Table of Contents Author Guidelines Submit a Manuscript
BioMed Research International
Volume 2015, Article ID 961085, 8 pages
http://dx.doi.org/10.1155/2015/961085
Research Article

New Prognostic Score for the Prediction of 30-Day Outcome in Spontaneous Supratentorial Cerebral Haemorrhage

1Department of Neurology, University of Debrecen, Clinical Center, Debrecen 4032, Hungary
2Division of Neuropathology, Institute of Pathology, University of Debrecen, Clinical Center, Debrecen 4032, Hungary
3Hygiene and Infection Control Services, Kenézy Hospital, Debrecen 4031, Hungary
4Department of Biomedical Laboratory and Imaging Science, University of Debrecen, Clinical Center, Debrecen 4032, Hungary
5Department of Neurology, Semmelweis University, Budapest 1083, Hungary

Received 17 October 2014; Revised 23 December 2014; Accepted 24 December 2014

Academic Editor: Alfredo Conti

Copyright © 2015 Rita Szepesi et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Abstract

Aims. The purpose of the present study was to evaluate predictors of outcome in primary supratentorial cerebral haemorrhage. Furthermore, we aimed to develop a prognostic model to predict 30-day fatality. Methods. We retrospectively analyzed a database of 156 patients with spontaneous supratentorial haemorrhage to explore the relationship between clinical and CT characteristics and fatal outcome within 30 days using multiple logistic regression analysis. The analyzed factors included volumetric data assessed by neuropathological and CT volumetry. A second CT scan in survivors or neuropathological ABC/2 volumetry in nonsurvivors was used along with the baseline CT to assess the growth index of haematoma. Results. Systolic blood pressure, serum potassium and glucose levels, platelet count, absolute and relative haematoma volumes, and presence and size of intraventricular haemorrhage statistically significantly predicted the fatal outcome within 30 days. Based on our results we formulated a six-factor scoring algorithm named SUSPEKT to predict outcome. Conclusions. After validation the SUSPEKT score may be applicable in general clinical practice for early patient selection to optimize individual management or for assessment of eligibility for treatment trials.

1. Introduction

Spontaneous intracerebral haemorrhage (ICH) accounts for 8% to 14% of all strokes and carries the highest mortality rate of major stroke subtypes [1]. The pharmacological and surgical interventions are less efficient than in case of ischaemic stroke [2]. Half of the mortality occurs within the first 2 days as a result of brain herniation [3]. Thirty-five to 52% of patients are dead by 1 month and only 20% of patients live independently at 6 months [4].

Predictors of mortality and functional outcome have been studied widely [57], but no treatment has shown a proven benefit so far. Accordingly, comprehensive analysis of the potential predictive factors of short term fatal outcome is of great importance. Reviewing the data of ICH patients admitted to the Neurology Department, University of Debrecen (UD), we retrospectively analyzed the relationship between clinical and CT characteristics and case fatality at 30 days.

The autopsy rate of our department is more than 90%; therefore we could incorporate also the results of brain autopsies into our analysis. This is a unique opportunity (compared with similar studies) in the era of declining brain autopsy rates [8]. Although all of our patients were immediately investigated by CT or MRI at acute admission (and repeated if required by the patient’s deteriorating condition), we also had opportunity to analyze the results of their brain autopsies, agreeing with the Agency for Healthcare Research and Quality U.S. Department of Health and Human Services: “clinical diagnoses, whether obtained from death certificates or hospital discharge data, contain major inaccuracies compared with autopsy diagnoses” [9].

Numerous pathological events might occur in the brain of ICH patients (e.g., enlargement of haemorrhage, secondary brainstem bleeding) between the first and second imaging [10] or between the last CT and death. On the other hand, there are ethical and financial limitations of daily CT/MRI during the agony phase for estimating the “final” pathological findings of our patients. So, in our study we combined the classical in vivo (image-based) volumetric evaluations with the measurements performed on the same formalin-fixed brain slices after the death of patients. Taking advantage of this opportunity, the ABC/2 method was also applied, using the initial CT as a point of reference, to assess the growth index of the haematoma in the nonsurvivor group as a novel approach. To the best of our knowledge, no remarkable previous attempts have been made to use this method on cadaver brains with ICH.

Furthermore, we developed a new scoring system to predict the risk of death within 30 days after primary supratentorial ICH using variables found to be independently and significantly associated with outcome by multiple logistic regression analysis. All of these variables can practically be reported within the first few hours after admission. After validation, this scoring system has the potential to be useful for early classification of patients.

