) (6–9). In RE-LY and ENGAGE AF-TIMI 48, two doses of dabigatran and edoxaban, respectively, were compared with warfarin. The meta-analysis was undertaken with both higher doses (dabigatran 150 mg twice daily and edoxaban 60 mg once daily) combined with the single doses studied in ROCKET AF (rivaroxaban 20 mg once daily) and ARISTOTLE (apixaban 5 mg twice daily). In total, 71 638 patients were included. However, in accordance with the Ruff et al. meta-analysis, we deleted the results with lower doses NOACs (i.e. dabigatran 110 mg, 6015 patients and edoxaban 30 mg, 7034 patients), yielding a
total of 58 634 patients, which were included in these analyses. These patients had an average age ranging from 70 to 73 years and a median follow-up time ranging from 1.8 to 2.8 years per study. From the individual publications, it could not be derived whether patients were (partly) recruited from primary care or whether the investigators included only patients who were on treatment for cardiology in outpatient clinics. The worldwide increasingly adopted GRADE methodology has been developed to make a clear separation between the quality of evidence and the strength of recommendations. It comprises a transparent process of moving from evidence to recommendations by using explicit, comprehensive criteria for downgrading or upgrading quality of evidence ratings. In the GRADE approach to the quality of evidence, a meta-analysis of RCTs starts from highquality evidence. The following factors, which might decrease the quality of evidence, are being assessed: risk of bias (study limitations), publication bias, inconsistency of results, indirectness of evidence, and imprecision. RESULTS Efficacy and safety Table 1 summarizes the effects on efficacy and safety of NOACs compared with VKAs for six patient outcomes, which are considered relevant by the AF guideline development group of the DCGP in consultation with AF patients. With respect to efficacy, patients randomized to NOACs had a lower risk of all-cause stroke and systemic embolism and of haemorrhagic stroke than those randomized to warfarin. The risk for ischaemic stroke was inconclusive. With respect to safety, assessment of the risk of major bleeding was inconclusive. Patients randomized to NOACs had a lower risk of intracranial haemorrhage and a higher risk of gastrointestinal bleeding than those randomized to warfarin. Risk of bias All included RCTs were non-inferiority studies, which adhered to high standards, including adequate conceal-
Table 1. Efficacy and safety outcomes, from meta-analysis (5). Pooled NOAC (events)
Pooled VKA (events)
Efficacy Stroke or systemic embolic events 911/29 312 1107/29 229 Ischaemic stroke 665/29 292 724/29 221 Haemorrhagic stroke 130/29 292 263/29 221 Safety Major bleeding 1541/29 287 1802/29 211 Intracranial haemorrhage 204/29 287 425/29 211 Gastrointestinal bleeding 751/29 287 591/29 211 Data are n/N; NOAC, new oral anticoagulant; RR, relative risk.
RR (95%CI)
P
Heterogeneity (I2)
0.81 (0.73–0.91) ⬍ 0.0001 0.92 (0.83–1.02)) 0.10 0.49 (0.38–0.64) ⬍ 0.0001
47% 32% 34%
0.86 (0.73–1.00) 0.48 (0.39–0.59) 1.25 (1.01–1.55)
83% 32% 74%
0.06 ⬍ 0.0001 0.043
Design
Risk of bias
Inconsistency
Indirectness
Imprecision
Other considerations
591/29 211 (2%)
751/29 287 (2.6%)
RR 1.25 (1.01–1.55)
RR 0.48 (0.39–0.59)
425/29 211 (1.5%)
RR 0.49 (0.38–0.64)
204/29 287 (0.7%)
263/29 221 (0.9%)
130/29 292 (0.44%)
RR 0.92 (0.83–1.02)
RR 0.86 (0.73–1.00)
724/29 221 (2.5%)
665/29 292 (2.3%)
RR 0.81 (0.73–0.91)
Relative effect (95%CI)
1541/29 287 1802/29 211 (5.3%) (6.2%)
1107/29 229 (3.8%)
VKA
911/29 312 (3.1%)
NOAC
5 more per 1000 (from 0 more to 11 more)
8 fewer per 1000 (from 6 fewer to 9 fewer)
CRITICAL
CRITICAL
Importance
VERY LOW
CRITICAL
MODERATE CRITICAL
CRITICAL
MODERATE CRITICAL
LOW
LOW
Quality
9 fewer per 1000 VERY LOW (from 17 fewer to 0 more)
5 fewer per 1000 (from 3 fewer to 6 fewer)
2 fewer per 1000 (from 4 fewer to 0 more)
7 fewer per 1000 (from 3 fewer to 10 fewer)
Absolute effect (95%CI)
No. of patients
aIn one of the four RCTs (RE-LY), patients were not blinded for the intervention (VKA or NOAC). The outcomes, however, were evaluated blindly. Moreover, in the results of previous placebo-controlled open-label RCTs assessing the effectiveness of warfarin did not differ from the results of double-blinded RCTs. Therefore, the risk of bias is low. bEffectiveness not evaluated in routine primary care. cA RRR of 10% was considered a considerable benefit for this outcome. The 95%CI of the pooled effect estimate overlaps the 0.90 margin. Therefore, the estimate is imprecise. dAll four RCTs have been funded by pharmaceutical companies. The meta-analysis, however, had not been commercially funded, although many