Y. Zhao*,1, W. Zhou*,2, M. Xing1, L. Zhang1, Y. Tong1, X. Lv3, Y. Ma2, W. Li1
1. Center of Clinical Big Data and Analytics of The Second Affiliated Hospital and Department of Big Data in Health Science School of Public Health, Zhejiang University School of Medicine, Hangzhou, 310058, Zhejiang, China; 2. Department of Biostatistics and Epidemiology, School of Public Health, China Medical University, Shenyang, China; 3. Peking University Institute of Mental Health (Sixth Hospital), National Clinical Research Center for Mental Disorders, NHC Key Laboratory of Mental Health (Peking University), 51 Huayuan North Road, Haidian District, Beijing, China; *Co-first Author
Corresponding Author: Wenyuan Li, Center of Clinical Big Data and Analytics of The Second Affiliated Hospital and Department of Big Data in Health Science School of Public Health, Zhejiang University School of Medicine, Hangzhou, 310058, Zhejiang, China Tel: +86-19186432600, E-mail: wenyuanli@zju.edu.cn. Yanan Ma, Department of Biostatistics and Epidemiology, School of Public Health, China Medical University, No. 77 Puhe Road, Shenyang North New Area, Shenyang 110122, Liaoning Province, China, Tel: +86-24-31939406, E-mail: ynma@cmu.edu.cn. Xiaozhen Lv, Peking University Institute of Mental Health (Sixth Hospital), National Clinical Research Center for Mental Disorders, NHC Key Laboratory of Mental Health (Peking University), 51 Huayuan North Road, Haidian District, Beijing, China, Tel: +86-10-62723725, E-mail: lvxiaozhen@bjmu.edu.cn.
J Prev Alz Dis 2024;
Published online May 28, 2024, http://dx.doi.org/10.14283/jpad.2024.96
Abstract
BACKGROUND AND OBJECTIVES: To identify cognitive decline trajectories in a Chinese elderly population, explore the associations between these trajectories and mortality, and further identify risk factors related to certain trajectories of cognitive decline.
DESIGN: Prospective cohort study.
SETTING: The group-based trajectory modeling and Cox proportional hazards models were conducted to explore the association between cognitive trajectory groups and mortality, while multinomial logistic regression models were constructed to estimate potential risk factors.
PARTICIPANTS: We included 7082 participants aged 65 years or above in three consecutive but non-overlapping cohorts of the Chinese Longitudinal Healthy Longevity Survey with the Chinese version of the Mini-Mental State Examination up to 6 years. Participants were subsequently followed for a median (IQR) of 2.89 (1.38-3.12) years to obtain their survival status and date of death.
MEASUREMENTS: Chinese version of the Mini-Mental State Examination was used to measure participants’ cognitive function.
RESULTS: Through use of group-based trajectory modeling, we determined three cognitive trajectory groups. Then, after adjusting for confounding factors, we found a monotonic and positive association between cognitive decline and mortality risk. Meanwhile, the association varied among elderly populations in different age groups and BMI categories, but did not differ by sex, smoking, drinking and exercising. Older seniors, females and those with poorer baseline cognitive function and less social participation tended to be more likely to be in the unfavorable trajectory groups.
CONCLUSION: We found that the faster the cognitive decline, the higher the mortality, especially among those aged 65-79 years and those overweight. Our findings suggested the importance of implement better monitoring of the cognitive function of the elderly population.
Key words: Cognitive function, trajectory group, group-based trajectory modeling, mortality, Chinese older adults.
Abbreviations: GBTM: Group-based trajectory modeling; MMSE: Mini-Mental State Examination; CLHLS: Chinese Longitudinal Healthy Longevity Survey; CMMSE: Chinese version of the Mini-Mental State Examination; BIC: Bayesian Information Criterion; BMI: Body mass index; AvePP: Average posterior probability; HR: Hazard ratios; CI: Confidence interval; OR: Odds ratio.
