I
I
ntraventrIcularntraventrIcular extension occurs in 30%–45% of extension occurs in 30%–45% ofpatients with ICH and is an independent predictor of
patients with ICH and is an independent predictor of
poor outcome.
poor outcome.1,3,25,261,3,25,26The presence of IVH signicantlyThe presence of IVH signicantly
increases the risk of death and IVH volume directly cor
increases the risk of death and IVH volume directly cor-
-responds to the likelihood of death.
responds to the likelihood of death.2727Recent reports haveRecent reports have
demonstrated that early expansion of IVH worsens out
demonstrated that early expansion of IVH worsens out-
-come
come2525 and intraventricular clot removal reduces inamand intraventricular clot removal reduces inam-
-mation, hydrocephalus, and long-term functional decits
mation, hydrocephalus, and long-term functional decits
in the setting of ICH.
in the setting of ICH.12,18,23,2812,18,23,28Thus, the accurate assessmentThus, the accurate assessment
of IVH volume and severity is critical, particularly to de
of IVH volume and severity is critical, particularly to de-
-termine the efcacy of novel treatments that aim to
termine the efcacy of novel treatments that aim to removeremove
Evaluation of intraventricular hemorrhage
Evaluation of intraventricular hemorrhage
assessment methods for predicting outcome following
assessment methods for predicting outcome following
intracerebral hemorrhage
intracerebral hemorrhage
Clinical article
Clinical article
B
BrianrianYY. H. Hwangwang, M.D., , M.D., SSaMuelaMuelS. BS. Bruceruce, B.a., , B.a., ggeoffreYeoffreYaappelBooMppelBooM, M.D., , M.D., M
MattHewattHew a. pa. piazzaiazza, B.a., , B.a., aaManDaManDaM. cM. carpenterarpenter, B.a., , B.a., ppaulaulr. gr. giganteigante, M.D., , M.D., c
cHriStopHerHriStopHerpp. K. Kellnerellner, M.D., , M.D., aanDrewnDrewff. D. Ducruetucruet, M.D., , M.D., MMicHaelicHael a. Ka. Kellnerellner, B.S., , B.S., r
rajeevajeev DDeBeB-S-Senen, , KKerrYerrY a. va. vaugHanaugHan, B.a., , B.a., ppHilipHilipM. MM. MeYerSeYerS, M.D., , M.D., anD
anD e. Se. SanDeranDer cconnollYonnollYjjrr., M.D.., M.D.
Department of Neuro
Department of Neurological Surgerylogical Surgery, Columbia University College of Physicians and S, Columbia University College of Physicians and Surgeonsurgeons, New Y, New York,ork,
New Y
New York ork
Object.
Object. Intraventricular hemorrhage (IVH) associated with intracerebral hemorrhage (ICH) is an independentIntraventricular hemorrhage (IVH) associated with intracerebral hemorrhage (ICH) is an independent predictor of poor outcome. Clinical methods for evaluating IVH, however, are not well established. This study sought predictor of poor outcome. Clinical methods for evaluating IVH, however, are not well established. This study sought to determine the best IVH
to determine the best IVH grading scale by evaluating the predictive accuracies of IVH, grading scale by evaluating the predictive accuracies of IVH, Graeb, and LeRoux scores inGraeb, and LeRoux scores in an independent cohort of ICH patients with IVH. Subacute IVH dynamics as well as the impact of external ventricu an independent cohort of ICH patients with IVH. Subacute IVH dynamics as well as the impact of external ventricu- -lar drain (EVD) placement on IVH and outcome were also investigated.
lar drain (EVD) placement on IVH and outcome were also investigated. Methods.
Methods.A consecutive cohort of 142 primary ICH patients with IVH was admitted to Columbia UniversityA consecutive cohort of 142 primary ICH patients with IVH was admitted to Columbia University Medical Center between February 2009 and
Medical Center between February 2009 and February 2011. Baseline demographics, clinical presentation, and hospiFebruary 2011. Baseline demographics, clinical presentation, and hospi- -tal course were prospectively recorded. Admission CT scans performed within 24 hours of onset were reviewed for tal course were prospectively recorded. Admission CT scans performed within 24 hours of onset were reviewed for ICH location, hematoma volume, and presence of IVH. Intraventricular hemorrhage was categorized according to ICH location, hematoma volume, and presence of IVH. Intraventricular hemorrhage was categorized according to IVH, Graeb, and LeRoux scores. For each patient, the
IVH, Graeb, and LeRoux scores. For each patient, the last scan performed within 6 days of ictus was simillast scan performed within 6 days of ictus was similarly evaluarly evalu- -ated. Outcomes at discharge were assessed using the modied Rankin Scale (mRS). R
ated. Outcomes at discharge were assessed using the modied Rankin Scale (mRS). Receiver operating characteristiceceiver operating characteristic analysis was used to determine the
analysis was used to determine the predictive accuracies of the grading scales for predictive accuracies of the grading scales for poor outcome (mRS scorepoor outcome (mRS score≥≥3).3).
Results.
