Variation in Restarting Antithrombotic Drugs at Hospital Discharge After Intracerebral Hemorrhage
Background and Purpose—Whether intracerebral hemorrhage (ICH) survivors should restart antithrombotic drugs is unknown. We analyzed the frequency of restarting antithrombotic drugs in ICH survivors who had taken prophylactic antithrombotic drugs in atrial fibrillation or after thromboembolic disease in 5 cohorts and explored factors associated with doing so.
Methods—We compared the characteristics and proportions of patients taking antithrombotic drugs at ICH onset and discharge in 4 hospital-based cohorts (Lille, France, n=542; Utrecht, The Netherlands, n=389; multicenter Clinical Relevance of Microbleeds in Stroke-2 (CROMIS-2) ICH, United Kingdom, n=667; and Amsterdam, The Netherlands, n=403) and 1 community-based study (Lothian, Scotland, n=137), using bivariate analyses. We sought characteristics associated with restarting using bivariate and multivariable logistic regression analyses.
Results—A total of 942 (44%) patients with ICH took antithrombotic drugs at hospital admission (no difference between cohorts). Antithrombotic drugs were restarted in 96 (20%) of the 469 survivors who had taken antithrombotic drugs for secondary prevention or atrial fibrillation, but this proportion differed when stratified by the cohort of origin (Lille, 18%; Utrecht, 45%; Lothian, 15%; CROMIS-2 ICH, 11%; Amsterdam, 20%; P<0.001) and by type of antithrombotic drug pre-ICH (14% in patients with previous antiplatelet drugs versus 26% in patients with previous vitamin K antagonists and 41% in patients with both drugs; P<0.001). We did not find other consistent, independent associations with restarting antithrombotic drugs.
Conclusions—The variation in clinical practice and lack of consistent associations with restarting antithrombotic drugs after ICH reflect current knowledge and support the need for randomized controlled trials to resolve this dilemma.
Nowadays, 40% of patients in high-income countries are taking antithrombotic drugs at the time they suffer a spontaneous intracerebral hemorrhage (ICH).1,2 This proportion has increased and is expected to increase further.3 However, it is unclear whether ICH survivors who took antithrombotic drugs should restart or avoid them.
For ICH survivors resuming antiplatelet drugs, there was no increase in the risk of recurrent ICH in observational cohort studies of survivors (496 in Germany4; 440 in Hong Kong5; 417 in Scotland6; 207 in the United States)7 although a multivariable analysis found an increased risk of recurrent ICH associated with aspirin in a subgroup of 104 lobar ICH survivors.8 In a Hong Kong cohort, aspirin was associated with a reduction in all serious vascular events in patients with standard indications for antiplatelet therapy.5 Because of these findings, a multicenter randomized controlled trial is investigating whether ICH survivors, who took antithrombotic drugs, should restart or avoid antiplatelet drugs (the Restart of Stop Antithrombotics Randomized Trial [RESTART]; http://www.RESTARTtrial.org; ISRCTN71907627).
For ICH survivors resuming anticoagulant drugs, some observational studies found a higher risk of bleeding but no effect on the risk of all serious vascular events,9–12 leaving uncertainty and controversy about what to do.13 A phase II randomized open-label clinical trial is investigating whether patients with atrial fibrillation who survived a recent anticoagulant-associated ICH should be treated with either apixaban or any/no antiplatelet drugs (Apixaban or Antiplatelet Drugs After Anticoagulation-Associated Cerebral Haemorrhage in Patients With Atrial Fibrillation [APACHE-AF]; NTR4526).
In the absence of randomized trials, European guidelines state that the use of antithrombotic drugs after ICH should be individualized, taking into account the risk of thromboembolic events and recurrent ICH.14 North-American guidelines focus on anticoagulant-associated ICH in patients with atrial fibrillation and state that anticoagulation after nonlobar ICH and antiplatelet drugs after all ICH might be considered.15
To establish whether there is variation in current practice, we investigated the frequency of restarting antithrombotic drugs at hospital discharge, and associations with doing so, in ICH survivors who had previously taken antithrombotic drugs for atrial fibrillation or secondary prevention of vascular events.
