Use of Antithrombotic Therapy and Long-Term Clinical Outcome Among Patients Surviving Intracerebral Hemorrhage
Background and Purpose—The effectiveness and safety of antithrombotic therapy (AT) among patients with a history of intracerebral hemorrhage remain uncertain. We therefore determined the prevalence of indication for AT among patients hospitalized with first-time intracerebral hemorrhage and examined the impact of subsequent AT use on the long-term clinical outcome.
Methods—We performed a population-based cohort study using nationwide Danish medical registries. Patients with risk of thromboembolism surviving the first 30 days after hospitalization because of intracerebral hemorrhage were identified and followed up. We estimated the hazard ratio of all-cause death, thromboembolic events, or major bleeding according to use of AT.
Results—We identified 6369 patients between 2005 and 2013. Among these patients, 2978 (47%) had indication for AT, and during the follow-up, (median: 2.3 year) 1281 (43%) died, 497 (17%) had a thromboembolic event, and 536 (18%) had major bleeding. Postdischarge use of oral anticoagulation therapy among patients with indication for oral anticoagulation therapy was associated with a significant lower risk of death (adjusted hazard ratio, 0.59; 95% confidence interval, 0.43–0.82) and thromboembolic events (adjusted hazard ratio 0.58; 95% confidence interval, 0.35–0.97) and no increased risk of major bleeding (adjusted hazard ratio 0.65; 95% confidence interval, 0.41–1.02). In contrast, use of platelet inhibitors among patients with indication for platelet inhibitors was not related to statistically significantly improved clinical outcome.
Conclusions—Approximately 1 of 2 patients surviving intracerebral hemorrhage had a high risk of thromboembolism. Postdischarge use of oral anticoagulation therapy was associated with a lower risk of all-cause mortality and thromboembolic events and no increased risk of major bleeding.
Clinicians are everyday challenged with the decision of resuming or initiating antithrombotic therapy (AT) among patients with previous hemorrhage, because of a high prevalence of ischemic risk factors or previous thromboembolic events among many of these patients. Patients with intracerebral hemorrhage (ICH) represent a special challenge because of the concerns about the risk of recurrent and potentially catastrophic ICH.1 Within recent years, there has been a remarkable development in the area of AT with introduction of new antithrombotic drugs and increased clinical attention to ensure that guideline recommendations are implemented.
Data on use of AT among patients surviving ICH have until recently been sparse, and there have therefore only been little evidence available for guiding clinical practice. As demonstrated in recent studies, it is therefore not surprising that there has been large variation in clinical practice, and many ICH survivors with a high risk of thromboembolism are not prescribed AT.2,3 The association between use of AT and clinical outcome among patients with previous intracranial hemorrhage/ICH have been examined in a few studies within recent years.4–7 Overall, the findings seem to encourage the use of AT among patients surviving ICH because AT was associated with a lower risk of thromboembolic events without an associated increased risk of major bleeding. However, questions remain, as the available data are primarily restricted to patients with atrial fibrillation resuming treatment with vitamin K antagonists. Additional population-based studies with long-term follow-up on well-characterized patient populations, including patients with other indications for AT than atrial fibrillation, are therefore needed to obtain a stronger evidence base for decision making about use of AT (including vitamin K antagonists, non–vitamin K antagonist oral anticoagulants, and platelet inhibitors) among ICH survivors.
We therefore conducted a nationwide follow-up study to examine the effectiveness and safety of AT among these patients.
The study was based on national Danish registries covering the entire population (≈5.6 million). The National Health Service provides tax-financed healthcare to all residents, and unambiguous individual-level linkage between registries is enabled by a unique 10-digit civil registration number, which is assigned to every citizen and used in all registries. According to Statistics Denmark, 89% of the population had Danish ancestry and 11% were immigrants or descendants in 2013.
The study was approved by the Danish Data Protection Agency (ID 1-16-02-371-13). According to Danish law, ethical approval is not required for registry-based studies.
The study population consisted of all Danes (≥18 years) admitted with first-time, acute, spontaneous (nontraumatic) ICH and surviving the first 30 days from January 1, 2005, to December 31, 2013.