Results.Seventy-three primary ICH patients (51%) had IVH. Median admission IVH, Graeb, and LeRSeventy-three primary ICH patients (51%) had IVH. Median admission IVH, Graeb, and LeRoux scoresoux scores were 13, 6, and 8, respectively. Median IVH, Graeb and LeRoux scores decreased to 9 (p = 0.005), 4 (p = 0.002), were 13, 6, and 8, respectively. Median IVH, Graeb and LeRoux scores decreased to 9 (p = 0.005), 4 (p = 0.002), and 4 (p = 0.003), respectively, within 6 days of ictus. Poor outcome w
and 4 (p = 0.003), respectively, within 6 days of ictus. Poor outcome was noted in 55 patients (75%). Areas under theas noted in 55 patients (75%). Areas under the receiver operating characteristic curve were similar among the IVH, Graeb, and LeRoux scores (0.745, 0.743, and receiver operating characteristic curve were similar among the IVH, Graeb, and LeRoux scores (0.745, 0.743, and 0.744, respectively) and within 6 days postictus (0.765, 0.722, 0.723, respectively). Moreover, the IVH, Graeb, and 0.744, respectively) and within 6 days postictus (0.765, 0.722, 0.723, respectively). Moreover, the IVH, Graeb, and LeRoux scores had similar maximum Youden Indices both at admission (0.515 vs 0.477 vs 0.440, respectively) and LeRoux scores had similar maximum Youden Indices both at admission (0.515 vs 0.477 vs 0.440, respectively) and within 6 days postictus (0.515 vs 0.339 vs 0.365, respectively). Patients who received EVDs had higher mean IVH within 6 days postictus (0.515 vs 0.339 vs 0.365, respectively). Patients who received EVDs had higher mean IVH volumes (23
volumes (23 ±±26 ml vs 926 ml vs 9±±11 ml, p = 0.003) and increased incidence of Glasgow Coma Scale scores11 ml, p = 0.003) and increased incidence of Glasgow Coma Scale scores <<8 (67% vs8 (67% vs
38%, p = 0.015) and hydrocephalus (82% vs 50%, p = 0.004) at admission but had similar outcome as those who did 38%, p = 0.015) and hydrocephalus (82% vs 50%, p = 0.004) at admission but had similar outcome as those who did not receive an EVD.
not receive an EVD. Conclusions.
Conclusions.The IVH, Graeb, and LeRoux scores predict outcome well with similarly good accuracy in ICHThe IVH, Graeb, and LeRoux scores predict outcome well with similarly good accuracy in ICH patients with IVH when assessed at admission and within 6 days after hemorrhage. Therefore, any of one of the patients with IVH when assessed at admission and within 6 days after hemorrhage. Therefore, any of one of the scores would be equally useful for assessing IVH severity and risk-stratifying ICH patients with regard to outcome. scores would be equally useful for assessing IVH severity and risk-stratifying ICH patients with regard to outcome. These results suggest that EVD placement may be benecial for patients with severe IVH, who have particularly poor These results suggest that EVD placement may be benecial for patients with severe IVH, who have particularly poor prognosis at admission, but a randomized clinical trial is needed to conclusively demonstrate its therapeutic value. prognosis at admission, but a randomized clinical trial is needed to conclusively demonstrate its therapeutic value. (DOI: 10.3171/2011.9.JNS10850)
(DOI: 10.3171/2011.9.JNS10850)
K
KeYeYwworDSorDS • • intraventricular intraventricular hemorrhage hemorrhage • • intracerebral intracerebral hemorrhage hemorrhage ••
Graeb
Graeb Score Score • • LeRoux LeRoux Score Score • • IVH IVH Score Score • • outcome outcome • • vascular vascular disordersdisorders
Abbrev
Abbreviations iations used in used in this pathis paper:per:AUROC = area under the ROCAUROC = area under the ROC curve; EVD = external ventricular drain; GCS = Glasgow Coma curve; EVD = external ventricular drain; GCS = Glasgow Coma Scale; ICH = intracerebral hemorrhage; IVH = intraventricular hem Scale; ICH = intracerebral hemorrhage; IVH = intraventricular hem- -orrhage; mRS = modified Rankin scale; ROC = receiver operating orrhage; mRS = modified Rankin scale; ROC = receiver operating characteristic; tPA = tissue plasminogen activator.
or reduce the effects of IVH. Methods to grade IVH sever-ity, however, are not well established.
The IVH,11Graeb,10 and LeRoux scores17were
devel-oped to estimate IVH severity based on gross hemorrhage size and the presence of dilation within each ventricle (Fig. 1). The scores are useful for IVH evaluation at admission and have been used to dene IVH severity20 or to select
patients for studies;21 however, their clinical effectiveness
to assess IVH or to predict outcome after ICH remains un-clear. The aim of this study was to prospectively evaluate the predictive accuracies of the IVH, Graeb, and LeRoux scores for functional outcome at discharge in a single-center, consecutive series of ICH patients with IVH. We also sought to investigate the subacute IVH dynamic, as well as the impact of EVD placement on IVH severity and outcome.
