Chapter 1: Subarachnoid Hemorrhage

The clinical and radiographic presentation of this case is consistent with high-grade (initially Hunt and Hess [HH] grade IV, then quickly progressing to grade V in transit to the tertiary care center) acute SAH. Airway, breathing, and circulation (ABC) have all been addressed, although the blood pressure is high at this time. The very first step in managing this patient is ventricular drain, the second step is ventricular drain, and the third step is ensuring that the ventricular drain you have just placed is working (ie, draining the hemorrhagic cerebrospinal fluid [CSF] adequately when the drain is kept open, and maintaining good waveforms when the drain is clamped). After ABC, placing external ventricular drain (EVD) is the most crucial, lifesaving, important early step for managing the patients with high-grade acute SAH with poor mental status and IVH. The presence of IVH complicates the natural course of both intracerebral hemorrhage (ICH) as well as SAH cases. IVH is often associated with development of an acute obstructive hydrocephalus, which may lead to vertical eye movement impairment and depressed level of arousal by its mass effect on the thalamus and midbrain. IVH is also associated with elevated intracranial pressure (ICP), which lowers the cerebral perfusion pressure (CPP) (by the principle of the equation, CPP = MAP − ICP) if the mean arterial pressure (MAP) remains constant. IVH has also been reported to be an independent risk factor for increased risk of developing symptomatic vasospasm. The mass effect and cerebral edema may rapidly progress to herniation syndrome and death. As such, the presence of IVH has been recognized as a significant risk factor of poor outcome for both ICH and SAH. ^1-3 Placing an EVD provides twofold benefits: (1) reliable (as long as the catheter tip is in the right location providing appropriate ICP waveforms without obstructing the ventricular catheter by any blood clot) measurements of the ICP, and (2) therapeutic drainage of the CSF in order to alleviate the intracranial hypertension (Figure 1-2).

It is important to note that the presence of IVH does not necessarily mean the ICP is abnormally elevated, and the placement of an EVD alone may not always lead to improved outcome even if the high ICP responds favorably to opening and lowering of the drain. ⁴ In the past, there were concerns regarding the potential harmful effect of EVD placement in treating the acute hydrocephalus in SAH cases. These concerns were mainly focused on the theoretical impact of suddenly lowering the ICP by EVD placement and eliminating the tamponade effect on ruptured aneurysmal wall, leading to an increased risk of rebleed in the acute phase. However, the clinical studies have failed to prove such a hypothesis and there is not sufficient evidence to believe that the CSF diversion by an EVD in treating acute hydrocephalus after SAH leads to a higher incidence of rerupturing of the unsecured aneurysms. ^{5, 6} It is wise, however, to avoid aggressively lowering the drain level immediately after placement. In managing SAH cases, whether to place an EVD or not is occasionally debatable. For instance, a patient with good-grade (eg, HH I or II) SAH who is awake, following commands with normal strength with no IVH, no acute hydrocephalus, and either absent or minimal volume of SAH (eg, classic Fisher groups 1 to 2) is not a candidate for EVD placement. On the other hand, a patient with a high HH grade, Fisher group 3 SAH, plus the radiographic evidence of severe IVH and acute hydrocephalus who is progressively getting worse in the level of arousal needs emergent placement of an EVD. These are extreme ends of the clinical spectrum of SAH, and the timing and indication for EVD could be debated for the cases that are somewhere in between these two extreme case scenarios. Acute hydrocephalus with IVH and clinical signs and symptoms of intracranial hypertension are all good indications for placing EVDs. It is also important to remember that even if the patient does not have any of the indications mentioned above, if the treating physician believes that there is a reasonable probability of developing these signs and symptoms in the near future, EVD placement should be considered. (Technical details and further management strategies are discussed in Chapter 22.) Despite the lack of "level 1" evidence of randomized data for improved outcomes, the use of an EVD is important as it can be helpful in managing ICP and CPP and is often lifesaving in certain SAH patients.

Changing neurologic status after EVD placement

Placement of an EVD frequently results in a significant improvement in neurologic status. Comatose patients may start to localize to painful stimulation and may even open their eyes. Although this is not always seen, when it happens, it may possibly indicate a favorable outlook (eg, a patient, who presents with HH grade V after aneurysmal SAH, wakes up after EVD placement and begins to follow verbal commands: if such patient remains awake and continues to follow commands throughout the course of his/her illness, then the patient is behaving like a low-grade HH [ie, grades I to III], not like a grade V who presents and remains in coma).

Patients with HH grade V have extremely poor prognosis. Many physicians and surgeons disclose such a poor prognosis to the patient's family and this often leads to withdrawal of life-sustaining care prior to any treatment. While the decision of treating versus not treating should be made based on the prognosis and for the best interest of the patient, the initial prognosis is mostly based on the bedside neurologic assessment, and physicians should be aware that the patient's clinical status may dramatically change after placing an EVD, which has significant implications for the prognosis. ⁷

There are several SAH grading systems worth mentioning here. In 1967, Hunt and Hess have reported 275 consecutive patients who were treated at the Ohio State University over a 12-year period. They believed that the intensity of the meningeal inflammatory reaction and the severity of neurologic deficit and the presence or absence of significant systemic disease should be taken into account when classifying SAH patients. From the original manuscript, their grading system (which is now known and widely used as the Hunt and Hess Grade) was a classification of patients with intracranial aneurysms according to surgical risk (Table 1-1). ⁸

Higher grades are associated with increased surgical risk for the repair of ruptured intracranial aneurysms. The Hunt and Hess original report included the presence of significant systemic disease (such as "hypertension, diabetes, severe arteriosclerosis, chronic pulmonary disease, and severe vasospasm seen on angiography") as a negative sign, and the presence of such disease resulted in placement of the patient in the next less favorable (higher surgical risk) grading category. ⁸ This grading system is not flawless as it can be challenging sometimes to differentiate between each category. For example, consider a patient with SAH with mild headache and nuchal rigidity versus another patient with moderate headache and nuchal rigidity (which means grades I and II, respectively, according to the original HH grading system). The only differentiating variable here would be the intensity of the headache, and that can be problematic as the intensity of headache is subjective and patients often cannot differentiate mild from moderate headache (most people would say "very bad" headache and cannot provide further details than that).

