Diagnostic criteria for dementia and Alzheimer disease

Transcription

1 ORIGINAL STUDY Brazilian Version of the Addenbrooke Cognitive Examination-revised in the Diagnosis of Mild Alzheimer Disease Viviane Amaral Carvalho, MSc,*w Maira Tonidandel Barbosa, MD, PhD,z and Paulo Caramelli, MD, PhD*w Objective: To investigate the accuracy of the Brazilian version of the Addenbrooke Cognitive Examination-revised (ACE-R) in the diagnosis of mild Alzheimer disease (AD). Background: The ACE-R is an accurate and brief cognitive battery for the detection of mild dementia, especially for the discrimination between AD and frontotemporal dementia. Methods: The battery was administered to 31 patients with mild AD and 62 age-matched and education-matched cognitively healthy controls. Both groups were selected using the Dementia Rating Scale and were submitted to the ACE-R. Depression was ruled out in both groups by the Cornell Scale for Depression in Dementia. The performance of patients and controls in the ACE-R was compared and receiver operator characteristic curve analysis was undertaken to ascertain the accuracy of the instrument for the diagnosis of mild AD. Results: The mean scores at the ACE-R were ± points for patients with AD and ± 7.90 points for controls. The cut-off score <78 yielded high diagnostic accuracy (receiver operator characteristic area under the curve = 0.947), with 100% sensitivity, 82.26% specificity, 73.8% positive predictive value, and 100% negative predictive value. Conclusions: The Brazilian version of the ACE-R displayed high diagnostic accuracy for the identification of mild AD in the studied sample. Key Words: dementia, cognitive evaluation, Addenbrooke Cognitive Examination-revised, cross-cultural adaptation, Brazil (Cog Behav Neurol 2010;23:8 13) Diagnostic criteria for dementia and Alzheimer disease (AD) require the objective evidence of cognitive and functional impairment, which is a challenge in very mild cases and also in populations with a heterogeneous educational background. 1,2 In this sense, there is still a need for specific instruments for the diagnosis of dementia validated in developing countries, including Brazil. 3 The Addenbrooke Cognitive Examination (ACE) 4 is a cognitive tool developed in the United Kingdom that assesses distinct cognitive domains and allows early detection of dementia. Moreover, it provides an index to differentiate AD from frontotemporal dementia, namely the VLOM ratio: (Verbal Fluency+Language)/(Orientation +Episodic memory/delayed recall). The Mini-Mental State Examination (MMSE) 5 is incorporated into the test and the composite battery is almost 3 times more efficient than the MMSE alone. 4 The ACE has been studied in the United Kingdom 6 9 and has also been validated in several other countries Mioshi et al 20 revised the ACE in 2006 and showed similar diagnostic accuracy and psychometric properties of the Addenbrooke Cognitive Examination-revised (ACE-R) inrelationtotheace.inadditiontothemmsescore and the 100 points for the total score, the ACE-R offers 5 subdomain scores (orientation/attention, memory, verbal fluency, language, and visuospatial). We have previously adapted and published the Brazilian version of the battery. 21 The aims of this study were to investigate the performance of patients with mild AD and cognitively healthy elderly volunteers in the Brazilian version of the ACE-R, and to determine its cut-off score/subscores and the corresponding sensitivity, specificity, and positive and negative predictive values in this sample. From the *Postgraduate Program in Neurology, University of São Paulo School of Medicine, Sa o Paulo (SP); wbehavioral and Cognitive Neurology Unit, Department of Internal Medicine, Faculty of Medicine of the Federal University of Minas Gerais; and zgeriatric Unit, Faculty of Medical Sciences and Mater Dei Hospital, Belo Horizonte (MG), Brazil. No funding was received for this work. Reprints: Paulo Caramelli, MD, PhD, Associate Professor of Neurology, Department of Internal Medicine, Faculty of Medicine, Federal University of Minas Gerais, Av. Prof. Alfredo Balena 190-Room 236, , Belo Horizonte (MG), Brazil ( caramelp@usp.br). Copyright r 2010 by Lippincott Williams & Wilkins MATERIALS AND METHODS The group with dementia was composed of 31 patients followed at the Behavioral and Cognitive Neurology Unit of the Hospital das Clı nicas of the Federal University of Minas Gerais or at the Geriatric Unit of the Faculty of Medical Sciences of Minas Gerais, both in Belo Horizonte, Brazil. All patients fulfilled the diagnostic criteria for probable AD according to the National Institute of Neurological and Communicative Disorders and Stroke/Alzheimer s Disease and Related 8 Cog Behav Neurol Volume 23, Number 1, March 2010

