In practice, a variety of grading systems are used to standardize the clinical classification, prognosis, and management of patients with acute neurological disorders. Many such grading systems for prognosis of spontaneous ICH have been developed in the past, but several of these scales used complex algebraic expressions for outcome prediction. An ideal grading scale should be simple to use, requiring no special statistical knowledge while maintaining efficacy and sensitivity. It is likely that the lack of a clinical grading system specific to basal ganglia bleed has resulted in disparities in clinical care, management, and prognosis.
Thus factors like age, GCS score, presence of comorbidities, hematoma volume, side of bleed, midline shift>5mm, and IVH were found to be statistically significant and were included in our decision-making and grouped to develop a modified ICH score. These factors impact patient treatment and prognosis, influencing the result either way. The patient's final score was calculated. It was analyzed statistically to find the best possible score below which the patient had at least 90 days survival rate irrespective of the type of management. Also, we tried to find out a minimum score above which surgery was advisable and a score above which the outcome was adverse despite of surgical management.
Age is an important factor that significantly affects post-operative recovery and response to conservative management. There is a greater mortality rate among the oldest-old regardless of ICH features. However, evidence from cohort studies are limited to sustenance this claim. According to Hegde et al., 28age was an independent predictor of mortality with age more than 70 years was associated with poor outcomes. However, because of the relatively younger age of incidence in the Indian population, we took 60 years as a cut-off value which was statistically proven significant.
A study by Gonzalez-Perez et al. assessed 30-day mortality and concluded that age, apart from increasing the prevalence of ICH, also increased the 30-day mortality rate.29 Sacco et al. also reported similar findings in a prospective analysis of 549 patients with ICH.30
On the other hand, a retrospective study published by Forti et al. in 2016 concluded that increasing age did not affect 1-year mortality in elderly subjects with acute supratentorial ICH.31 In our study, we found that patients more than 60 years of age had a significantly high mortality rate of 56.8%.
We also found a significant correlation between comorbidities like hypertension and diabetes with mortality in basal ganglia bleed patients. For ICH, hypertension is the most evidence-based and consistent risk factor. It was significantly associated with the worst clinical outcome.
Patients with significantly high blood pressure on admission and persistent inadequate blood pressure control adversely affect the prognosis in hypertensive intracerebral haemorrhage. 15
Likewise, DM also adversely affected the outcome in basal ganglia bleed patients. This could be attributed to the increased incidence of IVH and infectious complications in diabetic patients. Arboix et al. concluded that diabetes mellitus was an independent risk factor associated with mortality among ICH patients. In the diabetic and non-diabetic groups, the overall in-hospital mortality rate was 54.3% and 26.3% (P < 0.001), respectively.18 Similarly, Passero et al. also found in their study of 764 patients with ICH that high plasma glucose was associated with a greater incidence of infectious and cerebral complications and served as an independent predictor of 30-day and 3-month mortality.16 In addition to this, preictal history of heart disease has also been associated with the increase in the 30-day mortality rate in basal ganglia bleed patients.17
Left hemispheric sites of ICH were identified as significant and independent predictors of 30-day mortality19, which was consistent in our study.
GCS score32 hematoma volume 33, midline shift34, and intraventricular hemorrhage35, with or without hydrocephalus36 were independent risk factors associated with mortality and a significant determinant of surgical outcome.
In a study of 30-day mortality rates based on ICH volume, Broderick et al. 27 found that ICH volumes <30 cm3 had a 23% mortality rate for basal ganglia bleeds, ICH volumes between 30 and 59 cm3 had 64% mortality for deep haemorrhages, and ICH volumes >60 cm3 was associated with 93 percent deaths. Our mortality rates (19.4%) were similar to those reported by Broderick et al. for deep haemorrhage having hematoma volume <30 cc (23%). However, mortality rates were 92.1% for hematoma volume > 30cc in our study group. This difference in mortalities can be attributed to relatively more patients with hematoma volume greater than 60 cc in our third group ((volume >30cc).
Panchal et al. 37 reported a nearly10 percent survival rate in people with ICH with hematoma volume >30cc, consistent with our study.
OAC may be an independent predictor of mortality in basal ganglia bleed patients. Still, according to Houben et al., the addition of OAC use did not increase the predictive value of the original ICH score. 38However, in our study population, most of the patients were on antiplatelet, and the use of OAC was limited to a few patients only.
Likewise, smoking was also not associated with an increase in 30 days mortality in our study population, as published by Juvela39 and Thrift et al 40.
Management of Basal Ganglia bleeds
There is a consensus for management for cerebellar haemorrhages with mass effects or lobar haemorrhages in young patients with worsening neurologic status, in which emergent decompressive surgery is indicated. However, there are no clear-cut guidelines regarding the management of basal ganglia bleed as treatment protocol is decided considering multiple factors like size of bleed, intraventricular extension, GCS score, etc. Moreover, there is a paucity of evidence-based guidelines regarding its surgical management. There is no doubt that early diagnosis and attentive care of patients with basal ganglia bleeding are superior and crucial, given the likelihood of rapid deterioration in the first few hours after onset and early hematoma evacuation can reduce the physical and chemical damage to the surrounding brain.