The patients were identified from the Danish Stroke Registry (DSR). The DSR is a stroke-specific clinical registry. Reporting is mandatory for all hospital departments treating patients with acute stroke.8 The sensitivity and positive predictive value of the registration of patients in the DSR has been estimated to be >90%.9 The registry contains information on the subtype of stroke (ischemic stroke and ICH) and stroke severity but not more detailed information, that is, localization or volume. Information on vital status was obtained from The Danish Civil Registration System, which keeps daily updated electronic records on change of address, date of emigration, and changes in vital status.10 The criterion of at least 30-day survival was used to exclude patients who were moribund and therefore not candidates for AT.11,12
The patients with indication for AT were identified based on information from the DSR and the Danish National Registry of Patients (DNRP).13 The DNRP contains data, including diagnoses and surgical procedures, on all discharges from all nonpsychiatric hospitals in Denmark since 1977. Indication for AT was defined as a hospitalization history with myocardial infarction, ischemic stroke, peripheral arterial disease, venous thromboembolism, atrial fibrillation, or artificial heart valve.
The following are the major outcomes of this study:
All-cause mortality: this was assessed using information from the Danish Civil Registration System.
Thromboembolic events: this included hospitalization with myocardial infarction, ischemic stroke, peripheral arterial thrombosis, venous thromboembolism, and pulmonary embolism. Information was obtained from the DNRP and the DSR.
Major bleeding: this included hospitalization with recurrent ICH and other intracranial bleeding and gastrointestinal, urinary, and airway/lung bleeds. Information was obtained from the DSR and the DNRP.
Recurrent ICH: this was analyzed as part of the combined end point of major bleeding, but also as an individual end point, because of the serious nature of this type of bleeding.
Information on use of AT was obtained from the Danish National Database of Reimbursed Prescriptions. This registry contains information from 2004 and onward on all reimbursed prescription drugs dispensed at Danish pharmacies. The included antithrombotic medications are available by prescription only, except for low-dose aspirin; however, the potential for identification of individual-level use of low-dose aspirin from prescription data is high in Denmark because the prescribed proportion of individual-level low-dose aspirin sales is >90%.14
Preadmission use of AT was defined as at least one filled prescription within 90 days before the date of admission with ICH. Post-ICH use was determined using information from all filled prescriptions after the day of admission for each individual. The length of the individual prescriptions was estimated based on information on the package size and the daily defined dose as defined by the World Health Organization.15 Using this information, it was possible to estimate the length of time the individual patient was treated with AT during the follow-up period and to determine whether outcome events occurred during or off treatment. Data on in-hospital use of AT as well as information on the exact timing of potential shorter pauses in treatment (eg, because of dental procedure) was not available and therefore not accounted for.
Information on patient characteristics was obtained from the DSR, the DNRP, and the Danish National Database of Reimbursed Prescriptions and included age, sex, admission year, comorbidity (Charlson comorbidity score),16 stroke severity (Scandinavian Stroke Scale score),17 calendar year, CHA2DS2-VASc (Congestive heart failure, Hypertension, Age≥75 years, Diabetes mellitus, previous Stroke/transient ischemic attack, Vascular disease, Age 65–74 years, Sex category; age≥75 years and previous stroke carry doubled risk weight) score, HAS-BLED (Hypertension, Abnormal renal/liver function, Stroke, Bleeding history, Labile international normalized ratio, Elderly >65 years, Drug consumption/alcohol abuse) score, contraindications to oral anticoagulation therapy (OAT), body mass index, smoking habits, and alcohol intake. The CHA2DS2-VASc18 and HAS-BLED19 scores were computed for patients with atrial fibrillation. In the HAS-BLED score, the labile international normalized ratio component of the score was not included, because this information was not available. Both algorithms have been shown to predict thromboembolism and bleeding accurately.18,19 Furthermore, we registered whether the patients had any contraindications to AT that are registered specifically by the physician treating the patients during the initial ICH admission and reflects, for example, one or more of the following contraindications: recent surgery or bleeding, aspirin treatment, uncontrolled hypertension or hemorrhagic diathesis, dementia/alcoholism, lack of patient acceptance, and pregnant in first trimester and other.
We first estimated the prevalence of indication for AT among patients hospitalized with ICH and surviving the first 30 days after admission. Second, we computed cumulative incidence rates of the clinical outcomes. Patients were followed up from 30 days after admission until the time of the first outcome event (death, thromboembolic event, or major bleeding), date of emigration, and death (in the analyses on thromboembolic events and major bleeding) or the end of the study period, whichever came first.
Third, we examined possible predictors for post-ICH use of AT among all patients with indication for AT. These analyses were performed using separate multivariable logistic regression analyses for treatment with OAT and platelet inhibitors.
