Download: 
-
Globally, the population of the late-elderly (aged ≥75 years) was expected to rise from 275 million in 2020 to 768 million by 2050 (1), which might result in large disease burdens accompanied by increased life years of ill-health (2–3). However, the lack of data on multiple health measurements of older adults in large long-term cohorts increases the difficulty in studying healthy longevity. To fill the gap, analyses were conducted using data from the Chinese Longitudinal Healthy Longevity Survey (CLHLS) from 1998 to 2018. Separate logistic regression models were performed to identify factors associated with longevity and healthy longevity among 11,005 late-elderly participants, who were more likely to survive to 90 years old than younger adults. The corresponding population attributable fractions (PAFs) were also calculated. The incidence densities were 17.0/1,000 person-years for healthy longevity and 63.0/1,000 person-years for usual longevity. Considering early life factors, lifestyles, functional health, and diseases, the total weighted PAFs were 83.7% for longevity and 73.4% for healthy longevity, suggesting great potential for longevity and healthy longevity in China. Effective strategies may include targeting people through education and residence in early life, healthy lifestyles, disease prevention, and functional optimization in late life.
Study participants were recruited from the CLHLS study, which is the first study to investigate factors of healthy longevity in China covering 23 provincial-level administrative divisions (PLADs) (3). The study included 11,005 late-elderly participants who had the potential to survive to age 90 years by 2018, with complete measurement of outcomes and candidate factors for analysis (
Supplementary Figure S1 ). Healthy longevity was defined with reference to the WHO definition of healthy ageing and participants were classified into three categories (2): 1) healthy longevity: age at death ≥90 years, with good physical performance, cognitive function, mental health, visual function, and hearing function; 2) usual longevity: age at death ≥90 years, with at least one kind of functional impairment or disability mentioned above; 3) non-survival: age at death <90 years, irrespective of functional status (Supplementary Methods ). Candidate influencing factors, including demographic characteristics, lifestyles, functional status, self-rated health, and diseases, were collected through face-to-face interviews with all participants (Supplementary Methods ).Factors of longevity and healthy longevity were identified with separate logistic regression models via a bidirectional stepwise procedure and inclusion of candidate factors with P<0.1 in the raw estimates only adjusted for age: 1) non-longevity versus longevity; 2) usual longevity versus healthy longevity. As participants with different baseline ages will have different periods to age 90, age was included as a confounder as opposed to a risk factor in the analyses. The PAFs of modifiable factors were calculated by transforming them into binary or multifactorial variables according to previous studies and were adjusted for communality, which explains the overlap between factors (4-5). We performed a principal component analysis to calculate the communality for each factor and took into account how much each unobserved component explained each measured factor (5). Both age and sex were considered fixed variables. In this study, the PAF is interpreted as the probability gains of longevity or healthy longevity if all participants kept to healthy lifestyles or were in the absence of adverse health status factors (4). All analyses were finished with R 4.1.3 for Windows (R Foundation for Statistical Computing, Vienna, Austria). Statistical significance was set at two-tailed with a P-value <0.05.
Overall, 3,454 (31.4%) participants survived to age 90, of which 735 (6.7%) were classified as having healthy longevity and 2,719 (24.7%) were classified as having usual longevity. The corresponding incidence densities were 17.0/1,000 and 63.0/1,000 person-years, respectively. Rural residents and women were more likely to be classified as having longevity, but most of them were classified as having usual longevity, indicative of a lower likelihood of being classified as having healthy longevity. The incidences of longevity, healthy longevity, and usual longevity were higher in older adults with advanced ages (Table 1).
Groups All participants Longevity Healthy longevity Usual longevity No. of participants Person-years No. of events Incidence No. of events Incidence No. of events Incidence Total 11,005 43,154 3,454 80.0 735 17.0 2,719 63.0 Residence Urban 4,683 17,980 1,371 76.3 311 17.3 1,060 59.0 Rural 6,322 25,174 2,083 82.7 424 16.8 1,659 65.9 Sex Men 5,595 21,103 1,572 74.5 417 19.8 1,155 54.7 Women 5,410 22,051 1,882 85.3 318 14.4 1,564 70.9 Age group (years) 75–79 1,516 9,339 179 19.2 40 4.3 139 14.9 80–84 4,791 20,952 938 44.8 185 8.8 753 36.0 85–89 4,698 12,862 2,337 181.7 510 39.7 1,827 142.0 Table 1. Person-years of follow-up, number of outcome events, and incidence per 1,000 person years of observation by residence, sex, and age groups among Chinese late-elderly, 1998–2018.
