Abstract
Background
The EAT-Lancet Commission proposed the Planetary Health Diet to improve both human and planetary health. Evidence among older adults on its association with mortality and cardiovascular disease (CVD) remains limited. We therefore examined associations of the Planetary Health Diet Index (PHDI) with risks of all-cause mortality and CVD in a representative sample of U.S. middle-aged and older adults.
Methods
This cohort study used data from the U.S. Health and Retirement Study (HRS; 2012–2018) with dietary intake data from the 2013 Health Care and Nutrition Study (HCNS). Adherence to the PHDI was computed from 24-hour dietary recalls across 15 food groups (score range 0–140). All-cause mortality was ascertained via household surveys, and incident cardiovascular disease (CVD) was identified by self-reported physician-diagnosed heart disease. Cox proportional hazards models estimated hazard ratios (HRs) and 95% confidence intervals (CIs) for the association between quintiles of PHDI and risks of mortality and CVD, adjusting for demographic, lifestyle, and health-related covariates, including total energy intake.
Results
The total cohort included 7,873 participants (58.83% women; mean age 66.72 years) for the mortality analysis, and the sub-cohort included 5,454 participants (60.80% women; mean age 64.65 years) for the cardiovascular disease analysis. Higher PHDI adherence was associated with lower risks of total mortality. Participants in the highest PHDI quintile had a 59% lower risk of total mortality compared with those in the lowest quintile (HRs: 0.41; 95% CI: 0.32, 0.53). Each 1 SD increase in PHDI was associated with a 25% reduction in all-cause mortality (HRs: 0.75; 95% CI: 0.69, 0.81). PHDI was also inversely associated with CVD risk (HRs: 0.72; 95% CI: 0.58, 0.91).
Conclusions
In this cohort of U.S. older adults, higher adherence to the PHDI was associated with lower risks of all-cause mortality and cardiovascular disease.
Supplementary Information
The online version contains supplementary material available at 10.1186/s12937-026-01286-x.
Keywords: Planetary health diet index, Mortality, Cardiovascular disease, Epidemiology, Planetary health
Introduction
Mortality rates are high among older adults regardless of cause, and deaths due to chronic diseases remain substantial in this age group [1–3]. Dietary patterns play a critical role in shaping long-term health, influencing risks of all-cause mortality and chronic diseases such as cardiovascular disease (CVD) [4, 5]. Globally, poor diet quality is a leading risk factor for premature death, contributing to an estimated 11 million deaths annually, primarily from CVD, cancer, and type 2 diabetes [1, 2]. Conversely, adherence to diets abundant in plant-based foods and low in saturated fats and refined sugars has been consistently associated with lower risks of CVD incidence and mortality [6, 7]. Early prevention of CVD is therefore essential.
In this context, the EAT-Lancet diet, introduced in 2019, has been widely studied for its dual aim of promoting human health and environmental sustainability [4–6]. In this context, the EAT-Lancet diet, introduced in 2019, has been widely studied for its dual aim of promoting human health and environmental sustainability. With the global population approaching 10 billion, ensuring sustainable access to healthy diets for all is a major challenge. Several plant-forward dietary patterns—such as the Healthy Eating Index (HEI), the alternate Mediterranean diet score (aMED), and the Alternate Healthy Eating Index (AHEI-2010) have been associated with lower risks of chronic diseases, plausibly through lower glycemic load and anti-inflammatory properties [8–10]. Distinct from these indexes, the PHDI explicitly incorporates both human health and environmental sustainability considerations, positioning it as a relevant framework for the dual health–planet mandate.
The Planetary Health Diet is predominantly plant-based, emphasizes moderate total energy intake, and is rich in unsaturated fats while low in saturated fats, animal-source foods, ultra-processed products, and added sugars [11, 12]. Unlike strictly plant-based diets, it allows moderate amounts of animal products, such as fish and dairy, enhancing adaptability and feasibility across diverse populations [13].
Despite growing interest, evidence on the long-term associations of adherence to the PHDI with all-cause mortality and CVD remains limited. For example, among U.S. Black women, greater adherence to the Planetary Health Diet has been linked to lower risks of all-cause mortality and cause-specific mortality from CVD and cancer [14–16]. In addition, another study reported that higher PHDI adherence was associated with a reduced risk of CVD across three large cohorts of men and women. However, data specific to older adults—who face elevated risks of both mortality and CVD—are scarce. This gap is important because age-related changes in health status and lifestyle may affect dietary adherence and potentially modify diet–disease relationships [17].
We aimed to investigate whether higher adherence to the PHDI is associated with lower risks of all-cause and cause-specific mortality (including deaths due to CVD, cancer, infectious diseases, and neurological disorders) and with the incidence of CVD among older U.S. adults. By focusing on these outcomes, our study seeks to provide a comprehensive assessment of the potential health benefits associated with adherence to the PHDI in later life.
Methods
Study population
We used data from the Health and Retirement Study (HRS), a nationally representative longitudinal cohort of U.S. adults aged 50 years and older. The HRS collects detailed data on topics such as health, healthcare utilization, income and wealth, work and retirement, family connections, and cognitive function through biennial surveys conducted using a mixed-mode design (in-person and telephone interviews). In addition to its core surveys, the HRS includes supplementary sub-studies conducted between survey waves.
