08ournal of Neurology, Neurosurgery, and Psychiatry 1995;59:608-615 Heterogeneous cerebral glucose metabolism in normal pressure hydrocephalus Enrico Tedeschi, Steen G Hasselbalch, Gunhild Waldemar, Marianne Juhler, Peter H0gh, S0ren Holm, Lars Garde, Lisbet Lumholdt Knudsen, Leif Klinken, Flemming Gjerris, Olaf B Paulson Department of Neurology E Tedeschi S Hasselbalch G Waldemar P Hogh 0 B Poulson Department of Neurosurgery M Juhler L Garde F Gjerris Departnent of Neuroradiology, National University Hospital, Rigshospitalet, Copenhagen, Denmark L Lumholdt Knudsen Institute of Neuropathology, University of Copenhagen, Denmark L Klinken Department of Nuclear Medicine- CNR, University "Federico II", Naples, Italy E Tedeschi Correspondence to: Steen Hasselbalch, Department of Neurology, N2081, National University Hospital, Rigshospitalet, Blegdamsvej 9, DK-2100, Copenhagen, Denmark. Received 19 December 1994 and in final revised fonn 25 July 1995 Accepted 18 August 1995 Abstract The regional cerebral metabolic rate for glucose (rCMRgsu) has never been investi- gated in large consecutive groups of patients with normal pressure hydro- cephalus (NPH), a potentially treatable form of dementia with an unpredictable outcome after shunt surgery. Using PET and 18F-2-fluorodeoxyglucose, rCMRgsu was studied in 18 patients who fulfilled hydrodynamic criteria for NPH and in whom a biopsy of the frontal cortex was obtained. When compared with an age matched group of 11 healthy subjects, the patients with NPH showed a significant rCMRI,u reduction in all cortical and sub- cortical regions of interest. Individual metabolic patterns, however, disclosed a large topographical heterogeneity. Furthermore, histopathological examina- tion identified Alzheimer's disease or cerebrovascular disease in six cases, and no parenchymal disease or non-specific degenerative processes in the remaining 12. After separating the patients accord- ing to the histological diagnosis, the rCMROu patterns were still heteroge- neous, the abnormalities ranging from focal to diffuse in both subgroups. After shunt operation, 11 patients did not improve or worsened clinically. Six patients improved; of those, two had Alzheimer changes and two cerebro- vascular changes in their biopsy. The metabolic pattern of these six patients did not differ from the rest of the NPH group. The results indicate that the NPH syn- drome may be non-specifically associated with different degenerative disorders. The metabolic heterogeneity, together with the heterogeneous histopathological findings, indicate the necessity of re- evaluating the pathogenesis of the NPH syndrome, and may account for the high variability in the success rate of shunt surgery series. (JNeurolNeurosurg Psychiatry 1995;59:608-615) Keywords: cerebral glucose metabolism; normal pres- sure hydrocephalus The use of functional brain imaging for diag- nosing and studying the pathophysiology of dementia is increasing. The normal pressure hydrocephalus (NPH) syndrome is a poten- tially treatable form of dementia, clinically characterised by gait apraxia, urinary inconti- nence, and progressive dementia. It is also known as Hakim's triad or NPH syndrome.' 2 Although NPH symptoms are accepted as a clinical entity and the ventricular enlargement can be quantified on CT or MRI, it is sometimes not easy to differentiate in vivo atypical or endstage cases from other types of dementia, particularly from dementia of the Alzheimer type (DAT), probably due to a large overlap of clinical and radiological findings. Cerebrospinal fluid shunt surgery or CSF evacuation through repeated lumbar punctures can dramatically reverse the NPH syndrome, " but the success rate is highly vari- able. Even when applying strict hydrodynamic criteria in selecting patients for surgery, there is still a significant fraction of cases who do not benefit from CSF drainage.'46 Regional cerebral glucose (rCMRe1,) and oxygen (rCMRo2) metabolism have been extensively measured with PET and '8F-2-fluo- rodeoxyglucose (FDG) in DAT.7-'0 On the other hand, no PET studies of cerebral blood flow, rCMRo,, or rCMR51& have included an adequate number of consecutive patients to describe a specific pattern of metabolic derangement in NPH."'1-4 The cerebral blood flow findings from single photon emission computed tomography (SPECT) studies in NPH are very heterogeneous in the extent and location of the cortical deficits,'5-'8 and imag- ing of subcortical regions has recently offered interesting perspectives.'920 Considering the heterogeneity in the perfu- sion findings and the variability in the success rate of shunt surgery, the aim of the present study was to test the hypothesis that different PET-FDG patterns were present in patients with NPH, due to the possible coexistence of different underlying degenerative processes. We therefore measured rCMR?,j with PET- FDG in 18 patients fulfilling hydrodynamic criteria for NPH in whom a small brain biopsy was obtained from the right frontal region. We also aimed to match PET-FDG patterns with biopsy results, to achieve a thorough morpho- functional characterisation in patients with NPH. The metabolic alterations were evalu- ated in relation to the histological findings, the severity of the clinical symptoms, and the response to shunting. Materials and methods PATIENTS In a two year period, all patients younger than 80 years who were referred to the department of neurosurgery at Rigshospitalet Copenhagen 608 on August 29, 2020 by guest. Protected by copyright. http://jnnp.bmj.com/ J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.59.6.608 on 1 December 1995. Downloaded from
Transcript
08ournal ofNeurology, Neurosurgery, and Psychiatry 1995;59:608-615
Enrico Tedeschi, Steen G Hasselbalch, Gunhild Waldemar, Marianne Juhler, Peter H0gh,S0ren Holm, Lars Garde, Lisbet Lumholdt Knudsen, Leif Klinken, Flemming Gjerris,Olaf B Paulson
