Spontaneous Intracranial Hemorrhage: Which Patients Need Diagnostic Cerebral Angiography?
A Prospective Study of 206 Cases and Review of the Literature
Background and Purpose In spontaneous intracerebral hemorrhage (ICH), the site, age of the patients, and preexisting hypertension are important factors in determining the possibility of finding an underlying vascular abnormality by cerebral angiography. To what extent these three factors affect the indication for angiography remains controversial. A prospective study was carried out to correlate the angiographic findings with these three factors.
Methods Two hundred six consecutive spontaneous ICH cases with an age range from 5 to 79 years (median, 45) were investigated with CT and cerebral angiography over a 3-year period (April 1993 through March 1996). Exclusion criteria were (1) poor surgical risk or severely neurologically disabled patients, (2) refusal of angiography, (3) patients in whom severe coagulopathy accounted for the hemorrhage, (4) bleeding into tumor, or (5) subarachnoid hemorrhage–predominant cases.
Results Angiographic yield (the frequency of positive angiography in a defined patient group) was significantly higher in patients (1) at or below the median age of 45 than those above (53/105, 50%, versus 18/101, 18%; P<.001) and (2) without preexisting hypertension than those with (64/145, 44%, versus 5/58, 9%; P<.001). The correlation of age and preexisting hypertension to angiographic yield was independent (logistic regression coefficients −0.056 and −1.59 and SE 0.12 and 0.515, respectively, both P<.001). In patients of the younger age group without preexisting hypertension, angiographic yield was 48% in putaminal, thalamic, or posterior fossa ICH and 65% in lobar ICH. In the older hypertensive patients, the yields were 0% and 10%, respectively. However, in patients with isolated intraventricular hemorrhage, most were normotensive and the yield was high in both age groups (67% versus 63%).
Conclusion Diagnostic cerebral angiography should be considered for all spontaneous ICH patients except those over 45 years old with preexisting hypertension in thalamic, putaminal, or posterior fossa hemorrhage.
- spontaneous intracerebral hemorrhage
- intraventricular hemorrhage
- cerebral angiography
- computed tomography
Spontaneous intracerebral hemorrhage refers to those cases that occur in the absence of trauma. One of the major management tasks is to find out whether the hemorrhage is secondary to underlying structural vascular abnormality, such as arteriovenous malformation (AVM) and aneurysm, so that appropriate treatment can be carried out to prevent rebleeding. Despite great advances in imaging technology, such as CTA, MRI, and MRA, conventional catheter angiography remains the gold standard for diagnosis of vascular abnormality.1 2 3 Because this is an invasive procedure carrying a low but definite risk,4 5 6 the balance must be weighed between the diagnostic yield and the procedural risk. It has been documented that the site of hemorrhage, the patient’s age, and the presence of preexisting hypertension are important factors affecting the likelihood of finding a vascular abnormality by cerebral angiography in spontaneous ICH. However, the extent to which these three factors affect the indication for angiography remains controversial and only a few studies3 7 8 9 10 11 have addressed this issue.
Here we report the results of our prospective study of 206 consecutive spontaneous ICH cases investigated with CT and cerebral angiography and discuss the roles of site of hemorrhage, age of patients, and preexisting hypertension in the diagnostic yield of angiography.
Materials and Methods
All spontaneous ICH cases admitted to the neurosurgical service from April 1993 through March 1996 were included in this prospective observational study. The neurosurgical service is a tertiary referral center with a catchment population of 1.5 million, which amounts to approximately one quarter of the Hong Kong population. The diagnosis of spontaneous ICH was made on spontaneous acute onset of neurological symptoms and signs confirmed by CT within 24 hours. The ICH was classified according to the location of the hematoma into lobar, putaminal, thalamic, or caudate nucleus ICH, IVH, and posterior fossa hemorrhage (including cerebellar and brain stem). In cases in which parenchymal hematoma ruptured into the ventricle, the parenchymal hemorrhage was defined as the anatomical site of hemorrhage. IVH cases were defined as those without parenchymal hematoma.
Cerebral angiography was indicated when the presence of structural vascular abnormalities such as aneurysm or AVM was thought to affect patient management. Patients were excluded from the study for the following reasons: (1) poor surgical risk or severe neurological disability to the extent that even if underlying vascular abnormalities were found the management would not be affected; (2) refusal of angiography; (3) severe coagulopathy that accounted for the hemorrhage; (4) bleeding into tumor that was diagnosed by CT or MRI, whereas angiography was used for tumor vascularity or vascular displacement only; or (5) SAH-predominant cases.
