Article Text
PDF
Abstract
Background and purpose Little is known about the safety of long-term anticoagulation in patients with unruptured saccular intracranial aneurysms, although case reports of anticoagulation in patients with fusiform and saccular aneurysms have been published. The goal of this study was to estimate the rate of subarachnoid hemorrhage (SAH) due to aneurysm rupture in patients with intracranial saccular aneurysms receiving therapeutic doses of anticoagulation.
Methods This was a retrospective case series of patients with intradural aneurysms receiving therapeutic doses of anticoagulation (defined as a dose that achieves a target INR >2.0 or the equivalent dose of another anticoagulant). The primary outcome was SAH due to aneurysm rupture while receiving anticoagulation treatment. The rate of SAH due to aneurysm rupture was calculated among patients with aneurysms documented on imaging.
Results 42 patients with 48 intradural aneurysms on anticoagulation were identified. Their mean age was 67 years (range 40–83) and 28 (67%) were women. The mean aneurysm size was 5.1±3.6 mm and 40 (83%) aneurysms were in the anterior circulation. The mean±SD follow-up was 495±749 days per patient. Twelve patients had their aneurysms coiled, clipped or surgically wrapped before the period of anticoagulation. Three patients had clipping or coiling of their aneurysms during the period of anticoagulation. Six patients had a history of aneurysm rupture prior to anticoagulation, of which two were treated by coiling and four by clipping. Among the 48 aneurysms, 32 were never coiled or clipped prior to anticoagulation. No anticoagulated patients with known aneurysms developed SAH during 57 patient-years of follow-up. The patients with untreated aneurysms had 31 patient-years of follow-up and those who underwent clipping, coiling or surgical wrapping of their aneurysms had 26 patient-years of follow-up while receiving anticoagulation (for a combined total of 57 years of follow up).
Conclusions The risk of aneurysm rupture is not increased in patients receiving systemic anticoagulation. However, these results should be interpreted with caution given the small sample size in this study and the need for a prospective study to confirm these findings.
- Aneurysms
- intracranial aneurysm
- antithrombotics
- subarachnoid hemorrhage
- intracranial hemorrhage
- warfarin
- heparin
- aneurysm
- angiography
- angioplasty
- hemorrhage
- pharmacology
Statistics from Altmetric.com
- Aneurysms
- intracranial aneurysm
- antithrombotics
- subarachnoid hemorrhage
- intracranial hemorrhage
- warfarin
- heparin
- aneurysm
- angiography
- angioplasty
- hemorrhage
- pharmacology
Background
Intracranial aneurysms are common and are increasingly being detected by neuroimaging.1 The prevalence of aneurysms in the adult population varies between 0.8% and 8.9% in autopsy studies.2 3 The International Study of Unruptured Intracranial Aneurysms (ISUIA) found an overall rate of rupture for aneurysms of all sizes in the anterior and posterior circulations in patients with and without a history of rupture of another aneurysm to be 0.5% per year.4 Mortality associated with intracranial aneurysm rupture is estimated to be 24–67%.4–7 Among those who survive, morbidity is 40–44%.8 9 Approximately 60–65% of patients in one recent study achieved a modified Rankin scale of ≥2 12 months after aneurysmal subarachnoid hemorrhage (SAH).6
Although a subset of patients with intracranial aneurysms require anticoagulation for medical indications, the safety of anticoagulation in these patients has not been established. Anticoagulation treatment is associated with an increased risk of intracerebral hemorrhage10 11 and increased mortality.12–14 An elevated international normalized ratio (INR) at admission is also associated with a worse clinical outcome.15
Although fatal rupture of cerebral aneurysms has been reported following administration of tissue plasminogen activator and urokinase,16 17 fibrinolytics have also been given to patients with aneurysms without complication.18 19 A multicenter study of 22 patients with unruptured cerebral aneurysms detected by CT angiography (CTA) or MRI prior to the administration of intravenous tissue plasminogen activator did not reveal an increased risk of hemorrhage with thrombolytic therapy.20 Similarly, the use of aspirin was not found to be a risk factor for aneurysm rupture in a retrospective review of prospectively collected data from the ISUIA but, paradoxically, may confer a protective benefit.21
Despite several reports of aneurysm rupture following fibrinolysis, the likelihood of aneurysm rupture due to anticoagulation is unknown. The objective of this study was to estimate the risk of aneurysmal SAH in patients with intracranial aneurysms who are receiving therapeutic anticoagulation.
