ABSTRACT
Background
As populations age, extending healthspan, or years lived in good health, is a global priority. Most evidence on healthy lifestyle and prolonged healthspan comes from middle‐aged or comorbid populations, leaving it unclear whether benefits apply to healthy older adults. This study evaluates whether combined lifestyle behaviors are associated with disability‐free survival in community‐dwelling older adults.
Methods
The study included 11,287 Australian participants (median age 74 [IQR 72–77]) from the ASPirin in Reducing Events in the Elderly (ASPREE) study, with a median follow‐up of 6.6 years (IQR 5.5–7.9). Participants received one point for adherence to each of the following lifestyle factors: Mediterranean diet, moderate physical activity, non‐smoking, and moderate alcohol consumption and categorized as having low (0–1 factors), moderate (2 factors), or favorable (≥ 3 factors) lifestyle. The primary outcome was a composite endpoint comprised of the first occurrence of either death, dementia, or persistent physical disability. Associations of lifestyle categories with the composite endpoint and the individual components were examined, alongside effect modification by key demographic and health variables. Years gained in disability‐free survival and compression of morbidity were calculated.
Results
Compared to those with an unfavorable lifestyle, a moderate [HR 0.75 (95% CI 0.65–0.87)] and favorable [HR 0.60 (95% CI 0.52–0.70)] lifestyle were associated with a lower risk of the composite endpoint. Over a median of 6.6 years, a favorable lifestyle was prospectively associated with a 10% gain in years of healthspan and a moderate compression of morbidity. Associations did not differ across groups of age, sex, education, aspirin treatment, BMI, diabetes, and hypertension.
Conclusion
In healthy older adults, adherence to a healthy lifestyle was associated with a greater likelihood of surviving free from disability and dementia and was prospectively linked with a prolonged healthspan and a compression of morbidity, highlighting its potential importance in promoting healthy aging.
Trial Registration
ClinicalTrials.gov identifier: NCT01038583
Keywords: compression of morbidity, disability‐free survival, healthspan, lifestyle, older adults
Summary
- Key points
-
○Engaging in ≥ 3 healthy lifestyle behaviors (favorable lifestyle), including moderate physical activity, a Mediterranean diet, moderate alcohol consumption, and non‐smoking, was associated with a 40% lower relative risk of death, dementia, or disability compared with ≤ 1 healthy lifestyle behavior (unfavorable lifestyle).
-
○Participants with a favorable lifestyle were associated with approximately 10% more years lived in good health compared to those with an unfavorable lifestyle.
-
○The benefits were consistent across demographic and health subgroups.
-
○
- Why does this research matter?
-
○This study demonstrates that even in later life, adopting healthy lifestyle behaviors may be associated with extended healthspan, with a 40% lower risk of disability‐free survival loss (death, dementia, or persistent disability) and measurable gains in years lived in good health. These findings highlight that healthy lifestyle behaviors are associated with meaningful individual and population‐level benefits, reinforcing their importance as a cornerstone of healthy aging strategies.
-
○
In 11,287 older adults from the ASPirin in Reducing Events in the Elderly (ASPREE) cohort, healthy lifestyle behaviors were defined as Mediterranean diet, regular physical activity, moderate alcohol intake, and non‐smoking. Participants with three or more favorable behaviors had a 40% lower hazard of death, dementia, or persistent physical disability, and experienced a 10% longer healthspan and a compression of morbidity.
1. Introduction
The widening gap between healthspan, years lived in good health, and lifespan, total years lived, has increased by 13% globally over the past two decades [1]. This reflects an increasing period of life spent with disease or disability, spurring interest in lifestyle strategies to extend healthspan. Unhealthy lifestyle behaviors, such as smoking, excessive alcohol consumption, poor diet, and physical inactivity, negatively impact health during middle and older age [2]. However, the extent to which these factors collectively influence healthspan in older community dwelling populations remains less understood [3]. Most research has focused on middle‐aged or comorbid cohorts, leaving the impact of lifestyle among healthy older adults (70+ years) poorly understood. This raises the question of whether it is ever too late for healthy lifestyle behaviors to confer substantial benefits. In addition, the relative contributions of individual lifestyle factors in later life remain unclear. With the global population aged 70 years and above expected to double from 1 billion in 2020 to 2.1 billion by 2050 [4], clarifying the relationship between lifestyle and healthspan is essential for preventive strategies and policy development [5].
Disability‐free survival (DFS) is increasingly recognized as a valid, person‐centered measure of healthspan with significant public health relevance. While research in healthy older cohorts suggests a beneficial effect of combined lifestyle factors on outcomes like disability or death, significant gaps remain. Prior studies have relied on self‐reported [6] or administrative data [7] for endpoint definitions, with variability in how key components, particularly disability, are measured. Moreover, dementia, a critical determinant of functional dependence and institutionalization, has often been omitted from these composite endpoints [8]. Furthermore, most cohorts have included individuals with baseline comorbidities, where lifestyle changes may be less feasible or impactful.
The Aspirin in Reducing Events in the Elderly (ASPREE) study, a cohort of uniquely healthy older adults, provides an opportunity to address these gaps. ASPREE was the first large‐scale trial to employ an adjudicated composite endpoint consisting of the first occurrence of death, dementia, or persistent physical disability [9]. Leveraging ASPREE and its observational follow‐up, we investigate the association between a healthy lifestyle score, including physical activity, diet, smoking, and alcohol consumption, and a composite endpoint representing the inverse of disability‐free survival. Accordingly, we aimed to determine whether a healthy lifestyle is associated with a lower risk of death, dementia, or persistent physical disability in healthy older adults. This analysis extends prior work by using adjudicated outcomes, including dementia as a key endpoint, and estimating hazards, absolute risk reductions, and years gained to assess the benefits of a healthy lifestyle in community‐dwelling older adults.
2. Methods
2.1. Study Population and Trial Design
This analysis utilizes data from Australian participants who consented to participate in the ASPREE trial [9, 10, 11] (n = 16,703) and the observational follow‐up sub‐study; the ASPREE‐eXTension (ASPREE‐XT) [12] in which participants additionally consented to complete further surveys as part of the ASPREE Longitudinal Study of Older Persons (ALSOP) sub‐study (n = 14,892) [13]. The ASPREE trial was a large, randomized, double‐blind, placebo‐controlled trial evaluating the effect of 100 mg of aspirin on disability‐free survival in healthy adults aged 70 or older. Details of the trial and its primary findings have been published [9, 10, 11]. In brief, participants (70+ years) were in good health at enrolment, with no history of cardiovascular events, dementia, or major physical disabilities and were expected to live for at least five additional years (Figure S1). Data for this analysis includes information collected up to the second annual visit of the ASPREE‐XT study (2020). Baseline was defined as the date of aspirin randomization, with baseline data collected between March 2010 and December 2014. Annual visits were conducted at 12‐month intervals thereafter. The ALSOP sub‐study, involving ~90% of Australian ASPREE participants, collected medical and social data at baseline, 3 and 5 years [13].
