Neurological and Functional Outcome in Patients With Supratentorial Hemorrhages
A Prospective Study
Background and Purpose A prospective study was performed to evaluate neurological and functional outcome after spontaneous supratentorial bleeding. The aim of the study was to determine whether clinical or neuroradiological parameters could predict the outcome of these patients during the first hours of hospitalization.
Methods Two hundred seventy-nine patients—52 with thalamic, 87 with putaminal, and 140 with lobar hemorrhages—were followed prospectively and examined on admission and at 2 weeks, 3 months, and 6 months after onset. The patients underwent clinical (according to the Glasgow Coma Scale) and neuroradiological examinations on admission and were scored clinically and functionally (according to Stroke Severity score and Barthel Index) on the follow-up periods. Risk factors and the correlation between findings on admission and the latest clinical and functional results were calculated with the χ2 test, Pearson correlation test, and Student’s t test. Multivariate analysis was calculated with the stepwise regression test.
Results In all of the bleeding locations, lethal outcome was significantly correlated with size of the hematoma (P<.001) and Glasgow Coma Scale score on admission (P<.001). Intraventricular blood expansion was found to have a better prognosis in thalamic bleeding (P<.007) and a worse prognosis in lobar hemorrhage (P<.01). The functional outcome after 6 months was directly correlated with the size of the bleeding area in lobar and putaminal hemorrhages. No correlation was found in thalamic bleeding. A worse functional outcome was found in putaminocapsular bleeding (P=.004) and in patients with ischemic heart disease. A limited better recovery prognosis was found in patients with lobar hematoma in the temporal lobe (P=.052).
Conclusions The probability of lethal outcome can be calculated on admission in all patients with supratentorial bleeding and in correlation with the location and size of the bleeding area and level of consciousness. Intraventricular expansion of blood is a better prognostic factor in thalamic bleeding and a worse one in lobar hematoma. Functional outcome is correlated with size of the bleeding area and level of consciousness on admission in putaminal and lobar hemorrhages but has no correlation to thalamic hemorrhage.
Intracerebral hemorrhage accounts for approximately 10% of all strokes,1 with a mortality rate of 15% to 60% reported in various studies.2 3 4 5 6 7 Among intracerebral hemorrhages, spontaneous supratentorial bleedings are the most common, with 51% to 55% of them in the putamen, 14% to 47% in the lobar area, and 10% to 15% in the thalamus.4 5 7 8 9 10 11 12
In evaluations of the outcome of intracerebral hemorrhages,2 3 4 5 6 level of consciousness and size of hematoma2 3 4 5 6 10 11 12 13 14 15 16 were found to be predictive outcome parameters. Most of the studies were conducted to determine the volume of blood that can absolutely predict a lethal outcome. Some publications reported approximately 10 mL5 in subcortical bleeding and 50 mL8 or 80 mL5 in lobar as critical. The roles of intraventricular blood extension, exact location of bleeding, age, and sex in the outcome of these patients are controversial.
The prognosis of the late functional outcome after supratentorial bleeding has not been examined extensively. Total or partial disability was found to be correlated with level of consciousness or with blood volume of the hematomas in lobar,5 8 15 thalamic,11 13 and putaminal12 locations. Nevertheless, the studies are few in number, and the conclusions are not clear.
The purpose of this study is to evaluate, compare, and correlate clinical and neuroradiological data to predict the outcome and functional recovery potential of patients with spontaneous supratentorial hemorrhages.
Subjects and Methods
Two hundred seventy-nine consecutive patients participated in a prospective consecutive study during 1986 to 1993. All patients were hospitalized in our neurological department 15 minutes to 4 hours (mean, 90±140 minutes) after the onset of the first symptoms. The patients included had supratentorial bleeding, were of both sexes, and were older than 20 years. Patients with recurrent intracerebral hemorrhage, previous stroke, or neuroradiological evidence of arteriovenous malformation, brain tumors, or predominant subarachnoid hemorrhage and patients receiving anticoagulant therapy were excluded from the study.
All patients were examined clinically and evaluated by two neurologists. The follow-up period of each patient observed in this study ended after 6 months or by a fatal outcome. Neuroradiological measurements were estimated by a neuroradiologist who was not aware of the clinical status of the patients.