Methods
This was a retrospective single-center study which was approved by the institutional review board at our hospital.
Patient selection
Patients were identified in the hospital database via a search for the ICD 9 code for cerebral aneurysm (437.3) from 1 January 2000 to 5 January 2010. A keyword search for the word ‘aneurysm’ appearing in all radiology dictations was also used to identify additional patients that the ICD 9 code search may have missed.
Aneurysm identification and sizing
Medical records were reviewed to verify that the aneurysm was intracranial. One of the authors (NT) reviewed the available neuroimaging studies to verify the size of the aneurysms. The largest dimension was denoted as the size of the aneurysm, with aneurysm size tabulated based on catheter angiography, CTA or MRA.
Identification of patients receiving anticoagulation
All patients with aneurysms who were seen in the anticoagulation clinic at our center were identified. In addition, a computerized keyword search for ‘warfarin, coumadin, heparin, dalteparin, enoxaparin and anticoagulation’ was performed with the hospital's computerized medical records to identify patients with aneurysms who were taking anticoagulation but who were not seen in our anticoagulation clinic. Medication lists from clinic visits, hospital records and updated medication lists from discharge summaries were searched to determine the duration of anticoagulation following aneurysm diagnosis. The total time on anticoagulation was counted following the initial diagnosis of aneurysm by neuroimaging and the patient being concomitantly on anticoagulation. Thus, the follow-up time was restricted only to the time spent receiving anticoagulation with an aneurysm.
Identification of primary outcome
The primary outcome variable studied was SAH due to aneurysm rupture while receiving anticoagulation. Electronic medical records of clinic visits and discharge summaries were reviewed up to March 2011 to detect the presence of the primary outcome.
Inclusion criteria
Patients were included if they had at least one intradural aneurysm and received therapeutic doses of heparin, warfarin, dalteparin or enoxaparin for 24 h or more during the follow-up period. A therapeutic dose of warfarin was defined as a dose to achieve a target INR of >2.0 or a dose of another anticoagulant to achieve an equivalent therapeutic effect.
Exclusion criteria
Patients with mycotic aneurysms, fusiform aneurysms, dissecting aneurysms, pseudoaneurysms or an extradural aneurysm without another intradural aneurysm were excluded. Patients receiving prophylactic non-therapeutic doses of anticoagulation were excluded.
Results
Patient characteristics
A total of 560 patients were identified with the ICD 9 code for cerebral aneurysm of which 42 patients with 48 intradural aneurysms were found to be receiving therapeutic doses of anticoagulation following aneurysm diagnosis. The mean±SD age was 67±12 years (range 40–83) and 28 (67%) of the patients were women. There were 48% white patients, 31% black patients, 7% Hispanic, 2% Asian and in 12% of cases the race was not identified in the medical record. The indications for anticoagulation included atrial fibrillation (n=14), deep vein thrombosis (DVT) (n=12), pulmonary embolism (n=13), stroke (n=2), myocardial infarction (n=1), cervical internal carotid dissection (n=1), mitral regurgitation (n=2) and an artificial heart valve (n=1) (table 1). Four patients had more than one indication for anticoagulation.
Patient characteristics (N=42)
Thirty-nine patients (93%) were receiving warfarin, two patients (5%) were receiving intravenous heparin drips and one patient (2%) was receiving an intravenous heparin infusion prior to transitioning to enoxaparin. Of the patients treated with warfarin, eight were bridged to a therapeutic INR with enoxaparin, one was bridged with dalteparin and one was bridged with argatroban. The remainder of the patients treated with warfarin did not receive bridging anticoagulant therapy during the time of follow-up of their aneurysms. One of the patients treated with warfarin was taken off it for 5 days prior to coiling of her aneurysm and was given enoxaparin until the day of the procedure.
