Acute occlusion of the internal carotid artery or proximal MCA may result in a large supratentorial infarct encompassing the ipsilateral MCA territory, including portions of the frontal, temporal and parietal lobe. Such infarcts can result in a syndrome of lethal edema and herniation known as a malignant MCA infarct. Within minutes of acute ischemic injury, a decrease in oxygen and glucose causes failure of sodium–potassium ATPase, resulting in loss of the membrane potential and dissipation of ionic gradients. The subsequent rise of intracellular sodium ultimately induces cell death and cytotoxic edema. It has been estimated that approximately 1.9 million neurons die during each minute of ischemia [8]. Vasogenic edema then results from disruption of the blood–brain barrier and an increase in hydrostatic and osmotic pressure [9].

Both cytotoxic and vasogenic edema contribute to an increase in intracranial pressure (ICP) following AIS. Although the brain possesses a remarkable ability to maintain blood flow over a broad range of cerebral perfusion pressures, cerebral autoregulation is limited to ranges between 50 and 150 mmHg. Because cerebral perfusion pressure (CPP) is defined as the mean arterial pressure (MAP) minus the intracranial pressure (CPP = MAP − ICP), an increase in ICP from edema eventually reduces perfusion outside the range of autoregulation. In addition, a space-occupying lesion can increase the pressure gradient between the infratentorial and supratentorial compartment, leading to subfalcine and/or transtentorial herniation and a reduction in level of consciousness [10–14].

Malignant MCA infarction is characterized by a continuing cycle of cell death and edema through elevated ICP and diminished CPP. As time progresses and ICP rises, reduction in CPP leads to further cell death and secondary edema, resulting in further elevations in ICP and reduction in CPP. This cycle continues until herniation occurs, resulting in death. The goal of DC in malignant MCA stroke is to decrease intracranial pressure by allowing external expansion of edematous brain tissue into a compensatory space, to reduce the presence of herniation and brainstem compression, and to restore cerebral blood flow [14, 15].

Malignant MCA strokes occur in up to 10 % of patients with a supratentorial infarct and carry a mortality reaching 80 % in patients treated with standard medical therapy [16]. Patients with malignant MCA strokes present with features consistent with severe hemispheric infarct and space occupying edema that generally manifest between the second and fifth day following ischemic insult (Fig. 5.2) [4, 11, 12, 17–19].

Patients may initially present with reduced level of consciousness and undergo neurological deterioration over the next 1–2 days, often requiring mechanical ventilation secondary to diminished respiratory drive [1, 13, 18]. Other signs and symptoms can include headache, vomiting, pupil asymmetry, papilledema, gaze deviation, dense hemiparesis, and global aphasia. Although the National Institutes of Health Stroke Scale (NIHSS) is generally greater than 15 in these patients, the NIHSS may underestimate the severity of deficit in non-dominant hemisphere infarction [13, 20].

Accumulating radiologic data continue to provide better predictions about the evolution of malignant hemispheric infarcts, but variables such as hemorrhagic transformation, expansion of stroke volume, and spontaneous recannalization of occluded vessels make it difficult to reliably predict malignant MCA infarcts [14, 21]. Images that demonstrate greater than 50 % infarction of MCA territory on CT scan within 18 h have a sensitivity of 58 % and specificity of 94 % for the development of malignant MCA infarcts [22], and infarcts greater than 66 % of MCA territory within the same time span yield a sensitivity of 45 % and a specificity of 100 % [22].

Although CT scans can predict malignant infarcts with a high degree of specificity, their low sensitivity may not reveal patients that are at risk for a malignant MCA stroke. A study by Oppenheim demonstrated that when the initial infarct volume assessed by MRI was greater than 145 cm³ within 14 h, sensitivity was 100 % and specificity was 94 % [23]. As a result, the National Institute for Health and Clinical Excellence created criteria for consideration of hemicraniectomy that included CT evidence of an infarct greater than 50 % of the MCA territory or infarct volume greater than 145 cm³ on MRI [24].

Patients generally demonstrate improved outcomes when they undergo early surgical decompression before or soon after the development of neurological signs such as pupil asymmetry and/or impaired levels of consciousness. One-third of patients show neurological deterioration 24 h after onset of acute infarct, another third of patients within 24–48 h, and the majority of patients will demonstrate deterioration within 1 week following stroke [21]. Randomized controlled trials suggest that DC is able to reduce mortality and increase functional outcome within 48 h of malignant MCA infarct [4, 15, 25, 26], however more data is needed to determine the efficacy of DC after that time. These findings demonstrate that the timing of DC has a role in influencing outcome. Although data is emerging to allow prediction of malignant MCA infarct with greater accuracy, a combination of neurological signs, radiological findings, and clinical judgment are needed to determine the necessity and timing of decompression.

Decompressive hemicraniectomy was first described by Cushing in 1905 and used in the setting of AIS in 1956 [27, 28]. Surgical decompression allows the infarcted edematous tissue to swell outside the confines of the cranial vault in order to reduce external forces on the brain and decrease ICP. Under general anesthesia, a reverse question mark incision is fashioned and a craniectomy flap at least 12 cm is turned [1, 3, 14]. The dura is then opened widely. Removal of infarcted tissue for internal decompression is controversial due to the possibility of disrupting salvageable tissue around the area [1, 4, 15, 29–31]. However, depending on the degree of swelling, removal of some infarcted tissue may be necessary for adequate decompression. It is not uncommon for the brain to slowly expand out of the dural defect during surgery, which may make skin closure difficult.

