You are here

Severity of Gastrointestinal Bleeding in Patients Treated with Direct-Acting Oral Anticoagulants

Am J Med. 2018;131(5):573.e9-573.e15

Abstract

Background

Direct-acting oral anticoagulants (DOACs), which have gained approval for stroke prevention in nonvalvular atrial fibrillation and treatment of venous thromboembolism, have become increasingly preferred over warfarin given their predictable pharmacodynamics, lack of required monitoring, and superior outcomes. Direct-acting oral anticoagulants have been shown to be associated with an increased frequency of gastrointestinal bleeding compared with warfarin, but the severity and characteristics of gastrointestinal bleeding in these patients is poorly understood.

Methods

We retrospectively evaluated electronic medical records of patients with gastrointestinal bleeding (n = 8496) from 2010-2016. We identified 61 patients with gastrointestinal bleeding episodes while treated with DOACs (rivaroxaban, dabigatran, or apixaban) and 123 patients with gastrointestinal bleeding while taking warfarin. We randomly selected a control group of 296 patients with gastrointestinal bleeding who were not receiving anticoagulation treatment from the same sample. Outcomes included the need for hospitalization, blood transfusion, endoscopic or surgical intervention, and 30-day mortality.

Results

The DOAC and warfarin groups were similar in terms of age and underlying comorbidity (assessed using the Charlson Comorbidity Index), but the DOAC group had greater concomitant aspirin use. Gastrointestinal bleeding was classified as upper (n = 186), lower (n = 88), anorectal (n = 183), small bowel (n = 9), and indeterminate (n = 14). After adjusting for differences in baseline variables, the DOAC group had fewer hospitalizations and required fewer transfusions than the warfarin group. The DOAC and control groups were not statistically different for all outcomes. There were no significant mortality differences among groups.

Conclusion

Although prior studies have shown a higher frequency of gastrointestinal bleeding in patients treated with DOACs compared with warfarin, our data suggest that gastrointestinal bleeding in patients taking DOACs may be less severe. These differences occurred despite significantly greater concomitant aspirin use in the DOAC group compared with warfarin users.

Keywords: Apixaban, Aspirin, Bleeding, Dabigatran, Direct-acting oral anticoagulant, Gastrointestinal, Hospitalization, Intervention, Mortality, Rivaroxaban, Severity, Transfusion, Warfarin.

 

 

  • Patients with gastrointestinal bleeding treated with direct-acting oral anticoagulants had less severe bleeding as assessed by the need for hospitalization and blood transfusion compared with patients treated with warfarin.

  • Mortality rates were similar in patients treated with direct-acting oral anticoagulants, warfarin, and no anticoagulation.

 

Clinical Significance

 

The direct-acting oral anticoagulants (DOACs), which include the anti-factor Xa agents apixaban and rivaroxaban and the direct thrombin inhibitor dabigatran, have recently been approved for stroke prevention in nonvalvular atrial fibrillation as well as venous thromboembolism. These drugs are often preferred because of their predictable pharmacodynamics, lack of need for required monitoring, and superior outcomes. 1 2 3 Additionally, DOACs have a shorter time to peak effect (1-4 hours) compared with warfarin (4-5 days), shorter half-lives (5-17 hours vs 40 hours with warfarin), and fewer drug–drug interactions compared with warfarin. 4

Although DOACs are associated with a significant decrease in the incidence of intracerebral hemorrhage and hemorrhagic stroke compared with patients treated with warfarin, 1 2 3 5 6 DOACs have been associated with an approximate 50% increase in the incidence of gastrointestinal bleeding compared with warfarin. 1 2 3 7 This has raised considerable concern about the absolute incidence of gastrointestinal bleeding associated with DOACs, because warfarin alone is associated with a 3-fold increase in risk of major gastrointestinal bleeding when compared with placebo, a risk increase that is doubled with the addition of antiplatelet agents. 7 Furthermore, costs are approximately 50% higher in patients who experience major gastrointestinal bleeding while being treated with warfarin compared with controls. 8

Although it is accepted that the incidence of gastrointestinal bleeding in patients taking DOACs is increased, there are few data about the source and severity of gastrointestinal bleeding in patients treated with DOACs. Here, we hypothesized that DOACs not only may be associated with an increased incidence of gastrointestinal bleeding but also are associated with more severe gastrointestinal bleeding. Therefore, we examined the severity of gastrointestinal bleeding in patients treated with DOACs, warfarin, and no anticoagulation (control patients) as assessed by the need for hospitalization, blood transfusions, endoscopic or surgical intervention, and 30-day mortality.

