The 2002 NIMH Provisional Diagnostic Criteria for Depression of Alzheimer’s Disease (PDC-dAD): Gauging their Validity over a Decade Later

Abstract

Presented herein is evidence for criterion, content, and convergent/discriminant validity of the NIMH-Provisional Diagnostic Criteria for depression of Alzheimer’s Disease (PDC-dAD) that were formulated to address depression in Alzheimer’s disease (AD). Using meta-analytic and systematic review methods, we examined criterion validity evidence in epidemiological and clinical studies comparing the PDC-dAD to Diagnostic and Statistical Manual of Mental Disorders fourth edition (DSM-IV), and International Classification of Disease (ICD 9) depression diagnostic criteria. We estimated prevalence of depression by PDC, DSM, and ICD with an omnibus event rate effect-size. We also examined diagnostic agreement between PDC and DSM. To gauge content validity, we reviewed rates of symptom endorsement for each diagnostic approach. Finally, we examined the PDC’s relationship with assessment scales (global cognition, neuropsychiatric, and depression definition) for convergent validity evidence. The aggregate evidence supports the validity of the PDC-dAD. Our findings suggest that depression in AD differs from other depressive disorders including Major Depressive Disorder (MDD) in that dAD is more prevalent, with generally a milder presentation and with unique features not captured by the DSM. Although the PDC are the current standard for diagnosis of depression in AD, we identified the need for their further optimization based on predictive validity evidence.

Keywords

Alzheimer’s disease depression diagnosis DSM epidemiology meta-analysis NIMH Provisional Diagnostic Criteria validity

INTRODUCTION

The National Institute of Mental Health (NIMH) Provisional Diagnostic Criteria for depression of Alzheimer’s Disease (PDC-dAD) were formulated over a decade ago to enable better recognition of depression specific for Alzheimer’s disease (AD) [1]. The depression of AD (dAD) was hypothesized to be phenotypically different from depressive disorders seen in older adults without dementia. The PDC were derived from the Diagnostic and Statistical Manual of Mental Disorders- fourth edition-text revised (DSM-IV-TR), and were described to differentiate dAD from Major Depressive Disorder (MDD) and minor depression [1 –4].

In 2002, there was recognition by a panel of 21 experts in dementia and late-life depression from geriatric medicine, geriatric psychiatry, neurology, and neuropsychiatry, that the severity of signs and symptoms of dAD are milder, and encompass a wider range of symptoms than MDD [1]. A workshop convened by the NIMH addressed the heterogeneity in the epidemiological data and developed a set of specific diagnostic criteria for dAD that might be used by regulatory authorities for evaluation of a therapeutic intervention. They also concluded that the differences between dAD and MDD possibly include incidence and prevalence, history of depression (patient or family), frequency and duration of symptoms, suicidality, outcomes, and therapeutic responses. The description of symptoms in the PDC that followed the panel’s findings was phrased to avoid possible confounding of symptoms due to AD such as diminished function, language, and cognitive abilities. Similar to MDD according to DSM-IV-TR, at least one of depressed mood or anhedonia (decreased positive affect or pleasure) was required for the PDC. However, the PDC dropped the MDD item of diminished ability to think or concentrate [3]. Additionally, differentiating the PDC criteria was the requirement that there be three symptoms for diagnosis of depression, without requiring that they be persistently present for a 2-week time. In the PDC criteria, irritability and social isolation/withdrawal were added items. The panel proposed diagnostic criteria using the same approaches as were used in development of the DSM with exception of a field trial. A two-step approach was undertaken with a first group drafting the criteria, and the second group revising them. Also, they listed a series of exclusion criteria, particularly one that precludes diagnosis when symptoms are due to other psychiatric conditions such as major depressive disorder, bipolar disorder, bereavement, schizophrenia, schizoaffective disorder, psychosis of Alzheimer disease, anxiety disorder, or substance-related disorder.

From PDC’s inception, experts have repeatedly advocated for examination of their validity [5, 6]. Additionally, dementia experts have stressed that because currently we have no therapies for defeating AD, the treatment goal remains maintaining patients’ quality of life. Better recognition of depression is an important part of treatment [7]. These two points informed our current objective, where via a systematic review and meta-analysis we conducted a comprehensive appraisal of the validity-evidence evolving for the PDC, in line with the proposed approach for examining validity-evidence for diagnostic criteria [8]. Specifically, we examined the criterion validity of the PDC by comparing prevalence of depression by the PDC with prevalence by other diagnostic criteria and by assessing the degree of diagnostic overlap between PDC and DSM. We evaluated content validity of the PDC by comparing rates of symptom endorsement with those by DSM. Finally, we evaluated convergent/discriminant-related validity by examining the relationship of the PDC with depression scales. To achieve as comprehensiveness a review as possible, evidence emerging from reviews and anecdotal notes was included where appropriate.

METHODS

For a comprehensive systematic review, a literature search was conducted to assess the number and type of published studies on the PDC. Medical electronic search engines (PubMed and EMBASE) were queried on first week of November 2013 and updated December 10, 2015 using the “provisional diagnostic criteria for depression of AD” as the key term, for independent original studies without language limitations. The “related citations” function in the PubMed was used as a crawler to examine for possible studies failing to meet our search criteria. For EMBASE, only the keyword search tool was used to carry the search. The bibliography section of the retrieved published papers meeting our a priori set selection criteria was subsequently examined for additional studies. Within the systematic review, we then defined stricter criteria for conducting a meta-analysis based on the Meta-analysis Of Observational Studies in Epidemiology guideline [9] and aimed at surveying the evidence for the prevalence of dAD. See Supplementary Table 1 for the selection criteria.

Data were extracted by the first author and subsequently independently reviewed by two research assistants for accuracy. For all eligible studies, inferential and descriptive statistics and frequency were retrieved. Median and frequency estimates [Event Rate (ER)] were calculated for depression according to the PDC and for other comparable diagnostic criteria and assessment scales. Subsequently, the frequency estimates were pooled to derive an omnibus effect-size estimate ER at 95% confidence intervals (CI), and to allow examination of the confounding a priori selected factors [age, sex, and Mini-Mental State Examination (MMSE) [10]] on the slope of the ER via meta-regression using the DerSimonian and Laird approach [11]. Comprehensive Meta-Analysis (CMA: Ver. 2.0) [12] served as the statistical platform for all meta-analytic evaluations and graphical presentations. We anticipated that a full examination of heterogeneity would not be possible because of anticipated differences in study design, population, and outcomes [13]; however, the corresponding Q-values and I-square (I²) values were investigated. Where appropriate and with available data, an aggregate measure in terms of Hedges’ g or Cohen’s d was generated for examination of the group differences in terms of diagnosis. Consistent with other meta-analysis [14], for the graphical representation and meta-regression analyses, the frequency rates at the 95% CI were transformed.