For this analysis, we combined data from the 2012–2018 core HRS wave and the 2013 Health Care and Nutrition Study (HCNS), a sub-study that collected detailed dietary intake information via a validated food frequency questionnaire (FFQ). Participants with missing HCNS dietary data, implausible energy intake (< 600 or > 3,500 kcal/day), or participants with missing follow-up information or with follow-up time equal to 0 were excluded. After applying these exclusion criteria, the final analytic sample consisted of 7,873 participants. Furthermore, participants were excluded if they had a history of CVD at baseline, or with missing CVD records data, the subcohort analysis sample consisted of 5,454 participants (Fig. 1).
Fig. 1.
Flowchart of selection of the Health and Retirement Study (HRS) participants
Dietary assessment and PHDI
The PHDI scoring system is based on 15 food and nutrient categories: whole grains, tubers, vegetables, whole fruit, dairy foods, chicken and other poultry, nuts, soy foods, non-soy legumes, fish and shellfish, eggs, poultry, red/ processed meat, added sugars, added fat-unsaturated oils and added fat-saturated oils and trans fats. In this study, calculations of food and nutrient quantities were performed using the “Dietaryindex” R package, which first quantifies the daily intake (in grams) of various food and nutrient categories [18]. These definitions, together with the detailed calculation procedures, algorithm, and R code of the “Dietaryindex” package, are well documented and publicly available at https://github.com/jamesjiadazhan/dietaryindex. Each component is scored from 0 (least favorable) to 10 (most favorable) to reflect its health impact, with soy foods and non-soy legumes weighted at 0.5 and all other food groups weighted at 1. The final PHDI score was derived by summing the scores across all categories, ranging from 0 (no adherence) to 140 (perfect adherence). A higher score indicates greater alignment with the planetary health diet. Individual scores are proportionally calculated based on intake levels within the predefined range. Further details on the calculation of the PHDI are provided in Table S1 and Table S2 [see supplement information].
Health outcome measures
In this study, health outcomes include all-cause mortality and CVD outcomes, with follow-up spanning from April 2012 to June 2019. All-cause mortality data were collected through household surveys, with the earliest reported date of death used when multiple reports were available. Causes of death were classified into nontrauma, cardiovascular diseases, cancer, neurodegenerative diseases, respiratory diseases, infectious diseases, and other causes, following the International Classification of Diseases (ICD) system. Each participant provided follow-up time from the date of the baseline questionnaire until the date of death or the end of follow-up. Cardiovascular disease outcomes were identified based on self-reported diagnoses of myocardial infarction or stroke, verified by hospital records. In cases of inconsistent follow-up responses, the first recorded event was considered. The follow-up period for CVD outcomes was measured from the date of the baseline questionnaire until the date of diagnosis or the end of follow-up.
Covariates
Study covariates were obtained from either the 2012 core HRS survey or the 2013 HCNS survey, based on prior literature and established conceptual frameworks in nutritional epidemiology. Self-reported sociodemographic characteristics included: age, gender (male, female), race (White, Black, or others), education level (less than high school, high school graduate or equivalent, some college or college graduate), marital status (married, others-separated, divorced, widowed, never married), body mass index (BMI) (kg/m2) (underweight [< 18.5] and normal weight [18.5–24.9], overweight [25.0-29.9], obese [ ≧ 30.0]), total intake energy, physical activity (PA) (had/no; “had” defined as engaging in at least one type of vigorous, moderate, or mild physical activity), alcohol use (yes/no), smoker status (yes/never) and self-reported diagnosis high blood pressure, diabetes, and cancer diseases (yes/no).
Statistical analysis
Baseline characteristics were reported across quintiles of the PHDI. Continuous outcome variables are represented as means (standard deviation, SD) or medians (interquartile range) and were analyzed using analysis of variance (ANOVA). Categorical outcome variables are represented as proportions (%) and were analyzed for statistical significance using the chi-square test. This study used a cox proportional hazards regression model with time-varying covariates to estimate the hazard ratios (HRs) and 95% confidence intervals (CIs) of the PHDI associated with the all-cause mortality and cardiovascular disease. We employed three different models: a model adjusted for age; a model adjusted for age, gender, race, marital status, education level and BMI; and a model further adjusted for age, gender, race, marital status, education level, BMI, total intake energy, PA, smoking status, alcohol use, and self-reported hypertension status, diabetes, and cancer diseases. The fully adjusted model was used as the main model, and the results are based on this model unless otherwise stated. The covariates included in the model were identified from the literature to indicate potential confounding of diet-disease associations. Further, we estimated outcome risk per 1-SD increase in PHDI by modeling the score as a continuous variable. Restricted spline regression using four knots at the 5th, 35th, 65th, and 95th percentiles of the PHDI distribution (per Harrell’s method) was conducted to evaluate the associations with all-cause mortality and cardiovascular disease. In sensitivity analyses, we evaluated the associations between PHDI and all-cause mortality and CVD using deciles of PHDI adherenc; we employed Fine and Gray’s competing risk models to account for the possibility that death could preclude CVD diagnosis. Stratified analyses by sex and age group (< 65 vs. ≥65 years) were also conducted. All statistical analyses were carried out by R software (Version 4.1.2). All P-values were two-sided, and P < .05 was considered statistically significant.