Control of carbon dioxide levels using hyperventilation, with a goal pC02 of 30–32, as well as hyperosmolar therapy during surgery, may reduce cerebral herniation through the dural opening during surgery. After completion of the decompression, hemostasis is obtained to prevent postoperative epidural hematoma formation, a surgical drain is often left in place, and the temporalis muscle and skin flap are reapproximated. The bone flap is stored in a tissue bank or in a subcutaneous pocket fashioned in the patient’s abdomen.

Post-operative care after DC requires the use of a helmet when the patient is upright or out of bed. This may prevent injury to the exposed brain during falls, a high-risk event due to patient hemiparesis and/or neglect. Based on the degree of functional recovery and patient or family wishes, a cranioplasty may be performed in a delayed fashion. The preserved bone flap is replaced to both restore the normal cranial vault for protective purposes as well as for cosmesis. Bone flap replacement is often performed 6 weeks to 3 months after DC. The risks and benefits of bone flap replacement should be considered strongly in disabled patients as this procedure is associated with complication rates approaching 25 % [32].

There are limited studies evaluating the complications of decompressive hemicraniectomy, though life-threatening complications are generally uncommon. Along with the expected risk of bleeding and infection following any surgical procedure, hydrocephalus and subdural or epidural hematomas may occur. Sinking skin flap syndrome can occur in a delayed fashion in which midline shift to the contralateral side results in neurological symptoms including seizures, focal deficits, and paradoxical herniation [33–35]. This syndrome may be related to unopposed atmospheric pressure effects upon the skin and underlying brain once the swelling of infarcted tissue diminishes and may be worsened by cerebrospinal fluid diversion. Brain abscess following infarcts related to endocarditis may also be seen (Fig. 5.3). Additionally, bone flap necrosis following reinsertion of the autologous bone graft can appear in over 20 % of patients and may present a challenge during long-term follow-up [36].

A more likely complication arises when the craniectomy is not large enough to allow adequate expansion of edematous tissue, and as a result induces sheer stress at the margins of the bone flap and secondary venous insufficiency [37]. If compression of the swollen brain continues, herniation can occur adjacent to the bone margin. The diameter of the craniectomy should therefore provide adequate space to allow decompression of the infarcted area. Based on the observation that malignant MCA infarcts require an additional volume of at least 80 mL, studies suggest that the craniectomy should be at least 12 cm in diameter to allow sufficient expansion [20, 38].

There are multiple international randomized controlled trials evaluating the outcomes of DC in the setting of space-occupying malignant infarction. Three European randomized control trials, the Dutch HAMLET trial, the German DESTINY trial, and the French DECIMAL trial, compared hemicraniectomy to standard medical treatment [4, 15, 26, 38]. All three were stopped prematurely due to a significant decrease in mortality with DC. HAMLET, DESTINY, and DECIMAL all examined functional outcome as a primary outcome and mortality as a secondary outcome measure due to the emphasis on quality of life following malignant hemispheric infarction.

These trials included a total of 93 patients with an NIHSS of at least 16 and an impaired level of consciousness following a significant MCA stroke. Although they enrolled similar ages, there were slight differences among the three. DESTINY and HAMLET included patients between 18 and 60 years of age, while DECIMAL enrolled patients between 18 and 55 years old. There were also variations in the time to decompression following stroke onset. DECIMAL enrolled patients within 24 h of malignant MCA infarction and performed DC no later than 6 h after randomization. DESTINY mandated that DC be performed within 36 h from the onset of symptoms, and HAMLET enrolled patients that had evidence of space-occupying edema within 96 h of stroke onset.

The modified Rankin Scale (mRS) was used to assess functional outcome after stroke in all three trials (Table 5.1) [39]. The scale ranges from 0 to 6, with a score of 0 representing no symptoms and a score of 6 indicating death. For the purposes of statistical analysis, there is often a distinction made between moderate disability (mRS of 3) and moderately severe disability (mRS of 4) in clinical studies. The European trials initially defined a mRS score ≤ 3 as “favorable”, but post-hoc analysis also provided results between groups 0–4 and 5–6. In addition, the pooled analysis of the three trials defined a “favorable” outcome as a mRS ≤ 4.

The DESTINY trial reported that 47 % of patients receiving surgical treatment and 27 % of patients receiving medical treatment had favorable outcomes (mRS ≤ 3), although the difference was not statistically significant. When the authors included patients with moderately severe disability in the favorable group (mRS ≤ 4), there was a significant difference between surgical and medical therapy (77–33 %).

The DECIMAL trial reported that 50 % of surgical patients and 22 % of medical patients had favorable outcomes (mRS ≤ 3), and a statistically significant difference arose in favorable outcomes when dividing functional outcome between a mRS score of 0–4 and 5–6 (75 % in the surgical group versus 22 % in the medical group). In the HAMLET study, surgery had no effect on functional outcome before the study was stopped, although there was an increase in survival rate in the surgical treatment group.

In the combined analysis of the European studies [25], patients undergoing decompressive hemicraniectomy had an overall greater functional outcome when compared to medical therapy (43.1 % versus 21.4 % with a mRS ≤ 3, 74.5 % versus 23.8 % with a mRS ≤ 4). There was also no significant increase of severely disabled patients in the group undergoing surgical intervention (4 % in the surgical group versus 5 % in the medical group). To achieve survival with a mRS score of less than or equal to 3, the number needed to treat was four. To achieve survival with a mRS score less than or equal to 4, the number needed to treat was two.