To compare depression frequency rates between the PDC, DSM, and International Classification of Diseases (ICD), using the binary data (e.g., PDC positive, PDC negative versus DSM positive, DSM negative) an Odds Ratio (effect-size) was calculated via the CMA for each comparison. For the aggregate overlap between PDC and DSM-MDD, an average was calculated for PDC alone, PDC+DSM, and DSM alone from the included studies, and subsequently a chart was developed to present the percentage of overlap between diagnostic criteria. Of note, the ICD defines a typical depression episode, as mild, moderate, and severe, with the obligatory presence of three main symptoms, most of every day for at least 2 weeks. These symptoms include depressed mood (persistent sadness or low mood), decreased positive affect or pleasure, and fatigue or loss of energy. Other additional symptoms of the ICD may include change in weight, sleep impairment, psychomotor change, worthlessness, concentration issues, and suicidal ideation [15]. For the assessment of the frequency of individual PDC symptoms, data were retrieved from graphical representations and subsequently compiled for analysis. For retrieval of the data from graphs, two authors (AAS and PEL) independently rated the graphs for reported frequencies and subsequently averaged them before inclusion into the final analysis. We examined the PDC in relation to assessment scales that emerged as a result of our search, including Cornell Scale for Depression in Dementia (CSDD), Geriatric Depression Scale (GDS), Neuropsychiatric Inventory-Depression subscale (NPI), Cambridge Examination for Mental Disorder of the Elderly (CAMDEX), and Hamilton rating scale for Depression (HAM-D). This evidence allowed further examination of the frequency of depressive symptoms, as well as gauging differences between those with and without depression by PDC.

To complement our comprehensive systematic review of the validity-evidence for the PDC, bibliographic sections of the emerging studies were examined for psychometric properties, reviews, and extra notes.

RESULTS

Description of the included studies

The search of the electronic literature yielded six observational studies on AD using a cross-sectional design. No studies using a longitudinal design were found. Four of these studies were independent allowing examination of the frequency of depression [16 –19], and one was a case-series [20]. From the six studies, three independent observational studies, but not the case-series, reported enough data to run an aggregate measure for assessment of diagnostic agreement [17 , 21]. Similarly, three studies reported sufficient data to run an aggregate measure on the PDC symptoms [16 , 19]. For the use of the PDC in clinical trials, an ad-hoc search using both controlled and uncontrolled vocabulary revealed three extra studies. Further examination of the literature revealed five reviews and one book chapter on depression in AD also discussing the PDC. Two additional studies emerged to examine the PDC in nursing home settings, but with a mixed sample. Table 1 shows the list of all publications available to date (N = 25) that used the PDC for diagnosis or review.

Table 1

Published seminal works (N = 25) examining the PDC in AD or dementia-spectrum

Studies	Observational	Comparative	Clinical	Review	Book
			trial		chapter
AD
Rosenberg et al., [20]^*	X
Vilalta-Franch et al., [18]	X
Teng et al., [17]	X
Barca et al., [21]	X
Engedal et al., [19]	X
Chiu et al., [16]	X
Nakaaki et al., [23]		X
Lee et al., [24]		X
Starkstein et al., [25, 26]^**			X
Rao et al., [27]			X
Munro et al., (for DIADS-2) [28]			X
Banerjee et al., (HTA-SADD) [29]			X
Lyketsos &Olin, [30]				X
Lee &Lyketsos, [5]				X
Amore et al., [31]				X
Starkstein et al., [6]				X
Panza et al., [32]				X
Geda et al., /Smith et al.,				X
(ISTAART) [22]
Richly et al., 2012 [33]					X
Dementia (mixed) or MCI
Leontjevas et al., [34]		X
Verkaik et al., [35]	X
Leontjevas et al., [36]		X
da Gloria Portugal et al., [37]	X
Li et al., [38]	X
Jeon et al., [39]	X

^*Case series; ^**Non-specific to PDC.

Prevalence of depression by PDC and DSM

Four independent observational (prospective and retrospective) studies reported the frequency for depression in AD including ‘possible’ and ‘probable’ AD, with MMSE scores ranging from 0 to 30, using more than one diagnostic approach. These studies were carried out in various settings (e.g., hospital, nursing home) with the majority of participants being female (range 55.8 to 70.9%), and with an average age ranging from 75.2 to 83 [16 –19]. From these four studies, the majority reported individuals with AD attending outpatient clinics, with an exception of one study [19], included patients from five nursing homes and one department of geriatric psychiatry and two hospitals of geriatric psychiatry. The calculated median frequency rates for depression with the PDC was 36.9% [16 –19], while by DSM-MDD was 13.7%. One study showed that the frequency of depression by DSM minor depression alone was 22% [17]. In three studies [16 , 19], the median frequency by the ICD was 17.5%. In three studies [16 –18], the median frequency by the Neuropsychiatric Inventory (depression or dysphoria [NPI-Q] endorsed) was 50%. Other measures were inconsistently used and reported frequency of 9.8 to 49.7%. Finally, in one study PDC-dAD showed a higher frequency (44%) when compared with the DSM-established cutoff scores for screening measures such as the CSDD (30%) and the GDS (33%), and a lower rate compared to the NPI-Q depression item (50%) [17]. In other words, data from the studies examining the validity of the PDC with AD samples showed that depression frequency is more than 2-fold higher compared with diagnosis based on DSM-MDD or the ICD criteria in the same AD sample. See Table 2.

