During the study period, 725 patients with large hemispheric infarctions were admitted to our hospital. Of these, 131 patients were found to have a malignant infarction. Malignant MCA infarction was defined as the following; (1) visible signs of ischemia involving more than 2/3 of the MCA territory on brain computed tomography (CT) or a diffusion-weighted imaging (DWI) infarct volume > 145 cm³ with an acute onset of corresponding clinical signs and symptoms (such as hemiplegia, hemianopsia, hemianesthesia and progressive deterioration in consciousness); and (2) neuroradiological evidence of local brain swelling such as midline shifting of 5 mm or more indicating space-occupying edema.

For eligibility criteria see Table 1. Fifty-eight patients with a malignant infarction underwent decompressive craniectomy and received medical treatment (craniectomy group), and 73 patients received conservative care alone (conservative care group). A cut-off point of 70 years of age was set, and the study population was segregated into those above and those below this age limit.

We collected data such as age, sex, past medical history, distribution of the infarction, affected hemispheric dominancy, initial National Institutes of Health Stroke Scale (NIHSS) score, pre-operative NIHSS, measurement of the midline shifting at the septum pellucidum level, and functional recovery six months after surgery as assessed by the modified Rankin Scale (mRS) score.

All patients were treated within 48 hours of the stroke event. For the craniectomy group, the principal surgical technique was a large decompressive craniectomy with duroplasty. Briefly, a question mark-shaped skin flap that began at the ear was created and a wide craniotomy was performed on the affected side with partial removal of the frontal, temporal, and parietal bones, so that the floor of the middle fossa could be exposed and the bone flap would have a minimum of 12 cm diameter. The dura was opened in a "C" shape all over and 1 cm distant to the border of the craniotomy, and an augmented dural patch was inserted. The position of the temporalis muscle and skin flap was approximated and then re-secured. Infarcted brain tissue was not resected. After the operation, patients were transferred to a neurological intensive care unit. The bone flap was reimplanted six weeks to three months after surgery, having been stored under sterile conditions at -80℃.

In both groups of patients, standard medical treatment was based on published guidelines for the early management of patients with ischemic stroke. Endotracheal intubation or nasal probe was recommended to maintain adequate tissue oxygenation in patients with severely increased levels of ICP. Normoglycemia, normothermia and high perfusion pressure levels were obtained. To assist venous drainage, the head of the bed was elevated at an angle of 30 degrees. Administration of intravenous mannitol (0.25-0.5g/kg) was recommended in patients whose condition was rapidly worsening because of brain edema, without additional recommendations on loading doses.

The clinical status of each patient was rated both at admission and at the time of operation using NIHSS scoring. One-month and six-month mortality rates were defined as patient death one month and six months after surgery, respectively. At six months after surgery, each patient's functional outcome was assessed using the mRS score to rate physical disabilities. The six-month functional outcome was dichotomized between mRS score zero to three (favorable outcome) versus four to six (poor outcome).

A t-test was used to compare group variables. Pearson's Chi-square test was used to compare group mortality rates and age, and a logistic regression analysis was used to examine the likelihood of a better mRS score at six months after surgery. Univariate analyses were used to determine the effects of age, gender, hypertension, diabetes mellitus, atrial fibrillation, ischemic heart disease, hemorrhagic transformation, old infarction, midline shifting, and treatment modalities on mRS at six months after operation. Variables that were found to be significantly associated with better mortality rates and mRS score were subjected to multivariate analyses. Differences were regarded as significant if their probability values were p < 0.05. Statistical analyses were performed using SAS version 9.2.

A total of 58 patients received decompressive craniectomy. There were no surgical complications directly related to this procedure. In 17 patients of these patients, surgery failed to improve uncal herniation and they died within the one month of operation. And 28 patients died within six months of operation. Thus, the one-month mortality rate was 29.3%, and the six-month mortality rate was 48.3% in the craniectomy group. Favorable functional outcome at six months after surgery was noted in 19 patients(63.3%) in this group.

Seventy-three patients underwent conservative care. Their one-month mortality rate was 58.9%, and six-month mortality rate was 71.2%. There were eight patients(38.1%) with favorable functional outcome.

