Decompressive craniectomy versus conservative treatment: limits and possibilities in malignant stroke

ABSTRACT Background: Malignant infarction of the middle cerebral artery (MCA) occurs in a subgroup of patients with ischemic stroke and early decompressive craniectomy (DC) is one of its treatments. Objective: To investigate the functional outcome of patients with malignant ischemic stroke treated with decompressive craniectomy at a neurological emergency center in Northeastern Brazil. Methods: Prospective cohort study, in which 25 patients were divided into two groups: those undergoing surgical treatment with DC and those who continued to receive standard conservative treatment (CT). Functionality was assessed using the modified Rankin Scale (mRS), at follow-up after six months. Results: A favorable outcome (mRS≤3) was observed in 37.5% of the DC patients and 29.4% of CT patients (p=0.42). Fewer patients who underwent surgical treatment died (25%), compared to those treated conservatively (52.8%); however, with no statistical significance. Nonetheless, the proportion of patients with moderate to severe disability (mRS 4–5) was higher in the surgical group (37.5%) than in the non-surgical group (17.7%). Conclusion: In absolute values, superiority in the effectiveness of DC over CT was perceived, showing that the reduction in mortality was at the expense of increased disability.

Mortality in extensive ischemic stroke (EIS) in the territory of the middle cerebral artery (MCA) can reach up to 80% according to some controlled studies 1,2 . Transtentorial herniation is the most common cause of death 3 . In view of this scenario, the selection of patients for early prophylactic decompressive craniectomy (DC) surgery is an essential measure in the treatment of acute neurovascular syndrome in emergency neurological services worldwide 4 .
EIS is said to be malignant when there is neurological deterioration secondary to extensive cerebral infarction and consequent encephalic edema, which in turn usually occurs between the second and fifth day after the stroke 5,6,7 . The patient may present clinical signs of intracranial hypertension, such as headache, nausea, vomiting, anisocoria, epileptic crisis, mental confusion, lowering of consciousness level and coma 8,9,10 .
Two forms of malignant ischemic stroke management are recognized: isolated conservative treatment (CT) and surgical treatment associated to clinical measures, such as osmotherapy with mannitol or hypertonic saline 11,12,13,14 . The first multicenter randomized controlled trials that sought to evaluate the benefit of DC surgery were the HAMLET (Hemicraniectomy After Middle cerebral artery infarction with Life-threatening Edema trial) in 2006, the DECIMAL (DEcompressive Craniectomy In MALignant MCA Infarction) and the DESTINY I (Decompressive Surgery for the Treatment of Malignant Infarction of the Middle Cerebral Artery) in 2007. They established benefits in the survival of patients under 60 years of age undergoing early surgical decompression for malignant stroke, when compared to CT. However, such studies did not assess the quality of life of these survivors. Often, the survival of patients undergoing DC was accompanied by serious sequelae, such as tetraparesis, aphasia, chronic pain and depression. The choice of treatment in the EIS has become a real dilemma for neurologists, neurosurgeons and intensive care specialists, mainly with the advent of new knowledge from so-called Palliative Medicine 15,16,17 . Therefore, this study aimed to evaluate the functional outcome of patients with malignant ischemic stroke treated in a neurological emergency center in the Northeast of Brazil, submitted to DC surgery or not.