All studies were approved by ethics committees and registered, with written informed consent obtained from participants.
2.2. Lifestyle Score
Full details of the assessment of the lifestyle factors and classification are described in Table 1 and have been published previously [14].
TABLE 1.
Description and prevalence of lifestyle factors in the study population.
| Lifestyle factor | Points | Description | Prevalence, n (%) | Source and definition |
|---|---|---|---|---|
| Smoking | 0 | Current smoking | 289 (2.6) | The ASPirin in Reducing Events in the Elderly (ASPREE) trial baseline medical questionnaire collected smoking status, with participants reporting whether they were current, former, or never smokers. For the current analysis, participants were classified as either current smokers or non‐smokers, with former smokers grouped with non‐smokers to reflect current smoking behavior, regardless of past history. |
| 1 | Never or former smoker | 10,998 (97.4) | ||
| Alcohol consumption | 0 | None, low consumption (≤ 50 g/week), high consumption (> 100 g/week), former drinker | 8767 (77.7) | The ASPREE trial baseline medical questionnaire collected information about alcohol consumption. Participants reported the following detail relating to alcohol consumption: current, former or never; days of drinking per week and average standard drinks per day. Total current alcohol consumption (in grams/week) was calculated by multiplying the days of drinking per week, the average standard drinks per day and the equivalent grams of alcohol in one standard drink. For Australian participants, a standard drink was equivalent to 10 g. Aligned with previous methodologies [14, 15, 16] participants were categorized as either having moderate alcohol consumption or not, defined as those reporting between 51 and 100 g of alcohol per week (approximating to an average of 0.7–1.4 standard Australian alcoholic beverages a day), synonymous to previous cut‐offs [16]. |
| 1 | Moderate consumption (51–100 g/week) | 2520 (22.3) | ||
| Mediterranean diet | 0 | Lower median of the Med Diet score (≤ 11.3) | 5646 (50.0) | Dietary data were collected using a 54‐item food frequency questionnaire (FFQ) as part of the ALSOP year‐three medical questionnaire. A healthy diet was defined by the previously derived Mediterranean diet score (ASPREE‐MDS), details of which have been published previously [17]. Consumption was assessed across five categories, from “never/rarely” to “every day or several times a day.” A healthy diet was defined by high intake of vegetables, fruits, grains, and nuts/legumes, olive oil as the primary oil, and additional points for oily fish. Valid FFQs required at least 30 of 38 items completed; incomplete questionnaires were excluded. An “optimal” Mediterranean diet score excluded significant discretionary food intake. The ASPREE‐MDS score ranged from 0 to 18. Adherence was defined based on a score above the median (> 11.3), consistent with thresholds used in other cohort studies where adherence typically represents approximately 50% of the maximum score [18, 19, 20]. For instance, a score above 7 out of 14 on the MEDAS scale [19] and above 26.5 out of 55 on the MDS scale [21] are common thresholds for adherence. |
| 1 | Upper median of the Med Diet score (> 11.3) | 5641 (49.9) | ||
| Physical activity | 0 | Low physical activity: no or light activity weekly | 3749 (33.1) | The ALSOP baseline social questionnaire asked participants to report their typical weekly exercise and physical activity level, with options ranging from never/rarely to light, moderate, or vigorous activity. These categories were informed by World Health Organization (WHO) and Australian government guidelines, which recommend at least 30 min of moderate physical activity on five or more days per week for adults aged 65 and older [22, 23]. |
| 1 | High physical activity: moderate or vigorous activity weekly | 7591 (66.9) | ||
| Healthy lifestyle score categories, n (%) | ≤ 1 | Unfavorable | 1910 (16.9) | One point per lifestyle factor were assigned based on adherence to each of these four behaviors. The total lifestyle score ranged from 0 to 4 and was categorized into three groups: unfavorable (score ≤ 1), moderate (score = 2), and favorable (score ≥ 3). |
| 2 | Moderate | 4330 (38.4) | ||
| ≥ 3 | Favorable | 5047 (44.7) |
The lifestyle score was developed using four healthy lifestyle behaviors (moderate alcohol intake, non‐smoking, moderate physical activity, and a Mediterranean diet) that have been associated with a reduced incidence of chronic diseases [6, 15, 24, 25, 26, 27]. One point per healthy lifestyle factor was assigned based on adherence to each of these four behaviors (Table 1). The total lifestyle score ranged from 0 to 4 and was categorized into three groups: unfavorable (score ≤ 1), moderate (score = 2), and favorable (score ≥ 3). These categories have been used previously in ASPREE [14] and other cohort studies, such as the UK Biobank [15], and offer a clear, accessible public health message to encourage realistic lifestyle modifications.
2.3. Outcomes
The composite endpoint, representing the inverse of disability‐free survival, was comprised of the first occurrence of either dementia, persistent physical disability or death within an 8‐year time period. Methodological details have been published previously [10, 11] and further detail is provided in Table S1. Briefly, dementia was defined and adjudicated according to the Diagnostic and Statistical Manual for Mental Disorders, American Psychiatric Association (DSM‐IV) criteria [28]. Accrual of a persistent physical disability was defined as the first reported date of the loss of an activity of daily living (ADL) that was confirmed if still persistent approximately 6‐month later or by date of nursing home placement due to physical disability [29]. All deaths were confirmed by review of two independent sources and linkage with the National Death Indices.
2.4. Confounders
Covariates were selected a priori. Models were adjusted for age at baseline (continuous), sex, education (> 12 vs. ≤ 12 years), aspirin treatment, living circumstances (alone vs. with others), and neighborhood disadvantage (Index of Relative Socio‐economic Advantage and Disadvantage, IRSAD). These variables were chosen as they may influence both engagement in healthy lifestyle behaviors and the risk of mortality, dementia, or disability and are unlikely to lie on the causal pathway.
Subgroup analyses were conducted to explore potential effect modification by baseline health status, including body mass index (BMI < 25 vs. ≥ 25 kg/m2), hypertension, and diabetes mellitus. Diabetes mellitus was defined as self‐report of diabetes, fasting glucose ≥ 126 mg/dL, or use of glucose‐lowering medication. Hypertension was defined as systolic blood pressure ≥ 140 mmHg, diastolic blood pressure ≥ 90 mmHg, or use of antihypertensive medication. Full details on the ascertainment of these study measurements have been described previously [9, 13].