Table 2

Description of the epidemiological studies examining PDC

	Location		Diagnostic criteria			Scales	Demographic
Studies	Country/Setting	Total	PDC-	DSM-Major	ICD	Depression scales	Age	Sex	MMSE (Mean)
		n	dAD (%)	depression (%)	(%)		(Mean)	(F%)	Dep/Non-dep PDC
Vilalta-Franch	Spain/Memory clinic	491	27.4	13.4	4.9	CAMDEX: 9.8	75.2	70.9	17.4/17
et al., [18]^♦						NPI-depression: 43.7
Teng et al.,	United State of	101	44	14	NA	CSDD: 30	77.5	55.8	21/21.2
^*♦	America/AD					GDS: 33
	Research centers					NPI-Q-dysphoria: 50
Barca et al.,	Norway/Hospitals	112	53.6	34.8	47.3	NR	83	69.6	10/11.1
[21]^^♦	&nursing home					NR
Engedal et al.,	Norway/Hospitals &	112	53.6	34.8	47.3	NR	83	69.6	10/11.1
[19]^*	nursing home
Chiu et al.,	Taiwan/Dementia	302	29.8	9.3	17.5	NPI-depression: 54	77	67.2	13/16.3
[16]^*	Clinic					HAM-D10 : 49.7
						HAM-D17 : 13.2

CAMDEX, Cambridge Examination for Mental Disorder of the Elderly; CSDD, Cornell Scale for Depression in Dementia; NPI, Neuropsychiatric Inventory; GDS, Geriatric Depression Scale; HAM-D, Hamilton rating scale for Depression; NA, not applicable; NR, not reported. MDD, Major Depressive Disorder; ^*Reports PDC symptoms endorsement. ^**Duplicate data with Engedal et al., [19]. ^♦Studies reporting data for diagnostic agreement.

Considering the random effect model, the omnibus ER estimate for PDC was 0.377 [∼38%, 95% CI = 0.272 to 0.494]. This ER was heterogeneous [Q-value = 33.988; p < 0.001; between study variability I² = 91.173%; N = 4], and it was consistent with the value of the calculated median frequency (36.9%). The near 100% overlap between the median and the estimated aggregate omnibus ER suggests that the heterogeneity in those studies may not be a significant factor to warrant further investigation.

The omnibus ER estimate for the DSM was 0.162 [∼16%, 95% CI = 0.088 to 0.277]. The heterogeneity estimates for this ER were similar to those for the PDC [Q-value = 39.416; p < 0.001; I² = 92.389]. A similar approach was undertaken to assess the frequency of depression by the ICD based on three studies: the aggregate ER estimate was 0.177 [∼18%, 95% CI = 0.047 to 0.485]. See Fig. 1.

Fig.1

Forest-Plot presenting the event rates. The circles represent the individual event-rates, and the diamond is the aggregate event rate for each sub-group.

Prevalence modifiers

Using the mixed effect regression model on the PDC data showed that age significantly affected the meta-regression slope [Slope = 0.1456; SE = 0.079; p < 0.001], suggesting that the frequency of dAD increases with age. Given the current data, it is not possible to tease apart the young-old from the old-old. The meta-regression with the mean MMSE scores for depressed as covariate showed an inverse relationship and was non-significant [Slope = –0.0290; SE = 0.0676; p = 0.0668]. Meta-regression of sex (percent overall female in each study) as covariate on Logit (log-odds) ER was non-significant [slope = –0.0230; SE = 0.0443; p = 0.6032]. Age alone was the only factor explaining 87.2% of the 91.17% between study variability. Meta-regression of age, sex, and MMSE scores was not possible for the DSM data, given the lack of reported data in collected studies. However, as per one study, there was no significant difference between DSM depressed and non-depressed on age and on MMSE scores [18].

Diagnostic agreement

The diagnostic agreement between DSM and PDC varied considerably both at the individual [20], and at the group level [18]. Three studies reported degree of overlap between diagnostic criteria (see Fig. 2) [17 , 21]. In one sample, 43.6% of participant (61/140 cases) fulfilled both diagnosis of depression by PDC and DSM-MDD, 3.5% satisfied only DSM criteria, and 53% met PDC criteria alone [18]. A later study yielded similar results, 35% of subjects met PDC criteria alone whereas 65% met both PDC and DSM criteria [21]. Another group reported that in their sample, 44% meet PDC criteria (44/101), 14% meet DSM-major depression and PDC (14/101), and 22% meet DSM-minor depression and PDC (22/101), but only 8% of them meeting PDC alone (8/101) [17]. Consistent with the review by Geda et al. [22] (Supplementary data, appendix A by Smith and colleagues for the Neuropsychiatric Syndromes Professional Interest Area of ISTAART), we found that the aggregate of the percent overlaps for MDD and PDC, shows that the PDC are more inclusive than the DSM criteria in all studies (see Fig. 2). However, regarding the validity of the PDC, when DSM minor depression was considered, the presence of patients meeting PDC alone was somewhat reduced, which in part provides support for the dAD construct intended to capture a wide spectrum of symptoms of depression, some are unique to AD. In other words, the evidence supports the construct validity of the PDC.

Fig.2

Venn diagram showing the aggregate percent overlaps between PDC and DSM-MDD based on three studies [17 , 21].

Symptom presentation in PDC and DSM

Based on three studies reporting on the PDC symptoms [16 , 19], a total of 194 dAD patients provided endorsement for the 10 items on the PDC. After calculating the frequency of symptom endorsement from the pooled sample, the core symptoms (Depressed mood or Decreased positive affect) were the most prevalent for PDC and DSM. As seen in Fig. 3, the “Feelings of worthlessness, hopelessness, or excessive or inappropriate guilt” was more prevalent for the PDC than for DSM. With exception to the newly added items (social isolation and irritability) all other symptoms in comparison to PDC were more prevalent for DSM. Also, psychomotor change seems to allow differentiation between the PDC and DSM given the large difference in the percent endorsement, compared to other symptoms, where depressed by DSM endorsed more psychomotorsymptoms.

Fig.3

PDC symptoms endorsement from observational studies (N = 3, n = 194) and comparable DSM symptoms (N = 1, n = 39) [16 , 19]. Note the elevated feelings of worthlessness for PDC in contrast to DSM, in comparison to other symptoms.

Suicidal ideation and suicidal act are prima facie evidence of clinical depression. The point prevalence for suicidal thoughts by DSM (41%) was higher than the PDC. Also, for suicidal ideation, as rated by informant nurses, a comparable magnitude of suicidal thoughts, as measured by the CSDD, between depressed and non-depressed subjects by DSM, ICD, or PDC, was observed [19]. However, there were important differences in the percentage of expressed suicidal ideation that met DSM criteria (31%), ICD criteria (23%), and PDC (20%). Psychiatrists rated a lower frequency of suicidal thoughts for their patients within the PDC (27%) and ICD (28%) compared to DSM (41%).

