J7ournal ofNeurology, Neurosurgery, and Psychiatry 1994;57:1047-1056 1047 J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.57.9.1047 on 1 September 1994. Downloaded from Multiple system atrophy: natural history, MRI morphology, and receptor imaging with 123IBZM-SPECT

J B Schulz, T Klockgether, D Petersen, M Jauch, W Muiller-Schauenburg, S Spieker, K Voigt, J Dichgans

Abstract of the intermediolateral cell columns of the Sixteen patients with a clinical diagnosis spinal cord. All of these changes are not neces- of probable multiple system atrophy sarily found in each patient with MSA. In (MSA) were examined clinically by MRI some patients, degeneration affects mainly the and by 523I-iodobenzamide single photon basal ganglia, whereas in others the ponto- emission computed tomography (IBZM- cerebellar system or the spinal cord are more SPECT). The clinical records of another affected. Therefore, evidence of cell degenera- 16 patients were also analysed retrospec- tion and gliosis in at least three of the following tively. On the basis of their clinical pre- anatomical structures is thought to be suffi- sentation, patients were subdivided into cient for the neuropathological diagnosis of those with prominent parkinsonism MSA: striatum, substantia nigra, Purkinje (MSA-P, n = 11) and those with promi- cell layer of the cerebellar cortex, pontine nent cerebellar ataxia (MSA-C, n = 21). nuclei, inferior olives, and intermediolateral Autonomic symptoms were present in all cell columns of the spinal cord.'4 Striatum or patients and preceded the onset of motor substantia nigra must be one of these areas.5 symptoms in 63% of patients. Calculated By contrast with idiopathic Parkinson's syn- median lifetime and the median time to drome, there is no Lewy body pathology in become wheelchair bound after onset of MSA. Rather, a recent study had reported the disease were significantly shorter for presence ofglial cytoplasmatic inclusion bodies MSA-P than for MSA-C (lifetime: 4 0 v in MSA.6 The specificity of this finding 9-1 years; wheelchair: 3-1 v 5'0 years) remains to be proved in further studies. suggesting a better prognosis for cerebel- The clinical presentation of patients with lar patients. A significant loss of striatal MSA is highly variable, at least at the onset of dopamine receptors (below 2 SD thresh- the disorder: patients may present with old) was detected by IBZM-SPECT in parkinsonism, with a cerebellar syndrome, 63% of the patients (56% below 2 5 SD with progressive autonomic failure, or with threshold). There was no difference variable combinations of these syndromes. between patients with MSA-C and those Usually parkinsonian symptoms respond only with MSA-P in the proportion with sig- poorly to levodopa. Only a minority of nificant receptor loss and the extent of patients with MSA are correctly diagnosed dopamine receptor loss. Planimetric during life due to equivocal clinical sympto-

MRI evaluation showed cerebellar and matology and the lack of definite diagnostic http://jnnp.bmj.com/ brainstem atrophy in both groups. tests. Accordingly, some patients with MSA Atrophy was more pronounced in are simply diagnosed as having Parkinson's Department of patients with MSA-C than in those with disease. Other patients are categorised under Neurology MSA-P. Pontocerebeliar hyperintensities the umbrella terms Parkinson-plus syndrome J B Schulz and putaminal hypointensities on T2 or atypical parkinsonism. Patients with a com- T Klockgether S Spieker weighted MRI were found in both groups. bination of parkinsonism and autonomic fail- J Dichgans Pontocerebellar signal abnormalities ure are often referred to as having Shy-Drager on October 1, 2021 by guest. Protected copyright. Department of were more pronounced in MSA-C than in syndrome. On the other hand, MSA may be Neuroradiology MSA-P, whereas the rating scores for the underlying pathology in patients with D Petersen area but not for intensity of putaminal prominent cerebellar ataxia who received the K Voigt abnormalities were higher in MSA-P. diagnosis of sporadic olivopontocerebellar Division ofNuclear Medicine, MRI and IBZM-SPECT provide in vivo atrophy or idiopathic cerebellar ataxia during Department of evidence for combined basal ganglia and life. In an attempt to form a rational basis for Radiology, University pontocerebellar involvement in almost all the intra vitam diagnosis of MSA, Quinn7 pro- ofTubingen, Hoppe- patients in this series. Seyler-Strasse 3, posed clinical criteria for the diagnosis of D-72076 Tubingen, probable MSA; these include the presence of Federal Republic of (T Neurol Neurosurg Psychiatry 1994;57: 1047-1056) parkinsonism or cerebellar ataxia and severe Germany M Jauch symptomatic autonomic failure without W Muller-Schauenburg dementia and downgaze palsy. Dependent on Correspondence to: The term multiple system atrophy (MSA), whether parkinsonian or cerebellar symptoms Dr T Klockgether, first used by Graham and Oppenheimer in predominate the diagnosis is parkinsonian Department of Neurology, University of Tubingen, 1969,1 refers to a sporadically occurring adult (MSA-P) or cerebellar (MSA-C) (probable) 72076 Tubingen, Germany. onset neurodegenerative disease. Neuro- MSA. Received 29 March 1993 pathologically, MSA is characterised by the In recent years, modem imaging methods and in final revised form 15 November 1993. presence of olivopontocerebellar atrophy, such as MRI,'" SPECT," '3 and PET'4 15 Accepted 26 November 1993 striatonigral degeneration, and degeneration have increasingly been used in neurodegenera- 1048 Schulz, Klockgether, Petersen, Jauch, Midler-Schauenburg, Spieker, Voigt, Dichgans J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.57.9.1047 on 1 September 1994. Downloaded from tive diseases. Although these methods have standing. Orthostatic reduction in systolic also been applied to patients with MSA, their blood pressure of 30 mm Hg or more without diagnostic value in MSA remains uncertain. a sufficient increase of heart rate was consid- The reported abnormalities, such as putami- ered orthostatic hypotension. On the day of nal iron accumulation or loss of dopamine the clinical investigation MRI and IBZM- receptors, are found in other parkinsonian SPECT were performed. Two patients in syndromes. Also, most studies are restricted whom MRI could not be performed due to to patients with MSA who have parkinsonian claustrophobia or a metallic implant received symptomatology, whereas patients with cere- a high resolution cranial CT examination. bellar symptoms have attracted much less Those patients with probable MSA who attention. were not seen personally (n = 16) were In this study, we report clinical, MRI, and included only in the outcome evaluation. For "23I-iodobenzamide (IBZM) single photon these patients, age at onset of first notable SPECT data from a series of 16 patients with motor symptom, clinical type (parkinsonian v probable MSA of either parkinsonian (MSA- cerebellar), duration after onset of motor P) or cerebellar (MSA-C) type. The aim of symptoms until they became wheelchair the present study was twofold. Firstly, we bound, and age at death were determined wished to study the natural history and prog- from the clinical records. In many patients nosis of MSA and to compare MSA-C and autonomic symptoms antedated motor symp- MSA-P. Secondly, we wished to study the toms by several years. Autonomic symptoms abnormalities detected by MRI and IBZM- were often vague and of undetermined onset, SPECT in patients with MSA. In particular, however, so that it turned out to be impossi- we were interested in whether these imaging ble to use them for exact temporal definition procedures can identify abnormalities that are of onset of disease in the 16 patients studied common to all patients with MSA or whether retrospectively. they yield different results in cerebellar and parkinsonian patients. MRI MRI was performed with a 1'5 T supercon- ducting system (Magnetom, Siemens AG, Patients and methods Erlangen, Germany) with a circularly PATTENTS polarised head coil. Data were displayed on a The clinical records of all patients with 256 x 256 matrix. All patients were investi- parkinsonian symptoms (n = 850) and all gated by a standard examination programme patients with cerebellar degenerative disease that included sagittal and axial Ti weighted (n = 217) referred to our department between images (spin echo, TR = 600 ms, TE = 15 1983 and 1991 were reviewed. Of these, 32 ms, slice thickness = 4 mm), and axial pro- fulfilled all of the following four diagnostic cri- ton-density and T2 weighted images (dual teria for probable MSA: (a) progressive cere- spin echo, TR= 2100 ms, TE = 45 + 90 bellar ataxia and/or non-responsive or poorly ms). To improve slice profiles two subsequent responsive levodopa parkinsonism without sets of measurements of each sequence were evidence for a focal or non-focal symptomatic made, each with an interslice gap of 4 mm. origin of the disorder; (b) severe symptomatic Both sets of measurements together gave a autonomic failure with at least postural syn- complete series of images without a gap.

