Alzheimer’s & Dementia - (2016) 1-10

55
1 56
2 Featured Article 57
3 58
4 Mild cognitive impairment and risk of depression and anxiety: 59
5 60
6 a population-based study 61
7 62
8 63
9 Q4 Saira Saeed Mirzaa, M. Arfan Ikrama,b,c, Daniel Bosa,c, Raluca Mihaescua,d, Albert Hofmana, 64
10 Henning Tiemeiera,d,e,* 65
11 a
66
Department of Epidemiology, Erasmus Medical Center, Rotterdam, The Netherlands 67
12 b
Department of Neurology, Erasmus Medical Center, Rotterdam, The Netherlands
13 c 68
Department of Radiology, Erasmus Medical Center, Rotterdam, The Netherlands
14 d
Department of Psychiatry, Erasmus Medical Center, Rotterdam, The Netherlands
69
15 e
Department of Child and Adolescent Psychiatry, Erasmus Medical Center, Rotterdam, The Netherlands 70
16 71
17 72
18 73
Abstract Introduction: Many people with mild cognitive impairment (MCI) suffer from concomitant depres-
19 74
sion or anxiety. Whether MCI increases the risk of future depression or anxiety is unknown.
20 75
Methods: In the Rotterdam Study, cross-sectional (n 5 4168) and longitudinal associations
21 76
(n 5 2967) of MCI with Diagnostic and Statistical Manual of Mental Disorders—depressive and
22 77
anxiety disorders—were assessed (2002–2005 to 2009–2011).
23 78
Results: At baseline, 413 persons had MCI; 125 (22 MCI and 103 non-MCI) had a depressive dis-
24 79
order and 330 had an anxiety disorder (46 MCI and 284 non-MCI). In longitudinal depression anal-
25 80
ysis, of the 212 persons with prevalent MCI, 6 (2.8%) developed depression compared with 29 (1%)
26 81
in the nonexposed group. In longitudinal anxiety analysis, 11 (7.3%) of the 151 with prevalent MCI
27 82
developed anxiety, compared with 75 (3.4%) in nonexposed group. Persons with MCI had more
28 83
depressive and anxiety disorders and also a higher risk of developing depressive disorder, odds ratio
29 84
(OR) 3.13 (95% confidence interval [CI]: 1.26, 7.77), and anxiety disorder, OR 2.59 (95% CI: 1.31,
30 85
5.12).
31 86
Discussion: MCI is a risk factor for dementia and for depressive and anxiety disorders, suggesting
32 87
common pathological pathways for cognitive and psychiatric outcomes.
33 88
Ó 2016 The Alzheimer’s Association. Published by Elsevier Inc. All rights reserved.
34 89
35 90
36 Keywords: Mild cognitive impairment; Depression; Anxiety; Epidemiology; Longitudinal; Population based 91
37 92
38 93
39 94
1. Introduction clinical construct in which meaningful interventions are
40 95
possible [2]. 96
41 Dementia poses a high burden on society and health care,
42 Another important manifestation thought to be a part of 97
both in terms of suffering for patients and care givers and
43 dementia prodrome is the occurrence of affective disorders, 98
financial costs [1]. Because brain pathology is thought to
44 namely, depression and anxiety [3]. Depression and anxiety 99
accumulate for years before the onset of dementia, much
45 were shown to be highly prevalent in MCI [4,5]. However, 100
research has been dedicated to study this preclinical phase 101
46 most available literature built up on this association
47 of dementia. In this context, mild cognitive impairment 102
focused on the prognostic role of affective disorders in
48 (MCI) has been conceptualized as a transitional stage be- 103
MCI. Studies showed that MCI with comorbid affective
49 tween normal cognition and dementia and serves as a symptoms has an accelerated progression to dementia [6,7]. 104
50 105
There is also ample evidence suggesting that depression
51 106
in late life is associated with a 2-fold increased risk of de-
52 107
*Corresponding author. Tel.: 131 10 7043475; Fax: 131 10 7044657. mentia [8,9]. Although anxiety has not been associated 108
53
54 E-mail address: h.tiemeier@erasmusmc.nl with a higher risk of dementia [10], it has shown to be 109
http://dx.doi.org/10.1016/j.jalz.2016.06.2361
1552-5260/Ó 2016 The Alzheimer’s Association. Published by Elsevier Inc. All rights reserved.

FLA 5.4.0 DTD
JALZ2256_proof
9 August 2016
3:48 pm
ce

 2 S.S. Mirza et al. / Alzheimer’s & Dementia - (2016) 1-10

110 associated with cognitive decline [11]. Extending these find- administered correctly. Consequently, MCI was validly as- 171
111 ings, some recent studies have also shown that depression is sessed in 4198 participants. 172
112 a risk factor for MCI [12–14]. Studies investigating Between 2002 and 2005, depression and anxiety disor- 173
113 individual neuropsychiatric symptoms in relation to ders were assessed in the Rotterdam Study. Of the 4198 par- 174
114 175
incident MCI found that anxiety increases the risk of MCI ticipants with available MCI data, depression data were
115 176
[12]. These studies imply that depression and anxiety pre- available for 4168 participants, whereas anxiety data were
116 177
117 cede MCI in the chronological order of events. However, available for 4060 participants. At baseline, 125 participants 178
118 given that both MCI and affective symptoms are considered fulfilled the criteria for depressive disorders, whereas 330 179
119 to manifest during the preclinical stage of dementia, it is also participants met the criteria for anxiety disorders (Fig. 1). 180
120 not unlikely that MCI precedes depression and anxiety. Between 2009 and 2012, depressive disorders were reas- 181
121 However, the association of MCI in relation to risk of sessed in 3117 participants of the 3370 participants attending 182
122 depression or anxiety has never been investigated. There- the examination round (798 participants died during follow- 183
123 fore, we investigated the cross-sectional and longitudinal as- up, 4168 2 798 5 3370). After excluding 125 depression 184
124 sociations of MCI with depression and anxiety in a cases at baseline, and 25 persons who were diagnosed of 185
125 population-based cohort of older adults. incident dementia during the study period, depressive disor- 186
126 187
der data were available for 2967 persons for the analyses of
127 188
MCI and incident depressive disorders (response
128 2. Methods 189
129 rate 5 92%) (Fig. 1). 190
130 2.1. Setting Between 2009 and 2012, anxiety disorders were reas- 191
131 sessed in 2714 participants of the 3293 participants attending 192
132 This study was part of the Rotterdam Study, a population- the examination round (767 participants died during follow- 193
133 based cohort ongoing since 1990 in Ommoord, a district of up, 4060 2 767 5 3293). We excluded 330 anxiety cases at 194
134 Rotterdam [15]. In 1990, 7983 participants aged 55 years baseline, and 9 persons who were diagnosed with incident 195
135 or older were enrolled. In 2000, the original cohort was dementia during the study period. Therefore, 2375 partici- 196
136 expanded by additionally enrolling 3011 participants who pants were available for the analyses of MCI and incident 197
137 had become 55 years of age or moved to the district since anxiety disorders (response rate 5 82%) (Fig. 1). 198
138 the start of the study. Follow-up examinations including 199
139 home interviews and physical examinations at a research 200
140 2.3. Assessment of MCI 201
center take place every 3–4 years. The Rotterdam Study is
141 202
approved by the Medical Ethics Committee of the Erasmus MCI was assessed using the following criteria: (1) pres- 203
142
143 MC and by the Ministry of Health, Welfare, and Sport of The ence of subjective memory complaints, (2) presence of 204
144 Netherlands, implementing the “Wet Bevolkingsonderzoek: objective cognitive impairment, and (3) absence of dementia 205
145 ERGO (Population Studies Act: Rotterdam Study).” A writ- [16]. 206
146 ten informed consent was obtained from all the participants. Subjective memory complaints were assessed by inter- 207
147 view, which included three questions on memory (difficulty 208
148 remembering, forgetting what one had planned to, and diffi- 209
149 2.2. Study sample 210
culty finding words) and three questions on everyday func-
150 211
Between 2002 and 2005, the original cohort and the tioning (difficulty managing financing, problems using a
151 212
expanded cohort were reexamined, and an extensive neuro- telephone, and difficulty getting dressed). Persons answering
152 213
153 psychological test battery was implemented. Given that “yes” to at least one of these questions were scored positive 214
154 extensive neuropsychological testing is required to deter- on subjective memory complaints. Objective cognitive 215
155 mine MCI, 2002–2005 was set as baseline for MCI screening impairment was assessed using a cognitive test battery that 216
156 in our study. Of the 6061 study participants who underwent comprised letter-digit substitution task, Stroop test (reading, 217
157 examinations between 2002 and 2005, 192 participants were color naming, and interference subtasks) [17], verbal fluency 218
158 excluded because they were demented, 67 because they were test, and 15-word verbal learning test based on Rey’s recall of 219
159 not sufficiently screened for dementia, and another 250 par- words [18]. To obtain more robust measures, we calculated 220
160 ticipants because they did not answer the questions different compound scores for various cognitive domains 221
161 regarding subjective cognitive complaints. An additional including memory function, information processing speed, 222
162 1354 participants were excluded because they missed one and executive function. Briefly, compound score for memory 223
163 224
or more cognitive test scores or had unreliable test scores. was calculated as the mean Z score for the immediate and de-
164 225
165 Cognitive test results were considered unreliable if there layed recall of the 15-word verbal learning test. For informa- 226
166 had been any technical problems conducting the tests, if tion processing speed, average Z scores for the Stroop 227
167 there was refusal or insufficient motivation from the partic- reading and Stroop color-naming subtasks and the letter- 228
168 ipant to perform the tests, if there were any physical limita- digit substitution task were used. For calculating compound 229
169 tions to perform the tests, if there was any deviation from the score of executive function, Z scores of Stroop interference 230
170 instructions given to perform the tests, or if the tests were not subtask, the letter-digit substitution task, and the verbal 231

