Many of the cost-effectiveness analyses of apixaban against warfarin focused on Western populations but Asian evidence remains less clear. The present study aims to evaluate the cost-effectiveness of apixaban against warfarin in Chinese patients with non-valvular atrial fibrillation (NVAF) from a public institutional perspective in Hong Kong.
We used a Markov model incorporating 12 health state transitions, and simulated the disease progression of NVAF in 1,000 hypothetical patients treated with apixaban/warfarin. Risks of clinical events were based on the ARISTOTLE trial and were adjusted with local International Normalized Ratio control, defined as the time in therapeutic range. Real-life input for the model, including patients’ demographics and clinical profiles, post-event treatment patterns, and healthcare costs, were determined by a retrospective cohort of 40,569 incident patients retrieved from a Hong Kong-wide electronic medical database. Main outcome measurements included numbers of thromboembolic and bleeding events, life years, quality-adjusted life years (QALYs) and direct healthcare cost. When comparing apixaban and warfarin, treatment with incremental cost-effectiveness ratio (ICER) less than one local GDP per capita (USD 33,534 in 2014) was defined to be cost-effective.
In the lifetime simulation, fewer numbers of events were estimated for the apixaban group, resulting in reduced event-related direct medical costs. The estimated ICER of apixaban was USD 7,057 per QALY at base-case analysis and ranged from USD 1,061 to 14,867 per QALY under the 116 tested scenarios in deterministic sensitivity analysis. While in probabilistic sensitivity analysis, the probability of apixaban being the cost-effective alternative to warfarin was 96% and 98% at a willingness to pay threshold of USD 33,534 and 100,602 per QALY, respectively.
Citation: Li X, Tse VC, Lau WCY, Cheung BMY, Lip GYH, Wong ICK, et al. (2016) Cost-Effectiveness of Apixaban versus Warfarin in Chinese Patients with Non-Valvular Atrial Fibrillation: A Real-Life and Modelling Analyses. PLoS ONE 11(6): e0157129. https://doi.org/10.1371/journal.pone.0157129
Editor: Nanette H. Bishopric, University of Miami School of Medicine, UNITED STATES
Received: November 5, 2015; Accepted: May 25, 2016; Published: June 30, 2016
Copyright: © 2016 Li et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: All relevant data are within the paper and its Supporting Information files.
Funding: This study was supported by a grant from Bristol Myers Squibb and Pfizer (reference number: 260007295). EWC received financial support from Janssen, a division of Johnson and Johnson, Bristol Myers Squibb, Pfizer and Eisai; The Pharmaceutical Society of Hong Kong; The University of Hong Kong; Early Career Scheme and the General Research Fund, Research Grants Council, Hong Kong. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: EWC received financial support from Janssen, a division of Johnson and Johnson, Bristol Myers Squibb, Pfizer and Eisai; The Pharmaceutical Society of Hong Kong; The University of Hong Kong; Early Career Scheme and the General Research Fund, Research Grants Council, Hong Kong. There are no patents, products in development, or marketed products to declare. This does not alter the authors’ adherence to PLOS ONE policies on sharing data and materials.
Non-valvular atrial fibrillation (NVAF), the most common sustained cardiac arrhythmia, affects 1–2% of the general population worldwide and its prevalence increases with age [1, 2]. As one of the risk factors for stroke and systemic embolism, NVAF is associated with long-term morbidity, impaired quality of life, mortality and considerable financial burden. The healthcare cost for patients with NVAF is estimated to increase 1.6–3.1 fold after the occurrence of thromboembolic events, particularly in the first three years . Therefore, stroke prevention in patients with NVAF is important from both individual and societal perspectives.
Long-term anticoagulation management is required for stroke prophylaxis in patients with NVAF. Traditionally, warfarin (a vitamin K antagonist) has been the most commonly used anticoagulant . However warfarin poses challenges with respect to clinical management due to its narrow therapeutic range and drug metabolism affected by multiple factors such as genetic variables, diet and drug interactions. Apixaban (a novel direct factor Xa inhibitor) was therefore developed as an alternate anticoagulant. A multinational trial of ARISTOTLE (Apixaban for Reduction in Stroke and Other Thromboembolic Events in Atrial Fibrillation) demonstrated apixaban to be superior to warfarin in prevention of ischaemic stroke (21% reduction), bleeding (31% reduction) and mortality (11% reduction) without the need for monitoring . Recent subgroup analysis of the landmark trial showed consistent protective effect of apixaban on stroke, bleeding and mortality in East Asian and non-East Asian patients .
