Reversibility of brain morphology after shunt operations and preoperative clinical symptoms in patients with idiopathic normal pressure hydrocephalus

Authors


Correspondence: Dr Hiroaki Kazui MD PhD, Department of Psychiatry, Osaka University Graduate School of Medicine, D3, 2-2 Yamadaoka, Suita, Osaka 565-0871, Japan. Email: kazui@psy.med.osaka-u.ac.jp

Abstract

Aim

Brain deformations might prevent clinical symptoms from worsening in patients with idiopathic normal pressure hydrocephalus (iNPH). We investigated the relationship between reversibility of brain morphology after shunt operations and preoperative clinical symptoms in iNPH patients.

Methods

Using head magnetic resonance images with voxel-based morphometry, we measured the cerebrospinal fluid volume in the combined areas of the lateral and third ventricles and Sylvian fissure (the volume of the ventricles and Sylvian fissure (vVS)) and the volume of the subarachnoid space at high convexity and midline areas (vHCM) before and 1 year after lumboperitoneal shunt operations in 12 patients with shunt-responsive iNPH. We used the ratio of normalized vVS to normalized vHCM (nvVS/nvHCM) as an index of the severity of the brain deformation. The degree of reversibility of the brain morphology after the shunt operation was defined as the change ratio of the preoperative nvVS/nvHCM to the postoperative nvVS/nvHCM (CR-nvVS/nvHCM). Higher CR-nvVS/nvHCM values indicated more improvement in the brain deformation. In addition, we rated the severity of the white matter lesions on the preoperative magnetic resonance images based on the Fazekas scale. Dependency in activities of daily living, gait and cognition were evaluated before and 1 year after the shunt operations.

Results

After the shunt operations, the nvVS/nvHCM and nvVS decreased significantly, and nvHCM increased significantly. The CR-nvVS/nvHCM negatively correlated with the preoperative severity of dependency in activities of daily living, gait and cognitive impairments. The CR-nvVS/nvHCM negatively correlated with the Fazekas scale, but not with age, duration of the disease and cerebrospinal fluid pressure.

Conclusions

Reversibility of brain morphology, which varied among iNPH patients, would prevent clinical symptoms from worsening in iNPH patients. The presence of white matter lesions reduced the degree of reversibility of the brain deformations in iNPH patients.

Introduction

Normal pressure hydrocephalus (NPH) was first described in 1965 by Hakim et al. as a syndrome accompanied by a progressive triad of symptoms (gait disturbance, cognitive impairment and urinary incontinence) with ventricular dilatation and normal pressure of the cerebrospinal fluid (CSF) at lumbar puncture. [1] The symptoms of NPH can be improved by draining the CSF. Disproportionately enlarged subarachnoid-space hydrocephalus has recently been defined as idiopathic NPH (iNPH) with a number of characteristics such as a narrowed subarachnoid space at high convexity and midline areas (HCM) in addition to enlarged lateral and third ventricles and Sylvian fissures (ventricles and Sylvian fissure (VS)) on an magnetic resonance image (MRI).[2] In the cranium of patients with disproportionately enlarged subarachnoid-space hydrocephalus, there is an excessive pressure gradient between the ventricles and suprasylvian subarachnoid CSF spaces.[3] The excessive pressure gradient results in compression of the periventricular cerebral cortices and subcortices, leading to the occurrence of the triad of symptoms that is characteristic of iNPH.

The effects of the excessive pressure gradient on the occurrence of the triad of iNPH symptoms vary among individual iNPH patients. A recent epidemiological study in Japan reported that there were elderly subjects in the community who showed specific features of the combination of enlarged VS and narrowed subarachnoid spaces at HCM on MRI, but they had none of the apparent triad of symptoms of iNPH.[4] These subjects were called asymptomatic ventriculomegaly with features of iNPH on MRI (AVIM).[4] In addition, the subjects with AVIM had generalized cerebral hypoperfusion that was as severe as that in iNPH patients who exhibited the triad of iNPH symptoms.[5] Moreover, some of the AVIM subjects developed iNPH and exhibited the triad of symptoms over the following 4–8 years, but others did not.[4] These findings suggest that there are differences in individual resilience against brain compression and that subjects with AVIM might be iNPH patients with high brain resilience against brain compression. We suggest that the brain's plasticity is an important factor in brain resilience against brain compression.