FLA 5.4.0 DTD
JALZ2256_proof
9 August 2016
3:48 pm
ce

 S.S. Mirza et al. / Alzheimer’s & Dementia - (2016) 1-10 3

232 293
233 294
234 295
235 296
236 297
237 298
238 299
239 300
240 301
241 302
242 303
243 304
244 305
245 306
246 307
247 308
248 309
249 310
250 311
251 312
252 313
253 314
254 315
web 4C=FPO

255 316
256 317
257 318
258 319
259 Fig. 1. Flow diagram showing the study sample for both cross-sectional and longitudinal analyses of mild cognitive impairment with depression and anxiety. 320
260 321
261 322
262 fluency task were used. Persons were classified as cognitively 2.5. Assessment of anxiety disorders 323
263 impaired if they scored ,1.5 standard deviations of the age- 324
264 and education-adjusted means of the study population. During the home interview, an adapted version of the Mu- 325
265 Subsequently, we subclassified MCI into amnestic and non- nich version of the Composite International Diagnostic 326
266 amnestic MCI. Amnestic MCI was defined as persons with MCI Interview (M-CIDI) [22] was administered to all partici- 327
267 pants, to assess 1-year prevalence of the following anxiety 328
who had an impaired test score on memory irrespective of other
268 disorders according to the DSM-IV-TR (Text Revision) 329
domains. Non-amnestic MCI was defined as persons with MCI
269 criteria [23]: generalized anxiety disorder, panic disorder, 330
270 having normal memory function but an impaired test score on 331
executive function or information processing speed. agoraphobia, social phobia, and specific phobia as described
271 previously [24]. Participants were classified as positive for 332
272 333
anxiety disorders if they had at least one of the previously
273 334
mentioned anxiety disorders. The M-CIDI was specifically
274 2.4. Assessment of depressive disorders 335
275 designed to obtain DSM-IV diagnoses of mental disorders, 336
276 Depressive disorders were diagnosed during home inter- and the test-retest reliability for the anxiety disorders is satis- 337
277 view [19]. Participants were screened for symptoms of factory [22]. 338
278 depression with the Center for Epidemiological Studies— The clinicians conducting the SCAN (depression) inter- 339
279 Depression (CES-D) scale. Screen-positive persons (CES-D views and the research assistants conducting the CIDI (anx- 340
280 score
16) were invited for a semistructured clinical inter- iety) interview were blinded to the cognitive status of the 341
281 view with the Schedules for Clinical Assessment of Neuro- participants. 342
282 psychiatry (SCAN) [20]. This interview was conducted by a 343
283 trained clinician at the participant’s home 1 week to 2 months 344
284 2.6. Other assessments 345
(median time interval: 3 weeks) after the screening procedure
285 346
and the anxiety interview done simultaneously. We were able
286 347
287 to use the SCAN in this population-based setting because 2.6.1. Assessment of dementia for exclusion during the study 348
288 depression can be screened for with high sensitivity [21]. period 349
289 With a computerized Diagnostic and Statistical Manual of Participants were first screened for dementia at baseline 350
290 Mental Disorders (Fourth Edition) (DSM-IV)–based diag- and follow-up examinations using a three-step protocol 351
291 nostic algorithm, major depression, minor depression, and [25]. Screening was done using the Mini-Mental State Ex- 352
292 dysthymia during the past month were diagnosed. amination (MMSE) [26] and the Geriatric Mental Schedule 353

