Abstract
BACKGROUND:
Research shows that retirement age is associated with later-life cognition but has not sufficiently distinguished between retirement pathways. We examined how retirement age was associated with later-life dementia and mild cognitive impairment (MCI) for people who retired via the disability pathway (received a disability pension prior to old-age pension eligibility) and those who retired via the standard pathway.
METHODS:
The study sample comprised 7 210 participants from the Norwegian Trøndelag Health Study (HUNT4 70+, 2017–2019) who had worked for at least one year in 1967–2019, worked until age 55+, and retired before HUNT4. Dementia and MCI were clinically assessed in HUNT4 70+ when participants were aged 69–85 years. Historical data on participants’ retirement age and pathway were retrieved from population registers. We used multinomial regression to assess the dementia/MCI risk for women and men retiring via the disability pathway, or early (<67 years), on-time (age 67, old-age pension eligibility) or late (age 68+) via the standard pathway.
RESULTS:
In our study sample, 9.5% had dementia, 35.3% had MCI, and 28.1% retired via the disability pathway. The disability retirement group had an elevated risk of dementia compared to the on-time standard retirement group (relative risk ratio [RRR]: 1.64, 95% CI 1.14–2.37 for women, 1.70, 95% CI 1.17–2.48 for men). MCI risk was lower among men who retired late versus on-time (RRR, 0.76, 95% CI 0.61–0.95).
CONCLUSION:
Disability retirees should be monitored more closely, and preventive policies should be considered to minimize the dementia risk observed among this group of retirees.
Keywords: dementia, mild cognitive impairment, aging workforce, work disability, older workers
Introduction
The number of individuals with dementia is predicted to triple within the next 30 years (1). Given its devasting personal and societal costs and the current lack of effective pharmacological treatments, it is imperative to identify whether lifestyle and health risk behaviors can influence later-life cognitive impairment. The “use it or lose it” hypothesis posits that people can maintain cognitive function by engaging in cognitively-stimulating activities, such as paid work (2). Some studies suggest that later retirement may protect against cognitive impairment, with findings showing that employment and later retirement are negatively related to dementia (3–6), and cognitive decline accelerates post-retirement (7–11). For instance, a Swedish study (N=63 505) found a 2.9-fold lower risk of dementia among those who retired at age 66+ compared to their peers who retired at age 65 (3). The precise nature of the relationship between retirement age and later-life cognitive impairment is, however, still debated. A primary concern is the overlap between factors that influence whether people retire early, “on time”, or late (e.g., poor health, education, wealth (12)) with those related to later-life cognitive impairment (13). There is also evidence of reverse causality (i.e., cognitive impairment causes people to retire early) (5). Conclusions from a recent systematic review emphasized the need for more knowledge on the association between retirement and age-related cognitive decline (11).
To date, most retirement research has focused on people who follow the standard, work-to-retirement pathway. In many Western countries, however, many people retire via the “disability pathway”, that is, by exiting the workforce via a disability pension prior to becoming eligible for an old-age pension (14). Failing to account for variation in the pathways to retirement may cloud our understanding of the relationship between retirement age and later-life cognitive impairment.
Our study is based on data from Norway, where the context allows for considerable heterogeneity regarding retirement age and pathway. Anyone who has lived or worked legally in Norway for at least five years after age 16 is entitled to a retirement pension from the Norwegian state. The normal retirement age in Norway is 67, although collective agreements entitle many public and private sector employees to early retirement from age 62. A considerable proportion of Norwegian residents also leave the labor force before age 67 with a disability pension, which requires at least 50% reduced work capacity certified by a physician. At age 67, the old-age pension replaces the disability pension. More than a quarter (28%) of Norwegians aged 62–67 years receive disability benefits (15), with musculoskeletal and psychiatric problems being the most frequent health-related reasons for disability (16). In the current study, we shed light on the relationship between retirement age and later-life cognitive impairment by distinguishing between two different pathways to retirement (via disability or standard retirement) to help disentangle the relationship between retirement age and later-life cognitive impairment. Previous research has found that retirement age is more closely related to dementia and cognitive decline for men than for women (6, 10). Moreover, dementia is more common among women than among men (17, 18), particularly at older ages (18), whereas mild cognitive impairment (MCI) is more common in men (18–20). We therefore conduct sex- and diagnosis-specific analyses.
