Intracranial Pressure Response

The volume of blood escaping during rupture of an aneurysm varies from a negligible amount constituting a “warning leak” to massive amounts (≥150 mL) associated with immediate death. The pathophysiologic consequences depend on the volume and location of the bleeding, as well as the preexisting size of the cerebrospinal fluid (CSF) space into which the aneurysm ruptures and the patient’s age and premorbid condition. A general correlation can be expected between the volume of SAH and clinical grade, risk for vasospasm and other complications (e.g., increased intracranial pressure [ICP], seizures, hydrocephalus), and the extent of the physiologic changes (e.g., reduced cerebral blood flow [CBF] and metabolism), systemic alterations (e.g., hyponatremia, hypovolemia, hypermetabolism), catabolic state, cardiac arrhythmia, and cardiac wall motion abnormalities.

The ICP response to rebleeding of aneurysms is known in the subset of patients with SAH who have ICP monitors in place during their rebleeding, but these changes may not be representative of those that occur with the first hemorrhage. During rebleeding, ICP rises to diastolic blood pressure and CBF occurs only during systole.^8,9 Temporary circulatory arrest may help stop the aneurysmal bleeding but may also be associated with transient global ischemia, which would cause loss of consciousness. Because many patients do not lose consciousness, however, normal clotting probably also contributes to arrest of the hemorrhage. In addition, animal experiments have shown that the rate of blood flowing into the subarachnoid space determines the volume of SAH.¹⁰ High flow rates, which would theoretically occur with large aneurysm tears, produce large-volume SAH in a short time, whereas low flow rates, which might result from a small hole in the aneurysm, lead to slow accumulation of a small volume of SAH.

The results of ICP monitoring in 52 patients for a mean of 8 days after SAH showed that mean ICP rose as clinical grade worsened.¹¹ Mean ICP was 10 mm Hg in patients with clinical grades 1 and 2, 18 mm Hg in patients with clinical grades 2 and 3, and 29 mm Hg in those with clinical grades 3 to 5. Vasospasm, which was more common in patients with a poor clinical grade and larger SAH, was associated with a significant rise in ICP from a mean of 16 mm Hg in those without vasospasm to 29 mm Hg in those with vasospasm. This fact should be considered carefully in patients experiencing deterioration from vasospasm days after SAH. Substantial improvement in cerebral perfusion pressure can be achieved by ventricular drainage. Cerebral perfusion pressure is equal to mean arterial blood pressure minus ICP. There are no specific data regarding optimal cerebral perfusion pressure in patients with aneurysmal SAH, but guidelines for head injury suggest that it be maintained at values higher than 70 mm Hg. ICP is related to outcome; 80% of patients with pressure below 15 mm Hg do well, as opposed to good outcomes in just 15% of patients with ICP in excess of 15 mm Hg.¹²

Cerebral Blood Flow, Volume, and Metabolism

Numerous studies of CBF, cerebral blood volume, and cerebral metabolism have been conducted in patients with ruptured aneurysms. Almost all studies agree that CBF is globally decreased after SAH.¹² For example, in 30 patients with SAH, regional CBF decreased from a mean of 54 mL/100 g per minute in normal individuals to 42 mL/100 g per minute in patients with grade 1 to 2 and no vasospasm, 35 mL/100 g per minute in patients with grade 3 to 4 and no vasospasm, 36 mL/100 g per minute in patients with grade 1 to 2 and vasospasm, and 33 mL/100 g per minute in patients with grade 3 to 4 and vasospasm.¹³ The cerebral metabolic rate of oxygen (CMRO₂) showed a similar pattern, with progressive reductions being associated with deteriorating clinical grade and worsening vasospasm. Cerebral blood volume was markedly increased in patients with severe neurological deficits associated with severe vasospasm. It was concluded that vasospasm produces narrowing of the large, angiographically visible arteries at the base of the brain and that this narrowing is accompanied by compensatory dilation of the distal, intracerebral arterioles. Other studies have shown a lack of increased cerebral blood volume after SAH, thus suggesting impaired autoregulatory vasodilation.¹⁴

Mean CBF decreases with time after SAH; it reaches a nadir in 10 to 14 days, after which it slowly increases toward normal.^12,15 A relative hyperemia in relation to the reduced CMRO₂ occurs immediately after SAH.^16,17 In patients with poor grades, CBF and cerebral metabolism may remain depressed for weeks. In addition to global reductions in CBF and CMRO₂, regional perfusion defects can develop after SAH and can be correlated with areas of angiographically demonstrated severe vasospasm and ventricular dilation. Areas of low CBF are also present around intracerebral hematomas, although their size and importance may be overestimated.¹⁸ Regions of brain irrigated by vasospastic arteries have elevated oxygen extraction fractions. Positron emission tomography shows that as long as the area is ischemic and infarction has not developed, CMRO₂ remains normal but flow is reduced. The development of infarction is heralded by a fall in CMRO₂ with relatively increased CBF (relative hyperemia).¹⁶ The degree to which the reduced blood flow after SAH is due to hypovolemia and hypotension is unclear, but there is some evidence that the alterations in flow can be prevented by maintenance of normovolemia or hypervolemia with or without induced hypertension.^19,20

Little information is available on the pathogenesis of CBF and the metabolic changes that occur after SAH. In patients without vasospasm, intracerebral clots, or hydrocephalus studied in the first 4 days after SAH, CMRO₂ is decreased without accompanying changes in the oxygen extraction fraction, thus suggesting that the primary alteration is a reduction in CMRO₂ and that CBF falls because of decreased demand.¹⁶ This is usually attributed to a toxic effect of the subarachnoid blood, but a neural mechanism or an effect of global ischemia may be important. A relative hyperemia is usually present and is postulated to be due to intracranial circulatory arrest, transient global cerebral ischemia, and lactic acidosis occurring at the time of rupture. Mitochondrial respiration, sodium-potassium adenosine triphosphatase activity, and extracellular potassium and calcium levels are altered in the brain tissue of experimental animals exposed to subarachnoid blood, although the relationship of these changes to CBF and CMRO₂ has not been fully clarified.^21–23

The relationship of CBF to blood pressure and PaCO₂ is also altered after SAH. The response of CBF to changes in blood pressure at different times after SAH was studied in 38 patients.²⁴ Autoregulation was intact in good-grade patients but became progressively impaired in poor-grade patients and with development of vasospasm. Autoregulation is not lost in an all-or-none fashion. The degree of impairment tends to be worse as consciousness is more impaired, as vasospasm becomes more severe, and 5 to 10 days after SAH. Loss of the CBF response to changes in PaCO₂ occurs with more severe brain damage than that required to disturb autoregulation, and the combination of loss of autoregulation and variation in CBF with changes in PaCO₂ is termed vasomotor paralysis. After SAH, vasomotor paralysis may be observed in patients with clinical grades 4 and 5, usually with severe vasospasm. Measurements of CBF with intra-arterial xenon 133 in 38 patients with aneurysmal SAH found that responses to alterations in PaCO₂ were generally preserved, although they were reduced.²⁴ Impaired CO₂ reactivity was associated with increased ICP and high lactate levels in CSF, and poor clinical grade and vasospasm were associated with impaired CO₂ responsiveness. Transcranial Doppler studies have demonstrated impairment in CO₂ reactivity after SAH, even in patients with a good clinical grade.²⁵ This tends to occur during vasospasm and then subsequently resolves. Hyperventilation should be used with caution in patients with SAH. It may be useful for reducing increased ICP and increasing CMRO₂, but it may also increase the risk for ischemia by causing vasoconstriction.²⁶

Symptoms and Signs of Subarachnoid Hemorrhage

The hallmark of SAH is a sudden, usually severe, headache, although at most about 80% of patients who can give a history will recount such a headache.¹² Challenges arise in assessing the minority of patients in whom an atypical manifestation occurs, and for this reason clinicians must remain vigilant for the possibility of SAH because a missed diagnosis can result in catastrophic sequelae. The classically described cardinal symptom of SAH as a sudden, severe headache or the “worst headache of my life” may be overemphasized as a reliable initial diagnostic symptom, but all patients with headaches that are unusually severe or sudden in onset should be investigated for SAH. Byyny and coauthors reported that 4 of 27 patients seen in an emergency department with a sudden severe headache were found to have SAH on computed tomography (CT).²⁷ An additional 19% had SAH detectable only by lumbar puncture. The positive predictive value of a sudden severe headache was just 93%. Linn and colleagues reported that it may be the initial symptom in only a third of patients with SAH. In the same series, just half of the patients reported their headache to reach maximum severity instantaneously, with a fifth of the patients reporting it to escalate over a period of 1 to 5 minutes and the remainder over a period greater than 5 minutes.²⁸ Other ominous features include vomiting, onset with exertion, altered level of consciousness, meningismus, or focal neurological deficit. Absence of these clinical findings, however, does not rule out SAH.

