Clinical Profile and Predictors of Outcome in Spontaneous Intracerebral Hemorrhage from a Tertiary Care Centre in South India

Background. This article attempts to evaluate the clinical profile and outcome determinants following hypertensive SICH in a South Indian population. The study represents the largest series of SICH reported from a single centre in India. Materials and Methods. Prospective data collection and analysis of patients with SICH admitted to our centre between 1st January 2015 and 31st December 2018. The variables analysed include: age, sex, comorbidities, Glasgow coma score (GCS) on admission, radiological features, treatment modality, and outcome at three months. Modified Rankin score (mRS) was used to assess the outcome at discharge and three months. Results. Our study group of 905 patients included 638 males and 267 females and the mean age at presentation was 58.10 ± 12.76 years. The study group included 523 patients (57.8%) previously diagnosed hypertensive, of whom 36.3% (n = 190) were on irregular medication. The most frequent locations of hematoma were basal ganglia (478), thalamus (202), lobar (106), cerebellar (61), brainstem (31), and primary intraventricular haemorrhage (27). Secondary intraventricular extension was seen in 425 (47%) patients on admission. The mean volume of the clot on admission was 23.45 ± 19.79 ml, and clot progression was seen in only 46 (5.08%) cases. Surgical evacuation through craniotomy was done in 147 (16.8%) patients, and external ventricular drainage (EVD) was placed in 56 (6.2%) patients. Overall 3−month mortality was 30.1% (266 patients). On the last follow up a favorable outcome (mRS 0−3) was observed in 412 (45.53%) patients and a poor outcome (mRS 4−5) in 207 patients (22.87%). Independent predictors of mortality are Age >70 (p = <0.001, OR 4.806, 95% CI 3.064–7.54), admission GCS <8 (p = <0.001, OR7.684, 95% CI 5.055#x2013;11.68), and Hematoma volume >30 ml (p = <0.001, OR 2.45, 95% CI 1.626–3.691). Intraventricular haemorrhage was an additional poor outcome predictor (p < 0.015, CI 1.105–2.585). Surgical evacuation reduced mortality in the group, but morbidity rates remained the same. Conclusions. SICH predominantly affects a younger population in India in comparison to the Western society. Elderly age, poor GCS on admission, clot volume above 30 ml and intraventricular extension remain the most consistent predictors of death and poor outcome. Further studies are needed to assess the risk of SICH among hypertensive patients and to prognosticate the outcome after SICH using novel predictors, including biomarkers.


Introduction
Spontaneous Intracerebral Haemorrhage (SICH) is the second most common cause of stroke and accounts for 7.5−30% of all strokes [1,2]. Haemorrhagic stroke is generally associated with higher morbidity and mortality rates than ischemic stroke [3][4][5]. Only one-fi h of the patients regain functional independence a er SICH and between one-fourth to half of the patients succumb to the bleed [1,6]. Optimal management is controversial, and considerable debate exists primarily on the role of surgery in SICH [7]. In developing countries like India with severe resource constraints, treatment strategies need to be customised given the high morbidity and mortality associated with SICH. Published literature from India on outcome following SICH is limited [8][9][10][11][12][13][14][15][16][17][18]. e goal of this single institution based study was to analyse the clinical profile, to assess the functional outcome and to identify the outcome prognosticators a er SICH in a tertiary care hospital located in a coastal town in South India.

Materials and Methods
e study was approved by the Institute Ethics Committee, Manipal University (Approval No IEC 209/2015). All patients above the age of 18 years who were presented to the Emergency Department between 1 st January 2015 and 31 st December 2018 with computerised tomography (CT) evidence of SICH were included in the study. Patients with post-traumatic hematomas, intracranial space-occupying lesions with bleeds, haemorrhagic transformation of an ischemic stroke, vascular malformations, and aneurysms were excluded from the study. Demographic data, comorbidities, clinical, radiological data, and information about surgical interventions were recorded in an online registry. Volume of hematoma was measured using the axbxc/2 method [19]. Hematoma expansion was determined if sequential brain imaging was available, and defined as a relative parenchymal volume increase of more than 33% from initial to follow-up imaging within 3-72 hours [20]. All patients were admitted to an exclusive neurosurgical intensive care unit and started on antihypertensives (Amlodipine, Labetalol) and antiedema measures (Mannitol 20%). Patients with lobar hematoma and patients planned for surgery were started on antiepileptics (Phenytoin sodium 7 mg/kg in divided doses). All patients had a follow-up CT scan at 24 hours of admission or on clinical deterioration, whichever was earlier. Surgical evacuation was offered to all patients with supratentorial hematoma volume >30 ml or midline shi of >1 cm. Posterior fossa cerebellar hematomas with maximum diameter >3 cm were offered surgical intervention.
alamic or Basal Ganglia bleeds with intraventricular extension and significant hydrocephalus with a GCS<8 were managed with an External Ventricular Drain (EVD) followed by administration of Inj. Streptokinase 30,000 IU daily through the EVD. e EVD was retained for a maximum of five days following its insertion. e outcome determinants were mortality and morbidity as measured using the modified Rankin Scale (mRS) at discharge and three months. A mRS of 4 and 5 were considered as poor outcome and mRS of 0−3 as good outcome. Ninety day mRS was recorded in the outpatient clinic. In cases where the patient failed to visit the clinic, a telephonic mRS was recorded.