Introduction
It is estimated by the United Nations that, by 2050, the number of people aged over 60 will increase to around 2 billion worldwide (1). The number of older people is growing at a remarkable rate and will accelerate continuously, especially in developing countries. The aging of the population is gradually aggravating, and the incidence of chronic diseases, cognitive impairment, functional disability, and mental disorders is on the rise. It is considered that 15.07 million Chinese people age 60 and older are currently living with dementia, and the prevalence of dementia is believed to double at 5-year intervals (2), posing a huge threat to public health.
Cognitive decline refers to the gradual deterioration of cognitive functions such as memory, attention, and reasoning, often associated with aging but also influenced by various environmental and genetic factors (3-7). It is a dynamic process that varies across individuals. Findings from existing studies indicated that the majority of elderly individuals either did not experience cognitive decline or exhibited only a slow decline in cognitive function, while a minority experienced rapid cognitive decline (3-6). Emerging studies have also highlighted associations between cognitive decline and mortality rates (8-10). Hu et al. divided elderly individuals aged 80 and above into four groups based on their rates of cognitive decline using the GBTM model, followed up with each group and found that the trajectory of mortality risk roughly followed a hierarchical structure consistent with cognitive trajectories (8). Lv et al. divided the elderly population into three groups and discovered a positive association between the rate of change in MMSE scores over three years and subsequent three-year mortality rates using Cox proportional hazards models (9). However, current research on cognitive decline and mortality rates either lacks long-term observational data or lacks quantitative description of the associations.
Therefore, we hypothesized that longitudinal changes in cognitive function, represented by distinctive trajectories during the late life with differed risk of death, and conducted extensive analyses among a Chinese population with prolonged follow-up period. By leveraging longitudinal data and quantitative analyses, we aimed to provide a deeper understanding of how cognitive decline influences mortality outcomes among older adults. We also investigated whether certain sociodemographic characteristics, lifestyle factors, and health status may predispose individuals to follow certain trajectories of cognitive decline.
Methods
Study Design and Participants
The data used in this study were from the Chinese Longitudinal Healthy Longevity Study (CLHLS), a nationally representative survey of Chinese adults over 65 years. From the start of the project in 1998, CLHLS has been conducted in 866 highly diverse counties and cities selected from 23 provinces of China through a multi-stage random sampling method with unequal proportion, covering 85% of the elderly population in China (9, 11). Eight waves were conducted in 1998, 2000, 2002, 2005, 2008–2009, 2011–2012, 2014, and 2018–2019, respectively. Face-to-face interviews were conducted with a response rate of approximately 88 to 90 percent across all waves. The quality of the data has been verified in previous studies (12).
We identified 8323 participants who had three consecutive survey data available and were followed for at least three years after the third visit. That was, for an individual enrolled in 2002, the person would have two follow-up visits in 2005 and 2008, and then would be followed for 1095 days (3 years). We further excluded 154 participants who were younger than 65 years at baseline, 664 participants who did not complete the Chinese version of the Mini-Mental State Examination (CMMSE) scale at least once, and 423 participants who were already dementia at baseline. In the end, a total of 7082 participants were included in our analysis, with 3699 from the 2002 cohort, 1456 from the 2005 cohort and 1927 from the 2008 cohort (Fig S1).
The CLHLS study was approved by the research ethics committees of Peking University (IRB00001052-13074), and each participant signed an informed consent form for approval to participate in the study.
Outcome
The participants’ survival status and the date of death were identified through interviews with close family members or the village doctors in each cohort (9). We followed up for an additional three years after the last interview to obtain the specific time of death for each participant. The survival time was calculated by subtracting the date of the last interview from the date of death.
Cognitive function assessment
In this study, we used the CMMSE to assess the cognitive function of each participant, which evaluated five cognitive domains: orientation, registration, attention or calculation, recall and language. The validity and reliability of the CMMSE have been demonstrated in previous literature (11-13). The CMMSE scores ranged from 0 to 30, with lower scores indicating poorer cognitive function. Taking into account the educational background of the participants, we defined those with an MMSE score < 18 as poor cognitive function and those with a score ≥ 18 as normal cognitive function (14, 15).
