Abstract
High occupational attainment has been known as a marker of cognitive reserve. Previous studies in the general population have shown that high occupational attainment is associated with reduced risk of Alzheimer’s disease (AD). However, few studies have assessed the effect of occupational attainment on the clinical course of mild cognitive impairment (MCI). In this study, we evaluated whether individuals with high occupational attainment show more frequent progression from MCI to AD. Participants (n = 961) with MCI were recruited from a nationwide, hospital-based multi-center cohort, and were followed for up to 60 months (median: 17.64, interquartile range [12.36, 29.28]). We used Cox regression for competing risks to analyze the effect of occupational attainment on development of AD, treating dementia other than AD as a competing risk. Among the 961 individuals with MCI, a total of 280 (29.1%) converted to dementia during the follow-up period. The risk of progression to AD was higher in the individuals with high occupational attainment after controlling for potential confounders (hazard ratio = 1.83, 95% confidence interval = 1.25–2.69, p = 0.002). High occupational attainment in individuals with MCI is an independent risk factor for higher progression rate of MCI to AD. This result suggests that the protective effect of high occupational attainment against cognitive decline disappears in the MCI stage, and that careful assessment of occupational history can yield important clinical information for prognosis in individuals with MCI.
INTRODUCTION
Alzheimer’s disease (AD), the most common cause of dementia, is a progressive neurodegenerative disease with an increasing prevalence and social burden [1–3]. The concept of cognitive reserve has been suggested to explain differences between individuals in susceptibility to clinical manifestations of AD [4]. According to this theory, the brain actively attempts to cope with the neurodegenerative pathology by using preexisting compensatory mechanisms [5]. Experiences over the lifetime, such as educational and occupational attainment, can increase this reserve, which in turn delays or limits cognitive dysfunction. Previous studies in the general population have demonstrated that high educational or occupational attainment is associated with reduced risk of incident AD [6, 7]. However, once the dementia develops, patients with higher cognitive reserve appear to experience more rapid cognitive decline, which suggests that those with higher cognitive reserve have accumulated a greater degree of brain pathology by the time cognitive impairment is apparent [8–10].
Mild cognitive impairment (MCI) is an intermediate state between normal cognitive aging and dementia [11]. The clinical course of MCI is heterogeneous [12]. Although the risk of dementia is increased approximately three-fold in the population with MCI compared with cognitively normal subjects [13, 14], a substantial proportion of individuals with MCI remain in a stable state or revert to normal [15, 16]. It has been demonstrated that multiple factors including age, gender, vascular risk factors, and brain pathologies are involved in the prognosis of MCI [17–20]. However, only limited studies have investigated the effect of cognitive reserve on the clinical course of MCI [21]. It is still unclear whether higher occupational attainment would be a protective factor against cognitive decline in individuals with MCI or a risk factor for rapid cognitive decline as it is in AD.
In this study, we investigated the role of occupational attainment as a risk factor for the progression from MCI to AD in a large cohort collected through a coordinated national clinical network. Our hypothesis is that the individuals with MCI with higher occupational attainment will show frequent progression to AD than those with lower occupational attainment.
METHODS
Subjects
Subjects were recruited from the Clinical Research Center for Dementia of South Korea (CREDOS) study. CREDOS is a prospective, nationwide, hospital-based multi-center study with 56 participating hospitals [22, 23]. The CREDOS project was designed to assess the occurrence and risk factors of cognitive disorders.
