SEARCH

SEARCH BY CITATION

Keywords:

  • cognitive impairment;
  • dementia;
  • epidemiology;
  • old-age;
  • diagnosis

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Material and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Declaration of interest
  9. References

Objective

Progression from cognitive impairment (CI) to dementia is predicted by several factors, but their relative importance and interaction are unclear.

Method

We investigated numerous such factors in the AgeCoDe study, a longitudinal study of general practice patients aged 75+. We used recursive partitioning analysis (RPA) to identify hierarchical patterns of baseline covariates that predicted dementia-free survival.

Results

Among 784 non-demented patients with CI, 157 (20.0%) developed dementia over a follow-up interval of 4.5 years. RPA showed that more severe cognitive compromise, revealed by a Mini-Mental State Examination (MMSE) score < 27.47, was the strongest predictor of imminent dementia. Dementia-free survival time was shortest (mean 2.4 years) in such low-scoring patients who also had impaired instrumental activities of daily living (iADL) and subjective memory impairment with related worry (SMI-w). Patients with identical characteristics but without SMI-w had an estimated mean dementia-free survival time of 3.8 years, which was still shorter than in patients who had subthreshold MMSE scores but intact iADL (4.2–5.2 years).

Conclusion

Hierarchical patterns of readily available covariates can predict dementia-free survival in older general practice patients with CI. Although less widely appreciated than other variables, iADL impairment appears to be an especially noteworthy predictor of progression to dementia.

Significant outcomes
  • Combinations of five readily available factors – Mini-Mental State Examination (MUSE) score, performance in instrumental activities of daily living (iADL), subjective memory impairment, age, and type of cognitive deficit – were able to predict different dementia-free survival times in older general practice patients with cognitive impairment.
  • This information may be used for risk assessment and monitoring in clinical practice and research.
  • Older people who have low MMSE scores along with iADL impairment and subjective memory impairment with related worry should be closely monitored for cognitive changes, as they showed the highest risk of progression to dementia and the shortest mean dementia-free survival time.
Limitations
  • Participants were recruited by general practitioners (GP). Even though the vast majority of German elderly people have regular contact with a GP, generalizability of our sample may be limited.
  • Generalizability of our results may be limited also because of the study's relatively low response rate.
  • We used instruments chosen for their applicability to epidemiological studies. Such instruments are necessarily brief and may lack either sensitivity or specificity when compared with more extensive clinical and neuropsychological examinations.

Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Material and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Declaration of interest
  9. References

Dementia is among the most significant public health issues worldwide with huge economic costs [1, 2]. In 2010 alone, its worldwide costs were estimated at US$ 604 billion [1], and the number of people affected will likely increase from 24.3 million in the year 2001 to 81.1 million by 2040 [3].

Dementia – particularly Alzheimer's dementia – is often preceded by a transitional stage of subsyndromal cognitive impairments (CIs) [4]. Annual conversion rates from such impairments to dementia are approximately 5–10%, but there are also substantial numbers of cognitively impaired people who do not progress to dementia even after 10 years of follow-up [meta-analysis of 41 cohort studies [5]]. Therefore, there is strong interest in the identification of factors that can predict which persons with CI are most likely to progress (‘convert’) to dementia (e.g., [6-14]). These factors include combinations of different indicators (biomarkers, cognitive markers, sociodemographic factors, co morbidity, etc.) that may predict conversion to dementia.

A common outcome measure in such studies is the proportion of cognitively impaired individuals who progress to dementia within a given interval. A flexible alternative is measurement of the time elapsed to dementia onset. Knowledge about time to dementia onset and associated determinants might not only help patients and their relatives to prepare better for possible consequences of the disease (e.g., to discuss advance directives or care arrangements), but could be also used in clinical practice for risk assessment, patient monitoring, and planning more appropriate choice among available symptomatic treatment options (e.g., NMDA receptor antagonists).

Aims of the study

We sought to identify hierarchical patterns of predictors for dementia-free survival in a sample of elderly (75+ years) general practice patients with cognitive impairment. For practical reasons, we focused on readily available clinical measures rather than biomarkers requiring invasive or costly laboratory techniques.

Material and methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Material and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Declaration of interest
  9. References

Sample

We studied participants in the longitudinal German Study on Ageing, Cognition and Dementia in Primary Care Patients (AgeCoDe). These individuals were recruited by general practitioners (GP) near to six centers (Hamburg, Bonn, Düsseldorf, Leipzig, Mannheim, and Munich). The GP identified 22 701 patients aged 75 years or older. Of these, they referred 10 850 who they believed to be dementia-free and who met other eligibility criteria (Fig. 1). Eligible patients had at least one office visit in the prior 12 months, were not residents of nursing homes, and did not have an illness deemed likely to be fatal within 3 months. Deaf or blind patients or those who lacked sufficient facility in German were excluded. The protocol was approved by the ethics committees of all collaborating centers, and all participants provided written informed consent.

image

Figure 1. Sample attrition and sample.

