Neurocritical Care

8. Decompressive Craniectomy in Acute Stroke

A 48-year-old man was found by his wife on the bathroom floor. His speech was slurred, and the left side was weak. In an outside hospital, he was found to have profound left-sided weakness and neglect. The NIH Stroke Scale sum score was reportedly 17. A CT showed a right hyperdense MCA sign, and CTA showed bilaterally occluded carotid arteries. He received intravenous tPA. No endovascular intervention was available. The patient was transferred, and on arrival he does not open his eyes to pain, he has minimally reactive 4 mm pupils, but corneal reflexes are intact. There is a forced eye deviation to the right. He is localizing to pain only on the right side. Left arm and leg are flaccid and do not move after a noxious stimulus. He has Cheyne-Stokes breathing, but there are no marked hypoxemic episodes, and he seems to protect his airway well. He has atrial fibrillation, but with a normal ventricular response, and blood pressure is consistently within the normal range. A repeat CT scan shows an evolving infarct involving the right frontal, parietal, basal ganglia, and caudate nucleus areas. A significant midline shift is noted.

What do you do now?


FIGURE 8.1 CT scan (A, B) showing extensive decompressive craniectomy (with duraplasty using bovine pericardial graft) for swollen hemispheric infarct. Replacement of bone flap seen on repeat CT scan 6 months later (C).

Doing nothing knowing the patient will lapse into coma is not an option in a relatively young person. Further swelling of a major territorial infarct can be anticipated in acute carotid artery occlusion, and often these are patients who deteriorate beyond drowsiness. Medical management with osmotic diuretics is often ineffective, and patients may worsen rather quickly. Decompressive hemicraniectomy may result in a recovery that could potentially be meaningful for the patient. Yet, when it comes to the question of creating space to swell, preemptive removal of half the skull at the site of a newly developing hemispheric infarct may be perceived as overly aggressive. Responsible physicians will have to weigh in expected quality of life, social factors such as support from family members, age, and comorbidity.

What do we know from clinical trials? In a recent pooled analysis of (incompleted) randomized trials the natural history of a large hemispheric infarct (i.e., occlusion of the middle cerebral artery or carotid artery occlusion) was death in 60% and severe disability in nearly 30%. Comparison of the “natural history” with the outcome of patients undergoing decompressive hemicraniectomy remains seriously flawed due to unavoidable less aggressive care in non-surgically treated patients.

There is however good data showing that decompressive hemicraniectomy may be a life-saving procedure. There are also good physiologic arguments for decompressive surgery when performed early in the process. Apart from preventing permanent brainstem injury from direct compression, a reduction of intracranial pressure—even if marginally elevated—may improve cerebral blood flow and brain tissue oxygenation. These effects could allow an improved functional recovery in a proportion of survivors, as observed in recent trials.

The questions are: can we identify the best candidates for decompressive craniectomy, and what should trigger surgery in patients with hemispheric infarcts? Should this large hemicraniectomy be offered to all patients regardless of age, level of consciousness, involvement of vascular territories, or hemispheric dominance? Could early MRI predict clinical deterioration or does it only predict radiologic worsening? We have no satisfactory answers to most of these questions. Some criteria to help in a decision are shown in Table 8.1.

Timing of surgery remains undefined, but clinical deterioration is needed for most neurosurgeons to act. More than a few neurosurgeons confronted with a patient with a massive swollen infarct will still have to be convinced there is benefit to be gained from surgery. Other uncertainties are the technique of decompression—size of craniectomy, removal of additional bone from the squamous part of the temporal bone, extent of the durotomy, removal of the temporalis muscle, among other options.

