Author + information
- Received February 16, 2016
- Revision received June 2, 2016
- Accepted June 8, 2016
- Published online September 13, 2016.
- Jan Steffel, MDa,
- Robert P. Giugliano, MD, SMb,
- Eugene Braunwald, MDb,
- Sabina A. Murphy, MPHb,
- Michele Mercuri, MDc,
- Youngsook Choi, MDc,
- Phil Aylward, MDd,
- Harvey White, MDe,
- Jose Luis Zamorano, MDf,
- Elliott M. Antman, MDb and
- Christian T. Ruff, MD, MPHb,∗ ()
- aDepartment of Cardiology, University Heart Center Zurich, Zurich, Switzerland
- bTIMI Study Group, Cardiovascular Division, Brigham and Women’s Hospital, Boston, Massachusetts
- cDaiichi-Sankyo Pharma Development, Edison, New Jersey
- dFlinders Medical Centre, Adelaide, Australia
- eAuckland City Hospital, Auckland, New Zealand
- fRamón y Cajal University Hospital, Madrid, Spain
- ↵∗Reprint requests and correspondence:
Dr. Christian T. Ruff, TIMI Study Group, 350 Longwood Avenue, 1st Floor Offices, Boston, Massachusetts 02115.
Background Anticoagulation is often avoided in patients with atrial fibrillation who are at an increased risk of falling.
Objectives This study assessed the relative efficacy and safety of edoxaban versus warfarin in the ENGAGE AF–TIMI 48 (Effective Anticoagulation with Factor Xa Next Generation in Atrial Fibrillation–Thrombolysis In Myocardial Infarction 48) trial in patients with atrial fibrillation judged to be at increased risk of falling.
Methods We performed a pre-specified analysis of the ENGAGE AF–TIMI 48, comparing patients with versus without increased risk of falling.
Results Nine hundred patients (4.3%) were judged to be at increased risk of falling. These patients were older (median, 77 vs. 72 years; p < 0.001), and had a higher prevalence of comorbidities including prior stroke/transient ischemic attack, diabetes, and coronary artery disease. After multivariable adjustment, patients at increased risk of falling experienced more bone fractures caused by falling (adjusted hazard ratio [HRadj]: 1.88; 95% confidence interval [CI]: 1.49 to 2.38; p < 0.001), major bleeding (HRadj: 1.30; 95% CI: 1.04 to 1.64; p = 0.023), life-threatening bleeding (HRadj: 1.67; 95% CI: 1.11 to 2.50; p = 0.013), and all-cause death (HRadj: 1.45; 95% CI: 1.23 to 1.70; p < 0.001), but not ischemic events including stroke/systemic embolic event (HRadj: 1.16; 95% CI: 0.89 to 1.51; p = 0.27). No treatment interaction was observed between either dosing regimens of edoxaban and warfarin for the efficacy and safety outcomes. Treatment with edoxaban resulted in a greater absolute risk reduction in severe bleeding events and all-cause mortality compared with warfarin.
Conclusions Edoxaban is an attractive alternative to warfarin in patients at increased risk of falling, because it is associated with an even greater absolute reduction in severe bleeding events and mortality. (Effective aNticaoGulation with factor xA next Generation in Atrial Fibrillation [ENGAGE AF-TIMI 48]; NCT00781391)
Stroke is the most frequent serious complication in patients with atrial fibrillation (AF) (1). Oral anticoagulation (OAC) with direct factor IIa and factor Xa inhibitors has been shown to be as, if not more, effective than vitamin K antagonists (VKAs) for prevention of stroke or systemic embolic event (SEE) in patients with AF (2–5). Despite proven benefits, many patients do not receive OACs (6), and patients at an increased fall risk are at particular risk of undertreatment. This is largely driven by the perception that risks of bleeding, particularly intracranial hemorrhage (ICH), outweigh risk of stroke in these patients (7,8). Data for non-VKA OACs (NOACs) in patients at risk of falls are limited, although NOACs have consistently been linked with an approximately 50% reduced risk for intracranial bleeding as compared with VKAs (2–5), and may represent an attractive option for anticoagulation in patients with an increased risk of falls.