2. Subjects and Methods

2.1. Patient Population

Ethical permission for these investigations was granted by the Regional and Institutional Ethics Committee, University of Debrecen, Medical and Health Science Center. We retrospectively reviewed the data of 156 Caucasian patients (71 nonsurvivors and 85 survivors) with primary supratentorial ICH admitted to our Intensive Care Unit (ICU) in a 53-month period. The outcome was defined as 30-day fatality for any reason. All patients were older than 18 years of age and were transported to our ICU within 24 hours of stroke onset. If this point of time could not be ascertained, we used the last time when the patient was known to be well. All patients routinely underwent a baseline nonenhanced CT within 30 minutes of arrival, and the CT confirmed the ICH. A second CT was obtained on average on the 10th day after admission or when symptoms deteriorated. Exclusion criteria were traumatic ICH, subarachnoid haemorrhage (SAH), vascular malformation, tumour, haemorrhagic transformation of ischaemic stroke (on admission or any CT performed later), postthrombolytic haemorrhage in ischaemic stroke, infratentorial ICH, and primary intraventricular bleeding. We did not include subjects who had undergone neurosurgical evacuation or drainage. All patients were treated on specialized stroke units with multiparametric monitoring.

2.2. Data Collection

The following data were collected retrospectively from patients’ clinical notes partly based on the Debrecen Stroke Database [11, 12]: sex, age, current smoking, excessive alcohol consumption, systolic and diastolic arterial blood pressure, and pulse rate at arrival. Laboratory parameters were collected from the initial blood sampling: serum sodium, potassium, glucose levels, sedimentation rate, haemoglobin, white blood cell (WBC) and platelet counts, liver and kidney function tests, and coagulation parameters.

2.3. CT Analysis

Image analysis was carried out retrospectively by two consultant neuroradiologists of our Department of Biomedical Laboratory and Imaging Science, who were blinded to the outcome. CT scans were performed on two 16-slice MDCT (multidetector computed tomography) scanners (GE CT/e Dual, GE Lightspeed 16; GE Medical Systems). Slice thickness was 5 to 10 mms for supratentorial and 2.5 to 4 mms for infratentorial regions. Images were transferred to an offline image processing workstation as DICOM (Digital Imaging and Communications in Medicine) files. Haemorrhage segmentation was carried out using the 3D Slicer software package developed by Brigham and Women’s Hospital Surgical Planning Laboratory and MIT (Boston, Massachusetts, USA) [13]. This procedure allowed separation of the intracranial space from the skull and nonbrain structures; however, this method required verification and manual detachment of incorrectly labelled areas before performing volumetry with the built-in “Measurevol” module. The following variables were constructed: total intracranial volume; total haematoma volume; intraparenchymal haematoma volume; and intraventricular haematoma volume, each expressed as cm3. Additionally, relative volumes were defined as the ratio of total, intraparenchymal, and intraventricular haematoma volumes to intracranial volume yielding variables (without unit).

2.4. Neuropathological Analysis

Autopsies were performed within 48 hours after death. In our Neuropathology Laboratory, brains fixed in 10% formalin were cut into coronal slices. During the examination we verified the clinical diagnosis. Similarly to the in vivo diagnostic the ABC/2 method was used to estimate haematoma volumes [14, 15] based on the assumption that the volume of an intracranial haematoma can be approximated by an ellipsoid unless the haematoma is very irregular in shape. Ellipsoids can be described in terms of Cartesian coordinates using their three largest perpendicular axes [16]. Consecutive coronal slices were laid down with the posterior surface up, and maximal diameters of the haematoma on every slice were measured in mms: one in the horizontal direction and another perpendicular to that. These two largest diameters were used as A and B in the ABC/2 formula and were not necessarily on the same brain slice; each slice has been photo documented. Although slices were originally 10 mms thick, we measured the thickness of each slice at four different locations assuming that the brains may have shrunk during fixation. The precise thickness of a slice was obtained as the average of these measurements. The resulting numerical data were then summed to obtain the maximum diameter of the haematoma perpendicular to the previous ones, which was used as C in the ABC/2 formula. In cases when the haematoma did not fully penetrate a slice, that is, the first and last brain slice containing the bleeding, we measured the depth of the portion of haematoma by sticking a probe into it. Due to fixation in formalin and brain slicing in cases of ventricular extension it was not always exactly discernible which part of the haematoma was in the parenchyma and in the ventricles, respectively. Therefore the haematoma volume calculated by the ABC/2 method was equal to the sum of the intraparenchymal and intraventricular parts. Volumes of haemorrhages were converted from mm3 to cm3. In these cases the relative haematoma volume was calculated using the initial CT intracranial volumetry data (for methods see above).

