J Neurol (2004) 251 : 1383–1392 DOI 10.1007/s00415-004-0549-2
Kathy Dujardin Luc Defebvre Alain Duhamel Pascal Lecouffe Pascal Rogelet Marc Steinling Alain Destée
Received: 31 July 2003 Received in revised form: 23 February 2004 Accepted: 10 May 2004
K. Dujardin () · L. Defebvre · P. Rogelet · A. Destée Neurologie et Pathologie du Mouvement Neurologie A Hôpital Salengro Centre Hospitalier Universitaire 59037 Lille Cedex, France Tel.: +33-320/446-752 Fax: +33-320/446-680 E-Mail:
[email protected] A. Duhamel CERIM Lille University Faculty of Medicine Lille, France K. Dujardin Psychology Department Charles De Gaulle University Lille, France P. Lecouffe · M. Steinling Nuclear Medicine Unit Faculty of Medicine and Lille University Hospital Lille, France
ORIGINAL COMMUNICATION
Cognitive and SPECT characteristics predict progression of Parkinson’s disease in newly diagnosed patients
■ Abstract Objective To identify features in cognitive functioning and regional cerebral blood flow (rCBF) in newly diagnosed Parkinson’s disease (PD) patients and to determine whether these factors are able to predict the progression of the disease in general and the development of cognitive decline in particular. Methods 50 previously treatment-naive PD patients participated in the study. Cognitive assessment and SPECT were performed twice: at the time of diagnosis and then 3 years later. Six patients died or refused to continue. The Mattis dementia rating scale, the WAIS-R digit span test, a word list learning/recall test, a word fluency task and the Stroop wordcolour test were used to assess cognitive function. rCBF was measured in 10 pairs of regions of interest. Principal component analysis of the data from the final examination was used to determine which variables allowed the formation of patient subgroups. Thereafter, factorial discriminant analysis (FDA) was performed in order to obtain a predictive model of these final classes. Results A stepwise procedure enabled the identification of 3 clusters (26, 16 and 2 patients). As the patients in
the smallest cluster met the criteria for dementia at the final examination, they were discarded from further analyses. All the cognitive variables contributed to the constitution of the two other clusters. Age, educational level and all the rCBF parameters also contributed but to a lesser extent. Comparison of these groups showed reduced overall cognitive efficiency and an exacerbated subcorticofrontal syndrome in the 16-patient cluster. FDA showed that the best predictive model for the final classes was based on 7 variables: educational level, semantic and alternating word fluency, Stroop interference index and the right medial frontal, left parietal and left lenticular nucleus rCBF findings. Conclusion Even though both cognitive and rCBF parameters help predict the progression of newly diagnosed PD patients and bearing in mind the limitations of the SPECT method used here, it appears that the contribution of cognitive assessment is greater than that of rCBF measurement. ■ Key words cognitive function and rCBF · clinical evolution of Parkinson’s disease · longitudinal survey JON 1549
1384
Introduction Even though motor symptoms dominate the clinical presentation of Parkinson’s disease (PD) [32], the disease is usually associated with cognitive deficits. Overall cognitive efficiency is preserved but because basal ganglia and prefrontal cortex are closely interrelated through anatomofunctional circuits [3, 51], cognitive deficits in PD mainly concern tasks involving executive abilities [56]. Indeed, visuospatial processing has been shown to be perturbed in PD, although impairment was only observed in tasks requiring high attentional control or self-initiated response strategies [12, 17, 27]. Furthermore, covariance analyses showed that the visuospatial impairment was no longer observed when executive deficits were taken into account [5]. Episodic memory impairments are almost systematically reported but the observed deficit pattern is very specific and reflects difficulties in self-initiation of retrieval strategies [55]. The main cognitive impairments in PD concern tasks requiring planning and goal-directed behaviour [47], set-shifting [48], inhibition of a prepotent response [11, 25] and action coordination [20, 40]. This subcorticofrontal syndrome is observed from the first stages of the disease onwards, and its severity usually increases with progression of the disease [49]. However, the study of this change has been essentially based on comparisons of patient groups: longitudinal studies can thus generate some additional information about the heterogeneity of the patterns of change and can help to identify which characteristics are related to a risk of cognitive deterioration in the years following diagnosis. It should be noted that a certain proportion of patients will always meet the criteria for dementia, usually late in the course of the disease [18]. Even though such findings are still a matter of debate [10], some correlations have been found between PD patients’ cognitive performances (especially memory and executive functions) on one hand and both 11C-Snomifensine (a sensitive marker of striatal dopamine denervation) and [18F]-Fluoro-Dopa uptake (a marker of the pre-synaptic dopaminergic functions) on the other [41, 59]. However, positron emission tomography (PET) is not widely available in many nuclear medicine departments although single photon emission computed tomography (SPECT) is routinely used. Our group has already demonstrated the utility of SPECT measurements which increase the ability to distinguish progressive supranuclear palsy, dementia with Lewy bodies and Alzheimer’s disease from PD [22, 23] and also frontotemporal dementia from Alzheimer’s disease [16].In PD, perfusion abnormalities studied by SPECT may closely depend on the disease severity and cognitive status. Contrasting results have been reported [4, 21, 52, 63]. Consequently, for untreated PD patients, we thought that it would be useful to determine which cortical areas
are particularly involved and whether rCBF measurements besides cognitive assessment might generate additional useful information to predict the development of cognitive decline. Consequently, the aim of the present study was to try to identify particularities in cognitive functioning and SPECT rCBF measurements observed at the time of diagnosis which could influence or report on PD progression. In order to achieve this, a cohort of newly diagnosed PD patients underwent extensive examination soon after diagnosis and then again 3 years later.
