Abstract
Background
Increasingly, lifestyle factors in midlife are reported to impact health and functional status in old age. This work examines associations between dietary trajectories in middle age and subsequent impairments in physical function.
Method
Data are from 851 participants (61% men, mean age at first dietary assessment = 47 years, range 30–59 years) from the Baltimore Longitudinal Study of Aging. We used latent class analysis to derive dietary trajectories based on adherence to the Alternative Healthy Eating Index-2010 (AHEI), and further classified them based on tertiles, as poor (score <39.3), intermediate (39.3–48.9), or good (>48.9). Physical function was assessed with the Short Physical Performance Battery (SPPB). Random effects tobit regression models were used to examine associations between dietary trajectories and later physical function.
Results
Two latent classes of AHEI scores were generated and labeled “greatly improved” or “moderately improved.” In the greatly improved class, participants showed a trend in overall AHEI score from poor/intermediate to good diet categories across dietary assessments with age, over time. In the moderately improved class, the overall AHEI score shifted from poor to intermediate diet categories over time, and the prevalence of the good diet category remained low. Mean AHEI score between ages 30 and 59 years was higher in the greatly, than moderately, improved class. The moderately improved class had 1.6 points lower SPPB score (indicating poorer physical function) at older age than the greatly improved class (p = .022).
Conclusions
Findings suggest that improving diet quality in middle age may contribute to better physical function in older age.
Keywords: Alternative Healthy Eating Index, Baltimore Longitudinal Study of Aging, Diet trajectory, Physical performance, Short Physical Performance Battery
Diet is an important modifiable health behavior contributing to well-being over the life course. Dietary patterns have been associated with mortality and many chronic conditions, including cardiovascular disease, diabetes and cancer, as well as physical and cognitive function (1–4). Results from the Whitehall II prospective cohort study suggest that diet during middle age can affect long-term health outcomes; as middle-aged participants with better diet quality had lower incident cardiovascular disease and diabetes in later life, and lower all-cause mortality (5–7). Dietary patterns may change in response to life events such as disease diagnosis, social factors, or new nutrition knowledge (8). Therefore, it is important to measure change in dietary pattern during adulthood and assess its association with health outcomes in older age.
Physical function, an important indicator of overall health in older persons, has been shown to predict subsequent cognitive impairment, disability, poor quality of life, and mortality (9–12). Mounting evidence suggests that diet quality is associated with functional outcomes with aging. Results from the 1999–2002 National Health and Nutrition Examination Survey (NHANES) showed that better diet quality (measured with the Healthy Eating Index-2005) was associated with faster gait speed among older adults (4). The InCHIANTI study among men and women 65 years and older showed that higher adherence to a Mediterranean-style diet was associated with less decline in mobility and lower odds of frailty (13,14). The Whitehall II study showed that lower consumption of fruit and vegetables during midlife was associated with slower walking speed and lower grip strength in older age (15). A critical limitation of most previous work is the use of a single time point measure of diet or the use of an average measure of diet in middle age, neither of which considers change in diet which may occur in response to health events and/or change in dietary recommendations. Evaluating the association of change in dietary pattern over the life course with health outcomes may have more practical implications.
Our overall objective in conducting this study was to investigate the associations of dietary pattern trajectories in middle age with objective, performance-based measures of physical function in older age. We hypothesized that dietary patterns that improved over middle age would be inversely associated with impaired physical function in older age. We examined our hypothesis in the Baltimore Longitudinal Study of Aging (BLSA), which is implemented by the National Institute on Aging (NIA). These analyses included 851 men and women (61% men, 85% non-Hispanic White) between the ages of 30 and 59 years (mean age at first dietary assessment = 47 years, mean follow-up of 25.8 years, range 4.3–48.5 years) and adjusted for relevant confounders and covariates, to examine the associations between dietary pattern trajectories in middle age and physical function in older age. Specifically, using latent class analysis (LCA), we identified dietary pattern trajectories based on adherence to the Alternative Health Eating Index-2010 using dietary assessments collected over several years which allowed us examine change in dietary patterns over the life course.