2.5. Statistical Analysis

The growth index of haematomas was calculated in two steps according to the following logistic formulas (log indicates natural logarithm): , and , where hpb is haematoma per brain; hpb2 is ratio of total haematoma volume obtained by the second volumetry (follow-up CT scan in survivors, neuropathological ABC/2 method in nonsurvivors) to intracranial volume based on the follow-up CT (survivors) or the first CT (nonsurvivors); hpb1 is ratio of total haematoma volume to intracranial volume based on the first CT volumetry (survivors and nonsurvivors). Negative growth indices denote reduction of haematoma; positive growth indices denote haematoma expansion.

Variables were described using mean and SD (continuous variables) or category percentages (categorical variables) stratified for survivors and nonsurvivors. Associations between explanatory factors and the outcome of death within 30 days were evaluated using logistic regression. Curvatures in relationships were assessed and allowed for by transformation or by adding a squared term if this substantially improved model fit. Variables showing remarkable associations in unadjusted models were selected for multiple regression modeling unless they are collinear with each other, in which case elimination was carried out on grounds of clinical practicability. Left-out variables were assessed for their potential contribution by adding to the prefinal multiple model one by one and left in if found clinically remarkable or statistically significant. The final multiple logistic regression model was checked for goodness of fit using the Hosmer-Lemeshow test. The level of significance was set at . To create the scoring system, model-predicted probabilities of outcome were generated for all subjects, and statistics for the minimum, the 10th, 20th, …, 90th percentage points, and the maximum were derived from the resulting set of values. The coefficients of the final model were built into a spreadsheet-based calculation interface which places any patient, with given input data, in terms of probability-of-death percentage range, given the percentage points observed on our sample of patients. All analyses were performed using Stata Statistical Software (StataCorp. 2009).

3. Results

Of the 81 survivor patients, eight were excluded because they did not have a second CT. Of the 75 nonsurvivor patients we excluded two based on brain autopsy results: ruptured aneurysm of the middle cerebral artery and a metastatic tumour, respectively, had been confirmed. Twelve brains were unsuitable for volumetry due to fragmented, multilobar haematomas or having been damaged during removal from the skull, cases in which the ABC/2 method fails to accurately estimate haemorrhage volumes [17]. In addition, medical records were not sufficiently detailed in a total of nine cases. Thus, we analyzed the complete dataset of 59 nonsurvivor and 66 survivor patients.

The relationship between fatal outcome within 30 days and some selected discrete or continuous clinical variables is presented in Table 1. The relationship between age and greater odds of death was of borderline significance. Other variables, including sex, alcohol consumption, smoking, and pulse rate at admission, were not significantly associated with mortality. Higher systolic blood pressure at admission was a predictor of lethal outcome.

Table 1: Summary of epidemiological and clinical data.

There was no statistical significance regarding sex, current smoking, excessive alcohol consumption, diastolic arterial blood pressure, and pulse rate at arrival. Laboratory parameters were collected from the initial blood sampling: serum sodium, potassium, glucose levels, sedimentation rate, haemoglobin, white blood cell (WBC) and platelet counts, liver and kidney function tests, and coagulation parameters. Of these, serum potassium concentration was significantly associated with fatal outcome: high and low levels outside the normal range both represented elevated odds of death. We identified higher serum glucose concentration and lower platelet count as further predictors of fatal outcome.

Tables 2 and 3 summarize the differences in CT characteristics and volumetric findings between survivors and nonsurvivors. The mean time from symptom onset to initial CT was strongly and inversely related to occurrence of death within 30 days. In the survivor group, the mean time between baseline and second CT was 10.7 (SD 5.05) days; in the nonsurvivor group, the mean time between symptom onset and death was 7.7 (SD 6.4) days. Differences between the two groups in absolute volumes of total and intraparenchymal haematoma on baseline CT were strongly significant, similarly to relative haematoma volumes. The growth index of haematoma was found to be a very strong predictor of the outcome. The early presence of intraventricular haemorrhage and its volume was also strongly associated with fatal outcome. There was an approximate 43% decrease in the mean volume of total haemorrhage from baseline to follow-up CT in survivors and a 54% increase from initial CT to death in the other group and, consistently, there was a 42% decrease in survivors and a 56% increase in nonsurvivors in relative total haematoma volume. Intraventricular extension of the haematoma was more frequent in the nonsurvivor group both on initial CT and on follow-up examinations.