Randomized trials comparing surgery to conservative management have not demonstrated a clear-cut benefit for surgical intervention, whereas some have strongly advocated surgical management.
In 1995 a landmark multi-central trial for intracerebral haemorrhage (STITCH TRIAL) randomized 1033 patients with spontaneous supratentorial intracerebral haemorrhage within twenty-four hours to early surgery or initial conservative best medical therapy. It concluded that there is no benefit from early craniotomy in supratentorial intracerebral haemorrhage compared to initial conservative management. But there was significant crossover from conservative to surgical treatments thereby possibly reducing the unfavourable outcome percentage in conservatively managed group which could have been higher in the absence of crossover. This raises a question on the generalizability of the results of such trials. Furthermore, large numbers of patients were excluded in the STICH II trial (> 3300) at randomization because of either impaired level of consciousness or at risk of herniation, adding additional selection bias to the study.
Two prospective randomized trials and three meta-analyses have been completed in the recent past that compared surgery versus conservative treatment for ICH.41-43
For the past few years, various surgical options have been considered ranging from decompressive craniectomies to minimally invasive endoscopic approaches. A meta-analysis was done by Zhou et al. in 2012 44, and Gregson et al. 45 in 2013 suggested the superiority of minimally invasive techniques over craniotomy. Still, methodological issues with this analysis have been raised. In the STITCH trial, around 77% of cases, craniotomy was the surgical procedure of choice. The remaining patients had similar numbers of hematomas removed via burr hole, endoscopy, or stereotaxy. Majority of cases in our surgical managed group underwent craniotomy and hematoma evacuation with decompressive craniectomies limited to few patients only.
Various recent retrospective studies have advocated the role of decompressive craniectomies in patients with GCS score<8, large hematomas having midline shift, and ICP not responding to medical management. 46, 47,48 In our study, 234 (69.02%) patients had craniotomies, while 11(3.2%) underwent decompressive craniectomies with hematoma evacuation. The decision for decompressive craniectomy was taken intraoperatively when even after removing the hematoma, the brain was tensed and/or nonpulsatile, and it was impossible to replace the bone.
In our study ROC curve analysis clearly showed that surgery can be opted in basal ganglia bleed patients having modified ICH score of 6 and above and patients with score of 10 or more surgery was futile as there was 100 percent mortality at this score. Thus, apart from prognostication, our scale may also help in deciding the appropriate management strategy.
Various outcome predictor scales like the ICH score 6 have been developed in past for ICH that helps in predicting the functional outcome or mortality. While prognostication is undoubtedly essential to assess treatment benefits and risks and provide patients and families with information regarding the severity of illness, these scales failed to comment on management strategies. Our scale is different from these scales as rather than merely a prognostic value, it also helps in deciding the best possible management and surgical outcome in terms of mortality.
With regard to outcome in terms of 3 months mortality, our score offers excellent or out-standing discrimination between patients in comparison to all other scales. A 93% area under the ROC curve was found to be the highest of any study for spontaneous ICH.
Among various scales used for spontaneous ICH, Essen ICH and secondary ICH score (sICH) have the highest AUCs for in-hospital mortality, 3-month mortality, and 3-month poor functional outcome. Still, they do not incorporate any radiographic characteristics 49, limiting its usefulness as a therapeutic scale since radiological parameters are significant determinants of therapeutic intervention.
Existing ICH grading scales had limited usefulness in patients of ICH with intraventricular or subarachnoid extension. Patients with incorrectly predicted mortality (indicated by a disagreement between the maximum Youden Index and mortality for the given scale) have a higher proportion of IVH than patients with correctly predicted mortality.
This difference is statistically significant in a few scales like New ICH, original ICH (oICH), modified ICH-A (mICH-A). However, in our study, precedence was given to IVH with an additional point to IVH associated with hydrocephalus which was also proven statistically.
Furthermore, the importance of the etiological factor in causing spontaneous ICH in cortical-subcortical and infratentorial regions cannot be ruled out. ICH associated with vascular malformations has been shown to have lower mortality rates and improved functional outcomes than other types of spontaneous ICH. 50 Our study incorporated pure hypertensive basal ganglia bleed excluding bleed due to other etiological factors, thereby eliminating the confounding effect of different aetiologies.
Prognosis after subcortical bleed is often a fundamental question. There is no scale available specific to basal ganglia bleed that can be used to provide initial information regarding this. In addition to prognostication, a modified ICH score offers a framework for clinical decision-making and a reliable criterion for assessing the efficacy of new treatments. Thus improved standardization of clinical assessment with a grading scale such as the modified ICH Score is likely to provide more consistency in clinical care and clinical research for subcortical bleed; just as similar assessment scales have provided consistency in traumatic brain injury, aneurysmal SAH, and ischemic stroke. Our scale can thus help translate current research into effective interventions.
Limitations
The operated cases were not pure basal ganglia bleeds. It often extended to a cortical- subcortical junction or the cortex and or thalamus. The incidence of using oral anticoagulants was significantly less in our study group, so there is a possibility of altering the result with an increase in the number of patients on oral anticoagulants. Our findings are limited by some shortcomings. There could have been other possible variables which were statistically significant in influencing the outcome.