Methods ■ Participants Fifty patients with probable PD (24 females, 26 males) according to the United Kingdom Parkinson’s Disease Society Brain Bank criteria participated in the study [29]. None of the patients had received antiparkinsonian therapy prior to the study. After their first assessment, the majority of patients received antiparkinsonian treatment. Twenty-eight received L-dopa (with a mean (SD) dosage of 457 (180) mg) and 15 received an agonist only (Bromocriptine (n = 5) with a mean (SD) dosage of 29 (2.2) mg; Pergolide (n = 4) with a mean (SD) dosage of 3.9 (0–6) mg; Ropinirole (n = 3) with a mean (SD) dosage of 9.6 (2) mg; Piribedil (n = 2) with a dosage of 200 mg; and Lisuride, 2 mg (n = 1)). After 3 years of disease progression, 4 patients received a combination L-dopa (450 (238) mg) and Bromocriptine (26 (7.5) mg). Patients were exempt from psychotropic medications (anxiolytics, antidepressants, neuroleptics, etc.). Clinical examinations, neuropsychological evaluation and SPECT tests were administered to each patient twice: at the time of diagnosis and 3 years later. Two data sets were thus constituted for each patient, except for two patients who died before the last examination, two patients who refused to continue to participate, one patient whose symptoms became suggestive of multiple system atrophy and one who had suffered 2 strokes and had developed vascular dementia. For this final group of 44 patients,the median age was 66 years and the median disease duration at diagnosis was 1 year. ■ Clinical assessment Motor disability was evaluated during each session (after the patient had received his or her antiparkinsonian treatment) using the UPDRS motor subscale (UPDRS-III) with a score varying from 0 (optimal score) to 108 (worst score) [26]. ■ Neuropsychological assessment Owing to time restrictions, neuropsychological assessment was limited to certain tests known to be particularly sensitive to cognitive dysfunction in PD [56, 58]. Global cognitive efficiency was assessed using the Mattis dementia rating scale (Mattis DRS [42]) and the Mini mental state (MMS) examination score [28]. Immediate memory was assessed by the WAIS-R digit span test, giving a score for forward and backward digit span. Declarative memory for verbal material was evaluated using the French version of the Grober and Buschke [30] word list learning and recall test (G&B test), according to the procedure described by Pillon et al. [57, 58]. The patient was presented with a total of 16 words (presented four at a time) belonging to 16 different semantic categories.
1385
During the encoding phase, the subject was asked to point to and read each word aloud when its category cue was verbally provided. Next, the card was removed and immediate cued recall was tested by providing each category cue. If the subject was unable to recall the four items, the procedure was repeated until the correct answer was obtained. An immediate recall score determined the number of words correctly recalled during the first trial of immediate cued recall. The recall phase included three trials. Each trial consisted of a 2 minute period of free recall followed by cued recall for those items not retrieved at free recall. Items not retrieved at cued recall were given by the examiner. Performance was assessed by taking the total number of words (out of 48) correctly free-recalled and the total number of words (out of 48) correctly recalled after free and cued recall. After a 15 minute delay, free and cued recall were again measured. The delayed recall phase was followed by a recognition phase. Executive functions were evaluated using a word generation task and the Stroop word-colour test. Three different word generation tasks were administered. In the phonemic task, participants were instructed to name as many words beginning with the letter ‘P’ as they could in 1 minute. In the semantic task, participants were instructed to name as many animal nouns as they could in 1 minute. In the alternating task, they were instructed to alternatively name a word beginning with the letter ‘T’ and a word beginning with the letter ‘V’. For all tasks, people’s names and proper nouns were not allowed, as well as words stemming from the same root. Scores for each task were analysed separately. Performance was assessed by taking the number of different words named in 1 minute. A 50 item version of the Stroop word-colour test was applied. The procedure comprised two trials. In the first trial (simple condition), a 21 x 29.7 cm (A4) sheet of paper with 50 strings of five dots was presented to the subject. Each string was randomly printed in one of 3 colours (red, blue or green). The subject was instructed to name the colour of each string of dots as quickly as possible without error. In the second trial (interference condition), a 21 x 29.7 cm (A4) sheet of paper with 50 colour names printed in a colour different from the word itself was presented to the subject. Three colour names were used to construct the list (red, blue and green). Each word was pseudorandomly printed in one of these 3 colours with the restriction that the name had to be different from the ink in which it was printed. The subject was instructed to name the colour of the ink of each word as quickly as possible and without error. Performance was assessed by calculating an ‘interference cost’ index which was obtained by the difference between the time to complete the interference condition and the time to complete the simple condition.
