Prevention and management of vasospasm following SAH

Management of vasospasm seems to be a favourite topic among CICM exmainers. It has come up several times in the past papers and vivas. The following is a list of representative SAQs:

• Question 10 from the first paper of 2007
• Question 5 from the first paper of 2013
• Question 24 from the second paper of 2017

A chapter of Oh's Manual (Ch. 51, pp 568) is the canonic resource for these topics. The gospel of subarachnoid management seems to be this 2012 Guidelines Statement from the AHA. Another good resource is available from Expert Reviews - it is an article from 2015 which lists and discusses all the successfull and unsuccessful trials in this area. The Dabus-Noguiera article quoted in LITFL also offers some opinions about the weirder therapies for AH, such as fasudil, colforsin, IABP, partial aortic occlusion, and so forth.

As far as non-journal study resources go, the LITFL review of vasospasm and DCI is a treatment with satisfying levels of detail; with its authors' interest in neurocritical care being well known, its value is significant as a distillate of his expertise.

For the purpose of this short summary, all these sources have been combined and remixed.

In brief:

In detail:

Vasospasm and Delayed Cerebral Ischaemia (DCI)

This is a much-spoken-of complication of subarachnoid blood. Generally speaking, 70% of patients will develop this complication.

Firstly, the definitions. They seem to overlap somewhat. And there are thousands of them; just see this table from Nature (2011). Some of the definitions quoted below are derived from proposals made by a multidisciplinary research group, published in Stroke in 2010; others were made by researchers who pined for a clinically relevant definition.

In brief, you know it's vasospasm when...

• • A focal neurological deficit develops (and persists for longer than an hour)
  • The GCS drops by 2 points (and stays that low for longer than an hour)
  • A new stroke occurs (on CT/MR)
  • No better explanation for any of the above.

Or, alternatively, there are no symptoms, and your patient is having a silent vasospasm. Therein lies the disadvantage of clinical examination.

There are several techniques of imaging which can be used to recognise vasospasm:

Conventional 4 vessel DSA (Digital Subtraction Angiography):

• This is the gold standard for both diagnosis and management of vasospasm. One can confirm that vasospasm is occurring by a CT angiogram- or, if one were to go straight to DSA one could progress to some sort of definitive treatment. Verapimil and papaverine are the two most commonly used intra-arterial vasodilators.
• However, this requires a skilled interventional radiologist. It exposes the patient to contrast and it it in vasive, with a (not insignificant - around 1%) risk of atheroma embolism or vessel dissection..
• There is a small chance that this technique will lead to over-treatment: vessel narrowing may be detected, but this decrease in diameter may not reflect a decrease in flow, and may not warrant an injection of vasodilator.

On DSA assessment, the vasospasm may be graded:

• no vasospasm
• slight vasospasm, difficult to appreciate without control images
• Moderate vasospasm, obvious narrowing but still more than 50% of expected diameter
• Severe vasospasm, 50% or less of expected vessel diameter.

CTA/MRI:

• Together, the two modalities can be very useful in non-invasive diagnosis of vasospasm. The sensitivity is said to be 79%, and specificity 93%.
• The disadvantages are exposure to contrast and the need to confine an unstable ICU patient inside the Death Tunnel.
• Another disadvantage is that coils and clips cause artefact, which impairs one's assessment of the arterial diameter, and they can play havoc with the MRI images.

Transcranial Doppler (TCD):

• This is a totally non-invasive technique, which can be performed at the bedside, and that is its major advantage. Apparently, MCA flow velocities of more than 120cm/s are accurate in predicting ischaemia. Velocities over 200m/s practically scream "vasospasm". According to an old 1989 review, specificity is amazing, approaching 100%, but the sensisitivity is poor - only 58%.
• Unfortunately, every observer will arrive at a different flow velocity.
• Thus far, nobody can agree as to what slightly raised flow velocity means.
• Transcranial Doppler may be more accurate when MCA velocity is indexed to the velocity of the carotid artery. If the velocity is more than 3 times faster, there is probably vasospasm going on.
   

SPECT/PET

• These would be wonderful tools which would give us a glimpse of cerebral perfusion and metabolic resource utilisation. Unfortunately, nobody knows how to interpret them, and few centres can perfrom these studies. For lack of good validating evidence, they must for now remain experimental.

Radionuclide scan

• Can be used to obtain a picture of brain perfusion and metabolism and have
  shown variable correlation with vasospasm as assessed by more conventional
  methods.
• Disadvantages: They are resource heavy not easily available, radiation exposure, patient
  transport are issues.

