Age and Ageing 1996.25:209-21 3 Cognitive Decline in atients with Alzheimer's Disease, Vascular Dementia and Senile Dementia of Lewy Body Type C. BALLARD, A. ATEL, F. OYEBODE, G. WILCOCK Summary One hundred and twenty-four patients with DSM-III-R dementia were assessed with a standardized battery which included the Geriatric Mental State Schedule, the History and Aetiology Schedule, the Secondary Dementia Schedule and the CAMCOG. Eighty-nine patients were followed up for 1 year, 76 of whom completed a repeat CAMCOG. atients with Alzheimer's disease, vascular dementia and senile dementia of Lewy body type (SDLT) all had a similar degree of cognitive impairment at the time of the baseline interview. atients with Alzheimer's disease and vascular dementia each experienced a mean decline of 13 points on the CAMCOG in 1 year compared with a mean decline of 27 points in patients with SDLT. atients with SDLT had a significantly greater decline of verbal fluency than both the other groups. Women were significantly more impaired than men at the time of the baseline assessment but experienced a similar decline during the year of follow-up. Introduction Cognitive decline is an integral part of Alzheimer's disease. A patient may be expected to deteriorate by 3 to 4 points on a Mini-Mental State Examination [1] or approximately 12 points on a CAMCOG schedule [2] annually. There has been some suggestion that patients with mild impairment decline at a faster rate [3-5] and that women with Alzheimer's disease have worse cognitive functioning than men with a comparable duration of illness [6], although other studies have been unable to confirm these findings [2]. Early onset of Alzheimer's disease has also been suggested as a factor related to more rapid cognitive decline, with a two-fold difference evident between patients aged under- and over-65 years [7], although other reports have not found a significant association [2]. The presence of extrapyramidal symptoms was linked to more severe cognitive impairment in one report [8]. Samples have generally been small and the sample origins have differed between studies and this may partially explain some of the discrepancies. Less attention has been focused upon vascular dementia. Almkvist [9] has reviewed the literature comparing patients with vascular dementia and Alzheimer's disease. Controlling for the duration of illness, there appears to be no difference in the overall severity of cognitive impairment but patients with vascular dementia tend to have slower reaction times, more impaired verbal fluency and more difficulties with planning tasks. Senile dementia of Lewy Body type (SDLT) has received little attention but planning tasks have been found to be particularly severely impaired [10] and neuropathological studies have reported hippocampal sparing [11]. Treatment trials involving patients with vascular dementia have suggested an annual decline of approximately 4% on formal tests of cognitive function [12]. The rate of progression of SDLT is unknown. This study reports the results of a CAMCOG assessment of a sample of dementia patients in contact with clinical services at baseline and after 1 year of follow-up. Methods The case notes of consecutive referrals to four old-age psychiatry services in the West Midlands and a Memory Clinic in Bristol were reviewed. All patients aged 65 years or over who fulfilled the CAM DEX criteria for mild or moderate dementia [13] and had an informant in contact at least once a week were selected. All those who fulfilled the entry criteria were contacted together with their carer and asked if they would like to participate in the study. This procedure continued from April 1993 until the target number of patients had been recruited in December 1993. A comprehensive assessment was completed which included a Geriatric Mental State Schedule [14] and History and Aetiology Schedule [15]. Cognitive function was assessed using the CAMCOG schedule [13]. The total score was identified as well as the scores on each of the sub-sections [16] orientation, comprehension, expression, praxis, recent memory, visual