authors reported potential conflicts of interest. As there are only four phase III RCTs, a funnel plot was not reported. Three of the RCTs reported a statistically significant increase in adverse events (i.e. increase in gastrointestinal bleeding). eA RRR of 5% was considered a considerable benefit for this outcome. The 95%CI of the pooled effect estimate does not overlap the 0.95 margin. Therefore, the estimate is considered precise. f95%CIs of individual trials do not overlap. Inconsistency among trials (I2 83%) that could not be explained. gAdverse effects not assessed in routine primary care. hRelatively short follow-up (median 2.2 years). iA RRR of 10% was considered a considerable benefit for this outcome. The 95%CI of the pooled effect estimate does not overlap the 0.90 margin. Therefore, the estimate is considered precise. j95%CIs of individual trials do not overlap. Inconsistency among trials (I2 74%) that could not be explained. kA RRR of 10% was considered a considerable benefit for this outcome. The 95%CI of the pooled effect estimate does not include the 0.90 margin of considerable benefit, but does include the 1.10 margin of considerable harm. Therefore, the estimate is imprecise.
All-cause stroke and systemic embolism (follow-up mean 2.2 years; assessed with: clinical assessment) 4 Randomized No serious No serious Seriousb Seriousc Noned trials risk of inconsistency biasa Ischaemic and unspecified stroke (follow-up median 2.2 years; assessed with: clinical assessment) 4 Randomized No serious No serious Seriousb Seriousc Noned trials risk of inconsistency biasa Haemorrhagic stroke (follow-up median 2.2 years; assessed with: clinical assessment) 4 Randomized No serious No serious Seriousb No serious Noned trials risk of inconsistency imprecisione biasa Major bleeding (follow-up median 2.2 years; assessed with: clinical assessment) Seriousg,h Seriousc Noned 4 Randomized No serious Seriousf trials risk of biasa Intracranial bleeding (follow-up median 2.2 years; assessed with: clinical assessment) Seriousg,h No serious Noned 4 Randomized No serious No serious inconsistency imprecisioni trials risk of biasa Gastrointestinal bleeding (follow-up median 2.2 years; assessed with: clinical assessment) 4 Randomized No serious Seriousj Seriousg,h Seriousk Noned trials risk of biasa
No. of studies
Quality assessment
Table 2. Summary of the application of the different GRADE-criteria on the critical outcomes of NOACs resulting in a rating of the evidence (5).
Eur J Gen Pract Downloaded from informahealthcare.com by Biblioteka Uniwersytetu Warszawskiego on 01/23/15 For personal use only.
New oral anticoagulants 3
Eur J Gen Pract Downloaded from informahealthcare.com by Biblioteka Uniwersytetu Warszawskiego on 01/23/15 For personal use only.
4
W. Opstelten et al.
ment of allocation, blinded assessment of effects, a low rate of missing data, a well-supported non-inferiority margin, and an intention-to-treat analysis. However, only one RCT additionally performed a per-protocol analysis, which is recommended in non-inferiority trials to reduce the chance of falsely declaring non-inferiority (7,10). In one RCT, the patients were not blinded for allocated treatment, which could have introduced bias in the reporting or adjudication of events (6). This was compensated by the implementation of several validated procedures, including blinded evaluation of outcome effects. Moreover, the results of previous placebo controlled open-label RCTs assessing the effectiveness of warfarin did not differ from the results of double-blinded RCTs (11). In summary, we concluded that the included RCTs are at low risk of bias. Publication bias To our knowledge, the meta-analysis of Ruff et al. (5) included all available RCTs. There was no indication that they failed to report studies that showed no effect. Inconsistencies Some inconsistency among studies was found: two RCTs indicated a higher risk of major bleeding in patients on warfarin, whereas the other RCTs could not detect a risk difference (I2 83%) (6–9). Moreover, three RCTs reported an increased risk of gastrointestinal bleeding in patients on NOACs, whereas the fourth RCT indicated a statistically non-significant protection against this complication in patients randomized to NOACs (I2 74%) (6–9). These inconsistencies cannot be explained with the available data. Given that the inconsistency is present within the group of direct factor Xa inhibitors, a difference in drug properties seems an implausible explanation. A difference in patient characteristics does not seem responsible for the inconsistency either: in fact, the inconsistencies were observed between two studies, which included patients with similar baseline risks of stroke (6,8).
