Resumption of oral anticoagulation following traumatic injury and risk of stroke and bleeding in patients with atrial fibrillation: a nationwide cohort study

Resumption of oral anticoagulation following traumatic injury and risk of stroke and bleeding in... Abstract Aims We examined the risks of all-cause mortality, stroke, major bleeding, and recurrent traumatic injury associated with resumption of vitamin K antagonists (VKAs) and non-VKAs oral anticoagulants (NOACs) following traumatic injury in atrial fibrillation (AF) patients. Methods and results This was a Danish nationwide registry-based study (2005–16), including 4541 oral anticoagulant (OAC)-treated AF patients experiencing traumatic injury (defined as traumatic brain injury, hip fracture, or traumatic torso or abdominal injury). Within 90 days following discharge from traumatic injury, 60.6% resumed VKA (median age = 80, CHA2DS2-VASc = 4, HAS-BLED = 2), 16.7% resumed NOAC (median age = 81, CHA2DS2-VASc = 4, HAS-BLED = 2), and 22.7% did not resume OAC treatment (median age = 81, CHA2DS2-VASc = 4, HAS-BLED = 3). Switch from VKA to NOAC occurred among 9.5%. Since 2009, the trend in OAC resumption increased (P-value <0.0001), in particular with NOACs (P-value <0.0001). Follow-up started 90 days after discharge, and time-varying multiple Cox regression analyses were used for comparisons. Compared with non-resumption, VKA and NOAC resumption were associated with lower hazard [95% confidence interval (CI)] of all-cause mortality [hazard ratio (HR) 0.48 (0.42–0.53) and HR 0.55 (0.47–0.66), respectively] and ischaemic stroke [HR 0.56 (0.43–0.72) and HR 0.54 (0.35–0.82), respectively], increased major bleeding hazard [HR 1.30 (1.03–1.64) and HR 1.15 (0.81–1.63), respectively], and similar hazard of recurrent traumatic injury [HR 0.93 (0.73–1.18) and HR 0.87 (0.60–1.27), respectively]. Conclusion AF patients resuming VKA and NOAC treatment following traumatic injury have lower hazard of all-cause mortality and ischaemic stroke, increased hazard of major bleeding but without additional hazards of recurrent traumatic injury. Withholding OAC following a traumatic injury in AF patients may not be warranted. View largeDownload slide View largeDownload slide Atrial fibrillation, Traumatic injury, Oral anticoagulation, Warfarin, NOAC, Bleeding Introduction Patients with atrial fibrillation (AF) requiring oral anticoagulants (OAC) are a fragile group of patients and a clinical challenge if they experience a traumatic injury.1–5 For instance, if a patient with AF is admitted due to a head injury caused by a ground-level fall, the occurrence of death during the injury hospitalization is estimated to be 24%, whereas this constitutes 32% if patients are treated with OACs before the ground-level fall admission.6 After traumatic injury, doctors and patients face the decision of whether to resume OAC therapy, a complex decision that needs careful consideration of the balance between benefit (stroke prevention) and risk (bleeding) of OAC treatment.7 Resumption of OAC treatment will reduce the stroke risk, but increase the bleeding risk, which in these fragile patients may be a major concern in the perspective of a risk for a recurrent traumatic injury. Non-vitamin K antagonist (VKA) oral anticoagulants (NOACs) are frequently used among elderly AF patients8; however, following a traumatic injury it is uncertain to what extent NOACs and VKAs are used and what the risks of outcomes are with respectively NOAC and VKA resumption compared with non-resumption. This study focused on OAC-treated AF patients who experienced a traumatic injury, and we examined the OAC resumption pattern and risk of all-cause mortality, ischaemic stroke, major bleeding, and recurrent traumatic injury associated with NOAC and VKA resumption compared with no OAC resumption. Methods Data sources All residents of Denmark are at birth or immigration, given a unique personal identification number that enabled us to cross-link individual information from the following Danish nationwide registries: (i) The Danish national patient registry holds information about all hospital contacts since 1978. A hospital contact is coded with one primary, and if appropriate, one or more secondary diagnosis codes based on the International Classification of Diseases 10th revision (ICD-10).9 (ii) The Danish national prescription registry keeps information about all filled prescriptions since 1995 including the Anatomical Therapeutic Chemical (ATC) codes, the day the prescription was filled, the package size, and dosages.10 (iii) The Danish civil registration system registers vital status.11 Study population and outcomes We conducted an observational study from 1 January 2005 to 31 December 2016. The study population constituted AF patients treatment with VKA or NOAC [dabigatran (since August 2011), rivaroxaban (since February 2012), or apixaban (since December 2012)] the day before a traumatic injury hospital admission. Traumatic injury involved hospitalization with traumatic cranial or brain injury (fracture of the skull, injury of optic nerve, intracranial injury, or crushing or multiple injuries of head), traumatic hip fracture (femoral neck, pertrochanteric, or subtrochanteric fracture), or traumatic torso or abdominal injury (serious open wound of thoracic wall, abdominal wall, or pelvis; injury of blood vessels, intrathoracic organs, intra-abdominal organs, and pelvic organs; and traumatic pneumothorax or haemothorax). Validation studies of administrative data report high positive predictive values for AF, hip fracture, and traumatic brain injury.12–14 Traumatic torso or abdominal injury was defined as done in a previous study from 2015.15 A quarantine period of 90 days subsequent to the discharge day was introduced to allow patients to fill a new prescription of an OAC drug following the traumatic injury. The package size of an OAC agent runs for a maximum of 3 months, and it is plausible that if patients resumed OAC treatment they finished the package from before the traumatic injury at first. All patients were followed up for 90 days, and with the 90 days of quarantine period, we were able to categorize patients into resumption and non-resumption of treatment and thereby limit misclassification and avoid conditioning on the future.16 The exclusion criteria are shown in Figure 1. Figure 1 View largeDownload slide Selection of the study population. NOAC, non-vitamin K antagonist oral anticoagulant; OAC, oral anticoagulation; VKA, vitamin K antagonist. Figure 1 View largeDownload slide Selection of the study population. NOAC, non-vitamin K antagonist oral anticoagulant; OAC, oral anticoagulation; VKA, vitamin K antagonist. The three treatment groups were resumption of VKA, resumption of NOAC, or no OAC resumption, and ATC codes are listed in the Supplementary material online, Table S1. Follow-up began after the 90 days of quarantine period and was limited to the first 3 years after the 90 days of quarantine period (90 days + 3 years). Outcomes of interest during follow-up were the earliest of all-cause mortality, ischaemic stroke, major bleeding, and recurrent traumatic injury. Outcome status was last assessed on 31 December 2016. All ICD-10 and ATC codes used to define the study population, treatment groups, and outcomes are listed in the Supplementary material online, Table S1. Concomitant pharmacotherapy and co-morbidity Concomitant pharmacotherapy was identified from filled prescriptions for a period of 90 days prior to baseline. Co-morbidities were identified from diagnosis codes registered within a period of 10 years prior to baseline. ATC codes and ICD-10 codes used to define concomitant pharmacotherapy and co-morbidities are listed in the Supplementary material online, Table S1. All patients were assessed for risk of stroke with the CHA2DS2-VASc score17 [Congestive heart failure, Hypertension, Age ≥ 75 years (2 points), Diabetes mellitus, Stroke/TE/transient ischaemic attack (2 points), Vascular disease, Age 65–74 years, and female Sex] and risk of bleeding with the HAS-BLED score [Hypertension, Abnormal renal or liver function, Stroke/TE, previous Bleeding, Labile international normalized ratio (INR; left out because data are unavailable), Elderly (age ≥ 65 years), and concomitant drugs (antiplatelet agents, non-steroidal anti-inflammatory drugs)/alcohol abuse]. Statistical analysis Time trends in OAC resumption according to calendar year were analysed with Cochrane Armitage trend tests in the period 2005–16. Baseline was set as 90 days after discharge from traumatic injury. We estimated the outcome-specific absolute risk (cumulative incidence) by the Aalen–Johansen method separately in the treatment groups to account for competing risks of death.18 Stroke, major bleeding, and recurrent traumatic injury were not accounted for as competing events. Outcome-specific hazards in the three treatment groups were modelled by multiple Cox regression. Each outcome-specific model was adjusted for potential confounders as determined by published literature and associated expert knowledge. All models were adjusted for calendar year, age (continuous), sex, and type of traumatic injury. All-cause mortality models were further adjusted for CHA2DS2-VASc and HAS-BLED factors, stroke models for CHA2DS2-VASc factors, major bleeding models for HAS-BLED factors, and recurrent traumatic injury models for CHA2DS2-VASc and HAS-BLED factors, osteoporosis, dementia, and benzodiazepine usage. For each outcome, the two Cox regression models were obtained, one using only baseline information and another using dynamic updates of all time-varying covariates (treatment group, co-morbidity, concomitant pharmacotherapy, calendar year, and age). Based on the outcome-specific Cox regression models using only baseline information, we computed differences in standardized absolute 1-year risks between the three treatment groups (g-formula).19,20 The standardized risks for treatment Group ‘A’ were obtained as the average of the predicted absolute 1-year risks in a copy of the real data where the treatment variable was set to ‘A’. The risks were predicted by combining the outcome-specific Cox regression models.20 The standardized differences were supplied with 95% bootstrap confidence limits based on 1000 bootstrap data sets. The level of significance was set at 5%. Data management and statistical analyses were performed using SAS (version 9.4 for Windows, SAS Institute, Cary, NC, USA) and R [R Core Team (2016). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria]. Ethics Retrospective registry-based studies do not require approval from the Research Ethical Committee System. The Danish Data Protection Agency approved the use of data for this study (ref.no: 2007-58-0015/GEH-2014-012 I-Suite no: 02720). Results Study population At baseline, which was on Day 90 following discharge from traumatic injury, we included a total of 4541 AF patients, treated with OAC before admission (Figure 1). During the quarantine period of 90 days, 1695 patients were excluded as they suffered from one of the outcomes of interest. Supplementary material online, Table S2 reports the characteristics of the population at the time of discharge from traumatic injury, and Supplementary material online, Figure S1 illustrates outcome-specific cumulative incidences during the quarantine period of 90 days. Among the 4541 included patients at baseline (Table 1), the resumption pattern was as follows: non-resumption 22.7%, VKA 60.6%, and NOAC 16.7% (dabigatran, n = 333, 7.3%; rivaroxaban, n = 211, 4.7%; and apixaban, n = 213, 4.7%). The median age was 81, 80, and 81 years in the non-resumption, VKA, and NOAC groups, respectively. All groups had a median CHA2DS2-VASc score of 4, but the median HAS-BLED score was highest in the non-resumption group being 3. In the non-resumption group, 40.8% had suffered from a traumatic brain injury, whereas among patients resuming VKA and NOAC this was 27.4% and 29.6%, respectively. In addition, concomitant usage of aspirin (27.1%), adenosine diphosphate receptor antagonists (4.8%), and benzodiazepines (25.3%) were more frequent among non-resumption than patients resuming OAC treatment. Table 1 Patient characteristics at baseline [AuthorQuery id="AQ20" rid="20"]?