Hypertension

High blood pressure has consistently been reported as a major risk factor for ICH.^27–31 A meta-analysis and review by Areisen and colleagues estimated a crude odds ratio (OR) of 3.68 for hypertension and ICH in comparison to normotensive individuals.²⁴

Sturgeon and associates studied risk factors for ICH in a pooled cohort of the Atherosclerosis Risk in Communities study (ARIC) and the Cardiovascular Health Study (CHS).²¹ The ARIC cohort was recruited from 1987 to 1989 and included 15,792 men and women aged 45 to 64 years at baseline from four U.S. communities. The CHS cohort was recruited from 1989 to 1993 and included 5888 men and women 65 years or older at baseline from a sampling of four U.S. communities. Follow-up was in excess of 263,489 person-years. In this prospective study assessing baseline risk factors and subsequent occurrence of ICH, age, African American ethnicity (versus whites), and hypertension were positively associated with the development of ICH. Participants with systolic blood pressure of 160 mm Hg or greater or diastolic blood pressure of 110 mm Hg or greater had 5.55 (95% confidence interval [CI], 3.07 to 10.0) times the rate of ICH as nonhypertensive individuals.

Spontaneous ICH occurs predominantly in deep locations in the brain. The most common location is the putamen, followed by the subcortical white matter, cerebellum, and thalamus. In 100 unselected patients, Kase and associates found putaminal hemorrhage in 34, lobar hemorrhage in 24, thalamic hemorrhage in 20, cerebellar hemorrhage in 7, and pontine hemorrhage in 6.³²

Hemorrhages in the caudate nucleus, putamen, thalamus, brainstem, and cerebellum occur in the distribution of small perforating arteries with a diameter of 50 to 200 µm. These deep penetrating arteries are small nonbranching end arteries that arise directly from much larger arteries (e.g., middle cerebral artery, anterior choroidal artery, anterior cerebral artery, posterior cerebral artery, posterior communicating artery, cerebellar arteries, basilar artery). Their small size and proximal position predispose them to the development of microatheroma and lipohyalinosis. Electron microscopic studies suggest that most bleeding occurs at or near the bifurcation of affected arteries, where prominent degeneration of the media and smooth muscles can be seen.³³

The concept of ICH arising from rupture of miliary microaneurysms was first proposed by Charcot and Bouchard in 1868.³⁴ Studies by Russell³⁵ and by Cole and Yates³⁶ confirmed the occurrence of microaneurysms in an anatomic distribution closely correlated to that of hypertensive hemorrhages and further defined the epidemiology of microaneurysms. Russell identified a strong association between miliary aneurysms 300 to 900 µm in diameter and hypertension. A few aneurysms were observed in his control group of normotensive individuals (diastolic blood pressure <110 mm Hg), but 84% of this group were 60 years or older. Cole and Yates observed that microaneurysms 50 to 2500 µm in diameter were uncommon in those younger than 50 years, even in hypertensive subjects. Therefore, both hypertension and age appear to be major factors in the formation of microaneurysms.

Fisher has further defined the pathologic process affecting small cerebral arteries in hypertension and coined the term lipohyalinosis to specify a destructive vascular process previously referred to by a variety of names, including “fibrinoid necrosis,” “angionecrosis,” and “hyaline arterionecrosis.”^37–39 In Fisher’s view, raised arterial pressure alters the walls of small cerebral arterioles 80 to 300 µm in diameter and leads to focal subintimal fibrinoid deposition associated with the presence of fat-filled macrophages. As the process advances, the integrity of the elastica and media is lost, and the artery dilates locally to form a microaneurysm 500 to 1500 µm in diameter. Extravasation of red blood cells takes place through the damaged walls, and hemosiderin-filled macrophages are seen through and beyond the adventitia. Fisher found lipohyalinosis to be, by virtue of occlusion of the vascular lumen, the cause of many lacunar infarcts. He could not confirm that microaneurysms were the source of massive ICH.

Cerebral Amyloid Angiopathy

β-Amyloid protein is a peptide consisting of 39 to 43 amino acids. It is formed after cleavage of the amyloid precursor protein, a transmembranous protein of unknown function, by β- and γ-secretases.

The characteristic feature of cerebral amyloid angiopathy (CAA) is the deposition of β-amyloid protein in the media and adventitia of small cortical and leptomeningeal arteries.⁴⁰ The incidence of CAA rises steeply after the age of 70 and reportedly ranges from 23% to 48% in the 8th decade, from 37% to 46% in the 9th decade, and from 57% to 58% in the 10th decade.⁴⁰ The most significant feature of CAA is the presence of multiple hemorrhages in unusual locations in the absence of hypertension.⁴¹ Vice versa, hypertension is present in only a minority of patients. Hematomas are mostly subcortical or lobar and are more frequently found in the occipital and parietal lobes. A distinct feature, in sharp contrast to nonamyloid ICH, is the multiplicity of hemorrhages over time and location. As for amyloid angiopathy in other locations in the body, the involved vessels exhibit Congo red birefringence in polarized light.⁴¹ Features of vessels with CAA include a severe degree of amyloid deposition and coexistent fibrinoid necrosis.⁴² Apolipoprotein E ε3/4 and ε4/4 seem to be associated with more severe forms of CAA, and severe degrees of CAA were associated with ICH. The annual risk for recurrent hemorrhage was found to be 10.5%.¹⁶

Anticoagulant Therapy

The incidence of ICH in patients taking warfarin after myocardial infarction is 1% per year.⁴³ Sixty-one percent of ICHs occur in the first 6 months of anticoagulation therapy.⁴⁴ Long-term anticoagulation therapy increases the risk for ICH 8- to 11-fold.⁴⁵

The mortality rate of spontaneous ICH is as high as 67% in patients receiving oral anticoagulant therapy (OAT).^45–47 The incidence of OAT-related ICH (OAT-ICH) is expected to increase in the coming years as a result of an anticipated rise in the incidence of atrial fibrillation attributable to an aging population.

A number of factors contribute to the increased risk for ICH in this group of patients, including advanced age, previous cerebrovascular disease, hypertension, and concomitant use of aspirin.^32,43 In the Comparison of Warfarin and Aspirin for Symptomatic Intracranial Arterial Stenosis (WASID) trial,⁴⁷ which compared the effectiveness of aspirin (1300 mg/day) and warfarin (target international normalized ratio, 2.0 to 3.0) in preventing strokes in patients with transient ischemic attack or stroke caused by angiographically verified 50% to 99% stenosis of a major intracranial artery, major hemorrhage occurred in 3.2% of the aspirin group and 8.3% of the warfarin group. The study was terminated early because of a statistically significant higher rate of bleeding in the warfarin group and no statistically significant benefit in the primary end point of the study. In a review of eight placebo-controlled clinical trials for the prevention of stroke, Mayo and colleagues found that the risk for hemorrhagic stroke was 0.7% in 2981 patients treated with aspirin versus 0.37% in 2187 patients receiving placebo.⁴⁸

ICH during treatment with heparin is rare and occurs mostly in patients being treated for acute embolic cerebral infarction and uncontrolled hypertension. In most patients, the activated partial thromboplastin time is excessively prolonged.^49,50

The efficacy of fibrinolytic agents in the treatment of myocardial infarction is well known. ICH has been reported in 0.4% to 1.3% of patients with acute myocardial infarction treated with the single-chain tissue plasminogen activator (t-PA) alteplase.⁵¹ Thrombolysis in acute ischemic stroke increases the risk for severe, life-threatening hemorrhagic complications by up to 10-fold in comparison to controls. Intravenous t-PA was used in two studies, the European Cooperative Acute Stroke Study (ECASS)⁵² and the National Institute of Neurological Disorders and Stroke rt-PA Stroke Study,⁵³ with a therapeutic window of 6 and 3 hours, respectively. The National Institute of Neurological Disorders and Stroke rt-PA Stroke Study Group observed 2 patients (0.6%) with symptomatic and 1 patient (0.3%) with fatal hemorrhage in the placebo group (n = 312) and 20 patients (6.4%) with symptomatic and 9 patients (2.9%) with fatal hemorrhage in the recombinant tissue plasminogen activator (rt-PA) group (n = 312).⁵³ In both ECASS 1 and 2, rt-PA increased the risk for parenchymal hematoma (OR, 3.0 and 4.2).^52,54 However, the functional outcome at 3 months was better in the t-PA–treated group. Experimental focal cerebral ischemia causes a significant loss of the basal lamina components of cerebral microvessels.⁵⁵ The mechanisms for this microvascular damage may include degradation of plasmin-generated laminin, activation of matrix metalloproteinases, transmigration of leukocytes through the vessel wall, and other processes. This loss in vessel wall integrity is associated with the development of petechial hemorrhage. Revascularization, coupled with hypertension, may lead to life-threatening hemorrhagic complications.

