U N S O L IC I T E D R EV I E W

Non-vitamin K Oral Anticoagulants Versus Warfarin for Patients with Atrial Fibrillation: Absolute Benefit and Harm Assessments Yield Novel Insights Cyrus R. Kumana, Bernard M.Y. Cheung, David C.W. Siu, Hung-Fat Tse & Ian J. Lauder Department of Medicine, Queen Mary Hospital, The University of Hong Kong, Hong Kong, China

Keywords Atrial fibrillation; Drug acquisition costs; Hemorrhagic stroke/transformation; NNT; Non-vitamin K oral anticoagulants; Stroke/ systemic embolism. Correspondence Cyrus R. Kumana, Division of Clinical Pharmacology & Therapeutics, Department of Medicine, The University of Hong Kong, Queen Mary Hospital, Pokfulam, Hong Kong, China. Tel.: +852-2255-4248; Fax: +852-2816-2863; E-mail: [email protected]

doi: 10.1111/1755-5922.12173

SUMMARY Background and Objectives: Benefits and/or harms (including costs) of non-vitamin K oral anticoagulants (NOACs) versus warfarin therapy need appreciation in relative and absolute terms. Methods: Accordingly, we derived clinically relevant relative and absolute benefit/harm parameters for NOACs (apixaban, dabigatran, rivaroxaban, edoxaban) compared to warfarin from four clinical trials involving atrial fibrillation (AF) patients. For each trial, we tabulated patient numbers enduring four important outcomes and calculated unadjusted relative risk reduction (RRR) and number needed to treat (NNT)/year values (and 95% confidence intervals) for the NAOC compared to warfarin. These outcomes were as follows: stroke/systemic embolism (primary endpoint), hemorrhagic stroke, major bleeds, and death. We also addressed drug acquisition costs. Results: Each NOAC was noninferior to warfarin for primary-outcome prevention; RRRs were 12–33% and NNT/year values were 182–481, and all but one indicated statistically significant superiority. All the NOACs yielded statistically significant reductions in hemorrhagic stroke risk; RRRs were 42–74% and NNT/year values were 364–528. Major bleeding risk was comparable in both groups. Apixaban yielded a lower NNT/year for preventing death than for primary-outcome prevention. Compared to warfarin, NOAC acquisition costs were 70- to 140-fold greater. Conclusions: For the primary outcome, the absolute benefits of NOACs were modest (NNT/year values being large). Reduced hemorrhagic stroke rates with NOACs could be due to superior embolic infarct prevention and fewer consequential hemorrhagic transformations. Among apixaban recipients, the absolute mortality benefit exceeded that for the primary outcome, indicating prevention of additional unrelated deaths. The substantially greater NOAC acquisition costs need viewing against probable greater safety and the avoidance of monitoring bleeding risks.

Introduction Non-vitamin K oral anticoagulants (NAOCs) also known as novel oral anticoagulants have been developed based on the proposition that unlike warfarin their use to treat patients would not warrant routine monitoring of bleeding tendency, such as by determination of the international normalized ratio (INR). To date, at least four such marketed agents (apixaban, dabigatran, edoxaban, and rivaroxaban) have been tested for noninferiority to warfarin in four separate, large-scale, balanced, randomized clinical trials (ARISTOTLE,i RE-LY,ii ENGAGE,iii and ROCKETiv) entailing treatment of patients with atrial fibrillation (AF) at moderate-to-high risk of stroke [1–4]. In two of the trials, mitral stenosis that was not moderate to severe was not an exclusion criterion [1,3], but in the others all the patients had nonvalvular AF only. While the protocols of these trials differed in minor respects, they were all multicenter, multicountry, large-scale studies (recruiting in excess