Fig. 1 Regions of interest (1) medial frontal (Med-Fr), (2) lateral frontal (Lat-Fr), (3) posterior frontal (Post-Fr), (4) temporo-insular (Temp-Ins), (5) temporo-parietal (Temp-Par), (6) temporoparieto-occipital (Temp-Par-Occ), (7) parieto-occipital (Par-Occ), (8) occipital (Occ), (9) lenticular nucleus (Lent) and (10) thalamus (Thal) regions
ally and freehand adjusted. The size of the ROI depended on the region in question: this parameter ranged from 1.9 cm2 for the frontal ROI to 5.5 cm2 for the occipital ROI. Tracer uptake was expressed as a cortico-cerebellar activity ratio, so that mean cerebellar hemispheres uptake was used (slice located 0 mm above the orbito-mental plane) as the particular reference for each individual patient using a previously validated method [61]. ■ Data analysis
■ SPECT SPECT was performed using a brain-dedicated, fast-rotating SPECT system (the Tomomatic 564, Medimatic, Copenhagen, Denmark) 15 minutes after intravenous administration of a 555 Mbq bolus of Neurolite. The resolution of the camera was 9 mm FWMM. Patients lay with their eyes closed in a quiet, dark room. The head was carefully placed along the cantho-meatal line using a 3-laser light positioning system. Two series of 5 slices parallel to the cantho-meatal plane were obtained, with a thickness of 13 mm and an overlap of 6.5 mm. Each series required a duration of 10 minutes, and 2 million counts could be obtained from the middle slice. Reconstruction was performed using a filtered-back projection. We used an autogauss Filter (D. C. amplification: 80 %; truncation power: 3; acceptable noise level: 3 %). A correction for attenuation was performed using a value of 0.16 cm–1. Ten paired pairs of Regions of Interest (ROI) were designated in conformity with the Cabanis Atlas (1986) on a slice located 50 mm above the orbito-meatal plane: the medial frontal (Med-Fr), lateral frontal (Lat-Fr), posterior frontal (Post-Fr), temporo-insular (TempIns), temporo-parietal (Temp-Par), temporo-parieto-occipital (Temp-Par-Occ), parieto-occipital (Par-Occ), occipital (Occ), lenticular nucleus (Lent) and thalamus (Thal) regions (see Fig. 1). For each patient, the ROI were placed with a standard template but individu-
All statistical analyses were performed with SAS Software (SAS Institute Inc., Cary, NC 25513). We first analysed data from the final examination only, notably by using cluster analysis to identify homogeneous patients subgroups. This operation was carried out using the κ-means algorithm [31] and the variables taken into account were the UPDRS-III score and a subset of the most informative variables selected from the set of variables as a whole (i. e. cognitive and rCBF) by principal component analysis (PCA). The selected variables were standardized prior to performing cluster analysis, and the number of clusters was determined on the basis of 3 criteria: the cluster cubic criterion (CCC), the squared correlation ratio (R2) and the pseudo F statistic. The “optimal” number of clusters corresponded to a consensus among these 3 criteria: local CCC and pseudo F peaks combined with an inflection in the R2 increase [60]. A graphical check of the cluster separation was obtained by factorial discriminant analysis (FDA). In a second phase, we aimed to determine which of the parameters measured at the initial examination were the best predictors of the final patient classification. Bivariate analyses (the Wilcoxon rank sum test) were performed on all the data (demographical, neuropsychological and rCBF) from the initial examination. Those reaching a significance level greater than 0.20 were discarded. Subsequently, in order to select the best subset of predictive variables, a discriminant analysis using the stepwise option (which is a forward selection al-
1386
lowing elimination) was performed on the remaining parameters. Finally, we used FDA to obtain a predictive model of the final classes.
Results The first phase concerned the data from the final examination. As the rCBF variables did not give consistent information, only the UPDRS-III score and the neuropsychological variables were considered in this analysis. Of the 13 neuropsychological variables, PCA identified the following subgroups of correlated variables: (Stroop word-colour test: time to complete each condition and ‘interference cost’ index), (number of different words named in the alternating and phonemic word fluency tests, backward digit span), (G&B wordlist learning/recall test: number of words correctly recalled after free and delayed free recalls) and (G&B word list learning and recall test: number of words correctly recalled after free and cued recall and delayed free and cued recall). The variables thus retained for the cluster analysis were the UPDRS-III score and 7 neuropsychological variables: the “interference cost” index of the Stroop wordcolour test, the number of different words named in the alternating word fluency test, the number of words correctly free recalled, the number of words correctly recalled after free and cued recall in the G&B word list learning/recall test, the Mattis DRS score, the number of different words named in the semantic word fluency test and, lastly, the delayed free recall in the G&B word list learning/recall test. Three clusters with 26, 16, and 2 patients were identified (F = 15.49, CCC = 10.41, overall R2 = 0.42) and named as cluster 1, cluster 2 and cluster 3 respectively. Examination of the data for cluster 3 showed that it was atypical since these two individuals met the dementia criteria at the final examination: both had developed Alzheimer’s disease. Because of its insufTable 1 Median (range) values of all the parameters for each cluster
ficient size, this cluster was discarded from further analyses, which were thus restricted to the other two clusters. Since cognitive performance is largely dependent on age and education level, a stepwise, discriminant analysis (without a minimum significance level for both selection and elimination) was designed by adding these two variables to the previously selected ones in order to measure their influence on the cluster separation. Using this procedure, age was selected at step 7 with a low partial R2 (R2 = 0.02, F = 0.67, p = 0.42), and educational level was the last selected variable (partial R2 = 0.0001, F = 0.001, p = 0.96). The cluster separation was mainly explained by performance of the alternating word fluency task (partial R2 = 0.53, F = 45.48, p = 0.0001) and the Mattis DRS score (partial R2 = 0.156, F = 7.20, p = 0.011). Since automatic classification systematically identifies clusters which can overlap and/or have no clinical meaning, the quality of the cluster separation was assessed by performing FDA on the 8 variables retained for the cluster analysis. This confirmed the satisfactory separation of the 2 clusters: the R2 of the discriminant function was high (0.65) and the projection of individuals on the discriminant function plane showed graphically that the clusters were well separated. The clinical significance of the clusters was assessed using univariate and bivariate analysis. For each cluster, the median performance level and the range of values for all parameters are shown in Table 1. The Wilcoxon rank sum test for independent measures revealed significant differences between the groups for a large number of variables. Looking more closely at these results, cluster 2 thus appears to be constituted of 16 patients characterized by significantly lower scores for the Mattis DRS, all the Stroop wordcolour parameters (time to complete the first trial, time to complete the second trial, interference index), the free