C. BALLARD ET AL. memory, remote memory, attention and calculation, perception (agnosia), verbal fluency and abstract thinking. Information regarding a range of potential associations including deafness, education and the age of onset of the dementia was obtained from the History and Aetiology Schedule [15]. Depression was assessed using the Cornell Depression Scale [17] and psychotic symptoms were assessed using the Burns Symptom Checklist [18]. Dementia was diagnosed according the NINCDS ADRDA [19], DSM-III-R [20], the Hachinski scale [21], criteria for SDLT [22] and the HAS AGECAT [15]. The method of application is described elsewhere [18]. One year after the initial interview, each patient was re-interviewed and the CAMCOG schedule was again administered. The sub-scales were derived in the same manner as described for the initial interviews. Descriptive data are reported including the mean baseline total CAMCOG scores. These scores were compared between patients with the different dementias using the Mann- Whitney U test. A stepwise logistic regression analysis [23] was used to compare the CAMCOG sub-scores between patients with Alzheimer's disease and vascular dementia, Alzheimer's disease and SDLT, and vascular dementia and SDLT, respectively. In addition, a linear regression analysis was undertaken to explore the association of total CAMCOG scores and a range of demographic factors including sex, age, age at onset, Cornell Depression Score, deafness, number of years of education and the presence of psychotic symptoms. Spearman's rank correlation coefficients between each of the CAMCOG sub-scores and the overall CAMCOG score were included. These were calculated separately for patients with Alzheimer's disease, vascular dementia, and SDLT. The number of points of deterioration on the total CAMCOG score at 1 year's follow-up is described for patients with Alzheimer's disease, vascular dementia and SDLT. The magnitude of deterioration between the different dementias was compared using the Mann-Whitney U test. The magnitude of deterioration over 1 year was also evaluated using a linear regression analysis with sex, age, age at onset, Cornell Depression Scores, deafness, number of years of education and psychotic symptoms as the independent variables. Deterioration in each of the CAMCOG subscores was compared between patients with Alzheimer's disease and vascular dementia and between men and women using logistic regression analysis. Results Of the patients approached, 90.4% agreed to participate in the study. Of the 125 participants, 124 fulfilled the DSM-III-R criteria for dementia. The mean age of the subjects was 79.6 years, 91 were women and 33 men. Two were inpatients, 39 day patients and the remainder outpatients. Fifty-five patients had probable Alzheimer's disease, 33 possible Alzheimer's disease, 20 vascular dementia and 12 SDLT. Four patients fulfilled the DSM-III-R criteria for dementia but could not be classified further. One hundred and sixteen patients completed a baseline CAMCOG assessment. Six had impaired vision and two were unco-operative. Eighty-nine patients completed the year of followup. Twenty-one patients died during the course of the study, four moved to different areas, five dropped out and the follow-up information was incomplete for five patients. Seventy-six of the 89 patients had repeated CAMCOG assessments at 1 year, seven patients were unable to complete the repeat CAMCOG because of poor eyesight and six were unco-operative. Twentyeight of the patients not completing follow-up had Alzheimer's disease, five had vascular dementia and three had SDLT. Their mean baseline CAMCOG score was 38.58 (Alzheimer's disease mean 39.4, vascular dementia mean 39.2, SDLT mean 30.0). Fifty-three of the patients completing the repeat cognitive testing suffered from Alzheimer's disease, Table I. Comparing the baseline cognitive assessments between patients with the different dementias using a logistic regression analysis, statistics and probability (p) values vascular dementia senile dementia of Lewy body type Vascular dementia vs. senile dementia of Lewy body type CAMCOG score Total Orientation Comprehension Expression raxis Recent memory Visual memory Remote memory Attention and calculation erception (agnosia) Abstract thinking Verbal fluency 0.19 2.01 0.25 1.27 5.30 7.94 0.43 1.89 1.15 0.16 0.005* 0.39 0.51 0.17 0.28 2.10 0.11 0.79 0.55 5.41 1.82 0.03 0.15 0.74 0.37 0.85 1.51 1.56 1.29 0.40 0.05 4.18 2.09 0.33 0.20 0.01 0.37 0.21 0.04 0.15 0.93 0.61