Results
Participant characteristics
We included 7,873 h participants (58.83% women; mean age 66.72 years) in the total cohort analysis. The highest quintile group of PHDI had a lower proportion of individuals with a BMI ≥ 30 kg/m2, higher educational attainment, lower smoking rates, moderate alcohol consumption, higher levels of physical activity, and a lower baseline prevalence of hypertension. Additionally, participants in the higher quintiles of PHDI consumed more daily energy intake (Table 1 and Table S3). For the subcohort analysis, 5,454 participants (60.80% women; mean age 64.65 years) were eligible based on data availability. Their baseline characteristics are shown in Table 1 and Table S3. During a median follow-up of 6.5 years in the primary analysis, 845 deaths occurred. In the subcohort analysis, the median follow-up was 6.4 years, during which 938 incident CVD events were recorded.
Table 1.
Baseline characteristics of participants by quintiles of planetary health diet index (PHDI)
| Quartile of PHDI | ||||||||
|---|---|---|---|---|---|---|---|---|
| Overall | Quintile 1 | Quintile 2 | Quintile 3 | Quintile 4 | Quintile 5 | p value | ||
| Total cohort study | ||||||||
| Number of participants | 7873 | 1575 | 1574 | 1575 | 1574 | 1575 | ||
| Gender (%) | Male | 3241 (41.17) | 729 (46.29) | 670 (42.57) | 687 (43.62) | 600 (38.12) | 555 (35.24) | < 0.001 |
| Female | 4632 (58.83) | 846 (53.71) | 904 (57.43) | 888 (56.38) | 974 (61.88) | 1020 (64.76) | < 0.001 | |
| Age (years), mean (SD) | 66.72 (10.95) | 65.00 (11.21) | 66.61 (10.83) | 68.14 (10.83) | 67.61 (10.78) | 66.25 (10.84) | ||
| BMI (%) | < 25 kg/m2 | 1738 (22.08) | 354 (22.48) | 310 (19.70) | 311 (19.75) | 336 (21.35) | 427 (27.11) | < 0.001 |
| 25–29.9 kg/m2 | 2671 (33.93) | 490 (31.11) | 526 (33.42) | 551 (34.98) | 542 (34.43) | 562 (35.68) | ||
| >=30 kg/m2 | 3464 (44.00) | 731 (46.41) | 738 (46.89) | 713 (45.27) | 696 (44.22) | 586 (37.21) | ||
| Physical activity | ||||||||
| Vigorous (%) | Had | 3419 (43.43) | 548 (34.79) | 627 (39.83) | 648 (41.14) | 728 (46.25) | 868 (55.11) | < 0.001 |
| Moderate (%) | Had | 5809 (73.78) | 1064 (67.56) | 1123 (71.35) | 1150 (73.02) | 1196 (75.98) | 1276 (81.02) | < 0.001 |
| Mild (%) | Had | 6539 (83.06) | 1255 (79.68) | 1306 (82.97) | 1311 (83.24) | 1324 (84.12) | 1343 (85.27) | 0.001 |
| Race (%) | White | 5395 (68.53) | 1122 (71.24) | 1083 (68.81) | 1114 (70.73) | 1072 (68.11) | 1004 (63.75) | < 0.001 |
| Black | 1297 (16.47) | 290 (18.41) | 282 (17.92) | 263 (16.70) | 235 (14.93) | 227 (14.41) | ||
| Others | 1181 (15.00) | 163 (10.35) | 209 (13.28) | 198 (12.57) | 267 (16.96) | 344 (21.84) | ||
| Education level (%) | Less than high school | 1382 (17.55) | 347 (22.03) | 280 (17.79) | 264 (16.76) | 257 (16.33) | 234 (14.86) | < 0.001 |
| High school graduate or equivalent | 2576 (32.72) | 592 (37.59) | 571 (36.28) | 530 (33.65) | 481 (30.56) | 402 (25.52) | ||
| Some college or college graduate | 3915 (49.73) | 636 (40.38) | 723 (45.93) | 781 (49.59) | 836 (53.11) | 939 (59.62) | ||
| Marital status (%) | Married | 4614 (58.61) | 854 (54.22) | 914 (58.07) | 933 (59.24) | 945 (60.04) | 968 (61.46) | 0.001 |
| Others | 3259 (41.39) | 721 (45.78) | 660 (41.93) | 642 (40.76) | 629 (39.96) | 607 (38.54) | ||
| Alcohol use (%) | Yes | 4173 (53.00) | 762 (48.38) | 796 (50.57) | 842 (53.46) | 859 (54.57) | 914 (58.03) | < 0.001 |
| Smoker status (%) | Yes | 1711 (21.73) | 494 (31.37) | 386 (24.52) | 303 (19.24) | 263 (16.71) | 265 (16.83) | < 0.001 |
| Hypertension (%) | 4002 (50.83) | 771 (48.95) | 841 (53.43) | 857 (54.41) | 783 (49.75) | 750 (47.62) | < 0.001 | |
| Diabetes (%) | 1497 (19.01) | 239 (15.17) | 305 (19.38) | 311 (19.75) | 334 (21.22) | 308 (19.56) | < 0.001 | |
