Abstract
Background
To improve both human health and the health of our planet, the EAT-Lancet Commission proposed the planetary health diet (PHD).
Objectives
We aimed to evaluate associations of PHD with all-cause, cardiovascular disease (CVD), and cancer-specific mortality among United States Black females.
Methods
The Black Women’s Health Study is a prospective study of self-identified United States Black females. In 2001, 33,824 participants free of cancer and CVD completed a validated food frequency questionnaire. PHD Index (PHDI) was calculated based on reported consumption of 15 food groups, such as whole grains, nonstarchy vegetables, legumes, soy foods, added fat and trans fat, and red/processed meats. Deaths were identified through linkage to the National Death Index. Cox proportional hazards regression, stratified by age and adjusted for smoking status, body mass index, and other CVD risk factors, was used to calculate hazard ratios (HRs) for quintiles of PHDI in relation to all-cause, CVD-, and cancer-specific mortality.
Results
During 18 years of follow-up, we identified 3537 deaths, including 779 from CVD and 1625 from cancer. Females in the quintile representing the highest adherence to PHD were estimated to have an 18% reduction in risk of all-cause mortality [HR = 0.82, 95% confidence interval (CI): 0.71, 0.94] and 26% reduction in CVD-specific mortality (HR = 0.74, 95% CI: 0.55, 0.98), compared with those in the lowest quintile, with similar reductions observed for quintiles 2, 3, and 4. Among individuals under age 55, there was a significant trend of lower CVD mortality risk with a higher level of adherence to PHD (Ptrend = 0.004), and the HR for the highest compared with the lowest quintile was 0.43 (95% CI: 0.21, 0.87). PHDI was not associated with cancer-specific mortality.
Conclusions
Adherence to a diet that has been shown to benefit the planet was associated with a lower risk of mortality among Black females, primarily driven by a reduction in CVD-specific mortality risk.
Keywords: diet, mortality, cardiovascular disease, Black females, planetary health
Introduction
Unhealthy and unsustainably produced foods threaten not only human health but also planetary health by influencing climate change, freshwater use, nitrogen and phosphorous flows, biodiversity loss, and land-system change. Planetary health diet (PHD), a reference diet proposed by the EAT-Lancet Commission, is a diet that contributes to both human health and environmental sustainability. PHD consists of diverse plant-based foods, low amounts of animal source foods, unsaturated rather than saturated fats, as well as small amounts of refined grains, highly processed foods, and added sugar within an appropriate caloric intake with recommended intake range provided. Adherence to PHD can be measured by the Planetary Health Diet Index (PHDI) based on intake of 15 food groups, including whole grains, nonstarchy vegetables, starchy vegetables, fruits, nuts, legumes, soy foods, saturated added fat and trans fat, unsaturated added fat, added sugars, eggs, fish and shellfish, dairy foods, red/processed meats, and poultry [1,2].
In the United States, Black individuals experience a disproportionately high mortality rate, regardless of cause of death [3,4]. The higher rates of cardiovascular disease (CVD) experienced by Black individuals are a factor in the higher all-cause mortality [5]. Although smoking, obesity, and physical inactivity remain the top causes of CVD-specific mortality, some proportion of cardiovascular mortality may be attributed to suboptimal dietary habits [6]. High intake of macronutrients such as unsaturated fat, plant protein [7], and dietary fibers, food groups such as nuts [8], whole grain [9], fruits, vegetables, and soy foods [10], as well as dietary patterns such as Mediterranean diet [11] and plant-based diet [12] have all been associated with lower risk of CVD-specific and all-cause mortality, whereas high intake of sodium, added sugar, saturated fat, trans fat, and red/processed meat have been associated with increased risk of CVD and CVD-specific mortality [[13], [14], [15]].
Black individuals also experience a disproportionately high cancer-specific mortality rate [16]. Diet has been associated with risk of several cancers, including colorectal, liver, endometrial, and possibly with breast and lung cancer as well. There is strong epidemiological and mechanistic evidence for the beneficial effects of food groups such as whole grains, fruits and vegetables, dairy products, and fish as well as the adverse effects of food groups such as red/processed meat and added sugar [17].
Prior studies suggest that adherence to the PHD may reduce mortality [1,[18], [19], [20], [21]]. None of those studies, however, examined the association among Black individuals. Thus, we aimed to extend the evidence on possible benefits of the PHD by evaluating associations with all-cause, CVD-specific, and cancer-specific mortality in a large prospective cohort of Black females.
Methods
Study population
The Black Women’s Health Study (BWHS) is a longitudinal cohort of 59,000 self-identified Black females from across the United States who were aged 21–69 at study enrollment in 1995 [22]. Participants were asked to report race by checking all that apply from the following options: Black; White; Asian or Pacific Islander; and American Indian or Alaskan Native. Females who checked “Black” are included in the BWHS, regardless of whether or not they checked any other categories. Information on demographics, reproductive history, medical history, and behavioral factors such as smoking and alcohol consumption is collected via self-administered questionnaires every 2 years. Food frequency questionnaires (FFQs) were administered in 1995 and 2001. Deaths and causes of death are identified through linkage to the National Death Index (see next). Follow-up of the baseline cohort is complete for >85% of potential person-years.