The pathophysiology of ICH in patients maintained on a regimen of anticoagulation therapy is not well known, but various factors have been hypothesized. Hart and colleagues theorized that the use of OAT merely unmasks intracerebral bleeding that would otherwise remain asymptomatic, especially in patients with underlying hypertension or cerebrovascular disease.⁵⁶ This hypothesis is supported by the fact that gradient echo MRI indicates that microbleeding can be found even in neurologically normal individuals and is strongly associated with increased age and hypertension.⁵⁷ The Stroke Prevention in Reversible Ischemia Trial (SPIRIT) and the European Atrial Fibrillation Trial (EAFT) indicated that patients with primary underlying cerebrovascular disease had a remarkably higher risk for OAT-ICH.^58–60 Furthermore, the presence of white matter lesions, so-called leukoaraiosis, is an independent predictor of spontaneous ICH.⁶¹ Therefore, the underlying mechanism of spontaneous ICH and OAT-ICH may be the same, with OAT acting as an exacerbating factor. This may also explain why the distribution of locations in the brain where OAT-ICH occurs is no different from that seen in patients with spontaneous ICH.^46,47,62

OAT may also cause ICH directly. This is supported by the observation that higher intensities of anticoagulation clearly increase the risk for OAT-ICH.^{46,47,63–65} Oral anticoagulants interfere with the synthesis of vitamin K–dependent clotting factors, thereby resulting in low levels of factors VII, IX, and X and prothrombin. It is possible that adequate levels and functional forms of these clotting factors are essential to counteract the stress placed on blood vessels as part of normal daily activities and to prevent bleeding.^66,67

The dynamics of hematoma expansion in OAT-ICH remain to be firmly elucidated. Because of persistent coagulopathy, hematoma expansion in OAT-ICH may be more common and occur over a longer time frame than seen with spontaneous ICH. In a retrospective study of 47 patients with OAT-ICH, hematoma expansion was found in 28% of those evaluated within 24 hours of onset.⁶⁸ In another study, hematoma expansion up to the seventh day was found in 16% (9/57) of patients who were not receiving OAT versus 54% (7/13) of those who were.⁶⁹

Although the evolution of hematomas in patients managed with anticoagulation therapy is protracted, their hematomas are about twice the size of those in patients not receiving anticoagulation therapy, and their mortality rate increases to 60% to 65%.⁴³

Drug Abuse, Alcohol, and Smoking

Secondary Deterioration

A third to a fourth of the patients who are initially alert deteriorate in their level of consciousness within the first 24 hours.^93,94 Rather than the initial clinical signs, the most accurate predictor is large hematoma volume and ventricular extension.⁹⁴ Consequently, these patients need to be very closely monitored during the first 24 hours, especially if they are still awake and alert despite large hematoma volume. Expansion of the hematoma is the most common cause of neurological deterioration and death^93,94 within the first 3 hours after the ictus. Progression of the mass effect secondary to edema can essentially occur within two distinct time intervals: early within the first 2 days and late within the second and third weeks.⁹⁵

Baseline laboratory studies should essentially detect coagulopathies so that they can be treated accordingly. Therefore, these studies include a complete blood count, prothrombin time, partial thromboplastin time, and liver and renal function tests, as well as serum glucose, electrocardiography, and chest radiography.

The initial radiologic work-up consists of a plain CT scan of the brain. It is fast and easy to interpret. The location and size of the hematoma and the presence or absence of a mass effect, perilesional edema, and ventricular extension can be depicted (i.e., all the imaging criteria that are important in the acute phase to decide on a management plan). Contrast enhancement can offer the possibility of diagnosing associated abnormalities such as AVMs, aneurysms, and tumors, but this depends on the size of the masking blood clot. Blood on a CT scan is usually seen as hyperdense signal within 4 hours. Severely hemodiluted or anemic patients with hematocrits below 30, however, can show isodense to hypodense lesions. In patients with acute bleeding caused by ICH, MRI would not provide more useful information and is in fact frequently a source of confusion because of the variable interpretation of a blood clot producing signal inhomogeneities. After about 7 to 10 days, the high attenuation values of the hematoma start decreasing peripherally. Depending on its size, the entire hematoma may become isodense in 2 to 3 weeks and resolve completely to an area of decreased density within 2 months. Occasionally, a resolving hematoma may give the appearance of ring enhancement because of either increased vascularity at the periphery or disruption of the blood-brain barrier.

MRI, although superior in resolution and anatomic and functional detail, as already alluded to, is rarely instrumental in the acute evaluation of patients with ICH. The presence of blood and its confusing signal characteristics can mask an underlying structural lesion. MRI, however, often reveals associated lesions that may not be detected with CT. Thus, MRI in the acute setting should be reserved for highly atypical cases in which the clinical findings or the CT studies are unusual.

Putaminal Hemorrhage

Putaminal hemorrhage is the most common form of ICH and can be manifested in different ways, depending on the size and extent of the hematoma. Initially, there is an acute onset of headaches. However, these headaches are different from the thunderclap type noted in patients with SAH. They are soon followed by a gradual progression of focal neurological signs and, depending on the overall size, by a worsening level of consciousness. A marked deficit from the beginning is unusual. Common neurological findings are hemiparesis, hemisensory syndrome, homonymous hemianopia, horizontal gaze palsy, and either aphasia (dominant hemisphere) or hemineglect (nondominant hemisphere).^37,96 The pronounced neurological deficits associated with coma suggest large hematomas and carry a poor prognosis. Intraventricular extension of the hemorrhage is a sign of extensive parenchymal dissection or destruction and a bad prognosticator (Fig. 358-1).^97,98 In contrast, patients who are alert and have only limited motor deficits, normal extraocular movements, full visual fields, and a hematoma that does not extend laterally or upward out of the putaminal region fare better. This has been ascribed to reversible compression of capsular fibers as opposed to destruction.⁹⁷

FIGURE 358-1 Extension of a primarily putaminal hemorrhage into the rest of the basal ganglia, internal capsule, thalamus, and ventricle was evident at autopsy.

Thalamic Hemorrhage

Thalamic hemorrhages account for 10% to 15% of all ICHs (Fig. 358-2).^37,99 The bleeding originates from thalamic perforators of the posterior cerebral arteries and may extend laterally into the internal capsule, medially into the ventricles, superiorly into the corona radiata, and inferiorly into subthalamus and midbrain.¹⁰⁰

FIGURE 358-2 This medium-sized hypertensive thalamic hemorrhage, which spared the capsular fibers laterally, is well confined.

The signs and symptoms depend on the size and pattern of extension of the hematoma. In contrast to putaminal hemorrhages, thalamic bleeding will instantly result in gross neurological deficits with sensorimotor loss, a higher likelihood of vomiting, variable presence of headaches, and occasionally coma.^6,99,101 The ocular findings in patients with thalamic lesions are pathognomonic: upward gaze palsy, convergence gaze, miotic unreactive pupils because of compression of the midbrain tectum, and less commonly, retraction nystagmus on upward gaze and skew deviation (vertical misalignment of the eyes because of abnormal prenuclear vestibular input to the ocular motor nuclei).^6,37,99,101

Different characteristics of the so-called thalamic syndrome had already been described at the beginning of the 19th century by Dejerine and Roussy,¹⁰² followed by Lhermitte in 1925¹⁰³ and Baudouin and associates in 1930.¹⁰⁴ Fisher in 1959 emphasized language disorders and disturbances in ocular motility.¹⁰⁵

Kumral and colleagues presented an excellent study from 1995 based on 100 patients and provided a very detailed correlation of different symptoms (sensorimotor, oculomotor, and neurobehavioral) to four defined topographic types of thalamic hemorrhage¹⁰⁶:

Lobar Hemorrhage

Lobar cerebral hemorrhages constitute a separate group of supratentorial hemorrhages with heterogeneous etiology and include AVMs, trauma, anticoagulation therapy, and in the elderly, amyloidosis of the cerebral vessels.