A study reported that the volume of the CSF space in the HCM (vHCM) in iNPH patients was negatively correlated with the CSF volume in the Sylvian fissure, and it tended to be negatively correlated with the CSF volume in the lateral and third ventricles.[6] The volume of VS (vVS) decreases and the vHCM increases after shunt operations in iNPH patients.[3, 6] Therefore, the ratio of vVS to the vHCM (vVS/vHCM) in iNPH patients can be used as an index of morphological changes in the brain.[7] Moreover, the change in the ratio of vVS/vHCM before and after the shunt operation might be used as an index of the reversibility of the brain morphology in iNPH patients.

In this study, we hypothesized that high levels of reversibility of brain morphology could prevent worsening of the clinical triad of symptoms in iNPH patients. In other words, iNPH patients with higher reversibility would show milder clinical symptoms. To verify the hypothesis, we evaluated the association between the reversibility of brain morphology and the severity of the triad of symptoms before the shunt operation in iNPH patients. In addition, we examined the relationships between some demographic and neuroimaging factors and the reversibility of brain morphology in iNPH patients.

Methods

This study was approved by the ethics committee of Osaka University Medical Hospital (Osaka, Japan). We obtained written informed consent from all the subjects or their caregivers.

Subjects

We enrolled 12 iNPH patients from those who visited the neuropsychological clinic in the Department of Psychiatry of Osaka University Medical Hospital from March 2007 to February 2012 and those who completed the 1-year follow-up programme. Inclusion criteria of this study were as follows: (i) age >60 years; (ii) presence of more than one symptom of the triad of symptoms, including gradually developing gait disturbance, cognitive impairment and urinary disturbance; (iii) ventricular dilatation (Evans index >0.3) and narrowed subarachnoid spaces at high convexity without severe cortical atrophy on MRI; (iv) absence of any disease or condition that could cause the clinical symptoms and radiological findings; (v) no history or evidence of a condition that might cause secondary NPH, such as subarachnoid haemorrhage, head trauma injury or cerebral meningitis; (vi) normal CSF pressure at lumbar puncture and normal CSF content; and (vii) improvements in any of the triad of symptoms 1 year after the lumboperitoneal shunt operation. The mean age of the 12 iNPH patients in this study was 75.3 ± 4.5 years (range: 68–81 years). The mean duration of the disease before shunt placement was 2.8 ± 1.6 years. The mean duration of education was 11.3 ± 3.2 years. The mean Mini-Mental State Examination score was 21.9 ± 5.4. The mean CSF pressure at lumbar puncture was 141.8 ± 37.7 mmH2O.

Assessments

Dependence and severity of triad symptoms

Modified Rankin Scale (mRS)[8]

The mRS is a commonly used scale that determines dependency in activities of daily living. The scores range from 0 to 6, with 0 indicating perfect health without symptoms and 6 indicating death.

iNPH grading scale[9]

The iNPH grading scale is a validated tool that is used to assess the clinical triad of iNPH symptoms. It is rated according to the clinician's observations and interviews with patients and their caregiver to assess the severity of each component of the triad separately. The score for each domain ranges from 0 to 4, with higher scores indicating worse symptoms. We used gait and cognitive scores of the scale in this study.

Gait assessment

Timed up-and-go test[10]

The timed up-and-go test measures the total time for a subject to perform a series of movements that may include standing from a seated position in an armchair, walking forward 3 m and returning to the seated position. The timed up-and-go test has been used to evaluate walking ability and gait disturbances in iNPH patients.[9, 11, 12]

Ten-metre reciprocating walking test

The 10-m walking test measures the time required to walk a 10-m distance and return.

Gait status scale revised (GSSR)[9]

The GSSR focuses on the following 10 factors of gait disturbances: (i) postural stability, which is the response to sudden, strong posterior displacement that is produced by a pull on the shoulders (range: 0–4); (ii) independence of walking (range: 0–2); (iii) wide base gait (range: 0–1); (iv) lateral sway (range: 0–2); (v) petit pas gait (range: 0–2); (vi) festinating gait (range: 0–2); (vii) freezing of gait (range: 0–2); (viii) disturbed tandem walking (range: 0–1); (ix) shuffle (range: 0–1); and (x) duckfooted walking (range: 0–1). We used the total score of the 10 items of the GSSR that ranged from 0 to 18. A higher score reflected worse symptoms.