100

Cardiovascular Therapeutics 34 (2016) 100–106

of 6000–9000 patients into each treatment arm), and reported on a number of common, clinically relevant hard outcomes that were defined in similar terms. All four trials were industry-sponsored and recruited older patients (median ages ranged from 70 to 73 years) with a male preponderance of about 2/1. They all entailed the same target INR of 2–3 for those taking warfarin, and estimated mean durations in the therapeutic range were similar (55–65%). Three of the trials [2–4] also used double dummies and sham INR monitoring, and the median CHADS2 scores [5] of recruited patients were nearly 3.5 in one of them [2], and 2.1–2.8 in the remainder. Before advocating NOACs routinely in clinical practice, it is important to consider their purported effectiveness and safety in absolute terms, especially as they are still relatively unknown drugs for which long-term experience is lacking. There have been several systematic reviews, meta-analyses, and other reports that have addressed important absolute benefits and harms, often

ª 2016 John Wiley & Sons Ltd

C.R. Kumana et al.

using number needed to treat (NNT) and/or number needed to harm (NNH) [6–10]. They did not, however, facilitate comparison between trials as all of them omitted adjustments for treatment duration, some addressed only a limited number of outcomes, and several analyses predated the latest trial. In the short term, moreover, the burden from acquisition costs for NOACs was liable to be much greater, notwithstanding the savings accruing from the avoidance of INR monitoring and consequential inconveniences. Against this background, we set out to determine and compare the benefits and harms resulting from treatment involving these NOACs with reference to warfarin therapy expressed in both relative and absolute terms as well as acquisition costs.

Methods The endpoints that we selected for analysis were as follows: stroke (ischemic or hemorrhagic)/systemic embolism, any death, hemorrhagic stroke, and major bleeds, as these were considered most clinically relevant and most likely to provide critical insights. For each of the four trials, we tabulated patient numbers that had or did not have the four selected outcome events on each treatment. For each trial, we also derived relevant unadjusted values for relative risk reduction (RRR) and NNT (a measure of absolute benefit or harm) for the NOAC compared to warfarin therapy, together with corresponding 95% confidence intervals (CIs). In these trials, median patient follow-up periods on treatment varied from 1.6 to 2.8 years, and so it was important to take these into account when comparing absolute benefit (or harm). This was why we adjusted NNT values for treatment duration and expressed them as NNTs/ year, although we realize this involves assuming that beneficial (or harmful) effects of treatment are conferred uniformly during such periods. All such calculations were performed as previously described [11], using the raw data obtained from the published results of each trial and when necessary from their supplementary appendices [1–4]. For the RE-LY and ENGAGE trials, we only analyzed the treatments involving patients assigned to high-dose therapy (dabigatran 150 mg twice daily; edoxaban 60 mg daily), as only these dosages were being recommended based on their effectiveness for the primary outcome as compared to warfarin. The ENGAGE trial mostly entailed analysis of what was termed a “modified” intention-to-treat population, but for major bleeds the data analyzed referred to a so-called safety cohort consisting of very slightly fewer patients. In the ROCKET AF study, primaryoutcome data were presented as per protocol as well as on an intention-to-treat basis (median follow-up 1.6 and 1.9 years, respectively), whereas the reporting of other outcomes was largely confined to various so-called safety-as-treated populations. In the latter trial, patient numbers having different endpoints reported in per-protocol and intention-to-treat populations were not easy to disentangle, although the discrepancies were very minor. In this trial moreover, deaths from any cause (an important outcome) as well as patient numbers having hemorrhagic strokes were only available from the trial’s supplementary report, for which intention-to-treat details were lacking [4]. Whenever feasible, we nevertheless derived all parameters based on available intention-to-treat data. In the ENGAGE trial however, analysis of major bleeds was based on the so-called safety cohort, and for ROCKET AF both per-protocol and intention-to-treat parameters

ª 2016 John Wiley & Sons Ltd

NOACs Versus Warfarin in AF: Insights from NNT/Year

were derived for the primary outcome, whereas “safety-as-treated” populations were used for the other endpoints. Notwithstanding savings accruing from avoidance of INR monitoring and reduced consequential inconveniences, in the short term, at least the burden from acquisition costs of the NOACs was also liable to influence prescribing. We therefore also retrieved UK and local data for prevailing costs and/or cost savings pertaining to the use of these drugs based on the online British National Formulary and on locally available public hospital pricing, respectively.