Lobar hemorrhages usually occur in the subcortical white matter and have a predilection for the parietal, temporal, and occipital lobes.^107,108 In Ropper and Davis’ series of 26 patients with lobar hemorrhages, 11 (42%) were within the occipital lobe, 7 (27%) were in the temporal lobe, 4 (15%) were in the frontal lobe (Fig. 358-3), and 3 (12%) were in the parietal lobe.¹⁰⁸ The frequent occurrence of lobar hemorrhages in the parieto-occipital lobes has been attributed to the higher concentration of intracerebral microaneurysms reported in anatomic studies.¹⁰⁹

FIGURE 358-3 An extensive medial frontal intracerebral hematoma secondary to amyloid angiopathy caused this patient’s death.

Hypertension as a cause of lobar hemorrhage is unusual.^107,108,110 Only 31% of the patients reported by Ropper and Davis had chronic hypertension.¹⁰⁸ Kase and associates reported elevated blood pressure in only 50% of their patients on admission.¹⁰⁷ In a series reported by Broderick and colleagues, hypertension contributed almost equally to lobar and deep hemispheric, cerebellar, and pontine hemorrhages.¹⁹ Other causes of lobar hemorrhages are AVMs, tumors, anticoagulation therapy, blood dyscrasias, and CAA.^41,111,112 In a significant number of cases, no definite cause can be found.¹⁰⁷ CAA is probably the most common cause in nonhypertensive patients 70 years and older.

The clinical manifestations of lobar ICH depend on the location and size of the hematoma.¹⁰⁸ When compared with other forms of ICH, the frequency of associated hypertension and coma on admission is lower. The low incidence of coma is probably related to the peripheral location of the hematoma.¹⁰⁸ Most patients complain of headache and vomiting. Seizures are also frequent.^22,107,113 Hemiparesis is seldom pronounced.

The prognosis of patients with lobar ICH is relatively better than that of patients with other forms of ICH. Mortality rates range from 11% to 29%.^107,108,114 Functional outcome in survivors also tends to be better.¹¹⁴ In their series of 22 patients, Kase and coauthors reported good outcomes in those with hematoma volumes of less than 20 cm³.¹⁰⁷ Seventy percent survived after the surgical removal of hematomas that were 20 to 60 cm³. No patient with a hematoma volume greater than 60 cm³ survived.

Caudate Hemorrhage

Caudate hematomas represent approximately 5% to 7% of cases of ICH.⁹⁸ The most common cause of caudate hemorrhage has been arterial hypertension.^98,115–119

The head of the caudate nucleus receives its blood supply from Heubner’s artery and the anterior lenticulostriate and lateral lenticulostriate arteries, which also supply the anterior internal capsule and putamen.¹¹⁹ A rupture in these arteries causes parenchymal hemorrhage.

Patients have an abrupt headache and vomiting, followed by a decreased level of consciousness. They are usually disoriented, with evidence of neck stiffness.⁹⁸ Occasional patients suffer seizures and exhibit horizontal gaze paresis. On CT, ventricular extension of the hematoma into the frontal horn with secondary hydrocephalus is common. Occasionally, the hemorrhage extends into the anterior portion of the thalamus. Such patients have transient, but significant short-term memory deficits.⁹⁷ In Stein and colleagues’ series of 8 patients, most recovered fully with no significant neurological deficits.⁹⁸ Weisberg also reported a small series of caudate hemorrhages, but the outcome in this series was poor, and stupor and a massive amount of intraventricular hemorrhage were associated with poor outcome.¹¹⁵ Liliang and coworkers looked at clinical data from 36 consecutive patients with hypertensive caudate hemorrhage.¹²⁰ In this relatively large study, multivariate analysis and stepwise logistic regression revealed that hydrocephalus was the only independent prognostic factor for a poor outcome (P < .001).

Cerebellar Hemorrhage

The frequency of cerebellar hemorrhage ranges from 5% to 10%.^{37,49,121,122} One of the major differences from supratentorial hemorrhages is the entirely different prognosis once coma has occurred. If the diagnosis is made early and surgical intervention is prompt, coma is reversible after a cerebellar hemorrhage.^{109,123–126} A very recent study by Smajlović and colleagues on the 30-day prognosis and risk factors in 352 patients treated for ICH confirmed once again that when compared with all other ICHs, cerebellar hematomas had the best outcome after the first month; brainstem and multilobar ICHs had the worst.⁸⁷ The dentate nuclei are the most common substrate. The hematoma extends into the hemispheric white matter and often into the fourth ventricle, where it causes either brainstem compression or direct invasion (Fig. 358-4). Rarely, cerebellar hemorrhage involves only the vermis. Hypertension and anticoagulation are the two most important causative factors for cerebellar hemorrhage.^123,127

FIGURE 358-4 This extensive hypertensive cerebellar hemisphere hematoma started in the deep nuclei and extended. Brainstem compression, hydrocephalus, and death ensued.

Patients usually have headache and an inability to walk or stand. Vomiting is common and may or may not be associated with headache.^123,127 Other symptoms include dizziness, neck stiffness, dysarthria, tinnitus, and singultus. Loss of consciousness at the onset is rare. On admission to the hospital, about a third of patients are obtunded.¹²³ The early physical signs are appendicular or truncal ataxia, dysarthria, ipsilateral horizontal gaze palsy, peripheral facial palsy, nystagmus, and sixth nerve palsy. At least two of the three characteristic clinical signs—appendicular ataxia, ipsilateral gaze palsy, and peripheral facial palsy—were present in 73% of the patients reported by Ott and coworkers.¹²³ In two patients with cerebellar hematomas and a peripheral facial nerve palsy on the hematoma side, Messert observed spontaneous unilateral eye closure on the contralateral side. In an effort to compensate for gaze dissociations and extraocular motor palsies, the eye on the noninvolved side of the face was closed. In other words, the open eye is on the hematoma side.¹²⁸

The clinical course of cerebellar hemorrhage is unpredictable. These patients, whether alert or lethargic on admission, can deteriorate quickly to coma and die with no warning.^123,129 Although the prognosis and final outcome are largely related to the patient’s initial preoperative condition, even comatose patients can make a good recovery.¹²⁹ Little and colleagues reported on two groups of patients with cerebellar hemorrhage.¹³⁰ The first group had an abrupt onset, progressive course, and low level of consciousness. CT in this group of patients, who required surgery, showed cerebellar hematomas 3 cm or greater in diameter, obstructive hydrocephalus, and extension of hemorrhage into the fourth ventricle.^131,132 The second group of patients was awake and stable and had hematomas smaller than 3 cm in diameter. They were treated medically, with good outcomes.¹³⁰

Brainstem Hemorrhage

The pons is the most common location for nonvascular causes of ICH in the brainstem. Spontaneous nontraumatic midbrain and medullary hematomas are comparably rare. The first to evaluate a large number of pontine hemorrhages was Attwater in 1911.¹³³ After performing autopsies on 77 subjects with pontine hemorrhages, Attwater was able to differentiate between primary and secondary brainstem hemorrhages. He attributed some pontine hemorrhages to elevated ICP. Several years later in a monograph, Duret also described this phenomenon, which now bears his name as an eponym.¹³⁴ Under the so-called Duret hemorrhage, we understand a characteristic slit-like bleeding in the upper brainstem (mesencephalon and pons).¹³⁵ It needs to be mentioned, however, that this does not automatically imply spontaneous ICH as a cause. As Parizel and colleagues pointed out, they also occur in victims of craniocerebral trauma.¹³⁵

The location of a pontine hematoma depends on the causative perforating artery, which can be the basilar paramedian perforators or the short or long circumferential branches. Thus, these hematomas can be divided into paramedian, basal pontine, and lateral tegmental hemorrhages.