Odds ratios [95% confidence intervals (CIs)] for factors retained in the final logistic regression models are presented in Table 2. The PAFs for modifiable factors associated with longevity and healthy longevity are displayed in Table 3. Of all modifiable lifestyles, performing housework, watching TV or listening to the radio, and never smoking are the top three factors contributing to longevity (weighted PAFs were 6.7%, 4.9%, and 3.3%, respectively); never smoking or having quit, normal weight or overweight, and tea-drinking are the top three factors contributing to healthy longevity (weighted PAFs were 8.8%, 6.0%, and 4.9%, respectively). All modifiable lifestyles combined may explain 32.8% probability gains of longevity and 30.4% probability gains of healthy longevity. The probability gains reached 83.7% and 73.4% when all potentially modifiable factors, excluding age and sex, were considered. Among factors other than lifestyles, interventions at early ages also contribute to considerable probability gains of longevity and healthy longevity (weighted PAF of residence areas was 2.7% for longevity, weighted PAF of education years was 2.3% for healthy longevity); early prevention and usage of supporting tools or equipment to optimize functional abilities are also important in promoting longevity [weighted PAFs of chewing ability, activities of daily living (ADL) limitations, and cognitive impairment were 2.2%, 13.1%, and 9.3%, respectively] and healthy longevity (weighted PAFs of chewing ability, ADL limitations, and hearing function were 4.4%, 15.4%, and 20.5%, respectively).
Characteristics Longevity Healthy longevity OR (95% CI) P-value OR (95% CI) P-value Women 1.45 (1.30–1.61) <0.001 0.71 (0.58–0.87) 0.001 Rural 1.16 (1.05–1.28) 0.003 Not included Education years (per 1 year) Not included − 1.05 (1.03–1.08) <0.001 Smoking Never smokers 1 (ref) − 1 (ref) − Current smokers 0.86 (0.76–0.97) 0.016 0.79 (0.62–0.99) 0.044 Former smokers 0.83 (0.72–0.95) 0.007 1.14 (0.89–1.46) 0.310 Performing housework Not included Rarely or not 1 (ref) − − − Occasionally 1.25 (1.09–1.44) 0.002 − − Often 1.42 (1.29–1.57) <0.001 − − Raising domestic animals Rarely or not 1 (ref) − 1 (ref) − Occasionally 1.04 (0.88–1.22) 0.670 1.51 (1.13–2.01) 0.006 Often 1.23 (1.09–1.38) 0.001 1.35 (1.09–1.65) 0.005 Reading newspapers or book Not included Rarely or not 1 (ref) − − − Occasionally 0.94 (0.78–1.13) 0.490 − − Often 1.19 (1.03–1.38) 0.019 − − Watching TV or listening to the radio Not included Rarely or not 1 (ref) − − − Occasionally 1.11 (0.98–1.27) 0.110 − − Often 1.27 (1.14–1.43) <0.001 − − Garlic consumption Not included Rarely or not 1 (ref) − − − Occasionally 1.14 (1.04–1.26) 0.008 − − Often 1.05 (0.92–1.19) 0.510 − − Tea drinking Not included Rarely or not − − 1 (ref) − Occasionally − − 1.37 (1.10–1.71) 0.005 Often − − 1.21 (0.99–1.47) 0.062 Good chewing ability 1.14 (1.03–1.25) 0.008 1.25 (1.05–1.48) 0.014 ADL score (per 1 point) 0.84 (0.80–0.88) <0.001 0.85 (0.75–0.97) 0.018 MMSE score (per 1 point) 1.02 (1.01–1.03) <0.001 Psychological resources (per 1 point) 1.02 (1.01–1.03) 0.005 1.05 (1.03–1.08) <0.001 Hearing loss Not included 0.49 (0.33–0.74) 0.001 Heart disease 0.73 (0.61–0.87) 0.001 Cerebrovascular disease 0.77 (0.60–0.98) 0.036 Not included − Respiratory disease 0.79 (0.69–0.91) 0.001 Not included − BMI Not included Underweight − − 0.82 (0.68–0.98) 0.032 Normal − − 1 (ref) − Overweight − − 1.08 (0.81–1.43) 0.620 Obese − − 0.45 (0.22–0.89) 0.022 Self-rated health status Not included Good 1.07 (0.97–1.19) 0.19 − − So so 1 (ref) − − − Bad 0.76 (0.65–0.89) 0.001 − − Note: Data are adjusted OR (95% CI). A value higher than 1 indicates participants are more likely to be longevity or healthy longevity. Not included means factors were not selected or non-significant (P>0.05) in the stepwise logistic models and not retained in the final models.
Abbreviation: CI=confidence interval; OR=odds ratio; ADL=activities of daily living; MMSE=mini-mental state examination; BMI=body mass index.Table 2. Odds ratios (95% CIs) for factors associated with longevity and healthy longevity among Chinese late-elderly, 1998–2018.