Method
Study Description
The BLSA, initiated in 1958, is an ongoing prospective open cohort consisting of community-dwelling men and women (added in 1978) largely from the Washington DC–Baltimore area. Data are collected by staff-administered interview, clinical examination, and lab assessments. Once enrolled, participants are followed at varying intervals depending on their age (<60 years every 4 years, 60–79 years every 2 years, >80 years every year), typically in person at the NIA Clinical Research Unit or at a home visit for the most debilitated. A detailed description of the BLSA cohort has been provided elsewhere (16). A total of 851 men and women with valid diet record assessment between 30 and 59 years were included in the analysis. The study protocol was approved by the Institutional Review Board (IRB) of the National Institute of Environmental Health Sciences, and informed consent was obtained from participants at each visit. These analyses were determined to not involve human subjects by the George Washington University IRB.
Dietary Assessment and a Priori Dietary Pattern Score
We used dietary data collected between 1961 and 2008 using diet records. BLSA participants were instructed by trained personnel at the NIA Clinical Research Unit located at the Gerontology Research Center prior to 2003 and Medstar Harbor Hospital after 2003, on how to record daily dietary intake. Food pictures and portable scales were distributed to participants for portion size assessment. Participants completed their diet records at home and either brought them to the clinical center during their follow-up visits or sent them back by mail. Any uncertainty regarding the diet records was clarified by telephone. Dietary and nutrient intakes were derived from the diet record data using the University of Minnesota Nutrient Data System for Research program. Given the open nature of the cohort, “baseline” varied in terms of starting year and, therefore, was considered as the first visit between age 30 and 59 years with dietary assessment. We included participants with at least 3 intake days and used the first 7 intake days for those recording more than 7 days of intake. Change in dietary pattern over time from then through additional measures between age 30 and 59 years was estimated as dietary pattern trajectory for analysis.
Average intakes of specific food groups and nutrients across diet record days at each time point were calculated and used to derive the Alternative Healthy Eating Index-2010 (AHEI) score. The AHEI includes 11 food and nutrient components previously found to be associated with chronic disease (17). For each component, participants were assigned a score ranging from 0 to 10 according to their intake, where higher intake of vegetables, fruit, whole grains, nuts and legumes, and long-chain and polyunsaturated fatty acids, and lower intake of sugar-sweetened beverages and fruit juices, red/processed meat, trans fat and sodium, contributed to higher scores. The maximum score for the alcohol component was assigned for moderate intake of alcohol. The scoring for the AHEI-2010 potentially ranges from 0 to 110 points with higher scores indicating better diet quality.
Physical Function Assessment
Physical function was assessed using the Short Physical Performance Battery (SPPB), which includes 3 examinations of lower body performance: repeated chair stands, progressive standing balance, and usual gait speed. Each component is scored from 0 (worst) to 4 (best) (12). Repeated chair stand score was assessed by the time used for standing up from a chair 5 times. Standing balance score was generated based on the ability to stand in 3 different positions, from easy to difficult, side-by-side, semi-tandem, and full-tandem for 10 seconds each. Gait speed was assessed over 6 m. The SPPB score was calculated by summing the 3 component scores, and ranges from 0 to 12. Participants who had valid dietary assessment between 30 and 59 years and at least one SPPB score measured at 60 years and older were included in the analysis.
Covariates
Age at baseline, sex, race/ethnicity, education in years, smoking status, physical activity, and medical diseases were assessed through interview. Smoking status was categorized as nonsmoker, former smoker, or current smoker, based on the information obtained at the last visit before age 60 years. Education was classified as high school or below (≤12 years), any college (13–16 years), or post-college education (≥17 years). Questions about activities were asked and metabolic equivalents of task (MET) minutes per week were calculated (18). Mean physical activity between age 30 and 59 years was used and categorized into 4 groups: sedentary, MET minutes <50; low, MET minutes 50–249; moderate, MET minutes 250–499; and high, MET minutes ≥500 per week. Body mass index (BMI) was calculated as weight in kilograms divided by height in meters squared, which were measured by trained technicians, with calibrated scales, during each visit. Overweight was defined as BMI ≥25 kg/m2, and weight status at the last visit before 60 years was used for analysis. The number of diseases at baseline (age 30–59 years) was summed, and included chronic obstructive pulmonary disease, diabetes, cancer, Parkinson’s disease, bone diseases, dyslipidemia, renal disease, and cardiovascular diseases, based on ICD-9 codes.