Premonitory symptoms (warning leaks or sentinel hemorrhages), typically consisting of an unusually severe headache of sudden onset, sometimes associated with nausea, vomiting, and dizziness, are usually attributed to small hemorrhages from the aneurysm. In 1752 patients with ruptured aneurysms from three series, 20% had a history of a sudden, severe headache before the event leading to admission.^29–31 Other possible mechanisms for these headaches include hemorrhage into the aneurysm wall, acute expansion of the aneurysm sac, or ischemia. The importance of recognizing warning leaks has repeatedly been emphasized because the diagnosis may be delayed until catastrophic SAH occurs, which almost certainly makes the outcome worse than if the diagnosis had been made promptly and correctly.³⁰ Some authors discourage the use of terminology such as “sentinel bleed” or ““warning leak” and instead encourage clinicians to have a high index of suspicion in the pursuit of an accurate diagnosis of SAH to foster the mindset that the patient either has or has not experienced SAH.²⁸

Many other symptoms and signs can develop before rupture of an aneurysm, including hemiparesis, dysphasia, extraocular muscle impairment, visual loss, visual field defects, and localized headache. They depend on the size and anatomic location of the aneurysm.¹²

Ruptured aneurysms may produce distinct clinical features at specific sites. Transient bilateral lower extremity weakness may be due to rupture of an anterior cerebral artery aneurysm. SAH from a middle cerebral artery aneurysm is more likely to produce hemiparesis, paresthesia, hemianopia, and dysphasia. Sarner and Rose did not find that any particular aneurysm site had a higher propensity to induce coma.³² Seizures occur more commonly with aneurysms of the anterior circulation and probably with middle cerebral artery aneurysms. Third nerve palsy or unilateral retro-orbital pain suggests an aneurysm arising at the junction of the internal carotid and posterior communicating artery. Third nerve lesions also occur with aneurysms at the origin of the superior cerebellar artery. Carotid-ophthalmic artery aneurysms may produce unilateral visual loss or visual field defects. Focal neurological deficits after SAH may be due to a mass effect from the aneurysm, vasospasm, seizures, or hematomas in the brain or subdural spaces.

Terson reported vitreous hemorrhage and hemiparesis in association with SAH.¹² Vitreous and other intraocular hemorrhages may be seen on ophthalmoscopic examination in 3% to 13% of patients with SAH and are associated with poor clinical grades.^33,34 Because the prognosis for visual recovery is good with no treatment, vitrectomy is generally reserved for patients who fail to improve after months.

Numerous exertional activities have been temporally associated with aneurysm rupture. A cooperative study reported that a third of 2288 aneurysm ruptures occurred during sleep, a third during unspecified circumstances, and a third during various exertional activities, including lifting, emotional strain, defecation, coitus, coughing, and parturition.³⁵ Schievink and associates found that SAH occurred during stressful events in 43% of patients, nonstressful events in 34%, rest or sleep in 12%, and uncertain circumstances in 11%.³⁶ If one takes into account the fact that exertional activities probably occupy a small percentage of one’s lifetime, it seems likely that the fluctuations in blood pressure and changes in venous and CSF pressure that may occur with these activities increase the risk for rupture of an aneurysm. Head injury has only rarely been associated with aneurysmal SAH.

Computed Tomography

Non–contrast-enhanced cranial CT is the first investigation in patients with suspected SAH. The probability of detecting the hemorrhage is proportional to the volume of blood in the subarachnoid space, the time after hemorrhage, and the quality of the scan (Fig. 363-1). In a cooperative study, 3% of 1553 patients were found to have normal scans within 24 hours of confirmed SAH.³⁷ The incidence is probably lower with current CT scanners, but CT still has only about 93% sensitivity for SAH, thus highlighting the need for lumbar puncture when clinically indicated in the context of normal CT findings.²⁷ CT on the day of the ictus showed SAH in 92%, intraventricular hemorrhage in 20%, intracerebral hemorrhage in 19%, hydrocephalus in 16%, mass effect in 8%, aneurysm in 5%, subdural hemorrhage in 2%, and hypodense areas in 1%. With time, the incidence of normal CT scans increases, as well as the presence of areas of hypodensity, whereas hydrocephalus and hemorrhage decrease. By 5 days after SAH, 27% of the scans were normal and 58% showed hemorrhage. Intracerebral hemorrhage takes longer to resolve than SAH. Alert patients are significantly more likely than drowsy patients to have a normal scan or a thin, local collection of blood, and all other abnormalities evident on CT are more common in patients with a poor clinical grade. Additional stigmata of SAH include enlargement of the temporal horns in the absence of increases in other parts of the ventricular system.³⁸

FIGURE 363-1 A, Changes on computed tomography (CT) in patients with a ruptured aneurysm in relation to time after subarachnoid hemorrhage (SAH). The incidence of abnormalities on CT is highest initially and decreases thereafter such that at 5 days about 27% of patients have normal findings. B, Changes in cerebrospinal fluid (CSF) in patients with a ruptured aneurysm in relation to time after SAH. Xanthochromia appears within hours and is found almost universally after 12 hours.

(From Weir B. Headaches from aneurysms. Cephalalgia. 1994;14:79-87.)

The volume and location of subarachnoid blood on CT give prognostic information about the risk for vasospasm and outcome after SAH. A widely used system of grading SAH on CT was proposed by Fisher and coworkers (Fig. 363-2).^39,40 In a prospective study, these authors reported good correlation between the location and volume of the blood and the subsequent development of vasospasm. The degree of SAH on CT was an independent risk factor for death and disability in the International Cooperative Study on the Timing of Aneurysm Surgery.³⁷ Subsequent studies have suggested that initial clot volume and percentage of clot cleared per day are significant predictors of vasospasm whereas Fisher grade and initial clot density are less important.⁴¹ The Fisher scale may not correlate well with vasospasm because the categories are broad and not defined with modern imaging techniques.

FIGURE 363-2 Computed tomography shows different grades of subarachnoid hemorrhage (SAH) according to the scale of Fisher and colleagues.³⁹ Grade 1 is a scan with no SAH. A, Grade 2 is a scan showing a thin layer of subarachnoid blood less than 1 mm thick. This measurement, however, was taken from scans printed at different magnification, so it is not actually 1 mm. B, Grade 3 is a scan showing focal or diffuse subarachnoid blood thicker than 3 mm. Grade 4 is a scan showing intracerebral (C) or intraventricular (D) blood with or without subarachnoid blood.