Cognitive assessment

Frontal assessment battery (FAB)[13]

A simple tool designed to assess the frontal lobe function, the FAB comprises six subtests, and each explores one of the following functions related to the frontal lobes: (i) conceptualization and abstract reasoning; (ii) mental flexibility; (iii) motor programming and executive control of action; (iv) resistance to interference; (v) self-regulation and inhibitory control; and (vi) environmental autonomy. The FAB has been confirmed to have good internal consistency, optimal interrater reliability and concurrent validity. As an index representing overall frontal lobe function, we used the total score of the FAB, which ranges from 0 to 18.

Trail making test, part A (TMT-A)[14]

The TMT-A is a common neuropsychological test that is used to evaluate the psychomotor speed and to assess iNPH patients.[15] TMT-A requires a subject to connect randomly located numbers in a numerical order as rapidly and accurately as possible. In this study, the amount of time taken to complete the TMT-A was used in the analysis. We did not use the TMT-B because it was difficult to complete by many iNPH patients.

Attention/concentration subtests of the Wechsler Memory Scale Revised (WMS-R)

The Wechsler Memory Scale Revised is a comprehensive memory test that includes 12 subtests. In this study, mental control, digit span and visual memory span subtests, which make up the attention/concentration part, were performed. The weighted sum score of the three subtests was used in this analysis because the patients in this study were too old to calculate the age-corrected attention/concentration index.

Subtests of picture recognition and story recall of the Rivermead Behavioural Memory Test[16]

The Rivermead Behavioural Memory Test is a standardized, validated and reliable test for everyday memory. It consists of nine subtests. We used the subtests of picture recognition and the immediate and delayed recall of short story in this study.

Digit symbol substitution test and block design test of the Wechsler Adult Intelligence Scale-III

We used the digit symbol substitution and the block design subtests in the Wechsler Adult Intelligence Scale-III. The former test was used to measure the psychomotor speed, and the latter was used to evaluate visuoconstructive ability. These tests have been confirmed to be useful in iNPH patients.[17]

Shunt operation and programmable valve adjustment procedure

The patients underwent the lumboperitoneal shunt with the Codman–Hakim programmable valve systems (Johnson & Johnson K.K., Tokyo, Japan). In this study, adjustment of the shunt system's pressure was regulated as follows. The initial pressure of the shunt system was set 20–40 mmH2O lower than the CSF pressure. If the CSF pressure were under 100 mmH2O, we set the shunt at 100 mmH2O. Postoperatively, the pressure settings of the programmable valves were adjusted systematically, if necessary. If clinical improvement were absent or insufficient, the pressure setting was lowered by 20 mmH2O over a 1–2-week period. If adverse effects such as orthostatic headaches or subdural effusion were observed by computed tomography, the pressure setting was increased by 20 mmH2O. Pressure adjustments were repeated until the optimal pressure for each patient was attained.

MRI procedure

We performed MRI with a 1.5-T system (Signa Excite HD 12.x, General Electric Medical Systems, Milwaukee, WI, USA). A 3-D volumetric acquisition of a T1-weighted gradient echo sequence produced a gapless series of thin sagittal sections that covered the whole calvarium. The operating parameters were as follows: fields of view, 240.0 mm; matrix, 256 × 256; 124.0 × 1.4-mm contiguous sections; repetition time, 12.55 ms; echo time, 4.20 ms; and flip angle, 15°. In addition, axial T2-weighted fluid-attenuated inversion recovery images were obtained. The operating parameters were as follows: fields of view, 220.0 mm; matrix, 512 × 512; 21.0 × 5.0-mm contiguous sections; repetition time, 8002.00 ms; echo time, 141.41 ms; and flip angle, 90°.