This criticism has been actually predicted, and the original authors have mentioned it in their journal article: "It is recognized that such classifications are arbitrary and that the margins between categories may be ill-defined." ⁸ For this reason, it has been pointed out that the HH system has poor interobserver reliability and reproducibility. ⁹ Nevertheless, the HH grading system is widely used and numerous studies have shown that the higher grade (or sometimes called poor grade, which usually refers to HH grades IV and V) is associated with a poor outcome. ^10-13

Another grading system to consider is the one that is the most universally accepted system for patients presenting with altered level of consciousness, the Glasgow Coma Scale (GCS). In 1975, Jennet and Bond, from the University of Glasgow, reported a scale called Assessment of Outcome After Severe Brain Damage, a Practical Scale (Table 1-2). ¹⁴

The GCS is a more general grading system and was not developed specifically for SAH patients. However, studies show that for patients with aneurysmal SAH, the initial GCS score has positively correlated with long-term outcome. ¹⁵

In 1988, the World Federation of Neurosurgical Societies (WFNS) developed a grading system that incorporated both the GCS and bedside neurologic assessment focusing on any focal deficit (Table 1-3). ¹⁶

The HH and WFNS grading systems are by far the two most commonly used systems for grading patients with acute aneurysmal SAH. Despite the frequently raised criticisms regarding the inter-observer variability, the HH grade is used even more commonly than the WFNS scale (71% of reported studies from 1985 to 1992 used the HH grade compared to 19% that used the WFNS scale), ^{17, 18} and both grading systems have been shown to correlate reasonably well with the long-term outcome. ¹⁹

In 1980, Fisher and colleagues reported the relationship between the amount of SAH and risk of developing severe vasospasm (defined as delayed clinical symptoms and signs, Table 1-4). ²⁰

The Fisher group's grading system is based on the description of CT findings mainly focusing on the actual volume of blood in the subarachnoid space. There is a linear relationship between the amount of hemorrhage and the rate of developing symptomatic vasospasm. ²⁰ This grading system has been extensively studied and there are numerous clinical studies validating its usefulness. ^21-25 In multiple studies, the risk of developing symptomatic cerebral vasospasm appears to increase along with the increasing amount of acute hemorrhage in the subarachnoid space. The original report by Fisher et al does describe the low risk of vasospasm, and yet there is a clearly observed risk of vasospasm even for patients with minimal blood in the subarachnoid space and for those with intraparenchymal or intraventricular hemorrhage. ²⁰

It is important to understand that the Fisher scale actually did report some incidence of vaso-spasm in groups 1, 2, and 4. The group 3 had the highest incidence of vasospasm, but other groups also had vasospasms, just much lower in frequency. ²⁰ Like all other grading systems, the Fisher scale is not without limitations. There have been concerns in the literature reporting a low correlation between the Fisher grade and the incidence of symptomatic vasospasm (one of the recent studies showed about 50% correlation between the Fisher grade and vasospasm). ²⁶ Another criticism about the Fisher scale is its inevitable interpresonal variability in assessing the estimated blood volume. Also, according to the scale, all cases of CT head showing SAH with greater than 1 mm of vertical thickness is categorized as grade III, but this includes vast majority of patients with SAH who may not in fact have the same risk of developing vasospasm. ^{26, 27}

In light of these concerns, Claassen et al's group, from Columbia University, proposed another grading system (Table 1-5), the modified Fisher scale (mFS). ^{28, 29}

Note that the mFS incorporates the presence or absence of IVH, and if a patient has IVH, even if there is no blood in the subarachnoid space, the scale is 2 (as opposed to 1 [no blood seen] or 4 [minimal SAH and the presence of intraparenchymal hemorrhage or IVH] in the original Fisher scale). This scale emphasizes that the presence of IVH increases the risk of developing symptomatic vasospasm. This emphasis is stronger but not completely different from that of the Fisher scale as the original Fisher scale does report some incidence (although low) of vasospasm in those with IVH and absent or minimal SAH. Furthermore, the mFS uses a subjective description and coding of the hemorrhage by the use of "thick" or "thin" clots in the subarachnoid space, and the description of IVH did not take the exact amount of IVH into account (this scale takes the "presence" versus the "absence" of IVH into account, not how much IVH there is). The mFS emphasizes the importance of IVH as well as it highlights how the amount of hemorrhage once again plays an important role. Its grading system is easy and intuitive (unlike the classic Fisher scale in which group 4 actually has a lower incidence of vasospasm than lower grades), as the scale goes from 0 to 4, and the higher grade is the higher risk of developing delayed cerebral ischemia (DCI).

In order to minimize the interobserver variability in assessing the estimated volume of blood in the subarachnoid space, a volumetric quantification of Fisher grade 3 has been proposed and studied by Friedman and colleagues from the Mayo Clinic. ³⁰ However, while quantification of SAH may provide a more accurate assessment of the volume of blood in the subarachnoid space, it requires manual outlining of the hemorrhage volume, which can be time consuming and less reliable.