The following clinical parameters were noted: vascular risk factors; hypertension (>170/90 mm Hg; diabetes mellitus (glucose level >140 mg/dL preprandial on two examinations, glucose level >200 mg/dL postprandial, or HbA1c >8.5%); and ischemic heart disease (proven myocardial infarction, existence of multiple lesions on thallium heart isotope screen, or evidence of coronary disease on coronary artery catheterization).
No patient underwent any neurosurgical procedure. All patients were treated similarly during hospitalization. Therapy was based on the preservation of essential life functions and the prevention of secondary medical complications. Intracranial pressure was treated with intravenous mannitol or furosemide.
Clinical evaluations of all patients were done according to the Glasgow Coma Scale (GCS)17 score on admission and after 48 hours, according to the Stroke Severity (SS) score,18 and according to the Barthel Index (BI)19 on admission and at 2 weeks, 3 months, and 6 months after the event. The SS score consisted of an examination of mental function (0 to 11 points); motor nerve function, including involuntary movements and cerebellar ataxia (0 to 11 points); cranial nerve function (0 to 11 points); sensation (0 to 6 points); and reflexes (0 to 5 points).
Location and volume of the hematoma were estimated by brain CT scan. The examination was made with the use of Elsint CT 2400 Elite with a matrix of 512. We calculated volume by measuring the areas of high absorption with a tracing cursor in each cut demonstrating bleeding and multiplying by the cut thickness. The results were expressed in cubic centimeters. The blood volumes of the bottom and top pools were additionally calculated as a special segment.4 5 The location of the hematoma in thalamic hemorrhages was classified as midthalamic, thalamocapsular, paraventricular, posterior thalamic, and holothalamic bleedings.20 The classifications of putaminal hemorrhage included pure putaminal, putaminocapsular, putaminothalamic, putaminohemispheric, and putaminocapsulohemispheric bleedings.14 Lobar hematomas were classified as frontal, parietal, temporal, parieto-occipital, and huge lobar bleedings. Extension of blood into the ventricles was noted.
The data were analyzed with the χ2 test, Pearson correlation test, and Student’s t test. Multivariate analysis was calculated with the stepwise regression test. No significant exceptions were found.
Fifty-two patients (28 men and 24 women) with thalamic bleeding were included in the study (mean age, 66.6±10.6 years). The bleeding locations were as follows: 36 holothalamic or diffuse bleeding, 11 thalamocapsular, 8 midthalamic, 2 posterior thalamic, and one purely paraventricular thalamic bleeding.
The mortality rate was 36.6% (19 patients) during the 6-month follow-up period (Table 1⇓). Causes of patient mortality were as follows: direct result of the bleeding (10 patients), rebleeding (1 patient), late complications (8 patients), secondary infection (6 patients), and myocardial infarction (2 patients).
Eighty-seven patients (60 men and 27 women) were included (mean age, 69.9±13.6 years). The bleeding locations were as follows: 26 pure putaminal, 24 putaminocapsulohemispheric, 20 putaminocapsular, and 7 putaminothalamic.
Among the patients with fatal outcome, the putaminal location was as follows: putaminocapsulohemispheric (15 patients), putaminohemispheric (7 patients), and putaminothalamic (5 patients). No lethal outcome was observed in pure and putaminothalamic bleedings. The mortality rate after 6 months was 31% (27 patients) (Table 1⇑). Causes of mortality were as follows: direct result of the bleeding (23 patients) and late complications or secondary infection with sepsis (4 patients).
One hundred forty patients (66 men and 74 women) with lobar hematomas were included in this study (mean age, 66.8±11.1 years). The locations were as follows: 49 frontal lobe, 42 parietal lobe, 17 temporal lobe, 9 parieto-occipital lobe, and 23 massive multilobar bleeding.
The mortality rate after 6 months was 52% (73 patients) (Table 1⇑). All patients with massive multilobar hematomas had a fatal outcome. The causes of mortality were a direct result of the bleeding (60 patients) and late complications (13 patients), including intercurrent infectious disease (5 patients), myocardial infarction (4 patients), pulmonary embolus (1 patient), and pulmonary edema (1 patient). In 2 cases the cause could not be determined. (No postmortem examination was performed.)