Statistical Analysis
Tertile cutoffs of AHEI scores were first defined using dietary data from participants aged 50–59 years at baseline, as this age decade had the highest number of participants (Supplementary Table 1). These cutoff points were then used to categorize overall AHEI score for all dietary assessments at each visit between ages 30 and 59 years. Participants were classified as having poor, intermediate, or good diet at each dietary assessment if their overall AHEI scores were <39.3, between 39.3 and 48.9, or >48.9, respectively. LCA was used to develop profiles of within-person change in overall AHEI score across all visits with valid dietary data (mean number of dietary visits: 1.5, ranging 1–5). LCA hypothesizes the existence of subpopulations of adults with distinct dietary pattern trajectories, indicated by individual-level change in AHEI score over time. The analysis then aims to determine the number of subpopulations (“trajectory classes”) and provides estimates for each subpopulation: (i) prevalence of each trajectory and (ii) proportion of poor, intermediate, or good diet at each dietary assessment within each trajectory class. We identified 2 trajectory classes based on a combination of the Bayesian information criterion (BIC) (19), Lo-Mendell-Rubin adjusted likelihood ratio test (20), and scientific plausibility and meaningfulness of the resulting trajectory patterns.
Upon completing the LCA, we assigned participants into the class corresponding to the highest posterior probability of class membership. To account for missing data due to partial follow-up or missing visits between ages 30 and 59 years, the LCA model was fit using the full information maximum likelihood estimator, which is unbiased under the assumption of data missing at random (MAR) (21).
Sociodemographic characteristics were reported as mean (SD), or percentage, across the 2 AHEI latent classes, and t test and chi-squared tests were performed for comparison of continuous and categorical variables, respectively. The AHEI, overall and component scores, were also averaged over all dietary visits between age 30 and 59 years and compared across the 2 latent classes using t tests. Random effects tobit regression models were used to assess the association between dietary trajectory class and physical function, adjusting for age, sex, and race/ethnicity in Model 1, with additional adjustment for number of diseases in Model 2, with education additionally added in Model 3, and smoking status, physical activity, and overweight status additionally added in Model 4. Tobit regression (also termed “censored regression model”) was selected to address the ceiling and floor effects of the SPPB score (22). We included a participant-specific random effect to account for the within-subject correlation of SPPB scores over time. All analyses except tobit regression models were performed with SAS 9.4 (SAS institute, 2013). Tobit regression models were analyzed with Stata 13.1 (StataCorp, 2013) using the xttobit function with lower-censoring limit set at 0 and upper-censoring limit 12, corresponding respectively to the minimum and maximum of the SPPB score. Two-tailed tests of p <.05 were considered statistically significant.
Results
The 2 latent classes of AHEI scores generated through the LCA were labeled “greatly improved” and “moderately improved.” In the greatly improved class, the overall AHEI score showed an increasing trend moving from poor and intermediate to good diet categories across dietary assessments at increasing age categories (30–34, 35–39, 40–44, 45–49, 50–54, and 55–59 years) for each participant. In the moderately improved class, the overall AHEI score shifted from poor to intermediate diet categories and the prevalence of the good diet category remained low through all dietary assessments at different age categories (Figure 1).
Figure 1.
The prevalence of each diet quality category, based on Alternative Healthy Eating Index (AHEI) from the results of the 2-class latent class analysis (LCA) among men and women (since 1978) aged 30–59 years, in the Baltimore Longitudinal Study of Aging. Class 1 represents greatly improved; class 2 represents moderately improved.
The mean age at first dietary visit was 47 (SD = 8.8) years and 61% of individuals were men (Table 1). The majority were non-Hispanic White (80%), with education at the college level or above (93%). Based on the LCA, 35% (n = 295) were classified into the greatly improved class and the others (n = 556) into the moderately improved class. Compared to moderately improved, the greatly improved class included more women (p < .001), non-Hispanic Black or other race/ethnicities (p = .001), never smokers (p < .001), those with fewer chronic diseases at baseline (p < .001), and fewer who were overweight at last visit before 60 years (p < .001). The greatly improved class had higher mean SPPB score over visits with available physical function measurements at age 60 years and above (p < .001). There were no differences in education or physical activity level between the greatly and moderately improved classes (both p > .05).
Table 1.
