Journal of the Neurological Sciences 296 (2010) 712

Contents lists available at ScienceDirect

Journal of the Neurological Sciences

j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / j n s

Changes in the volumes of the brain and cerebrospinal uid spaces after shunt surgery in idiopathic normal-pressure hydrocephalus
Kotaro Hiraoka a,, Hiroshi Yamasaki a, Masahito Takagi a, Makoto Saito a, Yoshiyuki Nishio a, Osamu Iizuka a, Shigenori Kanno a, Hirokazu Kikuchi a, Takeo Kondo b, Etsuro Mori a
a b

Department of Behavioral Neurology and Cognitive Neuroscience, Tohoku University Graduate School of Medicine, 2-1 Seiryo-machi, Aoba-ku, Sendai 980-8575, Japan Department of Physical Medicine and Rehabilitation, Tohoku University Graduate School of Medicine, 2-1 Seiryo-machi, Aoba-ku, Sendai 980-8575, Japan

a r t i c l e

i n f o

a b s t r a c t
Objectives: To investigate volumetric changes of the brain and cerebrospinal uid (CSF) spaces after shunt surgery in shunt-responsive idiopathic normal-pressure hydrocephalus (iNPH), and correlations between the changes and postoperative clinical improvements. Methods: Twenty-one patients with shunt-responsive iNPH were studied. Magnetic resonance imaging (MRI) of the brain was performed before and 1 year after surgery, and clinical symptoms were assessed by the iNPH Grading Scale, a validated assessment tool of the triad of iNPH, the Modied Rankin Scale, the Timed Up and Go Test, and neuropsychological tests including the Mini-Mental State Examination. The volumes of the left cerebral hemisphere, infratentorial brain, ventricles, and suprasylvian and infrasylvian subarachnoid CSF spaces were measured using an MRI-based volumetric technique. Results: The volumes of the cerebral hemisphere and infratentorial brain did not change signicantly after shunt surgery (p = 0.231, 0.109, respectively). The volumes of the ventricles and infrasylvian subarachnoid CSF spaces were signicantly decreased (p b 0.0001, b 0.05, respectively), with a mean change rate of 26.1% and 4.5%, respectively. The volumes of the suprasylvian subarachnoid CSF spaces increased signicantly (p b 0.0001), with a mean change rate of 43.5%. The decrease in ventricular volumes was signicantly correlated with clinical improvement. 2010 Elsevier B.V. All rights reserved.

Article history: Received 7 December 2009 Received in revised form 22 June 2010 Accepted 23 June 2010 Keywords: Idiopathic normal-pressure hydrocephalus Cerebrospinal uid shunt Magnetic resonance imaging Volumetry

1. Introduction Idiopathic normal-pressure hydrocephalus (iNPH) is a syndrome manifesting as a triad of dementia, gait disturbance, and urinary impairment, with ventricular enlargement and normal cerebrospinal uid (CSF) pressure, and without preceding events such as subarachnoid hemorrhage and meningitis. It is characterized by clinical improvement following shunt placement [14]. Although the cause and pathophysiology of iNPH remain unclear, the conventional view is that it is due to obstruction of CSF circulation and/or absorption. The effectiveness of shunt placement for the treatment of iNPH has been established [14], but the mechanism of its effect has not yet been claried. Radiological studies have led to clarication of the morphological features of iNPH, and dilation of the ventricles and sylvian ssures as well as narrowing of the subarachnoid CSF spaces on the high convexity and interhemispheric ssure is seen on magnetic resonance imaging (MRI) [5]. However, there have been few studies investigating morphometric changes of the intracranial components

Corresponding author. Tel.: + 81 22 717 7358; fax: + 81 22 717 7360. E-mail address: khiraoka@mail.tains.tohoku.ac.jp (K. Hiraoka). 0022-510X/$ see front matter 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.jns.2010.06.021

after shunt surgery. Kitagaki et al. showed that the mean postoperative CSF volume of ve patients with iNPH was signicantly decreased in the sylvian space and in the ventricle, marginally decreased in the basal cistern, and signicantly increased in the suprasylvian space as compared with the preoperative volume [5]. In the study by Anderson et al. [6], the ventricular volumes of 10 (91%) out of 11 iNPH patients who underwent shunt surgery were decreased, with a mean change rate of 39%. Volume changes of the brain and the association between changes in volume of the intracranial components and clinical outcomes after shunt surgery remain to be claried. In this study, we investigated volumetric changes of the brain and CSF spaces and their association with clinical outcomes after shunt surgery. The CSF spaces were segmented into the ventricles and suprasylvian and infrasylvian subarachnoid spaces, as previous study [5] showed volume decrease of the ventricles, sylvian space, and basal cistern, and volume increase of the suprasylvian subarachnoid spaces. We assumed that cerebral volume increases postoperatively along with the clinical improvements and as the offset of ventricular volume decreases after shunt surgery. In addition, we expected to nd a correlation between the volumetric changes of the intracranial components and clinical improvement after shunt surgery. We aimed to elucidate the pathophysiology of