Potential risk factors and covariates
Potential risk factors of cognitive ageing and covariates of proportional risk regression included three main categories: sociodemographic characteristics, lifestyle factors, and health status factors. Data for these three categories of variables were collected via a self-administered questionnaire at baseline, with trained staff providing detailed information if needed. Among them, sociodemographic characteristics included age, sex (male/female), years of schooling, residence (city or town/rural), marital status (living with spouse/widowed or separated), and body mass index (BMI, underweight (<18.5 kg/m2)/normal (18.5 kg/m2-24 kg/m2)/overweight (>24 kg/m2)). Lifestyle factors included smoking (yes/no), drinking (yes/no), exercising (yes/no), and social activities (16). Health factors included psychological factors (17), hypertension (yes/no), diabetes (yes/no), heart diseases (yes/no) and cerebrovascular disease (yes/no). BMI was calculated by dividing weight in kilograms by height in meters squared.
Social activities consisted of seven variables, namely filed work, garden work, reading newspapers or books, raising domestic animals or pets, playing cards or mah-jongg, watching TV or listening to radio and taking part in some other activities. Scores ranged from 7 to 35, with higher scores indicating higher engagement (16).
Psychological factors, represented by psychological well-being, consisted of seven mental health-related questions: “Do you always look on bright side of things?”, “Do you always keep your belongings neat and clean?”, “Do you feel fearful or anxious?”, “Do you feel lonely and isolated?”, “Can you often make own decision?”, “Do you feel useless with age?” and “Are you always happy as younger?” The calculation was performed according to previous studies (17).
Statistical analysis
Descriptive analyses were presented as mean (standard deviation) and frequency (%) for continuous and categorical variables respectively. For missing values of continuous and categorical variables, we utilized the random forest imputation method. Additionally, correlation analysis was performed between independent variables to ensure no obvious collinearity.
Group-based trajectory modeling (GBTM) is a semi-parametric empirical model, which assumes the sample is composed of different subpopulations and divides the participants into different trajectory groups based on the maximum likelihood ratio (18, 19). We compared models of 1 to 4 trajectory groups to determine the optimal number of trajectory groups based on the log Bayes factor (2loge(B10)), calculated as 2(ΔBIC), where ΔBIC was the difference between the BIC of the alternative (more complex) model and the BIC of the null (simpler) model. This factor could be interpreted as the strength of evidence favoring the alternative model. Then, to determine the optimal trajectory shape that best described the observed trajectories, linear, quadratic, and cubic functions were all tried. The best-fitting trajectory groups were selected based on the log Bayes factor, average posterior probability (AvePP, above 0.7 indicated a better fit), group membership probability (each trajectory group should hold a group membership probability ≥ 5%) and statistical measures (the p-values of model parameters) (8, 20-24). Finally, three trajectory curves, respectively linear, quadratic, and linear, were considered to provide the best fitting result (Table S1 and Fig 1).
Figure 1. The optimal trajectory curves of cognitive decline over a six-year period
The solid lines (green: minimal decline; red: moderate decline; blue: rapid decline) represent mean estimated values, and the dashed lines represent 95% CIs
Cox proportional hazard models were constructed to determine the association of different cognitive function trajectory groups with mortality. Data were reported as adjusted hazard ratios (HR) and 95% confidence intervals (CI). Variable adjustments were accomplished via three stages: (1) we mainly adjusted for demographic characteristics including age, sex, years of schooling, residence, marital status, BMI and the baseline MMSE scores; (2) we added lifestyle factors including smoking, drinking, exercising, and social activities; (3) we additionally adjusted for health status including psychological factors, hypertension, diabetes, heart disease and cerebrovascular disease. In addition, the proportional hazards assumption was validated by tests based on Schoenfeld residuals, and the results showed that the assumption was not violated.
To assess the roles of potential risk factors in the different trajectory groups, multinomial logistic regression models were constructed to estimate odds ratios (ORs) and 95% CIs for the associations between risk factors and trajectory groups.
Tests for the interactions between changes in cognitive function and factors of interest were performed by individually adding an interaction term to the fully adjusted model. In this study, we explored six factors of interest: age, sex, BMI, smoking, drinking, and exercising. Subsequently, if the interaction term yielded statistical significance, we conducted stratified analyses based on the factor of interest.
Two-sided p-values < 0.05 were considered statistically significant. Data processing was conducted in Python (version 3.10.2), while the construction of the GBTM and all data analysis were performed in Stata.16.0 and R version 4.2.2.
