Oedema extension distance in intracerebral haemorrhage: Association with baseline characteristics and long-term outcome

Abstract

Introduction

Oedema extension distance is a derived parameter that may reduce sample size requirements to demonstrate reduction in perihaematomal oedema in early phase acute intracerebral haemorrhage trials. We aimed to identify baseline predictors of oedema extension distance and its association with clinical outcomes.

Patients and methods

Using Virtual International Stroke Trials Archive-Intracerebral Haemorrhage, first Intensive Blood Pressure Reduction in Acute Cerebral Hemorrhage Trial, and Minimally Invasive Surgery and rtPA for Intracerebral Hemorrhage Evacuation II datasets, we calculated oedema extension distance at baseline and at 72 h measured using computed tomography. Using linear regression, we tested for associations between baseline characteristics and oedema extension distance at 72 h. Ordinal regression (underlying assumptions validated) was used to test for associations between oedema extension distance at baseline and 72 h and oedema extension distance change between baseline and 72 h, and modified Rankin scale scores at 90 days, adjusted for baseline and 72 h prognostic factors.

Results

There were 1028 intracerebral haemorrhage cases with outcome data for analyses. Mean (standard deviation, SD) oedema extension distance at 72 h was 0.54 (0.26) cm, and mean oedema extension distance difference from baseline (EED_72–0) was 0.24 (0.18) cm. Oedema extension distance at 72 h was greater with increasing baseline haematoma volume and baseline oedema extension distance. Increasing age, lobar haemorrhage, and intraventricular haemorrhage were independently associated with EED_72–0. In multifactorial ordinal regression analysis, EED_72–0 was associated with worse modified Rankin scale scores at 90 days (odds ratio 1.96, 95% confidence interval 1.00–3.82).

Discussion

Increase in oedema extension distance over 72 h is independently associated with decreasing functional outcome at 90 days. Oedema extension distance may be a useful surrogate outcome measure in early phase trials of anti-oedema or anti-inflammatory treatments in intracerebral haemorrhage.

Keywords

Intracerebral haemorrhage oedema inflammation surrogate markers

Introduction

Perihaematomal oedema (PHE) develops rapidly and increases over several days following acute spontaneous intracerebral haemorrhage (ICH) and is proposed to reach maximal volume by two weeks.¹ The additional mass effect of PHE contributes to early neurological deterioration and poor outcome.^2,3 In addition, PHE may be a marker of secondary injury and a potential therapeutic target in ICH.⁴ A key-mediator of PHE is the innate immune response within the brain, characterised by the activation of resident microglia by damage-associated molecular patterns, infiltration of peripheral immune cells and the production of inflammatory mediators.⁵ These inflammatory mechanisms orchestrate tissue damage and blood–brain barrier breakdown, playing a key role in the development of PHE.⁶

PHE has been widely used as the main outcome measure in pre-clinical ICH studies targeting secondary injury and can be efficiently and reliably measured in both the experimental and clinical settings.^7–9 We have recently described a novel parameter, the oedema extension distance (EED), which has been employed by other groups.^10,11 It is relatively less dependent on haematoma volume and may reduce the sample size required in proof-of-concept trials by around 75% when compared to absolute or relative PHE volume.¹² Understanding the baseline determinants of EED and its association with clinical outcomes is required to establish the utility of EED as a surrogate outcome measure in ICH clinical trials.

In this study, we aimed to evaluate the EED in a large sample of ICH patients (taken from the Virtual International Stroke Trials Archive (VISTA),¹³ the first Intensive Blood Pressure Reduction in Acute Cerebral Hemorrhage Trial (INTERACT1),¹⁴ and the Minimally Invasive Surgery and rtPA for Intracerebral Hemorrhage Evacuation II (MISTIE II) trial¹⁵), to test for associations between EED, baseline clinical characteristics, and clinical outcome.

Patients and methods

We conducted a retrospective analysis of prospectively collected data from three sources: VISTA,¹³ INTERACT1,¹⁴ and MISTIE-II.¹⁵ Patients aged ≥18 years old with a supratentorial ICH were eligible for inclusion in our study. Those with high premorbid disability were excluded from the trials from which the datasets are derived. Ethical approval was not required for the re-use of these existing, anonymised clinical trial datasets.