Sociodemographic Characteristics Among Men and Women With Dietary Assessment Data at Age 30–59 Years in the Baltimore Longitudinal Study of Aging, 1961–2008
| Total | AHEI Class 1 (greatly improved) | AHEI Class 2 (moderately improved) | p | |
|---|---|---|---|---|
| N (%) | 851 | 295 (35) | 556 (65) | |
| Age at first dietary visit (baseline), mean (SD), y | 46.5 (8.8) | 46.1 (9.3) | 46.7 (8.4) | .34 |
| Men, n (%) | 516 (61) | 119 (40) | 397 (71) | <.001 |
| Race/ethnicity, n (%) | .001 | |||
| Non-Hispanic White | 678 (80) | 221 (75) | 457 (82) | |
| Non-Hispanic Black | 131 (15) | 49 (17) | 82 (15) | |
| Other | 42 (5) | 25 (8) | 17 (3) | |
| Education categories, n (%), n = 843 | .21 | |||
| High school and below | 55 (7) | 13 (4) | 42 (8) | |
| College | 244 (29) | 87 (30) | 157 (28) | |
| Post-college | 544 (64) | 192 (66) | 352 (64) | |
| Smoking status at last visit before 60, n (%) n = 849 | <.001 | |||
| Never | 368 (43) | 163 (56) | 205 (37) | |
| Former | 429 (51) | 118 (40) | 331 (56) | |
| Current | 52 (6) | 13 (4) | 39 (7) | |
| Overweight at last visit before 60 y, n (%), n = 842 | 527 (63) | 155 (53) | 372 (68) | <.001 |
| Number of chronic diseases at 30–59 y, n (%), n = 850 | <.001 | |||
| 0 | 140 (16) | 73 (25) | 67 (12) | |
| 1 | 341 (40) | 115 (39) | 226 (41) | |
| 2 | 252 (30) | 82 (28) | 170 (30) | |
| ≥3 | 117 (14) | 24 (8) | 93 (17) | |
| Mean physical activity at 30–59 y, n (%), n = 744 | .55 | |||
| Sedentary | 18 (2) | 4 (2) | 14 (3) | |
| Low | 182 (25) | 66 (26) | 116 (24) | |
| Intermediate | 320 (43) | 104 (41) | 216 (44) | |
| High | 224 (30) | 80 (31) | 144 (29) | |
| SPPBa, mean (SD), n = 410 | 10.8 (2.4) | 11.3 (1.7) | 10.5 (2.7) | <.001 |
Notes: AHEI = Alternative Healthy Eating Index; SD = standard deviation.
aAveraged Short Physical Performance Battery (SPPB) score over visits with available physical function measurements at 60 years and older.
The averaged AHEI score over all dietary visits for the greatly improved class was significantly higher than that for the moderately improved class (55 [SD = 7.2] vs 38 [SD = 7.2], p < .001) (Table 2). The averaged individual AHEI component scores were also higher for the greatly improved class than the moderately improved class (all p < .001).
Table 2.
Alternative Healthy Eating Index (AHEI) Mean (SD) for Averaged Full Score and Components Over All Dietary Visits, by Dietary Pattern Trajectory Class Among Men and Women (since 1978) Aged 30–59 Years in the Baltimore Longitudinal Study of Aging, 1961–2008
| AHEI Class 1 (greatly improved) | AHEI Class 2 (moderately improved) | p | |
|---|---|---|---|
| AHEI score for: | |||
| Vegetables | 3.0 (1.6) | 2.4 (1.3) | <.001 |
| Fruits | 2.2 (1.5) | 1.4 (1.2) | <.001 |
| Whole grains | 2.7 (2.0) | 1.4 (1.4) | <.001 |
| Sugar-sweetened beverages and fruit juice | 4.9 (3.4) | 3.2 (3.2) | <.001 |
| Legume and nuts | 5.8 (3.3) | 3.7 (2.9) | <.001 |
| Red and processed meat | 5.7 (3.0) | 2.2 (2.6) | <.001 |
| Trans fat | 6.0 (2.0) | 4.1 (2.0) | <.001 |
| Omega-3 fatty acid | 6.5 (3.1) | 4.7 (3.1) | <.001 |
| Polyunsaturated fatty acid | 6.4 (2.1) | 5.6 (1.7) | <.001 |
| Alcohol | 6.2 (3.1) | 4.8 (3.1) | <.001 |
| Sodium | 5.7 (2.7) | 4.6 (2.7) | <.001 |
| AHEI overall score | 55.2 (7.2) | 38.0 (7.2) | <.001 |
The available SPPB measures for each individual at 60 years and older ranged from 1 to 9, with a mean of 3.3. After adjusting for age, sex, race/ethnicity, education, smoking, physical activity, number of diseases, and overweight status, the moderately improved class had a SPPB score that was 1.6 points lower in older age than the greatly improved class (β = −1.59, 95% CI −2.96 to −0.23, p = .022) among 410 participants with available SPPB scores at age 60 years and older (Table 3). However, the association between dietary pattern trajectory and SPPB was attenuated after additional adjustment for birth decade in the sensitivity analysis (β = −0.48, 95% CI −1.58 to 0.62, p = .39, Supplementary Table 2).