Methods
Patient Selection and Data Collection
All patients with spontaneous ICH admitted to the neurological intensive care unit of Columbia University Medical Center between February 2009 and February 2011 were offered participation in the Columbia University In-tracerebral Hemorrhage Outcomes Project. The study was approved by the hospital’s institutional review board, and in all cases, informed consent was obtained from the pa-tient or surrogate. The diagnosis of ICH was established by CT scan. Patients were included in the analysis if they had spontaneous nontraumatic ICH with evidence of IVH on CT scans upon admission and within 24 hours of onset.11
Patients were excluded if the ICH was due to secondary causes such as trauma, aneurysm, or arteriovenous malfor-mation rupture, or hemorrhagic conversion of an infarct. Clinical Assessment and Management
All patients received standard neurological intensive care unit management according to our institutional pro -tocol, corresponding to the latest American Heart Asso-ciation guidelines.2,19 An EVD was placed in patients with
evidence of symptomatic hydrocephalus and a GCS score
< 8. The catheters were left in place until the total daily
amount of drained CSF became less than 100 ml and no ventriculomegaly or increased intracranial pressure was recorded after 48 hours of clamping. Surgical hematoma evacuation was strongly considered for patients with cere-bellar hemorrhage with clinical deterioration or those with brainstem compression and/or hydrocephalus. Surgery was also considered for patients with lobar clots greater than 30 ml and within 1 cm of the cortical surface. The decision to administer intrathecal tPA was based on the preference of the treating neurosurgeon. The protocol consisted of re-moval of approximately 3 ml of CSF followed by 1 mg of tPA delivered in 1 ml of saline, and a 2-ml saline ush. Patients received intrathecal tPA twice daily until the deci-sion was made to discontinue the therapy, based on nearly complete clearance of blood from the ventricles. Further details regarding the clinical management of the patients have been described separately.7 Baseline demographics,
medical history, clinical presentation, and clinical course, including EVD insertions, were prospectively recorded. Radiological Assessment
All CT scans were assessed using identical techniques. Two authors (G.A. and C.P.K.) analyzed the scans while blinded to radiologists’ reports and patient outcomes. Con-sensus was then reached for each scan. Admission CT scans performed within 24 hours of onset were reviewed for ICH location, ICH volume, and presence of intraven-tricular extension. Intravenintraven-tricular hemorrhage severity was assessed by determining the IVH,11 Graeb,10 and
Le-Roux scores17 according to previously published criteria.
Intracerebral hemorrhage volume was determined using the ABC/2 method.16Intraventricular hemorrhage volume
was measured as previously described, using a computer-based planimetric analysis program, Medical Image Pro-cessing, Analysis and Visualization (NIH).11 Scans were
evaluated for the presence of hydrocephalus and the of-cial radiology reports were consulted in questionable cases. Intraventricular hemorrhage was scored and ICH and IVH volumes were assessed based on the last scan performed during this period.
Fig. 1. Grading scales for assessing IVH severity. A: The IVH score (range 0–23). The IVH score is calculated using the
following equation: 3×(right lateral ventricle score+left lateral ventricle score+hydrocephalus score)+third ventricle score+ fourth ventricle score. B: The Graeb score (range 0–12). C: The LeRoux score (range 0–16).
Outcome Assessment
Survival and functional outcomes at discharge were as-sessed using the mRS, and patients were routinely followed up after discharge by the study neurosurgeon (E.S.C.). Poor outcome was dened as mRS scores≥3.
Statistical Analysis
Continuous variables were dichotomized based on clinical cutoff points or median values if they were not normally distributed. The chi-square or Fisher exact tests were used to nd signicant associations between categor-ical or dichotomized variables. The 2-sided Student t-test or the Mann-Whitney U-test was used for normally and nonnormally distributed continuous variables, respective -ly. Receiver operating characteristic analysis was used to determine the predictive accuracies of the LeRoux, Graeb, and IVH scores with regard to discharge poor outcome, as measured by the AUROC. The ROC analysis measures the ability of a scale to discriminate patients with a higher probability of poor outcome from those with a lower prob-ability of poor outcome by assigning the former a higher score, and has been used to determine a clinical grading scale’s predictive accuracy for outcome.4,5,9 An AUROC
greater than 0.7 is generally considered useful, and an AUROC between 0.8 and 0.9 indicates excellent accura-cy.8 Dependent AUROCs were compared
nonparametri-cally,13 and the maximum Youden Index was identied
for each scale to assess their prognostic performance. The maximum Youden Index is the optimization of a function that gives equal weight to sensitivity and specic -ity. It is a method of ROC analysis that offers an alter-native measure to AUROC; whereas AUROC concerns a variable’s discriminative utility at all cutoff points, the maximum Youden Index concerns only the cutoff point at which the combination of sensitivity and specicity is maximal. A probability value ≤ 0.05 was considered
sta-tistically signicant. Data analysis was performed with SPSS version 17.
Results
Baseline Characteristics
Of the 142 primary ICH patients enrolled during the study period, 73 (51.4%) met the inclusion criteria and were included in the analysis. Baseline demographic and admis-sion characteristics are outlined in Table 1. Patient charac-teristics were similar to previously reported larger cohorts with ICH.14,15,29,31,32 The mean age of the entire cohort was
62.6 years, 47% were women, and 80% had a history of hypertension. Fifty percent of patients had GCS scores <
8 at admission. Overall, 8% of patients received intrathe-cal tPA, and 53% and 15% underwent EVD placement and hematoma evacuation, respectively.