Angiography via the femoral approach (digital subtraction angiography) was carried out under local anesthesia once the patient’s condition had stabilized and within the same admission. However, in younger, uncooperative, or comatose patients the procedure was carried out under general anesthesia (in 35 of the 206 cases, or 17%). If the first angiography was inconclusive, a second study was carried out 6 weeks later.
A positive angiography was defined as the identification of a vascular abnormality accounting for the hemorrhage. Angiographic yield was defined as the frequency of positive angiography in a defined patient group.
The presence of preexisting hypertension was documented from the medical history, by evidence of chronic hypertension (cardiomegaly on chest x-ray and/or left ventricular hypertrophy on ECG), or by persistent hypertension (systolic blood pressure >140 mm Hg or diastolic blood pressure >90 mm Hg) more than 2 weeks after the onset of spontaneous ICH.
The χ2 test was used for comparison of angiographic yield and logistic regression for multivariant analysis. Significance was set at P<.05.
A total of 438 consecutive spontaneous ICH cases admitted to the neurosurgical service from April 1993 through March 1996 was studied. Seventy-nine cases were excluded because of SAH. One hundred fifty-three did not have angiography, 130 of them because of poor neurological condition or concurrent medical problems and the rest because of bleeding diatheses, bleeding into tumor, refusal of angiography, or loss of follow-up. Fifty-three of the non–SAH-excluded patients died within the same admission.
Two hundred six non-SAH cases received both CT and angiography. Among them, there were 129 males and 74 females aged 5 to 79 years (mean, 42 years; median, 45 years). All were ethnic Chinese. The distributions of age and site of hemorrhage are shown in Table 1⇓. Seventy-one patients (34%) had positive angiographic findings, which are shown in Table 2⇓. Most of the angiography studies were carried out in the acute stage (136 cases within 1 week, median 2 days; 45 within 2 to 4 weeks, median 21 days; 25 after more than 4 weeks, median 57 days). Eight patients underwent repeat angiography, with no change in angiographic diagnosis.
Fifty-eight cases had preexisting hypertension (28%); 16 of them were between ages 16 to 45 years, whereas 42 were older than 45 years. The angiographic yield was 45% (66/148) in normotensive patients in contrast to 9% (5/58) in hypertensive patients (P<.001). Among the 5 cases with hypertension and positive angiography, 2 were AVMs (1 patient age 37 years with IVH and 1 patient age 67 years with lobar ICH) and 3 were aneurysms (1 patient age 56 years with caudate ICH and 2 patients ages 45 and 61 years with lobar ICH). Angiographic yields decreased as patients’ ages increased (Table 3⇓). The angiographic yield was significantly greater in the age group of 45 years or younger (50% [53/105] versus 18% [18/101], P<.001). By logistic regression analysis, age and preexisting hypertension were both shown to be significant independent factors associated with a negative angiography (age: coefficient −0.056, standard error 0.012, P<.01; hypertension: coefficient −1.59, standard error 0.515, P<.01).
Site was also important. Of patients with IVH without parenchymal hematoma (ie, isolated IVH with and without SAH), 65% (11/17) had positive angiography. The angiographic yield was 6 of 9 patients younger than 45 years of age (67%) and 5 of 8 older than 45 years of age (63%). Only one 67-year-old patient had a history of hypertension and he had an AVM. Lobar ICH with IVH had a higher yield than lobar ICH without IVH (9/13, 65%, versus 35/77, 43%; P=.11), although the difference was not statistically significant. There was no increase in angiographic yield for putaminal or thalamic ICH with IVH compared with those without IVH.
Of the patients with lobar ICH 49% had positive angiography. Only 19% of cases had hypertension. When all lobar ICH cases were divided into four groups by age and by presence or absence of hypertension, there was a significant decrease in angiographic yield from the older hypertensive group (10%) compared with the younger normotensive group (65%).
