- Email: [email protected]
Effects of Anxiety Versus Depression on Cognition in Later Life
Effects of Anxiety Versus Depression on Cognition in Later Life E.J.M. Bierman, M.Sc., H.C. Comijs, Ph.D. C. Jonker, M.D., Ph.D., A.T.F. Beekman, M.D., Ph.D.
Objective: The authors investigated the relationship between anxiety and cognition in older persons, taking account of comorbid depression. Methods: Data were used from the Longitudinal Aging Study Amsterdam (LASA), a large epidemiological study of 3,107 elderly citizens in The Netherlands. Anxiety and depression were measured with the Hospital Anxiety and Depression Scale–Anxiety subscale and the Center for Epidemiologic Studies–Depression Scale. In measuring cognitive performance, general cognitive functioning was measured by means of Mini-Mental State Exam, episodic memory was measured with the Auditory Verbal Learning Test (AVLT), fluid intelligence by using the RAVEN, and information-processing speed by the coding task. Analysis of variance examined the association between anxiety symptoms and cognition in persons with and without depression. Results: Main effects of anxiety symptoms were found for learning and delayed recall of the AVLT. Depression symptoms showed significant main effects on almost all cognitive performance tests. Mild anxiety symptoms were associated with better cognitive performance, whereas severe anxiety symptoms were negatively associated with cognitive functioning. In contrast, depressive symptoms showed a linear association with cognition; more depression was associated with worse cognition. Conclusion: This study suggests that anxiety has a curvilinear relationship with cognition. Depressive symptoms, however, were always negatively associated with cognitive performance. (Am J Geriatr Psychiatry 2005; 13:686–693)
C
ognitive impairment is common in elderly persons, with prevalence rates ranging from 16.8% to 26.6% in the general population of older persons.1–5 Anxiety is also very common in later life. Results found by Beekman et al.6 showed that in the population age 65 or over, the prevalence of anxiety disorders is between 0.7% and 12%. Little research has been done on the association between anxiety and cognitive functioning in elderly
persons. Boenders et al.7 investigated memory performance in relation to anxiety and achievement motivation in a small elderly population who were referred to a geriatric outpatient clinic because of complaints about their memory. No association was found between anxiety and memory performance or severity of dementia. Wetherell et al.8 studied the cross-sectional association between cognitive performance and state-anxiety in cognitively intact elderly
Received July 5, 2004; revised December 17, 2004; accepted December 22, 2004. From the Dept. of Psychiatry, Institute for Research in Extramural Medicine (EMGO Institute) of the Vrije Universiteit Medical Centre, Amsterdam, The Netherlands. Send correspondence and reprint requests to Dr. E.J.M. Bierman, LASAEMGO, VU Medical Centre, Van der Boechorststraat 7,1081 BT, Amsterdam, The Netherlands. e-mail: [email protected] 䉷 2005 American Association for Geriatric Psychiatry
686
Am J Geriatr Psychiatry 13:8, August 2005
Bierman et al. subjects and found that higher state-anxiety was associated with poorer performance on some, but not all, cognition tests used. There are several reasons to assume an association between anxiety and cognitive functioning. Eysenck’s processing efficiency theory9 holds that anxiety interferes with cognitive performance by pree¨mpting some of the processing and storage resources of the working memory system. It posits that anxiety produces worry and other intrusive thoughts, which are verbal and therefore intrude with the phonological loop of working memory. Mendl10 provides a second possible explanatory model for the association between anxiety and cognitive functioning, citing the Yerkes and Dodson law.11 This law states that there exists an inverted U-shaped relationship between arousal and cognitive performance. Cognitive performance is best when an individual is in an optimal stress or arousal state, above or below which performance levels may drop. When investigating anxiety in older persons, comorbidity with depression cannot be disregarded, because it is very common, especially in later life.12 Merikangas et al.13 emphasized the importance of comorbidity, finding that mixed anxiety-anddepression has a greater stability over time than either disorder alone. An association between depressive symptoms and cognitive decline was found in several studies. Comparing depressed subjects with non-depressed control subjects, a metaanalysis of studies found reduced performance on almost all cognitive performance tests.14 According to Bassuk et al.,15 persons who are depressed may have (reversible) cognitive deficits as a consequence of motivational and attentional problems. Jorm16 stated that depression might also be a reaction to cognitive decline. Furthermore, Airiksinen et al.17 showed that cognitive impairment was associated with mixed anxiety-depressive disorders, considering this as one and the same disorder. Given the high correlations between presence of anxiety and depression, one may suppose a similar association between anxiety and cognitive performance as is found for depression and cognitive performance. In this study, we will investigate how anxiety symptoms are associated with cognitive performance in a large epidemiological study among elderly subjects. We will test the “inverted-U” hypothesis and attempt to separate the effect of anxiety symptoms on
Am J Geriatr Psychiatry 13:8, August 2005
cognitive functioning from the effect of depression symptoms on cognitive functioning.
METHODS Sample To investigate the association between anxiety and cognition in older persons, we used data from the Longitudinal Aging Study Amsterdam (LASA18). LASA is a longitudinal study of predictors and consequences of changes in well-being and autonomy in older persons. Details on sampling are described elsewhere19–21 and will briefly be summarized here. The sample was originally drawn for NESTOR-Living Arrangements and Social Network (NESTOR–LSN;22), for which random samples of older residents (age 55– 85) were drawn from population registers in 11 municipalities in three regions of The Netherlands. Ten months later, in 1992, all 3,805 respondents who participated in NESTOR–LSN were approached for LASA, of whom 3,107 (81.7%) took part;23 394 (10.4%) refused to participate because of lack of interest; 134 (3.5 %) were too ill or cognitively impaired to be interviewed; 126 (3.3%) died before being interviewed; and 44 (1.2%) could not be contacted. Attrition was related to age but not to sex. As expected, the older-old potential subjects were more often found to be too ill or cognitively impaired to participate. Additional testing of cognitive functioning was available for 2,615 subjects. This sample consisted of 1,279 men (48.9%) and 1,336 women (51.1%). Participants had an average age of 70.2 (standard deviation [SD]: 8.7) years. Education level was clustered into three categories: low (62.2%), medium (26.0%), and high (11.8%). All interviews were conducted at the respondents’ homes by specially trained and intensively supervised interviewers. Informed consent was obtained from each participant, and the study was approved by the Medical Ethics Committee of the VU Medical Centre. Respondents were interviewed once every 3 years; for this study, interviews from the first LASA measurement were used. Measures General cognitive functioning was measured by means of the Mini-Mental State Exam (MMSE24), a
687