>   Non-resumption  VKA  NOAC  P-value  n (%)  1031 (22.7)  2753 (60.6)  757 (16.7)    Men, n (%)  513 (49.8)  1347 (48.9)  309 (40.8)  <0.001  Age, median [IQR]  81 [75, 86]  80 [74, 85]  81 [74, 87]  <0.001  CHA2DS2-VASc, median [IQR]  4 [3, 5]  4 [3, 5]  4 [3, 5]  0.289  HAS-BLED, median [IQR]  3 [2, 3]  2 [2, 3]  2 [2, 3]  <0.001  Type of injury, n (%)        <0.001   Traumatic brain injury  421 (40.8)  755 (27.4)  224 (29.6)     Hip fracture  475 (46.1)  1530 (55.6)  426 (56.3)     Traumatic torso or abdominal injury  135 (13.1)  468 (17.0)  107 (14.1)    OAC before traumatic injury, n (%)      <0.001   VKA  930 (90.2)  2750 (99.9)  72 (9.5)     NOAC  101 (9.8)  3 (0.1)  685 (90.5)    Co-morbidities, n (%)   Stroke  263 (25.5)  588 (21.4)  217 (28.7)  <0.001   Myocardial infarction  109 (10.6)  247 (9.0)  62 (8.2)  0.181   Ischaemic heart disease  364 (35.3)  950 (34.5)  203 (26.8)  <0.001   Peripheral artery disease  83 (8.1)  183 (6.6)  35 (4.6)  0.016   Heart failure  327 (31.7)  849 (30.8)  194 (25.6)  0.010   Diabetes mellitus  141 (13.7)  409 (14.9)  112 (14.8)  0.646   Hypertension  535 (51.9)  1686 (61.2)  410 (54.2)  <0.001   Chronic kidney disease  102 (9.9)  214 (7.8)  58 (7.7)  0.088   Liver disease  14 (1.4)  47 (1.7)  14 (1.8)  0.676   Prior bleeding  466 (45.2)  802 (29.1)  233 (30.8)  <0.001   Alcohol misuse  45 (4.4)  119 (4.3)  50 (6.6)  0.027   Cancer  182 (17.7)  444 (16.1)  134 (17.7)  0.395   Dementia  92 (8.9)  157 (5.7)  89 (11.8)  <0.001   Osteoporosis  187 (18.1)  502 (18.2)  198 (26.2)  <0.001  Concomitant medication, n (%)   ADP receptor antagonists  50 (4.8)  41 (1.5)  13 (1.7)  <0.001   Aspirin  279 (27.1)  488 (17.7)  36 (4.8)  <0.001   Dipyridamole  15 (1.5)  9 (0.3)  0 (0.0)  <0.001   Non-steroid anti-inflammatory drug  63 (6.1)  174 (6.3)  43 (5.7)  0.808   Loop diuretics  411 (39.9)  1168 (42.4)  275 (36.3)  0.008   Beta-blockers  487 (47.2)  1397 (50.7)  427 (56.4)  0.001   Calcium channel blockers  197 (19.1)  652 (23.7)  148 (19.6)  0.002   Renin–angiotensin system blocker  369 (35.8)  1107 (40.2)  267 (35.3)  0.007   Digoxin  344 (33.4)  976 (35.5)  189 (25.0)  <0.001   Benzodiazepine  261 (25.3)  590 (21.4)  173 (22.9)  0.038    Non-resumption  VKA  NOAC  P-value  n (%)  1031 (22.7)  2753 (60.6)  757 (16.7)    Men, n (%)  513 (49.8)  1347 (48.9)  309 (40.8)  <0.001  Age, median [IQR]  81 [75, 86]  80 [74, 85]  81 [74, 87]  <0.001  CHA2DS2-VASc, median [IQR]  4 [3, 5]  4 [3, 5]  4 [3, 5]  0.289  HAS-BLED, median [IQR]  3 [2, 3]  2 [2, 3]  2 [2, 3]  <0.001  Type of injury, n (%)        <0.001   Traumatic brain injury  421 (40.8)  755 (27.4)  224 (29.6)     Hip fracture  475 (46.1)  1530 (55.6)  426 (56.3)     Traumatic torso or abdominal injury  135 (13.1)  468 (17.0)  107 (14.1)    OAC before traumatic injury, n (%)      <0.001   VKA  930 (90.2)  2750 (99.9)  72 (9.5)     NOAC  101 (9.8)  3 (0.1)  685 (90.5)    Co-morbidities, n (%)   Stroke  263 (25.5)  588 (21.4)  217 (28.7)  <0.001   Myocardial infarction  109 (10.6)  247 (9.0)  62 (8.2)  0.181   Ischaemic heart disease  364 (35.3)  950 (34.5)  203 (26.8)  <0.001   Peripheral artery disease  83 (8.1)  183 (6.6)  35 (4.6)  0.016   Heart failure  327 (31.7)  849 (30.8)  194 (25.6)  0.010   Diabetes mellitus  141 (13.7)  409 (14.9)  112 (14.8)  0.646   Hypertension  535 (51.9)  1686 (61.2)  410 (54.2)  <0.001   Chronic kidney disease  102 (9.9)  214 (7.8)  58 (7.7)  0.088   Liver disease  14 (1.4)  47 (1.7)  14 (1.8)  0.676   Prior bleeding  466 (45.2)  802 (29.1)  233 (30.8)  <0.001   Alcohol misuse  45 (4.4)  119 (4.3)  50 (6.6)  0.027   Cancer  182 (17.7)  444 (16.1)  134 (17.7)  0.395   Dementia  92 (8.9)  157 (5.7)  89 (11.8)  <0.001   Osteoporosis  187 (18.1)  502 (18.2)  198 (26.2)  <0.001  Concomitant medication, n (%)   ADP receptor antagonists  50 (4.8)  41 (1.5)  13 (1.7)  <0.001   Aspirin  279 (27.1)  488 (17.7)  36 (4.8)  <0.001   Dipyridamole  15 (1.5)  9 (0.3)  0 (0.0)  <0.001   Non-steroid anti-inflammatory drug  63 (6.1)  174 (6.3)  43 (5.7)  0.808   Loop diuretics  411 (39.9)  1168 (42.4)  275 (36.3)  0.008   Beta-blockers  487 (47.2)  1397 (50.7)  427 (56.4)  0.001   Calcium channel blockers  197 (19.1)  652 (23.7)  148 (19.6)  0.002   Renin–angiotensin system blocker  369 (35.8)  1107 (40.2)  267 (35.3)  0.007   Digoxin  344 (33.4)  976 (35.5)  189 (25.0)  <0.001   Benzodiazepine  261 (25.3)  590 (21.4)  173 (22.9)  0.038  Characteristics of the study population at the time of baseline. Table 1 Patient characteristics at baseline [AuthorQuery id="AQ20" rid="20"]?>   Non-resumption  VKA  NOAC  P-value  n (%)  1031 (22.7)  2753 (60.6)  757 (16.7)    Men, n (%)  513 (49.8)  1347 (48.9)  309 (40.8)  <0.001  Age, median [IQR]  81 [75, 86]  80 [74, 85]  81 [74, 87]  <0.001  CHA2DS2-VASc, median [IQR]  4 [3, 5]  4 [3, 5]  4 [3, 5]  0.289  HAS-BLED, median [IQR]  3 [2, 3]  2 [2, 3]  2 [2, 3]  <0.001  Type of injury, n (%)        <0.001   Traumatic brain injury  421 (40.8)  755 (27.4)  224 (29.6)     Hip fracture  475 (46.1)  1530 (55.6)  426 (56.3)     Traumatic torso or abdominal injury  135 (13.1)  468 (17.0)  107 (14.1)    OAC before traumatic injury, n (%)      <0.001   VKA  930 (90.2)  2750 (99.9)  72 (9.5)     NOAC  101 (9.8)  3 (0.1)  685 (90.5)    Co-morbidities, n (%)   Stroke  263 (25.5)  588 (21.4)  217 (28.7)  <0.001   Myocardial infarction  109 (10.6)  247 (9.0)  62 (8.2)  0.181   Ischaemic heart disease  364 (35.3)  950 (34.5)  203 (26.8)  <0.001   Peripheral artery disease  83 (8.1)  183 (6.6)  35 (4.6)  0.016   Heart failure  327 (31.7)  849 (30.8)  194 (25.6)  0.010   Diabetes mellitus  141 (13.7)  409 (14.9)  112 (14.8)  0.646   Hypertension  535 (51.9)  1686 (61.2)  410 (54.2)  <0.001   Chronic kidney disease  102 (9.9)  214 (7.8)  58 (7.7)  0.088   Liver disease  14 (1.4)  47 (1.7)  14 (1.8)  0.676   Prior bleeding  466 (45.2)  802 (29.1)  233 (30.8)  <0.001   Alcohol misuse  45 (4.4)  119 (4.3)  50 (6.6)  0.027   Cancer  182 (17.7)  444 (16.1)  134 (17.7)  0.395   Dementia  92 (8.9)  157 (5.7)  89 (11.8)  <0.001   Osteoporosis  187 (18.1)  502 (18.2)  198 (26.2)  <0.001  Concomitant medication, n (%)   ADP receptor antagonists  50 (4.8)  41 (1.5)  13 (1.7)  <0.001   Aspirin  279 (27.1)  488 (17.7)  36 (4.8)  <0.001   Dipyridamole  15 (1.5)  9 (0.3)  0 (0.0)  <0.001   Non-steroid anti-inflammatory drug  63 (6.1)  174 (6.3)  43 (5.7)  0.808   Loop diuretics  411 (39.9)  1168 (42.4)  275 (36.3)  0.008   Beta-blockers  487 (47.2)  1397 (50.7)  427 (56.4)  0.001   Calcium channel blockers  197 (19.1)  652 (23.7)  148 (19.6)  0.002   Renin–angiotensin system blocker  369 (35.8)  1107 (40.2)  267 (35.3)  0.007   Digoxin  344 (33.4)  976 (35.5)  189 (25.0)  <0.001   Benzodiazepine  261 (25.3)  590 (21.4)  173 (22.9)  0.038    Non-resumption  VKA  NOAC  P-value  n (%)  1031 (22.7)  2753 (60.6)  757 (16.7)    Men, n (%)  513 (49.8)  1347 (48.9)  309 (40.8)  <0.001  Age, median [IQR]  81 [75, 86]  80 [74, 85]  81 [74, 87]  <0.001  CHA2DS2-VASc, median [IQR]  4 [3, 5]  4 [3, 5]  4 [3, 5]  0.289  HAS-BLED, median [IQR]  3 [2, 3]  2 [2, 3]  2 [2, 3]  <0.001  Type of injury, n (%)        <0.001   Traumatic brain injury  421 (40.8)  755 (27.4)  224 (29.6)     Hip fracture  475 (46.1)  1530 (55.6)  426 (56.3)     Traumatic torso or abdominal injury  135 (13.1)  468 (17.0)  107 (14.1)    OAC before traumatic injury, n (%)      <0.001   VKA  930 (90.2)  2750 (99.9)  72 (9.5)     NOAC  101 (9.8)  3 (0.1)  685 (90.5)    Co-morbidities, n (%)   Stroke  263 (25.5)  588 (21.4)  217 (28.7)  <0.001   Myocardial infarction  109 (10.6)  247 (9.0)  62 (8.2)  0.181   Ischaemic heart disease  364 (35.3)  950 (34.5)  203 (26.8)  <0.001   Peripheral artery disease  83 (8.1)  183 (6.6)  35 (4.6)  0.016   Heart failure  327 (31.7)  849 (30.8)  194 (25.6)  0.010   Diabetes mellitus  141 (13.7)  409 (14.9)  112 (14.8)  0.646   Hypertension  535 (51.9)  1686 (61.2)  410 (54.2)  <0.001   Chronic kidney disease  102 (9.9)  214 (7.8)  58 (7.7)  0.088   Liver disease  14 (1.4)  47 (1.7)  14 (1.8)  0.676   Prior bleeding  466 (45.2)  802 (29.1)  233 (30.8)  <0.001   Alcohol misuse  45 (4.4)  119 (4.3)  50 (6.6)  0.027   Cancer  182 (17.7)  444 (16.1)  134 (17.7)  0.395   Dementia  92 (8.9)  157 (5.7)  89 (11.8)  <0.001   Osteoporosis  187 (18.1)  502 (18.2)  198 (26.2)  <0.001  Concomitant medication, n (%)   ADP receptor antagonists  50 (4.8)  41 (1.5)  13 (1.7)  <0.001   Aspirin  279 (27.1)  488 (17.7)  36 (4.8)  <0.001   Dipyridamole  15 (1.5)  9 (0.3)  0 (0.0)  <0.001   Non-steroid anti-inflammatory drug  63 (6.1)  174 (6.3)  43 (5.7)  0.808   Loop diuretics  411 (39.9)  1168 (42.4)  275 (36.3)  0.008   Beta-blockers  487 (47.2)  1397 (50.7)  427 (56.4)  0.001   Calcium channel blockers  197 (19.1)  652 (23.7)  148 (19.6)  0.002   Renin–angiotensin system blocker  369 (35.8)  1107 (40.2)  267 (35.3)  0.007   Digoxin  344 (33.4)  976 (35.5)  189 (25.0)  <0.001   Benzodiazepine  261 (25.3)  590 (21.4)  173 (22.9)  0.038  Characteristics of the study population at the time of baseline. From 2009 to 2016, an increase in overall resumption of OAC treatment occurred from 68.7% to 85.6%, respectively (P-value <0.0001 for increasing trend), and a rapid yearly increase of NOAC usage occurred since 2011 (P-value <0.0001 for increasing trend). In 2016, an NOAC was prescribed for 39.6% of patients following a traumatic injury (Figure 2). Figure 2 View largeDownload slide Time trends showing resumption pattern of OAC following a traumatic injury from 2005 to 2016. NOAC, non-vitamin K antagonist oral anticoagulant; OAC, oral anticoagulation; VKA, vitamin K antagonist. Figure 2 View largeDownload slide Time trends showing resumption pattern of OAC following a traumatic injury from 2005 to 2016. NOAC, non-vitamin K antagonist oral anticoagulant; OAC, oral anticoagulation; VKA, vitamin K antagonist. Oral anticoagulants resumption vs. non-resumption and associated risk of outcomes During the follow-up period of 3 years, 1438 died, 285 suffered from an ischaemic stroke, 442 from major bleeding of which 130 were intracranial bleedings (non-resumption, VKA resumption, and NOAC resumption experienced 37, 83, and 10 intracranial bleedings, respectively), and 357 from recurrent traumatic injury. The cumulative risks obtained with the Aalen–Johansen method without covariate adjustment are shown in Supplementary material online, Figure S2. Take home figure shows the standardized absolute risks. The 1-year standardized absolute risks and absolute risk differences are reported in Table 2. Table 2 Standardized absolute risks   At 1 year   No. of events/ no. at risk  Standardized absolute risk (95% CI)  Absolute risk difference (95% CI)a  Absolute risk difference (95% CI)a  All-cause mortality   Non-resumption  224/723  19.5% (17.5 to 21.8%)  Reference  —   VKA  364/2101  15.3% (14.4 to 16.7%)  −4.2% (−6.3 to −2.7%)a  Reference   NOAC  92/388  15.1% (12.5 to 16.9%)  −4.4% (−7.2 to −1.4%)a  −0.02% (−2.7 to 1.4%)  Stroke/thrombo-embolism   Non-resumption  45/696  4.0% (3.2 to 5.1%)  Reference  —   VKA  82/2056  3.4% (2.8 to 4.1%)  −0.6% (−1.7 to 0.4%)  Reference   NOAC  20/378  2.8% (1.8 to 4.1%)  −1.2% (−2.7 to 0.3%)  −0.6% (−1.9 to 0.8%)  Major bleeding   Non-resumption  48/695  4.2% (3.3 to 5.3%)  Reference  —   VKA  120/2022  4.8% (4.1 to 5.6%)  0.5% (−0.5 to 1.5%)  Reference   NOAC  20/379  3.6% (2.5 to 5.0%)  −0.6% (−2.1 to 1.1%)  −1.1% (−2.4 to 0.3%)  Recurrent traumatic injury   Non-resumption  51/693  4.4% (3.5 to 5.6%)  Reference  —   VKA  103/2026  4.4% (3.7 to 5.1%)  0.0% (−1.2 to 1.0%)  Reference   NOAC  27/373  4.0% (2.8 to 5.5%)  −0.4% (−2.0 to 1.2%)  −0.4% (−1.8 to 1.2%)    At 1 year   No. of events/ no. at risk  Standardized absolute risk (95% CI)  Absolute risk difference (95% CI)a  Absolute risk difference (95% CI)a  All-cause mortality   Non-resumption  224/723  19.5% (17.5 to 21.8%)  Reference  —   VKA  364/2101  15.3% (14.4 to 16.7%)  −4.2% (−6.3 to −2.7%)a  Reference   NOAC  92/388  15.1% (12.5 to 16.9%)  −4.4% (−7.2 to −1.4%)a  −0.02% (−2.7 to 1.4%)  Stroke/thrombo-embolism   Non-resumption  45/696  4.0% (3.2 to 5.1%)  Reference  —   VKA  82/2056  3.4% (2.8 to 4.1%)  −0.6% (−1.7 to 0.4%)  Reference   NOAC  20/378  2.8% (1.8 to 4.1%)  −1.2% (−2.7 to 0.3%)  −0.6% (−1.9 to 0.8%)  Major bleeding   Non-resumption  48/695  4.2% (3.3 to 5.3%)  Reference  —   VKA  120/2022  4.8% (4.1 to 5.6%)  0.5% (−0.5 to 1.5%)  Reference   NOAC  20/379  3.6% (2.5 to 5.0%)  −0.6% (−2.1 to 1.1%)  −1.1% (−2.4 to 0.3%)  Recurrent traumatic injury   Non-resumption  51/693  4.4% (3.5 to 5.6%)  Reference  —   VKA  103/2026  4.4% (3.7 to 5.1%)  0.0% (−1.2 to 1.0%)  Reference   NOAC  27/373  4.0% (2.8 to 5.5%)  −0.4% (−2.0 