In 2011, Ko and colleagues, from Columbia University, reported a study of volumetric analysis of SAH using a MIPAV (Medical Image Processing, Analysis, and Visualization; version 4.3; National Institutes of Heath, NIH) software package that automatically outlines the hemorrhage on CT at the click of a button. ³¹ This quantification analysis showed that patients with a higher volume of cisternal plus IVH clot burden developed a greater risk of developing DCI and poor outcome at 3 months (Figure 1-3). It also validated the modified Fisher scale as a reasonable grading system in predicting DCI that can be done easily at the bedside. However, it is important to note that although both the Fisher scale and the mFS have demonstrated the association between blood burden and DCI, a question still remained: Does the location and exact thresholds of blood volume matter? Ko and colleagues have reported:

Klimo and Schmidt have eloquently summarized a historical review of the literature on the relationship between the CT findings and the rate of developing cerebral vasospasm after aneurysmal SAH using different scales ³² :

The early phase of high-grade SAH is often complicated by the presence of ICP crisis. An ICP value out of the normal range (0 to 20 mm Hg) is considered abnormal, but the ICP alone as an absolute value may not always signify the need for an urgent treatment. A good example would be people with pseudotumor cerebri and high ICP but having normal daily activities. ICP also rises when patients cough or get suctioned. Such a rise, if it is induced and transient, does not necessarily need any treatment. In the setting of acute, high-grade SAH, however, abnormally elevated ICP is a major concern owing to its direct, negative impact on the cerebral perfusion pressure (CPP). With persistently low or decreasing CPP, a certain degree of ischemic insult is inevitable. A step-by-step algorithm for managing refractory ICP crisis is outlined below. This is a recommendation that reflects the latest medical treatment available in the literature.

With all other factors that are known to influence the ICP being constant, increasing the MAP leads to increased cerebral perfusion pressure. Increased cerebral perfusion pressure and blood flow may result in increased cerebral blood volume. Increased cerebral blood volume could lead to any one of the following scenarios:

1. Development of vasoconstriction would occur if autoregulation is intact in response to increased blood volume delivered. Such vasoconstriction may result in drop in ICP.

2. No vasoconstriction would occur if autoregulation is impaired (pressure-volume breakdown). Since blood volume is increased, there may be further rise in the ICP.

3. Mixed picture of all of the above as different parts of the brain have either impaired or intact autoregulation at a point prior to the pressure-volume breakdown.

It is important to emphasize that increasing the blood pressure with phenylephrine may either make the ICP go up or down. Clearly, if the MAP is too low to the point that the patient is in a state of shock, it makes sense to increase the MAP in order to maintain adequate end-organ perfusion. However, if the MAP is already 90 mm Hg, then increasing the MAP to 130 mm Hg with Neo-Synephrine may have a deleterious effect on the ICP. Finally, it is important to remember that every effort needs to be made to avoid the worst combination that is associated with devastating results: systemic hypotension and uncontrolled ICP.

(Figure 1-4)

The cardiac output (CO), hemoglobin concentration [Hgb], and the arterial saturation of oxygen determine the delivery of oxygen from the heart to all the end-organ systems, including the brain. The Pbto ₂ is significantly affected by the Do ₂ and the blood flow to the brain. Flow, by definition, is blood volume per given time. Therefore, there is a positive correlation between the blood volume reaching the brain tissue and Pbto ₂ measurements. Do ₂ has a linear relationship with the CO and Hgb levels as well as Sao ₂ (arterial oxygen saturation).

The oxygen content in the venous return from the brain back to the heart (measured by Sjvo ₂) depends on the Do ₂ and Vo ₂ (oxygen consumption). The product of the Do ₂ and O_EF (oxygen extraction fraction) determines the total Vo ₂:

Therefore, the oxygen extraction fraction is determined by the total consumption divided by the delivery of oxygen:

Thus, the oxygen extraction decreases when the delivery increases while the consumption remains the same. As the consumption of oxygen decreases, the extraction decreases as not as much oxygen is needed to be extracted from the vasculature into the brain cellular tissues. Understanding this relationship is crucial in applying this principle to the bedside information. If a patient has Pbto ₂, Sjvo ₂, and CO monitoring, one would observe that by increasing the cardiac output significantly, assuming the brain cellular metabolism is constant (eg, under pentobarbital coma for ICP control), the O_EF will decrease, and as a result Sjvo ₂ increases.

As one can see above, the brain cellular metabolism plays an important role in brain oxygen consumption, and therefore extraction, and affects the Pbto ₂.

There are ongoing debates regarding whether the partial pressure of oxygen tension measured in the brain means anything in brain-injured patients (or what it means short- or long-term). Frequently encountered criticisms are focused on the fact that the location of the probe may not be in the perfect location for monitoring the injury. The probe provides direct information, but it only provides the local information, and whatever the number it gives may not be applicable to the rest of the brain. It is also criticized that the absolute value of Pbto ₂ may not always correlate with the long-term outcome, making it difficult to determine what the critical or "dangerous" value is. From the traumatic brain injury literature, there are studies that have shown a positive correlation between the Pbto ₂ level and mortality and outcome both in pediatric as well as in the adult population. ^{39, 40} As long as the probe is not located in the isolated, dead brain tissue (eg, in the middle of a completed infarct, which will provide very low values of Pbto ₂, but that does not necessarily mean the rest of the brain is critically ischemic), it is reasonable to believe that persistent, severe (Pbto ₂ < 15 mm Hg) brain hypoxia is probably a bad sign.

As you can see above, significant brain hypoxia and neurochemical metabolic stress (lactate/pyruvate > 40) can be treated with a number of different methods. The basic mechanism of how these methods improve the Pbto ₂ and LPR is by increasing the Do ₂ and the flow to the brain. Another factor, brain cellular metabolism, plays a role in determining the Pbto ₂ as it influences the consumption and extraction of the delivered oxygen supply. Figure 1-5 illustrates how to address the factors that may contribute to the brain hypoxia and metabolic stress. A number of factors should be considered for estimating the brain cellular metabolism: fever, shivering, Co ₂ calorimetry for energy expenditure, pain/agitation, sedation/paralysis, and other variables that may affect the basic oxygen consumption and demand.