One patient presented with spontaneous SAH prior to receiving anticoagulation. A conventional angiogram at the time of the SAH did not reveal any aneurysm and the patient was diagnosed with ‘benign perimesencephalic hemorrhage’. Four years later the patient developed a subclavian vein DVT. Prior to starting anticoagulation for treatment of the DVT, a CTA showed a 1.5 mm right internal carotid aneurysm of the paraclinoid region. The patient was therefore included in our study and the aneurysm was classified as unruptured since no aneurysm was seen at the time of his SAH 4 years earlier. The patient received anticoagulation for 109 days without aneurysm rupture.
Aneurysm characteristics
Forty-eight intradural aneurysms were identified. The mean±SD aneurysm size was 5.1±3.6 mm (range 1.2–20). Three of the aneurysms were <2 mm. In four patients the size of the aneurysm could not be measured because vascular imaging was performed at another hospital and was not available for review. The highest quality imaging used to confirm the presence of an aneurysm was conventional angiogram in 13 patients, CTA in eight patients and MRA in 15 patients. Although in six patients the diagnosis of an aneurysm was made at another hospital and the type of imaging used was not recorded or available for review, we included these patients in the study.
Two patients with lesions which were either aneurysms or infundibula were excluded from the study. One patient had a 1.5 mm aneurysm or infundibulum of the paraclinoid carotid artery that was visualized on MRA and the second had a 3 mm supraclinoid carotid aneurysm or infundibulum discovered with CTA. Neither of the patients who were excluded experienced SAH while receiving anticoagulation. An additional patient with an arteriovenous malformation and a coexisting extranidal aneurysm of the anterior choroidal artery that was treated with coil embolization was included in our series.
Considering aneurysms of the posterior communicating artery segment of the internal carotid artery to be part of the posterior circulation, there were 40 aneurysms in the anterior circulation (83%) (table 2). In two patients the intracranial location could not be determined from available medical records.
Aneurysm location
Six of the 48 aneurysms had a prior history of rupture; two were treated by coiling and four by clipping prior to receiving anticoagulation. Five unruptured aneurysms were treated by coiling, four by clipping and one by wrapping before the period of anticoagulation.
Of the 48 aneurysms, 32 were not treated before or during the period of anticoagulation. Three patients had their unruptured aneurysms clipped or coiled during the period of follow-up. For the purposes of this study, these patients were considered to have treated aneurysms during the period of anticoagulation, even though for a period of time on anticoagulation during follow-up their aneurysms had not yet been clipped or coiled. One of these patients was receiving anticoagulation for 45 days prior to stent-assisted coiling of her aneurysm followed by 5 more days of anticoagulation. Another patient received anticoagulation for 1 year prior to clipping of her aneurysm and was treated with anticoagulation for 6 more years after clipping of the aneurysm. The third patient, who had two aneurysms, developed a DVT after clipping of the aneurysm in the left middle cerebral artery (MCA). The patient was treated with anticoagulation for 5 months after clipping of the MCA aneurysm and then anticoagulation was stopped before stent-assisted coiling of his second aneurysm, a 3 mm right ophthalmic aneurysm.
Patient outcome and follow-up
One patient had bleeding in the left basal ganglia while on warfarin with an INR of 2.7. The hemorrhage was attributed to metastatic cancer and was unrelated to the patient's right MCA aneurysm. One patient with a 6 mm anterior communicating artery aneurysm had a hemorrhagic stroke in the basal ganglia while receiving anticoagulation. The hemorrhage was attributed to hypertension.
No patient treated with anticoagulation with a diagnosis of an intracranial aneurysm developed symptomatic rupture of any aneurysm during follow-up. The mean±SD follow-up time for all patients was 495±749 days (figure 1). The total duration of follow-up for all patients while receiving anticoagulation was 57 years. For patients with aneurysms that were treated with endovascular coiling or surgical techniques, the total cumulative follow-up time was 26 years and the mean±SD follow-up time was 627±1047 days. The total follow-up time for patients with aneurysms that were never treated with neurosurgical techniques or endovascular coil embolization was 31 years with a mean±SD follow-up of 423±527 days.
Follow-up times (days) for individual patients receiving anticoagulation with intradural aneurysms.
Extradural intracranial aneurysms
As part of a separate analysis we followed seven patients with aneurysms of the cavernous segment of the internal carotid artery and one patient with a 1.5 mm petrous internal carotid aneurysm. Six of the patients had untreated cavernous aneurysms ranging in size from 1.5 to 6 mm and one patient had a 26 mm cavernous aneurysm that was treated with parent vessel occlusion prior to the period of anticoagulation. This group of patients with extradural aneurysms had a total of 10 years of follow-up and none experienced rupture of their aneurysm while receiving therapeutic doses of anticoagulation.