to 1.2%)  −0.4% (−1.8 to 1.2%)  a The corresponding absolute risk differences tested for significant differences between standardized absolute risks. Statistical significance was obtained when the 95% confidence limits for difference in absolute risks did not contain zero. Table 2 Standardized absolute risks   At 1 year   No. of events/ no. at risk  Standardized absolute risk (95% CI)  Absolute risk difference (95% CI)a  Absolute risk difference (95% CI)a  All-cause mortality   Non-resumption  224/723  19.5% (17.5 to 21.8%)  Reference  —   VKA  364/2101  15.3% (14.4 to 16.7%)  −4.2% (−6.3 to −2.7%)a  Reference   NOAC  92/388  15.1% (12.5 to 16.9%)  −4.4% (−7.2 to −1.4%)a  −0.02% (−2.7 to 1.4%)  Stroke/thrombo-embolism   Non-resumption  45/696  4.0% (3.2 to 5.1%)  Reference  —   VKA  82/2056  3.4% (2.8 to 4.1%)  −0.6% (−1.7 to 0.4%)  Reference   NOAC  20/378  2.8% (1.8 to 4.1%)  −1.2% (−2.7 to 0.3%)  −0.6% (−1.9 to 0.8%)  Major bleeding   Non-resumption  48/695  4.2% (3.3 to 5.3%)  Reference  —   VKA  120/2022  4.8% (4.1 to 5.6%)  0.5% (−0.5 to 1.5%)  Reference   NOAC  20/379  3.6% (2.5 to 5.0%)  −0.6% (−2.1 to 1.1%)  −1.1% (−2.4 to 0.3%)  Recurrent traumatic injury   Non-resumption  51/693  4.4% (3.5 to 5.6%)  Reference  —   VKA  103/2026  4.4% (3.7 to 5.1%)  0.0% (−1.2 to 1.0%)  Reference   NOAC  27/373  4.0% (2.8 to 5.5%)  −0.4% (−2.0 to 1.2%)  −0.4% (−1.8 to 1.2%)    At 1 year   No. of events/ no. at risk  Standardized absolute risk (95% CI)  Absolute risk difference (95% CI)a  Absolute risk difference (95% CI)a  All-cause mortality   Non-resumption  224/723  19.5% (17.5 to 21.8%)  Reference  —   VKA  364/2101  15.3% (14.4 to 16.7%)  −4.2% (−6.3 to −2.7%)a  Reference   NOAC  92/388  15.1% (12.5 to 16.9%)  −4.4% (−7.2 to −1.4%)a  −0.02% (−2.7 to 1.4%)  Stroke/thrombo-embolism   Non-resumption  45/696  4.0% (3.2 to 5.1%)  Reference  —   VKA  82/2056  3.4% (2.8 to 4.1%)  −0.6% (−1.7 to 0.4%)  Reference   NOAC  20/378  2.8% (1.8 to 4.1%)  −1.2% (−2.7 to 0.3%)  −0.6% (−1.9 to 0.8%)  Major bleeding   Non-resumption  48/695  4.2% (3.3 to 5.3%)  Reference  —   VKA  120/2022  4.8% (4.1 to 5.6%)  0.5% (−0.5 to 1.5%)  Reference   NOAC  20/379  3.6% (2.5 to 5.0%)  −0.6% (−2.1 to 1.1%)  −1.1% (−2.4 to 0.3%)  Recurrent traumatic injury   Non-resumption  51/693  4.4% (3.5 to 5.6%)  Reference  —   VKA  103/2026  4.4% (3.7 to 5.1%)  0.0% (−1.2 to 1.0%)  Reference   NOAC  27/373  4.0% (2.8 to 5.5%)  −0.4% (−2.0 to 1.2%)  −0.4% (−1.8 to 1.2%)  a The corresponding absolute risk differences tested for significant differences between standardized absolute risks. Statistical significance was obtained when the 95% confidence limits for difference in absolute risks did not contain zero. Take home figure View largeDownload slide Standardized absolute risks of outcomes associated with non-resumption vs. resumption of VKA and NOAC following traumatic injury. For each treatment option separately, the standardized absolute risks are based on the outcome-specific multiple Cox regression models and obtained as averages of the predicted risks had all patients received this treatment. NOAC, non-vitamin K antagonist oral anticoagulant; VKA, vitamin K antagonist. Take home figure View largeDownload slide Standardized absolute risks of outcomes associated with non-resumption vs. resumption of VKA and NOAC following traumatic injury. For each treatment option separately, the standardized absolute risks are based on the outcome-specific multiple Cox regression models and obtained as averages of the predicted risks had all patients received this treatment. NOAC, non-vitamin K antagonist oral anticoagulant; VKA, vitamin K antagonist. Figure 3 shows the adjusted hazard ratios (HRs) obtained from time-varying Cox regression models. Compared with non-resumption, VKA and NOAC resumption were associated with significantly lower hazard of all-cause mortality and ischaemic stroke and an increased associated hazard of major bleeding. For traumatic injury, the HR was not decreased or increased with resumption of VKA and NOAC vs. non-resumption. Predictors for the hazard rate of major bleeding following a traumatic injury are presented in Supplementary material online, Table S3. Figure 3 View largeDownload slide Adjusted hazard ratios of outcomes associated with non-resumption vs. resumption of VKA and NOAC following traumatic injury. Time-dependent multiple Cox regression models were used to allow patients to change treatment group during follow-up. NOAC, non-vitamin K antagonist oral anticoagulant; VKA, vitamin K antagonist. Figure 3 View largeDownload slide Adjusted hazard ratios of outcomes associated with non-resumption vs. resumption of VKA and NOAC following traumatic injury. Time-dependent multiple Cox regression models were used to allow patients to change treatment group during follow-up. NOAC, non-vitamin K antagonist oral anticoagulant; VKA, vitamin K antagonist. Type of traumatic injury The adjusted HRs were estimated in a subgroup analysis based on the type of traumatic injury. The results in the subgroups (see Supplementary material online, Table S4) according to traumatic brain injury, hip fracture, and traumatic torso and abdominal injury were generally similar to the results in Figure 3. Switch of oral anticoagulants following traumatic injury Among those resuming OAC treatment, it was most common to resume the same OAC treatment as before the traumatic injury (Table 1); 99.9% of the patients who resumed a VKA were also on treatment with a VKA before the traumatic injury, and among those treated with NOAC after the traumatic injury, 90.5% were treated with NOAC and 9.5% with VKA before the traumatic injury. We evaluated those patients treated with VKA before the traumatic injury who either resumed VKA treatment or switched to a NOAC after the traumatic injury (see Supplementary material online, Table S4). Compared with non-resumption, staying on VKA treatment or switching to NOAC were both associated with a significantly lower hazard of all-cause mortality and ischaemic stroke but an increased hazard of bleeding; however, for NOAC switchers, the lower ischaemic stroke and increased major bleeding hazards were not significant. Sensitivity analyses A subgroup analysis was conducted including only patients with HAS-BLED score ≥4 (see Supplementary material online, Table S3). The results were broadly similar to the main results in Figure 3, albeit the analysis was influenced by lack of power. When testing the impact of changing the quarantine period to 30, 60, or 180 days, the results were comparable to the main results (see Supplementary material online, Table S5 and Figure 3). Discussion In our nationwide observational study, including 4288 OAC-treated patients with AF who suffered a traumatic injury, the main findings were as follows: (i) overall 60.6% resumed a VKA, 16.7% resumed a NOAC, and 22.7% did not resume OAC treatment during the study period 2005–16; (ii) since 2009 an increase in OAC resumption has occurred; and (iii) compared with non-resumption, resumption of VKA or NOAC treatment was associated with significantly lower hazard of all-cause mortality and ischaemic stroke, increased hazard of major bleedings, and no difference in the hazard of recurrent traumatic injury. Since more than 50% of the study population comprised octogenarians (median age was 80–81 years), it is plausible that falls were the most common cause of the traumatic injury hospitalization4,5; however, we were not able to identify the specific cause of the traumatic injury in the Danish national patient registry. The results from this study are particularly interesting assuming that the traumatic injury was caused by a fall, because fall risk among OAC-treated AF patients is feared with regard to major and fatal bleedings.21–23 Underuse of OAC treatment has been reported in several countries,24,25 and two main reasons for a doctor’s non-prescription of VKA are concerns regarding fall and bleeding risks.21 Hylek et al.7 found that previous falls were the most frequent physician-cited reason for not prescribing OAC among AF patients aged 80 years or older. In our study, 22.7% did not resume OAC treatment, and these patients had a higher occurrence of traumatic brain injury at baseline compared with those patients resuming VKA and NOAC treatment. In a study by Liu et al.15, the authors found that suffering a traumatic brain injury was associated with low odds of VKA resumption compared with non-resumption. The authors discussed that the low odds of VKA resumption after a traumatic brain injury may be related to doctors’ fear of fall risk in fragile elderly AF patients and perhaps that doctors overestimate a patient’s predicted bleeding risk. In our study, patients not resuming OAC had the highest predicted bleeding risk (median HAS-BLED = 3), however, we found that the 1-year standardized absolute risks of recurrent traumatic injury ranged from 4.0% to 4.4% without a difference between OAC resumption vs. non-resumption. These results suggest that perhaps doctors may consider being less reluctant with OAC due to recurrent fall risk (and the risk of fall-related bleedings), since the fall risk does not seem to vary between resumption of OAC and non-resumption patients. Moreover, in our study, a subgroup analysis including only patients with baseline HAS-BLED score ≥4 showed that the hazard of recurrent traumatic injury did not differ significantly between VKA and NOAC resumption compared with non-resumption. However, we found an increased hazard of major bleeding with VKA and NOAC resumption compared with non-resumption, which indicates that doctors’ concerns regarding bleeding risk with OAC treatment are relevant. Our study showed that VKA and NOAC resumption was associated with significantly lower hazard of ischaemic stroke and all-cause mortality. A retrospective study by Albrecht et al.26 examined the risk of stroke and major bleeding with warfarin usage after a traumatic brain injury and concluded that the benefits in stroke reduction with warfarin usage outweighed the risk of major bleeding. Similarly, a subgroup analysis in our study including only patients suffering from a traumatic brain injury pointed to a lower associated hazard of ischaemic stroke and all-cause mortality with VKA and NOAC resumption; however, lack of power in the NOAC group for the ischaemic stroke outcome could not confirm a significantly lower ischaemic stroke HR with NOACs. Similarly to Albrecht et al., the hazard of major bleeding was significantly higher in our study among patients resuming VKA and NOAC treatment compared with non-resumption following a traumatic brain injury. Since 2009, an increase in OAC resumption occurred following traumatic injury. This may suggest that even though doctors fear for a patient’s fall and bleeding risks, the awareness of appropriate stroke prophylaxis has improved with the consequence that more patients resume OAC treatment following a traumatic injury.27 It was more common that patients treated with VKA prior to the traumatic injury switched to an NOAC after discharge (9.5%) than switching from an NOAC to a VKA (0.1%). Case reports have described problems with the lack of reversal for NOAC therapy among trauma patients taking NOACs.22,23 However, the increased NOAC use observed in our study could indicate that the benefits of NOACs as stroke prophylaxis might counterbalance doctors’ fear of a patient’s fall and bleeding risks. Moreover, idarucizumab for dabigatran reversal was first introduced in Denmark in March 2016.28 In our study, a switch from VKA to NOAC or staying on VKA after the traumatic injury had similar effect and safety profile compared with non-resumption. In clinical practice, an AF patient’s individual risk factors for stroke, bleeding, and recurrent fall may be weighed in the decision-making of withholding or resuming OAC treatment following a traumatic injury. However, the first key finding in our study was that withholding of OAC was generally not beneficial to AF patients. Second, in 2016, ∼15% did not still resume OAC treatment within 90 days after the traumatic injury. This implies that there is a need for increased awareness of resumption of OAC treatment following a traumatic injury. Furthermore, good clinical practice requires clinicians to evaluate and minimize a patient’s risk of recurrent falls, e.g. considerations regarding a patient’s co-morbidities and concomitant pharmacotherapy.29–31 