TABLE 8.1 Criteria for Decompressive Craniectomy in Large Hemispheric Stroke

Age less than 60 years

Rehabilitation opportunities

Patient able to cope with severe handicap

No major comorbidity

Any clinical deterioration in consciousness and need for intubation

Anticipated or documented multiple territorial involvement

Early (< 24 h) evidence of mass effect on CT scan

What did we do? Our patient underwent decompressive hemicraniectomy and developed considerable swelling outside the skull but without clinical deterioration (Figure 8.1AB). It is easy to imagine it would have caused severe deterioration with a swollen ischemic mass in a smaller confined space. Six months later (Figure 8.1C) his functional outcome was not truly favorable despite his being able to ambulate with a cane and being able to take oral intake. He had impaired judgment, was barely interactive, had depressed mood, impaired orientation, difficulty expressing his needs, and required assistance with bathing, toileting, and transfers from bed to chair. In summary, his outcome 6 months later can hardly be called satisfactory, and we do not know if more improvement is expected over time.

Medical management of large hemisphere strokes has been frustrating. Part of the problem is that these infarctions are complicated by an unrelenting swelling unresponsive to usual “antiedema” therapy. Mannitol and hypertonic saline have been inadequately evaluated for the treatment of ischemic brain edema, although clinical empirical experience is mixed, with some patients improving clinically and others progressing to development of brainstem involvement. Hyperventilation in patients intubated for airway protection has not been studied systematically, may negatively impact on cerebral oxygenation, and thus cannot be recommended except for very brief periods. Therapeutic hypothermia—using cooling devices—is increasing but its value in this clinical situation is just being studied in controlled clinical trials.

The main principle of neurocritical management is to avoid further brain injury. Therefore, attention should be directed to maintain adequate intravascular volume (hydrate with 0.9% saline avoiding hypotonic solutions and excessively positive fluid balance), treat fever (using a cooling device if necessary), treat aspiration pneumonitis (with broad-spectrum antibiotics until cultures are known), provide deep venous thrombosis prophylaxis (subcutaneous heparin three times a day) and control blood pressure (systolic blood pressure less than 180 mg and diastolic less than 105 mg). Swallowing precautions are needed and most patients need nasogastric feeding.

A guarded attitude toward these massive cerebral infarcts is understandable, but surgical treatment may be beneficial in some cases. There are some patients who are grateful for such aggressive care. But it is very difficult to know who those patients will be.


· Middle cerebral artery territory infarcts may be due to acute carotid occlusion and such infarcts may become more extensive and particularly severe.

· Malignant hemispheric swelling occurs in 30% of patients with large vessel occlusion.

· Medical management with osmotic diuretics is often unhelpful.

· Decompressive craniectomy should be considered in patients < 60 years, with mass effect on CT scan and evidence of early neurologic decline.

· Decompressive craniectomy may reduce mortality, but neurologic morbidity is considerable and an issue of utmost importance to address with proxy.

Further Reading

Bardutzky J, Schwab S. Antiedema therapy in ischemic stroke. Stroke 2007; 38: 3084–3094.

Huttner HB, Schwab S. Malignant middle cerebral artery infarction: clinical characteristics, treatment strategies, and future perspectives. Lancet Neurol 2009; 8:949–958.

Maramattom BV, Bahn MM, Wijdicks EFM. Which patient fares worse after early deterioration due to swelling from hemispheric stroke? Neurology 2004; 63: 2142–2145.

Rabinstein AA, Mueller-Kronast N, Maramattom BV et al. Factors predicting prognosis after decompressive hemicraniectomy for hemispheric infarction. Neurology 2006; 67:891–893.

Staykov D, Gupta R. Hemicraniectomy in malignant middle cerebral artery infarction. Stroke. 2011;42:513–516.

Thomalla G, Hartmann F, Juettler E. Prediction of malignant middle cerebral artery infarction by magnetic resonance imaging within 6 hours of symptom onset: a prospective multicenter observational study. Ann Neurol 2010; 68:435–445.

Vahedi K, Hofmeijer J, Juettler E et al; DECIMAL, DESTINY, and HAMLET investigators. Early decompressive surgery in malignant infarction of the middle cerebral artery: a pooled analysis of three randomized controlled trials. Lancet Neurol 2007; 6:215–222.