Results In these trials, patient numbers on each treatment that had or did not have the outcome events we selected to analyze were mostly from intention-to-treat data and are shown in Table 1. For ROCKET AF, most of the available details were from “safety-astreated” populations, and in ENGAGE patient numbers with major bleeds depended on a “safety cohort.” Using information from this table, we derived RRR and NNT/year values as shown in Table 2. A summary of the acquisition costs for warfarin and the NOACs is shown in Table 3. Depending on the dosage forms used, the figures show that in divergent parts of the world, the acquisition costs for NOACS at this time were about 70- to 140-fold those for warfarin.

Discussion This analysis used published raw data reported in the trials referred to, but whether such results apply outside the context of such studies has been questioned [7]. Nevertheless, consideration of absolute benefits and risks (using NNTs) together with relative benefits and risks (using RRRs) can provide important clinically relevant insights that could be just as applicable to patients encountered in clinical practice.

Stroke/Systemic Embolism Stroke/systemic embolism was the primary outcome in all of these trials and included both hemorrhagic and ischemic strokes. Table 2 shows that in all but one trial, remarkably the RRR and NNT/year values yielded statistically significant superiority. In ENGAGE (the exception), there was nevertheless a trend favouring high-dose edoxaban. In ROCKET AF moreover, superiority was only significant in the per-protocol analysis; based on the reported intention-to-treat analysis, there was only a trend. In these trials, the absolute accrued benefits of treatment with NOACs appeared modest in terms of NNT/year values (182–480) compared to the figure of 63 for simvastatin in the 4S trial of highrisk coronary artery disease patients [11]. Nonetheless, these findings (established noninferiority and probable superiority) should be viewed in the context of benefits from the avoidance of INR monitoring. The increase in acquisition costs for these drugs (approximately 70- to 140-fold) over and above those for warfarin seems striking but may be regarded as modest in absolute terms (Table 3). They should also be balanced against the savings enjoyed via avoidance of INR monitoring and the associated burden/inconvenience of attending anticoagulant clinics or other forms of follow-up. While both direct and indirect costs/charges

Cardiovascular Therapeutics 34 (2016) 100–106

101

C.R. Kumana et al.

NOACs Versus Warfarin in AF: Insights from NNT/Year

Table 1 Numbers with and without outcome events in the selected Non-vitamin K Oral Anticoagulant Trials ARISTOTLE

RE-LY

ENGAGE

ROCKET AF*

Outcome event

Treatment

No with event

No with no event

No with event

No with no event

No with event

No with no event

No with event

No with no event

Stroke systemic/embolism

NOAC Warfarin NOAC Warfarin NOAC Warfarin NOAC Warfarin

212 265 603 669 40 78 327 462

8908 8816 8517 8412 9080 9003 8793 8619

134 199 438 487 12 45 375 397

5942 5823 5638 5535 6064 5977 5701 5625

296 337 773 839 49 90 418 524

6739 6699 6262 6197 6986 6946 6594 6488

188 241 208 250 29 50 395 386

6770 6763 6853 6832 7032 7032 6716 6739

Any death Hemorrhagic stroke Major bleed

*For ROCKET AF, Stroke/systemic embolism refers to per-protocol population findings, and other outcomes to various safety-as-treated populations. In the warfarin arms of each study, respective numbers having a hemorrhagic stroke expressed as a proportion of those having a primary-outcome event (ischemic or hemorrhagic stroke/systemic embolism) were as follows: 78/265 (29%), 45/199 (23%), 90/337 (27%), and 50/241 (21%). Corresponding proportions in the novel anticoagulant arms of each study were as follows: 40/212 (19%), 12/111 (11%), 49/296 (17%), and 29/188 (15%). Based on the numbers of reported hemorrhagic stroke patients in these AF trials and median follow-up durations (see Table 2), respective rates of such strokes in those taking warfarin were crudely estimated to be about 4.8, 3.7, 4.6, and 4.4 per thousand patient years.

for routine INR monitoring and their impact on quality of life also need consideration, these endpoints were not a focus of our analysis.