Cognition and Anxiety Versus Depression frequently used screening instrument for measurement of global cognitive functioning. The scale consists of 23 items, and scores range from 0 to 30; higher scores indicate better cognitive functioning. Episodic memory was measured with a modified Dutch version of the Auditory Verbal Learning Test (AVLT25,26). This test consists of 15 words, which have to be learned during five trials. After each trial, the respondent is asked to recall as many words as possible. After a period of 20 minutes, he or she is asked to name the words that were learned before. The number of trials was limited to three because of scarcity of interview time. The total number of words the respondent has learned during the three trials is the Learning Score, which ranges from 0 to 45. The number of words reproduced after 20 minutes is the Delayed Recall Score, ranging from 0 to 15. The percentage of reproduced words in the Delayed Recall condition, divided by the total number of words reproduced in the Learning condition, is called the Retention Score. Higher scores on all three variables indicate better memory functioning. Fluid intelligence, defined as the ability to deal with new information, was measured by means of two subsets of 12 items (A and B) of Raven’s colored progressive matrices.27 Each item consists of a drawing (matrix) of a pattern from which a section is missing. On the bottom of the page, six patterns are printed, one of which fits into the missing section. The respondent has to choose which of the six alternatives fits best. The items increase in difficulty, and so do the two sections. The 24 items result in a total score range of 0 to 24, with higher scores indicating better cognitive performance. Information-processing speed was measured by means of an adjusted version of the Coding Task.28 This is a timed task in which the respondent has to combine as many characters as possible according to a given example. The example shows 15 combinations of two characters in a row of double boxes (the substitution key). The test itself shows rows of double boxes in which only the upper boxes contain characters and the lower boxes are empty. The respondent has to name the missing characters corresponding to the characters in the upper boxes, using the substitution key, as quickly and accurately as possible. The task consists of three identical 1-minute trials. The score on each trial is the number of correctly com-
688
pleted characters. The result of the last trial is included in the present study. Anxiety symptoms were measured with the Hospital Anxiety and Depression Scale–Anxiety subscale (HADS–A29). The anxiety subscale consists of seven items and is self rated by the respondents. Each answer is rated on a 4-point scale, ranging from 0: “rarely or never” to 3: “mostly or always.” Higher scores indicate more anxiety symptoms. Scores ranging from 0 to 7 were considered in the normal range; scores from 8 to 10 as mild anxiety symptoms; and scores 11 or higher as moderate/severe anxiety symptoms.30 We also used a dichotomous score based on the commonly-used cut-off score of 7.29 Depression symptoms were measured by means of the Center for Epidemiologic Studies–Depression Scale (CES–D31,32). The CES–D is a self-report scale consisting of 20 items; it was developed to measure depression symptoms in the community. Each answer is rated on a 4-point scale, ranging from 0: “rarely or never” to 3: “mostly or always.” Higher scores indicate more depression symptoms. We also used a dichotomous score based on the commonlyused cut-off score of 16.33 Confounders Possible confounders that may be associated with cognition and the independent variables include sociodemographics, physical health, alcohol consumption, level of education, and use of benzodiazepines. The sociodemographic variables age, gender, and education level were included as confounders. The level of education was classified into nine groups and divided over three levels; low education (elementary not completed, elementary education, lower vocational education), medium education (general intermediate education, intermediate vocational education, general secondary education), and high education (higher vocational education, college education, and university education). Physical health was reflected by number of chronic diseases. The number of chronic diseases respondents suffered from was assessed by asking them whether they had any of the following: chronic nonspecific lung disease, cardiac disease, peripheral atherosclerosis, stroke, diabetes mellitus, arthritis, malignant neoplasms, or other chronic diseases. The number ranged from 0 to 8. The accuracy of self-reports of
Am J Geriatr Psychiatry 13:8, August 2005
Bierman et al. these diseases, versus general-practitioner information, was shown to be adequate and independent of cognitive impairment.34 Alcohol consumption was assessed with a questionnaire developed by the Dutch central office of statistics (Central Bureau of Statistics)35 and classified according to Garretsen’s Indication of Present Alcohol Use36 into five categories (excessive, severe, moderate, light, and non-drinker). Finally, the use of benzodiazepines was taken into account as possible confounder. Use of prescribed drugs was assessed by asking the respondents which medications they used; responses were compared with information on the drug containers provided by the respondents. Drug use was classified into categories according to the Anatomic Therapeutic Chemical (ATC) classification. For the present study, we assessed number of prescribed drugs. For psychotropic drugs, a separate variable was computed, indicating use of antidepressant drugs and/or hypnotic drugs and/or and anxiolytic drugs (coded as Yes or No). Statistical Analyses For descriptors, elderly respondents were divided in four groups on the basis of the cut-off scores on the anxiety and depression scales: 1) respondents with neither anxiety nor depression symptoms (ANX–/DEP–); 2) respondents with anxiety but no depression symptoms (ANXⳭ/DEP–); 3) respondents with no anxiety but with depression symptoms (ANX–/DEPⳭ); and 4) respondents with both anxiety and depression symptoms (ANXⳭ/DEPⳭ). Analyses of variance (ANOVAs) were conducted to determine the main effects of both anxiety and depression symptoms and look into the interaction effect of mixed anxiety/depression; p values lower than 0.05 were regarded as statistically significant. To further clarify the role of depression symptoms in the association between anxiety and cognitive performance, we hereafter divided the sample into respondents with and without anxiety symptoms. ANOVAs were used to investigate the association between anxiety symptoms and cognitive performance. In the first model, we investigated the unadjusted association between anxiety symptoms and cognition; in the second model, the association between anxiety symptoms and cognition was adjusted for depression symptoms; and, in the final
Am J Geriatr Psychiatry 13:8, August 2005
model, the association was also adjusted for additional confounders. For the question of whether the association between anxiety and cognitive performance is an inverted U-shaped function, we performed an ANOVA with two cut-off points on the HADS–A scale (7 and 10), with 7 as the cut-off point between no-anxiety and the presence of anxiety and 10 being the cut-off point between mild anxiety symptoms and moderate/severe anxiety symptoms.30 To avoid the influence of comorbid depression symptoms, all respondents with depression symptoms (CES–D scores ⱖ16)31 were excluded from analysis. To compare the association between anxiety and cognition with the association between depression and cognition, we conducted analysis of variance on the association between depression and cognition, with the cut-off points of 16 and 24 on the CES–D scale, with 16 being the cut-off point between “no depression symptoms” and “mild depression symptoms” and 24 being the cut-off point between “mild depression symptoms” and “moderate/severe depression symptoms.” In this analysis, respondents with anxiety symptoms (HADS–A scores ⱖ7) were excluded from analysis. Also, we tested the curvilinearity of the association between anxiety symptoms (adjusted for depression symptoms) and cognition and between depression symptoms (adjusted for anxiety symptoms) and cognition by adding a quadratic term in the regression analyses. Finally, we conducted linear analyses to ensure that the association between depression symptoms and cognitive performance was linear.
RESULTS Table 1 shows the characteristics of these elderly respondents, divided in four groups: 1) with neither anxiety nor depression symptoms (ANX–/DEP–); 2) with anxiety but no depression symptoms (ANXⳭ/DEP–); 3) with no anxiety but only depression symptoms (ANX–/DEPⳭ); and 4) with both anxiety and depressive symptoms (ANXⳭ/DEPⳭ). Analysis of variance shows significant main effects of anxiety symptoms on Learning and Delayed Recall of the AVLT. Depression symptoms show significant effects on all cognitive performance tests on the AVLT
689
Cognition and Anxiety Versus Depression except Retention (Table 2). There were no significant interaction effects (data not shown). Comparison of performance on cognitive tests among those with and without anxiety symptoms (Table 3) shows no differences. After adjustment for depression symptoms, a significant difference in the scores on all cognitive tests but AVLT Retention was found between Anxiety and No Anxiety symptoms to the advantage of the respondents with no anxiety symptoms. After additional adjustment, this significant difference between the group with and without TABLE 1.