PDC and depression scales

Subjects meeting the PDC criteria have been reported to have a global NPI mean total score comparable to subjects diagnosed by DSM [PDC = 13.1, SD = 11.9; DSM = 14.3, SD = 12; Cohen’s d = 0.10] [18]. Similar results were observed for subjects not meeting either set of depression criteria [PDC no-depression = 9.4, SD = 11.1; DSM no-depression = 8.9, SD = 11; d = 0.04] [18]. Of note, the NPI assesses change in the presence, frequency, and severity of twelve neuropsychiatric symptoms for the period of four weeks, which is different from diagnostic approaches relying on presence of symptoms for the past two weeks. Also, the NPI only collects collateral information from the informant, which is a different method than from the diagnostic approach that corroborates clinical findings. For depression/dysphoria symptom, the presence or absence, or any uncertainties in the caregiver’s response or any inconsistencies between the response and other clinical information regarding sadness determines whether further exploration with eight subsequent items are required. Thus, the total NPI score would not be expected to differentiate different types of depression or severities, as it has only 1/12 of its total score coming fromdepression.

Additionally, significant group differences (depressed versus non-depressed) were observed for PDC on HAM-D [16, 19], NPI total score, and NPI caregiver burden, but not for DSM-MDD symptoms. For instance, one study found that PDC positives, but not the negatives, had significantly (p < 0.001) more depression symptoms (as assessed by HAM-D), elevated total neuropsychiatric score (mean score of 37.7 for depressed, and 8.2 for the non-depressed), and higher caregiver burden (mean score on NPI burden of 16.8 for the depressed, and 9.1 for the non-depressed) [16].

Psychometrically, using the CSDD, equal accuracy was observed by one study between diagnostic criteria, in that PDC showed on average 70% accuracy in comparison to 72% accuracy by DSM and 71.5% by the ICD in a sample of 112 AD patients [21]. Furthermore, at the low cutoff score of 6/7 on CSDD (meaning that a score of 7 and higher indicates a depressive disorder), the PDC showed substantially higher specificity (69%) than DSM (60%) or ICD (66%). Similarly, the PDC was shown to have higher specificity (85%) at the higher cutoff score of 9/10 in comparison to DSM (80%), and to ICD (83%) with a sample of ADpatients [21].

Now validated for the PDC, the Montgomery-Asberg Depression Scale (MADRS) and CSDD were shown to require lower cutoff scores, without loss of accuracy (cutoff score of 9 for MADRS, with sensitivity of 74.71% and specificity of 75%; and cutoff score of 11 for CSDD at the sensitivity of 80.39% and specificity of 66.67%) for PDC than DSM or ICD to allow optimal identification of dAD in 71 dementia patients. Additionally, larger area under the curve values were found for PDC than for DSM, using the CSDD or MADRS [38]. It is noteworthy that MADRS is a sensitive measure designed to detect change in the treatment and severity of a depression episode as per the DSM criteria in generalpopulation [40].

Furthermore, studies comparing PDC depressed to non-depressed on HAMD-D, CSDD, and MMSE scores found a large significance between group differences for HAM-D and CSDD (Cohen’s d = 1.3151 and 1.1154, respectively), but not for the MMSE score based on random effect model [Hedges’ g = –0.121; N = 4; 95% CI = –0.380 to 0.139; p = 0.363] [16, 19].

In aggregate, it appears that depression scales can detect dAD with high accuracy, though at lower cutoff points than for DSM-MDD. Also, group differences (depressed versus non-depressed) were observed for PDC but not for DSM-MDD, on HAM-D, NPI total score, and NPI caregiver burden. Herein, the validity-evidence (psychometric and group differences) supports PDC’s construct.

On functional assessment, two studies compared the PDC depressed to the non-depressed on activities of daily living [16, 19]. One study using the Physical self-maintenance scale found no significant difference between groups [19] whereas the other study used the Instrumental Activities of Daily Living scale and found poorer performance in thedepressed.

PDC, neuropsychology, and neuroimaging

Specific to AD, we have found two studies that provide limited external validity-evidence to the PDC-dAD; these contributed neuropsychological and neuroimaging findings comparing depressed to non-depressed.

The first study examined group differences via comprehensive neuropsychological assessment of 42 probable AD drug naïve patients (including 21 dAD, over the period of 2 years) [23]. The groups were balanced for age, education, gender, and duration of dementia and depression. They found no significant difference between PDC positive and PDC negative in terms of the scores on the Wechsler Adult Intelligence Scale-Revised, Wechsler Memory Scale-Revised (Logical memory I, Visual reproduction I), or the Rey Auditory Verbal Learning Test. In contrast to the PDC negatives, the PDC positives took a longer time to complete the Trail Making Test (TMT: Part B) (Bonferoni corrected Post-hoc test; p < 0.001, see Table 4 in the Nakaaki paper), a sensitive test of cognitive flexibility and divided attention. Subsequently, a significant difference was observed between groups in terms of the score obtained by subtracting the time taken to complete the TMT (Part A) from that taken for Part B (p < 0.001). Furthermore, consistent with others, reported in this manuscript earlier, they found no significant difference between AD groups on global cognition as assessed by the MMSE. In conclusion, the authors stated that the PDC positives, but not PDC negatives, showed greater impairment of attention and executive functions, specifically pertaining to divided attention. Additionally, they suggested that although the PDC were used in their study, PDC remains to be validated for use with AD patients.

A cross-sectional study used FDG-PET neuroimaging to show brain metabolic differences between 24 drug-naïve (for antidepressant) female PDC positive and PDC negative AD patients [24]. The groups were balanced for age, age of dementia onset, duration of dementia, education, and neurocognitive abilities, and consisted of patients with very mild and mild functional severity as assessed by clinical dementia rating scale. For assessment of cognition, eight neuropsychological tasks including verbal fluency, 15-item Boston Naming test, MMSE, word list memory, word list recall, word list recognition, constructional praxis, and constructional recall, using the Consortium to Establish a Registry for Alzheimer’s Disease battery, were used. This study showed that the PDC positives in contrast to the negatives had a statistically significant glucose hypometabolism in the right but not the left, superior frontal gyrus. In sum, the authors concluded that the frontal dysfunction is probably related to depressive symptoms.

The results of these studies seem consistent in terms of neuropsychological findings and share the same limitation. These studies have not compared PDC depressed to depressed by other diagnostic criteria, which limits our interpretation of them in light of PDC’s external validity, as defined by the extent to which the results can be generalized to other situations or people.