Hemorrhagic Strokes The most consistent, clinically and statistically significant finding common to ARISTOTLE, RE-LY, ENGAGE, and ROCKET AF was the reduction in hemorrhagic strokes; respective values for RRR were 49, 74, 46, and 42%, and for NNT/year they were 428, 364, 481, and 528. Several meta-analyses also emphasized that the substantial reduction in hemorrhagic strokes associated with NOAC rather than warfarin treatment was the main driver for the reduced risk of stroke/systemic embolism [6,7,10]. The relatively small absolute benefits indicated by the large NNT/year values, despite impressive RRRs, reflect the small absolute numbers of hemorrhagic strokes in patients taking warfarin (Table 1). From the latter table, it is also evident that among patients in the warfarin arms of these studies, numbers having hemorrhagic strokes expressed as a proportion of those having a primary-outcome event (ischemic or hemorrhagic stroke/systemic embolism) appear excessive (29, 23, 27, and 21%, respectively). Moreover, if noncerebral embolic events could have been eliminated from the denominator, among the remainder (stroke patients only) the proportions designated as hemorrhagic would have been even greater. Our crude estimates of respective hemorrhagic stroke rates reported in these AF trial patients taking warfarin (Table 1) were 4.8, 3.7, 4.6, and 4.4 per thousand patient years. By contrast, in Dutch population cohorts aged 65–74 and 75–84 years, hemorrhagic strokes per 1000 person years (In these cohorts, rates for unspecified strokes were negligible.) were estimated to be 0.6 and 1.8, respectively [12]. Obviously, our comparatively high rates could be a direct effect of vitamin K antagonism. Alternatively, they could be an indirect effect, possibly because many so-called hemorrhagic strokes were actually instances of

102

Cardiovascular Therapeutics 34 (2016) 100–106

mislabeled embolic strokes with hemorrhagic transformation. After all, patients recruited into these trials were from many diverse centers in diverse countries, where imaging facilities and the delays before they were utilized could have been very variable. Regrettably, it was not feasible to obtain hemorrhagic stroke rates in non-AF patients receiving warfarin from venous thromboembolism (VTE) trials involving these NOACs, many of which have been extensively cited in reviews [13,14]. This was because the relevant individual trials did not distinguish between different types of intracranial bleeds (traumatic vs. stroke), entailed far fewer patients than the AF trials, and did not exclude patients with AF, and treatment durations were generally much shorter (3–12 months) and variably documented. Table 1 shows that astonishingly, among NOAC-treated patients enduring a primary outcome in ARISTOTLE, RE-LY, ENGAGE, and ROCKET AF, respective hemorrhagic stroke rates were down to 19, 9, 17, and 15% compared to 29, 23, 27, and 21% for patients taking warfarin. While such encouraging results with NAOCs could well be due to inherently lower hemorrhagic stroke risks despite superior clot prevention efficacy, they could also be a reflection of fewer embolic strokes and a consequential reduction in instances of hemorrhagic transformation of the resulting infarcts. The latter possibility is supported by RE-LY trial findings, whereby high-dose dabigatran was associated with a greater RRR for hemorrhagic stroke than low-dose treatment (74% vs. 69%), although this difference was not statistically significant [1]. This paradoxical finding could be consistent with many labeled hemorrhagic strokes being hemorrhagic transformations; arguably, fewer embolic strokes on high-dose treatment may have resulted in fewer instances of this complication. Moreover, in dabigatrantreated mice, experimentally induced ischemic strokes were not associated with hemorrhagic transformation [15], which was in sharp contrast to what is known about warfarin-treated animals. In the ENGAGE trial however, low-dose edoxaban was associated with a nonsignificant higher rate of hemorrhagic stroke reduction