Transcranial Doppler (TCD)

TCD has been used for many years for monitoring SAH patients who are at risk for developing vaso-spasm. The TCD velocities rise in the segments of brain vasculature where vasospasm occurs. A mean flow velocity greater than 120 cm/s is commonly used as a reference point for "TCD spasm." ^{41, 42} Before firmly declaring whether TCD is useful or not, one should be aware of the frequent criticisms regarding the sensitivity and specificity of high velocities detected on TCD:

1. False positive: When the flow velocities rise, this does not necessarily mean the presence of true cerebral vasospasm. Hyperemia, a common cause of elevated velocity on TCD, is due to a number of different factors such as fever, hypervolemia, and even may be part of triple-H (hypertensive, hypervolemic, hemodilution) therapy. Consider a SAH patient who is getting 250 mL/h of normal saline infusion plus albumin 250 mL q6h for so-called prophylactic triple-H therapy for Fisher 3 with "lots of thick cisternal blood." This is bad medicine for two reasons: (1) prophylactic triple-H does not reduce the incidence of symptomatic vasospasm or improve outcomes, and may increase pulmonary complications (before saying that the heart and lungs can be abused in order to save the brain, we need to consider the following question: what would happen to the delivery of oxygen to the injured brain if we neglect to keep the lungs alive?) and (2) hypervolemia leading to high TCD velocities only increases the false-positive rate without having true cerebral vasospasm. This false rate then leads to more unnecessary and potentially harmful interventions. Hyperdynamic therapy may also increase the TCD velocities. There are physicians who utilize positive inotropic agents (eg, milrinone, dopamine, dobutamine, norepinephrine) and have a specific cardiac index goal for triple-H therapy. If you increase the cardiac output significantly, this will improve the flow, and an increase in flow leads to higher velocities. A recent article reports a poor correlation between the TCD spasm with clinically evident (symptomatic) spasm: "… of those with TCD spasm, only 28% had symptomatic spasm, and 34% had delayed cerebral ischemia." ⁴¹ Delayed cerebral ischemia (DCI) is defined as the presence of symptomatic vasospasm, infarction attributable to vasospasm, or both. ⁴³

2. False negative: The TCD is notorious for missing vasospasm occurring at the distal anterior cerebral artery (ACA) areas. The distal ACAs are technically challenging to insonate as they are far away from the TCD probe to navigate (typically sonographers try to insonate the top of carotid, M1 A1, then A2, then A3, and so on). As the TCD probe follows the proximal ACA to the distal segment, there may be an increased amount of noise and other vessels interfering, which leads to poor reception of the distal flow. The TCD is user-dependent. As such, there may be significant interobserver variability.

Despite these well-known and documented limitations, TCD is widely used and can be useful in caring for patients with SAH. The statistical evidence of significantly high false-positive and false-negative rates of the TCD is not surprising: The TCD only detects the flow velocities, and from the waveform analysis, a number of different pulsitility indices may be obtained. Any narrowing in a vessel would lead to a higher velocity. Consider the following clinical scenario: A sonographer insonates a major vessel (eg, the M1 segment of the middle cerebral artery [MCA]), and the waveform is consistent with an M1, and the signal is strong and consistent with an M1. After documenting normal velocities of a SAH patient every day for 7 days, all of a sudden the same segment's velocity rises by 200% (or steadily increasing velocity over time); this is more likely to be true vasospasm, meaning the M1 segment's cross-sectional caliber is not as large as it used to be. This is particularly true in the absence of metabolic changes such as, for example, fever or hyperemia, and if the TCD has been performed by the same sonographer. In order to minimize the false-positive rate of TCD due to fever, hyperemia, and hyperdynamic status, a Lindegaard ratio (LR) has been described: LR = mean TCD velocity of MCA/mean TCD velocity of ipsilateral, extracranial internal carotid artery. A sudden rise in MCA mean velocity out of proportion to the rise in ipsilateral carotid artery mean velocity argues against hyperemia-induced, false-positive vasospasm. A normal LR is < 1.7, and a value > 2.0 might indicate higher risk of vasospasm. Different institutions use different threshold points of LR for categorizing TCD spasm. An example: LR < 1.5 normal, 1.5 to 2.5 hyperemia, 2.5 to 3.5 mild vasospasm, 3.5 to 4.5 moderate vasospasm, > 4.5 severe vasospasm. The literature lacks in terms of which of these LR cutoff points are more accurate. The criticism regarding the low correlation between high velocity and high LR on TCD with symptomatic vasospasm or DCI is also not surprising: just because the vessels are narrow, that does not mean the patient would become symptomatic neurologically. This all depends on the brain's margin of tolerability toward the ischemia. The cerebral ischemic cascade is complex, and the compensatory mechanisms, in the setting of ischemia due to vasospasm, leads to a variety of clinical findings: from normal to focal deficit. A high cerebral vessel velocity and LR on the TCD and yet no symptoms does not mean the TCD did not work. It could mean that the brain's compensatory mechanism is working and therefore no symptoms. Is everyone with narrow vessels on cerebral angiograph symptomatic? Of course not. Does that mean the angiograph lies and is useless? No. The TCD can be useful as long as one can interpret it appropriately and understands its limitations. Another aspect regarding the TCD is that it is more useful when the trend is followed, rather than focusing on any absolute value at any given day.