METHODS
This was a prospective cohort study, in which all patients with MCA malignant ischemic stroke treated by the neurological emergency service of the General Hospital of Fortaleza City, Ceará State, Brazil, from November 1 st , 2015 to December 30 th , 2016, were evaluated. The inclusion criteria were diagnosis of ischemic stroke with an ictus of 48 hours or less, an extensive ischemic area measured with cranial CT (>50% of MCA territory), neurological deterioration and clinical signs of intracranial hypertension.
The initial sample of 29 patients was divided into two groups: 10 patients who were treated with DC surgery and 19 who continued receiving gold-standard conservative treatment in the stroke unit, with the support of the specialized multidisciplinary team, including neurologists, nurses, physiotherapists, occupational therapists, speech therapists and dentists. Two patients from the craniectomized group and two patients from the non-craniectomized group were lost from the follow-up, thus providing a final sample of 25 individuals.
The patients were monitored by the Neurology and Neurosurgery team during hospitalization and reassessed by telephone, after six months, by the Neurology team. The functional status of both groups was compared to the modified Rankin Scale (mRS). The favorable functional outcome of mild impairment (mRS 0-3) was adopted as the primary outcome. A mRS between 4 and 5 represents moderate to severe disability, which together with an mRS of 6, comprised the total of unfavorable outcomes. The magnitude of the neurological deficit was quantified with the standard National Institutes of Health Stroke Scale (NIHSS), stratified as mild (0-10), moderate (11)(12)(13)(14)(15)(16)(17)(18) and severe (>18). The level of consciousness was measured with the Glasgow Coma Scale (GCS) score, as severe (≤8), moderate (between 9 and 12) and mild (between 13 and 15).
Data were analyzed using the IBM SPSS version 23 software. Absolute and relative frequencies were calculated for the qualitative variables and mean and standard deviation for the quantitative variables. The quantitative measures were evaluated for normality using the Shapiro-Wilk test. Student's t-test, the Mann-Whitney test and the chisquared test were used to examine the association between variables. The results were presented in graphs and tables. For all the inferential procedures used, a significance level of 5% was adopted.

RESULTS
In the sample analyzed, most patients with EIS in both groups were women, representing 70.6% of the patients that received the conservative treatment and 50% of the patients that underwent craniectomy. Ages ranged from 21 to 67, with the majority corresponding to the age group between 21 and 59 (75.0% in CD group and 88.2% in CT group), as presented in Figure 1. Regarding origin, 62.5% of the surgical patients and 70.5% of the non-surgical patients lived in Fortaleza City, as presented in Table 1.
The most prevalent risk factors in both populations were systemic hypertension, sedentary lifestyle and smoking. Most patients treated in the referral service were sedentary (n=20). Smoking ranked as the second risk factor (n=16), followed by systemic hypertension (n=15), as shown in Table 1. Some patients presented more than one factor. Regarding the  previous functionality, evaluated with the modified Rankin Scale (mRS), Table 2 shows that most patients had no significant disability and that all the patients that had craniectomy were functionally independent regarding activities of daily living. One individual in the surgical group (3.5%) had a moderately severe disability (mRS=4) prior to the stroke. Most patients were able to reach the neurological referral service within the therapeutic window period for intravenous thrombolytic administration (4.5 hours after onset of symptoms), totaling 11 cases (38%). Nine patients (31%) arrived between 4.5-8 hours after the onset of symptoms, and nine (31%) patients arrived with an interval of more than 8 hours. Only one patient received thrombolytic therapy with alteplase. According to the etiological investigation of the stroke, 16 cases of the total sample had no defined etiology.
Seven patients had cardioembolic etiology and only one case presented large vessels mechanism (Table 1).
Regarding the clinical manifestations at the time of admission, all the patients presented motor and sensory symptoms. Twenty patients presented speech changes and visual changes. In addition, 64% of the patients had altered levels of consciousness. Only five patients had headaches, three had a seizure episode, and vomiting was reported in 20% of the cases (Table 3). At admission, 25% of the craniectomized patients and 23.5% of non-craniectomized patients arrived at the hospital with a Glasgow Coma Scale score considered severe (GCS<8); 62.5% of the surgical patients and 35.3% of the non-surgical patients had a moderate score (9-12), and 12.5% of the craniectomized patients and 41.2% of the non-craniectomized patients presented a mild GCS  Table 2. Clinical profile of the patients affected by malignant ischemic stroke of a neurological emergency center. (13)(14)(15). Regarding the magnitude of the neurological damage calculated by the NIHSS score, no individual was classified as mild NIHSS (0-10), 11 patients had moderate NIHSS (11)(12)(13)(14)(15)(16)(17)(18), and the majority of patients (n=14) had NIHSS considered severe >18). In relation to the cerebral hemisphere involved, the majority of craniectomized patients had their left side affected (n=6), whereas most of conservative patients had their right hemisphere more affected (n=9) ( Table 1). Regarding the functionality of the patients with malignant stroke at the 6-month follow-up, Figure 2 shows the percentage of all functional outcomes according to the mRS, whereas Table 4 divides the two groups into favorable outcome (mRS 0-3), unfavorable outcome (mRS 4-5) and death (mRS=6). Most patients who received conservative treatment evolved to death (52.8%). In addition, 37.5% (3/8) of the craniectomized patients and 29.4% (5/17) of the non-craniectomized patients achieved a favorable outcome of mild disability (mRS 0-3), whereas 37.5% (3/8) of the patients in the craniectomized group and 17.6% (3/17) of the non-craniectomized group had moderate to severe disability (mRS 4-5) after six months. Mortality (mRS=6) was 2.1% (2/8) among surgical patients and 52.9% (9/17) among those treated conservatively. Of those who died, 81.8% presented a cause of death related to brain herniation. One patient died of sudden cardiac arrest 10 months after discharge. The mean time between the stroke and death was 6.8 days.