Department ofNeurologyE TedeschiS HasselbalchG WaldemarP Hogh0 B PoulsonDepartment ofNeurosurgeryM JuhlerL GardeF GjerrisDepartnent ofNeuroradiology,National UniversityHospital,Rigshospitalet,Copenhagen, DenmarkL Lumholdt KnudsenInstitute ofNeuropathology,University ofCopenhagen, DenmarkL KlinkenDepartment ofNuclear Medicine-CNR, University"Federico II", Naples,ItalyE TedeschiCorrespondence to:Steen Hasselbalch,Department of Neurology,N2081, National UniversityHospital, Rigshospitalet,Blegdamsvej 9, DK-2100,Copenhagen, Denmark.Received 19 December 1994and in final revised fonn25 July 1995Accepted 18 August 1995
AbstractThe regional cerebral metabolic rate forglucose (rCMRgsu) has never been investi-gated in large consecutive groups ofpatients with normal pressure hydro-cephalus (NPH), a potentially treatableform of dementia with an unpredictableoutcome after shunt surgery. Using PETand 18F-2-fluorodeoxyglucose, rCMRgsuwas studied in 18 patients who fulfilledhydrodynamic criteria for NPH and inwhom a biopsy of the frontal cortex wasobtained. When compared with an agematched group of 11 healthy subjects, thepatients with NPH showed a significantrCMRI,u reduction in all cortical and sub-cortical regions of interest. Individualmetabolic patterns, however, discloseda large topographical heterogeneity.Furthermore, histopathological examina-tion identified Alzheimer's disease orcerebrovascular disease in six cases, andno parenchymal disease or non-specificdegenerative processes in the remaining12. After separating the patients accord-ing to the histological diagnosis, therCMROu patterns were still heteroge-neous, the abnormalities ranging fromfocal to diffuse in both subgroups. Aftershunt operation, 11 patients did notimprove or worsened clinically. Sixpatients improved; of those, two hadAlzheimer changes and two cerebro-vascular changes in their biopsy. Themetabolic pattern ofthese six patients didnot differ from the rest ofthe NPH group.The results indicate that the NPH syn-
drome may be non-specifically associatedwith different degenerative disorders. Themetabolic heterogeneity, together withthe heterogeneous histopathologicalfindings, indicate the necessity of re-evaluating the pathogenesis of the NPHsyndrome, and may account for the highvariability in the success rate of shuntsurgery series.
(JNeurolNeurosurg Psychiatry 1995;59:608-615)
Keywords: cerebral glucose metabolism; normal pres-sure hydrocephalus
The use of functional brain imaging for diag-nosing and studying the pathophysiology ofdementia is increasing. The normal pressurehydrocephalus (NPH) syndrome is a poten-tially treatable form of dementia, clinicallycharacterised by gait apraxia, urinary inconti-
nence, and progressive dementia. It is alsoknown as Hakim's triad or NPH syndrome.' 2Although NPH symptoms are accepted as aclinical entity and the ventricular enlargementcan be quantified on CT or MRI, it issometimes not easy to differentiate in vivoatypical or endstage cases from other types ofdementia, particularly from dementia of theAlzheimer type (DAT), probably due to alarge overlap of clinical and radiologicalfindings. Cerebrospinal fluid shunt surgery orCSF evacuation through repeated lumbarpunctures can dramatically reverse the NPHsyndrome," but the success rate is highly vari-able. Even when applying strict hydrodynamiccriteria in selecting patients for surgery, thereis still a significant fraction of cases who do notbenefit from CSF drainage.'46
Regional cerebral glucose (rCMRe1,) andoxygen (rCMRo2) metabolism have beenextensively measured with PET and '8F-2-fluo-rodeoxyglucose (FDG) in DAT.7-'0 On theother hand, no PET studies of cerebral bloodflow, rCMRo,, or rCMR51& have included anadequate number of consecutive patients todescribe a specific pattern of metabolicderangement in NPH."'1-4 The cerebral bloodflow findings from single photon emissioncomputed tomography (SPECT) studies inNPH are very heterogeneous in the extent andlocation of the cortical deficits,'5-'8 and imag-ing of subcortical regions has recently offeredinteresting perspectives.'920
Considering the heterogeneity in the perfu-sion findings and the variability in the successrate of shunt surgery, the aim of the presentstudy was to test the hypothesis that differentPET-FDG patterns were present in patientswith NPH, due to the possible coexistence ofdifferent underlying degenerative processes.We therefore measured rCMR?,j with PET-FDG in 18 patients fulfilling hydrodynamiccriteria for NPH in whom a small brain biopsywas obtained from the right frontal region. Wealso aimed to match PET-FDG patterns withbiopsy results, to achieve a thorough morpho-functional characterisation in patients withNPH. The metabolic alterations were evalu-ated in relation to the histological findings, theseverity of the clinical symptoms, and theresponse to shunting.
Materials and methodsPATIENTSIn a two year period, all patients younger than80 years who were referred to the departmentof neurosurgery at Rigshospitalet Copenhagen
608 on A
ugust 29, 2020 by guest. Protected by copyright.
http://jnnp.bmj.com
/J N
eurol Neurosurg P
sychiatry: first published as 10.1136/jnnp.59.6.608 on 1 Decem
Heterogeneous cerebral glucose metabolism in normal pressure hydrocephalus
with clinical and radiological findings sugges-tive of NPH were asked to enter the study.A final diagnosis of NPH was based on thefollowing previously validated25 criteria: (a)personal history and clinical evidence of twoor more symptoms of the NPH syndrome atthe neurological evaluation; (b) significantventricular enlargement on CT (Evans' ratio> 0-30),5 21 and (c) hydrodynamic criteria (seeCSF perfusion study and biopsy).
This report includes all the patients with afinal diagnosis of NPH, in whom a brain MRIstudy, PET-FDG, and a frontal biopsy wereobtained. A total of 18 consecutive patients(10 men, eight women, mean age 63-4, range46-76 years) was thus selected, comprisingonly patients with socalled "idiopathic" NPH,as no history of significant head trauma,meningitis, subarachnoid haemorrhage, orother causes known to be associated with thedevelopment ofNPH were reported.