Median (range)
Cluster 1 (26 subjects)
Cluster 2 (16 subjects)
Wilcoxon tests
Mattis DRS score* Stroop trial 1 (sec)* Stroop trial 2 (sec)* Stroop interference index* Free recall G&B* Total recall G&B Delayed free recall G&8 Delayed total recall G&B Long term retention G&B Phonemic fluency* Semantic fluency* Alternating fluency* Backward digit span UPDRS part 3
140 (129–144) 35 (25–54) 58 (35–96) 22 (9–57) 30.5 (21–41) 47 (41–46) 11 (7–15) 16 (14–16) 103.85 (65–157) 15 (8–24) 20 (12–29) 11 (6–17) 4 (2–6) 12.5 (3–34)
132.5 (118–137) 46.5 (31–93) 87 (57–254) 38 (16–161) 26 (12–34) 47 (45–48) 9 (3–14) 16 (14–16) 108 (50–162) 8.5 (1–18) 13 (7–19) 5.5 (0–10) 3 (3–5) 22 (6–40)
p < 0.0001 P < 0.001 p < 0.0001 p < 0.0001 P = 0.007 P = 0.49 P = 0.049 P = 0.68 p = 0.11 p < 0.0001 p < 0.0001 p < 0.0001 P = 0.022 P = 0.046
1387
recall score in the G&B word list learning and recall test, and all the word fluency task parameters. Moreover, there was a trend towards significantly lower scores in the backward digit span, the delayed free recall at the G&B word list learning and recall test as well as for the UPDRS-III score. Thus, in comparison with cluster 1, cluster 2 is a group of patients with a reduction in overall cognitive efficiency, an exacerbated subcorticofrontal syndrome and more severe motor dysfunction. The second phase concerned the prediction of these clusters on the basis of the demographic, neuropsychological and rCBF data collected at the initial examination. Bivariate analysis selected 25 variables, with a significance level set at 0.20 (see Table 2), A subset of 10 predictive variables was selected using a stepwise, discriminant analysis performed on the initial set of 25 variables. The selection of this subset was based on an inflection in the increase in R2 value. Thereafter, a predictive model for the final classes (clusters determined at the final examination) was elaborated using a score derived by FDA. This “S” score was a linear combination of the 10 variables previously selected. The results are presented in Table 3 and Fig. 2. The value of the square correlation ratio associated Table 2 Variables selected after the bivariate analyses Parameters
Wilcoxon test
Left Temp-Par Right Temp-Par Left Temp-Ins Right Temp-Ins Left Temp-Par-Occ Right Temp-Par-Occ Left Lent Left Lat-Fr Age Duration of education MMS Mattis DRS Stroop phase 1 Stroop phase 2 Stroop interference index Phonemic fluency Semantic fluency Alternating fluency Digit span forward Digit span backward Immediate recall G&B Free recall G&B Total recall G&B Delayed free recall G&B Delayed total recall G&B
p = 0.108 p = 0.045 p = 0.015 p = 0.082 p = 0.070 p = 0.064 p = 0.162 p = 0.178 p = 0.076 p = 0.001 p = 0.007 p = 0.002 p = 0.007 p < 0.001 p = 0.014 p < 0.001 p < 0.001 p < 0.001 p = 0.028 p = 0.076 p = 0.002 p = 0.007 p = 0.065 p = 0.005 p = 0.008
Table 3 Predictive variables selected by the stepwise discriminant analysis Variables (/100)
Correlation
Coefficient
Stroop interference index* Semantic fluency* Duration of education* MMS* Delayed total recall G&B Left Temp-Ins Mattis DRS Left Temp-Par-Occ Right Temp-Ins Left Lent Constant
–0.71 0.70 0.50 0.50 0.47 0.47 0.46 0.37 0.32 0.25
–2 12 22 89 21 18 –12 11 –23 –8 10.5
with S was 0.71. In Table 3, the correlations (expressed as absolute values) between the selected parameters and the S score measure the contribution of each variable to the prediction. Hence, the most discriminant parameters associated with the S score were: Stroop wordcolour test interference index, semantic fluency, educational level and MMS score. Table 3 shows the procedure used to compute this score on the basis of the 10 predictors. Fig. 1 represents the distribution of the discriminant score for each patient group. This shows that the 10 selected variables appear to be relevant for prediction of the final classification. We then derived a decision rule from the results shown on Fig. 2: if S < 0, then the subject belongs to cluster 2 (with a reduction in overall cognitive efficiency, an exacerbated subcorticofrontal syndrome and more severe motor dysfunction), if not, he or she belongs to cluster 1. Using this rule, only 2 patients were misclassified: both showed poor performance for the most discriminant parameter (the Stroop wordcolour test interference index) but performed normally in all the other tasks. Hence, patients classified as belonging to cluster 1 (namely the group of patients without cognitive deficit and with less severe motor symptoms 3 years after the diagnosis of PD) typically showed the following characteristics: a low score in the Stroop interference index, a high MMS score, a high number of animals named in the semantic fluency test and a high educational level. The discriminant score “S”was computed from the 10 variables shown in Table 3. To make the procedure easier, each raw value has to be divided by 100 before entering it in the equation. For example, if the raw scores of a subject are: Stroop interference index = 25, Semantic fluency = 10, Educational level = 10, MMS = 28, Delayed total recall G&B = 15, Left Temp-Ins = 80, Mattis DRS = 140, Left Temp-Par-Occ = 90, Right Temp-Ins = 90 and Left Lent = 80, the subject’s S score will be (0.25 * (–2)) + (0.10 * 12) + (0.10 * 22) + (0.28 * 89) – (0.15 * 21) + (0.80 * 18) + (1.40 * (–12)) + (0.90 * 11) + (0.90 *
1388
Fig. 2 Projection of the patients of both groups according to their discriminant score
(–23)) + (0.80 * (–8)) – 10.5 = 0.87, and he or she will thus belong to the cluster of subjects without cognitive impairment. We wish to emphasise that the discriminant score enables us to predict a patient’s progression on the basis of his or her initial examination. When applying the prediction rule, only 2 patients (5 %) were misclassified. However, since the same sample was used both to establish and evaluate the prediction rule, this observed misclassification rate could be biased, and tends to be optimistic. Only a further prospective study could estimate the true misclassification rate and then validate this type of predictive model. Nevertheless, statistical procedures for reducing this bias exist, and one such method was employed here: using cross-validation, we obtained a corrected misclassification rate of 12 % (5/42), i. e. quite close to the observed rate (5 %, 2/42).