One-month and six-month mortality rates showed a statistically significant difference between the treatment modalities of craniectomy and conservative care (p < 0.001, p = 0.007). These treatment modalities, however, were not statistically different for the six-month functional outcome (p = 0.076) (Table 4).

In patients older than 70 years of age, the craniectomy group had a 30.0% one-month mortality rate, while the conservative care group had a 60.8% one-month mortality rate. The six-month mortality rates were 60.0% and 74.5% for each group, respectively. We confirmed that 50.0% (four cases) and 23.1% (three cases) of patients yielded a favorable functional outcome in the craniectomy group and conservative care group, respectively.

For patients younger than 70 years of age, the craniectomy group showed a 28.9% one-month mortality rate, while the conservative care group showed a 54.5% one-month mortality rate. The six-month mortality rates were 42.1% and 63.6% for each group, respectively. There were 68.2%(15 cases) and 62.5% (five cases) patients with favorable functional outcome in the craniectomy group and conservative care group, respectively.

For patients both older than 70 and younger than 70 years of age, the mortality rates (one-month, six-month) were lower and the favorable functional outcome was higher in the craniectomy group compared to the conservative care group (Table 5).

Finally, age was not statistically significant among all patients with respect to the six-month mortality rate and the favorable functional outcome (p = 0.137, p = 0.077) (Table 6, Table 7).

Right to left hemisphere ratio was 85(64.9%) : 46(35.1%) (Table 2). Between the two hemisphere, there was no statistically significant difference for six-month mortality rate (p = 0.206, Table 6). However, the functional outcome after six months was significant (p = 0.032) (Table 7).

The mean midline shift for all patients was 7.34mm. When 10 mm was chosen as a cut-off point, midline shifting at the septum pellucidum level was significantly different with the six-month mortality rate and favorable functional outcome (p = 0.007, p = 0.004) (Table 6, Table 7).

The cut-off value for NIHSS scores was set up differently, depending on the location of the infarction site. The cut-off value was 18 points for a right-sided infarct, and 20 points for left-sided infarct. Eighty-four(64.1%) were lower than cut-off value, and 47 (35.9%) were higher. There was significant difference of NIHSS score in six-month mortality rate (p = 0.047, Table 6). But, no significant difference in functional outcome at six months after operation (p = 0.184, Table 7).

According to previous research regarding malignant infarctions, patients clinically presenting with severe hemispheric stroke syndrome, including hemiplegia, forced eye and head deviation, and progressive deterioration of consciousness within the first two days, have poor prognoses with mortality rates as high as 80%. Non-randomized studies found that decompressive surgery reduces these mortality rates and increases favorable functional outcomes in patients with large hemispheric infarctions when compared to conservative treatments. Schwab et al.20) reported that 73% of patients survived (mortality rate of 27%), whereas Sasaki et al.19) and Carter et al.2) documented the mortality rate of 33% and 0%, respectively.

Recently, three European randomized controlled trials (the French DECIMAL trial, the German DESTINY trial, and the Dutch HAMLET trial)7)10)23) demonstrated that decompressive craniectomy reduced the mortality rate by 49% at one year after stroke when compared with conservative care at the same time (78% in the craniectomy group, 29% in the conservative care group). And favorable functional outcome (zero to three in mRS) increased by 22% due to surgery (43% in the craniectomy group, 21% in the conservative care group).6-8)10)11)15)23)

In our study, the one-month mortality rate was 29.3% in the craniectomy group and 58.9% in the conservative care group. Six-month mortality rates were 48.3% and 71.2%, respectively. The favorable functional outcome was 63.3% in the craniectomy group and 38.1% in the conservative care group. Craniectomy reduced the mortality rate by 29.6% at one-month, 22.9% at six-months after surgery and increased the favorable functional outcome by 25.2% when compared with conservative care. Additionally, the mortality rates were significantly different (p < 0.001, p = 0.007) between two group (Table 4). This study reaffirms that hemicraniectomy for malignant MCA infarction is a life-saving treatment.

