Severity of Depression Impacts Imminent Conversion from Mild Cognitive Impairment to Alzheimer’s Disease

Abstract

Background:

Mild cognitive impairment (MCI) has been suggested to represent a prodromal stage of dementia and to confer a high risk for conversion to dementia Alzheimer’s type (DAT).

Objectives:

In this study, we examined the predictive value of depressive symptoms and neuropsychological variables on conversion of MCI to DAT.

Methods:

Neuropsychological and clinical follow-up data of 260 MCI patients seen at the Psychiatric Memory Clinic of the Medical University of Innsbruck between 2005 and 2015 were analyzed retrospectively. Depression was assessed using the Geriatric Depression Scale (GDS). Potential predictors of conversion from MCI to DAT were analyzed by logistic regression analyses and additional survival-analytic methods.

Results:

Of the 260 patients (mean age 71.5±7.7 years), 83 (32%) converted to DAT within a mean follow-up time of 3.2±2.2 years and estimated one-year conversion rate of 10.1%. The univariate analysis showed with few exceptions (gender, use of antidepressants, low GDS score) group differences at baseline in patients converted to DAT compared to stable MCI patients. Logistic regression analysis as well as survival analysis revealed moderate to severe depression together with higher age and specific cognitive deficits as predictors of conversion from MCI to DAT.

Conclusion:

Our results support the predictive value of different neuropsychological measures on the progression of DAT. In addition, we found a strong negative influence of depression on conversion to DAT in MCI patients. These results emphasize the importance of assessing depressive symptoms in the early stages of DAT when evaluating the conversion from MCI to DAT.

Keywords

Alzheimer’s disease depression mild cognitive impairment risk factors

INTRODUCTION

Over the last few years, the concept of dementia Alzheimer’s type (DAT) has changed significantly. There is increasing evidence that pathophysiologic changes already begin decades before cognitive symptoms qualifying for dementia or even mild cognitive impairment (MCI) occur [1]. Therefore, sporadic DAT may be a slowly progressing disease with a long preclinical phase which fluently merge to MCI and finally exacerbates in the clinically manifest dementia in old age.

MCI is described as a “significant cognitive impairment not yet fulfilling the diagnostic criteria of dementia” [2]. MCI is thought to represent a transitional state between clinically silent and early dementia stages. In individuals 65 years of age and older, the incidence of DAT is 1-2% per year in the general population, and approximately 15–20% in those with MCI diagnosed according to the original criteria by Petersen [3]. More recent meta-analysis of cohort studies the adjusted annual conversion rate from MCI to DAT ranges from approximately 4–10% depending on applied diagnostic criteria for MCI and study setting [4]. Currently, there is no consistent evidence for any effective intervention to prevent MCI patients from converting to DAT or to reduce the incidence of dementia [5, 6].

Therefore, as stated by numerous international Alzheimer’s Associations, the early identification of risk markers and prodromal symptoms of dementia is an increasingly important target for research and clinical practice. In recent years, late life depression has been suggested as a risk factor for DAT and may represent a prodrome of imminent dementia [7 –9]. Meta-analyses of studies on the prevalence of mental disorders in Western countries found prevalence rates of 2–42% for depressive disorders among people older than 65 years [10, 11]. Many older adults with depression present with cognitive deficits, particularly in episodic memory, psychomotor speed, and executive functioning [12 –14] which can persist even after remission of depression [15]. Vice versa, depressive symptoms are found in 16–27% of MCI patients [16 –18]. Johnson et al. [19] have analyzed data of patients with MCI and DAT and have described a so-called “depressive endophenotype” related to 5 specific items (feeling of worse memory problems, feeling downhearted and blue, feeling worthless, frequently feel like crying, and trouble concentrating) on the Geriatric Depression Scale (GDS) which was predictive for DAT. In accordance with the well-known reported variability of the cognitive profile in geriatric depression, research has yielded inconsistent findings on the association between depression and DAT. Therefore, the distinction of depression in old age as either comorbid disorder, a risk factor, a prodrome, or a neuropsychiatric symptom of early DAT is of relevance in clinical practice. Consequently, novel treatment approaches in preclinical DAT should include an exact evaluation of depressive symptoms and, if necessary, antidepressive therapy. In particular, the numerous other known risk factors of DAT such as diabetes [20–21], hypercholesterolemia [20], white matter pathology [22], or the presence of the apolipoprotein E (APOE) ɛ4 allele complicate the subdivision of high risk MCI patients with preclinical DAT from MCI non-converters.

We evaluated the impact of depressive symptoms measured by the GDS as well as neuropsychological and clinical variables as predictors of conversion from MCI to DAT in clinical patients of our memory-clinic. Therefore, we conducted a retrospective follow-up study using data from detailed neuropsychological and clinical examinations of MCI patients. The aim of the study was to determine the role of depressive symptoms in preclinical dementia and its impact on conversion to DAT. Further, we wanted to investigate whether the presence of depressive symptoms in the preclinical stages of dementia can be differentiated from those in the elderly not associated with DAT.