CT angiography (CTA) and CT perfusion (CTP)

CTA can provide a rapid, noninvasive method of visualizing the brain vasculature in patients suspected of having vasospasm (Figure 1-6). CTA can provide 2-D and 3-D images of the vessels and can be useful in detecting vasospasm and brain aneurysms without the risks that are associated with conventional angiography. It should be noted that the limitations of the CTA is that the false-negative rate can be significant for small brain aneurysms, particularly those that are < 3 mm in diameter. With the advent of the 64-section, multidetector CTA, however, the sensitivity of detecting small aneurysms, even for those that are < 3 mm in diameter, can be as high as 92%. ⁴⁴ Recently, even higher resolution may be obtained with more than 300-slice, fine-cut CTA softwares. CTA can also be helpful in following patients for suspected vasospasm. CTA is not without potential risk, especially in elderly patients who might be intravascularly volume depleted or those with baseline renal impairment. Efforts should be made to prevent contrast-induced nephropathy (CIN). One single-center, randomized trial revealed sodium bicarbonate showing a significantly reduced rate of CIN compared to sodium chloride infusion. ^{45, 46} However, relevant data in the literature are mixed and the renal protective effect in preventing the CIN appears to be secondary to the additional volume, rather than the actual content of the IV fluid. The following is suggested protocol for preventing the CIN for high-risk ICU patients who need to undergo CTA or conventional angiography.

1. Sodium chloride 1 mL/kg/h for at least 12 hours, or

2. Sodium bicarbonate 3 mL/kg/h for at least 12 hours, and/or

3. N-acetylcysteine, 600 mg IV prior and then 600 mg po bid postprocedure for 2 days.

CTP (Figure 1-7) provides three different perfusion maps.

1. Mean transit time (MTT) map: Time (in seconds) measured for the contrast material to reach the cerebral hemispheres. In the event of vasospasm, the areas that are supplied by the spasm vessel will have prolonged MTT.

2. Cerebral blood flow (CBF) map: The blood volume per given time. The areas that are supplied by the spasm vessel will have corresponding reduced CBF (usually the delayed MTT and reduced CBF occur concurrently).

3. Cerebral blood volume (CBV) map: The blood volume in the event of vasospasm shows either normal or increased blood volume if the cerebral autoregulation is intact (by compensatory self-dilatation of the brain vessels). This is the window of opportunity for appropriate medical and endovascular interventions. When and if the CBV map shows reduced volume, this may indicate complete infarction and hence irreversible ischemic injury.

Continuous Electroencephalography (cEEG)

cEEG may be helpful in detecting vasospasm in high-grade SAH patients in whom clinical examinations may be of limited use. Recent studies have shown the quantification of relative alpha (RA) signals on cEEG to be associated with detecting vasospasm—a positive association between the poor variability of an alpha rhythm on EEG and symptomatic vasospasm—remarkably, the onset of symptoms due to vasospasm may even be preceded by the change in RA on cEEG. ^{47, 48} Further details are to follow in the EEG section later. Despite the small sample sizes, studies report a rather high positive predictive value and a false-negative rate. cEEG monitoring is required for all high-grade SAH patients in order to be vigilant about the risk of nonconvulsive status epilepticus. Therefore, the use of cEEG in detecting vasospasm can be used in conjunction with other diagnostic tests in order to increase the sensitivity and specificity, which may have implications on the timing of performing angiography. ⁴⁹

Any intensivist managing aneurysmal SAH patients in an ICU must be aware of when to call for a conventional angiogram. The quick answer as to how to address this patient's acute hemiparesis and obtundation is stat angiogram. It is important to acknowledge that this is a neurologic emergency. A longer delay in getting the angiogram leads to a higher risk of irreversible ischemic damage. It is not wise to delay the angiogram with the clinical scenario provided above. Clearly, cerebral angiography is not a completely benign procedure. However, in an aneurysmal SAH patient, on day 7, who becomes acutely hemiparetic and obtunded with TCD, EEG, CTA, and CTP evidence of severe vasospasm and hypoperfusion, there should not be any delay in providing an angiogram. For clinically obvious cases, even obtaining CTA prior to conventional angiogram may not be necessary and only delay an effective treatment. There is a role for obtaining CTA/CTP if the index of suspicion is not high enough for severe vasospasm (ie, one can avoid getting an angiogram [which carries a potential risk] if the patient does not really have treatable vasospasm).

Triple-H Therapy

Triple-H therapy has been used for many years to improve and maintain adequate brain perfusion in patients with symptomatic vasospasm. Considering that SAH patients frequently tend to develop natriuresis and intravascular volume depletion, it makes physiologic and intuitive sense to provide adequate volume and avoid dehydration. This should be even more emphasized in the event of cerebral vasospasm and delayed cerebral ischemia as intravascular volume depletion could significantly contribute to devastating ischemic injuries. Hypervolemic therapy, when used prophylactically (ie, from bleed day 0 until day 14 in aneurysmal SAH regardless of the data from TCD, CTA, or neurologic examination), however, has failed to show improved CBF or CBV on a small (N = 82) but randomized clinical trial. In that study, the rate of developing symptomatic vasospasm was the same for both hypervolemic and normovolemic groups. ⁵⁰ There is a wide variety of clinical practice patterns regarding the use of triple-H therapy, but prophylactic hypervolemic therapy is not supported by any good evidence for any added benefit. Again, this does not mean that one should not provide adequate intravascular volume. It simply means, for a postoperative (clipping or coiling) SAH patient on bleed day 1, who is doing well without any signs of vasospasm, in an euvolemic state, adding additional fluid boluses is not likely to provide any additional benefit other than increasing the urine output. During the period of symptomatic vasospasm, however, nonrandomized studies indicate that additional fluid boluses can increase the CBF in the hypoperfused area even if the patient is considered to be in a euvolemic state. ⁵¹ Current diversity in clinical practice in using the triple-H therapy is secondary to the lack of "level 1" evidence in terms of when and how to use the therapy. However, it is not wise only to criticize the lack of data and not provide treatment. Until there are multiple randomized studies confirming what should be the gold standard, clinicians should be treating the patients with the best available knowledge, evidence, and, most of all, good judgment (because knowledge and evidence-based medical literature alone are often not sufficient to make a good decision).