Wijdicks EFM. Management of massive hemispheric cerebral infarct: Is there a ray of hope? Mayo Clin Proc 2000; 75:945–952.

9 Neurological Worsening After Subarachnoid Hemorrhage

A 44-year-old woman presented to the emergency department after having a thunderclap headache. Her past medical history was only significant for current smoking. Her neurological examination was normal, and she was alert (World Federation of Neurological Surgeons grade I). Head CT scan revealed diffuse hemorrhage in the subarachnoid cisterns with some blood in the lateral ventricles, which had preserved size (modified Fisher grade 4) (Figure 9.1A). She was admitted to our neurosciences ICU for further care.

Three hours later she became increasingly drowsy. Her blood pressure trended upward, and she developed sinus bradycardia. Her neurological examination still showed no focal deficits, but upward eye movements were limited. Repeat head CT scan confirmed the suspected hydrocephalus (Figure 9.1B). She improved back to normal after placement of a ventriculostomy catheter.

The following morning she underwent endovascular coiling of her ruptured anterior communicating artery aneurysm without complications. She remained well until postbleeding day 5, when she was noticed to be slightly confused. Mean blood flow velocities on transcranial Doppler had increased by 30% compared with the previous day, with maximal velocities in the 180 cm/s range. Her mild confusion fluctuated over the following 24 hours. Now, examining the patient on postbleeding day 6, she is at times restless and at times drowsy. Although still oriented to her situation when fully awake, she cannot answer minimally complex questions and gets easily distracted. Motor examination reveals a left pronator drift for the first time. Her ventriculostomy is draining well. She is febrile (38.6 degrees Celsius) and actually has been having high temperatures for several hours with poor response to acetaminophen. She is mildly hypertensive and has been polyuric overnight. Laboratory tests show no leukocytosis and a serum sodium concentration of 132 mmol/L (from 136 mmol/L the day before).

What do you do now?


FIGURE 9.1 A) Initial head CT scan showing SAH with aneurysmal pattern and some blood layering in the lateral ventricles consistent with a modified Fisher scale grade 4. B) Follow-up CT scan nearly 4 hours later shows hydrocephalus.

Patients with an aneurysmal subarachnoid hemorrhage (aSAH) often appear deceptively stable. They may “look great” only to acutely or gradually decline into a much worse neurologic state. The very moment when changes in neurologic condition occur are sometimes difficult to pinpoint (as with cerebral vasospasm) but in other instances changes are overwhelmingly clear (as with rebleeding). When caring for a patient with aSAH with worsening neurological condition, the differential diagnosis to be considered will also depend on the time from aneurysm rupture. This is clearly demonstrated by this patient, who declined early due to hydrocephalus and later because of cerebral vasospasm.

The major risks to the patient on the first day after aneurysmal rupture are rebleeding and acute hydrocephalus. It is hard to overlook a rebleeding because the clinical changes are dramatic. The patient suddenly becomes stuporous or comatose and the altered consciousness is accompanied by severe hypertension, tachypnea (or apnea), and tachycardia (or brief asystole). Motor responses change, and extensor posturing (mimicking a seizure to the untrained observer) may occur. In comatose patients with a poor grade aSAH, rebleeding may cause loss of pupillary and corneal reflexes, and nursing staff may see fresh blood in the ventriculostomy bag. This catastrophic event is markedly different from the presentation of hydrocephalus. Instead, as hydrocephalus develops patients become progressively less interactive, then drowsier, and finally unresponsive. While this progression may be rapid, it is not sudden, and alarms do not go off as patients only are mildly hypertensive and bradycardic. While patients are still arousable the only physical sign may be restricted eye movements in the vertical plane caused by the pressure of the expanded third ventricle over the tectum of the brainstem. Given the paucity of clinical clues, the recognition of acute hydrocephalus remains a challenge for physicians outside the neurosciences and many do not appreciate the dilated ventricles on CT scan.