In an autopsy review of 30 patients with pontine hemorrhages among 511 cases of ICH at Boston City Hospital, two thirds of the patients were comatose on initial evaluation and had massive hemorrhages that extended into the midbrain or fourth ventricle. Within 48 hours, 78% of the patients died. Fisher suggested that primary hemorrhage, by virtue of a pressure effect, causes the surrounding vessels to rupture and initiates a cascade of gradual enlargement of the hematoma.¹³⁶ The bleeding in hypertensive patients was attributed to leakage from tiny penetrating vessels damaged by lipohyalinosis and containing small microaneurysms.^49,109,136

Rupture of the paramedian perforating branches of the basilar artery is thought to be the cause of massive pontine hematomas (Fig. 358-5). The lesion usually begins in the midpons at the junction of the tegmentum and basis pontis and extends along the longitudinal axis of the brainstem into the midbrain, middle cerebellar peduncle, or the fourth ventricle.¹²¹

FIGURE 358-5 Large hypertensive pontine hematoma expanding and distorting the pons.

The clinical course of a hypertensive patient is typically one of rapid onset of coma. Awake patients may become symptomatic with headache, vomiting, and focal pontine signs such as facial or limb numbness, deafness, diplopia, quadriparesis, paraparesis, or hemiparesis. Occasionally, seizure (which can be a true convulsive episode), spasmodic decerebrate posturing, or violent shivering associated with autonomic dysfunction and rapidly developing hypothermia is reported. On examination, patients have an abnormal breathing pattern, apnea, cranial nerve and long-tract deficits, occasional decerebrate posturing, and multiple oculomotor findings.^121,137,138 Weakness of the pontine and bulbar musculature is invariably associated with large median pontine hemorrhages, but it is seldom appreciated because of the depressed level of consciousness.

Pinpoint (miotic) reactive pupils are a hallmark of pontine lesions. Among the possible various ophthalmic findings in patients with pontine hemorrhage are absent horizontal gaze movement when the paramedian pontine reticular formation is bilaterally damaged, one-and-a-half syndrome after a unilateral pontine tegmental or abducens nucleus lesion that results in a horizontal conjugate gaze palsy in one direction and internuclear ophthalmoplegia in the other,¹³⁹ and ocular bobbing. Fisher was the first to describe brisk, conjugate downward eye movements, followed by a slower “bobbing” back up to the primary position.¹⁴⁰

Massive pontine hemorrhages are always fatal, but death may not be instantaneous.¹³⁷ Some patients with medium-sized hematomas and most patients with small basal or lateral tegmental hematomas survive, but with various degrees of residual neurological deficits.

Brain Edema after Intracerebral Hemorrhage

One of the major challenges after primary ICH is the development of perihematomal edema, which forms rapidly after the ICH¹⁴¹ and contributes to a documented increase in perihematomal volume by approximately 75%. The causes of this perihematomal area of edema and cell death are not known decisively. Experimental data indicate that the presence of whole blood, but not intact red blood cells, induces the formation of edema. As red blood cells lyse, however, edema is observed, and the volume of the edema correlates with the volume of the lysed red blood cells. Hemoglobin and its degradation products induce edema formation and accumulation of reactive glial cells at the site of delivery.¹⁴² There are three distinct phases of edema formation after ICH. In the first hours after ICH, retraction of the clot begins. Intact red blood cells within the hematoma area have not been found to contribute to edema formation. As the coagulation cascade becomes activated over the following 24 to 48 hours, however, thrombin becomes activated and promotes edema formation and further disruption of the integrity of the blood-brain barrier.¹⁴³ The third phase of edema formation starts when red blood cells in the hematoma begin to lyse and hemoglobin and its degradation products are deposited into the brain parenchyma, thus initiating a potent inflammatory reaction.¹⁴⁴

In an observational study, Wu and colleagues studied 17 patients with spontaneous ICH treated medically.¹⁴⁵ Hematoma size and absolute and relative brain edema volumes were measured. Hematoma enlargement occurred in 4 of the 17 ICH patients (24%) within the first 24 hours. Hematoma sizes were reduced significantly at day 10 (P < .05) because of clot lysis. However, both absolute and relative brain edema increased gradually with time (P < .01). These results suggest that delayed brain edema after ICH may result from hematoma lysis. The authors concluded that reducing early hematoma growth and limiting clot lysis–induced brain toxicity could be potential therapies for ICH.

Steroids

The use of steroids in patients with ICH is controversial. Batjer and associates used steroids (4 mg dexamethasone intravenously every 6 hours) in their protocol.¹⁴⁸ The rationale for using steroids for the treatment of ICH is that they might lessen the damaging effects of cerebral edema, increased ICP, a disrupted blood-brain barrier, and stress. The first randomized study included 40 patients with ICH but showed no statistical difference in outcome associated with the use of steroids.¹⁴⁹ This study, however, lacked case uniformity and appropriate, relevant stratification. Therefore, a well-designed, randomized, placebo-controlled study was performed in 1987.¹⁵⁰ In their paper, Poungvarin and coworkers studied 93 patients 40 to 80 years old in a double-blind randomized design. Patients with documented primary supratentorial ICHs were randomly assigned to either dexamethasone (10 mg intravenously and then 5 mg every 6 hours) or placebo. The death rate at the 21st day was identical in both groups (dexamethasone versus placebo, 21 of 46 versus 21 of 47; P = .93). In contrast, the rate of complications (mostly infections and complications of diabetes) was 11 times higher in the dexamethasone group (P < .001). This led to early termination of the study.

Blood Pressure Management

The single most important factor in determining rapid expansion of an ICH is blood pressure. In studies of patients with hypertensive ICH, persistently elevated blood pressure increased the risk for hematoma progression.^151–153 In a retrospective review of 320 patients with hypertensive ICH, 10 showed rapid expansion of the hematoma on serial CT.¹⁵³ The consecutive scans were obtained an average of 1.7 and 48.9 hours after hemorrhage. Of the 10 patients with radiographic evidence of expansion, all had persistent hypertension, and half deteriorated neurologically. The average blood pressure in this group on admission was 179/110 mm Hg, and the average blood pressure recorded before deterioration was 190/121 mm Hg. The first 24 hours seems to be particularly critical. In a more recent retrospective study of 76 consecutive patients with hypertensive ICH, Ohwaki and colleagues observed that maximum systolic blood pressure was significantly associated with hematoma enlargement (P = .0074).¹⁵⁴ A target systolic blood pressure of 160 mm Hg or greater was significantly associated with hematoma enlargement when compared with a pressure of 150 mm Hg or less (P = .025).

The degree to which blood pressure should be controlled is controversial. Patients with a history of chronic hypertension have impaired autoregulation, and overzealous lowering of blood pressure can lower cerebral perfusion pressure and produce secondary ischemic damage, especially in those with a decreased level of consciousness, who may have elevated ICP. Some authors recommend lowering systolic blood pressure to less than 160 mm Hg¹⁵⁵; others recommend lowering it to normotensive levels but not below.¹⁵⁶ In a prospective, randomized trial of putaminal ICH in which craniotomy was compared with medical therapy, patients were initially treated with sodium nitroprusside (Nipride) to decrease systolic blood pressure by 25% during the first 24 hours. During the next 48 to 72 hours, blood pressure decreased to normotensive levels.¹⁴⁸ In this study, the mean admission systolic blood pressure was 234 mm Hg. Despite the tight control of blood pressure, the 6-month mortality rate was 77%. Because this study did not report the cause of death or the time of deterioration, it is difficult to determine whether the degree of blood pressure control was adequate.