Table 2: Baseline CT data.
Table 3: Total and relative haematoma volumes.
3.1. The SUSPEKT Scoring System

The final multiple logistic regression model showed statistically significant associations between 30-day case fatality and a number of variables: serum glucose (Sugar); total haematoma volume (Size); systolic blood Pressure; presence of intraventricular haemorrhage (Extension to the ventricular system); and serum potassium level (Kalium). Using these and adding age (lifeTime) which was of borderline significance we developed the six-factor scoring system SUSPEKT. The SUSPEKT scoring system in use to make predictions after primary supratentorial ICH is demonstrated in Tables 4 and 5. Values of represent means observed in the present study, hence the decimal fraction for intraventricular haemorrhage; denotes Euler’s number.

Table 4: Example 1 for the use of SUSPEKT score.
Table 5: Example 1 for the use of SUSPEKT score.

4. Discussion

Predictors of fatal outcome in primary ICH have been widely studied [6, 1824]. However no reliable and widely used scoring system has been established as yet. Our aim was to assess a relatively simple, reproducible, cost-effective predictive scoring system by analysis of a wide range of clinical and epidemiological data. Our results show that systolic blood pressure, serum potassium and glucose levels, and platelet count independently and statistically significantly predict the 30-day fatal outcome in primary supratentorial ICH. Therefore, we propose a predictive scoring system for clinical outcome in ICH using six parameters (serum glucose; total haematoma volume; systolic blood pressure; presence of intraventricular haemorrhage; serum potassium level; and age) with the acronym SUSPEKT. In the subsequent paragraphs we discuss the individual components of SUSPEKT score and some of the other analyzed candidate parameters which did not show significant correlation with outcome prediction.

Of the laboratory parameters, we found serum glucose (SU) level on admission to be significantly associated with lethal outcome and this result is also supported by previous studies [25, 26], while others reported that, in ICH, hyperglycaemia is not a significant and independent outcome predictor [6, 18, 27]. Recently Lee et al. confirmed in stroke patients by magnetic resonance imaging and spectroscopy study that acute hyperglycaemia adversely affects stroke outcome [28].

We also found that absolute and relative haematoma volumes (“S” for size), haematoma growth index, and presence and size of intraventricular haemorrhage are predictors of death within 30 days after ICH. We recommend the use of absolute haematoma volume since it is easier to calculate and its predictive value is similar to relative volume. Haematoma volume is a well-known essential predictor of outcome in patients with ICH [5, 6, 1820, 2933].

According to previous findings, high systolic [34] and mean arterial [27] blood pressure (P) on admission are associated with outcome, and admission diastolic blood pressure is associated with larger ICH volume [35]. We found only systolic blood pressure to strongly predict 30-day fatal outcome.

Presence and size of intraventricular haemorrhage (“E” for expansion) are also strongly associated with outcome in both overall and supratentorial cerebral haemorrhages [18, 20, 21, 26, 36, 37]. Haematoma expansion is a strong determinant of both mortality and poor functional outcome after ICH [7]. Brott et al. found a trend toward higher mortality at 4 weeks in patients with haemorrhage growth [38].

We observed a previously unreported, very strongly significant association between serum potassium (“K”) level and 30-day lethal outcome, which was related to both high and low levels outside the normal range, although others found that higher dietary potassium intake is associated with lower rates of stroke [39]. Similar (clinicopathological) observation has been published earlier by our group on patients with haemorrhagic transformation [40].

Our result could be explained by the fact that high and low potassium levels are associated with morbidity and mortality also through cardiac effects: ventricular fibrillation and cardiac arrest. At extreme serum potassium levels death probably occurred because of cardiac dysfunction and not directly due to intracranial mass effect.

Age (“T” for lifetime) had a borderline predictive value of 0.054 consistent with previous observations that the role of age in outcome after ICH is controversial. Some prior studies reported that age was significantly and independently correlated with fatal outcome, while others did not find this association [57, 1820, 2931, 41] or only for patients aged ≥80 years [6]. We found that the relationship between age and death within 30 days is of borderline significance.

We identified low platelet count as a predictor of fatal outcome and it is also reported that low platelet count and platelet dysfunction significantly correlate with haematoma growth [42]. Low haemoglobin showed no association with lethal outcome within 30 days. Although Kumar et al. [35] reported that anaemia and WBC count were associated with larger haematomas, none of these variables were predictors of 30-day mortality in concert with our results.

Time from symptom onset to initial CT was inversely and strongly associated with fatal outcome. This paradoxical relationship could be explained by cases of rapidly worsening complaints or severe signs caused by massive haemorrhages having presented to hospital sooner and, in contrast, slower response by patients experiencing mild impairment due to small haemorrhages.

