Background and Purpose Intracranial hemorrhage is an important complication of treatment with anticoagulants. We studied outcome of aneurysmal subarachnoid hemorrhage (SAH) occurring in patients on anticoagulant drugs because this may influence management of patients needing anticoagulant treatment but with increased risk of aneurysmal hemorrhage.
Methods From the prospective database of patients admitted with SAH to the Utrecht University Hospital, we compared 3-month outcome in patients with or without anticoagulant drugs by means of stratified analyses to adjust relative risks for biological and clinical differences between the two groups.
Results Death or dependency after SAH occurred in 14 of 15 patients on anticoagulant treatment and in 62 of 126 patients not being treated with anticoagulants (relative risk, 1.9; 95% confidence interval, 1.5 to 2.4).The patients on treatment with anticoagulants were more often comatose on admission; the frequencies of rebleeding, secondary ischemia, and hydrocephalus were not higher in patients on anticoagulants. In the stratified analysis, the worse outcome in the group on anticoagulant drugs was not essentially influenced by differences in sex, age, cardiovascular history, site of aneurysm, amount of cisternal blood, or extension of hemorrhage into the ventricles.
Conclusions The outcome of aneurysmal SAH in patients on anticoagulant drugs is extremely poor. The explanation for the worse prognosis in patients on anticoagulants lies in a worse clinical condition from the outset. The poor outcome urges a reconsideration of the balance of risks for anticoagulant treatment in patients with an unoperated intracranial aneurysm or with a family history of SAH and may lead to withholding treatment with anticoagulant drugs or to a preventive operation.
In a number of defined vascular disorders, the benefits of treatment with anticoagulants by far outweigh the risk of complications,1 among which intracranial hemorrhage is especially feared. The balance of risks for anticoagulant treatment might be different in patients at increased risk of hemorrhagic stroke, such as patients who are known to have an intracranial aneurysm or patients in whom an aneurysm is detected during the diagnostic workup after an ischemic stroke. These aneurysms are not rare: more than 5% of patients older than 50 years harbor an intracranial aneurysm.2 3 In patients evaluated after ischemic stroke this proportion is probably even higher, because patients with atherosclerotic vessel diseases and patients with intracranial aneurysms have common risk factors, such as smoking and hypertension.4 Another category of patients at increased risk of SAH are patients with a positive family history for SAH.5
Outcome after SAH is a crucial factor in the balance of risks regarding whether to begin anticoagulant treatment or whether to perform preventive surgery of the intracranial aneurysm before anticoagulant treatment is instituted. In general, the outcome of SAH is poor: approximately 40% of patients die, and half the patients who survive are handicapped.6 7 8 9 The prolonged bleeding expected in patients on anticoagulants might result in an even worse prognosis than usual. To investigate the influence of anticoagulant treatment on outcome after SAH, we compared in a consecutive series of patients the outcome of SAH in patients on anticoagulant treatment with that of patients not being treated with anticoagulants.
Subjects and Methods
From the database of patients admitted with SAH to the Utrecht University Hospital, we selected all patients undergoing anticoagulant treatment at the time of hemorrhage and compared baseline characteristics and eventual outcome with those in a consecutive series of patients not being treated with anticoagulants, selected from the same database. The diagnosis of SAH from rupture of an intracranial saccular aneurysm was based on an aneurysmal pattern of hemorrhage on CT10 or, in patients with a normal CT scan, xanthochromic cerebrospinal fluid and an aneurysm on angiography or autopsy. Patients with other causes for the SAH were excluded. The patients on anticoagulant drugs were selected from the period between January 1989 and June 1995. We decided in advance to select five times as many patients for the reference group as for the anticoagulated group because no substantially greater precision of the relative risk estimate may be expected at higher ratios. This proved to be the case after the patients from 1993 through 1995 had been included. For the present study, patients with a perimesencephalic hemorrhage were excluded from both series, since these patients invariably have a good prognosis.11 The inclusion criterion for the anticoagulated series was a history (provided by the patient, next of kin, or general practitioner) of treatment with anticoagulants supported by an INR above 1.5. The anticoagulated state was antagonized immediately after admission in our hospital in all patients. Inclusion criteria for the reference group were a statement by the patients that they were was not receiving anticoagulant treatment (in these patients INR was not always assessed) or, in patients with a depressed level of consciousness, an INR lower than 1.1. Patients with an INR between 1.1 and 1.5 were excluded from both groups. CT scanning was done in all patients, but angiography was not performed if the clinical condition was judged to preclude surgery.