COGNITIVE DECLINE IN THREE TYES OF DEMENTIA Table II. A linear regression analysis of variables associated with the severity of cognitive impairment: T-values and probability (p) Table IV. Linear regression analysis of variables associated with the magnitude of cognitive decline over 1 year: t statistic and p values Variable Variables Female sex Age at onset of dementia Age of patient at assessment Cornell depression score Deafness Number of years of education One or more psychotic symptom 2.47-0.63 0.28 0.43-0.51-0.01» 0.94 0.78 0.96 Female sex Age of onset of dementia Age of patient at assessment Cornell depression score Deafness Number of years of education One or more psychotic symptom 0.11 0.24 0.03-0.83 0.89 0.59 0.91 0.81 0.93 0.60 14 had vascular dementia and seven had SDLT. Twenty-one were men and 55 were women. Overall, the mean CAMCOG scores of the sample were 43.86 with a standard deviation of 19.30. The range was 1-86. atients with Alzheimer's disease had a mean score of 42.68 (SD 17.94), patients with vascular dementia had a mean score of 44.50 (SD 24.30) and patients with SDLT had a mean score of 47.67 (SD 18.0). None of these differences was significant (Mann- Whitney U test: vascular dementia: U = 862.5; z = -; p = 0.90, Alzheimer's disease vs. SDLT: U = 441; z = -0.92; p = 0.36, vascular dementia vs. SDLT: U = 105.5; z = 0.56; p = ). Stepwise logistic regression analyses showed that patients with vascular dementia had significantly better recent memory ( = 5.30; p = 0.02) and significantly better visual memory ( = 7.94; p = 0.005) than patients with Alzheimer's disease. atients with SDLT had significantly better recent memory ( = 5.41; p = 0.02) than patients with Alzheimer's disease and significantly better visual memory ( = 4.1; p = 0.004) than patients with vascular dementia. These results are shown in more detail in Table I. Table III. Correlation coefficients (earson) between CAMCOG sub-scores and total CAMCOG scores Orientation 0.69 Comprehension 0.74 Expression 0.77 raxis 0.68 Recent memory 0.58 Visual memory 0.55 Remote memory 0.69 Attention and calculation 0.49 erception 0.61 Abstract thinking 0.55 Verbal fluency 0.56 Senile dementia of Lewy body type. Alzheimer's Vascular disease dementia SDLT* 0.84 0.74 2.92 0.77 0.36 0.60 0.68 0.79 0.09 0.54 The linear regression analysis looking at the variables sex, age, age at onset, Cornell Depression score, deafness, number of years of education and psychotic symptoms showed only an association between male sex and better total scores on the CAMCOG schedule (t = 2.47; p = 0.01). Men had a mean CAMCOG score of 50.85 compared with a mean of 41.34 by women. The results of the linear regression are shown in more detail in Table II. Examining the correlations between the sub-scores and the overall CAMCOG scores, in patients with Alzheimer's disease, only one of the 11 sub-scores had a correlation of < with the total score. Ten of the 11 sub-scales had a correlation of >0.70 with the total CAMCOG score amongst patients with vascular dementia. For patients with SDLT, there were poor correlations between the overall CAMCOG score and abstract thinking (R = 0.09) and recent memory (R = 0.36) respectively. The correlations are shown in more detail in Table III. At 1-year follow-up, patients with Alzheimer's disease showed a mean deterioration of 13.21 (SD 12.61) points on the total CAMCOG schedule, compared with a deterioration of 13.29 (SD 13.48) points for patients with vascular dementia and 27.00 (SD 19.77) points for patients with SDLT. There was a trend for patients with SDLT to have a greater deterioration in their overall cognitive function than for patients with Alzheimer's disease (Mann Whitney U test: U = 102; z = 1.93; p = 0.05) and patients with vascular dementia (Mann-Whitney U test: U = 363.5; z = 1.79; p = ). There was no significant difference in the overall deterioration of cognitive function between patients with Alzheimer's disease and vascular dementia (Mann-Whitney U test: U = 363.5; z = 0.11; p = 0.90). None of the variables assessed in the linear regression analysis was significantly associated with a greater rate of cognitive deterioration; this is shown in Table IV. There were no significant differences between the deterioration of any of the sub-scores in patients with Alzheimer's disease and vascular dementia. atients with SDLT, however, showed a significantly greater deterioration in verbal fluency ( = 5.63; p = 0.02)