In addition to improved clinical efficacy and safety compared against standard treatment, cost-effectiveness is also an important consideration for new treatments in the setting of increased healthcare costs globally. Apixaban has been found to be cost-effective against warfarin in Europe [8–10], Australia  and USA . Notably, the current CE literature all focused on Western populations. Although Asians and Westerns share similar risk factor profiles, they might not be the same due to more history of stroke, increased tendency to bleed and suboptimal international normalised ratio (INR) control when treated with warfarin . These factors may have an impact on the cost-effective manner of anticoagulation therapies. Hence, it is important to incorporate local data to evaluate CE in the Asian population. The objective of the present study is to evaluate the cost-effectiveness of apixaban versus warfarin for stroke prophylaxis in Chinese patients with NVAF, from the public institutional perspective in Hong Kong.
The research protocol was approved by the Institutional Review Board of the University of Hong Kong/Hospital Authority Hong Kong West Cluster, Hong Kong. No written informed consent was obtained from patients for their clinical records to be used in this study, due to this is a retrospective study based on electronic medical database. All of the patient records/information was anonymized and de-identified prior to analysis.
This study compares the cost-effectiveness of apixaban against warfarin in Chinese patients with NVAF using a Markov model. Data sources and the retrieval procedure are illustrated in Fig 1. Lifetime cost and health outcomes were estimated from Markov cohort simulations. A retrospective cohort study was conducted to obtain real-life data for the model inputs. Clinical event risks were based on the ARISTOTLE trial  and incorporated by local INR control. Where local data were not available, data retrieved from comprehensive literature review and expert opinion were used. Baseline inputs are listed in Tables A and B in S1 File. Model inputs were entered, cross-checked and analysed independently by two researchers.
A previously developed Microsoft Excel Markov model incorporating 12 health states was adapted to simulate lifetime disease progression in 1,000 patients treated by apixaban (5mg BID) or warfarin (any strength). The health states modeled include NVAF, ischaemic stroke, systemic embolism, myocardial infarction, bleeding and death. A simplified schematic representation of the model structure and the health states transitions starting from NVAF are shown in Figures A and B in S1 File. A detailed description of the model can be found in previous publications [8, 10, 15]. For each treatment group, the number of events in each health state, life years (LYs), quality-adjusted life years (QALYs) and direct healthcare cost were calculated at the end of each model cycle and accumulated across lifetime.
The target population of the model are patients with NVAF who are suitable for warfarin therapy. Patient profiles were determined by a retrospective cohort analysis of incident patients with NVAF in the Clinical Data Analysis and Reporting System (CDARS). In brief, CDARS is an electronic medical database in Hong Kong that covers a population of over 7 million utilising public healthcare services and accounts for 80% of inpatient records in the territory . De-identified electronic patients’ records, including demographics, diagnoses, procedures, drug descriptions, laboratory tests and date of consultation, hospital admission and discharge, were all imported to CDARS for audit and research purposes.
At the time of the study, CDARS was unable to identify patients with NVAF directly. Target patients were identified by first retrieving patients with documented diagnosis of atrial fibrillation from Jan 1st 2010 to Dec 31st 2013 using ICD-9 diagnosis codes (Table C in S1 File for ICD-9 codes used). Among these patients, those with diagnoses of valvular heart disease, hyperthyroidism or had undergone valvular replacement surgery within 1 year prior to the index date were further excluded according to corresponding ICD-9 codes. The follow-up of each patient was commenced from the index date until the occurrence of any clinical events, death or July 31st 2014, whichever came first. This resulted in the identification of 40,569 eligible patients with 76,196 patient-years of follow-up. The key demographic and clinical profiles of these patients are shown in Table 1.
Risk of clinical events
The clinical event rates of apixaban and warfarin were obtained from the ARISTOTLE trial . As the risk of ischaemic stroke and bleeding associated with warfarin are highly dependent on the quality of INR control [17, 18], the model was developed to adjust the event risks by local INR control. Time in therapeutic range (TTR) distribution was defined as the indicator of INR control and collected from the CDARS retrospective cohort study.