Limitations and strengths In our observational study, it was methodologically necessary to include a quarantine period of 90 days16; however, changing the quarantine period to 30, 60, or 180 days (see Supplementary material online, Table S5) did not alter our results. The data from the Danish administrative registries did not enable us to estimate when OAC resumption would be safe during the quarantine periods. A limitation was residual confounding with unmeasured confounders, such as INR, time in therapeutic range, haemoglobin levels, serum creatinine levels, and body mass index. Another limitation may be confounding by indication, since it was possible that patients not resuming OAC treatment were frailer with a higher risk of all-cause mortality, stroke, major bleeding, and recurrent traumatic injury than those resuming VKA and NOAC treatment. Importantly, our conclusions support associations. The main strength in our study, given by the Danish administrative registries, was the opportunity to examine a research question that would be challenging to perform as a randomized controlled trial. This study was able to include a large nationwide population and evaluated both VKA and NOAC resumption by filled prescriptions as directed by doctor, and the Danish registries have been validated intensely.9,12,13 Conclusions Among the OAC-treated patients with AF who experienced a traumatic injury, we conclude that within 90 days after the traumatic injury approximately 3 of 4 resumed OAC. Resumption of VKA or NOAC treatment was associated with significantly lower hazards of all-cause mortality and ischaemic stroke, at the cost of an increased hazard of major bleedings, but without any additional hazard of recurrent traumatic injury, compared with non-resumption. Our study suggests that resuming OAC following traumatic injury may result in better clinical outcomes, including lower risk of stroke. Supplementary material Supplementary material is available at European Heart Journal online. Funding This work was supported by the Velux Foundations and Boehringer-Ingelheim. The sponsors had no influence on the study design, interpretation of results, or the decision to submit the manuscript for publication. G.H.G. is supported by an unrestricted clinical research scholarship from the Novo Nordisk Foundation. Conflict of interest: L.S.: research funding from Boehringer-Ingelheim. E.L.F.: research funding from Lundbeck Foundation, Janssen Pharmaceuticals, and BMS. M.L.: speaker fees from BMS and Bayer. K.G.: research funding from BMS. G.Y.H.L.: consultant for Bayer/Janssen, BMS/Pfizer, Biotronik, Medtronic, Boehringer-Ingelheim, Microlife, and Daiichi-Sankyo and speaker for Bayer, BMS/Pfizer, Medtronic, Boehringer-Ingelheim, Microlife, Roche, and Daiichi-Sankyo. C.T.P.: research contracts with Bayer and Biotronic and received speaker fees from Bayer and BMS. G.H.G.: research funding from Boehringer-Ingelheim, Pfizer, BMS, AstraZeneca, and Bayer and declares ownership of stocks in Lundbeck A/S, Novo Nordisk A/S, and ALK Abello Pharmaceuticals. J.B.O.: received speaker fees from Bayer, BMS, and Boehringer-Ingelheim and research funding from BMS and The Capital Region of Denmark. References 1 Schnabel RB, Yin X, Gona P, Larson MG, Beiser AS, McManus DD, Newton-Cheh C, Lubitz SA, Magnani JW, Ellinor PT, Seshadri S, Wolf PA, Vasan RS, Benjamin EJ, Levy D. 50 year trends in atrial fibrillation prevalence, incidence, risk factors, and mortality in the Framingham Heart Study: a cohort study. Lancet  2015; 386: 154– 162. Google Scholar CrossRef Search ADS PubMed  2 Healey JS, Oldgren J, Ezekowitz M, Zhu J, Pais P, Wang J, Commerford P, Jansky P, Avezum A, Sigamani A, Damasceno A, Reilly P, Grinvalds A, Nakamya J, Aje A, Almahmeed W, Moriarty A, Wallentin L, Yusuf S, Connolly SJ. Occurrence of death and stroke in patients in 47 countries 1 year after presenting with atrial fibrillation: a cohort study. Lancet  2016; 388: 1161– 1169. Google Scholar CrossRef Search ADS PubMed  3 Hadjizacharia P, O’Keeffe T, Brown CVR, Inaba K, Salim A, Chan LS, Demetriades D, Rhee P. Incidence, risk factors, and outcomes for atrial arrhythmias in trauma patients. Am Surg  2011; 77: 634– 639. Google Scholar PubMed  4 Majdan M, Mauritz W. Unintentional fall-related mortality in the elderly: comparing patterns in two countries with different demographic structure. BMJ Open  2015; 5: e008672. Google Scholar CrossRef Search ADS PubMed  5 Park YJ, Ro YS, Shin SD, Song KJ, Lee SC, Kim YJ, Kim JY, Hong KJ, Kim JE, Kim MJ, Kim SC. Age effects on case fatality rates of injury patients by mechanism. Am J Emerg Med  2016; 34: 515– 520. Google Scholar CrossRef Search ADS PubMed  6 Inui TS, Parina R, Chang DC, Inui TS, Coimbra R. Mortality after ground-level fall in the elderly patient taking oral anticoagulation for atrial fibrillation/flutter: a long-term analysis of risk versus benefit. J Trauma Acute Care Surg  2014; 76: 642– 650. Google Scholar CrossRef Search ADS PubMed  7 Hylek EM, D’Antonio J, Evans-Molina C, Shea C, Henault LE, Regan S. Translating the results of randomized trials into clinical practice. Stroke  2006; 37: 1075– 1080. Google Scholar CrossRef Search ADS PubMed  8 Staerk L, Fosbøl EL, Gadsbøll K, Sindet-Pedersen C, Pallisgaard JL, Lamberts M, Lip GYH, Torp-Pedersen C, Gislason GH, Olesen JB. Non-vitamin K antagonist oral anticoagulation usage according to age among patients with atrial fibrillation: Temporal trends 2011–2015 in Denmark. Sci Rep  2016; 6: 31477. Google Scholar CrossRef Search ADS PubMed  9 Schmidt M, Schmidt SAJ, Sandegaard JL, Ehrenstein V, Pedersen L, Sørensen HT. The Danish National Patient Registry: a review of content, data quality, and research potential. Clin Epidemiol  2015; 7: 449– 490. Google Scholar CrossRef Search ADS PubMed  10 Kildemoes HW, Sørensen HT, Hallas J. The Danish National Prescription Registry. Scand J Public Health  2011; 39: 38– 41. Google Scholar CrossRef Search ADS PubMed  11 Pedersen CB. The Danish Civil Registration System. Scand J Public Health  2011; 39: 22– 25. Google Scholar CrossRef Search ADS PubMed  12 Frost L, Andersen LV, Vestergaard P, Husted S, Mortensen LS. Trend in mortality after stroke with atrial fibrillation. Am J Med  2007; 120: 47– 53. Google Scholar CrossRef Search ADS PubMed  13 Hundrup YA, Høidrup S, Obel EB, Rasmussen NK. The validity of self-reported fractures among Danish female nurses: comparison with fractures registered in the Danish National Hospital Register. Scand J Public Health  2004; 32: 136– 143. Google Scholar CrossRef Search ADS PubMed  14 Shore AD, Mccarthy ML, Serpi T, Gertner M. Validity of administrative data for characterizing traumatic brain injury-related hospitalizations. Brain Inj  2005; 19: 613– 621. Google Scholar CrossRef Search ADS PubMed  15 Liu X, Baumgarten M, Smith G, Gambert S, Gottlieb S, Rattinger G, Albrecht J, Langenberg P, Zuckerman I. Warfarin usage among elderly atrial fibrillation patients with traumatic injury, an analysis of United States Medicare fee-for-service enrollees. J Clin Pharmacol  2015; 55: 25– 32. Google Scholar CrossRef Search ADS PubMed  16 Staerk L, Lip GYH, Olesen JB, Fosbøl EL, Pallisgaard JL, Bonde AN, Gundlund A, Lindhardt TB, Hansen ML, Torp-Pedersen C, Gislason GH. Stroke and recurrent haemorrhage associated with antithrombotic treatment after gastrointestinal bleeding in patients with atrial fibrillation: nationwide cohort study. BMJ  2015; 351: h5876. Google Scholar CrossRef Search ADS PubMed  17 Olesen JB, Lip GYH, Hansen ML, Hansen PR, Tolstrup JS, Lindhardsen J, Selmer C, Ahlehoff O, Olsen A-MS, Gislason GH, Torp-Pedersen C. Validation of risk stratification schemes for predicting stroke and thromboembolism in patients with atrial fibrillation: nationwide cohort study. BMJ  2011; 342: d124– d124. Google Scholar CrossRef Search ADS PubMed  18 Aalen OO, Johansen S. An empirical transition matrix for non-homogeneous Markov chains based on censored observations. Scand J Stat  1978; 5: 141– 150. 19 Sato T, Matsuyama Y. Marginal structural models as a tool for standardization. Epidemiology  2003; 14: 680– 686. Google Scholar CrossRef Search ADS PubMed  20 Benichou J, Gail MH. Estimates of absolute cause-specific risk in cohort studies. Biometrics  1990; 46: 813– 826. Google Scholar CrossRef Search ADS PubMed  21 Kakkar AK, Mueller I, Bassand J-P, Fitzmaurice DA, Goldhaber SZ, Goto S, Haas S, Hacke W, Lip GYH, Mantovani LG, Turpie AGG, Eickels M, van, Misselwitz F, Rushton-Smith S, Kayani G, Wilkinson P, Verheugt FWA. Risk profiles and antithrombotic treatment of patients newly diagnosed with atrial fibrillation at risk of stroke: perspectives from the international, observational, prospective GARFIELD registry. PLoS One  2013; 8: e63479. Google Scholar CrossRef Search ADS PubMed  22 Stöllberger C, Ulram A, Bastovansky A, Finsterer J. Traumatic fatal cerebral hemorrhage in an old patient with a history of multiple sclerosis under dabigatran: a case report and review of the literature. J Geriatr Cardiol  2015; 12: 83. Google Scholar PubMed  23 Stöllberger C, Finsterer J. Fatal consequences of climbing a ladder under apixaban and drunken. Neurol Neurochir Pol  2016; 50: 200– 202. Google Scholar PubMed  24 Sabouret P, Bricard M, Hermann M-A, Cotté F-E, Deret-Bixio L, Rushton-Smith S. Discrepancy between guidelines for stroke prevention in atrial fibrillation and practice patterns in primary care. The nationwide French AFIGP survey. Arch Cardiovasc Dis  2015; 108: 544– 553. Google Scholar CrossRef Search ADS PubMed  25 Alamneh EA, Chalmers L, Bereznicki LR. Suboptimal use of oral anticoagulants in atrial fibrillation: has the introduction of direct oral anticoagulants improved prescribing practices? Am J Cardiovasc Drugs  2016; 16: 183– 200. Google Scholar CrossRef Search ADS PubMed  26 Albrecht JS, Liu X, Baumgarten M, Langenberg P, Rattinger GB, Smith GS, Gambert SR, Gottlieb SS, Zuckerman IH. Benefits and risks of anticoagulation resumption following traumatic brain injury. JAMA Intern Med  2014; 174: 1244– 1251. Google Scholar CrossRef Search ADS PubMed  27 Kirchhof P, Benussi S, Kotecha D, Ahlsson A, Atar D, Casadei B, Castella M, Diener H-C, Heidbuchel H, Hendriks J, Hindricks G, Manolis AS, Oldgren J, Popescu BA, Schotten U, Putte BV, Vardas P, Agewall S, Camm J, Esquivias GB, Budts W, Carerj S, Casselman F, Coca A, Caterina RD, Deftereos S, Dobrev D, Ferro JM, Filippatos G, Fitzsimons D, Gorenek B, Guenoun M, Hohnloser SH, Kolh P, Lip GYH, Manolis A, McMurray J, Ponikowski P, Rosenhek R, Ruschitzka F, Savelieva I, Sharma S, Suwalski P, Tamargo JL, Taylor CJ, Gelder ICV, Voors AA, Windecker S, Zamorano JL, Zeppenfeld K. 2016 ESC guidelines for the management of atrial fibrillation developed in collaboration with EACTS. Eur Heart J  2016; 37: 2893– 2962. Google Scholar CrossRef Search ADS PubMed  28 Pollack CVJ, Reilly PA, Eikelboom J, Glund S, Verhamme P, Bernstein RA, Dubiel R, Huisman MV, Hylek EM, Kamphuisen PW, Kreuzer J, Levy JH, Sellke FW, Stangier J, Steiner T, Wang B, Kam C-W, Weitz JI. Idarucizumab for dabigatran reversal. N Engl J Med  2015; 373: 511– 520. Google Scholar CrossRef Search ADS PubMed  29 Jørgensen TSH, Hansen AH, Sahlberg M, Gislason GH, Torp-Pedersen C, Andersson C, Holm E. Falls and comorbidity: the pathway to fractures. Scand J Public Health  2014; 42: 287– 294. Google Scholar CrossRef Search ADS PubMed  30 Torstensson M, Hansen AH, Leth-Møller K, Jørgensen TSH, Sahlberg M, Andersson C, Kristensen KE, Ryg J, Weeke P, Torp-Pedersen C, Gislason G, Holm E. Danish register-based study on the association between specific cardiovascular drugs and fragility fractures. BMJ Open  2015; 5: e009522. Google Scholar CrossRef Search ADS PubMed  31 de Vries OJ, Peeters G, Elders P, Sonnenberg C, Muller M, Deeg DJH, Lips P. The elimination half-life of benzodiazepines and fall risk: two prospective observational studies. Age Ageing  2013; 42: 764– 770. Google Scholar CrossRef Search ADS PubMed  Published on behalf of the European Society of Cardiology. All rights reserved. © The Author 2017. For permissions, please email: journals.permissions@oup.com. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices) http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png European Heart Journal Oxford University Press

Resumption of oral anticoagulation following traumatic injury and risk of stroke and bleeding in patients with atrial fibrillation: a nationwide cohort study

Loading next page...