STUDY PROGRAMMEOn admission of the patients a trained neuro-logist performed a clinical and neurologicalevaluation paying special attention to gait,sphincter function, and cognitive status. Gaitwas rated on a point scale (1 = normal,2 = abnormal but possible without support,3 = need of cane, 4 = need of support ofanother person, 5 = bedridden); likewise,urinary incontinence was rated as 1 = none,2 = rare, 3 = occasional, 4 = constant,5 = catheter. The presence of dementia wasevaluated according to DMS III-R criteria,22 itsseverity was assessed with the mini mentalstate examination (MMSE23), and the globaldeterioration scale (GDS24). Scores for MMSEwere also obtained in the control subjects.
In all patients and control subjects astandard laboratory test battery was per-formed. Both groups underwent cranial MRIand PET-FDG and the patients with NPHalso had a CSF perfusion study and a brainbiopsy. Three to nine months after shuntoperation, the clinical response to the opera-tion was assessed. A trained neurologist scoredthe patients as described, and changes in theclinical condition were semiquantitativelyevaluated by calculating the change in the sumof scores. The clinical status was graded asworsened, unchanged, slightly improved, orconsiderably improved.
All subjects and patients (with theirrelatives) gave informed consent to participatein the study, which was approved by thelocal ethics committee (protocol No V.100.1500/90).
CONTROL SUBJECTSEleven healthy subjects (six men, five women,mean age 62 (range 53-75 years), recruited byadvertisement, volunteered for the study. Thesubjects had no family history of dementia andno personal history of any psychiatric orneurological disease, severe head trauma, anymetabolic or endocrine disturbances, cancer,severe hypertension, heart disease, or alcoholor drug misuse. Neurological and neuro-psychological examination disclosed no
abnormalities (MMSE scores > 27). The brainMRI was described as normal by a trainedradiologist.
MRIUsing conventional spin echo sequences(TR = 2000 ms, TE = 30-90 ms) on a 0 3Tesla MR tomograph (FONAR, USA), theproton density weighted images of ninetransverse brain slices were obtained. Slicethickness was 10 mm, with an interslice dis-tance of 12 mm. In one patient the MRI wastechnically inadequate because of movementartefacts; instead, a cranial CT was performed.
PET-FDGThe subject was placed supine, in quiet, dimsurroundings, with eyes closed and earsunplugged from 30 minutes before to40 minutes after the intravenous injection of180-250 MBq FDG. Starting simultaneouslywith FDG injection, 1-5 ml blood sampleswere drawn from the radial artery withincreasing time intervals, then placed on iceand centrifuged. Plasma activity was countedin a gamma counter (COBRA 50003, PackardInstruments, Downers Grove, Illinois).Some 40 to 50 minutes after injection, the
subject was moved to the couch of theTherascan 3128 PET scanner (Atomic energyLtd, Canada) and three scans were performed,in wobble mode, for a total scan time of 45minutes. To avoid head movements duringthe scan, the head was fixed by an individuallymoulded cushion made of a two componentfoam. Nine consecutive slices, 12 mm thick,with an image plane resolution of 10-12 mmwere thus obtained. Correction for dead time,randoms, scatter, and attenuation was per-formed as previously described.25 The cor-rected count rate was 0.6-1-46 x 10 countsper slice. Calibration of the PET camera andcross calibration with the gamma counterwere performed at the end of the day.
Regional CMRIU was measured by theautoradiographic approach first described bySokoloff et al,26 and later modified by Brooks.27The following fixed values for the rate con-stants were used: k, = 0105., k2 = 0 126, k3 =0-075, and k4 = 0-0068. The lumped constantwas fixed at 0-56. Plasma glucose concen-trations during the study were obtained byaveraging several measurements performedwith a Beckman glucose analyser (BeckmanInstruments Inc, USA).
POSITIONINGThe canthomeatal plane was used as a refer-ence for MRI and PET-FDG, and the mid-planes of the nine slices were placed at thesame levels above the canthomeatal plane forboth studies. The transaxial MR slices weretilted according to the canthomeatal plane,identified on the sagittal MR TI weighted(TR = 500 ms, TE = 25 ms) image whichdisplayed a small vaseline filled tube fixed tothe skin. The same positioning was obtainedfor the PET-FDG study using a laser beamparallel to the canthomeatal line. As slicethickness was roughly the same (MRI: 10 mm,
609 on A
ugust 29, 2020 by guest. Protected by copyright.
http://jnnp.bmj.com
/J N
eurol Neurosurg P
sychiatry: first published as 10.1136/jnnp.59.6.608 on 1 Decem
Tedeschi, Hasselbalch, Waldemar, J7uhler, Hogh, Holm, et al
PET: 12 mm), the volume of the braincovered by these scans was easily comparable.
DATA ANALYSISAlthough automatic CO registration of MRIand PET-FDG was not achievable at the timeof the study, the use of a common positioningtechnique and an atlas derived region ofinterest definition, made it possible to easilyidentify brain structures on MRI and cor-responding functional areas on PET-FDG.With reference to an anatomic atlas,28 ninebrain levels, parallel to the canthomeatal planeseparated by 10 mm were selected. In each ofthese levels, a template of symmetric corticaland subcortical regions of interest wasdefined, as previously described in detail.29 Foreach level, the region of interest template wasdrawn on the MRI slice and then applied tothe PET image, using a VAX station. Itwas only allowed to redraw or adjust the pre-cise location of central grey matter structures,white matter regions, and the outer bound-aries of the brain, as minute positioning errorsbetween MRI and PET had the greatestinfluence on these areas. The regions of interestwere incorporated into larger regions of mean-ingful anatomical localisation, which couldcomprise several slices.29 The weighted meanmetabolism value of the larger regions ofinterest was calculated from the mean pixelvalue and the area of the region for each slicein which the region of interest appeared.Furthermore, asymmetry ratios within indi-vidual subjects were calculated between left-right and anterior-posterior correspondingregions of interest. In particular, for eachregion of interest, a side to side asymmetryindex (SAI) in respect to the contralateralregion, was calculated:
SAII (%) = 100 x ([Mj(R) - Mi(j]/Mi(max))where MR and ML are the weighted meanmetabolic values in the right and the leftcounterparts of the two regions of interestrespectively, and Mmax is the highest of thesetwo values. Frontotemporal and frontoparietalratios (anterior/posterior ratios) were calcu-lated, to detect possible internal asymmetries,also along the anterior-posterior axis of thebrain.