212 C. BALLARD ET AL. Table V. Comparison of depression on CAMCOG sub-scores between patients with different dementias using a logistic regression analysis, scores and probability (p) vascular dementia Alzheimer SDLT 's disease vs. Vascular SDLT dementia vs. CAMCOG deterioration Total Orientation Comprehension Expression raxis Recent memory Visual memory Remote memory Attention and calculation erception (Agnosia) Abstract thinking Verbal fluency 0.02 0.13 0.13 1.66 0.33 0.25 0.19 0.00 2.38 0.98 0.72 0.71 0.61 0.41 0.20 0.98 3.36 0.49 0.47 0.77 1.50 1.11 5.63 0.05 0.48 0.49 0.38 0.29 3.30 3.56 0.99 1.00 3.37 1.97 6.18 1.07 0.71 0.32 10.94 0.06 0.32 0.31 0.16 0.01* 0.30 0.40 0.0009* SDLT = senile dementia of Lewy body type. than did patients with Alzheimer's disease and patients with vascular dementia ( = 10.95; p = 0.0009) and a significantly greater deterioration in remote memory than patients with vascular dementia ( = 6.18; p = 0.01). These results are shown in more detail in Table V. Discussion The sample was not representative of dementia patients in the community although in view of the high participation rate it should be representative of patients with mild to moderate dementia in contact with clinical services. A standardized cognitive assessment was undertaken at baseline and at 1-year follow-up. Approximately 70% of the patients were re-interviewed at that time. A similar proportion of patients with each type of dementia dropped out of the study. Only three patients with SDLT did not complete the study, and they had a lower mean CAMCOG score than those remaining in the study. This should not have been a major bias for betvveen-group comparisons as the comparisons between baseline and 1-year CAMCOG scores were only made for the 76 subjects completing the study. The study included patients with Alzheimer's disease, vascular dementia and SDLT, a comparison not previously made. The overall level of cognitive deterioration was approximately 13 points on the CAMCOG schedule in 1 year amongst patients with Alzheimer's disease. This was similar to the 12 points described by Burns et al. [2] on the same schedule. The baseline cognitive scores were similar in patients with the three different dementia types, although the linear regression analysis showed male sex to be significantly associated with higher CAMCOG scores. This is similar to the study by Henderson and Buckwalter [6] and is of particular interest in view of recent suggestions that oestrogens may modulate the progression of cognitive impairment [24]. Sex was not however an important factor in explaining the magnitude of cognitive deterioration over the 1-year follow-up. It is possible that if gender does exert an influence it may do so at an earlier stage of the disorder, and a plateau may then be reached. At the time of the baseline interview, patients with vascular dementia had significantly less impairment of recent memory and visual memory than patients with Alzheimer's disease, although there was no significant difference in the overall level of cognitive impairment or verbal fluency. This supports the conclusions of Almkvist [9] regarding the overall severity of cognitive impairment, but differences in individual cognitive domains require further clarification. Although patients with Alzheimer's disease and SDLT had a similar level of cognitive impairment at the time of the baseline interview, patients with SDLT showed significantly less impairment of recent memory. This is consistent with neuropathological data suggesting hippocampal sparing [11]. In addition, over the year of follow-up, patients with SDLT showed a trend towards a greater overall decline in cognitive function, with a deterioration of 27 points on the CAMCOG score. The small number of SDLT patients followed up for 1 year must be considered when interpreting the data. atients with SDLT had significantly greater deterioration in their verbal fluency than patients with Alzheimer's disease or vascular dementia. This is consistent with work [10] suggesting that frontal lobe functioning is predominantly affected in SDLT patients.