| Cancer (%) | 923 (11.72) | 157 (9.97) | 191 (12.13) | 199 (12.63) | 201 (12.77) | 175 (11.11) | 0.077 | |
| Subcohort study (CVD) | ||||||||
| Number of participants | 5454 | 1091 | 1091 | 1090 | 1091 | 1091 | ||
| Gender (%) | Male | 2138 (39.20) | 505 (46.29) | 446 (40.88) | 450 (41.28) | 386 (35.38) | 351 (32.17) | < 0.001 |
| Female | 3316 (60.80) | 586 (53.71) | 645 (59.12) | 640 (58.72) | 705 (64.62) | 740 (67.83) | ||
| Age (years), mean (SD) | 64.65 (9.39) | 62.76 (9.24) | 64.51 (9.36) | 65.80 (9.31) | 65.64 (9.33) | 64.52 (9.42) | < 0.001 | |
| BMI (%) | < 25 kg/m2 | 1150 (21.09) | 233 (21.36) | 193 (17.69) | 200 (18.35) | 225 (20.62) | 299 (27.41) | < 0.001 |
| 25–29.9 kg/m2 | 1859 (34.09) | 347 (31.81) | 376 (34.46) | 375 (34.40) | 362 (33.18) | 399 (36.57) | ||
| >=30 kg/m2 | 2445 (44.83) | 511 (46.84) | 522 (47.85) | 515 (47.25) | 504 (46.20) | 393 (36.02) | ||
| Physical activity | ||||||||
| Vigorous (%) | Had | 2780 (50.97) | 455 (41.70) | 507 (46.47) | 531 (48.72) | 585 (53.62) | 702 (64.34) | < 0.001 |
| Moderate (%) | Had | 4630 (84.89) | 853 (78.19) | 896 (82.13) | 928 (85.14) | 957 (87.72) | 996 (91.29) | < 0.001 |
| Mild (%) | Had | 5125 (93.97) | 998 (91.48) | 1018 (93.31) | 1029 (94.40) | 1036 (94.96) | 1044 (95.69) | < 0.001 |
| Race (%) | White | 3646 (66.85) | 765 (70.12) | 728 (66.73) | 743 (68.17) | 726 (66.54) | 684 (62.69) | < 0.001 |
| Black | 922 (16.91) | 206 (18.88) | 214 (19.62) | 190 (17.43) | 165 (15.12) | 147 (13.47) | ||
| Others | 886 (16.24) | 120 (11.00) | 149 (13.66) | 157 (14.40) | 200 (18.33) | 260 (23.83) | ||
| Education level (%) | Less than high school | 882 (16.17) | 214 (19.62) | 179 (16.41) | 169 (15.50) | 170 (15.58) | 150 (13.75) | < 0.001 |
| High school graduate or equivalent | 1749 (32.07) | 419 (38.41) | 400 (36.66) | 349 (32.02) | 304 (27.86) | 277 (25.39) | ||
| Some college or college graduate | 2823 (51.76) | 458 (41.98) | 512 (46.93) | 572 (52.48) | 617 (56.55) | 664 (60.86) | ||
| Marital status (%) | Married | 3512 (64.39) | 659 (60.40) | 683 (62.60) | 716 (65.69) | 723 (66.27) | 731 (67.00) | 0.005 |
| Others | 1942 (35.61) | 432 (39.60) | 408 (37.40) | 374 (34.31) | 368 (33.73) | 360 (33.00) | ||
| Alcohol use (%) | Yes | 3350 (61.42) | 613 (56.19) | 637 (58.39) | 688 (63.12) | 682 (62.51) | 730 (66.91) | < 0.001 |
| Smoker status (%) | Yes | 1427 (26.16) | 421 (38.59) | 320 (29.33) | 251 (23.03) | 224 (20.53) | 211 (19.34) | < 0.001 |
| Hypertension (%) | 2821 (51.72) | 529 (48.49) | 609 (55.82) | 619 (56.79) | 541 (49.59) | 523 (47.94) | < 0.001 | |
| Diabetes (%) | 989 (18.13) | 154 (14.12) | 216 (19.80) | 206 (18.90) | 204 (18.70) | 209 (19.16) | 0.004 | |
| Cancer (%) | 622 (11.40) | 108 (9.90) | 123 (11.27) | 140 (12.84) | 130 (11.92) | 121 (11.09) | 0.278 | |
PHDI and risks of mortality and CVD
A statistically significant inverse association was observed between the PHDI and total mortality (HRsQ5 vs. Q1: 0.41; 95% CI: 0.32, 0.53). In multivariable-adjusted models, the PHDI remained inversely associated with total mortality, whether defined as quintiles (Table 2). The PHDI was also inversely associated with CVD risk (HRsQ5 vs. Q1: 0.72; 95% CI: 0.58, 0.91) (Table 2). When assessed continuously, each 1 SD increase in the PHDI was associated with a 25% lower risk of all-cause mortality (HRs: 0.75; 95% CI: 0.69, 0.81). Cause-specific mortality risks associated with each 1 SD increase in PHDI were as follows: cardiovascular mortality (HRs: 0.70; 95% CI: 0.61, 0.80), cancer mortality (HRs: 0.77; 95% CI: 0.67, 0.89), neurodegenerative mortality (HRs: 0.87; 95% CI: 0.43, 1.76), and infectious mortality (HRs: 0.68; 95% CI: 0.56, 0.83) (Figure S1) [see supplement information].