There was no correlation between smoking and 30-day mortality. In a Japanese study smoking was an independent risk factor for death in both the ischaemic and haemorrhagic stroke [36], while other reports found no association [31, 37, 43]. Ikehara et al. reported that heavy alcohol consumption is significantly associated with increased mortality from haemorrhagic stroke for men [44]; however, in the present study we found no association between alcohol consumption and fatal outcome in ICH, which is in accordance with the results of Hansagi et al. [45].

Several prognostic models have been developed to predict fatal outcome after ICH, the most reliable and simplest being the ICH score by Hemphill et al. [6, 19, 2224]. Prognostic scores limited to supratentorial haemorrhages are also known [18, 20, 21]. These studies focused on several factors: age [6, 19, 20, 22], sex [20], level of consciousness (GCS, NIHSS consciousness score) [6, 18, 19, 22], volume or location of haemorrhage [6, 1820, 23], presence of intraventricular haemorrhage [6, 1821, 24], presence of hydrocephalus on initial computed tomography scan [23], subarachnoid extension of haemorrhage [24], high NIHSS total score [22, 24], focal neurological deficit on admission [23], and pulse pressure [6, 24] were predictors of poor clinical prognosis; low temperature at admission and low NIHSS total score were predictors of good clinical outcome (modified Rankin score ≤ 2) [24]. Compared with the previously used prognostic score, undoubtable forcefulness of SUSPEKT score is that it includes serum potassium level which is a novelty and hyperglycaemia which can be corrected [2]. Another advantage of our scoring system is that it consists of only 6 parameters and except haematoma volume these can be managed and influenced by conservative (i.e., nonsurgical) therapeutic interventions.

Early prognostication after ICH is riddled with uncertainties. Given this and the self-fulfilling pessimistic general opinion of poor outcome, reliable early prognostication after ICH can help to avoid a fatalistic approach by the wider care team including family members and enables evidence based DNR (do not resuscitate) orders. Thus, deliberated guideline-concordant therapy is essential for all ICH patients [2]. Early prognostication helps clinicians to reach an objective opinion of predictable outcome.

Our findings are limited by some shortcomings. Prehospital deaths, exclusion of large, multilobar haematomas, and missing second CT in the survivor group may have led to possible selection bias.

5. Conclusion

We have demonstrated that systolic blood pressure, serum potassium and glucose levels, platelet count, absolute and relative haematoma volumes, haematoma growth index, and presence and size of intraventricular haemorrhage independently predict 30-day fatal outcome in primary supratentorial ICH. We developed the SUSPEKT score which after validation on large independent datasets may be applicable in clinical practice for patient selection to optimize individual management and for assessment of eligibility for clinical trials.

Conflict of Interests

The authors declare that there is no conflict of interests regarding the publication of this paper.

Authors’ Contribution

Rita Szepesi, Ibolya Katalin Széll, and Tibor Hortobágyi made equal contributions to this work and are equally considered to be first authors.

Acknowledgments

This work has been supported by the National Brain Research Program, Hungary (KTIA_13_NAP-A-II/7 and KTIA-NAP-13-1-2013-0001), and by the Nonprofit Foundation of the Neurology Department, University of Debrecen.