Results
Descriptive characteristics
The baseline characteristics of the participants were presented in Table 1. At baseline, a total of 7082 elderly participants including 3445 men and 3637 women aged 65 years or older (mean age was 77.06 years old), were enrolled in this research. According to the CMMSE scale, the mean CMMSE scores of the three follow-up visits were 27.34 (SD=2.93), 26.11 (SD=5.16), 24.49 (SD=6.95) respectively. Subsequently, during a median (IQR) follow-up of 2.89 (1.38-3.12) years, 1828 deaths occurred (Fig S2).
Table 1. The baseline characteristics of participants under different MMSE trajectory curves
Note: Continuous variables were presented as mean (standard deviation), and categorical variables were expressed as frequency (%). The analysis of variance tests were used for continuous variables with homogeneous variances, the Mann-Whitney U tests were used for continuous variables with unequal variances and the chi-square tests were used for categorical variables. Note: There were missing data on years of schooling (n = 18), BMI (n = 31), smoking (n = 6), drinking (n = 9), exercise (n = 9), social activities (n = 1), psychological factors (n = 233), which were then imputed with random forest. Abbreviations: BMI: Body Mass Index, MMSE: Mini-Mental State Examination
Cognitive function trajectory modeling
Based on the fitting curve results, we categorized the cognitive function trajectories into three groups. (1) The trajectory curve of a group showed a smooth and higher MMSE score, representing a slower downward trend (minimal decline, 79.9%) with the MMSE score slightly changing from 27.83 to 27.01 in the three follow-up visits. (2) Another group showed a moderate downward trend (moderate decline, 13.5%) with the MMSE score falling from 24.87 to 18.41 in the three follow-up visits. (3) The remaining group showed a faster downward trend (rapid decline, 6.6%), with the MMSE scores dropping from 25.73 points to 4.33 points in the three follow-up visits. Meanwhile, descriptive statistics were performed on the participants within different cognitive trajectory groups, as shown in Table 1.
Association between cognitive function trajectory groups and mortality
Table 2 showed the association between different cognitive trajectory groups and mortality, which indicated that the faster the cognitive decline, the higher the mortality rate. Compared with the minimal decline group, mortality was 47% higher in the moderate decline group (HR = 1.47, 95% CI = (1.22, 1.78)), while mortality was 124% higher in the rapid decline group (HR = 2.24, 95% CI = (1.77, 2.84)).
Table 2. Association between cognitive trajectory groups and mortality under stratified models
Note: model 1: we adjusted for age, sex, education, residence, marital status, BMI, baseline MMSE score; model 2: on the basis of model 1, we added smoking, drinking, exercising, and social activities; model 3: we additionally adjusted for psychological factors, hypertension, diabetes, heart disease and cerebrovascular disease. Abbreviations: HR: hazard ratio, CI: confidence interval.
Associations of potential risk factors with trajectory groups
Table 3 presented the multinomial logistic regression that related risk factors to different cognitive trajectory groups, with the minimal decline as the reference. We found that age, sex, baseline MMSE scores, social activities were significantly associated with the other two trajectory groups. Older adults, females were more likely than younger individuals and males to be in the less favorable trajectory groups, while subjects with better baseline levels of cognitive function and greater participation in social activities tended to be in the favorable trajectory groups in comparison with the minimal decline group.