Table 2.
HRs (95% CIs) of quintiles of planetary health diet index (PHDI) for all-cause mortality and cardiovascular disease
| Quartile of PHDI | ||||||
|---|---|---|---|---|---|---|
| Quintile 1 | Quintile 2 | Quintile 3 | Quintile 4 | Quintile 5 | p for trend | |
| All-cause mortality | ||||||
| Cases | 228 | 185 | 176 | 161 | 95 | |
| Model 1 | 1.00 (ref.) | 0.69 (0.57, 0.84) | 0.57 (0.47, 0.70) | 0.56 (0.45, 0.68) | 0.35 (0.27, 0.44) | < 0.001 |
| Model 2 | 1.00 (ref.) | 0.71 (0.58, 0.86) | 0.59 (0.48, 0.72) | 0.58 (0.48, 0.72) | 0.36 (0.28, 0.46) | < 0.001 |
| Model 3 | 1.00 (ref.) | 0.74 (0.60, 0.90) | 0.63 (0.51, 0.77) | 0.64 (0.52, 0.79) | 0.41 (0.32, 0.53) | < 0.001 |
| Cardiovascular disease | ||||||
| Cases | 195 | 200 | 221 | 183 | 139 | |
| Model 1 | 1.00 (ref.) | 0.94 (0.77, 1.15) | 0.99 (0.82, 1.21) | 0.82 (0.67, 1.01) | 0.63 (0.51, 0.79) | < 0.001 |
| Model 2 | 1.00 (ref.) | 0.95 (0.78, 1.16) | 1.01 (0.83, 1.22) | 0.91 (0.74, 1.11) | 0.67 (0.53, 0.83) | < 0.001 |
| Model 3 | 1.00 (ref.) | 0.97 (0.80, 1.19) | 1.05 (0.86, 1.28) | 0.82 (0.67, 1.01) | 0.72 (0.58, 0.91) | 0.007 |
Boldface indicates that the result is statistically significant
For all-cause mortality, the association was most pronounced among participants aged ≥ 65 years (HRs per SD: 0.76; 95% CI: 0.70, 0.83), those who were physically active (HRs per SD: 0.70; 95% CI: 0.63, 0.78), and individuals with higher educational attainment (HRs per SD: 0.69; 95% CI: 0.61, 0.77). Similarly, in the CVD subgroup analyses, significant inverse associations with the PHDI were observed among participants aged 50–65 years (HRs per SD: 0.82; 95% CI: 0.73, 0.92), females (HRs per SD: 0.91; 95% CI: 0.83, 1.00), individuals with obesity (BMI ≥ 30 kg/m2) (HRs per SD: 0.86; 95% CI: 0.78, 0.96), and those who were physically active (HRs per SD: 0.91; 95% CI: 0.84, 0.98). And for all‐cause mortality, the association was weaker but still significant among participants with obesity (BMI ≥ 30 kg/m2) (HR per SD: 0.81; 95% CI: 0.71, 0.88). Additionally, in both all‐cause mortality and CVD analyses, higher PHDI scores were inversely associated with mortality among smokers and drinkers (Fig. 2, Table S4, and Table S5).
Fig. 2.
HRs (95% CIs) for PHDI per SD increase in relation to all-cause mortality and cardiovascular disease across subgroups. Models were all adjusted for the covariates including age, gender, race, education level, marital status, BMI, total intake energy, PA, alcohol use, smoker status and self-reported diagnosis high blood pressure, diabetes, and cancer diseases
Spline analysis of PHDI in relation to all-cause mortality and cancer
To further evaluate the potential non-linear associations between PHDI and mortality, we employed a four-knot restricted cubic spline model, and the food-group–specific analyses were conducted as exploratory without multiplicity correction. The analysis revealed a monotonic decrease in the spline curve, indicating an inverse relationship between PHDI and both all-cause mortality and CVD outcomes (Fig. 3). Furthermore, the analysis indicated that higher intakes of whole grains, whole fruits, vegetable, nuts, soy foods, and added unsaturated fats were associated with a lower risk of all-cause mortality (Figure S2) [see supplement information] and CVD outcomes (Figure S3) [see supplement information]. In contrast, greater consumption of tubers, red/processed meats, eggs, added saturated fats, and added sugars was associated with an increased risk of both all-cause mortality (Figure S2) and CVD outcomes (Figure S3). Overall, these findings highlight the protective effects of higher adherence to the PHDI on both total mortality and CVD, with consistent benefits observed across diverse subgroups.