MATERIALS AND METHODS

Study design

This was a retrospective, observational study to assess neuropsychological and clinical measures in patients with MCI as potential predictors for conversion to DAT. Data were collected from files of MCI patients who had visited the Memory Clinic (Department of Psychiatry) at the Medical University of Innsbruck between 2005 and 2015. All patients completed a neuropsychological assessment and a clinical interview at baseline and at least one follow-up visit until a minimum period of 12 months. Only those patients for which data of at least one follow-up visit were available were included into retrospective analysis. Information on somatic comorbidities, the APOE genotype as well as current psychotropic medication with antidepressants (ATDs), benzodiazepines (BZDs), and antipsychotics (APs) was obtained from medical records, patients, or caregivers. Inclusion criteria were a diagnosis of MCI (amnestic MCI or amnestic MCI multi-domain type), German language proficiency, and age of ≥60 years. Exclusion criteria included acute neurological diseases (e.g., stroke, Parkinson’s disease, normal pressure hydrocephalus), serious or unstable medical diseases (e.g., cardiac disease, hypertension, current infection, neoplasia, clinically significant hepatic, renal, pulmonary, metabolic, or endocrine disturbances), or the presence of a comorbid psychiatric disorder (including delirium, alcohol or drug dependency, and schizophrenia. No patient had received antidementive therapy prior to the follow-up visit.

MCI was diagnosed according to the criteria of Petersen [2] if patients reported subjective memory complaints over the previous 6 months, and showed an impaired memory function (verbal or figural) in the neuropsychological assessment of >1.5 standard deviation (SD) related to age and education and additionally had a Clinical Dementia Rating Scale (CDR) score of 0.5. At follow-up, progression from MCI to DAT was defined as follows: presence of subjective memory complaints over the previous 6 months, impaired neuropsychological function of >2 SD or more corrected for age and education in one memory function (verbal or figural memory) and at least one other cognitive domain, deficits in activities of daily living assessed with a clinical interview, and a CDR scoreof ≥1.

This retrospective study of clinical data was approved by the Ethics Committee of the Medical University of Innsbruck, Austria. Due to retrospective analysis of data, patients were exempt from signing an informed consent.

Neuropsychological measures

As part of the standard clinical evaluation procedures all patients had undergone a neuropsychological assessment at baseline and annual follow-up visits. They were tested on the verbal memory (word list learning, word list delayed recall) and recognition subtests of the “Consortium to Establish a Registry for Alzheimer’s Disease (CERAD) battery [23], figural memory (free recall, CERAD), object naming (Boston Naming Test [BNT] –short version, CERAD), categorical verbal fluency (animals/min, s-words/min, CERAD), planning (CLOX Test part 1), divided attention, and cognitive flexibility (Trail Making Test-A and B) [24]. As cut-off levels for univariate analysis and multiple logistic regression analysis, we set a more strict cut-off defined as z >1 SD below the age-corrected normative mean for memory functions and a more liberal one of z >0 for object naming because of the higher sensitivity and specificity of impairment in memory functions in preclinical stages of DAT and the inclusion of solely the amnestic subtypes of MCI in the study. In addition, the Mini-Mental State Examination (MMSE) [25] was applied.

Depressive symptoms were assessed using the 30-items GDS version [26]. The GDS questions are answered with “yes” or “no”. The cumulative score is rated on a scoring grid. The grid sets a range of 0–9 as “normal”, 10–14 as “mildly depressed”, 15–20 as “moderately depressed”, and 21–30 as “severely depressed”.

Statistical methods

Patient characteristics of converters from MCI to DAT and non-converters (diagnosis of MCI both at baseline and follow-up) were compared by means of the respective two-sample tests (t-test, Mann-Whitney U-test, or Chi-square test for normally distributed, non-normally distributed metric variables and nominal variables, respectively). Deviations from normality were checked by assessing the skewness of the respective variables considering values >0.5 or < –0.5 as deviations from a symmetric distribution requiring non-parametric testing.

To identify predictors of conversion from MCI to DAT, a two-step procedure was used. In a first step, all potential predictors were considered one by one (univariate analysis). Variables considered included socio-demographics (age, gender, education), depression (GDS), cognition (MMSE, Word List Delayed Recall, Visual Memory, CLOX1, BNT), psychotropic medication (ATD, AP, BZD), and APOE. Simple logistic regression was used for significance testing. In addition, odds ratios (OR) were determined as a measure of effect size.