Radiographic Characteristics
Table 1 lists radiographic characteristics of the study cohort. Mean ICH and IVH volumes were 27±33 ml and
17 ± 22 ml, respectively. Sixty-seven percent had
radio-graphic evidence of hydrocephalus at admission. Median IVH, Graeb, and LeRoux scores were 13, 8, and 6,
respectively. Within 6 days, the incidence of hydrocephalus de -creased to 28% (p = 0.016). Median IVH, Graeb, and Le-Roux scores signicantly decreased to 9 (p = 0.005), 4 (p = 0.002), and 4 (p = 0.003), respectively. Thirty-four patients (46.5%) had an increase in their IVH severity according to IVH, Graeb, and LeRoux scores.
Clinical Outcome and Change in IVH Severity
Twenty-ve patients (34%) died during initial admis-sion. Poor outcome (mRS score ≥ 3) was noted in 55
pa-tients (75%). Increase or decrease in IVH severity within 6 postbleed days was not associated with outcome (p = 0.762 and p = 0.513, respectively).
Comparison of IVH Grading Scales in Predicting Outcome Figure 2 left shows the ROC curves for the admission IVH, Graeb, and LeRoux scores with regard to discharge functional outcome. The 3 scores demonstrated good ac-curacy for predicting outcome. The AUROCs of the IVH (0.745, 95% CI 0.589–0.900), Graeb (0.743, 95% CI 0.601– 0.886) and LeRoux (0.744, 95% CI 0.604–0.884) scores were not statistically different (p = 0.646). The 3 scores were also good at predicting outcome when assessed again within 6 days of admission with similar AUROCs (0.765, 95% CI 0.640–0.891 [IVH] vs 0.722, 95% CI 0.596–0.848 [Graeb] vs 0.723, 95% CI 0.597–0.850 [LeRoux]; Fig. 2 right). Furthermore, the 3 scores had similar maximum Youden Indices and associated sensitivities and specicities when assessed at admission and within 6 days postic -tus (Table 2). Admission IVH score of 9 (IVH volume of approximately 6.0 ml) and 7 (IVH volume of approximate-ly 3.3 ml) were associated with maximum Youden Index, and thus were identied as an optimal cutoff point for poor outcome.
Impact of EVD on IVH Severity and Outcome
At admission, patients who received an EVD had sim-ilar demographic characteristics as compared with those who did not (Table 3). The proportion of patients with a GCS score<8 was signicantly higher in the EVD group
(p = 0.015). The EVD group also had signicantly worse radiographic ndings with higher mean IVH volumes compared with the non-EVD group (23 ± 26 ml vs 9 ±
11 ml, respectively; p = 0.003) and increased incidence of hydrocephalus (82% vs 50%, respectively; p = 0.004). In-traventricular hemorrhage severity, as assessed by IVH (17 vs 8, respectively; p <0.001), Graeb (11 vs 3, respectively;
p < 0.001) and LeRoux scores (7 vs 3, respectively; p <
0.001) were signicantly worse in the EVD group as com-pared with the non-EVD group (Table 3). Within 6 days of admission, the median IVH (17 vs 10; p = 0.001), Graeb (7 vs 5; p <0.001) and LeRoux scores (11 vs 5, p <0.001)
signicantly decreased in the EVD group compared with the admission scores. During the same period, there was no signicant change in median IVH (8 vs 9, p = 0.058), Graeb (3 vs 4, p = 0.182) or LeRoux scores (3 vs 4, p = 0.091) among patients who did not undergo EVD place-ment. The EVD and non-EVD groups had similar postmission IVH, Graeb, and LeRoux scores (Table 3). In ad-dition, there was no signicant difference in the incidence
of poor outcome between the EVD and non-EVD groups (79.5% vs 70.6%, p = 0.251).
Discussion
Comparison of the IVH, Graeb, and LeRoux Scores
In this study, we have demonstrated that the admis-sion IVH, Graeb, and LeRoux scores predict functional outcome at discharge with good accuracy. The 3 scores retained their predictive accuracy when assessed within 6 days postictus, suggesting that they are robust against potentially confounding factors, such as discrepancies in medical or surgical management among patients. Although the scores were not intended for predicting outcome per se, their predictive accuracies reect their ability to assess IVH severity and risk-stratify patients with regard to out-come. Therefore, any 1 of the 3 scores can be considered for IVH evaluation in the setting of ICH and assessment of hemorrhage progression and treatment response, as well as for incorporation into ICH grading scales for improved outcome prediction and risk-stratication.