Fatality and mortality
For thromboembolic and bleeding events, annual case-fatality rates were estimated from the CDARS retrospective cohort study. All-cause mortality was modelled based on life tables in the general population, adjusting by age, gender and the occurrence of NVAF and relevant co-morbidities.
Healthcare resource consumption regarding anticoagulation management was reflected by the frequency of routine care and monitoring visits, proportion of patients experiencing dyspepsia and proportion of patients requiring annual renal monitoring whilst on treatment from the CDARS retrospective cohort (Table A in S1 File).
Treatment pattern post-events
Treatment patterns post-event were determined from the CDARS retrospective cohort study, published literature or expert opinion wherever appropriate (Table A in S1 File). Following all bleeding events, a certain number of patients were assumed to switch to aspirin as second line treatment. The rate of treatment switch (± 90 days of the event index date) was determined by retrieving prescription records from CDARS and literature. The remaining patients were modelled to have their original treatment interrupted for 6 weeks then restarted  (i.e. as shown in Table A in S1 File, post intracranial haemorrhage, 44% of patients switched to aspirin and the remaining 56% discontinued their original treatment temporarily and restarted the same treatment 6 weeks later). Following occurrence of ischaemic stroke and systemic embolism, all patients were assumed to continue the original treatment as advised by experts.
Utility for NVAF was adopted from Ho et al , which was the only study to assess the health utility of Hong Kong patients with NVAF on oral anticoagulants. Due to the absence of population-specific utility estimations, a UK-based utility catalogue was adopted for other health states . Marginal disutility was assumed to be associated with INR monitoring when treated with warfarin due to the required blood test . All utility inputs were discounted at 3.5% per year and detailed in Table B in S1 File.
The analysis was conducted from a public institutional perspective in Hong Kong so only direct medical costs were considered. Costs are presented to the value of US dollars in 2014 (1 USD to 7.75 HKD) and discounted at 3.5% per year according to cost-effectiveness analyses (CEA) guideline .
The total cost components are outlined in Table B in S1 File. Daily drug costs for warfarin and apixaban were based on local retail prices (internal communication with Bristol-Myers Squibb, Hong Kong, November 2014). Annual monitoring and management costs were determined by multiplying the cost for blood and renal tests per visit by number of clinical visits per patient from the CDARS retrospective cohort. Similarly, acute care costs associated with clinical events were calculated by multiplying daily inpatient charges in public hospital [25, 26] by the average hospital length of stay estimated from the same cohort. Long-term care costs include medication costs only.
We compared the cost-effectiveness of apixaban versus warfarin by assessing the differences in lifetime cost and clinical benefits. Base-case analysis was performed based on model inputs described earlier. Treatment with incremental cost-effective ratio (ICER) less than the willingness-to-pay (WTP) threshold of one local GDP per capita (USD 33,534 in 2014 ) for each QALY gained was considered a cost-effective alternative as recommended by the WHO-CHOICE guideline .
The effect of uncertainties in the base-case analysis was evaluated by deterministic and probabilistic sensitivity analyses (SA). In deterministic SA, each of the 116 input parameters varied from a predefined range (95% CI or SD, as listed in Tables A and B in S1 File) while all the others were constant to project its effect on ICER. All tested parameters were ranked according to ICER variation and the top 15 most influential parameters were presented in Tornado diagrams.
In probabilistic SA, probability distributions were assigned to each parameter (gamma distribution to event rates, beta distribution to utilities and log-normal distribution to cost) and varied simultaneously in 2,000 iterations for the pair-wise comparisons. The cost and health outcomes were further assessed using cost-effectiveness acceptability curves (CEACs). Two willingness-to-pay (WTP) thresholds, one-time and three-times of local GDP per capita (USD 33,534 and USD 100,602 in 2014 ) were applied to explore the proportion of acceptable trials for each treatment.
From a cohort of 1,000 patients over a lifetime, the use of apixaban was predicted to result in 116 fewer NVAF-related events (12 fewer ischaemic strokes, 15 fewer haemorrhagic strokes, 12 fewer other intracranial haemorrhages, 25 fewer other major bleeds, 2 fewer myocardial infarctions) and 31 fewer event-related deaths than patients treated with warfarin in the base-case analysis (Table 2). Compared with warfarin, 36 patients need to be treated with apixaban to avoid one additional stroke, and apixaban is the dominant treatment in reducing bleeding events.