FLA 5.4.0 DTD
JALZ2256_proof
9 August 2016
3:48 pm
ce

 4 S.S. Mirza et al. / Alzheimer’s & Dementia - (2016) 1-10

354 (GMS) organic level [27]. Second, screen positives (MMSE line were excluded for the depression analyses, and simi- 415
355 ,26 or GMS organic level .0) subsequently underwent an larly, cases of anxiety at baseline were excluded before 416
356 examination and informant interview with the Cambridge performing the anxiety analyses. We also excluded persons 417
357 Examination for Mental Disorders in the Elderly [28]. Par- with incident dementia diagnosed during the study period. 418
358 419
ticipants who were suspected of having dementia, if neces- Furthermore, we examined the longitudinal associations of
359 420
sary, had further neuropsychological testing. Additionally, subtypes of MCI with incident depressive and anxiety disor-
360 421
361 the total cohort was continuously monitored for dementia ders. 422
362 through computerized linkage between the study database For all analyses, two models were fitted. In both cross- 423
363 and digitized medical records from general practitioners sectional and longitudinal analyses, model 1 was adjusted 424
364 and the Regional Institute for Outpatient Mental Health for age and sex only. Model 2 was additionally adjusted 425
365 Q1 Care. Third, a consensus panel led by a neurologist decided for educational level, BMI, smoking status, serum total 426
366 on the final diagnosis in accordance with the standards using cholesterol, HDL cholesterol, hypertension, diabetes melli- 427
367 the DSM-III-R (Third Edition Revised) criteria for dementia tus 2, myocardial infarction, stroke, and cohort. The longi- 428
368 and the NINCDS-ADRDA for Alzheimer’s disease [29]. tudinal analyses were additionally adjusted for time 429
369 interval between the two assessments of depression and 430
370 431
2.6.2. Covariates anxiety.
371 432
Age, sex, education, body mass index (BMI), smoking As a first sensitivity analysis, we also investigated the
372 433
373 status, total serum cholesterol, high-density lipoprotein cross-sectional and longitudinal associations of MCI with 434
374 (HDL) cholesterol, prevalent disease including hypertension continuously assessed depressive symptoms. In a second 435
375 and diabetes mellitus type 2, and cardiovascular events sensitivity analysis, we adjusted the longitudinal analysis 436
376 including myocardial infarction (MI) and stroke were of depressive disorders for baseline depressive symptoms 437
377 considered potential confounders for the association of as assessed by CES-D score. Third, to investigate if the as- 438
378 MCI with depression and anxiety. Education, vascular fac- sociations of MCI with depression and anxiety were largely 439
379 tors, and cardiovascular events are implicated as risk factors explained by the presence of subjective memory com- 440
380 for MCI and are also associated with depression and anxiety plaints, we also tested the cross-sectional and longitudinal 441
381 [16,30–33]. associations of objective cognitive impairment (no subjec- 442
382 443
Education was assessed during the home interview, and tive memory complaints and thus not formally defined as
383 444
participants were classified into three groups: low educa- MCI) with depression and anxiety. Fourth, to better explore
384 445
385 tional level (primary or unfinished secondary or lower voca- if the observed associations are attributed to the construct 446
386 tional training), intermediate level (secondary or of MCI, or can be explained by cognitive impairment, we 447
387 intermediate or higher vocational training), and high level adjusted both cross-sectional and longitudinal analyses 448
388 (completed college or university). BMI was calculated as for the MMSE score as a global measure of cognitive func- 449
389 weight in kilograms/height in square meters. We defined tion. 450
390 smokers as never, former, or current. Serum total cholesterol Finally, we explored differences in depression and anx- 451
391 and HDL cholesterol (millimoles per liter) were measured iety profiles between participants with MCI who pro- 452
392 by an automated enzymatic procedure (Boehringer Man- gressed to dementia and those who did not change their 453
393 nheim System). Hypertension was defined as a blood pres- MCI status. 454
394 455
sure
140/90 mmHg or use of blood pressure–lowering The following covariates had missing values, which
395 456
medication, prescribed for the indication of hypertension were dealt with using multiple imputations and chained
396 457
397 [34]. Diabetes mellitus was defined as a fasting serum equations (ICE command Stata): education (1.5%), BMI 458
398 glucose level
7.0 mmol/L, nonfasting serum glucose level (1.5%), total cholesterol (1.7%), HDL cholesterol (1.7%), 459
399
11.1 mmol/L, or use of antidiabetic medication [35]. At and MI (0.6%). 460
400 study entry, history of MI and stroke was assessed using For all tests, a significance level of 0.05 was used. All an- 461
401 home interviews and confirmed by reviewing medical re- alyses were performed using Stata Software, Version 13 462
402 cords [36,37]. (Stata Corp, College Station, TX, USA). 463
403 464
404 465
405 2.7. Statistical analyses 3. Results 466
406 We first assessed the cross-sectional associations of MCI Characteristics of participants included in the analyses of 467
407 468
with depressive disorders and anxiety disorders using logis- depressive disorders are summarized in Table 1. Numbers
408 469
409 tic regression. In secondary analyses, we also assessed the for the analyses of anxiety disorders were similar but slightly 470
410 associations of subtypes of MCI (amnestic and nonamnestic) less, as shown in the Supplementary Table. 471
411 with depressive and anxiety disorders. Table 2 presents the results for the cross-sectional associ- 472
412 Subsequently, we examined the longitudinal associations ations of MCI and its subtypes, with depressive and anxiety 473
413 of MCI with depressive and anxiety disorders using logistic disorders; 5.3% (n 5 22) of persons with MCI (n 5 413) 474
414 regression. For these analyses, cases of depression at base- had comorbid depression (major depression 5 12, 475