In posterior fossa ICH (cerebellar and brain stem), the 6 patients with AVM were normotensive and the oldest patient was 39 years old. There were 6 patients who had preexisting hypertension; the youngest was 48 years old. None of them had positive angiography. In putaminal ICH, only one 11-year-old boy had AVM and one 62-year-old man had an aneurysm; both were normotensive. There were 18 patients with preexisting hypertension (44%), and 11 of them were older than 45 years. In thalamic ICH, 15 patients were hypertensive; 12 of them were older than 45 years. All 6 angiography- positive cases were normotensive and the oldest was 39 years old. Grouping putaminal, thalamic, and posterior fossa ICH cases together and dividing them into four groups according to age (≤45 and >45) and whether they had preexisting hypertension (Table 4⇓) showed that the angiographic yield was 48% in the younger normotensive group, 7% in the older normotensive group, and 0% in all hypertensive patients (P<.001).
Spontaneous ICH, or classic apoplexy, has been documented in autopsy studies to be more common in the striatum and posterior fossa12 and has been associated with preexisting hypertension. However, even in the presence of preexisting hypertension as much as 36% in a series of 144 autopsy cases showed a specific structural lesion causing the brain hemorrhage.10 The likelihood of demonstrating a vascular lesion by cerebral angiography was significantly greater in lobar hemorrhage compared with deep-seated hemorrhage.3 7 8 11 These studies also suggested a negative impact of age and preexisting hypertension on angiographic yield. On the basis of these data, some authors have suggested that in hypertensive patients with deep-seated brain hemorrhage, cerebral angiography was not indicated,11 whereas others have recommended that angiography should be considered for all patients with spontaneous ICH, regardless of site or history of hypertension.3
Our study contained the largest number of patients studied to date, with prospective data on the site of hemorrhage, age, and preexisting hypertension. Patients were selected by referral to the neurosurgical service and by our exclusion criteria. Patients who were considered too old and/or medically unfit were excluded by the referring physicians. Our results represented the positive angiographic rate in a selective group of spontaneous ICH patients who had been referred for possible neurosurgical intervention.
Our data demonstrate for the first time that both age and the presence of preexisting hypertension are two independent negative factors correlated to angiographic yield. However, the 18% angiographic yield in the older age group and the 9% yield in the hypertensive group should be regarded as clinically significant and therefore outweighing the procedural risk of angiography. Further analysis of angiographic yield in patients older than 45 years showed that there was no trend of decrease toward older age groups (Table 3⇑). As the incidence of AVM decreased with age, the incidence of aneurysm increased. However, in hypertensive patients with putaminal, thalamic, and/or posterior fossa hemorrhage, angiographic yield was 0% (Table 4⇑). Younger hypertensive patients with these types of hemorrhage also showed no vascular lesions on angiography: There were only 10 patients in this group and the youngest of them was 34 years of age. It was therefore not possible to draw any age-related conclusions for this group of patients. Previous studies showed conflicting data. Angiographic yield in patients with putaminal, thalamic, and/or posterior fossa hemorrhage varied from 0% to 50%.3 11 The relationship between angiographic findings, age, and preexisting hypertension in these two previous studies was not analyzed because of the small number of patients in the group.
The impact of age and preexisting hypertension on angiographic yield was most significant in lobar brain hemorrhage. However, in patients with IVH, the high angiographic yield of 65% (Table 4⇑), which is consistent with previous studies,11 13 14 was not affected by age. The influence of hypertension could not be determined because only one case had preexisting hypertension and this patient had an AVM.
More recently, CT, CTA, MRI, and MRA have been employed as noninvasive screening procedures for the investigation of intracranial vascular abnormality.15 16 17 18 Both CTA and MRA have good sensitivity, approaching 90%, for aneurysms larger than 3 mm in comparison with conventional angiography.19 20 21 22 23 Others have suggested that 5 mm is the critical size for MRA.6 23 MRI is superior in demonstrating the exact anatomic relationships of AVM,24 25 26 but a small lesion in the presence of a hematoma can be missed.2 24 27 Hence, these noninvasive imaging modalities can be used for screening in spontaneous ICH, particularly in older hypertensive patients. MRI can be used for investigation of possible AVM and MRA or CTA for aneurysm. However, a negative result cannot completely exclude a vascular lesion, and conventional cerebral angiography may be required for definitive diagnosis and treatment planning.1 2 3
In summary, our study showed that among a group of spontaneous ICH patients who were referred for possible neurosurgical intervention, age and preexisting hypertension were two independent negative factors correlated to the angiographic yield. The angiographic yield was significantly higher in patients (1) at or below the age of 45 years and (2) without preexisting hypertension. Analysis of angiographic yield of different sites of hemorrhage together with these two factors showed that (1) lobar ICH had at least a 10% yield in the older patients with preexisting hypertension and up to a 64% yield in the younger normotensive patients; (2) putaminal, thalamic, and posterior fossa ICH grouped together had a yield of 48% in normotensive patients at or younger than 45 years, 7% in normotensive patients older than 45 years of age, but 0% in all patients with preexisting hypertension; and (3) isolated IVH patients had 63% yield in the older and 67% yield in the younger groups.