Table 3.
Association Between Dietary Pattern Trajectory (moderately vs greatly improved) During Ages 30–59 Years and Short Physical Performance Battery (SPPB) Score at Age 60 Years or Older, Among Men and Women (since 1978) in the Baltimore Longitudinal Study of Aging
| β | SE | 95% CI | p | |
|---|---|---|---|---|
| Model 1a | −2.12 | 0.66 | −3.41, −0.82 | .001 |
| Model 2b | −2.04 | 0.67 | −3.34, −0.73 | .002 |
| Model 3c | −2.01 | 0.66 | −3.31, −0.71 | .002 |
| Model 4d | −1.59 | 0.70 | −2.96, −0.23 | .022 |
Notes: a Adjusting for age, sex, and race/ethnicity. b Adjusting for age, sex, race/ethnicity, and number of diseases. c Adjusting for age, sex, race/ethnicity, number of diseases, and education. d Adjusting for age, sex, race/ethnicity, number of diseases, education, smoking, physical activity, and overweight status.
Discussion
We classified individuals into 2 dietary pattern trajectory classes based on their reported dietary intakes during ages 30–59 years. Our results indicate that improving diet quality in middle age was associated with better physical function at older age.
The mean of averaged overall AHEI score across diet visits among all participants was 44 out of 110, which is similar to the general U.S. population. In comparison, the representative NHANES reported mean AHEI scores of 40–48 from 1999 to 2012 for adults aged 20 years and older (23). Among individual components of the AHEI, trans fat, polyunsaturated fatty acids, and omega-3 fatty acids yielded a relatively higher score than other components in the BLSA cohort, indicating that the quality of fat intake was better than that of other food groups. Intakes of fruit, vegetables, and whole grains were relatively low in this cohort. This was also seen in the NHANES data, where AHEI scores for trans fat, polyunsaturated fatty acids, and red and processed meat were higher than other components, while that for whole grains was the lowest (23). However, intakes of omega-3 fatty acids and nuts and legumes appeared to be better in the BLSA cohort than in NHANES. Therefore, whole grains, fruits, and vegetables intake should be a primary focus to improve dietary quality in the BLSA cohort.
Change in pattern of dietary intake during adulthood could influence many chronic conditions in later life, although few studies have examined dietary pattern trajectories and health outcomes. A recent analysis from the Whitehall II study showed an improving trajectory in AHEI score among individuals free of dementia, and the AHEI score was significantly higher among dementia-free individuals after 10 years since the first dietary assessment relative to those diagnosed with dementia during this time. Importantly, however, no association was seen between AHEI score assessed at baseline and incident dementia (24). Women in the Medical Research Council National Survey of Health and Development with higher adherence to a dietary pattern characterized by high intake of fruit, vegetables, and dairy, based on 3 dietary assessments at middle age, had lower BMI, waist circumference, and blood pressure at older age; and men with higher adherence to a mixed pattern (including mixed dishes, fruit, vegetables, and desserts) at middle age had lower waist circumference and blood pressure at older age (25). An analysis of Framingham Offspring data found that, compared to individuals who did not change their eating pattern across 3 dietary assessments, those whose eating pattern moved towards poorer quality had twice the risk of being overweight or obese at the last assessment (26). A similar LCA on dietary trajectory in the China Health and Nutrition Survey cohort demonstrated that individuals with consistently good quality diet over time had lower glycated hemoglobin than those with declining diet quality or consistently poor quality diet (27). Together, these results suggest that the study of dietary pattern trajectories may be more meaningful than single dietary measurements or average dietary pattern score, as it can show that improvements in diet at particular life stages can affect health outcomes.
