Risk factors for unfavorable outcome after spontaneous intracerebral hemorrhage in elderly patients

Background: Spontaneous intracerebral hemorrhage (SICH) of the elderly is a devastating form of stroke with a high morbidity and economic burden. There is still a limited understanding of the risk factors for an unfavorable outcome where a surgical therapy may be less meaningful. Thus, the aim of this study is to identify factors associated with unfavorable outcome and time to death in surgically treated elderly patients with SICH. Methods: We performed a single-center retrospective study of 70 patients (age > 60 years) with SICH operated between 2008 and 2020. Functional outcome was assessed by modified Rankin Scale. Various clinical and neuroradiological variables including type of neurosurgical treatment, anatomical location of hemorrhage, volumetry and distribution of hemorrhage were assessed. Univariate and multivariate logistic regression models were performed. Length of stay (LOS) and hospital costs are presented. Results: The overall mortality (mean follow-up time of 22 months) in this study was 32/70 patients (45.71%), 30-days mortality was 8/70 (11.42%), and 12-months mortality was 22/70 (31.43%). Average LOS was 73.5 days with a median of 58, 766 € estimated in hospital costs per patient. Multivariate analysis for 12-months mortality was significant for intraventricular hemorrhage (IVH) (p = 0.007, HR = 1.021, 95% CI = 1.006 – 1.037). ROC analysis for 12-months mortality for IVH volume > = 7 cm 3 presented an are under the curve of 0.658. Conclusions: We identified IVH volume > 7 cm 3 as an independent prognostic risk factor for mortality in elderly patients after SICH. This may help clinicians in decision-making for this critical and growing subgroup of patients.


Introduction
Spontaneous intracerebral hemorrhage (SICH) is a devastating neurological disorder and the second most common type of stroke with an incidence between 15.9 and 36.9 per 100, 000 persons, depending on the type of population [1,2].Despite the improvement in neurosurgical techniques, neuroanaesthesia, neuroimaging, and intensive care medicine, reported mortality ranges up to 55%, demonstrating the severity of this disease [1,3,4].The extent and type of treatment remain controversial and whether patients benefit from neurosurgical treatment including external ventricular drainage (EVD) insertion, hematoma evacuation or a more conservative approach on an intensive care unit (ICU) with optimized neurological treatment is highly debated.So far, previous efforts to answer some of the questions related to SICH management have been unsuccessful.For instance, the STICH trial for early surgery in lobar and deep-seated intracerebral hemorrhage (ICH) did not find any overall benefit for surgical treatment compared to conservative treatment [5].The subsequent STICH II trial addressing early surgery in supratentorial lobar hemorrhage only, assuming a more favorable outcome for this subgroup, similarly did not show any superiority of surgical treatment compared to conservative management [6].Further studies investigating different neurosurgical approaches for different Abbreviations: AUC, area under the curve; CCT, cranial computed tomography; CI, confidence interval; COPD, chronic obstructive pulmonary disease; CSF, cerebrospinal fluid; DM, Diabetes mellitus; EVD, external ventricular drain; GCS, Glasgow coma scale; HR, hazard ratio; ICH, intracerebral hemorrhage; ICP, intracranial pressure; ICU, intensive care unit; IVH, intraventricular hemorrhage; LOS, length of stay; MRS, modified Rankin Scale; ROC, receiver operating characteristic curve; SAH, subarachnoid hemorrhage; SICH, spontaneous intracerebral hemorrhage; STICH, Early surgery versus initial conservative treatment in patients with spontaneous supratentorial intracerebral haematomas in the International Surgical In the elderly, however, this question becomes particularly important when it comes to the quality of survival and economic burden to the health care systems.The demographic development at least in high-and middle-income countries will enforce the treating physicians to even more critically envisage this discussion in the future as increasing age is associated with a higher incidence, higher state of nursing dependency and also a higher risk of mortality in the long run [3,10].
In order to address some of these questions and provide data that may help in decision-making, we thought to analyze our patient population aiming at: i) presenting our experience in surgical treatment of elderly SICH patients, ii) determining clinical and neuroradiological risk factors for short-term and long-term mortality and iii) discussing the necessity of surgical treatment for this cohort.