Age, Sex, and Risk Factors
In putaminal and thalamic bleedings, no significant correlation was found between age or sex and the BI or SS score (P=.29 to .45). In lobar hematomas, higher SS scores were found in men (P=.02). No sex or age dependence in the BI results was observed.
Risk factors found in this study were as follows: hypertension (165 patients; 59%), cigarette smoking (146 patients; 52%), diabetes mellitus (84 patients; 30%), ischemic heart disease (62 patients; 22%), and hyperlipidemia (47 patients; 18%).
No significant correlation was found between each of the vascular risk factors and the outcome. Hypertension and ischemic heart disease had a greater tendency toward a reciprocal correlation with poor outcome. Ischemic heart disease was found to have a limited reciprocal correlation with BI after 6 months only in patients with lobar hematomas (P=.053).
Neuroradiological and Clinical Examinations
In thalamic hemorrhage, the mean value of hemorrhage volume was 16.2±12.0 cm3 (range, 2.8 to 46 cm3) (Table 2⇓). In putaminal hemorrhage, the mean value of hemorrhage volume was 26.1±20.1 cm3 (range, 3.1 to 61 cm3). In lobar hematomas, the mean value of hemorrhage volume was 55.3±55.2 cm3 (range, 2.7 to 239 cm3) (frontal lobe, 36.6±22.8 cm3; parietal lobe, 44.9±35.5 cm3; temporal lobe, 16.2±13.2 cm3; and parieto-occipital lobe, 16.4±8.2 cm3). The GCS results for each of the bleeding locations are shown in Table 2⇓.
Lethal and Functional Outcome in the Three Groups
Lethal outcome. With the use of the stepwise logistic regression test, the probability of lethal outcome was calculated as follows (Fig 1⇓):
with radiological parameters (+1 in cases with diffuse holothalamic bleeding) and
with clinical parameters. The probability of survival was calculated as follows:
(+0.65 in cases with intraventricular hemorrhage; −0.65 in cases without intraventricular hemorrhage).
Significant correlation was found between low GCS score or size of hematoma and lethal outcome (P<.001 for each) and between intraventricular penetration and better prognosis for survival (P=.02), especially during the first 48 hours (P=.0007). The relative risks of probability of death in bleeding without intraventricular expansion versus bleeding with intraventricular expansion were 1.16 for blood volume of 7 cm3 and 2.73 for blood volume of 25 cm3.
Functional outcome. No correlation was found between GCS score, size of hematoma, or intraventricular expansion and SS score and BI in any stage of examination (P=.1 to .54).
Lethal outcome. The probability of lethal outcome was calculated as follows:
A significant correlation was found between low GCS score or size of hematoma and lethal outcome. A high significance for mortality was found in putaminocapsulothalamic bleeding. Intraventricular extension was not found to be a risk factor for poor outcome. The relative risks of probability of death, according to a GCS score of 5 versus 11, were 5.3 for a blood volume of 5 cm3 and 1.0 for a blood volume of 25 cm3.
Functional outcome. A significant correlation was found between size of bleeding and BI or SS score in putaminothalamic, putaminocapsular, and putaminocapsulothalamic locations (P=.04, P=.02, and P=.02, respectively). A poor functional outcome was found in putaminocapsular and putaminothalamic bleedings (P=.044 and P=.051, respectively). No correlation was found between the expansion of blood into the ventricle and SS score or BI.
Lethal outcome. The probability of lethal outcome was calculated as follows (Fig 2⇓):
Size of bleeding area (P<.0036), GCS score (P<.0001), and intraventricular extension were significantly correlated with lethal outcome. The relative risks of probability of death, according to a GCS score of 5 versus 11, were 4.2 for a blood volume of 30 cm3 and 1.7 for a blood volume of 100 cm3.
Functional outcome. Significant correlations were found between GCS score or size of bleeding on admission and SS score (P<.001 for each) or BI (P<.05 and P<.001, respectively). The correlation between intraventricular extension and SS score or BI was also significant (P<.01 for each). Temporal lobe bleeding was noted to have a somewhat better prognosis compared with other locations (P<.052) (Figs 3⇓ and 4⇓).