Methods

Study Population

We performed a retrospective case–control study of all adult patients (age ≥18 years) seen at the Medical University of South Carolina with gastrointestinal bleeding from January 1, 2010 to January 1, 2016. Patients were identified by an International Classification of Diseases, Ninth Revision Clinical Modification code search for gastrointestinal bleeding, which included the following: 569.3 (rectal and anal hemorrhage), 578.0 (hematemesis), 578.1 (blood in stool), and 578.9 (hemorrhage of gastrointestinal tract, unspecified). The electronic medical records of our study population were manually reviewed by 2 authors (MMB and TS) to confirm the diagnosis of gastrointestinal bleeding. If multiple encounters were identified, we analyzed the first encounter. We excluded patients with a clinical diagnosis of cirrhosis (owing to propensity for variceal bleeding), patients with severe thrombocytopenia (platelet count <50 × 10 3 ), patients with hematologic disorders, and patients treated with clopidogrel or new antiplatelet agents.

Gastrointestinal bleeding was identified in 8496 patients ( Figure ). The resulting study sample included 61 patients with gastrointestinal bleeding while being treated with DOACs (apixaban, rivaroxaban, or dabigatran) and 123 patients treated with warfarin. We then randomly selected a control group of 296 patients with gastrointestinal bleeding not receiving anticoagulation ( Figure ). The study cohort thus included 480 unique patients with gastrointestinal bleeding in the outpatient setting, the emergency department, or admitted to the hospital.

ajm14390-fig-0001
 

Figure
CONSORT diagram of patients. Shown is the flow of patients, illustrating treatment and control groups. Patients receiving no anticoagulation were randomly selected from the cohort of all patients with gastrointestinal bleeding. DOACs = direct-acting oral anticoagulants.

 

The study was approved by the institutional review board at the Medical University of South Carolina.

Data Collection and Definitions

We abstracted more than 60 unique variables for each patient at the time of presentation, including demographic, clinical, and historical data, such as the presence of previous gastrointestinal bleeding, medical comorbidities, indication for anticoagulation, concomitant aspirin use, endoscopic evaluation, source of bleeding, blood transfusions, hospital days, 30-day mortality, and complete laboratory data. Each subject's overall morbidity status was assessed using the Charlson Comorbidity Index (CCI), a formal scoring system including 22 medical comorbid conditions. 9 Clinical and laboratory variables reported were recorded at the time of presentation.

The diagnosis of gastrointestinal bleeding was made and confirmed by the documentation of witnessed hematemesis, melena, hematochezia, or any combination thereof by medical personnel accordging to the history and physical examination of the patient, with or without a positive fecal occult blood test results and with or without a significant drop in hemoglobin or hematocrit. Bleeding was characterized as upper, lower, anorectal, small bowel, or indeterminate on the basis of history and clinical presentation, history, and endoscopic findings. The finding of an endoscopic lesion documented to be consistent with bleeding in a specific location of the gastrointestinal tract assigned the source of bleeding to that portion of the gastrointestinal tract. With the exception of hematemesis, the source of bleeding was considered to be indeterminate if no endoscopy was performed. Indications for anticoagulation were determined according to review of history and clinical features.

Laboratory values included were those closest to the onset of bleeding. Laboratory values beyond 5 days from the onset of bleeding were excluded. Most patients who underwent endoscopy had the procedure performed acutely or early after the onset of gastrointestinal bleeding. Deferred endoscopic evaluation was included in this analysis only if the procedure occurred within 3 months and provided an explanation of the bleeding event.

Outcomes measured were need for hospitalization, number of hospital days, need for blood transfusion, number of transfusions, bleeding requiring intervention to achieve hemostasis, and 30-day mortality. Intervention included either endoscopic or surgical.