K. Hiraoka et al. / Journal of the Neurological Sciences 296 (2010) 712

iNPH and the mechanism of shunt effect by precisely measuring morphometric changes after shunt surgery in patients with denite iNPH. 2. Methods 2.1. Subjects and study design Patients who had been diagnosed as having iNPH and who had undergone shunt surgery between March 2005 and November 2007 were enrolled in a prospective follow-up program. The patients were diagnosed by neurologists as probable iNPH based on the diagnostic criteria published by the Guidelines Committee of Idiopathic Normal Pressure Hydrocephalus, the Japanese Society of Normal Pressure Hydrocephalus [7], i.e., (1) individuals who develop symptoms in their 60 s or older, (2) the presence of more than one of the triad of gait disturbance, cognitive impairment, and urinary incontinence, (3) MRI features of iNPH (i.e., both ventricular dilation (Evans Index N 0.3) and narrowing of the subarachnoid CSF spaces on the high convexity and interhemispheric ssure), (4) CSF pressure of 200 mm H2O or less and normal CSF content, (5) positive CSF tap test, (6) clinical symptoms that cannot be completely explained by other neurological or non-neurological diseases, and (7) no obvious preceding diseases that could possibly cause ventricular dilation, including subarachnoid hemorrhage, meningitis, head injury, congenital hydrocephalus, and aqueductal stenosis. Routine laboratory tests and single photon emission computed tomography (SPECT) of the brain were performed for differential diagnosis, and MRI and clinical evaluations were performed as a baseline assessment. After conrmation that there was no contraindication to the surgical procedure, patients who fullled the criteria underwent ventriculoperitoneal (VP) or lumboperitoneal (LP) shunt surgery. After surgery, the patients were followed up for 1 year, and the pressure settings of their programmable valves were adjusted in the outpatient clinic. One year after surgery, the patients were re-admitted to the hospital for re-assessment of their clinical symptoms and for radiological investigations. In this study, only baseline and 1-year follow-up data were used. Only data from those patients who completed the 1-year follow-up program and who had denite iNPH (i.e., those who showed clinical improvements as described below) were included in order to avoid contamination of the results with data from patients with unspecied diagnosis. Clinical improvement after shunt surgery was dened as a decrement of more than 1 point in total score on the iNPH Grading Scale (iNPHGS) [8] compared to the baseline score. The iNPHGS is a validated tool for assessment of the clinical triad. It is rated according to the clinician's observation and interview with the patient and his or her caregiver in order separately to assess the severity of each component of the triad. The score for each domain ranges from 0 to 4, and higher scores indicate worse symptoms. Patients who dropped out from the 1-year follow-up program and patients who did not show clinical improvements 1 year after surgery were excluded from volumetry. The procedure followed the clinical study guidelines of the Ethics Committee of Tohoku University Hospital and was approved by the Internal Review Board. Written informed consent was obtained from the patients or their families. 2.2. Clinical assessment Clinical symptoms were assessed at baseline and 1 year after shunt surgery. In addition to the iNPHGS, the modied Rankin Scale (mRS) [9], Timed Up and Go (TUG) test [10], Mini-Mental State Examination (MMSE) [11], Frontal Assessment Battery (FAB) [12], and the modied version of the Neuropsychiatric Inventory (NPI) [1315] were administered. In the TUG test, patients are timed while they rise