For the individual characterisation of themetabolic patterns in patients and controls,the absolute CMRg,IU and the within subjectside to side asymmetry indices and the ante-rior-posterior ratios were calculated dividingthe cerebal cortex into three symmetric mainregions: (a) frontal, left, and right, (b) temporal,left, and right, and (c) parietal, left and right.A normality range (mean (2 SD)) for bothabsolute CMRg,U values and internal asymme-try ratios was defined from the control materialand regions outside the range were consideredabnormal. A "metabolic impairment score",expressing both the absolute reduction in corti-cal rCMR,gu and the presence of within subjectcortical asymmetries was then assigned to eachindividual PET-FDG pattern. In particular, 1point was given for any region resulting inabnormal absolute rCMRg,u or in internal
asymmetry and 1-5 points for any region withboth absolute rCMRg,u reduction andincreased internal asymmetry. Metabolicimpairment score values (range 0-8) thusenabled us to pinpoint abnormal regions inevery patient with NPH. Also, ratios betweenrCMRgiu values in the association cortex(middle frontal gyrus and angular-supramar-ginal parietal gyrus) v primary cortex (precen-tral and postcentral gyri) and v subcorticalgrey matter (basal ganglia regions of interest)were obtained in every patient and controlsubject and correlated with the histologicalfindings.
CSF PERFUSION STUDY AND BIOPSYThe patients underwent a lumboventricularperfusion test with measurement of theresistance to CSF outflow2' and a 24 hourmonitoring of intracranial pressure. Increasedresistance to CSF outflow (> 10 mm Hg/ml/min) with a normal or slightly raised intracra-nial pressure (< 15 mm Hg) were required forthe final diagnosis of NPH. One to four weekslater, during the placement of a ventriculoperi-toneal shunt through the same burr hole, asmall (< 1 cm3) biopsy was taken from theright superior frontal cortex and underlyingwhite matter, avoiding the tissue affected bythe earlier intracranial pressure measurementand without electrocoagulating the tissuebefore its excision.
Five ,um thick slices of the formalin fixed,paraffin embedded biopsy specimens werestained with standard dyes (haematoxylin-eosin, perodic acid-Schiffs, and Congo red)and immunohistochemical techniques (mark-ers for glial fibrillary acidic protein, ubiquitin,tau, and # amyloid). Brain tissue was thenexamined and the prevalent pathologicalchanges present were described.
STATISTICS
Regional CMRF,U values and internal ratioswere compared by Student's two sampleunpaired t tests between the NPH and controlgroups. A Bonferroni correction for the effectof multiple comparisons was applied, and aneffective P value < 0-0025 was consideredsignificant. Simple linear regression analysis(Pearson product moment correlation) wasperformed in the patients with NPH betweenabsolute rCMRg,u values in cortical and sub-cortical grey and white matter and the neuro-logical evaluation scores, and also betweenmetabolic impairment scores and resistance tooutflow values. The statistical significancelimit was set at P = 0 05.
ResultsTable 1 presents the clinical, MRI and histo-logical findings for the patients. Gait apraxiawas the most frequent symptom (89%). Eightpatients had incontinence, ranging from rareto constant. The mean (SD) MMSE and GDSscores were 22-3 (5-5) and 4 4 (1-2), respec-tively, indicating mild to moderate dementiain the group. Eleven patients (of a total of 14evaluations) were demented according to the
610 on A
ugust 29, 2020 by guest. Protected by copyright.
http://jnnp.bmj.com
/J N
eurol Neurosurg P
sychiatry: first published as 10.1136/jnnp.59.6.608 on 1 Decem
Heterogeneous cerebral glucose metabolism in normal pressure hydrocephalus
Table 1 Clinical, MRI and histopathological data in the group ofpatients with NPH
Age! White ClinicalCase sex Gait Incontinence GDS MMSE matter lesions RoXz Biopsy shunt response
1 63/M 2 2 3 27 None 15 AD No change2 46/F 1 1 4 13 None 25 No sign path Worse3 74/F 3 1 2 27 Present 29 Non-specific Considerable4 52/M 2 1 6 26 None 16 No sign path No change5 76/F 3 4 - 19 Present 29 Non-specific Considerable6 60/M 2 1 4 26 None 12 AD Slight7 76/F 3 3 6 17 None 18 AD Worse8 70/F 2 2 6 20 None 20 No sign path Worse9 68/M 2 2 3 25 Present 29 CVD Slight10 70/F 2 3 6 27 None 20 AD Considerable11 74/M 3 2 4 25 Present 24 CVD Considerable12 57/M 2 1 4 - Present 17 No sign path No change13 53/M 1 1 4 21 None 20 No sign path No change14 53/M 2 1 4 28 None 16 Non-specific No change15 49/F 2 1 4 29 Present 20 No sign path No change16 64/M 2 1 4 23 Present 15 No sign path No change17 67/M 4 4 6 12 Present (CT) 16 No sign path No change18 69/F 4 1 5 14 Present 16 No sign path Not followed
Figures in the gait and incontinence columns represent the scores at the clinical evaluation (see Methods for definition); GDS =global deterioration scale24; MMSE = mini mental state examination scores23; WM lesions = periventricular/subcortical white mat-ter abnormalities on MRI (on CT in case 17); R,,, = resistance to CSF outflow (mm Hg/mltmin) measured by a lumboventricularperfusion test; biopsy = histopathological changes in the right frontal brain tissue; AD = Alzheimer's disease; CVD = cerebrovas-cular disease; No sign path = no significant pathology; Non-specific = non-specific changes; clinical shunt response = clinical eval-uation of the response to shunt operation, evaluated three to nine months after shunt operation; no change = no change insymptoms (gait, incontinence, and dementia); slight = slight improvement in symptoms; considerable = considerable improve-ment in symptoms.