Methods
Design and data sources
We employed a historical cohort design. Our starting point was clinical assessments of dementia and MCI conducted as part of the fourth wave of the Norwegian Trøndelag Health Study (HUNT4 70+, 2017–2019), a longitudinal study of health in the general population. Participants were aged 69 to 86 years at the time of cognitive assessment. Their HUNT4 70+ data can be linked with their data in the Norwegian population registers using the unique personal identification number assigned to all Norwegian residents. The population registers contain data on the income, employment, and sociodemographic characteristics of the entire population. In the current study, we linked data on participants’ later life cognitive status with historical data on participants’ employment, income, and sociodemographic characteristics from the population registers, as well as data on their pre-retirement health from HUNT1 (1984–86) and HUNT2 (1995–97).
Sample population
All adult residents of the former Nord-Trøndelag County were invited to participate in each of the four waves of HUNT (21, 22). HUNT4 participants aged 70+ years were invited to participate in the HUNT4 70+ sub-study; in total, 9 930 (51.2%) individuals aged 69–105 years participated. Our sample was comprised of 7 210 HUNT4 70+ participants (a) born 1933–1949 who were (b) employed for at least one year during 1967–2019, (c) employed until age 55+ years, (d) had retired at the time of participation and the cognitive assessment in HUNT4 70+, and for whom (e) data on retirement age and cognitive impairment diagnosis were available (see Figure 1 for the sampling scheme). Participants provided informed, written consent.
Figure 1.
Sample selection showing those invited to participate in HUNT4 70+, non-participants, participants included in HUNT4 70+, those excluded from our study sample, and those included in the final study sample (n=7 210).
Measures
Retirement age and pathway
Using register data, we defined retirement age as the age at which a person received income from a retirement or disability pension and his/her income from paid work dropped below the basic income threshold (“Grunnbeløp”) as determined by the Norwegian national insurance scheme. In 2019, the basic income threshold was NOK 98 866 (~10 000 EUR). By using register data, we were able to avoid some of the problems of self-reported retirement age (e.g., social desirability, recall bias, item-non response bias) (23). Norwegian policy allows people to work and receive an old-age pension. Compared to the first year of pension receipt, our definition of retirement age thus allowed us to identify more precisely when people almost or completely stopped working. We classified participants as belonging to one of four retirement pathway groups: disability (retired <67 years, received disability benefits), early standard (retired <67 years, no disability benefits), on-time standard (retired at age 67), or late standard (retired at age 68+ years).
Dementia and MCI diagnosis
Each HUNT4 70+ participant was assessed for cognitive impairment by two trained medical doctors, from a group of nine, using a comprehensive protocol (for details on the assessment, see (18)). Following the DSM-5 diagnostic criteria (24), participants were categorized as having no cognitive impairment, MCI (mild neurocognitive disorder), or dementia (major neurocognitive disorder). Participants were aged 69–86 years at the time of assessment.
Potential confounders
Retirement age and pathway are related to a number of socioeconomic, health, occupational, and social factors (12), many of which are also risk factors for dementia (for a recent review of risk factors for dementia, see (13)). We therefore examined whether the relationship between retirement age and dementia/MCI changed after adjusting for education, occupational physical demands, number of children, midlife marital status, and midlife lifestyle and health (smoking, alcohol consumption, obesity, cardiovascular disease (CVD), insufficient physical activity, hypertension, diabetes, having no close friends). Measures are described in Appendix 1.
Statistical analyses
Stata v. 17 was used to conduct all analyses. Multiple imputation with 20 iterations was used to impute missing data on exposure and adjustment variables. Inverse-probability weighting (IPW) was used to account for non-response in HUNT4 70+ and to correct for bias due to skewed participation related to age, sex, and education. Details on the imputation and IPW procedures are provided in Appendix 1.
After analyzing the descriptive characteristics of the sample, we checked whether the likelihood of being included in the sample depended on sex. We used linear regression to test whether years in employment and retirement age differed between men and women and logistic regression to examine how sex was related to disability retirement, dementia, and MCI, controlling for age and education.
We then conducted a series of multinomial logistic regression analyses to assess the relationship between dementia/MCI and retirement age for women and men separately. Although MCI is considered an intermediate stage between normal cognition and dementia, most people with MCI will not progress to dementia even after 10 years of follow-up according to a meta-analysis (25). We therefore used multinomial as opposed to ordinal logistic regression. First, we examined the association by using retirement age as a continuous variable with 1-year increments. The four regression models successively adjusted for measures of potential confounding related to age (grouped as 70–74, 75–79, 80–85) (Model 1); education and occupational physical demands (Model 2); family measures including number of children and marital status (Model 3); and lifestyle and health measures, including smoking, alcohol consumption, obesity, insufficient physical activity, hypertension, diabetes, CVD, and having at least one close friend (Model 4). Interactions between retirement age, sex, disability retirement, and occupational physical demands on the risk of dementia and MCI were tested by adding an interaction term (e.g., sex*retirement age) to Model 1.