FLA 5.4.0 DTD
JALZ2256_proof
9 August 2016
3:48 pm
ce

 S.S. Mirza et al. / Alzheimer’s & Dementia - (2016) 1-10 5

476 Table 1 537

477 Baseline characteristics of the study population included in the analyses of MCI and depressive disorders 538
478 Cross-sectional analysis of depressive disorders, Longitudinal analysis of depressive disorders, 539
479 N 5 4168 N 5 2796 540
480 541
Characteristics No MCI, n 5 3755 MCI, n 5 413 No MCI, n 5 2755 MCI, n 5 212
481 542
482 Age, years 71.4 (7.1) 73.3 (7.5) 70.0 (6.4) 70.0 (6.2) 543
483 Women 2183 (58.1) 215 (52.1) 1608 (58.4) 108 (50.9) 544
484 BMI, kg/m2 27.6 (4.1) 27.7 (4.1) 27.7 (4.0) 27.7 (4.0) 545
Educational level
485 546
Low 341 (9.2) 71 (17.3) 197 (7.3) 30 (14.3)
486 547
Intermediate 2826 (76.5) 284 (69.3) 2100 (77.5) 147 (70.0)
487 High 528 (14.3) 55 (13.4) 412 (15.2) 33 (15.7) 548
488 Smoking 549
489 Never 1120 (29.8) 111 (26.9) 838 (30.4) 58 (27.4) 550
490 Former 2067 (55.0) 230 (55.7) 1536 (55.7) 126 (59.4) 551
491 Current 568 (15.1) 72 (17.4) 381 (13.8) 28 (13.2) 552
492 Total, cholesterol, mmol/L 5.6 (1.0) 5.4 (0.9) 5.7 (1.0) 5.5 (1.0) 553
493 HDL cholesterol, mmol/L 1.5 (0.4) 1.4 (0.4) 1.5 (0.4) 1.4 (0.4) 554
494 Hypertension 3024 (80.5) 344 (83.3) 2156 (78.3) 161 (75.9) 555
Diabetes mellitus 508 (13.5) 74 (17.9) 316 (11.5) 31 (14.6)
495 556
Myocardial infarction 252 (6.7) 50 (12.3) 157 (5.7) 22 (10.5)
496 557
Stroke 112 (3.0) 22 (5.3) 53 (1.9) 11 (5.2)
497 558
498 Abbreviations: BMI, body mass index; HDL, high-density lipoprotein; MCI, mild cognitive impairment. 559
499 NOTE. Values are means (standard deviation) or counts (percentage). 560
500 561
501 562
502 563
503 564
504 dysthymia 5 3, minor depression 5 7), compared with 2.7% agoraphobia 5 16; because of overlap in diagnoses, 565
505 (n 5 103) of those without MCI (n 5 3755). Participants with these numbers cannot be simply added up to 46), compared 566
506 MCI were more likely to have prevalent depressive disorders with 7.7% (n 5 284) of those without MCI (n 5 3662). In the 567
507 than persons without MCI, odds ratio (OR) 1.94 (95% confi- cross-sectional analyses, participants with MCI were more 568
508 dence interval [CI]: 1.20, 3.15). The model was significant likely to have prevalent anxiety disorders than persons 569
509 overall (P , .001) and explained 4.6% of the total variance. without MCI, OR 1.70 (95% CI: 1.19, 2.42). The model 570
510 Female sex, diabetes mellitus 2, and past and current smoking was significant overall (P , .001) and explained 4.6% 571
511 572
were significant predictors of depressive disorders. In second- of the total variance. Age and female sex were significant
512 573
ary analyses, we found similar associations for both amnestic predictors of anxiety disorders. In analyses with subtypes
513 574
514 and nonamnestic MCI with depressive disorders, compared of MCI, participants with nonamnestic MCI were 575
515 with those without MCI: amnestic, OR 2.02 (95% CI: 0.95, more likely to have an anxiety disorder, OR 2.00 (95% CI: 576
516 4.29), and nonamnestic, OR 1.90 (95% CI: 1.05, 3.40). 1.30, 3.04) but not those with amnestic MCI, OR 1.28 577
517 In all, 11.6% (n 5 46) of persons with MCI (n 5 398) had (95% CI: 0.69, 2.38). 578
518 a comorbid anxiety disorder (generalized anxiety disorder In the total sample, only 1% of (n 5 44) participants 579
519 [GAD] 5 14, social phobia 5 11, special phobia 5 9, and had both a depressive and an anxiety disorder at baseline. 580
520 581
521 582
522 Table 2 583
523 Cross-sectional association of MCI with DSM depressive and anxiety disorders 584
524 Analyses of depressive disorders, N 5 4168 Analyses of anxiety disorders, N 5 4060 585
525 586
ORs (95% CIs) ORs (95% CIs)
526 587
y z
527 MCI Cases/N* Model 1 P Model 2 P Cases/N* Model 1y P Model 2z P 588
528 No MCI 103/3755 Reference Reference 284/3662 Reference Reference 589
529 MCI 22/413 2.02 (1.25, 3.26) .004 1.94 (1.20, 3.15) .007 46/398 1.74 (1.24, 2.44) .001 1.70 (1.19, 2.42) .004 590
530 Amnestic 8/161 2.11 (1.00, 4.44) .05 2.02 (0.95, 4.29) .07 14/160 1.35 (0.76, 2.39) .30 1.28 (0.69, 2.38) .42 591
531 Non-amnestic 14/252 1.97 (1.10, 3.53) .02 1.90 (1.05, 3.40) .03 32/238 2.00 (1.34, 2.98) .001 2.00 (1.30, 3.04) .001 592
532 Abbreviations: CIs, confidence intervals; DSM, Diagnostic and Statistical Manual of Mental Disorders; MCI, mild cognitive impairment; ORs, odds ratios.
593
533 *Cases/N: number of depression or anxiety cases/total number of participants in respective groups; P: P values. 594
534 y
Model 1: adjusted for age and sex. 595
535 z
Model 2: additionally adjusted for educational level, body mass index, smoking, total cholesterol, high-density lipoprotein cholesterol, hypertension, 596
536 diabetes mellitus 2, myocardial infarction, stroke, and cohort. 597

FLA 5.4.0 DTD
JALZ2256_proof
9 August 2016
3:48 pm
ce

 6 S.S. Mirza et al. / Alzheimer’s & Dementia - (2016) 1-10

598 Table 3 659

599 Longitudinal association of MCI with incident DSM depressive and anxiety disorders 660
600 Analyses of depressive disorders, N 5 2967 Analyses of anxiety disorders, N 5 2375 661
601 662
ORs (95% CIs) ORs (95% CIs)
602 663
y z y
603 MCI Cases/N* Model 1 P Model 2 P Cases/N* Model 1y P Model 2z P 664
604 No MCI 29/2755 Reference Reference 75/2224 Reference Reference
665
605 MCI 6/212 2.96 (1.21, 7.25) .02 3.13 (1.26, 7.77) .01 11/151 2.54 (1.31, 4.94) .006 2.59 (1.31, 5.12) .006 666
606 Amnestic 1/80 1.58 (0.21, 11.9) .66 1.79 (0.23, 13.70) .57 5/67 2.92 (1.12, 7.62) .03 3.22 (1.21, 8.55) .02 667
607 Non-amnestic 5/132 3.66 (1.39, 9.65) .009 3.77 (1.40, 10.13) .008 6/84 2.31 (0.97, 5.52) .06 2.24 (0.90,5.56) .08 668
608 669
Abbreviations: CIs, confidence intervals; DSM, Diagnostic and Statistical Manual of Mental Disorders; MCI, mild cognitive impairment; ORs, odds ratios.
609 670
NOTE. For these analyses, cases of dementia that occurred during the study period were excluded. These were 25 cases in the analyses of depressive disorders
610 and 9 for the analyses of anxiety disorders. The ORs represent the relative risk of any depressive disorder (major depressive disorder, dysthymia, or minor
671
611 depression) or anxiety disorder (generalized anxiety disorder, panic disorder, or phobias) .9 years. 672
612 *Cases/N: number of depression or anxiety cases/total number of participants in respective groups; P: P values. 673
613 y
Model 1: adjusted for age and sex. 674
z
614 Model 2: additionally adjusted for educational level, body mass index, smoking, total cholesterol, high-density lipoprotein cholesterol, hypertension, 675
615 diabetes mellitus 2, myocardial infarction, stroke, cohort, and time interval between two assessments. 676
616 677
617 678
618 679
619 Results for the longitudinal analyses of MCI with depres- baseline depressive symptoms, MCI was not associated 680
620 sive and anxiety disorders are presented in Table 3; 2.8% with a clear increase in the risk of developing incident 681
621 (n 5 6) of persons with MCI at baseline (n 5 212) devel- depressive symptoms (fully adjusted b MCI 0.73 [95% CI: 682
622 oped incident depression (major depression 5 4 and mi- 20.11, 1.58]; b amnestic MCI 0.41 [95% CI: 20.93, 683
623 nor depression 5 2) compared with 1% (n 5 29) 1.76]; and b nonamnestic MCI 0.93 [95% CI: 20.12, 1.97]). 684
624 of those without MCI at baseline (n 5 2755). We found In the longitudinal analysis of MCI and depressive disor- 685
625 that MCI at baseline was associated with a higher relative ders, results did not change meaningfully after adjusting for 686
626 risk of incident depressive disorders, OR 3.13 (95% CI: baseline depressive symptoms, OR 2.51 (95% CI: 0.97, 687
627 1.26, 7.77). The model was borderline significant 6.48). Both amnestic (OR 1.95 [95% CI: 0.25, 15.30]) and 688
628 689
(P 5 .06) and explained 6.2% of the total variance. Age nonamnestic MCI (OR 2.71 [95% CI: 0.96, 7.66]) contrib-
629 690
and female sex were significant predictors of incident uted to the observed association with depressive disorders.
630 691
631 depressive disorders. In secondary analyses, we found In our data, only 1% of participants had both prevalent 692
632 that nonamnestic MCI was associated with a higher depressive and anxiety disorders. 693
633 relative risk of depressive disorders, OR 3.77 (95% CI: Persons with objective cognitive impairment (n/ 694
634 1.40, 10.13). N 5 180/2967) were not only more likely to have both 695
635 In all, 7.3% (n 5 11) of persons with MCI at baseline depressive (OR 1.68 [95% CI: 0.95, 3.00]) and anxiety 696
636 developed anxiety (GAD 5 4, social phobia 5 1, special (OR 1.38 [95% CI: 0.90, 2.13]) disorders but also had a 697
637 phobia 5 1, and agoraphobia 5 5), compared with 3.4% higher risk of developing depressive (OR 2.13 [95% CI: 698
638 (n 5 75) of those without MCI at baseline (n 5 2224). 0.73, 6.22]) and anxiety (OR 2.11 [95% CI: 1.00, 4.45]) dis- 699
639 MCI at baseline was also associated with a higher relative orders. Results did not change meaningfully when addi- 700
640 701
risk of anxiety disorders, OR 2.59 (95% CI: 1.31, 5.12). tionally adjusting for MMSE score, neither in the cross-
641 702
The model was significant overall (P , .001) and explained sectional nor in the longitudinal analysis. Persons with
642 703
643 7% of the total variance. Age, female sex, education, BMI, MCI had a higher risk of developing both depressive (OR 704
644 diabetes mellitus 2, and HDL cholesterol were significant 3.14 [95% CI: 1.22, 8.05]) and anxiety (OR 2.51 [95% 705
645 predictors for incident anxiety disorders. Both amnestic CI: 1.25, 5.03]) disorders even if analyses were adjusted 706
646 MCI, OR 3.22 (95% CI: 1.21, 8.55), and nonamnestic for MMSE score. 707
647 MCI, OR 2.24 (95% CI: 0.90, 5.56), predicted the risk of Of the 413 participants with MCI at baseline, 33 devel- 708
648 anxiety after adjusting for all potential confounders. There oped dementia. Of those who developed dementia, 6 709
649 were no participants who developed both a depressive and (18%) had depression (major depression 5 1 and depressive 710
650 an anxiety disorder. symptoms 5 5) and 2 (6%) had an anxiety disorder at base- 711
651 In the cross-sectional sensitivity analysis of MCI and line, compared with 36 (9.5%) with depression (depressive 712
652 713
depressive symptoms, participants with MCI were more disorder 5 21 and depressive symptoms 5 15) and 12%
653 714
likely to have depressive symptoms (fully adjusted b for with anxiety among those who did not progress to dementia.
654 715
655 MCI 2.33 [95% CI: 1.62, 3.03]; b amnestic MCI 1.96 Depressive symptoms scores as assessed by CES-D were not 716
656 [95% CI: 0.89, 3.02]; and b nonamnestic MCI 2.56 [95% significantly different between the MCI and non-MCI 717
657 CI: 1.69, 3.44]). In the longitudinal analysis adjusted for groups at baseline. 718
658 719