Patients and Methods
We gathered data from 4 hospital-based cohorts of adults with spontaneous first-ever or recurrent ICH (Lille, France, n=542; Utrecht, The Netherlands, n=389; multicenter CROMIS-2 ICH, United Kingdom, n=667; and Amsterdam, The Netherlands, n=403) and 1 community-based study (Lothian, Scotland, n=137) that collected information on antithrombotic treatment at the time of ICH and at hospital discharge. We excluded patients with purely extra-axial intracranial hemorrhage, patients with ICH attributable to a secondary cause of ICH (intracranial vascular malformation, head trauma, tumor or hemorrhagic transformation within a cerebral infarct, or congenital coagulopathies), and patients treated with heparin.
Lille Cohort, France
The Prognosis of Intra-Cerebral Hemorrhage (PITCH) cohort study prospectively recruited consecutive adults admitted with ICH to the emergency department of Lille University Hospital from November 2004 to April 2009 in Lille, France.16
Utrecht Cohort, The Netherlands
The Utrecht cohort was extracted from a prospective database, including consecutive adults admitted with ICH to the University Medical Center Utrecht, the Netherlands, from January 2007 to July 2012.
Lothian Cohort, Scotland
The Lothian Audit of the Treatment of Cerebral Hemorrhage (LATCH) prospectively ascertained all residents in the Lothian Healthboard region of Scotland (mid-2010 population aged 16 years and above was 695 335) aged ≥16 years when diagnosed with ICH confirmed by brain imaging or pathology between June 2010 and May 2011.
CROMIS-2 ICH Cohort, United Kingdom
The Clinical Relevance of Microbleeds in Stroke (CROMIS-2) ICH cohort is an ongoing multicenter observational study, prospectively recruiting adults with imaging-confirmed symptomatic spontaneous ICH. For this analysis, we included patients with complete clinical data, recruited between August 2010 and August 2013.
Amsterdam Cohort, The Netherlands
The Amsterdam ICH cohort was extracted from a prospective database (the Amsterdam Stroke Database), including consecutive adults admitted with ICH to the Academic Medical Center from January 2006 to May 2013.
We collected demographic characteristics, history of arterial hypertension, atrial fibrillation, transient ischemic attack, ischemic stroke, ICH, ischemic heart disease (myocardial infarction, history of cardiac stent, unstable angina, or acute coronary syndrome), peripheral vascular disease, symptomatic deep vein thrombosis, and pulmonary embolism. Definitions have been given elsewhere16 and were consistent across the 5 cohorts. For each cohort, ≥1 study investigator with expertise in stroke reviewed diagnostic brain images for ICH location. ICH was classified as (1) lobar when exclusively involving the cerebral hemispheres superficial to the deep gray matter structures; (2) nonlobar when exclusively involving lenticular or caudate nuclei, thalamus, internal or external capsule, brain stem, or cerebellum; (3) unclassifiable when involving both lobar and nonlobar cerebral regions; and (4) multiple if there was >1 ICH at presentation. We classified antithrombotic drug use according to type (antiplatelet, anticoagulant, or both) and assessed the use of these drugs at presentation and hospital discharge. Case fatality rate was recorded at 30 days.
We investigated characteristics associated with antithrombotic drug use at presentation and characteristics of survivors associated with restarting antithrombotic drugs using bivariate analyses (χ2, Fisher exact test if cell counts <5, Kruskal–Wallis test). We performed multivariable logistic regression analyses with fixed entry of variables (selected when P≤0.2 in bivariate analyses). We included 1 variable for approximately every 10 patients restarting antithrombotic drugs. We performed statistical analyses with the SPSS 22.0 package for windows.