Second, to simplify our results, we assessed the relationship between dementia/MCI and the disability, early standard, and late standard retirement groups relative to the on-time standard retirement group (reference). We analyzed women and men separately, statistically controlled for age, and then adjusted for potential confounders in successive models as described above.
We report the relative risk ratios (RRR) and 95% confidence intervals (CI) for dementia and MCI associated with 1) retiring one life year later, or 2) disability, early standard, and late standard retirement relative to on-time standard retirement.
To check the robustness of the findings, we re-ran the analyses using only complete cases (i.e., without multiple imputation) and without IPW (n=4 885, 67.8%). To reduce the potential influence of reverse causation, we also re-ran analyses based on only those participants who had retired more than five years prior to participation in HUNT4 70+ (n=5 856, 81.2%) (cf. (6)).
Results
Table 1 displays the characteristics of the sample population. At age 69–86, n=687 (9.5%) participants were diagnosed with dementia and an additional n=2 548 (35.3%) were diagnosed with MCI. More than one quarter (28.1%) of the sample population retired early via the disability pathway.
Table 1.
Characteristics of the sample population, stratified by sex and retirement pathway. The HUNT Study, Norway
| Standard retirement (n=5 186) |
Disability retirement (n=2 024) |
|||
|---|---|---|---|---|
| Women | Men | Women | Men | |
| (n=2480) | (n=2706) | (n=1156) | (n=868) | |
| Registers | ||||
| Age in 2018, mean (SD) | 76.0 (4.2) | 76.1 (4.1) | 76.3 (4.1) | 76.3 (4.1) |
| Education, n (%) | ||||
| Primary | 453 (18.3) | 399 (14.8) | 337 (29.2) | 246 (28.3) |
| Secondary | 1 382 (55.7) | 1 513 (55.9) | 607 (52.5) | 522 (60.1) |
| Tertiary | 645 (26.0) | 794 (29.3) | 212 (18.3) | 100 (11.5) |
| Married, n (%) | 2 199 (88.7) | 2 404 (88.8) | 1 001 (86.6) | 754 (86.9) |
| No. of children, n (%) | ||||
| 0 | 123 (5.0) | 181 (6.7) | 41 (3.6) | 81 (9.3) |
| 1–3 | 1 881 (75.9) | 2 122 (78.4) | 858 (74.2) | 621 (71.5) |
| 4+ | 476 (19.2) | 403 (14.9) | 257 (22.2) | 166 (19.1) |
| Age at retirement, mean (SD) | 66.7 (2.5) | 66.8 (2.7) | 61.5 (3.0) | 61.5 (3.0) |
| Years employed 1967–2019, mean (SD) | 31.9 (9.1) | 40.1 (5.3) | 26.5 (8.4) | 34.8 (5.7) |
| Occupational physical demands, n (%) | ||||
| Low | 871 (35.1) | 1 009 (37.3) | 258 (22.3) | 168 (19.4) |
| Intermediate | 527 (21.3) | 1 095 (40.5) | 295 (25.5) | 429 (49.4) |
| High | 1 015 (40.9) | 587 (21.7) | 527 (45.6) | 254 (29.3) |
| HUNT1 & HUNT2 | ||||
| Hypertension, n (%) | 1 001 (40.4) | 1 398 (51.7) | 471 (40.7) | 489 (56.3) |
| Daily smoking, n (%) | 588 (23.7) | 668 (24.7) | 366 (31.7) | 263 (30.3) |
| Insufficient physical activity, n (%) | 1 057 (42.6) | 912 (33.7) | 551 (47.7) | 341 (39.3) |
| Obese, n (%) | 345 (13.9) | 264 (9.8) | 222 (19.2) | 156 (18.0) |
| High alcohol consumption, n (%) | 132 (5.3) | 306 (11.3) | 38 (3.3) | 89 (10.3) |
| Diabetes, n (%) | 28 (1.1) | 41 (1.5) | 22 (1.9) | 32 (3.7) |
| CVD, n (%) | 24 (1.0) | 78 (2.9) | 30 (2.6) | 74 (8.5) |
| No close friends, n (%) | 27 (1.1) | 85 (3.1) | 21 (1.8) | 35 (4.0) |
| HUNT4 70+ | ||||
| Dementia, n (%) | 190 (7.7) | 229 (8.5) | 135 (11.7) | 133 (15.3) |
| MCI, n (%) | 784 (31.6) | 984 (36.4) | 400 (34.6) | 380 (43.8) |
Note. HUNT: Trøndelag Health Study; The four surveys were: HUNT1: 1984–1986; HUNT2: 1995–1997; HUNT4 70+: 2017–2019; CVD: cardiovascular disease; MCI: mild cognitive impairment.