The criteria for MCI in the CREDOS study [24] are as follows: (1) presence of memory complaints that raise concern about a change in cognition; (2) intact function in activities of daily living (ADL) except performing complex functional tasks; (3) objective cognitive impairment (at least –1.0 standard deviation below age- and education-adjusted norms) in more than one cognitive domain on standardized neuropsychological testing [25, 26]; (4) Clinical Dementia Rating (CDR) of 0.5 [27]; and (5) not demented according to Diagnostic and Statistical Manual of Mental Disorders (DSM)-IV-TR criteria [28]. MCI was classified into amnestic and non-amnestic subtypes and included in the analysis as a covariate because amnestic MCI had a higher progression rate to AD in a previous study [29]. Clinical classification of amnestic MCI and non-amnestic MCI was based on memory function, which was considered abnormal when the performance in the delayed recall item was lower than –1.0 standard deviation compared to age- and education-matched norms [30]. All diagnoses of patients were confirmed according to current diagnostic criteria (MCI –criteria for the clinical and cognitive syndrome) from the National Institute of Aging and Alzheimer’s Association (NIA-AA) [31]. A reliable informant familiar with the daily activities of the subjects was required.
Exclusion criteria were (1) history of significant hearing or visual impairment rendering participation in the interview difficult; (2) neurological disorders (e.g., territorial infarction, intracranial hemorrhage, brain tumor, hydrocephalus, multiple sclerosis, Parkinson’s disease, or Huntington’s disease); (3) major psychiatric disorders (e.g., schizophrenia, mental retardation or mania); and (4) physical illnesses that could interfere with the clinical study (e.g., severe cardiac disease, respiratory diseases, uncontrolled diabetes or malignancy). Patients were eligible for analysis if they had at least one longitudinal clinical review after baseline.
There were 978 MCI subjects drawn from the CREDOS cohort (November 2005 to May 2012) and screened for eligibility. Of these, 17 subjects with missing data were excluded (Fig. 1). Therefore, a total of 961 participants with MCI were included in our analysis. These subjects were followed for up to 60 months (median: 17.64, interquartile range [12.36, 29.28]). This study was approved by the institutional review boards of the participating centers, and signed informed consent was obtained from all participants and informants.
Clinical evaluation
We evaluated all participants with a complete medical interview, neurological examinations, extensive neuropsychological tests, and magnetic resonance imaging (MRI) scans. We obtained a complete blood count, blood chemistry tests, vitamin B12/folate, syphilis serology, and thyroid function tests to exclude common secondary causes of dementia. Brain magnetic resonance imaging (MRI) scans were obtained with transaxial T2 and T1-weighted scans, and fluid-attenuated inversion recovery slices. Absence of structural lesions such as brain tumors or traumatic brain injuries was confirmed by MRI. Visual rating for white matter hyperintensity (WMH) was conducted in accordance with the protocol developed for CREDOS [32, 33]. The patients were classified into two groups (mild and moderate/severe) based on their WMH around the lateral ventricles or in deep white matter.
A standardized neuropsychological battery, the Seoul Neuropsychological Screening Battery-Dementia Version (SNSB-D) [26] was administered to all participants at baseline and at follow up visits. The Korean version of the Mini-Mental State Examination (K-MMSE) [34] and the 15-item Geriatric Depression Scale (GDS-15) [35] were additionally administered at each evaluation.
Occupational attainment was rated as described by Garibotto and colleagues [9, 36]. Using the information provided by a self-report from respondents concerning their main occupation during their working life, patients were classified into two groups: low attainment group (no occupation, unskilled/skilled laborer, housewife, tradesman, lower level civil servant, employee, self-employed small business, and office or sales personnel) and high attainment group (mid-level civil servant or management, head of a small business, academician or specialist in a subordinate position, senior civil servant or management, senior academic position, and self-employed with high degree of responsibility). Coding was carried out by two investigators, and showed 92.4% agreement. Disagreements in coding were resolved by consensus between the two investigators. The data analysis was based on the coding results after agreement was reached. These investigators were blinded to the dementia progression outcomes. Current employment was entered as a covariate, patients who had an occupation at enrollment were coded with 1, and the others werecoded 0.