DMS III-R criteria. Eight out of 18 patientswith NPH thus showed the complete Hakim'striad. Eight patients showed MRI hyper-intensities in the white matter areas, and inone patient white matter lesions were visibleon CT (MRI not available). In six patientsspecific degenerative changes were found inthe cerebral parenchyma: cerebrovasculardisease in two, typical Alzheimer's diseasechanges in four. Of the remaining 12 patients,nine did not show any pathological abnormal-ity, while in three patients (3, 5, and 14) non-
specific degenerative processes were described(subacute encephalomalacia or cortical degen-eration).
Nine months after shunt operation, clinicalfollow up showed that the clinical status was
improved in six patients, and in four of these,the improvement was considerable. The group
Table 2 Absolute values and internal asymmetry ratios of regional glucose metabolismmeasured by PET in patients with normal pressure hydrocephalus and in control subjects
Values are means (SD). CMR,,U = regional glucose metabolism (amol. 100 g- '.min 1; mean ofleft and right symmetric ROIs); SAL = side to side asymmetry index in (%), negative valuesindicate higher metabolism in left than in the right hemispheric ROI.*P < 0-05; **P < 0-001; ***P < 0-0001 (Student's two sample t test). Values in parenthesis are Pvalues before Bonferroni correction for multiple comparisons. NS = non-significant (afterBonferroni correction).
of patients that responded to shunt operationincluded two patients with non-specificchanges in their biopsy, two patients withAlzheimer's disease, and two patients withcerebrovascular disease. In eight patientsno change was seen after shunt operation,whereas symptoms progressed in threepatients.
Table 2 shows absolute rCMR5,1 and theasymmetry indices. The NPH group as awhole showed a pronounced reduction inrCMRg,u compared with the control group inthe entire cortex (20%), in the frontal (20%),temporal (22%), parietal (19%), and occipital(15%) lobes, and in the central grey (28%)and white matter (28%). The side by sideasymmetry indices and anterior/posteriorratios showed no significant differences in anyregion of interest between the two groups; thecoefficient of variation, however, was on aver-age, two to four times higher in the side byside asymmetry index of the regions of interestof the patients with NPH, indicating aninhomogeneous rCMRgu reduction.
Figure 1 shows the cortical metabolicpatterns of all patients. Whereas rCMRf& wasglobally decreased in the NPH group, this wasnot the case on an individual level, where focalas well as diffuse significant reductions wereobserved, and where significantly abnormalside by side asymmetry indices and anterior/posterior ratios were oddly distributedthroughout the cortex. In fact the observedmetabolic impairment score values rangedfrom 0 (cases 3 and 16) to 7 and were deter-mined by abnormal cortical regions localisedwithout a recognisable pattem (fig 1). When,based of the biopsy findings, patients withspecific degenerative processes were separatedfrom those with no disease or non-specificchanges, no characteristic distribution of themetabolic abnormalities was evident in eitherof the two subgroups. Both included meta-bolic pattems ranging from focal reductions inrCMRFu or abnormal asymmetry ratios tosevere diffuse symmetric or asymmetricdecrease in rCMRf& (fig 2). Of the two
611
on August 29, 2020 by guest. P
rotected by copyright.http://jnnp.bm
j.com/
J Neurol N
eurosurg Psychiatry: first published as 10.1136/jnnp.59.6.608 on 1 D
patients with no abnormal regions (metabolicimpairment score = 0), patient 3 showed non-specific changes and patient 16 no disease atall. The analysis of the metabolic ratiosbetween association cortex and primary cortexor subcortical areas in the patients with NPHshowed no significant differences. In particu-lar, the metabolic ratio between associationand primary cortex in the four patients with"DAT-NPH" was 0 93 (0 04) whereas that in
patients with "pure NPH" was 0-97 (0 04) (P >0 05); the metabolic ratio between associationcortex and central grey matter in the patientswith "DAT-NPH" was 0-96 (0 07) and that inthe patients with "pure NPH" was 0 95 (0 09)(P > 0-05).No significant correlation was found
between the absolute metabolism in eithercortical or white matter regions and the severityof the specific symptoms of NPH (gait apraxia
Figure 2 Absolute rCMRg,,, images at comparable brain levels of a control subject (left), a patient with NPH with no relevant pathology in the biopsy(case 2, middle), and a patient with NPH with histopathological changes associated with DAT (case 7, right). A focal (frontal) reduction in glucosemetabolism is evident in the middle scan and the scan on the right shows diffuse CMRJi,, reduction.
612 on A
ugust 29, 2020 by guest. Protected by copyright.
http://jnnp.bmj.com
/J N
eurol Neurosurg P
sychiatry: first published as 10.1136/jnnp.59.6.608 on 1 Decem
Heterogeneous cerebral glucose metabolism in normal pressure hydrocephalus
and urinary incontinence). The severity ofdementia, assessed by GDS and MMSE, was
significantly correlated with the rCMR81ureduction in central grey matter (P < 005and P < 0-001, respectively). Likewise, thedecrease in frontal cortical metabolism was
significantly correlated with MMSE scores
(P < 0-005). White matter metabolism was
not correlated with either of the two dementiaevaluation scales. Also, no correlation was
found between the metabolic impairmentscore and the resistance to outflow values,neither when all patients were analysedtogether, nor when the two groups (non-specific and specific changes in the biopsies)were analysed separately. When shunt respon-
ders were compared with non-responders, no
particular metabolic pattern was evident, andaccordingly, in both groups MIS values rangedfrom the lowest (0) to the highest values (7).
DiscussionThe present study, the first PET-FDG reporton a prospective group of patients with NPH,showed heterogeneous patterns of corticalmetabolic abnormalities, associated withspecific degenerative processes (DAT andcerebrovascular disease) in some cases, andto non-specific or no cortical disease inothers. The observed metabolic heterogeneity,together with the heterogeneity in thehistopathological findings, supports thehypothesis that NPH may be non-specificallyrelated to different degenerative disorders andindicates the necessity of a reassessment of thepathogenic mechanisms of the syndrome.