We examined the association between postdischarge use of AT and the clinical outcomes using Cox proportional hazards regression, adjusting for the variables presented in Table 1. We separated the patients in 2 groups: group 1 consisted of patients with potential indication for OAT because of atrial fibrillation, artificial heart valve, or venous thromboembolism and group 2 consisted of patients with potential indication for resumption of platelet inhibitors, including previous acute myocardial infarction, ischemic stroke (without atrial fibrillation), and peripheral arterial disease. We used multiple imputations to impute the missing values among the covariates assuming that data were missing at random (Markov chain Monte Carlo method). We created 20 data sets based on the aforementioned covariates. The outcome measures were averaged across the 20 imputations correcting for between- and within-imputation variation. Postdischarge use of AT was included as a time-dependent variable. Using this approach, the patients were only considered exposed during the exact number of days of available treatment provided by the individual filled prescriptions.
Furthermore, we repeated the analysis with the follow-up period beginning at 1 year after admission, to counteract any delayed influence on the outcome from their primary ICH admission, and we also repeated the analysis with restriction to patients who were treated with AT before their ICH event. In addition, we made an analysis examining the association of time between first filled prescription for AT after ICH and the clinical outcome to explore when it is safe to resume AT therapy. Finally, we did a propensity score–matched sensitivity analysis on the association between post-ICH OAT use and the clinical outcomes. Here, we matched patients using OAT within the first 180 days after ICH with patients not using OAT using the nearest neighbor–matching method. We aimed to match each OAT user with a nonuser with similar propensity score using replacement. The propensity score was based on the covariates listed in Table 1. An absolute standardized difference <10% and a variance ratio between 0.8 and 1.25 were considered to support the assumption of balance between the groups (Figures I and II in the online-only Data Supplement). The matching was followed by Cox proportional hazards regression stratified on the matched pairs and with additional adjustment for CHA2DS2-VASc and HAS-BLED scores to account for any residual confounding.
We identified 6369 incident ICH patients surviving at least 30 days, of whom 2978 (47%) had indication for AT at the time of admission. Of these, 2298 (77%) had only a single indication for AT, whereas the rest had ≥2 indications. Patient characteristics are presented in Table 1. Among patients with atrial fibrillation, 94% had a CHA2DS-VASc score ≥2 and 50% had a HAS-BLED score ≥4.
Use of AT
The proportion of patients who filled a prescription for OAT within 180 days after ICH admission was reduced by >55% compared with the proportion of patients using OAT at the time of admission for ICH (Table 2).
Post-ICH use of OAT was primarily initiated within the first 3 to 6 months during which time ≤34% had reimbursed their AT prescription. Within 5 years, ≈48% of all patients had filled at least one prescription for AT. Previous use of OAT, a high Scandinavian Stroke Scale score, and smoking history were significant predictors for resumption of OAT (Table 3). Significant predictors of resumption of platelet inhibitors included pre-ICH use of platelet inhibitors and a high Scandinavian Stroke Scale score.
Among patients having indication for AT and surviving the first 30 days after their initial ICH event, 1281 (43%) died, 497 (17%) developed a thromboembolic event, and 536 (18%) had a major bleeding during a median follow-up period of 2.3 years. Among patients receiving post-ICH AT, 91 (6.1%) experienced a recurrent ICH as their first major bleeding event, whereas 113 patients (7.6%) who did not receive any AT post-ICH also experienced a recurrent ICH. The distribution of post-ICH major bleeding types can be seen in Table I in the online-only Data Supplement.
Postdischarge use of OAT among patients with indication for OAT was associated with a significant lower risk of death (adjusted hazard ratio [HR], 0.59; 95% confidence interval [CI], 0.43–0.82) and thromboembolic events (adjusted HR, 0.58; 95% CI, 0.35–0.97) and a no increased risk of major bleeding (adjusted HR, 0.65; 95% CI, 0.41–1.02), including no increased risk of recurrent ICH (adjusted HR, 0.90; 95% CI, 0.44–1.82; Table 4). In contrast, use of platelet inhibitors among patients with indication for platelet inhibitors was not related to statistically significantly improved clinical outcome, including death (adjusted HR, 0.96; 95% CI, 0.80–1.15), thromboembolic events (adjusted HR, 0.99; 95% CI, 0.75–1.31), and major bleeding (adjusted HR, 0.89; 95% CI, 0.68–1.16). In the land mark analysis in which the follow-up period started 1 year after ICH admission, postdischarge use of OAT was also associated with improved clinical outcome; however, the difference only reached statistical significance for major bleeding (see Table II in the online-only Data Supplement). In contrast, post-ICH OAT use among pre-ICH OAT users was associated with significantly reduced risk of both death (adjusted HR, 0.51; 95% CI, 0.33–0.78) and thromboembolic events (adjusted HR, 0.45; 95% CI, 0.22–0.90) and no increased risk of major bleeding (adjusted HR, 0.71; 95% CI, 0.40–1.26; see Table III in the online-only Data Supplement).