References

  1. W. Rosamond, K. Flegal, G. Friday et al., “Heart disease and stroke statistics—2007 Update: a report from the American Heart Association Statistics Committee and Stroke Statistics Subcommittee,” Circulation, vol. 115, no. 5, pp. e69–e171, 2007. View at Publisher · View at Google Scholar · View at PubMed · View at Scopus
  2. T. Steiner, R. Al-Shahi Salman, R. Beer et al., “European Stroke Organisation (ESO) guidelines for the management of spontaneous intracerebral hemorrhage,” International Journal of Stroke, vol. 9, no. 7, pp. 840–855, 2014. View at Publisher · View at Google Scholar · View at PubMed
  3. J. P. Broderick, H. P. Adams Jr., W. Barsan et al., “Guidelines for the management of spontaneous intracerebral hemorrhage: a statement for Health Professionals from a Special Writing Group of the Stroke Council, American Heart Association,” Stroke, vol. 30, no. 4, pp. 905–915, 1999. View at Publisher · View at Google Scholar · View at Scopus
  4. S. Kazui, H. Naritomi, H. Yamamoto, T. Sawada, and T. Yamaguchi, “Enlargement of spontaneous intracerebral hemorrhage: incidence and time course,” Stroke, vol. 27, no. 10, pp. 1783–1787, 1996. View at Publisher · View at Google Scholar · View at Scopus
  5. J. P. Broderick, T. G. Brott, J. E. Duldner, T. Tomsick, and G. Huster, “Volume of intracerebral hemorrhage: a powerful and easy-to-use predictor of 30-day mortality,” Stroke, vol. 24, no. 7, pp. 987–993, 1993. View at Publisher · View at Google Scholar · View at Scopus
  6. J. C. Hemphill III, D. C. Bonovich, L. Besmertis, G. T. Manley, and S. C. Johnston, “The ICH score: a simple, reliable grading scale for intracerebral hemorrhage,” Stroke, vol. 32, no. 4, pp. 891–897, 2001. View at Publisher · View at Google Scholar · View at Scopus
  7. S. M. Davis, J. Broderick, M. Hennerici et al., “Hematoma growth is a determinant of mortality and poor outcome after intracerebral hemorrhage,” Neurology, vol. 66, no. 8, pp. 1175–1181, 2006. View at Publisher · View at Google Scholar · View at PubMed · View at Scopus
  8. K. Petros and C. Wittekind, “Autopsy—a procedure of medical history?” Medizinische Klinik—Intensivmedizin und Notfallmedizin, vol. 109, no. 2, pp. 115–120, 2014 (German). View at Publisher · View at Google Scholar · View at PubMed · View at Scopus
  9. K. G. Shojania, E. C. Burton, K. M. McDonald, and L. Goldman, “The autopsy as an outcome and performance measure,” Evidence Report/Technology Assessment (Summary), no. 58, pp. 1–5, 2002. View at Google Scholar · View at Scopus
  10. B. Fülesdi, K. R. Kovács, D. Bereczki, P. Bágyi, I. Fekete, and L. Csiba, “Computed tomography and transcranial doppler findings in acute and subacute phases of intracerebral hemorrhagic stroke,” Journal of Neuroimaging, vol. 24, no. 2, pp. 124–130, 2014. View at Publisher · View at Google Scholar · View at PubMed · View at Scopus
  11. D. Bereczki, L. Mihálka, I. Fekete et al., “The Debrecen Stroke Database: demographic characteristics, risk factors, stroke severity and outcome in 8088 consecutive hospitalised patients with acute cerebrovascular disease,” International Journal of Stroke, vol. 4, no. 5, pp. 335–339, 2009. View at Publisher · View at Google Scholar · View at PubMed · View at Scopus
  12. L. Mihálka, I. Fekete, T. Csépány, L. Csiba, and D. Bereczki, “Basic characteristics of hospital stroke services in Eastern Hungary,” European Journal of Epidemiology, vol. 15, no. 5, pp. 461–466, 1999. View at Publisher · View at Google Scholar · View at Scopus
  13. S. Pieper, B. Lorensen, W. Schroeder, and R. Kikinis, “The NA-MIC Kit: ITK, VTK, pipelines, grids and 3D slicer as an open platform for the medical image computing community,” in Proceedings of the 3rd IEEE International Symposium on Biomedical Imaging: Nano to Macro, pp. 698–701, IEEE, Arlington, Va, USA, April 2006. View at Publisher · View at Google Scholar
  14. R. U. Kothari, T. Brott, J. P. Broderick et al., “The ABCs of measuring intracerebral hemorrhage volumes,” Stroke, vol. 27, no. 8, pp. 1304–1305, 1996. View at Publisher · View at Google Scholar · View at Scopus
  15. J. M. Gebel, C. A. Sila, M. A. Sloan et al., “Comparison of the ABC/2 estimation technique to computer-assisted volumetric analysis of intraparenchymal and subdural hematomas complicating the GUSTO-1 trial,” Stroke, vol. 29, no. 9, pp. 1799–1801, 1998. View at Publisher · View at Google Scholar · View at Scopus