Fig. 3.
Restricted cubic splines illustrating the non-linear associations between HRs (95% CIs) for all-cause mortality (A) and PHDI, and HRs (95% CIs) for cardiovascular disease (B) and PHDI. Models were all adjusted for the covariates including age, gender, race, education level, marital status, BMI, total intake energy, PA, alcohol use, smoker status and self-reported diagnosis high blood pressure, diabetes, and cancer diseases
Sensitivity analyses
To ensure the robustness of our findings, we conducted multiple sensitivity analyses. Decile-based adherence analyses (Supplementary Table S6) and competing risk models (Supplementary Table S7) yielded results consistent with the primary findings. Stratified analyses further confirmed the stability of the associations across sex and age groups, with similar patterns observed for middle-aged and older adults (Supplementary Table S8) as well as for males and females (Supplementary Table S9).
Discussion
In this prospective cohort of U.S. older adults from the Health and Retirement Study, higher PHDI adherence was associated with lower risks of all-cause mortality and CVD. These findings are consistent with prior research indicating that adherence to the EAT-Lancet dietary recommendations is associated with lower mortality risk and improved health outcomes [4]. our study further extends the existing literature by revealing that higher adherence to the PHDI is associated with the greatest reductions in hazard ratios for CVD, thereby underscoring the critical role of dietary patterns in mitigating these health risks.
Several previous large-scale prospective cohort studies, including those conducted in the United States [12], Sweden [19], Denmark [20], and the United Kingdom [21], have reported significant inverse associations between adherence to planetary health diets and health risks. Prior analyses of large U.S. cohorts with long-term follow-up have shown that participants with the highest PHDI adherence experienced lower risks of all-cause mortality, with similar reductions observed for cardiovascular, cancer, respiratory, and neurodegenerative mortality [5]. Similar patterns have been observed in European cohorts: higher adherence to the EAT-Lancet diet in EPIC-Oxford was linked to lower risks of all-cause mortality, ischemic heart disease, and diabetes, and the Malmö Diet and Cancer cohort likewise showed lower risks of all-cause, cancer, and cardiovascular mortality with greater adherence. These findings align with our results in U.S. older adults, reinforcing the generalizability of the inverse association across populations and settings [19].
Our study found that higher adherence to the PHDI was most strongly associated with reductions in CVD mortality and infectious disease mortality. Given that cardiovascular diseases remain the leading cause of death globally, these findings reinforce the relevance of plant-forward, nutrient-rich dietary patterns in lowering CVD-related mortality risk [22]. In our analysis, the trend across PHDI quintiles for CVD outcomes appeared less linear than for all-cause mortality. Although the underlying reasons are unclear, factors such as differences in disease management behaviors or medical treatment among older adults may contribute to this pattern; however, these possibilities warrant further investigation. The observed associations may be partly explained by biological mechanisms proposed in prior studies, including improvements in lipid metabolism, reductions in systemic inflammation, lower blood pressure, and enhanced insulin sensitivity [23–25]. These mechanisms remain hypothetical in the context of our study, as we did not directly measure physiological intermediates, but they provide plausible pathways through which plant-forward dietary patterns could contribute to lower CVD and infectious disease mortality risks. Our study also shows an inverse association between PHDI and infectious disease mortality, an outcome seldom explored in dietary research [26]. Although the mechanisms cannot be inferred from our data, prior evidence suggests that plant-forward diets rich in essential nutrients may support immune function and reduce vulnerability to infections [27–30]. These potential pathways—including immune and inflammatory regulation—remain speculative, as our study did not assess relevant biomarkers or inflammatory indices such as the Dietary Inflammatory Index (DII). Future research is needed to clarify the biological mechanisms underlying this association [31–33].
A key strength of this study is the relatively large sample size and long follow-up period of the HRS cohort, which offered sufficient data to examine the associations between PHDI and all-cause mortality as well as CVD outcomes with reasonable statistical precision. The use of a continuous scoring system for the PHDI allowed for a more nuanced characterization of dietary adherence, which may have contributed to a more detailed assessment of its associations with health outcomes. In addition, the inclusion of a broad set of covariates—covering sociodemographic characteristics, lifestyle factors, selected health conditions, and dietary components—helped reduce, although not eliminate, the potential influence of confounding. These analytic approaches support the credibility of the observed associations, while acknowledging that limitations such as self-reported diet and disease history, unverified CVD outcomes, incomplete cause-of-death coding, and variability in baseline timing may still introduce uncertainty [34].
Despite these strengths, certain limitations must be acknowledged. First, CVD outcome and dietary intake were self-reported, which are subject to measurement errors and recall bias. Second, while we adjusted for multiple confounders, residual confounding from unmeasured factors such as genetic predispositions, stress, and sleep quality cannot be entirely ruled out. Third, our study focused on an older U.S. population, which may limit the generalizability of our findings to younger or more diverse populations. Future research should aim to replicate these findings in other demographic groups and explore potential modifications in dietary recommendations based on individual health profiles.