 
/lp/ou_press/resumption-of-oral-anticoagulation-following-traumatic-injury-and-risk-5oYffEoSuo
Publisher
Oxford University Press
Copyright
Published on behalf of the European Society of Cardiology. All rights reserved. © The Author 2017. For permissions, please email: journals.permissions@oup.com.
ISSN
0195-668X
eISSN
1522-9645
D.O.I.
10.1093/eurheartj/ehx598
Publisher site
See Article on Publisher Site

Abstract

Abstract Aims We examined the risks of all-cause mortality, stroke, major bleeding, and recurrent traumatic injury associated with resumption of vitamin K antagonists (VKAs) and non-VKAs oral anticoagulants (NOACs) following traumatic injury in atrial fibrillation (AF) patients. Methods and results This was a Danish nationwide registry-based study (2005–16), including 4541 oral anticoagulant (OAC)-treated AF patients experiencing traumatic injury (defined as traumatic brain injury, hip fracture, or traumatic torso or abdominal injury). Within 90 days following discharge from traumatic injury, 60.6% resumed VKA (median age = 80, CHA2DS2-VASc = 4, HAS-BLED = 2), 16.7% resumed NOAC (median age = 81, CHA2DS2-VASc = 4, HAS-BLED = 2), and 22.7% did not resume OAC treatment (median age = 81, CHA2DS2-VASc = 4, HAS-BLED = 3). Switch from VKA to NOAC occurred among 9.5%. Since 2009, the trend in OAC resumption increased (P-value <0.0001), in particular with NOACs (P-value <0.0001). Follow-up started 90 days after discharge, and time-varying multiple Cox regression analyses were used for comparisons. Compared with non-resumption, VKA and NOAC resumption were associated with lower hazard [95% confidence interval (CI)] of all-cause mortality [hazard ratio (HR) 0.48 (0.42–0.53) and HR 0.55 (0.47–0.66), respectively] and ischaemic stroke [HR 0.56 (0.43–0.72) and HR 0.54 (0.35–0.82), respectively], increased major bleeding hazard [HR 1.30 (1.03–1.64) and HR 1.15 (0.81–1.63), respectively], and similar hazard of recurrent traumatic injury [HR 0.93 (0.73–1.18) and HR 0.87 (0.60–1.27), respectively]. Conclusion AF patients resuming VKA and NOAC treatment following traumatic injury have lower hazard of all-cause mortality and ischaemic stroke, increased hazard of major bleeding but without additional hazards of recurrent traumatic injury. Withholding OAC following a traumatic injury in AF patients may not be warranted. View largeDownload slide View largeDownload slide Atrial fibrillation, Traumatic injury, Oral anticoagulation, Warfarin, NOAC, Bleeding Introduction Patients with atrial fibrillation (AF) requiring oral anticoagulants (OAC) are a fragile group of patients and a clinical challenge if they experience a traumatic injury.1–5 For instance, if a patient with AF is admitted due to a head injury caused by a ground-level fall, the occurrence of death during the injury hospitalization is estimated to be 24%, whereas this constitutes 32% if patients are treated with OACs before the ground-level fall admission.6 After traumatic injury, doctors and patients face the decision of whether to resume OAC therapy, a complex decision that needs careful consideration of the balance between benefit (stroke prevention) and risk (bleeding) of OAC treatment.7 Resumption of OAC treatment will reduce the stroke risk, but increase the bleeding risk, which in these fragile patients may be a major concern in the perspective of a risk for a recurrent traumatic injury. Non-vitamin K antagonist (VKA) oral anticoagulants (NOACs) are frequently used among elderly AF patients8; however, following a traumatic injury it is uncertain to what extent NOACs and VKAs are used and what the risks of outcomes are with respectively NOAC and VKA resumption compared with non-resumption. This study focused on OAC-treated AF patients who experienced a traumatic injury, and we examined the OAC resumption pattern and risk of all-cause mortality, ischaemic stroke, major bleeding, and recurrent traumatic injury associated with NOAC and VKA resumption compared with no OAC resumption. Methods Data sources All residents of Denmark are at birth or immigration, given a unique personal identification number that enabled us to cross-link individual information from the following Danish nationwide registries: (i) The Danish national patient registry holds information about all hospital contacts since 1978. A hospital contact is coded with one primary, and if appropriate, one or more secondary diagnosis codes based on the International Classification of Diseases 10th revision (ICD-10).9 (ii) The Danish national prescription registry keeps information about all filled prescriptions since 1995 including the Anatomical Therapeutic Chemical (ATC) codes, the day the prescription was filled, the package size, and dosages.10 (iii) The Danish civil registration system registers vital status.11 Study population and outcomes We conducted an observational study from 1 January 2005 to 31 December 2016. The study population constituted AF patients treatment with VKA or NOAC [dabigatran (since August 2011), rivaroxaban (since February 2012), or apixaban (since December 2012)] the day before a traumatic injury hospital admission. Traumatic injury involved hospitalization with traumatic cranial or brain injury (fracture of the skull, injury of optic nerve, intracranial injury, or crushing or multiple injuries of head), traumatic hip fracture (femoral neck, pertrochanteric, or subtrochanteric fracture), or traumatic torso or abdominal injury (serious open wound of thoracic wall, abdominal wall, or pelvis; injury of blood vessels, intrathoracic organs, intra-abdominal organs, and pelvic organs; and traumatic pneumothorax or haemothorax). Validation studies of administrative data report high positive predictive values for AF, hip fracture, and traumatic brain injury.12–14 Traumatic torso or abdominal injury was defined as done in a previous study from 2015.15 A quarantine period of 90 days subsequent to the discharge day was introduced to allow patients to fill a new prescription of an OAC drug following the traumatic injury. The package size of an OAC agent runs for a maximum of 3 months, and it is plausible that if patients resumed OAC treatment they finished the package from before the traumatic injury at first. All patients were followed up for 90 days, and with the 90 days of quarantine period, we were able to categorize patients into resumption and non-resumption of treatment and thereby limit misclassification and avoid conditioning on the future.16 The exclusion criteria are shown in Figure 1. Figure 1 View largeDownload slide Selection of the study population. NOAC, non-vitamin K antagonist oral anticoagulant; OAC, oral anticoagulation; VKA, vitamin K antagonist. Figure 1 View largeDownload slide Selection of the study population. NOAC, non-vitamin K antagonist oral anticoagulant; OAC, oral anticoagulation; VKA, vitamin K antagonist. The three treatment groups were resumption of VKA, resumption of NOAC, or no OAC resumption, and ATC codes are listed in the Supplementary material online, Table S1. Follow-up began after the 90 days of quarantine period and was limited to the first 3 years after the 90 days of quarantine period (90 days + 3 years). Outcomes of interest during follow-up were the earliest of all-cause mortality, ischaemic stroke, major bleeding, and recurrent traumatic injury. Outcome status was last assessed on 31 December 2016. All ICD-10 and ATC codes used to define the study population, treatment groups, and outcomes are listed in the Supplementary material online, Table S1. Concomitant pharmacotherapy and co-morbidity Concomitant pharmacotherapy was identified from filled prescriptions for a period of 90 days prior to baseline. Co-morbidities were identified from diagnosis codes registered within a period of 10 years prior to baseline. ATC codes and ICD-10 codes used to define concomitant pharmacotherapy and co-morbidities are listed in the Supplementary material online, Table S1. All patients were assessed for risk of stroke with the CHA2DS2-VASc score17 [Congestive heart failure, Hypertension, Age ≥ 75 years (2 points), Diabetes mellitus, Stroke/TE/transient ischaemic attack (2 points), Vascular disease, Age 65–74 years, and female Sex] and risk of bleeding with the HAS-BLED score [Hypertension, Abnormal renal or liver function, Stroke/TE, previous Bleeding, Labile international normalized ratio (INR; left out because data are unavailable), Elderly (age ≥ 65 years), and concomitant drugs (antiplatelet agents, non-steroidal anti-inflammatory drugs)/alcohol abuse]. Statistical analysis Time trends in OAC resumption according to calendar year were analysed with Cochrane Armitage trend tests in the period 2005–16. Baseline was set as 90 days after discharge from traumatic injury. We estimated the outcome-specific absolute risk (cumulative incidence) by the Aalen–Johansen method separately in the treatment groups to account for competing risks of death.18 Stroke, major bleeding, and recurrent traumatic injury were not accounted for as competing events. Outcome-specific hazards in the three treatment groups were modelled by multiple Cox regression. Each outcome-specific model was adjusted for potential confounders as determined by published literature and associated expert knowledge. All models were adjusted for calendar year, age (continuous), sex, and type of traumatic injury. All-cause mortality models were further adjusted for CHA2DS2-VASc and HAS-BLED factors, stroke models for CHA2DS2-VASc factors, major bleeding models for HAS-BLED factors, and recurrent traumatic injury models for CHA2DS2-VASc and HAS-BLED factors, osteoporosis, dementia, and benzodiazepine usage. For each outcome, the two Cox regression models were obtained, one using only baseline information and another using dynamic updates of all time-varying covariates (treatment group, co-morbidity, concomitant pharmacotherapy, calendar year, and age). Based on the outcome-specific Cox regression models using only baseline information, we computed differences in standardized absolute 1-year risks between the three treatment groups (g-formula).19,20 The standardized risks for treatment Group ‘A’ were obtained as the average of the predicted absolute 1-year risks in a copy of the real data where the treatment variable was set to ‘A’. The risks were predicted by combining the outcome-specific Cox regression models.20 The standardized differences were supplied with 95% bootstrap confidence limits based on 1000 bootstrap data sets. The level of significance was set at 5%. Data management and statistical analyses were performed using SAS (version 9.4 for Windows, SAS Institute, Cary, NC, USA) and R [R Core Team (2016). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria]. Ethics Retrospective registry-based studies do not require approval from the Research Ethical Committee System. The Danish Data Protection Agency approved the use of data for this study (ref.no: 2007-58-0015/GEH-2014-012 I-Suite no: 02720). Results Study population At baseline, which was on Day 90 following discharge from traumatic injury, we included a total of 4541 AF patients, treated with OAC before admission (Figure 1). During the quarantine period of 90 days, 1695 patients were excluded as they suffered from one of the outcomes of interest. Supplementary material online, Table S2 reports the characteristics of the population at the time of discharge from traumatic injury, and Supplementary material online, Figure S1 illustrates outcome-specific cumulative incidences during the quarantine period of 90 days. Among the 4541 included patients at baseline (Table 1), the resumption pattern was as follows: non-resumption 22.7%, VKA 60.6%, and NOAC 16.7% (dabigatran, n = 333, 7.3%; rivaroxaban, n = 211, 4.7%; and apixaban, n = 213, 4.7%). The median age was 81, 80, and 81 years in the non-resumption, VKA, and NOAC groups, respectively. All groups had a median CHA2DS2-VASc score of 4, but the median HAS-BLED score was highest in the non-resumption group being 3. In the non-resumption group, 40.8% had suffered from a traumatic brain injury, whereas among patients resuming VKA and NOAC this was 27.4% and 29.6%, respectively. In addition, concomitant usage of aspirin (27.1%), adenosine diphosphate receptor antagonists (4.8%), and benzodiazepines (25.3%) were more frequent among non-resumption than patients resuming OAC treatment. Table 1 Patient characteristics at baseline [AuthorQuery id="AQ20" rid="20"]?