cope or presyncope and/or pronounced uri- Special care was taken to position the head http://jnnp.bmj.com/ nary incontinence or retention not due to and neck symmetrically in the head coil. other causes; (c) absence of any neurodegen- Twenty-four age and sex matched healthy erative disorders in relatives and no evidence persons served as controls (mean age 58-0 of consanguinity of parents; (d) absence of (15-6) years). Neurological examination of dementia according to DSM-III R criteria, these persons was normal and routine evalua- generalised tendon areflexia, or predominant tion of their MRIs gave no pathological find- downgaze supranuclear palsy. Based on their ings. clinical presentation, patients were subdivided Quantitative MRI evaluation of cerebel- on October 1, 2021 by guest. Protected copyright. into two groups, one with predominant cere- lum, pons, brainstem, and spinal cord was bellar symptoms (MSA-C, n = 21) and performed under blinded conditions. Regions another with predominant parkinsonian pre- of interest of any form, location, and size sentation (MSA-P, n = 1 1). could be outlined with a computer program 16 patients were personally interviewed and (fig 3). For quantitative planimetric evalua- clinically examined by one of us (JBS) with a tion of areas we used an image analyser and a standardised examination procedure. Their computerised interpretation program.20 21 This mean age was 60-1 (SD 8-3) years. Severity of technique permits measurements of the cerebellar symptoms was rated on a scale amount of tissue within a region of interest ranging from 0 (absent) to 5 (most severe).'6 without lining out the border between tissue We also used the Hoehn and Yahr scale17 and and CSF by hand. Furthermore, it takes into the unified Parkinson's disease rating scale account partial volume effects that often (UPDRS)'8 for rating the severity of parkin- occur at the surface of the cerebellum. For sonian patients, and the Kurtzke EDSS scale'9 each case the individual signal intensities of for estimation of global disability state. CSF and of the cerebellar tissue were defined. Clinical examination included measurement Those measurements gave two sets of normal of blood pressure and heart rate in the supine distributed brain tissue and CSF grey values. position and one and five minutes after active Pixels with values equal or higher than the Multiple system atrophy 1049 J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.57.9.1047 on 1 September 1994. Downloaded from mean tissue value were counted as tissue, middle cerebellar peduncles22 were rated to be whereas pixels with values equal or lower than absent (grade 0), mild (grade 1), moderate the mean CSF value were counted as CSF. (grade 2), or strong (grade 3). The remaining pixels were designated as con- taining both tissue and CSF. It was assumed SPECT that the difference in signal intensity between IBZM-SPECT was performed two hours after grey and white matter could be ignored for intravenous injection of 185 MBq 123IBZM this purpose. Those voxels with signal intensi- (3-iodo-6-methoxybenzamide; Cygne BV, ties between the mean values of brain tissue Netherlands/Du Pont de Nemour, Germany). and CSF were linearly weighted according to If patients were taking medication, this was their tissue content. This allowed us to finally stopped 24 hours before SPECT scanning. summarise the number of effective tissue pixels SPECT data were acquired with a dual head and to obtain corrected planimetric values for rotating gamma Camera (Picker Digital cerebellar structures (see fig 3). To compen- Dyna, AP Collimator) and a PCS 512 com- sate for individual variations in head size we puter system (gamma 11). For data acquisi- measured the total area of the posterior fossa tion the following standard programme was on the midsagittal slice and related each plani- used: two complete circles (360° each camera metric measurement to the respective poste- head), 64 frames/circle, 64 x 64 matrix, slice rior fossa area. There were no statistically thickness 5-5 mm, pixel resolution 5-5 mm, significant differences in the area of the poste- full width at half maximum (FWHM) of the rior fossa between controls and patients with reconstructed line spread function 20 mm. MSA (mean (SD); controls 4922 (379) pixels; The acquisition time was 40 s/frame. MSA-P 4910 (445) pixels; MSA-C: 4834 We superimposed MRI and SPECT with (480) pixels). external landmarks. The orbital meatal line, The areas of the cerebellar vermis and the as shown by a laser system on the SPECT sys- basis pontis were measured on the midsagittal tem, was marked by a pencil line and oil filled slice. As we obtained a complete series of tubes. This line represents the collimator mid- images without a gap, the midsagittal slice was line and the first MRI image. After prefiltering easily identifiable by showing the cerebral the raw data (METZ filter) and overlaying a aquaeduct in its complete craniocaudal RAMP filter, SPECT reconstruction was extent. The basis pontis was delineated as an done for axial, coronal, and sagittal slices elliptical structure: the ventral border was including attenuation correction with an obvious and the dorsal border was determined attenuation coefficient of 0 139 cm-'. Images by the medial lemniscus, which appeared as a from MRI (256 x 256, 16 bit count depth) narrow strip of reduced signal intensity. The were transmitted via ethernet, converted into areas of the cerebellar hemispheres were cal- gamma 11 file format (256 x 256, 8 bit count culated on the first parasagittal slice lateral to depth), and internally reduced to 128 x 128 the lateral border of the middle cerebellar (8 bit count depth) for further procedures. peduncle. As there were no systematic size After landmark based slice identification and differences between the hemispheres of the scale adaptation by zooming, the in plane two sides, we calculated the mean of both matching was done under visual control, with a sides for further analysis. We also measured region of interest contour shifting as an inter- the area of medulla oblongata in the horizontal mediate step. To identify the correct anatomi-