anxiety disappeared for the Delayed Recall score on the AVLT. For investigating the possible inverted U-shaped relationship between anxiety and cognitive performance, we divided the respondents into three groups by their scores on anxiety and depression symptoms. Figure 1 shows the relationship between the level of anxiety symptoms and depression symptoms and performance on cognitive tests. In all cognitive tests, we found a curvilinear relationship with level of anxiety symptoms. This was in
Characteristics of the Study Sample
N Age, years Gender, N (%) Men Women Education level, N (%) Low Medium High Use of benzodiazepines, N (%) Alcohol consumption, N (%) No Light Moderate High Excessive Number of chronic diseases HADS–A CES–D MMSE Coding task Raven Progressive Matrices AVLT Learning Delayed recall Retention
Anxiety– / Depression–
Anxietyⴐ / Depression–
Anxiety– / Depressionⴐ
Anxietyⴐ / Depressionⴐ
2,108 69.76 (8.6)
94 69.07 (9.0)
136 73.44 (8.6)
185 70.65 (8.7)
1,085 (51.5) 1,023 (48.5)
34 (36.2) 60 (63.8)
52 (38.2) 84 (61.8)
63 (34.1) 122 (65.9)
1271 (60.3) 571 (27.1) 264 (12.5) 219 (10.4)
59 (62.8) 26 (27.7) 9 (9.6) 18 (19.1)
92 (67.6) 26 (19.1) 18 (13.2) 10 (20.6)
128 (69.2) 44 (23.8) 13 (7.0) 26 (30.8)
420 (19.9) 1113 (52.8) 464 (22.0) 77 (3.7) 21 (1.0) 1.27 (1.2) 1.47 (1.7) 5.10 (4.1) 27.32 (2.4) 25.76 (7.8) 18.12 (4.0)
22 (23.4) 52 (55.3) 18 (19.1) 2 (2.1) — 1.71 (1.2) 8.93 (2.7) 10.00 (4.1) 27.68 (2.2) 27.64 (7.3) 18.44 (3.8)
33 (24.3) 72 (52.9) 27 (19.9) 3 (2.2) 1 (0.7) 1.83 (1.3) 3.60 (1.8) 20.19 (4.4) 26.56 (2.7) 23.57 (7.5) 16.61 (4.0)
59 (31.9) 90 (48.6) 25 (13.5) 10 (5.4) 1 (0.5) 2.04 (1.29) 10.33 (3.2) 25.02 (8.0) 26.86 (2.5) 23.53 (6.8) 17.03 (4.3)
18.66 (6.2) 5.08 (2.8) 61.66 (25.9)
19.93 (6.0) 5.57 (2.6) 65.40 (22.9)
17.24 (6.2) 4.53 (2.7) 57.93 (26.4)
17.97 (6.4) 4.89 (2.7) 60.99 (26.7)
Note: Values are mean (standard deviation), unless otherwise indicated. HADS–A: Hospital Anxiety and Depression Scale–Anxiety subscale; CES–D: Center for Epidemiologic Studies–Depression Scale; MMSE: MiniMental State Exam; AVLT: Auditory Verbal Learning Test.
TABLE 2.
Main Effects of Anxiety and Depression on Cognitive Performance Main Effect: Anxiety
MMSE Coding task Raven Progressive Matrices AVLT Learning Delayed recall Retention
Main Effect: Depression
F[1, 2433]⳱3.45 F[1, 2433]⳱2.57 F[1, 2433]⳱1.58
p⳱0.06 p⳱0.11 p⳱0.21
F[1, 2433]⳱15.14 F[1, 2433]⳱21.91 F[1, 2433]⳱21.05
p ⬍0.01 p ⬍0.01 p ⬍0.01
F[1, 2433]⳱5.02 F[1, 2433]⳱4.68 F[1, 2433]⳱3.55
p⳱0.03 p⳱0.03 p⳱0.06
F[1, 2433]⳱7.47 F[1, 2433]⳱4.82 F[1, 2433]⳱1.93
p⳱0.01 p⳱0.03 p⳱0.17
Note: MMSE: Mini-Mental State Exam; AVLT: Auditory Verbal Learning Test.
690
Am J Geriatr Psychiatry 13:8, August 2005
Bierman et al. TABLE 3.
Analyses of Variance for Cognitive Performance in Anxiety and No-Anxiety Symptom Groups
MMSE Coding task Raven Progressive Matrices AVLT Learning Delayed recall Retention
F
p
Fa
pa
Fb
pb
F[1, 2523]⳱0.79 F[1, 2476]⳱2.26 F[1, 2467]⳱3.50
p⳱0.37 p⳱0.13 p⳱0.06
F[1, 2511]⳱12.11 F[1, 2465]⳱10.41 F[1, 2457]⳱9.21
p ⬍0.01 p ⬍0.01 p ⬍0.01
F[1, 2486]⳱6.86 F[1, 2440]⳱4.18 F[1, 2436]⳱3.76
p⳱0.01 p⳱0.04 p⳱0.05
F[1, 2532]⳱0.03 F[1, 2532]⳱0.22 F[1, 2532]⳱0.52
p⳱0.87 p⳱0.64 p⳱0.47
F[1, 2520]⳱11.38 F[1, 2520]⳱8.36 F[1, 2520]⳱3.65
p ⬍0.01 p ⬍0.01 p⳱0.06
F[1, 2495]⳱4.38 F[1, 2495]⳱2.61 F[1, 2495]⳱1.02
p⳱0.04 p⳱0.11 p⳱0.31
Note: MMSE: Mini-Mental State Exam; AVLT: Auditory Verbal Learning Test. a Adjusted for depression symptoms. b Adjusted for depression symptoms and other confounders (age, gender, education level, alcohol use, benzodiazepine use, and number of chronic diseases.
FIGURE 1.
Different Effects of Anxiety and Depression on Cognition in Elderly Subjects
28.0 AVLT learning/memory
21.5
MMSE
27.5 27.0 26.5 26.0 25.5
Anxiety Depression
18.5 17.5
Anxiety Depression
6.5 AVLT delayed recall
28 Coding Task
19.5
16.5
29
27 26 25 24 23
6.0 5.5 5.0 4.5
22
4.0
19.5
67.5
AVLT retention
19.0 RAVEN
20.5
18.5 18.0 17.5 17.0 16.5
65.5 63.5 61.5 59.5 57.5
No
Mild
Severe
No
Mild
Severe
Note: No anxiety/depression: HADS–A ⱕ7 and CES–D ⱕ16; Mild: 7 ⬎ HADS–A ⱕ 10 and 16 ⬎ CES–D ⱕ 24; Severe: HADS–A ⬎ 10 and CES–D ⬎ 24. MMSE: Mini-Mental State Exam;24 Coding task: Alphabet Coding Task 15;28 Raven: Raven Progressive Matrices test;27 AVLT: Auditory Verbal Learning Test.25,26
Am J Geriatr Psychiatry 13:8, August 2005
691
Cognition and Anxiety Versus Depression contrast with the declining scores on cognitive performance tests with increasing scores on the depression scale. The curvilinear relationship only reached the level of significance in Learning and Memory from the AVLT (F[2, 2199]⳱3.20; p⳱0.04). The other domains were not significant (MMSE: F[2, 2192]⳱0.46; p⳱0.63; Coding Task: F[2, 2155]⳱2.21; p⳱0.11; RAVEN: F[2, 2155]⳱1.85; p⳱0.16; AVLT Delayed Recall: F[2, 2199]⳱2.74; p⳱0.07; AVLT Retention: F[2, 2199]⳱ 0.99; p⳱0.37). Finally, the curvilinear relationship was tested with a quadratic term in the regression analyses. With anxiety symptoms, we found a quadratic relationship in all cognitive performance tests, reaching the level of significance for the MMSE, the Coding Task, and Learning and Memory from the Auditory Verbal Learning Test. For depression symptoms, we found no significant curvilinear relationship in all cognitive performance tests. Actually, linear analyses showed that the association between depression symptoms and cognitive performance was linear in all cognitive performance tests.