  16. G. C. Newman, “Clarification of abc/2 rule for ICH volume,” Stroke, vol. 38, no. 3, p. 862, 2007. View at Publisher · View at Google Scholar · View at PubMed · View at Scopus
  17. H. B. Huttner, T. Steiner, M. Hartmann et al., “Comparison of ABC/2 estimation technique to computer-assisted planimetric analysis in warfarin-related intracerebral parenchymal hemorrhage,” Stroke, vol. 37, no. 2, pp. 404–408, 2006. View at Publisher · View at Google Scholar · View at PubMed · View at Scopus
  18. S. Tuhrim, J. M. Dambrosia, T. R. Price et al., “Intracerebral hemorrhage: external validation and extension of a model for prediction of 30-day survival,” Annals of Neurology, vol. 29, no. 6, pp. 658–663, 1991. View at Publisher · View at Google Scholar · View at PubMed · View at Scopus
  19. J. L. Ruiz-Sandoval, E. Chiquete, S. Romero-Vargas, J. J. Padilla-Martínez, and S. González-Cornejo, “Grading scale for prediction of outcome in primary intracerebral hemorrhages,” Stroke, vol. 38, no. 5, pp. 1641–1644, 2007. View at Publisher · View at Google Scholar · View at PubMed · View at Scopus
  20. D. R. Lisk, W. Pasteur, H. Rhoades, R. D. Putnam, and J. C. Grotta, “Early presentation of hemispheric intracerebral hemorrhage: prediction of outcome and guidelines for treatment allocation,” Neurology, vol. 44, no. 1, pp. 133–139, 1994. View at Publisher · View at Google Scholar · View at Scopus
  21. S. Tuhrim, D. R. Horowitz, M. Sacher, and J. H. Godbold, “Volume of ventricular blood is an important determinant of outcome in supratentorial intracerebral hemorrhage,” Critical Care Medicine, vol. 27, no. 3, pp. 617–621, 1999. View at Publisher · View at Google Scholar · View at Scopus
  22. C. Weimar, J. Benemann, and H. C. Diener, “German stroke study collaborations. Development and validation of the essen intracerebral hemorrhage score,” Journal of Neurology, Neurosurgery & Psychiatry, vol. 77, no. 5, pp. 601–605, 2006. View at Publisher · View at Google Scholar · View at PubMed
  23. M. Shaya, A. Dubey, C. Berk et al., “Factors influencing outcome in intracerebral hematoma: a simple, reliable, and accurate method to grade intracerebral hemorrhage,” Surgical Neurology, vol. 63, no. 4, pp. 343–348, 2005. View at Publisher · View at Google Scholar · View at PubMed · View at Scopus
  24. R. T. F. Cheung and L.-Y. Zou, “Use of the original, modified, or new intracerebral hemorrhage score to predict mortality and morbidity after intracerebral hemorrhage,” Stroke, vol. 34, no. 7, pp. 1717–1722, 2003. View at Publisher · View at Google Scholar · View at PubMed · View at Scopus
  25. R. Fogelholm, K. Murros, A. Rissanen, and S. Avikainen, “Admission blood glucose and short term survival in primary intracerebral haemorrhage: a population based study,” Journal of Neurology, Neurosurgery and Psychiatry, vol. 76, no. 3, pp. 349–353, 2005. View at Publisher · View at Google Scholar · View at PubMed · View at Scopus
  26. L. G. Stead, A. Jain, M. F. Bellolio et al., “Emergency department hyperglycemia as a predictor of early mortality and worse functional outcome after intracerebral hemorrhage,” Neurocritical Care, vol. 13, no. 1, pp. 67–74, 2010. View at Publisher · View at Google Scholar · View at PubMed · View at Scopus
  27. S. Tetri, S. Juvela, P. Saloheimo, J. Pyhtinen, and M. Hillbom, “Hypertension and diabetes as predictors of early death after spontaneous intracerebral hemorrhage: clinical article,” Journal of Neurosurgery, vol. 110, no. 3, pp. 411–417, 2009. View at Publisher · View at Google Scholar · View at PubMed · View at Scopus
  28. S.-H. Lee, B. J. Kim, H.-J. Bae, J. S. Lee, B.-J. Park, and B.-W. Yoon, “Effects of glucose level on early and long-term mortality after intracerebral haemorrhage: the Acute Brain Bleeding Analysis Study,” Diabetologia, vol. 53, no. 3, pp. 429–434, 2010. View at Publisher · View at Google Scholar · View at PubMed · View at Scopus
  29. O. G. Nilsson, A. Lindgren, L. Brandt, and H. Säveland, “Prediction of death in patients with primary intracerebral hemorrhage: a prospective study of a defined population,” Journal of Neurosurgery, vol. 97, no. 3, pp. 531–536, 2002. View at Publisher · View at Google Scholar · View at PubMed · View at Scopus