Discussion This longitudinal follow-up of a cohort of 44 consecutive patients at an early stage of Parkinson’s disease shows that 3 years after diagnosis, a certain proportion of patients have reduced overall cognitive efficiency, an exacerbated subcorticofrontal syndrome and more severe motor symptoms as measured by the motor UPDRS score.At the time of diagnosis, the best predictors of this progression were certain specific, cognitive scores such as the interference index from the Stroop word-colour test and the number of animal nouns named at the semantic word fluency task, as well as more general vari-
ables such as the educational level and the MMS score. Other indices also contribute to this prediction to a lesser extent: there were other specific, cognitive scores such as delayed recall in the G&B test, the Mattis DRS score and SPECT measurements such as the Left TempIns, Left Temp-Par-Occ, Right Temp-Ins and Left Lent rCBF. Hence, even though at the time of diagnosis none of the patients had clinical symptoms or a cognitive profile suggestive of a particular disease progression,our results suggest that particular observation of specific aspects of cognitive assessment and rCBF could allow early detection of patients with idiopathic PD, i. e. those at risk of more rapid degradation of their cognitive and motor status. The cognitive scores which predict this change relate to specific parameters such as executive function (Stroop word-colour test interference index) and semantic and episodic memory (semantic fluency and delayed recall at the G&B test respectively) although the overall cognitive efficiency (MMS and Mattis DRS Scores), as well as demographical variables such as the educational level, also appeared to have an influence.At the perfusion level, it appears that low rCBF in the temporo-insular areas should be considered as a predictive factor. Three years after diagnosis, even though none of the patients included in our analysis had dementia or met the criteria for Alzheimer’s disease (NINCDS-ADRDA criteria [44]) or dementia with Lewy bodies [43], a certain proportion (16/44=36.36 %) presented with more pronounced cognitive deficits and more severe motor symptoms. The main predictor of this progression was the Stroop word-colour test interference index, which re-
1389
flects the ability to resist interference and is usually considered to reflect the integrity of the frontal cortex [15, 50, 64]. Recent studies have even suggested that the Stroop interference effect leads to specific activation of the anterior cingulate, inferior and middle frontal areas as well as the insula [37, 53, 62]. It can thus be suggested that the patients with more rapid cognitive decline suffer from more significant degradation of the associative nigro-striato-cortical circuit relating the associative area of the striatum to frontal areas involved in executive function [51]. However, the other predictors seem to reflect dysfunction of either the temporo-limbic circuits (e. g. the semantic word fluency task and long term retention at the G&B memory test) or more diffuse disturbances as evidenced by the MMS and the Mattis DRS scores. In this respect, it is interesting to note that all the rCBF indices retained after the statistical analysis involved posterior regions (left and right temp-insular and left temp-par-occ areas). Consequently, there seems to be some association between the cognitive and rCBF parameters suggested here as predictive factors. Nevertheless, as can be seen in Table 4, the differences between the patient clusters in terms of the values of predicting indices were very small. Only four parameters showed a significant impairment within the cluster, with faster degradation. However, whatever the parameter, a trend towards lower performance or perfusion level appears. Table 4 shows that the cluster 1 patients had very similar performance to that of healthy controls except in the Stroop word-colour test interference index, where they clearly appear to be more disturbed by the interference condition. In contrast, cluster 2 patients performed worse than controls. One should bear in mind that all the predictive rCBF indices involve posterior regions. None of the frontal rCBF indices were selected, even though PD patients in the second cluster presented a cognitive pattern suggesting frontal dysfunction. In PD, perfusion abnormalities studied by SPECT may closely depend on the disease severity and cognitive status. Contrasting results have been reported: in nondemented PD patients mean Table 4 Median (range) of the parameters predicting the cluster 1 and 2’s separation at the first examination. In the last column, the values of a group of 40 healthy control subjects representative of the patients population are given for information
cortical and regional uptake ratios were not significantly different from controls [63] whereas other studies revealed a local frontal [52] or parietal [63] hypoperfusion contralateral to the motor symptoms in hemiparkinsonian patients, with a more pronounced defect in advanced patients with a bilateral symptom distribution [4, 22]; in demented PD patients, a significant decrease was observed in all cortical areas, with lower uptake in temporal and parietal areas [4] or in frontal areas [63]. However, clinical practice demonstrates that there is no strong association in PD between progressive cognitive impairment (subcortico-frontal syndrome) and focal hypometabolism, in contrast to what is reported in patients with progressive supranuclear palsy at earlier stages of the disease [21]. However, none of the patients included in our analysis had dementia at the final evaluation and so, consequently, it can be suggested that frontal rCBF indices would only be selected if a certain proportion of patients met the dementia criteria. In the present study, the bilateral temporo-insular and temporo-parieto-occipital areas appeared to be the most discriminant rCBF indices, suggesting, from a perfusion point of view, a possible progression for the cluster 2 patients towards a pattern which has been reported in Alzheimer’s disease, i. e. bilateral, decreased perfusion of the temporal and parietal regions [36, 46]. Only a longer, longitudinal study could provide the answer to such a question. A progression towards dementia with Lewy bodies in which temporo-parietal and occipital reduced perfusion (and also sometimes frontal hypoperfusion) have been reported [23, 24] can here be excluded since, according to the criteria proposed by McKeith et al. [43], the delay between the parkinsonian syndrome and dementia cannot be longer than 2 years when establishing this diagnosis; yet, in our study, the examination period lasted 3 years. The interest of the lenticular nucleus region seems more limited; admittedly, our system had poor resolution, and we occasionally encountered difficulties when measuring Neurolite uptake in low volume subcortical regions (e. g. the thalamus and lenticular nucleus).