The cost of treating one patient with apixaban was USD 1,228 more than warfarin (Table 2). This was predominantly attributed to the higher drug price of apixaban, in contrast to lower costs related to management, monitoring and clinical events. Compared with warfarin, the net cost, net health benefits and incremental cost-effective ratios of apixaban were detailed in Table 3.
Deterministic sensitivity analyses
As illustrated in Fig 2, tested model inputs had limited impact on the ICER. The ICER ranged from USD 1,061 to 14,867 per QALY gained under 116 tested scenarios, all of which were within the threshold of USD 33,534 per QALY (one local GDP per capita). Uncertainties in drug costs, the risk of cardiovascular hospitalisation for both treatments, risk of ischaemic stroke for apixaban and risk of intracranial haemorrhage for warfarin had greatest impact on the ICER.
Probabilistic sensitivity analyses
The probabilistic sensitivity analysis demonstrated that apixiban is more cost-effective against warfarin over a life-time horizon (Fig 3). ICER was below the threshold of USD 33,534 per QALY (one local GDP per capita) in 96% of the trials. When the upper range of the WHO threshold of USD 100,602 per QALY gained (three-times of local GDP per capita) was applied, 98% of the trials supported the cost-effectiveness of apixaban.
The uncertainties in the cost-effective estimates are further summarised in the cost-effectiveness acceptability curve in Fig 4. This demonstrates that apixaban was a superior treatment choice representing the maximum net benefit over warfarin, when payers are willing to pay USD 7,500 or above for one QALY gained.
Summary of findings
In this study, our principal findings were as follows: (i) fewer thromboembolic events were estimated for the apixaban group in the lifetime cohort simulations, resulting in corresponding increased LYs, QALYs and reduced event-related healthcare costs; (ii) compared with warfarin, apixaban presented an ICER of USD 1,061–14,867 per QALY gained which is within the common acceptable WTP of one local GDP per capita as recommended by the WHO; (iii) both deterministic and probabilistic sensitivity analyses support the cost-effectiveness of apixaban over a range of uncertainties. This study utilises the strength of CDARS, which collects clinical data in the public healthcare setting across the whole territory. As far as we are aware, this is the only study which utilises real-life, Hong Kong-specific cost, patient demographics and clinical profiles, treatment patterns and case fatality rates to compare the cost-effectiveness of apixaban versus warfarin.
The greater health benefits and cost-effectiveness of apixaban findings in this study are in line with other modeling CE studies conducted in western developed countries [8–12]. Lee et al demonstrated apxiaban as a dominant strategy (both cost and life year savings) from the US Medicare perspective , while the remaining four studies all found apixaban more cost-effective against warfarin in Australia, Netherlands, Sweden and United Kingdom at ICERs of AUD 13,679, EUR 10,576, SEK 33,458, GBP 11,909 per QALY respectively, applying locally relevant WTP thresholds.
In light of the widely reported cost-effectiveness of apixaban in Caucasians, evaluation of the Asian population is still limited. However, from a clinical perspective, Asian and Western patients with AF are not necessarily the same. Firstly, among the five components of the CHADS2 (congestive heart failure, hypertension, age≥75 years, diabetes mellitus and stroke) score, similar risk factor profiles were shared, except previous stroke is more prevalent in Asians (29% vs. 18%) . Secondly, stroke risk is higher in Asians when treated with warfarin as reflected in the subgroup analysis of ARISTOTLE trial (3.4% vs. 1.4%) [6, 7]. Thirdly, despite a lower INR in general, warfarin users in Asia are more likely to bleed than non-Asians (3.8% vs. 3.0%) [6, 7]. Although the mechanism for these differences remains to be determined, they may all challenge the proper use and cost-effectiveness of warfarin and its alternatives in Asia. For the first time, our analysis showed that apixaban is cost-effective in a Chinese population. This is an important ethnic group in Asia with a high prevalence of stroke  and rapidly increasing ageing population , which highlights the prominence of this study.
