A randomized controlled study of operative versus nonoperative treatment for large spontaneous supratentorial intracerebral hemorrhage
Correspondence Address: Source of Support: None, Conflict of Interest: None DOI: 10.4103/neuroindia.NI_151_16
Source of Support: None, Conflict of Interest: None
Keywords: Mortality, spontaneous intracerebral hematoma, surgical management
Spontaneous intracerebral hemorrhage (ICH) accounts for approximately 4–14% of all strokes and is associated with a high morbidity and mortality. Between 32–50% of the patients die within the first month and only 20% are independent 6 months after the intracerebral bleeding. Traditionally, these patients are managed by medical management alone and or by removing the hematoma surgically. Medical treatment is recommended for patients with a small cerebral hemorrhage, while the best therapeutic option for those with a medium-to-large sized spontaneous supratentorial hemorrhage remains controversial. Some studies have reported lack of efficacy of surgical evacuation whereas others have shown a lower mortality in patients who underwent surgical removal of hemorrhage., The efficacy of any medical or surgical treatment is yet to be proven by a large randomized trial. The aim of this study was to randomize the patients of spontaneous supratentorial intracerebral hemorrhage (SSICH) into the standard medical management (MM) alone or the combined surgical as well as medical management (SM) groups and to compare the outcome of these two management approaches.
A total of 482 patients with a diagnosis of SSICH were admitted to the emergency centre of King George's Medical University (KGMU), Lucknow from September 2013 to August 2015. The patients were evaluated using the following inclusion and exclusion criteria.
The inclusion criteria included an age between 20–65 years, the volume of hemorrhage on non-contrast computed tomography (NCCT) of the head >30 cc, the Glasgow Coma Scale (GCS) score of 4–14, and the symptoms appearing less than 72 hours prior to the diagnostic CT scan. The exclusion criteria included an infratentorial extension of hemorrhage, the presence of coagulation disorders, pregnancy, intra-tumoral hemorrhage, clinico-radiological (CT scan) suspicion of hemorrhage due to cerebral-aneurysm/arteriovenous malformation (AVM), or other vascular anomalies, and any concurrent serious illness that would interfere with the safety of the surgery.
A total of 61 patients fulfilling the inclusion criteria were included in the study, of which 27 were randomized to receive standard MM and the remaining 34 were randomized to receive SM. Randomization was done by computer generated random binary numbers, and to maintain concealment of allocation, a computer operator's (a nonmedical staff) help was taken. A proper written consent after explaining the details of the study and randomization was obtained. Ethical clearance was obtained from the institutional ethics committee of KGMU. All randomized patient's demographic profile consisting of the ictal duration, comorbidities, vital parameters, GCS, pupillary status, and basic blood parameters were noted. In the non-contrast computed tomographic (NCCT) scan of the head, the hematoma volume, midline shift, location (deep or superficial), and intraventricular extension were noted.
Conventional standard medical therapy included maintaining the airway by placing a nasopharyngeal tube, an endotracheal intubation, or a tracheostomy as and when required, O2 by mask or ventilator support, head end elevation by 30°, hydration with input-output charting, control of blood pressure, reduction of intracranial pressure (ICP) using mannitol and furosemide, prophylactic antiepileptic therapy to prevent seizures, prophylactic antibiotics, appropriate feeding, physiotherapy, and management of any associated morbidity, if present. All the patients were kept in the intensive care unit (ICU) with close monitoring of vital parameters and blood coagulation profile by a team consisting of neurosurgeons, neurophysicians, and critical care experts. Serum sodium, potassium, calcium, and magnesium were checked daily and necessary corrections were done. Acid base imbalance was monitored and corrected. The surgical management included a craniotomy, middle frontal gyrus approach, hematoma evacuation, and achievement of hemostasis. No thrombolytic agents were used in any of the patients. External ventricular drain (EVD) was inserted if ventricles were exposed during the surgery, which was removed in 2–3 days or when cerebrospinal fluid (CSF) became clear. During the follow-up, a NCCT head with complete neurological examination was performed and the modified Rankin Scale (mRS) grade was noted. Primary and secondary outcome measures were mortality and dependency (mRS grade) at 3 months, respectively. Patients with a mRS of ≤3 were considered to have a good functional outcome, whereas those with a mRS of >3 were considered to have a poor outcome.