Statistical Analysis

Descriptive statistics were used to describe both overall and by study group analyses as means and standard deviations or as medians and 25th, 75th quartiles for continuous variables, and frequencies and percentages for categorical variables. Differences by group were assessed using χ 2 tests or Fisher's exact tests as appropriate for categorical variables, and Student's t tests, Wilcoxon rank sum tests, or general linear models as appropriate for continuous variables. Multivariable logistic regression was used to test group differences for 4 outcomes (hospitalization, need for transfusion, need for intervention, and 30-day mortality). Adjusted odds ratios, 95% confidence intervals (CI), and P values were reported. Linear mixed models were used to assess group differences in number of hospital days and number of required transfusions. Because the distribution of number of hospital days had several extreme outliers, the variable was winsorized before analysis. Beta estimates, standard errors, and P values were reported. Adjusted odds ratios, 95% CI, and P values for pairwise comparisons of mortality between DOAC versus warfarin and DOAC versus control were not reported owing to zero deaths in the DOAC group. All models were run for overall and pairwise differences and adjusted for age, sex, and CCI score. All analyses were performed using SAS version 9.4 (SAS Institute, Cary, NC). Hypothesis tests were 2-sided, and an α level of .05 was used to denote statistical significance.

Results

In our study population of 480 patients with gastrointestinal bleeding, there were 61 patients (13%) who were treated with DOACs (rivaroxaban = 29, dabigatran = 23, apixaban = 9), 123 patients (26%) who were treated with warfarin, and 296 patients (62%) who were not treated with anticoagulation ( Figure ). The source of gastrointestinal bleeding was characterized as upper in 186 patients, lower in 88 patients, anorectal in 183 patients, small bowel in 9 patients, and indeterminate in 14 patients.

The DOAC group was older, less likely to be female, more likely to be Caucasian, and more often had atrial fibrillation as the indication for anticoagulation compared with the warfarin and control groups ( Table 1 ). In pairwise comparisons (not shown in Table 1 ), the DOAC group had fewer females ( P = .008), was more likely to be Caucasian ( P < .001), more often had atrial fibrillation as the indication for anticoagulation ( P < .001), and had more aspirin use ( P < .001) than the warfarin group. The DOAC group was older ( P = .002), more likely to be Caucasian ( P < .001), and had a trend toward greater morbidity in terms of CCI in comparison with the control group. The warfarin group was older ( P = .02), more likely to be female ( P = .03), had greater morbidity in terms of CCI ( P = .05), and had less aspirin use ( P < .001) than the control group.

Characteristics Total
(n = 480)
DOAC
(n = 61)
Warfarin
(n = 123)
Control
(n = 296)
P Value
Age, mean ± SD 64 ± 13 68 ± 13 66 ± 14 63 ± 13 .002
Female, n (%) 233 (48.5) 23 (37.7) 72 (58.5) 138 (46.6) .02
Race, n (%)         <.001
Caucasian 248 (51.7) 48 (78.7) 63 (51.2) 137 (46.3)  
African-American 223 (46.5) 13 (21.3) 57 (46.3) 153 (51.7)  
Other 9 (1.9) 0 3 (2.4) 6 (2)  
Charlson Comorbidity Index, mean ± SD 4.1 ± 2.3 4.5 ± 2.0 4.3 ± 2.0 3.9 ± 2.4 .057
Indication for anticoagulation, n (%)         <.001
Atrial fibrillation/flutter 100 (20.8) 55 (90.2) 45 (36.6) 0  
DVT/PE 56 (11.7) 5 (8.2) 51 (41.5) 0  
AVR/MVR 17 (3.5) 0 17 (13.8) 0  
Other 11 (2.3) 1 (1.6) 10 (8.1) 0  
None 296 (61.7) 0 0 296 (100)  
Aspirin use, n (%) 163 (35.5) 26 (42.6) 17 (13.8) 120 (43.6) <.001
DOAC = direct-acting oral anticoagulant.

Table 1Baseline Demographic and Clinical Characteristics

 

There were no overall significant differences in the source of bleeding among the 3 groups ( Table 2 ). The warfarin group had more upper gastrointestinal bleeding and less anorectal bleeding than the control group ( Table 2 ).