EEG:

• The EEG can demonstrate findings which are strongly associated with vasospasm. Not only that, but it can detect areas of ischaemia. Most interestingly, it can do so early - i.e. before there are other fatures. However, deriving meaning from the squiggly lines requires a skilled interpreter. In the hands of such a guru, sensitivity and specificity for detecting vasospasm would be very high.

Risk factors for vasospasm

The classical risk factors are as follows:

• Thick subarachnoid blood: As Fisher (of the Fisher scale) had found in a series of 47 patients, anything thicker than 1mm all but guarantees vasospasm. This is probably the most important risk factor.
• Smoking (OR = 4.7) - probably as a result of some sort of vascular endothelial confusion.
• Cocaine use (OR = 6.4) - even though being intubated prevented them from snorting it.
• Youth (under 50 years of age = almost twice the risk of the over-50s) - likely owing to the higher elasticity and thus greater "spasmability" of young blood vessels; this is supported by the finding that old crusty vessels are protective from vasospasm, presumably with the atheroma acting as some sort of naturally secreted stent.
• SSRI and statin use (OR =2.0 and 2.75) - nobody knows why

A recent analysis of 370 patients has added several more risk factors:

• Rebleeding (OR = 1.16)
• Hypertension (OR = 1.92)
• Daily alcohol intake (OR = 1.23)
• Leukocytosis, WCC > 10.0 (OR = 1.31 - for infarction, rather than angiographic vasospasm)
• Hyperglycaemia(OR = 1.32 - for infarction, rather than angiographic vasospasm)
• QTc > 450msec (OR = 1.52)
• LVH on ECG (OR = 2.71)
• ST depression on ECG (OR = 1.29)

Also: there is no difference in incidence of vasospasm between the clipped group and the coiled group.

That's important to know.

Pathophysiology of vasospasm

One could (and perhaps one day will) devote a massive excess of attention to this fascinating area.

For now, I will limit myself to quoting LITFL's article by Oliver Flower, which mentions the following:

• oxyHb contacts intraluminal wall of vessel
• calcium dependent and independent pathways
• free radical mechanisms
• imbalance in vasoactive agents: NO, endothelin, arachadonic acid
• neuronal mechanisms leading to increased vascular tone, endothelial tone and apoptosis

The reference for this is likely an article by Kolias and Belli from 2008 (Journal of Neuroscience Research).

Prevention and management of vasospasm

Nimodipine for the prevention of vasospasm

The BRANT trial from England has demonstrated a massively decreased risk of SAH-induced stroke in patients receiving nimpodipine - they were 34% less likely to develop stroke. One ought to continue nimodipine for 21 days to get the optimal effect. The usue of nimodipine undergoes a more thorough dissection in the following chapter (Evidence for the use of nimodipine in SAH)

Nicardipine for the prevention of vasospasm

This -drug the other calcium channel blocker for vasospasm - also appears to have some sort of cerebral vessel selectivity. Trials from the 1990s support its use as a means of preventing vasospasm; however there was no heroic stroke reduction, and both the treatment and placebo groups had the same outcomes at 3 months. In more recent research focus has shifted to using it intrathecally, or intrarterially, or as a slow-release implant. A 2013 meta-analysis of RCTs (5 of them) has finally found some outcome benefit, but only after combining all the intravenius intrathecal and intraarterial data, which seems like a strange strategy.

The rise and fall of "Triple H therapy"

Since its debut in 1990, "Triple H" therapy for subarachnoid haemorrhage has been used in ICUs worldwide as a means of protecting the SAH patients from vasospasm-induced stroke.

Hypertension, hypervolaemia, hemodilution. In order to prevent hypoperfusion injury, the prevailing theory was that the blood pressure should be high, the blood volume should be hyperexpanded, and the hematocric should be low (to decrease blood viscosity). This theory seems to have arisen from the simple relationships of fluid dynamics, where pressure gradients and fluid viscosity are the governing determinants of flow rate.

• Hypertension was supposed to improve the flow though the spasming vessels by overcoming the normal mechanisms of cerebral autoregulation, and forcefully driving an increase in cerebral blood flow by the sheer effect of the pressure gradient. Of course, the vessels of the brain have some degree of control over what is happening to their diameter, and normally cerebral autoregulation of bloodflow will compensate for systemic increases in mean arterial pressure. The key idea was that induced hypertension would eventually “break” this compensation. Ideally, one would continue this treatment for the same duration as the nimodipine (i.e. for about 21 days).
  Ultimately, this component of Triple H therapy ended up being the only one broadly supported by a recent multidisciplinary consensus conference. Among a group of neurocritical care specialists and neurosurgeons, hypertension "was considered reasonable by most participants", even though there was a great variety in opinion regarding how hight he blood pressure should be, which blood pressure variable to aim for (MAP vs SBP?) and which agents we should use to achieve it.
  If pushed, most would agree that the "correct" blood pressure is the pressure at which your neurological deficits disappear.