DISCUSSION
The average age of patients in our sample was 50.0±10.9, whereas in the study by Vital et al. 18 it was 63 (32-83). Patients were allocated to the groups within 48 hours of the stroke, with this time being 30 hours for the DECIMAL study, 96 hours for the HAMLET, 36 hours for the DESTINY I and 48 hours for the DESTINY II. The mean time between the onset of symptoms and admission to the emergency room was 7 hours (range 2-25). In the studies by Bongiorni et al. 19 at the Hospital de Clínicas de Porto Alegre, Brazil, the average was 13.2±1.4 hours. The mean NIHSS admission score in the DECIMAL study was 22.5 for the surgery group and 23.4 for the non-surgery group, whereas in the present sample it was 20.7 and 19.9, respectively, showing that the patients of the present study were in a less serious condition upon arrival at the neurological referral service.
In the study by Mattos et al. 20 , patients with GCS<8 in the preoperative examination showed a tendency for poor outcomes, which represented a quarter of the craniectomized patients in the present study. The mean ictus-craniectomy interval for the patients of the present study was 39.7   The cardioembolic mechanism was the most prevalent in the present sample, found in 37.5% in the surgical group and 29.5% in the non-surgical group. The mean time between symptom onset and therapeutic decision among patients with hemispheric infarction in the study by Fiorot Junior et al. 21 was 33 hours for craniectomized patients and 37 hours for non-craniectomized patients, whereas, in our study, 26.5 and 31 hours, respectively. In the DESTINY I study, these values were 24.4 ± 6.9 and 23.8 ± 7.8 hours, showing that the time for making decisions about the type of treatment in our hospital is similar to international data.
In the analysis of functionality between the two therapeutic groups, the present study showed that the favorable outcome (mRS 0-3) after 6 months was better among the patients who underwent DC surgery when compared to those on CT, numerically corresponding to 37.5% (3/8) of the craniectomized patients compared to 29.4% (5/17) of the non-craniectomized patients. The proportion of patients with moderate to severe disability (mRS 4-5) was higher in the surgical group (37.5%) than in the non-surgical group (17.7%), proving that survival is accompanied by a reduction in functionality. Furthermore, the number of deaths (mRS=6) among the patients who underwent DC (25% mortality) was lower when compared to those on CT (52.9% mortality). Studies of DC in patients with traumatic brain injury showed that the procedure, even after clinical signs of brain herniation, can still lead to better functional outcomes compared to conservative treatment 22 . Although this has not shown such benefit in stroke studies, in our sample, even with a rate of 37.5% of patients in the CD group with anisocoria, CD brought a favorable outcome (mRS) for 37.5% of them. In the international literature, the DESTINY I study, with the inclusion of 32 patients, revealed a statistically significant reduction in mortality: 15 (88%) of the 17 patients randomized to hemicraniectomy versus 7 (47%) of the 15 patients randomized to conservative therapy survived after 30 days (p≤0.02).
In the study by Kim et al. 23 on the attitude of patients as to their disability after MCA hemispheric stroke, 55% of patients considered an mRS score of 3 to be the worst disability acceptable.
In sum, it is concluded that although DC surgery is a procedure that reduces mortality, the increased survival is accompanied by serious functional disability. Accordingly, it is important for the medical team to inform family members about the risks, susceptibilities and likely postoperative functional outcome for patients, considering that many of them will require lifelong special and individualized care.