2.5. Statistical Analyses
Participants included in the analysis had completed both baseline and year‐3 ALSOP questionnaires and were followed for up to 8 years. Details of the analytic sample can be viewed in Figure S1.
Multivariable Cox proportional hazards models were used to estimate hazard ratios (HRs) and 95% confidence intervals (CIs) for associations between lifestyle categories and the composite endpoint and its individual components (death, dementia, and disability) within an 8‐year time period. Model 1 adjusted for age (continuous), sex, and aspirin allocation. Model 2 included additional adjustments for living circumstances (alone vs. with others), education (< 12 vs. ≥ 12 years), and socioeconomic status (relative socio‐economic advantage and disadvantage [IRSAD] deciles). Among the covariates included in adjusted analyses, only IRSAD had missingness (< 0.5%). Given the negligible extent of missing data and its unlikely influence on model estimates, complete‐case analysis was used and multiple imputation was not considered necessary. Absolute risks were calculated for each lifestyle category using observed incidence rates and hazard ratios from the Cox proportional hazards model. For the reference group (unfavorable lifestyle, score ≤ 1), absolute risk was calculated as: Absolute Risk = 1 − exp. (−Incidence Rate × Time), where the incidence rate was expressed per 1000 person‐years. For the moderate and favorable lifestyle groups, hazard ratios were applied to the reference group's incidence rate to adjust the risk, and the same formula was used to compute their absolute risks. CIs for the absolute risks were derived using the Delta Method. Restricted mean survival time (RMST) was used to estimate years gained in healthspan and compression of morbidity, defined here as time lived free of dementia or disability, over a follow‐up period of up to 8 years, with comparisons made across lifestyle categories.
In secondary analyses, we investigated the association between the individual lifestyle factors and the composite endpoint as well as the association between healthy lifestyle categories and each component of the composite endpoint separately.
To investigate whether a priori selected factors modified the association between the lifestyle score and the composite end point, we performed analyses stratified by median age (≥ 74/< 74 years), sex (male/female), education (≥ 12/< 12‐year), BMI (≥ 25 vs. < 25 kg/m2), baseline diagnosis of type 2 diabetes (yes/no) and baseline diagnosis of hypertension (yes/no). Interaction was tested by including a cross‐product term along with the main effect terms in the models. Additionally, as the data is derived from a clinical trial of low‐dose aspirin, which observed higher all‐cause mortality among older adults who received daily aspirin than among those who received placebo [10], we performed stratified analyses by aspirin treatment allocation.
2.6. Sensitivity Analysis
In sensitivity analyses, we excluded former alcohol consumers and former smokers (who quit < 15‐year ago), given the potential for changes in these behaviors due to underlying health conditions and resulting reverse causality. Because declining function often precedes dementia and persistent physical disability and may influence lifestyle behaviors, we also conducted a sensitivity analysis excluding all participants with ≤ 4 years of follow‐up, including those who experienced an event.
To account for differences in the relative contribution of each lifestyle factor and to evaluate whether weighting improved model performance, we created a weighted lifestyle score. Each lifestyle factor was multiplied by its β‐coefficient from an adjusted Cox model, summed, and scaled to 100. Categories (unfavorable, moderate, favorable) were defined using the same distributional cut points as the unweighted score, and Cox models were re‐run using this weighted score. Finally, because dietary data were only available at year three rather than baseline, we conducted sensitivity analyses using a lifestyle score recalculated at year three, incorporating smoking, alcohol use, physical activity, and diet, and treating year three as the new baseline.
Statistical analyses were conducted using Stata (version 17), with significance set at p < 0.05. Cumulative incidence plots were generated in R (version 4.3.3). The proportional hazards assumption was assessed and met using log–log Kaplan–Meier survival plots.
3. Results
Baseline characteristics in the analytic sample and by lifestyle categories are shown in Table 2 and characteristics by the composite endpoint, the analytic cohort and the total ASPREE cohort can be viewed in Table S2. A total of 11,287 Australian participants were included in the analysis and followed for a total of 8 years (median = 6.6 years (IQR: 5.5–7.9)). Median age was 73 years (IQR: 71.7–77.3), and 54.2% were female. Over 74,716 person‐years of follow‐up, 1239 events occurred for the composite endpoint (dementia, disability, or death), corresponding to an event rate of 16.6 per 1000 person‐years. For the sub‐components, there were 397 dementia cases (incidence rate: 5.3 per 1000 person‐years across 74,707 person‐years), 385 disability cases (5.2 per 1000 person‐years across 74,590 person‐years), and 670 deaths (8.8 per 1000 person‐years across 75,754 person‐years). Participants who reached the composite endpoint were older, more often male, living alone, less educated, more likely to have diabetes, hypertension, frailty or pre‐frailty, and have depressive symptoms (Table S2).
TABLE 2.
Baseline characteristics of the study population according to lifestyle score categories.