2 Cog Behav Neurol Volume 23, Number 1, March 2010 ACE-R in the Diagnosis of Mild AD Disorders Association 1. The patients were classified as having mild dementia according to the Diagnostic and Statistical Manual of Mental Disorders 3rd editionrevised 22 and the Clinical Dementia Rating Scale score of Moreover, all patients with AD scored between 100 and 122 points, both inclusive, on the Dementia Rating Scale (DRS). 24,25 The control group was composed of 62 healthy agematched and education-matched elderly volunteers selected in the community, especially in senior groups. All subjects scored Z123 points on the DRS, 24,25 had no complaints of cognitive decline or history of neurologic or psychiatric illness, and were not using of any drug that could affect the central nervous system. Neither group fulfilled diagnostic criteria for depression, with scores of <8 points on the Cornell Scale for Depression in Dementia (CSDD). 26,27 The study was approved by the Ethics Committee of the Hospital das Clínicas of the University of Sa o Paulo School of Medicine and by the Ethics Committee of the Federal University of Minas Gerais. All participants signed the written informed consent. Demographic data, and CSDD, DRS, and ACE-R scores were compared across different groups by w 2 (sex), the Student t test (normal distribution values), and the Mann-Whitney test (non-normal distribution values). The Kolmogorov-Smirnov normality test analyzed the scores distribution in each continuous variable. The diagnostic accuracy of the battery was investigated using the receiver operator characteristic curve analysis. Sensitivity, specificity, and positive and negative predictive values were also calculated using these discriminant analyses. Statistical analyses were either performed using the Statistical Package for the Social Sciences 15.0 or MedCalc software, both for Windows. RESULTS Table 1 displays the demographic characteristics, the CSDD and DRS scores, and the ACE-R composite or total score and its subdomain scores for AD and control groups. Both groups were matched for age and education (T = 0.144, P = and U = 946.0, P = 0.901, respectively), although sex distribution was significantly different between groups (w 2 = 5.688, P = 0.017). The performance in the ACE-R was significantly different, with patients presenting lower scores in all subdomains and in the composite score (P<0.001). The results of mood assessment, as determined by CSDD scores, showed greater scores for patients with AD versus controls (P<0.001), although not representing diagnosis of depression. In the control group, ACE-R total score was significantly correlated with education (r = 0.524, P<0.001) and age (r = 0.328, P = 0.009), whereas in the AD group only education seemed to influence the final score of the battery (r = 0.665, P<0.001). When combining both groups, there was also a significant correlation only between the years of schooling (r = 0.399, P<0.001) and the ACE-R total score. The sex of the patient did not affect the ACE-R total score, either in patients with AD (T = 1.677, P = 0.104) or in patients with AD and controls combined (T = 1.226, P = 0.223). However, sex did show an association with the battery composite score among controls (T = 2.217, P = 0.030). Table2showstheACE-Randitscut-offscoresfor 5 subdomains defined by the best relationship between sensitivity and specificity for the discrimination of AD from controls. The cut-off score determined for the composite battery in the whole sample was <78 points (sensitivity = 100%, specificity = 82.26%, positive predictive value = 73.8%, and negative predictive value = 100%). ACE-R total score presented high diagnostic accuracy TABLE 1. Comparison of Demographic Data, CSDD, DRS, ACE-R Total and Subscores Between AD and Controls Controls, N = 62 AD Patients, N = 31 Mean SD Median Mean SD Median P