plane at the level of the inferior olivary com- cal sites axial SPECT and MRI images were http://jnnp.bmj.com/ plex and the cervical spinal cord on the hori- superimposed adding then two to three con- zontal plane at the level of the dens. Also the secutive SPECT images. On these sum diameter (mm) of the middle cerebellar images, basal ganglia and prefrontal regions of peduncle was measured in a horizontal plane interest were drawn. The size of the regions of with a distance algorithm. interest varied from patient to patient and The degree of cortical atrophy and the ranged from 20 to 30 pixels (6 to 9 cm2). Care severity of signal abnormalities on T2 was taken that the striatal region of

interest, on October 1, 2021 by guest. Protected copyright. weighted images were rated independently by which had an elongated shape, was clearly two of us (JBS, DP). To ensure uniform and within the borders of the striatum, as shown anonymous quality of the images, all images by the superimposed MRI, and that the stri- were documented on x ray films without any atal and prefrontal regions of interest of an identification data. The degree of cortical individual patient had comparable sizes. The atrophy and hypointensities of the putamen, prefrontal region of interest had a band-like pallidum, substantia nigra, and red and cere- shape following isocontours. The regions of bellar dentate nucleus were rated on a scale interest were used for calculating counts per ranging from 0 to 3. Grade 0 indicated com- pixel in the basal ganglia and prefrontal cor- plete absence, grade 1 mild, grade 2 moder- tex. The ratio of these values (BG/FC ratio) ate, and grade 3 strong atrophy or then represents a semiquantitative measure of hypointensity respectively. The area of signal the relative density of basal ganglia dopamine hypointensity in the putamen was also rated. receptors labelled by IBZM.23 Grade 0 indicated no hypointensity, grade 1 Because we did not obtain the permission hypointensity limited to the lateral margin of of the local ethics committee to perform the putamen, grade 2 hypointensity extending IBZM-SPECT in healthy controls, we com- through part of the body of the putamen, and pared the results of the patients with MSA grade 3 hypointensity throughout the entire with a group of untreated de novo patients putamen." Hyperintensities of the pons and with idiopathic Parkinson's syndrome (Hoehn 1050 Schulz, Klockgether, Petersen, J7auch, Muiller-Schauenburg, Spieker, Voigt, Dichgans J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.57.9.1047 on 1 September 1994. Downloaded from Figure 1 Survival the BG/FC ratio (1-76 (SD 0 1)) in our idio- function (A) andffunction to become wheelchair pathic Parkinson's syndrome group was iden- bound (B) ofpatients with tical to that reported by Brucke and MSA-C and MSA-P, as colleagues'2 in their healthy controls. calculated by the Kaplan- Meier life table method.23 Squares represent patients STATISTICAL ANALYSIS with MSA-P(n = 11), Parametric variables (age, quantitative mor- triangles patients with 4) 4 phometric MRI data of infratentorial struc- MSA-C (n = 21). Filled symbols represent -z tures, and IBZM-SPECT data) were completed cases, who died L- statistically analysed by an analysis of variance (A) or became wheelchair C/U for each of the dependent variables. The bound (B); open symbols censored cases, who were , analysis of significant main effects was per- lost duringfollow up or formed by Duncan's multiple range test. have not died or become Ordinal variables (MRI rating scores) were wheelchair bound. analysed by Kruskal-Wallis one way analysis of variance by ranks followed by the Mann- Whitney U test with Bonferroni adjustment B according to the number of tests performed. 100 Survival analysis was performed with the 4- method described by Kaplan and Meier.24 .0 E 80 co

0. c Results .' 60 Eleven out of 32 patients with MSA were E a) women and 21 were men. The mean age at o 40 the onset of motor symptoms was 54-1 (SD 7X1); range 40 to 66 years), with no differ- 52X8 .1 ences between the subgroups (MSA-C ~0> 20 years; MSA-P 55-7 years). The calculated median lifetime24 after onset of motor symp-

o o toms was 8-9 years. It was significantly shorter C 0 L 2 4 6 8 in MSA-P (4 0 years) than in MSA-C (9 1 Latency from onset of symptoms (y) years; fig 1A). Similarly, the calculated median time to become wheelchair bound was and Yahr stage I and II; mean age 53-8(10-0) shorter in MSA-P (3 1 years) than in MSA-C years). All patients with idiopathic (5 0 years; fig 1B). Parkinson's syndrome underwent an apomor- Tables 1 and 2 give the clinical data of the phine test and were subsequently treated with 16 personally examined patients. The mean levodopa. Only those patients were included duration of motor symptoms was shorter in who had a positive test and MSA-P (mean (SD) 3 0 (2X1) years) than in showed pronounced clinical improvement MSA-C (6-6 (2-6); p < 0 05). All patients had after a three month treatment with levodopa. dysautonomia. The most frequent autonomic We believe that it is justified to use levodopa symptoms were orthostatic hypotension

responsive patients with idiopathic (94%), and urinary incontinence or retention http://jnnp.bmj.com/ Parkinson's syndrome as controls because two (94%). All male patients complained of erec- earlier studies'2 '3 failed to show any differ- tile dysfunction. Another common sign was ences of the BG/FC uptake ratios of IBZM in vocal cord dysfunction (81%) with nocturnal healthy persons and untreated patients with stridor or hoarseness. Three patients required idiopathic Parkinson's syndrome. Moreover, tracheotomy. In one patient a combined cen-