Variables included baseline clinical characteristics (age, sex, hypertension, smoking, diabetes mellitus, hypercholesterolaemia, atrial fibrillation, past history of stroke or transient ischaemic attack, smoking status, baseline blood pressure, and neurological impairment based on scores on the National Institutes of Health Stroke Scale), medication at baseline (antiplatelet agent(s), anticoagulants, and lipid lowering/statin use), imaging parameters (baseline and 72 h haematoma and oedema volumes, haematoma location, presence of intraventricular haemorrhage (IVH)), and outcome data at 90 days.

All datasets used equivalent techniques to calculate PHE. As described in Yang et al.,¹⁶ PHE volumes were calculated independently by two trained neurologists, blind to clinical data, treatment, and date and sequence of scan, using computer-assisted multi-slice planimetric and voxel threshold techniques. A semi-automated threshold-based approach (range 5–33 Hounsfield Units) was applied with adjustment to identify regions of PHE to estimate volumes (cm³) from slice thickness separate to boundaries of blood. Inter-reader reliability was tested with re-analysis after 30 and 60% of the scans were read by both readers to assess for drift.¹⁶

The EED was calculated from the haematoma and PHE volumes, as outlined elsewhere¹² (Figure 1) at baseline (EED₀) and 72 h (EED₇₂), and these were used to calculate change from baseline to 72 h (EED_72–0). The relationship between PHE volume and EED is described by the equation below

PHE volume = \frac{4}{3} π (3 r_{h}^{2} r_{eed} + 3 r_{h} r_{eed}^{2} + r_{eed}^{3})

where r_h is the radius of the haematoma and r_eed is the EED.

Figure 1.

Example of a CT scan demonstrating delineation of the region of PHE (outlined in green) and ICH (outlined in red). The EED is the difference between the radius (r_e) of a sphere (shown in green) equal to the combined volume of PHE and ICH and the radius of a sphere (shown in red) equal to the volume of the ICH alone (r_h).¹² EED: oedema extension distance; ICH: intracerebral haemorrhage; PHE: perihaematomal oedema.

Statistical methodology

Summary statistics were used to inform sample size calculations for trials with EED as an endpoint. Pearson correlation coefficients were used to test for correlations between PHE and EED with haematoma volumes at baseline and 72 h with visual representation using box plots; a comparison of correlation coefficients drawn from the same sample was performed. Multifactorial linear regression was used to test for associations between EED_72–0 and baseline clinical and imaging characteristics. Ordinal regression models were used to determine independent associations of EED and outcome (90-day modified Rankin scale (mRS) scores), with EED₀, EED₇₂, and EED_72–0 considered in separate models. Backward elimination was used to select the final model for linear and ordinal regression, with variables being removed manually after inspection of model fit. A Cox regression model was also used to examine mortality over 90 days. Data are reported with odds ratios (ORs) or hazard ratios (HRs), as appropriate, with 95% confidence intervals (CIs). A sensitivity analysis excluding those patients with acute neurosurgical intervention was performed. All analyses were undertaken using Stata Statistical Software: Release 14 (StataCorp 2015).

Results

Of a total of 1373 ICH patients in the pooled dataset (286 INTERACT1, 987 VISTA, and 100 MISTIE-II), 50 were excluded due to infratentorial ICH, 71 as their 72 h follow-up scan was performed <48 h or >96 h from ictus, and 224 due to incomplete data (see Supplemental Figure 1 for details). Thus, 1028 ICH patients comprised the final study population. Table 1 shows that patients in the MISTIE-II dataset had larger haematoma and oedema volumes, a higher proportion of vascular risk factors (hypertension and hypercholesterolaemia), and a higher proportion of lobar ICH and IVH than those in the other datasets.

Table 1.

Population baseline characteristics.

