We retrospectively studied consecutive patients with a diagnosis of S-ICH admitted to the Stroke Unit at our single hospital (a university hospital serving a population of 1500000) between January 2000 and December 2011. Patients aged greater than 40 years were selected because of the likelihood of a secondary etiology in younger patients. In addition, we excluded patients with an infratentorial hemorrhage, because small changes in hemorrhage size or location are believed to have greater impacts on survival than supratentorial hemorrhages in these patients. Patient related data were independently extracted from a computerized database (PACS; Marosis m-view, Marotech Co., Seoul, Korea) by three clinical research coordinators using a structured form. To preserve patient confidentiality, patient identifiers were omitted from the collated data set. The inclusion criteria applied were a diagnosis of supratentorial S-ICH, confirmed by computerized tomography (CT) or magnetic resonance imaging, and admission to a stroke unit within 24 hours of symptom onset. As recommended by the Stroke Council of the American Heart Association1,2), selected patients underwent conventional angiography to identify secondary causes. Patients with a hemorrhage secondary to head trauma, a ruptured cerebral aneurysm, an arteriovenous malformation, a tumor, bleeding diathesis, or a hemorrhagic infarction were excluded. This study was approved by the institutional review boards of our hospital (2012-SCMC-065-00).

Hemodialysis was performed using Gambro AK 95® units (Gambro, Stockholm, Sweden) equipped with a Polyamix® dialysis membrane (Gambro) using Bicart® (Gambro) dialysis solution. Hemodialysis was performed three times a week at 4 hours/day at a mean blood flow rate of 150-200 mL/min. The anticoagulant (heparin) loading dose was 1000 IU and was intravenously infused, and the maximum maintenance dose was 200 IU per hour. In addition, nafamostat mesylate was intravenously infused at an hourly dosage of 35 mg without a loading dose.

The baseline variables recorded were selected by reviewing the literature21,25,29). Demographic variables included; age, gender, cigarette smoking and alcohol intake histories, body mass index (BMI), level of consciousness, pupillary abnormalities, limb weakness, and underlying disease. Smoking status was classified as never smoker, ex-smoker (a lifelong smoking history of ≥6 months), or current smoker (≥10 cigarettes/day for ≥6 months), and alcohol intake was dichotomized as < or ≥46 g/day 3 days/week12). BMI was calculated by dividing weight in kilograms by height in meters squared. Pupillary abnormalities at admission were classified as 0, 1, or 2 based on pupil reactivity scores17). Levels of consciousness at admission were assessed using the Glasgow Coma Scale (GCS)27); GCS scores of 3-8, 9-12, and 13-15 were defined as 'severely altered', 'moderately altered', or 'mildly altered or normal' conscious states, respectively. Blood pressure at admission was not viewed as a potential prognostic variable because it is often markedly elevated during the first 1 to 2 days after a severe stroke.

CT variables, that is, side of hemorrhage, location of hemorrhage, volume of hematoma, lateral shift of cerebral midline structures, and ventricular extension of hemorrhage, were also included in the analysis. Hematoma volume was estimated from axial CT scans using the formula A×B×C/2, where A is greatest diameter on the largest hemorrhage slice, B is maximal diameter perpendicular to this, and C is vertical hematoma depth18). Hematoma locations were defined based on sites of origin, such as, basal ganglia, thalamus, lobal, or ventricle alone. Ventricular extension of hemorrhage included a pure intraventricular hemorrhage and secondary spread of cerebral parenchymal hemorrhage into ventricular space. A CT scan was performed whenever neurological status changed and 48 hours after the initial CT scan if neurological status was stable. Hematoma enlargement was defined as an increase in hematoma volume of ≥25% versus initial CT volume. Laboratory parameters obtained at admission were dichotomized during the analysis; these were hemoglobin, leukocyte count, platelet count, C-reactive protein (CRP) level, blood urea nitrogen (BUN), creatine, prothrombin time and activated partial prothrombin time. International Normalized Ratio (INR), plasma total cholesterol level, aspartate aminotransferase (AST), alanine aminotransferase (ALT), blood plasma glucose, and hemoglobin A1c.