Delayed vasospasm occurs days later, typically starting 3 to 5 days after the hemorrhage to reach a peak around day 7 before resolving by days 10 to 12. Contrary to a common assumption of trainees, the first manifestation of vasospasm is usually not a focal deficit. Instead, diminished alertness and lucidity tend to be the presenting symptoms of this complication. Patients developing cerebral vasospasm are often febrile and have developed hyponatremia, which are factors that can also diminish alertness. Consequently the diagnosis of symptomatic cerebral vasospasm is far from straightforward, and good clinical judgment and experience are necessary to recognize it. Some patients are at higher risk for ischemic damage from cerebral vasospasm after aSAH and these risk factors are listed in Table 9.1. Useful modalities for the screening of vasospasm and diagnosis of delayed cerebral ischemia are summarized in Table 9.2.

Transcranial Doppler (TCD) is useful to monitor for cerebral vasospasm, especially when trends from serial measurements are documented. We suspect cerebral vasospasm when the mean blood flow velocity in the M1 segment of the middle cerebral artery exceeds 120 cm/s and consider it severe when this measurement is greater than 200 cm/s. We have been increasingly using a combination of CT angiogram and CT perfusion in patients with suspected vasospasm. Conventional angiography is reserved for patients who are refractory to medical therapy and might be candidates for endovascular treatment. However, all these techniques have limitations. Cerebral vasospasm is primarily caused by endothelial dysfunction and involves first and foremost the microcirculation. TCD and angiograms are very sensitive for the detection of vasospasm in the large arterial segments but much less accurate when cerebral vasospasm is more distal. Thus, patients may develop ischemic lesions, particularly in deep brain regions, despite having normal or near-normal velocities on TCD and vessel diameters on cerebral angiogram. CT perfusion scans only partially overcome this limitation because their interpretation in practice relies on side-to-side comparison, which loses value in common cases of diffuse, bilateral cerebral vasospasm. CT perfusion may show hypoperfusion much more clearly in the cortex than in the deep white matter. Since ischemia can occur in the absence of documented vasospasm (either because we do not have the right tools to identify it or because there are mechanisms other than reductions in arterial luminal diameter causing the ischemia) the term “delayed cerebral ischemia” is more appropriate than “symptomatic cerebral vasospasm.”

TABLE 9.1 Risk Factors for the Development of Delayed Cerebral Ischemia after aSAH

Extensive clot in subarachnoid cisterns on admission CT scan*

Intraventricular hemorrhage on CT scan within first 24 hours*

Young age

Active smoking

Cocaine use

Poor clinical neurologic examination at onset

*Factors considered in the modified Fisher scale: grade 0, no SAH or IVH; grade 1, thin SAH without IVH; grade 2, thin SAH with IVH; grade 3, thick SAH without IVH; grade 4, thick SAH with IVH.

TABLE 9.2 Modalities for Diagnosis and Monitoring of Vasospasm and Delayed Ischemic Damage