Studies evaluating CT angiography (CTA) for detection of intracranial aneurysms report sensitivities of 77% to 97% and specificities of 87% to 100%. Sensitivity diminishes significantly with aneurysms measuring less than 3 mm in maximal diameter, and in such cases sensitivities have been reported in the range of 40% to 91%.^42–47 CTA has been demonstrated to be useful in determining suitability of the aneurysm for endovascular treatment in more than 95% of cases.⁴⁸

We use CTA as the preliminary study for evaluation of patients with spontaneous SAH. Catheter-based angiography is indicated when contemplating endovascular treatment, when no source of hemorrhage is identified on CTA, when further anatomic information is required before surgery, and in some cases as the first investigation when there is concern about the administration of contrast material because of allergies or poor renal function.

Lumbar Puncture

Lumbar puncture is indicated for the diagnosis of SAH when findings on CT are normal. CT may show normal results if the SAH is very small or an inordinate amount of time has elapsed between hemorrhage and the CT scan. Contraindications to lumbar puncture include abnormal blood clotting, increased ICP as a result of a space-occupying lesion, suspected spinal arteriovenous malformation, and infection at the lumbar puncture site. When bedside lumbar puncture is not feasible, fluoroscopically assisted lumbar puncture is recommended. Risks associated with lumbar puncture include neurological deterioration from rebleeding of the aneurysm or cerebral herniation. Data from two studies reported that 17 (10%) of 165 patients with SAH who underwent lumbar puncture experienced deterioration within 24 hours.^49,50 It was thus recommended that CT always be performed before lumbar puncture in the setting of suspected SAH. Limited accessibility to a CT scanner and a high level of suspicion for infectious meningitis may make lumbar puncture the initial diagnostic test in patients who do not have focal neurological deficits or a depressed level of consciousness. A review of the literature on lumbar puncture and meningitis concluded that there was no evidence to recommend CT before lumbar puncture in patients with suspected acute meningitis unless atypical features or focal neurological findings were present.⁵¹

Interpretation of CSF results can sometimes be challenging. Theoretically, any erythrocytes in CSF represent hemorrhage. It is possible that a small hemorrhage from an aneurysm could occur directly into the brain parenchyma or into a loculated CSF space and that erythrocytes would not be detected in lumbar CSF, although this is very rare. During CSF access, however, erythrocytes may be introduced artifactually into the CSF sample (traumatic tap). There are many criteria for differentiating a traumatic tap from SAH (Table 363-2), but none are very reliable. A declining erythrocyte count in subsequent tubes is an unreliable indicator of a traumatic tap.⁵² Xanthochromia is a yellow discoloration in a centrifuged CSF sample caused by release of hemoglobin and its breakdown products from hemolysis of erythrocytes. It is a very reliable sign of SAH in CSF obtained more than 12 hours after SAH. The most sensitive test for detection of xanthochromia, however, is spectrophotometry, but most laboratories only visually inspect uncentrifuged CSF. If the CSF is not centrifuged, it may be discolored by erythrocytes that are present as a result of either SAH or a traumatic tap. Furthermore, when erythrocytes are introduced into the CSF sample during a traumatic tap, they will eventually lyse and produce xanthochromia. Therefore, CSF samples should be kept at 4°C, centrifuged immediately, and examined for xanthochromia in timely fashion. The time for xanthochromia to appear after SAH is variable, but it persists longer than intact erythrocytes do (see Fig. 363-1). The sensitivity of visual inspection for detecting xanthochromia in 81 patients more than 12 hours after SAH was less than 50%, thus highlighting the utility of spectrophotometry.⁵³ If spectrophotometry for hemoglobin and bilirubin is negative on CSF obtained more than a few hours after the onset of symptoms, angiography is probably not necessary except under unusual circumstances. If findings on CT are normal, erythrocytes are present but not xanthochromia, and 12 or more hours has elapsed after the ictus, we generally perform at least CTA or magnetic resonance angiography (MRA) and occasionally catheter angiography. About 70% of patients seen 3 or more weeks after suspected SAH have xanthochromia. CT will usually be normal, and if CSF is clear at this time, CTA or catheter-based angiography should be performed.

TABLE 363-2 Characteristics of Cerebrospinal Fluid after Subarachnoid Hemorrhage and Traumatic Lumbar Puncture*

* The presence of any erythrocytes or xanthochromia should lead to the tentative diagnosis of a ruptured aneurysm.

Magnetic Resonance Imaging

Magnetic resonance imaging (MRI; fluid-attenuated inversion recovery [FLAIR] and proton density sequences) has been shown to be as sensitive as CT in detecting acute SAH, but MRI remains impractical in most cases as an initial investigation.⁵⁴ It is, however, useful for investigating atypical hemorrhage patterns (e.g., craniocervical junction, parenchymal hemorrhage) or when there has been a delay from the onset of symptoms to initial imaging. For the detection of subacute hemorrhage, gradient echo T2 is the most sensitive sequence, with sensitivities of 94% in the acute phase and 100% in the subacute phase. The next most sensitive sequence is FLAIR, with values of 81% and 87% for the acute and subacute phases, respectively.⁵⁵

There is concern about performing MRI postoperatively in patients with aneurysm clips, and a fatality has been reported.⁵⁶ Modern aneurysm clips are alloys of cobalt, nickel, molybdenum, and chromium, with or without small amounts of iron.⁵⁷ They are not ferromagnetic and should not move in the magnet. Titanium clips are either pure titanium or alloys of titanium, vanadium, and aluminum and are not ferromagnetic. Some institutions test the clips by checking whether they move in the magnet before implantation. Stainless steel clips should not be used.

Catheter Cerebral Angiography

Catheter-based cerebral angiography remains the “gold standard” for the investigation of spontaneous SAH, but it is not always required. Technologic advances in cross-sectional imaging and the ability to create three-dimensional volume-rendered images from 64-slice or greater multidetector CT scanners make catheter angiography unnecessary in some cases. When the pattern of hemorrhage is consistent with the discovered aneurysm and no additional confounding factors are present, the surgeon may be confident in operating without performing catheter-based angiography. Complex aneurysms, endovascular treatment, and associated vascular lesions (i.e., arteriovenous malformation or dural arteriovenous fistula) mandate catheter angiography.

Catheter angiography is also indicated for spontaneous SAH with no underlying cause detected on CTA. A complete angiogram consists of a six-vessel study (including both external carotid arteries) and may include provocative maneuvers as clinically relevant (i.e., cross-compression to study the anterior communicating artery, Alcock’s maneuver to study the posterior communicating artery). A rapidly deteriorating neurological condition may preclude catheter angiography. If neck or back pain or a lower extremity neurological deficit is prominent, searching for a spinal arteriovenous malformation, aneurysm, or neoplasm with spinal MRI or angiography, or both, may be indicated. If the initial angiogram is negative and the clinical and CT patterns are not consistent with benign perimesencephalic hemorrhage, we usually obtain another angiogram about 7 days later. Angiogram-negative benign perimesencephalic hemorrhage is typically manifested as a severe headache without instantaneous onset, and the blood seen on CT is symmetrical about the midline and localized primarily to the prepontine, interpeduncular, crural, and ambient cisterns.

In a review of 2899 procedures, the rate of neurological complications with catheter angiography was 1.3%; 0.7% were transient, 0.2% were reversible, and 0.5% were permanent. Neurological complications were significantly more common in patients 55 years or older, in patients with cardiovascular disease, and when fluoroscopic times were 10 minutes or longer.⁵⁸ Allergic reactions to contrast medium occur in less than 1 in 50,000 studies, and about 1 in a million patients die as a result of such a reaction.⁵⁹

Rupture of an aneurysm during angiography is uncommon. In the first cooperative study, in which 5484 patients were studied by angiography, 7 patients (0.13%) rebled during angiography and 12 (0.22%) 10 minutes to 24 hours later.³⁵ Based on subsequent studies, it is estimated that 3% of patients show extravasation of dye when angiography is performed for investigation of SAH.⁶⁰ If extravasation was observed on an angiogram, the mortality was 70%. It has been suggested that the rate of rebleeding associated with cerebral angiography can be reduced by avoiding catheter angiography within the first 6 hours after SAH, but this remains unproven.^60,61 Because rebleeding is probably most common immediately after the first hemorrhage, whether angiography causes the rebleeding is unknown. We do not delay angiography because a patient is seen immediately after SAH has occurred.