On the basis of our results, we suggest that diagnostic cerebral angiography should be considered for all spontaneous ICH patients except those older than 45 years with preexisting hypertension in thalamic, putaminal, or posterior fossa hemorrhage.
Selected Abbreviations and Acronyms
|CTA||=||computed tomographic angiography|
|MRA||=||magnetic resonance angiography|
- Received November 27, 1996.
- Revision received April 14, 1997.
- Accepted May 2, 1997.
- Copyright © 1997 by American Heart Association
Fleckenstein JL, Oshinski JN. Magnetic resonance angiography of cerebrovascular disease. In: Batjer HH, Caplan LR, Friberg L, Greenlee RG Jr, Kopitnik TA Jr, Young WI, eds. Cerebrovascular Disease. Philadelphia, Pa: Lippincott-Raven; 1997:221-237.
Grinsburg MI, Chason DP, Mendelsohn DB. Computed tomography of cerebrovascular disease. In: Batjer HH, Caplan LR, Friberg L, Greenlee RG Jr, Kopitnik TA Jr, Young WI, eds. Cerebrovascular Disease. Philadelphia, Pa: Lippincott-Raven; 1997:161-183.
Halpin SFS, Britton JA, Byrne JA, Clifton A, Hart G, Moore A. Prospective evaluation of cerebral angiography and computed tomography in cerebral haematoma. J Neurol Neurosurg Psychiatr. 1994;57:1180-1186.
Hankey GJ, Warlow CP, Sellar RJ. Cerebral angiographic risk in mild cerebral vascular disease. Stroke. 1990;21:209-222.
Hayward RD, O’Reilly GV. Intracerebral haemorrhage: accuracy of computed transverse axial scanning in predicting the underlying etiology. Lancet. 1976;1:120-126.
Loes DJ, Smoker WRK, Biller J, Cornell S. Nontraumatic lobar intracerebral hemorrhage: CT/angiographic correlation. Am J Neuroradiol. 1987;8:1027-1030.
McCormick WF, Rosenfield DB. Massive brain hemorrhage: a review of 144 cases and an examination of their causes. Stroke. 1973;4:946-954.
Toffol GJ, Biller J, Adams HP Jr, Smoker WRK. The predicted value of arteriography in nontraumatic intracerebral hemorrhage. Stroke. 1986;17:881-883.
Russell DS, in discussion. The pathology of spontaneous intracerebral hemorrhage. Proc R Soc Med. 1953;47:689-704.
Little JR, Blomquist GA Jr, Ethier R. Intraventricular hemorrhage in adults. Stroke. 1987;8:143-149.
Androux RW, Russell EJ, Meyer JR. MRI of cerebrovascular disease. In: Batjer HH, Caplan LR, Friberg L, Greenlee RG Jr, Kopitnik TA Jr, Young WI, eds. Cerebrovascular Disease. Philadelphia, Pa: Lippincott-Raven; 1997:185-210.
Atlas SW, Hurst RW. Intracranial vascular malformations and aneurysms. In: Atlas SW, ed. Magnetic Resonance Imaging of the Brain and Spine. 2nd ed. Philadelphia, Pa: Lippincott-Raven; 1996:489-556.
Huston J, Nichols DA, Luetmer PH, Goodwin JT, Meyer FB, Weibers DO, Weaver AL, Winn HR. Blinded prospective evaluation of sensitivity of MR angiography to known intracranial aneurysms: importance of aneurysm size. Am J Neuroradiol. 1994;15:1607-1617.
Surgai Y, Hamamoto Y, Ookubo T, So K. Angiographical frequency of unruptured incidental intracranial aneurysms. Neurol Surg. 1994;22:429-432.