	Total N = 1028	Cohort			p-value
	Total N = 1028	VISTA N = 725	INTERACT1 N = 241	MISTIE-II N = 62	p-value
Mean age (SD)	64.7 (12.1)	65.6 (12.1)	62.7 (12.3)	62.5 (10.9)	0.002
Male (%)	656 (63.8)	466 (64.3)	151 (62.7)	39 (62.9)	0.89
Hypertension (%)	843 (82.0)	610 (84.1)	178 (73.9)	55 (88.7)	0.001
Diabetes mellitus (%)	170 (16.5)	132 (18.2)	22 (9.1)	16 (25.8)	<0.001
Previous anticoagulation (%)	10 (1.0)	–	3 (1.2)	7 (11.3)	<0.001
Mean baseline SBP mmHg (SD)	173.0 (37.0)	173.0 (39.0)	178.5 (30.0)	143.5 (30.0)^a	<0.001
Median baseline NIHSS (IQR)	13.0 (9.0)	13.0 (8.0)	10.0 (10.0)	19.5 (7.0)^a	<0.001
Median baseline haematoma volume cm³ (IQR)	13.7 (19.8)	14.6 (19.1)	9.5 (12.4)	39.0 (24.0)	<0.001
Median 72 h haematoma volume cm³ (IQR)	15.0 (22.9)	17.1 (25.8)	9.9 (12.6)	19.6 (30.3)	<0.001
Median baseline oedema volume cm³ (IQR)	9.2 (12.9)	8.9 (12.3)	7.5 (9.1)	27.1 (20.8)	<0.001
Median 72 h oedema volume cm³ (IQR)	21.8 (27.8)	23.5 (30.9)	15.3 (21.4)	30.0 (16.4)	<0.001
Presence of IVH at baseline (%)	320 (31.1)	225 (31.0)	60 (24.9)	35 (56.5)	0.11
Index haematoma location (%)
Supratentorial lobar	159 (15.5)	113 (15.6)	22 (9.1)	24 (38.7)	<0.001
Supratenorial deep	869 (84.5)	612 (84.4)	219 (90.9)	38 (61.3)	<0.001

INTERACT1: first Intensive Blood Pressure Reduction in Acute Cerebral Hemorrhage Trial; IVH: intraventricular haemorrhage; MISTIE II: Minimally Invasive Surgery and rtPA for Intracerebral Hemorrhage Evacuation II; NIHSS: National Institute of Health Stroke Score; SBP: systolic blood pressure; SD: standard deviation; TIA: transient ischaemic attack; VISTA: Virtual International Stroke Trials Archive.

^aRecordings at randomisation.

Overall, mean (SD) EED₀ and EED₇₂ were 0.30 cm (0.14) and 0.54 cm (0.26), respectively, and mean EED_72–0 was 0.24 cm (0.18). The relationship between EED₀ and EED₇₂ is demonstrated by scatterplot in Supplemental Figure 2. Table 2 indicates the sample sizes required to detect reductions in the various PHE volume parameters. For example, to detect a 10% effect size with 90% power, 3179, 2673, and 1182 ICH patients are needed for the different endpoints of difference from baseline to 72 h in PHE volume, relative PHE volume, and EED, respectively. At median values for haematoma volume and EED, a given percentage reduction in EED is expected to give rise to a similar reduction in PHE volume using the model described in Figure 1.

Table 2.

Clinical trial sample size calculations using PHE as the primary outcome.

	80% power			90% power
Reduction in measure (%)	PHE Vol (72–0)	Relative PHE (72–0)	EED_72–0	PHE Vol (72–0)	Relative PHE (72–0)	EED_72–0
5	9485	7988	3532	12,698	10,694	4728
10	2374	1997	883	3179	2673	1182
15	1056	888	392	1413	1188	525
20	594	499	221	795	668	296
25	380	320	141	508	428	189
30	264	222	98	353	297	131
35	194	163	72	260	218	96
40	148	125	55	199	167	74

EED: oedema extension distance; PHE: perihaematomal oedema.

Number of patients required in each arm of a clinical trial with difference in PHE between baseline and 72 h as the primary outcome, assuming α = 0.05 and either 80 or 90% power to detect a range of reductions in each measure. Calculations based on data from the conservative arms of VISTA, INTERACT1, and MISTIE-II. Mean (SD) for each measure was PHE volume (72–0), 15.27 ml (18.78 ml); relative PHE (72–0), 0.86 ml (0.97 ml); EED_72–0, 0.24 cm (0.18 cm).