FLA 5.4.0 DTD
JALZ2256_proof
9 August 2016
3:48 pm
ce

 S.S. Mirza et al. / Alzheimer’s & Dementia - (2016) 1-10 7

720 4. Discussion as these comparative analyses lacked sufficient power. Sec- 781
721 ond, depressive or anxiety disorders might affect those indi- 782
722 In this population-based study, persons with MCI were viduals who are genetically or environmentally more 783
723 more likely to have prevalent depressive and anxiety disor- vulnerable to develop these conditions. Perhaps in vulner- 784
724 ders compared with those without MCI. Moreover, MCI 785
able persons, depressive or anxiety disorders precipitate
725 was associated with a higher relative risk of incidence of 786
earlier, or their risk is amplified by having MCI. Third, the
726 each depressive and anxiety disorder. Non-amnestic MCI 787
727 incidence of depression or anxiety in persons with MCI 788
was associated with a higher relative risk of incident depres- might represent an early stage of dementia. This implies
728 sive disorders, whereas both subtypes of MCI were associ- 789
729 that depression or anxiety appears as an intermediate stage 790
ated with a higher relative risk of incident anxiety disorders.
730 between MCI and dementia. Studies have shown that depres- 791
731 The prevalence of depression (2%) and anxiety (8%) in sion in late life is a prodrome of dementia [47]. Fourth, both 792
732 our population was comparable with the previously re- MCI and psychiatric symptoms can result from a common 793
733 ported prevalence in The Netherlands [38,39]. Prevalence etiological factor(s) where cognitive symptoms precede 794
734 of depression in our population also falls within the wide emotional symptoms. 795
735 range of 1%–20% reported by population-based studies. Vascular pathology is a possible explanation for the 796
736 However, methodological variability might explain the 797
observed associations; however, we carefully adjusted for
737 798
differences [40]. Also, the prevalence of anxiety disorders vascular risk factors and comorbidity. This indicates that
738 799
739 in elderly has been reported to range from 3.2% to 14.2% other mechanisms than vascular pathology probably play 800
740 [40,41]. Interestingly, incidence of affective disorders in a role, although the presence of subclinical cardiovascular 801
741 MCI patients in our study was higher than the incidence disease cannot be ruled out. Some studies have also sug- 802
742 of dementia (3%), which is a documented consequence gested that the atrophy of hippocampus results in both 803
743 of MCI. cognitive impairment and depression [48]. There is limited 804
744 We found that persons with MCI had a higher prevalence evidence as to the pathophysiology of anxiety disorders in 805
745 of depressive and anxiety disorders compared with those context of early stages of neurodegeneration, but some 806
746 without MCI, which is in line with the existing literature studies have highlighted the role of caudate nucleus pathol- 807
747 [42–46]. Studies have shown that both depression and ogy as a possible link between MCI and anxiety [45]. The 808
748 809
anxiety are highly prevalent in MCI [4,5]. There are no caudate nucleus plays a vital role in executive function,
749 810
prospective studies to investigate the association of MCI shows typical pathological features in early stages of de-
750 811
751 with depressive or anxiety disorders. The only evidence mentia [49], and has been implicated in anxiety disorders 812
752 about an association of MCI with depression and anxiety [50,51]. In Alzheimer’s disease patients with comorbid 813
753 is derived from either cross-sectional studies or a few short anxiety, the role of bilateral entorhinal cortex, amygdala, 814
754 follow-up studies, which reported depression a risk factor anterior parahippocampal gyri, left superior temporal 815
755 for MCI, that is, depression precedes the development of gyrus, and insula has been implicated [5,52]. However, 816
756 MCI [12,13,45]. pathological changes in both hippocampus and the 817
757 We found that MCI increased the risk of incident depres- caudate nucleus could very well be the consequences of 818
758 sion and anxiety. A lack of longitudinal studies on this sub- underlying vascular pathology. Another common pathway 819
759 ject hampered comparison of our results. However, several could be the dysregulation of neurotransmitters. Both 820
760 potential explanations gave rise to our hypothesis and can 821
altered serotonergic activity observed in anxiety and
761 822
account for the observed associations. First, both depression depression, and the low adrenergic activity observed in
762 823
763
and anxiety could occur as a “reactive” response to the un- depression, are also associated with cognitive disorders 824
764 derlying cognitive impairment, and symptoms could worsen [53–55]. 825
765 with the increasing cognitive impairment. Experiencing Studies have also argued that MCI is a psychiatric entity. 826
766 forgetfulness, gradual inability to perform everyday tasks, This hypothesis proposes that in MCI, although the under- 827
767 and a fear of developing dementia could be intimidating lying pathology is neurodegenerative, the manifestation is 828
768 enough to trigger severe symptoms of anxiety or depression largely in the form of psychiatric symptoms [56,57]. 829
769 in vulnerable individuals with MCI. Possibly, psychiatric Persons experiencing a decline in their cognitive abilities 830
770 symptoms would affect those persons with MCI more, and being aware of this decline are likely to react with an 831
771 who are more “aware” of their condition, and are able to alteration in mood or behavior and sleeping or eating 832
772 quantitate the decline in their cognitive abilities. This might habits. 833
773 834
be particularly true for symptoms of anxiety and could To our knowledge, this is the first longitudinal study to
774 835
775 explain the finding of the prominent association between investigate MCI in relation to incident anxiety and depres- 836
776 nonamnestic MCI with prevalent anxiety disorders in our sion. We also tested objective cognitive impairment sepa- 837
777 study. Previously, an equally high prevalence of anxiety in rately in association to incident depression and anxiety. 838
778 both subtypes of MCI has been reported but only in small This enabled us to disentangle if the observed associations 839
779 clinical sample [45]. Nevertheless, the results regarding sub- were only driven by subjective memory complaints 840
780 types of MCI in our study should be interpreted with caution, component of MCI and thus possibly by preexisting 841