>   Non-resumption  VKA  NOAC  P-value  n (%)  1031 (22.7)  2753 (60.6)  757 (16.7)    Men, n (%)  513 (49.8)  1347 (48.9)  309 (40.8)  <0.001  Age, median [IQR]  81 [75, 86]  80 [74, 85]  81 [74, 87]  <0.001  CHA2DS2-VASc, median [IQR]  4 [3, 5]  4 [3, 5]  4 [3, 5]  0.289  HAS-BLED, median [IQR]  3 [2, 3]  2 [2, 3]  2 [2, 3]  <0.001  Type of injury, n (%)        <0.001   Traumatic brain injury  421 (40.8)  755 (27.4)  224 (29.6)     Hip fracture  475 (46.1)  1530 (55.6)  426 (56.3)     Traumatic torso or abdominal injury  135 (13.1)  468 (17.0)  107 (14.1)    OAC before traumatic injury, n (%)      <0.001   VKA  930 (90.2)  2750 (99.9)  72 (9.5)     NOAC  101 (9.8)  3 (0.1)  685 (90.5)    Co-morbidities, n (%)   Stroke  263 (25.5)  588 (21.4)  217 (28.7)  <0.001   Myocardial infarction  109 (10.6)  247 (9.0)  62 (8.2)  0.181   Ischaemic heart disease  364 (35.3)  950 (34.5)  203 (26.8)  <0.001   Peripheral artery disease  83 (8.1)  183 (6.6)  35 (4.6)  0.016   Heart failure  327 (31.7)  849 (30.8)  194 (25.6)  0.010   Diabetes mellitus  141 (13.7)  409 (14.9)  112 (14.8)  0.646   Hypertension  535 (51.9)  1686 (61.2)  410 (54.2)  <0.001   Chronic kidney disease  102 (9.9)  214 (7.8)  58 (7.7)  0.088   Liver disease  14 (1.4)  47 (1.7)  14 (1.8)  0.676   Prior bleeding  466 (45.2)  802 (29.1)  233 (30.8)  <0.001   Alcohol misuse  45 (4.4)  119 (4.3)  50 (6.6)  0.027   Cancer  182 (17.7)  444 (16.1)  134 (17.7)  0.395   Dementia  92 (8.9)  157 (5.7)  89 (11.8)  <0.001   Osteoporosis  187 (18.1)  502 (18.2)  198 (26.2)  <0.001  Concomitant medication, n (%)   ADP receptor antagonists  50 (4.8)  41 (1.5)  13 (1.7)  <0.001   Aspirin  279 (27.1)  488 (17.7)  36 (4.8)  <0.001   Dipyridamole  15 (1.5)  9 (0.3)  0 (0.0)  <0.001   Non-steroid anti-inflammatory drug  63 (6.1)  174 (6.3)  43 (5.7)  0.808   Loop diuretics  411 (39.9)  1168 (42.4)  275 (36.3)  0.008   Beta-blockers  487 (47.2)  1397 (50.7)  427 (56.4)  0.001   Calcium channel blockers  197 (19.1)  652 (23.7)  148 (19.6)  0.002   Renin–angiotensin system blocker  369 (35.8)  1107 (40.2)  267 (35.3)  0.007   Digoxin  344 (33.4)  976 (35.5)  189 (25.0)  <0.001   Benzodiazepine  261 (25.3)  590 (21.4)  173 (22.9)  0.038    Non-resumption  VKA  NOAC  P-value  n (%)  1031 (22.7)  2753 (60.6)  757 (16.7)    Men, n (%)  513 (49.8)  1347 (48.9)  309 (40.8)  <0.001  Age, median [IQR]  81 [75, 86]  80 [74, 85]  81 [74, 87]  <0.001  CHA2DS2-VASc, median [IQR]  4 [3, 5]  4 [3, 5]  4 [3, 5]  0.289  HAS-BLED, median [IQR]  3 [2, 3]  2 [2, 3]  2 [2, 3]  <0.001  Type of injury, n (%)        <0.001   Traumatic brain injury  421 (40.8)  755 (27.4)  224 (29.6)     Hip fracture  475 (46.1)  1530 (55.6)  426 (56.3)     Traumatic torso or abdominal injury  135 (13.1)  468 (17.0)  107 (14.1)    OAC before traumatic injury, n (%)      <0.001   VKA  930 (90.2)  2750 (99.9)  72 (9.5)     NOAC  101 (9.8)  3 (0.1)  685 (90.5)    Co-morbidities, n (%)   Stroke  263 (25.5)  588 (21.4)  217 (28.7)  <0.001   Myocardial infarction  109 (10.6)  247 (9.0)  62 (8.2)  0.181   Ischaemic heart disease  364 (35.3)  950 (34.5)  203 (26.8)  <0.001   Peripheral artery disease  83 (8.1)  183 (6.6)  35 (4.6)  0.016   Heart failure  327 (31.7)  849 (30.8)  194 (25.6)  0.010   Diabetes mellitus  141 (13.7)  409 (14.9)  112 (14.8)  0.646   Hypertension  535 (51.9)  1686 (61.2)  410 (54.2)  <0.001   Chronic kidney disease  102 (9.9)  214 (7.8)  58 (7.7)  0.088   Liver disease  14 (1.4)  47 (1.7)  14 (1.8)  0.676   Prior bleeding  466 (45.2)  802 (29.1)  233 (30.8)  <0.001   Alcohol misuse  45 (4.4)  119 (4.3)  50 (6.6)  0.027   Cancer  182 (17.7)  444 (16.1)  134 (17.7)  0.395   Dementia  92 (8.9)  157 (5.7)  89 (11.8)  <0.001   Osteoporosis  187 (18.1)  502 (18.2)  198 (26.2)  <0.001  Concomitant medication, n (%)   ADP receptor antagonists  50 (4.8)  41 (1.5)  13 (1.7)  <0.001   Aspirin  279 (27.1)  488 (17.7)  36 (4.8)  <0.001   Dipyridamole  15 (1.5)  9 (0.3)  0 (0.0)  <0.001   Non-steroid anti-inflammatory drug  63 (6.1)  174 (6.3)  43 (5.7)  0.808   Loop diuretics  411 (39.9)  1168 (42.4)  275 (36.3)  0.008   Beta-blockers  487 (47.2)  1397 (50.7)  427 (56.4)  0.001   Calcium channel blockers  197 (19.1)  652 (23.7)  148 (19.6)  0.002   Renin–angiotensin system blocker  369 (35.8)  1107 (40.2)  267 (35.3)  0.007   Digoxin  344 (33.4)  976 (35.5)  189 (25.0)  <0.001   Benzodiazepine  261 (25.3)  590 (21.4)  173 (22.9)  0.038  Characteristics of the study population at the time of baseline. Table 1 Patient characteristics at baseline [AuthorQuery id="AQ20" rid="20"]?>   Non-resumption  VKA  NOAC  P-value  n (%)  1031 (22.7)  2753 (60.6)  757 (16.7)    Men, n (%)  513 (49.8)  1347 (48.9)  309 (40.8)  <0.001  Age, median [IQR]  81 [75, 86]  80 [74, 85]  81 [74, 87]  <0.001  CHA2DS2-VASc, median [IQR]  4 [3, 5]  4 [3, 5]  4 [3, 5]  0.289  HAS-BLED, median [IQR]  3 [2, 3]  2 [2, 3]  2 [2, 3]  <0.001  Type of injury, n (%)        <0.001   Traumatic brain injury  421 (40.8)  755 (27.4)  224 (29.6)     Hip fracture  475 (46.1)  1530 (55.6)  426 (56.3)     Traumatic torso or abdominal injury  135 (13.1)  468 (17.0)  107 (14.1)    OAC before traumatic injury, n (%)      <0.001   VKA  930 (90.2)  2750 (99.9)  72 (9.5)     NOAC  101 (9.8)  3 (0.1)  685 (90.5)    Co-morbidities, n (%)   Stroke  263 (25.5)  588 (21.4)  217 (28.7)  <0.001   Myocardial infarction  109 (10.6)  247 (9.0)  62 (8.2)  0.181   Ischaemic heart disease  364 (35.3)  950 (34.5)  203 (26.8)  <0.001   Peripheral artery disease  83 (8.1)  183 (6.6)  35 (4.6)  0.016   Heart failure  327 (31.7)  849 (30.8)  194 (25.6)  0.010   Diabetes mellitus  141 (13.7)  409 (14.9)  112 (14.8)  0.646   Hypertension  535 (51.9)  1686 (61.2)  410 (54.2)  <0.001   Chronic kidney disease  102 (9.9)  214 (7.8)  58 (7.7)  0.088   Liver disease  14 (1.4)  47 (1.7)  14 (1.8)  0.676   Prior bleeding  466 (45.2)  802 (29.1)  233 (30.8)  <0.001   Alcohol misuse  45 (4.4)  119 (4.3)  50 (6.6)  0.027   Cancer  182 (17.7)  444 (16.1)  134 (17.7)  0.395   Dementia  92 (8.9)  157 (5.7)  89 (11.8)  <0.001   Osteoporosis  187 (18.1)  502 (18.2)  198 (26.2)  <0.001  Concomitant medication, n (%)   ADP receptor antagonists  50 (4.8)  41 (1.5)  13 (1.7)  <0.001   Aspirin  279 (27.1)  488 (17.7)  36 (4.8)  <0.001   Dipyridamole  15 (1.5)  9 (0.3)  0 (0.0)  <0.001   Non-steroid anti-inflammatory drug  63 (6.1)  174 (6.3)  43 (5.7)  0.808   Loop diuretics  411 (39.9)  1168 (42.4)  275 (36.3)  0.008   Beta-blockers  487 (47.2)  1397 (50.7)  427 (56.4)  0.001   Calcium channel blockers  197 (19.1)  652 (23.7)  148 (19.6)  0.002   Renin–angiotensin system blocker  369 (35.8)  1107 (40.2)  267 (35.3)  0.007   Digoxin  344 (33.4)  976 (35.5)  189 (25.0)  <0.001   Benzodiazepine  261 (25.3)  590 (21.4)  173 (22.9)  0.038    Non-resumption  VKA  NOAC  P-value  n (%)  1031 (22.7)  2753 (60.6)  757 (16.7)    Men, n (%)  513 (49.8)  1347 (48.9)  309 (40.8)  <0.001  Age, median [IQR]  81 [75, 86]  80 [74, 85]  81 [74, 87]  <0.001  CHA2DS2-VASc, median [IQR]  4 [3, 5]  4 [3, 5]  4 [3, 5]  0.289  HAS-BLED, median [IQR]  3 [2, 3]  2 [2, 3]  2 [2, 3]  <0.001  Type of injury, n (%)        <0.001   Traumatic brain injury  421 (40.8)  755 (27.4)  224 (29.6)     Hip fracture  475 (46.1)  1530 (55.6)  426 (56.3)     Traumatic torso or abdominal injury  135 (13.1)  468 (17.0)  107 (14.1)    OAC before traumatic injury, n (%)      <0.001   VKA  930 (90.2)  2750 (99.9)  72 (9.5)     NOAC  101 (9.8)  3 (0.1)  685 (90.5)    Co-morbidities, n (%)   Stroke  263 (25.5)  588 (21.4)  217 (28.7)  <0.001   Myocardial infarction  109 (10.6)  247 (9.0)  62 (8.2)  0.181   Ischaemic heart disease  364 (35.3)  950 (34.5)  203 (26.8)  <0.001   Peripheral artery disease  83 (8.1)  183 (6.6)  35 (4.6)  0.016   Heart failure  327 (31.7)  849 (30.8)  194 (25.6)  0.010   Diabetes mellitus  141 (13.7)  409 (14.9)  112 (14.8)  0.646   Hypertension  535 (51.9)  1686 (61.2)  410 (54.2)  <0.001   Chronic kidney disease  102 (9.9)  214 (7.8)  58 (7.7)  0.088   Liver disease  14 (1.4)  47 (1.7)  14 (1.8)  0.676   Prior bleeding  466 (45.2)  802 (29.1)  233 (30.8)  <0.001   Alcohol misuse  45 (4.4)  119 (4.3)  50 (6.6)  0.027   Cancer  182 (17.7)  444 (16.1)  134 (17.7)  0.395   Dementia  92 (8.9)  157 (5.7)  89 (11.8)  <0.001   Osteoporosis  187 (18.1)  502 (18.2)  198 (26.2)  <0.001  Concomitant medication, n (%)   ADP receptor antagonists  50 (4.8)  41 (1.5)  13 (1.7)  <0.001   Aspirin  279 (27.1)  488 (17.7)  36 (4.8)  <0.001   Dipyridamole  15 (1.5)  9 (0.3)  0 (0.0)  <0.001   Non-steroid anti-inflammatory drug  63 (6.1)  174 (6.3)  43 (5.7)  0.808   Loop diuretics  411 (39.9)  1168 (42.4)  275 (36.3)  0.008   Beta-blockers  487 (47.2)  1397 (50.7)  427 (56.4)  0.001   Calcium channel blockers  197 (19.1)  652 (23.7)  148 (19.6)  0.002   Renin–angiotensin system blocker  369 (35.8)  1107 (40.2)  267 (35.3)  0.007   Digoxin  344 (33.4)  976 (35.5)  189 (25.0)  <0.001   Benzodiazepine  261 (25.3)  590 (21.4)  173 (22.9)  0.038  Characteristics of the study population at the time of baseline. From 2009 to 2016, an increase in overall resumption of OAC treatment occurred from 68.7% to 85.6%, respectively (P-value <0.0001 for increasing trend), and a rapid yearly increase of NOAC usage occurred since 2011 (P-value <0.0001 for increasing trend). In 2016, an NOAC was prescribed for 39.6% of patients following a traumatic injury (Figure 2). Figure 2 View largeDownload slide Time trends showing resumption pattern of OAC following a traumatic injury from 2005 to 2016. NOAC, non-vitamin K antagonist oral anticoagulant; OAC, oral anticoagulation; VKA, vitamin K antagonist. Figure 2 View largeDownload slide Time trends showing resumption pattern of OAC following a traumatic injury from 2005 to 2016. NOAC, non-vitamin K antagonist oral anticoagulant; OAC, oral anticoagulation; VKA, vitamin K antagonist. Oral anticoagulants resumption vs. non-resumption and associated risk of outcomes During the follow-up period of 3 years, 1438 died, 285 suffered from an ischaemic stroke, 442 from major bleeding of which 130 were intracranial bleedings (non-resumption, VKA resumption, and NOAC resumption experienced 37, 83, and 10 intracranial bleedings, respectively), and 357 from recurrent traumatic injury. The cumulative risks obtained with the Aalen–Johansen method without covariate adjustment are shown in Supplementary material online, Figure S2. Take home figure shows the standardized absolute risks. The 1-year standardized absolute risks and absolute risk differences are reported in Table 2. Table 2 Standardized absolute risks   At 1 year   No. of events/ no. at risk  Standardized absolute risk (95% CI)  Absolute risk difference (95% CI)a  Absolute risk difference (95% CI)a  All-cause mortality   Non-resumption  224/723  19.5% (17.5 to 21.8%)  Reference  —   VKA  364/2101  15.3% (14.4 to 16.7%)  −4.2% (−6.3 to −2.7%)a  Reference   NOAC  92/388  15.1% (12.5 to 16.9%)  −4.4% (−7.2 to −1.4%)a  −0.02% (−2.7 to 1.4%)  Stroke/thrombo-embolism   Non-resumption  45/696  4.0% (3.2 to 5.1%)  Reference  —   VKA  82/2056  3.4% (2.8 to 4.1%)  −0.6% (−1.7 to 0.4%)  Reference   NOAC  20/378  2.8% (1.8 to 4.1%)  −1.2% (−2.7 to 0.3%)  −0.6% (−1.9 to 0.8%)  Major bleeding   Non-resumption  48/695  4.2% (3.3 to 5.3%)  Reference  —   VKA  120/2022  4.8% (4.1 to 5.6%)  0.5% (−0.5 to 1.5%)  Reference   NOAC  20/379  3.6% (2.5 to 5.0%)  −0.6% (−2.1 to 1.1%)  −1.1% (−2.4 to 0.3%)  Recurrent traumatic injury   Non-resumption  51/693  4.4% (3.5 to 5.6%)  Reference  —   VKA  103/2026  4.4% (3.7 to 5.1%)  0.0% (−1.2 to 1.0%)  Reference   NOAC  27/373  4.0% (2.8 to 5.5%)  −0.4% (−2.0 to 1.2%)  −0.4% (−1.8 to 1.2%)    At 1 year   No. of events/ no. at risk  Standardized absolute risk (95% CI)  Absolute risk difference (95% CI)a  Absolute risk difference (95% CI)a  All-cause mortality   Non-resumption  224/723  19.5% (17.5 to 21.8%)  Reference  —   VKA  364/2101  15.3% (14.4 to 16.7%)  −4.2% (−6.3 to −2.7%)a  Reference   NOAC  92/388  15.1% (12.5 to 16.9%)  −4.4% (−7.2 to −1.4%)a  −0.02% (−2.7 to 1.4%)  Stroke/thrombo-embolism   Non-resumption  45/696  4.0% (3.2 to 5.1%)  Reference  —   VKA  82/2056  3.4% (2.8 to 4.1%)  −0.6% (−1.7 to 0.4%)  Reference   NOAC  20/378  2.8% (1.8 to 4.1%)  −1.2% (−2.7 to 0.3%)  −0.6% (−1.9 to 0.8%)  Major bleeding   Non-resumption  48/695  4.2% (3.3 to 5.3%)  Reference  —   VKA  120/2022  4.8% (4.1 to 5.6%)  0.5% (−0.5 to 1.5%)  Reference   NOAC  20/379  3.6% (2.5 to 5.0%)  −0.6% (−2.1 to 1.1%)  −1.1% (−2.4 to 0.3%)  Recurrent traumatic injury   Non-resumption  