In the ENGAGE AF–TIMI 48 (Effective Anticoagulation with Factor Xa Next Generation in Atrial Fibrillation–Thrombolysis In Myocardial Infarction 48) trial, 2 dosing regimens of the oral direct factor Xa inhibitor edoxaban were compared with well-managed warfarin (median time in therapeutic range, 68.4%) for prevention of stroke/SEE in patients with AF (2). Both higher dose edoxaban regimen (HDER) and lower dose edoxaban regimen (LDER) were noninferior to warfarin for prevention of stroke and SEE in patients with AF. Both were associated with significantly less bleeding, including ICH. HDER is approved in the United States, Europe, and several Asian countries for use in patients with AF and venous thromboembolism. Because of an increase in ischemic stroke with LDER compared with warfarin, the LDER was not submitted for regulatory approval. The current pre-specified subgroup analysis was designed to assess risk and compare relative efficacy and safety of edoxaban versus warfarin in patients identified to be at an increased risk of falls.
Population and procedures
The design and results of the phase 3 ENGAGE AF–TIMI 48 trial have been reported previously (2,9). Briefly, patients with AF at moderate-to-high risk of stroke (n = 21,105) were randomized to receive warfarin continually adjusted to an international normalized ratio of 2.0 to 3.0, HDER (60 mg once daily), or LDER (30 mg once daily). In patients weighing ≤60 kg, a calculated creatinine clearance of 30 to 50 ml/min using the Cockcroft-Gault equation (10), or who required concomitant use of verapamil, quinidine, or dronedarone, edoxaban dose was reduced by 50% (i.e., HDER, 60 mg ≥30 mg; LDER, 30 mg ≥15 mg). Key exclusion criteria included patients with a contraindication for anticoagulation; presence of a condition associated with high risk of bleeding (e.g., history of intracranial, intraocular, spinal, retroperitoneal, or intra-articular bleeding); overt gastrointestinal bleeding or active ulcer within the previous year; uncontrolled hypertension (blood pressure >170/100 mm Hg); or hemorrhagic disorder, dual-antiplatelet therapy, or concomitant use of medications that increase the risk of bleeding (including chronic nonaspirin nonsteroidal anti-inflammatory drug use for ≥4 days/week) (9).
Based on previous studies (11–19), investigators categorized patients as having an increased risk of falling if they had any of the following 8 criteria at randomization: 1) prior history of falls; 2) lower extremity weakness; 3) poor balance; 4) cognitive impairment; 5) orthostatic hypotension; 6) use of psychotropic drugs; 7) severe arthritis; or 8) dizziness. In addition, a modified Charlson Comorbidity Index (CCI) (20) was calculated post hoc to independently assess the burden of comorbidities in patients who were and were not deemed to be at an increased risk of falling. The following adjustments were made to fit the CCI to the data available from the ENGAGE-AF TIMI 48 database: a Rankin score 3 to 5 was used instead of “hemiplegia”; rheumatoid arthritis was used for “rheumatic disease”; and patients with severe renal and severe hepatic disease, and patients with active cancer were excluded from ENGAGE AF–TIMI 48.
The primary efficacy endpoint of ENGAGE AF–TIMI 48 was stroke/SEE; other efficacy endpoints included ischemic stroke, hemorrhagic stroke, all-cause, and cardiovascular mortality (2). The principal safety endpoint was major bleeding as per the International Society on Thrombosis and Haemostasis definition (21); other safety endpoints included life-threatening bleeding, fatal bleeding, clinically relevant nonmajor bleeding, and ICH (9). All cardiovascular, cerebrovascular, bleeding, and death events were adjudicated by an independent committee blinded to randomized treatment assignment. The 3 net clinical endpoints were: 1) combination of death/stroke/SEE or major bleed (primary net endpoint); 2) death, disabling stroke, or life-threatening bleed (secondary net clinical endpoint); and 3) death, stroke, SEE, or life-threatening bleed (tertiary net clinical endpoint). Investigators reported bone fractures and fractures caused by falls using the electronic case record form; these endpoints were not independently adjudicated.
Baseline characteristics were compared with the Wilcoxon rank sum test and chi-square test for continuous and categorical variables, respectively. Efficacy analyses were performed in the intention-to-treat study population and bleeding outcomes in the safety population (including patients who had received at least 1 dose of study drug). Event rates are expressed per 100 patient-years. Absolute risk reductions (ARR) were calculated for the difference in event rates between the warfarin and edoxaban groups, stratified by risk of falling.
Hazard ratios (HRs) with 95% confidence intervals (CIs) comparing risk of falling were calculated using a Cox proportional hazards model. A multivariable model for fall risk adjusted for differences in baseline characteristics between patients with versus without fall risk as outlined in Table 1, including sex, age, weight, AF type, VKA naive, history of stroke/transient ischemic attack, hypertension, coronary artery disease, congestive heart failure, diabetes, and creatinine clearance <50 ml/min at randomization. Aspirin use at randomization and randomization group were also included in this model.