In a second step, the potential predictors were analyzed jointly by means of multiple logistic regression analysis. As the time elapsed between baseline and follow-up varied considerably between patients, we also added this variable to the list of potential predictors. Significant predictors were identified by backward variable elimination. Two versions of the logistic regression analysis were run: one in which all continuous independent variables were categorized (usually dichotomized) to facilitate understanding for the medical reader, and a second one without categorization of continuous variables, which is more satisfactory from a statistical point of view [27].

To better represent the longitudinal nature of the data, an additional analysis by survival-analytic methods was performed. Due to the naturalistic study design (where diagnoses were updated only when patients visited the memory clinic), the exact times of conversion were unknown. We therefore estimated these times based on MMSE scores at baseline (t₁) and follow-up (t₂). For short to medium follow-up times (less than 3 years), we assumed MMSE scores to show an approximately linear course. Based on this assumption, we determined the time when the MMSE score reached the value 25, i.e., the cut-off value for dementia and used it as an estimate of the true conversion time. For longer follow-up times (≥ 3 years), we used an alternative method since approximation of the MMSE by a straight line might be too inaccurate. Instead we traced the course of the MMSE back from time t₂ assuming an average reduction of 3 points per year, as is suggested as a rough approximation by the literature [28]. The estimate of the conversion time was again defined as the time when the MMSE reached the value 25. Using this convention, we analyzed the conversion data by means of a Cox regression model using the same set of independent variables as in the logistic regression described above. In this model, the survival time was defined as the estimated conversion time in converters and as the time to follow-up in non-converters (assuming the latter time to be subject to right-censoring). Moreover, we determined conversion rates from MCI to DAT using the Kaplan-Meier method.

All statistical tests were performed at a 0.05 level of significance. Significance testing was performed without adjustment for multiplicity as such adjustments are uncommon in the context of multiple regression analyses (logistic regression, Cox regression). We note that in principle this may lead to an inflation of the type-one error.

RESULTS

Demographic data

Two hundred and sixty patients diagnosed with MCI according to the diagnostic criteria of Petersen [2, 3] were included in the study. Of these, 83 (32%) converted to DAT (converters) and 177 remained stable (non-converters) within a mean follow-up time of 3.2±2.2 years (mean±SD). At baseline, converters and non-converters had a similar gender distribution. Age and GDS score were significantly higher and education years and MMSE score were lower in converters compared to non-converters. Detailed subject characteristics are presented in Table 1.

Table 1

Clinical and demographic patient characteristics of the patients sample and group comparison of MCI converters and non-converters at baseline (total sample, n = 260)

	Group			Comparison
Variable	Mean±SD or n (%)			Test statistic	p-value
	Total	Converters	Non-converters
	n = 260	n = 83	n = 177
Age (y)	71.5±7.7	74.3±7.3	70.2±7.7	t = 4.04	<0.001
Male	80 (30.8%)	27 (32.5%)	53 (29.9%)	χ² = 0.18	0.674
Female	180 (69.2%)	56 (67.5%)	124 (70.1%)
APOE (% ɛ4-carriers)	44.9%	48.6%	30.1%	χ² = 7.02	0.010
Median (Interquartile range)^a
Education (y)	10.0	9.0	10.0	Z = 2.71^b	0.007
	(8.0–12.0)	(8.0–11.0)	(8.0–12.0)
GDS (total score)	9.0	11.0	8.0	Z = 2.76^b	0.006
	(5.0–14.0)	(6.0–16.0)	(5.0–12.0)
MMSE (total score)	28.0	26.5	29.0	Z = 8.02^b	<0.001
	(27.0–29.0)	(26.0–28.0)	(28.0–29.0)

^aDue to deviations from a normal distribution the Mann-Whitney U-test was used. ^bDue to deviations from a normal distribution the Mann-Whitney U-test was used.

Predictors of conversion from MCI to dementia (logistic regression)

Findings of the univariate and multivariate logistic regression analyses for the identification of predictors of conversion to DAT are shown in Table 2.

Table 2

Predictors of conversion to dementia – univariate and multiple logistic regression