The data was analyzed using the Statistical Package for Social Sciences version 15 (SPSS Inc.); chi-square test and Student's t-test were used for data analysis. Logistic regression analysis was employed to assess the effect of multiple variables in the outcome of spontaneous ICH. The confidence level of the study was kept at 95%, hence a P value less than 0.05 indicated a statistically significant association.
The outcome in both the groups were studied with relation to the patient's age and sex, personal and past history, time between the ictus and first CT scan, the GCS score at admission, the best motor response at admission, pupillary asymmetry, size of hematoma, location of hematoma, intraventricular extension of hematoma, midline shift, and ICH score at admission. [Table 1] demonstrates that the distribution of the baseline characteristics of the medical and surgical groups were comparable with respect to age, sex, GCS on admission, and CT scan findings such as the hematoma volume, location, intraventricular extension, and midline shift.
On univariate analysis, it was found that mortality had significant association with various parameters [Table 2].
Postoperative NCCT head showed <10 cc of residual blood in 76.5% (26 out of 34) of the operated patients, whereas 8 patients (23.5%) had more than 10 cc of residual blood. Postoperative residual blood ranged from less than 5 ml to 20 ml, with a mean volume of 8.9 ml [Table 3].
On binary logistic regression analysis of the treatment arm, assessing parameters such as the GCS, motor response, hematoma volume, midline shift, IVH, location of hematoma, pupillary asymmetry, and ICH score (where mortality was taken as a dependent variable), it was found that only the treatment arm, hematoma volume, and location of hematoma had significant association with outcome [Table 4].
This multivariate model showed an accuracy of 85.2%, thus showing a high discriminant value. The sensitivity of the model was 90.9% and the specificity was 70.6%, and hence, it was considered to be a more sensitive than specific model [Table 5].
At 3 months, only 17 patients (MM-4/SM-13) had survived, out of which 2 patients (50%) in the MM group and 3 patients (23.1%) in the SM group had a favourable outcome (mRS grade ≤3) but showed no significant difference (P = 0.301). The mean mRS grade in the MM group was 3.5 ± 1.290, and in the SM group, it was 3.9 ± 0.862. The univariate and multivariate analysis of all the parameters with mRS showed no significant association (P > 0.05).
Although spontaneous intracerebral hemorrhage accounts for only 15% of all strokes, it is one of the most disabling forms of stroke. Many studies have shown that the level of disability and mortality after ICH depends on the GCS score at presentation, hemorrhage size, ventricular extension, and patient's age., Surgery has been typically done in younger patients with worse or deteriorating GCS scores and slightly larger hemorrhages. There is no convincing evidence of benefit from any medical treatment, and the role of surgery remains controversial. Several clinical and radiological factors such as age, level of consciousness, hypertension, volume of the hematoma, volume of peripheral edema, midline shift on the initial CT scan, and intraventricular spread of the bleeding appear to be markers of a poor prognosis after spontaneous ICH.,,,,, However, only a few studies have attempted to identify factors related to a favourable functional outcome in patients suffering from spontaneous supratentorial ICH.,, Lisk et al., observed that the GCS score, hemorrhage volume, age, and gender were important predictors of a poor outcome in patients with spontaneous ICH. Daverat et al., on logistic regression analysis, found five independent predictors of a satisfactory outcome at 6 months, namely, the age, hemorrhage size, intraventricular spread of the hemorrhage, limb paresis, and CSF pathway communication disorders. According to Castellanos et al., a good outcome in a medium-to-large sized ICH can be predicted on admission by three readily assessable factors, i.e., the Canadian stroke scale (CSS) score, ICH location, and fibrinogen levels. A low severity of neurological deficit assessed by the CSS, cortical location of the haemorrhage, and low fibrinogen levels accurately predicted a good outcome in 85% of the patients in their study.