Conclusion
Our findings show that higher PHDI scores were associated with lower risks of all-cause mortality and CVD outcomes in an older U.S. population. These results describe a pattern in which closer adherence to the PHDI coincides with more favorable health outcomes. Further work is needed to examine these associations in different populations and to improve approaches for assessing adherence to planetary health–aligned dietary patterns.
Supplementary Information
Acknowledgements
The authors thank the HRS project members for their substantial work on data collection and patient participation.
Abbreviations
- PHDI
The Planetary Health Diet Index
- CVD
Cardiovascular disease
- HRS
The Health and Retirement Study
- HRs
Hazard ratios
- CIs
Confidence intervals
- HCNS
Health Care and Nutrition Study
- FFQ
Food frequency questionnaire
- ICD
The International Classification of Diseases
- BMI
Body mass index
- PA
Physical activity
- SD
Standard deviation
- ANOVA
Analysis of variance
Authors’ contributions
Zhaoting Bu: Methodology, Data analysis, Writing- Original draft preparation and Writing- Editing. Zhiyong Li: Data analysis, Writing- Editing and Writing- Reviewing. Xiaoyue Liu: Methodology, Writing- Original draft preparation and Writing- Editing. Chenan Liu: Writing- Original draft preparation and Writing- Reviewing. Sanyu Ge: Writing- Original draft preparation. Bing Yin: Writing- Reviewing and Supervision. Yue Chen: Software. Xin Zheng: Writing- Reviewing. Li Deng: Writing- editing and Software. Hong Zhao: Writing- editing. Yi Li: Writing- editing. Hanping Shi: Conceptualization, Writing- Reviewing. All authors contributed to the article and approved the submitted version.
Funding
This work was supported by the National Key Research and Development Program (2022YFC2009600, 2022YFC2009601) and Laboratory for Clinical Medicine, Capital Medical University (2023-SYJCLC02).
Data availability
The datasets generated during and/or analysed during the current study are available from the corresponding author on reasonable request.
Declarations
Ethics approval and consent to participate
Data from the HRS are freely available to the public and can be accessed on their website, thus obviating the need for consent from the medical ethics committee.
Competing interests
The authors declare no competing interests.
Footnotes
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Zhaoting Bu, Zhiyong Li and Xiaoyue Liu contributed equally to this work.
References
- 1.Springmann M, Wiebe K, Mason-D’Croz D, et al. Health and nutritional aspects of sustainable diet strategies and their association with environmental impacts: a global modelling analysis with country-level detail. Lancet Planet Health. 2018;2:e451–61. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Carson JAS, Lichtenstein AH, Anderson CAM, et al. Dietary cholesterol and cardiovascular risk: A science advisory from the American heart association. Circulation. 2020;141:e39–53. [DOI] [PubMed] [Google Scholar]
- 3.Garrison-Desany HM, Ladd-Acosta C, Hong X, et al. Addressing the smoking-hypertension paradox in pregnancy: insight from a multiethnic US birth cohort. Precis Nutr. 2023;2:e00035. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Li J, Lee DH, Hu J, et al. Dietary inflammatory potential and risk of cardiovascular disease among men and women in the U.S. J Am Coll Cardiol. 2020;76:2181–93. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Makker K, Wang X. Early life origins of cardio-metabolic outcomes in Boston birth cohort: review of findings and future directions. Precis Nutr. 2023;2(3):e00050. [PMC free article] [PubMed]
- 6.Baden MY, Liu G, Satija A, et al. Changes in Plant-Based diet quality and total and Cause-Specific mortality. Circulation. 2019;140:979–91. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Hooper L, Martin N, Jimoh OF, et al. Reduction in saturated fat intake for cardiovascular disease. Cochrane Database Syst Rev. 2020;5:Cd011737. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Grant MG, Pratt C, Wong RP, et al. Implementing the National Heart, Lung, and Blood Institute's Strategic Vision in the Division of Cardiovascular Sciences-2022 Update. Circ Res. 2022;131(8):713–24. [DOI] [PMC free article] [PubMed]
- 9.Chiuve SE, Fung TT, Rimm EB, et al. Alternative dietary indices both strongly predict risk of chronic disease. J Nutr. 2012;142(6):1009–18. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Hou W, Zhang M, Ji Y, et al. A prospective birth cohort study of maternal prenatal cigarette smoking assessed by self-report and biomarkers on childhood risk of overweight or obesity. Precis Nutr. 2022;1:e00017. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Dalile B, Kim C, Challinor A, et al. The EAT-Lancet reference diet and cognitive function across the life course. Lancet Planet Health. 2022;6:e749–59. [DOI] [PubMed] [Google Scholar]