Methods
Fig. 1 represents a flow chart overview about the selection of the study population.272 patients, who underwent a neurosurgical treatment for ICH between 2008 and 2020 at the neurosurgical department of the Medical University of Vienna, were retrospectively evaluated.152 patients were excluded due ICH caused by a neurovascular pathology (e. g., subarachnoid hemorrhage (SAH) due to ruptured aneurysm, arteriovenous malformation), trauma, tumor bleeding, or a constitutional pathology with a consecutive higher risk of bleeding (e.g., leukemia, full anticoagulation after heart surgery).From the remaining 120 patients, 50 patients were excluded due to age at onset under 60 years.This is in line with the definition of elderly age within the STICH II trial.
The following variables were assessed: patient ´s demographics, date and type of neurosurgical treatment, duration of EVD placement, amount of daily drained cerebrospinal fluid (CSF), length of stay (ICU and general ward), date of death, cause of death, and various comorbidities (e.g., arterial hypertension, atrial fibrillation, diabetes (DM) and chronic obstructive pulmonary disease (COPD).A GCS (Glasgow coma scale) was assessed at admission and a cranial computed tomography (CCT) was performed for all patients.Neurosurgical treatment was determined by the treating neurosurgeon based on clinical and neuroimaging parameters and contained EVD insertion alone, EVD insertion combined with hematoma evacuation, hematoma evacuation without EVD insertion, and lumbar drain insertion.
In case of intraventricular involvement of the hemorrhage, suspected intracranial pressure (ICP) elevation or suspicion of hydrocephalus a uni-or bilateral external ventricular drain (EVD) insertion was performed by standard means through a burr hole at the Kocher´s point as an entry gateway.
Hematoma evacuation was performed in supratentorial or infratentorial cerebellar ICH with significant space-occupying mass effect justifying clot removal.Standard hematoma evacuation was accomplished via an osteoplastic craniotomy.In case of life-threatening brain herniation due to massive space-occupying bleeding a decompressive hemicraniectomy with blood removal was performed.In case of suspicion of a vascular malformation a diagnostic cerebral angiography was performed.Further neurosurgical techniques like endoscopic hematoma evacuation or intraventricular thrombolysis were not utilized.For radiological evaluation, the last CCT scan before neurosurgical intervention was reviewed, including parameters such as anatomical location of hemorrhage (lobar, basal ganglia/thalamus, infratentorial, purely intraventricular), distribution of hemorrhage in ventricles with the intraventricular hemorrhage (IVH) score, and volumetry of parenchymal hemorrhage, ICH, IVH and intraventricular CSF by using the Brainlab smartbrush software (Brainlab AG, Munich, Germany) (Fig. 2) [10][11][12].Latest follow up was assessed with the modified Rankin Scale (mRS) until June 2022.The information on date and cause of death was derived from patient ´s medical record at the Medical University of Vienna and by the Austrian register of deaths.The estimation of costs per patient during hospital stay was calculated by the amount of days of hospitalization (ICU and normal ward) multiplied with the average cost for a day (1120 euro) in hospital in Austria in 2020.[13] Statistical analysis was performed using DATAtab online statistic calculator (DATAtab Team (2023), DATAtab e.U.Graz, Austria, https://datatab.de)and SPSS Statistics v. 29 (IBM, Armonk, NY, USA).The categorial variables were presented as counts and percentage, while metric parameters were displayed as median and range.Fisher´s exact test and chi-square test were performed for categorial variables and Mann-Whitney-U test was used for continuous variables.Univariate and multivariate logistic regressions were performed.Statistical significance was set at p < 0.05 for all performed tests.
The study protocol was approved by the local Ethics Committee of the Medical University of Vienna (EK 1055/2022, 15.02.2022).No formal consent was required for this type of study.

Results
A total of 120 patients with SICH underwent neurosurgical treatment between 2008 and 2020 at our institution.Out of this cohort, 70 patients were 60 years or older at the time of onset, thus fulfilling the inclusion criteria.At the time of onset, 38/70 patients were males (54.29%) with a median age of 67 years (range: 60-88 years).The median length of stay (LOS) on general ward was 20 days (range: 0-218 days) and on the ICU was 59 days (range: 0-177 days).The median length of hospital stay was 73.5 days (range: 4-243 days) with an estimated cost per patient of 58, 766 euros (range: 4, 480 -272, 160 euros).The median mRS at last follow-up was 5.
Within the overall mortality group, EVD insertion alone was the most common (18/32, 56.25%, p = 0.038).All other procedures did not differ between groups.When comparing the location of hemorrhage, a statistically significant difference could be observed for lobar hemorrhage and basal ganglia/thalamus hemorrhage (p = 0.021 and p = 0.020, respectively) in the overall mortality group, whereas the infratentorial bleeding showed no significant difference (p = 0.933).Except for basal ganglia/thalamus hemorrhage in the 12-months mortality group (p = 0.020), there was no difference in hematoma location in all other mortality groups.
In our study population, 47/70 (67.14%) patients suffered from parenchymal bleeding combined with IVH, leading to a significant higher overall mortality (p = 0.021).However, no significant differences could be observed when comparing the 30-days, 12-months, and 24months mortality.
In more detail, the higher the intraventricular blood distribution in the left ventricle the higher the risk for overall mortality (p = 0.045), which interestingly was not significant for the right ventricle (p = 0.059) (Table 3).Furthermore, higher IVH blood distribution in left ventricle is more often seen in 12-months and 24-months mortality, but not in 30days mortality (p = 0.028, p = 0.021 and p = 0.381).This statistical significance did not occur for IVH distribution in the right ventricle for all mortality groups.In addition, the larger the blood distribution within the III. and IV.ventricle the higher the mortality risk with p-values reaching p = 0.045 and p = 0.044, respectively.
A ROC curve model for IVH for 12-months mortality showed an area under the curve (AUC) of 0.658 (95% CI = 0.511 -0.805) with a cutoff Values are expressed as numbers (%) or as median (range).value for IVH volume of 7 cm 3 (p = 0.38).Kaplan-Meier analysis for 12 months and overall period (mean follow-up time 22 months) revealed a significant decrease in survival for patients with an IVH volume > 7 cm 3 (p = 0.048 and p = 0.024, respectively).