The GCS score is now a standard neurological assessment tool that is reproducible and reliable. It has been associated with the ICH outcome in most of the prediction models, as in the University of California, San Francisco (UCSF) ICH cohort.,, Barbara et al., suggested that improved outcomes can be achieved with early surgery if the decision to operate is made within 8 hours of ictus, with hematomas of 20–50 ml, for patients with a GCS of 9 and more, or for patients aged 50–69 years. In particular, they suggested that, when the GCS is below 9, early surgery does not significantly improve outcome. This argument suggests that, once the GCS has dropped to below 9, irretrievable damage has already occurred and surgery will not be successful in rescuing the patient. Juvela et al., found that there were no significant differences in the mortality or morbidity rates between the two treatment groups (conservative and surgical group); however, the mortality rate of the surgical group with GCS scores of 7–10 was significantly lower than that of the corresponding conservative group. In a retrospective, nonrandomized study in Japan, Kanaya and Kuroda compared the effects of surgical evacuation of the hematoma on mortality in 3638 patients with a putaminal hemorrhage to those of medical management in 3372 patients. On the basis of the results, they recommended surgical treatment if the hematoma size is more than 30 ml and the level of consciousness is somnolent to semi-comatose.
The Surgical Trial in Intracerebral Hemorrhage (STICH) studied 1033 patients with spontaneous supratentorial ICH enrolled within 72 hours after stroke. For enrolment, the ICH diameter more than 2cm and a GCS of 5 or more was considered. Hemorrhages caused by a vascular abnormality, brain tumor, or trauma were excluded, along with ICH located in the cerebellum or in the brain stem. The STICH investigators concluded that “patients with spontaneous supratentorial ICH in neurosurgical units showed no overall benefit for early surgery when compared with initial conservative management.” However, the trial showed that patients with superficial hemorrhages (distance to cortical surface less than 1cm) benefited from surgery. A good outcome was observed in 26% of the surgical and 24% of the medically treated group. Even the mortality after 6 months was nearly identical, 36% versus 37%, respectively, and the dependency of the patients was 72% in the medical group and 67% for the surgical group at 6 months. Furthermore, the trial showed that a poor initial GCS (less than 9) was associated with a poor outcome regardless of the surgical or nonsurgical treatment. The results of the STICH trial still discussed and interpreted in different ways. In our study, the mortality rate at 3 months was significantly lower in the surgical group (61.8%) compared to the conservative group (85.2%) [P = 0.043], and the mortality rate was significantly higher (80.5%) in patients with a GCS of 4–8 compared to a GCS of 9–14 (55%) both in the surgical as well as conservative group (P = 0.037). However, the mortality rate in the surgical group (65.2%) was found to be significantly lower in the group with GCS 4–8 compared to the conservative therapy group (100%) [P = 0.005]. The survival rate at 3 months was found to be directly proportional to the motor response at the time of admission (P = 0.021).
The ICH volume has been consistently associated with the outcome in the ICH prediction models., Larsen et al., observed that the acute mortality of ICH was 27%, and determinant for the immediate prognosis was the level of consciousness and the volume of hematoma. The crucial size was 50 ml with a mortality of 90% for larger hematomas and only 10% for hematomas smaller than that. In another study by Mitra et al., an age of more than 60 years, a GCS of 6 or less at the time of admission, an ICH volume greater than 30 ml, a midline shift seen on the CT scan of more than 3 mm, and the presence of intraventricular hemorrhage (IVH) and hydrocephalus had an adverse impact on outcome. In our study, the mortality rate at 3 months was found to be directly proportional to the volume of hematoma (P = 0.039) with 100% mortality for patients with a volume >90 ml. The mortality was 56.4% in patients with volume in the range of 31–60 ml whereas it was 81% in patients with the volume in the range of 61–90 ml. The mortality rate in the group of 31–60 ml volume was significantly lower in the surgical group (35.7%) than in the conservative group (77.8%) (P = 0.016).
The mass effect that results from the volume of the hematoma, the edematous tissue surrounding the hematoma, and obstructive hydrocephalus with subsequent herniation remain the chief secondary causes of death in the first few days after an intracerebral hemorrhage. Fogelholm et al., described that the most important independent predictors of death within the first 28 days were unconsciousness on admission and >6 mm lateral shift of cerebral midline structures. Nag et al., found that the midline shift is prognostically poor only when coexisting with other factors causing a mass effect such as ventricular compression by the hematoma. In our study, the mortality rate at 3 months was higher with an increase in the midline shift irrespective of the group (MM/SM); however, the mortality rate was significantly higher in the conservative group (90.9%) with a midline shift of more than 5 mm compared with the surgical group (63.3%), which was statistically significant (P = 0.023).