DISCUSSION The present study sought to investigate whether anxiety and depression have different effects on cognition in a large sample of community-living elderly persons. We found that mild anxiety symptoms seem to be beneficial, whereas severe anxiety negatively influences cognitive performance. Depression symptoms always negatively influence cognitive functioning. The association between anxiety symptoms and cognitive performance seems to be a curvilinear relationship. These findings support the YerkesDodson law, which states that optimum performance should be seen at levels of moderate arousal.11 These findings are strongest in the AVLT; it may be that this test is more sensitive to mental effort than the other cognitive measures. Our results also support Eysenck’s processingefficiency theory, in which he states that anxiety interferes with cognitive performance by pree¨mpting some of the processing and storage resources of the working memory system. Moderate levels of anxiety can have an arousing function that services behavioral performance up to a point, the point of disability
692
coming when the anxiety symptoms become so severe as to overwhelm information-processing resources. Moderate levels of depression would not be expected to have the initially beneficial arousing qualities, and, in addition to drawing attention (processing) recourses, depression may hinder motivation. We found that the relationship between anxiety symptoms and cognitive performance was strongly influenced by depressive symptoms. Correcting for depressive symptoms changed the effect on cognitive performance from negative to positive in most cognitive-performance tests, indicating a beneficial effect on cognition, especially for mild anxiety symptoms. The findings should be placed in the context of the strengths and limitations of this study. Strengths of the current study are that it is community-based, including large numbers of elderly subjects, with data available on relevant confounders and testing in four domains of cognition. However there are also limitations. First, selective non-response of the most frail participants is an inevitable problem in communitybased research among elderly subjects. This might have underestimated the effects of both anxiety and depression on cognitive functioning, because the subjects with the most severe anxiety and depression symptoms frequently drop out of studies like this. The second limitation is the cross-sectional nature of his study, in which it is impossible to determine causality because it is not possible to study whether the anxiety and depression symptoms proceed the cognitive decline. Longitudinal research is necessary to clarify this issue. Longitudinal research also makes it possible to investigate the influence of the duration of the anxiety symptoms on cognitive performance. Furthermore, our data are mainly based on selfreport measurements. No objective clinical judgment or diagnoses have been made. Finally, not all cognitive domains were included in this study. To investigate whether the curvilinear pattern was due to resource effects, measuring crystallized cognitive abilities would have been helpful. One might expect fewer differences among the three anxiety groups on a crystallized measure than on fluid measures, as we used in our study. The effects of anxiety and depression on executive functioning might be elaborated in further research. In conclusion, this study suggests that the association between anxiety and cognition in older people
Am J Geriatr Psychiatry 13:8, August 2005
Bierman et al. is different from the association between depression and cognition. Depression has a linear association with deteriorating cognition. In contrast, mild anxiety appears to be an adaptive response, enhancing cog-
nition when it is most needed. The idea that depression and anxiety may have different roles in adapting to stress is potentially very important, but it needs to be corroborated in future research.
References 1. Graham IE, Rockwood K, Beattie BL, et al: Prevalence and severity of cognitive impairment with and without dementia in an elderly population. Lancet 1997; 349(9068):1793–1796 2. Unverzagt FW, Gao S, Baiyewu 0, et al: Prevalence of cognitive impairment: data from The Indianapolis Study of Health and Aging. Neurology 2001; 57:1655–1662 3. Hanninen T, Koivisto K, Reinikainen KJ, et al: Prevalence of ageing-associated cognitive decline in an elderly population. Age Ageing 1996; 25:201–205 4. Ritchie K, Artero S, Touchon J: Classification criteria for mild cognitive impairment: a population-based validation study. Neurology 2001; 56:37–42 5. Schroder J, Kratz B, Pantel J, et al: Prevalence of mild cognitive impairment in an elderly community sample. J Neural Transm (suppl)1998; 54:51–59 6. Beekman ATF, Bremmer MA, Deeg DJ, et al: Anxiety disorders in later life: a report from The Longitudinal Aging Study, Amsterdam. Int J Geriatr Psychiatry 1998; 13:717–726 7. Koenders ME, Passchier J, Teuns G, et al: Trait-anxiety and achievement motivation are positively correlated with memory performance in patients who visit a geriatric outpatient clinic with amnestic symptoms. Psychol Rep 1993; 73:1227–1231 8. Wetherell JL, Reynolds CA, Gatz M, et al: Anxiety, cognitive performance, and cognitive decline in normal aging. J Gerontol B Psychol Sci Soc Sci 2002; 57:246–255 9. Eysenck MW, Calvo MG: Anxiety and performance: the processing-efficiency theory. Cognition and Emotion 1992; 6:409–434 10. Mendl M: Performing under pressure: stress and cognitive function. Applied Animal Behaviour Science 1999; 65:221–244 11. Yerkes RM, Dodson JD: The relation of strength of stimulus to rapidity of habit-formation. J Comp Neurol Psychol 1908; 18: 459–482 12. Salzman C, Lebowitz BD (eds): Anxiety in the Elderly: Treatment and Research. New York, Springer Publishing, 1991 13. Merikangas KR, Zhang H, Avenevoli S, et al: Longitudinal trajectories of depression and anxiety in a prospective community study: The Zurich Cohort Study. Arch Gen Psychiatry 2003; 60: 993–1000 14. Christensen H, Griffiths K, Mackinnon A, et al: Quantitative review of cognitive deficits in depression and Alzheimer-type dementia. J Int Neuropsychol Soc 1997; 3:631–651 15. Bassuk SS, Berkman LF, Wypij D: Depressive symptomatology and incident cognitive decline in an elderly community sample. Arch Gen Psychiatry 1998; 55:1073–1081 16. Jorm AF: Is depression a risk factor for dementia or cognitive decline? a review. Gerontology 2000; 46:219–227 17. Airaksinen E, Larsson M, Lundberg I, et al: Cognitive functions in depressive disorders: evidence from a population-based study. Psychol Med 2004; 34:83–91 18. Deeg DJH, Knipscheer CPM, Van Tilburg W: Autonomy and Well-Being in the Aging Population: Concepts and Design of The Longitudinal Aging Study, Amsterdam. Bunnik, Netherlands, NIG Trend Studies No.7, Netherlands Institute for Gerontology, 1993 19. Beekman ATF, Deeg DJ, van Tilburg T, et al: Major and minor
Am J Geriatr Psychiatry 13:8, August 2005
depression in later life: a study of prevalence and risk factors. J Affect Disord 1995; 36:65–75 20. Deeg DJH, Westendorp-de Seriere M: Autonomy and Well-Being in the Aging Population, I: Report From The Longitudinal Aging Study, Amsterdam, 1992–1993. Amsterdam, The Netherlands, VU University Press, 1994 21. Knipscheer CPM, Jong-Gierveld J de, Tilburg T van: Living Arrangements and Social Network of Older Adults. Amsterdam, The Netherlands, VU University Press, 2004 22. Tilburg W van, Dijkstra PA, Broese van Grounau MI (eds): The Primary Social Network in the NESTOR Program: Living Arrangements and Social Networks in Older Adults. Amsterdam, The Netherlands, VU, Uitgeverij, 1992 23. Smits CH, de Vries MZ (eds): Procedures and results of the field work, in Autonomy and Well-Being in the Aging Population: Report From The Longitudinal Aging Study, Amsterdam, 1992–1993. Amsterdam, The Netherlands, Vrije Universiteit Press, 1994 24. 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; 12:189–198 25. Rey A: L’Examen Clinique en Psychologie (The Clinical Examination in Psychology). Paris, France, Presse Universitaire de France, 1964 26. Dee1man BG, Brouwer WH, van Zomeren AH, et al: Functiestoornissen na trauma capitis (Cognitive Impairment After Trauma Capitis), in Neuropsychologie in Nederland (Neuropsychology in the Netherlands). Edited by Jennekens-Schinke A, Diamant JJ, Diesfeldt HFA, et al. Deventer, The Netherlands, 1980, pp 253–281 27. Raven JC (ed): Manual for the Coloured Progressive Matrices Test (Revised). Windsor, UK, NFRE-Nelson, 1995 28. Savage RD (ed): Alphabet Coding Task 15. Western, Australia, Murdoch University, 1984 29. Zigmund AS, Snaith RP: The Hospital Anxiety and Depression Scale. Acta Psychiatr Scand 1993; 67:361–370 30. Snaith RP: The Hospital Anxiety And Depression Scale. Health and Quality of Life Outcomes 2003; 1:1–29 31. Radloff LS: The CES–D Scale: a self-report depression scale for research in the general population. Appl Psychol Meas 1997; 3:385–401 32. Beekman A TF, Deeg DJ, Van Limbeek J, et al: Criterion validity of the Center for Epidemiologic Studies Depression Scale (CES–D): results from a community-based sample of older subjects in The Netherlands. Psychol Med 1997; 27:231–235 33. Berkman LF, Berkman CS, Kasl S, et al: Depressive symptoms in relation to physical health and functioning in the elderly. Am J Epidemiol 1986; 124:372–388 34. Kriegsman DM, Penninx BW, van Eijk JT, et al: Self-reports and general-practitioner information on the presence of chronic diseases in community-dwelling elderly: a study on the accuracy of patients’ self-reports and on determinants of inaccuracy. J Clin Epidemiol 1996; 49:1407–1417 35. Netherlands Central Bureau of Statistics: Health Interview Questionnaire. Heerlen, CBS, 1989 36. Garretsen HFL (ed): Probleem Drinkers (Problem Drinkers). Lisse, The Netherlands, Swets & Zeitlinger, 1983