The internal review board of the Lille University Hospital and the medical ethics committee of the University Medical Center of Utrecht and of the Academic Medical Center of Amsterdam decided that no consent was needed for the respective cohorts because of their observational design. LATCH was approved by the National Health Service Lothian Caldicott Guardian (patients in National Health Service Lothian are informed about the use of their data for audit, and information leaflets about LATCH were distributed to inform patients and carers about their right to opt out). The CROMIS-2 ICH study was approved by the National Research Ethics Service Committee, London (Queen Square), and all patients gave written informed consent.
In all cohorts, there were 2138 patients (median age, 72 years; interquartile range, 59–80; 53% men) with spontaneous ICH, of whom 942 (44%) were taking antithrombotic drugs at presentation: 559 (26%) took antiplatelet drugs, 333 (16%) took vitamin K antagonist drugs, and 50 (2%) took both. No patient took novel oral anticoagulants at the time of ICH. Predominant indications for antithrombotic drugs were ischemic stroke or transient ischemic attack (n=312), atrial fibrillation (n=298), and ischemic heart disease (n=189); 173 patients had ≥2 reasons for being on antithrombotic drugs, and 177 patients had taken antithrombotic drugs without atrial fibrillation or previous thromboembolic disease.
The 5 cohorts were similar in most baseline characteristics although patients in the Lothian cohort were older, and patients in CROMIS-2 ICH had lower case fatality rate at 30 days (Table I in the online-only Data Supplement). After excluding patients without atrial fibrillation or thromboembolic disease before ICH, the proportions of patients taking antithrombotic drugs at presentation were similar in the 5 cohorts: Lille 38%, Utrecht 44%, Lothian 40%, CROMIS-2 ICH 40%, and Amsterdam 33% (P=0.052). Multivariable analyses showed that patients taking antithrombotic drugs at presentation were older (odds ratio per 1-year increase, 1.04; 95% confidence interval [CI], 1.03–1.05), more likely to be men (odds ratio, 1.42; 95% CI, 1.17–1.72), and hypertensive (odds ratio, 2.80; 95% CI, 2.29–3.41) but less likely to have had previous ICH (odds ratio, 0.56; 95% CI, 0.36–0.89) than patients not taking antithrombotic drugs (Table II in the online-only Data Supplement).We could not detect differences between patients taking different types of antithrombotic drugs (Table 1).
Patients taking antithrombotic drugs had higher case fatality rate at 30 days, overall (36%; 95% CI, 33%–39%) and when stratifying by the type of antithrombotic drug at presentation (35%; 95% CI, 31%–39% in patients taking antiplatelet drugs; 38%; 95% CI, 33%–43% in patients taking vitamin K antagonists; and 40%, 95% CI, 26%–54% in patients taking both drugs), when compared with patients not taking antithrombotic drugs (23%; 95% CI, 21%–26%).
Frequency of Restarting Antithrombotic Drugs at Hospital Discharge
Of the 942 patients who had taken antithrombotic drugs at presentation, 595 (63%) survived to hospital discharge. After excluding 126 survivors who had taken antiplatelet drugs for primary prevention in sinus rhythm, we investigated the frequency of restarting antithrombotic drugs at hospital discharge in the remaining 469 survivors, whose characteristics differed across cohorts (Table 2). Antithrombotic drugs were restarted in 96 patients (20%; 95% CI, 17%–24%) at hospital discharge (Figure). In the 236 survivors who took antiplatelet drugs at presentation, 32 (14%) restarted antiplatelet agents and 204 patients discontinued antithrombotic drugs. In the 206 survivors who took vitamin K antagonists at presentation, 29 (14%) restarted vitamin K antagonists, 23 (11%) restarted antiplatelet drugs, and 1 (0.5%) restarted both. Among the 27 survivors who took antiplatelet drugs and vitamin K antagonists at presentation, 8 (30%) restarted antiplatelet drugs, 2 (7%) switched to vitamin K antagonists, and 1 (4%) restarted both. The proportions in whom antithrombotic drugs were restarted differed according to cohort of origin (P<0.001; Table 2) and to baseline antithrombotic drug regimen: 14% (95% CI, 9%–18%) in patients who took antiplatelet drugs, 26% (95% CI, 20–32%) in patients who took vitamin K antagonists, and 41% (95% CI, 21–61%) in patients who took both drugs (P<0.001; Figure).