Of the HUNT4 70+ participants born 1933–1949 who had cognitive data and who retired before participation in HUNT4 70+ (n=8 217), 214 (2.6%) women versus 16 (0.2%) men were excluded because they had worked for less than one year in 1967–2019. A further 1,060 (12.9%) women versus 353 (4.3%) men were excluded because they stopped working before age 55. Women included in the study sample worked on average 8.6 fewer years and retired on average 0.5 years earlier than men, both p<0.001. Women were also more likely than men to retire via the disability pathway in an age- and education-adjusted model, odds ratio 1.38, 95% CI 1.24–1.54. In age- and education-adjusted models, a higher proportion of men (n=362, 10.1%) than women (n=325, 8.9%) were diagnosed with dementia, RRR 1.38, 95% CI 1.16–1.63. Similarly, a higher proportion of men (n=1 184, 32.6%) than women (n=1 364, 38.2%) were diagnosed with MCI, RRR 1.39, 95% CI 1.26–1.54.
Delaying retirement and risk of dementia and MCI
The two-way interactions between retirement age and the disability pathway (p=0.005) and between sex and the disability pathway (p=0.040) were statistically-significant predictors of dementia. For MCI, there were statistically significant two-way interactions between retirement age and sex (p=0.020), retirement age and the disability pathway (p=0.007), and between sex and the disability pathway (p=0.018). There were no statistically significant interactions between retirement age and occupational physical demands on either dementia or MCI (all ps>0.05).
Given the above interactions between retirement age, sex, and the disability pathway, further analyses were stratified by both sex and retirement pathway. For men who retired via the disability pathway, each year of postponed retirement was associated with a lower risk of dementia after adjusting for potential confounders (Model 4: RRR 0.92, Table 2). Retiring one life year later via the on-time or late standard pathway was associated with a lower risk of MCI for men, also after adjusting for potential confounders (Model 4: RRR 0.92, Table 2), but was not associated with dementia risk. For women, retiring one life year later via the disability pathway was not associated with a lower risk of dementia or MCI after adjusting for potential confounders (Model 4). Retiring one life year later via the early standard pathway to retirement was unrelated to either dementia or MCI in women and men (Table 2).
Table 2.
Relative risk ratio (RRR) and 95% confidence intervals (CI) for later-life dementia and MCI associated with retiring one life year later when following different retirement pathways, by sex.
| Women (n=3 636) | Men (n=3 574) | |||||
|---|---|---|---|---|---|---|
| Dementia RRR (95% CI) |
MCI RRR (95% CI) |
Dementia RRR (95% CI) |
MCI RRR (95% CI) |
|||
| Model 1 | ||||||
| Disability | 0.95 (0.89–1.02) | 0.94 (0.91–0.99) | 0.91 (0.84–0.97) | 0.95 (0.90–0.99) | ||
| Early standard <67 years | 1.09 (0.92–1.28) | 1.00 (0.89–1.12) | 0.84 (0.69–1.04) | 0.92 (0.83–1.03) | ||
| On-time or late standard 67+ years | 1.05 (0.95–1.17) | 1.05 (0.98–1.12) | 0.98 (0.90–1.07) | 0.92 (0.87–0.97) | ||
| Model 2 | ||||||
| Disability | 0.97 (0.90–1.04) | 0.96 (0.92–1.00) | 0.92 (0.85–0.99) | 0.95 (0.91–1.00) | ||
| Early standard <67 years | 1.11 (0.93–1.31) | 1.03 (0.92–1.16) | 0.91 (0.73–1.13) | 0.98 (0.88–1.09) | ||
| On-time or late standard 67+ years | 1.06 (0.96–1.17) | 1.05 (0.98–1.12) | 0.99 (0.90–1.08) | 0.92 (0.87–0.98) | ||
| Model 3 | ||||||
| Disability | 0.98 (0.91–1.05) | 0.96 (0.92–1.00) | 0.92 (0.85–0.99) | 0.96 (0.91–1.01) | ||
| Early standard <67 years | 1.10 (0.93–1.31) | 1.04 (0.92–1.17) | 0.92 (0.74–1.14) | 0.99 (0.89–1.11) | ||
| On-time or late standard 67+ years | 1.07 (0.96–1.19) | 1.04 (0.97–1.12) | 0.98 (0.90–1.08) | 0.92 (0.87–0.98) | ||
| Model 4 | ||||||
| Disability | 0.98 (0.91–1.06) | 0.96 (0.92–1.00) | 0.92 (0.85–0.99) | 0.95 (0.91–1.01) | ||
| Early standard <67 years | 1.07 (0.89–1.28) | 1.04 (0.92–1.17) | 0.93 (0.75–1.15) | 1.00 (0.89–1-12) | ||
| On-time or late standard 67+ years | 1.07 (0.97–1.19) | 1.04 (0.97–1.12) | 0.98 (0.90–1.08) | 0.92 (0.86–0.98) | ||
Note. 95% CI not including 1 are in bold. Potential confounders were added in successive models: age (model 1); education and occupational physical demands (model 2); marital status and number of children (model 3); smoking, high alcohol consumption, obesity, physical inactivity, hypertension, diabetes, CVD, and having at least one close friend (model 4).