In theory, there might be at least two conflicting trends in the immediate hours after ICH, namely, the development of perilesional microcirculatory insufficiency with a propensity to render the brain ischemic and, on the other hand, a propensity to rehemorrhage. The former would necessitate increased perfusion and the latter strict blood pressure control. Qureshi and colleagues looked at the rate of 24-hour blood pressure decline and mortality after spontaneous ICH.¹⁵⁷ One hundred five patients with ICH were included in this retrospective study. Logistic regression analysis showed that the rate of decline in blood pressure in the first 24 hours was an independent predictor of mortality but did not affect the functional outcome of survivors. More recently, The Antihypertensive Treatment of Acute Cerebral Hemorrhage (ATACH) trial set out to answer the question about the optimal blood pressure range in the first 24 hours after ICH. This trial is a multicenter open-labeled pilot trial to determine the tolerability and safety of three escalating goals of antihypertensive therapy (110 to 140 mm Hg, 140 to 170 mm Hg, 170 to 210 mm Hg) for acute hypertension in 60 subjects with supratentorial ICH. The initial results of this trial were presented at the 2008 International Stroke Conference¹⁵⁸: “Aggressive systolic blood pressure reduction to 110-140 mmHg in the first 24 hours using intravenous nicardipine was well tolerated with a low risk of hematoma expansion, neurological deterioration and in-hospital mortality. The results favor pharmacological reduction of systolic blood pressure in patients with acute ICH.”

Experimental rat studies suggest that although transient alterations in blood flow occur within minutes of hemorrhage, the severe alterations in perihematomal microcirculation that cause ischemia are not maximal until 4 hours thereafter.^159,160 In addition, the formation of edema is not maximal until after 6 to 8 hours.¹⁶¹ However, the period associated with the maximal risk for progression of hemorrhage in the presence of persistent hypertension is 3 to 6 hours.^{23,153,162,163} Therefore, an argument can be made to reduce blood pressure dramatically during the first 4 hours after hemorrhage to decrease the risk for rehemorrhage and then to raise blood pressure slowly to perfuse ischemic areas.

The discussion whether an ischemic penumbra in ICH exists is ongoing. At present, a majority of researchers seem to favor that there is no ischemic penumbra, but rather a perilesional area with depressed metabolism (metabolic penumbra?) that can be detected with positron emission tomography,¹⁶⁴ an area of reduced metabolic demand as detected by diffusion- and perfusion-weighted MRI,¹⁶⁵ or only a perilesional area with no significant changes at all. This was suggested after sampling in eight dogs and serially measuring regional cerebral blood flow (CBF) with radiolabeled microspheres, cerebral oxygen extraction, the cerebral metabolic rate of oxygen consumption, glucose utilization, and lactate production.¹⁶⁶

The last entry to date demonstrated vasogenic edema and only a mild perfusion deficit above the threshold for ischemia in an MRI rat ICH model, thus making a perihematomal penumbra unlikely.¹⁶⁷ A prospective clinical study with perfusion-weighted MRI performed during the treatment of 18 ICH patients, however, confirmed the presence of a hypoperfused area around the ICH that disappeared completely after 1 week.¹⁶⁸ Similar results were obtained from a single-photon emission CT study that compared early uptake within hours of hemorrhage and 6 to 9 months postictally.¹⁶⁹ An interesting microdialysis study in 2006 revealed that the immediate zone around an evacuated ICH exhibits a similar biochemical pattern as the penumbra zone surrounding focal traumatic brain contusions.¹⁷⁰

Intracranial Pressure Management

Elevated ICP is considered a major contributor to mortality after ICH and should be treated vigorously.¹⁷¹ ICP may be managed with osmotherapy, controlled hyperventilation, surgery, and other measures. A therapeutic goal for all treatment of elevated ICP is an ICP of less than 20 mm Hg and cerebral perfusion pressure greater than 70 mm Hg. Mannitol effectively and safely decreases ICP and can be used alone or with urea to potentiate its effect. In a study involving ICP-monitored patients, aggressive medical treatment with urea, mannitol, or both adequately controlled ICP and was associated with better outcomes and lower mortality rates.¹⁷² The beneficial effect of sustained hyperventilation on ICP is unresolved and has not been examined systematically in patients with this condition. In theory, reduction of ICP by hyperventilation ceases when the pH of cerebrospinal fluid reaches equilibrium. If hyperventilation is instituted for elevated ICP, PCO₂ should be maintained between 30 and 35 mm Hg. Induced barbiturate coma and therapeutic hypothermia are considered last ditch options and not part of a standardized algorithm for the treatment of elevated ICP in patients with ICH. Short-acting barbiturates such as thiopental have been shown to effectively reduce elevated ICP.^173,174 The effect is presumably mediated through reduction of CBF and cerebral blood volume. In addition to reducing the volume of the normal brain, barbiturates reduce brain swelling and may act as free radical scavengers.^1,2 Moderate hypothermia (32°C to 35°C) may be neuroprotective and decreases ICP by lowering cerebral blood volume (by lowering metabolic demand).¹⁷⁵

Seizures after Intracerebral Hemorrhage

Seizure activity can result in neuronal injury and worsening of an already critically ill patient and must be treated aggressively. The incidence of immediate seizures after supratentorial parenchymal ICH in published series ranges from 1.4% to 17%.^176–183

In a prospective, observational study of 761 consecutive patients with nontraumatic, nonaneurysmal ICH, Passero and associates found the cumulative actuarial risk of experiencing seizures to be 7.2% in the first 5 days and 8.1% within 30 days.¹⁸³ A lobar location of hemorrhage was also found to be an independent predictor of early seizures, as was previously recognized.^{176,178–180} Prophylactic antiepileptic therapy may be considered for 1 month. It should be tapered and discontinued if no seizure activity occurs during treatment, although data supporting this therapy are lacking.

Glycerol for Intracerebral Hemorrhage

A double-blind, randomized, placebo-controlled study of 216 patients examined the use of glycerol for the treatment of ICH.¹⁸⁴ One hundred seven patients received active treatment and 109 were given placebo. Treatment consisted of 400 mL of 10% glycerol in saline administered over a period of 4 hours on 6 consecutive days. In experimental animals, glycerol reduces cerebral edema without a rebound effect and increases cerebral perfusion. In this study, however, glycerol had no effect on outcome and caused subclinical hemolysis in some patients.

Hemodilution for Intracerebral Hemorrhage

Hemodilution has also been studied in patients with ICH because it increases CBF and decreases blood viscosity in experimental animals. One prospective randomized study sought to answer this question.¹⁸⁵ In this study, 164 patients with ICH within 12 hours of onset and a hematocrit higher than 35% were randomized to either hemodilution or control. Eighty-three patients were treated and 81 served as controls. Therapy consisted of removal of 350 mL of blood and infusion of 350 mL of dextran 40 in normal saline. There was no difference in outcome when the hematocrit level was decreased an average of 13%.¹⁸⁵

Recombinant Activated Factor VII

As demonstrated by multiple studies, the volume of the hematoma is a critical determinant of mortality and functional outcome after ICH,^163,186 and early hematoma growth is an important cause of neurological deterioration.^{23,162,187,188} Early hematoma growth occurs in the absence of coagulopathy and appears to result from continued bleeding or rebleeding at multiple sites within the first few hours after onset.¹⁸⁹ Recombinant activated factor VII (rFVIIa) is approved to treat bleeding in patients with hemophilia, and it has been reported to reduce bleeding in patients without coagulopathy as well.¹⁹⁰ Furthermore, a dose escalation safety study found that doses of rFVIIa ranging from 5 to 160 µg/kg of body weight were not associated with a high frequency of thromboembolic complications in patients with acute ICH.¹⁹¹ In 2005, Mayer and coauthors published the results of a double-blind, placebo-controlled trial in which 399 patients with ICH were randomly assigned to receive a single intravenous dose of 40, 80, or 160 µg/kg of rFVIIa or placebo.¹⁹² Treatment was instituted within 4 hours after the onset of symptoms. Mortality at 90 days was 29% in patients who received placebo versus 18% in the three rFVIIa groups combined (P = .02). Serious thromboembolic adverse events, mainly myocardial or cerebral infarction, occurred in 7% of rFVIIa-treated patients versus 2% of those given placebo (P = .12). From these data, Mayer and associates concluded that treatment with rFVIIa within 4 hours after the onset of ICH limits growth of the hematoma, reduces mortality, and improves functional outcomes at 90 days despite a small increase in the frequency of thromboembolic adverse events. To confirm the treatment advantage with respect to outcome, a phase III trial labeled with the acronym FAST (Factor Seven for Acute Hemorrhagic Stroke) was conducted by the same group.¹⁹³ In this trial, 841 patients with ICH were randomly assigned to receive placebo (268 patients), 20 µg of rFVIIa per kilogram of body weight (276 patients), or 80 µg of rFVIIa per kilogram (297 patients) within 4 hours after onset of the ictus. Surprisingly, despite a reconfirmed significant reduction in volume of the hemorrhage, rFVIIa therapy did not improve survival or functional outcome. The proportion of patients with a poor clinical outcome at 90 days did not differ between groups (24% in the placebo group, 26% in the group receiving 20 µg/kg of rFVIIa, and 29% in the group receiving 80 µg/kg).