DOAC = direct-acting oral anticoagulant; EGD = esophagogastroduodenoscopy; GI = gastrointestinal; INR = international normalized ratio; IQR = interquartile range.

Table 2Variables Related to Origin, Presentation, and Evaluation of Bleeding Events

 

Hemoglobin and hematocrit values at presentation were lower in the warfarin group compared with both the DOAC and control groups. There were no significant differences in hemoglobin and hematocrit values between the DOAC and control groups ( Table 2 ). Platelet counts were similar in all groups, and as expected, the international normalized ratio (INR) was higher in patients taking warfarin than in the other groups ( Table 2 ).

One or more endoscopy procedures were performed in 285 of the 480 patients (59.3%). The warfarin group had more frequent enteroscopies and less frequent colonoscopies compared with the control group ( Table 2 ).

Culprit lesions were those typically seen in patients with gastrointestinal bleeding ( Table 3 ). Culprit lesions were identified in 59% of patients (285 of 480), including 44% of DOAC patients (27 of 61), 48% of warfarin patients (59 of 123), and 67% of control patients (199 of 296). In the entire study population the most commonly identified culprit lesions were hemorrhoids (21.7%), peptic ulcers (13.5%), diverticula (6.5%), and arteriovenous malformations (5.0%).

Bleeding Source, Culprit Lesion Total
(n = 480)
DOAC
(n = 61)
Warfarin
(n = 123)
Control
(n = 296)
Upper GI        
Peptic ulcer disease 65 (13.5) 8 (13.1) 17 (13.8) 40 (13.5)
Arteriovenous malformations 17 (3.5) 2 (3.3) 4 (3.3) 11 (3.7)
Mallory-Weiss tears 5 (1.0) 0 0 5 (1.7)
Esophagitis/gastritis 7 (1.4) 1 (1.6) 2 (1.6) 4 (1.4)
Esophageal/gastric cancer 5 (1.0) 0 1 (0.8) 4 (1.4)
Other 5 (1.0) 1 (1.6) 2 (1.6) 2 (0.7)
No lesion found on endoscopy 36 (7.5) 9 (14.8) 13 (10.6) 14 (4.7)
Unknown lesion (no endoscopy) 46 (9.6) 4 (6.6) 19 (15.4) 23 (7.8)
Total 186 (38.8) 25 (41.0) 58 (47.2) 103 (34.8)
Lower GI        
Diverticula 31 (6.5) 2 (3.3) 10 (8.1) 19 (6.4)
Colorectal cancer 7 (1.5) 0 0 7 (2.4)
Inflammatory bowel disease 7 (1.5) 0 3 (2.4) 4 (1.4)
Postpolypectomy bleed 5 (1.0) 0 0 5 (1.7)
Radiation proctitis 4 (0.8) 0 1 (0.8) 3 (1.0)
Ischemic colitis 4 (0.8) 0 1 (0.8) 3 (1.0)
Other/nonspecific colitis 3 (0.6) 0 1 (0.8) 2 (0.7)
Other 2 (0.4) 0 0 2 (0.7)
Arteriovenous malformations 1 (0.2) 1 (1.6) 0 0
No lesion found on endoscopy 10 (2.1) 2 (3.3) 3 (2.4) 5 (1.7)
Unknown lesion (no endoscopy) 14 (2.9) 3 (4.9) 4 (3.3) 7 (2.4)
Total 88 (18.3) 8 (13.1) 23 (18.7) 57 (19.2)
Anorectal        
Hemorrhoids 104 (21.7) 10 (16.4) 14 (11.4) 80 (27.0)
Anal fissure 6 (1.3) 0 0 6 (2.0)
No lesion found on endoscopy 26 (5.4) 3 (4.9) 5 (4.1) 18 (6.1)
Unknown lesion (no endoscopy) 47 (9.8) 12 (19.7) 16 (13.0) 19 (6.4)
Total 183 (38.1) 25 (40.1) 35 (28.5) 123 (41.6)
Small bowel        
Arteriovenous malformations 6 (1.3) 1 (1.6) 3 (2.4) 2 (0.7)
Peptic ulcer disease 1 (0.2) 1 (1.6) 0 0
No lesion found on endoscopy 2 (0.4) 0 1 (0.8) 1 (0.3)
Unknown lesion (no endoscopy) 0 0 0 0
Total 9 (1.9) 2 (3.3) 4 (3.25) 3 (1.0)
Indeterminate        
No lesion found on endoscopy 9 (1.9) 0 2 (1.6) 7 (2.4)
Unknown lesion (no endoscopy) 5 (1.0) 1 (1.6) 1 (0.8) 3 (1.0)
Values are number (percentage). DOAC = direct-acting oral anticoagulant; GI = gastrointestinal.