51/693  4.4% (3.5 to 5.6%)  Reference  —   VKA  103/2026  4.4% (3.7 to 5.1%)  0.0% (−1.2 to 1.0%)  Reference   NOAC  27/373  4.0% (2.8 to 5.5%)  −0.4% (−2.0 to 1.2%)  −0.4% (−1.8 to 1.2%)  a The corresponding absolute risk differences tested for significant differences between standardized absolute risks. Statistical significance was obtained when the 95% confidence limits for difference in absolute risks did not contain zero. Table 2 Standardized absolute risks   At 1 year   No. of events/ no. at risk  Standardized absolute risk (95% CI)  Absolute risk difference (95% CI)a  Absolute risk difference (95% CI)a  All-cause mortality   Non-resumption  224/723  19.5% (17.5 to 21.8%)  Reference  —   VKA  364/2101  15.3% (14.4 to 16.7%)  −4.2% (−6.3 to −2.7%)a  Reference   NOAC  92/388  15.1% (12.5 to 16.9%)  −4.4% (−7.2 to −1.4%)a  −0.02% (−2.7 to 1.4%)  Stroke/thrombo-embolism   Non-resumption  45/696  4.0% (3.2 to 5.1%)  Reference  —   VKA  82/2056  3.4% (2.8 to 4.1%)  −0.6% (−1.7 to 0.4%)  Reference   NOAC  20/378  2.8% (1.8 to 4.1%)  −1.2% (−2.7 to 0.3%)  −0.6% (−1.9 to 0.8%)  Major bleeding   Non-resumption  48/695  4.2% (3.3 to 5.3%)  Reference  —   VKA  120/2022  4.8% (4.1 to 5.6%)  0.5% (−0.5 to 1.5%)  Reference   NOAC  20/379  3.6% (2.5 to 5.0%)  −0.6% (−2.1 to 1.1%)  −1.1% (−2.4 to 0.3%)  Recurrent traumatic injury   Non-resumption  51/693  4.4% (3.5 to 5.6%)  Reference  —   VKA  103/2026  4.4% (3.7 to 5.1%)  0.0% (−1.2 to 1.0%)  Reference   NOAC  27/373  4.0% (2.8 to 5.5%)  −0.4% (−2.0 to 1.2%)  −0.4% (−1.8 to 1.2%)    At 1 year   No. of events/ no. at risk  Standardized absolute risk (95% CI)  Absolute risk difference (95% CI)a  Absolute risk difference (95% CI)a  All-cause mortality   Non-resumption  224/723  19.5% (17.5 to 21.8%)  Reference  —   VKA  364/2101  15.3% (14.4 to 16.7%)  −4.2% (−6.3 to −2.7%)a  Reference   NOAC  92/388  15.1% (12.5 to 16.9%)  −4.4% (−7.2 to −1.4%)a  −0.02% (−2.7 to 1.4%)  Stroke/thrombo-embolism   Non-resumption  45/696  4.0% (3.2 to 5.1%)  Reference  —   VKA  82/2056  3.4% (2.8 to 4.1%)  −0.6% (−1.7 to 0.4%)  Reference   NOAC  20/378  2.8% (1.8 to 4.1%)  −1.2% (−2.7 to 0.3%)  −0.6% (−1.9 to 0.8%)  Major bleeding   Non-resumption  48/695  4.2% (3.3 to 5.3%)  Reference  —   VKA  120/2022  4.8% (4.1 to 5.6%)  0.5% (−0.5 to 1.5%)  Reference   NOAC  20/379  3.6% (2.5 to 5.0%)  −0.6% (−2.1 to 1.1%)  −1.1% (−2.4 to 0.3%)  Recurrent traumatic injury   Non-resumption  51/693  4.4% (3.5 to 5.6%)  Reference  —   VKA  103/2026  4.4% (3.7 to 5.1%)  0.0% (−1.2 to 1.0%)  Reference   NOAC  27/373  4.0% (2.8 to 5.5%)  −0.4% (−2.0 to 1.2%)  −0.4% (−1.8 to 1.2%)  a The corresponding absolute risk differences tested for significant differences between standardized absolute risks. Statistical significance was obtained when the 95% confidence limits for difference in absolute risks did not contain zero. Take home figure View largeDownload slide Standardized absolute risks of outcomes associated with non-resumption vs. resumption of VKA and NOAC following traumatic injury. For each treatment option separately, the standardized absolute risks are based on the outcome-specific multiple Cox regression models and obtained as averages of the predicted risks had all patients received this treatment. NOAC, non-vitamin K antagonist oral anticoagulant; VKA, vitamin K antagonist. Take home figure View largeDownload slide Standardized absolute risks of outcomes associated with non-resumption vs. resumption of VKA and NOAC following traumatic injury. For each treatment option separately, the standardized absolute risks are based on the outcome-specific multiple Cox regression models and obtained as averages of the predicted risks had all patients received this treatment. NOAC, non-vitamin K antagonist oral anticoagulant; VKA, vitamin K antagonist. Figure 3 shows the adjusted hazard ratios (HRs) obtained from time-varying Cox regression models. Compared with non-resumption, VKA and NOAC resumption were associated with significantly lower hazard of all-cause mortality and ischaemic stroke and an increased associated hazard of major bleeding. For traumatic injury, the HR was not decreased or increased with resumption of VKA and NOAC vs. non-resumption. Predictors for the hazard rate of major bleeding following a traumatic injury are presented in Supplementary material online, Table S3. Figure 3 View largeDownload slide Adjusted hazard ratios of outcomes associated with non-resumption vs. resumption of VKA and NOAC following traumatic injury. Time-dependent multiple Cox regression models were used to allow patients to change treatment group during follow-up. NOAC, non-vitamin K antagonist oral anticoagulant; VKA, vitamin K antagonist. Figure 3 View largeDownload slide Adjusted hazard ratios of outcomes associated with non-resumption vs. resumption of VKA and NOAC following traumatic injury. Time-dependent multiple Cox regression models were used to allow patients to change treatment group during follow-up. NOAC, non-vitamin K antagonist oral anticoagulant; VKA, vitamin K antagonist. Type of traumatic injury The adjusted HRs were estimated in a subgroup analysis based on the type of traumatic injury. The results in the subgroups (see Supplementary material online, Table S4) according to traumatic brain injury, hip fracture, and traumatic torso and abdominal injury were generally similar to the results in Figure 3. Switch of oral anticoagulants following traumatic injury Among those resuming OAC treatment, it was most common to resume the same OAC treatment as before the traumatic injury (Table 1); 99.9% of the patients who resumed a VKA were also on treatment with a VKA before the traumatic injury, and among those treated with NOAC after the traumatic injury, 90.5% were treated with NOAC and 9.5% with VKA before the traumatic injury. We evaluated those patients treated with VKA before the traumatic injury who either resumed VKA treatment or switched to a NOAC after the traumatic injury (see Supplementary material online, Table S4). Compared with non-resumption, staying on VKA treatment or switching to NOAC were both associated with a significantly lower hazard of all-cause mortality and ischaemic stroke but an increased hazard of bleeding; however, for NOAC switchers, the lower ischaemic stroke and increased major bleeding hazards were not significant. Sensitivity analyses A subgroup analysis was conducted including only patients with HAS-BLED score ≥4 (see Supplementary material online, Table S3). The results were broadly similar to the main results in Figure 3, albeit the analysis was influenced by lack of power. When testing the impact of changing the quarantine period to 30, 60, or 180 days, the results were comparable to the main results (see Supplementary material online, Table S5 and Figure 3). Discussion In our nationwide observational study, including 4288 OAC-treated patients with AF who suffered a traumatic injury, the main findings were as follows: (i) overall 60.6% resumed a VKA, 16.7% resumed a NOAC, and 22.7% did not resume OAC treatment during the study period 2005–16; (ii) since 2009 an increase in OAC resumption has occurred; and (iii) compared with non-resumption, resumption of VKA or NOAC treatment was associated with significantly lower hazard of all-cause mortality and ischaemic stroke, increased hazard of major bleedings, and no difference in the hazard of recurrent traumatic injury. Since more than 50% of the study population comprised octogenarians (median age was 80–81 years), it is plausible that falls were the most common cause of the traumatic injury hospitalization4,5; however, we were not able to identify the specific cause of the traumatic injury in the Danish national patient registry. The results from this study are particularly interesting assuming that the traumatic injury was caused by a fall, because fall risk among OAC-treated AF patients is feared with regard to major and fatal bleedings.21–23 Underuse of OAC treatment has been reported in several countries,24,25 and two main reasons for a doctor’s non-prescription of VKA are concerns regarding fall and bleeding risks.21 Hylek et al.7 found that previous falls were the most frequent physician-cited reason for not prescribing OAC among AF patients aged 80 years or older. In our study, 22.7% did not resume OAC treatment, and these patients had a higher occurrence of traumatic brain injury at baseline compared with those patients resuming VKA and NOAC treatment. In a study by Liu et al.15, the authors found that suffering a traumatic brain injury was associated with low odds of VKA resumption compared with non-resumption. The authors discussed that the low odds of VKA resumption after a traumatic brain injury may be related to doctors’ fear of fall risk in fragile elderly AF patients and perhaps that doctors overestimate a patient’s predicted bleeding risk. In our study, patients not resuming OAC had the highest predicted bleeding risk (median HAS-BLED = 3), however, we found that the 1-year standardized absolute risks of recurrent traumatic injury ranged from 4.0% to 4.4% without a difference between OAC resumption vs. non-resumption. These results suggest that perhaps doctors may consider being less reluctant with OAC due to recurrent fall risk (and the risk of fall-related bleedings), since the fall risk does not seem to vary between resumption of OAC and non-resumption patients. Moreover, in our study, a subgroup analysis including only patients with baseline HAS-BLED score ≥4 showed that the hazard of recurrent traumatic injury did not differ significantly between VKA and NOAC resumption compared with non-resumption. However, we found an increased hazard of major bleeding with VKA and NOAC resumption compared with non-resumption, which indicates that doctors’ concerns regarding bleeding risk with OAC treatment are relevant. Our study showed that VKA and NOAC resumption was associated with significantly lower hazard of ischaemic stroke and all-cause mortality. A retrospective study by Albrecht et al.26 examined the risk of stroke and major bleeding with warfarin usage after a traumatic brain injury and concluded that the benefits in stroke reduction with warfarin usage outweighed the risk of major bleeding. Similarly, a subgroup analysis in our study including only patients suffering from a traumatic brain injury pointed to a lower associated hazard of ischaemic stroke and all-cause mortality with VKA and NOAC resumption; however, lack of power in the NOAC group for the ischaemic stroke outcome could not confirm a significantly lower ischaemic stroke HR with NOACs. Similarly to Albrecht et al., the hazard of major bleeding was significantly higher in our study among patients resuming VKA and NOAC treatment compared with non-resumption following a traumatic brain injury. Since 2009, an increase in OAC resumption occurred following traumatic injury. This may suggest that even though doctors fear for a patient’s fall and bleeding risks, the awareness of appropriate stroke prophylaxis has improved with the consequence that more patients resume OAC treatment following a traumatic injury.27 It was more common that patients treated with VKA prior to the traumatic injury switched to an NOAC after discharge (9.5%) than switching from an NOAC to a VKA (0.1%). Case reports have described problems with the lack of reversal for NOAC therapy among trauma patients taking NOACs.22,23 However, the increased NOAC use observed in our study could indicate that the benefits of NOACs as stroke prophylaxis might counterbalance doctors’ fear of a patient’s fall and bleeding risks. Moreover, idarucizumab for dabigatran reversal was first introduced in Denmark in March 2016.28 In our study, a switch from VKA to NOAC or staying on VKA after the traumatic injury had similar effect and safety profile compared with non-resumption. In clinical practice, an AF patient’s individual risk factors for stroke, bleeding, and recurrent fall may be weighed in the decision-making of withholding or resuming OAC treatment following a traumatic injury. However, the first key finding in our study was that withholding of OAC was generally not beneficial to AF patients. Second, in 2016, ∼15% did not still resume OAC treatment within 90 days after the traumatic injury. This implies that there is a need for increased awareness of resumption of OAC treatment following a traumatic injury. Furthermore, good clinical practice requires clinicians to evaluate