Variable	Category	N	Univariate logistic regression			Multiple logistic
						regression
			Conversion rate	Odds ratio	p-value	Adjusted OR	p– value
Age	<70	119	21.8%
	≥70	141	40.4%	2.43	0.002^**	2.78	0.011^*
Gender	Male	80	33.8%
	Female	180	31.1%	0.89	0.674	0.73	0.802
GDS Baseline	Overall				0.003^**		0.004^**
	<5	55	23.6%	1.00	–	1.00	–
	5–9	77	27.3%	1.21	0.638	1.56	0.370
	10–14	62	32.3%	1.54	0.301	1.91	0.264
	15–19	40	42.5%	2.39	0.053(^*)	6.00	0.005^**
	≥20	22	54.5%	3.46	0.011^*	5.28	0.028^*
MMSE Baseline	≥28	162	15.4%
	≤ 27	91	62.6%	9.19	<0.001^**	8.33	<0.001^**
Word List Delayed Recall	Z > –1	171	21.1%
	Z≤–1	68	60.3%	5.69	<0.001^**	1.84	0.162
Visual Memory	Z > –1	181	21.5%
	Z≤–1	69	60.9%	5.66	<0.001^**	4.33	<0.001^**
CLOX1	>10	213	26.8%
	≤ 10	47	55.3%	3.39	<0.001^**	1.66	0.274
Boston Naming Test	Z > 0	147	20.4%
	Z≤0	104	50.0%	3.90	<0.001^**	3.36	0.002^**
Antidepressants	No	167	28.7%
	Yes	93	37.6%	1.50	0.141	0.40	0.035
Sedatives	No	209	28.7%
	Yes	51	45.1%	2.04	0.026^*	1.10	0.851
Neuroleptics	No	236	28.6%
	Yes	24	62.5%	4.12	0.001^**	3.73	0.032^*
APOE	ɛ 22,32,33	137	27.0%
	ɛ 4 carrier	67	44.9%	2.20	0.008^**	1.31	0.553
Myocardial infarction	No	242	29.8%
	Yes	18	61.1%	3.71	0.009^**	1.09	0.895
Diabetes mellitus II	No	240	30.8%
	Yes	20	45.0%	1.75	0.197	2.03	0.277

^*p < 0.05; ^**p < 0.01; (^*)p < 0.10. OR, odds ratio.

The univariate analysis yielded a number of variables associated with conversion to DAT. Among socio-demographic factors, higher age and lower education were significantly associated with higher conversion rates. Baseline level of depression, assessed with the GDS, was also significantly associated with conversion, with gradually increasing conversion rates for increasing levels of depression. As expected, all measures of cognitive functioning showed highly significant associations with conversion to DAT (higher conversion rates in patients with lower scores). Among psychotropic medications, both the use of AP and the use of BZD were related to a higher probability of conversion in the univariate analysis, whereas the use of ATD was not.

Results of the multiple logistic regression analysis are presented in the right-most column of Table 2. Age remained a significant predictor of conversion to dementia with an almost three-fold higher risk of conversion in patients aged 70+ compared to those aged below 70 years. Similarly, depression remained a significant predictor of conversion to DAT. Patients with higher baseline GDS scores showed a significantly increased risk of conversion, attaining odds ratios above 5 for the two highest GDS categories compared to the lowest GDS category. Among neuropsychological variables, MMSE, visual memory, and naming (BNT) emerged as significant predictors of conversion, whereas the effect of CLOX1 and Word List Delayed Recall did not attain significance. Regarding psychotropic medication, the regression analysis revealed that use of ATD at baseline went along with a significantly reduced conversion rate to dementia, a finding not observed in the univariate analysis. Apparently, this is mainly due to the fact that in the multiple logistic regression the effects of the other correlated variables were taken into consideration. The effect of AP use on conversion remained significant whereas medication with BZDs lost its significance. Neither APOE genotype nor any of the comorbid diseases considered showed a significant effect in the multiple regression analysis. Time elapsed between baseline and follow-up did not show a significant effect on the conversion rate in the logistic regression analysis (Wald χ² = 1.60, p = 0.206). This somewhat surprising finding was obviously due to the naturalistic design of our study giving rise to slightly, but not significantly shorter time intervals in converters (2.7±1.9 years) than in non-converters (3.4±2.3 years).

When considering the variables age, education, and GDS score as well as neuropsychological measures as continuous rather than as categorical variables, both univariate and multivariate results remain almost unchanged. In particular, all statistically significant findings of the first analysis remain significant.

Results of survival analysis

Findings of the survival analysis are shown in Fig. 1 (Kaplan Meier graph) and Table 3. The estimated one-year conversion rate from MCI to DAT was 10.1% (95% confidence interval (CI): 7.5–13.7%) while the 5-year conversion rate amounted to 38.3% (95% CI: 30.9–45.7%). Results of the univariate and multiple Cox regression analyses (left and right hand side of Table 3, respectively) matched well with the corresponding results of the logistic regression. In the multiple Cox regression, nearly the same predictors showed statistical significance as in the logistic regression, in particular age, GDS, important neuropsychological measures, and use of ATD (decreased hazard of conversion in patients treated with ATD). In addition, a significant effect of APOE was observed with increased probability of conversion in ɛ4 allele carriers (p = 0.047). Details can be found in Table 3.

Fig.1

Estimated proportion of patients not converting to DAT in relation to time elapsed from baseline.