| Characteristics | Total study cohort (n = 11,287) | Lifestyle score categories | ||
|---|---|---|---|---|
| Unfavorable (n = 1910) | Moderate (n = 4330) | Favorable (n = 5047) | ||
| Age (years), median (IQR) | 73.9 (71.7–77.3) | 74.6 (71.9–78.5) | 74.3 (71.8–77.9) | 73.3 (71.5–76.3) |
| Female sex, n (%) | 6113 (54.2) | 1098 (57.5) | 2276 (52.6) | 2739 (54.3) |
| Living alone, n (%) | 3460 (30.7) | 688 (36.0) | 1372 (31.7) | 1400 (27.7) |
| < 12 years of education, n (%) | 5358 (47.5) | 1066 (55.8) | 2233 (51.6) | 2059 (40.8) |
| IRSAD decile, median (IQR) | 6.0 (4.0–9.0) | 6.0 (3.0–9.0) | 6.0 (4.0–9.0) | 7.0 (4.0–9.0) |
| Aspirin allocation, n (%) | 5618 (49.8) | 915 (47.9) | 2161 (49.9) | 2542 (50.4) |
| Medical history, n (%) | ||||
| Diabetes | 1028 (9.1) | 237 (12.4) | 425 (9.8) | 366 (7.3) |
| Hypertension | 8352 (74.0) | 1533 (80.3) | 3273 (75.6) | 3546 (70.3) |
| Dyslipidemia | 7599 (67.3) | 1281 (67.1) | 2923 (67.5) | 3395 (67.3) |
| Frail/pre‐frail | 3981 (35.3) | 915 (47.9) | 1646 (38.0) | 1420 (28.1) |
| CES‐D‐10, score 8+ | 4790 (42.4) | 962 (50.4) | 1865 (43.1) | 1963 (38.9) |
| Prescribed medications, n (%) | ||||
| Statins | 3385 (48.8) | 642 (48.8) | 1361 (49.2) | 1382 (48.3) |
| Antihypertensives | 5787 (83.4) | 1153 (87.7) | 2320 (83.8) | 2314 (80.9) |
| Physical examination | ||||
| BMI kg/m2, mean (SD) | 27.9 (4.5) | 29.2 (5.1) | 28.1 (4.5) | 27.2 (4.0) |
| Waist circumference (cm), mean (SD) | 96.8 (12.4) | 100.1 (13.3) | 97.7 (12.2) | 94.9 (11.8) |
| Systolic BP (mm Hg), mean (SD) | 139.6 (16.2) | 140.8 (16.3) | 139.5 (16.2) | 139.2 (16.2) |
| Diastolic BP (mm Hg), mean (SD) | 77.3 (9.8) | 77.5 (10.0) | 77.2 (9.9) | 77.3 (9.7) |
| Pathology | ||||
| HDL (mmol/L), mean (SD) | 1.6 (0.5) | 1.5 (0.4) | 1.6 (0.4) | 1.6 (0.5) |
| Non‐HDL (mmol/L), mean (SD) | 3.7 (0.9) | 3.7 (1.0) | 3.7 (1.0) | 3.7 (0.9) |
| Creatinine (μmol/L), mean (SD) | 79.7 (18.9) | 81.0 (21.2) | 80.4 (18.3) | 78.6 (18.4) |
| eGFR (mL/min/1.73 m2), median (IQR) | 74.3 (64.0–84.2) | 73.0 (61.1–83.3) | 73.4 (63.1–83.9) | 75.5 (65.6–85.0) |
| Healthy lifestyle factors, n (%) | ||||
| No current smoking | 10,998 (97.4) | 1717 (89.9) | 4248 (98.1) | 5033 (99.7) |
| Regular physical activity | 7576 (67.1) | 91 (4.8) | 2791 (64.5) | 4694 (93.0) |
| Healthy diet | 5641 (50.0) | 17 (0.9) | 1286 (29.7) | 4338 (86.0) |
| Moderate alcohol consumption | 2520 (22.3) | 16 (0.8) | 335 (7.7) | 2169 (43.0) |
Abbreviations: μmol/L, micromoles per liter; BMI, body mass index; BP, blood pressure; CES‐D‐10, Center for Epidemiologic Studies Short Depression Scale; cm, centimeters; eGRF, estimated glomerular filtration rate; Higher score, less disadvantage; IQR, interquartile range; IRSAD, Participant Index of Relative Socio‐economic Advantage and Disadvantage; mL/min, milliliter per minute; mmol/L, millimoles per liter; n, sample size; SD, standard deviation.
Missing observations: IRSAD, 24 (0.21%); BMI, 52 (0.46%); waist circumference, 101 (0.89%); HDL, 288 (2.54%); non‐HDL, 288 (2.54%); creatinine, 332 (2.93%); eGFR, 332 (2.54%).
Participants in the unfavorable lifestyle group were older, more likely to be female, living alone, less educated, and more likely to have vascular risk factors, markers of poor kidney health, frailty/pre‐frailty, and depressive symptoms (Table 2). Lifestyle factor prevalence is detailed in Table 1. A total of 16.9% of participants adhered to ≤ 1 factor (unfavorable), 38.4% adhered to 2 factors (moderate), and 44.7% adhered to ≥ 3 factors (favorable). Characteristics by dementia, persistent physical disability and all‐cause mortality are shown in Table S3.
3.1. Association Between Lifestyle Categories and the Composite Endpoint
Rates of the composite endpoint by lifestyle categories are shown in Table 3 and Figure 1A. Compared to the unfavorable lifestyle group, individuals with a moderate lifestyle had a 25% lower hazard of the composite endpoint (HR: 0.75 [95% CI: 0.65, 0.87]), and those with a favorable lifestyle had a 40% lower hazard (HR: 0.60 [95% CI: 0.52, 0.70]) (Table 3). Associations did not substantially differ between the part‐ and full‐adjusted models.
TABLE 3.
Hazard ratios of the composite endpoint, all‐cause mortality, dementia, and persistent physical disability according to healthy lifestyle score.
| Healthy lifestyle score categories | N 11,287 | Events/person years | Event rate a | Hazard ratio (95% CI) b | |
|---|---|---|---|---|---|
| Model 1 | Model 2 | ||||
| Composite endpoint | |||||
| Unfavorable (≤ 1) | 1910 | 308/12,418 | 24.8 | 1.00 (Reference) | 1.00 (Reference) |
| Moderate (2) | 4330 | 515/28,705 | 17.9 | 0.75 (0.65, 0.86) | 0.75 (0.65, 0.87) |
| Favorable (≥ 3) | 5047 | 416/33,593 | 12.4 | 0.58 (0.50, 0.68) | 0.60 (0.52, 0.70) |
| Sub‐components of the composite endpoint | |||||
| Persistent physical disability | |||||
| Unfavorable (≤ 1) | 1910 | 118/12,386 | 9.5 | 1.00 (Reference) | 1.00 (Reference) |
| Moderate (2) | 4330 | 157/28,641 | 5.5 | 0.60 (0.47, 0.76) | 0.60 (0.48, 0.77) |
| Favorable (≥ 3) | 5047 | 110/33,562 | 3.3 | 0.40 (0.31, 0.52) | 0.42 (0.32, 0.54) |
| Dementia | |||||
| Unfavorable (≤ 1) | 1910 | 81/12,514 | 6.5 | 1.00 (Reference) | 1.00 (Reference) |
| Moderate (2) | 4330 | 164/28,652 | 5.7 | 0.91 (0.70, 1.19) | 0.93 (0.71, 1.21) |
| Favorable (≥ 3) | 5047 | 152/33,540 | 4.5 | 0.80 (0.61, 1.05) | 0.81 (0.61, 1.07) |
| All‐cause mortality | |||||
| Unfavorable (≤ 1) | 1910 | 167/12,571 | 13.1 | 1.00 (Reference) | 1.00 (Reference) |
| Moderate (2) | 4330 | 294/29,084 | 10.1 | 0.81 (0.67, 0.98) | 0.82 (0.67, 0.99) |
| Favorable (≥ 3) | 5047 | 209/33,918 | 6.2 | 0.58 (0.47, 0.71) | 0.61 (0.49, 0.75) |
Rates per 1000‐person years.