The clinical condition on admission was graded by means of the WFNS grading scale12 ; in the analysis we dichotomized between presence and absence of coma (WFNS grades 1, 2, 3, and 4 versus WFNS grade 5). The amount of subarachnoid blood was assessed by means of the grading scale described by Hijdra and coworkers.13 With this method, each of the 10 basal cisterns and fissures (frontal interhemispheric fissure, quadrigeminal cistern, and the paired suprasellar cisterns, ambient cisterns, basal sylvian fissures, and lateral sylvian fissures) is graded separately on a semiquantitative scale, according to the amount of extravasated blood: 0, no blood; 1, small amount of blood; 2, moderately filled with blood; and 3, completely filled with blood. The total amount of subarachnoid blood (total sum score) was calculated by adding the 10 scores and could range from 0 to 30; this variable was dichotomized at the median in the analysis. This grading system has good interobserver agreement14 and is a reliable predictor for outcome in general and for the development of secondary ischemia in particular. Delayed cerebral ischemia was diagnosed if (1) a decrease in level of consciousness or new focal signs emerged gradually and (2) CT excluded other causes. Rebleeding was diagnosed if (1) there was a sudden decrease in level of consciousness, with or without new focal signs, and (2) CT showed an increase of subarachnoid, intracerebral, or intraventricular blood in comparison with the previous scan. Hydrocephalus was diagnosed if patients had (1) a gradual decrease in level of consciousness and (2) enlargement of the lateral or third ventricles compared with a previous scan. Outcome at 3 months was graded by means of the Glasgow Outcome Scale.15 In the outcome analysis we defined poor outcome as death or dependency.
In the comparison of outcome for patients on and off anticoagulant drugs, we used relative risks with matching 95% confidence intervals. Because baseline characteristics differed for the two groups, we used stratified analysis to calculate Mantel-Haenszel adjusted relative risks and corresponding 95% confidence intervals.16 Stratified analysis was used for the following characteristics: age, sex, cardiovascular history, amount of cisternal blood, intraventricular extension of hemorrhage, and site of aneurysm (vertebrobasilar versus carotid circulation).
We found 15 patients in whom aneurysmal rupture had occurred during treatment with anticoagulants. The reasons for anticoagulant treatment were ischemic heart disease (9 patients); atrial fibrillation (2); valve prosthesis (1); recurrent venous thrombosis (1); recent hip replacement (1); and recent bilateral knee prostheses (1). The reference group consisted of 126 patients with aneurysmal rupture not being treated with anticoagulants. The baseline characteristics of the 15 patients on anticoagulant drugs and those of the untreated reference series are listed in Table 1⇓. The patients on treatment with anticoagulants were more often men, were older, more often had a cardiovascular history, and more often were comatose on admission. They tended to have a larger amount of subarachnoid blood, more often had intraventricular as well as intraparenchymal extension of the hemorrhage, and more often had a posterior circulation aneurysm. The proportions of patients with hydrocephalus, delayed cerebral ischemia, and rebleeding were similar in both series, but the patients on treatment underwent surgery less frequently (Table 2⇓).
Outcome was significantly worse in the patients on treatment with anticoagulants: 14 of 15 patients had a poor outcome compared with only 62 of 126 patients not on treatment with anticoagulants (Table 2⇑). In the stratified analysis, the worse outcome in the anticoagulated group was not markedly influenced by differences in sex, age, site of aneurysm, amount of cisternal blood, extension of hemorrhage into the ventricles or parenchyma, or cardiovascular history; the adjusted relative risks were essentially the same as the crude relative risks.