A well-defined subgroup of warfarin (Coumadin)-treated ICH patients might benefit from rFVIIa therapy. At the present time, there are no clear-cut recommendations, and the treatment modalities used are diverse.¹⁹⁴ Successful administration of rFVIIa for treatment of spontaneous ICH in a patient taking oral anticoagulants for a prosthetic heart valve has previously been reported.¹⁹⁵

Neuroprotective Agents

There has been some interest in the use of neuroprotective agents. Because ICH causes focal cerebral ischemia,^160,196,197 neuroprotectants could interrupt the excitotoxic cascade and prevent neuronal death. In an experimental rat model of ICH, muscimol (γ-aminobutyric acid antagonist) and MK 801 (N-methyl-D-aspartate [NMDA] antagonist) increased the rats’ tolerance to larger hematomas. The area of white matter around the basal ganglia was also better preserved. In another study,¹⁶⁰ pretreatment with nimodipine (Ca²⁺ channel blocker) and D-CPP-ene (NMDA receptor blocker) significantly reduced the ischemic volume and brain edema, respectively, at 24 hours in comparison to no treatment.

Summary

The four randomized medical trials for ICH, two testing steroids,¹⁴⁹ one glycerol,¹⁸⁴ and one hemodilution,¹⁸⁵ did not show significant benefit for any therapy. General measures to control blood pressure, reduce ICP, prevent seizures, and maintain systemic health are important in preventing progression of the hemorrhage, edema, and brain ischemia. Despite disappointing results thus far, these parameters should be controlled aggressively. The optimal reduction in blood pressure needed to perfuse the brain adequately yet prevent rebleeding or progression is controversial. However, the patient’s state of consciousness is the best guide to prognosis^1,198 and may help determine the degree of blood pressure and ICP control that should be instituted. If neurological deterioration occurs or ICP cannot be controlled, surgical evacuation should be considered.

Experimental Rationale

The physiologic effect of ICH on the brain is multifactorial. In clinical practice, patients are particularly susceptible to neurological deterioration within the first 24 hours after hemorrhage,^23,94 especially in the first 4 to 6 hours.^23,162,163 Neurological status is, however, often affected disproportionate to the anatomic extent of the lesion. Furthermore, studies have failed to correlate elevated ICP with clinical findings.¹⁹⁹ Animal models suggest that the hematoma’s effect on local rather than global CBF is responsible for progressive ischemia.^{159,160,196,200,201} The results of pathologic and experimental studies have indicated that a penumbra of progressive tissue damage and edema develops in regions immediately surrounding a hematoma.^202,203 Mechanical injury^203,204 as a result of elevated local tissue pressure, decreased CBF,²⁰⁵ reduced plasma levels,¹⁶¹ and inflammation related to accumulated clotting proteins²⁰⁶ and protease induction²⁰⁷ have all been implicated as mediators of this form of secondary injury. Clinical neurological deterioration, which occurs in approximately a third of patients with ICH,⁹³ can develop as a direct consequence of this process.

Nath and coworkers studied the effect of various volumes of blood injected into the caudate nucleus of rats.¹⁹⁶ Larger volumes produced larger areas of ischemia at 1 minute, and CBF decreased significantly in areas near the hematoma and ipsilateral frontal lobe. Cerebral perfusion pressure, however, remained unchanged, thus implying the presence of a local squeezing effect on the microcirculation rather than a generalized alteration in perfusion pressure.

Nehls and colleagues studied the amount of ipsilateral caudate blood flow that was below ischemic levels.¹⁵⁹ At 5 minutes, 11.5% of the caudate volume was associated with blood flow of less than 25 mL/kg per minute. The ischemic area increased to a maximum of 38.9% at 4 hours. This time course suggested that interventions that reduce hematoma size might decrease ischemia. Therefore, in a subsequent study, the caudate balloon was deflated after 10 minutes in group 1 and after 24 hours in group 2. Group 1 had a higher mean CBF in the caudate nucleus and cortex and smaller areas of ischemia than did group 2. Group 1 also had a better neurological outcome. In further studies, the potential for limiting local ischemia was found to be less if the balloon was deflated after 2.5 hours.¹⁶⁰ A pig model using a clot rather than a balloon showed that removal of the clot at 3 hours markedly reduced perihematomal edema and mass effect at 24 hours.¹⁶¹ Eliminating the hematoma prevented its serum proteins from diffusing into the adjacent white matter, thereby preventing subsequent edema. These experiments imply that early evacuation of ICH may decrease the ischemia caused by the hematoma by improving local blood flow, preventing the formation of edema, and preventing local mass effect.

Ropper and Zervas compared lesions made from whole blood, centrifuged blood, and inert plastic in the caudate nucleus.²⁰⁰ In all animals, regional CBF about the hematoma decreased. Relative hyperperfusion in one or both cortices was associated with whole blood on the second and third days, immediately with centrifuged blood, and never with inert plastic. These findings imply that sheer destruction of the caudate nucleus is not entirely responsible for changes in blood flow whereas certain elements within blood may be.

Jenkins and associates compared equal volumetric caudate injections of blood, oil (equal in viscosity to that of blood), and cerebrospinal fluid.²⁰¹ They evaluated CBF and ischemic cell damage by light microscopy. CBF was immediately reduced adjacent to the lesion in all groups. With blood, however, CBF was reduced over a greater radius and throughout the ipsilateral cortex. In addition, at 4 hours, ischemic damage was present with both blood and oil but not with cerebrospinal fluid. This finding implies that both tissue pressure and vasoactive substances are components in the decreased regional CBF and that both play a role in ischemia at 4 hours.

In rats whose hematomas were “contained” within the caudate nucleus, global cerebral perfusion pressure was unaffected, whereas “unconfined” extension into the ventricles or subarachnoid space reduced cerebral perfusion pressure globally. This finding may explain the poorer outcome in this subgroup and may indicate a need to monitor ICP.¹⁶⁰

In summary, ICH causes alterations in CBF. Larger hematomas produce a stronger, immediate effect. The effect of both vasoactive substances and local tissue pressure increases during the first 4 hours, and the effect persists for 24 hours. This change in regional CBF produces histologic ischemia and a poorer neurological outcome, which is reversible to some extent. Animal studies suggest that early surgery may help limit secondary ischemia and improve outcome.

Effect of Age

Age has a predictive role in the outcome and mortality rate of patients with ICH.^{162,208–211} It is an important factor that affects microglia and astrocyte reactions and plasticity. Brain swelling was more severe in old rats than in young rats at 3 days after ICH (P < .05).²¹² There were also more severe neurological deficits in the older rats at 1 day after ICH that persisted for the 4 weeks (P < .05).²¹² In a prospective outcome study of conservatively treated patients, old age was the most important predictor of a poor outcome.²¹¹ In a prospective, randomized trial that compared surgical and conservative treatment, surgically treated patients younger than 60 years had a significantly lower mortality rate than did those older than 60 years (25% versus 65%).²⁰⁹ In surgically treated patients older than 60 years, 67% had poor outcomes (activities of daily living [ADL] score >3), as compared with 50% of the patients younger than 60 years.²⁰⁸ The relationship between age and outcome is even more pronounced with thalamic hemorrhages. Patients who were younger than 59, 60 to 69 and older than 70 years had 59%, 33%, and 17% good or excellent outcomes, respectively. In a retrospective study of patients identified as having a “rapidly progressive” hematoma by serial CT, age older than 65 years was associated with 100% mortality in those whose deterioration prompted surgery. Age is thus an important element in any treatment decision, and age older than 60 years implies a poor prognosis, regardless of treatment.