693
Objective: The authors investigated the relationship between anxiety and cognition in older persons, taking account of comorbid depression. Methods: Data were used from the Longitudinal Aging Study Amsterdam (LASA), a large epidemiological study of 3,107 elderly citizens in The Netherlands. Anxiety and depression were measured with the Hospital Anxiety and Depression Scale–Anxiety subscale and the Center for Epidemiologic Studies–Depression Scale. In measuring cognitive performance, general cognitive functioning was measured by means of Mini-Mental State Exam, episodic memory was measured with the Auditory Verbal Learning Test (AVLT), fluid intelligence by using the RAVEN, and information-processing speed by the coding task. Analysis of variance examined the association between anxiety symptoms and cognition in persons with and without depression. Results: Main effects of anxiety symptoms were found for learning and delayed recall of the AVLT. Depression symptoms showed significant main effects on almost all cognitive performance tests. Mild anxiety symptoms were associated with better cognitive performance, whereas severe anxiety symptoms were negatively associated with cognitive functioning. In contrast, depressive symptoms showed a linear association with cognition; more depression was associated with worse cognition. Conclusion: This study suggests that anxiety has a curvilinear relationship with cognition. Depressive symptoms, however, were always negatively associated with cognitive performance. (Am J Geriatr Psychiatry 2005; 13:686–693)
C
ognitive impairment is common in elderly persons, with prevalence rates ranging from 16.8% to 26.6% in the general population of older persons.1–5 Anxiety is also very common in later life. Results found by Beekman et al.6 showed that in the population age 65 or over, the prevalence of anxiety disorders is between 0.7% and 12%. Little research has been done on the association between anxiety and cognitive functioning in elderly
persons. Boenders et al.7 investigated memory performance in relation to anxiety and achievement motivation in a small elderly population who were referred to a geriatric outpatient clinic because of complaints about their memory. No association was found between anxiety and memory performance or severity of dementia. Wetherell et al.8 studied the cross-sectional association between cognitive performance and state-anxiety in cognitively intact elderly
Received July 5, 2004; revised December 17, 2004; accepted December 22, 2004. From the Dept. of Psychiatry, Institute for Research in Extramural Medicine (EMGO Institute) of the Vrije Universiteit Medical Centre, Amsterdam, The Netherlands. Send correspondence and reprint requests to Dr. E.J.M. Bierman, LASAEMGO, VU Medical Centre, Van der Boechorststraat 7,1081 BT, Amsterdam, The Netherlands. e-mail: [email protected] 䉷 2005 American Association for Geriatric Psychiatry
686
Am J Geriatr Psychiatry 13:8, August 2005
Bierman et al. subjects and found that higher state-anxiety was associated with poorer performance on some, but not all, cognition tests used. There are several reasons to assume an association between anxiety and cognitive functioning. Eysenck’s processing efficiency theory9 holds that anxiety interferes with cognitive performance by pree¨mpting some of the processing and storage resources of the working memory system. It posits that anxiety produces worry and other intrusive thoughts, which are verbal and therefore intrude with the phonological loop of working memory. Mendl10 provides a second possible explanatory model for the association between anxiety and cognitive functioning, citing the Yerkes and Dodson law.11 This law states that there exists an inverted U-shaped relationship between arousal and cognitive performance. Cognitive performance is best when an individual is in an optimal stress or arousal state, above or below which performance levels may drop. When investigating anxiety in older persons, comorbidity with depression cannot be disregarded, because it is very common, especially in later life.12 Merikangas et al.13 emphasized the importance of comorbidity, finding that mixed anxiety-anddepression has a greater stability over time than either disorder alone. An association between depressive symptoms and cognitive decline was found in several studies. Comparing depressed subjects with non-depressed control subjects, a metaanalysis of studies found reduced performance on almost all cognitive performance tests.14 According to Bassuk et al.,15 persons who are depressed may have (reversible) cognitive deficits as a consequence of motivational and attentional problems. Jorm16 stated that depression might also be a reaction to cognitive decline. Furthermore, Airiksinen et al.17 showed that cognitive impairment was associated with mixed anxiety-depressive disorders, considering this as one and the same disorder. Given the high correlations between presence of anxiety and depression, one may suppose a similar association between anxiety and cognitive performance as is found for depression and cognitive performance. In this study, we will investigate how anxiety symptoms are associated with cognitive performance in a large epidemiological study among elderly subjects. We will test the “inverted-U” hypothesis and attempt to separate the effect of anxiety symptoms on
Am J Geriatr Psychiatry 13:8, August 2005
cognitive functioning from the effect of depression symptoms on cognitive functioning.
METHODS Sample To investigate the association between anxiety and cognition in older persons, we used data from the Longitudinal Aging Study Amsterdam (LASA18). LASA is a longitudinal study of predictors and consequences of changes in well-being and autonomy in older persons. Details on sampling are described elsewhere19–21 and will briefly be summarized here. The sample was originally drawn for NESTOR-Living Arrangements and Social Network (NESTOR–LSN;22), for which random samples of older residents (age 55– 85) were drawn from population registers in 11 municipalities in three regions of The Netherlands. Ten months later, in 1992, all 3,805 respondents who participated in NESTOR–LSN were approached for LASA, of whom 3,107 (81.7%) took part;23 394 (10.4%) refused to participate because of lack of interest; 134 (3.5 %) were too ill or cognitively impaired to be interviewed; 126 (3.3%) died before being interviewed; and 44 (1.2%) could not be contacted. Attrition was related to age but not to sex. As expected, the older-old potential subjects were more often found to be too ill or cognitively impaired to participate. Additional testing of cognitive functioning was available for 2,615 subjects. This sample consisted of 1,279 men (48.9%) and 1,336 women (51.1%). Participants had an average age of 70.2 (standard deviation [SD]: 8.7) years. Education level was clustered into three categories: low (62.2%), medium (26.0%), and high (11.8%). All interviews were conducted at the respondents’ homes by specially trained and intensively supervised interviewers. Informed consent was obtained from each participant, and the study was approved by the Medical Ethics Committee of the VU Medical Centre. Respondents were interviewed once every 3 years; for this study, interviews from the first LASA measurement were used. Measures General cognitive functioning was measured by means of the Mini-Mental State Exam (MMSE24), a
687
Cognition and Anxiety Versus Depression frequently used screening instrument for measurement of global cognitive functioning. The scale consists of 23 items, and scores range from 0 to 30; higher scores indicate better cognitive functioning. Episodic memory was measured with a modified Dutch version of the Auditory Verbal Learning Test (AVLT25,26). This test consists of 15 words, which have to be learned during five trials. After each trial, the respondent is asked to recall as many words as possible. After a period of 20 minutes, he or she is asked to name the words that were learned before. The number of trials was limited to three because of scarcity of interview time. The total number of words the respondent has learned during the three trials is the Learning Score, which ranges from 0 to 45. The number of words reproduced after 20 minutes is the Delayed Recall Score, ranging from 0 to 15. The percentage of reproduced words in the Delayed Recall condition, divided by the total number of words reproduced in the Learning condition, is called the Retention Score. Higher scores on all three variables indicate better memory functioning. Fluid intelligence, defined as the ability to deal with new information, was measured by means of two subsets of 12 items (A and B) of Raven’s colored progressive matrices.27 Each item consists of a drawing (matrix) of a pattern from which a section is missing. On the bottom of the page, six patterns are printed, one of which fits into the missing section. The respondent has to choose which of the six alternatives fits best. The items increase in difficulty, and so do the two sections. The 24 items result in a total score range of 0 to 24, with higher scores indicating better cognitive performance. Information-processing speed was measured by means of an adjusted version of the Coding Task.28 This is a timed task in which the respondent has to combine as many characters as possible according to a given example. The example shows 15 combinations of two characters in a row of double boxes (the substitution key). The test itself shows rows of double boxes in which only the upper boxes contain characters and the lower boxes are empty. The respondent has to name the missing characters corresponding to the characters in the upper boxes, using the substitution key, as quickly and accurately as possible. The task consists of three identical 1-minute trials. The score on each trial is the number of correctly com-