ª 2016 John Wiley & Sons Ltd

ª 2016 John Wiley & Sons Ltd

11 ( 2 to 22) 74 (49 to 86)

238 (119 to ∞) 428 (273 to 987) 120 (85 to 200)

10 (0 to 20) 49 (25 to 65)

30 (19 to 39)

6 ( 8 to 19)

33 (17 to 47)

303 (169 to 1501)

20 (4 to 34)

475 ( 427 to 153)

228 ( 2269 to 108) 364 (250 to 666)

182 (118 to 397)

20 (9 to 30)

8 ( 2 to 17) 46 (22 to 62)

12 ( 3 to 25)

%RRR (95% CI)

185 (119 to 420)

299 ( 2038 to 139) 481 (306 to 1123)

481 ( 2398 to 219)

NNT/year (95% CI)

%RRR (95% CI)

%RRR (95% CI)

NNT/year (95% CI)

Edoxaban‡ 60 mg daily

Dabigatran‡ 150 mg 9 2/day

Apixaban 5 mg 9 2/day NNT/year (95% CI)

ENGAGE AF TIMI 48 (14,071) 2.8

RE-LY (12,098) 2.0

ARISTOTLE (18,201) 1.8

3 ( 18 to 11)

17 (0 to 32) 42 (8 to 64)

21 (5 to 35)

%RRR (95% CI)

Rivaroxaban 20 mg daily

ROCKET AF* (14,171) 1.6 or 1.9†

1166 ( 178 to 256)**

257 (131 to 6437) 528 (288 to 3144)

217 (121 to 1031)

NNT/year (95% CI)

*Analyses shown for this trial were not intention-to-treat, and that shown for the primary endpoint was per-protocol; based on the reported intention-to-treat analysis, respective RRR and NNT/year values and corresponding CIs were as follows: 12 ( 5 to 35) % and 367 ( 1031 to 161). †Only for the intention-to-treat population. Negative RRR or NNT/year values indicate harm; ∞ refers any value considered to near infinity. ‡Parameters for dabigatran 110 mg 9 2/day and edoxaban 30 mg daily were not derived. §Dosages were reduced (based on varying criteria) for renal insufficiency, low body weight, and co-treatment with putative interacting drugs. ¶Ischemic or hemorrhagic. **This point estimate (indicating minimal harm) appears anomalous, because it falls between a low confidence ( 178) indicating harm and a high value (256) indicating some benefit.

Stroke¶/systemic embolus [primary endpoint] Any death Hemorrhagic stroke Major bleed

NOAC Standard dosage§

Trial (Patient nos) Median follow-up (years)

Table 2 RRR and NNT/year values derived from randomized controlled trials of treatment with NOACs versus Warfarin in AF patients

C.R. Kumana et al. NOACs Versus Warfarin in AF: Insights from NNT/Year

Cardiovascular Therapeutics 34 (2016) 100–106

103

C.R. Kumana et al.

NOACs Versus Warfarin in AF: Insights from NNT/Year

12. Hollander M, Koudstaal PJ, Bots ML, et al.

17. Cohen D. Dabigatran: How the drug company

23. Mani H, Lindhoff-Last E. New oral

Incidence, risk, and case fatality of first ever

withheld important analyses. BMJ 2014;2014:

anticoagulants in patients with nonvalvular

stroke in the elderly population. The Rotterdam

g4670.

atrial fibrillation: A review of pharmacokinetics,

Study. J Neurol Neurosurg Psychiatry 2002;74:317–321. 13. Yeh CH, Gross PL, Weitz JI. Evolving use of new oral anticoagulants for treatment of venous thromboembolism. Blood 2014;124:1020–1028. 14. Dobesh PP, Fanikos J. New oral anticoagulants