particular for the use of NOACs in daily practice. Several factors, such as reduced patient compliance in daily care compared to research context, (temporary) worsening of renal function, and interactions with concomitant medication, may affect the blood level of NOACs and, therefore, its therapeutic effects. Without regular monitoring of coagulation, any deviation beyond the therapeutic range will not be noticed. This may expose patients to inadequate protection against ischaemic strokes or increased risk of adverse events. A recent observational study reported that dabigatran might be as effective and safe in daily practice and research context (15). However, since the study was not randomized, confounding by indication may have biased these results. Imprecision For all relevant outcomes, we considered a relative risk reduction (RRR) of 10% as clinically relevant. However, for haemorrhagic stroke, which is associated with severe impact on the quality of life, we put the limit at 5% (12,13). With respect to these limits, only the risk reductions of haemorrhagic stroke and intracranial haemorrhage were sufficiently precise. Table 2 summarizes the application of the different GRADE-criteria on the critical outcomes of NOACs resulting in a rating of the evidence.
DISCUSSION Bearing in mind all these considerations, the AF guideline development group of the DCGP is reluctant to recommend broadly the use of NOACs in the elderly. For now, we think that these drugs should only be prescribed if the following conditions are met: age younger than 80 years, few comorbid conditions, good renal function (GFR ⬎ 50 ml/min), and good expected drug adherence (16). Using these criteria, GPs could discuss the choice in antithrombotic therapy with their patients according to the principles of shared decision making.
Conclusion Indirectness of evidence The inclusion and exclusion criteria of the RCTs permitted entrance of patients with varying severity of illness. However, it has been shown that many AF patients who use oral anticoagulants for prevention of thromboembolic complications do not fulfil the eligibility criteria of the trials that demonstrated effectiveness and safety of these drugs. In a case-control study, 40% of users of VKAs presenting with bleeding had one or more exclusion criteria for the trials, supporting the use of anticoagulation for their condition (14). Although this may be equally true for both VKAs and NOACs, this has consequences in
NOACs are alternatives of VKAs because of their lower risk of harm and ease of use. Outside research context, however, the experience with NOACs is still limited. We feel additional research in primary care setting should be performed to assess the effectiveness and safety of NOACs before they can be recommended for use in primary care.
ACKNOWLEDGEMENT The authors thank Holger Schünemann, MD, PhD, epidemiologist, for his comment on the manuscript.
New oral anticoagulants Declaration of interest: The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.
Eur J Gen Pract Downloaded from informahealthcare.com by Biblioteka Uniwersytetu Warszawskiego on 01/23/15 For personal use only.
REFERENCES 1. Heeringa J, van der Kuip DA, Hofman A, Kors JA, van Herpen G, Stricker BH, et al. Prevalence, incidence and lifetime risk of atrial fibrillation: The Rotterdam study. Eur Heart J. 2006;27:949–53. 2. Lip GY, Nieuwlaat R, Pisters R, Lane DA, Crijns HJ. Refining clinical risk stratification for predicting stroke and thromboembolism in atrial fibrillation using a novel risk factor-based approach: The Euro heart survey on atrial fibrillation. Chest 2010;137:263–72. 3. Hart RG, Pearce LA, Aguilar MI. Meta-analysis: Antithrombotic therapy to prevent stroke in patients who have nonvalvular atrial fibrillation. Ann Intern Med. 2007;146:857–67. 4. Guyatt GH, Oxman AD, Vist GE, Kunz R, Falck-Ytter Y, Alonso-Coello P, et al. GRADE: An emerging consensus on rating quality of evidence and strength of recommendations. Br Med J. 2008;336:924–6. 5. Ruff CT, Giugliano RP, Braunwald E, Hoffman EB, Deenadayalu N, Ezekowitz MD, et al. Comparison of the efficacy and safety of new oral anticoagulants with warfarin in patients with atrial fibrillation: A meta-analysis of randomised trials. Lancet 2014;383:955–62. 6. Connolly SJ, Ezekowitz MD, Yusuf S, Eikelboom J, Oldgren J, Parekh A, et al. Dabigatran versus warfarin in patients with atrial fibrillation. N Engl J Med. 2009;361:1139–51. 7. Patel MR, Mahaffey KW, Garg J, Pan G, Singer DE, Hacke W, et al. Rivaroxaban versus warfarin in nonvalvular atrial fibrillation. N Engl J Med. 2011;365:883–91. 8. Granger CB, Alexander JH, McMurray JJ, Lopes RD, Hylek EM, Hanna M, et al. Apixaban versus warfarin in patients with atrial fibrillation. N Engl J Med. 2011;365:981–92.