Model 1 adjusted for age, sex and aspirin treatment allocation; Model 2 adjusted for model 1 plus living status, education, socioeconomic status (IRSAD).
FIGURE 1.
Cumulative incidence of the composite endpoint, persistent physical disability, dementia, and all‐cause mortality. Figure A shows the cumulative incidence of the composite endpoint (death from any cause, dementia, or persistent physical disability) according to lifestyle categories. First events that counted toward the primary end point during the trial included 513 deaths, 379 cases of dementia, and 347 cases of persistent physical disability. Figures B, C, and D show cumulative incidences of all events of persistent physical disability, dementia, and death, respectively, according to lifestyle categories. Cumulative incidences for dementia and persistent physical disability are adjusted with death treated as a competing risk. The graph stops at year 8 because only a small number of participants (< 100 participants in lifestyle sub‐groups) with an event reached year 9. Differences between lifestyle groups were assessed using the log‐rank test for the composite endpoint (χ 2(2) = 92.6, p < 0.001) and mortality (χ 2(2) = 60.1, p < 0.001), and Gray's test for dementia (χ 2(2) = 7.29, p = 0.026) and persistent physical disability (χ 2(2) = 67.29, p < 0.001).
Absolute risks and restricted mean survival times are detailed in Table S4. Over a median follow‐up of 6.6 years (IQR: 5.5–7.9), the absolute risk of the composite endpoint was 20.0% (95% CI: 19.2, 20.8) for the unfavorable group, 14.9% (95% CI: 14.4, 15.3) for the moderate group, and 10.6% (95% CI: 10.3, 10.9) for the favorable group. Participants with a favorable lifestyle gained approximately 10% (0.63 years) more time of healthy lifespan compared to the unfavorable group over the median follow‐up period.
3.2. Association Between Lifestyle, Persistent Physical Disability, All‐Cause Mortality, and Dementia
Rates of persistent physical disability, dementia, and all‐cause mortality are shown in Table 3 and Figure 1B–D. For persistent physical disability, compared with those in the unfavorable lifestyle group, individuals in the moderate lifestyle group had a 40% lower hazard [HR 0.60 (95% CI 0.48, 0.77)] and individuals in the favorable lifestyle group had a 58% lower hazard (HR 0.42 [95% CI 0.32, 0.54]) (Table 3 and Figure 1B). A similar exposure‐response association was observed for all‐cause mortality, albeit with reduced effect size (Table 3, Figure 1D). Absolute risks followed a similar exposure‐response trend (Table S4). There was no association between lifestyle score categories and incident dementia (Table 3 and Figure 1C). Over a median 6.6 years, a favorable lifestyle was associated with 0.41 additional years free from dementia or disability compared to an unfavorable lifestyle. The strongest association was observed for disability‐free years (+0.45 years), while dementia‐free years were extended by 0.24 years.
3.3. Further Analyses
Table 4 and Table S5 present adjusted hazard ratios for the association between each individual lifestyle factor and the composite endpoint and sub‐components, respectively. After adjustment for covariates and mutual adjustment for the other lifestyle factors, each healthy lifestyle factor remained independently associated with a lower hazard of the composite endpoint, with the largest effect size for non‐smoking, followed by moderate/high physical activity, moderate alcohol consumption, and a Mediterranean diet. A similar pattern was observed for persistent physical disability. Adherence to a Mediterranean diet and non‐smoking were significantly associated with reduced hazard of all‐cause mortality, while only moderate alcohol consumption was linked to reduced dementia hazard. Results remained consistent in models with and without mutual lifestyle factor adjustment.
TABLE 4.
Hazard ratios of the composite endpoint in relation to individual lifestyle factors.
| Lifestyle factor | n | Events/person years | Event rate b | Hazard ratio (95% CI) a | |
|---|---|---|---|---|---|
| Model 1 | Model 2 | ||||
| Smoking | |||||
| Smoker | 289 | 49/1868 | 26.2 | 1.00 (Reference) | 1.00 (Reference) |
| Non‐smoker | 10,998 | 1190/72,848 | 16.3 | 0.55 (0.42, 0.74) | 0.58 (0.44, 0.78) |
| Alcohol consumption | |||||
| None/low/high | 8767 | 1022/57,981 | 17.6 | 1.00 (Reference) | 1.00 (Reference) |
| Moderate | 2520 | 217/16,735 | 11.4 | 0.80 (0.69, 0.93) | 0.83 (0.71, 0.96) |
| Mediterranean diet | |||||
| Low | 5646 | 717/37,152 | 19.3 | 1.00 (Reference) | 1.00 (Reference) |
| High | 5641 | 522/37,564 | 13.9 | 0.83 (0.74, 0.93) | 0.86 (0.77, 0.97) |
| Physical activity | |||||
| ≤ Low | 3711 | 509/24,427 | 20.8 | 1.00 (Reference) | 1.00 (Reference) |
| ≥ Moderate | 7576 | 730/50,289 | 13.5 | 0.76 (0.69, 0.87) | 0.79 (0.71, 0.89) |
Model 1 adjusted for age, sex, aspirin treatment allocation, living status, socioeconomic status; Model 2 is adjusted for model 1 and mutually adjusted for other lifestyle factors.
Rates per 1000‐person years.
Stratified analyses revealed no significant differences in associations between the lifestyle categories and the composite endpoint by sex, age (< 74/≥ 74), education (≥ 12/< 12‐year), BMI (< 25/≥ 25), aspirin treatment, diabetes, or hypertension (p‐interaction > 0.05; Table S6). Results from the weighted lifestyle score aligned closely with those from the unweighted score (Table S7), reinforcing the practicality of the unweighted score.
Sensitivity analyses (excluding former smokers/drinkers, participants with follow‐up ≤ 4 years, using year‐3 lifestyle score) did not alter the results (Table S7).
4. Discussion
In this cohort of initially healthy older Australians, we found that engaging in multiple healthy lifestyle behaviors was associated with prolonged healthspan. Reporting participation in ≥ 3 healthy lifestyle factors was associated with a 10% gain in healthy lifespan over a median follow‐up of 6.6 years. The similar gains in morbidity (dementia and/or disability) and overall survival (0.41 vs. 0.42 years) suggest that a favorable lifestyle may not only extend lifespan but also compress morbidity, delaying disability and cognitive impairment until closer to death. While modest at the individual level, these gains may have significant implications at a population level, delaying admission to care facilities, reducing disability, and alleviating pressure on healthcare systems. Across aging populations, small increases in healthspan could translate into significant additional years of healthy life, underscoring the scalability and public health impact of promoting healthy lifestyle behaviors, even among older adults. The association between engagement in a healthy lifestyle on healthspan persisted irrespective of demographic and socioeconomic factors and was not modified by selected confounders including age, sex, education, and cardiometabolic factors. Our findings underscore the broad applicability of lifestyle interventions across different demographic and health profiles, supporting their value as foundational strategies for healthy aging.