Table 1 Clinical data ofpersonally examined patients on October 1, 2021 by guest. Protected copyright. Age of Duration of Treatment onset First symptoms (y) Parkinsonian Cerebellar Pyramidal Kurtzke Levodopa at the time of Case Sex (y) symptom MO AF symptoms symptoms signs score response IBZM-SPECT MSA with predominant parkinsonian symptoms (MSA-P): 1 M 43 Erectile dysfunction 1 6 HRS AS and G Yes 3 0 NT None 2 M 56 VCP 1 2 HRS AS and G No 4 0 Poor LD 375, Dep, Brom 3 M 53 Erectile dysfunction 2 2 HRS AS and G, AL Yes 2-0 Poor None 4 M 66 Erectile dysfunction 2 20 HRS, Tr AS and G, AL No 4 0 Poor LD 375, Dep, Mem 5 F 58 Urinary dysfunction 4 5 HRS, AC No Yes 9 0 Poor LD 625, Dep, Uis 6 F 55 Bradykinesia 4 1 HRS, AC No No 9-5 Poor LD 500, Dep, Lis 7 M 59 Erectile dysfunction 7 10 HRS AS and G, AL No 9 5 Poor LD 375, Dep, Mem MSA with predominant cerebellar symptoms (MSA-C): 8 M 59 Erectile dysfunction 2 4 No AS and G, AL Yes 3 5 NT None 9 F 49 Sleep apnoea syndrome 4 6 No AS and G, AL No 6-0 NT None 10 M 55 AS and G 7 4 HRS AS and G, AL No 7-5 NT None 11 M 51 AS and G 5 1 HRS AS and G, AL No 5-0 NT None 12 M 44 AS and G 6 6 No AS and G, AL Yes 7-0 NT None 13 M 40 Erectile dysfunction 7 8 No AS and G, AL Yes 8-5 NT None 14 M 56 VCP 9 10 HRS AS and G, AL No 8-0 None None 15 M 62 AS and G 9 5 HRS AS and G, AL No 8-5 NT None 16 F 59 AS and G 10 2 HRS, Tr AS and G, AL Yes 9 0 NT None VCP = vocal cord paralysis; AS and G = ataxia of stance and gait; MO = motor; AF = autonomic failure; HRS = hypokinetic-rigid syndrome; Tr = tremor; AC = antecollis; AL = ataxia of limbs; NT = not tested; LD = levodopa + dopa decarboxylase inhibitor (daily dose in mg); Dep = L-deprenyl; Brom = ; Lis = , Mem = . Multiple system atrophy 1051 J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.57.9.1047 on 1 September 1994. Downloaded from Table 2 Details ofautonomic failures in personally examined patients Orthostatic Erectile Urinary Case hypotension dysfunction dysfunction Stridor Anhidrosis Diarrhoea 1 + Yes R, I NS No + 2 + Yes R, SUC T No No 3 + Yes R, I No No No 4 + + Yes R, I No No No 5 + + * I,SUC NS No No 6 + + * R, I, SUC T No No 7 + + Yes R, I, SUC NS No No 8 (+) Yes No NS No No 9 (+) * I SAS No No 10 + Yes R NS No + 11 + Yes R, SUC NS + No 12 No Yes I NS + No 13 + + Yes R, SUC NS + No 14 (+ ) Yes I T + No 15 + + Yes R, I, SUC NS No No 16 + * I, SUC No No No Orthostatic hypotension: (+) = orthostatic reduction in systolic blood pressure by 30 mmHg or more, + = symptomatic hypotension with presyncope, + + = symptomatic hypotension with syncope; erectile dysfinction: *= female patient; urinary dysfunction: R = retention, I = incontinence, SUC = suprapubic urinary catheter; stridor: NS = nocturnal stridor, T = tracheotomy after bilateral vocal cord paralysis, SAS = sleep apnoea syndrome.

tral and obstructive sleep apnoea syndrome in MSA-P (mean (SD) 50 (24) mm Hg) than was diagnosed. Disturbances of sweating or in MSA-C (37 (17) mm Hg; p = 023), but bowel function were infrequent. the difference did not reach significance. Motor symptoms occurred in various com- binations. Ten patients had a combination of IBZM-SPECT parkinsonian and cerebellar features (63%), The BG/FC ratio, which serves as a semi- seven without (44%) and three with pyrami- quantitative measure of dopamine receptor dal signs (19%). Three patients had cerebellar binding in the striatum, was significantly and pyramidal symptoms without parkinson- reduced in MSA (1-50 (0d16) compared with ism (19%). A pure cerebellar syndrome, pure idiopathic Parkinson's syndrome (1-76 parkinsonism, or a combination of parkinson- (010)). There were no differences between ism and pyramidal signs occurred in one MSA-C (1.49 (0-14) and MSA-P (1-52 patient each. In all patients with MSA-P, the (0.20)). Three of the seven patients with first motor symptom was bradykinesia, in all MSA-P and seven of the nine patients with patients with MSA-C it was ataxia of stance MSA-C had BG/FC ratios that were more and gait. The first motor symptom, whether than 2 SD below the normal range (fig 2). If a parkinsonian or cerebellar, determined the threshold of 2-5 SD was chosen, three clinical type during the entire course of the patients with MSA-P and only six with MSA- disease. C were outside the normal range. The BG/FC In 10 patients (63%) symptoms of dysau- ratios showed a weak, negative correlation tonomia preceded motor signs (table 2), in with the Kurtzke score (r = -0A49; p = 0-05), some of them by more than 10 years. In the but no significant correlation with the scores remaining six patients in whom the disease rating severity of parkinsonism. They were

started with parkinsonism or ataxia, auto- not significantly correlated with the rating http://jnnp.bmj.com/ nomic symptoms occurred within one to eight scores for putaminal hypointensities. years after the onset of motor symptoms. The serial order in which autonomic symptoms MRI occurred was random. Autonomic symptoms All measures of cerebellar and brainstem size were similar in both groups and their duration (fig 3) were significantly different by analysis did not differ significantly between groups of variance between the patient and control (mean (SD) MSA-P 6-6 (6 7) years; MSA-C groups (cerebellar vermis: F2,,6 = 58-7, hemi- 5*1 (2 8) years). The orthostatic reduction in spheres: F2,36 = 36-7, middle cerebellar on October 1, 2021 by guest. Protected copyright. systolic blood pressure was more pronounced peduncle: F2,36 = 79-6, basis pontis: F,,36 = 57-2, medulla oblongata: F2,6 = 20'7; p < 0-0001), whereas there were no differences in Figure 2 IBZM binding spinal cord size and the estimated degree of as detected by SPECT in 2.0 A--L cortical atrophy (controls: median 0, range 0 idiopathic Parkinson's to 2; MSA-P 0'5, range 0 to 1; MSA-C 0, syndrome (IPS), MSA-C, and MSA-P. BGIFC ratio A range 0 to 2). Further analysis of different 1.8 - AA v = basal ganglia to frontal -AA--A v main effects by Duncan's multiple range test cortex ratio ofIBZM 0 A A 0 -v showed that the size of cerebellar and brain- binding. Solid line A 0 v stem structures was reduced in the MSA-P represents median ofIPS _ dashed line 2 SD. ) 1.6 A and MSA-C groups compared with controls group, LL ------__ _------El-- Filled symbols represent o A (p < 0.001) and that the amount of reduction median value ofeach v group. Shaded symbols was greater in MSA-C than in MSA-P (fig 4). represent individual 1.4 _ On average, the middle cerebellar peduncle patients with MSA-C showed the most pronounced size reduction without parkinsonism of all infratentorial brain structures (fig 3). Its (n = 4) or patients with MSA-P without ataxia 1.2 diameter was reduced by more than 2 SD (n = 2) respectively. IPS MSA-C McSA-P compared with controls in all but two patients 1 052 Schulz, Klockgether, Petersen, Jauch, Muller-Schauenburg, Spieker, Voigt, Dichgans J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.57.9.1047 on 1 September 1994. Downloaded from