DISCUSSION

This is the first comprehensive quantitative review evaluating validity-evidence of the Provisional Diagnostic Criteria for dAD. We report a higher frequency of depression by the PDC compared to DSM-MDD, and ICD 10. This is both consistent with the aim of the NIMH convened panel of experts to capture depression in AD, and to have validity-evidence supporting the framework. The assumption is based on dAD being a distinct clinical entity, which includes milder forms of depression not recognized by standard criteria. The data on the diagnostic agreement between PDC and DSM similarly support PDC as a broader diagnostic construct which covers a wider spectrum of severity of depression in AD, thus resulting in a portion of patients only meeting these criteria, in studies that examined the overlap between the PDC and DSM-MDD. Of note, there is only a very low frequency of subjects evaluated as meeting DSM but not PDC criteria, which supports the criterion validity of the PDC. However, there exist a dilemma; although PDC have an inherent overlap with DSM-MDD, one of PDC’s exclusion criteria is that the symptoms are not better accounted for by, among other disorders, MDD. To our knowledge, differential diagnostic issues between dAD and MDD have not been systematically examined. As well, PDC were based upon DSM-MDD items. Thus, diagnostic overlap between the PDC and DSM, when most items are similar, is expected. In terms of individual symptoms, their frequency is similar across rating instruments within the PDC and DSM diagnostic construct with exception to feelings of worthlessness and psychomotor changes. It appears that the new symptoms, irritability and social withdrawal, which were added to the PDC are useful, in that only individuals with mild-moderate severity AD endorse them. Also, in terms of group differences (depressed versus non-depressed), PDC differed from DSM in terms of overall neuropsychiatric symptom burden, where PDC depressed showed more depressive symptoms, elevated total NPI score, and higher caregiver burden.

The psychometric evidence suggests that the PDC requires a lower cutoff point on well-established depression scales for dementia samples. At those cutoff points, they capture dAD with high accuracy. It remains to be seen whether the cases identified with the lower cutoff (false positives by DSM cutoff) truly suffer from depression. Evidence emerging from suicidal symptoms supports that the PDC, in aggregate, includes milder forms of clinical depression, with overall the lowest rate of suicidal ideation compared to other diagnostic criteria. It may be important to note here that non-demented normal older adults are less likely to verbalize suicidal thoughts and may experience death ideation, in the form of passive wish to die. This would explain the overall lower frequency in comparison to other diagnostic items. The finding of low rates of endorsement of suicidality by patients meeting the PDC may reflect lifespan changes in death ideation, or it may indicate depression experienced in parallel with diminished self-awareness and insight. In sum, the current validity study of the PDC, in line with what was suggested in the literature as validity evidence [8], has examined concurrent validators (including psychological tests), prevalence difference, delineations from other diagnostic criteria, and external validators.

Limitation of the PDC

There have been expressed concerns about the temporal specifications of the PDC [25]. In psychometric terms, the PDC’s inter-rater reliability, stability, and predictive validity are not determined and still need to be established. In addition, it is speculated that there may be subtypes of depression that PDC can identify but there are no data available to support such sub-classification. It is speculated that the PDC core criteria may not be suitable for AD patients with greater cognitive impairment given that depression symptoms may be masked by somatic complaints or not easily expressed. In severe AD, there is lower reporting of sad mood [33], one of the two core items needed for diagnosis of depression. While this observation may have been sample dependent, self-awareness or ability to rate one’s emotional state at a later stage of dementia is likely lower [41]. On the other hand, at the severe stage of AD, depressive symptoms such as sad mood tend to overlap with highly prevalent behavioral symptoms [42], suggesting that being overly concerned to extend the usability of any depression diagnostic criteria to this stage of illness maybe misplaced. Notwithstanding these limitations, the current review showed that depressed mood and decreased positive affect were the most commonly reported symptoms among those positively diagnosed with dAD.

Limitation of the study and current literature

Here we compared the symptoms of PDC with those of DSM-IV, or DSM-IV-TR but not to the ICD-9 due to shortage of data. Our search engines for this review were restricted to PubMed and EMBASE, which are the most elaborate medical search engines. Whether other databases would have added value to our search strategy is not known. Furthermore, within our search strategy, we omitted examination of depression as diagnosed by PDC in dementia spectrum or severity, something that could have provided added external validity given that there are only limited external validity criteria studies addressing the utility of the PDC. There is a specific gap in the literature related to some external validators (e.g., family history, history of the syndrome, and course of the illness). Additionally, although some scales have been suggested and validated for use with the PDC, no short and easy to administer scale has been validated for screening of the PDC-dAD in tertiary settings, other than for use in nursing homes [38]. Identifying PDC dAD patients who would benefit from pharmacological and non-pharmacological interventions remains largely unresolved. Consistent with the goal of the International Society to Advance Alzheimer’s Research and Treatment [22], it is recognizable that there is a need for neurobiological studies of the PDC-dAD in order to facilitate treatment development and implementation. Furthermore, the link between primary AD pathology and depression by the PDC remains to be clarified.

Future research direction

To date, only pharmacological compounds (e.g., selective serotonin reuptake inhibitors) are used for treatment of depression diagnosed by the PDC-dAD [27 –29], and no non-pharmacological approaches to management of dAD are reported. New compounds should be investigated applying the PDC and also neurobiological studies using the PDC may point to new compounds. In addition, it is important to acknowledge that treatment response may vary as a function of depression severity [43, 44], and that dAD severity needs to be determined to facilitate treatment selection. In line with previous work, even if predictors of these were known, different treatment approaches might need to be attempted before the optimum treatment is identified [45].

Moreover, in the current literature, there is no discussion of the quality of the evidence that substantiates the proposed PDC items. This limitation suggests the need for future examination of the evidence leading to refinement of specific PDC symptoms.