Image analysis and assessment

For analysis of the voxel-based morphometry, we used Statistical Parametric Mapping 8 software (Wellcome Trust Centre for Neuroimaging, University College London, London, UK) running under MATLAB v.2010b (MathWorks, Inc., Natick, MA, USA). The CSF space was automatically segmented from the source data with the Montreal Neurological Institute template, and it was normalized by diffeomorphic anatomical registration through exponentiated lie algebra, which is a suite of tools used to obtain accurate intersubject registration of brain images.[18] The regions of interest (ROI) were set in the VS (VS-ROI) and in the HCM (HCM-ROI) (Fig. 1). The ROI were defined as the largest areas with differences in the CSF density between patients with iNPH and those with Alzheimer's disease or normal elderly controls.[7] We then measured the volumes of the segmented CSF space with the VS-ROI and HCM-ROI.[7] The volume of the VS-ROI was deemed to be vVS and the volume of the HCM-ROI was deemed to be vHCM in this study. Moreover, the vVS and vHCM were normalized (nvVS and nvHCM) by dividing the values by the intracranial volume. In addition, we used the ratio of nvVS to nvHCM (nvVS/nvHCM) as an index of the severities of the brain deformations. Larger nvVS/nvHCM values indicated severe deformation. The volumetric change rates from before and after the shunt operation were defined as the change ratio of nvVS/nvHCM (CR-nvVS/nvHCM) or the ratio of the preoperative nvVS/nvHCM to the postoperative nvVS/nvHCM. The CR-nvVS/nvHCM means reversibility of brain morphology.

Figure 1.

The regions of interest that were generated from the results of voxel-based morphometry (red, lateral and third ventricles and Sylvian fissure area; blue, high convexity and midline areas).

We rated the extent of periventricular hyperintensity (PVH) and deep white matter hyperintensity (DWMH) in patients' T2-weighted fluid-attenuated inversion recovery MRI according to the Fazekas scale.[19] PVH and DWMH were ranked with four grades. To ensure interrater reliability, two investigators (T.W. and H.Ka.) blind to the patients' clinical data independently scored these scales. We used the data scored by T.W. in the analyses.

Statistical analyses

The clinical data and CSF volumes were compared before and after the shunt operation with Wilcoxon signed-rank tests. The association between CR-nvVS/nvHCM and the gait and cognitive assessment scores before the shunt operation was analyzed with Spearman's rank correlation tests. The associations between CR-nvVS/nvHCM and age, duration of the disease, CSF pressure or the preoperative scores of the Fazekas scale in PVH and DWMH on the T2-weighted fluid-attenuated inversion recovery images were also analyzed in the same manner. All statistical analyses were performed with STATISTICA v.06J for Windows (StatSoft, Inc., Tulsa, OK, USA). P-values less than 0.05 were considered statistically significant.

Results

Changes in clinical assessments and CSF volumes after the shunt operation

The scores of the mRS, iNPH grading scale gait and cognition, GSSR, timed up-and-go test, 10-m walking test, FAB, and Wechsler Adult Intelligence Scale-III digit symbol substitution and block design were significantly improved 1 year after the shunt operation (Table 1). After the shunt operation, the nvVS and the nvVS/nvHCM decreased significantly, and the nvHCM increased significantly (P < 0.01) (Table 2).