Discussion
Estimates of the rate of rupture of aneurysms in the general population vary and appear to depend on the size and location of the aneurysm and whether the patient has previously experienced rupture of another aneurysm. ISUIA estimated that, for aneurysms <10 mm, the rate of aneurysm rupture in patients without a history of SAH is <0.05% per year, and 0.5% per year in patients who have a prior history of aneurysmal SAH.4 The prospective arm of the study showed a 0.52% annual rate of rupture (5-year rupture rate of 2.6%) for anterior circulation aneurysms (not including posterior communicating artery aneurysms) that were 7–12 mm and a 5-year rate of rupture of 0% for aneurysms <7 mm in patients without a prior history of rupture.7 One meta-analysis, which included data from the ISUIA, estimated the rate to be 0.49% per year for aneurysms in the anterior circulation. This meta-analysis included patients with SAH and included posterior communicating artery aneurysms with the anterior circulation aneurysms.22
Whether anticoagulation increases the risk of rupture of an aneurysm is a question that commonly enters clinical practice. Despite the frequency of anticoagulation use and aneurysms in the general population, there are few clinical data to inform the clinician on the safety of starting anticoagulation in this group of patients. A retrospective case series reported one patient who developed a fusiform intracranial dissecting aneurysm during 3 months of anticoagulation during which time no SAH was observed.23 Another case report described a patient with an 8 cm anterior communicating artery aneurysm who was treated with anticoagulation for 6 months and also did not rupture.24 Finney and Gholston reported a case of a 51-year-old man with a history of Sturge–Weber syndrome in whom a 2 cm aneurysm of the middle cerebral artery ruptured while he was receiving treatment with warfarin for a myocardial infarction with a prothrombin time of 51 s (normal range 11–16 s).25 The patient underwent craniotomy for evacuation of a hematoma and clipping of the aneurysm. No arteriovenous malformation was found. The patient survived with a severe hemiplegia.
None of the 42 patients with intradural aneurysms in our study who were receiving anticoagulation for 57 combined years experienced symptomatic aneurysm rupture. The risk of aneurysm rupture found in our study was therefore no higher than the risk of rupture in patients with unruptured aneurysms in the general population described in the above referenced studies. Furthermore, some of the patients in our study had a prior history of rupture, which may have additionally predisposed them to a higher than expected risk of rupture which was not borne out (although treatment of those aneurysms with endovascular and surgical techniques prior to starting anticoagulation could have decreased the risk of re-rupture while on anticoagulation).
Several factors may have underestimated the risk of rupture of intracranial aneurysms while receiving anticoagulation. Small aneurysm size is associated with a lower rate of rupture. Our study included three aneurysms that were <2 mm, and the majority of aneurysms in this study were small (<7 mm). There was a potential selection bias for inclusion in this study because medical care providers may have decided not to anticoagulate patients with large aneurysms; this group of patients would not have come to our attention. The inclusion of these ‘low-risk’ patients may have led us to underestimate the risks of anticoagulation since small aneurysms are less prone to SAH.
A shortcoming is that patients who died before reaching the hospital or who could have been taken to another hospital would not have come to our attention. Also, given the small number of patients in this study, we cannot exclude the possibility of a moderate increase in risk of SAH in patients with aneurysms who take anticoagulation. Another caveat is that, in theory, if rupture of an aneurysm occurs while receiving anticoagulation, it is likely to be more fatal.
Conclusions
The risk of aneurysm rupture in patients on therapeutic anticoagulation in this study is no higher than previously published data on the risk of SAH in patients who are not anticoagulated. Future prospective studies should be performed in order to estimate this risk better and substantiate our conclusions.
Acknowledgments
The authors thank Amy Deipolyi for her proof reading of the manuscript.
References
Footnotes
Competing interests None.
Patient consent A waiver of informed consent was obtained for the completion of this study from the IRB of our hospital. The study is retrospective and involved reviewing medical records, with no risk to the patient.
Provenance and peer review Not commissioned; externally peer reviewed.