Table 3Source of Bleeding and Culprit Lesions by Treatment Group

 

Outcomes

The warfarin group had more frequent hospitalizations, longer hospital stays, and more frequent need for transfusion compared with the DOAC and control groups ( Table 4 ). In patients receiving transfusion there were no significant differences between the groups in number of packed red blood cells transfused ( Table 4 ). There were no differences among the 3 groups in need for endoscopic or surgical intervention, and although no patients in the DOAC group died, there were no differences among the 3 groups in 30-day mortality ( Table 4 ).

Outcome Total
(n = 480)
DOAC
(n = 61)
Warfarin
(n = 123)
Control
(n = 296)
Overall
P Value
Hospitalization, n (%) 234 (48.8) 31 (50.8) 78 (63.4) 125 (42.4) .009
Hospital days, mean ± SD 6.3 ± 5.4 6.1 ± 4.8 7.8 ± 6.1 5.4 ± 4.9 .007
Required transfusion, n (%) 145 (30.2) 17 (27.9) 50 (42.0) 78 (26.4) .02
Units of blood transfused, * mean ± SD 3.7 ± 2.7 3.7 ± 3.3 3.9 ± 2.4 3.6 ± 2.7 .93
Required intervention, n (%) 41 (8.5) 4 (6.6) 12 (9.8) 25 (8.5) .65
30-day mortality, n (%) 14 (2.9) 0 3 (2.4) 11 (3.7) .38

* In patients requiring transfusion.

DOAC = direct-acting oral anticoagulant.

Table 4Clinical Outcomes (Unadjusted)

 

After adjusting for the baseline variables of age, sex, and CCI, the DOAC group had significantly fewer hospitalizations and less need for transfusion than the warfarin group ( Table 5 ). The warfarin group had significantly more hospitalizations, more hospital days, and required blood transfusion more frequently than the control group ( Table 5 ). The DOAC group had no significant differences in need for hospitalization, number of hospital days, need for transfusion, or number of transfusions compared with the control group. There were no significant differences between any of the groups in the need for endoscopic or surgical intervention or 30-day mortality ( Table 5 ).

CI = confidence interval; DOAC = direct-acting oral anticoagulant; OR = adjusted odds ratio; SE = standard error.
*In patients requiring transfusion.

Table 5Clinical Outcomes (Adjusted for Age, Sex, and Charlson Comorbidity Index)

 

Discussion

Here, we have shown that patients being treated with DOACs seem to have less severe gastrointestinal bleeding than patients treated with warfarin. Although several studies have shown that the frequency of gastrointestinal bleeding is greater in patients treated with DOACs compared with warfarin, 1 2 3 our observation that gastrointestinal bleeding associated with the use of DOACs versus warfarin is less likely to require hospitalization and blood transfusion is new. Further, the differences in outcomes were demonstrated despite the DOAC group having more frequent concomitant aspirin use compared with the warfarin group.