Our results show that improving diet quality in middle age was associated with better physical function at older ages, consistent with previous studies that have shown protective effects of good diet quality in midlife on physical function at older ages (13–15). Biological mechanisms for these associations are likely through higher exposure to beneficial nutrients which are associated with lower risk of physical decline. For example, oxidative damage caused by free radicals has been associated with physical functional impairment (28), and dietary intakes of antioxidants such as vitamin C and β-carotene and plasma antioxidants such as α-tocopherol and γ-tocopherol have been shown to be significantly protective for knee extension and SPPB score among older adults (29). Dietary intakes of omega-3 fatty acids have been positively correlated with bone mineral density and muscle strength, and negatively correlated with time needed to complete repeated chair stands among older individuals (30); as well as with faster gait speed, longer waking distance, less time used for “timed up and go” test and higher grip strength in men and women aged 50 years and older (31). Other nutrients such as protein, vitamin D, vitamin E, and folate have also been reported to have beneficial associations with physical performance among older adults (32–34).
To our knowledge, this is the first study to investigate the association between dietary trajectory in midlife and physical function at older age. The BLSA cohort has a long follow-up time for dietary and physical function measures, with repeated measures. LCA is an advanced statistical method that allows evaluation of change in dietary pattern with within-subject repeated measures of diet through middle age, and modeling of the trajectory of dietary pattern and repeated measures of SPPB over time. Our study also has limitations. Participants in the BLSA cohort are mainly non-Hispanic White with high levels of education, and women were not added into study until 1978, hence the results may not be generalizable to the general U.S. population. Our analyses focused on dietary determinants in middle age as a predictor of physical performance measures in older age. However, in addition to diet, other lifestyle factors including physical activity, sedentary behavior, and smoking status are also associated with functional declines in older age (35–37). Several of these behaviors also may “track” together such that an individual who follows a better quality diet, may also have higher physical activity levels, lower sedentary behaviors, and may be a nonsmoker. We attempted to account for this by adjusting for these variables in our models, however, unmeasured and residual confounding cannot be completely ruled out.
Due to the open cohort nature of the study, these analyses included participants who were born from the 1900s to 1950s. Birth cohort was correlated with many of the covariates, and as the change of dietary pattern over time (dietary trajectory) was analyzed in the model, the dietary trajectory classes may reflect cohort effects, thus, including birth cohort may constitute an overadjustment (Supplementary Table 3).
Conclusions
Improving diet quality in midlife was significantly associated with better physical function at older age. Diet as a modifiable health factor plays an important role in preventing aging-related chronic diseases. As the AHEI score was generally low indicating poor diet quality, improving diet quality at all life stages may be an important way to protect against or delay the onset of functional decline. However, studies with long-term dietary intervention are needed to establish causality.
Funding
This study was supported by National Institute on Aging, grant no. R01AG051752 to S.A.T. The sponsoring institution did not interfere with the collection, analysis, presentation, or interpretation of the data reported here.
Supplementary Material
Acknowledgments
We thank the BLSA participants and the National Institute on Aging for data collection and management.
Conflict of Interest
None declared.
Author Contributions
S.A.T. designed the study and contributed to analysis and interpretation, and to the drafting of the manuscript. Q.-L.X. and Y.J. performed the statistical analysis for the project and, in addition, Y.J. also contributed to the drafting of the manuscript. L.F., K.L.T., E.M.S, Q.-L.X., and T.T. provided scientific guidance, contributed to analysis and interpretation, and critically revised the manuscript.