from an arm chair, walk 3 m, turn, walk back, and sit down again, and the number of steps is counted. 2.3. Shunt surgery and programmable valve adjustments The operative procedure either VP shunt or LP shunt was selected according to the condition of the patient's cervical and lumbar vertebrae when investigated by spinal MRI and the preference of the patient and his or her family. In cases of VP shunt, a catheter was placed in the anterior horn of the right lateral ventricle, and in cases of LP shunt surgery, a catheter was placed in the lumbar subarachnoid space. The abdominal-side catheter was guided to the abdomen subcutaneously and inserted in the peritoneal cavity. We used the Codman-Hakim programmable valve system in all shunt surgery. The initial pressure for the shunt system was set before surgery, based on the patient's height and weight [16,17]. Postoperatively, patients were followed up in outpatient clinics; the pressure settings of their programmable valves were adjusted step by step and clinical improvements and adverse effects were noted. If clinical improvement was absent or insufcient, the pressure setting was lowered by 1030 mm H2O over a period of 1 2 weeks. If orthostatic headache or subdural effusion/hematoma was observed by computed tomography (CT), the pressure setting was increased by 30 mm H2O. Pressure adjustments were repeated until optimal pressure for each patient was attained. 2.4. MR volumetry Brain MRI was performed at baseline and 1 year after surgery. Axial, three-dimensional spoiled gradient echo (SPGR) images were obtained for volumetry. The images were generated with a 1.5-T MRI unit (Signa Horizon LX CV/i; GE Healthcare, Milwaukee, WI). Operating parameters were as follows: eld of view, 250 mm; matrix, 256 256; contiguous sections, 108 1.5 mm; repetition time, 20 ms; echo time, 4.1 ms; and ip angle, 30. Axial T2-weighted images and uid-attenuated inversion recovery (FLAIR) images were also obtained for diagnosis. The volumes of the left cerebral hemisphere, infratentorial brain, ventricles, and subarachnoid CSF spaces on the left side were measured on the MR image (Fig. 1). The reason for not measuring the right cerebral hemisphere and subarachnoid space on the right side was that there were artefacts caused by the shunt valve on the postoperative images in the cases of VP shunt surgery. The subarachnoid CSF spaces were segmented into two parts suprasylvian and infrasylvian as mentioned below. The MRI data sets of all images were transmitted to a personal computer from the MRI unit. Measurements were performed by one investigator (K.H.), who was blinded to (1) all clinical information, (2) the order (pre-surgery and post-surgery) of MRIs, and (3) the time of enrolment. The right side of cerebral convexity and circumferential tissues such as dura mater, cranium, subcutaneous fat, and scalp were masked in both the pre-surgery and post-surgery images, sparing the right lateral ventricle for volumetry, so that the investigator could not discriminate by means of artefacts between the pre-surgery and postsurgery images in cases of VP shunt surgery. The MRI data sets were analyzed using ImageJ 1.37 (NIH, Washington, USA), based on built-in functions [18]. For volumetry, we used a combination of semiautomatic segmentation technique through density thresholding and manual tracing, thereby avoiding partial voluming and observer bias. The volume of each structure was obtained by automatically counting the number of pixels within the segmented regions and then multiplying the number by voxel size (0.982 1.50 = 1.44 mm3). The reliability and validity of this method have been established and described elsewhere [19]. The left cerebral hemisphere and infratentorial brain were outlined and the ventricles were extracted with surrounding brain parenchyma by tracing with a manually driven mouse cursor. The boundary between the cerebral hemispheres and infratentorial brain

K. Hiraoka et al. / Journal of the Neurological Sciences 296 (2010) 712

Fig. 1. Segmentation of the intracranial components. The intracranial components were segmented into the left cerebral hemisphere, infratentorial brain (both sides), bilateral ventricles, and suprasylvian and infrasylvian subarachnoid CSF spaces (left side).