Diagnosis of incident dementia
Evidence of impairment in ADL supported by objective information from a caregiver was required for the diagnosis of incident dementia during follow-up. A standardized neuropsychological battery (SNSB-D) was used to confirm the cognitive impairment. Incident dementia cases also were required to meet DSM-IV-TR criteria for dementia [28]. In addition, specific diagnostic criteria were used for dementia subclassification after full neurological review. A diagnosis of probable AD was made according to the criteria from the National Institute of Neurological and Communicative Diseases and Stroke-Alzheimer’s Disease and Related Disorders Association (NINCDS-ADRDA) [37]. Diagnosis of subcortical vascular dementia (SVD) was made in accordance with the criteria from the National Institute of Neurological Disorders and Stroke-Association Internationale pour la Recherche et l’Enseignement en Neurosciences (NINDS-AIREN) criteria [38] and imaging criteria proposed by Erkinjuntti [39]. Other forms of incident dementia such as Lewy body dementia, frontotemporal dementia, progressive supranuclear palsy, and normal pressure hydrocephalus were diagnosed by standard neurological criteria. Onset of dementia was marked as the date on which the clinical symptoms and neuropsychological findings of the subject first allowed the diagnosis of dementia to be made. In this study, each participant was assigned only one predominant etiological cause of dementia.
Statistical analysis
For non-normally distributed continuous variables, data are presented as median and interquartile range, and group comparisons are performed by the non-parametric Wilcoxon rank-sum test. Categorical variables are summarized as frequencies and proportions, and compared with the Chi-square test. We used Cox regression for competing risk to analyze the effect of candidate risk factors on progression to AD, treating dementia other than AD as a competing risk. The time variable was defined as time from MCI diagnosis to probable AD onset, or time to dementia other than AD. Durations at risk were recorded from baseline to last follow-up or to a diagnosis of a dementia. The assumptions of proportional hazards were confirmed by the residuals-based test, and no violation of the assumption was found for any variable [40]. Results were considered statistically significant at a two-tailed threshold of p < 0.05. R 3.1.0 public statistics software (http://www.r-project.org) was used for all statistical analyses.
RESULTS
Subject characteristics
The clinical and demographic characteristics of the participants at baseline are presented in Table 1. High occupational attainment was recorded in 202 cases (21.0%) and low occupational attainment in 759 cases (79.0%). We recruited 330 (34.3%) male subjects and 631 (65.7%) female subjects (male: female, 1 : 1.91). The median age was 71 (interquartile range: 66–76). 783 (81.5%) participants were classified as amnestic MCI. There were significant differences between occupational groups with respect to gender, age, education, K-MMSE score, and GDS-15. Individuals with high occupational attainment had a high proportion of male gender, older age, higher education, higher K-MMSE score, and lower GDS-15 score. No differences were found in the proportion of current occupational status, diabetes, hypertension, WMH, APOE ɛ4 genotype, and amnestic MCI subtype.
Longitudinal progression of subjects
Among the 759 participants with low occupational attainment, 212 (27.9%) progressed to dementia, 472 (62.2%) had stable MCI, and 75 (9.9%) reverted to a normal state (Fig. 1). Among the 212 patients with low occupational attainment who progressed to dementia, 197 (92.9%) were classified as AD, and 15 (7.1%) were diagnosed with dementia other than AD (6 vascular dementia, 3 Lewy body dementia, 4 frontotemporal dementia, 1 progressive supranuclear palsy, and 1 normal pressure hydrocephalus). Among the 202 participants (21.0%) with high occupational attainment, 68 (33.7%) progressed to dementia, 116 (57.4%) remained in the range of MCI, and 18 (8.9%) reverted to a normal state. Among the 68 patients in the high occupational attainment group who progressed to dementia, 61 subjects (89.7%) were diagnosed with AD, and 7 (10.3%) progressed to dementia other than AD (4 vascular dementia and 3 Lewy body dementia). These proportions are uncorrected for variable survival times and variable observation periods.
Risk factors for progression from MCI to dementia
Table 2 presents the results of univariable survival analyses of risks for progression to AD, estimated by Cox regression with consideration of the competing risk for developing non-AD dementia. Patients with high occupational attainment, older age, lower K-MMSE score, and amnestic MCI were more likely to progress from MCI to AD in these univariable survival analyses (p < 0.05).