Download figure to PowerPoint

Among a randomly selected sample of 6619 of these referred patients, 3327 individuals participated in the study. These respondents were slightly younger than those who refused (M = 80.1/SD = 3.6 years vs. 80.8/3.8; = −6.607, < 0.001) and included an increased proportion of men (34.5% vs. 31.1%; χ² = 6.028, d.f. = 1, = 0.01). Baseline assessments subsequently showed that 109 respondents had dementia or were younger than age 75 years. Another 20 participants did not complete their baseline assessment or were lost to follow up without information regarding a subsequent onset of dementia, leaving an analysis pool of 3198 persons (Fig. 1).

Data collection and assessment procedures

Data collection began on January 23, 2003. Baseline and subsequent follow-up assessments were performed in patients' homes by trained physicians and psychologists. The main assessment instrument was the Structured Interview for Diagnosis of Dementia of Alzheimer type, Multi-infarct Dementia and Dementia of other Aetiology according to DSM-III-R, DSM-IV, and ICD-10 (SIDAM) [15, 16]. The SIDAM includes a 55-item test battery that covers several domains of cognitive function (orientation, memory, abstract reasoning, verbal abilities and calculation, constructional abilities, aphasia, and apraxia). The battery includes the 30-item Mini-Mental State Examination (MMSE) [17] – a well-established brief assessment of cognitive status that may be also used for prediction of dementia survival [18]. Also included are a third-party rating of psychosocial impairment with a 14-item scale for the assessment of activities of daily living (SIDAM ADL Scale) and a summary clinical diagnostic impression.

Being dementia-free at baseline, participants could not have cognitive deficits sufficient to impair basic ADL, but some nonetheless had difficulties with instrumental ADL (iADL, for example, using the telephone or handling routine finances). We assessed the latter using the iADL Scale of Lawton and Brody [19]. We identified depressive features using the 15-item version of the Geriatric Depression Scale [20] with a cutoff ≥6 [21]. Subjective memory impairment (SMI) was evaluated prior to cognitive testing by asking two questions: ‘Do you feel as if your memory is becoming worse?’(yes/no); and if yes, ‘Does this worry you?’ (yes/no). We thereby categorized subjects into three groups: no SMI; SMI but without related worry; and SMI with related worry (SMI-w). A standardized interview provided information on sociodemographic characteristics, impairment in hearing and vision, medications, smoking history, alcohol consumption, and family history of dementia. Alcohol consumption was classified according to guidelines of the World Health Organization [22]. Patients were divided into non-smokers and former and current smokers. Participants' GP completed questionnaires about medical diagnoses and took blood samples for genetic analysis.

We used essentially identical procedures at three follow-up waves separated by 18 months. The final assessment was completed on November 13, 2009. If a participant had died in the interim or was otherwise unavailable, we obtained information about his/her cognitive status by interviewing an informant (usually a relative) using the structured Global Deterioration Scale [23] and ‘Changes in Performance of Everyday Activities’ and ‘Changes in Habits’ subscales from the Blessed Dementia Scale [24]. Dates of death were obtained from relatives or from official registry offices.

Diagnostic groups

Both at baseline (for exclusion) and at subsequent evaluations, dementia status [DSM-IV criteria; [25]] was agreed at consensus conferences that included the interviewer and an experienced geriatrician or geriatric psychiatrist. If SIDAM results were unavailable, we assigned follow-up dementia diagnoses when ratings were ≥4 on the Global Deterioration Scale [23] and/or  >8 on the Blessed Dementia Rating subscales [24].

Among the dementia-free patient sample, we identified individuals who were unimpaired in basic ADL but whose SIDAM cognitive test battery yielded one or more domain scores that fell below age- and education-specific norms by at least one standard deviation [26]. This operationally defined category of patients with cognitive impairment (CI) encompassed all participants who met current diagnostic criteria for mild cognitive impairment [MCI; [27]], viz.

  1. Evidence of cognitive decline
  2. Not demented according to DSM-IV criteria
  3. Preserved basic activities of daily living, with little or no impairment in complex instrumental functions.