Associations With Restarting Antithrombotic Drugs at Hospital Discharge
Bivariate analyses comparing survivors in whom antithrombotic drugs were restarted or not at hospital discharge found that restarting was more likely in younger patients, in patients with a history of hypertension and atrial fibrillation, and less likely in patients with a history of ischemic stroke or transient ischemic attack (Table III in the online-only Data Supplement). Initial clinical severity, measured by the Glasgow Coma Scale, was not associated with the decision to restart antithrombotic drugs (P=0.47; Table III in the online-only Data Supplement); in bivariate analyses of the 3 cohorts with available data, median ICH volume did not differ between survivors who restarted or stopped antithrombotic drugs (10.0 versus 12.3 mL; P=0.41). When stratifying by type of antithrombotic drug at presentation, the bivariate comparison did not identify significant associations with the decision to restart antithrombotic drugs for patients who had taken antiplatelet drugs before ICH, but found that patients who had taken anticoagulant drugs before ICH and who restarted antithrombotic drugs were more likely to be younger and to have nonlobar ICH (Tables IV and V in the online-only Data Supplement). In multivariable analyses, the only characteristic that was consistently and independently associated with restarting antithrombotic drugs, whether grouped together or stratified by pre-ICH antithrombotic regimen, was the cohort of origin (Table 3). Younger age was associated with the decision to restart overall, but this effect could not be detected for specific pre-ICH antithrombotic regimens.
One fifth of patients who were taking antithrombotic drugs at the time of ICH for atrial fibrillation or secondary prevention of vascular events restarted an antithrombotic drug at hospital discharge. The proportion restarting these drugs was highest in patients who were taking vitamin K antagonists before ICH. The only consistent association with the decision to restart antithrombotic drugs (overall and stratified by type) was the cohort of origin, suggesting significant variation in practice.
Overall, 44% of patients experienced ICH while on antithrombotic drugs, highlighting how common the clinical dilemma of whether to restart or stop such drugs in ICH survivors is in high-income countries. The proportion restarting antithrombotic drugs seems low considering their proven benefit for prevention of thromboembolic disease and the shortage of evidence that they increase the risk of recurrent ICH.4–12,17 Apart from the influence of the cohort of origin, restarting antithrombotic drugs was more likely in younger patients. Increasing age is a strong risk factor for both ischemic and hemorrhagic events, so perhaps physicians are more concerned about the hazards of restarting antithrombotic drugs in older age groups, even though older patients are at higher risk of thromboembolic events.
Strengths of our study included its large sample size, sufficient to power multivariable analyses adequately, and its inclusion of 5 prospective cohorts representing different European settings. Consistent with recommendations on observational studies,18 the characteristics of some of the cohorts make them likely to be representative, including the community-based design of the Lothian cohort and the similarity of the baseline characteristics of the Lille cohort to those of a population-based registry.16 In the CROMIS-2 ICH cohort, the case fatality rate was low because patients were recruited at any time ≤30 days after ICH onset, but this will not have influenced our analyses of restarting antithrombotic drugs at hospital discharge. Among the survivors, there was some variation between cohorts (Table 2), but we could not relate this to our main findings. We do not have data on the precise date of restarting antithrombotic drugs, but only on antithrombotic drug status at hospital discharge, so we cannot make inferences about the resumption of these drugs in the community. The influence of brain microbleeds on the decision to restart could not be evaluated because data were not available for all 5 cohorts. We could not examine the influence of congestive heart failure, hypertension, age, diabetes mellitus, prior stroke or transient ischemic attack or thromboembolism (CHADS2)19 and hypertension, abnormal renal or liver function, stroke history, bleeding predisposition, labile international normalized ratio, elderly, drugs/alcohol usage (HAS-BLED)20 scores on the decision to restart anticoagulant drugs because not all of the variables were available. Of the variables shared by the 2 scores that were available (hypertension, previous ischemic stroke, and age), only increasing age was associated with the decision to stop antithrombotic drugs, implying that there is greater concern about recurrent ICH than about future ischemic events. Finally, this study was observational, so conclusions cannot be drawn on the effects of treatment.