was the plane of the intercollicular level of the midbrain. The boundaries between the right and left hemispheres were traced manually in the middle of the corpus callosum, septum pellucidum, body of fornix, and anterior and posterior commissures. The ventral boundary of the infratentorial brain was the plane including the apex of the odontoid process. The ventricles included the bilateral lateral ventricles, third ventricle, cerebral aqueduct of Sylvius, and fourth ventricle. For the segmentation of the subarachnoid CSF spaces, the ventricles were blacked out from the original MR images. The external boundary of the subarachnoid CSF spaces was outlined by tracing the dura mater with a manually driven mouse cursor. The boundary between the right and left halves of the subarachnoid spaces was determined on the basis of landmarks such as the falx cerebri and cerebral aqueduct of Sylvius. The subarachnoid CSF spaces were segmented into upper and lower parts along the intermediate plane between the plane through the nasion and the external occipital protuberance and the cranial top point of the dura mater, so that the upper part included the subarachnoid CSF spaces above the sylvian ssure, and the lower part included the subarachnoid CSF spaces in the sylvian ssure and below it. Subsequently, each intracranial component was automatically segmented on the extracted images using density threshold set at a range between minimum and maximum pixel values. For volumetry of the cerebral hemisphere and infratentorial brain, the maximum value was the largest pixel value of the brain, and the minimum value was half the mean pixel value of the gray matter (the caudate head) and the mean value of the CSF, since the gray matter and CSF are predominant constituents of the brainextrabrain interface. For volumetry of the ventricles and subarachnoid CSF spaces, the maximum value was half the value of the mean pixel value of the gray matter (the caudate head) and the mean value of the CSF, and the minimum value was the minimum pixel value of the CSF (lateral ventricles). The testretest reliabilities of the volumetry were expressed as intraclass correlation coefcients (ICCs), which were calculated from MRIs repeated after a 1-week interval and volumetry by a single rater (K.H.) under blind conditions for 10 subjects. The ICCs for the cerebral hemisphere, infratentorial brain, ventricles, and suprasylvian and infrasylvian subarachnoid CSF spaces were 0.987, 0.992, 0.995, 0.987, and 0.984, respectively. Coefcients of variation (standard deviation/mean, where standard deviation indicates the square-root value of the arithmetic mean of 10 variance estimates) were 0.662%, 0.775%, 1.593%, 1.720%, and 1.661% for the cerebral hemisphere, infratentorial brain, ventricles, and suprasylvian and infrasylvian subarachnoid CSF spaces, respectively.

2.5. Statistical analysis The volumetric change rate of each intracranial component was dened as the percentage change, which was calculated as postoperative volume minus baseline volume divided by baseline volume (100). The changes of iNPHGS, mRS, MMSE, FAB, and modied NPI were calculated by subtracting baseline scores from postoperative scores. The change rates of time and steps in TUG test were calculated as postoperative value minus baseline value divided by baseline value (100). Two-tailed Student's t test was used for comparison of volume change rates of each intracranial component between VP shunt cases and LP shunt cases. Paired, two-tailed Student's t test was used for comparison of baseline and postoperative volumes of each intracranial component. Spearman's correlation was used for exploratory correlation analysis between volumetric change rates of intracranial components and change (rate) of clinical assessments. All statistical analyses were performed on the statistical software package SPSS, version 11.0.1 J (SPSS Inc., Chicago, IL). The statistically signicant level was set at p b 0.05. The signicance level for multiple comparisons was not corrected because of the explorative nature of this study. 3. Results During the study period, 38 patients with probable iNPH underwent shunt surgery. Of these 38 patients, 12 dropped out due to: ischemic stroke (n = 3), malignancy (n = 2), severe infectious disease (n = 2), chronic subdural hematoma (n = 2), withdrawal of consent (n = 1), institutionalization (n = 1), and femoral fracture (n = 1). Therefore 26 patients completed the 1-year follow-up program and were re-assessed 1 year after surgery. Of these 26 patients, 21 showed clinical improvements and were selected for

Table 1 Demographic characteristics of the patients (n = 21). Sex (male/female) Operative procedure (VP/LP) 8/13 15/6 Mean Age at baseline (years) Education (years) Duration of disease (years) Interval between operation and postoperative MRI (months) 76.2 9.7 3.1 12.8 SD 3.6 2.9 1.4 0.8

VP, ventriculoperitoneal; LP, lumboperitoneal; MRI, magnetic resonance imaging; SD, standard deviation.

10

K. Hiraoka et al. / Journal of the Neurological Sciences 296 (2010) 712 Table 4 Volumetric results for the patients (n = 21). p value 0.001 0.030 0.022 0.002 b 0.001 b 0.001 b 0.001 b 0.001 b 0.001 b 0.001 Baseline Mean SD Cerebral hemisphere Infratentorial brain Ventricles Subarachnoid Suprasylvian CSF spaces Infrasylvian Total 446.3 131.6 Post-op Mean SD 0.231 0.109 p value Change rate (%) Mean 0.6 1.3 SD 2.2 3.5

Table 2 Results of clinical assessment at baseline and 1 year after surgery (n = 21). Baseline Mean TUG test MMSE FAB iNPHGS Time (s) Steps (/30) (/18) Gait (/4) Cognition (/4) Urination (/4) Total (/12) (/6) (/144) 20.8 27.5 20.5 8.9 2.5 2.6 1.9 6.9 3.0 9.1 SD 11.0 12.9 5.3 2.9 0.6 0.8 1.1 1.9 0.9 6.6 Post-op Mean 14.3 22.3 23.0 11.5 1.7 1.8 0.8 4.1 2.0 4.3 SD 6.8 8.6 6.2 3.7 0.9 1.0 0.9 2.2 0.9 3.9