Some subjects, however, did not meet all criteria for MCI because the latter's criterion for ‘evidence of cognitive decline’ typically requires SMI (a ‘complaint’), as well as objective evidence of impairment and not all our CI patients had SMI. For present purposes, we included these individuals as well as those who met full criteria for MCI to encompass all patients with CI. By contrast, we did follow the usual MCI approach of differentiating patients' cognitive syndromes into four groups: those with a single-domain amnestic deficit; those with a relatively isolated deficit in a non-memory domain (single non-amnestic deficit); those with impairment in at least two domains that included a clear deficit in memory (multidomain amnestic deficit); and those with multiple impairments but little or no memory difficulty (multidomain non-amnestic deficit).

APOE genotyping

Leukocyte DNA was isolated with the Qiagen blood isolation kit according to the instructions of the manufacturer (Qiagen, Hilden, Germany). The apolipoprotein E (APOE) genotype was studied as described elsewhere [28]. Patients were divided by APOE status into those with or without at least oneε4 allele.

Statistical analysis

Our primary approach followed principles of survival analysis; that is, we followed up an at-risk cohort, analyzing for variables that predicted time to event (here, dementia). To do this, we used recursive partitioning analysis (RPA) to identify variables that dichotomized the at-risk pool into groups with a statistically significant contrast in their predicted dementia-free survival time. The method is hierarchical in that it first identifies the categorical distinction that creates the most extreme contrast in dementia-free survival time. The two resulting groups are then evaluated for further (secondary) covariate-based distinctions that again create subcategories with the strongest contrast in predicted dementia-free survival time. These subcategories are again explored for further (tertiary) covariate-based distinctions, etc. The process is continued for each branch of the categorical tree until there are no further statistically significant subgroup distinctions. To create dichotomies in continuous or quasi-continuous variables, the RPA routine used splitting rules to identify critical values that resulted in subgroups with the strongest discrimination in dementia-free survival time.

Mean dementia-free survival time was estimated for each categorical subgroup using the Kaplan–Meier survival method. This method censored participants at their point of last evaluation if they died or dropped out of the study without dementia or if they had not developed dementia by the end of follow-up. We used the log rank test to evaluate differences in the survival distributions of the subgroups.

Finally, we used multivariable Cox proportional hazards regression to estimate hazard ratios (HRs) and Wald 95% confidence intervals (95% CI) associated with each baseline characteristic that had been evaluated in the RPA. The Cox HRs cannot be compared directly against the results from RPA, but they provide a familiar measure of the association between the subgroups' individual components and dementia risk.

All analyses employed an alpha level for statistical significance of 0.05 (two-tailed). The analyses were performed using the R-package rpart (version 3.1-41) and Predictive Analytics Software (version 19.0; IBM Corp., Armonk, NY, USA).

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Material and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Declaration of interest
  9. References

Characteristics of the sample

Of the 3198 dementia-free GP patients at baseline, 2414 (75.5%) were classified as cognitively unimpaired, while 784 (24.5%) met criteria for CI without dementia. The mean age of patients with CI was 79.9 years (SD = 3.8), and 541 patients (69.0%) were female. Some 207 (26.4%) of the CI patients had a predominantly amnestic deficit (single-domain amnestic deficit: = 96/12.2%; multidomain amnestic deficit: = 111/14.2%), while 577 (73.6%) had a non-amnestic deficit (single non-memory deficit: = 480/61.2%; multidomain non-amnestic deficit: = 97/12.4%). A total of 157 (20.0%) developed dementia during the follow-up period. The remaining 627 patients remained dementia free at the end of follow-up III or until they died or dropped out of the study (Fig. 1).

Hierarchical patterns of predictors of dementia-free survival

The mean dementia-free survival time among the patients with baseline CI, as estimated by Kaplan–Meier survival analysis, was 4.8 years (95% CI = 4.7–5.0). RPA identified a cutoff of 27.47 on the MMSE score as the strongest single discriminator of dementia-free survival (Fig. 2). We found no further variables that allocated patients with MMSE scores ≥ 27.47 into groups with significantly different dementia-free survival times. These higher-scoring patients (group A) had the longest estimated mean dementia-free survival time (M = 5.5 years) and the lowest cumulative risk of progression to dementia (7.7%). Among those who scored lower on the MMSE, RPA next identified impairment in at least one domain of iADL as the strongest discriminator of dementia-free survival. Those with iADL impairment were further differentiable by the presence or absence of SMI with related worry (SMI-w). Thus, the 30 patients who had low MMSE scores along with iADL impairment and SMIw (group F) had the shortest estimated mean dementia-free survival time (M = 2.4 years) and the highest cumulative risk of progression to dementia (73.3%). Mean dementia-free survival time was longer (3.8 years) and cumulative risk of progression was lower (34.3%) in other low MMSE subjects who had iADL impairment and SMI but without related worry or those with no SMI (group E). Among patients who had MMSE scores < 27.47 but unimpaired iADL, estimated dementia-free survival time was differentiable by subtype of cognitive deficit (amnestic vs. non-amnestic types). A final distinction by age (≥79.5 vs. <79.5 years) was evident among the subgroup who had amnestic syndromes. The resulting categories (groups B–D) showed results intermediate between groups A and E. Altogether, RPA identified six groups (A–F) of cognitively impaired patients with different estimated mean dementia-free survival times. The distinction among the estimates for the six groups was robust (log rank: χ² = 154.255, d.f. = 5, < 0.001).

image

Figure 2. Recursive partitioning analysis of hierarchical predictors of dementia-free survival in GP patients with cognitive impairment.