References
- 1. Burr ML, Gilbert J, Holliday RA, et al. Effects of changes in fat, fish, and fibre intakes on death and myocardial reinfarction: Diet and Reinfarction Trial (DART). Lancet. 1989;334:757–761 doi: 10.1016/s0140-6736(89)90828-3 [DOI] [PubMed] [Google Scholar]
- 2. Tabung FK, Brown LS, Fung TT. Dietary patterns and colorectal cancer risk: a review of 17 years of evidence (2000–2016). Curr Colorect Cancer Rep. 2017;13:440–454 doi: 10.1007/s11888-017-0390-5 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Carlos S, De La Fuente-Arrillaga C, Bes-Rastrollo M, et al. Mediterranean diet and health outcomes in the SUN cohort. Nutrients. 2018;10:439. doi: 10.3390/nu10040439 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Xu B, Houston DK, Locher JL, et al. Higher Healthy Eating Index-2005 scores are associated with better physical performance. J Gerontol A: Biomed Sci Med Sci. 2011;67:93–99 doi: 10.1093/gerona/glr159 [DOI] [PubMed] [Google Scholar]
- 5. Akbaraly TN, Ferrie JE, Berr C, et al. Alternative Healthy Eating Index and mortality over 18 y of follow-up: results from the Whitehall II cohort. Am J Clin Nutr. 2011;94:247–253. doi: 10.3945/ajcn.111.013128 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Brunner EJ, Mosdøl A, Witte DR, et al. Dietary patterns and 15-y risks of major coronary events, diabetes, and mortality. Am J Clin Nutr. 2008;87:1414–1421. doi: 10.1093/ajcn/87.5.1414 [DOI] [PubMed] [Google Scholar]
- 7. McNaughton SA, Mishra GD, Brunner EJ. Dietary patterns, insulin resistance, and incidence of type 2 diabetes in the Whitehall II Study. Diabetes Care. 2008;31:1343–1348. doi: 10.2337/dc07-1946 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Krondl M, Coleman P. Social and biocultural determinants of food selection. Prog Food Nutr Sci. 1986;10:179–203. [PubMed] [Google Scholar]
- 9. Chou CH, Hwang CL, Wu YT. Effect of exercise on physical function, daily living activities, and quality of life in the frail older adults: a meta-analysis. Arch Phys Med Rehabil. 2012;93:237–244. doi: 10.1016/j.apmr.2011.08.042 [DOI] [PubMed] [Google Scholar]
- 10. Rosano C, Simonsick EM, Harris TB, et al. Association between physical and cognitive function in healthy elderly: the Health, Aging and Body Composition Study. Neuroepidemiology. 2004;24:8–14 doi: 10.1159/000081043 [DOI] [PubMed] [Google Scholar]
- 11. Reuben DB, Rubenstein LV, Hirsch SH, Hays RD. Value of functional status as a predictor of mortality: results of a prospective study. Am J Med. 1992;93:663–669. doi: 10.1016/0002-9343(92)90200-u [DOI] [PubMed] [Google Scholar]
- 12. Guralnik JM, Ferrucci L, Simonsick EM, Salive ME, Wallace RB. Lower-extremity function in persons over the age of 70 years as a predictor of subsequent disability. N Engl J Med. 1995;332:556–561. doi: 10.1056/NEJM199503023320902 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Milaneschi Y, Bandinelli S, Corsi AM, et al. Mediterranean diet and mobility decline in older persons. Exp Gerontol. 2011;46:303–308. doi: 10.1016/j.exger.2010.11.030 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Talegawkar SA, Bandinelli S, Bandeen-Roche K, et al. A higher adherence to a Mediterranean-style diet is inversely associated with the development of frailty in community-dwelling elderly men and women. J Nutr. 2012;142:2161–2166. doi: 10.3945/jn.112.165498 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Sabia S, Elbaz A, Rouveau N, Brunner EJ, Kivimaki M, Singh-Manoux A. Cumulative associations between midlife health behaviors and physical functioning in early old age: a 17-year prospective cohort study. J Am Geriatr Soc. 2014;62:1860–1868. doi: 10.1111/jgs.13071 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Shock NW. Normal human aging: the Baltimore Longitudinal Study of Aging. US Department of Health and Human Services, Public Health Service, National Institutes of Health, National Institute on Aging, Gerontology Research Center; 1984.
- 17. Chiuve SE, Fung TT, Rimm EB, et al. Alternative dietary indices both strongly predict risk of chronic disease. J Nutr. 2012;142:1009–1018. doi: 10.3945/jn.111.157222 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Ainsworth BE, Haskell WL, Whitt MC, et al. Compendium of physical activities: an update of activity codes and MET intensities. Med Sci Sports Exerc. 2000;32:S498–S504. doi: 10.1097/00005768-200009001-00009 [DOI] [PubMed] [Google Scholar]
- 19. Nylund KL, Asparouhov T, Muthén BO. Deciding on the number of classes in latent class analysis and growth mixture modeling: a Monte Carlo Simulation Study. Struct Eq Model. 2007;14:535–569 doi: 10.1080/10705510701575396 [DOI] [Google Scholar]