Table 3. The roles of potential risk factors in the different trajectory groups by multinomial logistic regression (reference = the minimal decline)
Note: We adjusted for age, sex, education, residence, marital status, BMI, baseline MMSE score, smoking, drinking, exercising, social activities, psychological factors, hypertension, diabetes, heart disease and cerebrovascular disease. **: p<0.01; *: p<0.05
Subgroup analysis
We performed subgroup analysis to further investigate the joint associations of the risk factors of interest and different cognitive trajectory groups with the risk of mortality. We found that the associations between different trajectory groups and mortality varied between different age groups (p-interaction = 0.018) and BMI categories (p–interaction = 0.007), but did not differ by sex, smoking, drinking and exercising (Fig 2). In the age-stratified model, among individuals aged 65-79, compared to the minimal decline group, the HRs of all-cause death was 1.92 (95%CI, 1.31-2.83) for those in the moderate decline group, while for the rapid decline group, the HRs was 5.15 (95%CI, 3.08-8.62). Among individuals aged 80 or above, the HRs was 1.55 (95%CI, 1.34-1.81) for those in the moderate decline group compared to the minimal group, and for those in the rapid decline group, the HRs was 2.47 (95%CI, 2.10-2.89). In the BMI-stratified model, among individuals classified as underweight, the HRs of all-cause death was 1.43 (95%CI, 1.13-1.82) for those in the moderate decline group compared to the minimal decline group, while for the rapid decline group, the HRs was 2.08 (95%CI, 1.59-2.71). Among the population with normal weight, the HRs was 1.63 (95%CI, 1.29-2.07) for individuals in the moderate decline group compared to the minimal decline group, and 2.11 (95%CI, 1.59-2.82) for those in the rapid decline group. Among the overweight population, compared to the minimal decline group, the HRs of all-cause death was 2.37 (95%CI, 1.60-3.51) for those in the moderate decline group, and 3.50 (95%CI, 2.27-5.40) for those in the rapid decline group.
Figure 2. Joint association of predictors interested and trajectory groups with the risk of mortality
We adjusted for age, sex, education, residence, marital status, BMI, baseline MMSE score, smoking, drinking, exercising, social activities, psychological factors, hypertension, diabetes, heart disease and cerebrovascular disease.
Discussion
In this study, we identified three distinct cognitive trajectory groups and observed a consistent and positive association between cognitive decline and mortality risk after accounting for confounding factors. The associations varied across age subgroups and BMI subgroups. Older seniors, females, and those with lower baseline cognitive function and reduced social engagement were more likely to be in the less favorable cognitive trajectory groups.
Our study provided new evidence and offered insights into the intricate relationships between cognitive decline and mortality among a Chinese population, complementing previous research globally (25-28). First, in contrast to another study (8), which primary focused on cognitive trajectories in a Chinese population age 80 and above, our study population encompassed a wider range of ages, addressing a notable gap and demonstrating that the associations between cognitive decline and death risk exhibited unique characteristics in different age groups. Second, building upon the research by Lv et al., we extended the follow-up period to observe the cognitive decline trajectories of the elderly population over a longer duration (9). This extended observation period enhanced the stability of the obtained trajectory groups, allowing for a more comprehensive exploration of the relationship between cognitive decline trajectories and mortality. Furthermore, our study quantitatively assessed the associations between cognitive function decline and mortality rates, a crucial aspect that was overlooked in some prior research (8, 10), providing valuable insights into the associations between cognitive decline and mortality outcomes.
In our study, elderly individuals were categorized into three distinct trajectory groups based on their longitudinal cognitive status: one group showed minimal decline, suggesting relatively stable cognitive function over time; another group exhibited moderate cognitive decline, indicating more notable changes in cognitive function; and the third group exhibited rapid cognitive decline warranting immediate attention. This confirmed our hypothesis that changes in late-life cognitive function could be represented by more than one trajectory (29). This cognitive decline was closely associated with elevated mortality rates, as individuals in the rapidly declining group demonstrated a 124% higher mortality rate in comparison to those with minimal decline. This underscored the critical role of cognitive assessments as early indicators of overall health risks for healthcare practitioners, which was consistent with previous studies employing GBTM methodology (30) or nonnormal Growth Mixture Models (GMM) (31).
By employing multinomial logistic regression analysis, we further identified age, gender, education, baseline cognitive status, and lifestyle choices as key determinants of cognitive decline in this aging population. The gender disparity in cognitive decline was noteworthy, with a significantly lower proportion of males in both moderate and rapid decline groups, suggesting that males may possess certain advantages in maintaining cognitive function in later life. It should be noted that in our study population, men were more likely to have higher levels of education than women, especially those living in rural areas (32), which may contribute to these differences. Additionally, our results highlighted the role of social activities in promoting better cognitive function, aligning with the established concept of the bidirectional relationship between intellectual activities and cognitive function (33). These findings were further supported by a US nationwide cohort (34), revealing that more frequent leisure-time physical activity was associated with better cognitive function. In our stratified analyses, we observed consistent associations between cognitive decline and mortality rates across different age and BMI groups, with the association appearing particularly pronounced in the 65-79 age group and among individuals classified as overweight. However, considering the relatively small sample sizes within these subgroups, further studies with larger sample sizes are necessary to validate our findings. Nevertheless, our findings emphasized the necessity for the identification of high-risk populations, sustained monitoring, and personalized strategies for cognitive health maintenance and intervention (35).