Patients were treated in accord with institutional treatment algorithms. All patients (medical and surgical) were cared for in a neurosurgical intensive care unit until considered stable enough to move to a general unit. Treatment was administered according to current practices, but was not rigidly regimented as primary attending neurosurgeons were allowed to exercise medical judgment.

Surgical intervention was undertaken in selected patients exhibiting clinical deterioration and a moderate or larger amount of lobar hemorrhage, and in patients with basal ganglionic or thalamic hemorrhage with a hematoma volume of >30 mL, in patients with an expanding hematoma, and in patients with evidence of progressive neurological deterioration. Surgical approaches were individualized based on S-ICH location and size. The techniques used were craniotomy, craniectomy, CT-guided stereotactic hematoma evacuation, and/or extraventricular drainage. Patients with a deep-seated S-ICH usually underwent stereotactic evacuation when hematoma volume was from 30 to 60 mL by preoperative CT. Immediately after initial hematoma aspiration or evacuation, patients underwent repeat CT to check catheter placement. When repeat CT showed remnant hematoma, a urokinase (6000 U) injection was considered through the catheter to facilitate aspiration, and a CT scan was repeated 12 hours later. Patients with a hemorrhage of 10 to 30 mL and a severe neurological deficit (limb weakness) were also treated by stereotactic evacuation. On the other hand, when hematoma volume exceeded 60 mL, evacuation was performed via craniotomy or decompressive craniectomy. The attending neurosurgeon decided whether to implant a bone-flap (craniotomy or craniectomy) depending on the intraoperative presence of cerebral swelling after S-ICH removal. Conservative treatment mainly involved the management of hypertension and intracranial pressure. Blood pressure was controlled using several medications. Mean arterial pressure was maintained in the range 100 to 140 mm Hg using an antihypertensive (e.g., labetalol, amolodipine, or nicardipin). In addition, a hypertonic agent (e.g., glycerol or mannitol) was administered when CT indicated a mass effect or when clinical symptoms indicated an elevated intracranial pressure.

Patients were followed up for at least 6 months after S-ICH. Functional outcomes were assessed using the Glasgow Outcome Scale (GOS) in 6 months after ictus. GOS scores were defined as follows15) : 1=dead, 2=vegetative state, 3=severe disability (able to follow commands, but unable to live independently), 4=moderate disability (able to live independently, but unable to return to work or school), 5=good recovery (able to return to work or school).

For statistical purposes, patients were assigned to two outcome category groups, namely, a "good functional recovery" group (GOS score 4 to 5), members of which were functionally independent, or a "poor functional recovery" group (GOS score 1 to 3). Information on mortality and functional outcome was obtained from medical records, information obtained from family members was obtained by conducting telephone interviews, and information from patients was obtained by conducting direct examinations at our outpatient department after discharge.

Our primary aim was to estimate 30-day mortality and functional recovery at 6-month post-S-ICH in S-ICH patients on chronic hemodialysis for ESRD patients. We defined 30-day mortality as death within 30 days of a radiological diagnosis of S-ICH, and functional recovery was defined as described by GOS at 6 months after a radiological diagnosis of S-ICH.

SPSS version 12.0 (SPSS Institute, Inc., Chicago, IL, USA) was used for the analysis, and the Student's t test and a nonparametric test were used to analyze continuous and ordinal variables (e.g., mean levels of BUN and creatine), respectively. Discrete variables, such as, gender, smoking, chronic alcohol intake, ventricular extension of hemorrhage, and hydrocephalus and categorical variables, such as, age subgroup, BMI, GCS subgroup, pupillary abnormality, underlying conditions, and location of hemorrhage, were analyzed using Pearson's χ² test. Univariate analysis was used initially to identify relationships between potential prognostic factors and 30-day mortality and functional recovery at 6 months post-S-ICH using the χ² test. Subsequently, multivariate logistic regression analysis was used to identify variables independently associated with the above-mentioned outcomes. Kaplan-Meier survival analysis was used to determine cumulative mortalities. Statistical significance was accepted for p values of <0.05, and all tests were 2-tailed.