The presence of intraventricular blood has been strongly associated with an impaired consciousness at presentation. Simple comparison between the ICH patients with or without an IVH extension suggests that mortality is substantially increased if IVH is present. Larsen et al., observed that intraventricular hemorrhage was a bad prognostic sign in the ganglionic-thalamic hematomas. In a study by Hallevi et al., the patients with an IVH were twice as likely to have a poor functional outcome (discharge mRS >3) when compared to patients without an IVH (P = 0.001). In the present study, the mortality rate was higher in patients having an intraventricular extension of the hematoma irrespective of the treatment group (MM/SM) [P = 0.132] and was significantly higher in the MM group (100%) compared to the SM group (80.6%) [P = 0.02].
In many studies, the most favorable outcome was recorded in patients with a subcortical hematoma and the worst outcome was recorded in the presence of brainstem hematomas., Castellanos et al., observed that the cortical location of the bleeding was an independent predictor of a good short-term ICH outcome. In the present study, the patients with a basal ganglia hematoma had significantly higher mortality rates at 3 months (78.8%) compared to patients with a lobar hematoma (67.8%) in both the groups (MM and SM) [P = 0.005]; however, the patients with a basal ganglia hematoma in the MM group had significantly higher mortality rates (91.6%) compared to the SM group (67.8%) [P = 0.036].
Portenoy et al., observed that a low GCS score, coma, ataxic respiration, abnormal pupils, acute hypertension, large hemorrhage size, and intraventricular extension of blood were associated with a poor outcome. Multivariate analysis using the technique of logistic regression identified three variables, the GCS score, hemorrhage size, and intraventricular extension of blood, which were most predictive of the outcome. In the present study, the mortality rate was higher in patients having asymmetrical pupils (94.1%) compared to patients having symmetrical pupils (63.6%), irrespective of the group (MM or SM) (P = 0.017);however, the patients having symmetrical pupils in the conservative group had significantly higher mortality rates (80%) as compared to the surgical group (50%) (P = 0.039). ICH score is a grading scale for intracerebral hemorrhage developed by Hemphill et al., in 2001. They observed that factors independently associated with a 30-day mortality were the GCS score, age ≥80 years, infratentorial origin of the ICH, the ICH volume, and the presence of intraventricular hemorrhage. They noticed that all 26 patients with an ICH score of 0 survived and all 6 patients with an ICH score of 5 died. The thirty-day mortality increased steadily with the ICH score (P = 0.005). In our study, the mortality rate increased with an increase in the ICH score (P = 0.023) with a 90% mortality in patients with an ICH score of 4, and the mortality rate being reduced to 50% in patients with an ICH score of 2. Interestingly, however, the mortality rate in patients with an ICH score of 3 was significantly lower in the SM group as compared to the MM group (P = 0.043).
Zuo et al., observed that gross total removal of the hematoma is an effective method to decrease the ICH induced injury to brain tissue due to decreased perihematomal edema formation and secondary injury by coagulation end products-activated inflammatory cascade. In the present study, out of 34 patients, the mortality rate at 3 months was significantly higher (100%) when a residual hematoma volume of more than 10 cc was present compared to the patients in whom the hematoma volume was less than 10 cc (50%) [P = 0.011].
A number of studies have shown a direct relationship of the hematoma volume with the clinical outcome in ICH.,,,,, In a study by Davis et al., the percentage hematoma growth, the initial ICH volume, the GCS score, and the presence of an IVH had a significant association with mortality (P < 0.05). In addition, the percentage hematoma growth, the initial ICH volume, the GCS score, and the age predicted the predicted the functional outcome (mRS) with statistical significance (P < 0.05). According to their observations, hematoma growth is an independent determinant of both mortality and functional outcome after intracerebral hemorrhage. Rost et al., concluded that the size of the hemorrhage was frequently used in clinical decisions in patients with ICH, and scores predicting mortality as well as good functional outcome were developed using ICH volumes categorized as <30 cm 3, 30–60 cm 3, and >60 cm 3. Nag et al., showed that the initial hematoma volume was an independent predictor of the clinical outcome and a hematoma volume >30 cm 3 was a bad prognostic factor with a higher National Institutes of Health Stroke Scale (NIHSS) score on admission and an early mortality. Except for brain stem hematoma, no other location of stroke was found to be prognostically significant with respect to the functional recovery.