Factors PAFs for longevity (%) PAFs for healthy longevity (%) Raw Weighted† Raw Weighted† Lifestyle factors Smoking* 9.8 3.3 21.4 8.8 Tea drinking (often or occasionally) Not included Not included 12.0 4.9 Garlic consumption (often or occasionally) 7.5 2.5 Not included Not included Performing housework (often or occasionally) 19.9 6.7 Not included Not included Reading newspapers or books (often) 3.4 1.1 Not included Not included Raising domestic animals (often or occasionally) 4.8 1.6 9.8 4.0 Watching TV or listening to the radio (often or occasionally) 14.5 4.9 Not included Not included BMI (normal weight or overweight; 18.5–27.9 kg/m2) Not included Not included 14.7 6.0 Other potentially modifiable factors Residence (rural) 7.9 2.7 Not included Not included Education years (≥6) Not included Not included 5.5 2.3 Chewing ability (good) 6.4 2.2 10.7 4.4 ADL limitations (no) 38.8 13.1 37.5 15.4 Cognitive impairment (no) 27.6 9.3 Not included Not included Mental health (good) 9.7 3.3 17.3 7.1 Hearing function (good) Not included Not included 49.8 20.5 Heart disease (no) 24.8 8.4 Not included Not included Cerebrovascular disease (no) 27.1 9.1 Not included Not included Respiratory disease (no) 19.0 6.4 Not included Not included Self-rated health status (good or so so) 26.7 9.0 Not included Not included Combined factors All modifiable lifestyle factors 47.5 32.8 46.8 30.4 All potentially modifiable factors 94.1 83.7 88.3 73.4 Note: All models were adjusted for age, sex, and all other potentially modifiable factors. Not included means factors were not selected or non-significant (P>0.05) in the stepwise logistic models and not retained in the final models.
Abbreviation: CI=confidence interval; PAF=population attributable fraction; BMI=body mass index; ADL=activities of daily living.
* Based on the results of separate logistic models, never smoking was considered in the calculation of PAFs for longevity, and never or quit smoking was considered in the calculation of PAFs for healthy longevity.
† Weighted PAFs were the relative contributions of each factor to the overall PAF of all potentially modifiable factors when adjusted for communality.Table 3. Population attributable fractions for modifiable factors of longevity and healthy longevity among the late-elderly in China, 1998–2018.
-
In this large cohort study of 11,005 late-elderly participants, incidences of longevity and healthy longevity in China were provided and 19 modifiable factors for longevity and healthy longevity were identified. The weighted PAFs of longevity and healthy longevity are 32.8% and 30.4%, respectively, for all modifiable lifestyle factors and would increase to 83.7% for longevity and 73.4% for healthy longevity when all potentially modifiable factors were considered. Longevity and healthy longevity should be considered from a life course perspective, in which education and residence in early life and lifestyles in late life would continuously influence both the health status (including functional status and diseases) of a person and the probability of longevity and healthy longevity. Although functional status and diseases are probably irreversible in late life, the benefits are still substantial if supporting environments are strengthened to optimize the intrinsic capacity and functional abilities of late-elderly populations.
In this study, the probability of healthy longevity in China was suggested to be similar of that in younger Americans (6–7). Sex and 19 modifiable factors associated with longevity and healthy longevity in China were identified. These results are consistent with previous findings. The PAF for modifiable lifestyles of healthy longevity is comparable to a previous study that investigated successful aging in the United Kingdom when not adjusted for communality (4). Sociodemographically, women, rural residents, and participants with higher education levels were more likely to be classified as having longevity and/or healthy longevity (6). Regarding lifestyles, never smoking and engaging in leisure activities were beneficial to longevity and/or healthy longevity in late-elderly participants, which may be explained by their effects on reducing the risk of cognitive impairment, ADL disability, and mortality (4,6,8); garlic consumption and tea drinking may protect the late-elderly participants by the antidotal effect, inhibiting the growth of cancer cells, and reducing inflammatory levels (9); late-elderly participants who were underweight or obese may suffer from higher risk of chronic diseases and disabilities, reducing the likelihood of healthy longevity (10). Regarding other health measurements, late-elderly participants with good chewing ability, ADL independence, cognitive function, mental health, and hearing function were more likely to reach longevity and/or healthy longevity. These late-elderly participants may suffer from slower functional declines, and good chewing ability also enabled the adequate intake of essential nutrients. Conversely, late-elderly participants with heart disease, cerebrovascular disease, respiratory disease, and bad self-rated health status were more likely to suffer from functional declines and less likely to reach longevity (11).
Several limitations need to be considered when interpreting these findings. First, this study did not include data on genetic information and biomarkers, which need to be further investigated in future studies. Second, the inclusion of Chinese late-elderly participants might limit the generalizability of these findings to other racial groups.
In summary, incidences of longevity and healthy longevity in China in a large cohort of late-elderly participants were provided and revealed that education and residence in early life, healthy lifestyles, disease prevention, and functional optimization in late life together contribute to longevity and healthy longevity. The findings of this study reinforced the importance of a life course perspective in health management and may be applied to develop intervention strategies against the burdens attributed to the ageing population in China.
-
No conflicts of interest.
HTML
| Citation: |
|