No significant intergroup differences were observed between demographic characteristics, that is, age, gender, chronic alcohol intake, body mass index, or new limb weakness. However, ESRD patients had a lower mean GCS score (p=0.028), and more ESRD patients had nonreactive pupils (p=0.040) and underlying diseases (p=0.013) (Table 1).

In terms of radiologic characteristics, no significant intergroup difference was found for sidedness or hemorrhagic volume, midline shift, location of hemorrhage, or hydrocephalus. However, ventricular extension of hemorrhages (p=0.010), and hematoma enlargement within 48 hours were more prevalent in the ESRD patients (Table 2).

In terms of laboratory findings, no significant intergroup difference was found for leukocytosis (≥10×10³/µL), hepatopathy (AST or ALT >40 IU/L), or elevated hemoglobin A1c (≥6.0%). However, anemia (<11.7 g/dL in male, <10.7 g/dL in female) (p<0.001), thrombocytopenia (<150×10³/µL) (p=0.032), elevated INR (>1.2) (p=0.004), elevated CRP (>5.0 mg/L) (p=0.002), elevated blood plasma glucose (>126 mg/dL) (p<0.001), elevated BUN (>24 mg/dL) (p<0.001), elevated serum creatine (>1.4 mg/dL) (p<0.001), and abnormal findings of urine analysis (p<0.001) were more prevalent in S-ICH patients with ESRD (Table 3). In these patients, mean BUN was 54.2 mg/dL (range from 10.7 mg/dL to 145.9 mg/dL) and mean serum creatine was 7.7 mg/dL (range from 2.9 mg/dL to 27.7 mg/dL). Urine output was nil in 43 S-ICH patients with ESRD (42.2%); proteinuria was present in 49 (48.0%), glucosuria in 24 (23.5%), and hematuria in 17 (16.7%) (Table 3).

Regarding treatment modalities, 1005 (64.5%) patients were treated conservatively and 553 (35.5%) were treated surgically. No significant intergroup difference was found in terms of treatment modality type (p=0.667). Furthermore, of the 553 S-ICH patients that underwent surgical treatment, no significant difference was found between S-ICH patients with or without ESRD in terms of surgical techniques (p=0.605) or time to surgery (p=0.311) (Table 4).

For all study subjects, 30-day, 3-month, and 6-month mortalities were 16.4%, 25.7%, and 27.3%, respectively. Overall, 256 patients died within 30 days, 401 patients within 3 months, and 425 patients within 6 months of S-ICH. A significant and meaningful intergroup difference was found between 30-day (p<0.001), 3-month (p<0.001), and 6-month (p<0.001) mortalities (Table 5).

In terms of functional recovery, at 30 days after S-ICH, GOS scores were similar (p=0.545), but at 3 months (p=0.046) and 6 months after S-ICH GOS scores were significantly different (p=0.013) (Table 5).

In patients without ESRD, univariate analysis showed that the following variables were significantly associated with death within 30 days of S-ICH; GCS score, pupillary abnormality, underlying disease, ventricular extension of hemorrhage, hemorrhagic volume, thrombocytopenia, and treatment modality. In S-ICH patients with ESRD in addition to these factors, age, anemia, and C-reactive protein were also found to be associated with death within 30 days of S-ICH.