18. Reilly PA, Lehr T, Haertter S, et al. The effect of dabigatran plasma concentrations and patient characteristics on the frequency of ischemic stroke and major bleeding in atrial fibrillation patients. J Am Coll Cardiol 2014;63:321–328. 19. Chin PKL, Wright DFB, Patterson DM, Doogue PM, Begg EJ. A proposal for dose- adjustment

safety, efficacy, quality of life, and costeffectiveness. Drug Des Devel Ther 2014;8:789– 798. 24. St€ ollberger C, Finsterer J. Concerns regarding the use of dabigatran for stroke prevention in Atrial Fibrillation. Pharmaceuticals 2012;5:155– 168.

for the treatment of venous thromboembolism:

of dabigatran etexilate in atrial fibrillation

understanding differences and similarities. Drugs

guided by thrombin time. Brit J Clin Pharmacol

Desmeules JA. Role of P-glycoprotein in the

2014;74:2015–2032.

2014;78:599–609.

uptake/efflux transport of oral vitamin K

15. Ferdinand B, Mirceska A, Pfeilschifter J, et al.

20. Shafeeq H, Tran TH. New oral anticoagulants

No influence of dabigatran anticoagulation on

for atrial fibrillation: Are they worth the risk? P

hemorrhagic transformation in an experimental

T. 2014;39:54–64.

model of ischemic Stroke. PLoS ONE 2012;7: e40804. 16. Harper P, Young L, Merriman E. Bleeding risk

21. Thachil J. The newer direct oral anticoagulants: A practical guide. Clin Med 2014;14:165–175.

25. Gschwind L, Rollason V, Daali Y, Bonnabry P,

antagonists and rivaroxaban through the Caco-2 cell model. Basic Clin Pharmacol Toxicol 2013;113:259–265. 26. Rachidi S, Aldin ES, Greenberg C, Sachs B, Streiff M, Zeidan AM. The use of novel oral

22. Gregory YH. Lip, Giancarlo Agnelli. Edoxaban:

anticoagulants for thromboprophylaxis after

with dabigatran in the frail elderly. N Engl J

A focused review of its clinical pharmacology.

elective major orthopedic surgery. Expert Rev

Med 2012;366:864–866.

Eur Heart J 2014;35:1844–1855.

Hematol 2013;6:677–695.

106

Cardiovascular Therapeutics 34 (2016) 100–106

ª 2016 John Wiley & Sons Ltd

C.R. Kumana et al.

NOACs Versus Warfarin in AF: Insights from NNT/Year

and edoxaban had a greater overall absolute impact on preventing deaths than stroke/systemic embolism, although for the later agent the difference was not statistically significant. Both drugs therefore appeared to be preventing deaths additional to those accruing from primary-endpoint prevention. This is all the more surprising, as by no means all trial patients in the warfarin treatment arms having a stroke/systemic embolism would have died during their follow-up. Similarly, not all NOAC-treated patients having a stroke/systemic embolism would be expected to survive during follow-up. Assuming that up to 50% of patients enduring a stroke/systemic embolism while taking warfarin actually survive, then the benefits of taking these drugs (in terms of preventing deaths) could only extend to the remaining 50% who die. Conceivably, therefore, apixaban and edoxaban prevented a considerable number of deaths over and above those prevented by virtue of stroke/systemic embolism avoidance. Possible mechanisms include prevention of fatalities from non-stroke-related bleeding, myocardial infarction, and even pulmonary emboli (from the right atrium perhaps).

Conclusions Based on our analysis of unadjusted raw AF trial data, recourse to NOACs demonstrated noninferiority to warfarin and possible superiority too. All four drugs resulted in significant reductions in “hemorrhagic strokes,” possibly consistent with prevention of embolic infarcts and consequential hemorrhagic transformations. The overall risks of major bleeds during treatment with warfarin and NOACs appeared comparable, while fatal and critical-site bleeding (predefined subgroups) were less frequent with NOACs. Compared to warfarin, apixaban showed a lower overall mortality, and corresponding NNT/year values suggest a greater overall effect on the prevention of mortality than stroke/systemic embolism, which implies prevention of additional unrelated fatal events too. Greater NOAC acquisition costs need viewing against probably greater efficacy and safety, and avoidance of INR monitoring.

Ethics This is a review article, and no ethical approval is required.