For the present analysis, individuals who had already been diagnosed with CVD or cancer prior to dietary assessment in 2001 (n = 1608) were considered ineligible. The analysis was based on dietary intake collected from the 2001 FFQ because the 1995 FFQ did not ask for consumption of soy foods, a food group in the PHD. We excluded individuals who did not complete the 2001 FFQ (n = 17,832), had implausible total energy intake values (<500 or >3800 kcal/d) (n = 2688), or left 10 or more FFQ items blank (n = 2919). We also excluded individuals who had missing data on the key food items needed to calculate the PHDI (n = 129). The final analytic population was 33,824 (Figure 1). The study protocol was approved by the Boston University Medical Campus Institutional Review Board. Study participants provided informed consent by completing questionnaires and returning them to the BWHS.
FIGURE 1.
Flow chart of participants included in this analysis. CVD, cardiovascular disease; FFQ, food frequency questionnaire; PHDI, Planetary Health Diet Index.
Exposure (dietary assessment)
The BWHS FFQ is a slightly modified version of Block short-form FFQ [23], with several food items added after participants were asked to write in foods regularly eaten that did not appear on the original Block short-form, such as cantaloupe, mustard greens, liver, and cornbread. The BWHS 2001 FFQ had 85 line items. On the FFQ, participants were asked to report how frequently they consumed each food item in the last year (using multiple choice questions with 9 choices ranging from never or <1 serving/mo to 2+ servings/d) and approximate portion size consumed. A medium portion size was defined for each line item and 4 options were provided for size: small ( half of the medium portion), medium, large (1.5 times the medium portion size), and super (2 times the medium portion size). FFQ responses (and nutrient estimates) were validated with data from 3 24-h diet records [24].
The maximum score for each food group (10) was based on the greatest beneficial health effect of that food group, and the minimum score (0) was based on the least health benefit. Thus, for foods groups like soy foods and nonstarchy vegetables, a high consumption per day would receive a high score, whereas for food groups like starchy vegetables and red/processed meat, a high consumption would receive a score close to zero. The scores for each food group were then summed, with a weight of 0.5 for soy foods and legumes and a weight of 1 for all other food groups, to calculate a total score, ranging from 0 (nonadherence) to 140 (total adherence) (Table 1). Higher PHDI reflects higher adherence to PHD. Detailed descriptions on scoring methods can be found elsewhere [1].
TABLE 1.
Food groups included in the PHD, with PHDI scoring criteria and distribution of food group intake PHDI in the study population.
| Food groups included in the PHD | Examples of food items in each group | Intake (g/d) cutoffs for receiving minimum or maximum PHDI score for individual food groups |
Median food group consumption in the BWHS in g/d (min, max) | ||
|---|---|---|---|---|---|
| g/d cutoff for a minimum score of 0 | g/d cutoff for a maximum score of 10 | Weight applied | |||
| Whole grain | High fiber, bran, or granola cereals, shredded wheat, dark breads such as wheat, rye, and pumpernickel | 0 | ≥75 | 1 | 14.6 (0.0, 264.0) |
| Starchy vegetables | Potatoes (e.g., chips and fries) and cassava | ≥200 | ≤50 | 1 | 19.2 (0.0, 421.5) |
| Nonstarchy vegetables | Vegetables not including potato or other starchy vegetables (e.g., tomatoes, broccoli, collard greens, cole slaw, and carrots) | 0 | ≥300 | 1 | 121.9 (0.0, 2336.4) |
| Whole fruits | Apples, pears, bananas, cantaloupe, canned fruit, orange, and grapefruit | 0 | ≥200 | 1 | 128.0 (0.0, 2820.0) |
| Dairy foods | Milk, cheese, yogurt, and ice cream | ≥1000 | ≤250 | 1 | 61.4 (0.0, 1353.1) |
| Red/processed meat | Beef, lamb, pork, processed meat such as bacon, sausage, ham, and lunch meat | ≥100 | ≤14 | 1 | 34.8 (0.0, 1056.0) |
| Poultry | Chicken, duck, goose, and ostrich | ≥100 | ≤29 | 1 | 47.3 (0.0, 976.0) |
| Eggs | Eggs | ≥120 | ≤13 | 1 | 8.8 (0.0, 396.0) |
| Fish and shellfish | Fish (e.g., dark meat fish and tuna fish) and shellfish (e.g., shrimp, clams, and crabs) | 0 | ≥28 | 1 | 32.0 (0.0, 1003.5) |
| Nuts | Peanuts, peanut butter, walnuts, almonds, hazelnuts, pecan, cashews, and pistachios | 0 | ≥50 | 1 | 1.1 (0.0, 124.0) |
| Nonsoy legumes | Dry beans (e.g., pinto, kidney, black beans), lentils, and peas (e.g., black-eyed peas) | 0 | ≥100 | 0.5 | 14.2 (0.0, 636.0) |
| Soy foods | Soybeans, soy milk, tofu, and soy burgers | 0 | ≥50 | 0.5 | 0.0 (0.0, 19.2) |
| Added fat—unsaturated oils | Olive, soybean, rapeseed, sunflower, and peanut oil | ≤3.5% of TEI | ≥21% of TEI | 1 | 22.0 (4.1, 50.2) |
| Added fat—saturated oils and trans fat | Palm oil, coconut oil, dairy fat (butter), margarine, lard, and tallow | ≥10% of TEI | 0% of TEI | 1 | 11.4 (1.8, 32.4) |
| Added sugar | Sugar from fruit juice, soft drinks, jam, ice cream, cakes, and sugar eaten separately | ≥25% of TEI | ≤5% of TEI | 1 | 10.4 (0.3, 80.0) |
Abbreviations: BWHS, Black Women’s Health Study; kcal/d, kcal/day; max, maximum; min, minimum; no., number; PHD, planetary health diet; PHDI, Planetary Health Diet Index; TEI, total energy intake.