5
9. Giugliano RP, Ruff CT, Braunwald E, Murphy SA, Wiviott SD, Halperin JL, et al. Edoxaban versus warfarin in patients with atrial fibrillation. N Engl J Med. 2013;369:2093–104. 10. Piaggio G, Elbourne DR, Altman DG, Pocock SJ, Evans SJ. Reporting of noninferiority and equivalence randomized trials: An extension of the CONSORT statement. J Am Med Assoc. 2006;295:1152–60. 11. Ezekowitz MD, Connolly S, Parekh A, Reilly PA, Varrone J, Wang S, et al. Rationale and design of RE-LY: Randomized evaluation of long-term anticoagulant therapy, warfarin, compared with dabigatran. Am Heart J. 2009;157:805–10, 810 e1–2. 12. Olesen JB, Lip GY, Hansen ML, Hansen PR, Tolstrup JS, Lindhardsen J, et al. Validation of risk stratification schemes for predicting stroke and thromboembolism in patients with atrial fibrillation: Nationwide cohort study. Br Med J. 2011; 342(31):d124. 13. Friberg L, Rosenqvist M, Lip GY. Net clinical benefit of warfarin in patients with atrial fibrillation: A report from the Swedish atrial fibrillation cohort study. Circulation 2012; 125:2298–307. 14. Levi M, Hovingh GK, Cannegieter SC, Vermeulen M, Buller HR, Rosendaal FR. Bleeding in patients receiving vitamin K antagonists who would have been excluded from trials on which the indication for anticoagulation was based. Blood 2008; 111:4471–6. 15. Larsen TB, Rasmussen LH, Skjoth F, Due KM, Callreus T, Rosenzweig M, et al. Efficacy and safety of dabigatran etexilate and warfarin in ‘real-world’ patients with atrial fibrillation. J Am Coll Cardiol. 2013;61:2264–73. 16. van den Donk M, Opstelten W. Important changes in the Dutch College of General Practitioners (NHG) practice guideline ‘Atrial fibrillation’: Revised practice guideline issued [in Dutch]. Ned Tijdschr Geneeskd. 2013;157:A6697.
Background Paper
Eur J Gen Pract Downloaded from informahealthcare.com by Biblioteka Uniwersytetu Warszawskiego on 01/23/15 For personal use only.
New oral anticoagulants for nonvalvular atrial fibrillation in the elderly: Limited applicability in primary care
Wim Opstelten, Maureen van den Donk, Ton Kuijpers & Jako Burgers Dutch College of General Practitioners, Utrecht, the Netherlands
KEY MESSAGE: • New oral anticoagulants (NOACs) are alternatives of vitamin K antagonists (VKAs) because of their effectiveness, lower risk of harms, and ease of use • The effectiveness and safety of NOACs has been assessed in randomized controlled trials, but the experience with these drugs is still limited outside research context. • As long as additional data about the benefits and harms of NOACs in primary care setting are lacking, general practitioners should be cautious in prescribing these drugs, especially in elderly patients with multiple comorbidity.
ABSTRACT Background: Based on the results from randomized controlled trials (RCTs), new oral anticoagulants (NOACs) seem attractive alternatives to vitamin K antagonists (VKAs) because of their effectiveness, safety, and ease of use. However, the use of NOACs in unselected elderly patients with atrial fibrillation (AF) in primary care is arguable. Objectives: To assess the evidence for the effectiveness and safety of NOACs compared with VKAs in elderly patients with nonvalvular AF in primary care. Methods: Starting from the meta-analysis of Ruff et al. (Lancet 2014;383:955–62), we used the GRADE-approach to make a transparent and explicit judgement of the quality of evidence. Results: The meta-analysis reviewed four non-inferiority RCTs, including 58 634 AF patients with an average age of 70–73 years. Inconsistency of results, indirectness of evidence, and imprecision of risk reductions resulted in downgrading of the quality of evidence available from these studies. The quality of evidence for a decrease in all-cause stroke and systemic embolism (RR: 0.81; 95%CI: 0.73–0.91) for elderly patients using NOACs compared to VKAs in routine primary care was low. The quality of evidence for a lower risk for haemorrhagic stroke (RR: 0.49; 95% CI: 0.38–0.64) and for a lower risk of intracranial bleeding (RR: 0.48; 95% CI: 0.39–0.59) was moderate. Conclusion: There is uncertainty about effectiveness and safety of NOACs in unselected elderly patients with AF in primary care. Therefore, the balance between benefit and harm is still unclear. For this reason, routine use of NOACs is not recommended in elderly patients in primary care.