Comparison of retirement pathway groups
Figure 2 and Table S1 (Appendix 2) display the results of the multinomial logistic regression analyses of the relationship between retirement group (disability, early standard [<67 years of age], late standard [after 67 years of age]; reference: on-time standard [at 67 years of age]) and dementia/MCI risk. The disability retirement group had an increased risk of dementia compared to the on-time standard group (Figure 2, Table S1). Results were similar for women and men and were apparent after adjusting for potential confounders (Model 4; women: RRR 1.64, 95% CI 1.14–2.37; men: RRR 1.70, 95% CI 1.17–2.48). For men, the late standard group had a lower risk of MCI than the on-time standard group, also after adjusting for potential confounders (Model 4; RRR 0.76, 95% CI 0.61–0.95). There was no evidence that dementia risk differed between the late standard and on-time standard groups for women or men. There was also no evidence that dementia/MCI risk differed between the early and on-time standard groups after adjusting for potential confounders (Figure 2, Table S1).
Figure 2.
Results from multinomial logistic regression showing relative risk ratios with corresponding 95% confidence intervals for the associations of retirement pathway with dementia and MCI risk in A) women and B) men. On-time standard retirement served as the reference category. Analyses were performed on a multiple imputed dataset and were IPW weighted. Potential confounders were added in successive models: age (Model 1); education and occupational physical demands (Model 2); marital status and number of children (Model 3); smoking, high alcohol consumption, obesity, physical inactivity, hypertension, diabetes, CVD, and having at least one close friend (Model 4).
The pattern of the results of the supplementary analyses based on (a) the complete cases without IPW and (b) only those participants who retired more than five years prior to participation in HUNT4 70+ were largely consistent with the results described above. Some effects that were statistically significant in the main analyses were not statistically significant in the supplementary analyses, which is not wholly unexpected given the smaller sample sizes. The full results of the supplementary analyses are available in Appendix 2.
Discussion
In this historical cohort study, we found that early retirement via the disability pathway - but not early retirement via the standard pathway - was associated with a higher risk of later-life dementia for women and for men. The association between disability retirement and later-life dementia remained even after adjusting for a range of potential confounders. Later retirement via the standard pathway was associated with a reduced MCI risk in men but was not associated with a lower dementia risk. We interpret our results as an indication that the overall association between retirement age and later-life dementia is primarily driven by an increased risk of later-life dementia among those who exit the workforce early (i.e., prior to becoming eligible for an old-age pension) due to work disability. Many Western countries, including Norway (26), have reduced incentives for early retirement. The current results suggest that reducing early retirement without improving people’s underlying work ability is unlikely to meaningfully affect the incidence of later-life cognitive impairment (see also (27)). Here we note that work ability depends on both an individual’s physical and mental capacities, and also on the specific demands and resources of their work context. Our results also suggest that the cognitive health of people who exit the workforce due to disability should be monitored more closely, as they appear to be at higher risk for later-life dementia. People who exit the workforce due to disability may be an appropriate target group for interventions that reduce the modifiable risk factors for later-life cognitive impairment (e.g., physical inactivity, smoking).
The observed associations between disability retirement and dementia could be linked to several health-related mechanisms. For instance, psychiatric disorders such as major depressive disorder have been linked to a higher dementia risk (28), although it is worth noting that depression also may be a part of the prodromal phase of dementia (28, 29). Multimorbidity, including neuropsychiatric, cardiovascular, sensory impairment, and cancer multimorbidity, is also linked to a higher dementia risk (30). Although we did not have data on the reason for disability in our study population, both psychiatric and cardiovascular diseases are common causes of disability among individuals aged 60–67 in Norway (16).
We did not find consistent evidence that retirement age was related to cognitive impairment for women and men who followed the standard pathway to retirement. There was, however, some indication that an older age of retirement among men retiring via the standard pathway after age 67 years was related to a lower risk for MCI. The stronger relationship between retirement age and men’s later-life cognition is consistent with previous studies (6, 10). Gender differences in post-retirement activities could potentially explain our findings (i.e., retired women might be more likely to engage in activities that stimulate cognitive health than retired men) (10), although this requires further investigation.