Effect of Hematoma Volume

Expansion of the hematoma is the most common cause of underlying neurological deterioration within the first 3 hours after the onset of hemorrhage. In the era before CT, the volume of ICH could be inferred only from the angiographic distortion of vessels, thereby yielding imprecise estimates of the size and quality of the hematoma. CT made it possible to quantify the volume of hematomas and elucidate their characteristics, which helped delineate their natural history and clinical outcome. In experimental animal models of ICH, larger volumes were associated with larger areas of ischemia and poorer outcomes.²¹⁴ In many retrospective* and prospective studies,²⁰⁹ the volume of hematomas based on CT measurements is a strong predictor of functional outcome and death in humans. The progression of hematomas has been examined by serial CT imaging, which helps predict further growth and potential clinical deterioration.^23,94

Hematomas can be measured by direct computer imaging or estimated with simple calculations. The volume can be calculated by recording the largest diameter seen on CT, the diameter orthogonal to it, and the number of 1-cm slices on which the hemorrhage can be seen. The total volume can be estimated by using the formula for an ellipsoid, V = × π × ABC ÷ 8, where A, B, and C represent the respective diameters of the three dimensions.¹⁶³ Simplified, this equation yields an approximation of ABC/2, which has proved to be quite accurate.²¹⁶

Supratentorial Hematomas

In 1977, Hier and colleagues reviewed 5000 CT scans to correlate the volume of putaminal hematomas with clinical findings and prognosis.⁹⁷ The authors defined three groups. Patients with small hematomas (<35 cm³) showed mild to moderate hemiparesis or hemisensory loss, preservation of higher cortical function, and a good prognosis, regardless of treatment. Patients with moderate hematomas (mean, 120 cm³) had classic flaccid hemiplegia, hemisensory defect, lateral gaze preference, homonymous hemianopia, and either aphasia or apractagnosia. Massive hematomas (>200 cm³) produced coma, fixed and dilated pupils, papilledema, absent eye movements, bilateral fixed plantar response, and rapid death. These correlations suggested that patients with moderate hematomas might be candidates for a controlled clinical comparison of surgical and conservative treatment.

In a retrospective review of 188 patients with supratentorial ICH, Broderick and coworkers showed that hematoma volume was the strongest predictor of the 30-day mortality rate and functional outcome for all locations (putaminal, thalamic, and subcortical).¹⁶³ The 30-day mortality rates for deep hemorrhages less than 30 mL, between 30 and 60 mL, and greater than 60 mL were 23%, 64%, and 93%, respectively. Mortality rates for lobar hemorrhages at these volumes were 7%, 60%, and 71%, respectively. Half the patients who died did so within the first 2 days. Of 71 patients surviving with hematoma volumes larger than 30 cm³, only 1 (1.4%) was independent at 30 days. In contrast, of the 91 patients who survived with hematoma volumes of less than 30 cm³, 16 (18%) were independent. Combining hematoma volumes with admission Glasgow Coma Scale (GCS) scores proved to be a 97% sensitive and 97% specific test for 30-day mortality. For patients with GCS scores lower than 8 and hematoma volumes greater than 60 cm³, the probability of death at 30 days was 91%, whereas it was 19% for those with GCS scores higher than 9 and hematoma volumes of less than 30 cm³. Even though this study did not aim to evaluate the effectiveness of treatment, operative removal was associated with a decreased 30-day mortality rate, although overall surgical morbidity and mortality rates were not significantly different from those associated with conservative treatment.

Volpin and coworkers retrospectively reviewed the outcome of medical treatment and craniotomy in 132 supratentorial ICHs with respect to hematoma volume, regardless of location.²¹⁷ When compared with conservative treatment, surgery decreased the mortality rate of comatose patients with hematoma volumes between 26 and 85 cm³, but the probability of discharge with a severe deficit was high. All patients with hematoma volumes greater than 85 cm³ died, irrespective of treatment, and all patients with hematoma volumes of less than 26 cm³ survived without surgery.

In contrast, a retrospective comparison of surgery and conservative therapy in 182 patients with putaminal hemorrhage showed that the size of the hematoma on CT was a statistically significant predictor of outcome regardless of the treatment modality.¹¹⁶ Localized hematomas or those extending into either limb of the internal capsule (groups I to III) were compared with those that extended into both limbs of the internal capsule, the thalamus, or both (groups IV and V). The 30-day mortality rate was significantly lower in groups I to III than in groups IV and V (12% versus 57%).

A randomized, prospective study comparing endoscopic removal of supratentorial ICH with medical management found surgery to be beneficial for hematomas of all volumes, especially subcortical hematomas.²⁰⁹ Interestingly, patients with hematoma volumes of less than 50 cm³ had better outcomes after surgery than after conservative treatment (25% versus 0% with an ADL score of 1), although the mortality rate was the same. For large hematomas (>50 cm³), there was no difference in functional outcome between the two groups, but the mortality rate was lower in the surgical group than in the conservative group (48% versus 90%). This study suggests that surgical evacuation may play a lifesaving role in patients with large hematomas by sparing viable local brain function by decreasing mass effect, progressive edema, or impaired cerebral perfusion. The overall lower surgical mortality rate (30%) in this study in comparison to others may reflect surgical technique and is discussed later in this chapter.

Large-volume thalamic hematomas are more devastating than similarly sized subcortical or putaminal hematomas. Of 29 patients with thalamic hemorrhages, those with volumes greater than 10 mL or with a maximal diameter greater than 3 cm had ADL scores of 4 or 5.²¹⁶ Thalamic hematomas with a long axis greater than 3 cm are associated with poor outcomes.^99,101,115 In a comparison of 75 patients with thalamic hemorrhages who underwent either stereotactic aspiration or conservative treatment, 31 of 40 (78%) surgical patients with hematoma diameters larger than 3.3 cm returned to useful activity in 6 months.²¹⁰ This finding suggests that the less invasive nature of stereotactic aspiration may improve therapeutic outcomes irrespective of the size and location of the hematoma.

Hematoma volume also seems to be related to the risk for deterioration. In a retrospective study of 182 African Americans, the presence of ICH greater than 30 mL increased the risk for deterioration and death in the first 24 hours by 6.78- and 6.66-fold, respectively.⁹⁴ Of 46 noncomatose patients, 15 deteriorated during the initial 24 hours. In this study, hematoma volume was a better early predictor of poor outcome than was the admission GCS score.

Understanding the natural time course of an acute ICH and its effect on clinical deterioration is critical to therapeutic decision making. Before CT was available, ICH was considered to be a monophasic event that stopped quickly as a result of clotting and tamponade by the surrounding regions.²¹⁹ This impression is incorrect, as demonstrated by CT scans showing that hematomas expand over time.^23,188 The radiographic progression of a hematoma and its correlation with the clinical course have been well studied. Rehemorrhage typically occurs within the first 6 hours of the primary ictus.^162,163 Most of the extravasation of contrast material in the angiograms of patients with ICH is seen within 3 to 6 hours of onset.^220,221 If deterioration occurs later than 6 hours after hemorrhage, other factors—edema, hydrocephalus, a new intraventricular hemorrhage, or a metabolic abnormality—must be contributing.

In a study of 419 patients, Fujii and coworkers obtained the first CT scan within 24 hours of onset and a follow-up scan 24 hours after admission.¹⁸⁷ They found that the hematomas had enlarged in 14.3% of the patients. In another study, Kazui and colleagues performed sequential CT scans in 204 patients with acute ICH.¹⁸⁸ Hematomas enlarged 40% in 20% of the patients. This was seen when the scans were obtained early; none of the patients showed an increase in hematoma size after 24 hours.

To define the progression over time more closely, Brott and associates prospectively studied 103 patients with ICH at all locations who underwent CT within 3 hours of onset.²³ The patients were rescanned 1 and 20 hours after the first scan. On the 1-hour follow-up scan, 26% of patients exhibited hematoma growth (>33% enlargement). An additional 12% showed growth between 1 and 20 hours. Therefore, 38% of the patients exhibited hematoma progression within 24 hours of hemorrhage. Of these patients, 33% deteriorated within the first hour, and an additional 25% deteriorated within the next 20 hours. Thus, the clinical condition of more than 50% of all patients showing progression on serial CT deteriorated. This finding implies that early hematoma evacuation may not only reduce perihematomal ischemia^160,196,197 and the toxic effect of blood products^160,200,201 but also contain potential hemorrhagic progression.