To our knowledge, this is the rst study to directly compare the 3 most commonly used IVH grading scales for their abilities to predict outcome after ICH with
intra-ventricular extension. With rapidly growing interest in nov-el therapies for IVH in the setting of ICH, simple and ac-curate assessment of IVH is becoming critical for system-atic evaluation and monitoring of disease progression and clinical efcacy of treatments. However, there is currently no widely accepted, standardized method for measuring IVH severity or volume. The recently developed IVH score is different from the Graeb and LeRoux scores in several aspects. First, unlike the Graeb and LeRoux scores, the IVH score was specically developed in a cohort of ICH patients with IVH, and hence, may better account for the characteristics of IVH in this specic population. Further-more, the IVH score may be more sensitive to the differ-ences in severity as it assesses hemorrhage size in thirds rather than halves. The score also considers hydrocephalus, an independent predictor of outcome after ICH,6separately
from IVH volume in each ventricle. This may enable the IVH score to evaluate the severity of IVH with better ac-curacy as compared with the other scales because hemor-rhage burden and ventricular expansion do not necessarily always go hand-in-hand. Nevertheless, in our cohort, the difference in how hydrocephalus is evaluated by each scale did not translate into improved predictive accuracy for the IVH score. Also in our cohort, the potential advantages of the IVH score failed to translate into improved predictive
TABLE 1: Demographic, admission, and radiographic data of the study population according to functional outcome (mRS score) at discharge*
Parameter Overall mRS Score<3 mRS Score≥3 p Value
no.ofpatients 73 18 55
demographic characteristics
mean age (yrs)±SD 62.6±16.6 50.6±16.4 65.4±15.4 0.002†
female 31 (42.5) 6 (33.3) 28 (50.9) 0.756
hx of HTN 58 (79.5) 10 (55.6) 48 (87.3) 0.467 admission characteristics
median GCS score (IQR) 7 (4–13) 15 (14–15) 6 (4–11) <0.001† median ICH score (IQR) 3 (2–3.5) 7 (3–10) 14 (7–17) <0.001† mean ICH volume (ml)±SD 27.35±32.96 13.37±23.79 30.67±34.10 0.077 mean IVH volume (ml)±SD 16.92±21.84 4.41±5.15 19.94±23.25 <0.001† bleeding in lt lateral ventricle 61 (83.6) 8 (44.4) 53 (96.4) 0.008† bleeding in rt lateral ventricle 59 (80.8) 10 (55.6) 49 (89.1) 0.449 bleeding in 3rd ventricle 53 (72.6) 6 (33.3) 47 (85.5) 0.016† bleeding in 4th ventricle 44 (60.3) 5 (27.8) 39 (70.9) 0.037† hydrocephalus 49 (67.1) 5 (27.8) 44 (80.0) 0.005† infratentorial ICH 13 (17.8) 1 (5.6) 12 (21.8) 0.440 mean midline shift (mm)±SD 3.65±4.97 2.07±3.37 4.03±5.23 0.188 median IVH score (IQR) 13 (6–17) 7 (3–10) 14 (7–17) 0.004† median LeRoux score (IQR) 8 (3–12) 3 (2–6) 9 (4–12) 0.005† median Graeb score (IQR) 6 (3–8) 3 (2–4) 6 (4–9) 0.005† hospital course
surgical hematoma evacuation 11 (15.1) 1 (5.6) 10 (18.2) 0.679 EVD placement 39 (53.4) 4 (22.2) 35 (63.6) 0.071
intrathecal tPA 6 (8.2) 0 (0) 6 (10.9) 0.588
* HTN = hypertension; IQR = interquartile range. † Statistically signifcant.
accuracy for outcome. Given the comparable performances of the 3 scores, one important feature that distinguishes the IVH score from the others is its ability to rapidly estimate IVH volume without the need for time-consuming calcu-lations that often involve computer-assisted planimetric analysis. This is important because, once validated as a re-liable and accurate volume estimation tool, the IVH score will allow quick and routine clinical assessment of IVH and total volume (IVH and ICH volume) at admission and during treatment course. This will be useful to researchers as clinical trials move toward novel therapies that address both ICH and IVH volumes in recognition of their inde-pendent effects on outcome.12
Effects of Alterations in IVH Severity and Volume on Outcome
Intraventricular hemorrhage volume is directly related to the risk of poor outcome after ICH, with volumes of >
20 ml often leading to death.24,25,30 In our cohort,
admis-sion IVH volume of 6.0 ml was associated with a signi-cant increase in the likelihood of poor functional outcome. This suggests that even relatively small IVH can lead to clinically signicant ventricular dysfunction, intracranial hypertension, reduced cerebral perfusion, and secondary cerebral injuries.11The lower threshold for poor outcome at
postadmission may be due to the prolonged exposure of the
ventricles to blood, which worsens IVH-mediated injuries and neurological outcome.22,23
Intraventricular hemorrhage is a dynamic condition and its course depends on many factors, including mean arterial pressure, baseline ICH volume, and treatment.25
Nearly half of our cohort experienced an increase in their IVH severity, suggesting that IVH progression is common beyond the rst 24 hours despite standard neurocritical management. Although hydrocephalus resolved and IVH grades improved in a signicant number of patients, the overall mean IVH volume did not signicantly change. Furthermore, increase or decrease in IVH severity with-in 6 days of hemorrhage was not signicantly associated with outcome. This suggests that subacute worsening in IVH volume may not signicantly worsen outcome once patients are medically stabilized and hydrocephalus and intracranial pressure are optimally managed. Surprisingly, a decrease in IVH severity also did not have signicant inuence on outcome. It is possible that IVH either causes a considerable amount of its damage in the early periods, or IVH severity within the rst 24 hours determines the ex-tent of cerebral injuries to come. Studies have reported that early changes in IVH severity may signicantly inuence outcome. Steiner et al.25 have reported that 17% of ICH
patients had IVH expansion within 24 hours of symptom onset and that the growth was associated with severe dis-ability and death. We did not investigate the incidence or degree of IVH growth in the rst 24 hours. Further studies are needed to determine the IVH dynamics and its clinical impact on hospital course and outcome. Continued inves-tigation into the IVH pathophysiology may also help iden-tify an optimal timing and degree of intervention.