_0.g~~~~~~~~~ 0-s* TXT * ;t T * * rTA '*}t'N5t _ '_~o+_1UV ~ ~ - T t:?v t id~~~~06 L0 0T Co401(

_U Ver Vermis CH MCP BP MO CSC Figure 4 MRI morphology inAMSA-C and MSA-P. Sizes of anatomical structures were determined planimetrically and are given as percentages of the respective control values (mean (SD)). Vermis = anterior and posterior cerebellar vermis; CH = cerebellar 1 hemispheres; M'ICP = middle cerebellar peduncles; BP = basis pontis; MO = medulla oblongata; CSC = cervical spinal cord. Comparison ofMSA-P and MSA-C was by one way analysis of vanrance and Wr . gsubsequentt il ! Duncan's multiple range test: **p <0 01; ***p < 0*001.

ffit--F1Analysis of the degree (H2,27 = 104) and area (H2,27 = 8 4) of putaminal hypointensitiei "'''!A~'-v:v;.g,-g>'A....-f'-' identified the presence of significant group effects (p < 0-01). The rating scores for th: 09 degreev ...... of putaminal hypointensities in MSA- C (median 2, range 1 to 3) and MSA-P (median 2, range 2 to 3) were significantly higher than those of controls (median 0, range 0 to 2; p < 0-05; figs 5 and 6A). The area of the hypointensities was greater in MSA-P (median 2, range 1 to 3) than in MSA-C 1 to < 1W _ g -(median_ 1-5, range 2; p 0-05; fig 6B) and both groups differed significantly from controls (median 0, range 0 to 2; p < 0-05). No significant differences in the signal intensi- 4 ties of the pallidum, red nucleus, substantia nigra, or dentate nucleus were noted. Putaminal hypointensity exceeding hypo-

intensity of the pallidum has been proposed as http://jnnp.bmj.com/ an easily accessible indicator of putaminal abnormality.9102526 This was true in 12 of 14 r_s s s patients with MSA, but also in five of 14 con- trol subjects. The rating scores of putaminal hypointensities were not correlated with the Kurtzke scale, the diameter of the middle cerebellar peduncle, or the BG/FC ratio from rs /E IBZM-SPECT.; on October 1, 2021 by guest. Protected copyright. Significant group effects (p < 0-001) for hyperintensities within the pons (H2,27 1 6-5) , '~: ~ and middle cerebellar peduncle (H2,2 = 21 0) were identified. Subsequent mean contrasts Figure 3 MRI (Ti weighted spin echo images l ST, TR =600 ms, TE =15ms,sl showed that both patient groups had higher thickness = 4 mm) of infraten,rtoial brain structures showing the posteriorfossa in the scores than controls and that patients with midsagittal plane (upper left) aznd axial images at the level of the middle cerebellar MSA-C had higher scores than patients with peduncles (upper right), inferioor olivary complex (lower left) and dens axis (lower right). MSA-P (figs 6B and 7). (A)MRI of a 34 year-old healtthy male. The dashed lines outline the size of the postenror fossa, the anterior and posteriorr vermis, the ventral part of the pons, the maximum diameter of the middle cerebella'r peduncles and the area of medulla oblongata and the cervical spinal cord as describecI in the text. (B) MRI of a 69 year oldfemale patient (case Discussion 16) with MSA-C. There is severe atrophy of the anterior and posterior cerebellar vermis, MSA is principally a neuropathological entity the basis pontis, the middle cere?bellar peduncles, and the medulla oblongata. encompassing striatonigral degeneration, with MSA. The diameter was negatively cor- olivopontocerebellar atrophy, and degenera- related with the Kurtzke score (r = -0-71; p tion of the intermediolateral cell columns of = 0-004); there was no correlation with the spinal cord.1 Clinically patients with MSA BG/FC ratios of IBZM-SPECT or rating present with severe symptomatic autonomic scores for putaminal hypointensities. failure and parkinsonian or cerebellar symp- Multiple system atrophy 1053

toms. This has led to the clinical distinction J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.57.9.1047 on 1 September 1994. Downloaded from between striatonigral degeneration (parkin- sonian) and olivopontocerebellar atrophy (cerebellar) type.7 A major problem in clinical studies of MSA is the uncertainty of the clini- ., cal diagnosis. A definite neuropathological j diagnosis was made in only two of our patients who died in the course of the study. ijF Although we have used stringent clinical inclusion criteria, some of our patients may have other degenerative diseases. For exam- ple, parkinsonism in combination with auto- nomic failure may also occur in idiopathic Parkinson's disease.27-29 On the other hand, patients with cerebellar ataxia or parkinson- ism without severe dysautonomia were not included in the present study although they * may have unrecognised MSA. Most recent studies of MSA come from researchers with a special interest in Parkinson's disease.8914153>2 This has proba- bly led to an over-representation of patients 2^'9 with parkinsonian MSA in clinical and neu- ropathological studies of patients with MSA. In the present study, both patients with pre- dominant parkinsonian and predominant cerebellar features were included allowing us -q~~~~~~~~~~~~~~7 e to compare their clinical characteristics, prog- Figure 5 Hypointensities of basal ganglia in MRI (T2 weighted spin echo imag,es 1 ST, nosis, and results of imaging procedures. TR = 2100 ms, TE = 45 + 90 ms, slice thickness = 4 mm). Upper left: 53y,ear old In all patients, the initial motor symptom, healthy male. Normal distribution ofhypointensities in the basal ganglia. Absence of whether parkinsonian or cerebellar, deter- hypointensity in putamen, normal hypointensity in pallidum (putaminal rating sccores: intensity 0, area 0). Upper right: 50 year old male patient with MSA-C (case 12 mined the clinical type in the entire course of Hypointensity at the dorsolateral margin of the putamen (intensity 2, area 1). Lc mwer left: the disease. Nevertheless, with disease pro- 68 year old male patient with MSA-P (case 4). The strong hypointensity extends through gression an overlap of parkinsonian and cere- part ofthe body ofputamen (intensity 3; area 2). Lower right: 66 year old male MSA-P (case 7). Hypointensity extending throughout the entire putamen and inwtesity bellar symptoms occurs. Pyramidal symptoms exceeding that in globus paUlidus (intensity 3; area 3). Note the hyperintensity at the may add to the complex clinical picture in lateral putaminal border. MSA, but are usually not predominant fea- tures. By taking a detailed autonomic history we found that autonomic symptoms, in par- Figure 6 Severity of ticular erectile dysfunction, preceded the putaminal hypointensities onset of motor symptoms in more than 60% (A) and pontine and of our patients, in some of them by more than cerebellar hyperintensities C ._ (B) on T2 weighted MRIs. 10 years. Although it is impossible to prove a) Degree ofsignal that these often vague autonomic symptoms http://jnnp.bmj.com/ abnormalities was rated C c 2 * occurred as a result of neuronal degeneration independently by two examiners according to the Ca) in the context of MSA, this finding suggests criteria outlined in the 0 that MSA may have a presymptomatic phase methods. Median values of several years. are given. In a review 213 *p <005; **p > 0-01 v 0(>4 1 of pathologically estab- controls, Kruskal- Wallis lished cases of MSA Quinn32 reported a one way analysis of CC median survival after onset of first symptoms variance by ranksfollowed on October 1, 2021 by guest. Protected copyright. by the Mann-Whitney U of five to six years, whereas we found a longer test. survival of about nine years in our patients. Putamen PutamEen Possible explanations for the difference intensity area between the study of Quinn and the present one is that all of Quinn's patients were diag- 3 - B nosed by pathology, whereas we may have C ._ included patients with different disorders, 0 such as idiopathic Parkinson's disease. It is * 4- _ * * also difficult to pinpoint precisely onset of dis- a) ease in many patients. More important is the a) finding that calculated median lifetime and C CD the median time to become wheelchair bound were significantly shorter in MSA-P than