Figure 2 demonstrates a positive correlation between (a) baseline (r = 0.78; p < 0.001) and (b) 72 h PHE and haematoma volumes (r = 0.80; p < 0.001), confirming that PHE volume is strongly influenced by haematoma size. In contrast, EED is much less strongly correlated with haematoma volume at baseline (r = 0.35; p < 0.001) and 72 h (r = 0.30; p < 0.001). A comparison of the respective correlation coefficients demonstrated significant difference at both baseline and 72 h (p < 0.001). Table 3 shows results of a multivariable linear regression model used to identify factors associated with EED_72–0. Larger baseline haematoma volume was associated with higher EED_72–0 (p < 0.001), whereas increasing age (p < 0.001), lobar location (versus deep) (p = 0.018), and presence of IVH (versus not) (p = 0.003) were associated with lower EED_72–0. Table 4 shows the results of ordinal regression analysis; larger EED_72–0 (OR 1.96, 95% CI 1.00–3.82), presence of IVH, diabetes mellitus, prior use of antiplatelets and anticoagulation, and 72 h haematoma volume were independently associated with a poor 90-day outcome (mRS score 3–6). Lobar haemorrhage location and male sex were predictive of a good 90-day outcome (mRS score 0–2). At 90 days 128 (12.5%) patients had died. EED_72–0 was not associated with death by 90 days (HR 2.21, 95% CI 0.90–5.47) (Supplementary Table 1).

Figure 2.

Box plots showing the relationship between oedema volume and haematoma volume, at baseline (a: r = 0.78, p < 0.001) and 72 h (b: r = 0.80, p < 0.001) and EED and haematoma volume, at baseline (c: r = 0.35, p < 0.001) and 72 h (d: r = 0.30, p < 0.001).

Table 3.

Table describing baseline factors associated with (square root) EED_72–0.

Predictor variable	Coefficient	P value	Lower 95% CI	Upper 95% CI
Age	−0.001	<0.001	−0.001	<−0.001
Lobar haemorrhage (versus deep haemorrhage)	−0.018	0.018	−0.033	−0.003
Baseline haematoma volume	0.001	<0.001	<0.001	0.001
Presence of baseline IVH (versus not)	−0.016	0.003	−0.027	−0.006

CI: confidence interval; EED: oedema extension distance; IVH: intraventricular haemorrhage.

Table 4.

Results of testing for associations with mRS at 90 days using an ordinal regression model.

Predictor variable	Odds ratio	Standard error	P value	Lower 95% CI	Higher 95% CI
Male sex	0.78	0.09	0.035	0.61	0.98
Age	1.05	0.006	<0.001	1.04	1.06
Diabetes	1.65	0.27	0.002	1.20	2.29
Antiplatelet therapy	1.70	0.30	0.003	1.20	2.41
Anticoagulation	2.74	0.90	0.002	1.44	5.20
Statin	0.74	0.15	0.124	0.50	1.09
Lobar haemorrhage	0.59	0.11	0.003	0.41	0.84
Haematoma volume 72 h	1.04	0.004	<0.01	1.04	1.06
Presence of IVH at 72 h	2.68	0.34	<0.001	2.09	3.45
EED_72–0	1.96	0.67	0.049	1.00	3.82

CI: confidence interval; EED: oedema extension distance; IVH: intraventricular haemorrhage; mRS: modified Rankin scale.

Of the 1028 patients, 82 (8.0%) underwent acute neurosurgical intervention (41 VISTA and 41 MISIE-II). Of the 41 VISTA patients, 26 (63.4%) underwent ventricular drainage, 2 (4.9%) underwent haematoma evacuation, 5 (12.2%) had craniotomies, and in 8 (19.5%), the procedure details were unavailable. The MISTIE-II patients who had undergone neurosurgical intervention were in the minimally invasive surgery treatment arm of the trial. A multivariable ordinal regression sensitivity analysis was performed excluding patients that underwent acute neurosurgery and larger EED_72–0 was associated with 90-day outcome (OR 1.95, 95% CI 0.97–3.92) with a similar OR point estimate, but it was no longer statistically significant (p = 0.06) (Supplementary Table 2).

Discussion

Our analysis of a large dataset of over 1000 ICH patients has shown that EED_72–0, as a derived parameter, can markedly reduce the sample size required in early phase ICH clinical trials targeting PHE, as compared with conventional measures of absolute or relative PHE volumes. We show that higher EED_72–0 is associated with higher baseline haematoma volume, while a lower EED_72–0 was associated with advancing age, lobar location, and IVH. Furthermore, EED_72–0 was independently associated with a worse 90-day outcome, as measured by mRS.