HRs with 95% CIs comparing the relative efficacy and safety of edoxaban versus warfarin for the subgroup of increased fall risk were calculated with the Cox proportional hazards models, with treatment as a covariate along with the stratification factors of CHADS2 score and dose-adjustment status. Adjusted HRs (adj HRs) and 95% CI comparing the relative efficacy and safety of edoxaban versus warfarin were calculated with the addition of adjustment for differences in baseline characteristics across randomized treatment groups in patients at increased risk of falling (Table 2: diastolic blood pressure, diabetes, and weight).
All tests were 2-sided with a p value <0.05 considered to be significant. The TIMI Study Group has an independent copy of the trial database and conducted all analyses. Analyses were performed with use of Stata/SE version 13.1 (Stata Corp., College Station, Texas).
Baseline characteristics in patients with and without an increased risk of falls are summarized in Table 1. Nine hundred patients (4.3%) were judged at randomization to be at increased risk of falling. These patients were on average older; more frequently female; and had a higher incidence of comorbidities, including coronary artery disease, diabetes, and moderate renal insufficiency. Patients with increased fall risk had a higher stroke risk (CHA2DS2-VASc), and more frequent history of stroke or transient ischemic attack. In contrast, these patients were less likely to have a history of congestive heart failure or to be VKA naive. Among patients randomized to warfarin, median time in therapeutic range (i.e., percentage of time patients had an international normalized ratio of 2.0 to 3.0) was not different between the 2 groups (67.3% vs. 68.5%; p = 0.15). A significantly higher percentage of patients at an increased risk of falls received a dose reduction of edoxaban (according to the predefined criteria [body weight ≤60 kg; creatinine clearance 30 to 50 ml/min; or concomitant use of verapamil, quinidine, or dronedarone]) compared with those not at an increased risk (38.0% vs. 24.8%; p < 0.001). Patients at increased risk of falling had a higher mean CCI (3.2 vs. 2.8; p < 0.001).
Baseline characteristics of patients at increased fall risk across randomized treatment assignment are summarized in Online Table 1. Patient characteristics were similar among the 3 treatment groups, except for lower diastolic pressure and higher prevalence of patients with weight <50 kg in the warfarin arm, and higher prevalence of diabetes in the HDER arm. There were no differences in baseline characteristics per randomization group in patients not deemed to be at increased risk for falls (Online Table 2).
Outcomes in patients with versus without increased fall risk
Outcomes are summarized in Table 2. On univariate analysis, an increased fall risk was associated with increased risk of ischemic and bleeding events. After multivariable adjustment, patients at increased risk of falling had a significantly increased risk of bone fractures caused by falling (adj HR: 1.88; 95% CI: 1.49 to 2.38; p < 0.001), major bleeding (adj HR: 1.30; 95% CI: 1.04 to 1.64; p = 0.023), life-threatening bleeding (adj HR: 1.67; 95% CI: 1.11 to 2.50; p = 0.013), and all-cause death (adj HR: 1.45; 95% CI: 1.23 to 1.70; p < 0.001) and all 3 net clinical endpoints. In contrast, the risk of ischemic events, including stroke/SEE (adj HR: 1.16; 95% CI: 0.89 to 1.51; p = 0.267), ischemic stroke (adj HR: 1.09; 95% CI: 0.81 to 1.48; p = 0.570), and myocardial infarction (adj HR: 0.86; 95% CI: 0.55 to 1.34; p = 0.51), was not elevated in patients at increased fall risk.
A higher degree of frailty as assessed by the CCI was associated with an elevated risk for ischemic and bleeding events, including stroke/SEE, major bleeding, and all-cause mortality (all p < 0.001) (Online Table 3), but not fatal bleeding or ICH. Importantly, an increased fall risk remained associated with worse outcomes, both ischemic and bleeding, even after adjusting for CCI (Online Table 4). Even after adjusting for both the CCI and differences in baseline parameters, increased fall risk remained associated with a higher risk for bleeding events, bone fractures caused by falling, and mortality, but not stroke/SEE (Online Table 5).