The present study has a number of strengths, including the large and diverse general population sample, rich survey data from three time points starting in 1984, register data on retirement age and sociodemographic characteristics, and thorough clinical assessments of dementia and MCI in the whole study sample. Nevertheless, the present study also has several limitations. Our sample was limited to those younger than age 86, which probably explains why the prevalence of dementia was not higher among the women in our sample (18). More women than men were excluded from the sample due to insufficient employment data, and stay-at-home women (and men) were also excluded. This may imply that our sample of women are not necessarily representative of the population of women aged 69–85 in Norway. Our results may not generalize to other countries due to differences in, for instance, disability and old-age retirement policies and benefits or norms for post-retirement life. A recent study found that post-retirement memory decline was faster in countries characterized by less generous welfare systems (31). Our results may not generalize to future generations given changes in the nature of work (increasing automatization, shift towards less physically demanding work)(32, 33) and the increase in women’s education and workforce participation. Further, excluding participants who retired within five years of the cognitive assessment did not modify our main findings, but we cannot rule out the possibility of reverse causality. Although no participants were excluded due to health issues, individuals who chose not to participate in the HUNT Study had, on average, higher mortality and higher prevalence of several chronic diseases than participants (34), suggesting that health-related selection bias may be present. We focused on the overall association between retirement age and cognitive impairment. It is presumed, however, that the consequences of working and retirement on later-life health and cognition depends on many different individual, occupational and societal factors (35), such as education (36), psychosocial working conditions (37), and the cognitive demands of the job (38, 39). Future research should explore whether there are particular subgroups for whom, or conditions under which, early or later retirement causally affects later-life cognitive impairment.
Supplementary Material
Acknowledgements
The HUNT Study is a collaboration between the HUNT Research Centre, Faculty of Medicine and Health Sciences, Norwegian University of Science and Technology (NTNU), the Trøndelag County Council, Central Norway Health Authority, and the Norwegian Institute of Public Health. We are grateful to the participants in the HUNT study and to the study management for the use of these data. EZ was responsible for the original draft and CB for the final draft of the manuscript. All authors contributed to the conceptualization, drafting, and reviewing of the final manuscript. EZ, BHS, and BB did the data curation and formal analyses. BHS, VS, AKH, CB, AJ, GS, HPK, JRH, SET, and YS acquired funding. EZ, BHS, BB, AKH, and VS worked with methodology. BHS, VS, and AKH were project administrators.
Declaration of Sources of Funding
This work was supported by the National Institutes of Health (grant number R01AG069109–01). This work was also partly supported by the Research Council of Norway through its Centers of Excellence funding scheme (project number 2262700). The funding agencies played no role in the design, execution, analysis or interpretation of data, or the writing of this manuscript.
Footnotes
Ethics approval
This study was approved by the HUNT study board of directors and the Regional Committee for Medical and Health Research Ethics. All participants in the HUNT Study gave a written informed consent upon participation.
Conflicts of interest
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
Data availability
The data used for this study were derived from The Trøndelag Health Study (HUNT), https://www.ntnu.edu/hunt. Any research group with a Principal Investigator affiliated with a Norwegian research institute can apply for access to analyze HUNT data. This means that research groups from non-Norwegian countries must find a collaboration partner in Norway to be able to use HUNT material. Each project needs to be approved by the HUNT Data Access Committee (DAC), Regional Committee for Medical and Health Research Ethics, and in some cases also the Data Inspectorate. Due to participant confidentiality, participant data are not publicly available.