As haematoma and PHE volumes are closely correlated, PHE is highly variable and necessitates large sample sizes for clinical trials with oedema as an outcome measure. Relative PHE volume has been considered as a means of reducing the variability introduced by the correlation between haematoma and oedema volumes, but is typically disproportionately large for smaller haematoma volumes, and thus an unsuitable parameter for clinical trials.⁸ PHE is strongly influenced by the intensity of the parenchymal inflammatory response, which diffuses in a linear fashion from the haematoma border. Conversely, EED is relatively less dependent on haematoma volume and likely to be more representative of the pathophysiological processes related to PHE than total PHE volume.¹⁷

In performing a sensitivity analysis by excluding patients with acute neurosurgical intervention, the p-value for the association between EED_72–0 and 90-day mRS increased from 0.049 to 0.06. The majority of neurosurgical patients in VISTA had ventricular drainage only (26/41; 63.4%), which would not be expected to significantly alter the PHE. Patients in MISTIE were randomised to either undergo tPA-augmented minimally invasive surgery to evacuate the haematoma or standard medical care. Our analysis demonstrates a clear correlation between haematoma volume and PHE volume, suggesting that a surgical reduction in haematoma volume will also reduce PHE volume, which was confirmed in MISTIE-II.⁹ However, given that EED is relatively less dependent on haematoma volume (unlike PHE volume) we would not expect EED to be reduced in patients undergoing haematoma evacuation, relative to non-surgical patients. Further analysis of our dataset confirms this, with no difference in EED_72–0 between surgical and non-surgical patients, either in the whole combined dataset (surgical mean EED_72–0 0.24 cm (SD 0.20 cm) versus non-surgical mean EED_72–0 0.24 cm (0.18 cm); T-test p = 0.86) or within the MISTIE II dataset alone (surgical 0.16 cm (0.16 cm) versus non-surgical 0.14 cm (0.10 cm); p = 0.44). The slight increase in the p-value from 0.049 to 0.06 on excluding surgical cases is thus likely to be related to reduced power (a type II statistical error), rather than heterogeneity in the association between EED_72–0 and 90-day mRS by whether or not neurosurgery was performed.

We have identified larger haematoma volume as predictors of EED_72–0. Older age, lobar (versus deep) haemorrhage location, and IVH were predictors of smaller EED_72–0. Older age is an independent predictor of poor outcome after ICH and is used in clinical grading scales.¹⁸ However, this might be due to any number of pathophysiological mechanisms, such as impaired coagulation cascade, inflammatory or astrocyte responses, cerebral atrophy, or reduced functional reserve. A study of experimentally induced ICH found worse neurological outcome and larger PHE volumes in 18-month-old rats compared to 3-month-old rats.¹⁹ However, a retrospective analysis of 219 consecutive ICH patients assessed with sequential CT failed to find any association of PHE volume and age.²⁰ Older age emerged as a predictor of a smaller EED₇₂ in our study involving a much larger ICH cohort, which may suggest that older people have impaired, or delayed, inflammatory response in ICH.²¹

We found that lobar location of ICH was predictive of smaller EED_72–0. The existing literature is conflicted in regards to haematoma location and PHE volume,^22–24 which may be due to confounding and chance related to study design and sample size, as well as the inclusion of larger haematomas which may not be easily classified as deep or lobar. Moreover, no previous study has used EED as a measure of PHE which is relatively less dependent on haematoma size and likely to provide a more reliable assessment of the association between ICH location and oedema by adjusting for haematoma volume. The capacity for water to diffuse through brain tissue can be driven by any change in the microstructure which alters diffusivity such as mechanical compression, membrane damage, inflammation, and shifts in water content in either the intra- or extra-cellular space. A potential biological explanation for higher EED in deep ICH is that deep haematomas are typically adjacent to densely packed white matter tracts. Myelin is a major diffusion barrier to water, a property that is exploited in diffusion tractography. This may preferentially facilitate the movement of water along the direction of white matter tracts, allowing greater propagation of oedema than might be seen with a lobar haemorrhage surrounded by grey matter. Further studies using diffusion tensor MRI may help to test this hypothesis further. Finally, although EED_72–0 was not statistically associated with mortality, it had a large HR in the model and requires confirmation in a larger dataset.