Efficacy of edoxaban versus warfarin in patients with increased fall risk
No treatment interaction for efficacy was observed in patients stratified by increased fall risk (Figure 1). A sensitivity analysis adjusting for differences in baseline characteristics (diastolic pressure, diabetes, and weight) across randomized treatment groups in patients with an increased fall risk demonstrated consistent results. There were no significant treatment interactions for increased risk of falling regarding the net clinical outcome in patients on HDER versus warfarin (Figure 1). Relative efficacy with LDER compared with warfarin was qualitatively similar regardless of fall risk (Online Figure 1).
Safety of edoxaban versus warfarin in patients with increased fall risk
There was no significant treatment-subgroup interaction for the safety endpoints (Figure 2). Results were consistent after multivariable adjustment for differences in baseline diastolic pressure, diabetes, and weight across randomized treatment groups in patients with an increased fall risk (Online Table 1). No fatal bleeding occurred in patients with increased fall risk treated with edoxaban, versus 4 in the warfarin group.
Because the absolute rate of severe bleeding events was substantially higher in patients at increased risk of falling, HDER demonstrated greater ARR in patients with versus without increased risk of falling for hemorrhagic stroke (ARR, −64/10,000 patient-years [95% CI: −146 to +18] vs. −20/10,000 patient-years [95% CI: −32 to −8]), ICH (ARR, −175 events/10,000 patient-years [95% CI: −297 to −52] vs. −39 events/10,000 patient-years [95% CI: −56 to −22]), life-threatening bleeding (ARR, −106 events/10,000 patient-years [95% CI: −230 to +18] vs. −31 events/10,000 patient-years [95% CI: −47 to −15]), and all-cause mortality (ARR, −66 [95% CI: −331 to +199] vs. −34 [95% CI: −75 to +7]) (Central Illustration). The calculated number needed to treat using HDER over warfarin to avoid 1 severe bleeding event or death was lower in patients with versus without increased risk of falls (Central Illustration). The results with LDER were similar (Online Figure 2), with an even greater ARR in bleeding events for patients with versus without increased risk of falling (Online Figure 3).
With aging of the population, there will be an increase in the number of frail patients with AF. Anticoagulation is often not prescribed to such patients because of their perceived fall risk. Our data indicate that patients at increased risk of falling are at elevated absolute risk of thromboembolic and bleeding events. After adjusting for differences in baseline variables, an increased risk of falling remained significantly associated with increased risk of bone fractures because of falling, bleeding events, and mortality (all-cause and cardiovascular). This remained true even after adjusting for frailty, as assessed by the modified CCI. Taken together, these results indicate an incremental value of fall risk assessment beyond established risk factors and frailty to estimate the propensity for falling and the risk of bleeding complications and mortality. In contrast, the risk of falling was not associated with ischemic events (stroke/SEE, myocardial infarction) after adjusting for differences in baseline variables. The 8 factors used to evaluate an “increased risk of falling” in ENGAGE AF–TIMI 48 are easily assessable, in contrast to many of the available frailty and fall risk scores, which are based on up to 70 (or more) laborious and time-consuming clinical and functional tests (22–26). As such, estimating fall risk using the criteria in ENGAGE AF–TIMI 48 may represent a versatile tool to quickly gauge these patients in daily clinical practice.
In prior studies, increased fall risk has not been consistently linked with an increased risk of severe bleeding or worse outcome. In a study of medical patients discharged on anticoagulation, the risk of major bleeding over the ensuing 12 months was similar in those with and without a high risk of falls (8.0 vs. 6.8 per 100 patient-years; p = 0.64) (27). In contrast, and similar to our results, patients with a prior history of falls on anticoagulation in the Loire Valley Atrial Fibrillation Project had an increased risk for major bleeding (HR: 3.32; 95% CI: 1.23 to 8.91; p < 0.02), but not ICH. Notably, stroke and SEE as well as all-cause mortality were independently increased in these patients (28).
In this analysis, we observed a consistent relative efficacy and safety of edoxaban compared with warfarin in patients with an increased risk of falling. Together with the results of other high-risk subgroup analyses from ENGAGE AF–TIMI 48, such as the elderly (29), patients on antiplatelet therapy (30), renal dysfunction (31), or with congestive heart failure (32), our results support the overall favorable efficacy and enhanced safety of edoxaban in these difficult to treat high-risk patients. Because patients at increased risk of falling demonstrate a higher risk of events, treatment with edoxaban resulted in a greater ARR for the most severe bleeding events (i.e., intracranial, life-threatening, or fatal bleeding as compared with patients not at increased risk of falls). Because anticoagulation for AF is usually a life-long treatment, the number needed to treat becomes even more favorable for HDER. Fewer than 6 patients would need to be treated for 10 years with HDER instead of warfarin to prevent either an ICH or the occurrence of 1 secondary net clinical outcome (i.e., 1 of the most severe clinical events, such as death, disabling stroke, or life-threatening bleeding). This effect was even more pronounced in LDER, because of even lower bleeding rates, although this regimen is not yet approved. Notably, there were no fatal bleeding events in either edoxaban regimen among patients at increased risk of falls. A higher proportion of patients at an increased risk of falling qualify for a dose adjustment that has previously been shown to provide similar efficacy with even greater relative safety in eligible patients (33).