Surgically treated patients with hematomas smaller than 50 ml made a significantly better functional recovery than patients of the medically treated group; however, both the groups had a comparable mortality rate. Patients with larger hematomas showed significantly lower mortality rates after surgery; however, they did not have better functional recovery rates than the medically treated group. The outcome of surgical patients with putaminal or thalamic hemorrhage was no better than for those who underwent medical treatment; however, there was a trend toward better quality of life and chance of survival in the operated group. Zuccarello et al., reported no significant difference in mortality at 3 months. Analysis of the secondary outcome measures at 3 months showed a nonsignificant trend toward a better outcome in the surgical treatment group versus the medical treatment group for the median Glasgow Outcome Scale (GOS), Barthel Index, and Rankin Scale, as well as a significant difference in the NIHSS score (P < 0.05). Kim et al., observed that, on univariate analysis, a positive effect of the surgical treatment in reducing mortality at 90 days (P = 0.002), as well as on the GOS score (P = 0.006) and mRS at 90-days (P = 0.023) was present. However, on multivariate logistic analysis, there was a significant difference only in the mortality (P = 0.036) between the groups at 3 months.
Many studies have shown that there were no significant differences in the morbidity between the two treatment groups (conservative and surgical group).,,,, Morgenstern et al., in 1998 could only demonstrate a modest early mortality benefit from surgery with a dependency rate of 35% and 47% for the medical and surgical groups, respectively, at 6 months. A randomized study by Batjer et al., showed that none of the patients were capable of returning to the prestroke activity at 6 months, and only 19% were capable of independent life at home with no significant difference between the groups. The STICH–II trial concluded that early surgery does not increase the rate of death or disability at 6 months and might have a small but clinically relevant survival advantage for patients with spontaneous superficial intracerebral hemorrhage without IVH. The prognosis-based Rankin scale showed a favorable outcome in 47% of the patients in the early surgery group, and in 44% of those in the initial conservative treatment group (P > 0.05).
In the present study, the patients randomized to the SM group were more severely disabled (76.9%) compared to those assigned to the MM group (50%) at 3 months, although the results were not statistically significant (P = 0.301). Overall, 17 patients (13+4) were alive at 3 months. Among them, only 5 patients (3 with SM+2 with MM) were able to look after themselves (mRS grade ≤3). We were not able to assess the predictive factors responsible for independency at 3 months due to the small numbers of surviving patients at 3 months (n = 17), and on univariate and multivariate analysis of mRS with various parameters, these factors were found to be insignificant in the two groups (P > 0.05).
In our study, both the univariate and multivariate analysis showed results in favor of SM in comparison to MM in terms of mortality (P < 0.05), whereas there was no significant difference in the functional outcome at 3 months (P > 0.05). Therefore, a randomised study recruiting a large number of patients is required to establishing a protocol and a recommendation, which can assist in the formulation of universally acceptable guidelines.
In SSICH patients, the mortality was found to be significantly associated with the treatment arm, hematoma volume and location of the hematoma (deep or superficial). Our results were in favor of surgical treatment among patients presenting with GCS 4–8, with a hematoma volume 31–60 ml, a midline shift of more than 5 mm, an intraventricular extension of the hematoma, and without pupillary asymmetry. During surgery, the target should be gross total removal of the hematoma.
We sincerely thank Dr. Mukta Meel and Vivaan for their constant support and help during the study and in preparation of the manuscript including its submission.
Financial support and sponsorship
Department of Neurosurgery, King George's medical university, Lucknow - 226 003, Uttar Pradesh, India.
Conflicts of interest
There are no conflicts of interest.
[Table 1], [Table 2], [Table 3], [Table 4], [Table 5]