688
pleted characters. The result of the last trial is included in the present study. Anxiety symptoms were measured with the Hospital Anxiety and Depression Scale–Anxiety subscale (HADS–A29). The anxiety subscale consists of seven items and is self rated by the respondents. Each answer is rated on a 4-point scale, ranging from 0: “rarely or never” to 3: “mostly or always.” Higher scores indicate more anxiety symptoms. Scores ranging from 0 to 7 were considered in the normal range; scores from 8 to 10 as mild anxiety symptoms; and scores 11 or higher as moderate/severe anxiety symptoms.30 We also used a dichotomous score based on the commonly-used cut-off score of 7.29 Depression symptoms were measured by means of the Center for Epidemiologic Studies–Depression Scale (CES–D31,32). The CES–D is a self-report scale consisting of 20 items; it was developed to measure depression symptoms in the community. Each answer is rated on a 4-point scale, ranging from 0: “rarely or never” to 3: “mostly or always.” Higher scores indicate more depression symptoms. We also used a dichotomous score based on the commonlyused cut-off score of 16.33 Confounders Possible confounders that may be associated with cognition and the independent variables include sociodemographics, physical health, alcohol consumption, level of education, and use of benzodiazepines. The sociodemographic variables age, gender, and education level were included as confounders. The level of education was classified into nine groups and divided over three levels; low education (elementary not completed, elementary education, lower vocational education), medium education (general intermediate education, intermediate vocational education, general secondary education), and high education (higher vocational education, college education, and university education). Physical health was reflected by number of chronic diseases. The number of chronic diseases respondents suffered from was assessed by asking them whether they had any of the following: chronic nonspecific lung disease, cardiac disease, peripheral atherosclerosis, stroke, diabetes mellitus, arthritis, malignant neoplasms, or other chronic diseases. The number ranged from 0 to 8. The accuracy of self-reports of
Am J Geriatr Psychiatry 13:8, August 2005
Bierman et al. these diseases, versus general-practitioner information, was shown to be adequate and independent of cognitive impairment.34 Alcohol consumption was assessed with a questionnaire developed by the Dutch central office of statistics (Central Bureau of Statistics)35 and classified according to Garretsen’s Indication of Present Alcohol Use36 into five categories (excessive, severe, moderate, light, and non-drinker). Finally, the use of benzodiazepines was taken into account as possible confounder. Use of prescribed drugs was assessed by asking the respondents which medications they used; responses were compared with information on the drug containers provided by the respondents. Drug use was classified into categories according to the Anatomic Therapeutic Chemical (ATC) classification. For the present study, we assessed number of prescribed drugs. For psychotropic drugs, a separate variable was computed, indicating use of antidepressant drugs and/or hypnotic drugs and/or and anxiolytic drugs (coded as Yes or No). Statistical Analyses For descriptors, elderly respondents were divided in four groups on the basis of the cut-off scores on the anxiety and depression scales: 1) respondents with neither anxiety nor depression symptoms (ANX–/DEP–); 2) respondents with anxiety but no depression symptoms (ANXⳭ/DEP–); 3) respondents with no anxiety but with depression symptoms (ANX–/DEPⳭ); and 4) respondents with both anxiety and depression symptoms (ANXⳭ/DEPⳭ). Analyses of variance (ANOVAs) were conducted to determine the main effects of both anxiety and depression symptoms and look into the interaction effect of mixed anxiety/depression; p values lower than 0.05 were regarded as statistically significant. To further clarify the role of depression symptoms in the association between anxiety and cognitive performance, we hereafter divided the sample into respondents with and without anxiety symptoms. ANOVAs were used to investigate the association between anxiety symptoms and cognitive performance. In the first model, we investigated the unadjusted association between anxiety symptoms and cognition; in the second model, the association between anxiety symptoms and cognition was adjusted for depression symptoms; and, in the final
Am J Geriatr Psychiatry 13:8, August 2005
model, the association was also adjusted for additional confounders. For the question of whether the association between anxiety and cognitive performance is an inverted U-shaped function, we performed an ANOVA with two cut-off points on the HADS–A scale (7 and 10), with 7 as the cut-off point between no-anxiety and the presence of anxiety and 10 being the cut-off point between mild anxiety symptoms and moderate/severe anxiety symptoms.30 To avoid the influence of comorbid depression symptoms, all respondents with depression symptoms (CES–D scores ⱖ16)31 were excluded from analysis. To compare the association between anxiety and cognition with the association between depression and cognition, we conducted analysis of variance on the association between depression and cognition, with the cut-off points of 16 and 24 on the CES–D scale, with 16 being the cut-off point between “no depression symptoms” and “mild depression symptoms” and 24 being the cut-off point between “mild depression symptoms” and “moderate/severe depression symptoms.” In this analysis, respondents with anxiety symptoms (HADS–A scores ⱖ7) were excluded from analysis. Also, we tested the curvilinearity of the association between anxiety symptoms (adjusted for depression symptoms) and cognition and between depression symptoms (adjusted for anxiety symptoms) and cognition by adding a quadratic term in the regression analyses. Finally, we conducted linear analyses to ensure that the association between depression symptoms and cognitive performance was linear.
RESULTS Table 1 shows the characteristics of these elderly respondents, divided in four groups: 1) with neither anxiety nor depression symptoms (ANX–/DEP–); 2) with anxiety but no depression symptoms (ANXⳭ/DEP–); 3) with no anxiety but only depression symptoms (ANX–/DEPⳭ); and 4) with both anxiety and depressive symptoms (ANXⳭ/DEPⳭ). Analysis of variance shows significant main effects of anxiety symptoms on Learning and Delayed Recall of the AVLT. Depression symptoms show significant effects on all cognitive performance tests on the AVLT
689
Cognition and Anxiety Versus Depression except Retention (Table 2). There were no significant interaction effects (data not shown). Comparison of performance on cognitive tests among those with and without anxiety symptoms (Table 3) shows no differences. After adjustment for depression symptoms, a significant difference in the scores on all cognitive tests but AVLT Retention was found between Anxiety and No Anxiety symptoms to the advantage of the respondents with no anxiety symptoms. After additional adjustment, this significant difference between the group with and without TABLE 1.
anxiety disappeared for the Delayed Recall score on the AVLT. For investigating the possible inverted U-shaped relationship between anxiety and cognitive performance, we divided the respondents into three groups by their scores on anxiety and depression symptoms. Figure 1 shows the relationship between the level of anxiety symptoms and depression symptoms and performance on cognitive tests. In all cognitive tests, we found a curvilinear relationship with level of anxiety symptoms. This was in
Characteristics of the Study Sample
N Age, years Gender, N (%) Men Women Education level, N (%) Low Medium High Use of benzodiazepines, N (%) Alcohol consumption, N (%) No Light Moderate High Excessive Number of chronic diseases HADS–A CES–D MMSE Coding task Raven Progressive Matrices AVLT Learning Delayed recall Retention
Anxiety– / Depression–
Anxietyⴐ / Depression–
Anxiety– / Depressionⴐ
Anxietyⴐ / Depressionⴐ
2,108 69.76 (8.6)
94 69.07 (9.0)
136 73.44 (8.6)
185 70.65 (8.7)
1,085 (51.5) 1,023 (48.5)
34 (36.2) 60 (63.8)
52 (38.2) 84 (61.8)
63 (34.1) 122 (65.9)
1271 (60.3) 571 (27.1) 264 (12.5) 219 (10.4)
59 (62.8) 26 (27.7) 9 (9.6) 18 (19.1)
92 (67.6) 26 (19.1) 18 (13.2) 10 (20.6)
128 (69.2) 44 (23.8) 13 (7.0) 26 (30.8)
420 (19.9) 1113 (52.8) 464 (22.0) 77 (3.7) 21 (1.0) 1.27 (1.2) 1.47 (1.7) 5.10 (4.1) 27.32 (2.4) 25.76 (7.8) 18.12 (4.0)
22 (23.4) 52 (55.3) 18 (19.1) 2 (2.1) — 1.71 (1.2) 8.93 (2.7) 10.00 (4.1) 27.68 (2.2) 27.64 (7.3) 18.44 (3.8)
33 (24.3) 72 (52.9) 27 (19.9) 3 (2.2) 1 (0.7) 1.83 (1.3) 3.60 (1.8) 20.19 (4.4) 26.56 (2.7) 23.57 (7.5) 16.61 (4.0)
59 (31.9) 90 (48.6) 25 (13.5) 10 (5.4) 1 (0.5) 2.04 (1.29) 10.33 (3.2) 25.02 (8.0) 26.86 (2.5) 23.53 (6.8) 17.03 (4.3)
18.66 (6.2) 5.08 (2.8) 61.66 (25.9)
19.93 (6.0) 5.57 (2.6) 65.40 (22.9)
17.24 (6.2) 4.53 (2.7) 57.93 (26.4)
17.97 (6.4) 4.89 (2.7) 60.99 (26.7)
Note: Values are mean (standard deviation), unless otherwise indicated. HADS–A: Hospital Anxiety and Depression Scale–Anxiety subscale; CES–D: Center for Epidemiologic Studies–Depression Scale; MMSE: MiniMental State Exam; AVLT: Auditory Verbal Learning Test.