Table 1. Results of the clinical assessments before and after shunt operation
Clinical assessmentsBefore shuntOne year after shuntP-value
Mean ± SDMean ± SD
  1. Wilcoxon signed-rank test. FAB, frontal assessment battery; GSSR, Gait status scale revised; iNPHGS, idiopathic normal pressure hydrocephalus grading scale; MMSE, Mini-Mental State Examination; mRS, modified Rankin Scale; RBMT, Rivermead Behavioural Memory Test; TMT-A, trail making test, part A; TUG, timed up-and-go test; WAIS-III, Wechsler Adult Intelligence Scale-III; WMS-R, Wechsler Memory Scale Revised.
mRS2.6 ± 1.01.9 ± 0.90.018
iNPHGS gait2.1 ± 0.71.5 ± 0.80.018
iNPHGS cognition2.5 ± 0.71.8 ± 0.90.033
GSSR7.1 ± 4.83.8 ± 3.80.005
TUG time (s)16.7 ± 8.412.5 ± 3.60.019
10-m walking test time (s)27.5 ± 15.419.2 ± 5.50.008
MMSE21.9 ± 5.424.9 ± 3.20.074
FAB9.5 ± 2.712.7 ± 1.80.002
TMT-A (s)145.6 ± 127.4114.9 ± 100.80.308
WMS-R attention/concentration raw score42.4 ± 10.846.7 ± 15.50.236
RBMT picture recognition7.5 ± 2.79.1 ± 1.50.128
RBMT story immediate recall6.2 ± 4.17.3 ± 3.10.799
RBMT story delayed recall4.0 ± 3.84.9 ± 4.00.575
WAIS-III digit symbol substitution test raw score24.1 ± 12.031.8 ± 11.40.019
WAIS-III block design test raw score18.2 ± 10.226.1 ± 10.40.012
Table 2. CSF volume before and after shunt operation
 CSF volumeP-value
Before shunt (n = 12)After shunt (n = 12)
Mean ± SDMean ± SD
  1. Wilcoxon signed-rank test. CSF, cerebrospinal fluid; nvHCM, normalized volume of the high convexity and midline areas; nvVS, normalized volume of the area of the lateral and third ventricles and Sylvian fissure.
nvVS0.44 ± 0.090.39 ± 0.070.002
nvHCM0.09 ± 0.020.11 ± 0.020.002
nvVS/nvHCM5.21 ± 1.973.50 ± 0.980.002

Correlation between the CR-nvVS/nvHCM and the severity of the clinical symptoms

The CR-nvVS/nvHCM was significantly correlated with the preoperative scores of all of the gait and cognitive assessments except for MMSE and Wechsler Adult Intelligence Scale-III block design test (Table 3 and Fig. 2). All of the significant correlations indicated an association between higher CR-nvVS/nvHCM and milder impairments before the shunt operation.

Figure 2.

Correlation between the CR-nvVS/nvHCM and the severity of the representative clinical data. Spearman's rank correlation test. CR-nvVS/nvHCM, change rate of the proportion of the normalized volume of the area of the lateral and third ventricles and Sylvian fissure to the normalized volume of the high convexity and midline areas between before and after the shunt operation; FAB, frontal assessment battery; GSSR, gait status scale revised; TUG, timed up-and-go test; WMS-R, Wechsler Memory Scale Revised.

Table 3. Correlations between CR-nvVS/nvHCM and the results of the clinical assessments before the shunt operation
Clinical assessmentsrsP-value
  1. Spearman's rank correlation test. CR-nvVS/nvHCM, change rate of the proportion of the normalized volume of the area of the lateral and third ventricles and Sylvian fissure to the normalized volume of the high convexity and midline areas between before and after the shunt operation; FAB, frontal assessment battery; GSSR, gait status scale revised; iNPHGS, idiopathic normal pressure hydrocephalus grading scale; MMSE, Mini-Mental State Examination; mRS, modified Rankin Scale; RBMT, Rivermead Behavioural Memory Test; TMT-A, trail making test, part A; TUG, timed up-and-go test; WAIS-III, Wechsler Adult Intelligence Scale-III; WMS-R, Wechsler Memory Scale Revised.
mRS−0.640.024
iNPHGS gait−0.610.033
iNPHGS cognition−0.700.011
GSSR−0.600.041
TUG−0.810.001
10-m walking test−0.590.042
MMSE0.470.127
FAB0.780.003
TMT-A−0.83<0.001
WMS-R attention/concentration raw score0.820.002
RBMT picture recognition0.670.025
RBMT story immediate recall0.720.012
RBMT story delayed recall0.610.046
WAIS-III digit symbol substitution test raw score0.730.016
WAIS-III block design test raw score0.290.413

Correlation between the CR-nvVS/nvHCM or the nvVS/nvHCM and the preoperative data

In the PVH and DWMH assessments with the Fazekas scale, the intraclass correlational coefficients, which assessed the interrater reliability, were 0.92 and 0.93, respectively. These results indicated high reliability of the scoring of the white matter lesions.

The mean PVH score of the Fazekas scale of the patients was 2.3 ±0.8, and the mean DWMH score was 2.1 ± 0.8. There were significant negative correlations between the CR-nvVS/nvHCM and the PVH score and the DWMH score of the Fazekas scale (Table 4). Moreover, there was a significant negative correlation between the preoperative nvVS/nvHCM ratio and preoperative PVH (r = −0.64, P < 0.05) as well as a negative trend between the preoperative nvVS/nvHCM ratio and preoperative DWMH (r = −0.55, P < 0.07). However, there were no significant correlations between the CR-nvVS/nvHCM and age, duration of the disease or CSF pressure (Table 4).