There have been numerous attempts to identify features in the first hours after intracerebral hemorrhages that can predict outcome. Studies of supratentorial intracerebral hemorrhage found that depressed level of consciousness, hematoma volume, and intraventricular extension of blood are predictive parameters.2 5 6 11 12 21 22 In our patients outcome was related to lobar, putaminal, or thalamic location of hemorrhage.
A direct correlation between the size of hematoma or depressed level of consciousness and outcome was found in all hemorrhage locations. Previous studies that examined thalamic,11 13 23 24 25 26 27 28 29 putaminal,12 and lobar hematomas8 9 14 15 demonstrated similar findings.
Massaro et al8 have suggested that depressed consciousness may be a marker of intraventricular extension in large lobar hematomas that relates to their outcome. However, the use of blood volume as a single parameter, which was reported in some of the publications, could not be reproduced in our study. Both parameters—size of hematoma and scoring of consciousness level—must be used together to predict outcome in each of the locations of supratentorial bleeding. In the present study we found a significant difference of predicting factors for late functional outcome among the bleeding location groups. Whereas in lobar and putaminal hematomas the functional and neurological recovery course can be predicted by hematoma size and clinical status on admission, in thalamic bleeding no clinical or radiological findings on admission were found to be significant parameters for the recovery prognosis. A significant role for intraventricular expansion in the recovery prognosis was demonstrated only in lobar hematomas. It seems that the evaluation of subcortical bleeding as a unique group may lead to errors and misinterpretations in results and conclusions. Therefore, thalamic hemorrhages must be analyzed separately from putaminal bleedings. The difference between some of the results of previous studies and ours can be explained by different methods of evaluation of the recovery status. While the previous studies used activities of daily living11 12 or the Rankin Scale,9 with a four- to five-stage recovery, including stages of vegetative state and death, or a categorization of patients according to conditions of ability or disability,22 we used gradient scores of the BI and the SS score to grade the functional status of the surviving patients.
The poor recovery prognosis of putaminal bleeding in the putaminocapsular region may be secondary to the anatomic location of this bleeding, which damages the corticospinal tract in the internal capsule. The higher level of significant recovery prognosis in temporal lobe hematoma can be explained by the fact that this region contains fewer anatomic structures, the deficits of which are represented in the BI. Possibly the BI is not the optimal evaluation method to be used in temporal lobe lesions; therefore, other methods of evaluation must be used.
The prognostic role of intraventricular extension4 21 30 31 32 33 34 35 36 37 is controversial. We demonstrated a clear difference in the outcome of intraventricular extension related to hemorrhage locations. In lobar hematomas we found a significant direct correlation, as have others.8 In the putamen there was no influence of intraventricular bleeding on the outcome, which has been infrequently noted.12 In thalamic hemorrhages the intraventricular extension predicts a significantly better outcome. Kwak et al11 found a tendency toward poorer prognosis in intraventricular hemorrhage without showing statistical significance.
The better prognosis of intraventricular extension of thalamic hemorrhage found in this study was not limited to small hematomas.
The lack of correlation between vascular risk factors and outcome that was demonstrated in our patients was previously noted,8 although it remains controversial.21 30 Tuhrim et al38 reported that mean arterial pressure is the main risk factor for a poor outcome in intracerebral hemorrhage. Weisberg et al14 proposed an opposite opinion, concluding that most small hemorrhages and good outcome were due to hypertension, while large hemorrhages with poor outcome were often due to causes other than hypertension.
In conclusion, a significant difference was demonstrated in early outcome as well as in late functional recovery potential between lobar, putaminal, and thalamic bleedings. The volume of the hematoma and the grading according to the GCS on admission can predict survival outcome in lobar and putaminal bleeding but not in thalamic hemorrhage. For patients with intraventricular hemorrhage, expansion of blood would be suggested as a marker for poor outcome in lobar hematomas and a better outcome in thalamic hemorrhage.
We would like to give special thanks to the statistical laboratories of the Mathematics and Statistical Departments of Tel Aviv University for the statistical analysis in this study and to Judy Brandt for her assistance with English editing and expert word processing.
- Received May 29, 1995.
- Revision received September 21, 1995.
- Accepted September 21, 1995.