Previous PET-FDG studies on NPH havereported diffusely reduced preoperativeCMRg,U.1113 14 Jagust et al II measured rCMRI1uin two patients and stated that PET-FDGcould differentiate NPH from DAT-that is,diffuse rCNR,u reduction in the first case v
predominant temporoparietal changes in thesecond. By contrast, a patient with globalasymmetric reduction and improvement aftershunt operation, reported by Kaye et al,13could not be differentiated from patients withDAT on the basis of the PET-FDG pattern.Friedland14 described another patient, pre-
operatively diagnosed as probable Alzheimer'sdisease, who showed a diffuse hypometa-bolism, most severe in the parietal cortex, thatreversed after shunt surgery, in agreementwith clinical and radiological improvement.One PET study of CMRo, and cerebral bloodflow in seven patients with chronic NPHfound a reduction of CMRo, in the cortex, butno regional analysis was performed."2 Thesefew PET studies and earlier reports of globalreduction in CMRI,U and CMRo230 31 do notseem to provide sufficient data to define any
specific pattern of metabolic derangement inNPH.Our group of patients with NPH as a whole
showed a pronounced diffuse reduction ofCMR8IU when compared with age matchedhealthy subjects. The PET-FDG individual(metabolic impairment score) analysis cor-
rectly identified 16 out of 18 patients as abnor-
mal (sensitivity: 89%) and nine out of 11healthy controls as normal (specificity: 82%).The distribution of rCMRg, deficits within theNPH group was indeed heterogenous, and acharacteristic metabolic pattern was not evi-dent. Such heterogeneity, together with thatreported in previous SPECT studies,'5-18 20suggests the presence of different disorders inpatients with "NPH". When grouped accord-ing to the histological findings, NPH sub-groups still showed a variety of metabolicpatterns. In the 12 patients with a metabolicabnormality in their right frontal lobe, it canbe reasonably assumed that the metabolicdeficit was the expression of the histologicalchange identified in the brain biopsy. In theother cases, undetected degenerative disordersin other areas of the brain may be responsiblefor the heterogenous metabolic pattern. Toexclude this possibility in the subgroup withno specific disease, which could represent the"true idiopathic" NPH patients, we also con-sidered only the eight patients (2, 5, 12-15,17, and 18) with a metabolic deficit in theright frontal region: the metabolic heterogene-ity was still evident, suggesting that evenpatients with "biopsy verified true idiopathic"NPH exhibit a heterogeneous PET-FDGpattern. In patients with DAT, CMReU reduc-tions may asymmetrically involve all neocorticalassociation areas,1032-35 and various focal corti-cal blood flow abnormalities have recentlybeen shown with SPECT-HMPAO.'6 Like-wise, dementia in patients with cerebrovascu-lar disease has consistently been associatedwith varied focal reductions in cerebral bloodflow or CMR,U. 0333' In conclusion, in agree-ment with other reports in the medical litera-ture, heterogenous rCMR.,1 patterns werefound in the two subgroups with specific dis-ease but heterogeneity was also present in thepatients with NPH with no specific disease inthe brain biopsy. This conclusion was alsosupported by the fact that the metabolic ratios(association cortex/primary cortex and associ-ation cortex/central grey matter) did not sig-nificantly differ between the four patients withDAT and the 12 with no specific disease intheir biopsy-that is, the two subgroups bothpresented variable PET-FDG pattems. How-ever, a comparison to validate a method fordifferentiating the two diseases should bemade with much larger groups, not consistingof patients who are likely to have both of them.The presence of different degenerative
disorders in the biopsies of our cases may beeasily explained by admitting that thesepatients developed NPH in association withanother primary degenerative process. Ourpatients were referred to Rigshospitalet witha clinicoradiological suspicion of NPH anddid not differ in clinical presentation fromother NPH groups described in the medicalliterature 2 6 20 30 3l 3843 The diagnosis and theindication for a shunt operation were thenconfirmed according to well defined hydro-dynamic criteria.45 The measurement of resis-tance to outflow (Ro,t) as a hydrodynamicvariable for CSF disturbances is reproducibleand comparable when different infusion and
613 on A
ugust 29, 2020 by guest. Protected by copyright.
http://jnnp.bmj.com
/J N
eurol Neurosurg P
sychiatry: first published as 10.1136/jnnp.59.6.608 on 1 Decem
Tedeschi, Hasselbalch, Waldemar, Juhkler, Hogh, Holm, et al
perfusion methods are used,9 and the normalrange is well defined.44 B wave activity was alsoregistered, but R,,,,t is considered a more validvariable for obstruction of CSF, and is there-fore used as the main diagnostic criteria forNPH. Different degenerative disorders,including DAT, progressive supranuclearpalsy, and cerebrovascular disease, may them-selves lead to a deficit in CSF reabsorptionand have been previously described in patientswith "NPH".45 S' Moreover, the pathophysio-logy of idiopathic NPH at present is not fullyunderstood and as no unequivocal pathologi-cal findings in large groups have beenobserved,48 it is not surprising to find con-comitant degenerative processes in some ofthese patients. In line with this notion, it isinteresting that out of six responders to shuntoperation, four showed specific degenerativedisease in their biopsy (two Alzheimer's dis-ease and two cardiovascular disease). Thisfinding suggests that the symptomatology ofthese patients may derive from a combinationof NPH and a specific degenerative disorder.When shunted, the patients may still improveclinically when NPH contributes significantlyto the symptoms. Thus, the finding of a spe-cific degenerative disorder in a brain biopsyfrom a patient who shows typical symptoms ofNPH (or the finding of coexisting typicalsymptoms of another degenerative disorder)should not automatically defer the patientfrom shunting, and in our study shuntingseemed to have temporarily halted theinevitable deterioration of two patients withAlzheimer's disease.From the metabolic patterns it was not
possible preoperatively to point out thosepatients who would benefit from a shuntoperation, which again was probably due tothe heterogenous nature of NPH, but largerpatient groups should be studied to draw anyfirm conclusions.The lack of correlation between clinical
symptoms typical for NPH and the distribu-tion of cortical metabolic abnormalities sug-gests that heterogeneous PET-FDG patternsdo not reflect the stages of progression of asingle disease, but are rather due to the under-lying concurrent degenerative disorders. TheMMSE scores did correlate with metabolismin the frontal cortex and in central grey matterregions, whereas GDS scores correlated onlywith central grey matter metabolism. What-ever the underlying pathology may be, this isnot surprising since metabolic abnormalitiesin areas thought to be involved in cognitivefunctions should lead to a decreased perfor-mance in the dementia evaluation tests.