The presence of baseline IVH was also associated with a lower EED_72–0. Although there is no clear explanation for this association, it may be that IVH leads to higher intraventricular pressure which may be transmitted to the brain parenchyma, reducing the production of PHE. The presence of blood adjacent to the ventricular ependyma may alter the biological processes in the underlying brain parenchyma such that less oedema is generated over the first 72 h. For example, the inflammatory response may progress more quickly in the intraventricular space and ‘anti-inflammatory’ cytokines (e.g. interleukin-1 receptor antagonist, transforming growth factor beta) associated with repair and recovery may be generated earlier, influencing the brain parenchyma adjacent to the ventricle. Further work in experimental models is needed to test these hypotheses.

Although our study was strengthened by having a large, ethnically diverse (60% Caucasian, 34% Asian) sample with robust measures of PHE, we acknowledge several limitations that include the development of the EED parameter, which required the assumption of an ellipsoid-shaped haematoma and oedema, which is present in only 70% of patients.⁴ Peak PHE volume may have an independent effect on outcome in ICH,²⁵ but we were unable to assess this in our cohort. Furthermore, our sample size was still small for examination of modest, but still clinically important, associations relevant to the serious disease of ICH. Finally, our cohort was younger and with a greater proportion of lobar ICH compared to population based studies,²⁶ this may limit our study’s representativeness.

Conclusion

We have previously shown that the use of EED as the outcome measure for PHE in clinical trials markedly reduces the sample size required. We have now confirmed this in a much larger dataset and demonstrated that EED_72–0 is significantly associated with mRS at 90 days. Although validation in prospective studies is desired, our study strengthens the case for the use of EED_72–0 as a surrogate outcome measure for early phase clinical trials of anti-oedema treatments in ICH.

Supplemental Material

Supplemental material for Oedema extension distance in intracerebral haemorrhage: Association with baseline characteristics and long-term outcome

Supplemental Material for Oedema extension distance in intracerebral haemorrhage: Association with baseline characteristics and long-term outcome by Robert Hurford, Andy Vail, Calvin Heal, Wendy C Ziai, Jesse Dawson, Santosh B Murthy, Xia Wang, Craig S Anderson, Daniel F Hanley, Adrian R Parry-Jones and on behalf of the VISTA-ICH Collaborators: the STRONG STAR Consortium in European Stroke Journal

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: RH is supported by an Association of British Neurologists’ Clinical Research Training fellowship. ARP-J is supported by a National Institute for Health Research Clinician Scientist Award.

Informed consent

Not applicable.

Ethical approval

Not applicable.

Guarantor

ARP-J.

Contributorship

Study concept and design: RH, AV, CH, and ARP-J. Acquisition of data: RH, ARP-J, WCZ, JD, SBM, XW, CSA, and DFH. Analysis and interpretation of data: RH, AV, CH, and ARP-J. Drafting of the manuscript: RH and ARP-J. Critical revision of the manuscript for important intellectual content: CH, WCZ, JD, SBM, XW, CSA, and DFH. Statistical analysis: RH, AV, and CH. Study supervision: ARP-J.

Acknowledgements

VISTA-ICH Steering Committee: Daniel F. Hanley (Chair), Kenneth S. Butcher, Stephen Davis, Barbara Gregson, Kennedy R. Lees, Patrick Lyden, Stephan Mayer, Keith Muir, and Thorsten Steiner.

References

Balami

Buchan

AM.

Complications of intracerebral haemorrhage. Lancet Neurol 2012; 11: 101–118.

Zazulia

Diringer

Derdeyn

, et al. Progression of mass effect after intracerebral hemorrhage. Stroke 1999; 30: 1167–1173.

Appelboom

Bruce

Hickman

, et al. Volume-dependent effect of perihaematomal oedema on outcome for spontaneous intracerebral haemorrhages. J Neurol Neurosurg Psychiatry 2013; 84: 488–493.

Wang

Dore

Inflammation after intracerebral hemorrhage. J Cereb Blood Flow Metab 2007; 27: 894–908.