  30. M. Gomis, A. Ois, A. Rodríguez-Campello et al., “Outcome of intracerebral haemorrhage patients pre-treated with statins,” European Journal of Neurology, vol. 17, no. 3, pp. 443–448, 2010. View at Publisher · View at Google Scholar · View at PubMed · View at Scopus
  31. S. Juvela, “Risk factors for impaired outcome after spontaneous intracerebral hemorrhage,” Archives of Neurology, vol. 52, no. 12, pp. 1193–1200, 1995. View at Publisher · View at Google Scholar · View at Scopus
  32. M. Reinhard, F. Neunhoeffer, T. A. Gerds et al., “Secondary decline of cerebral autoregulation is associated with worse outcome after intracerebral hemorrhage,” Intensive Care Medicine, vol. 36, no. 2, pp. 264–271, 2010. View at Publisher · View at Google Scholar · View at PubMed · View at Scopus
  33. J. M. Gebel Jr., E. C. Jauch, T. G. Brott et al., “Relative edema volume is a predictor of outcome in patients with hyperacute spontaneous intracerebral hemorrhage,” Stroke, vol. 33, no. 11, pp. 2636–2641, 2002. View at Publisher · View at Google Scholar · View at Scopus
  34. S. Tuhrim, J. M. Dambrosia, T. R. Price et al., “Prediction of intracerebral hemorrhage survival,” Annals of Neurology, vol. 24, no. 2, pp. 258–263, 1988. View at Publisher · View at Google Scholar · View at PubMed · View at Scopus
  35. M. A. Kumar, N. S. Rost, R. W. Snider et al., “Anemia and hematoma volume in acute intracerebral hemorrhage,” Critical Care Medicine, vol. 37, no. 4, pp. 1442–1447, 2009. View at Publisher · View at Google Scholar · View at PubMed · View at Scopus
  36. K. S. Wong, “Risk factors for early death in acute ischemic stroke and intracerebral hemorrhage: a prospective hospital-based study in Asia. Asian Acute Stroke Advisory Panel,” Stroke, vol. 30, no. 11, pp. 2326–2330, 1999. View at Publisher · View at Google Scholar · View at Scopus
  37. H. Ueshima, S. R. Choudhury, A. Okayama et al., “Cigarette smoking as a risk factor for stroke death in Japan NIPPON DATA80,” Stroke, vol. 35, no. 8, pp. 1836–1841, 2004. View at Publisher · View at Google Scholar · View at PubMed · View at Scopus
  38. T. Brott, J. Broderick, R. Kothari et al., “Early hemorrhage growth in patients with intracerebral hemorrhage,” Stroke, vol. 28, no. 1, pp. 1–5, 1997. View at Publisher · View at Google Scholar · View at Scopus
  39. L. D'Elia, G. Barba, F. P. Cappuccio, and P. Strazzullo, “Potassium intake, stroke, and cardiovascular disease: a meta-analysis of prospective studies,” Journal of the American College of Cardiology, vol. 57, no. 10, pp. 1210–1219, 2011. View at Publisher · View at Google Scholar · View at PubMed · View at Scopus
  40. L. Kerenyi, L. Kardos, J. Szász et al., “Factors influencing hemorrhagic transformation in ischemic stroke: a clinicopathological comparison,” European Journal of Neurology, vol. 13, no. 11, pp. 1251–1255, 2006. View at Publisher · View at Google Scholar · View at PubMed · View at Scopus
  41. J. Garibi, G. Bilbao, I. Pomposo, and C. Hostalot, “Prognostic factors in a series of 185 consecutive spontaneous supratentorial intracerebral haematomas,” British Journal of Neurosurgery, vol. 16, no. 4, pp. 355–361, 2002. View at Publisher · View at Google Scholar · View at Scopus
  42. W. C. Ziai, M. T. Torbey, T. S. Kickler, S. Oh, A. Bhardwaj, and R. J. Wityk, “Platelet count and function in spontaneous intracerebral hemorrhage,” Journal of Stroke and Cerebrovascular Diseases, vol. 12, no. 4, pp. 201–206, 2003. View at Publisher · View at Google Scholar · View at PubMed · View at Scopus
  43. A. G. Thrift, J. J. McNeil, and G. A. Donnan, “The risk of intracerebral haemorrhage with smoking. The melbourne risk factor study group,” Cerebrovascular Diseases, vol. 9, no. 1, pp. 34–39, 1999. View at Publisher · View at Google Scholar · View at Scopus
  44. S. Ikehara, H. Iso, H. Toyoshima et al., “Alcohol consumption and mortality from stroke and coronary heart disease among Japanese men and women: the Japan Collaborative Cohort Study,” Stroke, vol. 39, no. 11, pp. 2936–2942, 2008. View at Publisher · View at Google Scholar · View at PubMed · View at Scopus
  45. H. Hansagi, A. Romelsjo, M. G. De Verdier, S. Andreasson, and A. Leifman, “Alcohol consumption and stroke mortality: 20-year follow-up of 15,077 men and women,” Stroke, vol. 26, no. 10, pp. 1768–1773, 1995. View at Publisher · View at Google Scholar · View at Scopus