Sex (Male:Female)* 22:40 13:18 w 2 = Age, yearsw T = Education, years U = CSDD (38) U = <0.001 DRS (144)w T = <0.001 ACE-R Attention and Orientation (18) ) 15 U = <0.001 Memory (26) ) 10 U = <0.001 Fluency (14) ) 6 U = <0.001 Language (26) ) 21 U = <0.001 Visuospatial (16) ) 13 U = <0.001 MMSE (30) ) 23 U = <0.001 ACE-R total score ) 66 U = <0.001 Maximum points for each instrument or ACE-R domains are in parenthesis in the left column. *w 2 test. wstudent t test; Mann-Whitney test for the other variables. ACE-R indicates Brazilian version of the Addenbrooke Cognitive Examination-revised; CSDD, Cornell Scale for Depression in Dementia; DRS, Dementia Rating Scale; MMSE, Mini-Mental State Examination; SD, standard deviation. r 2010 Lippincott Williams & Wilkins 9

3 Carvalho et al Cog Behav Neurol Volume 23, Number 1, March 2010 TABLE 2. Cut-off Scores for the Brazilian Version of the ACE-R ACE-R AUC SD Maximum Points Cut-off Sensitivity (%) Specificity (%) PPV (%) NPV (%) Attention and Orientation < Memory < Fluency < Language < Visuospatial < MMSE < ACE-R total score < ACE-R indicates Brazilian version of the Addenbrooke Cognitive Examination-Revised; AUC, area under the curve; MMSE, Mini-Mental State Examination; NPV, negative predictive value; PPV, positive predictive value; SD, standard deviation. [receiver operator characteristic area under the curve (AUC)=0.947] similar to what has been found for the subdomains Attention and Orientation (AUC=0.919), Memory (AUC=0.938), and MMSE (AUC=0.940). The mean time of administration of the ACE-R was 16 minutes for controls and 20 minutes for patients. The positive and negative predictive values were calculated for different prevalence rates 28 and these data are shown in Table 3. To investigate the ACE-R cut-off scores for different educational levels, we divided the samples into <11 years [group of low education (GLE)] and Z11 years of schooling [group of high education (GHE)]. Table 4 shows the demographic data from these 2 subgroups. In GLE, only the years of schooling were significant correlated with ACE-R total score (r = 0.383; P = 0.010). Age (r = 0.174; P = 0.259) and sex of the patients (T = 0.768; P = 0.447) did not influence the battery composite score. In GHE, none of the variables correlated with the battery composite score (age: r = 0.189, P = and sex: T = 0.956; P = 0.344). However, education almost reached the statistical values for significance (r = 0.279; P = 0.052). The mean ± SD of the ACE-R total score was ± 7.26 points in GLE and ± 7.65 points in GHE. Figures 1 and 2 present the ROC Curves for these 2 groups of education and Table 5 shows the different ACE-R cut-off scores for these 2 groups of education. DISCUSSION The Brazilian version of the ACE-R proved to be an accurate and brief cognitive tool for the diagnosis of mild AD in our milieu. The battery composite score and 3 subdomain scores (Attention and Orientation, Memory, and TABLE 3. Sensitivity, Specificity, Positive Predictive Values, and Negative Predictive Values at Different Prevalence Rates PPV (NPV) at Different Prevalence Rates Sensitivity Specificity 6% 12% 18% 24% < NPV indicates negative predictive value; PPV, positive predictive value Fluency) yielded excellent discrimination between patients and controls. These findings suggest that the battery is able to detect the most common cognitive domains affected in the early stages of AD. Therefore, the ACE-R seems to be an efficient instrument for administration in elderly persons with heterogeneous educational levels. The cut-off score identified for the composite battery (<78) points presented high sensitivity and specificity (100% and 82.26%, respectively) for the detection of mild AD. Although the first British study grouped several types of cortical dementia, our cut-off score is far from the one previously found, which was 88 (sensitivity = 94% and specificity = 89%) or 82 (sensitivity = 84% and specificity = 100%). 20 However, both studies presented similar capacities to discriminate patients from controls, and the Brazilian score reached 100% sensitivity. Only 3 years have passed since the publication of the ACE-R 20 and a fairly reduced number of studies with the application of this new version in patients with AD are already available. 