INTRODUCTION Atrial fibrillation (AF) is a common disorder with an increasing prevalence in our ageing society. The prevalence in patients older than 65 years is over 8% and the lifetime risk to develop AF at the age of 55 years amounts 24% in men and 22% in women (1). AF increases the risk of ischaemic stroke, which can be determined using the CHA2DS2-VASc score (2). Almost all AF patients older than 65 years are eligible for treatment with oral anticoagulants. Until recently, vitamin K
antagonists (VKAs) were available as potent anticoagulant drugs, lowering the risk of an ischemic stroke by approximately 60% (3). The major side effect of these drugs, however, is the increased risk of bleeding. In addition, VKAs have a narrow therapeutic window and food and drugs influence their anticoagulant effect, which necessitates frequent coagulation monitoring using the international normalized ratio (INR) and dose adjustment. Moreover, many patients perceive the repeated venepuncture as a burden. Recently, new oral anticoagulants (NOACs) became available for treatment
Correspondence: W. Opstelten, Dutch College of General Practitioners, PO Box 3231, 3502 GE Utrecht, the Netherlands. Tel: ⫹ 31 30 282 35 00. Fax: ⫹ 31 30 282 35 01. Email: [email protected] (Received 28 February 2014; accepted 15 November 2014) ISSN 1381-4788 print/ISSN 1751-1402 online © 2014 Informa Healthcare DOI: 10.3109/13814788.2014.989986
Eur J Gen Pract Downloaded from informahealthcare.com by Biblioteka Uniwersytetu Warszawskiego on 01/23/15 For personal use only.
2
W. Opstelten et al.
of nonvalvular AF. These drugs seem promising, as clinical trials showed similar efficacy as VKAs (i.e., warfarin), fewer side effects so far, and regular INR monitoring is not needed. However, whether or not use of NOACs is justified for the broad spectrum of unselected elderly patients in primary care is arguable. During the recent revision of the AF guideline of the Dutch College of General Practitioners (DCGP), we evaluated the evidence for the efficacy and safety of NOACs in primary care according to the Grading of recommendations assessment, development and evaluation (GRADE) methodology, and, subsequently, formulated a recommendation for the use of these drugs (4,16).
MATERIAL AND METHODS The meta-analysis of Ruff et al. was used as a basis for determining the impact of NOACs on elderly patients with AF (5). This meta-analysis was not funded by industry, although individual authors reported potential conflicts of interest. It included four large phase III randomized controlled trials (RCTs): the randomized evaluation of long-term anticoagulation therapy (RE-LY; dabigatran), the rivaroxaban once daily oral direct factor Xa inhibition compared with vitamin K antagonism for prevention of stroke and embolism trial in atrial fibrillation (ROCKET AF), the apixaban for reduction of stroke and other thromboembolic events in atrial fibrillation (ARISTOTLE) trial, and the effective anticoagulation with factor Xa next generation in atrial fibrillation—thrombolysis in myocardial infarction 48 (ENGAGE AF-TIMI 48) trial (edoxaban) (6–9). In RE-LY and ENGAGE AF-TIMI 48, two doses of dabigatran and edoxaban, respectively, were compared with warfarin. The meta-analysis was undertaken with both higher doses (dabigatran 150 mg twice daily and edoxaban 60 mg once daily) combined with the single doses studied in ROCKET AF (rivaroxaban 20 mg once daily) and ARISTOTLE (apixaban 5 mg twice daily). In total, 71 638 patients were included. However, in accordance with the Ruff et al. meta-analysis, we deleted the results with lower doses NOACs (i.e. dabigatran 110 mg, 6015 patients and edoxaban 30 mg, 7034 patients), yielding a
total of 58 634 patients, which were included in these analyses. These patients had an average age ranging from 70 to 73 years and a median follow-up time ranging from 1.8 to 2.8 years per study. From the individual publications, it could not be derived whether patients were (partly) recruited from primary care or whether the investigators included only patients who were on treatment for cardiology in outpatient clinics. The worldwide increasingly adopted GRADE methodology has been developed to make a clear separation between the quality of evidence and the strength of recommendations. It comprises a transparent process of moving from evidence to recommendations by using explicit, comprehensive criteria for downgrading or upgrading quality of evidence ratings. In the GRADE approach to the quality of evidence, a meta-analysis of RCTs starts from highquality evidence. The following factors, which might decrease the quality of evidence, are being assessed: risk of bias (study limitations), publication bias, inconsistency of results, indirectness of evidence, and imprecision. RESULTS Efficacy and safety Table 1 summarizes the effects on efficacy and safety of NOACs compared with VKAs for six patient outcomes, which are considered relevant by the AF guideline development group of the DCGP in consultation with AF patients. With respect to efficacy, patients randomized to NOACs had a lower risk of all-cause stroke and systemic embolism and of haemorrhagic stroke than those randomized to warfarin. The risk for ischaemic stroke was inconclusive. With respect to safety, assessment of the risk of major bleeding was inconclusive. Patients randomized to NOACs had a lower risk of intracranial haemorrhage and a higher risk of gastrointestinal bleeding than those randomized to warfarin. Risk of bias All included RCTs were non-inferiority studies, which adhered to high standards, including adequate conceal-
Table 1. Efficacy and safety outcomes, from meta-analysis (5). Pooled NOAC (events)
Pooled VKA (events)
Efficacy Stroke or systemic embolic events 911/29 312 1107/29 229 Ischaemic stroke 665/29 292 724/29 221 Haemorrhagic stroke 130/29 292 263/29 221 Safety Major bleeding 1541/29 287 1802/29 211 Intracranial haemorrhage 204/29 287 425/29 211 Gastrointestinal bleeding 751/29 287 591/29 211 Data are n/N; NOAC, new oral anticoagulant; RR, relative risk.