Despite methodological differences, similar cohort studies of older adults highlight the benefits of healthy lifestyle behaviors on disability‐free survival. For example, a study of Japanese older adults (n = 9910) reported that healthy behaviors (non‐smoking, daily walking, and fruit/vegetable consumption) extended years free of disability and death by 17.1 months over 10 years of follow‐up [7]. Similarly, two studies using data from the Swedish National Study of Ageing and Care in Kungsholmen cohort (≥ 60 years old) reported that favorable lifestyle behaviors, including active leisure and strong social networks, extended disability‐free survival by 3.26 years among older adults with diabetes (n = 2216) [30], delayed disability onset by 3.5 years and extended lifespan by 2.8 years in the broader cohort including those without diabetes, over 15 years of follow‐up [31]. Consistent with these findings, our study is the largest cohort of healthy older adults to identify the cumulative benefits of four common healthy behaviors on a rigorously adjudicated outcome comprising disability, death, and dementia. Importantly, we also report absolute risks and gains in healthspan, supporting public health strategies aimed at promoting healthy aging.
Studies investigating the association between a comparable lifestyle composite and mortality [26, 32, 33] or disability [6, 34, 35] have previously been reported and reflect our findings here. The lack of an association between the lifestyle categories and incident dementia contrasts with findings from previous studies demonstrating a strong relationship between favorable lifestyle and reduced risk of dementia [15, 36]. This discrepancy may reflect the particularly healthy profile of the cohort and the relatively short follow‐up period, potentially insufficient in capturing the long latency period typically associated with dementia development. Additionally, other lifestyle factors not included in the composite, such as engaging in socially or cognitively stimulating activities, may contribute to reducing dementia risk [37].
The stronger association between a favorable lifestyle and reduced disability risk, compared to dementia and mortality, may be attributed to multiple factors. Lifestyle factors such as physical activity and diet directly influence functional health, muscle strength, and mobility, which are closely linked to the risk of disability [38]. Furthermore, in this healthier older cohort, lifestyle may have a greater influence on preventing early physical limitations than on dementia, which typically manifests at a later stage.
Our study has several notable strengths. First, the ASPREE cohort comprises a large, contemporary, and well‐characterized group of adults aged ≥ 70 years, all in relatively good health at baseline [13]. Second, the rigorous methodologies employed in ASPREE allowed for the use of adjudicated, objective endpoints, enhancing the accuracy and reliability of findings. In contrast, previous studies relied on self‐reported measures or administrative data, which are prone to bias and variability. Additionally, we evaluated both weighted and unweighted lifestyle scores, providing a nuanced comparison of their predictive utility. By reporting relative and absolute measures, we facilitate broader interpretation and practical application for public health messaging. Lastly, our inclusion of gains in mean years of healthspan across lifestyle categories offers a tangible metric for understanding economic and societal benefits, further strengthening the relevance of our findings for policy and intervention development.
This study has several limitations. First, classifying moderate alcohol consumption as low‐risk may be debated given emerging evidence of potential harms, particularly for cancer [39]. However, this approach aligns with prior ASPREE analyses and other studies reporting a J‐shaped association with health outcomes. Moderate alcohol intake may also reflect greater social engagement and higher socioeconomic status, as observed in ASPREE [16].
Second, individual lifestyle factors may contribute unevenly to the composite endpoint. Nonetheless, close agreement between the weighted and unweighted scores supports the robustness of our findings. The unweighted score also offers a practical and simple tool for clinical and public health use, avoiding complex weighting calculations. Thirdly, due to the absence of baseline dietary data, year‐three dietary data was used. Although dietary behavior may shift over time by factors such as oral health, income, marital status, medication use, or changes in residence [40], the absence of non‐community‐dwelling participants likely minimized major changes in this sample. Supporting this, sensitivity analyses using year‐three scores across the other three lifestyle factors produced similar results. Fourth, although the absolute number of current smokers and events was sufficient to provide reasonably precise estimates, the very low prevalence of smoking in this cohort (2.9%) limits the generalizability of our smoking findings to populations with higher smoking rates and constrains the extent to which its contribution can be scaled relative to other lifestyle factors.
Further limitations include the reliance on self‐reported lifestyle variables, which introduce misclassification. The ASPREE cohort comprised healthy, older, mostly white, education and adherent participants with access to universal healthcare, which may limit generalizability. The absence of mid‐life lifestyle data also restricts assessment of long‐term behavioral patterns. Finally, unmeasured factors such as social engagement, cognitive activity, and environmental context may also influence healthspan.
5. Conclusion
In a large, well‐characterized population of healthy older people, adherence to each individual healthy lifestyle factor was associated with a reduced risk of the composite endpoint, while adhering to at least three healthy factors resulted in a 10% prolongation of healthspan. These findings highlight the importance of maintaining a healthy lifestyle, even after 70 years of age.
Author Contributions
Catherine Robb, Prudence R. Carr, Jocasta Ball, and John J. McNeil conceived and designed the study, designed the analysis plan, supervised the analysis, and interpreted the results. Catherine Robb conducted the analysis, wrote the first draft, and revised the manuscript. Galina Polekhina reviewed the analysis, interpreted the results, and reviewed the manuscript. All further co‐authors interpreted the results and reviewed the manuscript. We thank the ASPREE trial staff, participants, and general practitioners involved in the study.
Funding
This work was supported by the National Institute on Aging and the National Cancer Institute at the National Institutes of Health (grant numbers U01AG029824, U19AG062682); the National Health and Medical Research Council of Australia (grant numbers 334047, 1127060); Monash University; and the Victorian Cancer Agency. J.J.M. is supported through an NHMRC Leadership Fellowship (IG 1173690).
Disclosure
The funding bodies played no role in the study design; analysis, interpretation; manuscript preparation, review, or approval; or the decision to submit for publication.
Conflicts of Interest
The authors declare no conflicts of interest.
Supporting information
Table S1: Description of the ascertainment of the three components of the primary composite end point, including all‐cause mortality.
Figure S1: Flow of participant inclusion and ASPREE study eligibility criteria.
Table S2: Baseline characteristics of the ASPREE study population, study cohort, and according to the composite endpoint.
Table S3: Baseline characteristics of the study participants by persistent physical disability, dementia, and all‐cause mortality.