4- a MSA-C. These results imply that the progno- * sis is better in MSA-C than in MSA-P and cc justify the division into MSA-C and MSA-P at least for clinical and prognostic purposes. Pons Middle cere-bellar The reasons why parkinsonian patients have a pedunc-le worse prognosis are unclear. Possibly, parkin- 1054 Schulz, Klockgether, Petersen, jauch, Midler-Schauenburg, Spieker, Voigt, Dichgans

the synaptic cleft.3536 A further consideration J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.57.9.1047 on 1 September 1994. Downloaded from is whether our SPECT results are altered by regional atrophy of the striatum or the pre- frontal cortex. Our method of data evaluation leads to an estimation of the relative density of striatal receptors labelled by IBZM (as com- pared with the prefrontal cortex) rather than to an estimation of total basal ganglia radioac- tivity. Therefore, shrinkage of the striatum itself would not lead to a reduced BG/FC ratio. On the other hand, cortical atrophy would lead to an understimation of the FC value and an overestimation of the BG/FC ratio. This is improbable in our patients because the MRIs suggest that there is no sig- nificant degree of cortical atrophy in patients with MSA. This is in accord with neuropatho- logical findings showing that degeneration of the cerebral cortex is not a typical neu- ropathological feature of MSA.7 Although there was a statistically significant reduction of IBZM binding in patients with MSA compared with controls, only 63% of our patients had BG/FC ratios that were more than 2 SD below the normal ratio (56% below the 2-5 SD threshold). These results closely correspond to those of postmortem receptor binding studies37 38 and to a recent PET study of MSA using "IC- to label Figure 7 MRI (T2 weighted spin echo images 1-5T, TR = 2100 ms, TE = 45 + 90 ms, slice thickness = 4 mm) of upper pons (left side) and middle cerebellar peduncles (right dopamine receptors."5 Although only patients side). The upper panel shows images of a 53 year old healthy male without signal with parkinsonian MSA were included in this abnormalities. The lower panel shows hyperintensities of the transverse pontinefibres study the authors found a modest reduction of between tegmentum and the base ofpons (left) and in the middle cerebellar peduncles (right) in a 53 year oldfemale patient with MSA-C. D2 receptor binding. We found, surprisingly, that the proportion of patients with pathologi- cal IBZM-SPECT and the extent of receptor loss were not different between MSA-C and MSA-P and were sometimes greater in MSA- sonian symptomatology and the more severe P. In particular, individual patients without orthostatic hypotension in patients with any parkinsonian symptoms may have a MSA-P are the cause of earlier immobility. pathological BG/FC ratio in IBZM-SPECT, Loss of striatal neurons is an integral part of whereas IBZM binding may be normal in the degenerative process in MSA.1432 other patients with predominant parkinson- Dopamine D2 receptors are located on striatal ism. These findings indicate that striatal interneurons and on a subset of striatal pro- degeneration occurs in patients with MSA http://jnnp.bmj.com/ jection neurons, mainly on those projecting to independently of their clinical presentation. the external pallidal segment.3 Loss of these It is well known that an extensive loss of receptors may thus be taken as an indicator of striatal neurons, as in the late stages of degeneration of these striatal neurons. Huntington's disease, may lead to parkinson- Dopamine D2 receptors are imaged by IBZM- ian symptoms. The moderate extent of recep- SPECT.'2 23 Pharmacologically, IBZM is a tor loss in MSA and the lack of difference specific dopamine D2 antagonist with high between parkinsonian and cerebellar patients affinity for the receptor. Experimental studies suggest that loss of D2 receptors alone does on October 1, 2021 by guest. Protected copyright. in animals show that IBZM binds to areas in not explain the occurrence of parkinsonian the brain with high concentrations of symptoms in MSA. Indeed, Brooks and col- dopamine D2 receptors.2334 Both in animals leagues14 have reported significantly reduced and in humans, IBZM can be displaced from striatal uptake of the positron emitting tracers its binding sites by compounds that are '8F-dopa and "IC-nomifensine, which are known to specifically interact with dopamine markers for dopamine storage capacity and D2 receptors.'234 Reduction of striatal IBZM integrity of dopamine uptake sites respec- binding should thus reflect a loss of striatal tively. Also, a recent PET study reported a dopamine D, receptors. By contrast with moderate loss of DI receptors in the striatum PET, however, the IBZM-SPECT method of patients with striatonigral degeneration.39 yields only semiquantitative data and the sen- Neuropathological studies show that the palli- sitivity of the method has not yet been clearly dal nuclei also undergo severe degeneration in determined. Also, we need to consider MSA,4 which can result in parkinsonian whether the medication of our patients inter- symptoms.40 These findings suggest that the fered with tracer binding. We believe that this development of parkinsonism poorly respon- is unlikely because medication was stopped 24 sive to levodopa in MSA is due to widespread hours before SPECT scanning, allowing ade- neuronal degeneration within the basal gan- quate time for dopaminergic drugs to leave glia rather than to the loss of a single popula- Multiple system atrophy 1055