Bae and colleagues reported similar results from their retrospective study of 320 patients who underwent serial CT for ICH.¹⁵³ Three percent of patients showed rapid hematoma progression. The follow-up scan was performed at a mean of 48.9 hours (range, 10.5 to 149 hours), and 50% of the patients showing progression deteriorated before 24 hours had elapsed. In this study, the most important risk factor for progression was persistent hypertension.

When Qureshi and coworkers retrospectively reviewed 182 African American patients to identify independent predictors of early deterioration and death, 23% of the patients with GCS scores higher than 12 showed early deterioration (mean, 7.9 hours).⁹⁴ An ICH volume greater than 30 mL and ventricular extension were independent predictors of early deterioration (OR, 6.78 and 4.67, respectively) and death (OR, 6.66 and 4.23, respectively).

The goals of surgical evacuation of a hematoma are to reduce the mass effect, block the release of neuropathic products from the hematoma, and prevent prolonged interaction between the hematoma and normal tissue, which can initiate pathologic processes.²²² From the data available it is clear that within the spectrum of ICH there are some patients (with large or space-occupying ICHs) who require surgery for neurological deterioration and others with small hematomas who should be managed conservatively. There is equipoise about the management of patients between these two extremes. Some patients have a penumbra of functionally impaired but potentially viable tissue around the ICH. Surgical removal of the clot may improve the function and recovery in this penumbra.¹⁶⁹ As suggested by the aforementioned studies, early surgery may play an important role in preventing secondary deterioration, and the timing of surgery becomes an important question that needs to be elucidated.

Infratentorial Hematomas

With cerebellar hemorrhages, identifying the appropriate clinical progression is paramount to guiding surgical evacuation. When hematomas are near the brainstem, however, irreversible deterioration can occur without warning. Many studies recommend surgery for all hematomas greater than 3 cm in diameter.^96,123,232 Smaller lesions have a more benign course. However, the patient’s clinical profile must still be monitored carefully because spatial accommodation is, by necessity, more limited here than supratentorially.

Cerebellar hemorrhage tends to progress rapidly and to cause death because of its proximity to the brainstem. It has been recommended that patients with GCS scores of 13 or less or with hemorrhages of 4 mL or greater should undergo surgical evacuation.²³³ Other authors, however, contend that rapid progression of particular cerebellar and cranial nerve syndromes (see earlier) represents a surgical emergency despite these criteria.^92,234

Surgical Techniques

In 1903, Cushing first removed an intracerebral hematoma by craniotomy.⁸ In 1932, Bagley first described indications for removal based on location.⁹ In 1950, Fazio, a neurologist, suggested surgical removal based on pathologic studies.²³⁵

In 1961, McKissock and coworkers published the first pessimistic view of surgical treatment of ICH.¹ They reported a 51% mortality rate in 244 operated patients and a 100% mortality rate in comatose patients. Since then, outcome studies have yielded controversial results. Operative mortality can range from 20% to 90% in comatose patients with deep ganglionic or thalamic bleeding.^1,148,236 Most studies, however, have either failed or lacked the power to stratify patients appropriately or to identify the optimal timing of surgery in a well-designed, randomized, prospective manner (see earlier discussion). Because of this controversy, various less invasive methods of removal are practiced around the world: simple aspiration, stereotactic aspiration, fibrinolytic treatment, mechanically assisted aspiration, and endoscopy.²³⁷ In certain circumstances, some of these techniques may be more efficacious for deep putaminal or thalamic hemorrhages. Others are beneficial for subcortical hematomas. These techniques and the use of standard craniotomy for ICH are detailed in this section.

Table 358-2 summarizes the clinical retrospective studies that have compared surgery and medical management.

Craniotomy

Several technical points must be considered when planning a craniotomy for ICH. In the preoperative period, an arterial line, intravenous fluids, and correction of electrolytes are crucial. During intubation, hypertension must be avoided, and medical therapy must be maximized based on the patient’s neurological condition.

We treat a putaminal hemorrhage as a microsurgical lesion requiring meticulous surgical technique to avoid adding insult to injury (Figs. 358-6 and 358-7). Accordingly, we favor a transcisternal-transsylvian-transinsular approach.²⁴⁷ The hematoma often extends to within millimeters of the insular cortex, and a small 2-cm insular corticotomy is often all that is needed to evacuate the largest hematoma. The operating microscope is used routinely, and its improved illumination and magnification make identification and bipolar coagulation of cavity wall bleeders straightforward. Alternatively, we rely on a malleable graduated suction device to suction and handle tissue.

FIGURE 358-6 A, Admission computed tomographic scans of a 54-year-old patient with an acute hypertensive putaminal hemorrhage, grade 1/5 hemiparesis, and mild confusion. B, Computed tomographic scans of the same patient on postoperative day 1.

FIGURE 358-7 A and B, Admission computed tomographic scans of a 71-year-old patient with warfarin-induced hemorrhage into the putamen, globus pallidus, and claustrum. He was completely hemiplegic and lethargic. C and D, Computed tomographic scans of the same patient 8 days after surgery.

For several reasons we rarely use self-retaining retractors. First, steady retraction is deleterious to brain parenchyma. Second, the “static” retraction provided is nonergonomic and ignores the fact that the evacuation process relies on constant change in angle, depth, and orientation. The center of the hematoma is removed first. The remaining marginal clot then collapses and can likewise be evacuated, with particular attention paid to bleeding points and possible subtle pathologic findings such as small tumors, cryptic AVMs, and cavernous angiomas. All tissue is sent for histologic analysis.

Extreme caution is used to avoid traumatic manipulation of the capsular fibers at the depth of the surgical cavity. Hemostasis is ensured by elevating systolic pressure temporarily to identify potential rebleeding sites.²⁴⁷ Otherwise, blood pressure is maintained postoperatively in the normal range for the specific patient. If the hematoma extends significantly into the temporal lobe, the transtemporal approach can be used. The transfrontal approach is rarely used and is almost obsolete because the surgical track is, by necessity, deep. Suzuki and Sato once reported that the functional prognosis was better with the transsylvian than with the transcortical approach.²⁴⁸ Kanaya and Kuroda advocated transcortical approaches for large hematomas.²⁴⁹ The general surgical principles for evacuating hematomas at other locations also follow commonsense strategies: corticotomies are placed near the epicenter of the ICH, their length is minimized, and above all, eloquent tissue is avoided.

For infratentorial hematomas, a suboccipital craniotomy with the patient in the prone or lateral position is standard. Most commonly (i.e., for unilateral hemorrhage in the deep cerebellar nuclei), a paramedian straight incision is used. We strongly advocate a bone flap craniotomy rather than a craniectomy; the latter is slower, produces less pleasing cosmetic results, and may cause craniectomy headaches. A ventriculostomy is often necessary to relieve hydrocephalus.²⁵⁰ Figure 358-8 shows an example of evacuation of a cerebellar hematoma, coupled with postoperative intrathecal clot lysis with rt-PA, which resulted in prompt resolution of the intraventricular hematoma.

FIGURE 358-8 A, A 53-year-old man with a hypertensive left cerebellar hemorrhage, extension into the third and fourth ventricles, and obstructive hydrocephalus. He was increasingly somnolent with divergent gaze. Emergency placement of bilateral ventriculostomies was done in the operating room with immediate subsequent evacuation of the hematoma via a left paramedian small craniotomy, 2.5 by 2.5 cm. The clot was evacuated with forceps, irrigation, and bipolar coagulation. B-E, Postoperative computed tomographic CT scans show good evacuation of the intracerebral hematoma but persistent intraventricular hematoma in the third ventricle. A single intraventricular dose of 7.5 mg of alteplase (recombinant tissue plasminogen activator) was given. F, Successful clot lysis. He never required a shunt and was weaned off the ventriculostomy within 1 week.

Typically, lobar hemorrhages are the most rewarding to treat surgically, particularly when addressed early. This is only a generalization, however. Age, medical history, and neurological grade at admission are among the most powerful prognosticators, as discussed earlier.