Statistical Methods.
Statistical analyses were performed with SPSS 24.0. Descriptive statistics including mean, median and standard deviation were computed for baseline characteristics. Chi-Square test was used to compare categorical variables, and Student t-test and Mann Whitney test was applied to calculate the p-value for continuous variables for univariate statistics. Wilcoxon signed-rank test was used to compare mRS at discharge and 90 days. Predictors of poor outcome and death at three months were analyzed using logistic regression analysis. Variables with a p-value less than 0.05 at univariate level were considered significant.

Results
In total, 1052 patients with spontaneous ICH requiring hospitalization were identified within the study period. Based on our inclusion and exclusion criteria, 905 patients were included in the study. Twenty patients were lost to follow up, and their mRS at 90 days could not be obtained.

Patient Demographics.
e mean age of our study group was 58.10 ± 12.76 (18 years to 93 years), and 638 (70.5%) of them were males. Nearly one-third of the patients (266 patients 29.39%) were below the age of 50 years.

Risk Factors.
History of arterial hypertension was present in 57.8% (푛 = 523) of patients. Of these 523 hypertensive patients, nearly one-third 36.3% (푛 = 190) were on irregular treatment, and their mean duration of hypertension was 5.23 ± 4.64 years. e remaining 333 patients were newly diagnosed hypertensive on presentation. A definitive history of diabetes was available for 245 (27.1%) patients with a mean duration of 7.1 years. e mean random blood glucose levels on admission were 162.49 ± 69.05 dl. At the time of the ICH, a total of 83 (9.2%) patients were on antiplatelet medications (Table 1).

Surgery.
Surgical evacuation of the clot through craniotomy was performed in 147 (16.8%) patients. External ventricular drainage for intraventricular haemorrhage was performed in 56 (6.2%) patients.

Outcome.
e overall mortality was 30.1% (266 patients). Of these, 116 (12.81%) patients succumbed during initial hospitalization, and 150 (16.57%) patients died within three months of follow up. For survivors at three months, the median mRS was 4 (IQR 3−6). Functional status at discharge and three months are presented in Table 4. Among 639 survivors at three months, 207 (33.4%) had a mRS 4-5 (poor outcome), and 412 had a favourable outcome (mRS 0−3), and twenty patients were lost to follow up. We observed that following discharge, many of the patients showed a gradual recovery with rehabilitative measures. e number of patients with favourable outcome ( mRS <=3) increased from 196 to 412 at the end of 3 months. e number of patients with poor outcome (mRS 4-5) dropped to 207 at the end of 3 months (Table 4). Wilcoxon Signed Rank Test showed this recovery pattern to be statistically significant (푝 =< 0.001) ( Table 4).

Prognostic Factors.
Summary of the analysis of the various factors related to mortality and outcome are shown in Table  1.
Step wise Binary logistic regression of significant factors on T 1: Demographic, etiological, clinical and radiological characteristics of patients with outcome and mortality. GCS-glasgow coma Score, ICH-intracerebral hemorrhage, IVE-intraventricular haemorrhage, EVD-external ventricular drain.

푛 = 905
Alive ( [21]. e mean age of patients in our study was 58.10 ± 12.76, and twenty-seven patients (2.9%) were above 80 years in our group. is comparatively younger age of incidence has been reported uniformly across India and appears to be a characteristic feature of the SICH in univariate analysis was performed. Age > 70 years, GCS < 8, Volume > 30 ml, presence of hydrocephalus, and hematoma growth were significant factors predicting mortality. Similarly the factors predicting poor outcome (mRS 4-5) were Age > 70 years, GCS < 8, Volume > 30 ml, and intraventricular extension. e summary of the above parameters with their significance and Confidence interval are tabulated in (Table 5).