In summary, we found a very heterogeneouspattern of glucose metabolism reduction in agroup of consecutive idiopathic patients withNPH diagnosed according to well definedcriteria. The presence of different pathologicalprocesses supports the hypothesis that NPHmay be non-specifically related to multipledegenerative disorders with a common clinicalpresentation. The present study also showedthat different parenchymal degenerations,associated with the development of NPH syn-
drome, were characterised by heterogeneousPET-FDG patterns, and that even patientswith no specific histopathological changes,which may represent the patients with "trueidiopathic" NPH, did not present a typicalpattern of metabolic impairment. Although aclear explanation for the pathogenic mecha-nisms of NPH and the degenerative processesleading to dementia is not achievable yet, wethink that the integration of morphofunctionalbrain imaging modalities is required for a betterunderstanding of these diseases.
The excellent technical assistance of Karin Stahr, Pia Tejmer,and Gerda Thomsen is gratefully acknowledged. This studywas supported by grants from the Danish Medical ResearchCouncil; the Lundbeck Foundation; the Danish HospitalFoundation for Medical Research; Region of Copenhagen, theFaroe Islands, and Greenland; the Kathrine and ViggoSkovgaard's Foundation; and the Foundation of 17-12-1981.
1 Hakim S, Adams RD. The special clinical problem ofsymptomatic hydrocephalus with normal cerebrospinalfluid pressure. Observations on cerebrospinal fluidhydrodynamics. _7 Neurol Sci 1965;2:307-27.
2 Gjerris F, Borgesen SE. Current concepts of measurementof cerebrospinal fluid absorption and biomechanics ofhydrocephalus. Adv Tech Stand Neurosurg 1992;19:145-77.
3 Black PM. The normal pressure hydrocephalus syndrome.In: Scott RM, ed. Hydrocephalus (concepts of neurosurgery,III). Baltimore: Williams and Wilkins, 1990:109-14.
4 Borgesen SE, Gjerris F, Sorensen SC. Intracranial pressureand conductance to outflow of CSF in normal pressurehydrocephalus..7 Neurosurg 1979;50:489-93.
5 Borgesen SE, Gjerris F. The predictive value of con-ductance to outflow of CSF in normal pressurehydrocephalus. Brain 1982;105:65-86.
6 Vanneste J, Augustijn P, Dirven C, Tan WF, GoedhartZD. Shunting normal pressure hydrocephalus: do thebenefits outweigh the risks? A multicenter study andliterature review. Neurology 1992;42:54-9.
7 Frackowiack RSJ, Pozzilli C, Legg NJ, et al. Regional cere-bral oxygen supply and utilization in dementia. A clinicaland physiological study with oxygen- 15 and positrontomography. Brain 1981;104:753-78.
8 Hoffman JM, Guze BH, Baxter LR, et al. "'F-Fluoro-deoxyglucose (FDG) and positron emission tomography(PET) in aging and dementia. A decade of studies. EurNeurol 1989;29(suppl 3):16-24.
9 Rapoport SI. Positron emission tomography in Alzheimer'sdisease in relation to disease pathogenesis: a criticalreview. Cerebrovasc Brain Metab Rev 199 1;3:297-335.
10 Mazziotta JC, Frackowiack RSJ, Phelps ME. The use ofpositron emission tomography in the clinical assessmentof dementia. Sem NuclMed 1992;XXII:233-46.
12 Brooks DJ, Beaney RP, Powell M, et al. Studies on cerebraloxygen metabolism, blood flow, and blood volume, inpatients with hydrocephalus before and after surgicaldecompression, using positron emission tomography.Brain 1986;109:613-28.
13 Kaye JA, Grady CL, Haxby JV, et al. Plasticity in the agingbrain. Reversibility of anatomic, metabolic, and cognitivedeficits in normal pressure hydrocephalus followingshunt surgery. Arch Neurol 1990;47:1336-41.
14 Friedland RP. "Normal"-pressure hydrocephalus and thesaga of the treatable dementias. J7AMA 1989;262:2577-81.
15 Vorstrup S, Christensen J, Gjerris F, et al. Cerebral bloodflow in patients with normal-pressure hydrocephalusbefore and after shunting. 7 Neurosurg 1987;66:379-87.
16 Graff-Radford NR, Rezai K, Godersky JC, et al. Regionalcerebral blood flow in normal pressure hydrocephalus.JNeurol Neurosurg Psychiatry 1987;50: 1589-96.
17 Moretti JL, Sergent A, Louarn F, et al. Cortical perfusionassessment with 123I-isopropyl amphetamine (123I-IAMP) in normal pressure hydrocephalus (NPH). Eur3Nucl Med 1988;14:13-9.
18 GranadoJM, Daz F, Alday R. Evaluation of brain SPECTin the diagnosis and prognosis of the normal pressurehydrocephalus syndrome. Acta Neurochir 1991;112:88-91.
19 Kimura M, Tanaka A, Yoshinaga S. Significance ofperiventricular hemodynamics in normal pressure hydro-cephalus. Neurosurgery 1992;30:701-4.
20 Waldemar G, Schmidt JF, Delecluse F, et al. High resolutionSPECT with [9""Tc]-d,l-HMPAO in normal pressurehydrocephalus before and after shunt operation. .7 NeurolNeurosurg Psychiatry 1993;52:655-64.