Download figure to PowerPoint

Table 1 shows HRs from multivariable Cox proportional hazards regression that identified eight individual variables associated with increased dementia risk. The HRs with confidence intervals that excluded the null value of 1.0 included all variables that contributed to the distinct risk groups from RPA and also identified depression, impaired hearing, and APOE ε4 status as risk factors.

Table 1. Cox proportional hazards regression of time to incident dementia in general practice patients with cognitive impairment (= 745)a
Baseline characteristicsWaldd.f.HR (95% CI)P value
  1. HR, hazard ratio; CI, confidence interval; iADL, instrumental activities of daily living; SMI, subjective memory impairment; MMSE, Mini-Mental State Examination.

  2. a

    Missing data for 39 (5.0%) subjects.

  3. b

    A score ≥6 on the Geriatric Depression Scale [20].

  4. c

    At least one close family member (mother, father, or sibling) suffered from dementia.

  5. d

    Difficulties in at least one domain of the iADL Scale [19].

  6. e

    The higher the MMSE score, the better the cognition.

  7. f

    Amnestic subtypes of cognitive deficit include single-domain amnestic deficit and multidomain amnestic deficit.

  8. g

    Non-amnestic subtypes of cognitive deficit include single non-memory deficit and multidomain non-amnestic deficit.

Age, every additional year11.86911.08 (1.03–1.13)0.001
Gender, male vs. female0.68310.81 (0.49–1.34)0.41
Living situation, living alone vs. with others2.84711.42 (0.95–2.14)0.09
Family status, single/widowed/divorced vs. married1.17411.52 (0.71–3.23)0.28
Comorbidity
Diabetes mellitus2.58711.35 (0.94–1.96)0.11
Hypertension0.06011.05 (0.70–1.59)0.81
Cardiac arrhythmia1.08210.80 (0.53–1.21)0.30
Coronary heart disease1.10211.25 (0.82–1.90)0.29
Myocardial infarction2.38510.59 (0.31–1.15)0.12
Peripheral arterial obstructive disease0.03510.95 (0.53–1.69)0.85
Carotic artery stenosis (>80%)0.79910.61 (0.21–1.79)0.37
Transient ischemic attack2.64211.55 (0.91–2.64)0.10
Stroke1.25511.43 (0.77–2.65)0.26
Hyperlipidemia1.12311.27 (0.82–1.95)0.29
Hypercholesterolemia0.49211.16 (0.76–1.78)0.48
Parkinson's disease2.47912.04 (0.84–4.93)0.12
Epilepsy3.44610.14 (0.02–1.12)0.06
Hyperthyroidism0.02510.94 (0.45–1.96)0.87
Hypothyroidism0.06310.88 (0.34–2.32)0.80
Traumatic brain injury1.13911.26 (0.82–1.94)0.29
Depressionb4.38711.60 (1.03–2.48)0.04
Impairment in vision0.00511.02 (0.65–1.58)0.95
Impairment in hearing4.02411.44 (1.01–2.05)0.05
Smoking
Former smoker vs. non-smoker2.66510.69 (0.45–1.08)0.10
Current smoker vs. non-smoker0.03310.94 (0.47–1.87)0.86
Alcohol consumption
Normal drinking vs. no drinking0.43211.13 (0.78–1.65)0.51
Risky/harmful drinking/addiction vs. no drinking0.54010.47 (0.06–3.59)0.46
Family history of dementiac0.04211.04 (0.69–1.58)0.84
APOE ε4 allele10.81111.81 (1.27–2.58)0.001
iADL impairmentd15.45512.26 (1.51–3.40)<0.001
Use of antidementia drugs0.65711.19 (0.78–1.84)0.42
Subjective memory impairment (SMI)
SMI without related worry vs. no SMI0.22911.11 (0.73–1.69)0.63
SMI with related worry vs. no SMI6.78311.82 (1.16–2.85)0.009
MMSE scoree, ever additional point28.03110.81 (0.75–0.87)<0.001
Amnesticfvs. non-amnestic subtypes of cognitive deficitg17.51512.10 (1.48–2.97)<0.001