Table 3

Predictors of conversion to dementia – results of univariate and multiple Cox regression

Variable	Category (unit)	Univariate Cox regression		Multiple Cox regression
		Hazard ratio	p-value	Adjusted HR	p-value
Age	Years	1.092^a	<0.001^***	1.100^a	<0.001^**
Gender	Female versus male	0.846	0.467	0.654	0.124
GDS Baseline	Raw score	1.037^a	0.021^*	1.063^a	0.004^**
MMSE Baseline	Raw score	0.647^b	<0.001^**	0.807^b	0.002^**
Word List Delayed Recall	Z-Score	0.498^b	<0.001^**	0.672^b	0.001^**
Visual Memory	Z-Score	0.562^b	<0.001^**	0.781^b	0.054
CLOX1	Raw score	0.815^b	<0.001^**	1.063	0.259
Boston Naming Test	Z-Score	0.584^b	<0.001^**	0.757^b	0.023^*
Antidepressants	Yes versus no	1.529	0.057	0.463	0.014^*
Sedatives	Yes versus no	1.948	0.007^**	1.125	0.658
Neuroleptics	Yes versus no	2.834	<0.001^**	1.542	0.203
APOE	ɛ4 carrier versus non-ɛ4	1.691	0.026^*	1.774	0.047^*
Myocardial infarction	Yes versus no	2.991	0.003^**	1.680	0.181
Diabetes mellitus II	Yes versus no	1.568	0.205	1.287	0.538

^*p < 0.05; ^**p < 0.01; (^*)p < 0.10. HR, hazard ratio. ^aincrease of hazard per unit increase of the independent variable. ^bdecrease of hazard per unit increase of the independent variable.

DISCUSSION

In this study, we examined the impact of a number of sociodemographic, neuropsychological, clinical, and pharmacologic variables on the risk for conversion from MCI to DAT in a longitudinal follow-up study. Our results underscore that, next to established risk factors for DAT such as the presence of an APOE ɛ4 allele, higher age, and memory impairment, moderate to severe depressive symptoms are predictive for the imminent conversion to DAT. We found a clear positive association of depression severity and risk of conversion to DAT, suggesting a linear relationship between severity of depression and the risk of developing DAT starting at a GDS score of 15, which corresponds to moderate depression. These findings confirm previous results showing an increased risk of developing dementia in correlation with increasing severity of depression [29]. Also in accordance with Chen et al., we found that mild depressive symptoms are frequent in MCI patients but do not necessarily increase the risk for conversion to DAT.

Further, we found that antidepressant treatment can lower dementia risk in MCI patients, whereas psychotropic medication with AP showed the opposite effect.

One of the most important finding of this study was that a higher baseline level of depression was associated with a higher risk of conversion to DAT, leading to increasing levels of depression. This result remained significant independent of other significant and already established risk factors for DAT such as age, lower MMSE score, and the APOE genotype to the multiple logistic regression analysis.

Our findings are in line with those of Teng et al. who compared differences of neuropsychiatric symptoms in converting and non-converting MCI patients and demonstrated a higher frequency of depression and apathy in converters compared to stable MCI patients within 2 years of follow-up. A large population-based cohort study including over 2.4 million persons also found an increased risk for developing dementia in persons with depression [31]. In contrast to our study, however, depression was measured as present or absent without taking the severity of depressive symptoms measured by a rating scale into account. In addition, there was no information on current antidepressant treatment and the intake of sedatives or antipsychotics. Comparison of baseline MMSE scores and results of all other measured cognitive functions showed significantly lower scores in MCI converters compared to non-converters. After the regression analysis only MMSE scores, impaired visual memory and deficits in naming remained predictive for conversion to DAT. These results are in keeping with previous studies which reported that MCI subjects with lower MMSE scores and impaired memory are more likely to progress to DAT [32, 33]. Interestingly, regression analysis revealed a MMSE score of 27 and lower as a cut-off which is more reliable than that used in other studies. Our finding corresponds to O’Bryant et al. who reported a MMSE cut-off score of 27 to be appropriate for correct discrimination of healthy and MCI patients especially with higher education [34].

The impact of APOE ɛ4 on the progression from MCI to DAT remains to be discussed controversially. The majority of association studies of DAT support the view that the APOE ɛ4 allele is as the most important genetic risk factor in late onset DAT [35, 36]. An accelerated progression from MCI to DAT in ɛ4 homozygotes compared with other APOE genotypes has been reported [37]. Partly controversial, we found a higher rate of conversion from MCI to DAT in APOE ɛ4 carriers, carrying one or two APOE ɛ4 alleles in multiple cox regression but not the multiple logistic regression analysis. In the latter analysis, the impact of the APOE ɛ4 alleles appears to be outweighed by cognitive deficits, higher age or moderate to severe depression in MCI patients. This could be explained by the overrepresentation of the ɛ4 allele in our high-risk population of amnestic MCI patients.