21 Evans WA. An encephalographic ratio for estimatingventricular enlargement and cerebral atrophy. ArchNeurol 1942;47:931-7.
22 DSM III-R. Diagnostic and statistical manual of mental disor-
614 on A
ugust 29, 2020 by guest. Protected by copyright.
http://jnnp.bmj.com
/J N
eurol Neurosurg P
sychiatry: first published as 10.1136/jnnp.59.6.608 on 1 Decem
Heterogeneous cerebral glucose metabolism in normal pressure hydrocephalus
ders. 3rd (revised) ed, Washington: American PsychiatricAssociation 1987.
23 Folstein MF, Folstein SE. McHugh PR. Mini-mental state:a practical method for grading the cognitive state ofpatients for the clinician. Jf Psychiat Res 1975;12: 189-98.
24 Reisberg B, Ferris SH, de Leon M, et al. The global deteri-oration scale for assessment of primary degenerativedementia. AmI Psychiatry 1982;139:1136-9.
25 Cooke BE, Evans AC, Fanthome EO, et al. Performancefigures and images from the Therascan 3128 positronemission tomography. IEEE Trans Nucl Sci 1984;NS31:640-4.
26 Sokoloff L, Reivich M, Kennedy C, et al. The C-14-deoxyglucose method for the measurement of localcerebral glucose utilization: theory, procedure, andnormal values in the conscious and anesthetized albinorat. Jf Neurochem 1977;28:897-916.
27 Brooks RA. Alternative formula for glucose utilisationusing labelled deoxyglucose. Jf Nucl Med 1982;23:538-9.
28 Aquilonius SM, Eckemras SA. A colour atlas of the humanbrain. Adapted to computed tomography. Stockhoim:Esselte Studium, 1980.
29 Waldemar G, Hasselbalch S, Andersen AR, et al. 99-Tc-d,l-HMPAO and SPECT of the brain in normal aging.Cereb Blood Flow Metab 1991;11:508-21.
30 Grubb RI, Raichle ME, Gado MH, et al. Cerebral bloodflow, oxygen utilization and blood volume in dementia.Neurology 1977;27:905-10.
31 Lying-Tunnell U, Lindblad BS, Maimund HO, Persson B.Cerebral blood flow and metabolic rate of oxygen, glu-cose, lactate, pyruvate, ketone bodies and amino acids.Acta Neurol Scand 1981;63:337-50.
32 Benson FD, PET/Dementia: an update. Neurobiol 1988;9:87-8.
33 Duara R, Barker W, Loewenstein D, et al. Sensitivity andspecificity of positron emission tomography and magneticresonance imaging studies in Alzheimer's disease andmulti-infarct dementia. Eur Neurol 1989;29(suppl 3):9-15.
34 Meyer E. The use of positron emission tomography in theearly diagnosis of senile dementia. In: Bergner M,Reisberg B, eds. Diagnosis and treatment ofsenile dementia.Berlin: Springer-Verlag, 1989;243-58.
35 Powers WJ, Perimutter JS, Videen TO, et al. Blindedclinical evaluation of positron emission tomography fordiagnosis of probable Alzheimer's disease. Neurology1992;42:765-70.
36 Waldemar G, Bruhn P, Kristensen M, Johnsen A,Paulson OB, Lassen NA. Heterogeneity of neocortical
blood flow deficits in dementia of the Alzheimer type: a[99'Tc]-d,l-HMPAO SPECT study. Neurol NeurosurgPsychiatry 1994;57:285-95,
37 Benson DF, Kuhl DE, Hawkins R, et al. The fluoro-deoxyglucose "IF scan in Alzheimer's disease and multi-infarct dementia. Arch Neurol 1983;40:711-4.
38 Greenberg JO, Shenkin HA, Adam R. Idiopathic normalpressure hydrocephalus-a report of 73 patients. Jf NeurolNeurosurg Psychiatry 1973;40:336-41.
39 Jacobs L, Conti D, Kinkel WR, Manning EJ. "Normal-pressure" hydrocephalus. Relationship of clinical andradiographic findings to improvement following shuntsurgery. JAMA 1976;235:510-2.
40 Fisher CM. The clinical picture in occult hydrocephalus.Clin Neurosurg 1977;24:270-84.
41 Fisher CM. Hydrocephalus as a cause of disturbances ofgait in the elderly. Neurology 1982;32:1358-63.
42 Meyer JS, Tachibana H, Hardenberg JP, et al. Normalpressure hydrocephalus. Influences on cerebral hemody-namic and cerebrospinal fluid pressure-chemical autoreg-ulation. SurgNeurol 1984;21:195-203.
43 Larsson A, Wikkelso C, Bilting M, Stephensen H. Clinicalparameters in 74 consecutive patients shunt operated fornormal pressure hydrocephalus. Acta Neurol Scand199 1;84:475-82.
44 Albech M, Gjerris F, B0rgesen SE, et al. The resistance tocerebrospinal fluid outflow in healthy volunteers.Neurosurg 1991;74:597-600.
45 Coblentz JM, Mattis S, Zingesser LH, et al. Preseniledementia. Clinical aspects and evaluation of cerbrospinalfluid hydrodynamics. Arch Neurol 1973;29:299-308.
46 Sohn RS, Siegel BA. Gado M, Torack RM. Alzheimer'sdisease with abnormal cerebrospinal fluid flow. Neurology1973;23: 1058-65.
50 Bradley WG, Whittemore AR. Watanabe AS, et al.Association of deep white matter infarction with chroniccommunicating hydrocephalus. Implications regardingthe possible origin ofnormal pressure hydrocephalus. Am_JNeuro Radiol 1991;12:31-9.
51 Kirkpatrick JB, Hayman LA. White matter lesion in MRimaging of clinically healthy brains of elderly subjects:possible pathologic basis. Radiology 1987;162:509-1 1.
615
on August 29, 2020 by guest. P
rotected by copyright.http://jnnp.bm
j.com/
J Neurol N
eurosurg Psychiatry: first published as 10.1136/jnnp.59.6.608 on 1 D