References
- 1.GBD 2019 Dementia Forecasting Collaborators. Estimation of the global prevalence of dementia in 2019 and forecasted prevalence in 2050: an analysis for the Global Burden of Disease Study 2019. Lancet Public Health. 2022;7:e105–e125. doi: 10.1016/s2468-2667(21)00249-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Hultsch DF, Hertzog C, Small BJ, Dixon RA. Use it or lose it: engaged lifestyle as a buffer of cognitive decline in aging? Psychol Aging. 1999;14:245–263. doi: 10.1037//0882-7974.14.2.245 [DOI] [PubMed] [Google Scholar]
- 3.Sundström A, Rönnlund M, Josefsson M. A nationwide Swedish study of age at retirement and dementia risk. Int J Geriatr Psychiatry. 2020;35:1243–1249. doi: 10.1002/gps.5363 [DOI] [PubMed] [Google Scholar]
- 4.Lupton MK, Stahl D, Archer N, Foy C, Poppe M, Lovestone S, et al. Education, occupation and retirement age effects on the age of onset of Alzheimer’s disease. Int J Geriatr Psychiatry. 2010;25:30–36. doi: 10.1002/gps.2294 [DOI] [PubMed] [Google Scholar]
- 5.Grotz C, Letenneur L, Bonsang E, Amieva H, Meillon C, Quertemont E, et al. Retirement age and the age of onset of Alzheimer’s disease: results from the ICTUS study. PLoS One. 2015;10:e0115056. doi: 10.1371/journal.pone.0115056 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Dufouil C, Pereira E, Chêne G, Glymour MM, Alpérovitch A, Saubusse E, et al. Older age at retirement is associated with decreased risk of dementia. Eur J Epidemiol. 2014;29:353–361. doi: 10.1007/s10654-014-9906-3 [DOI] [PubMed] [Google Scholar]
- 7.Xue B, Cadar D, Fleischmann M, Stansfeld S, Carr E, Kivimäki M, et al. Effect of retirement on cognitive function: the Whitehall II cohort study. Eur J Epidemiol. 2018;33:989–1001. doi: 10.1007/s10654-017-0347-7 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Finkel D, Andel R, Gatz M, Pedersen NL. The role of occupational complexity in trajectories of cognitive aging before and after retirement. Psychol Aging. 2009;24:563–573. doi: 10.1037/a0015511 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Bonsang E, Adam S, Perelman S. Does retirement affect cognitive functioning? J Health Econ. 2012;31:490–501. doi: 10.1016/j.jhealeco.2012.03.005 [DOI] [PubMed] [Google Scholar]
- 10.Atalay K, Barrett GF, Staneva A. The effect of retirement on elderly cognitive functioning. J Health Econ. 2019;66:37–53. doi: 10.1016/j.jhealeco.2019.04.006 [DOI] [PubMed] [Google Scholar]
- 11.Meng A, Nexø MA, Borg V. The impact of retirement on age related cognitive decline - a systematic review. BMC Geriatr. 2017;17:160. doi: 10.1186/s12877-017-0556-7 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Scharn M, Sewdas R, Boot CRL, Huisman M, Lindeboom M, van der Beek AJ. Domains and determinants of retirement timing: A systematic review of longitudinal studies. BMC Public Health. 2018;18:1083. doi: 10.1186/s12889-018-5983-7 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Livingston G, Huntley J, Sommerlad A, Ames D, Ballard C, Banerjee S, et al. Dementia prevention, intervention, and care: 2020 report of the Lancet Commission. Lancet. 2020;396:413–446. doi: 10.1016/S0140-6736(20)30367-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Eurostat. Disability and social participation in Europe. Luxembourg: Office for Official Publications of the European Communities; 2001. [Google Scholar]
- 15.Statistics Norway. Recipients of disability benefit. Oslo: Statistics Norway. [Google Scholar]
- 16.Norwegian Labour and Welfare Administration. Mottakere av uføretrygd etter hoveddiagnose (primærdiagnose), alder og kjønn. [Google Scholar]
- 17.Chêne G, Beiser A, Au R, Preis SR, Wolf PA, Dufouil C, et al. Gender and incidence of dementia in the Framingham Heart Study from mid-adult life. Alzheimers Dement. 2015;11:310–320. doi: 10.1016/j.jalz.2013.10.005 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Gjøra L, Strand BH, Bergh S, Borza T, Braekhus A, Engedal K, et al. Current and Future Prevalence Estimates of Mild Cognitive Impairment, Dementia, and Its Subtypes in a Population-Based Sample of People 70 Years and Older in Norway: The HUNT Study. J Alzheimers Dis. 2021;79:1213–1226. doi: 10.3233/JAD-201275 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Petersen RC, Roberts RO, Knopman DS, Geda YE, Cha RH, Pankratz VS, et al. Prevalence of mild cognitive impairment is higher in men. The Mayo Clinic Study of Aging. Neurology. 2010;75:889–897. doi: 10.1212/WNL.0b013e3181f11d85 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Roberts RO, Geda YE, Knopman DS, Cha RH, Pankratz VS, Boeve BF, et al. The incidence of MCI differs by subtype and is higher in men: the Mayo Clinic Study of Aging. Neurology. 2012;78:342–351. doi: 10.1212/WNL.0b013e3182452862 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Krokstad S, Langhammer A, Hveem K, Holmen TL, Midthjell K, Stene TR, et al. Cohort Profile: the HUNT Study, Norway. Int J Epidemiol. 2013;42:968–977. doi: 10.1093/ije/dys095 [DOI] [PubMed] [Google Scholar]