Table S4: Absolute risks, restricted mean survival times (yrs saved), and people prevented of the composite endpoint and sub‐components (death, dementia, and persistent physical disability) in relation to lifestyle categories.
Table S5: Hazard ratios of dementia, persistent physical disability, and all‐cause mortality in relation to individual lifestyle factors.
Table S6: Hazard ratios of the composite end point according to lifestyle categories within demographic, anthropometric, and health subgroups.
Table S7: Hazard ratios of the composite end point according to lifestyle categories in sensitivity analyses.
Acknowledgments
The authors have nothing to report. Open access publishing facilitated by Monash University, as part of the Wiley ‐ Monash University agreement via the Council of Australasian University Librarians
References
- 1. Garmany A. and Terzic A., “Global Healthspan‐Lifespan Gaps Among 183 World Health Organization Member States,” JAMA Network Open 7, no. 12 (2024): e2450241, 10.1001/jamanetworkopen.2024.50241. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Collaborators GBDRF , “Global, Regional, and National Comparative Risk Assessment of 79 Behavioural, Environmental and Occupational, and Metabolic Risks or Clusters of Risks, 1990‐2015: A Systematic Analysis for the Global Burden of Disease Study 2015,” Lancet 388, no. 10053 (2016): 1659–1724, 10.1016/S0140-6736(16)31679-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. McLaughlin D., Adams J., Almeida O. P., et al., “Are the National Guidelines for Health Behaviour Appropriate for Older Australians? Evidence From the Men, Women and Ageing Project,” Australasian Journal on Ageing 30, no. Suppl 2 (2011): 13–16, 10.1111/j.1741-6612.2010.00498.x. [DOI] [PubMed] [Google Scholar]
- 4. World Health Organization , “Fact Sheet: Ageing and Health,” 2021, https://www.who.int/news‐room/fact‐sheets/detail/ageing‐and‐health.
- 5. Beard J. R., Officer A., de Carvalho I. A., et al., “The World Report on Ageing and Health: A Policy Framework for Healthy Ageing,” Lancet 387, no. 10033 (2016): 2145–2154, 10.1016/S0140-6736(15)00516-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Artaud F., Dugravot A., Sabia S., Singh‐Manoux A., Tzourio C., and Elbaz A., “Unhealthy Behaviours and Disability in Older Adults: Three‐City Dijon Cohort Study,” BMJ 347 (2013): f4240, 10.1136/bmj.f4240. [DOI] [PubMed] [Google Scholar]
- 7. Zhang S., Tomata Y., Discacciati A., et al., “Combined Healthy Lifestyle Behaviors and Disability‐Free Survival: The Ohsaki Cohort 2006 Study,” Journal of General Internal Medicine 34, no. 9 (2019): 1724–1729, 10.1007/s11606-019-05061-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Rizzuto D., Bellocco R., Kivipelto M., Clerici F., Wimo A., and Fratiglioni L., “Dementia After Age 75: Survival in Different Severity Stages and Years of Life Lost,” Current Alzheimer Research 9, no. 7 (2012): 795–800, 10.2174/156720512802455421. [DOI] [PubMed] [Google Scholar]
- 9. McNeil J. J., Woods R. L., Nelson M. R., et al., “Baseline Characteristics of Participants in the ASPREE (ASPirin in Reducing Events in the Elderly) Study,” Journals of Gerontology. Series A, Biological Sciences and Medical Sciences 72, no. 11 (2017): 1586–1593, 10.1093/gerona/glw342. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. McNeil J. J., Nelson M. R., Woods R. L., et al., “Effect of Aspirin on All‐Cause Mortality in the Healthy Elderly,” New England Journal of Medicine 379, no. 16 (2018): 1519–1528, 10.1056/NEJMoa1803955. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. McNeil J. J., Woods R. L., Nelson M. R., et al., “Effect of Aspirin on Disability‐Free Survival in the Healthy Elderly,” New England Journal of Medicine 379, no. 16 (2018): 1499–1508, 10.1056/NEJMoa1800722. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Ernst M. E., Broder J. C., Wolfe R., et al., “Health Characteristics and Aspirin Use in Participants at the Baseline of the ASPirin in Reducing Events in the Elderly‐eXTension (ASPREE‐XT) Observational Study,” Contemporary Clinical Trials 130 (2023): 107231, 10.1016/j.cct.2023.107231. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. McNeil J. J., Woods R. L., Ward S. A., et al., “Cohort Profile: The ASPREE Longitudinal Study of Older Persons (ALSOP),” International Journal of Epidemiology 48, no. 4 (2019): 1048–1049h, 10.1093/ije/dyy279. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Robb C., Carr P. R., Ball J., et al., “Association of a Healthy Lifestyle With Mortality in Older People,” BMC Geriatrics 23, no. 1 (2023): 646, 10.1186/s12877-023-04247-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Lourida I., Hannon E., Littlejohns T. J., et al., “Association of Lifestyle and Genetic Risk With Incidence of Dementia,” Journal of the American Medical Association 322, no. 5 (2019): 430–437, 10.1001/jama.2019.9879. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Neumann J. T., Freak‐Poli R., Orchard S. G., et al., “Alcohol Consumption and Risks of Cardiovascular Disease and All‐Cause Mortality in Healthy Older Adults,” European Journal of Preventive Cardiology 29, no. 6 (2022): e230–e232, 10.1093/eurjpc/zwab177. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Clayton‐Chubb D., Vaughan N. V., George E. S., et al., “Mediterranean Diet and Ultra‐Processed Food Intake in Older Australian Adults‐Associations With Frailty and Cardiometabolic Conditions,” Nutrients 16, no. 17 (2024), 10.3390/nu16172978. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Tognon G., Lissner L., Saebye D., Walker K. Z., and Heitmann B. L., “The Mediterranean Diet in Relation to Mortality and CVD: A Danish Cohort Study,” British Journal of Nutrition 111, no. 1 (2014): 151–159, 10.1017/S0007114513001931. [DOI] [PubMed] [Google Scholar]
- 19. Trichopoulou A., Kouris‐Blazos A., Wahlqvist M. L., et al., “Diet and Overall Survival in Elderly People,” BMJ 311, no. 7018 (1995): 1457–1460, 10.1136/bmj.311.7018.1457. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Trichopoulou A., Orfanos P., Norat T., et al., “Modified Mediterranean Diet and Survival: EPIC‐Elderly Prospective Cohort Study,” BMJ 330, no. 7498 (2005): 991, 10.1136/bmj.38415.644155.8F. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Panagiotakos D. B., Pitsavos C., and Stefanadis C., “Dietary Patterns: A Mediterranean Diet Score and Its Relation to Clinical and Biological Markers of Cardiovascular Disease Risk,” Nutrition, Metabolism, and Cardiovascular Diseases 16, no. 8 (2006): 559–568, 10.1016/j.numecd.2005.08.006. [DOI] [PubMed] [Google Scholar]
- 22. Department of Health and Aged Care , “Physical Activity and Exercise Guidelines for all Australians,” (2021), https://www.health.gov.au/health‐topics/physical‐activity‐and‐exercise/physical‐activity‐and‐exercise‐guidelines‐for‐all‐australians.