Strengths of this study include the relatively large sample size, as well as repeated cognitive assessments that enabled trajectory modeling across 6 years. Examining the trajectory profiles of multiple cognitive tests using a data-driven approach allowed for a robust comparison across cognitive domains regarding their associations with mortality and therefore helped to inform targeted strategies for healthy aging. In addition, we visualized the developmental patterns of cognitive changes on a timeframe along with detailed trajectory statistics for each distinct subgroup, which highlighted the continuous and dynamic traits of cognitive decline.
Our study was also subjected to a few limitations. First, we did not include several potential confounding covariates such as inflammation status, sleep quality, and medication usage due to data availability. Second, similar to any observational studies, we cannot rule out the possibility of residual confounding. Third, since our study was conducted among a Chinese population, our findings might not be applicable to populations of different ethnicity or socioeconomic backgrounds.
Conclusions
We divided three distinct trajectory groups based on longitudinal cognitive status among a Chinese population, and found a monotonic and positive association between faster cognitive decline and higher mortality rates. Further exploration showed that the association varied across different age and BMI groups. In addition, older seniors and females were more likely to be in the unfavorable trajectory groups, while, conversely, participants with better baseline cognitive function and more social activities were less likely to be in these groups. Our findings highlight the importance of longitudinal monitoring of cognitive status among the elderly and encourage comprehensive personalized intervention strategies to maintain cognitive health.
Ethics approval and consent to participate: Peking University Research Center approved the CLHLS for Healthy Ageing and Development/National Development Institute, and all participants provided written informed consent.
Declaration of Competing Interest: The authors have no conflicts of interest to declare.
Contributions: XL, YM and WL conceived and directed the research. YZ, WZ and MX contributed to the data management and data analysis. YZ and WZ drafted the manuscript. XL, YM, WL, YT, MX and LZ revised the manuscript. All authors have approved the final manuscript. YZ and WZ contributed equally as co-first authors. XL, YM and WL contributed equally as the co-corresponding authors.
Acknowledgments: We acknowledge the Center for Healthy Aging and Development Studies for providing us with the data from CLHLS.
Funding: This study was supported by the National Natural Science Foundation of China (82203984, 82003539), Key Laboratory of Intelligent Preventive Medicine of Zhejiang Province (2020E10004), Leading Innovative and Entrepreneur Team Introduction Program of Zhejiang (2019R01007), Key Research and Development Program of Zhejiang Province (2020C03002), and Healthy Zhejiang One Million People Cohort (K-20230085).