- 22.Åsvold BO, Langhammer A, Rehn TA, Kjelvik G, Grøntvedt TV, Sørgjerd EP, et al. Cohort Profile Update: The HUNT Study, Norway. Int J Epidemiol. 2022. doi: 10.1093/ije/dyac095 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Eyjólfsdóttir HS, Baumann I, Agahi N, Lennartsson C. How to Measure Retirement Age? A Comparison of Survey and Register Data. Journal of Population Ageing. 2021;14:143–161. doi: 10.1007/s12062-019-09254-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Guha M Diagnostic and Statistical Manual of Mental Disorders: DSM-5 (5th edition). Reference Reviews. 2014;28:36–37. doi: 10.1108/RR-10-2013-0256 [DOI] [Google Scholar]
- 25.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;119:252–265. doi: 10.1111/j.1600-0447.2008.01326.x [DOI] [PubMed] [Google Scholar]
- 26.Midtsundstad T Flexible retirement in Norway. Flexible Retirement Following the Scandinavian Model. German Bundestag, Berlin; 2014. [Google Scholar]
- 27.Denier N, Clouston SAP, Richards M, Hofer SM. Retirement and Cognition: A Life Course View. Adv Life Course Res. 2017;31:11–21. doi: 10.1016/j.alcr.2016.10.004 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Byers AL, Yaffe K. Depression and risk of developing dementia. Nat Rev Neurol. 2011;7:323–331. doi: 10.1038/nrneurol.2011.60 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Bennett S, Thomas AJ. Depression and dementia: cause, consequence or coincidence? Maturitas. 2014;79:184–190. doi: 10.1016/j.maturitas.2014.05.009 [DOI] [PubMed] [Google Scholar]
- 30.Grande G, Marengoni A, Vetrano DL, Roso-Llorach A, Rizzuto D, Zucchelli A, et al. Multimorbidity burden and dementia risk in older adults: The role of inflammation and genetics. Alzheimers Dement. 2021;17:768–776. doi: 10.1002/alz.12237 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Mäcken J, Riley AR, Glymour MM. Cross-national Differences in the Association Between Retirement and Memory Decline. J Gerontol B Psychol Sci Soc Sci. 2021;76:620–631. doi: 10.1093/geronb/gbaa223 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Nabe-Nielsen K, Holtermann A, Gyntelberg F, Garde AH, Islamoska S, Prescott E, et al. The effect of occupational physical activity on dementia: Results from the Copenhagen Male Study. Scand J Med Sci Sports. 2021;31:446–455. doi: 10.1111/sms.13846 [DOI] [PubMed] [Google Scholar]
- 33.Hai Y, Xue Y, Wang YH. Does Long-Term Shift Work Increase the Risk of Dementia? A Systematic Review and Meta-Analysis. Am J Alzheimers Dis Other Demen. 2022;37:15333175221141535. doi: 10.1177/15333175221141535 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Langhammer A, Krokstad S, Romundstad P, Heggland J, Holmen J. The HUNT study: participation is associated with survival and depends on socioeconomic status, diseases and symptoms. BMC Med Res Methodol. 2012;12:143. doi: 10.1186/1471-2288-12-143 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Kuhn A The complex effects of retirement on health.: IZA World of Labor; 2018. doi:doi: 10.15185/izawol.430 [DOI] [Google Scholar]
- 36.Hale JM, Bijlsma MJ, Lorenti A. Does postponing retirement affect cognitive function? A counterfactual experiment to disentangle life course risk factors. SSM Popul Health. 2021;15:100855. doi: 10.1016/j.ssmph.2021.100855 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Fleischmann M, Xue B, Head J. Mental Health Before and After Retirement—Assessing the Relevance of Psychosocial Working Conditions: The Whitehall II Prospective Study of British Civil Servants. The Journals of Gerontology: Series B. 2019;75:403–413. doi: 10.1093/geronb/gbz042 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Carr DC, Willis R, Kail BL, Carstensen LL. Alternative Retirement Paths and Cognitive Performance: Exploring the Role of Preretirement Job Complexity. The Gerontologist. 2019;60:460–471. doi: 10.1093/geront/gnz079 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Jung D, Lee J, Meijer E. Revisiting the effect of retirement on Cognition: Heterogeneity and endowment. The Journal of the Economics of Ageing. 2022;21:100361. doi: 10.1016/j.jeoa.2021.100361 [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The data used for this study were derived from The Trøndelag Health Study (HUNT), https://www.ntnu.edu/hunt. Any research group with a Principal Investigator affiliated with a Norwegian research institute can apply for access to analyze HUNT data. This means that research groups from non-Norwegian countries must find a collaboration partner in Norway to be able to use HUNT material. Each project needs to be approved by the HUNT Data Access Committee (DAC), Regional Committee for Medical and Health Research Ethics, and in some cases also the Data Inspectorate. Due to participant confidentiality, participant data are not publicly available.