Among psychotropic medications, solely the use of AP was related to a higher probability of converting to dementia in multiple regression analysis. The higher rate of antipsychotic medication prescription in converters may be associated with the development of neuropsychiatric symptoms before the onset of full blown DAT [38]. In this context, Cummings and Apostolova reported a prevalence of 35–75% of neuropsychiatric symptoms in MCI patients [40]. Studies on the prevalence of sleep disturbances in MCI patients revealed an association of sleep/night-time disturbances and an increase risk to develop dementia [41]. These are frequently treated with antipsychotics such as prothipendyl or quetiapine in the elderly [41, 42]. Thus, our results do not necessarily suggest direct causal influence of antipsychotics on conversion from MCI to DAT.

Further, prior studies found a higher conversion rate in MCI patients with anxiety symptoms [43]. In support of this, we found a higher prevalence of BZD intake in MCI converters compared to non-converters in the univariate analysis. This association, however, lost its significance in the more meaningful multiple regression analysis, with concurs with studies finding no direct associations of BZD use and neuropathological changes such as amyloid-β deposits related to DAT [44]. On the other hand, groups have described a higher rate of BZD intake in patients before the development of DAT [45]. Again, these findings may be explained by the use of BZDs in treating prodromes of DAT such as anxiety, depression, and insomnia without being directly linked with BZD treatment itself.

Our findings also suggest that antidepressive treatment may lower the risk of conversion to DAT in MCI patients. Results of the regression analysis revealed that use of ATD at baseline went along with a significantly reduced conversion rate to DAT. Antidepressants may have a beneficial effect in patients with depression and cognitive deficits reflecting a possible “preclinical dementia” subtype of late-life depression. The relevance of this distinction is reflected in the diversity of MCI converters compared to non-converters in regard to ATD usage. The protective effect of ATDs may not only be related to current depression but may have an independent positive effect on cognitive functions in the elderly. Treating depression may have positive effects by reducing risk factors for DAT such as social and physical inactivity. In support of this, antidepressants have been shown to improve not only depression but also cognitive functions in the elderly and may also have limited neuroprotective effects [46, 47]. Furthermore, Kessing et al. have shown that a long-term treatment particularly with tricyclics is associated with a reduced risk of DAT in patients with severe depression [48].

Our study had several limitations. Firstly, we had no age-matched control group with solely geriatric depression and without MCI. Secondly, we cannot rule out that some MCI non-converters will convert to DAT in the following years or were not included into our retrospective analysis because of missing follow-up visit. However, the follow-up period of 3.2 years is relatively long and adding the variable time elapsed to follow-up to the list of predictors in the logistic regression analysis showed no significant influence on conversion in the multivariate model. Additionally, we have no information regarding the history and course of depression as well as on specifics of psychopharmacological treatment, such as for instance dose, length, and outcome of treatment. On the other hand, next to the considerable sample size, we consider the detailed clinical and neuropsychological examination of a representative MCI-patient sample a strength.

The patient group studied is of highly relevance in respect to the development of new and early therapeutic strategies for DAT. Based on our results, future studies should include especially severely depressed patients with cognitive deficits to specify clinical markers for preclinical dementia risk. To date, severe depression is frequently mentioned as an exclusion criterion in studies assessing prevention strategies or treatment for early DAT. Instead, we recommend to consider antidepressive treatment in MCI patients as possible strategy for delaying conversion to DAT.

DISCLOSURE STATEMENT

Authors’ disclosures available online (http://j-alz.com/manuscript-disclosures/16-1135r3).

References

Jack

Jr , Knopman

, Jagust

, Shaw

, Aisen

, Weiner

, Petersen

, Trojanowski

(2010) Hypothetical model of dynamic biomarkers of the Alzheimer’s pathological cascade. Lancet Neurol 9, 119–128.

Petersen

, Doody

, Kurz

, Mohs

, Morris

, Rabins

, Ritchie

, Rossor

, Thal

, Winblad

(2001) Current concepts in mild cognitive impairment. Arch Neurol 58, 1985–1992.

Petersen

, Smith

, Waring

, Ivnik

, Tangalos

, Kokmen

(1999) Mild cognitive impairment: Clinical characterization and outcome. Arch Neurol 56, 303–308.

Mitchell

, Shiri-Feshki

(2009) Rate of progression of mild cognitive impairment to dementia–meta-analysis of 41 robust inception cohort studies. Acta Psychiatr Scand 119, 252–265.

Middleton

, Yaffe

(2009) Promising strategies for the prevention of dementia. Arch Neurol 66, 1210–1215.

Cooper

, Li

, Lyketsos

, Livingston

(2013) Treatment for mild cognitive impairment: Systematic review. Br J Psychiatry 203, 255–264.

Ownby

, Crocco

, Acevedo

, John

, Loewenstein

(2006) Depression and risk for Alzheimer disease: Systematic review, meta-analysis, and metaregression analysis. Arch Gen Psychiatry 63, 530–538.