- Copyright © 1995 by American Heart Association
Portenoy RK, Lipton RB, Berger AR, Lesser ML, Lantos G. Intracerebral hemorrhage: a model for the prediction of outcome. J Neurol Neurosurg Psychiatry. 1987;50:976-979.
Helweg-Larsen S, Sommer W, Strange P, Lester J, Boysen G. Prognosis for patients treated conservatively for spontaneous intracerebral hematomas. Stroke. 1984;15:1045-1048.
Weisberg LA. Computed tomography in intracranial hemorrhage. Arch Neurol. 1979;363:422-426.
Radberg JA, Olsson JE, Radberg CT. Prognostic parameters in spontaneous intracerebral hematomas with special reference to anticoagulant treatment. Stroke. 1991;22:571-576.
Brott T, Thalinger K, Hertzberg V. Hypertension as a risk factor for spontaneous intracerebral hemorrhage. Stroke. 1986;17:1978-1983.
Massaro AR, Sacco RL, Mohe JP, Fowlkes MA, Tatemichi TK, Price TR, Hier DB, Wolf PA. Clinical discriminators of lobar and deep hemorrhages: the stroke data bank. Neurology. 1991;41:1881-1885.
Kase CS, Williams JP, Wyatt DA, Mohr JP. Lobar intracerebral hematomas: clinical and CT analysis of 22 cases. Neurology. 1982;32:1146-1150.
Kwak R, Kadoya S, Suzuki T. Factors affecting the prognosis in thalamic hemorrhage. Stroke. 1983;14:493-500.
Waga S, Yamamoto Y. Hypertensive putaminal hemorrhage: treatment and results: is surgical treatment superior to conservative one? Stroke. 1983;14:480-485.
Walshe TM, Davis KD, Fisher CM. Thalamic hemorrhage: a computed tomographic-clinical correlation. Neurology. 1977;27:217-222.
McKissock W, Richardson A, Taylor J. Primary intracerebral haemorrhage. A controlled trial of surgical and conservative treatment in 180 unselected cases. Lancet. 1961;2:221-226.
Jannett B, Teasdale G. Management of head injuries. Philadelphia, Pa: FA Davis Co; 1981:77-84.
Weisberg LA. Thalamic hemorrhage: clinical CT correlations. Neurology. 1986;36:1382-1386.
Berlit P, Tornow K. Outcome of intracerebral hemorrhage: clinical and CT findings in 326 patiens. Eur J Neurol. 1994;1:299-234.
Barraquer-Bordas I, Illa I, Escartin A, Ruscalleda J, Martivilalta L. Thalamic hemorrhage. A study of 23 patients with diagnosis by computed tomography. Stroke. 1981;12:524-527.
Congia S, Cannas A, Tacconi P, Borghero G, Montaldo S, Porcella A. Thalamic hemorrhages: 30 cases studied: clinico-tomodensitometric correlations. Acta Neurol. 1992;14:22-28.
Kawahara N, Sato K, Muraki M, Tanaka K, Kaneko M, Uemura K. CT classification of small thalamic hemorrhages and their clinical implications. Neurology. 1986;36:165-172.
Douglas MA, Haerer AF. Long term prognosis of hypertensive intracerebral hemorrhage. Stroke. 1982;13:488-491.
Young WB, Lee KP, Pessin MS, Kwan ES, Rand WM, Caplan LP. Prognostic significance of ventricular blood in supratentorial hemorrhage: a volumetric study. Neurology. 1990;40:616-619.
Nakahara A, Nishimura T, Miura N, Kagama M, Kitamura K. A study of hypertensive intracerebral hemorrhage—responsible factors for surgical intervention of HICH in basal ganglia. Neurol Surg. 1978;6:647-655.
Stein RW, Caplan LR, Hier DB. Outcome of intracranial hemorrhage: role of blood pressure and location and size of lesions. Ann Neurol. 1983;14:132-133. Abstract.
Kameyama M. Vascular lesions of the thalamus—incidence and clinicopathologic features. Adv Neurol Sci. 1964;8:821-842.
Ross YBWEM, Hassan D, Vermeuleu M. Outcome in patients with large intraventricular haemorrhages: a volumetric study. J Neurol Neurosurg Psychiatry. 1995;58:622-624.