Patients at increased fall risk are at risk of undertreatment with anticoagulants in spite of an elevated risk for thromboembolism (7,8). These patients are at increased risk for ischemic events by virtue of their age (average age 77 years in this study) and burden of comorbidities. However, despite the proven benefit of OAC, the accompanying increased risk of bleeding in these patients results in routine undertreatment with anticoagulation (6). The risk of ischemic events including stroke is often overlooked or underestimated, yet an ischemic stroke in these patients may result in loss of independence and quality of life, and is associated with substantial morbidity and mortality (34). Several studies have indicated an overall benefit from anticoagulation in patients at increased risk of falling (35,36), indicating that the risk of severe bleeding is counterbalanced by a similar if not greater reduction in the risk for stroke. In clinical practice, it is imperative to conduct an individual risk-benefit assessment before initiation of anticoagulant treatment in these high-risk patients.
In the absence of a uniform definition, the determination that a patient was at increased fall risk was based on several recognized risk factors (11–19). To correlate investigator-derived judgment with an objective measure of fall risk or frailty, the available established scores were evaluated. All turned out to be unsuitable because various relevant parameters could not be obtained retrospectively from our database (e.g., gait instability, strength by handgrip, and so forth). However, patients at increased risk of falls did show a higher rate of falls leading to bone fractures, internally validating the prospective evidence-based “falls risk” assessments by the investigators. Moreover, patients deemed to be at increased risk of falls had a higher modified CCI, which is linked to a higher risk of falls (37–39). Although a clear pattern of a larger ARR was observed with both HDER and LDER, the magnitude of this effect remains elusive because of the small number of events, with some ARR CI crossing the zero boundary. It is also possible that the greater absolute benefit of edoxaban compared with warfarin in patients at risk of falling is accounted for by other factors, such as age, but our study had limited power to explore subgroup-subgroup analyses. Because of the lack of a placebo arm, our study is unable to address whether patients with AF at increased risk of falling generally profit from anticoagulation. Similarly, the lack of association of an independently increased risk for ischemic events in our patients at risk of falling does not imply the lack of a treatment benefit, because they were all anticoagulated. Any subgroup analysis must be interpreted as hypothesis generating. Although baseline parameters among treatment groups differed only slightly in 3 variables (percentage of patients with diabetes, body weight ≤50 kg, diastolic blood pressure), “increased risk of falling” was neither randomized nor stratified, leaving the possibility of unmeasured residual confounding and bias through multiple testing. Finally, a selection/referral bias cannot be excluded, such that patients at highest risk of falling may not be included in the study.
In the ENGAGE AF–TIMI 48 study, increased risk of falling is independently associated with an elevated risk for major bleeding, fractures caused by a fall, and mortality, even after adjusting for frailty and differences in baseline variables. Data from this pre-specified subgroup analysis demonstrate a consistent relative efficacy and enhanced safety of edoxaban compared with well-managed warfarin in these high-risk patients. Because of the higher rate of severe bleeding complications in patients at increased risk of falling, treatment with edoxaban resulted in a greater ARR in severe bleeding events and all-cause mortality in these patients. Hence, our data indicate that edoxaban may represent an attractive option of anticoagulation in this difficult to treat patient population.
COMPETENCY IN PATIENT CARE AND PROCEDURAL SKILLS: Patients with AF at risk of falling face higher rates of bleeding, bone fractures, and mortality related in part to patient age. Anticoagulation with edoxaban was associated with lower rates of severe bleeding and mortality than with warfarin.
TRANSLATIONAL OUTLOOK: Future studies should address dosing regimens of anticoagulants in patients with AF at risk of falling that achieve optimum efficacy and safety.
For supplemental tables and figures, please see the online version of this article.