- 23. World Health Organistation , “WHO Guidelines on Physical Activity and Sedentary Behaviour,” (2020), https://www.who.int/publications/i/item/9789240015128. [PubMed]
- 24. Carr P. R., Weigl K., Jansen L., et al., “Healthy Lifestyle Factors Associated With Lower Risk of Colorectal Cancer Irrespective of Genetic Risk,” Gastroenterology 155, no. 6 (2018): 1805–1815.e5, 10.1053/j.gastro.2018.08.044. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Chakravarty E. F., Hubert H. B., Krishnan E., Bruce B. B., Lingala V. B., and Fries J. F., “Lifestyle Risk Factors Predict Disability and Death in Healthy Aging Adults,” American Journal of Medicine 125, no. 2 (2012): 190–197, 10.1016/j.amjmed.2011.08.006. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Knoops K. T., de Groot L. C., Kromhout D., et al., “Mediterranean Diet, Lifestyle Factors, and 10‐Year Mortality in Elderly European Men and Women: The HALE Project,” Journal of the American Medical Association 292, no. 12 (2004): 1433–1439, 10.1001/jama.292.12.1433. [DOI] [PubMed] [Google Scholar]
- 27. Brinkman S., Voortman T., Kiefte‐de Jong J. C., et al., “The Association Between Lifestyle and Overall Health, Using the Frailty Index,” Archives of Gerontology and Geriatrics 76 (2018): 85–91, 10.1016/j.archger.2018.02.006. [DOI] [PubMed] [Google Scholar]
- 28. Association AP , “Diagnostic and Statistical Manual of Mental Disorders: DSM IV,” 1994.
- 29. Katz S., “Assessing Self‐Maintenance: Activities of Daily Living, Mobility, and Instrumental Activities of Daily Living,” Journal of the American Geriatrics Society 31, no. 12 (1983): 721–727, 10.1111/j.1532-5415.1983.tb03391.x. [DOI] [PubMed] [Google Scholar]
- 30. Shang Y., Wu W., Dove A., et al., “Healthy Behaviors, Leisure Activities, and Social Network Prolong Disability‐Free Survival in Older Adults With Diabetes,” Journals of Gerontology. Series A, Biological Sciences and Medical Sciences 77, no. 10 (2022): 2093–2101, 10.1093/gerona/glac054. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. Wu W., Xu W., Englund S., Shang Y., Pan K. Y., and Rizzuto D., “Can Health Behaviours Prolong Survival and Compress the Period of Survival With the Disability? A Population‐Based Cohort Study,” Age and Ageing 50, no. 2 (2021): 480–487, 10.1093/ageing/afaa143. [DOI] [PubMed] [Google Scholar]
- 32. Rizzuto D., Orsini N., Qiu C., Wang H. X., and Fratiglioni L., “Lifestyle, Social Factors, and Survival After Age 75: Population Based Study,” BMJ 345 (2012): e5568, 10.1136/bmj.e5568. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33. Sakaniwa R., Noguchi M., Imano H., et al., “Impact of Modifiable Healthy Lifestyle Adoption on Lifetime Gain From Middle to Older Age,” Age and Ageing 51, no. 5 (2022): afac080, 10.1093/ageing/afac080. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34. Rahman M. M., Jagger C., Leigh L., et al., “The Impact of Education and Lifestyle Factors on Disability‐Free Life Expectancy From Mid‐Life to Older Age: A Multi‐Cohort Study,” International Journal of Public Health 67 (2022): 1605045, 10.3389/ijph.2022.1605045. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35. Jacob M. E., Yee L. M., Diehr P. H., et al., “Can a Healthy Lifestyle Compress the Disabled Period in Older Adults?,” Journal of the American Geriatrics Society 64, no. 10 (2016): 1952–1961, 10.1111/jgs.14314. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36. Dhana K., Evans D. A., Rajan K. B., Bennett D. A., and Morris M. C., “Healthy Lifestyle and the Risk of Alzheimer Dementia: Findings From 2 Longitudinal Studies,” Neurology 95, no. 4 (2020): e374–e383, 10.1212/WNL.0000000000009816. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37. Wu Z., Pandigama D. H., Wrigglesworth J., et al., “Lifestyle Enrichment in Later Life and Its Association With Dementia Risk,” JAMA Network Open 6, no. 7 (2023): e2323690, 10.1001/jamanetworkopen.2023.23690. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38. Dhuli K., Naureen Z., Medori M. C., et al., “Physical Activity for Health,” Journal of Preventive Medicine and Hygiene 63, no. 2 Suppl. 3 (2022): E150–E159, 10.15167/2421-4248/jpmh2022.63.2S3.2756. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39. Anderson B. O., Berdzuli N., Ilbawi A., et al., “Health and Cancer Risks Associated With Low Levels of Alcohol Consumption,” Lancet Public Health 8, no. 1 (2023): e6–e7, 10.1016/S2468-2667(22)00317-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40. Drewnowski A. and Shultz J. M., “Impact of Aging on Eating Behaviors, Food Choices, Nutrition, and Health Status,” Journal of Nutrition, Health & Aging 5, no. 2 (2001): 75–79. [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Table S1: Description of the ascertainment of the three components of the primary composite end point, including all‐cause mortality.
Figure S1: Flow of participant inclusion and ASPREE study eligibility criteria.
Table S2: Baseline characteristics of the ASPREE study population, study cohort, and according to the composite endpoint.
Table S3: Baseline characteristics of the study participants by persistent physical disability, dementia, and all‐cause mortality.
Table S4: Absolute risks, restricted mean survival times (yrs saved), and people prevented of the composite endpoint and sub‐components (death, dementia, and persistent physical disability) in relation to lifestyle categories.
Table S5: Hazard ratios of dementia, persistent physical disability, and all‐cause mortality in relation to individual lifestyle factors.
Table S6: Hazard ratios of the composite end point according to lifestyle categories within demographic, anthropometric, and health subgroups.
Table S7: Hazard ratios of the composite end point according to lifestyle categories in sensitivity analyses.