Several randomized trials have evaluated the incidence of bleeding, and specifically gastrointestinal bleeding, in patients treated with DOACs compared with warfarin. 1 2 3 7 In a study randomizing patients anticoagulated with rivaroxaban versus warfarin for atrial fibrillation, there was no difference in major bleeding but a lower incidence of fatal bleeding with rivaroxaban. 1 Rivaroxaban was shown to have a higher likelihood of major gastrointestinal bleeding compared with warfarin (hazard ratio of 1.66, P < .0001). 1 In a study randomizing patients anticoagulated with apixaban versus warfarin for atrial fibrillation, there was a lower incidence of major bleeding with apixaban but no difference in major gastrointestinal bleeding. 2 Finally, in a study randomizing patients treated with dabigatran 110 mg twice daily versus dabigatran 150 mg twice daily versus warfarin for atrial fibrillation, there was a lower incidence of major bleeding with low-dose dabigatran versus warfarin, and a higher incidence of major gastrointestinal bleeding only in the high-dose dabigatran group when compared with warfarin (hazard ratio 1.50; 95% CI, 1.19-1.89; P < .001). 3 Each of these studies evaluated major bleeding as clinically apparent bleeding with ≥2 g/dL drop in hemoglobin, transfusion of ≥2 U of packed red blood cells, or bleeding resulting in death. 1 2 3 None of the above studies provided information regarding the source of gastrointestinal bleeding or the culprit lesion responsible for gastrointestinal bleeding. Our study differs from the aforementioned studies in its specificity in identification of both the source and culprit lesions of gastrointestinal bleeding.

The increased incidence of major gastrointestinal bleeding with DOACs versus warfarin in these studies has raised some caution in their use, despite other proven benefits. Our data showing that gastrointestinal bleeding seems to be less severe with DOACs versus warfarin should mitigate some of this concern and support the use of DOACs over warfarin. In addition to the data supporting the superiority of DOACs over warfarin with regard to stroke prevention in patients with atrial fibrillation, reduction in intracranial hemorrhage, and reduced overall bleeding in some studies, there are other practical reasons that make these drugs preferable. Warfarin requires frequent monitoring, which at its best provides an inconvenience to patients and cost to the healthcare system, and at its worst provides a barrier to patient compliance and safe drug use. Warfarin has a number of both food and drug interactions that make control of anticoagulation more difficult, require more frequent monitoring, and increase the risk of adverse events due to over- or under-anticoagulation. Direct-acting oral anticoagulants, on the other hand, have little to no significant food or drug interactions and require no monitoring.

The finding of more severe gastrointestinal bleeding in patients treated with warfarin compared with DOACs in our study may be partly explained by the inherent difficulty in maintaining therapeutic drug levels in patients treated with warfarin, resulting in supra-therapeutic INRs and thus propensity for severe bleeding. The mean INR of 3.1 in the warfarin group is outside the generally accepted therapeutic range and supports this hypothesis. The longer length of hospital stay in patients taking warfarin may be related to warfarin's long half-life requiring more time for dissipation of the drug and subsequent reinstitution when necessary. The mechanism underlying potentially less severe bleeding in patients taking DOACs may be related to DOACs' relatively short half-life; once gastrointestinal bleeding is recognized, DOACs can be stopped, and their anticoagulation effect is attenuated fairly quickly.

Concerns about increased incidence of gastrointestinal bleeding associated with DOACs have been coupled with a general reluctance to use drugs that cannot be monitored or reversed. In 2015, however, a clinical trial revealed andexanet as an effective reversal agent for apixaban and rivaroxaban. 10 Further, a similar trial investigating idarucizumab as a reversal agent for dabigatran led to US Food and Drug Administration approval of this agent in 2015 for patients in need of urgent reversal of anticoagulation from dabigatran. 11 It should be noted that no patient in the cohort of patients receiving DOACs in our study received a reversal agent.

We recognize limitations of this study. This is an observational, retrospective, single-center study, which may limit the generalizability of its findings and may introduce potential bias in the results. For example, our study included patients with all types of gastrointestinal bleeding, and minor bleeding may have been more often reported in patients treated with DOACs because of the novelty of the drugs. This could introduce a bias toward the presence of more episodes of less severe bleeding with DOACs. We tend to discount this as a potential bias, however, because essentially all patients receiving anticoagulation are extensively counseled about the risk of gastrointestinal bleeding, and we speculate that all are similarly likely to report it. Finally, we cannot exclude the possibility of a type 2 error, because our study cohort is not large enough to detect small (but potentially meaningful) differences in 30-day mortality or gastrointestinal bleeding requiring intervention in patients treated with DOACs versus warfarin.