Keep

Hua

Intracerebral haemorrhage: mechanisms of injury and therapeutic targets. Lancet Neurol 2012; 11: 720–731.

Keep

Xiang

Ennis

, et al. Blood-brain barrier function in intracerebral hemorrhage. Acta Neurochir Suppl 2008; 105: 73–77.

Frantzias

Sena

Macleod

, et al. Treatment of intracerebral hemorrhage in animal models: meta-analysis. Ann Neurol 2011; 69: 389–399.

Volbers

Staykov

Wagner

, et al. Semi-automatic volumetric assessment of perihemorrhagic edema with computed tomography. Eur J Neurol 2011; 18: 1323–1328.

Mould

Carhuapoma

Muschelli

, et al. Minimally invasive surgery plus recombinant tissue-type plasminogen activator for intracerebral hemorrhage evacuation decreases perihematomal edema. Stroke 2013; 44: 627–634.

10.

Putaala

Sharma

, et al. Persistent hyperglycemia is associated with increased mortality after intracerebral hemorrhage. J Am Heart Assoc 2017; 6: e005760.

11.

Sharma

Strbian

, et al. Natural history of perihematomal edema and impact on outcome after intracerebral hemorrhage. Stroke 2017; 48: 873–879.

12.

Parry-Jones

Wang

Sato

, et al. Edema extension distance: outcome measure for phase II clinical trials targeting edema after intracerebral hemorrhage. Stroke 2015; 46: e137–e140.

13.

Ali

Bath

Brady

, et al. Development, expansion, and use of a stroke clinical trials resource for novel exploratory analyses. Int J Stroke 2012; 7: 133–138.

14.

Anderson

Huang

Wang

, et al. Intensive blood pressure reduction in acute cerebral haemorrhage trial (INTERACT): a randomised pilot trial. Lancet Neurol 2008; 7: 391–399.

15.

Hanley

Thompson

Muschelli

, et al. Safety and efficacy of minimally invasive surgery plus alteplase in intracerebral haemorrhage evacuation (MISTIE): a randomised, controlled, open-label, phase 2 trial. Lancet Neurol 2016; 15: 1228–1237.

16.

Yang

Arima

, et al. Prognostic significance of perihematomal edema in acute intracerebral hemorrhage: pooled analysis from the Intensive Blood Pressure Reduction in Acute Cerebral Hemorrhage Trial studies. Stroke 2015; 46: 1009–1013.

17.

Parry-Jones

Wang

Sato

, et al. Edema extension distance: outcome measure for phase II clinical trials targeting edema after intracerebral hemorrhage. Stroke 2015; 46: e137–e140.

18.

Rost

Smith

Chang

, et al. Prediction of functional outcome in patients with primary intracerebral hemorrhage: the FUNC score. Stroke 2008; 39: 2304–2309.

19.

Gong

Hua

Keep

, et al. Intracerebral hemorrhage: effects of aging on brain edema and neurological deficits. Stroke 2004; 35: 2571–2575.

20.

Staykov

Wagner

Volbers

, et al. Natural course of perihemorrhagic edema after intracerebral hemorrhage. Stroke 2011; 42: 2625–2629.

21.

Sieber

Claus

Witte

, et al. Attenuated inflammatory response in aged mice brains following stroke. PLoS One 2011; 6: e26288.

22.

McCarron

Alberts

MJ.

Location characteristics of early perihaematomal oedema. J Neurol Neurosurg Psychiatry 2006; 77: 378–380.

23.

Arima

Wang

Huang

, et al. Significance of perihematomal edema in acute intracerebral hemorrhage: the INTERACT trial. Neurology 2009; 73: 1963–1968.

24.

Gebel

JMJ

Jauch

Brott

, et al. Natural history of perihematomal edema in patients with hyperacute spontaneous intracerebral hemorrhage. Stroke 2002; 33: 2631–2635.

25.

Volbers

Giede-Jeppe

Gerner

, et al. Peak perihemorrhagic edema correlates with functional outcome in intracerebral hemorrhage. Neurology 2018; 90: e1005–e1012.

26.

Samarasekera

Fonville

Lerpiniere

, et al. Influence of intracerebral hemorrhage location on incidence, characteristics, and outcome: population-based study. Stroke 2015; 46: 361–368.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.20 MB