Impact of an EVD on Outcome in ICH Patients With IVH Current treatment for IVH in ICH patients is EVD-as-sisted drainage but its usefulness remains controversial.21
In our cohort, patients who underwent EVD placement had signicantly higher IVH volume and incidence of hy-drocephalus at admission with signicantly worse IVH, as assessed by the IVH grading scales. Based on our manage-ment protocol, it is likely that a majority of patients with
Fig. 2. Comparison of ROC curves and AUROCs of admission (left) and postadmission (right) IVH scores, Graeb scores, and
LeRoux scores as predictors of discharge poor functional outcome (mRS score ≥3).
TABLE 2: Performance of the IVH, Graeb, and LeRoux scores for predicting poor outcome (mRS score≥3)
Score Youden Index Sensitivity Specifcity admission
IVH 0.515 0.729 0.786
Graeb 0.477 0.763 0.643
LeRoux 0.440 0.797 0.714
w/in 6 days postictus
IVH 0.515 0.729 0.786
Graeb 0.339 0.339 1
hydrocephalus in the EVD group were symptomatic from ventricular expansion. Interestingly, patients who under-went EVD placement and those who did not had similar IVH severity within 6 days of hemorrhage, suggesting that an EVD was effective at alleviating IVH and ventricular obstruction. Although EVD insertion was not associated with outcome, it is important to recognize that the EVD group had similar mortality rates and functional outcome as those who did not receive an EVD, despite having a signicantly worse prognosis at admission, which is con -sistent with previous reports.6,30Although this prospective
observational study was not designed to isolate the impact of EVD on outcome, our results suggest EVD placement, in conjunction with neurocritical care and appropriate surgical and medical interventions, may be benecial in a subset of ICH patients with IVH with particularly dismal prognosis based on admission characteristics. A random-ized clinical trial is needed to determine whether EVD insertion improves outcome in ICH patients with IVH. Study Limitations
This study has several limitations. First, only the ad-mission scans and the last scan performed within 6 days were used for our study. All the patients received at least 1 scan during this time period but they were not performed on the same postbleed day. It is possible that not all wors-ening or improvement in IVH was included in our analysis.
Second, we estimated the IVH volume based on the IVH score. Although this method has been shown to be accu-rate,11it is novel and is yet to be conclusively validated. In
addition, this study may have been underpowered for dif -ferentiating the predictive accuracies of the scales, as well as detecting small difference among subgroups of patients. Our results must be validated in future studies with larger cohorts and longer follow-up periods that are sufciently powered to control for the heterogeneity of the study popu-lation and treatments.
Conclusions
The IVH score, Graeb score, and LeRoux score as-sessed at admission predict discharge functional outcome in ICH patients with IVH with good accuracy. Therefore, any 1 of the 3 scores may be considered for IVH assess-ment in the setting of ICH or for improving the perfor-mance of ICH grading scales that incorporates IVH se-verity for outcome prediction. Although IVH progression is common after the rst 24 hours even with standard optimal neurocritical care, its impact on clinical outcome needs to be further elucidated. Future research is also needed to better understand IVH pathophysiology, which may help dene optimal timing and degree of interven -tion. Although EVD placement remains one of the main treatments of IVH and acute hydrocephalus, its impact
TABLE 3: Demographic, admission, and outcome data of study population according to EVD status
Parameter NoEVD EVD pValue
no.ofpatients 34 39
demographic characteristics
mean age (yrs)±SD 65.0±16.3 60.4±16.7 0.234
female(%) 14(41.2) 20(51.3) 0.388
hx of HTN (%) 26 (76.5) 32 (82.1) 0.556
admission characteristics
GCS score<8(%) 13(38.2) 26(66.7) 0.015*
infratentorial ICH (%) 5 (14.7) 8 (20.5) 0.518 mean ICH volume (ml)±SD 31.38±36.79 23.85±29.25 0.334 mean IVH volume (ml)±SD 9.21±11.30 23.44±26.24 0.003*
hydrocephalus (%) 17 (50.0) 32 (82.1) 0.004*
median IVH score (IQR) 8 (4–14) 17 (10–17) <0.001* median LeRoux score (IQR) 3 (2–9) 11 (5–14) <0.001* median Graeb score (IQR) 3 (2–6) 7 (5–10) <0.001* postadmission characteristics (within 6 days postictus)
median IVH score (IQR) 9 (6–16) 10 (6–17) 0.701 median LeRoux score (IQR) 4 (2–10) 5 (2–10) 0.446 median Graeb score (IQR) 4 (2–8) 5 (2–7) 0.668 clinical course and outcome (%)
intrathecal tPA 0 6 (15.4) 0.027*
hematoma evacuation 3 (8.8) 8 (20.5) 0.202
in-hospital deaths 10 (29.4) 15 (38.5) 0.416
in-hospital poor outcome (mRS score≥3) 24 (70.6) 31 (79.5) 0.251 * Statistically signifcant.