- 12.Hirvonen K, Bai Y, Headey D, Masters WA. Affordability of the EAT-Lancet reference diet: a global analysis. Lancet Glob Health. 2020;8:e59–66. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.van Selm B, Frehner A, de Boer IJM, et al. Circularity in animal production requires a change in the EAT-Lancet diet in Europe. Nat Food. 2022;3:66–73. [DOI] [PubMed] [Google Scholar]
- 14.Venegas Hargous C, Orellana L, Strugnell C, et al. Adapting the planetary health diet index for children and adolescents. Int J Behav Nutr Phys Act. 2023;20:146. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Sawicki CM, Ramesh G, Bui L, et al. Planetary health diet and cardiovascular disease: results from three large prospective cohort studies in the USA. Lancet Planet Health. 2024;8:e666–74. [DOI] [PubMed] [Google Scholar]
- 16.Willett W, Rockström J, Loken B, et al. Food in the anthropocene: the EAT-Lancet commission on healthy diets from sustainable food systems. Lancet. 2019;393:447–92. [DOI] [PubMed] [Google Scholar]
- 17.Carcelén-Fraile MDC, Déniz-Ramírez NDP, Sabina-Campos J, et al. Exercise and nutrition in the mental health of the older adult population: a randomized controlled clinical trial. Nutrients. 2024;16(11):1741. [DOI] [PMC free article] [PubMed]
- 18.Zhan JJ, Hodge RA, Dunlop AL, et al. Dietaryindex: a user-friendly and versatile R package for standardizing dietary pattern analysis in epidemiological and clinical studies. Am J Clin Nutr. 2024;120:1165–74. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Zhang S, Stubbendorff A, Olsson K, et al. Adherence to the EAT-Lancet diet, genetic susceptibility, and risk of type 2 diabetes in Swedish adults. Metabolism. 2023;141:155401. [DOI] [PubMed] [Google Scholar]
- 20.Ibsen DB, Christiansen AH, Olsen A, et al. Adherence to the EAT-Lancet diet and risk of stroke and stroke subtypes: A cohort study. Stroke. 2022;53:154–63. [DOI] [PubMed] [Google Scholar]
- 21.Lu X, Wu L, Shao L, et al. Adherence to the EAT-Lancet diet and incident depression and anxiety. Nat Commun. 2024;15:5599. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Chen Y, Zheng X, Wang Y, et al. Association between dietary quality and accelerated aging: a cross-sectional study of two cohorts. Food Funct. 2024;15:7837–48. [DOI] [PubMed] [Google Scholar]
- 23.Koch CA, Kjeldsen EW, Frikke-Schmidt R. Vegetarian or vegan diets and blood lipids: a meta-analysis of randomized trials. Eur Heart J. 2023;44:2609–22. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.de Luis Roman DA, Primo D, O IZ, et al. Adiponectin gene variant rs3774261, effects on lipid profile and adiponectin levels after a high polyunsaturated fat hypocaloric diet with mediterranean pattern. Nutrients. 2021;13(6):1811. [DOI] [PMC free article] [PubMed]
- 25.Liu C, Liu T, Zhang Q, et al. New-Onset age of nonalcoholic fatty liver disease and cancer risk. JAMA Netw Open. 2023;6:e2335511. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Zheng X, Chen Y, Lin SQ, et al. The relationship between different fatty acids intake and the depressive symptoms: A population-based study. J Affect Disord. 2024;357:68–76. [DOI] [PubMed] [Google Scholar]
- 27.Barrea L, Muscogiuri G, Frias-Toral E, et al. Nutrition and immune system: from the mediterranean diet to dietary supplementary through the microbiota. Crit Rev Food Sci Nutr. 2021;61:3066–90. [DOI] [PubMed] [Google Scholar]
- 28.Filippini T, Fairweather-Tait S, Vinceti M. Selenium and immune function: a systematic review and meta-analysis of experimental human studies. Am J Clin Nutr. 2023;117:93–110. [DOI] [PubMed] [Google Scholar]
- 29.Fantacone ML, Lowry MB, Uesugi SL, et al. The effect of a multivitamin and mineral supplement on immune function in healthy older adults: a double-blind, randomized, controlled trial. Nutrients. 2020;12(8):2447. [DOI] [PMC free article] [PubMed]
- 30.Shiekh PA, Singh A, Kumar A. Exosome laden oxygen releasing antioxidant and antibacterial cryogel wound dressing OxOBand alleviate diabetic and infectious wound healing. Biomaterials. 2020;249:120020. [DOI] [PubMed] [Google Scholar]
- 31.Song M, Garrett WS, Chan AT. Nutrients, foods, and colorectal cancer prevention. Gastroenterology. 2015;148:1244–e12601216. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Shah RV, Steffen LM, Nayor M, et al. Dietary metabolic signatures and cardiometabolic risk. Eur Heart J. 2023;44:557–69. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Zheng X, Ge YZ, Ruan GT, et al. Association between the dietary inflammatory index and All-Cause mortality in adults with obesity. Ann Nutr Metab. 2023;79(5):434–47. [DOI] [PubMed] [Google Scholar]
- 34.Chen Y, Zheng X, Liu C, et al. Anthropometrics and cancer prognosis: a multicenter cohort study. Am J Clin Nutr. 2024;120:47–55. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The datasets generated during and/or analysed during the current study are available from the corresponding author on reasonable request.