Recently, three trials published their pooled analysis of 93 patients,7)10)23) all less than 61 years of age, indicating that hemicraniectomy reduced the mortality rate by 49% at one year after stroke when compared with optimal medical treatments. This confirms the results from previous non-randomized studies and case reports that were analyzed in this review.6-8)10)11)15)23) The question remains as to whether similar results would be expected in an elderly population as the incidence of stroke increases with old age. Jüttler et al.9) suggested that about 50% of all patients with malignant MCA infarcts are older than 60 years of age.

In the present study, 27 patients were younger than 60 years of age, while 104(79.4%) were older than 60 years. However, 60 patients(45.8%) were under 70 years of age and 71(54.2%) were older than 70 years of age. For this reason, we segregated the patients with malignant MCA infarction into two group with a determining cut-off point of 70 years of age in order to balance the groups more equally (Table 8).

In previous reports, age was a key factor in determining who would benefit from decompressive surgery after malignant cerebral infarction, and some investigators have reported poorer outcomes in elderly patients. Preexisting disabilities and severe comorbid conditions are also important factors that increase poor outcomes, and both are more prominent in elderly patients.13)17)21)22)

We have shown that patients older than 70 years have higher mortality rates (both at one month and six months after surgery) and lower favorable functional outcomes compared with the patients under the age of 70 years. When comparing patients older than 70 years of age who received craniectomy to patients of the same age who received conservative care, the mortality rate was 30.8%, a 14.5% of rate reduction, while the favorable functional outcome rate increased by 26.9% in the former group. In the patients younger than 70 years of age, these results were similar.

Among all patients, age was not statistically significant for six-month mortality rate (p = 0.137, Table 6). In other words, no significant difference was observed between patients younger than 70 years of age and those older than 70 years of age in terms of mortality rate. We conclude that age itself has little effect on prognosis. Therefore, elderly patients with malignant infarctions will have better outcomes with decompressive craniectomy compared to conservative care only.

Preexisting factors that are also predictors of poor survival (diabetes mellitus, hypertension, atrial fibrillation, ischemic heart disease, a previous history of stroke and hemorrhagic transformation) did not significantly influence mortality rates or functional outcomes in our study. We assume that this was caused by the relatively small number of patients (n = 131) that were included in this study, in comparison to the larger numbers that have been included in other stroke reports.2)3-5)11)12)14)17)25)26)

Offering life-saving treatment for large dominant hemispheric infarcts is controversial, mainly because surgery may leave patients with an unacceptable quality of life. The hemispheric location of the infarct did not have prognostic relevance in our study, as demonstrated by other series. Gupta et al.6) reviewed the functional outcomes of 27 patients with decompression of the dominant hemisphere and 111 patients who had non-dominant infarcts and found that the outcomes were similar. In the study by Kilincer et al.,12) half of the patients had a dominant hemispheric infarction with global aphasia pre-operatively, and six of seven patients in the good outcome group had a dominant hemispheric infarction and most of the patients showed considerable improvements in their aphasia.

In our study, the hemispheric location of the infarct had no effect on mortality rates but did have an effect on the functional outcomes. The larger the infarction was in size, the more the brain stem was compressed, indicating that mortality is more influenced by the size of infarct rather than the hemispheric location. Therefore, we believe that the location of the infarction should not be an exclusive criterion for surgery.

The NIHSS score is commonly obtained in acute stroke patients. A score of zero indicates no clinically relevant neurological abnormalities. If a patient scores more than 20, it usually indicates a dense paralysis with impaired consciousness.2) NIHSS score ≥20 for left-sided infarctions, or ≥18 for right-sided infarctions may predict which patients will develop malignant edema and/or have a poor outcome.18) Lam et al.16) concluded that a NIHSS score of > 22 is predictive of high mortality.

In our study, patients who presented with NIHSS score of ≥20 for left hemispheric infarction, or ≥18 for the right hemispheric infarction in the presurgical exam demonstrated a tendency toward poor outcomes. The NIHSS score was significantly related to mortality rates but not to the functional outcomes.

There are several limitations in our study. Although we present important data about decompressive surgery, it is a non-randomized retrospective study with results that need to be confirmed by larger randomized trials. Additionally, our follow-up period of six months is rather short, thus a longer follow-up evaluation is necessary.