Another important differentiator and strength of our study is the use of real-life, local INR control data in our model. It is well documented in the published literature that good INR control, although important for good clinical outcome , is unusual in both clinical trial  and real-life settings . The ARISTOTLE trial reported that only 50% of warfarin patients have TTR ≥ 66.0% . Hong Kong practice is not an exception. Consistent with our findings from the CDARS retrospective cohort study, which found a median TTR of 40.9%, Ho et al also reported unsatisfactory INR control among patients in Hong Kong with similar TTR (median: 38.8%) . As TTR is negatively correlated with bleeding and thromboembolic events , suboptimal INR control in real-life may worsen the performance of warfarin . Potential health benefits offered by apixaban may be greater than that in the clinical trial setting for patients with poor TTR  when compared to warfarin. By incorporating real-life INR control for event risks adjustment, our study provides a more realistic picture compared to previous studies where the risks were based on clinical trials [11, 12] or tested in scenario analyses [9, 10].
Of note, variations in healthcare system and services have been well recognised worldwide  and this is another key reason why country-specific CEA needs to be conducted. This study summarised local relevant information and will contribute to anticoagulation policy consideration in Hong Kong and other Asian countries with similar healthcare systems.
There are limitations in this modeling analysis. Firstly, clinical efficacy and safety parameters were based on the landmark trial with 1.8 years of follow-up. They were assumed to be constant over the lifetime in the model, but might not reflect the long-term performance of treatments in real terms. Secondly, following a major bleeding event, treatment options were limited to switching to aspirin or restarting the original anticoagulant in 6 weeks. In real-life clinical practice, some patients may cease treatment or choose other anticoagulants. However, aspirin is perceived as a safer alternative to oral anticoagulants in the Chinese population  given the particular concerns of increased bleeding risk among the Chinese [37–40]. As indicated in our previous work , more than twice the number of NVAF patients was receiving antiplatelet drugs including aspirin, compared to oral anticoagulants in Hong Kong, despite having a higher risk of stroke (CHADS2≥2). In addition, the cut-off point of assessing treatment switch was restrict to 90 days, based on Witt et al’s study on the association of warfarin interruption and gastrointestinal tract bleeding . This may be arbitrary as there are no published guidelines on the optimal anticoagulation approach for those experiencing a thromboembolic event at the time our study was conducted . Thirdly, in the baseline analysis, we optimistically assumed no renal monitoring was required for patients on apixaban based on the subgroup analysis of ARISTOTLE  which may not be the case in real clinical practice. The subgroup analysis was pre-specified for patients with impaired renal function and showed that apixaban was more effective than warfarin in preventing stroke or systemic embolism and reducing mortality, irrespective of renal function. Considering the median age of 78.9 years in this cohort of patients, we analysed two other scenarios—assuming 50% and 100% of patients will have their renal function evaluated annually, respectively, whilst on apixaban. Both scenarios show that the frequency of renal monitoring had very limited impact on the ICERs (USD 7,270 and USD 7,482 per QALY saved respectively) therefore baseline conclusions remains robust. Fourthly, treatment attributes such as convenience of drug administration, food and drug interaction were not included in this model, which may underestimate the benefits of apixaban. Lastly, the estimate for acute event cost per episode was restricted to general ward charges only and the costs of emergency and intensive care or rehabilitation have not been included. These would have led to underestimates of event related costs for both groups.
In conclusion, the current comprehensive modeling assessment based on real-life, local clinical, economic evidence and ARISTOTLE trial, demonstrate that apixaban is a cost-effective alternative against warfarin for the prevention of stroke and bleeding events in Chinese patients with NVAF. Our results suggest that apixaban is relatively more favourable in terms of health benefits and incremental cost.
Figure A. Schematic representation of Markov model. NVAF: non-valvular atrial fibrillation. Figure B. Health states transitions from NVAF. CRNM Bleed: Clinically relevant non-major bleed; ICH: Intracerebral haemorrhage; NVAF: Non-valvular atrial fibrillation; NVAF without original anticoagulation: patients with NVAF who discontinue the initially assigned anticoagulation treatment but have not yet experienced any of the events including ischaemic stroke, haemorrhagic stroke, myocardial infarction, systematic embolism or death. Table A. Base-case and sensitivity analyses inputs for clinical events, deaths, post-event treatment and anticoagulation management. Table B. Base-case and sensitivity analyses inputs for utility and cost. Table C. The ICD-9 codes (diagnosis codes unless other specified) used in this study. Supplemental References.