The ENGAGE AF–TIMI 48 study was supported by Daiichi-Sankyo Pharma Development. Dr. Steffel has received consultant and/or speaker fees from Amgen, AstraZeneca, Atricure, Bayer, Biotronik, Biosense Webster, Boehringer-Ingelheim, Boston Scientific, Bristol-Myers Squibb, Daiichi-Sankyo, Cook Medical, Medtronic, Novartis, Pfizer, Roche, Sanofi, Sorin, St. Jude Medical, and Zoll; grant support through his institution from Bayer Healthcare, Biosense Webster, Biotronik, Boston Scientific, Daiichi-Sankyo, Medtronic, and St. Jude Medical; and is codirector of CorXL. Dr. Giugliano has received consulting fees from the American College of Cardiology, Boehringer-Ingelheim, Bristol-Myers Squibb, Janssen, Daiichi-Sankyo, Merck, Pfizer, Portola, and Sanofi; and grant support to his institution from Daiichi-Sankyo, Merck, Johnson & Johnson, Sanofi, and AstraZeneca. Dr. Braunwald has received consulting fees from The Medicines Company and Theravance; lecture fees from Menarini and Medscape; grant support through his institution from Daiichi-Sankyo, AstraZeneca, Merck & Co., Novartis, and GlaxoSmithKline; served as an unpaid consultant for Merck and Novartis; and provided uncompensated lectures for Merck & Co. Ms. Murphy has received grant support through her institution from Daiichi-Sankyo. Dr. Mercuri is an employee of Daiichi-Sankyo; and has a pending patent related to the clinical properties of edoxaban. Dr. Choi is an employee of Daiichi-Sankyo. Dr. Aylward has received research support from Daiicho-Sankyo. Dr. White has received a research grant from Sanofi, Eli Lilly and Company, National Institutes of Health, Merck Sharpe & Dohme, AstraZeneca, GlaxoSmithKline, George Institute, Omthera Pharmaceuticals, Pfizer New Zealand, Intarcia Therapeutics Inc., Elsai Inc., DalGen Products and Services, and Daiichi-Sankyo Pharma Development; and consulting fees from AstraZeneca. Dr. Zamorano is an investigator on the Daiichi-Sankyo Engage trial. Dr. Antman has received grant support through his institution from Daiichi Sankyo. Dr. Ruff has received consulting fees from Bayer, Daiichi-Sankyo, Portola, and Boehringer Ingelheim; and grant support through his institution from Daiichi-Sankyo, AstraZeneca, Eisai, and Intarcia. Parts of this paper were presented as a poster at the American Heart Association Scientific Sessions 2015 in Orlando, Florida.
- Abbreviations and Acronyms
- atrial fibrillation
- absolute risk reduction
- Charlson Comorbidity Index
- confidence interval
- higher dose edoxaban regimen
- hazard ratio
- intracranial hemorrhage
- lower dose edoxaban regimen
- non-vitamin K oral anticoagulant
- oral anticoagulation
- systemic embolic event
- vitamin K antagonist
- Received February 16, 2016.
- Revision received June 2, 2016.
- Accepted June 8, 2016.
- 2016 American College of Cardiology Foundation
- Camm A.J.,
- Lip G.Y.,
- De Caterina R.,
- et al.
- Giugliano R.P.,
- Ruff C.T.,
- Braunwald E.,
- et al.
- Ruff C.T.,
- Giugliano R.P.,
- Antman E.M.,
- et al.
- Rice L.A.,
- Ousley C.,
- Sosnoff J.J.
- Muir S.W.,
- Gopaul K.,
- Montero Odasso M.M.
- van Hateren K.J.,
- Kleefstra N.,
- Blanker M.H.,
- et al.
- Schulman S.,
- Kearon C.
- Rolfson D.B.,
- Majumdar S.R.,
- Tsuyuki R.T.,
- Tahir A.,
- Rockwood K.
- Nyberg L.,
- Gustafson Y.
- Kato E.T.,
- Giugliano R.P.,
- Ruff C.T.,
- et al.
- Xu H.,
- Ruff C.T.,
- Giugliano R.P.,
- et al.
- Bohula E.A.,
- Giugliano R.P.,
- Ruff C.T.,
- et al.
- Magnani G.,
- Giugliano R.P.,
- Ruff C.T.,
- et al.
- Saposnik G.,
- Black S.E.,
- Hakim A.,
- et al.
- Ma M.K.,
- Yap D.Y.,
- Yip T.P.,
- et al.
- Li M.,
- Tomlinson G.,
- Naglie G.,
- et al.