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Material and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Declaration of interest
  9. References

We used recursive partitioning analysis (RPA) to investigate distinctions in dementia-free survival in a sample of 784 cognitively impaired elderly general practice patients. This method identifies groups with various clinical characteristics that create categorical hierarchies or ‘trees’ of dementia-free survival. Searching among numerous clinical characteristics, RPA identified six categories with different estimated mean dementia-free survival times of dementia ranging from 2.4 to 5.5 years. The strongest predictor of imminent dementia was a score on the MMSE below 27.47. In patients with such low scores (but not in others), the next strongest predictor of dementia onset was impairment in instrumental activities of daily living (iADL). Patients with impaired iADL were further differentiable into those with a subjective sense of memory impairment together with related worry or concern (SMI-w) vs. others who lacked SMI or who had SMI without apparent concern. Patients who had intact iADL were differentiable by type of cognitive syndrome (amnestic or non-amnestic). Finally, among the latter, dementia-free survival was further differentiable by age (≥79.5 vs. < 79.5 years).

Our finding that an objective cognitive deficit (here, a subthreshold MMSE score) is a strong predictor of imminent dementia corroborates results of others (e.g., [8, 29-31]). This finding is also consistent with the requirement of objective cognitive impairment in the 2004 consensus criteria for MCI [27]. However, an important distinction is the primacy of objective cognitive deficit in our approach, whereas the foremost (‘gateway’) criterion for the diagnosis of MCI is the presence of a cognitive complaint. We also observed that a cognitive complaint was a predictor of dementia onset, but only when accompanied by an associated sense of worry or concern (SMI-w) and only in patients who had impaired iADL. The importance of iADL impairment as a predictor of dementia onset has been noted by others [32-34], but such impairment is not mentioned in the consensus criteria for MCI beyond acknowledgement that the diagnosis can be made in the presence of minimal impairment in complex instrumental functions [27]. Our results suggest that impairment in iADL is second only to objective cognitive impairment as a predictor of dementia outcomes. Our finding that a cognitive complaint predicted dementia onset only when accompanied by a subjective sense of worry or concern (SMI-w) agrees with previous observations in the AgeCoDe study [35]. Those results suggested that SMI-w is a stronger predictor of conversion to dementia than SMI without worry or concern, and it has been speculated that this concern reflects the patient's sense that his or her cognitive decline represents the beginning of a severe cognitive disorder rather than ‘normal’ aging [35]. We also observed that an amnestic vs. non-amnestic subtype of cognitive deficit was a predictor of dementia outcomes. This result conforms to the categorization of MCI in the 2004 consensus document [27], but with the important distinction that our RPA results suggested that the amnestic/non-amnestic dichotomy was important only in patients with intact iADL. Finally, RPA identified older age as a predictor of dementia outcome, but only in patients with low MMSE score whose iADL were intact and whose cognitive deficits were primarily in the realm of memory. Despite these restrictions, the distinction is probably important because it applies to one of the largest groups of patients with CI (332 of 784; 42.3%) identified by RPA. This finding is again of potential interest because age has not historically been considered in the MCI construct [27, 36].

In addition to those five variables that contributed to the distinct RPA groups with different estimated mean dementia-free survival times of dementia, Cox proportional hazards regression also identified depression, impaired hearing, and APOEε4 status as associated with increased dementia risk. Recent studies suggest a complex relationship particularly between depression and the development of cognitive deficits/dementia. The possible underlying mechanisms at play include, but are not restricted to, shared confounding or risk factors (e.g., vascular diseases). Depression, for example, may lead to cognitive deficits only in presence of a genetic susceptibility factor (interaction effect); alternatively, depression may directly lead to cognitive deficits (causal relation) as it may cause hippocampus damage via hypercortisolemia, or depressive symptoms may be a reaction to self-perceived cognitive deficits in a neurodegenerative process (reverse causality) (for an overview, see [37, 38]). Despite the need to validate these mechanisms [37], a closer monitoring of cognitive changes in elderly people with depressive symptoms [38] may generally contribute to improve the early detection of dementia.