We thank our colleague in the Department of Pharmacology and Pharmacy of the University of Hong Kong–Dr. Shweta Anand for literature search and review; and Ms. Lisa Wong for the proofreading our manuscript.
Conceived and designed the experiments: EWC XL IW. Performed the experiments: XL WL. Analyzed the data: XL WL. Contributed reagents/materials/analysis tools: EWC IW. Wrote the paper: XL VT WL BC GL IW EWC.
- 1. Camm AJ, Kirchhof P, Lip GYH, Schotten U, Savelieva I, Ernst S, et al. Guidelines for the management of atrial fibrillation: The Task Force for the Management of Atrial Fibrillation of the European Society of Cardiology (ESC). European Heart Journal. 2010;31(19):2369–429. pmid:20802247
- 2. Tse HF, Wang YJ, Ahmed Ai-Abdullah M, Pizarro-Borromeo AB, Chiang CE, Krittayaphong R, et al. Stroke prevention in atrial fibrillation—an Asian stroke perspective. Heart rhythm: the official journal of the Heart Rhythm Society. 2013;10(7):1082–8.
- 3. Boccuzzi SJ, Martin J, Stephenson J, Kreilick C, Fernandes J, Beaulieu J, et al. Retrospective study of total healthcare costs associated with chronic nonvalvular atrial fibrillation and the occurrence of a first transient ischemic attack, stroke or major bleed. Current medical research and opinion. 2009;25(12):2853–64. pmid:19916729
- 4. De Caterina R, Husted S, Wallentin L, Andreotti F, Arnesen H, Bachmann F, et al. Vitamin K antagonists in heart disease: current status and perspectives (Section III). Position paper of the ESC Working Group on Thrombosis—Task Force on Anticoagulants in Heart Disease. Thrombosis and haemostasis. 2013;110(6):1087–107. pmid:24226379
- 5. Holbrook AM, Pereira JA, Labiris R, McDonald H, Douketis JD, Crowther M, et al. Systematic overview of warfarin and its drug and food interactions. Archives of internal medicine. 2005;165(10):1095–106. pmid:15911722
- 6. Granger CB, Alexander JH, McMurray JJV, Lopes RD, Hylek EM, Hanna M, et al. Apixaban versus Warfarin in Patients with Atrial Fibrillation. New Engl J Med. 2011;365(11):981–92. pmid:21870978
- 7. Goto S, Zhu J, Liu L, Oh BH, Wojdyla DM, Aylward P, et al. Efficacy and safety of apixaban compared with warfarin for stroke prevention in patients with atrial fibrillation from East Asia: a subanalysis of the Apixaban for Reduction in Stroke and Other Thromboembolic Events in Atrial Fibrillation (ARISTOTLE) Trial. American heart journal. 2014;168(3):303–9. pmid:25173541
- 8. Stevanovic J, Pompen M, Le HH, Rozenbaum MH, Tieleman RG, Postma MJ. Economic evaluation of apixaban for the prevention of stroke in non-valvular atrial fibrillation in the Netherlands. Plos One. 2014;9(8):e103974. pmid:25093723
- 9. Lanitis T, Kongnakorn T, Jacobson L, De Geer A. Cost-effectiveness of apixaban versus warfarin and aspirin in Sweden for stroke prevention in patients with atrial fibrillation. Thrombosis research. 2014;134(2):278–87. pmid:24935675
- 10. Dorian P, Kongnakorn T, Phatak H, Rublee DA, Kuznik A, Lanitis T, et al. Cost-effectiveness of apixaban vs. current standard of care for stroke prevention in patients with atrial fibrillation. Eur Heart J. 2014;35(28):1897–906. pmid:24513791
- 11. Ademi Z, Pasupathi K, Liew D. Cost-effectiveness of apixaban compared to warfarin in the management of atrial fibrillation in Australia. European journal of preventive cardiology. 2015;22(3):344–53. pmid:24281250
- 12. Lee S, Mullin R, Blazawski J, Coleman CI. Cost-effectiveness of apixaban compared with warfarin for stroke prevention in atrial fibrillation. Plos One. 2012;7(10):e47473. pmid:23056642
- 13. Chiang CE, Wang KL, Lip GY. Stroke prevention in atrial fibrillation: an Asian perspective. Thrombosis and haemostasis. 2014;111(5):789–97. pmid:24500243
- 14. January CT, Wann LS, Alpert JS, Calkins H, Cigarroa JE, Cleveland JC Jr., et al. 2014 AHA/ACC/HRS guideline for the management of patients with atrial fibrillation: executive summary: a report of the American College of Cardiology/American Heart Association Task Force on practice guidelines and the Heart Rhythm Society. Circulation. 2014;130(23):2071–104. pmid:24682348
- 15. Lip GY, Kongnakorn T, Phatak H, Kuznik A, Lanitis T, Liu LZ, et al. Cost-effectiveness of apixaban versus other new oral anticoagulants for stroke prevention in atrial fibrillation. Clinical therapeutics. 2014;36(2):192–210.e20. pmid:24508420
- 16. The Government of Hong Kong Special Administrative Region. Hong Kong the facts: public health 2013 [cited 2014 September 22nd]. Available: http://www.gov.hk/en/about/abouthk/factsheets/docs/public_health.pdf.