Devanand

, Sano

, Tang

, Taylor

, Gurland

, Wilder

, Stern

, Mayeux

(1996) Depressed mood and the incidence of Alzheimer’s disease in the elderly living in the community. Arch Gen Psychiatry 53, 175–182.

Ritchie

, Carriere

, Berr

, Amieva

, Dartigues

, Ancelin

, Ritchie

(2016) The clinical picture of alzheimer’s disease in the decade before diagnosis: Clinical and biomarker trajectories. J Clin Psychiatry 77, e305–311.

10.

Alexopoulos

, Kiosses

, Klimstra

, Kalayam

, Bruce

(2002) Clinical presentation of the “depression-executive dysfunction syndrome” of late life. Am J Geriatr Psychiatry 10, 98–106.

11.

Volkert

, Schulz

, Harter

, Wlodarczyk

, Andreas

(2013) The prevalence of mental disorders in older people in Western countries - a meta-analysis. Ageing Res Rev 12, 339–353.

12.

Koenig

, Blazer

(1992) Epidemiology of geriatric affective disorders. Clin Geriatr Med 8, 235–251.

13.

Herrmann

, Goodwin

, Ebmeier

(2007) The cognitive neuropsychology of depression in the elderly. Psychol Med 37, 1693–1702.

14.

Austin

, Mitchell

, Goodwin

(2001) Cognitive deficits in depression: Possible implications for functional neuropathology. Br J Psychiatry 178, 200–206.

15.

Bhalla

, Butters

, Mulsant

, Begley

, Zmuda

, Schoderbek

, Pollock

, Reynolds

3rd , Becker

(2006) Persistence of neuropsychologic deficits in the remitted state of late-life depression. Am J Geriatr Psychiatry 14, 419–427.

16.

Lee

, Potter

, Wagner

, Welsh-Bohmer

, Steffens

(2007) Persistent mild cognitive impairment in geriatric depression. Int Psychogeriatr 19, 125–135.

17.

Van der Mussele

, Fransen

, Struyfs

, Luyckx

, Marien

, Saerens

, Somers

, Goeman

, De Deyn

, Engelborghs

(2014) Depression in mild cognitive impairment is associated withprogression to Alzheimer’s disease: A longitudinal study. JAlzheimers Dis 42, 1239–1250.

18.

Barnes

, Alexopoulos

, Lopez

, Williamson

, Yaffe

(2006) Depressive symptoms, vascular disease, and mild cognitive impairment: Findings from the Cardiovascular Health Study. Arch Gen Psychiatry 63, 273–279.

19.

Johnson

, Hall

, O’Bryant

(2013) A depressive endophenotype of mild cognitive impairment and Alzheimer’s disease. PLoS One 8, e68848.

20.

, Wang

, Zhang

, Xu

, Gao

, Fang

, Yan

, Zhou

, Chongqing Ageing Study Group (2011) Vascular risk factors promote conversion from mild cognitive impairment to Alzheimer disease. Neurology 76, 1485–1491.

21.

Cooper

, Sommerlad

, Lyketsos

, Livingston

(2015) Modifiable predictors of dementia in mild cognitive impairment: A systematic review and meta-analysis. Am J Psychiatry 172, 323–334.

22.

Defrancesco

, Marksteiner

, Deisenhammer

, Kemmler

, Djurdjevic

, Schocke

(2013) Impact of white matter lesions and cognitive deficits on conversion from mild cognitive impairment to Alzheimer’s disease. J Alzheimers Dis 34, 665–672.

23.

Rosen

, Mohs

, Davis

(1984) A new rating scale for Alzheimer’s disease. Am J Psychiatry 141, 1356–1364.

24.

Reitan

(1955) The relation of the trail making test to organicbrain damage. J Consult Psychol 19, 393–394.

25.

Folstein

, Folstein

, McHugh

(1975) “Mini-mental state”. A practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res 12, 189–198.

26.

Yesavage

, Brink

, Rose

, Lum

, Huang

, Adey

, Leirer

(1982) Development and validation of a geriatric depressionscreening scale: A preliminary report. J PsychiatrRes 17, 37–49.

27.

MacCallum

, Zhang

, Preacher

, Rucker

(2002) On the practice of dichotomization of quantitative variables. Psychol Methods 7, 19–40.

28.

Morris

, Edland

, Clark

, Galasko

, Koss

, Mohs

, van Belle

, Fillenbaum

, Heyman

(1993) The consortium to establish a registry for Alzheimer’s disease (CERAD). IV. Rates of cognitive change in the longitudinal assessment of probable Alzheimer’s disease. Neurology 43, 2457–2465.

29.

Chen

, Hu

, Wei

, Qin

, McCracken

, Copeland

(2008) Severity of depression and risk for subsequent dementia: Cohort studies in China and the UK. Br J Psychiatry 193, 373–377.