Clot Volume.
Volume of the clot is also a crucial radiological predictor of outcome [2]. It has been reported that each mL increase in baseline clot volume is associated with a 1% increased risk of mortality [31]. It has also been shown that clinically significant hematoma growth occurs in up to 1/4 of ICH patients [32] and for each 10% increase in hematoma volume growth, the risk of death increases by 5% [31]. e mean volume of the clot in the mortality group was 32.45 ± 24.39 ml, and the mean volume amongst the survivors was 17.55 ± 15.59 ml (푝 < 0.001). We dichotomized our supratentorial bleed patients into two groups of volume of 30 ml and observed that 125/234 (53.6%) with clot volume >30 ml were dead at three months. Hematoma expansion was seen in 46 (5.08%) patients in our study. e mean duration of time of onset to the first CT scan in these 46 patients was 6 : 20 hours (2-17 : 15 hours). e rate of expansion is lesser than other reported studies [28]. e precise reason for the low incidence of clot expansion in our series is not clear. e probable reasons could be relatively younger age at presentation, intensive blood pressure reduction measures, no patients with anticoagulants and lower incidence of patients on antiplatelet medications.

Site of Hematoma.
Infratentorial hematomas, especially brain stem hematomas, are known to carry a poor outcome compared to supratentorial hematoma. In our series, 17 of the 31 patients with brain stem hematoma expired, and the median mRS at 90 days for the remaining 14 patients was 6. Of the 61 patients with cerebellar hematoma 19 expired and the median mRS at 90 days for the remaining 42 patients was 3. In the INTERACT 2 trial, of the 2,066 patients included in the analyses, the involvement of posterior limb of internal capsule and thalamus and infratentorial sites increased risks of death or major disability [33]. In our series of 780 patients with supratentorial clots, lobar hematomas had the best outcome (푝 < 0.001).

Intraventricular Extension. Intraventricular extension
(IVE) is seen in around 40-60% of patients with spontaneous ICH and is known to be a significant predictor of 30-day mortality and long term outcome [9,12,21]. Subgroup analysis from the STICH 1 Trial data concluded that the absence of IVH resulted in better outcomes. (31.4% vs. 15.1%; 푝 < 0.001) the Indian subcontinent [9,12,22,23]. Increasing age is also associated with increased morbidity and elderly population above 70 years had a high mortality rate (46.1%) in our series. Several authors have shown that older adults with acute ICH experienced the worse outcomes compared with their younger counterparts, including death, dependency, and overall quality of life [5,[24][25][26].

Hypertension.
Hypertension is the most common risk factor in all the studies relating to SICH. Feldmann et al. have reported a relative risk of 3.9 for ICH in patients with hypertension [27]. A definite history of hypertension could be elicited in only 57.8% of our patients, and nearly third of these patients were noncompliant hypertensive on irregular medication. e remaining patients had never under gone a prior medical evaluation and were diagnosed to have raised blood pressure on presentation. is observation is in contrast to the higher rates of preexisting hypertension (70−80%) observed in other Indian studies [9,12,22]. e adverse effect of high blood pressure is assumed to be through hematoma expansion, which was, however, observed only in 5.08% patients in our series. is increase in hematoma volume did not correlate with mortality or outcome in our series (푝 = 1.227). We do follow a policy of aggressive blood pressure reduction as it has been well proven that aggressive BP lowering is beneficial in reducing haematoma growth, however it seldom translates to the better clinical outcome [28].

Diabetes and Hyperglycemia.
It is proposed that high blood glucose at admission contributes to poor outcome, due to exacerbation of cerebral oedema and cerebral damage. A recent meta-analysis by Zheng et al. in 2018, concluded that hyperglycemia was associated with poor functional outcome in patients with ICH [29]. However, the pool of available evidence about blood glucose variability and ICH is still limited, and random blood glucose has not been a predictor of mortality in the Indian ICH studies [9,12,30]. In our series, multivariate logistic regression analysis failed to demonstrate an association between blood glucose and poor outcome.