Twenty percent to 30% of patients have multiple aneurysms.¹² A combination of clinical and radiologic features can identify the ruptured aneurysm in 90% to 95% of cases.^62–65 A review of 69 patients with multiple aneurysms generated the following algorithm to predict which aneurysms would bleed: (1) exclude extradural aneurysms, (2) study CT for the presence of focal SAH, (3) look for focal spasm or a mass effect on the angiogram, (4) pick the larger or more irregularly shaped aneurysm, (5) examine the patient for focal neurological signs, (6) consider repeating the angiogram at a later date to look for change in aneurysm size or for focal angiographic signs, and (7) choose the aneurysm that has the highest chance of rupture (anterior communicating artery aneurysm).⁶⁶ Overall, the most proximal and largest aneurysm usually ruptures. If there are two aneurysms on the same artery, it is generally the proximal one that is ruptured. In some cases, MRI has provided additional evidence of localizing value. In about two thirds of patients with multiple aneurysms, all lesions will be able to be clipped through a single craniotomy, and this may be advisable depending on the age and condition of the patient and the location of the aneurysms. In exceptional circumstances and despite the best diagnostic aids, it may not be possible to determine preoperatively which aneurysm bled. The literature contains cases in which recurrent SAH developed in patients with multiple aneurysms after the unruptured one was clipped.

No cause of SAH will be found in 9% to 30% of patients undergoing angiography for SAH, but there are numerous other causes of SAH (see Table 363-1). We perform cranial MRI and MRA in such patients, although the yield is low. In 15 series published between 1978 and 1988, 253 of 1218 patients underwent repeat angiography after an initially negative study, and an aneurysm was found in 11%.⁶⁷ The aneurysm may be missed initially if it thromboses totally after bleeding. Review of the initial studies by another neuroradiologist may be helpful. The anterior communicating artery complex probably harbors the most missed aneurysms. Studies have shown that there is a subgroup of patients with angiogram-negative SAH in whom blood is located predominately in the prepontine and perimesencephalic cisterns. Repeat angiography is probably unnecessary in this situation if a good-quality initial angiogram does not show a posterior circulation aneurysm.⁶⁸

Residual aneurysm was detected on 8% of postoperative angiograms obtained within days of surgery on 2416 patients reported in nine series.⁶⁹ Residual aneurysms hemorrhage at a rate of about 0.02%/yr, although the natural history of these rests is not known with certainty.^70,71 The advantage of knowing about an unexpected residual aneurysm intraoperatively is that an attempt can be made to obliterate it at the time. This must be weighed against the risk associated with further clip manipulation and angiography itself. The incidence of unexpected major arterial occlusion was about 6% in the series just reviewed.⁶⁹ Occlusion is also best detected by intraoperative angiography before permanent ischemia develops. Several series have identified characteristics that increase the yield of intraoperative angiography, such as giant aneurysms and those arising at the ophthalmic, anterior communicating, or middle cerebral arteries or at the basilar bifurcation.^69,72,73 Angiography is virtually always performed at the completion of endovascular aneurysm treatment, and residual aneurysms are seen frequently. Follow-up of residual aneurysms after surgery or endovascular treatment, usually with MRI/MRA, is indicated if treatment would be undertaken should growth occur.

Clinical Grading

Several clinical grading scales have been developed, including the Botterell, Hunt and Hess, and World Federation of Neurosurgical Societies (WFNS) scales (Table 363-3).^74–76 None are universally accepted despite numerous analyses.^77–79 Challenges in the development of a universal grading scale include significant interobserver and intraobserver variability and omission of important additional features that may be predictive of outcome but are too complex to include in a clinical grading scale.⁷⁹

Clinical grading is useful for estimating prognosis, for standardizing assessment to facilitate communication between physicians, and possibly for improving outcome measures in multicenter studies. Finally, repeated standardized assessment with some type of semiquantitative neurological scale is essential to detect deterioration in the patient’s condition. The neurological grade may best be determined after the patient is resuscitated and has undergone ventricular drainage if necessary. Neurological grade is an independent predictor of outcome after aneurysmal SAH.^37,80 Assessment of level of consciousness with the Glasgow Coma Scale (GCS), which is the basis of the WFNS scale, has less interobserver variability.^81,82 The WFNS scale was based on the observation that in a large clinical trial, the clinical features that best predicted outcome were the level of consciousness and the presence of a focal neurological deficit. The GCS is probably the most useful aspect of the grading scales.⁷⁹

General Care

Initial emergency care of a patient with SAH includes assessment of airway, breathing, and circulatory function. A brief neurological evaluation of the level of consciousness, cranial nerves, and motor function will determine whether emergency surgical interventions (placement of an external ventricular drain and evacuation of an intracerebral hematoma) are required. Intubation should be considered for patients with a GCS score of less than 8. Other than lifesaving procedures such as those to reduce severely increased ICP, repair of the aneurysm becomes the focus of treatment to reduce the risk for rebleeding. Secondary benefits of urgent aneurysm repair include safer use of treatments of vasospasm. The decision to treat and the modality used for aneurysm repair (endovascular or clipping) are based on multiple factors, including neurological grade, patient age, location and size of the aneurysm, aneurysm morphology, and the medical condition of the patient. The family history should be documented. Screening of other family members may be indicated if there are first-degree relatives with aneurysms. Diseases associated with aneurysms, such as coarctation of the aorta, polycystic kidney disease, fibromuscular dysplasia, and sickle cell disease, as well as cocaine use and smoking, should be elicited.

Typical admitting orders are shown in Table 363-4. Most patients are admitted to an intensive care or high-intensity observation unit. Bed rest in a dark room, limited visitors, and minimal stimulation are advocated by some practitioners but have not been proved to reduce rerupture rates. Once the aneurysm is repaired, early mobilization is encouraged as tolerated in an effort to minimize the complications associated with bed rest.⁸³ Adequate analgesia should be ensured and excessive painful stimuli avoided because pain can increase cerebral oxygen use by up to 30%.⁸⁴ Intermittent pneumatic compression devices are used routinely, as well as low-molecular-weight heparin prophylaxis, usually beginning 24 hours after aneurysm repair once a postoperative CT scan confirms no unexpected findings.

TABLE 363-4 Admitting Orders for Patients with Aneurysmal Subarachnoid Hemorrhage

Rebleeding

Aneurysm rebleeding has historically been an important cause of mortality after SAH. In the International Cooperative Study on the Timing of Aneurysm Surgery, the peak risk for rebleeding was within the first 24 hours of SAH.⁹⁴ Four percent of patients rebled within 24 hours of SAH. The rate declined thereafter to 1.5% per day, with a cumulative risk of 19% in the first 2 weeks. In another series, 17% of patients rebled in the first 24 hours after the initial SAH despite maintenance of systolic blood pressure below 150 mm Hg.⁹⁵ Rebleeding risk was maximal immediately after the SAH, with 39% of patients who rebled within 24 hours rebleeding within 2 hours. Rebleeding was documented in 7% of 574 patients registered in the Columbia University SAH Outcomes Project between 1996 and 2002. Seventy-three percent occurred within 3 days of ictus. The Hunt-Hess grade on admission and maximum aneurysm diameter were independent predictors of rebleeding.⁹⁶ Other series found an increased risk for rebleeding in patients with poor clinical grade, high blood pressure, short interval from SAH to admission, large aneurysm size, advanced age, and female sex.^95,97 Rebleeding was associated with a mortality in excess of 75% in several series.^96,98