- 17. Connolly SJ, Pogue J, Eikelboom J, Flaker G, Commerford P, Franzosi MG, et al. Benefit of oral anticoagulant over antiplatelet therapy in atrial fibrillation depends on the quality of international normalized ratio control achieved by centers and countries as measured by time in therapeutic range. Circulation. 2008;118(20):2029–37. pmid:18955670
- 18. Reynolds MW, Fahrbach K, Hauch O, Wygant G, Estok R, Cella C, et al. Warfarin anticoagulation and outcomes in patients with atrial fibrillation: a systematic review and metaanalysis. Chest. 2004;126(6):1938–45. pmid:15596696
- 19. Claassen DO, Kazemi N, Zubkov AY, Wijdicks EF, Rabinstein AA. Restarting anticoagulation therapy after warfarin-associated intracerebral hemorrhage. Arch Neurol. 2008;65(10):1313–8. pmid:18852344
- 20. Sabir I, Khavandi K, Brownrigg J, Camm AJ. Oral anticoagulants for Asian patients with atrial fibrillation. Nat Rev Cardiol. 2014;11(5):290–303. pmid:24614113
- 21. Ho JCS, Chang AM, Yan BP, Yu CM, Lam YY, Lee VWY. Dabigatran Compared With Warfarin for Stroke Prevention With Atrial Fibrillation: Experience in Hong Kong. Clinical cardiology. 2012;35(12):E40–5. pmid:23108916
- 22. Sullivan PW, Slejko JF, Sculpher MJ, Ghushchyan V. Catalogue of EQ-5D scores for the United Kingdom. Medical decision making: an international journal of the Society for Medical Decision Making. 2011;31(6):800–4.
- 23. Gage BF, Cardinalli AB, Owens DK. The effect of stroke and stroke prophylaxis with aspirin or warfarin on quality of life. Archives of internal medicine. 1996;156(16):1829–36. pmid:8790077
- 24. National Institute for Health and Care Excellence. Guide to the methods of technology appraisal 2013. 2013 [cited 2015 July 7th]. Available: https://www.nice.org.uk/article/pmg9/chapter/Foreword.
- 25. Hong Kong Hospital Authority. Hospital Authority Ordinance (Chapter 113): Revisons to list of charges 2013 [cited 2014 October 20th]. Available: http://www.ha.org.hk/visitor/ha_visitor_index.asp?Content_ID=10045&Lang=ENG.
- 26. Hong Kong Government Gazette. Provison for Public Charges—Non-Eligible Persons 2003 [cited 2014 September 22th]. Available: https://www.ha.org.hk/haho/ho/cs/207298en.pdf.
- 27. The World Bank. Hong Kong GDP per capita 1960–2014 [cited 2014 November 26th]. Available: http://www.tradingeconomics.com/hong-kong/gdp-per-capita.
- 28. The Commission on Macroeconomics and Health of the World Health Organization. Macroeconomics and health: investing in health for economic development 2001 [cited 2014 Oct 10th]. Available: http://whqlibdoc.who.int/publications/2001/924154550x.pdf.
- 29. Feigin VL, Forouzanfar MH, Krishnamurthi R, Mensah GA, Connor M, Bennett DA, et al. Global and regional burden of stroke during 1990–2010: findings from the Global Burden of Disease Study 2010. The Lancet. 2014;383(9913):245–55.