Our study has some limitations. First, even though more than 90% of German elderly people have regular contact with a GP [39], our sample may be not representative. Generalizability of our results may be questionable also because of the study's relatively low response rate. Older non-respondents in particular may have included disproportionate numbers of patients with CI. Second, we assigned dementia onsets by convention at the midpoint between times of assessment. This procedure may not only reduce precision in dating of onset; it may also introduce bias because of differential mortality or non-response among participants who developed dementia during the interval. We attempted to control for such mortality or response bias by retrospective interviews of an informant, but we cannot be certain of the degree of control obtained by this method. Third, if SIDAM results were unavailable, we assigned follow-up dementia diagnoses based on Global Deterioration Scale ratings [23] and Blessed Dementia Rating subscales [24]. We chose this procedure because it reduced sample attrition. We are aware, however, that dementia diagnoses assigned based on neuropsychological test results may be more accurate than diagnoses assigned based on rating scales. We cannot ascertain the lack of differential misclassification. Fourth, RPA identified patient subgroups that could not be differentiated further by their estimated dementia-free survival times. It is possible, however, that studies with larger samples might have provided sufficient statistical power to identify other discriminators of dementia-free survival within these subgroups. Finally, we used instruments chosen for their applicability to epidemiological studies. Such instruments are necessarily brief and may lack either sensitivity or specificity when compared with more extensive clinical and neuropsychological examinations.

Nonetheless, these brief instruments provided data sufficient to suggest a plausible hierarchy of variables that identify groups with near- and longer-term progression to dementia. It may be, in fact, that the RPA-generated hierarchy served to identify groups in order of probabilities that their members had a prodrome of Alzheimer's dementia (AD). AD affects not only memory, but also language, praxis functions, and executive functions. Patients with disordered language, praxis, or executive functions (‘Werkzeugstörungen’) may therefore have been more likely to have Alzheimer's disease [40]. Furthermore, the fact that their illness was sufficiently advanced to provoke such features could imply that dementia was imminent. Patients who sensed this – and therefore worried about it – may have been most likely to progress rapidly. Others who lacked iADL impairment may still have had prodromal AD, but presumably in a milder state and therefore further from dementia onset. Among the latter individuals, those with amnestic features were probably more likely to have very early Alzheimer's disease and would therefore have been more likely to progress. Finally, we identified a relatively low-risk group who lacked iADL impairment and also lacked an amnestic picture. Even so, some of these may in fact have had very early manifestations of AD. Because age is the strongest risk factor for AD, it stands to reason that the oldest patients in this group may again have been enriched for Alzheimer's pathology and therefore more likely (eventually) to progress.

Importantly, our results tend to affirm rather than refute the construct of MCI and its diagnostic criteria [27, 36]. At the same time, our findings suggest that empirical inquiry may produce some refinements of these criteria. Because the existing criteria have clear value, it will be important to test and attempt to validate our conclusions with further empirical inquiry. In the mean time, the present findings of an empirical hierarchy of predictors of dementia-free survival may be of interest. Notably, our results relied on five easily identifiable characteristics that could be readily ascertained, if desired, in clinical practice for risk assessment and patient monitoring. As the method of RPA seemed to show broad applicability for identifying clinical predictors of dementia-free survival, further studies might also apply this method with inclusion of biomarkers or other, more specialized, variables to explore the (hierarchical) role of the latter in assessing dementia-free survival time. Results of this last approach might then be used in more specialized settings such as memory clinics.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Material and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Declaration of interest
  9. References

Funding: This publication is part of the German Research Network on Dementia (KND) and the German Research Network on Degenerative Dementia (KNDD) and was funded by the German Federal Ministry of Education and Research (grants KND: 01GI0102, 01GI0420, 01GI0422, 01GI0423, 01GI0429, 01GI0431, 01GI0433, 01GI0434; grants KNDD: 01GI0710, 01GI0711, 01GI0712, 01GI0713, 01GI0714, 01GI0715, 01GI0716, 01ET1006B). Tobias Luck was supported in writing the publication by a research fellowship of the German Research Foundation (grant: Lu 1730/1-1).

Further members of the AgeCoDe Study Group are as follows: Heinz-HaraldAbholz, Wolfgang Blank, Hendrik van den Bussche (Principal Investigator 2002–2011), Sandra Eifflaender-Gorfer, Annette Ernst, Kathrin Heser, Hanna Kaduszkiewicz, Teresa Kaufeler, MirjamKöhler, Alexander Koppara, Carolin Lange, Manfred Mayer, Julia Olbrich, Anna Schumacher, Janine Stein, Susanne Steinmann, FranziskaTebarth, Klaus Weckbecker, Dagmar Weeg, Steffen Wolfsgruber, and Thomas Zimmermann.

We want to thank all participating patients and their general practitioners for their kind collaboration. Professor Ingmar Skoog generously provided comments on a draft of the manuscript.