54.4 443.8 57.0 12.5 133.2 13.4

124.1 24.1 33.5 11.8 103.1 838.5

90.8 25.9 b 0.0001 26.1 19.7 44.8 9.5 b 0.0001 43.5 38.8 4.5 3.4 9.7 3.2

mRS Modied NPI

18.2 97.5 13.6 b 0.05 89.9 810.2 89.0 b 0.001

Post-op, postoperative results; SD, standard deviation; TUG test, Timed Up and Go test; MMSE, Mini-Mental State Examination; FAB, Frontal Assessment Battery; iNPHGS, iNPH Grading Scale; mRS, modied Rankin Scale; NPI, Neuropsychiatric Inventory.

Unit: cm3. SD, standard deviation; post-op, postoperative results; CSF, cerebrospinal uid. The total indicates the sum of volumes of the cerebral hemisphere, infratentorial brain, ventricles, and subarachnoid CSF spaces.

volumetry. All these 21 patients were right-handed. Their demographic characteristics are summarized in Table 1, and the results of their clinical assessments at baseline and 1 year after shunt surgery are shown in Table 2. Initially, we analyzed the volumetric changes of VP cases and LP cases separately (Table 3). As both cases showed similar volumetric changes, VP cases and LP cases were brought together for further analysis. Volumetric results for the patients are shown in Table 4. The volumes of the cerebral hemisphere and infratentorial brain did not change signicantly after shunt surgery (p = 0.231 and 0.109, respectively). The ventricular volumes decreased signi cantly (p b 0.0001) after shunt surgery, with a mean change rate of 26.1% (range 5.9% to 82.0%). There was a signicant increase in suprasylvian subarachnoid CSF spaces (p b 0.0001), with a mean change rate of 43.5%, and the infrasylvian subarachnoid CSF spaces showed a signicant decrease (p b 0.05), with a mean change rate of 4.5%. The total volumes of the cerebral hemisphere, infratentorial brain, ventricles, and suprasylvian and infrasylvian subarachnoid CSF spaces showed a signicant decrease (p b 0.001), with a mean change rate of 3.4%. The results of the correlation analysis between volumetric changes and changes in clinical assessments are shown in Table 5. The change rates of the ventricular volume showed a signicant correlation with changes in scores of the MMSE, FAB, scores for the gait domain in the iNPHGS, and the mRS, and change rates related to time in the TUG test (p b 0.05). The correlations indicated that the patients whose ventricular volume decreased most showed more clinical improvements than the patients with less substantial ventricular volume decrease. Change rates in the volume of the cerebral hemisphere, infratentorial brain, and suprasylvian and infrasylvian subarachnoid CSF spaces did not show a signicant correlation with changes in clinical parameters.
Table 3 Volumetric change rates for patients following VP shunt surgery (n = 15) and LP shunt surgery (n = 6). Change rate (%) VP shunt surgery (n = 15) Mean Cerebral hemisphere Infratentorial brain Ventricles Subarachnoid CSF spaces 0.9 1.0 27.5 41.2 4.4 SD 2.3 3.7 21.4 44.2 10.6 LP shunt surgery (n = 6) Mean 0.1 1.9 22.5 49.4 5.0 SD 2.0 3.2 15.6 21.8 7.7 0.35 0.63 0.61 0.67 0.90 p