- 30. Suzman R, Beard J. Global health and ageing 2011 [cited 2015 May 10th]. Available: http://www.who.int/ageing/publications/global_health/en/.
- 31. Morgan CL, McEwan P, Tukiendorf A, Robinson PA, Clemens A, Plumb JM. Warfarin treatment in patients with atrial fibrillation: Observing outcomes associated with varying levels of INR control. Thrombosis research. 2009;124(1):37–41. pmid:19062079
- 32. White HD, Gruber M, Feyzi J, Kaatz S, Tse HF, Husted S, et al. Comparison of outcomes among patients randomized to warfarin therapy according to anticoagulant control: results from SPORTIF III and V. Archives of internal medicine. 2007;167(3):239–45. pmid:17296878
- 33. Ho CW, Ho MH, Chan PH, Hai JJ, Cheung E, Yeung CY, et al. Ischemic stroke and intracranial hemorrhage with aspirin, dabigatran, and warfarin: impact of quality of anticoagulation control. Stroke. 2015;46(1):23–30. pmid:25406148
- 34. Wan Y, Heneghan C, Perera R, Roberts N, Hollowell J, Glasziou P, et al. Anticoagulation control and prediction of adverse events in patients with atrial fibrillation: a systematic review. Circulation Cardiovascular quality and outcomes. 2008;1(2):84–91. pmid:20031794
- 35. Albertsen IE, Rasmussen LH, Overvad TF, Graungaard T, Larsen TB, Lip GY. Risk of stroke or systemic embolism in atrial fibrillation patients treated with warfarin: a systematic review and meta-analysis. Stroke. 2013;44(5):1329–36. pmid:23482597
- 36. Clancy CM, Cronin K. Evidence-based decision making: global evidence, local decisions. Health Aff (Millwood). 2005;24(1):151–62.
- 37. Shen AY, Yao JF, Brar SS, Jorgensen MB, Chen W. Racial/ethnic differences in the risk of intracranial hemorrhage among patients with atrial fibrillation. Journal of the American College of Cardiology. 2007;50(4):309–15. pmid:17659197
- 38. van Asch CJJ, Luitse MJA, Rinkel GJE, van der Tweel I, Algra A, Klijn CJM. Incidence, case fatality, and functional outcome of intracerebral haemorrhage over time, according to age, sex, and ethnic origin: a systematic review and meta-analysis. The Lancet Neurology. 2010;9(2):167–76. pmid:20056489
- 39. Wang C, Yang Z, Wang C, Wang Y, Zhao X, Liu L, et al. Significant underuse of warfarin in patients with nonvalvular atrial fibrillation: results from the china national stroke registry. J Stroke Cerebrovasc Dis. 2014;23(5):1157–63. pmid:24291013
- 40. Yasaka M, Lip GY. Impact of non-vitamin k antagonist oral anticoagulants on intracranial bleeding in Asian patients with non-valvular atrial fibrillation. Circulation Journal. 2014;78(10):2367–72. pmid:25213003
- 41. Chan EW, Lau WCY, Siu CW, Lip GYH, Leung WK, Anand S, et al. Effect of suboptimal anticoagulation treatment with antiplatelet therapy and warfarin on clinical outcomes in patients with nonvalvular atrial fibrillation: A population-wide cohort study. Heart Rhythm. 2016 Mar 28. pii: S1547-5271(16)30150-3.
- 42. Witt DM, Delate T, Garcia DA, Clark NP, Hylek EM, Ageno W, et al. Risk of thromboembolism, recurrent hemorrhage, and death after warfarin therapy interruption for gastrointestinal tract bleeding. Archives of internal medicine. 2012;172(19):1484–91. pmid:22987143
- 43. Warren J Manning DES, Gregory YH Lip. Atrial fibrillation: Anticoagulant therapy to prevent embolization: UpToDate; 2015 [cited 2015 July 14th]. Available: http://www.uptodate.com/contents/atrial-fibrillation-anticoagulant-therapy-to-prevent-embolization.
- 44. Hohnloser SH, Hijazi Z, Thomas L, Alexander JH, Amerena J, Hanna M, et al. Efficacy of apixaban when compared with warfarin in relation to renal function in patients with atrial fibrillation: insights from the ARISTOTLE trial. Eur Heart J. 2012;33(22):2821–30. pmid:22933567