References
1. Bloom DE, Luca DL. Chapter 1 – The Global Demography of Aging: Facts, Explanations, FutureIn Handbook of the Economics of Population Aging, Piggott J, Woodland A, eds. North-Holland, 2016;pp. 3-56. DOI: https://doi.org/10.1016/bs.hespa.2016.06.002
2. Jia L, Du Y, Chu L, et al. Prevalence, risk factors, and management of dementia and mild cognitive impairment in adults aged 60 years or older in China: a cross-sectional study. The Lancet Public Health 2020;5, e661-e671. DOI: https://doi.org/10.1016/S2468-2667(20)30185-7
3. Breitve MH, Chwiszczuk LJ, Hynninen MJ, et al. A systematic review of cognitive decline in dementia with Lewy bodies versus Alzheimer’s disease. Alzheimers Res Ther 2014;6, 53. DOI: https://doi.org/10.1186/s13195-014-0053-6
4. Ma C, Li M, Wu C. Cognitive Function Trajectories and Factors among Chinese Older Adults with Subjective Memory Decline: CHARLS Longitudinal Study Results (2011-2018). Int J Environ Res Public Health 2022;19. DOI: https://doi.org/10.3390/ijerph192416707
5. Hayden KM, Reed BR, Manly JJ, et al. Cognitive decline in the elderly: an analysis of population heterogeneity. Age Ageing 2011;40, 684-9. DOI: https://doi.org/10.1093/ageing/afr101
6. Pietrzak RH, Lim YY, Ames D, et al. Trajectories of memory decline in preclinical Alzheimer’s disease: results from the Australian Imaging, Biomarkers and Lifestyle Flagship Study of ageing. Neurobiol Aging 2015;36, 1231-8. DOI: https://doi.org/10.1016/j.neurobiolaging.2014.12.015
7. Small BJ, Backman L. Longitudinal trajectories of cognitive change in preclinical Alzheimer’s disease: a growth mixture modeling analysis. Cortex 2007;43, 826-34. DOI: https://doi.org/10.1016/S0010-9452(08)70682-8
8. Hu X, Gu S, Sun X, et al. Cognitive ageing trajectories and mortality of Chinese oldest-old. Archives of Gerontology and Geriatrics 2019;82, 81-87. DOI: https://doi.org/10.1016/j.archger.2019.01.018
9. Lv X, Li W, Ma Y, et al. Cognitive decline and mortality among community-dwelling Chinese older people. BMC Med 2019;17, 63. DOI: https://doi.org/10.1186/s12916-019-1295-8
10. Marioni RE, Proust-Lima C, Amieva H, et al. Cognitive lifestyle jointly predicts longitudinal cognitive decline and mortality risk. Eur J Epidemiol 2014;29, 211-9. DOI: https://doi.org/10.1007/s10654-014-9881-8
11. Zeng Y, Feng Q, Hesketh T, Christensen K, Vaupel JW. Survival, disabilities in activities of daily living, and physical and cognitive functioning among the oldest-old in China: a cohort study. Lancet 2017;389, 1619-1629. DOI: https://doi.org/10.1016/S0140-6736(17)30548-2
12. Yi Z, Jr D, Vlosky DA, Gu D. Healthy Longevity in China: Demographic, Socioeconomic, and Psychological Dimensions, 2008. DOI: https://doi.org/10.1007/978-1-4020-6752-5
13. Folstein MF, Folstein SE, McHugh PR. “Mini-mental state”. A practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res 1975;12, 189-98. DOI: https://doi.org/10.1016/0022-3956(75)90026-6
14. Wang J, Li T, Lv Y, et al. Fine Particulate Matter and Poor Cognitive Function among Chinese Older Adults: Evidence from a Community-Based, 12-Year Prospective Cohort Study. Environ Health Perspect 2020;128, 67013. DOI: https://doi.org/10.1289/EHP5304
15. Zhu A, Chen H, Shen J, et al. Interaction between plant-based dietary pattern and air pollution on cognitive function: a prospective cohort analysis of Chinese older adults. Lancet Reg Health West Pac 2022;20, 100372. DOI: https://doi.org/10.1016/j.lanwpc.2021.100372
16. Zhang Y, Fu S, Ding D, et al. Leisure Activities, APOE epsilon4, and Cognitive Decline: A Longitudinal Cohort Study. Front Aging Neurosci 2021.13, 736201. DOI: https://doi.org/10.3389/fnagi.2021.736201
17. Zhang Q, Wu Y, Han T, Liu E. Changes in Cognitive Function and Risk Factors for Cognitive Impairment of the Elderly in China: 2005-2014. Int J Environ Res Public Health 2019;16. DOI: https://doi.org/10.3390/ijerph16162847