We found that hemorrhage from a ruptured, saccular aneurysm is devastating in patients undergoing anticoagulant treatment; 14 of the 15 patients had died or were dependent as a result of the hemorrhage. Treatment with anticoagulant drugs doubles the already high risk of poor outcome in patients with SAH. This worsening in outcome proved not to be influenced by potential confounders, such as age or cardiovascular history.
The explanation for the worse prognosis in patients on anticoagulants lies in the worse clinical condition from the outset, which is expressed on CT by a tendency toward a more extensive extravasation of blood in the basal cisterns and higher frequency of intraparenchymal extension of the hemorrhage. Anticoagulants seem to increase the extent of the bleeding at the time of aneurysmal rupture and thereby provoke a greater impact of the hemorrhage, which emerges clinically as a higher percentage of patients with coma on admission. A more abundant bleeding related to treatment with anticoagulants has also been found in patients with primary intracerebral hemorrhage, but in intracerebral hemorrhage outcome is only slightly worse in patients on treatment with anticoagulants.17 In these patients, the tamponade of the hemorrhage by the parenchyma may limit the extravasation. We did not collect data on the use of aspirin or other NSAIDs. The prolonged bleeding time in patients who use aspirin or other NSAIDs may result in an effect on hemorrhage from a ruptured aneurysm similar to the use of anticoagulant drugs. In a recent study, however, it was found that use of NSAIDs preceding aneurysmal SAH does not affect final outcome and that NSAIDs taken shortly after the hemorrhage may even improve outcome by reducing the risk of secondary ischemia.18 Therefore, we do not believe that lack of data on NSAID use has biased our results.
The study group and the reference group differ in two aspects that at first glance might have biased the results. First, surgical treatment of the aneurysm was performed less frequently in patients on treatment with anticoagulants. In our institution the decision to operate depends almost exclusively on the clinical condition. Clipping of the aneurysm is performed as early as possible in patients with a normal level of consciousness on admission; in other patients operation is deferred until 10 days after the hemorrhage and performed only if patients have recovered to at least WFNS grade III. The proportion of patients who underwent surgical treatment is therefore merely an indicator of the clinical condition and not a bias in final outcome. Second, the study group was selected from the patients admitted between 1989 and 1995, and the reference group was selected from those admitted between 1993 and 1995. During the period 1989 through 1995 all patients were treated with nimodipine and hypervolemia, and the overall results in our institution have not changed during these years (G.J.E.R., unpublished data, 1996). The longer period for selection of the study group therefore cannot explain the substantial difference in outcome between the two series of patients.
With the increasing use of CT and MR techniques, asymptomatic aneurysms are found by chance in a growing number of patients. The finding of an unruptured aneurysm may create a dilemma in itself, even in patients not being treated with anticoagulants. In younger patients the choice for elective surgery of unruptured intracranial aneurysms is usually straightforward, but with increasing age the risks of surgery can outweigh the risk of aneurysmal rupture. The risks of surgery are often increased in patients at risk of cardiovascular diseases, which applies to most patients in whom anticoagulants are indicated. The present study shows that in patients with an indication for anticoagulants, the decision for surgery of an unruptured aneurysm depends not only on the increased perioperative mortality and morbidity from the cardiovascular disease that led to the indication for anticoagulants but also on the devastating outcome of SAH if it occurs during treatment with anticoagulants. This expected poor outcome can have a major impact on the decision.19
An increased risk of SAH is present not only in patients known to have an aneurysm but also in patients with inherited disorders such as polycystic kidney disease, Marfan's syndrome, or Ehlers-Danlos syndrome and in asymptomatic, first-degree relatives of patients with SAH.5 If an indication to start anticoagulant treatment emerges in any of these categories of patients, the poor chance of recovery after SAH that occurs during anticoagulant treatment should at least be incorporated in weighing the benefits and risks of this treatment and may even favor screening these patients for an intracranial aneurysm.