References

Funding The study was funded by departmental resources.

Conflict of interest All authors have completed the Unified Competing Interest form at www.icmje.org/coi_disclosure.pdf (available on request from the corresponding author) and declare as follows: HFT reports personal fees from Daiichi Sankyo, Pfizer, BMS, Bayer, and BI in the previous 3 years; there were no other relationships or activities that could appear to have influenced the submitted work. In the previous 3 years, none of the other coauthors had supported, engaged in activities, or had financial relationships with any organizations that might have an interest in the submitted work.

Author Contributions CR Kumana contributed to conception and design, acquisition, analysis, and interpretation, drafted manuscript, and critically revised manuscript; BMY Cheung contributed to conception and design, and acquisition, and critically revised manuscript; DCW Siu contributed to conception and critically revised manuscript; HF Tse contributed to design and acquisition, and critically revised manuscript; IJ Lauder contributed to analysis and interpretation and critically revised manuscript. All authors gave final approval and agree to be accountable for all aspects of work ensuring integrity and accuracy of the manuscript. Notes i

Apixaban for Reduction in Stroke and Other Thromboembolic Events. ii Randomised Evaluation Long-Term Anticoagulation Therapy. iii Refers to ENGAGE AF-TIMI 48 or Effective Anticoagulation with Factor Xa Next Generation in Atrial Fibrillation –Thrombolysis In Myocardial Infarction 48. iv Rivaroxaban Once daily Compared with vitamin K antagonists for prevention of stroke and Embolism Trial in Atrial Fibrillation.

5. Gage BF, Waterman AD, Shannon W, Boechler

risk of intracranial hemorrhage traditional and

M, Rich MW, Radford MJ. Validation of clinical

bayesian meta-analysis and mixed treatment

1. Connolly SJ, Ezekowitz MD, Yusuf S, et al.

classification schemes for predicting stroke:

comparison of randomized trials of new oral

Dabigatran versus warfarin in patients with

Results from the National Registry of Atrial

anticoagulants in atrial fibrillation. JAMA Neurol

atrial fibrillation, and supplementary appendix.

Fibrillation. JAMA 2001;285:2864–2870.

N Engl J Med 2009;361:1139–1151.

6. Miller CS, Grandi SM, Shimony A, Filion KB,

2. Patel MR, Mahaffey KW, Garg J, et al.

2013;70:1486–1490. 9. Caldeira D, Barra M, Fausto J, Costa J, Ferreira

Eisenberg MJ. Meta-analysis of efficacy and

JJ, Pinto FJ. Intracranial hemorrhage risk with

Rivaroxaban versus warfarin in nonvalvular

safety of new oral anticoagulants (dabigatran,

the new oral anticoagulants: A systematic

atrial fibrillation, and supplementary appendix.

rivaroxaban, apixaban) versus warfarin in

review and meta-analysis. J Neurol

N Engl J Med 2011;365:883–891.

patients with atrial fibrillation. Am J Cardiol

3. Granger CB, Alexander JH, McMurray JJ, et al. Apixaban versus warfarin in patients with atrial

2012;110:453–460. 7. Dentali F, Riva N, Crowther M, Turpie AG, Lip

2014;33:555–560. 10. Ruff CT, Giugliano RP, Braunwald E, et al. Comparison of the efficacy and safety of new

fibrillation, and supplementary appendix. N

GY, Ageno W. Efficacy and safety of the novel

oral anticoagulants with warfarin in patients

Engl J Med 2011;365:981–992.

oral anticoagulants in atrial fibrillation: A

with atrial fibrillation: A meta-analysis of

4. Giugliano RP, Ruff CT, Braunwald E, et al. Edoxaban versus warfarin in patients with atrial fibrillation, and supplementary appendix. N Engl J Med 2013;369:2093–2104.