One may conclude:

1. There are randomized data arguing against the use of prophylactic hypervolemic therapy in patients with no signs of vasospasm in SAH. ⁴⁹

2. There are no randomized data for use of triple-H therapy in improving the long-term outcome in SAH (regardless of whether triple-H therapy was used as a prophylaxis or as a treatment).

3. There are nonrandomized, observational studies showing augmentation of CBF in hypoperfused brain regions after triple-H therapy (or different parts of the triple-H therapy: ie, hypervolemic therapy alone [normal saline boluses], ⁵¹ or hyperdynamic therapy [increasing cardiac index without significantly increasing the blood pressure], ⁵² or hypertensive therapy alone, or a combination of all of the above).

The following is a reasonable algorithm for managing patients with risk of developing symptomatic vasospasm and delayed cerebral ischemia.

Cerebral Angiography

(Figure 1-8)

There are a number of different vasodilators available for IA therapy of vasospasm in patients with aneurysmal SAH. These include papaverine, nicardipine, verapamil, and more recently milrinone. These vasodilators often produce an immediate result in increasing the vessel calibers, but there is a limitation: the positive effect may not last long. It is not uncommon to see patients with vasospasm who respond to initial IA vasodilators become symptomatic the same day or the next day, requiring further multiple angiographic treatments. Unless the interventional suite is located in the middle of the NeuroICU with the team actually staying in the angiography suite 24/7, there may be some delay between the redetection of the symptomatic vasospasm and the actual time of reangiography. If the patient has aspiration pneumonia on a high-maintenance ventilator setting with multiple vasopressors, the presence of an anesthesiologist is needed for reangiography, and this can be another source of delay. Balloon angioplasty (BA), despite being more aggressive and associated with potentially fatal complication, may provide a longer lasting effect. It is true that BA does not guarantee that the vasospasm will not ever occur again. There are clearly angioplasty-refractory vasospasms. However, it is usually fair to state that, in general, BA is superior in terms of the durability compared to injecting a few milligrams of any of the vasodilators mentioned above. It is difficult to quantify how long these vasodilators might work. Every practitioner has different thresholds for choosing different methods to address severe vasospasm. Ten milligrams of IA verapamil to the proximal M1 vaso-spasm may be perfectly fine for several days in some patients, but the positive effect may only last for 2 hours in others. Furthermore, the response rate among patients is unpredictable. IA infusion of vasodilation therapy may be considered safer in terms of performing the procedure compared to BA, as there is little to no risk of rupture of the vessel. However, if IA infusion of vasodilation therapy fails (and if it fails in the middle of night when no angiographic team members are available in-house), then there is a significant risk of stroke while preparing for repeat angiography. Another important aspect to consider is the operator's comfort level with each therapy. All of these pros and cons should be taken into account when choosing a therapy.

Intrathecal (IT) infusion and basal cistern implants of calcium channel blockers

Recently, injecting L-type dihydropyridine calcium channel blockers such as nicardipine via EVD as an intrathecal (IT) therapy for vasospasm has been reported. ^54-56 Although generally considered as a new therapy, one of the first human cases dates back to early 1990s when IT nicardipine was injected prophylactically with positive results. ⁵⁴ These reports were case series with small sample sizes, and there is currently lack of good evidence at this time for routinely implanting this therapy. Given anecdotal reports of positive findings in both preventing and treating vasospasm, further studies are warranted to investigate the safety and efficacy of this therapy. As more Pbto ₂ and other multimodality brain monitoring probes are used in clinical practice, there may be more information about how intraventricular use of vasodilators may improve the cerebral blood flow and whether it has more impact on the proximal versus distal vasculature. A recent randomized double-blind phase II study reported the use of prolonged-release implants of nicardipine and showed a reduced incidence of vasospasm and improved clinical outcome 1 year after SAH. ⁵⁷ This study used nicardipine-releasing pellet implantation into the basal cisterns (10 implants placed directly onto the proximal vessels at the basal cisterns) after blood clots were washed out for both study and control groups. The implant group had a significantly reduced incidence of angiographic and symptomatic vasospasm with better short- and long-term outcome. Despite rather convincing data, it is important to remember that this therapy requires surgical clipping of the aneurysm, thorough washout of the fresh blood clots, followed by multiple implantation of nicardipine pellets.

Intra-aortic Balloon Counterpulsation Therapy

Patients with high-grade SAH may also have depressed cardiac function—a typical model for this is that of the neurogenic stunned myocardium phenomenon (sometimes described as neurogenic stress cardiomyopathy or takotsubo cardiomyopathy) with mid to moderately elevated troponin and severely depressed ejection fraction (EF) and reversible left ventricular wall motion abnormality (see below for further details). Depressed cardiac function poses an additional challenge in managing symptomatic vasospasm. Triple-H therapy (hypertensive, hypervolemic, hemodilution) can contribute to developing severe pulmonary edema, and yet patients require induced hypertension and hypervolemia, which increases the afterload and results in further cardiac injury. Intra-aortic balloon counterpulsation was first described in human cerebral vasospasm cases in the mid to late 1990s in order to "allow continuation of triple-H therapy and to maintain adequate cerebral perfusion." ⁵⁸ Imaging studies of both animals and humans reported significantly augmented cerebral blood flow in vasospasm cases by using different methods of brain perfusion scans. ^{59, 60} Inflation at the beginning of diastole and deflation at the end of diastole have been shown to increase the cardiac perfusion, reduce afterload, maintain cardiac performance, and optimize end-organ perfusion including the brain. This therapy is not used routinely, and there are no large studies demonstrating safety and outcome benefits in vasospasm patients. Nevertheless, it is reasonable to be aware and consider this therapy when patients are having severely depressed cardiac function and symptomatic vasospasm refractory to other less invasive medical therapy (for more details, refer to Chapter 35).