and minimize a patient’s risk of recurrent falls, e.g. considerations regarding a patient’s co-morbidities and concomitant pharmacotherapy.29–31 Limitations and strengths In our observational study, it was methodologically necessary to include a quarantine period of 90 days16; however, changing the quarantine period to 30, 60, or 180 days (see Supplementary material online, Table S5) did not alter our results. The data from the Danish administrative registries did not enable us to estimate when OAC resumption would be safe during the quarantine periods. A limitation was residual confounding with unmeasured confounders, such as INR, time in therapeutic range, haemoglobin levels, serum creatinine levels, and body mass index. Another limitation may be confounding by indication, since it was possible that patients not resuming OAC treatment were frailer with a higher risk of all-cause mortality, stroke, major bleeding, and recurrent traumatic injury than those resuming VKA and NOAC treatment. Importantly, our conclusions support associations. The main strength in our study, given by the Danish administrative registries, was the opportunity to examine a research question that would be challenging to perform as a randomized controlled trial. This study was able to include a large nationwide population and evaluated both VKA and NOAC resumption by filled prescriptions as directed by doctor, and the Danish registries have been validated intensely.9,12,13 Conclusions Among the OAC-treated patients with AF who experienced a traumatic injury, we conclude that within 90 days after the traumatic injury approximately 3 of 4 resumed OAC. Resumption of VKA or NOAC treatment was associated with significantly lower hazards of all-cause mortality and ischaemic stroke, at the cost of an increased hazard of major bleedings, but without any additional hazard of recurrent traumatic injury, compared with non-resumption. Our study suggests that resuming OAC following traumatic injury may result in better clinical outcomes, including lower risk of stroke. Supplementary material Supplementary material is available at European Heart Journal online. Funding This work was supported by the Velux Foundations and Boehringer-Ingelheim. The sponsors had no influence on the study design, interpretation of results, or the decision to submit the manuscript for publication. G.H.G. is supported by an unrestricted clinical research scholarship from the Novo Nordisk Foundation. Conflict of interest: L.S.: research funding from Boehringer-Ingelheim. E.L.F.: research funding from Lundbeck Foundation, Janssen Pharmaceuticals, and BMS. M.L.: speaker fees from BMS and Bayer. K.G.: research funding from BMS. G.Y.H.L.: consultant for Bayer/Janssen, BMS/Pfizer, Biotronik, Medtronic, Boehringer-Ingelheim, Microlife, and Daiichi-Sankyo and speaker for Bayer, BMS/Pfizer, Medtronic, Boehringer-Ingelheim, Microlife, Roche, and Daiichi-Sankyo. C.T.P.: research contracts with Bayer and Biotronic and received speaker fees from Bayer and BMS. G.H.G.: research funding from Boehringer-Ingelheim, Pfizer, BMS, AstraZeneca, and Bayer and declares ownership of stocks in Lundbeck A/S, Novo Nordisk A/S, and ALK Abello Pharmaceuticals. J.B.O.: received speaker fees from Bayer, BMS, and Boehringer-Ingelheim and research funding from BMS and The Capital Region of Denmark. References 1 Schnabel RB, Yin X, Gona P, Larson MG, Beiser AS, McManus DD, Newton-Cheh C, Lubitz SA, Magnani JW, Ellinor PT, Seshadri S, Wolf PA, Vasan RS, Benjamin EJ, Levy D. 50 year trends in atrial fibrillation prevalence, incidence, risk factors, and mortality in the Framingham Heart Study: a cohort study. Lancet  2015; 386: 154– 162. Google Scholar CrossRef Search ADS PubMed  2 Healey JS, Oldgren J, Ezekowitz M, Zhu J, Pais P, Wang J, Commerford P, Jansky P, Avezum A, Sigamani A, Damasceno A, Reilly P, Grinvalds A, Nakamya J, Aje A, Almahmeed W, Moriarty A, Wallentin L, Yusuf S, Connolly SJ. Occurrence of death and stroke in patients in 47 countries 1 year after presenting with atrial fibrillation: a cohort study. Lancet  2016; 388: 1161– 1169. Google Scholar CrossRef Search ADS PubMed  3 Hadjizacharia P, O’Keeffe T, Brown CVR, Inaba K, Salim A, Chan LS, Demetriades D, Rhee P. Incidence, risk factors, and outcomes for atrial arrhythmias in trauma patients. Am Surg  2011; 77: 634– 639. Google Scholar PubMed  4 Majdan M, Mauritz W. Unintentional fall-related mortality in the elderly: comparing patterns in two countries with different demographic structure. BMJ Open  2015; 5: e008672. Google Scholar CrossRef Search ADS PubMed  5 Park YJ, Ro YS, Shin SD, Song KJ, Lee SC, Kim YJ, Kim JY, Hong KJ, Kim JE, Kim MJ, Kim SC. Age effects on case fatality rates of injury patients by mechanism. Am J Emerg Med  2016; 34: 515– 520. Google Scholar CrossRef Search ADS PubMed  6 Inui TS, Parina R, Chang DC, Inui TS, Coimbra R. Mortality after ground-level fall in the elderly patient taking oral anticoagulation for atrial fibrillation/flutter: a long-term analysis of risk versus benefit. J Trauma Acute Care Surg  2014; 76: 642– 650. Google Scholar CrossRef Search ADS PubMed  7 Hylek EM, D’Antonio J, Evans-Molina C, Shea C, Henault LE, Regan S. Translating the results of randomized trials into clinical practice. Stroke  2006; 37: 1075– 1080. Google Scholar CrossRef Search ADS PubMed  8 Staerk L, Fosbøl EL, Gadsbøll K, Sindet-Pedersen C, Pallisgaard JL, Lamberts M, Lip GYH, Torp-Pedersen C, Gislason GH, Olesen JB. Non-vitamin K antagonist oral anticoagulation usage according to age among patients with atrial fibrillation: Temporal trends 2011–2015 in Denmark. Sci Rep  2016; 6: 31477. Google Scholar CrossRef Search ADS PubMed  9 Schmidt M, Schmidt SAJ, Sandegaard JL, Ehrenstein V, Pedersen L, Sørensen HT. The Danish National Patient Registry: a review of content, data quality, and research potential. Clin Epidemiol  2015; 7: 449– 490. Google Scholar CrossRef Search ADS PubMed  10 Kildemoes HW, Sørensen HT, Hallas J. The Danish National Prescription Registry. Scand J Public Health  2011; 39: 38– 41. Google Scholar CrossRef Search ADS PubMed  11 Pedersen CB. The Danish Civil Registration System. Scand J Public Health  2011; 39: 22– 25. Google Scholar CrossRef Search ADS PubMed  12 Frost L, Andersen LV, Vestergaard P, Husted S, Mortensen LS. Trend in mortality after stroke with atrial fibrillation. Am J Med  2007; 120: 47– 53. Google Scholar CrossRef Search ADS PubMed  13 Hundrup YA, Høidrup S, Obel EB, Rasmussen NK. The validity of self-reported fractures among Danish female nurses: comparison with fractures registered in the Danish National Hospital Register. Scand J Public Health  2004; 32: 136– 143. Google Scholar CrossRef Search ADS PubMed  14 Shore AD, Mccarthy ML, Serpi T, Gertner M. Validity of administrative data for characterizing traumatic brain injury-related hospitalizations. Brain Inj  2005; 19: 613– 621. Google Scholar CrossRef Search ADS PubMed  15 Liu X, Baumgarten M, Smith G, Gambert S, Gottlieb S, Rattinger G, Albrecht J, Langenberg P, Zuckerman I. Warfarin usage among elderly atrial fibrillation patients with traumatic injury, an analysis of United States Medicare fee-for-service enrollees. J Clin Pharmacol  2015; 55: 25– 32. Google Scholar CrossRef Search ADS PubMed  16 Staerk L, Lip GYH, Olesen JB, Fosbøl EL, Pallisgaard JL, Bonde AN, Gundlund A, Lindhardt TB, Hansen ML, Torp-Pedersen C, Gislason GH. Stroke and recurrent haemorrhage associated with antithrombotic treatment after gastrointestinal bleeding in patients with atrial fibrillation: nationwide cohort study. BMJ  2015; 351: h5876. Google Scholar CrossRef Search ADS PubMed  17 Olesen JB, Lip GYH, Hansen ML, Hansen PR, Tolstrup JS, Lindhardsen J, Selmer C, Ahlehoff O, Olsen A-MS, Gislason GH, Torp-Pedersen C. Validation of risk stratification schemes for predicting stroke and thromboembolism in patients with atrial fibrillation: nationwide cohort study. BMJ  2011; 342: d124– d124. Google Scholar CrossRef Search ADS PubMed  18 Aalen OO, Johansen S. An empirical transition matrix for non-homogeneous Markov chains based on censored observations. Scand J Stat  1978; 5: 141– 150. 19 Sato T, Matsuyama Y. Marginal structural models as a tool for standardization. Epidemiology  2003; 14: 680– 686. Google Scholar CrossRef Search ADS PubMed  20 Benichou J, Gail MH. Estimates of absolute cause-specific risk in cohort studies. Biometrics  1990; 46: 813– 826. Google Scholar CrossRef Search ADS PubMed  21 Kakkar AK, Mueller I, Bassand J-P, Fitzmaurice DA, Goldhaber SZ, Goto S, Haas S, Hacke W, Lip GYH, Mantovani LG, Turpie AGG, Eickels M, van, Misselwitz F, Rushton-Smith S, Kayani G, Wilkinson P, Verheugt FWA. Risk profiles and antithrombotic treatment of patients newly diagnosed with atrial fibrillation at risk of stroke: perspectives from the international, observational, prospective GARFIELD registry. PLoS One  2013; 8: e63479. Google Scholar CrossRef Search ADS PubMed  22 Stöllberger C, Ulram A, Bastovansky A, Finsterer J. Traumatic fatal cerebral hemorrhage in an old patient with a history of multiple sclerosis under dabigatran: a case report and review of the literature. J Geriatr Cardiol  2015; 12: 83. Google Scholar PubMed  23 Stöllberger C, Finsterer J. Fatal consequences of climbing a ladder under apixaban and drunken. Neurol Neurochir Pol  2016; 50: 200– 202. Google Scholar PubMed  24 Sabouret P, Bricard M, Hermann M-A, Cotté F-E, Deret-Bixio L, Rushton-Smith S. Discrepancy between guidelines for stroke prevention in atrial fibrillation and practice patterns in primary care. The nationwide French AFIGP survey. Arch Cardiovasc Dis  2015; 108: 544– 553. Google Scholar CrossRef Search ADS PubMed  25 Alamneh EA, Chalmers L, Bereznicki LR. Suboptimal use of oral anticoagulants in atrial fibrillation: has the introduction of direct oral anticoagulants improved prescribing practices? Am J Cardiovasc Drugs  2016; 16: 183– 200. Google Scholar CrossRef Search ADS PubMed  26 Albrecht JS, Liu X, Baumgarten M, Langenberg P, Rattinger GB, Smith GS, Gambert SR, Gottlieb SS, Zuckerman IH. Benefits and risks of anticoagulation resumption following traumatic brain injury. JAMA Intern Med  2014; 174: 1244– 1251. Google Scholar CrossRef Search ADS PubMed  27 Kirchhof P, Benussi S, Kotecha D, Ahlsson A, Atar D, Casadei B, Castella M, Diener H-C, Heidbuchel H, Hendriks J, Hindricks G, Manolis AS, Oldgren J, Popescu BA, Schotten U, Putte BV, Vardas P, Agewall S, Camm J, Esquivias GB, Budts W, Carerj S, Casselman F, Coca A, Caterina RD, Deftereos S, Dobrev D, Ferro JM, Filippatos G, Fitzsimons D, Gorenek B, Guenoun M, Hohnloser SH, Kolh P, Lip GYH, Manolis A, McMurray J, Ponikowski P, Rosenhek R, Ruschitzka F, Savelieva I, Sharma S, Suwalski P, Tamargo JL, Taylor CJ, Gelder ICV, Voors AA, Windecker S, Zamorano JL, Zeppenfeld K. 2016 ESC guidelines for the management of atrial fibrillation developed in collaboration with EACTS. Eur Heart J  2016; 37: 2893– 2962. Google Scholar CrossRef Search ADS PubMed  28 Pollack CVJ, Reilly PA, Eikelboom J, Glund S, Verhamme P, Bernstein RA, Dubiel R, Huisman MV, Hylek EM, Kamphuisen PW, Kreuzer J, Levy JH, Sellke FW, Stangier J, Steiner T, Wang B, Kam C-W, Weitz JI. Idarucizumab for dabigatran reversal. N Engl J Med  2015; 373: 511– 520. Google Scholar CrossRef Search ADS PubMed  29 Jørgensen TSH, Hansen AH, Sahlberg M, Gislason GH, Torp-Pedersen C, Andersson C, Holm E. Falls and comorbidity: the pathway to fractures. Scand J Public Health  2014; 42: 287– 294. Google Scholar CrossRef Search ADS PubMed  30 Torstensson M, Hansen AH, Leth-Møller K, Jørgensen TSH, Sahlberg M, Andersson C, Kristensen KE, Ryg J, Weeke P, Torp-Pedersen C, Gislason G, Holm E. Danish register-based study on the association between specific cardiovascular drugs and fragility fractures. BMJ Open  2015; 5: e009522. Google Scholar CrossRef Search ADS PubMed  31 de Vries OJ, Peeters G, Elders P, Sonnenberg C, Muller M, Deeg DJH, Lips P. The elimination half-life of benzodiazepines and fall risk: two prospective observational studies. Age Ageing  2013; 42: 764– 770. Google Scholar CrossRef Search ADS PubMed  Published on behalf of the European Society of Cardiology. All rights reserved. © The Author 2017. For permissions, please email: journals.permissions@oup.com. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices)

Journal

European Heart JournalOxford University Press

Published: Nov 18, 2017

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off