The finding that large portions of individuals meeting PDC also meet DSM-MDD criteria means that these individuals are significantly affected by depression. The finding that a large portion meets PDC but not DSM criteria also has criterion validity because the PDC were developed to capture dAD, a more prevalent depressive entity [16–19 , 21], distinct from DSM-MDD (or from late-life depression as the Neuropsychiatry Symptoms Interest group of the Alzheimer’s Association suggests) [46]. However important questions remain regarding individuals only meeting PDC in relation to those meeting only other sets of diagnostic criteria. The evidence accrued in this study indicates that when minor and major depression criteria are combined, the degree of overlap increases. This suggests that a fraction of individuals remain diagnosed by the PDC alone, and we cannot infer whether they are depressed or otherwise. From the psychometric data, the PDC+/DSM- people seem to suffer a softer depression given that we observed: 1) the non-significant differences for global cognition between depressed and non-depressed; 2) overall frequency of depressive symptoms endorsed; and 3) lower rate of suicidal ideation. However, it would be desirable to have evidence of scores from commonly used depression scales, not examined to date, and their cutoff for the PDC for further validation. Promising research avenues would be head-to-head examination of the PDC versus DSM studies in AD where criterion validity is examined concurrently via determination of quality of life and treatment response. It is also currently of considerable interest given the publication of new diagnostic criteria by the International Working Group and by the National Institute on Aging working group commissioned by the Alzheimer’s Association that dAD evolves as a function of disease stage, from prodromal AD/MCI to full-blown dementia, and that a revision of the PDC will be needed to accommodate the non-dementia stages of AD [47 –49]. The emergence of the DSM-V diagnostic approach to cognitive impairment should be taken into consideration, given that the PDC relies on diagnosis of dementia by DSM. The DSM-V dementia criteria, major Neurocognitive Disorder (NCD) due to AD is specific to dementia and not to depression diagnosis co-existing with dementia. These new DSM criteria acknowledge depression symptoms, but not diagnosis of depression that co-exists in AD. The general descriptor for NCD can be behavioral (mood), but to the extent that it may not be clinically significant. Moreover, the DSM specifier, Major NCD due to AD requires significant cognitive decline (2 or more cognitive domains), not due to other mental disorder, that interferes with the person’s independence.

Given current status of the PDC in the literature, future studies investigating the PDC could: 1) further examine group differences between depressed and non-depressed on functional status; 2) allow differentiation between young-old and the old-old on PDC symptomology; and 3) examine the difference between rates of depression in institutions versus at nursing homes. Additionally, because of the distinctiveness of depression in AD in relation to late-life major depression [22, 46] and because of the inclusivity and breadth of the PDC, studies are needed to promote a better understanding of the underlying etiology.

The PDC, like other criteria sets, use a categorical approach to the diagnosis of depression. Given the recent interest in depression assessment using a dimensional approach, we acknowledge that there may be value in investigating this approach in the assessment of dAD. This would require a dimensional definition of the dAD construct and the development of appropriate continuous measures. The advantage of this approach could be a more fine-grained characterization of dAD symptomatology, improved insight into the pathophysiology of dAD, and potentially more individualized treatments. This could also enhance patient and caregiver satisfaction because of greater consideration of their perception of symptoms. This in turn ultimately holds promise for improving treatment outcomes. Such an undertaking would be in line with the new Research Domain Criteria by the NIMH [50] (https://www.nimh.nih.gov/research-priorities/rdoc/index.shtml).

Moreover, since the development of the new provisional diagnostic algorithm, the Mild Behavioral Impairment (MBI) [51], that is broader and captures a range of syndromes, further examination of the PDC-dAD, and whether the PDC-dAD can be captured by the MBI, or that the PDC can work with MCI and allow better differential diagnosis at the prodromal stage is warranted.

Conclusion

The published evidence supports the criterion-,content-, and convergent validity of the PDC but dimensions of reliability and predictive validity remain to be determined. Although not a globally accepted standard, PDC is a set of diagnostic criteria that is clinically useful for identifying depression in AD. Similar to the DSM-MDD that went over several mitigations/changes/revisions, the current PDC-dAD can now be utilized for research and clinical purpose until further modification or future studies to suggest otherwise. For diagnosis of depression in AD, the PDC will allow those currently unable to reach diagnostic threshold by either the DSM-MDD or ICD criteria for depression in order to obtain optimal treatment, now to get one diagnosis and treatment. Although, given the differences between DSM-MDD and PDC-dAD, it is important to assess the efficacy of treatment for dAD, as others highlighted [20], with large sample size. With this validity-evidence in mind, we recognized that the PDC are needed to define depression as a disorder in AD (and not just a specifier). Whether PDC can be used for dementia-spectrum remains to be determined.

Footnotes

ACKNOWLEDGMENTS

Part of this work has been presented in preliminary form at the 65th Annual Meeting of the American Academy of Neurology, March 15–23, 2013, San Diego, and published as part of a PhD dissertation. The authors wish to extend gratitude to the research assistants Hae Jung Min and Peter Chan, who revised the data independently for accuracy.

Authors’ disclosures available online ().

Appendix

References

Olin

, Schneider

, Katz

, Meyers

, Alexopoulos

, Breitner

, Bruce

, Caine

, Cummings

, Devanand

, Krishnan

, Lyketsos

, Lyness

, Rabins

, Reynolds

3rd , Rovner

, Steffens

, Tariot

, Lebowitz

(2002) Provisional diagnostic criteria for depression of Alzheimer disease. Am J Geriatr Psychiatry 10, 125–128.

Olin

, Katz

, Meyers

, Schneider

, Lebowitz

(2002) Provisional diagnostic criteria for depression of Alzheimer disease: Rationale and background. Am J Geriatr Psychiatry 10, 129–141.

Olin

, Schneider

, Katz

, Meyers

, Alexopoulos

, Breitner

, Bruce

, Caine

, Cummings

, Devanand

, Jeste

, Krishnan

, Lyketsos

, Lyness

, Rabins

, Iii

, Rovner

, Steffens

, Unutzer

, Lebowitz

(2003) Provisional diagnostic criteria for depression of Alzheimer’s disease: Description and review. Expert Rev Neurother 3, 99–106.

American Psychiatric Association (2000) Diagnostic and Statistical Manual of Mental Disorders (DSM)- Fourth edition, Text revised, American Psychiatric Association, Washington, DC.

Lee

, Lyketsos

(2003) Depression in Alzheimer’s disease: Heterogeneity and related issues. Biol Psychiatry 54, 353–362.

Starkstein

, Mizrahi

, Power

(2008) Depression in Alzheimer’s disease: Phenomenology, clinical correlates and treatment. Int Rev Psychiatry 20, 382–388.

Winblad

, Amouyel

, Andrieu

, Ballard

, Brayne

, Brodaty

, Cedazo-Minguez

, Dubois

, Edvardsson

, Feldman

, Fratiglioni

, Frisoni

, Gauthier

, Georges

, Graff

, Iqbal

, Jessen

, Johansson

, Jonsson

, Kivipelto

, Knapp

, Mangialasche

, Melis

, Nordberg

, Rikkert

, Qiu

, Sakmar

, Scheltens

, Schneider

, Sperling

, Tjernberg

, Waldemar

, Wimo

, Zetterberg

(2016) Defeating Alzheimer’s disease and other dementias: A priority for European science and society. Lancet Neurol 15, 455–532.