NeuroFlo device

This is an intra-aortic catheter with two small balloons designed to augment CBF during the acute phase of ischemic brain injury. The distal balloon is placed above the renal arteries and the proximal balloon is placed below the renal arteries. Partial occlusion of the descending abdominal aorta by inflation of the balloons leads to redirected flow, providing increased CBF as a result (Figures 1-9 and 1-10).

In March 2005, the US Food and Drug Administration approved this device for clinical use under the humanitarian device exemption program after a pilot study in 2004 showed feasibility. In the pilot study, 1 hour of partial aortic occlusion in 17 patients with acute ischemic stroke led to improved blood flow and brain perfusion along with reduced neurologic deficits. ⁶¹ A randomized controlled trial, SENTIS (Safety and Efficacy of NeuroFlo for Treatment of Ischemic Stroke), is currently in progress for more data in acute stroke. A single-arm, observational study on symptomatic vasospasm after aneurysmal SAH has been reported in 24 patients, showing increased mean flow velocity and improvement in NIH stroke scale 20 minutes post-procedure with sustained clinical benefit on a 30-day follow-up in most patients. ⁶²

Cerebral (or renal) salt wasting (CSW) syndrome

Altered plasma and cerebrospinal fluid (CSF) concentration of natriuretic peptide has been thought to be the etiology of this syndrome, which is seen after any severe brain injury but more commonly described in patients with aneurysmal SAH. Sodium wasting is accompanied by free water loss leading to intravascular volume depletion. If CSW and intravascular volume depletion is treated like the syndrome of inappropriate antidiuretic hormone (SIADH) with water restriction, systemic hypotension and end-organ hypoperfusion may occur without improving hyponatremia. One should attempt to accurately assess intravascular volume status as the first step. Treatment should focus on replacing the sodium and targeting euvolemia for overall volume state. Oral salt tablets (2 to 4 g po q4 to 8h) and isotonic saline via IV infusion is reasonable as the first step. If salt wasting gets worse, then 2% to 3% continuous IV infusion of hypertonic saline can be given (start 2% to 3% at 50 mL/h, then titrate up or down). Mineralocorticoids promote sodium absorption at the level of the distal tubule in the kidney and can be used to treat CSW (fludrocortisone 0.05 to 0.2 mg po qd).

A wrong (and dangerous) fluid management would be to misdiagnose CSW as SIADH and perform water restriction with the use of a vasopressin receptor antagonist. Such therapy would lead to profound intravascular volume depletion in the setting of ICP crisis, brain hypoxia, brain metabolic crisis, and symptomatic vasospasm. This can lead to devastating ischemic brain injury. Recently, the term cerebral salt wasting syndrome has been challenged, with the term renal salt wasting syndrome being suggested as there have been cases where the same syndrome occurred in the absence of any brain disease. ⁶³ In the setting of acute cerebral injury such as SAH, cerebral, or renal, salt wasting syndrome may be more common than what has been reported.

Syndrome of inappropriate antidiuretic hormone Secretion (SIADH)

SIADH is another important differential diagnosis for hyponatremia in the NeuroICU. The same etiologies that can lead to CSW can also cause SIADH (eg, SAH, ischemic stroke, traumatic brain injury [TBI], tumor). Differentiating SIADH from CSW can be challenging. A traditional teaching regarding this is the concept of difference in volume status: SIADH has either a euvolemic or hypervolemic state and salt wasting syndrome has volume depletion.

Applying volume restriction and promoting free water loss (aquauresis) by antagonizing the vasopressin receptor (V2 receptors in the kidney) with IV conivaptan (20 mg IV loading over 30 minutes followed by 20 mg/day, may increase to 40 mg/day) or po tolvaptan (15 mg po qd, may increase to 30 mg/day with maximum of 60 mg/day) are appropriate therapies. As these aquauretic agents lead to effective free water loss, one needs to be cautious about aggressive fluid restriction.

With all the laboratory findings for both CSW and SIADH being similar, only volume status can be a hint for differentiating these two syndromes. However, there is no one gold standard parameter that is believed to be always accurate in assessing volume status. Assessing intravascular volume status is a daily challenge for any ICU. As such, for serum hyponatremia occurring in neurologic patients (eg, aneurysmal SAH), it should really be treated with one good old therapy: salt. Hyponatremia is often well tolerated and only patients who are symptomatic, or patients who have a low threshold for recurrent seizures, or patients with severe and worsening hyponatremia (< 125 mEq/L typically) should be treated. If hyponatremia is to be treated in a patient who needs adequate CPP and good intravascular volume (eg, in the middle of active symptomatic vasospasm), giving salt by hypertonic saline is preferred over aggressive volume management. Losing significant free water might be appropriate for SIADH, but can be harmful for patients without SIADH who really are experiencing CSW.

In general, the following tips can be helpful:

1. SIADH is a euvolemic or hypervolemic state due to free water retention, urine osmolality > 100 mOsm/kg, urine Na > 40 mmol/L.

2. FENa (fraction of sodium excretion) < 1% for volume depletion due to CSW, and > 1% for volume overloaded SIADH.

3. Serum uric acid level is normal (3.6 to 8.3 mg/dL) or high for hyponatremia due to dehydration without CSW or SIADH.

4. Serum uric acid level is low for both CSW and SIADH.

5. After successfully treating serum hyponatremia, the renal transport system problem may continue for CSW, so the serum uric acid level remains low (as fractional excretion of urate remains high) for CSW but may be normal for SIADH.