4. Discussion In this study, we did not include subjects in whom no improvement was noted after shunt surgery. While the exclusion of patients with iNPH who showed no improvement because of treatment failure or other problems possibly causes a selection bias, the exclusion of those with an ambiguous diagnosis (including misdiagnosis and comorbidity of other diseases) makes it possible to obtain uncontaminated data from patients with a denite diagnosis, which is suitable for the pathophysiological study of iNPH. Initially, we hypothesized that the volume of the brain increases after shunt surgery as the symptoms of iNPH improve and ventricular volume decreases. However, the results revealed that the volumes of the cerebral hemisphere and infratentorial brain did not change signicantly after shunt surgery. Previous longitudinal MRI studies of Alzheimer's disease have indicated that the 1-year whole-brain volume decrease was 0.98% to 2.8% [20,21], whereas the 1-year brain volume decrease in normal aging was 0.4% to 0.45% [20,22]. In this study, the rate of hemispheric volume change after shunt surgery was compatible with the 1-year brain volume decrease in normal aging. The results suggest that the volumes of the cerebral hemisphere and infratentorial brain do not increase after shunt surgery. The volume of suprasylvian subarachnoid CSF increased, whereas the volumes of the infrasylvian subarachnoid CSF spaces and ventricles decreased signicantly. If the CSF in each part is proportionally drained by shunt surgery, the CSF volumes in both parts should have been reduced. The contrasting changes may suggest the preoperative existence of a pressure gradient, i.e., the pressure of the ventricles and infrasylvian subarachnoid CSF spaces is likely to be higher than that of the suprasylvian subarachnoid CSF spaces. Shunt surgery may relieve the pressure gradient, evacuate CSF from the ventricles and the infrasylvian subarachnoid CSF spaces, and consequently increase the CSF in the suprasylvian subarachnoid spaces. The existence of a transmantle pressure gradient in iNPH, which is a pressure gradient between the ventricles and subarachnoid CSF spaces, has not yet been established, as it has been found in some studies [23,24] and not in others [25]. However, the ndings of this study support the existence of such a pressure gradient. The pressure gradient is against Pascal's principle, which states that pressure exerted anywhere in a conned uid is transmitted equally in all directions throughout the uid. However, the principle may not be applicable in cases in which CSF ows through the complex subarachnoid spaces synchronized with heart beats. The total volume of cerebral hemisphere, infratentorial brain, ventricles, and subarachnoid CSF spaces decreased signicantly after shunt surgery. This may seem anomalous considering that the total intracranial volume is unalterable in the elderly. The decrease of the

Suprasylvian Infrasylvian

Two-tailed Student's t test. VP, ventriculoperitoneal; LP, lumboperitoneal; SD, standard deviation; CSF, cerebrospinal uid.

K. Hiraoka et al. / Journal of the Neurological Sciences 296 (2010) 712 Table 5 Correlation matrix contrasting volumetric change rates against clinical changes and change rate (n = 21). (Change rate) TUG test Time (Change rate) Cerebral hemisphere Infratentorial brain Ventricles Subarachnoid CSF spaces 0.04 0.12 0.45 0.03 0.03 Steps 0.12 0.00 0.47 0.25 0.15 0.10 0.11 0.49 0.02 0.04 0.24 0.13 0.58 0.10 0.31 (Change) MMSE FAB iNPHGS Gait 0.02 0.03 0.48 0.16 0.01 Cognition 0.20 0.12 0.35 0.20 0.16 Urination 0.25 0.08 0.32 0.13 0.15 Total 0.02 0.01 0.50 0.05 0.02 0.09 0.05 0.46 0.02 0.18 mRS

11

Modied NPI

Suprasylvian Infrasylvian

0.06 0.04 0.26 0.24 0.27

Spearman's correlation, p b 0.05. TUG test, Timed Up and Go test; MMSE, Mini-Mental State Examination; FAB, Frontal Assessment Battery; iNPHGS, iNPH Grading Scale; mRS, modied Rankin Scale; NPI, Neuropsychiatric Inventory. Change rate (%) = (postoperative value preoperative value) / preoperative value 100. Change = postoperative value preoperative value.

total volume may be explained by volume increase of the venous sinuses in the cranium. In most of our cases, expansion of the crosssection of the sagittal sinus after shunt surgery was observed by visual assessment of MRIs, although the volume of the venous sinuses was not measured in this study. The results were similar whether the catheters were inserted in the lateral ventricles (VP shunt surgery) or in the lumbar subarachnoid space (LP shunt surgery), which indicates that the shunt effect on CSF dynamics is the same whether CSF is drained from the lateral ventricles or lumbar subarachnoid space. Relief of the abnormal pressure gradient by shunt surgery altered the distribution of CSF in the cranium, and deformation of the cerebral hemisphere caused by the pressure of CSF in the ventricles was reduced, which may contribute to the improvement in symptoms. The clinical improvements after shunt surgery correlated with the decrease of ventricular volumes. As postoperative clinical assessments and MRI were performed in a single time point for each patient, it cannot be concluded that the clinical symptoms improved in proportion to the ventricular volume decrease in individual patients. However, the correlation may suggest an association between the ventricular volume decrease and improvement of the symptoms. With reference to the pathophysiology of iNPH, the correlation may also suggest some association between the ventricular enlargement and the manifestation of symptoms. Some previous studies have raised concerns about the role of frontal lobe dysfunction in the pathophysiology of iNPH [2632]. A study of iNPH by Momjian et al. showed autoregulation disturbance of cerebral blood ow in paraventricular white matter, which may be caused by the abnormal pressure of the CSF in the ventricles [33]. The abnormal pressure of the CSF in the ventricles towards the cerebrum deforms the cerebrum, which probably leads to neural dysfunction, especially of the frontal lobe. In conclusion, the ndings in this study suggest the existence of a transmantle pressure gradient, which enlarges the ventricles, deforms the cerebral hemispheres, and causes symptoms in iNPH. Shunt surgery may relieve the abnormal pressure gradient and reduce the deformation of the cerebrum, which may contribute to the improvement in symptoms.