Ruptured aneurysms should be obliterated by endovascular treatment or clipping as soon as feasible. Early aneurysm occlusion reduces the risk for rebleeding, facilitates treatment of vasospasm by increasing the safety of hemodynamic manipulations, and allows the patient to be mobilized more quickly, thereby avoiding the complications related to bed rest. For surgery, there is evidence that early clipping is associated with a better outcome.^37,99,100 Endovascular treatment of ruptured aneurysms in good-grade patients with relatively small, narrow-necked, anterior circulation aneurysms resulted in better 1-year outcomes than did neurosurgical clipping in one study.^71,101 Endovascular treatment is associated with a lower incidence of epilepsy.⁷¹ However, the incidence of early rebleeding after coiling of a ruptured aneurysm is higher than after surgery and is about 1.4% in the acute phase, with mortality approaching 100%. Independent risk factors for rebleeding after endovascular treatment were intracerebral hematoma and small aneurysm size.¹⁰² Late rebleeding, defined as rebleeding from a coiled aneurysm less than 1 month after coiling, has been reported to occur in approximately 0.3%/yr with a mortality of 0.2%/yr.¹⁰³ Rebleeding after surgical clipping is less frequent and varied from 0.14% of 715 patients monitored for an average of 8 years to about 1% in Yasargil and colleagues’ series of more than 1350 aneurysms observed for an unspecified time.^104,105

Antifibrinolytic drugs for the treatment of patients with SAH had been abandoned, but there is renewed interest. Randomized trials have shown that antifibrinolytic drugs reduce the risk for rebleeding but increase the risk for cerebral infarction and as a result had no overall effect on outcome.^106,107 A Cochrane review of the use of antifibrinolytics in patients with SAH concluded that they were of no benefit.¹⁰⁸ Antifibrinolytics increase mortality even in good-grade patients with small-volume SAH,¹⁰⁹ and they have been found to be harmful even when given for as little as 4 days.¹¹⁰ Recent observational studies suggest that short-term antifibrinolytic treatment, which reduces rebleeding, could improve outcome when combined with modern SAH treatment, but randomized, blinded trials are needed to prove the efficacy of such treatment.^111,112

Hydrocephalus

The frequency of acute ventricular dilation after SAH is 20%.^37,113 Ventricular drainage may be required on an emergency basis as a lifesaving measure to relieve acute hydrocephalus and decrease ICP in patients with a depressed level of consciousness. Factors associated in several studies with clinically important hydrocephalus within the first days after SAH include increasing age, preexisting or postoperative hypertension, intraventricular hemorrhage, additional SAH, posterior circulation aneurysms, use of antifibrinolytic drugs, hyponatremia, and depressed level of consciousness (equivalent to worsening clinical grade).^114,115 The pathogenesis of ventricular dilation is probably multifactorial and related to blockade of CSF circulation either within the ventricular system (aqueduct of Sylvius, outlets of the fourth ventricle) or in the subarachnoid space (tentorial incisura or basal cisterns) or related to increased resistance to outflow of CSF at the arachnoid granulations. Acutely, such blockages must be due to blood clots, which leads to proliferation of macrophages, arachnoid cells, and fibroblasts after several weeks.

It is worthwhile quantifying the degree of hydrocephalus by measuring the ventriculocranial ratio and determining whether the ventricles are larger than the 95th percentile for age (Fig. 363-3 and Table 363-5).¹¹⁶ Insertion of a ventricular drain is usually indicated for patients with a ventriculocranial ratio 20% to 25% greater than the 95th percentile for age and a depressed level of consciousness. This applies to most grade 4 and 5 patients.¹¹⁷ Risks associated with ventricular drainage include rebleeding, infection, and intracerebral hematoma along the catheter track. Fountas and colleagues reviewed the literature on ventricular drainage and rebleeding and concluded that studies had not controlled adequately for factors in addition to ventricular drainage that might affect rebleeding, so whether ventricular drainage increased the risk for rebleeding could not be determined.¹¹⁸ If ventricular drainage is needed, it is probably best to lower ICP just enough to maintain adequate cerebral perfusion pressure.

FIGURE 363-3 Computed tomography showing ventriculomegaly. The ventriculocranial ratio (A/B) is calculated by taking the ratio of the width of the ventricles and the width of the inner diameter of the skull at the level of the foramen of Monro at a point behind the head of the caudate nuclei where the lateral walls of the ventricles are parallel.¹¹⁶

TABLE 363-5 Upper 95% Confidence Value for Ventriculocranial Ratio by Age

Data from van Gijn J, Hijdra A, Wijdicks EFM, et al. Acute hydrocephalus after aneurysmal subarachnoid hemorrhage. J Neurosurg. 1985;63:355-362.

If a drain is not in place preoperatively, we usually insert one at the time of surgery to achieve optimal brain relaxation.¹¹⁹ The catheter is left in place postoperatively to monitor ICP, drain CSF as necessary, and administer fibrinolytic agents if indicated. In addition, if an endovascular approach that requires antiplatelet agents is being considered, a ventricular drain may best be placed before endovascular treatment. Some physicians change the catheter or convert it to a permanent shunt if it cannot be removed within 5 to 7 days, unless substantial intraventricular blood is present. A randomized trial showed no difference in the need for permanent CSF diversion or the length of intensive care or hospital stay between gradually decreasing versus immediate closure and discontinuation of external ventricular drainage.¹²⁰

Chronic hydrocephalus develops in 10% to 21% of patients surviving aneurysmal SAH.¹¹⁵ Factors associated with shunt-dependent hydrocephalus after aneurysmal SAH were increasing age, female sex, poor Hunt and Hess grade on admission, thick SAH on initial CT scans, intraventricular hemorrhage, radiologically confirmed hydrocephalus at the time of admission, location of the ruptured aneurysm in the distal posterior circulation, clinical vasospasm, and endovascular treatment.¹¹⁵ Data on rates of hydrocephalus and vasospasm after clipping and coiling are conflicting and would best be examined in studies in which patients are randomized to the treatment methods.⁷¹

Patients should undergo a CT 1 month after SAH to assess ventricular size. Indications for permanent CSF diversion are subjective and include lack of improvement from a neurological plateau or deterioration in ventricular dilation, often with periventricular radiolucencies, rounding of the frontal horns, and obliteration of the cortical sulci. Endoscopic third ventriculostomy is an alternative to permanent shunting that requires further investigation.

Intraventricular Hemorrhage and Increased Intracranial Pressure

A major intraventricular hemorrhage complicates aneurysm rupture in 13% to 28% of patients in clinical series and 37% to 54% of subjects in autopsy series.^37,117 Small amounts of intraventricular blood, such as blood layered in the occipital horns, are seen more commonly. In 91 cases of intraventricular hemorrhage, the aneurysm involved the anterior cerebral artery in 40%, the internal carotid artery in 25%, and the middle cerebral artery in 21%.¹¹⁷ Aneurysms of the anterior communicating artery and the basilar termination are the ones that most commonly cause large, primarily intraventricular hemorrhage. Intraventricular hemorrhage is an independent risk factor for death, disability, vasospasm, and acute and chronic hydrocephalus after SAH.^{114,115,121–123}

More than 50% of patients with large intraventricular hemorrhages are admitted with poor grades, and the mortality in such cases exceeds 64%.¹¹⁷ Ventricular size predicts survival, in addition to known prognostic factors such as age, clinical grade, and hypertension. A fourth ventricle that is dilated and packed with clot is a particularly ominous sign.¹²⁴ Intraventricular thrombolysis with recombinant tissue plasminogen activator seems to assist in the acute management of patients with large aneurysmal intraventricular hemorrhages by speeding clearance of aneurysmal intraventricular hemorrhage, normalizing ICP, and reducing ventricular catheter obstruction.^125–127 A randomized trial is needed to confirm these findings, establish the safety of treatment, and determine whether treatment affects outcome. A ruptured aneurysm should be repaired before administering tissue plasminogen activator; otherwise, there would be potential for lysis of the clot in the ruptured aneurysm with catastrophic rebleeding.