30.

Teng

, Lu

, Cummings

(2007) Neuropsychiatric symptoms are associated with progression from mild cognitive impairment to Alzheimer’s disease. Dement Geriatr Cogn Disord 24, 253–259.

31.

Katon

, Lin

, Williams

, Ciechanowski

, Heckbert

, Ludman

, Rutter

, Crane

, Oliver

, Von Korff

(2010) Comorbid depression is associated with an increased risk of dementia diagnosis in patients with diabetes: A prospective cohort study. J Gen Intern Med 25, 423–429.

32.

Griffith

, Netson

, Harrell

, Zamrini

, Brockington

, Marson

(2006) Amnestic mild cognitive impairment: Diagnostic outcomes and clinical prediction over a two-year time period. J Int Neuropsychol Soc 12, 166–175.

33.

Tabert

, Manly

, Liu

, Pelton

, Rosenblum

, Jacobs

, Zamora

, Goodkind

, Bell

, Stern

, Devanand

(2006) Neuropsychological prediction of conversion to Alzheimer disease in patients with mild cognitive impairment. Arch Gen Psychiatry 63, 916–924.

34.

O’Bryant

, Humphreys

, Smith

, Ivnik

, Graff-Radford

, Petersen

, Lucas

(2008) Detecting dementia with themini-mental state examination in highly educated individuals. Arch Neurol 65, 963–967.

35.

Viticchi

, Falsetti

, Vernieri

, Altamura

, Altavilla

, Luzzi

, Bartolini

, Provinciali

, Silvestrini

(2014) ApolipoproteinE genotype and cerebrovascular alterations can influenceconversion to dementia in patients with mild cognitive impairment. J Alzheimers Dis 41, 401–410.

36.

Kim

, Basak

, Holtzman

(2009) The role of apolipoprotein E in Alzheimer’s disease. Neuron 63, 287–303.

37.

Paulson

, Igo

(2011) Genetics of dementia. Semin Neurol 31, 449–460.

38.

, Caracciolo

, Wang

, Santoni

, Winblad

, Fratiglioni

(2013) Accelerated progression from mild cognitive impairment to dementia among APOE epsilon4epsilon4 carriers. J Alzheimers Dis 33, 507–515.

39.

Masters

, Morris

, Roe

(2015) “Noncognitive” symptoms of early Alzheimer disease: A longitudinal analysis. Neurology 84, 617–622.

40.

Apostolova

, Cummings

(2008) Neuropsychiatric manifestations in mild cognitive impairment: A systematic review of the literature. Dement Geriatr Cogn Disord 25, 115–126.

41.

Beaulieu-Bonneau

, Hudon

(2009) Sleep disturbances in older adults with mild cognitive impairment. Int Psychogeriatr 21, 654–666.

42.

Mann

, Haastert

, Bohmdorfer

, Fruhwald

, Iglseder

, Roller-Wirnsberger

, Meyer

(2013) Prevalence and associationsof potentially inappropriate prescriptions in Austrian nursinghome residents: Secondary analysis of a cross-sectional study. Wien Klin Wochenschr 125, 180–188.

43.

Mah

, Binns

, Steffens

, Alzheimer’s Disease Neuroimaging Initiative (2015) Anxiety symptoms in amnestic mild cognitive impairment are associated with medial temporal atrophy and predict conversion to Alzheimer disease. Am J Geriatr Psychiatry 23, 466–476.

44.

Chung

, Nakajima

, Shinagawa

, Plitman

, Chakravarty

, Iwata

, Caravaggio

, Pollock

, Gerretsen

, Graff-Guerrero

, Alzheimer’s Disease Neuroimaging Initiative (2016) Benzodiazepine use attenuates cortical beta-amyloid and is not associated with progressive cognitive decline in nondemented elderly adults: A pilot study using F18-florbetapir positron emission tomography. Am J Geriatr Psychiatry 24, 1028–1039.

45.

Billioti de Gage

, Moride

, Ducruet

, Kurth

, Verdoux

, Tournier

, Pariente

, Begaud

(2014) Benzodiazepine use and risk of Alzheimer’s disease: Case-control study. BMJ 349, g5205.

46.

Caraci

, Copani

, Nicoletti

, Drago

(2010) Depression and Alzheimer’s disease: Neurobiological links and common pharmacological targets. Eur J Pharmacol 626, 64–71.

47.

Dranovsky

, Hen

(2006) Hippocampal neurogenesis: Regulation by stress and antidepressants. Biol Psychiatry 59, 1136–1143.

48.

Kessing

, Forman

, Andersen

(2011) Do continued antidepressants protect against dementia in patients with severe depressive disorder? Int Clin Psychopharmacol 26, 316–322.