Acknowledgment This work was partly supported by a Research Grant from the Ministry of Health, Labour and Welfare of Japan (2008-Nanchi-17).

References
[1] Boon AJ, Tans JT, Delwel EJ, Egeler-Peerdeman SM, Hanlo PW, Wurzer HA, et al. The Dutch normal-pressure hydrocephalus study. How to select patients for shunting? An analysis of four diagnostic criteria. Surg Neurol 2000;53:2017. [2] Raftopoulos C, Deleval J, Chaskis C, Leonard A, Cantraine F, Desmyttere F, et al. Cognitive recovery in idiopathic normal pressure hydrocephalus: a prospective study. Neurosurgery 1994;35:397404. [3] Weiner HL, Constantini S, Cohen H, Wisoff JH. Current treatment of normal-pressure hydrocephalus: comparison of ow-regulated and differential-pressure shunt valves. Neurosurgery 1995;37:87784. [4] Krauss JK, Droste DW, Vach W, Regel JP, Orszagh M, Borremans JJ, et al. Cerebrospinal uid shunting in idiopathic normal-pressure hydrocephalus of the elderly: effect of periventricular and deep white matter lesions. Neurosurgery 1996;39:2929. [5] Kitagaki H, Mori E, Ishii K, Yamaji S, Hirono N, Imamura T. CSF spaces in idiopathic normal pressure hydrocephalus: morphology and volumetry. AJNR Am J Neuroradiol 1998;19:127784. [6] Anderson RC, Grant JJ, de la Paz R, Frucht S, Goodman RR. Volumetric measurements in the detection of reduced ventricular volume in patients with normal-pressure hydrocephalus whose clinical condition improved after ventriculoperitoneal shunt placement. J Neurosurg 2002;97:739. [7] Ishikawa M, Hashimoto M, Kuwana N, Mori E, Miyake H, Wachi A, et al . Guidelines for management of idiopathic normal pressure hydrocephalus. Neurol Med Chir (Tokyo) 2008;48:S1S23 Suppl. [8] Kubo Y, Kazui H, Yoshida T, Kito Y, Kimura N, Tokunaga H, et al. Validation of grading scale for evaluating symptoms of idiopathic normal-pressure hydrocephalus. Dement Geriatr Cogn Disord 2008;25:3745. [9] Rankin J. Cerebral vascular accidents in patients over the age of 60. III Diagnosis and treatment Scott Med J 1957;2:25468. [10] Podsiadlo D, Richardson S. The timed Up & Go: a test of basic functional mobility for frail elderly persons. J Am Geriatr Soc 1991;39:1428. [11] Folstein MF, Folstein SE, McHugh PR. Mini-mental state. A practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res 1975;12: 18998. [12] Dubois B, Slachevsky A, Litvan I, Pillon B. The FAB, a Frontal Assessment Battery at bedside. Neurology 2000;55:16216. [13] Cummings JL, Mega M, Gray K, Rosenberg-Thompson S, Carusi DA, Gornbein J. The Neuropsychiatric Inventory: comprehensive assessment of psychopathology in dementia. Neurology 1994;44:230814. [14] Hirono N, Mori E, Ikejiri Y, Imamura T, Shimomura T, Hashimoto M, et al. Japanese version of the Neuropsychiatric Inventorya scoring system for neuropsychiatric disturbance in dementia patients. No To Shinkei 1997;49:26671. [15] Mori S, Mori E, Iseki E, Kosaka K. Efcacy and safety of donepezil in patients with dementia with Lewy bodies: preliminary ndings from an open-label study. Psychiatry Clin Neurosci 2006;60:1905. [16] Miyake H, Ohta T, Kajimoto Y, Nagao K. New concept for the pressure setting of a programmable pressure valve and measurement of in vivo shunt ow performed using a microowmeter. Technical note J Neurosurg 2000;92:1817. [17] Miyake H, Kajimoto Y, Tsuji M, Ukita T, Tucker A, Ohmura T. Development of a quick reference table for setting programmable pressure valves in patients with idiopathic normal pressure hydrocephalus. Neurol Med Chir (Tokyo) 2008;48: 42732 discussion 32. [18] Rasband W. ImageJ. [http://rsb.info.nih.gov/ij/] website National Institutes of Health, Bethesda, Maryland, USA; 1997.