Median (range)
Cluster 1 (26 subjects)
Cluster 2 (16 subjects)
Healthy Controls*
Stroop interference index* Semantic fluency* Duration of education* MMS* Delayed total recall G&B Left Temp-Ins Mattis DRS Left Temp-Par-Occ Right Temp-Ins Left Lent
63 (26–166) 18 (11–25) 12 (5–17) 29.7 (–) 16 (13–16) 94 (84–112) 141 (125–144) 90 (80–104) 96 (86–115) 97 (81–121)
110 (23–215) 13 (3–18) 7.5 (7–12) 29.1 (–) 15 (12–16) 89 (79–97) 136 (123–140) 87 (78–95) 92 (81–108) 92 (82–107)
20 (15–32) 19 (11–32) 12 (7–18) 29.7 (28–30) 16 (14–16) > 80 142 (138–144) > 80 > 80 > 80
1390
Few longitudinal follow-ups of PD patients have been conducted [35, 38, 39, 54, 65] and all were interested in predicting dementia. The largest longitudinal study to date is probably that performed by a group in Leeds [33] including 89 PD patients and 50 healthy controls assessed at regular intervals over a period of ten years. The objectives of this study differed from ours, since the Leeds group was mainly interested in the incidence of dementia in PD. Examination of their results suggests, however, that the progression of our cohort is very similar to that reported by Hughes et al. Indeed, their survival analysis (see their figure p. 1599) showed that the incidence of dementia was very low after three years of disease (about 10 %) but that it increased dramatically after about 8 years of PD progression. Although the data from the demented patients were not considered in the present study, it is necessary to bear in mind that our initial cluster analysis identified a small group of 2 patients (4.55 % of the cohort) who did meet the dementia criteria and were discarded from further analyses because of an insufficient group size. However, the application of our discriminant score equation to them revealed that both obtained negative scores: –2.75 and –1.95, respectively, i. e. high values very far from the limit between both groups and that predict belonging to the cluster of subjects with cognitive impairment. Hence, identifying patients with a faster progression (i. e. more
severe subcortico-frontal cognitive dysfunction and motor symptoms) three years after diagnosis can perhaps constitute a way of detecting those who are more likely to develop dementia in the subsequent years: clinicians could then increase their surveillance in order to treat these patients adequately as soon as the first signs of dementia appear. Indeed, in a very recent study, Woods and Tröster [6] reported that a prodromal frontal/executive dysfunction whose characteristics are rather comparable to what we observed, can predict the incidence of dementia in PD. As underlined by these authors, dementia in PD is associated, among others with an increase in caregiver distress and the social burden [1, 2], and an early identification of incident dementia may help to set up an appropriate medical and psychosocial management of this disorder. Furthermore, the idea that treatments such as acetylcholinesterase inhibitors can slow down cognitive and behavioural disorders in demented patients [13, 19, 45] also represents an important challenge since an early administration seems to increase their efficacy. In the future, it will be essential to perform the same longitudinal study with PD patients using a dopamine transporter marker (e. g. beta-CIT): this will allow more objective analysis of the neurochemical consequences of degeneration in the nigrostriatal dopaminergic system.
References 1. Aarsland D, Larsen JP, Karlsen K, Lim NG, Tandberg E (1999) Mental symptoms in Parkinson’s disease are important contributors to caregiver distress. Int J Geriatr Psychiatry 14(10):866–874 2. Aarsland D, Larsen JP, Tandberg E, Laake K (2000) Predictors of nursing home placement in Parkinson’s disease; a population-based, prospective study. J Am Geriatr Soc 48(8):938–942 3. Alexander GE, De Long MR, Strick PL (1986) Parallel organization of functionally segregated circuits linking basal ganglia and cortex. Ann Rev Neurosci 9:357–381 4. Antonini A, De Notaris R, Benti R, De Gaspari D, Pezzoli G (2001) Perfusion ECD/SPECT in the characterization of cognitive deficits in Parkinson’s disease. Neurol Sci 1:45–46 5. Bondi MW, Kaszniak AW, Bayles KA, Vance KT (1993) Contribution of frontal system dysfunction to memory and perceptual abilities in Parkinson’s disease. Neuropsychology 7:89–102
6. Booij J, Tissingh G, Winogrodzka A, von Royen EA (1999) Imaging of the dopaminergic neurotransmission system using single photon emission tomography and positron emission tomography in patients with parkinsonism. Eur J Nucl Med 26:171–182 7. Brooks DJ (1997) PET and SPECT studies in Parkinson’s disease. Bailliere’s Clinical Neurology 6:69–87 8. Brooks DJ (1998) The early diagnosis of Parkinson’s disease. Ann Neurol 44(Suppl 1):10–18 9. Brooks DJ, Salmon EP, Mathias CJ, Quinn N, Leenders KL, Bannister R, Marsden CD, Frackowiak RS (1990) The relationship between locomotor disability, autonomic dysfunction and the integrity of the striatal dopaminergic system in patients with multisystem atrophy, pure autonomic failure, and Parkinson’s disease, studied with PET. Brain 113:1539–1552