RR (95%CI)
P
Heterogeneity (I2)
0.81 (0.73–0.91) ⬍ 0.0001 0.92 (0.83–1.02)) 0.10 0.49 (0.38–0.64) ⬍ 0.0001
47% 32% 34%
0.86 (0.73–1.00) 0.48 (0.39–0.59) 1.25 (1.01–1.55)
83% 32% 74%
0.06 ⬍ 0.0001 0.043
Design
Risk of bias
Inconsistency
Indirectness
Imprecision
Other considerations
591/29 211 (2%)
751/29 287 (2.6%)
RR 1.25 (1.01–1.55)
RR 0.48 (0.39–0.59)
425/29 211 (1.5%)
RR 0.49 (0.38–0.64)
204/29 287 (0.7%)
263/29 221 (0.9%)
130/29 292 (0.44%)
RR 0.92 (0.83–1.02)
RR 0.86 (0.73–1.00)
724/29 221 (2.5%)
665/29 292 (2.3%)
RR 0.81 (0.73–0.91)
Relative effect (95%CI)
1541/29 287 1802/29 211 (5.3%) (6.2%)
1107/29 229 (3.8%)
VKA
911/29 312 (3.1%)
NOAC
5 more per 1000 (from 0 more to 11 more)
8 fewer per 1000 (from 6 fewer to 9 fewer)
CRITICAL
CRITICAL
Importance
VERY LOW
CRITICAL
MODERATE CRITICAL
CRITICAL
MODERATE CRITICAL
LOW
LOW
Quality
9 fewer per 1000 VERY LOW (from 17 fewer to 0 more)
5 fewer per 1000 (from 3 fewer to 6 fewer)
2 fewer per 1000 (from 4 fewer to 0 more)
7 fewer per 1000 (from 3 fewer to 10 fewer)
Absolute effect (95%CI)
No. of patients
aIn one of the four RCTs (RE-LY), patients were not blinded for the intervention (VKA or NOAC). The outcomes, however, were evaluated blindly. Moreover, in the results of previous placebo-controlled open-label RCTs assessing the effectiveness of warfarin did not differ from the results of double-blinded RCTs. Therefore, the risk of bias is low. bEffectiveness not evaluated in routine primary care. cA RRR of 10% was considered a considerable benefit for this outcome. The 95%CI of the pooled effect estimate overlaps the 0.90 margin. Therefore, the estimate is imprecise. dAll four RCTs have been funded by pharmaceutical companies. The meta-analysis, however, had not been commercially funded, although many authors reported potential conflicts of interest. As there are only four phase III RCTs, a funnel plot was not reported. Three of the RCTs reported a statistically significant increase in adverse events (i.e. increase in gastrointestinal bleeding). eA RRR of 5% was considered a considerable benefit for this outcome. The 95%CI of the pooled effect estimate does not overlap the 0.95 margin. Therefore, the estimate is considered precise. f95%CIs of individual trials do not overlap. Inconsistency among trials (I2 83%) that could not be explained. gAdverse effects not assessed in routine primary care. hRelatively short follow-up (median 2.2 years). iA RRR of 10% was considered a considerable benefit for this outcome. The 95%CI of the pooled effect estimate does not overlap the 0.90 margin. Therefore, the estimate is considered precise. j95%CIs of individual trials do not overlap. Inconsistency among trials (I2 74%) that could not be explained. kA RRR of 10% was considered a considerable benefit for this outcome. The 95%CI of the pooled effect estimate does not include the 0.90 margin of considerable benefit, but does include the 1.10 margin of considerable harm. Therefore, the estimate is imprecise.