We then conducted a multivariable survival analysis using Cox regression with competing risk for developing non-AD dementia. In this analysis, occupational attainment was entered as an independent variable in the survival analysis model along with all candidate risk factors including gender, age, education, current employment status, diabetes, hypertension, K-MMSE score, GDS-15, WMH, and amnestic MCI as covariates. The resulting HRs with CI and p-values are displayed in Table 3. Subjects with high occupational attainment had a significantly higher hazard ratio (HR) for progression to AD (HR = 1.83, 95% confidence interval [CI] = 1.25–2.69, p = 0.002) after controlling gender, age, education, current employment status, diabetes, hypertension, K-MMSE score, GDS-15, WMH, and amnestic MCI. Figure 2 displays the greater hazard for and more rapid onset of AD among individuals with MCI with high occupational attainment. In addition, significantly higher HRs for AD were observed for female gender, increasing age and education, absence of hypertension, lower K-MMSE score, and amnestic MCI (p < 0.05, Table 3). Current employment status was not associated with progression to AD in individuals with MCI.
In addition, we conducted sensitivity analyses excluding participants with ‘no occupation’ or ‘housewife’ status or both in the low occupational attainment group. Subjects with high occupational attainment still had a significantly higher HR for progression to AD in these analyses (exclusion of ‘no occupation’: n = 930, HR = 1.84, 95% CI = 1.25–2.71, p = 0.002; exclusion of ‘housewife’: n = 648, HR = 1.78, 95% CI = 1.16–2.74, p = 0.009; exclusion of both ‘no occupation’ and ‘housewife’: n = 617, HR = 1.76, 95% CI = 1.14–2.73, p = 0.011).
DISCUSSION
In this study, we investigated the effects of occupational attainment on progression to AD in a prospective cohort of individuals with MCI examined for up to five years. In confirmation of our hypothesis, we found that rate of progression from MCI to AD was increased in the patients with high occupational attainment.
Previous studies conducted in the general population reported different results. Several population-based studies reported that high occupational attainment reduces the risk of incident AD in the non-demented elderly [6, 41–44]. A twin study also found a similar result [44]. Such findings have been taken to suggest that occupational experiences created a reserve against the onset of dementia [46]. However, we found that higher occupational attainment was not a protective factor against cognitive decline in our selected population of patients who already demonstrated MCI. In this group, as in AD patients, high occupational attainment was, rather, a risk factor for likelihood of cognitive decline, as we had hypothesized. Thus, our results suggest that the protective effects of high occupational attainment against cognitive decline disappear in the MCI stage. At the point of time that MCI is revealed, compensatory capacity of cognitive reserve which masking MCI manifestation has been exhausted and cognitive decline progress more rapidly due to higher cumulated burden of neuropathology such as amyloid beta-peptide, tau protein, loss of neurons, or others. Our results also add convergent validity to the proposal that MCI is in most patients who progress to dementia a prodrome of AD (92% or 258 of 280 cases who progressed to dementia developed AD).
To the best of our knowledge, there are few studies that have investigated the role of occupational attainment in the progression of individuals with MCI to dementia. Allegri et al. reported that lack of occupation is a risk factor for progression from MCI to dementia [21]. However, they did not consider the characteristics of occupation, the sample size was limited (n = 239), and brain imaging was not performed routinely. In this study, we included a relatively large sample of approximately 1,000 subjects with MCI in multicenter clinics. Moreover, we considered various confounding variables such as WMH by assessing brain MRI in all subjects. We also controlled for depressive symptoms at baseline by employing the GDS-15.
Previous neuroimaging studies have shown that the individuals with MCI plus high occupational attainment have lower brain glucose metabolism [9], and higher acetylcholinesterase (AChE) activity [47]. These results are consistent with the brain reserve hypothesis [48] that more AD pathology is needed to cause noticeable dementia in persons with higher levels of compensatory premorbid ability. Taken together with our results, these findings suggest that the pathological burden exceeds the cognitive reserve effect of occupational attainment by the time that MCI is apparent.