tion of neurons as in idiopathic Parkinson's out any cerebellar signs. In particular, shrink- J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.57.9.1047 on 1 September 1994. Downloaded from syndrome. age of the middle cerebellar peduncles was Abnormalities on MRI in MSA include found in all cerebellar patients and in all but infratentorial atrophy and signal abnormalities two parkinsonian patients. In a pathological on T2 weighted images within the putamen study, Feamley and Lees4 found pontocere- and the pontocerebellar system. The infraten- bellar involvement in seven out of 10 cases of torial atrophy in MSA affects the cerebellar striatonigral degeneration. In the remaining vermis, cerebellar hemispheres, middle cere- three, cell counts of the cerebellar cortex and bellar peduncles, pons, and lower brainstem. pontine nuclei were in the low normal range. This pattern of macroscopic changes closely The present finding of distributed imaging corresponds with the changes described in abnormalities in patients with MSA-inde- neuropathological studies of olivopontocere- pendent of their clinical presentation-sup- bellar atrophy.4142 With the same technique of ports the concept that striatonigral in vivo morphometric evaluation we found degeneration and (sporadic) olivopontocere- similar changes in patients with autosomal bellar atrophy are closely interrelated and dominant cerebellar ataxias.43 Thus olivopon- should be combined under the heading of tocerebellar atrophy is not a specific morpho- MSA. logical finding in MSA. The infratentorial It would be desirable to complement clini- atrophy in MSA is associated with signal cal criteria for the diagnosis of MSA by imag- hyperintensities within the pons and middle ing criteria that help to differentiate MSA cerebellar peduncles. These abnormalities, from other movement disorders. By the neu- which were first described by Savoiardo et al'2 ropathological definition of MSA' 7 32 these in a heterogeneous group of cerebellar criteria should include imaging abnormalities patients, most probably reflect degeneration within both the basal ganglia and the ponto- of pontocerebellar fibres. These fibres origi- cerebellar system. Unfortunately, imaging nate in the pontine nuclei, cross the midline procedures to visualise spinal cord involve- by taking a transverse course within the pons, ment in MSA have not yet been developed. and ascend within the contralateral middle The present data show that single criteria, cerebellar peduncle. Our finding that infraten- such as pronounced putaminal hypointensity, torial signal abnormalities are restricted to this decreased IBZM binding, or cerebellar atro- fibre tract, and that the middle cerebellar phy are not sufficient because they lack both peduncles exhibit the largest size reduction, specificity and sensitivity. Further studies may are strong evidence that the degeneration of show whether a particular combination of the pontocerebellar tract represents the core imaging criteria will lead to improved diagnosis pathology of olivopontocerebellar atrophy. of MSA. Several authors have described the occur- rence of putaminal hypointensities in patients 1 Graham JG, Oppenheimer DR. Orthostatic hypotension with parkinson plus syndromes.8-112526 These and nicotine sensitivity in a case of multiple system atro- phy. J Neurol Neurosurg Psychiatry 1969;32:28-34. abnormalities are usually ascribed to accumu- 2 Bannister R, Oppenheimer DR. Degenerative disease of lation of iron, although this view has been the nervous system associated with autonomic failure. Brain 1972;95:457-74. questioned by Brooks and colleagues.44 The 3 Spokes EG, Bannister R, Oppenheimer DR. Multiple sys- finding that putaminal hypointensities occur tem atrophy with autonomic failure-clinical, histo- logical and neurochemical observations on four cases. in patients with parkinson plus syndromes J7 Neurol Sci 1979;43:59-82. with poor response to dopaminergic drugs has 4 Fearnley JM, Lees AJ. Striatonigral degeneration. A clini- copathological study. Brain 1990;113:1823-42. http://jnnp.bmj.com/ led to the suggestion that these changes reflect 5 Gray F, Vincent D, Hauw Jj. Quantitative study of lateral damage to postsynaptic striatal dopamine horn cells in 15 cases of multiple system atrophy. Acta Neuropathol (Berl) 1988;75:513-8. receptors. The lack of correlation between 6 Papp MI, Kahn JE, Lantos PL. Glial cytoplasmatic inclu- putaminal MRI abnormalities and IBZM sions in the CNS of patients with multiple system atro- phy (striatonigral degeneration, olivopontocerebellar binding, however, contradicts this view. By atrophy and Shy-Drager syndrome). Jf Neurol Sci 1989; contrast, our findings suggest that putaminal 94:79-100. 7 Quinn N. Multiple system atrophy-the nature of the iron accumulation and loss of striatal beast. J Neurol Neurosurg Psychiaty 1989;52 (special