Table IV in the online-only Data Supplement presents the results on the association between time of first filled prescription for AT after ICH and the clinical outcome. Both early (within 30 days from inclusion) and late (>180 days from inclusion) initiation of OAT after ICH was associated with a significantly lower risk of death. Similar patterns were found for thromboembolic events and major bleeding, although the statistical precision was modest in some of the analyses. Initiation of OAT within 31 to 180 days was not associated with a favorable outcome, which may be a chance finding because of the few patients starting OAT within this timeframe.
To further test the robustness of the primary analyses, we also performed a propensity score–matched analysis including 192 patients who started OAT therapy within 180 days after ICH and 192 matching nonusers of OAT. The characteristics of the OAT users and the matched nonusers were well balanced (Figures I and II in the online-only Data Supplement). Use of OAT was associated with a lower risk of death (adjusted HR, 0.48; 95% CI, 0.30–0.75) and thromboembolic events (adjusted HR, 0.48; 95% CI, 0.25–0.96) and no increased risk of major bleeding (adjusted HR, 0.65; 95% CI, 0.35–1.20).
Almost half of the patients surviving the first 30 days after their ICH event had indication for AT in our nationwide study. The outcome for this population was in general poor and included a high absolute risk of all-cause mortality as well as ischemic and major bleeding events. Overall, the use of AT post ICH, in particular OAT, was associated with a lower risk of death and thromboembolic events and no increased risk of major bleeding, including recurrent ICH, although not reaching statistical significance in all analyses. Early start/resumption of AT after ICH seemed to be safe.
The high proportion of patients with indication for AT in our study confirms and extends recent findings.2,3 Pasquini et al3 reported that 44% of ICH patients in their study used AT at the time of hospital admission, whereas Pennlert et al2 in a nationwide Swedish study found that 39% of the patients, who eventually were discharged alive after ICH, were using AT at the time of admission. The high absolute risk of thromboembolic events and death among the ICH survivors with indication for AT in our study is also in accordance with other recent studies.2,4,6,7,20
A high proportion of the patients receiving AT at the time of ICH did not resume AT despite their high risk of thromboembolic events. This is in accordance with the low proportion of OAT use among ICH survivors with atrial fibrillation, which has been reported in other studies.3,6 The proportion receiving OAT post ICH varies somewhat between the studies, which may both reflect differences between the studies with regard to definition of study population and follow-up period and reflect variation in clinical practice.3 Worth noting in our study was the marked relative increase in the use of non–vitamin K antagonist oral anticoagulants when comparing the post- and pre-ICH period. This likely reflects that some of the patients who had previously been treated with vitamin K oral anticoagulants were shifted to non–vitamin K antagonist oral anticoagulants in anticipation of a lower risk of recurrent ICH. However, although the relative increase was large (5-fold), the proportion of ICH survivors who used non–vitamin K antagonist oral anticoagulants remained low.
The association between post-ICH use of OAT and the clinical outcome observed in our study is in accordance and extend findings from other recent studies linking OAT with an improved clinical outcome among patients surviving intracranial hemorrhage and traumatic brain injury.4–7,12 The studies differ in design, type of patients included, sample size, and level of details of the available data on patient characteristics and care (including drug use) and outcomes. The individual studies have specific set of strengths and limitations and are therefore complimentary from a methodological point of view. It is consequently remarkable, how consistent the results are across the studies. Altogether, the studies provide support for the hypothesis that depending on appropriate patient selection, OAT therapy can be used safely in patients with high risk of thromboembolism and also in patients at increased risk of ICH or other intracranial hemorrhage. The studies furthermore indicate that early start of OAT therapy can be used without offsetting patient safety.6,12 Still, it is essential to be aware that all of the data presented to date are observational, and although the studies indicate that a nihilistic therapeutic approach is not warranted, they should not be interpreted as a recommendation for unreflected and aggressive use of OAT in all patients with a high predicted risk of thromboembolism surviving ICH. The decision of whether to use or avoid OAT should in any case depend on a thorough assessment of the individual patient’s thrombohemorrhagic risk and should reflect the patient’s preferences and informed choice.