ª 2016 John Wiley & Sons Ltd

systematic review and meta-analysis of the literature. Circulation 2012;126:2381–2391. 8. Chatterjee S, Sardar P, Biondi-Zoccai G, Kumbhani DJ. New oral anticoagulants and the

randomised trials. Lancet 2014;383:955–962. 11. Kumana CR, Cheung BM, Lauder IJ. Gauging the impact of statins using number needed to treat. JAMA 1999;282:1899–1901.

Cardiovascular Therapeutics 34 (2016) 100–106

105

C.R. Kumana et al.

NOACs Versus Warfarin in AF: Insights from NNT/Year

12. Hollander M, Koudstaal PJ, Bots ML, et al.

17. Cohen D. Dabigatran: How the drug company

23. Mani H, Lindhoff-Last E. New oral

Incidence, risk, and case fatality of first ever

withheld important analyses. BMJ 2014;2014:

anticoagulants in patients with nonvalvular

stroke in the elderly population. The Rotterdam

g4670.

atrial fibrillation: A review of pharmacokinetics,

Study. J Neurol Neurosurg Psychiatry 2002;74:317–321. 13. Yeh CH, Gross PL, Weitz JI. Evolving use of new oral anticoagulants for treatment of venous thromboembolism. Blood 2014;124:1020–1028. 14. Dobesh PP, Fanikos J. New oral anticoagulants

18. Reilly PA, Lehr T, Haertter S, et al. The effect of dabigatran plasma concentrations and patient characteristics on the frequency of ischemic stroke and major bleeding in atrial fibrillation patients. J Am Coll Cardiol 2014;63:321–328. 19. Chin PKL, Wright DFB, Patterson DM, Doogue PM, Begg EJ. A proposal for dose- adjustment

safety, efficacy, quality of life, and costeffectiveness. Drug Des Devel Ther 2014;8:789– 798. 24. St€ ollberger C, Finsterer J. Concerns regarding the use of dabigatran for stroke prevention in Atrial Fibrillation. Pharmaceuticals 2012;5:155– 168.

for the treatment of venous thromboembolism:

of dabigatran etexilate in atrial fibrillation

understanding differences and similarities. Drugs

guided by thrombin time. Brit J Clin Pharmacol

Desmeules JA. Role of P-glycoprotein in the

2014;74:2015–2032.

2014;78:599–609.

uptake/efflux transport of oral vitamin K

15. Ferdinand B, Mirceska A, Pfeilschifter J, et al.

20. Shafeeq H, Tran TH. New oral anticoagulants

No influence of dabigatran anticoagulation on

for atrial fibrillation: Are they worth the risk? P

hemorrhagic transformation in an experimental

T. 2014;39:54–64.

model of ischemic Stroke. PLoS ONE 2012;7: e40804. 16. Harper P, Young L, Merriman E. Bleeding risk

21. Thachil J. The newer direct oral anticoagulants: A practical guide. Clin Med 2014;14:165–175.

25. Gschwind L, Rollason V, Daali Y, Bonnabry P,

antagonists and rivaroxaban through the Caco-2 cell model. Basic Clin Pharmacol Toxicol 2013;113:259–265. 26. Rachidi S, Aldin ES, Greenberg C, Sachs B, Streiff M, Zeidan AM. The use of novel oral

22. Gregory YH. Lip, Giancarlo Agnelli. Edoxaban:

anticoagulants for thromboprophylaxis after

with dabigatran in the frail elderly. N Engl J

A focused review of its clinical pharmacology.

elective major orthopedic surgery. Expert Rev

Med 2012;366:864–866.

Eur Heart J 2014;35:1844–1855.

Hematol 2013;6:677–695.

106

Cardiovascular Therapeutics 34 (2016) 100–106

ª 2016 John Wiley & Sons Ltd

Non-vitamin K Oral Anticoagulants Versus Warfarin for Patients with Atrial Fibrillation: Absolute Benefit and Harm Assessments Yield Novel Insights.

Benefits and/or harms (including costs) of non-vitamin K oral anticoagulants (NOACs) versus warfarin therapy need appreciation in relative and absolut...
566B Sizes 0 Downloads 9 Views