Diagnostic modality

Parameter evaluated

Catheter angiography

Large vessel spasm

Transcranial Doppler

Large vessel spasm (circle of Willis) VMR with CO2 challenge

CT angiography

Large vessel spasm

CT perfusion

Cerebral perfusion

MR angiography

Large vessel spasm


Cerebral perfusion and early ischemia


Cerebral perfusion

Jugular oximetry

Regional brain oxygenation

Brain tissue O2

Brain tissue O2

VMR, vasomotor reactivity

When we suspect a patient is having delayed cerebral ischemia we initiate hemodynamic augmentation therapy. The former approach of the “triple H” (hypervolemia, hypertension, hemodilution) has fallen out of favor for good reasons. Hypervolemia is not sufficient to produce a sustained increase in cerebral blood flow and perfusion. Furthermore, it can impair brain oxygenation and it is the main cause of cardiopulmonary complications in these patients. Hemodilution, if excessive, can compromise oxygen-carrying capacity and result in insufficient brain oxygen delivery. Thus, we rely mostly on inducing hypertension after ensuring a normovolemic state. Most frequently we use phenylephrine or norepinephrine, depending on the initial heart rate and the cardiac status, and we aim to increase the mean arterial pressure by 20–25% as the first step. If symptoms persist, we keep raising the blood pressure, sometimes reaching mean arterial pressures of 140 mmHg. When induced hypertension fails to yield clinical improvement or patients cannot tolerate this medical treatment (e.g., patients with advanced coronary artery disease, heart failure from chronic hypertension, ischemic cardiomyopathy or apical ballooning syndrome), we pursue endovascular therapies, i.e., angioplasty when possible or intra-arterial infusion of a calcium channel blocker.


FIGURE 9.2 A) Cerebral angiogram (right carotid injection) shows severe right middle cerebral artery vasospasm. B) Good angiographic result after treatment with balloon angioplasty.

Our patient was treated with hemodynamic augmentation followed by angioplasty of the right middle cerebral artery (Figure 9.2). Her neurological deficits improved, and she recovered favorably over the subsequent days. Eight weeks later she had returned to work as a teacher.

It is no surprise that “good grade” patients with aSAH may rapidly become “poor grade” patients. The causes are well documented. They just need to be recognized and treated rapidly. Care provided by a dedicated team of neuroscience nurses and skilled physicians with expertise in the management of aSAH is essential to reduce the morbidity of this disease.


· The causes of neurological decline in aSAH relate to the time from aneurysm rupture.

· During the first few hours consider rebleeding if the decline is catastrophic and hydrocephalus if the patient drifts into stupor.

· Many patients with aSAH may benefit from a ventriculostomy.

· Cerebral vasospasm typically occurs after the third day from aneurysm rupture. It often presents with subtle changes in cognition and attention before focal deficits are noted. It remains very difficult to confidently diagnose this condition and there are no accurate non-invasive ways to do it.

· TCD, CT perfusion scan, and cerebral angiogram are useful to monitor and document vasospasm, but the diagnosis of delayed cerebral ischemia remains primarily clinical.

· Induced hypertension is the most useful medical treatment to reverse ischemic symptoms. Endovascular therapy is necessary when symptoms are refractory.

Further Reading

Frontera JA, Fernandez A, Schmidt JM et al. Clinical response to hypertensive hypervolemic therapy and outcome after subarachnoid hemorrhage. Neurosurgery 2010; 66:35–41.

Jun P, Ko NU, English JD, Dowd CF, Halbach VV, Higashida RT, Lawton MT, Hetts SW. Endovascular treatment of medically refractory cerebral vasospasm following aneurysmal subarachnoid hemorrhage. AJNR Am J Neuroradiol 2010; 31:1911–1916.

Hasan D, Vermeulen M, Wijdicks EFM et al. Management problems in acute hydrocephalus after subarachnoid hemorrhage. Stroke 1989; 20:747–753.

Kumar R, Friedman JA. Subarachnoid hemorrhage:the first 24 hours.A surgeon’s perspective. Neurocrit Care 2011;14:287–290.

Rabinstein AA. Secondary brain injury after aneurysmal subarachnoid hemorrhage: more than vasospasm. Lancet Neurol 2011;10:593–595.

Rabinstein AA, Friedman JA, Weigand SD et al. Predictors of cerebral infarction in aneurysmal subarachnoid hemorrhage. Stroke 2004; 35:1862–1866.

Rabinstein AA, Lanzino G, Wijdicks EFM. Multidisciplinary management and emerging therapeutic strategies in aneurysmal subarachnoid haemorrhage. Lancet Neurol 2010; 9:504–519.

Rabinstein AA, Wijdicks EFM. Cerebral vasospasm in subarachnoid hemorrhage. Curr Treat Options Neurol 2005; 7:99–107.