Declaration of interest

  1. Top of page
  2. Abstract
  3. Introduction
  4. Material and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Declaration of interest
  9. References

Steffi G. Riedel-Heller serves as editor of Psychiatrische Praxis and has received honoraria from serving on the scientific advisory board of Novartis. Frank Jessen serves on scientific advisory boards for AC Immune, UCB, Schwabe, GE Healthcare, Lilly, and Janssen Cilag and has received honoraria from Esai, Pfizer, Novartis, Lilly, GE Healthcare, and Schwabe. He has received commercial research support from Schwabe. Wolfgang Maier received research grants from the following companies, respectively, is member of the Advisory Boards or draws a fee for speech from: AstraZeneca, Eli Lilly, Janssen Cilag, Lundbeck, and Pfizer. All other authors report no disclosures. All authors declare no conflict of interest in relation to this study.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Material and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Declaration of interest
  9. References
  • 1
    Alzheimer's Disease International. World Alzheimer Report 2010 – The Global Economic Impact of Dementia. London: Alzheimer's Disease International, 2010.
  • 2
    Leicht H, Heinrich S, Heider D et al. Net costs of dementia by disease stage. Acta Psychiatr Scand 2011;124:384395.
  • 3
    Ferri CP, Prince M, Brayne C et al. Global prevalence of dementia: a Delphi consensus study. Lancet 2005;366:21122117.
  • 4
    Busse A, Hensel A, Gühne U, Angermeyer MC, Riedel-Heller SG. Mild cognitive impairment: long-term course of four clinical subtypes. Neurology 2006;67:21762185.
  • 5
    Mitchell AJ, Shiri-Feshki M. Rate of progression of mild cognitive impairment to dementia - meta-analysis of 41 robust inception cohort studies. Acta Psychiatr Scand 2009;19:252265.
  • 6
    Chapman RM, Mapstone M, McCrary JW et al. Predicting conversion from mild cognitive impairment to Alzheimer's disease using neuropsychological tests and multivariate methods. J Clin Exp Neuropsychol 2011;33:187199.
  • 7
    Cui Y, Liu B, Luo S et al. Identification of conversion from mild cognitive impairment to Alzheimer's disease using multivariate predictors. PLoS ONE 2011;6:e21896.
  • 8
    Devanand DP, Liu X, Tabert MH et al. Combining early markers strongly predicts conversion from mild cognitive impairment to Alzheimer's disease. Biol Psychiatry 2008;64:871879.
  • 9
    Ewers M, Walsh C, Trojanowski JQ et al. Prediction of conversion from mild cognitive impairment to Alzheimer's disease dementia based upon biomarkers and neuropsychological test performance. Neurobiol Aging 2012;33:12031214.
  • 10
    Furney SJ, Kronenberg D, Simmons A et al. Combinatorial markers of mild cognitive impairment conversion to Alzheimer's disease - cytokines and MRI measures together predict disease progression. J Alzheimers Dis 2011;26 (Suppl 3):395405.
  • 11
    Gomar JJ, Bobes-Bascaran MT, Conejero-Goldberg C, Davies P, Goldberg TE. Utility of combinations of biomarkers, cognitive markers, and risk factors to predict conversion from mild cognitive impairment to Alzheimer disease in patients in the Alzheimer's disease neuroimaging initiative. Arch Gen Psychiatry 2011;68:961969.
  • 12
    Heister D, Brewer JB, Magda S, Blennow K, Mcevoy LK. Predicting MCI outcome with clinically available MRI and CSF biomarkers. Neurology 2011;77:16191628.
  • 13
    Kantarci K, Weigand SD, Przybelski SA et al. Risk of dementia in MCI: combined effect of cerebrovascular disease, volumetric MRI, and 1H MRS. Neurology 2009;72:15191525.
  • 14
    Landau SM, Harvey D, Madison CM et al. Comparing predictors of conversion and decline in mild cognitive impairment. Neurology 2010;75:230238.
  • 15
    Zaudig M, Hiller W. SIDAM - A structured interview for the diagnosis of dementia of the Alzheimer type, multi-infarct (or vascular) dementia and dementias of other aetiology according to DSM-III-R, DSM-IV and ICD-10 (SIDAM-manual) [in German]. Bern: Huber, 1996.
  • 16
    Zaudig M, Mittelhammer J, Hiller W et al. SIDAM - A structured interview for the diagnosis of dementia of the Alzheimer type, multi-infarct dementia and dementias of other aetiology according to ICD-10 and DSM-III-R. Psychol Med 1991;21:225236.
  • 17
    Folstein MF, Folstein ME, McHugh PR. Mini-Mental-State: a practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res 1975;12:189198.
  • 18