The strength of our study include the prospective population-based design with complete follow-up, limiting the risk of bias and the large sample size.
Although we included a variety of well-known confounders, we cannot exclude the possibility that our results remain influenced by confounding because of the observational nature of the study. Potentially important predictors of AT use among ICH survivors include the location (lobar versus nonlobar) and the size of the ICH because these factors are associated with the risk of recurrent ICH.21,22 Other factors that may both influence use of AT and the clinical outcome include the amount of microbleeds, the intensity of the OAT (as reflected by international normalized ratio levels), and the functional level of the patients (eg, modified Rankin Scale); however, except for data on stroke severity at the time of admission, data on these factors were not available. Furthermore, the data available did not fully allow that the definitions of bleeding proposed by the Bleeding Academic Research Consortium were used; however, the definition used in our study did closely resemble a type 2 bleeding or higher as defined by Bleeding Academic Research Consortium.23
In conclusion, indication for AT was found among almost half of all surviving ICH patients. Postdischarge initiation or resuming of OAT among patients with indication was associated with lower mortality and lower risk of thromboembolic events and was not associated with an increased risk of major bleeding, including recurrent ICH. Randomized clinical trial data are warranted to confirm these findings.
Sources of Funding
This study was partly supported by a research grant from Bristol-Myers Squibb (BMS) and Pfizer to Aarhus University.
Drs Hansen, Brandes, and Husted are members of an advisory board for BMS/Pfizer. Dr Damgaard has received speaker’s honoraria from Bayer, Boehringer-Ingelheim, BMS, and Pfizer. Dr Johnsen is member of advisory boards for BMS/Pfizer and Bayer and has received speaker’s honoraria from Bayer, Boehringer-Ingelheim, BMS, and Pfizer. The other authors report no conflicts.
BMS and Pfizer had no influence on the study design, data collection, the presentation of data, or the conclusions made.
The online-only Data Supplement is available with this article at http://stroke.ahajournals.org/lookup/suppl/doi:10.1161/STROKEAHA.116.012945/-/DC1.
- Received February 5, 2016.
- Revision received April 27, 2016.
- Accepted May 3, 2016.
- © 2016 American Heart Association, Inc.
- Eckman MH,
- Rosand J,
- Knudsen KA,
- Singer DE,
- Greenberg SM
- Pennlert J,
- Asplund K,
- Carlberg B,
- Wiklund PG,
- Wisten A,
- Åsberg S,
- et al
- Pasquini M,
- Charidimou A,
- van Asch CJ,
- Baharoglu MI,
- Samarasekera N,
- Werring DJ,
- et al
- Nielsen PB,
- Larsen TB,
- Skjøth F,
- Gorst-Rasmussen A,
- Rasmussen LH,
- Lip GY
- Chao TF,
- Liu CJ,
- Liao JN,
- Wang KL,
- Lin YJ,
- Chang SL,
- et al
- Mainz J,
- Krog BR,
- Bjørnshave B,
- Bartels P
- Majeed A,
- Kim YK,
- Roberts RS,
- Holmström M,
- Schulman S
- Schmidt M,
- Schmidt SA,
- Sandegaard JL,
- Ehrenstein V,
- Pedersen L,
- Sørensen HT
- Schmidt M,
- Hallas J,
- Friis S
- 15.↵ATC/DDD Index 2015. WHO Collaborating Centre for Drug Statistics Methodology. http://www.whocc.no/atc_ddd_index/. Accessed July 22, 2015.
- 17.↵Multicenter trial of hemodilution in ischemic stroke–background and study protocol. Scandinavian Stroke Study Group. Stroke. 1985;16:885–890.
- Olesen JB,
- Lip GY,
- Hansen ML,
- Hansen PR,
- Tolstrup JS,
- Lindhardsen J,
- et al
- Olesen JB,
- Lip GY,
- Hansen PR,
- Lindhardsen J,
- Ahlehoff O,
- Andersson C,
- et al
- Flynn RW,
- MacDonald TM,
- Murray GD,
- MacWalter RS,
- Doney AS
- Passero S,
- Burgalassi L,
- D’Andrea P,
- Battistini N
- Broderick JP,
- Brott TG,
- Duldner JE,
- Tomsick T,
- Huster G
- Mehran R,
- Rao SV,
- Bhatt DL,
- Gibson CM,
- Caixeta A,
- Eikelboom J,
- et al