Kendell

, Jablensky

(2003) Distinguishing between the validity and utility of psychiatric diagnoses. Am J Psychiatry 160, 4–12.

Stroup

, Berlin

, Morton

, Olkin

, Williamson

, Rennie

, Moher

, Becker

, Sipe

, Thacker

(2000) Meta-analysis of observational studies in epidemiology: A proposal for reporting. Meta-analysis Of Observational Studies in Epidemiology (MOOSE) group. JAMA 283, 2008–2012.

10.

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.

11.

DerSimonian

, Laird

(1986) Meta-analysis in clinical trials. Control Clin Trials 7, 177–188.

12.

Borenstein

, Hedges

, Higgins

, Rothstein

(2005) Comprehensive meta-analysis [Version 2.0]. Biostat, Englewood, NJ.

13.

Berlin

(1995) Invited commentary: Benefits of heterogeneity in meta-analysis of data from epidemiologic studies. Am J Epidemiol 142, 383–387.

14.

Rutledge

, Reis

, Linke

, Greenberg

, Mills

(2006) Depression in heart failure a meta-analytic review of prevalence, intervention effects, and associations with clinical outcomes. J Am Coll Cardiol 48, 1527–1537.

15.

WHO (1992) The ICD-10 Classification of Mental and Behavioural Disorders: Clinical descriptions and diagnostic guideline. In WHO, ed. World Health Organization, Geneva, p. 100 http://www.who.int/classifications/icd/en/bluebook.pdf.

16.

Chiu

, Steffens

, Chen

, Hsu

, Huang

, Lai

(2012) Depression in Taiwanese patients with Alzheimer’s disease determined by the National Institutes of Mental Health Provisional Criteria. Int Psychogeriatr 24, 1299–1305.

17.

Teng

, Ringman

, Ross

, Mulnard

, Dick

, Bartzokis

, Davies

, Galasko

, Hewett

, Mungas

, Reed

, Schneider

, Segal-Gidan

, Yaffe

, Cummings

(2008) Diagnosing depression in Alzheimer disease with the national institute of mental health provisional criteria. Am J Geriatr Psychiatry 16, 469–477.

18.

Vilalta-Franch

, Garre-Olmo

, Lopez-Pousa

, Turon-Estrada

, Lozano-Gallego

, Hernandez-Ferrandiz

, Pericot-Nierga

, Feijoo-Lorza

(2006) Comparison of different clinical diagnostic criteria for depression in Alzheimer disease. Am J Geriatr Psychiatry 14, 589–597.

19.

Engedal

, Barca

, Laks

, Selbaek

(2011) Depression in Alzheimer’s disease: Specificity of depressive symptoms using three different clinical criteria. Int J Geriatr Psychiatry 26, 944–951.

20.

Rosenberg

, Onyike

, Katz

, Porsteinsson

, Mintzer

, Schneider

, Rabins

, Meinert

, Martin

, Lyketsos

, Depression of Alzheimer’s Disease Study (2005) Clinical application of operationalized criteria for ‘Depression of Alzheimer’s Disease’. Int J Geriatr Psychiatry 20, 119–127.

21.

Barca

, Engedal

, Selbaek

(2010) A reliability and validity study of the cornell scale among elderly inpatients, using various clinical criteria. Dement Geriatr Cogn Disord 29, 438–447.

22.

Geda

, Schneider

, Gitlin

, Miller

, Smith

, Bell

, Evans

, Lee

, Porsteinsson

, Lanctôt

, Rosenberg

, Sultzer

, Francis

, Brodaty

, Padala

, Onyike

, Ortiz

, Ancoli-Israel

, Bliwise

, Martin

, Vitiello

, Yaffe

, Zee

, Herrmann

, Sweet

, Ballard

, Khin

, Alfaro

, Murray

, Schultz

, Lyketsos

(2013) Neuropsychiatric symptoms in Alzheimer’s disease: Past progress and anticipation of the future. Alzheimers Dement 9, 602–608.

23.

Nakaaki

, Murata

, Sato

, Shinagawa

, Tatsumi

, Hirono

, Furukawa

(2007) Greater impairment of ability in the divided attention task is seen in Alzheimer’s disease patients with depression than in those without depression. Dement Geriatr Cogn Disord 23, 231–240.

24.

Lee

, Choo

, Jhoo

, Kim

, Youn

, Lee

, Kang

, Lee

, Kang

, Woo

(2006) Frontal dysfunction underlies depressive syndrome in Alzheimer disease: A FDG-PET study. Am J Geriatr Psychiatry 14, 625–628.

25.

Starkstein

, Jorge

, Mizrahi

, Robinson

(2005) The construct of minor and major depression in Alzheimer’s disease. Am J Psychiatry 162, 2086–2093.

26.

Starkstein

, Mizrahi

, Garau

(2005) Specificity of symptoms of depression in Alzheimer disease: A longitudinal analysis. Am J Geriatr Psychiatry 13, 802–807.

27.

Rao

, Spiro

, Rosenberg

, Lee

, Rosenblatt

, Lyketsos

(2006) An open-label study of escitalopram (Lexapro) for the treatment of ‘Depression of Alzheimer’s disease’ (dAD). Int J Geriatr Psychiatry 21, 273–274.

28.

Munro

, Longmire

, Drye

, Martin

, Frangakis

, Meinert

, Mintzer

, Porsteinsson

, Rabins

, Rosenberg

, Schneider

, Weintraub

, Lyketsos

(2012) Cognitive outcomes after sertaline treatment in patients with depression of Alzheimer disease. Am J Geriatr Psychiatry 20, 1036–1044.

29.

Banerjee

, Hellier

, Dewey

, Romeo

, Ballard

, Baldwin

, Bentham

, Fox

, Holmes

, Katona

, Knapp

, Lawton

, Lindesay

, Livingston

, McCrae

, Moniz-Cook

, Murray

, Nurock

, Orrell

, O’Brien

, Poppe

, Thomas

, Walwyn

, Wilson

, Burns

(2011) Sertraline or mirtazapine for depression in dementia (HTA-SADD): A randomised, multicentre, double-blind, placebo-controlled trial. Lancet 378, 403–411.

30.