Multivariate logistic regression analysis was performed on variables found to be significant by univariate analysis (Table 6). In the S-ICH patients without ESRD, multivariate analysis showed that GCS score (3 to 8 vs. 9 to 12; p=0.038, and 3 to 8 vs. 13 to 15; p<0.001), pupillary abnormality (0 reactive vs. 2 reactive; p<0.001), treatment modality (surgery vs. conservative treatment; p=0.006, and craniotomy vs. other surgical modalities; p<0.001), underlying disease (ischemic heart disease vs. other disease; p=0.034, and atrial fibrillation vs. other disease; p=0.018) were independently associated with 30-day mortality. In the S-ICH patients with ESRD, multivariate analysis showed that age (≥70 years vs. 40 to 54 years; p=0.019), GCS score (3 to 8 vs. 13 to 15; p<0.001, and 9 to 12 vs. 13 to 15; p=0.014), pupillary abnormality (0 reactive vs. 1 reactive; p=0.032, and 0 reactive vs. 2 reactive; p<0.001), ventricular extension of hemorrhage (present vs. absent; p=0.017), hemorrhagic volume (≥60 mL vs. <30 mL; p=0.002, and 30-59 mL vs. <30 mL; p=0.029), anemia (present vs. absent; p=0.008), and treatment modality (surgery vs. conservative treatment; p=0.034), were independently associated with high 30-day mortality. Associations between other variables and 30-day mortality by univariate analysis were not confirmed by multivariate analysis. In S-ICH patients with ESRD, duration of hemodialysis was not associated with 30-day mortality (p=0.199).

Actually, underlying disease, which was independently associated with high 30-day mortality in S-ICH patients without ESRD, was not associated with high 30-day mortality in S-ICH patients with ESRD. Conversely, several other variables, such as, ventricular extension of hemorrhage, hemorrhagic volume, and anemia, not associated with 30-day mortality in S-ICH patients without ESRD, were found to be significantly and independently associated with 30-day mortality in S-ICH with ESRD.

Of the variables examined, the following were found to be significantly associated with functional recovery in S-ICH patients with and without ESRD by univariate analysis; age, GCS score, location of hemorrhage, hemorrhagic volume, and treatment modality. However, pupillary response, underlying disease, ventricular extension of hemorrhage, anemia, thrombocytopenia, and CRP, which were all found to be associated with 30-day mortality, were not found to be associated with functional recovery in either group by univariate analysis.

In patients without ESRD, multivariate logistic regression analysis showed that the following five variables were independently associated with functional recovery at 6 months after S-ICH : age (40 to 54 years vs. ≥70 years; p<0.001, and 40 to 54 years vs. 55 to 69 years; p=0.038), GCS score (9 to 12 vs. 3 to 8; p=0.009, 13 to 15 vs. 3 to 8; p<0.001, and 13 to 15 vs. 9 to 12; p=0.041), location of hemorrhage (thalamus vs. other regions; p=0.022, lobal vs. other regions; p=0.042, and ventricle only vs. other regions; p=0.008), volume of hemorrhage (<30 mL vs. ≥60 mL; p=0.001), and treatment modality (conservative treatment vs. surgery; p=0.039, other surgical modalities vs. craniotomy; p<0.001, and stereotactic aspiration vs. other surgical modalities; p=0.005).

In S-ICH patients with ESRD, multivariate logistic regression analysis showed that the following four variables were independently associated with functional recovery at 6 months after S-ICH; GCS score (9 to 12 vs. 3 to 8; p<0.001, and 13 to 15 vs. 3 to 8; p<0.001), hemorrhagic volume (<30 mL vs. ≥60 mL), treatment modality (conservative treatment vs. surgery; p<0.001), and duration of hemodialysis treatment (<50 months vs. ≥50 months; p=0.027). Although age and location of hemorrhage were associated with good functional recovery by univariate analysis in S-ICH patients with ESRD, multivariate analysis showed no such association. Thus, age and location of hemorrhage were associated with functional recovery in S-ICH patients without ESRD, but not in S-ICH patients with ESRD (Table 7).