Discussion
In this study of SICH in the elderly, a mortality rate of 11.42%, 31.43%, and 45.71% was observed after 30 days, 12 months, and overall, respectively.While these results go in line with previous studies reporting similar mortality rates, we were able to identify IVH volume > 7 cm 3 as a negative prognostic factor for 12-months survival.[5,6,[14][15][16][17] In addition, we could show that the average estimated in hospital cost per patient in Austria was 58, 766 € with a mean LOS of 73.5 days.
SICH is a devastating neurological disorder with a higher mortality and morbidity compared to other types of stroke and the escalation of treatment at presentation needs to be outweighed against the expected duration and quality of life [18].So far, studies have been unable to show an advantage of a neurosurgical intervention over conservative treatment in terms of morbidity [5][6][7]9,19].In view of the growth of an aging population and the increasing economic pressure on the health care systems, this fact needs to be specifically addressed and a better understanding of the short-and long-term mortality is critical [17].
Furthermore, details on prognostic factors need to be scrutinized.In terms of clinical factors, single variable analysis and univariate cox regression analysis in our study identified DM as a negative risk factor for short-term and long-term mortality (Table 1 and Table 4).Other comorbidities as well as treatment with anticoagulation and antiplatelet Values are expressed as numbers (%) or as median (range).therapy before ictus had no significant impact on death during follow up (Table 4).Interestingly, Øie et al. and Fric-Shamji et al. identified oral anticoagulation as a risk factor for severe disability and mortality [17,20].Oral anticoagulation may facilitate hematoma expansion after initial SICH, which itself is associated with higher mortality [21].In comparison to our cohort Øie et al. and Fric-Shamji et al., had a greater study population [17,20].Other reports in the literature showed conflicting results where elevated blood glucose levels did not affect mortality rate, while others including two meta-analyses by Guo et al. and Wu et al. showed a higher mortality rate for hyperglycemia in ICH [3,[22][23][24][25][26][27].A potential explanation was the downregulation of Aquaporin-4 expression and subsequent increase of vasogenic brain edema.Others have suggested that hyperglycemia contributes to secondary neuronal injuries by exacerbating the cerebral inflammatory milieu and oxidative stress to the brain.[24,25,27] In terms of radiological factors location and extent of hemorrhage are obvious factors of interest.Here, Hemphill et al. introduced the ICH Score, a grading scale for ICH that allows risk stratification on the time of onset [28].Besides GCS, age, and infratentorial location the ICH volume over 30 cm 3 was determined as a variable associated with 30-day mortality.However, Hemphill et al. applied the straightforward ABC/2 method for volumetry.This method measures the volume of a 3-dimensional object (in our case the hemorrhage) where A is the greatest diameter, B the diameter orthogonal to A, and C the approximate number of CT slices multiplied by the thickness.[29,30] While this method and its clinical application seems to be easy, the accuracy decreases the higher the inhomogeneity in the measured objects.This is true for the complex 3-dimensional-patterns of intracerebral bleedings leading to more unspecific and inaccurate results.Therefore, we used the Brainlab smartbrush software for volumetry, which allowed us not only to measure accurately the 3-dimensional pattern but also to divide the ICH into different compartments and to independently asses their volume (Fig. 1).This method was previously used in another study from our institution, but except for some volumetric studies on other pathologies, no studies were found in literature utilizing the Brainlab smartbrush software to assess ICH volumes [11,12].Interestingly, our results showed that the volume of parenchymal hematoma had no impact on the mortality rate, whereas the IVH volume was a significant negative prognostic factor (Table 2 and Table 4).This result seems to be contradictory to literature, where a larger hematoma volume is described as an independent risk factor for high morbidity and mortality [17,31].A potential explanation maybe that the volume of a hematoma is not linearly correlated with its space occupying effect across ages with more atrophic brains in the elderly [22].Most studies report a hematoma volume > 30 ml/cm 3 as a risk factor for mortality, without distinguishing between IVH and parenchymal volume.[3,22,28] Presence of IVH itself is associated with higher mortality (Table 2) but most of the studies in literature did not asses the IVH volume [17,27,32].Young et al. determined a volume of > 20 cm 3 as a risk factor for poor outcome and Al-Khaled et al. showed intraventricular extension of hemorrhage to be associated with higher in -hospital mortality [33,34].In our ROC analysis a cut-off value of => 7 cm 3 IVH was identified as a negative prognostic factor for 12-months survival.Our results thus may suggest, that a more conservative management strategy in elderly patients with IVH and SICH may be more meaningful.
All of these clinical, radiological and outcome considerations need to be taken into account when assessing the economic impact and cost effectiveness, which will be even more critical in future.In our analysis, the average hospitalization costs per patient amounted to 58, 766 € with a range from 4, 480 € to 272, 160 €.Considering that we calculated the costs in a very conservative estimation (average cost per day of 1120 euro multiplied with days of hospitalization), the real costs are certainly much higher, due to the fact, that no difference in calculation was made between normal ward and ICU [13].A recent cost utility analysis comparing minimal invasive surgery to conservative treatment from Vardanyan et al. showed, that surgery was not cost-effective compared to optimal medical treatment [35].Similarly, Specogna et al. showed also higher costs for surgical treatment compared to medical treatment [36].
Despite multiple randomized trials, optimal treatment for SICH remains still unclear with no clear evidence for surgical intervention.Chen et al. investigated predictors of surgical intervention in patients with SICH identifying several independent predictors including younger age [37].However, age itself is reported as risk factor for higher mortality [16].In light of the fact that the economic aspects for treating SICH will increasingly influence decision-making in future, our data would favor a more conservative treatment strategy in SICH of the elderly with the identified negative prognostic factors.