COGNITIVE DECLINE IN THREE TYES OF DEMENTIA 213 Compared with patients with Alzheimer's disease there was a poor correlation between abstract thinking and the total CAMCOG score in SDLT patients. This might form the basis of a diagnostic instrument if a reference range could be established. Acknowledgement We thank the MRC for support. Rejerences 1. Morris J, Edlund S, Clark MD, et al. The Consortium to Establish a Register for Alzheimer's Disease (CERAD): art IV. Rate of cognitive change in the longitudinal assessment of probable Alzheimer's disease. Neurology 1993;43:2457-65. 2. Burns A, Jacoby R, Levy R. rogression of cognitive impairment in Alzheimer's disease. J Am Geriatr Soc 1991;39:39-45. 3. Thai LJ, Grundman M, Klauber MR. Dementia: characteristics of a referred population and factors associated with progression. Neurology 1988;33:1083-90. 4. Katzman R, Brown T, Thai LJ, et al. Comparison of rate of annual change for mental status score in 4 independent studies of patients with Alzheimer's disease. Ann Neurol 1988;24:384-9. 5. Mortimer JA, Ebbitt B, Jun S, Finch MD. redictors of cognitive and functional progression in patients with probable Alzheimer's disease. Neurology 1992;42:1689-96. 6. Henderson VW, Buckwalter JG. Cognitive deficits of men or women with Alzheimer's disease. Neurology 1994;44:90-6. 7. Jacobs D, Sano M, Mardker K. Age at onset of Alzheimer's disease: relation to pattern of cognitive dysfunction and rate of decline. Neurology 1994;44: 1215-20. 8. Richards M, McLoughlin D, Levy R. The relationship between extrapyramidal signs and cognitive function in patients with moderate to severe Alzheimer's disease. Int J Geriatr sychiatry 1995;10:395-9. 9. Almkvist O. Neuropsychological deficits in vascular dementia in relation to Alzheimer's disease: reviewing evidence for functional similarity or divergence. Dementia 1994;S:203-9. 10. Sahgal A, Galloway H, McKeith IG, Edwardson JA, Lloyd S. A comparative study of attentional deficits in senile dementia of Alzheimer's and Lewy body types. Dementia 1992;3:350-4. 11. Gibb WRG, Esiri MM, Lees AJ. Clinical and pathological features of diffuse Lewy body disease (Lewy body dementia). Brain 1987; 110:1131-53. 12. Meyer JS, Rodgers RL, McClintic K, Mortel KF, Lofti R. A randomized clinical trial of aspirin therapy in multiinfarct dementia, jf Am Geriatr Soc 1989;37:549-55. 13. Roth M, Tym E, Mountjoy C, et al. CAMDEX: a standardised instrument for the diagnosis of mental disorder in the elderly with special reference to the early detection of dementia. Br J sychiatry 1986; 149: 698-709. 14. Copeland JRM, Kelleher MJ, Kellet JM. A semistructured clinical interview for the assessment of diagnosis and mental state in the elderly: The Geriatric Mental State: I. Development. sychol Med 1976;6:439-49. 15. Dewey ME, Copeland JRM, Lobo A, Saz, Dia JL. Computerised diagnosis from a standardised history schedule: a preliminary communication about the organic section of the HASAGECAT system. Int J Geriatric sychiatry 1992,7:443-6. 16. Hooper J, Bucks R. The Cognitive rofile: development of a cognitive assessment and profile based on the CAMCOG. resented to the sychologists' Special Interest Group for the Elderly. Southmead Hospital, Bristol, 1993. 17. Alexopoulos GS, Abrams RC, Young RC, Shamoian CA. Cornell Scale for Depression in Dementia. Biol sychiatry 1988,23:271-34. 18. Ballard CG, Saad K, Gahir M, Solis M, Coope B, Wilcock G. The prevalence and phenomenology of psychotic symptoms in dementia sufferers. Int J Geriatr sychiatry 1995;10:477-85. 19. McKhann G, Drachman D, Folstein M, Katzman R, rice D, Stadlan EM. Clinical diagnosis of Alzheimer's disease: Report of the NINCDSADRDA Work Group under the auspices of Department of Health and Human Services Task Forces on Alzheimer's Disease. Neurology 1984;34:939-44. 20. American sychiatric Association. Diagnostic and statistical manual of mental disorders. Revised 3rd edn. Washington DC: American sychiatric Association. 21. Hachinski VC, Ilief LD, Zilhka E, Cerebral bloodflowin dementia. Arch Neurol 1975;32:632-7. 22. McKeith IG, erry RH, Fairbairn AF, Jabeen S, erry EK. Operational criteria for senile dementia of Lewy body type. sychol Med 1992;22:911-22. 23. SSS/CT. Statistical package for the social sciences. Chicago, 1988. 24. Henderson VW, aganini-hill A, Emanuel CK, Dunn ME, Buckwalter JG. Estrogen replacement therapy in older women. Arch Neurol 1994;51:896-900. Authors' addresses C. Ballard MCR Neurochemical athology Unit, Newcastle General Hospital, Westgate Road, Newcastle upon Tyne NE4 6BE A. atel, F. Oyebode Queen Elizabeth sychiatric Hospital, Birmingham B15 2QZ G. Wilcock Department of Health Care of the Elderly, Frenchay Hospital, Bristol BS16 1LE Received in revised form 2 October 1995