• Hypervolaemia was supposed to increase the circulating fluid volume, thereby allowing enhanced filling of the cerebral circulation. One is puzzled as to how this is supposed to work, given the known distribution of blood volume (with more than 70% of it tending to slosh around in the venous capacitance vessels rather than arteries). It certainly doesn't seem to do what it's supposed to, according to empirical evidence.

• Haemodilution is supposed to improve laminar flow though the cerebral vessels by decreasing the viscosity of the blood. This may actually be relevant in patients with polycythaemia, or patients who have a subarachnoid haemorrhage in addition to an actual hyperviscosity syndrome.

Myburgh has offered this strategy a nice overview in CCR(2005) and couldn’t recommend it on the balance of evidence. Cerebral autoregulation thresholds are too variable among patients, he says, and probably vary regionally among different parts of the cerebral circulation.

Do the experts agree? No of course they do not. Myburg’s opinion is contradicted by Dhar et al, whose prospective physiological study found that Triple H therapy (all three independently) improve cerebral oxygen delivery to vulnerable regions. The debate about this is ongoing. Thus far, the 2009 iteration of the Stroke  guidelines for management of SAH has concluded that induced hypertension as part of Triple H therapy is a “reasonable approach”, even though the enthusiastic evidence for its use is far from satisfactory.

The most recent spoonfulls of evidence added into this cauldron have been generally unsupportive of "Triple H" therapy. In 2010, a review of the literature found no evidence to support the use of the full package, or any of its components individually. In 2014, the Japanese have confirmed a lack of improvement in clinical outcdomes in a prospective multicentre study. With such a heterogeneously lumpy porridge of evidence, it is little wonder that a 2014 survey of management protocols has found "striking variability in the practice patterns of European physicians", with 44% of responders using "Triple-H" therapy in spite of the patchiness of the evidence in support of it.

Intrathecal thrombolysis

Spilled blood causes vasospasm(it is thought), so one might think that causing a premature breakdown of this blood is beneficial in the prevention of vasospasm. The injection of urokinase followed by some unusual head movements was viewed as a good idea by some Japanese investigators, and had yielded a 50% decrease in vasospasm and associated complications. It is still viewed as an "experimental" therapy.

Statins for subarachnoid haemorrhage

There was some interest in the idea that statin therapy might somehow decrease the morbidity from subarachnoid haemrorrhage . The STASH trial has compared placebo to an 80mg dose of simvastatin, and was forced to conclude that the findings "do not support a beneficial effect of simvastatin in patients with SAH "

Magnesium infusion

Given the smooth-muscle-relaxing effects of magnesium cations, its application to cerebral vasospasm seems no great logical leap. The MASH-2 trial from 2012 was a multi-center investigation of 64mmol of MgSO₄ per day, randomised among 606 patients, which somberly concluded that "intravenous magnesium sulphate does not improve clinical outcome after aneurysmal subarachnoid haemorrhage".

Endothelin receptor antagonists

Clazosentan (a relative of bosentan) is an endothelin-1 receptor antagonist, and it has been investigated in the aptly named Clazosentan to Overcome Neurological iSChemia and Infarct OccUrring after Subarachnoid hemorrhage (CONSCIOUS-1) trial. The complications of clazosentan therapy (such as anaemia and hypotension) were "manageable". Its use, at least in in severe and moderate vasospasm, demonstrated "a trend toward improvement", as opposed to actual improvement. In short, these drugs only merit the most lukewarm of recommendations.

Intra-arterial vasodilators

Nicardipine (as above) has been used intra-arterially to selectively dilate the spasming vessels. Similarly, papaverine ( phosphodiesterase inhibitor) and verapimil (another calcium channel blocker) have been used with some effect, and remain in routine use today. The most recent Critical Care Guidelines on the Endovascular Management of Cerebral Vasospasm have found papaverine to be the most extensively studied agent, and the one with the greatest amount of support behind it; less data exists in support of verapimil and nicardipine. Milrinone has also showed some promise, both as an intrarterial agent and as an inthrathecal irrigation; there may be an additive effect with nimdipine in the setting of refractory vasospasm.

Balloon angioplasty

If one does not wish to expose the patient to intraarterial vasodilators, one may instead address the spasm in a very direct and brutally mechanical way, but dilating the spasmed vessel with a balloon. The abovementioned guidelines statement had identified 27 studies. There seems to be some disagreement as to which is better (intraarterial papaverine vs. balloonoplasty studies have a substantial degree of heterogeneity in their reported outcomes). There is no disagreement, however, about the risks - balloons have a known association with arterial rupture and fatal haemorrhage.