FLA 5.4.0 DTD
JALZ2256_proof
9 August 2016
3:48 pm
ce

 8 S.S. Mirza et al. / Alzheimer’s & Dementia - (2016) 1-10

842 depression- or anxiety-related complaints [58]. In addi- 903

843 tion, we also adjusted for cognitive score to test the extent RESEARCH IN CONTEXT 904
844 to which cognitive impairment drives the associations. 905
845 Other strengths include a large population-based sample, 906
846 1. Systematic review: Authors reviewed literature using 907
adjusting for several potential confounders, and a robust
847 Pubmed. In recent years, mild cognitive impairment 908
continuous monitoring of dementia that enabled to
848 909
exclude incident cases occurring during the study period. (MCI) has been of considerable interest because it
849 910
850 However, 1354 subjects from our population (n 5 6061) potentially marks the preclinical phase of dementia.
911
851 did not have all cognitive tests or had unreliable cognitive Affective disorders are another common manifesta- 912
852 test results. This might have led to some selection, as per- tion in the preclinical phase of dementia. Cross- 913
853 sons with absent cognitive test scores might be cognitively sectional studies report a higher prevalence of af- 914
854 more compromised than those who performed all cogni- fective disorders also in persons with MCI, but it is 915
855 tive tests. This could have led to an underestimation of as- unknown if MCI also increases the risk of future 916
856 sociations. However, the unreliable test scores were only affective disorders. 917
857 partly because of refusal or insufficient motivation from 918
858 2. Interpretation: MCI was associated not only with a 919
the participant. In other participants, they were a result higher prevalence of Diagnostic and Statistical
859 920
of technical problems in tests conduction, physical limita-
860 Manual of Mental Disorders—depressive and anxi- 921
861 tions to perform the tests, deviation from the instructions ety disorders—but also with a higher risk of devel- 922
862 given to perform the tests, or incorrectly administered oping these disorders. MCI is thus not only a 923
863 tests. Thus, unreliable test scores were not always related potential marker of preclinical dementia but also a 924
864 to participant characteristics. In addition, we did not have risk factor for depression and anxiety. This suggests 925
865 sufficient number of cases to test the association between shared etiology between dementia and affective dis- 926
866 amnestic MCI and incident depression and therefore the 927
orders.
867 estimates might not be precise. Additionally, residual con- 928
868 founding because of unknown or unmeasured confounders 3. Future directions: MCI is a possible harbinger of 929
869 might be present. many nondementia mental disorders. Our results 930
870 also direct joint etiological research to unravel the 931
In conclusion, MCI is a possible precursor of depressive
871 pathways involved in the pathogenesis of these disor- 932
and anxiety disorders. Therefore, MCI should not just be re-
872 933
garded as a potential transition stage between normal aging ders.
873 934
874 and dementia but a forerunner of both cognitive and psychi- 935
875 atric outcomes. Our results suggest a shared etiology between 936
876 neurodegenerative and psychiatric disorders and open ave- 937
877 nues for etiological research to unravel the common biolog- 938
878 ical pathways underlying cognitive and psychiatric disorders. 939
879 References 940
880 941
881 Acknowledgments 942
[1] World Health Organization and Alzheimer’s Disease International.
882 Dementia: a public health priority; 2012. Q3 943
Q2 The Rotterdam Study is funded by Erasmus Medical Center,
883 [2] Petersen RC, Roberts RO, Knopman DS, Boeve BF, Geda YE, 944
Rotterdam, The Netherlands Organization for the Health
884 Ivnik RJ, et al. Mild cognitive impairment: ten years later. Arch Neurol 945
885 Research and Development (ZonMw), the Ministry of Edu- 2009;66:1447–55. 946
886 cation, Culture and Science, the Ministry for Health, Welfare [3] Lyketsos CG, Carrillo MC, Ryan JM, Khachaturian AS, Trzepacz P, 947
and Sports, the Research Institute for Diseases in the Elderly Amatniek J, et al. Neuropsychiatric symptoms in Alzheimer’s disease.
887 948
Alzheimers Dement 2011;7:532–9.
888 (014-93-015, RIDE2), and The Netherlands Genomics 949
[4] Geda YE, Smith GE, Knopman DS, Boeve BF, Tangalos EG, Ivnik RJ,
889 Initiative/Netherlands Consortium for Healthy Ageing proj- et al. De novo genesis of neuropsychiatric symptoms in mild cognitive 950
890 ect no. 050-060-810. The work of Henning Tiemeier is sup- impairment (MCI). Int Psychogeriatr 2004;16:51–60. 951
891 ported by Vidi (grant 017.106.370). None of the funding [5] Monastero R, Mangialasche F, Camarda C, Ercolani S, Camarda R. A 952
892 organizations or sponsors were involved in the study design; systematic review of neuropsychiatric symptoms in mild cognitive 953
893 in collection, analysis, or interpretation of data; in writing impairment. J Alzheimers Dis 2009;18:11–30. 954
894 the report; and in making decision to submit the article for
[6] Peters ME, Rosenberg PB, Steinberg M, Norton MC, Welsh- 955
895 Bohmer KA, Hayden KM, et al. Neuropsychiatric symptoms as risk 956
publication. Financial disclosures: The authors disclose the factors for progression from CIND to dementia: the Cache County
896 957
897 conflicts of interest as none. Study. Am J Geriatr Psychiatry 2013;21:1116–24.
958
[7] Rosenberg PB, Mielke MM, Appleby BS, Oh ES, Geda YE,
898 959
Lyketsos CG. The association of neuropsychiatric symptoms in MCI
899 Supplementary data with incident dementia and Alzheimer disease. Am J Geriatr Psychia- 960
900 try 2013;21:685–95. 961
901 Supplementary data related to this article can be found at [8] Barnes DE, Yaffe K, Byers AL, McCormick M, Schaefer C, 962
902 http://dx.doi.org/10.1016/j.jalz.2016.06.2361. Whitmer RA. Midlife vs late-life depressive symptoms and risk of 963