Table 4. Correlations between CR-nvVS/nvHCM and demographic/Fazekas data
Baseline datarsP-value
  1. Spearman's rank correlation test. CR-nvVS/nvHCM, change rate of the proportion of the normalized volume of the area of the lateral and third ventricles and Sylvian fissure to the normalized volume of the high convexity and the midline areas between before and after the shunt operation; CSF, cerebrospinal fluid; DWMH, deep white matter hyperintensities; PVH, periventricular hyperintensities.
Age−0.340.283
Duration of the disease0.140.667
CSF pressure0.250.441
Fazekas score of PVH−0.620.032
Fazaekas score of DWMH−0.610.033

Discussion

We confirmed that gait disturbance and cognitive impairment improved 1 year after shunt operations in this study's 12 iNPH patients. We also confirmed that the vVS decreased and the vHCM increased after shunt operations in the iNPH patients, results that are consistent with those of previous studies.[3, 6] However, the degree of improvement in the brain deformation after the shunt operations varied among the iNPH patients. The CR-nvVS/nvHCM was significantly correlated with the preoperative scores of almost all of the gait and cognitive assessments in this study. These findings indicated that iNPH patients who exhibited greater improvement in their brain deformation after the shunt operations, which might mean the iNPH patients with more plastic brains, had fewer clinical symptoms before the shunt operation. In other words, morphological changes in the brain in iNPH patients might prevent clinical symptoms from worsening in response to the excessive pressure gradient in the cranium of patients with disproportionately enlarged subarachnoid space hydrocephalus. This interpretation is consistent with that of subjects with AVIM,[4] as discussed in the Introduction. The mechanisms underlying the prevention of symptoms worsening as the result of morphological changes in the brain are unknown. This is an important issue that needs to be resolved in a future study.

In this study, we examined whether the reversibility of brain morphology was correlated with age, duration of the disease, CSF pressure, severity of PVH or severity of DWMH in iNPH patients. The reversibility of brain morphology was negatively correlated with the severity of the PVH and DWMH before the shunt operation. In addition, the degree of PVH and DWMH of the brain was negatively associated with the preoperative nvVS/nvHCM ratio. Thus, PVH and DWMH changes on MRI would be important factors indicating reduced plasticity of the brain. This is supported by findings that severity of PVH and DWMH were significantly and negatively correlated with the width of the anterior horns of the lateral ventricles in iNPH patients.[20] These results indicated that the iNPH patients with more severe PVH and DWMH could have smaller preoperative brain deformations and less improvement in brain deformation after the shunt operation. The PVH and DWMH of iNPH patients on MRI pathologically reflected scattered arteriosclerotic changes, microinfarcts, leptomeningeal fibrosis and glial reactions in the periventricular regions and deep white matter.[21, 22] Such pathological changes in the periventricular regions and deep white matter appear to make the brain less plastic, and they contribute to less deformation and more irreversibility of the brain morphology.

Our study had several limitations. First, the sample size of this study was small. Second, we assessed the associations between reversibility of brain morphology and only five factors that could influence the reversibility of brain, including the severity of periventricular and deep white matter lesions, age, duration of illness and CSF pressure. Third, we did not assess urinary dysfunction in detail because methods that objectively evaluate urinary dysfunction have not been established. These issues should be taken into consideration before the findings can be generalized.

The results of this study indicated the importance of plasticity of the brain in preventing the worsening of clinical symptoms in iNPH patients. Further studies are needed to clarify the factors that influence the plasticity of the brain.

Acknowledgments

This study was supported in part by the Research Committee of Normal Pressure Hydrocephalus and Related Disorders, Studies on the Etiology, Pathogenesis and Therapy from the Japanese Ministry of Health, Labour and Welfare (Tokyo, Japan) and by the Japanese Ministry of Education, Culture, Sports, Science and Technology (Tokyo, Japan) (24591708 and 24591710).

Ancillary