TABLE 2.
Main Effects of Anxiety and Depression on Cognitive Performance Main Effect: Anxiety
MMSE Coding task Raven Progressive Matrices AVLT Learning Delayed recall Retention
Main Effect: Depression
F[1, 2433]⳱3.45 F[1, 2433]⳱2.57 F[1, 2433]⳱1.58
p⳱0.06 p⳱0.11 p⳱0.21
F[1, 2433]⳱15.14 F[1, 2433]⳱21.91 F[1, 2433]⳱21.05
p ⬍0.01 p ⬍0.01 p ⬍0.01
F[1, 2433]⳱5.02 F[1, 2433]⳱4.68 F[1, 2433]⳱3.55
p⳱0.03 p⳱0.03 p⳱0.06
F[1, 2433]⳱7.47 F[1, 2433]⳱4.82 F[1, 2433]⳱1.93
p⳱0.01 p⳱0.03 p⳱0.17
Note: MMSE: Mini-Mental State Exam; AVLT: Auditory Verbal Learning Test.
690
Am J Geriatr Psychiatry 13:8, August 2005
Bierman et al. TABLE 3.
Analyses of Variance for Cognitive Performance in Anxiety and No-Anxiety Symptom Groups
MMSE Coding task Raven Progressive Matrices AVLT Learning Delayed recall Retention
F
p
Fa
pa
Fb
pb
F[1, 2523]⳱0.79 F[1, 2476]⳱2.26 F[1, 2467]⳱3.50
p⳱0.37 p⳱0.13 p⳱0.06
F[1, 2511]⳱12.11 F[1, 2465]⳱10.41 F[1, 2457]⳱9.21
p ⬍0.01 p ⬍0.01 p ⬍0.01
F[1, 2486]⳱6.86 F[1, 2440]⳱4.18 F[1, 2436]⳱3.76
p⳱0.01 p⳱0.04 p⳱0.05
F[1, 2532]⳱0.03 F[1, 2532]⳱0.22 F[1, 2532]⳱0.52
p⳱0.87 p⳱0.64 p⳱0.47
F[1, 2520]⳱11.38 F[1, 2520]⳱8.36 F[1, 2520]⳱3.65
p ⬍0.01 p ⬍0.01 p⳱0.06
F[1, 2495]⳱4.38 F[1, 2495]⳱2.61 F[1, 2495]⳱1.02
p⳱0.04 p⳱0.11 p⳱0.31
Note: MMSE: Mini-Mental State Exam; AVLT: Auditory Verbal Learning Test. a Adjusted for depression symptoms. b Adjusted for depression symptoms and other confounders (age, gender, education level, alcohol use, benzodiazepine use, and number of chronic diseases.
FIGURE 1.
Different Effects of Anxiety and Depression on Cognition in Elderly Subjects
28.0 AVLT learning/memory
21.5
MMSE
27.5 27.0 26.5 26.0 25.5
Anxiety Depression
18.5 17.5
Anxiety Depression
6.5 AVLT delayed recall
28 Coding Task
19.5
16.5
29
27 26 25 24 23
6.0 5.5 5.0 4.5
22
4.0
19.5
67.5
AVLT retention
19.0 RAVEN
20.5
18.5 18.0 17.5 17.0 16.5
65.5 63.5 61.5 59.5 57.5
No
Mild
Severe
No
Mild
Severe
Note: No anxiety/depression: HADS–A ⱕ7 and CES–D ⱕ16; Mild: 7 ⬎ HADS–A ⱕ 10 and 16 ⬎ CES–D ⱕ 24; Severe: HADS–A ⬎ 10 and CES–D ⬎ 24. MMSE: Mini-Mental State Exam;24 Coding task: Alphabet Coding Task 15;28 Raven: Raven Progressive Matrices test;27 AVLT: Auditory Verbal Learning Test.25,26
Am J Geriatr Psychiatry 13:8, August 2005
691
Cognition and Anxiety Versus Depression contrast with the declining scores on cognitive performance tests with increasing scores on the depression scale. The curvilinear relationship only reached the level of significance in Learning and Memory from the AVLT (F[2, 2199]⳱3.20; p⳱0.04). The other domains were not significant (MMSE: F[2, 2192]⳱0.46; p⳱0.63; Coding Task: F[2, 2155]⳱2.21; p⳱0.11; RAVEN: F[2, 2155]⳱1.85; p⳱0.16; AVLT Delayed Recall: F[2, 2199]⳱2.74; p⳱0.07; AVLT Retention: F[2, 2199]⳱ 0.99; p⳱0.37). Finally, the curvilinear relationship was tested with a quadratic term in the regression analyses. With anxiety symptoms, we found a quadratic relationship in all cognitive performance tests, reaching the level of significance for the MMSE, the Coding Task, and Learning and Memory from the Auditory Verbal Learning Test. For depression symptoms, we found no significant curvilinear relationship in all cognitive performance tests. Actually, linear analyses showed that the association between depression symptoms and cognitive performance was linear in all cognitive performance tests.
DISCUSSION The present study sought to investigate whether anxiety and depression have different effects on cognition in a large sample of community-living elderly persons. We found that mild anxiety symptoms seem to be beneficial, whereas severe anxiety negatively influences cognitive performance. Depression symptoms always negatively influence cognitive functioning. The association between anxiety symptoms and cognitive performance seems to be a curvilinear relationship. These findings support the YerkesDodson law, which states that optimum performance should be seen at levels of moderate arousal.11 These findings are strongest in the AVLT; it may be that this test is more sensitive to mental effort than the other cognitive measures. Our results also support Eysenck’s processingefficiency theory, in which he states that anxiety interferes with cognitive performance by pree¨mpting some of the processing and storage resources of the working memory system. Moderate levels of anxiety can have an arousing function that services behavioral performance up to a point, the point of disability
692
coming when the anxiety symptoms become so severe as to overwhelm information-processing resources. Moderate levels of depression would not be expected to have the initially beneficial arousing qualities, and, in addition to drawing attention (processing) recourses, depression may hinder motivation. We found that the relationship between anxiety symptoms and cognitive performance was strongly influenced by depressive symptoms. Correcting for depressive symptoms changed the effect on cognitive performance from negative to positive in most cognitive-performance tests, indicating a beneficial effect on cognition, especially for mild anxiety symptoms. The findings should be placed in the context of the strengths and limitations of this study. Strengths of the current study are that it is community-based, including large numbers of elderly subjects, with data available on relevant confounders and testing in four domains of cognition. However there are also limitations. First, selective non-response of the most frail participants is an inevitable problem in communitybased research among elderly subjects. This might have underestimated the effects of both anxiety and depression on cognitive functioning, because the subjects with the most severe anxiety and depression symptoms frequently drop out of studies like this. The second limitation is the cross-sectional nature of his study, in which it is impossible to determine causality because it is not possible to study whether the anxiety and depression symptoms proceed the cognitive decline. Longitudinal research is necessary to clarify this issue. Longitudinal research also makes it possible to investigate the influence of the duration of the anxiety symptoms on cognitive performance. Furthermore, our data are mainly based on selfreport measurements. No objective clinical judgment or diagnoses have been made. Finally, not all cognitive domains were included in this study. To investigate whether the curvilinear pattern was due to resource effects, measuring crystallized cognitive abilities would have been helpful. One might expect fewer differences among the three anxiety groups on a crystallized measure than on fluid measures, as we used in our study. The effects of anxiety and depression on executive functioning might be elaborated in further research. In conclusion, this study suggests that the association between anxiety and cognition in older people
Am J Geriatr Psychiatry 13:8, August 2005
Bierman et al. is different from the association between depression and cognition. Depression has a linear association with deteriorating cognition. In contrast, mild anxiety appears to be an adaptive response, enhancing cog-
nition when it is most needed. The idea that depression and anxiety may have different roles in adapting to stress is potentially very important, but it needs to be corroborated in future research.