The most likely diagnosis for this patient is a vasospasm that has already been in progress since bleed day 5. When the trend of the TCD velocities of one vessel continues to rise along with corresponding CTA evidence of luminal narrowing, it is reasonable to think that this patient has always had vasospasm—an asymptomatic vasospasm in this case—until now. She had had mild to moderate spasm that was not enough to cause any neurologic deficit. The hemodynamic and blood chemistry tests indicate that she has now developed severe serum hyponatremia and intravascular volume depletion. With reduced intravascular volume status, the CBF was reduced, and what was previously asymptomatic vasospasm is now becoming clinically significant. It is reasonable to think that the CSW syndrome is playing a major role here and as she lost sodium there was concurrent free water loss. The next thing to do for her is probably not an emergent angiogram. Angiography is not a completely benign procedure. Since she became symptomatic with salt and free water loss, the most logical and safe thing to do next is to provide salt and water (perhaps with a higher MAP target to augment CBF rather than ignoring it). With restoration of sodium, intravascular volume, and short-term hypertensive therapy, the patient's symptoms are likely to disappear rather rapidly.

Neurogenic stunned myocardium is a physiologically interesting phenomenon that is associated with a number of different disease states. For neurointensivists, a good example of this syndrome is in the setting of high-grade aneurysmal SAH. There are a number of terminologies that are considered synonymous with this syndrome and that share a similar proposed pathophysiology: neurogenic stunned myocardium, takotsubo cardiomyopathy, broken-heart syndrome, contraction band necrosis syndrome, and Gebrochenes-Herz syndrome among others. Despite different terms and variations between each of these syndromes, there is a common denominator that links all of these phenomena: mental stress. This type of stress is mediated by the brain—whether it is emotion driven (such as in broken-heart syndrome in which shocking news such as sudden death of a family member led to a cardiogenic shock and death, or literally "scared to death," either with good or bad emotions leading to a sudden onset of cardiac failure) with absolutely no structural brain pathology or with a structural abnormality such as acute SAH with intense ICP elevation, inflammation, and sympathetic surge; these are all cerebrally induced triggers. The condition is thought not to be induced by coronary artery atherosclerosis or plaque rupture as in typical acute coronary syndrome. The adrenergically mediated sympathetic surge has been pointed out as the main etiology and mechanism for stunning of the heart. Typically, patients present with clinical features that are consistent with a cardiogenic shock: systemic hypotension and forward/systolic failure with severely depressed ejection fraction (eg, new-onset heart failure with ejection fraction of 10% to 20% in a previously healthy young female). Unlike ST-elevation myocardial infarction (STEMI)cases, ECG may only show nonspecific T-wave ("cerebral T"; Figure 1-11) or ST-segment abnormalities. However, ECG can show ST-segment elevation that looks exactly the same as typical STEMI. Therefore, ECG alone cannot distinguish between STEMI/non ST-elevation myocardial infarction (NSTEMI) and neurogenic stunned myocardium.

In the ED, this syndrome can be fairly challenging as the treatment for NSTEMI and neurogenic stunned myocardium is different. An echocardiogram can be helpful. What is typically seen is transient (as the clinical features, laboratory abnormalities, and ECG findings often resolve after several days) wall motion abnormality: the normal base of the left ventricle but akinetic everywhere else including the apex. Normal base, abnormal apex would not be a bad phrase to remember. The apex is not spared in classic takotsubo cardiomyopathy. On left ventriculogram (Figure 1-12), the apex of the left ventricle is diseased (not spared) and dilated—a classic description associates the shape of the heart in this syndrome to that of an octopus trap jar (in Japanese, "takotsubo"). Remember the apex of the heart is the left inferior and the base of the heart is the right superior portion. The regional wall motion abnormalities are to be seen in the nonvascular area. While variations do occur (eg, midventricular ballooning rather than apical ballooning and apical sparing pattern), it is important to remember the classic description and understand the pathophysiology behind it.

A recent study reported a retrospective analysis of 350 patients with acute SAH and highly abnormal cardiac enzymes. ⁶⁴ This study revealed a few characteristic differences between a true coronary event versus a neurogenic event prior to obtaining an echocardiogram:

1. Troponins were at least 10-fold higher with infarcted heart than stunned myocardium (2.8 versus 0.22, P < .001).

2. Neurogenic stunned heart syndrome is reversible: usually within 4 to 5 days significant improvement can be seen in cardiac output, EF, and normalization of cardiac enzymes.

3. Echocardiogram showing significant inconsistencies with ECG abnormalities (ie, severely depressed EF but only nonspecific T wave inversions and cerebral T waves) is more consistent with neurogenic stunned myocardium.

4. Troponin < 2.8 ng/mL and EF < 40% in the setting of acute aneurysmal SAH is consistent with stunned myocardium rather than a true coronary event.

These tips are only true in general and, as such, care for each patient must be individualized. Nevertheless, it is helpful to remember how to differentiate these two entities in the emergency department in order to prioritize the procedures for managing the patients. Typical stunned myocardium patients should be treated with securing of the ruptured aneurysm in order to avoid rebleed and then providing appropriate hemodynamic support (avoid the use of pure alpha₁-adrenergic receptor agonist). The use of inotropic agents in order to support the reduced contractility and low EF while paying close attention to keep the patient euvolemic is essential (neglecting to address intravascular volume depletion in the setting of stunned myocardium would require a higher than necessary amount of vasopressors and may cause worsening of the cardiac injuries). Initiating prophylactic triple-H therapy in a patient with neurogenic stunned myocardium during the first few days of SAH may be more harmful than beneficial.