dopamine receptors represent pathological supplement):78-89. on October 1, 2021 by guest. Protected copyright. 8 Drayer BP, Olanow W, Burger P, Johnson GA, Herfkens processes that are independent of each other. R, Riederer S. Parkinson plus syndrome: Diagnosis The topography of MRI abnormalities was using high field MR imaging of brain iron. Radiology 1986;159:493-8. related to the clinical symptomatology. 9 Pastakia B, Polinsky R, Di Chiro G, Simmons JT, Brown R, Quantitative and semiquantitative MRI evalu- Wener L. Multiple system atrophy: MR imaging. Radiology 1986;159:499-502. ation gave significant differences between the 10 Savoiardo M, Strada L, Girotti F, et al. MR imaging in MSA-P and the MSA-C groups. Putaminal progressive supranuclear palsy and Shy-Drager syn- drome. J Comput Assist Tomogr 1989;13:555-60. hypointensities were more pronounced in 11 Stern MB, Braffman BH, Skolnick BE, Hurtig HI, patients with predominant parkinsonism, Grossman RI. Magnetic resonance imaging in Parkinson's disease and parkinsonian syndromes. whereas infratentorial atrophy and pontocere- Neurology 1989;39: 1524-6. bellar hyperintensities were more severe in 12 Brucke T, Podreka I, Angelberger P, et al. Dopamine D2 receptor imaging with SPECT: Studies in different patients with predominant cerebellar ataxia. neuropsychiatric disorders. J Cereb Blood Flow Metab Putaminal or infrantentorial MRI changes 1991;11:220-8. 13 Schwarz J, Tatsch K, Arnold G, et al. '23I-iodobenzamide- were, however, not confined to one group of SPECT predicts dopaminergic responsiveness in patients patients. There were patients without obvious with de novo parkinsonism. Neurology 1992;42:556-61. 14 Brooks DJ, Salmon EP, Mathias CJ, et al. The relationship parkinsonian signs who had pronounced puta- between locomotor disability, autonomic dysfunction, minal hypointensity; on the other hand, and the integrity of the striatal dopaminergic system in patients with multiple system atrophy, pure autonomic infratentorial atrophy or pontocerebellar failure and Parkinson's disease, studied with PET. Brain hyperintensities were present in patients with- 1990;113: 1539-52. 1056 'Schulz, Klockgether, Petersen, jauch, Miller-Schauenburg, Spieker, Voigt, Dichgans J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.57.9.1047 on 1 September 1994. Downloaded from 15 Brooks DJ, Ibanez V, Sawle GV, et al. Striatal D2 receptor clinicopathological studv. Neurol Neurosurg Psychiatry status in patients with Parkinson's disease, striatonigral 1992;55: 181-4. degeneration, and progressive supranuclear palsy, 32 Quinn N. Multiple system atrophy. In: Marsden CD, measured with llC-raclopride and positron emission Fahn S, eds. Movement disorders III. London. tomography. Ann Neurol 1992;31:184-92. Butterworths, 1994:262-87. 16 Klockgether T, Schroth G, Diener H-C, Dichgans J. 33 Harrison MB, Wiley RG, Wooten GF. Changes in D2 but Idiopathic cerebellar ataxia of late onset: natural history not D, receptor binding in the striatum following selec- and MRI morphology. _7 Neurol Neurosurg Psychiatnv tive lesion of striatopallidal neurons. Brain Re, 1990;53:297-305. 1 992;590:305-10. 17 Hoehn MM, Yahr MD. Parkinsonism: onset, progression, 34 Kung HF, Guo Y-Z, Billings JB, et al. Preparation and and mortality. Neurology 1967;17:427-42. biodistribution of 1251-IBZM: a potential CNS D-2 18 Fahn S, Elton RL, Members of the UPDRS Development dopamine receptor imaging agent. 7 Nucl Med Biol Committee. Unified Parkinson's Disease Rating Scale. 1988;15: 195-201. In: Fahn C, Marsden CD, Calne DB, Goldstein M, eds. 35 Drewe J, Mazer N, Abisch E, Krummen K, Keck M Recent developments in Parkinson's disease. Vol 2. Differential effect of food on kinetics of bromocriptine in Florham Park, New Jersey: Macmillan Healthcare a modified release capsule and a conventional formnula- Information, 1987:153-64. tion. Eur_7 Clin Pharmacol 1988;35:535-41 19 Kurtzke JF. Rating neurologic impairment in multiple 36 Delforge J, Loc'h C, Hantraye P, et al. Kinetic analysis of sclerosis: an expanded disability status scale (EDSS). [76Br]bromolisuride binding to dopamine D2 receptors Neurology 1983;33:1444-52. studied by PET. Cereb Blood Flow Metab 1991 11: 20 Schroth G, Naegele T, Klose U, Mann K. Petersen D. 914-25. Reversible brain shrinkage in abstinent alcoholics, 37 Quik M, Spokes E, MacKay A, Bannister R. Alterations in measured by MRI. Neuroradiology 1988;30:385-9. 3H- binding in human caudate nucleus, sub- 21 Klose U, Naegele T, Schroth G, Mann K. stantia nigra and frontal cortex in the Shy-Drager syn- Volumenbestimmung aus dreidimensionalen MR- drome and Parkinson's disease. Neurol Sci Datensatzen. In: Nuesslin F, ed. Medizinische Physik. 1979;43:429-37. Tubingen: Gulde, 1988:590-4. 38 Cortes R, Camps M, Gueye B, Probst A, Palacios JM. 22 Savoiardo M, Strada L, Girotti F, et al. Olivoponto- Dopamine receptors in the human brain: autoradio- cerebellar atrophy: MR dignosis and relationship to graphic distribution of DI and D2 sites in Parkinson multisystem atrophy. Radiology 1990;174:693-6. syndrome of different etiology. Brainl Res 1989;483 23 Kung HF, Pan S, Kung MP, Kasliwal R, Reilly J, Alavi A. 30-8. In vitro and in vivo evaluation of ["'2I]IBZM: a potential 39 Shinotoh H, Inoue 0, Hirayama K, et al. Dopamine Dl CNS D2-dopamine receptor imaging agent. _7 Nucl Med receptors in Parkinson's disease and striatonigral degen- 1989;30:88-92. eration: a positron emission tomography study. Neurol 24 Kaplan EL, Meier P. Nonparametric estimation from Neurosurg Psychiatry 1993;56:467-72. incomplete observations. _7 Am Stat Assoc 1 958;53: 40 DeLong MR, Georgopoulos AP. Motor functions of the 457-81. basal ganglia. In: Brookhart JM, Mountcastle VB, eds. 25 Rutledge JN, Hilal SK, Silver AJ. Defendini R, Fahn S. Handbook of physiology, section 1 the nervous system, Vol. Study of movement disorders and brain iron by MR. II motor control, Part 2. Bethesda, Maryland: American American Journal ofNeuroradiology 1987;8:397-41 1. Phvsiological Society, 1981:101 7-61. 26 Olanow CW. Magnetic resonance imaging in parkinson- 41 Dejerine J, Thomas A. L'atrophie olivo-ponto-cere- ism. Neurol Clin 1992,10:405-20. belleuse. Nouvelle Iconographie de la Salp&triere 1900;13: 27 Goetz CG, Lutge W, Tanner CM. Autonomic dysfunction 330-70. in Parkinson's disease. Neurology 1986;36:73-5. 42 Greenfield JG. The spino-cerebellar degeneration.. 28 Koller WC, Vetere-Overfield B, Williamson A, Busenbark Springfield: Charles C Thomas, 1954. K, Nash J, Parrish D. Sexual dysfunction in Parkinson's 43 Wullner U, Klockgether T, Petersen D, Naegele T, disease. Clin Neuropharmacol 1990;13:461-3. Dichgans J. Magnetic resonance imaging (MRI) in 29 Sandroni P, Ahlskog JE, Fealey RD, Low PA. Autonomic hereditary and idiopathic ataxia. Neurology 1993;43: involvement in extrapyramidal and cerebellar disorders. 318-25. Clin Auton Res 199 1;1:147-55. 44 Brooks DJ, Luthert P, Gadian D, Marsden CD. Does 30 Hughes AJ, Colosimo C, Kleedorfer B, Daniel SE, Lees signal-attenuation on high-field T2-weighted MRI of the AJ. The dopaminergic response in multiple system brain reflect regional cerebral iron deposition? atrophy. Neurol Neurosurg Psychiatry 1 992;55: 1009-13. Observations on the relationship between regional cere- 31 Hughes AJ, Daniel SE, Kilford L, Lees AJ. The accuracy bral water proton T2-values and iron levels. .7 Neuroi of clinical diagnosis of idiopathic Parkinson's disease: A Neurosurg Psychiatry 1989;52:108-1 1. http://jnnp.bmj.com/ on October 1, 2021 by guest. Protected copyright.