Some patients with acute hydrocephalus have markedly increased ICP and require ventricular drainage as a lifesaving measure.¹¹³ Increased ICP may occur in the absence of ventricular dilation.¹²⁸ The infection rate after ventricular drain placement is about 10%.¹²⁹

Causes of increased ICP acutely after SAH include acute hydrocephalus, intraventricular or intracerebral hemorrhage, brain swelling, ischemic brain edema, and increased resistance to CSF outflow, probably from blockade of the arachnoid villi by blood. There is some evidence that the blood-brain barrier is disrupted after SAH, and this could contribute to the edema and brain swelling.¹³⁰

Treatment of increased ICP after SAH includes removal of sizable intracranial hematomas, ventricular drainage, assisted ventilation, sedation, pharmacologic paralysis, maintenance of normal body temperature and sodium, prevention of seizures, administration of 20% mannitol or 23.4% hypertonic saline boluses intravenously with or without concomitant furosemide, and short periods of hyperventilation to overcome acute elevations in pressure. The increased ICP associated with endotracheal suctioning can be avoided by pretreatment with 100% oxygen and intravenous lidocaine, 1 to 1.5 mg/kg.

Intracerebral Hemorrhage

Aneurysms arising from the distal anterior cerebral arteries are the most likely to produce intracerebral hematomas. Because of the relative rarity of these aneurysms, however, intracerebral hematomas are more commonly seen with aneurysms of the middle and anterior communicating arteries, which are associated with clots in about 67% and 62% of cases in autopsy series, respectively. These percentages are lower in patients surviving their ruptured aneurysms. Intracerebral hematomas complicated 34% of the aneurysm cases reported by Pasqualin and associates.¹³¹ The sites of hemorrhage vary depending on the location of the aneurysm and the direction of rupture of the aneurysm into the brain parenchyma. The pattern is usually distinctive but does not always differ sufficiently from that of hypertensive intracerebral hemorrhage to allow an accurate diagnosis based on CT or MRI. Indications for angiography must be based on clinical suspicion.

A retrospective review of patients from 11 medical centers identified 132 patients with intracerebral hematoma secondary to ruptured aneurysms.¹³² Discriminant function analysis showed that in order of importance, size and location of the hematoma, aneurysm location, and extent of the midline shift were factors contributing to prediction of survival. Hematoma size was also a strong predictor of clinical grade. About 40% of hematomas were frontal and 40% were temporal. Patients with temporal lobe clots have the greatest capacity for clinical recovery.

Craniotomy for evacuation of hematoma is generally indicated in patients with a depressed or deteriorating level of consciousness, with or without signs of herniation. The aneurysm should be obliterated at the time of clot removal. An alternative to aneurysm clipping at the time of hematoma evacuation is endovascular coiling followed by clot evacuation.¹³³ In the majority of patients we consider this a less desirable approach because of the need for two procedures and the risk for rebleeding after clot removal before endovascular repair.

Seizures

Seizures occur at or around the time of SAH in up to 20% of patients.¹³⁴ They are not usually witnessed by medical personnel, so it is difficult to differentiate these episodes from posturing or movements precipitated by acutely increased ICP. In-hospital seizures or abnormal movements within hours of SAH are probably associated with rather than the cause of rebleeding.¹³⁵ Patients who have seizures should undergo investigations to rule out an underlying new event, such as rebleeding, cerebral ischemia, hydrocephalus, or metabolic disturbances such as hyponatremia. There are several reasons to treat seizures in patients with SAH. Seizures are associated with brain swelling secondary to increased cerebral blood volume, which may be poorly tolerated after SAH, when intracranial compliance may be low. Seizures increase cerebral oxygen consumption and may cause hypoxemia, hypercapnia, acidosis, aspiration, and pneumonia. The rise in blood pressure that can accompany a seizure may increase the risk for rebleeding. Early prevention of seizures might reduce the risk for late epilepsy by preventing kindling, although this has not generally been demonstrable clinically after SAH ^136,137 or head injury.¹³⁸ Early seizures within a week of SAH do not seem to be a risk factor for late epilepsy.^{135,136,139,140}

The incidence of epilepsy, defined as two or more seizures at least 1 week after SAH, in patients who survived after surgery for ruptured aneurysms is 3% to 10%.^{136,141–143} In one large series, 72% of seizures began within 1 year and 94% within 2 years.¹³⁹ Risk factors for the late development of epilepsy have been determined in several series and commonly include younger age, middle cerebral artery aneurysms, intracerebral or subdural hematoma, poor initial grade, postoperative focal neurological deficit with cortical infarction, medial temporal lobe retraction, history of seizures, and shunt-dependent hydrocephalus.^136,139,142

Prophylactic antiepileptic drug administration after SAH is used very arbitrarily in neurosurgical centers and is associated with increased in-hospital complications and worse outcome.⁸⁰ Our current practice is to administer antiepileptic drugs to patients who have had seizures before or associated with SAH and to poor-grade patients. Endovascular treatment of ruptured aneurysms is associated with a lower incidence of epilepsy than surgical clipping is; our only indication for prophylaxis in these cases is previous seizures.⁷¹

Fever

Fever (temperature >38.3°C) occurred in 54% of 580 patients in one series (Fig. 363-4).¹⁴¹ Other studies found temperatures higher than 38.5°C 8 days after SAH, cumulative fever burden, and fever unresponsive to treatment to be associated with a poor outcome.^146–148 Mechanisms by which fever alters the physiology of patients with SAH include increases in cerebral edema and ICP,^149,150 exacerbation of ischemia,¹⁵¹ decreasing arteriojugular difference in oxygen content,¹⁴⁹ and alteration in the level of consciousness.¹⁴⁵ Infection can be identified in 75% of febrile SAH patients.¹⁵² The remainder of fevers may be central or neurogenic in origin. Standard fever therapy with a target of normothermia includes acetaminophen, ibuprofen, and cooling blankets. Fever refractory to conventional therapy may be considered for core temperature–controlled surface or endovascular cooling therapies.^153–155 Antishivering therapies may also be used concomitantly. No randomized, controlled trials of treatment of fever after SAH have been conducted, but sufficient evidence exists at present to warrant treatment.

Anemia

Anemia was the second most common medical complication after SAH (see Fig. 363-4).¹⁴⁵ Hemoglobin levels lower than 9 mg/dL developed in 36% of 580 patients and led to transfusion. Anemia after SAH is usually attributed to the combined effects of an SAH-related reduction in red blood cell mass, immobility, phlebotomy, and hemodilution.¹⁵⁶ Conflicting data have been published on what hemoglobin level is appropriate for patients with SAH, and no randomized clinical trials have been performed. Anemia was associated with a poor outcome¹⁴⁵ and higher hemoglobin levels with a better outcome in some studies,¹⁵⁷ and transfusion was associated with vasospasm and a poor outcome in others.¹⁵⁸ These observational data are correlative, and it remains unclear whether the need for blood transfusion reflects more severe illness or whether blood transfusions themselves contribute to poorer outcomes. A transfusion threshold of 7 g/dL was at least as effective in critically ill patients as a threshold of 10 g/dL, although few patients with brain injuries were included.¹⁵⁹

There are many potential mechanisms to explain the detrimental effects of red blood cell transfusions, including alterations in vasoconstriction caused by a depleted supply of nitric oxide,¹⁶⁰ proinflammatory effects of stored erythrocytes affecting immune function,¹⁶¹ reduced deformability of transfused erythrocytes, and alterations in the ability of stored erythrocytes to bind and release oxygen, both of which lead to ischemia.¹⁶² Transfused erythrocytes are generally depleted of leukocytes to decrease the generation of interleukin-1 (IL-1), IL-6, IL-8, and tumor necrosis factor-α.¹⁶³