on clinical outcome remains controversial. Although an EVD was associated with improvement in outcome in patients with severe IVH with particularly poor progno-sis at admission, a randomized clinical trial is needed to conclusively demonstrate its therapeutic benets in ICH patients with IVH.
Disclosure
Brian Y. Hwang was supported by a Doris Duke Clinical Research Fellowship.
Author contributions to the study and manuscript preparation include the following. Conception and design: Appelboom, Hwang, Carpenter, CP Kellner, Ducruet, Meyers, Connolly. Acquisition of data: Bruce, Piazza, Carpenter, Deb-Sen, Vaughan. Analysis and interpretation of data: Appelboom, Hwang, Bruce, Piazza, Carpenter, Gigante, CP Kellner, Deb-Sen, Vaughan. Drafting the article: Appelboom, Hwang, Bruce, Carpenter, Gigante, CP Kellner, M Kellner, Deb-Sen, Vaughan, Connolly. Critically revising the article: all authors. Reviewed submitted version of manuscript: all authors. Approved the final version of the manuscript on behalf of all authors: Appelboom. Statistical analysis: Appelboom, Bruce, Piazza. Administrative/technical/material support: Appelboom, Hwang, Carpenter, CP Kellner, MA Kellner, Deb-Sen. Study supervision: Appelboom, Hwang, Carpenter, CP Kellner, Meyers, Connolly.
References
1. Bhattathi ri PS, Gregson B, Prasad KS, Mendelow AD: Intra-ventricular hemorrhage and hydrocephalus after spontaneous intracerebral hemorrhage: results from the STICH trial. Acta Neurochir Suppl96:65–68, 2006
2. Broderick J, Connolly S, Feldmann E, Hanley D, Kase C, Krieger D, et al: Guidelines for the management of spontaneous intracerebral hemorrhage in adults: 2007 update: a guideline from the American Heart Association/American Stroke Asso-ciation Stroke Council, High Blood Pressure Research Council, and the Quality of Care and Outcomes in Research Interdisci-plinary Working Group. Stroke38:2001–2023, 2007
3. Broderick JP, Brott T, Tomsick T, Miller R, Huster G: Intrace-rebral hemorrhage more than twice as common as subarach-noid hemorrhage. J Neurosurg78:188–191, 1993
4. Cheung RT, Zou LY: Use of the original, modied, or new intracerebral hemorrhage score to predict mortality and mor-bidity after intracerebral hemorrhage. Stroke34:1717–1722, 2003
5. Chuang YC, Chen YM, Peng SK, Peng SY: Risk stratication for predicting 30-day mortality of intracerebral hemorrhage. Int J Qual Health Care21:441–447, 2009
6. Diringer MN, Edwards DF, Zazulia AR: Hydrocephalus: a previously unrecognized predictor of poor outcome from su -pratentorial intracerebral hemorrhage. Stroke29:1352–1357, 1998
7. Ducruet AF, Hickman ZL, Zacharia BE, Grobelny BT, Narula R, Guo KH, et al: Exacerbation of perihematomal edema and sterile meningitis with intraventricular administration of tis-sue plasminogen activator in patients with intracerebral hem-orrhage. Neurosurgery66:648–655, 2010
8. Gianni ni EG, Zaman A, Kreil A, Floreani A, Dulbecco P, Tes-ta E, et al: Platelet count /spleen diameter ratio for the noninva-sive diagnosis of esophageal varices: results of a multicenter, prospective, validation study. Am J Gastroenterol 101: 2511–2519, 2006
9. Godoy DA, Piñero G, Di Napoli M: Predicting mortal ity in spontaneous intracerebral hemorrhage: can modication to original score improve the prediction? Stroke37:1038–1044, 2006
10. Graeb DA, Robertson WD, Lapointe JS, Nugent RA, Harr ison
PB: Computed tomographic diagnosis of intraventricular hem-orrhage. Etiology and prognosis. Radiology143:91–96, 1982 11. Hallevi H, Dar NS, Barreto AD, Morales MM, Marti n-Schild
S, Abraham AT, et al: The I VH score: a novel tool for estimat-ing intraventricular hemorrhage volume: clinical and research implications. Crit Care Med 37:969–974, e1, 2009
12. Hanley DF: Intraventricular hemorrhage: severity factor and treatment target in spontaneous intracerebral hemorrhage. Stroke40:1533–1538, 2009
13. Hanley JA, McNeil BJ: A method of comparing the areas under receiver operating characterist ic curves derived from the same cases. Radiology148:839–843, 1983
14. Hemphill JC III, Bonovich DC, Besmertis L, Manley GT, Johnston SC: The ICH score: a simple, reliable grading scale for intracerebral hemorrhage. Stroke32:891–897, 2001 15. Hemphill JC III, Farrant M, Neill TA Jr: Prospective validation