FLA 5.4.0 DTD
JALZ2256_proof
9 August 2016
3:48 pm
ce

 S.S. Mirza et al. / Alzheimer’s & Dementia - (2016) 1-10 9

964 dementia: differential effects for Alzheimer disease and vascular de- Schedule. I. Development and reliability. Psychol Med 1976; 1025
965 mentia. Arch Gen Psychiatry 2012;69:493–8. 6:439–49. 1026
966 [9] Gao Y, Huang C, Zhao K, Ma L, Qiu X, Zhang L, et al. Depression as a [28] Roth M, Tym E, Mountjoy CQ, Huppert FA, Hendrie H, Verma S, et al. 1027
967 risk factor for dementia and mild cognitive impairment: a meta- CAMDEX. A standardised instrument for the diagnosis of mental dis- 1028
analysis of longitudinal studies. Int J Geriatr Psychiatry 2013; order in the elderly with special reference to the early detection of de-
968 1029
28:441–9. mentia. Br J Psychiatry 1986;149:698–709.
969 1030
[10] de Bruijn RF, Direk N, Mirza SS, Hofman A, Koudstaal PJ, [29] McKhann G, Drachman D, Folstein M, Katzman R, Price D,
970 Tiemeier H, et al. Anxiety is not associated with the risk of dementia Stadlan EM. Clinical diagnosis of Alzheimer’s disease: report of the 1031
971 or cognitive decline: the Rotterdam Study. Am J Geriatr Psychiatry NINCDS-ADRDA Work Group under the auspices of Department of 1032
972 2014;22:1382–90. Health and Human Services Task Force on Alzheimer’s Disease. 1033
973 [11] Brodaty H, Heffernan M, Draper B, Reppermund S, Kochan NA, Neurology 1984;34:939–44. 1034
974 Slavin MJ, et al. Neuropsychiatric symptoms in older people with [30] Glaesmer H, Riedel-Heller S, Braehler E, Spangenberg L, Luppa M. 1035
975 and without cognitive impairment. J Alzheimers Dis 2012;31:411–20. Age- and gender-specific prevalence and risk factors for depressive 1036
976 [12] Geda YE, Roberts RO, Mielke MM, Knopman DS, Christianson TJ, symptoms in the elderly: a population-based study. Int Psychogeriatr 1037
977 Pankratz VS, et al. Baseline neuropsychiatric symptoms and the risk 2011;23:1294–300. 1038
of incident mild cognitive impairment: a population-based study. [31] Musselman DL, Evans DL, Nemeroff CB. The relationship of depres-
978 1039
Am J Psychiatry 2014;171:572–81. sion to cardiovascular disease: epidemiology, biology, and treatment.
979 1040
[13] Spira AP, Rebok GW, Stone KL, Kramer JH, Yaffe K. Depressive Arch Gen Psychiatry 1998;55:580–92.
980 symptoms in oldest-old women: risk of mild cognitive impairment [32] Thurston RC, Rewak M, Kubzansky LD. An anxious heart: anxiety 1041
981 and dementia. Am J Geriatr Psychiatry 2012;20:1006–15. and the onset of cardiovascular diseases. Prog Cardiovasc Dis 2013; 1042
982 [14] Steenland K, Karnes C, Seals R, Carnevale C, Hermida A, Levey A. 55:524–37. 1043
983 Late-life depression as a risk factor for mild cognitive impairment or [33] Vink D, Aartsen MJ, Schoevers RA. Risk factors for anxiety and 1044
984 Alzheimer’s disease in 30 US Alzheimer’s disease centers. J Alz- depression in the elderly: a review. J Affect Disord 2008;106:29–44. 1045
985 heimers Dis 2012;31:265–75. [34] Whitworth JA. 2003 World Health Organization (WHO)/International 1046
986 [15] Hofman A, Darwish MS, van Duijn CM, Franco OH, Goedegebure A, Society of Hypertension (ISH) statement on management of hyperten- 1047
987 Ikram MA, et al. The Rotterdam Study: 2014 objectives and design up- sion. J Hypertens 2003;21:1983–92. 1048
date. Eur J Epidemiol 2013;28:889–926. [35] Alberti KG, Zimmet PZ. Definition, diagnosis and classification of
988 1049
[16] de Bruijn RF, Akoudad S, Cremers LG, Hofman A, Niessen WJ, van diabetes mellitus and its complications. Part 1: diagnosis and classifi-
989 1050
der Lugt A, et al. Determinants, MRI correlates, and prognosis of cation of diabetes mellitus provisional report of a WHO consultation.
990 mild cognitive impairment: the Rotterdam Study. J Alzheimers Dis Diabet Med 1998;15:539–53. 1051
991 2014;42:S239–49. [36] Leening MJ, Kavousi M, Heeringa J, van Rooij FJ, Verkroost-van HJ, 1052
992 [17] Houx PJ, Jolles J, Vreeling FW. Stroop interference: aging effects as- Deckers JW, et al. Methods of data collection and definitions of cardiac 1053
993 sessed with the Stroop Color-Word Test. Exp Aging Res 1993; outcomes in the Rotterdam Study. Eur J Epidemiol 2012;27:173–85. 1054
994 19:209–24. [37] Wieberdink RG, Ikram MA, Hofman A, Koudstaal PJ, Breteler MM. 1055
995 [18] Hoogendam YY, Hofman A, Van der Geest JN, van der Lugt A, Trends in stroke incidence rates and stroke risk factors in Rotterdam, 1056
996 Ikram MA. Patterns of cognitive function in aging: the Rotterdam the Netherlands from 1990 to 2008. Eur J Epidemiol 2012; 1057
997 Study. Eur J Epidemiol 2014;29:133–40. 27:287–95. 1058
[19] Luijendijk HJ, van den Berg JF, Dekker MJ, van Tuijl HR, Otte W, [38] Beekman AT, Bremmer MA, Deeg DJ, van Balkom AJ, Smit JH, de
998 1059
Smit F, et al. Incidence and recurrence of late-life depression. Arch Beurs E, et al. Anxiety disorders in later life: a report from the Longi-
999 1060
Gen Psychiatry 2008;65:1394–401. tudinal Aging Study Amsterdam. Int J Geriatr Psychiatry 1998;
1000 [20] Wing JK, Babor T, Brugha T, Burke J, Cooper JE, Giel R, et al. SCAN. 13:717–26. 1061
1001 Schedules for Clinical Assessment in Neuropsychiatry. Arch Gen Psy- [39] Beekman AT, de Beurs E, van Balkom AJ, Deeg DJ, Van DR, van TW. 1062
1002 chiatry 1990;47:589–93. Anxiety and depression in later life: co-occurrence and communality 1063
1003 [21] Beekman AT, Deeg DJ, Van LJ, Braam AW, De Vries MZ, van of risk factors. Am J Psychiatry 2000;157:89–95. 1064
1004 Tilburg W. Criterion validity of the Center for Epidemiologic Studies [40] Steffens DC, Skoog I, Norton MC, Hart AD, Tschanz JT, 1065
1005 Depression scale (CES-D): results from a community-based sample of Plassman BL, et al. Prevalence of depression and its treatment in an 1066
1006 older subjects in The Netherlands. Psychol Med 1997;27:231–5. elderly population: the Cache County study. Arch Gen Psychiatry 1067
1007 [22] Wittchen HU, Lachner G, Wunderlich U, Pfister H. Test-retest reli- 2000;57:601–7. 1068
ability of the computerized DSM-IV version of the Munich- [41] Ritchie K, Artero S, Beluche I, Ancelin ML, Mann A, Dupuy AM,
1008 1069
Composite International Diagnostic Interview (M-CIDI). Soc Psychi- et al. Prevalence of DSM-IV psychiatric disorder in the French elderly
1009 1070
atry Psychiatr Epidemiol 1998;33:568–78. population. Br J Psychiatry 2004;184:147–52.
1010 [23] American Psychiatric Association. Diagnostic and Statistical Manual [42] Feldman H, Scheltens P, Scarpini E, Hermann N, Mesenbrink P, 1071
1011 of Mental Disorders. 4th ed. Washington, DC: American Psychiatric Mancione L, et al. Behavioral symptoms in mild cognitive impair- 1072
1012 Association; 1994. ment. Neurology 2004;62:1199–201. 1073
1013 [24] Hek K, Direk N, Newson RS, Hofman A, Hoogendijk WJ, Mulder CL, [43] Lyketsos CG, Lopez O, Jones B, Fitzpatrick AL, Breitner J, 1074
1014 et al. Anxiety disorders and salivary cortisol levels in older adults: a DeKosky S. Prevalence of neuropsychiatric symptoms in dementia 1075
1015 population-based study. Psychoneuroendocrinology 2013;38:300–5. and mild cognitive impairment: results from the cardiovascular health 1076
1016 [25] Schrijvers EM, Verhaaren BF, Koudstaal PJ, Hofman A, Ikram MA, study. JAMA 2002;288:1475–83. 1077
1017 Breteler MM. Is dementia incidence declining?: Trends in dementia [44] Palmer K, Berger AK, Monastero R, Winblad B, Backman L, 1078
incidence since 1990 in the Rotterdam Study. Neurology 2012; Fratiglioni L. Predictors of progression from mild cognitive impair-
1018 1079
78:1456–63. ment to Alzheimer disease. Neurology 2007;68:1596–602.
1019 1080
[26] Folstein MF, Folstein SE, McHugh PR. “Mini-mental state.” A prac- [45] Rozzini L, Vicini CB, Conti M, Delrio I, Borroni B, Trabucchi M, et al.
1020 tical method for grading the cognitive state of patients for the clinician. Neuropsychiatric symptoms in amnestic and nonamnestic mild cogni- 1081
1021 J Psychiatr Res 1975;12:189–98. tive impairment. Dement Geriatr Cogn Disord 2008;25:32–6. 1082
1022 [27] Copeland JRM, Kelleher MJ, Kellett JM, Gourlay AJ, Gurland BJ, [46] Rozzini L, Chilovi BV, Peli M, Conti M, Rozzini R, Trabucchi M, et al. 1083
1023 Fleiss JL, et al. A semi-structured clinical interview for the assesment Anxiety symptoms in mild cognitive impairment. Int J Geriatr Psychi- 1084
1024 of diagnosis and mental state in the elderly. The Geriatric Mental State atry 2009;24:300–5. 1085