Hall et al. showed that a higher education level delayed the onset of accelerated cognitive decline, but that once it begins it was more rapid in individuals with a higher education level [49]. Scarmeas et al. reported similar results for the faster cognitive decline in AD patients with higher education [50]. In addition, a previous CREDOS study reported that the protective effects of education against progression to AD disappeared in late-stage MCI [51]. Thus, studies have demonstrated separate effects for education and occupational attainment, and suggested that each of these life experiences contributes independently to cognitive reserve. Our finding is in line with these previous studies [52].
In our data, 72.1% of baseline MCI subjects had stable MCI or reverted to a normal state at the end of follow up. Ganguli et al. reported 40–70% of individuals with MCI remained MCI or reverted to normal over each 2-year interval [16] and Palmer et al. reported 11% of individuals with MCI remained stable, while 25% of patients had improved at 3 years after baseline assessment [15]. However, Manly et al. reported results similar to those in our study: 76.8% of subjects remained in MCI status or reverted to normal [14]. Some possible reasons for these discrepancies are differences in clinical setting, mean follow-up length, as well as subject characteristics such as age, sex, or co-morbidity.
We did not find a significant association between AD progression and current employment status. This result suggests that the environmental stimulus associated with active employment has no effect on loss of cognitive reserve in individuals with MCI.
Our finding highlights the importance of assessment of cognitive reserve for each patient in the clinical setting. The individuals with MCI plus high occupational attainment will present with AD-related clinical symptoms more often than individuals with low occupational attainment. Therefore, early interventions, supported by cognitive testing at more frequent intervals could be considered for those patients [46].
In addition, our finding suggests that a mismatch in occupational attainment across groups could lead to differences in the observed rates of decline in clinical trials of cognitive enhancer candidate drugs. Moreover, as we described above, occupational attainment has its own independent effect on the rate of cognitive decline, separate from the effects of the other covariates commonly used in clinical trials such as education. Therefore, occupational attainment should be considered when matching subjects in clinical trials of novel drugs for MCI as well as for dementia [46]. However, rating of occupational attainment is not simple because of its complex nature and lack of agreement about definitions and categories. In the future, the development of standardized tools to reflect occupational history in clinical trials will be a research priority.
We employed Cox regression for analysis of competing risk. This statistical analysis was developed for situations in which observation of one outcome may obscure observation of the other. In our study, we assigned each participant only one predominant etiological cause of dementia. Therefore, the effect of occupational attainment on AD could be obscured by the incidence of dementia other than AD. Previous studies did not consider the competing risk of dementia other than AD; they simply excluded other forms of dementia in their analysis [53]. However, simple exclusion of competing events could result in an overestimate of the cumulative incidences. These rigorous statistical steps could be a strength of our study.
There are several limitations that should be mentioned. First, we could not obtain APOE ɛ4 allele status from all of the subjects (Table 1). Nevertheless, the APOE ɛ4 carrier status was not associated with occupational attainment in our data, so this genotype is unlikely to be a confounding factor. Second, the generalizability of participants should be noted as a possible limitation. It has been known that the progression rate of cognitive decline in MCI is higher in studies that recruit from memory clinics than from community-based samples [20]. Replication of our results in community-based participants could be helpful. Third, we employed a dichotomous grouping of occupational attainment, which does not fully reflect the complex nature of occupation. Occupational attainment could be defined in various ways including physical demands, social relations or task complexity [54].
In conclusion, our study demonstrated that high occupational attainment in individuals with MCI is independently associated with an increased risk of progression to AD. Thus, our results suggest that careful assessment of occupational attainment in patients with incident MCI can yield important clinical information for prognosis, and that standardized measures of occupational attainment should be adopted in future clinical trials in MCI.
Footnotes
ACKNOWLEDGMENTS
This study was supported by a grant of the Korea Healthcare Technology R&D Project, Ministry of Health and Welfare, Republic of Korea (HI10C2020) and Basic Science Research Program through the National Research Foundation of Korea funded by the Ministry of Education, Science and Technology (NRF-2014R1A2A1A10052419).