All-cause stroke and systemic embolism (follow-up mean 2.2 years; assessed with: clinical assessment) 4 Randomized No serious No serious Seriousb Seriousc Noned trials risk of inconsistency biasa Ischaemic and unspecified stroke (follow-up median 2.2 years; assessed with: clinical assessment) 4 Randomized No serious No serious Seriousb Seriousc Noned trials risk of inconsistency biasa Haemorrhagic stroke (follow-up median 2.2 years; assessed with: clinical assessment) 4 Randomized No serious No serious Seriousb No serious Noned trials risk of inconsistency imprecisione biasa Major bleeding (follow-up median 2.2 years; assessed with: clinical assessment) Seriousg,h Seriousc Noned 4 Randomized No serious Seriousf trials risk of biasa Intracranial bleeding (follow-up median 2.2 years; assessed with: clinical assessment) Seriousg,h No serious Noned 4 Randomized No serious No serious inconsistency imprecisioni trials risk of biasa Gastrointestinal bleeding (follow-up median 2.2 years; assessed with: clinical assessment) 4 Randomized No serious Seriousj Seriousg,h Seriousk Noned trials risk of biasa
No. of studies
Quality assessment
Table 2. Summary of the application of the different GRADE-criteria on the critical outcomes of NOACs resulting in a rating of the evidence (5).
Eur J Gen Pract Downloaded from informahealthcare.com by Biblioteka Uniwersytetu Warszawskiego on 01/23/15 For personal use only.
New oral anticoagulants 3
Eur J Gen Pract Downloaded from informahealthcare.com by Biblioteka Uniwersytetu Warszawskiego on 01/23/15 For personal use only.
4
W. Opstelten et al.
ment of allocation, blinded assessment of effects, a low rate of missing data, a well-supported non-inferiority margin, and an intention-to-treat analysis. However, only one RCT additionally performed a per-protocol analysis, which is recommended in non-inferiority trials to reduce the chance of falsely declaring non-inferiority (7,10). In one RCT, the patients were not blinded for allocated treatment, which could have introduced bias in the reporting or adjudication of events (6). This was compensated by the implementation of several validated procedures, including blinded evaluation of outcome effects. Moreover, the results of previous placebo controlled open-label RCTs assessing the effectiveness of warfarin did not differ from the results of double-blinded RCTs (11). In summary, we concluded that the included RCTs are at low risk of bias. Publication bias To our knowledge, the meta-analysis of Ruff et al. (5) included all available RCTs. There was no indication that they failed to report studies that showed no effect. Inconsistencies Some inconsistency among studies was found: two RCTs indicated a higher risk of major bleeding in patients on warfarin, whereas the other RCTs could not detect a risk difference (I2 83%) (6–9). Moreover, three RCTs reported an increased risk of gastrointestinal bleeding in patients on NOACs, whereas the fourth RCT indicated a statistically non-significant protection against this complication in patients randomized to NOACs (I2 74%) (6–9). These inconsistencies cannot be explained with the available data. Given that the inconsistency is present within the group of direct factor Xa inhibitors, a difference in drug properties seems an implausible explanation. A difference in patient characteristics does not seem responsible for the inconsistency either: in fact, the inconsistencies were observed between two studies, which included patients with similar baseline risks of stroke (6,8).
particular for the use of NOACs in daily practice. Several factors, such as reduced patient compliance in daily care compared to research context, (temporary) worsening of renal function, and interactions with concomitant medication, may affect the blood level of NOACs and, therefore, its therapeutic effects. Without regular monitoring of coagulation, any deviation beyond the therapeutic range will not be noticed. This may expose patients to inadequate protection against ischaemic strokes or increased risk of adverse events. A recent observational study reported that dabigatran might be as effective and safe in daily practice and research context (15). However, since the study was not randomized, confounding by indication may have biased these results. Imprecision For all relevant outcomes, we considered a relative risk reduction (RRR) of 10% as clinically relevant. However, for haemorrhagic stroke, which is associated with severe impact on the quality of life, we put the limit at 5% (12,13). With respect to these limits, only the risk reductions of haemorrhagic stroke and intracranial haemorrhage were sufficiently precise. Table 2 summarizes the application of the different GRADE-criteria on the critical outcomes of NOACs resulting in a rating of the evidence.
DISCUSSION Bearing in mind all these considerations, the AF guideline development group of the DCGP is reluctant to recommend broadly the use of NOACs in the elderly. For now, we think that these drugs should only be prescribed if the following conditions are met: age younger than 80 years, few comorbid conditions, good renal function (GFR ⬎ 50 ml/min), and good expected drug adherence (16). Using these criteria, GPs could discuss the choice in antithrombotic therapy with their patients according to the principles of shared decision making.
Conclusion Indirectness of evidence The inclusion and exclusion criteria of the RCTs permitted entrance of patients with varying severity of illness. However, it has been shown that many AF patients who use oral anticoagulants for prevention of thromboembolic complications do not fulfil the eligibility criteria of the trials that demonstrated effectiveness and safety of these drugs. In a case-control study, 40% of users of VKAs presenting with bleeding had one or more exclusion criteria for the trials, supporting the use of anticoagulation for their condition (14). Although this may be equally true for both VKAs and NOACs, this has consequences in
NOACs are alternatives of VKAs because of their lower risk of harm and ease of use. Outside research context, however, the experience with NOACs is still limited. We feel additional research in primary care setting should be performed to assess the effectiveness and safety of NOACs before they can be recommended for use in primary care.