of the ICH Score for 12-month fu nctional outcome. Neurology 73:1088–1094, 20 09
16. Kothari RU, Brott T, Broderick JP, Barsan WG, Sauerbeck LR, Zuccarello M, et al: The ABCs of measuring intracerebral hemorrhage volumes. Stroke27:1304–1305, 1996
17. LeRoux PD, Haglund MM, Newell DW, Grady MS, Winn HR: Intraventricular hemorrhage in blunt head trauma: an analysis of 43 cases. Neurosurgery31:678–685, 1992
18. Mayfrank L, Kim Y, Kissler J, Delsing P, Gilsbach JM, Schröder JM, et al: Morphological changes following exper-imental intraventricular haemorrhage and intraventricular brinolytic treatment with recombinant tissue plasminogen activator. Acta Neuropathol100:561–567, 2000
19. Morgenstern LB, Hemphill JC III, Anderson C, Becker K, Broderick JP, Connolly ES Jr, et al: Guidelines for the management of spontaneous intracerebral hemorrhage: a guide -line for healthcare professionals from the American Heart Association/American Stroke Association. Stroke 41:2108– 2129, 2010
20. Nieuwkamp DJ, de Gans K, Rinkel GJ, Algra A: Treatment and outcome of severe intraventricular extension in patients with subarachnoid or intracerebral hemorrhage: a systematic review of the literature. J Neurol247:117–121, 2000
21. Nishikawa T, Ueba T, Kajiwara M, Miyamatsu N, Yamashita K: A priority treatment of the intraventricular hemorrhage (IVH) should be performed in the patients suffering intracerebral hemorrhage with large IVH. Clin Neurol Neurosurg 111: 450–453, 2009
22. Pang D, Sclabassi RJ, Horton JA: Lysis of intraventricular blood clot with urokinase in a canine model: Part 2. In vivo safety study of intraventricular urokinase. Neurosurgery19: 547–552, 1986
23. Pang D, Sclabassi RJ, Horton JA: Lysis of intraventricular blood clot with urokinase in a canine model: Part 3. Effects of intraventricular urokinase on clot lysis and posthemorrhagic hydrocephalus. Neurosurgery19:553–572, 1986
24. Ruscalleda J, Peiró A: Prognostic factors in intraparenchyma-tous hematoma with ventricular hemor rhage. Neuroradiology 28:34–37, 1986
25. Steiner T, Diringer MN, Schneider D, Mayer SA, Begtrup K, Broderick J, et al: Dynamics of intraventricular hemorrhage in patients with spontaneous intracerebral hemorrhage: risk factors, clinical impact, and effect of hemostatic therapy with recombinant activated factor VII. Neurosurgery59:767–774, 2006
26. Tuhri m S, Dambrosia JM, Price TR, Mohr JP, Wolf PA, Hier DB, et al: Intracerebral hemorrhage: external validation and extension of a model for prediction of 30-day survival. Ann Neurol29:658–663, 1991
27. Tuhrim S, Horowitz DR, Sacher M, Godbold JH: Volume of ventricular blood is an important determinant of outcome in supratentorial intracerebral hemorrhage. Crit Care Med27: 617–621, 1999
28. Wagner KR, Xi G, Hua Y, Zuccarello M, de Cour ten-Myers GM, Broderick JP, et al: Ultra-early clot aspiration after ly-sis with tissue plasminogen activator in a porcine model of intracerebral hemorrhage: edema reduction and blood-brain barrier protection. J Neurosurg90:491–498, 1999
29. Weimar C, Benemann J, Diener HC: Development and valida-tion of the Essen Intracerebral Haemorrhage Score. J Neurol Neurosurg Psychiatry77:601–605, 2006
30. Young WB, Lee KP, Pessin MS, Kwan ES, Rand WM, Caplan LR: Prognostic signicance of ventricular blood in supratento -rial hemorrhage: a volumetric study. Neurology40:616–619, 1990
31. Zahuranec DB, Gonzales NR, Brown DL, Lisabeth LD, Lon-gwell PJ, Eden SV, et al: Presentation of int racerebral haemor-rhage in a community. J Neurol Neurosurg Psychiatry 77: 340–344, 2006
32. Zahuranec DB, Morgenstern LB, Sánchez BN, Resnicow K, White DB, Hemphill JC III: Do-not-resuscitate orders and pre-dictive models after intracerebral hemorrhage. Neurology75: 626–633, 2010
Manuscript submitted May 31, 2010. Accepted September 14, 2011.
Current affiliation for Dr. Hwang: Department of Neurosurgery, Johns Hopkins University School of Medicine, Baltim ore, Maryland. Please include this information when citing this paper: published online October 14, 2011; DOI: 10.3171/2011.9.JNS10850.
Address correspondence to:Geoffrey Appelboom, M.D., Depart-ment of Neurosurgical Surgery, Columbia University Medical Cen-ter, New York, New York 10032. email: [email protected].