Morbidity and Mortality
Less than half of the patients with ICH survive one year, and less than a third survive five years [40]. Poon et al. 's systematic review and meta-analysis of 122 longitudinal cohort studies reporting long-term (>30 days) outcome a er spontaneous 'primary' ICH have shown a 1-year survival of 46.0% and a 5-year survival of 29.2% following SICH [31]. e 1-month case fatality a er ICH has remained unchanged for several decades at around 40%, but the outcome in the longer term is less clear [5]. In addition to the risk of mortality and poor functional outcome, survivors are also at a considerable risk of recurrent ICH and vaso occlusive events, including ischemic strokes. ree of our patients presented again with SICH at other sites and two developed ischemic infarcts. Our mortality rate at the end of 3 months was 30.1%, which is marginally better than other quoted Indian studies [9,12]. With a median mRS score of 4 at 90 days follow up our overall outcome too appears to be comparable to other reported Indian studies.

Limitations
e major limitations of our study are short follow up (90 days) and the fact that the outcome assessment is restricted to mRS. Other important parameters, including cognitive disability etc. have not been assessed. Biomarkers and other novel predictors like neurotophil-lymphocyte ratio was not evaluated.

Conclusion
Spontaneous Intracerebral Hemorrhage predominantly affects a younger population in India in comparison to the Western population. Age > 70 years, GCS < 8 on admission, clot volume above 30 ml, and intraventricular extension and hematoma growth and hydrocephalus remain the most consistent predictors of death and poor outcome following spontaneous intracerebral hematoma. Further studies are needed to assess the risk of SICH among hypertensive patients and to prognosticate outcome a er SICH using novel predictors, including bio markers.
Data Availability e data used to support the findings of this study are available from the corresponding author upon request.

Conflicts of Interest
e authors declare that they have no conflicts of interest regarding the publication of this paper.
Funding e authors did not receive any funding for this study. and the presence of hydrocephalus lowered the likelihood of favourable outcome still further to 11.5% (푝 = 0.031) [7]. Delayed Intraventricular haemorrhage (dIVH) too is known to adversely affect the outcome. In the INTERACT 2 study, dIVH had greater odds of 90-day death or major disability versus initial IVH [34]. In our cohort, 425 patients (47%) had IVE of hematoma, and IVH was an independent predictor of the poor outcome (푝 = 0.015). Hydrocephalus was observed in 236 patients (26.1%) in our group and was also an independent predictor of mortality (푝 = 0.014).
is probably can be attributed to the significant number of small to moderate thalamic bleeds with IVE in our study.

Surgery. Role of surgery in SICH remains controversial.
A meta-analysis published by Prasad et al. a er the STICH 1 trail concluded that surgery added to medical management reduces the odds of being dead or dependent compared with medical management alone [35]. In Troberg's, study supposedly the most extensive study on long-term functional outcome with the longest follow-up (up to 10.8 years) of survival and functional status a er surgery for ICH, 31% of all operated ICH patients were deceased a er 1 year and only 24% of patients available for assessment of longterm functional outcome were independent in activities of daily life [33]. Preictal heart disease and reduced level of consciousness before surgery were the most consistent predictors of mortality regardless of follow-up time. Our study, however, showed no significant influence of clot evacuation on mortality (푝 = 0.621).

ICH Scoring.
Several scores for predicting functional outcome and mortality a er ICH have been developed. e ICH score proposed by Hemphill has stood the test of time for prognosticating SICH [8]. For SICH scores of 0, 1, 2, 3, 4, and 5 our mortality rates were 7%, 16%, 33%, 62%, 95%, and 50% in comparison to Hemphill et al. 0%, 13%, 26%, 72%, 94%, and 100%. e Essen ICH score is determined only by clinical variables (age, the severity of neurological deficits, and level of consciousness) and has the advantage of not requiring the measurement of ICH volume [36]. e FUNC score additionally takes into consideration the occurrence of preICH cognitive impairment [37]. More recently, scores like the BAT score have been designed for identifying predictors of hematoma expansion [38]. We prefer to use the ICH score due to its ease in application. In an earlier publication, Hegde et al. attempted to validate the ICH score in an Indian setting and suggested reducing the age cut off from 80 years to 70 years in the original SICH score [18].
is was influenced by the fact that the mean age of the affected group in our study as well as in other Indian studies is much younger compared to the Western population. However, we do agree with Pinho et al. that even though the use of prognostic scores is recommended other factors must also be weigh when evaluating individual patients and an early subjective clinical judgment by experienced clinicians is not inferior to the application of formal prognostic scores in predicting outcome [24,39].