Previous studies on characteristics associated with the use of antithrombotic drugs after ICH found conflicting results: one study failed to identify characteristics associated with the use of antithrombotic drug after ICH,6 whereas another showed that the use of antiplatelet agents was more frequent in patients with a history of ischemic heart disease and diabetes mellitus.7 Neither of these studies was dedicated to the decision to restart in the group of patients with indications for antithrombotic drugs.
The variation in clinical practice in these cohorts suggests clinical equipoise about whether to restart antithrombotic drugs after ICH. Additional observational studies of long-term outcome are justified, as are randomized controlled trials for ICH survivors who have comorbidities that justify the prescription of antiplatelet (eg, RESTART, http://www.RESTARTtrial.org, ISRCTN71907627) or (novel) oral anticoagulant drugs,21 (eg, APACHE-AF, NTR4526).
The CROMIS-2 ICH study is supported by investigators at multiple centers throughout the United Kingdom and is supported by the National Institute for Health Research Clinical Research Network. The Lothian Audit of the Treatment of Cerebral Hemorrhage was supported by coinvestigators and a community-based network of collaborators, to whom we are grateful.
Sources of Funding
Dr Charidimou is funded by the Greek State Scholarship Foundation, the Stroke Association and the British Heart Foundation. He undertook this work at University College London/University College London Hospitals, who received funding from the Department of Health’s National Institute for Health Research Biomedical Research Centers funding scheme. Dr Samarasekera is funded by an Medical Research Council/Stroke Association Clinical Research Training Fellowship. Dr Werring receives research support for the CROMIS-2 study from the Stroke Association and British Heart Foundation and undertook this work at University College London/University College London Hospitals, who received a proportion of funding from the Department of Health’s National Institute for Health Research Biomedical Research Centers funding scheme. Dr Klijn is supported by a Clinical Established Investigator grant from the Netherlands Heart Foundation (2012 T077) and an Aspasia grant from the Dutch Research Council (ZonMw; 015.008.048). Dr Salman is funded by a Medical Research Council senior clinical fellowship.
Dr Werring has served on advisory boards for Bayer and Daiichi Sankyo United Kingdom. The other authors report no conflicts.
The online-only Data Supplement is available with this article at http://stroke.ahajournals.org/lookup/suppl/doi:10.1161/STROKEAHA.114.006202/-/DC1.
- Received May 20, 2014.
- Revision received June 24, 2014.
- Accepted July 1, 2014.
- © 2014 American Heart Association, Inc.
- Béjot Y,
- Cordonnier C,
- Durier J,
- Aboa-Eboulé C,
- Rouaud O,
- Giroud M
- Smolina K,
- Wright FL,
- Rayner M,
- Goldacre MJ
- Flynn RW,
- MacDonald TM,
- Murray GD,
- MacWalter RS,
- Doney AS
- Yung D,
- Kapral MK,
- Asllani E,
- Fang J,
- Lee DS
- Molina CA,
- Selim MH
- Steiner T,
- Kaste M,
- Forsting M,
- Mendelow D,
- Kwiecinski H,
- Szikora I,
- et al
- Morgenstern LB,
- Hemphill JC III.,
- Anderson C,
- Becker K,
- Broderick JP,
- Connolly ES Jr.,
- et al
- Flaherty ML,
- Adeoye O,
- Sekar P,
- Haverbusch M,
- Moomaw CJ,
- Tao H,
- et al