In summary, our study found that gastrointestinal bleeding in patients treated with DOACs was associated with fewer hospitalizations and less need for transfusions compared with patients treated with warfarin. Concerns about the increased incidence of gastrointestinal bleeding with DOACs compared with warfarin in several randomized trials may be at least partially offset by our findings that gastrointestinal bleeding associated with DOACs was less severe and more often clinically insignificant compared with warfarin. Because we recognize that this is a relatively small cohort, we believe that larger multicenter studies would be useful to further assess the severity of bleeding in patients taking DOACs and warfarin.

Funding: None.

Conflict of Interest: None.

Authorship: MMB: study concept and design; acquisition of data; review of clinical records; analysis and interpretation of data; drafting of the manuscript; critical revision of the manuscript for important intellectual content. JCN: analysis and interpretation of data; choice of statistical methods; statistical analysis; formulation of tables and figures. TS: review of clinical records; analysis and interpretation of data; drafting of the manuscript. DCR: study concept and design; analysis and interpretation of data; drafting of the manuscript; critical revision of the manuscript for important intellectual content; supervision of study conduct.

References

  • 1 M.R. Patel, K.W. Mahaffey, J. Garg, et al. Rivaroxaban versus warfarin in nonvalvular atrial fibrillation. N Engl J Med. 2011;365:883-891
  • 2 C.B. Granger, J.H. Alexander, J.J.V. McMurray, et al. Apixaban versus warfarin in patients with atrial fibrillation. N Engl J Med. 2011;365:981-992
  • 3 S.J. Connolly, M.D. Ezekowitz, S. Yusuf, et al. Dabigatran versus warfarin in patients with atrial fibrillation. N Engl J Med. 2009;361:1139-1151
  • 4 C.H. Yeh, P.L. Gross, J.I. Weitz. Evolving use of new oral anticoagulants for treatment of venous thromboembolism. Blood. 2014;124(7):1020-1028
  • 5 S. Schulman, C. Kearon, A.K. Kakkar, et al. Dabigatran versus warfarin in the treatment of acute venous thromboembolism. N Engl J Med. 2009;361:2342-2352
  • 6 R.P. Giugliano, C.T. Ruff, E. Braunwald, et al. Edoxaban versus warfarin in patients with atrial fibrillation. N Engl J Med. 2013;369(22):2093-2104
  • 7 J. Desia, J.M. Kolb, J.I. Weitz, J. Alsenberg. Gastrointestinal bleeding with the new oral anticoagulants–defining the issues and the management strategies. Thromb Haemost. 2013;110:205-212
  • 8 S.R. Ghate, J. Biskupiak, X. Ye, W.J. Kwong, D.I. Brixner. All-cause and bleeding related health care costs in warfarin-treated patients with atrial fibrillation. J Manag Care Pharm. 2011;17(9):672-684
  • 9 H. Quan, B. Li, C. Couris, et al. Updating and validating the charlson comorbidity index and score for risk adjustment in hospital discharge abstracts using data from 6 countries. Am J Epidemiol. 2011;173(6):676-682
  • 10 D.M. Siegal, J.T. Curnutte, S.J. Connolly, et al. Andexanet alfa for the reversal of factor Xa inhibitor activity. N Engl J Med. 2015;373:2413-2424
  • 11 C.V. Pollack Jr, P.A. Reilly, J. Eikelboom, et al. Idarucizumab for dabigatran reversal. N Engl J Med. 2015;373:511-520

© 2018 Elsevier Inc. All rights reserved.

Practitioners and researchers must always rely on their own experience and knowledge in evaluating and using any information, methods, compounds or experiments described herein. Because of rapid advances in the medical sciences, in particular, independent verification of diagnoses and drug dosages should be made. To the fullest extent of the law, no responsibility is assumed by Elsevier for any injury and/or damage to persons or property as a matter of products liability, negligence or otherwise, or from any use or operation of any methods, products, instructions, or ideas contained in the material herein.

This e-print is distributed with the support of BMS.  

Reproduced by:

Elsevier España, S.L.U.
(A member of Elsevier)
Av. Josep Tarradellas, 20-30
08029 Barcelona
Tel.: 932 000 711
Fax: 932 091 136