10. Broussolle E, Dentresangle C, Landais P, Garcia-Larrea L, Pollak P, Croisile B, Hibert O, Bonnefoi F, Galy G, Froment JC, Comar D (1999) The relation of the putamen and caudate nucleus 18-F Dopa uptake to motor and cognitive performances in Parkinson’s disease. J Neurol Sci 166:141–151 11. Brown RG, Marsden CD (1988) An investigation of the phenomenon of “set” in Parkinson’s disease. Mov Disord 3(2):152–161 12. Brown RG, Marsden CD (1990) Cognitive function in Parkinson’s disease: from description to theory. Trends Neurosci 13(1):21–29 13. Bullock R, Cameron A (2002) Rivastigmine for the treatment of dementia and visual hallucinations associated with Parkinson’s disease: a case series. Curr Med Res Opin 18(5):258–264 14. Cabanis EA, Tamraz J, Iba-Zizen MT (1986) Imagerie par résonance magnétique (IRM) de la tête à 0.5 Tesla. Atlas de corrélations anatomiques normales dans 3 dimensions, selon l’orientation du plan neuro-oculaire (PNO). Feuillets de Radiologie 26:309–416
1391
15. Carter CS, Mintun M, Cohen JD (1995) Interference and facilitation effects during selective attention: an H2150 PET study of Stroop task performance. Neuroimage 2(4):264–272 16. Charpentier P, Lavenu I, Defebvre L, Duhamel A, Lecouffe P, Pasquier P, Steinling M (2000) Alzheimer’s disease and fronto-temporal dementia are differentiated by discriminant analysis applied to 99mTcHmPao SPECT. J Neurol Neurosurg Psychiatry 69: 661–663 17. Crevits L, De Ridder K (1997) Disturbed striatoprefrontal mediated visual behavior in moderate to severe parkinsonian patients. J Neurol Neurosurg Psychiatry 63:296–299 18. Cummings JL (1988) The dementias of Parkinson’s disease; prevalence, characteristics, neurobiology, and comparison with dementia of the Alzheimer type. Eur Neurol 28(Suppl 1):15–23 19. Cummings JL (2003) Use of cholinesterase inhibitors in clinical practice: evidence-based recommendations. Am J Geriatr Psychiatry 11(2):131–145 20. Dalrymple-Alford JC, Kalders AS, Jones RD, Watson RW (1994) A central executive deficit in patients with Parkinson’s disease. J Neurol Neurosurg Psychiatry 57:360–367 21. D’Antona R, Baron JC, Samson Y, Serdaru M, Viader F, Agid Y (1985) Subcortical dementia. Frontal cortex hypometabolism detected by positron tomography in patients with progressive supranuclear palsy? Brain 108: 785–799 22. Defebvre L, Lecouffe P, Destée A, Houdart P, Steinling M (1995) Tomographic measurements of regional cerebral blood flow in progressive supranuclear palsy and Parkinson’s disease. Acta Neurol Scand 92:235–241 23. Defebvre L, Leduc V, Duhamel A, Lecouffe P, Pasquier F, Lamy C, Steinling M, Destée A (1999) Technetium HMPAO SPECT study in dementia with Lewy bodies, Alzheimer’s disease and idiopathic Parkinson’s disease. J Nucl Med 40:956–962 24. Donnemiller E, Heilmann J, Wenning GK, Berger W, Decristoforo C, Moncayo R, Poewe W, Ransmayr G (1997) Brain perfusion scintigraphy with 99mTc-HMPAO or 99mTc-ECD and 123Iβ-CIT single-photon emission tomography in dementia of the Alzheimertype and diffuse Lewy body disease. Eur J Nucl Med 24:320–325 25. Dujardin K, Degreef JF, Rogelet P, Defebvre L, Destée A (1999) Impairment of the supervisory attentional system in early untreated patients with Parkinson’s disease. J Neurol 246(9): 783–788
26. Fahn S, Elton RL and members of the UPDRS development committee (1987) Unified Idiopathic Parkinson’s disease Rating Scale. In: Fahn S, Marsden CD, Calne D, Goldstein M (eds) Recent developments in Parkinson’s disease (Vol. 2). Florham Park, New Jersey, Mac Millan Healthcare Information, USA, pp 153–164 27. Flowers KA, Robertson C (1995) Perceptual abnormalities in Parkinson’s disease: top-down or bottom-up processes? Perception 24:1201–1221 28. Folstein MF, Folstein SE, McHugh PR (1975) Mini Mental State: a practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res 12:189–198 29. Gibb WRG, Lees AJ (1988) The prevalence of the Lewy body to the pathogenesis of idiopathic Parkinson’s disease. J Neurol Neurosurg Psychiatry 51:745–752 30. Grober E, Buschke H (1987) Genuine memory deficits in dementia. Dev Neuropsychol 3:13–36 31. Hartigan JA (1975) Clustering algorithms. John Wiley & Sons 32. Hoehn MM, Yahr MD (1967) Parkinsonism: onset, progression and mortality. Neurology 17:427–442 33. Hughes TA, Ross H, Musa S, Bhattacherjee S, Nathon RN, Mindham RH, Spokes EG (2000) A ten year study of the incidence of and factors predicting dementia in Parkinson’s disease. Neurology 54:1596–1602 34. Ichise M, Kim YJ, Ballinger JR, Vines D, Erami SS, Tanaka F, Lang AE (1999) SPECT imaging of pre- and postsynaptic dopaminergic alterations in Ldopa-untreated PD. Neurology 52: 1206–1214 35. Jacobs DM, Marder K, Cote LJ, Sano M, Stern Y, Mayeux R (1995) Neuropsychological characteristics of preclinical dementia in Parkinson’s disease. Neurology 45(9):1691–1696 36. Jagust WH, Johnson KA, Holman BL (1995) SPECT perfusion imaging in the diagnosis of dementia. J Neuroimag 5:545–552 37. Leung HC, Skudlarski P, Gatenby JC, Peterson BS, Gore JC (2000) An Eventrelated Functional MRI Study of the Stroop Color Word Interference Task. Cerebral Cortex 10(6):552–560 38. Levy G, Jacobs DM, Tang MX, Cote LJ, Louis ED, Alfaro B, Mejia H, Stern Y, Marder K (2002) Memory and executive function impairment predict dementia in Parkinson’s disease. Mov Disord 17(6):1221–1226 39. Mahieux F, Fénelon G, Flahault A, Manifacier MJ, Michelet D, Boller F (1998) Neuropsychological prediction of dementia in Parkinson’s disease. J Neurol Neurosurg Psychiatry 64: 178–183