20,29,30 The authors have proposed some improvements in the revised version, especially the redistribution of subdomain scores and the substitution of some tasks. For these reasons, we are unable to compare the present cut-off scores with those of other studies. However, 2 earlier investigations have shown ACE-R raw scores obtained by patients with mild AD. Mioshi et al 20 found the mean total score to be 3 points higher than ours. Similarly, these same investigators 29 observed a median value of the total score slightly higher (71) than the score found in this study (66). These minor discrepancies may be secondary to distinct age, education, or sample sizes between the studies. The mean age and schooling of our AD sample were ± 6.74 and 9.97 ± 5.19, respectively. Further research including populations with younger age and lower educational levels must be conducted, considering the diversity of our population and also the possible influences of these factors on the performance in cognitive tests. 30,31 A significant difference in scores in the CSDD 26 among groups was observed. However, it is important to reiterate that none of the subjects included in this study met the diagnostic criteria for depression. Although patients with these characteristics were not included here, future studies comparing the performance of patients with AD with and without depressive symptoms are 10 r 2010 Lippincott Williams & Wilkins

4 Cog Behav Neurol Volume 23, Number 1, March 2010 ACE-R in the Diagnosis of Mild AD TABLE 4. Samples Divided by Schooling <11 11 or More AD Patients Controls P AD Patients Controls P Age (7.98) (6.61) T = * (5.48) (6.65) T = * Education 5.47 (2.32) 5.52 (2.03) U = w (3.14) (2.95) U = w Sex w 2 = z w 2 = z *Student t test. wmann-whitney test. zw 2 test. AD indicates Alzheimer disease. recommended, considering the high prevalence of these symptoms in patients with dementia. 32 It is also important to ascertain the influence of depressive factors on ACE-R performance, as has been shown for other cognitive tests. 33 In this study, ACE-R total scores have been influenced by education and age in control and only by the years of formal schooling in the patients with AD. These results are in agreement with a normative study of the Brazilian ACE-R conducted with a larger sample (n = 114) of healthy volunteers. 34 Sex interfered only with control group performance in this study. However, we did not find any association between sex and ACE-R total score in the sample of 114 healthy elderly individuals mentioned above. When controls and patients were considered as a group, only education showed significant correlation with the test. We also divided the sample into 2 groups according to educational level (GHE and GLE). As expected, GHE performed better than GLE. Two cut-off scores were identified, with the GHE maintaining the same initial value (<78 points), whereas in the GLE a lower cut-off score for the ACE-R (<68 points) was found. The significant influence of education on the performance in the battery was also confirmed in a large study conducted in India using the first version of the ACE, in which subjects were stratified by different levels of education. 35 We hypothesized that this phenomenon is a result of the relatively high demands of some parts of the test, which could be more difficult for people with modest cognitive acquisition during their childhood and professional life. Both the ACE-R total score and the MMSE (ACE-R) had high diagnostic accuracy. The AUC of the former was slightly higher than the latter (0.947 vs , respectively). Bier et al, 10 Alexopoulos et al, 11 and Stokholm et al 19 also showed a similar feature for the ACE. Beyond the ACE-R total score, Attention and Orientation, Memory, and Fluency subscales also showed similar levels of accuracy to the MMSE. Nonetheless, we consider that the ACE-R FIGURE 1. ROC curve of the ACE-R as a test for detecting early AD in low education group. ACE-R indicates Addenbrooke s Cognitive Examination; AD, Alzheimer disease; ROC, receiver operator characteristic. FIGURE 2. ROC curve of the ACE-R as a test for detecting early AD in high education group. ACE-R indicates Addenbrooke s Cognitive Examination; AD, Alzheimer disease; ROC, receiver operator characteristic. r 2010 Lippincott Williams & Wilkins 11