5. Conclusions Shunt surgery changed the distribution of CSF in the cranium, but did not change brain volume. It also reduced the deformation of the cerebral hemisphere, which may result from the pressure of CSF in the ventricles and may contribute to the development of symptoms.

12

K. Hiraoka et al. / Journal of the Neurological Sciences 296 (2010) 712 [27] Iddon JL, Pickard JD, Cross JJ, Grifths PD, Czosnyka M, Sahakian BJ. Specic patterns of cognitive impairment in patients with idiopathic normal pressure hydrocephalus and Alzheimer's disease: a pilot study. J Neurol Neurosurg Psychiatry 1999;67:72332. [28] Stolze H, Kuhtz-Buschbeck JP, Drucke H, Johnk K, Illert M, Deuschl G. Comparative analysis of the gait disorder of normal pressure hydrocephalus and Parkinson's disease. J Neurol Neurosurg Psychiatry 2001;70:28997. [29] Miyoshi N, Kazui H, Ogino A, Ishikawa M, Miyake H, Tokunaga H, et al. Association between cognitive impairment and gait disturbance in patients with idiopathic normal pressure hydrocephalus. Dement Geriatr Cogn Disord 2005;20:716. [30] Lenfeldt N, Larsson A, Nyberg L, Andersson M, Birgander R, Eklund A, et al. Idiopathic normal pressure hydrocephalus: increased supplementary motor activity accounts for improvement after CSF drainage. Brain 2008;131:290412. [31] Ogino A, Kazui H, Miyoshi N, Hashimoto M, Ohkawa S, Tokunaga H, et al. Cognitive impairment in patients with idiopathic normal pressure hydrocephalus. Dement Geriatr Cogn Disord 2006;21:1139. [32] Sakakibara R, Hattori T, Yasuda K, Yamanishi T. Micturitional disturbance after acute hemispheric stroke: analysis of the lesion site by CT and MRI. J Neurol Sci 1996;137:4756. [33] Momjian S, Czosnyka Z, Owler BK, Czosnyka M, Pena A, Pickard JD. Pattern of white matter regional cerebral blood ow and autoregulation in normal pressure hydrocephalus. Brain 2004;127:96572.

[19] Mori E, Hirono N, Yamashita H, Imamura T, Ikejiri Y, Ikeda M, et al. Premorbid brain size as a determinant of reserve capacity against intellectual decline in Alzheimer's disease. Am J Psychiatry 1997;154:1824. [20] Fotenos AF, Snyder AZ, Girton LE, Morris JC, Buckner RL. Normative estimates of cross-sectional and longitudinal brain volume decline in aging and AD. Neurology 2005;64:10329. [21] Chan D, Janssen JC, Whitwell JL, Watt HC, Jenkins R, Frost C, et al. Change in rates of cerebral atrophy over time in early-onset Alzheimer's disease: longitudinal MRI study. Lancet 2003;362:11212. [22] Enzinger C, Fazekas F, Matthews PM, Ropele S, Schmidt H, Smith S, et al. Risk factors for progression of brain atrophy in aging: six-year follow-up of normal subjects. Neurology 2005;64:170411. [23] Conner ES, Foley L, Black PM. Experimental normal-pressure hydrocephalus is accompanied by increased transmantle pressure. J Neurosurg 1984;61:3227. [24] Hoff J, Barber R. Transcerebral mantle pressure in normal pressure hydrocephalus. Arch Neurol 1974;31:1015. [25] Stephensen H, Tisell M, Wikkelso C. There is no transmantle pressure gradient in communicating or noncommunicating hydrocephalus. Neurosurgery 2002;50: 76371. [26] Sakakibara R, Kanda T, Sekido T, Uchiyama T, Awa Y, Ito T, et al. Mechanism of bladder dysfunction in idiopathic normal pressure hydrocephalus. Neurourol Urodyn 2008;27:50710.