References 1. Graham IE, Rockwood K, Beattie BL, et al: Prevalence and severity of cognitive impairment with and without dementia in an elderly population. Lancet 1997; 349(9068):1793–1796 2. Unverzagt FW, Gao S, Baiyewu 0, et al: Prevalence of cognitive impairment: data from The Indianapolis Study of Health and Aging. Neurology 2001; 57:1655–1662 3. Hanninen T, Koivisto K, Reinikainen KJ, et al: Prevalence of ageing-associated cognitive decline in an elderly population. Age Ageing 1996; 25:201–205 4. Ritchie K, Artero S, Touchon J: Classification criteria for mild cognitive impairment: a population-based validation study. Neurology 2001; 56:37–42 5. Schroder J, Kratz B, Pantel J, et al: Prevalence of mild cognitive impairment in an elderly community sample. J Neural Transm (suppl)1998; 54:51–59 6. Beekman ATF, Bremmer MA, Deeg DJ, et al: Anxiety disorders in later life: a report from The Longitudinal Aging Study, Amsterdam. Int J Geriatr Psychiatry 1998; 13:717–726 7. Koenders ME, Passchier J, Teuns G, et al: Trait-anxiety and achievement motivation are positively correlated with memory performance in patients who visit a geriatric outpatient clinic with amnestic symptoms. Psychol Rep 1993; 73:1227–1231 8. Wetherell JL, Reynolds CA, Gatz M, et al: Anxiety, cognitive performance, and cognitive decline in normal aging. J Gerontol B Psychol Sci Soc Sci 2002; 57:246–255 9. Eysenck MW, Calvo MG: Anxiety and performance: the processing-efficiency theory. Cognition and Emotion 1992; 6:409–434 10. Mendl M: Performing under pressure: stress and cognitive function. Applied Animal Behaviour Science 1999; 65:221–244 11. Yerkes RM, Dodson JD: The relation of strength of stimulus to rapidity of habit-formation. J Comp Neurol Psychol 1908; 18: 459–482 12. Salzman C, Lebowitz BD (eds): Anxiety in the Elderly: Treatment and Research. New York, Springer Publishing, 1991 13. Merikangas KR, Zhang H, Avenevoli S, et al: Longitudinal trajectories of depression and anxiety in a prospective community study: The Zurich Cohort Study. Arch Gen Psychiatry 2003; 60: 993–1000 14. Christensen H, Griffiths K, Mackinnon A, et al: Quantitative review of cognitive deficits in depression and Alzheimer-type dementia. J Int Neuropsychol Soc 1997; 3:631–651 15. Bassuk SS, Berkman LF, Wypij D: Depressive symptomatology and incident cognitive decline in an elderly community sample. Arch Gen Psychiatry 1998; 55:1073–1081 16. Jorm AF: Is depression a risk factor for dementia or cognitive decline? a review. Gerontology 2000; 46:219–227 17. Airaksinen E, Larsson M, Lundberg I, et al: Cognitive functions in depressive disorders: evidence from a population-based study. Psychol Med 2004; 34:83–91 18. Deeg DJH, Knipscheer CPM, Van Tilburg W: Autonomy and Well-Being in the Aging Population: Concepts and Design of The Longitudinal Aging Study, Amsterdam. Bunnik, Netherlands, NIG Trend Studies No.7, Netherlands Institute for Gerontology, 1993 19. Beekman ATF, Deeg DJ, van Tilburg T, et al: Major and minor
Am J Geriatr Psychiatry 13:8, August 2005
depression in later life: a study of prevalence and risk factors. J Affect Disord 1995; 36:65–75 20. Deeg DJH, Westendorp-de Seriere M: Autonomy and Well-Being in the Aging Population, I: Report From The Longitudinal Aging Study, Amsterdam, 1992–1993. Amsterdam, The Netherlands, VU University Press, 1994 21. Knipscheer CPM, Jong-Gierveld J de, Tilburg T van: Living Arrangements and Social Network of Older Adults. Amsterdam, The Netherlands, VU University Press, 2004 22. Tilburg W van, Dijkstra PA, Broese van Grounau MI (eds): The Primary Social Network in the NESTOR Program: Living Arrangements and Social Networks in Older Adults. Amsterdam, The Netherlands, VU, Uitgeverij, 1992 23. Smits CH, de Vries MZ (eds): Procedures and results of the field work, in Autonomy and Well-Being in the Aging Population: Report From The Longitudinal Aging Study, Amsterdam, 1992–1993. Amsterdam, The Netherlands, Vrije Universiteit Press, 1994 24. 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; 12:189–198 25. Rey A: L’Examen Clinique en Psychologie (The Clinical Examination in Psychology). Paris, France, Presse Universitaire de France, 1964 26. Dee1man BG, Brouwer WH, van Zomeren AH, et al: Functiestoornissen na trauma capitis (Cognitive Impairment After Trauma Capitis), in Neuropsychologie in Nederland (Neuropsychology in the Netherlands). Edited by Jennekens-Schinke A, Diamant JJ, Diesfeldt HFA, et al. Deventer, The Netherlands, 1980, pp 253–281 27. Raven JC (ed): Manual for the Coloured Progressive Matrices Test (Revised). Windsor, UK, NFRE-Nelson, 1995 28. Savage RD (ed): Alphabet Coding Task 15. Western, Australia, Murdoch University, 1984 29. Zigmund AS, Snaith RP: The Hospital Anxiety and Depression Scale. Acta Psychiatr Scand 1993; 67:361–370 30. Snaith RP: The Hospital Anxiety And Depression Scale. Health and Quality of Life Outcomes 2003; 1:1–29 31. Radloff LS: The CES–D Scale: a self-report depression scale for research in the general population. Appl Psychol Meas 1997; 3:385–401 32. Beekman A TF, Deeg DJ, Van Limbeek J, et al: Criterion validity of the Center for Epidemiologic Studies Depression Scale (CES–D): results from a community-based sample of older subjects in The Netherlands. Psychol Med 1997; 27:231–235 33. Berkman LF, Berkman CS, Kasl S, et al: Depressive symptoms in relation to physical health and functioning in the elderly. Am J Epidemiol 1986; 124:372–388 34. Kriegsman DM, Penninx BW, van Eijk JT, et al: Self-reports and general-practitioner information on the presence of chronic diseases in community-dwelling elderly: a study on the accuracy of patients’ self-reports and on determinants of inaccuracy. J Clin Epidemiol 1996; 49:1407–1417 35. Netherlands Central Bureau of Statistics: Health Interview Questionnaire. Heerlen, CBS, 1989 36. Garretsen HFL (ed): Probleem Drinkers (Problem Drinkers). Lisse, The Netherlands, Swets & Zeitlinger, 1983
693