Limitations
Besides the limitation of our study rendered from the inherent nature of its retrospective design, the main limitation is derived from the small sample size and the lack of a cohort of patients managed without any surgical intervention.Despite these limitations, our study adds some preliminary evidence, that could help clinicians in decision-making.In view of the demographic trend and its challenging development, further studies are needed to elaborate new therapeutic concepts and identify new or additional variables in management of SICH in elderly patients.

Conclusion
The overall mortality of SICH in elderly patients is high despite extensive neurosurgical treatment.IVH volume over 7 cm 3 at presentation was identified as a negative prognostic factor and should be taken into account when treatment decisions need to be taken.
in the design of the study, wrote application for ethical approval, collected the data and helped to draft the manuscript.AC performed statistical analysis and helped to draft the manuscript.CD helped drafting the manuscript, revised it, supervised the research, and approved the final manuscript.CD, JH, AR and KR supervised the research.All authors read an approved the final manuscript.
We declare that this manuscript is original, has not been published before and is not currently being considered for publication elsewhere.
We declare that all procedures were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki Declaration and its later amendments or comparable ethical standards.The study protocol was approved by the local Ethics Committee of the Medical University of Vienna (EK 1055/2022, 15.02.2022).No formal consent was required for this type of study.

Fig. 2 .
Fig. 2. Example of using Brainlab smartbrush software for volumetry measurement in a CCT scan.Volumetric analysis shows a SICH with parenchymal component, IVH as well as normal CSF.First row: Volumetry of parenchymal bleeding highlighted in yellow.Second row: The volume of IVH is highlighted in purple.Third row: Volumetry of normal CSF highlighted in green.CCT cerebral computed tomography, CSF cerebrospinal fluid, ICH intracerebral hemorrhage, IVH intraventricular hemorrhage, SICH spontaneous intracerebral hemorrhage.
Trial in Intracerebral Haemorrhage.

Table 1
Mortality rate during follow-up in relation to sex, age at onset and comorbidities.

Table 2
Mortality rate during follow-up in relation to operation and location of hemorrhage and volumetric parameters in patients and its relation to different periods of mortality rate.

Table 3
IVH score system byHijdra et al.showing the blood distribution between ventricles in relation to different periods of mortality rate.

Table 4
Univariate and multivariate cox regression analysis of predictors for 12-months mortality after SICH in surgically treated elderly patients.