18. Alvarez-Bustos A, Carnicero-Carreno JA, Sanchez-Sanchez JL, et al. Associations between frailty trajectories and frailty status and adverse outcomes in community-dwelling older adults. J Cachexia Sarcopenia Muscle 2022;13, 230-239. DOI: https://doi.org/10.1002/jcsm.12888
19. Nagin DS. Group-based trajectory modeling: an overview. Ann Nutr Metab 2014;65, 205-10. DOI: https://doi.org/10.1159/000360229
20. Andruff H, Carraro N, Thompson AJ, Gaudreau P, Louvet B. Latent Class Growth Modelling: A Tutorial, 2009. DOI: 10.20982/tqmp.05.1.p011
21. Jones B, Nagin D, Roeder K. A SAS Procedure Based on Mixture Models for Estimating Developmental Trajectories. Sociological Methods & Research 2001;29. DOI: https://doi.org/10.1177/0049124101029003005
22. Yu J, Feng Q, Yu J, Zeng Y, Feng L; Late-Life Cognitive Trajectories and their Associated Lifestyle Factors. J Alzheimers Dis 2020;73, 1555-1563. DOI: https://doi.org/10.3233/JAD-191058
23. Zhang W, Chen Y, Chen N.Body mass index and trajectories of the cognition among Chinese middle and old-aged adults. BMC Geriatr 2022;22, 613. DOI: https://doi.org/10.1186/s12877-022-03301-2
24. Wang S, Li W, Li S, et al. Association between plant-based dietary pattern and biological aging trajectory in a large prospective cohort. BMC Medicine 2023;21. DOI: https://doi.org/10.1186/s12916-023-02974-9
25. Yaffe K, Peltz CB, Ewing SK, et al. Long-term Cognitive Trajectories and Mortality in Older Women. J Gerontol A Biol Sci Med Sci 2016;71, 1074-80. DOI: https://doi.org/10.1093/gerona/glw003
26. Taniguchi Y, Kitamura A, Ishizaki T, et al. Association of trajectories of cognitive function with cause-specific mortality and medical and long-term care costs. Geriatr Gerontol Int 2019;19, 1236-1242. DOI: 10.1111/ggi.13802
27. Wu Z, Woods RL, Chong T, et al. Cognitive Trajectories in Community-Dwelling Older Adults and Incident Dementia, Disability and Death: A 10-Year Longitudinal Study. Front Med (Lausanne) 2022;9, 917254. DOI: https://doi.org/10.3389/fmed.2022.917254
28. Li C, Ma Y, Hua R, Zheng F, Xie W. Long-term physical activity participation trajectories were associated with subsequent cognitive decline, risk of dementia and all-cause mortality among adults aged >/=50 years: a population-based cohort study. Age Ageing 2022;51. DOI: https://doi.org/10.1093/ageing/afac071
29. Nie J, Qiu Q, Phillips M, et al. Early Diagnosis of Mild Cognitive Impairment Based on Eye Movement Parameters in an Aging Chinese Population. Front Aging Neurosci 2020;12, 221. DOI: https://doi.org/10.3389/fnagi.2020.00221
30. Su J, Xiao X. Factors leading to the trajectory of cognitive decline in middle-aged and older adults using group-based trajectory modeling: A cohort study. Medicine 2022;101, e31817. DOI: https://doi.org/10.1097/MD.0000000000031817
31. Yu W, Chen R, Zhang M, et al. Cognitive decline trajectories and influencing factors in China: A non-normal growth mixture model analysis. Archives of Gerontology and Geriatrics 2021;95, 104381. DOI:10.1016/j.archger.2021.104381
32. Zhou DF, Wu CS, Qi H, et al. Prevalence of dementia in rural China: impact of age, gender and education. Acta Neurol Scand 2006;114, 273-80. DOI: https://doi.org/10.1111/j.1600-0404.2006.00641.x
33. Li X, Zhang J, Hou R, et al. Bidirectional associations of intellectual and social activities with cognitive function among middle-aged and elderly adults in China. J Affect Disord 2022;319, 83-89. DOI: https://doi.org/10.1016/j.jad.2022.09.031
34. Gillum RF, Obisesan TO. Physical activity, cognitive function, and mortality in a US national cohort. Annals of epidemiology 2010;20, 251-257. DOI: https://doi.org/10.1016/j.annepidem.2010.01.005
35. Ophey A, Roheger M, Folkerts AK, Skoetz N, Kalbe E. A Systematic Review on Predictors of Working Memory Training Responsiveness in Healthy Older Adults: Methodological Challenges and Future Directions. Front Aging Neurosci 2020;12, 575804. DOI: https://doi.org/10.3389/fnagi.2020.575804
© Serdi 2024