40. Malapani C, Pillon B, Dubois B, Agid Y (1994) Impaired simultaneous cognitive task performance in Parkinson’s disease: a dopamine-related dysfunction. Neurology 44:319–326 41. Marié RM, Barré L, Dupuy B, Viader F, Defer G, Baron JC (1999) Relationship between striatal dopamine denervation and frontal executive tests in Parkinson’s disease. Neurosci Lett 260:77–80 42. Mattis S (1976) Mental status examination for organic mental syndrome in the elderly patient. In: Bellak L, Karasy TE (eds) Geriatric Psychiatry. New York, Grune and Stratton, pp 77–121 43. McKeith IG, Galasko D, Kosaka K, Perry EK, Dickson DW, Hansen LA, Salmon DP, Lowe J, Mirra SS, Byrne EJ, Lennox G, Quinn NP, Edwardson JA, Ince PG, Bergeron C, Burns A, Miller BL, Lovestone S, Collerton D, Jansen EN, Ballard C, de Vos RA, Wilcock GK, Jellinger KA, Perry RH (1996) Consensus guidelines for the clinical and pathologic diagnosis of dementia with Lewy bodies (DLB): report of the consortium on DLB international workshop. Neurology 47(5):1113–1124 44. McKhann G, Drachman D, Folstein M, Katzman R, Price D, Stadlan EM (1984) Clinical diagnosis of Alzheimer’s disease: report of the NINCDS-ADRDA Work Group under the auspices of Department of Health and Human Services Task Force on Alzheimer’s disease. Neurology 34:939–944 45. Mori S (2002) Responses to donepezil in Alzheimer’s disease and Parkinson’s disease, Ann N Y Acad Sci 977:493–500 46. Neary D, Snowden JS, Shields RA, Burjan AWI, Northen B (1987) Single photon emission tomography using 99mTc-HM-PAO in the investigation of dementia. J Neurol Neurosurg Psychiatry 50:1101–1109 47. Owen AM, James M, Leigh PN, Summers BA, Marsden CD, Quinn NP, Lange KW, Robbins TW (1992) FrontoStriatal Cognitive Deficits at Different Stages of Parkinson’s Disease. Brain 115:1727–1751 48. Owen AM, Roberts AC, Hodges JR, Summers BA, Polkey CE, Robbins TW (1993) Contrasting mechanisms of impaired attentional set-shifting in patients with frontal lobe damage or Parkinson’s disease. Brain 116: 1159–1175 49. Owen AM, Sahakian BJ, Hodges JR, Summers BA, Polkey CE, Robbins TW (1995) Dopamine-Dependent Frontostriatal Planning Deficits in Early Parkinson’s Disease. Neuropsychology 9(1):126–140
1392
50. Pardo JV, Pardo PJ, Janer KW, Raichle ME (1990) The anterior cingulate cortex mediates processing selection in the Stroop attentional conflict paradigm. Proc Natl Acad Sci USA 87(1): 256–259 51. Parent A, Hazrati LN (1995) Functional anatomy of the basal ganglia. I. The cortico-basal ganglia-thalamo-cortical loop. Brain Res Rev 20:91–127 52. Perlmutter JS, Raichle ME (1985) Regional blood flow in hemiparkinsonism. Neurology 35:1127–1134 53. Peterson BS, Skudlarski P, Gatenby JC, Zhang H, Anderson AW, Gore JC (1999) An fMRI study of Stroop WordColor Interference: evidence for cingulate subregions subserving multiple distributed attentional systems. Biol Psychiatry 45(10):1237–1258 54. Picirilli M, D’Alessandro P, Finali G, Piccini GL, Agostini L (1989) Frontal lobe dysfunction in Parkinson’s disease: prognostic value for dementia? Eur Neurol 29:71–76 55. Pillon B, Deweer B, Agid Y, Dubois B (1993) Explicit memory in Alzheimer’s, Huntington’s and Parkinson’s diseases. Arch Neurol 50:374–379
56. Pillon B, Dubois B, Agid Y (1996) Testing cognition may contribute to the diagnosis of movement disorders. Neurology 46:324–334 57. Pillon B, Ertle S, Deweer B, Sarazin M, Agid Y, Dubois B (1996) Memory for spatial location is affected in Parkinson’s disease. Neuropsychologia 34(1):77–85 58. Pillon B, Gouider-Khouja N, Deweer B, Vidaihet M, Malapani C, Dubois B, Agid Y (1995) Neuropsychological pattern of striatonigral degeneration: comparison with Parkinson’s disease and progressive supranuclear palsy. J Neurol Neurosurg Psychiatry 58: 174–179 59. Rinne JO, Portin R, Ruottinen H, Nurmi E, Bergman J, Haaparanta M, Solin O (2000) Cognitive impairment and the brain dopaminergic system in Parkinson disease: [18F]fluorodopa positron emission tomographic study. Arch Neurol 57:470–475 60. Sarle WS (1993) SAS technical report A-108: cubic clustering criterion. SAS Institute Inc, Cary, NC
61. Steinling M, Mazingue A, Kassiastis P, Gaudet Y, Fialdes P, Duhamel A, Dubois P, Vergnes R (1988) Le HmPaO-Tc comme indicateur du débit sanguin cérébral local: étude quantifiée comparée à la méthode par inhalation du Xenon 133. Ann Radiol 4:229–237 62. Swick D, Jovanovic J (2002) Anterior cingulate cortex and the Stroop task; neuropsychological evidence for topographic specificity. Neuropsychologia 40(8):1240–1253 63. Wang SJ, Liu RS, Liu HC, Lin KN, Shan DE, Liao KK, Fuh JL, Lee LS (1993) Technetium99m hexamethylpropylene amine oxime single photon emission tomography of the brain in early Parkinson’s disease: correlation with dementia and lateralization. Eur J Nucl Med 4:339–344 64. West R, Bell MA (1997) Stroop colorword interference and electroencephalogram activation: evidence for age-related decline of the anterior attention system. Neuropsychology 11(3):421–427 65. Woods SP, Troster AI (2003) Prodromal frontal/executive dysfunction predicts incident dementia in Parkinson’s disease. J Int Neuropsychol Soc 9(1): 17–24