Lyketsos

, Olin

(2002) Depression in Alzheimer’s disease: Overview and treatment. Biol Psychiatry 52, 243–252.

31.

Amore

, Tagariello

, Laterza

, Savoia

(2007) Subtypes of depression in dementia. (Suppl 1). Arch Gerontol Geriatr 44, 23–33.

32.

Panza

, Frisardi

, Capurso

, D’Introno

, Colacicco

, Imbimbo

, Santamato

, Vendemiale

, Seripa

, Pilotto

, Capurso

, Solfrizzi

(2010) Late-life depression, mild cognitive impairment, and dementia: Possible continuum? Am J Geriatr Psychiatry 18, 98–116.

33.

Richly

, Manes

, Bustin

(2012) Depression and Alzheimer’s disease. In Depression in Neurologic Disorders: Diagnosis and Management, Kanner

, ed. Wiley, Somerset, NJ, pp. 177–188.

34.

Leontjevas

, van Hooren

, Mulders

(2009) The Montgomery-Asberg Depression Rating Scale and the Cornell Scale for Depression in Dementia: A validation study with patients exhibiting early-onset dementia. Am J Geriatr Psychiatry 17, 56–64.

35.

Verkaik

, Francke

, van Meijel

, Ribbe

, Bensing

(2009) Comorbid depression in dementia on psychogeriatric nursing home wards: Which symptoms are prominent? Am J Geriatr Psychiatry 17, 565–573.

36.

Leontjevas

, Gerritsen

, Vernooij-Dassen

, Smalbrugge

, Koopmans

(2012) Comparative validation of proxy-based Montgomery-Asberg depression rating scale and cornell scale for depression in dementia in nursing home residents with dementia. Am J Geriatr Psychiatry 20, 985–993.

37.

da Gloria Portugal

, Coutinho

, Almeida

, Barca

, Knapskog

, Engedal

, Laks

(2012) Validation of Montgomery-Asberg Rating Scale and Cornell Scale for Depression in Dementia in Brazilian elderly patients. Int Psychogeriatr 24, 1291–1298.

38.

, Jeon

, Low

, Chenoweth

, O’Connor

, Beattie

, Brodaty

(2015) Validity of the geriatric depression scale and the collateral source version of the geriatric depression scale in nursing homes. Int Psychogeriatr 27, 1495–1504.

39.

Jeon

, Li

, Low

, Chenoweth

, O’Connor

, Beattie

, Liu

, Brodaty

(2015) The clinical utility of the Cornell Scale forDepression in Dementia as a routine assessment in nursing homes. Am J Geriatr Psychiatry 23, 784–793.

40.

Montgomery

, Asberg

(1979) A new depression scale designed to be sensitive to change. Br J Psychiatry 134, 382–389.

41.

Feher

, Larrabee

, Crook

3rd (1992) Factors attenuating the validity of the Geriatric Depression Scale in a dementia population. J Am Geriatr Soc 40, 906–909.

42.

Thorpe

(2009) Depression vs. dementia: How do we assess? Canad Rev Alzheimers Dis Other Demen 12, 17–22.

43.

Bruce

, Ten Have

, Reynolds

3rd , Katz

, Schulberg

, Mulsant

, Brown

, McAvay

, Pearson

, Alexopoulos

(2004) Reducing suicidal ideation and depressive symptoms in depressed older primary care patients: A randomized controlled trial. JAMA 291, 1081–1091.

44.

Alexopoulos

, Reynolds

3rd , Bruce

, Katz

, Raue

, Mulsant

, Oslin

, Ten Have

, Group

(2009) Reducing suicidal ideation and depression in older primary care patients: 24-month outcomes of the PROSPECT study. Am J Psychiatry 166, 882–890.

45.

Sepehry

, Lee

, Hsiung

GYR

, Jacova

(2012) Stay the course-is it justified? Lancet 379, 220–220.

46.

Nowrangi

, Lyketsos

, Rosenberg

(2015) Principles and management of neuropsychiatric symptoms in Alzheimer’s dementia. Alzheimers Res Ther 7, 12.

47.

Dubois

, Feldman

, Jacova

, Dekosky

, Barberger-Gateau

, Cummings

, Delacourte

, Galasko

, Gauthier

, Jicha

, Meguro

, O’Brien

, Pasquier

, Robert

, Rossor

, Salloway

, Stern

, Visser

, Scheltens

(2007) Research criteria for the diagnosis of Alzheimer’s disease: Revising the NINCDS-ADRDA criteria. Lancet Neurol 6, 734–746.

48.

Albert

, DeKosky

, Dickson

, Dubois

, Feldman

, Fox

, Gamst

, Holtzman

, Jagust

, Petersen

, Snyder

, Carrillo

, Thies

, Phelps

(2011) The diagnosis of mild cognitive impairment due to Alzheimer’s disease: Recommendations from the National Institute on Aging-Alzheimer’s Association workgroups on diagnostic guidelines for Alzheimer’s disease. Alzheimers Dement 7, 270–279.

49.

McKhann

, Knopman

, Chertkow

, Hyman

, Jack

Jr , Kawas

, Klunk

, Koroshetz

, Manly

, Mayeux

, Mohs

, Morris

, Rossor

, Scheltens

, Carrillo

, Thies

, Weintraub

, Phelps

(2011) The diagnosis of dementia due to Alzheimer’s disease: Recommendations from the National Institute on Aging-Alzheimer’s Association workgroups on diagnostic guidelines for Alzheimer’s disease. Alzheimers Dement 7, 263–269.

50.

Sharp

, Fowler

, Salas

, Nielsen

, Allen

, Oldham

, Kosten

, Mathew

, Madan

, Frueh

, Fonagy

(2016) Operationalizing NIMH Research Domain Criteria (RDoC) in naturalistic clinical settings. Bull Menninger Clin 80, 187–212.

51.

Ismail

, Smith

, Geda

, Sultzer

, Brodaty

, Smith

, Aguera-Ortiz

, Sweet

, Miller

, Lyketsos

, Area INSPI (2016) Neuropsychiatric symptoms as early manifestations of emergent dementia: Provisional diagnostic criteria for mild behavioral impairment. Alzheimers Dement 12, 195–202.

52.

McKhann

, Drachman

, Folstein

, Katzman

, Price

, Stadlan

(1984) Clinical diagnosis of Alzheimer’s disease: Report of the NINCDS-ADRDA Work Group under the auspices of Department of Health and Human Services Task Force on Alzheimer’s Disease. Neurology 34, 939–944.