    Hötte SD, Lankers D, Kissler S, Freyberger HJ, Schröder SG. MMSE for survival prognostics in dementia [in German]. Psychiatr Prax 2010;37:7883.
  • 19
    Lawton MP, Brody EM. Assessment of older people: self-maintaining and instrumental activities of daily living. Gerontologist 1969;9:179186.
  • 20
    Sheikh JI, Yesavage JA. Geriatric depression scale (GDS): recent evidence and development of a shorter version. In: Brink TL, ed. Clinical gerontology: a guide to assessment and intervention. New York: Haworth, 1986:165173.
  • 21
    Gauggel S, Birkner B. Validity and reliability of a German version of the Geriatric Depression Scale (GDS) [in German]. Z Klin Psychol 1999;28:1827.
  • 22
    World Health Organization – Regional Office Europe for Europe. Alcohol and health – Implications for Public Health Policy. Report of a WHO Working Group, Oslo, 1995, October.
  • 23
    Reisberg B, Ferris SH, de Leon MJ, Crook T. The Global Deterioration Scale for assessment of primary degenerative dementia. Am J Psychiatry 1982;139:11361139.
  • 24
    Blessed G, Tomlinson BE, Roth M. The association between quantitative measures of dementia and of senile changes in the cerebral grey matter of elderly subjects. Br J Psychiatry 1968;114:797811.
  • 25
    APA – American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders – DSM-IV-TR, 4th edn (Text Revision). Washington, DC: American Psychiatric Association, 2000.
  • 26
    Luck T, Zaudig M, Wiese B, Riedel-Heller SG. SIDAM: age- and education-specific reference values for the cognitive test section according to the new CASMIN educational classification [in German]. Z Gerontopsych Gerontopsychiat 2007;20:3138.
  • 27
    Winblad B, Palmer K, Kivipelto M et al. Mild cognitive impairment – beyond controversies, towards a consensus: report of the International Working Group on Mild Cognitive Impairment. J Intern Med 2004;256:240246.
  • 28
    Hixson JE, Vernier DT. Restriction isotyping of human apolipoprotein E by gene amplification and cleavage with HhaI. J Lipid Res 1990;31:545548.
  • 29
    Van Den Kommer TN, Bontempo DE, Comijs HC et al. Classification models for early identification of persons at risk for dementia in primary care: an evaluation in a sample aged 80 years and older. Dement Geriatr Cogn Disord 2009;28:567577.
  • 30
    Maioli F, Coveri M, Pagni P et al. Conversion of mild cognitive impairment to dementia in elderly subjects: a preliminary study in a memory and cognitive disorder unit. Arch Gerontol Geriatr 2007;44(Suppl 1):233241.
  • 31
    Pozueta A, Rodríguez-Rodríguez E, Vaqquez-Higuera JL et al. Detection of early Alzheimer's disease in MCI patients by the combination of MMSE and an episodic memory test. BMC Neurol 2011;11:78.
  • 32
    Pérès K, Chrysostome V, Fabrigoule C, Orgogozo JM, Dartigues JF, Barberger-Gateau P. Restriction in complex activities of daily living in MCI: impact on outcome. Neurolog 2006;67:461466.
  • 33
    Luck T, Luppa M, Angermeyer MC, Villringer A, König HH, Riedel-Heller SG. Impact of Impairment in Instrumental Activities of Daily Living and Mild Cognitive Impairment on Time to Incident Dementia – Results of the Leipzig Longitudinal Study of the Aged (LEILA75+). Psychol Med 2011;41:10871097.
  • 34
    Tabert MH, Albert SM, Borukhova-Milov L et al. Functional deficits in patients with mild cognitive impairment: prediction of AD. Neurology 2002;58:758764.
  • 35
    Jessen F, Wiese B, Bachmann C et al. Prediction of dementia by subjective memory impairment: effects of severity and temporal association with cognitive impairment. Arch Gen Psychiatry 2010;67:414422.
  • 36
    Albert MS, Dekosky ST, Dickson D et al. 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 2011;7:270279.
  • 37
    Panza F, Frisardi V, Capurso C et al. Late-life depression, mild cognitive impairment, and dementia: possible continuum? Am J Geriatr Psychiatry 2010;18:98116.
  • 38
    Shahnawaz Z, Reppermund S, Brodaty H et al. Prevalence and characteristics of depression in mild cognitive impairment: the Sydney Memory and Ageing Study. Acta Psychiatr Scand 2013;127:394402.
  • 39
    Linden M, Gilberg R, Horgas AL, Steinhagen-Thiessen E. Die Inanspruchnahme medizinischer und pflegerischer Hilfe im hohen Alter [in German]. In: Baltes PB, Mayer KU, eds. Die Berliner Altersstudie. Berlin: Akademie Verlag, 1996:475495.
  • 40
    Lauter H, Meyer JE. Clinical and nosological concepts of senile dementia. In: Müller C, Ciompi L, ed. Senile dementia. Bern: Hans Huber, 1968:1326.