Until further studies evaluating the optimum transfusion threshold are conducted, decisions to transfuse should take into account such factors as vasospasm and concomitant coronary heart disease. We currently consider transfusing patients when the hemoglobin concentration is 8 mg/dL or less in the absence of symptomatic vasospasm and when the hemoglobin is less than 10 mg/dL in the presence of symptomatic vasospasm. Alternative strategies such as erythropoietin, which has been reported to possess neuroprotective properties, may be considered in the future.¹⁶⁴

Hyperglycemia

Hyperglycemia (glucose >200 g/dL) occurred in 30% of SAH patients and was a predictor of poor functional outcome and mortality at 3 months (see Figs. 363-4 and 363-5).¹⁴⁵ Predictors of hyperglycemia in one study were higher Hunt and Hess grade, elevated admission Acute Physiology, Age, and Chronic Health Evaluation (APACHE) II physiologic subscores, older age, and history of diabetes mellitus.¹⁶⁵ Strict glucose control has been associated with reductions in ICP, duration of mechanical ventilation, length of hospital stay, use of vasopressors, and frequency of seizures and diabetes insipidus in critically ill neurological patients.¹⁶⁶ Conversely, hypoglycemia is obviously potentially detrimental.¹⁶⁷ No prospective trials are available to guide treatment of patients with SAH. At present, tight glucose control (5 to 7 mmol/L) is a reasonable goal.

Respiratory Complications

Respiratory complications are more common with advancing age and poor clinical grade. They have been reported to cause almost 50% of deaths from medical causes,¹⁴⁴ although more recent studies have suggested that respiratory complications had no independent impact on neurological outcome 3 months after SAH.¹⁴⁵ The most common respiratory complications are pulmonary edema, pneumonia, atelectasis, aspiration, pneumothorax, asthma, and pulmonary emboli.

Indications for intubation and controlled ventilation in patients with aneurysmal SAH include gross abnormalities in respiratory rate, decreased tidal volume, decreased inspiratory force, hypoxia, and hypercapnia. Depressed level of consciousness is an indication because of an inability to protect the airway and clear secretions, inability to breathe deeply or sigh to prevent atelectasis, and loss of the protective pharyngeal reflexes that normally prevent atelectasis. Most grade 4 and all grade 5 patients require intubation and ventilation.

A small number of patients with SAH suffer immediate respiratory arrest or life-threatening arrhythmias from which they die unless emergency life support is instituted. In a series of 245 patients with SAH, 15% required emergency intubation and ventilation.¹⁶⁸

Pulmonary edema may complicate the course of patients with aneurysmal SAH at any time and may be cardiogenic (increased pulmonary venous pressure secondary to left-sided heart failure) or noncardiogenic (neurogenic or secondary to pulmonary insults such as aspiration and shock). Delayed cases are usually cardiogenic and caused by fluid overload during hemodynamic therapy for vasospasm.¹⁶⁹ In patients dying of SAH, pulmonary edema is more common and was diagnosed clinically in 34% and at autopsy in 71%.¹⁷⁰ Eight percent of 477 patients in one series had acute neurogenic pulmonary edema on admission.¹⁷¹ Clinical grade and volume of SAH were significantly higher than in patients without pulmonary edema, and the outcome was worse.^172,173 Survival can be achieved in about 50% of cases. A sudden increase in ICP is probably a prerequisite in all cases. SAH may be associated with marked sympathetic hyperactivity that causes systemic hypertension and pulmonary vasoconstriction. If the constriction extends to the pulmonary veins, pressure in the pulmonary capillaries will be increased and, by a hydrostatic mechanism, cause transudation of low-protein fluid into the lungs.^170,174 A second theory is that a primary increase in lung capillary permeability allows protein-rich plasma to leak into the lungs. The cause of increased capillary permeability may involve neural pathways from the central nervous system or pressure changes in the lung, thus creating some overlap with the hydrostatic theory. In any case, neurogenic pulmonary edema is characterized by rapid onset, association with severe neurological injury often involving the hypothalamus, suppression by adrenergic blockers, high protein content in the edema fluid, and resemblance to epinephrine-induced pulmonary edema. Treatment includes immediate intubation and ventilation, adequate oxygenation, positive end-expiratory pressure, furosemide, and measures to reduce increased ICP.

Most pneumonia occurs in intubated patients. Prevention of pneumonia in these cases should include removal of nasogastric and endotracheal tubes as soon as indicated, avoidance of unanticipated extubation and gastric overdistention, strict hand washing, semirecumbent positioning of the patient, maintenance of adequate nutrition, oral intubation, and proper ventilator care.¹⁷⁵ Gastrointestinal stress ulcer prophylaxis with sucralfate does not eliminate the protective antibacterial effect of acidic gastric acid secretions. Such prophylaxis may decrease gastric bacterial growth and reduce the risk for pneumonia. Exposure to antibiotics is a risk factor for ventilator-associated pneumonia because it results in colonization of the patient with antibiotic-resistant bacteria. The use of antibiotics for inappropriate or unproven indications or for prophylactic purposes should be considered carefully.

In patients with head injury, tracheostomy and percutaneous feeding gastrostomy shorten the duration of mechanical ventilation.¹⁷⁶ Tracheostomy should be considered when an SAH patient has been intubated for 10 to 14 days and extubation is not imminent.

Cardiovascular Complications

Some type of cardiovascular abnormality develops in most patients with SAH. Hypertension (systolic blood pressure >160 mm Hg) and hypotension (systolic blood pressure <100 mm Hg) are common and occur in 27% and 18%, respectively.¹⁴⁵ Other cardiovascular events include life-threatening arrhythmias (8%), myocardial ischemia (6%), and successful resuscitation from cardiac arrest (4%).

Electrocardiographic (ECG) abnormalities are very common after SAH. The spectrum of ECG changes includes ST-segment alteration, T-wave changes, prominent U waves, QT prolongation, conduction abnormalities, and sinus bradycardia and tachycardia.^177,178 These ECG changes were not associated with overall morbidity and mortality in one study.¹⁷⁷

Neurogenic stunned myocardium, a severe cardiac injury that can occur in patients with SAH, is characterized histologically by contraction band necrosis, which is a reversible cardiac pathology often found in patients who die after SAH. This histologic change is characteristic of heart muscle exposed to excessive catecholamines and intracellular calcium and leads to a hypercontracted state.

Echocardiography shows left ventricular wall motion abnormalities with moderate to severely reduced ejection fractions in 8% of patients studied within 6 days of SAH.¹⁷⁹ Echocardiographic dysfunction is more common in patients who are poor grade, have had episodes of pulmonary edema and hypotension requiring intravenous pressors, and have elevated cardiac enzymes.^179–181 Symmetrical T-wave inversion and severe QT prolongation may be specifically associated with the subpopulation of SAH patients with ventricular dysfunction.^179,182 Echocardiography is recommended for SAH patients with hypotension, pulmonary edema, acute ECG changes (Q waves, ST-segment elevation, or other changes of acute ischemia), symmetrical T-wave inversion, or severe QT prolongation. Some patients with SAH and left ventricular dysfunction have tako-tsubo, or apical ballooning syndrome, which shares some features with the general population of SAH patients with ventricular dysfunction.¹⁸³

Elevated troponin I has been reported in up to 68% of SAH patients.^180,184 Levels peak about 2 days after SAH. Left ventricular wall motion abnormalities seen on echocardiography are more common in patients with elevated troponin, and elevated troponin levels are associated with disability and death after SAH, as well as worse neurological grade, systolic and diastolic cardiac dysfunction, and pulmonary congestion.^180,184

Patients with SAH should have continuous monitoring of cardiac rhythm, as well as baseline and follow-up ECG studies and frequent monitoring of electrolytes, particularly potassium. Atrial flutter and fibrillation occur in up to 4% of patients and are associated with poor outcome.¹⁸⁵ Hypokalemia should be avoided because it may aggravate arrhythmias and has been associated with a prolonged QT interval, ventricular fibrillation, and torsades de pointes,^186,187 and echocardiography should be performed in high-risk patients to optimize care.