5 Carvalho et al Cog Behav Neurol Volume 23, Number 1, March 2010 TABLE 5. Cut-off Scores by Education Groups Group Cut-off Sensitivity Specificity AUC (SD) <11 (control/patients: 29/15) < (0.021) 11 or more (control/patients: 33/16) < (0.033) Comparison of the ROC curves: Z = 0.946; P value = AUC indicates area under the curve; ROC, receiver operator characteristic; SD, standard deviation. offers the advantage of exploring different patterns of cognitive deterioration and thus may be more suitable for the evaluation of distinct dementia neuropsychologic profiles. Bak and Mioshi 36 emphasize that the MMSE focuses on the evaluation of skills related to verbal memory, attention, and visuospatial functions, disregarding executive functioning and more sophisticated cognitive abilities. Furthermore, memory and language are investigated superficially. These authors also noted that the test does not offer qualitative information between different cognitive domains and disease profiles. Although ACE-R administration takes a few minutes more than the MMSE and its diagnostic accuracy is similar, we believe that the ACE-R can provide more information about cognitive status than the MMSE, in addition to its advantage in detecting frontotemporal dementia. Moreover, owing to the high prevalence and the variability of etiologies for dementia in the elderly population, there is a growing consensus on the need for routine evaluation of cognitive functioning by health professionals and primary medical care. 37 We consider that even if only a few minutes are available for cognitive testing in a clinical setting, the ACE-R may offer additional information about the dementia feature and major data for differential diagnosis. In conclusion, the ACE-R seems to be a very useful toolforabriefassessmentofsomeofthemaincognitive domains affected in cortical dementias also in developing countries as well. ACKNOWLEDGMENTS The authors thank Professor John Hodges and Dr Eneida Mioshi for their kind support for this study. They also thank all the volunteers who agreed to participate and the Institutions that permitted the research, especially the SESC (Servic o Social do Come rcio de Minas Gerais) and Minas Tenis Clube - the Silver Head Project. REFERENCES 1. McKhann G, Drachman D, Folstein M, et al. 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. 1984;34: Teng EL. Cultural and educational factors in the diagnosis of dementia. Alzheimer Dis Assoc Disord. 2002;16:S77 S Nitrini R, Caramelli P, Bottino CMC, et al. Diagno stico de doenc a de alzheimer no brasil: avaliac ão cognitiva e funcional. recomendac o es do departamento científico de neurologia cognitiva e do envelhecimento da academia brasileira de neurologia. Arq Neuropsiquiatr. 2005;63: Mathuranath PS, Nestor PJ, Berrios GE, et al. A brief cognitive test battery to differentiate Alzheimer s disease and frontotemporal dementia. Neurology. 2000;55: Folstein MF, Folstein SE, McHugh PR. Mini-mental state. A practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res 1975;12: Dudas RB, Berrios GE, Hodges JR. The Addenbrooke s cognitive examination (ACE) in the differential diagnosis of early dementias versus affective disorder. Am J Geriatr Psychiatry. 2005;13: Bak TH, Rogers TT, Crawford LM, et al. Cognitive bedside assessment in atypical parkinsonian syndromes. J Neurol Neurosurg Psychiatry. 2005;76: Galton CJ, Erzinc lioglu S, Sahakian BJ, et al. A comparison of the Addenbrooke s cognitive examination (ACE), conventional neuropsychological assessment, and simple MRI-based medial temporal lobe evaluation in the early diagnosis of Alzheimer s disease. Cogn Behav Neurol. 2005;18: Davies RR, Dawson K, Mioshi E, et al. Differentiation of semantic dementia and Alzheimer s disease using the Addenbrooke s cognitive examination (ACE). Int J Geriatr Psychiatry. 2008;4: Bier JC, Donckels V, Van Eyll E, et al. The French Addenbrooke s cognitive examination is effective in detecting dementia in a Frenchspeaking population. Dement Geriatr Cogn Disord. 2005;19: Alexopoulos P, Greim B, Nadler K, et al. Validation of the Addenbrooke s cognitive examination for detecting early Alzheimer s disease and mild vascular dementia in a German population. Dement Geriatr Cogn Disord. 2006;22: García-Caballero A, García-Lado I, González-Hermida J, et al. Validation of the Spanish version of the Addenbrooke s cognitive examination in a rural community in Spain. Int J Geriatr Psychiatry. 