- 20. Lo Y, Mendell NR, Rubin DB. Testing the number of components in a normal mixture. Biometrika. 2001;88:767–778 doi: 10.1093/biomet/88.3.767 [DOI] [Google Scholar]
- 21. Rubin DB. Inference and missing data. Biometrika. 1976;63:581–592 doi: 10.1093/biomet/63.3.581 [DOI] [Google Scholar]
- 22. Tobin J. Estimation of relationships for limited dependent variables. Econometrica. 1958;26:24–36 doi: 10.2307/1907382. [DOI] [Google Scholar]
- 23. Wang DD, Li Y, Chiuve SE, Hu FB, Willett WC. Improvements in US diet helped reduce disease burden and lower premature deaths, 1999–2012; overall diet remains poor. Health Affairs. 2015;34:1916–1922 doi: 10.1377/hlthaff.2015.0640 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Akbaraly TN, Singh-Manoux A, Dugravot A, Brunner EJ, Kivimäki M, Sabia S. Association of midlife diet with subsequent risk for dementia. J Am Med Assoc. 2019;321:957–968. doi: 10.1001/jama.2019.1432 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. McNaughton SA, Mishra GD, Stephen AM, Wadsworth ME. Dietary patterns throughout adult life are associated with body mass index, waist circumference, blood pressure, and red cell folate. J Nutr. 2007;137:99–105. doi: 10.1093/jn/137.1.99 [DOI] [PubMed] [Google Scholar]
- 26. Pachucki MA. Food pattern analysis over time: unhealthful eating trajectories predict obesity. Int J Obes (Lond). 2012;36:686–694. doi: 10.1038/ijo.2011.133 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. Batis C, Mendez MA, Sotres-Alvarez D, Gordon-Larsen P, Popkin B. Dietary pattern trajectories during 15 years of follow-up and HbA1c, insulin resistance and diabetes prevalence among Chinese adults. J Epidemiol Community Health. 2014;68:773–779. doi: 10.1136/jech-2013-203560 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. De la Fuente M. Effects of antioxidants on immune system ageing. Eur J Clin Nutri. 2002;56:S5. doi: 10.1038/sj.ejcn.1601476 [DOI] [PubMed] [Google Scholar]
- 29. Cesari M, Pahor M, Bartali B, et al. Antioxidants and physical performance in elderly persons: the Invecchiare in Chianti (InCHIANTI) study. Am J Clin Nutr. 2004;79:289–294. doi: 10.1093/ajcn/79.2.289 [DOI] [PubMed] [Google Scholar]
- 30. Rousseau JH, Kleppinger A, Kenny AM. Self-reported dietary intake of omega-3 fatty acids and association with bone and lower extremity function. J Am Geriatr Soc. 2009;57:1781–1788. doi: 10.1111/j.1532-5415.2008.01870.x [DOI] [PubMed] [Google Scholar]
- 31. Wu IC, Chang HY, Hsu CC, et al. Association between dietary fiber intake and physical performance in older adults: a nationwide study in Taiwan. PLoS ONE. 2013;8:e80209. doi: 10.1371/journal.pone.0080209 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32. Bartali B, Frongillo EA, Bandinelli S, et al. Low nutrient intake is an essential component of frailty in older persons. J Gerontol A Biol Sci Med Sci. 2006;61:589–593. doi: 10.1093/gerona/61.6.589 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33. Houston DK, Cesari M, Ferrucci L, et al. Association between vitamin D status and physical performance: the InCHIANTI study. J Gerontol A Biol Sci Med Sci. 2007;62:440–446. doi: 10.1093/gerona/62.4.440 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34. Ng TP, Aung KC, Feng L, Scherer SC, Yap KB. Homocysteine, folate, vitamin B-12, and physical function in older adults: cross-sectional findings from the Singapore Longitudinal Ageing Study. Am J Clin Nutr. 2012;96:1362–1368. doi: 10.3945/ajcn.112.035741 [DOI] [PubMed] [Google Scholar]
- 35. Gil-Salcedo A, Dugravot A, Fayosse A, et al. Healthy behaviors at age 50 years and frailty at older ages in a 20-year follow-up of the UK Whitehall II cohort: a longitudinal study. PLoS Med. 2020;17:e1003147. doi: 10.1371/journal.pmed.1003147 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36. DiPietro L, Jin Y, Talegawkar S, Matthews CE. The joint associations of sedentary time and physical activity with mobility disability in older people: the NIH-AARP Diet and Health Study. J Gerontol A Biol Sci Med Sci. 2018;73:532–538. doi: 10.1093/gerona/glx122 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37. Brunner EJ, Welch CA, Shipley MJ, Ahmadi-Abhari S, Singh-Manoux A, Kivimäki M. Midlife risk factors for impaired physical and cognitive functioning at older ages: a cohort study. J Gerontol A Biol Sci Med Sci. 2017;72:237–242. doi: 10.1093/gerona/glw092 [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.