Venous Thromboembolism

Symptomatic deep venous thrombosis (DVT) develops in about 2% of patients with SAH, although a negative impact on outcome was not shown in one study.¹⁴⁵ Asymptomatic DVT occurs in 19% to 50% of patients undergoing neurosurgical procedures. Risk factors for venous thromboembolism include increased age, heart failure, previous venous thromboembolism, direct trauma to the lower limbs, varicose veins, use of oral contraceptives, pregnancy and the puerperium, obesity, malignancy, infection, duration of surgery longer than 4 hours, and limb weakness or paralysis.^188,189 Patients with deficiencies in antithrombin III, protein C, or protein S and with various genetic clotting factor abnormalities, such as factor V Leiden, are also at risk for venous thromboembolism.¹⁸⁹ Graduated compression stockings and pneumatic compression devices decrease the incidence of DVT in neurosurgical patients, as detected by the radiolabeled fibrinogen technique, and are recommended for all patients.^190,191 They work by intermittently squeezing the lower limbs, compressing the veins, and preventing thrombosis from venous stasis. There is evidence that they induce a systemic fibrinolytic state. Other proven methods of prophylaxis include low-dose subcutaneous heparin, adjusted-dose subcutaneous heparin, low-molecular-weight heparin, and oral anticoagulants. It is reasonable to assume that these drugs have a low therapeutic index in patients undergoing craniotomy. As the dose is increased, the risk for venous thromboembolism will decrease, but the risk for catastrophic intracranial bleeding will increase. We routinely place patients on a regimen of low-molecular-weight heparin 24 hours after aneurysm obliteration.

If DVT or pulmonary embolism develops, options include anticoagulation or placement of an inferior vena cava filter. Anticoagulation is usually considered to be safe when a week or more has passed since the craniotomy, although this is based on a very limited number of patients described in the medical literature.¹⁹² A filter may still be a better temporizing measure if the need for other invasive procedures, such as further angiography or shunting, has not been determined. Filters reduce the risk for pulmonary embolism, although whether they reduce mortality is less certain.¹⁹³

Fluid and Electrolyte Disturbances

Euvolemia should be the goal in a patient with SAH in the absence of vasospasm. Daily fluid balance should not be greater than 500 mL positive after correcting for hypovolemia. Normal saline or lactated Ringer’s solution administered at a rate of up to 1.5 mL/kg per hour is recommended. Fluid management in patients with vasospasm is discussed in Chapter 364.

The most common electrolyte disturbances reported after aneurysmal SAH are hyponatremia (14% to 40%), hypomagnesemia (37%), hypokalemia (27%), and hypernatremia (20%).^{145,166,194,195} Hyponatremia may result from the syndrome of inappropriate antidiuretic hormone (SIADH), cerebral salt wasting, or a combination of the two.¹⁹⁶ Opinions differ on the relative frequency of each, and in a series of 316 patients with spontaneous SAH, it was reported that of the 57% in whom hyponatremia developed, the cause was SIADH in 70%, salt wasting in 7%, and hypovolemic hyponatremia (some probably had salt wasting) in 21%.¹⁹⁶ Hyponatremia is not thought to have major prognostic significance.^145,195 The symptoms and signs of hyponatremia include deterioration in the level of consciousness, onset of or exacerbation of focal neurological deficits, seizures, and asterixis.

The first principle of fluid and sodium management is to make an accurate diagnosis (Table 363-8). The key differentiating features between SIADH and salt wasting is volume status. Acute change in body weight is a good measure of volume status, but there are numerous others, including changes in hematocrit, the blood urea nitrogen–to-creatinine ratio, and cardiac filling pressure.¹⁹⁷ The primary treatment of salt wasting is adequate water and sodium replacement to maintain at least normovolemia and normal serum sodium. Natriuresis can be prevented by administering mineralocorticoids such as fludrocortisone acetate, but a randomized, controlled trial showed that this had no effect on the risk for cerebral ischemia or outcome, probably because administering large quantities of sodium and water is equally effective.^198,199 Volume replacement should be guided by measures of total body water such as weight. Ventricular filling pressures are less reliable. In patients with SIADH, fluid restriction with careful management during the vasospastic period may be appropriate. An alternative approach when fluid restriction is potentially detrimental includes the judicious use of hypertonic saline and salt supplements. SIADH should be reversed slowly because rapid salt and water replacement may lead to fatal osmotic demyelination.²⁰⁰

Diabetes insipidus complicates the course of about 0.04% of patients with aneurysmal SAH.²⁰¹ Most cases are associated with ruptured anterior communicating artery aneurysms. Treatment consists of replacement of fluid losses with hypotonic solutions such as 5% dextrose in water or 0.45% sodium chloride. In the chronic phase, exogenous replacement of antidiuretic hormone may be required.

Magnesium is a potentially neuroprotective agent that acts as a glutamate receptor antagonist and calcium channel blocker. Magnesium showed a trend toward a reduced risk for delayed cerebral ischemia in 283 patients with SAH randomized to magnesium or placebo administration.²⁰² Another randomized, placebo-controlled, dose-adapted trial failed to demonstrate a difference in the incidence of delayed ischemic neurological deficits or secondary ischemia.¹⁶⁷ Larger phase III clinical trials are ongoing.²⁰³

Gastrointestinal Complications

There is a marked increase in urinary catecholamine excretion within the first 3 days after SAH that is probably due to sympathetic hyperactivity. This, as well as increased cortisol, glucagon, and cytokine release, probably contributes to the increased oxygen consumption, carbon dioxide production, and metabolic expenditure noted in SAH patients.²⁰⁴ In addition to a hypermetabolic state, patients with SAH have a catabolic response marked by negative nitrogen balance and impaired ability to use exogenous nitrogen. These changes are similar to those found in patients with other acute, severe neurological illnesses such as head injury.²⁰⁵ This response causes weight loss and may increase the risk for infection and impair wound healing.²⁰⁶ The primary therapy to counteract it is early nutritional support. Few studies have investigated this treatment in patients with SAH, but after stroke there is evidence that early institution of nutritional support improves survival.²⁰⁷ In patients who cannot swallow because of impaired consciousness or focal neurological deficits, nutritional support is best achieved by gastrostomy or jejunostomy tubes rather than by nasogastric tubes.²⁰⁷ It is probably preferable to feed enterally instead of parenterally because parenteral nutrition may be associated with loss of the intestinal mucosal bacterial barrier and thus increased risk for sepsis. Enteral feeding avoids the risk for sepsis from central venous catheters, reduces the incidence of hyperglycemia, and is less expensive. Parenteral nutrition is indicated for the occasional patients who cannot meet their nutritional needs with enteral feeding and in whom this situation is likely to persist for at least 5 days.

Mucosal ulceration (stress ulceration) in the stomach and upper gastrointestinal tract may occur after SAH and can cause serious and sometimes fatal bleeding. Studies specific to patients with SAH are limited but suggest that such lesions occur after SAH just as they do after other serious neurological injuries. Accepted treatment includes neutralization of gastric acid secretion with antacids or drugs that block acid production, such as histamine receptor antagonists (cimetidine, ranitidine) or proton pump inhibitors (omeprazole).²⁰⁸ The use of drugs to prevent gastrointestinal bleeding remains controversial, and recent reviews suggest that prophylaxis can be limited to patients at increased risk, such as those ventilated for more than 48 hours or those with coagulopathy.²⁰⁸ These studies may not apply to brain-injured patients, so withholding prophylaxis is controversial.

Stool softeners are administered routinely to prevent straining, impaction, and abdominal pain. Nimodipine may be associated with gastrointestinal pseudo-obstruction and ileus.²⁰⁹ This usually resolves with conservative measures but may progress to perforation and require surgery.