To summarize, anticoagulants are widely used in the prevention of thromboembolic cardiovascular disease, and aneurysms are not rare in the age group of patients for whom treatment with anticoagulants is considered. If a patient eligible for treatment with anticoagulants is known to be at increased risk of SAH or to have an unoperated intracranial aneurysm, the poor outcome of aneurysmal SAH urges a reconsideration of the balance of risks, leading to withholding treatment with anticoagulants or to a preventive operation.
Selected Abbreviations and Acronyms
|INR||=||international normalized ratio|
|NSAIDs||=||nonsteroidal anti-inflammatory drugs|
|WFNS||=||World Federation of Neurological Surgeons|
This study was partially funded by a clinical investigator grant from the University Hospital Utrecht to Dr Rinkel. We would like to thank J. van Gijn, MD, for his comments on a previous version of the manuscript.
- Received July 2, 1996.
- Revision received October 15, 1996.
- Accepted October 21, 1996.
- Copyright © 1997 by American Heart Association
Teunissen LL, Rinkel GJE, Algra A, van Gijn J. Risk factors for subarachnoid hemorrhage: a systematic review. Stroke. 1996;27:544-549.
Bromberg JEC, Rinkel GJE, Algra A, et al. Subarachnoid haemorrhage in first and second degree relatives of patients with subarachnoid haemorrhage. BMJ. 1995;311:288-289.
Anderson CS, Jamrozik KD, Broadhurst RJ, Stewart-Wynne EG. Predicting survival for 1 year among different subtypes of stroke: results from the Perth Community Stroke Study. Stroke. 1994;25:1935-1944.
Broderick JP, Brott TG, Duldner JE, Tomsick T, Leach A. Initial and recurrent bleeding are the major causes of death following subarachnoid hemorrhage. Stroke. 1994;25:1342-1347.
Inagawa T, Tokuda Y, Ohbayashi N, Takaya M, Moritake K. Study of aneurysmal subarachnoid hemorrhage in Izumo City, Japan. Stroke. 1995;26:761-766.
Bamford JM, Sandercock PAG, Dennis MS, Burn JPS, Warlow CP. A prospective study of acute cerebrovascular disease in the community: the Oxfordshire Community Stroke Project 1981-1986, 2: incidence, case fatality rates and overall outcome at one year of cerebral infarction, primary intracerebral and subarachnoid haemorrhage. J Neurol Neurosurg Psychiatry. 1990;53:16-22.
Rinkel GJE, Wijdicks EFM, Vermeulen M, Ramos LMP, Tanghe HL, Meiners LC, van Gijn J. Nonaneurysmal perimesencephalic subarachnoid hemorrhage: CT and MR patterns that differ from aneurysmal rupture. AJNR Am J Neuroradiol. 1991;12:829-834.
Hijdra A, Brouwers PJAM, Vermeulen M, van Gijn J. Grading the amount of blood on computed tomograms after subarachnoid hemorrhage. Stroke. 1990;21:1156-1161.
Brouwers PJAM, Dippel DW, Vermeulen M, Lindsay KW, Hasan D, van Gijn J. Amount of blood on computed tomography as an independent predictor after aneurysm rupture. Stroke. 1993;24:809-814.
Kleinbaum DG, Kupper LL, Morgenstern H. Epidemiology Research. Belmont, Calif: Lifetime Learning Publications; 1982:345.
Franke CL, de Jonge J, van Swieten JC, Op de Coul AAW, van Gijn J. Intracerebral hematomas during anticoagulant treatment. Stroke. 1990;21:726-730.
Crevel H, van Habbema JDF, Braakman R. Decision analysis of the management of incidental intracranial aneurysms. Neurology. 1986;36:1335-1339.