2006;21: Sarasola D, Calcagno ML, Sabe L, et al. Utilidad del Addenbrooke s cognitive examination en espanol para el diagnostico de demencia y para la diferenciacon entre la enfermedad de alzheimer y la demencia frontotemporal. Rev Argent Neuropsicol. 2004;4: Sarasola D, Luján-Calcagno M, Sabe L, et al. El Addenbrooke s cognitive examination en espan ol para el diagnóstico de demencia y para la diferenciacio n entre enfermedad de alzheimer y demencia frontotemporal. Rev Neurol. 2005;41: Roca M, Torralva T, Lo pez P, et al. Diferenciacio n entre demencias en estadio inicial y depresión utilizando la versio n española del Addenbrooke s cognitive examination. Rev Neurol. 2008;46: Mathuranath PS, Hodges JR, Mathew R, et al. Adaptation of the ACE for a Malayalam speaking population in southern India. Int J Geriatr Psychiatry. 2004;19: Shelley BP, Al-Khabouri J, Hussein SS, et al. Frontotemporal dementia in Oman: cognitive behavioural profile and neuroimaging characteristics. a prospective hospital-based study. J Neurol Sci. 2007;260: Newman J. Brief assessment of cognitive mental status in Hebrew: Addenbrooke s cognitive examination. Israel Medic Assoc. 2005;7: Stokholm J, Vogel A, Johannsen P, et al. Validation of the Danish Addenbrooke s cognitive examination as a screening test in a memory clinic. Dement Geriatr Cogn Disord. 2009;27: Mioshi E, Dawson K, Mitchell J, et al. The Addenbrooke s cognitive examination revised (ACE-R): a brief cognitive test battery for dementia screening. Int J Geriatr Psychiatry. 2006;21: Amaral-Carvalho V, Caramelli P. Brazilian adaptation of the Addenbrooke s cognitive examination-revised (ACE-R). Dement Neuropsychol. 2007;1: r 2010 Lippincott Williams & Wilkins

6 Cog Behav Neurol Volume 23, Number 1, March 2010 ACE-R in the Diagnosis of Mild AD 22. American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders (DSM). 3rd ed. Washington: American Psychiatric Association; Morris JC. Clinical dementia rating: a reliable and valid diagnostic and staging measure for dementia of the Alzheimer type. Int Psychogeriatr. 1997;9: Mattis S. Dementia Rating Scale: Professional Manual. Florida: Psychological Assessment Resources; Porto SC, Charchat HF, Caramelli P, et al. Dementia rating scale- DRS-in the diagnosis of patients with Alzheimer s dementia. Arq Neuropsiquiatr. 2003;61: Alexopoulos GS, Abrams RC, Young RC, et al. Cornell scale for depression in dementia. Biol Psychiatr. 1988;23: Alexopoulos GS, Abrams RC, Young RC, et al. Use of the Cornell scale in nondemented patients. J Am Geriatr Soc. 1988;36: Nitrini R, Bottino CMC, Albala C, et al. Prevalence of dementia in Latin America: a collaborative study of population-based cohorts. Int Psychogeriatr. 2009;21: Mioshi E, Kipps CM, Dawson K, et al. Activities of daily living in frontotemporal dementia and alzheimer disease. Neurology. 2007;68: Van Hooren SA, Valentijn AM, Bosma H, et al. Cognitive functioning in healthy older adults aged 64-81: a cohort study into the effects of age, sex, and education. Aging Neuropsychol Cogn. 2007;14: Machado TH, Fichman HC, Santos EL, et al. Normative data for healthy elderly on the phonemic verbal fluency task FAS. Dement Neuropsychol. 2009;3: Starkstein SE, Mizrahi R. Depression in Alzheimer s disease. Expert Rev Neurother. 2006;6: Herrmann LL, Goodwin GM, Ebmeier KP. The cognitive neuropsychology of depression in the elderly. Psychol Med. 2007;37: Amaral-Carvalho V. Addenbrooke s Cognitive Examination-Revised (ACE-R): Cross-cultural adaptation, normative data for cognitively preserved elderly and its applicability as a brief cognitive evaluation tool for mild probable Alzheimer s disease patients. (Master s thesis). Sa o Paulo: University of São Paulo, School of Medicine; Mathuranath PS, Cherian JP, Mathew R, et al. Mini mental state examination and the Addenbrooke s cognitive examination: effect of education and norms for multicultural population. Neurol India. 2007;55: Bak TH, Mioshi E. A cognitive bedside assessment beyond the MMSE: the Addenbrooke s cognitive examination. Pract Neurol. 2007;7: Brodaty H, Howarth GC, Mant A, et al. General practice and dementia. A national survey of Australian GPs. Med J Aust. 1994; 160: r 2010 Lippincott Williams & Wilkins 13