FLA 5.4.0 DTD
JALZ2256_proof
9 August 2016
3:48 pm
ce

 10 S.S. Mirza et al. / Alzheimer’s & Dementia - (2016) 1-10

1086 [47] Mirza SS, de Bruijn RF, Direk N, Hofman A, Koudstaal PJ, Ikram MA, [53] Lai MK, Tsang SW, Esiri MM, Francis PT, Wong PT, Chen CP. Differ- 1147
1087 et al. Depressive symptoms predict incident dementia during short- but ential involvement of hippocampal serotonin1A receptors and re- 1148
1088 not long-term follow-up period. Alzheimers Dement 2014;10:S323–9. uptake sites in non-cognitive behaviors of Alzheimer’s disease. Psy- 1149
1089 [48] Enache D, Winblad B, Aarsland D. Depression in dementia: epidemi- chopharmacology (Berl) 2011;213:431–9. 1150
ology, mechanisms, and treatment. Curr Opin Psychiatry 2011;24:461–72. [54] Lin SH, Lee LT, Yang YK. Serotonin and mental disorders: a concise
1090 1151
[49] Barber R, McKeith I, Ballard C, O’Brien J. Volumetric MRI study of review on molecular neuroimaging evidence. Clin Psychopharmacol
1091 1152
the caudate nucleus in patients with dementia with Lewy bodies, Alz- Neurosci 2014;12:196–202.
1092 heimer’s disease, and vascular dementia. J Neurol Neurosurg Psychi- [55] Zubenko GS, Moossy J, Kopp U. Neurochemical correlates of major 1153
1093 atry 2002;72:406–7. depression in primary dementia. Arch Neurol 1990;47:209–14. 1154
1094 [50] Grahn JA, Parkinson JA, Owen AM. The cognitive functions of the [56] Lyketsos CG. Lessons from neuropsychiatry. J Neuropsychiatry Clin 1155
1095 caudate nucleus. Prog Neurobiol 2008;86:141–55. Neurosci 2006;18:445–9. 1156
1096 [51] Whiteside SP, Port JD, Deacon BJ, Abramowitz JS. A magnetic reso- [57] Lyketsos CG, Kozauer N, Rabins PV. Psychiatric manifestations of 1157
1097 nance spectroscopy investigation of obsessive-compulsive disorder neurologic disease: where are we headed? Dialogues Clin Neurosci 1158
1098 and anxiety. Psychiatry Res 2006;146:137–47. 2007;9:111–24. 1159
1099 [52] Hashimoto H, Monserratt L, Nguyen P, Feil D, Harwood D, [58] Yates JA, Clare L, Woods RT, Matthews FE. Subjective mem- 1160
Mandelkern MA, et al. Anxiety and regional cortical glucose meta- ory complaints are involved in the relationship between mood
1100 1161
bolism in patients with Alzheimer’s disease. J Neuropsychiatry Clin and mild cognitive impairment. J Alzheimers Dis 2015;
1101 1162
Neurosci 2006;18:521–8. 48:S115–23.
1102 1163
1103 1164
1104 1165
1105 1166
1106 1167
1107 1168
1108 1169
1109 1170
1110 1171
1111 1172
1112 1173
1113 1174
1114 1175
1115 1176
1116 1177
1117 1178
1118 1179
1119 1180
1120 1181
1121 1182
1122 1183
1123 1184
1124 1185
1125 1186
1126 1187
1127 1188
1128 1189
1129 1190
1130 1191
1131 1192
1132 1193
1133 1194
1134 1195
1135 1196
1136 1197
1137 1198
1138 1199
1139 1200
1140 1201
1141 1202
1142 1203
1143 1204
1144 1205
1145 1206
1146 1207

FLA 5.4.0 DTD
JALZ2256_proof
9 August 2016
3:48 pm
ce