ACKNOWLEDGEMENT The authors thank Holger Schünemann, MD, PhD, epidemiologist, for his comment on the manuscript.
New oral anticoagulants Declaration of interest: The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.
Eur J Gen Pract Downloaded from informahealthcare.com by Biblioteka Uniwersytetu Warszawskiego on 01/23/15 For personal use only.
REFERENCES 1. Heeringa J, van der Kuip DA, Hofman A, Kors JA, van Herpen G, Stricker BH, et al. Prevalence, incidence and lifetime risk of atrial fibrillation: The Rotterdam study. Eur Heart J. 2006;27:949–53. 2. Lip GY, Nieuwlaat R, Pisters R, Lane DA, Crijns HJ. Refining clinical risk stratification for predicting stroke and thromboembolism in atrial fibrillation using a novel risk factor-based approach: The Euro heart survey on atrial fibrillation. Chest 2010;137:263–72. 3. Hart RG, Pearce LA, Aguilar MI. Meta-analysis: Antithrombotic therapy to prevent stroke in patients who have nonvalvular atrial fibrillation. Ann Intern Med. 2007;146:857–67. 4. Guyatt GH, Oxman AD, Vist GE, Kunz R, Falck-Ytter Y, Alonso-Coello P, et al. GRADE: An emerging consensus on rating quality of evidence and strength of recommendations. Br Med J. 2008;336:924–6. 5. Ruff CT, Giugliano RP, Braunwald E, Hoffman EB, Deenadayalu N, Ezekowitz MD, et al. Comparison of the efficacy and safety of new oral anticoagulants with warfarin in patients with atrial fibrillation: A meta-analysis of randomised trials. Lancet 2014;383:955–62. 6. Connolly SJ, Ezekowitz MD, Yusuf S, Eikelboom J, Oldgren J, Parekh A, et al. Dabigatran versus warfarin in patients with atrial fibrillation. N Engl J Med. 2009;361:1139–51. 7. Patel MR, Mahaffey KW, Garg J, Pan G, Singer DE, Hacke W, et al. Rivaroxaban versus warfarin in nonvalvular atrial fibrillation. N Engl J Med. 2011;365:883–91. 8. Granger CB, Alexander JH, McMurray JJ, Lopes RD, Hylek EM, Hanna M, et al. Apixaban versus warfarin in patients with atrial fibrillation. N Engl J Med. 2011;365:981–92.
5
9. Giugliano RP, Ruff CT, Braunwald E, Murphy SA, Wiviott SD, Halperin JL, et al. Edoxaban versus warfarin in patients with atrial fibrillation. N Engl J Med. 2013;369:2093–104. 10. Piaggio G, Elbourne DR, Altman DG, Pocock SJ, Evans SJ. Reporting of noninferiority and equivalence randomized trials: An extension of the CONSORT statement. J Am Med Assoc. 2006;295:1152–60. 11. Ezekowitz MD, Connolly S, Parekh A, Reilly PA, Varrone J, Wang S, et al. Rationale and design of RE-LY: Randomized evaluation of long-term anticoagulant therapy, warfarin, compared with dabigatran. Am Heart J. 2009;157:805–10, 810 e1–2. 12. Olesen JB, Lip GY, Hansen ML, Hansen PR, Tolstrup JS, Lindhardsen J, et al. Validation of risk stratification schemes for predicting stroke and thromboembolism in patients with atrial fibrillation: Nationwide cohort study. Br Med J. 2011; 342(31):d124. 13. Friberg L, Rosenqvist M, Lip GY. Net clinical benefit of warfarin in patients with atrial fibrillation: A report from the Swedish atrial fibrillation cohort study. Circulation 2012; 125:2298–307. 14. Levi M, Hovingh GK, Cannegieter SC, Vermeulen M, Buller HR, Rosendaal FR. Bleeding in patients receiving vitamin K antagonists who would have been excluded from trials on which the indication for anticoagulation was based. Blood 2008; 111:4471–6. 15. Larsen TB, Rasmussen LH, Skjoth F, Due KM, Callreus T, Rosenzweig M, et al. Efficacy and safety of dabigatran etexilate and warfarin in ‘real-world’ patients with atrial fibrillation. J Am Coll Cardiol. 2013;61:2264–73. 16. van den Donk M, Opstelten W. Important changes in the Dutch College of General Practitioners (NHG) practice guideline ‘Atrial fibrillation’: Revised practice guideline issued [in Dutch]. Ned Tijdschr Geneeskd. 2013;157:A6697.