Outcome ascertainment
Deaths from the start of follow-up in 2003 through the end of follow-up in 2021 were identified from the National Death Index, reports from next of kin, and the United States Postal Service. Cause-specific mortality was obtained from either the National Death Index or state-issued death certificate and was classified based on International Classification of Disease Code of I00–I09, I11, I13, and I20–I51 for CVD-specific mortality and C00–C97 for cancer-specific mortality. We identified a total of 3537 deaths, including 779 CVD deaths and 1625 cancer deaths.
Covariates
Covariates were selected a priori based on previous evidence on factors associated with diet and mortality. The selected covariates included age (years), total energy intake (kcal/d), years of education (≤12, 13–15, 16, ≥17 y), BMI (in kg/m2) (<25, 25–<30, 30–<35, ≥35), smoking status (never smoker, past smoker, current smoker <15 cigarettes/d, current smoker ≥15 cigarettes/d), alcohol consumption (<1, 1–6, ≥7 drinks/d), physical activity (<7, 7–<20, ≥20 metabolic equivalents hours/wk), neighborhood socioeconomic status (SES) in quintiles, marital status (married/livingtogether, separated/divorced/widowed, or never married), and region of residence (Northeast, South, Midwest, or West). All covariates were taken from the 2001 questionnaire except for marital status (from 1999 questionnaire) and education level (from 1995 questionnaire). BMI was calculated as weight in kilograms (kg) divided by the square of height in meters (m). SES was assessed by linking geocoded residential addresses to 6 variables from the United States Census that had been selected through principal components analysis. Detailed descriptions of the scoring method are available elsewhere [25].
Statistical approach
Cox proportional hazard models, stratified by age and questionnaire cycle, were used to estimate hazard ratios (HRs) and 95% confidence intervals (CIs) for the association between PHDI (quintiles) and mortality, all-cause and separately for CVD- and cancer-specific mortality. To avoid potential bias because of reverse causation (individuals who are close to death may change their dietary patterns), we began follow-up for mortality in 2003, 2 years after the completion of the FFQ. Multivariable models were adjusted for total energy intake, education, BMI, smoking status, alcohol consumption, physical activity, neighborhood SES, marital status, and region of residence. Missing information on covariates was low (generally <2%) and included as a separate indicator category in models.
To evaluate potential effect measure modification of the association between PHDI and mortality, additional analyses were stratified on age at risk (<55, ≥55 y), smoking status (never, past, and current), neighborhood SES (below/above the mean score), educational level (≤12, 13–15, ≥16 y), and BMI (<25, 25–<30, 30–<35, ≥35). All analyses were performed using SAS 9.4.
Results
In the current study population, the mean PHDI was 81.42, with a range from 36.80 to 126.58. The distributions of food group intakes and food group PHDIs can be found in Table 1. At baseline, participants with higher PHDI had an older age, a lower BMI, were less likely to smoke, had higher attained education, more physical activity, and lived in neighborhoods characterized as having a higher SES. To investigate the gradient of age in the quintiles of PHDI, we examined the food consumption by individual food groups in the study population by median baseline age. We found that the high adherence to PHD among older participants was driven by higher consumption of plant-based foods and lower consumption of meat, consistent with previous BWHS findings [26]. Alcohol consumption was distributed similarly across quintiles of PHDI (Table 2).
TABLE 2.
Age-standardized characteristics of Black Women’s Health Study participants by quintiles of PHDI, 2001.
| Planetary Health Diet Index (PHDI) |
|||||
|---|---|---|---|---|---|
| Quintile 1 PHDI <71.2 (N = 6780) |
Quintile 2 PHDI 71.2–<78.2 (N = 6784) |
Quintile 3 PHDI 78.2–<84.4 (N = 6742) |
Quintile 4 PHDI 84.4–<91.4 (N = 6741) |
Quintile 5 PHDI ≥91.4 (N = 6777) |
|
| Age | 40.5 (8.8) | 43.0 (9.6) | 44.4 (10.1) | 46.5 (10.6) | 48.7 (10.9) |
| Region of residence, % (n) | |||||
| Northeast | 22 (1481) | 23 (1581) | 25 (1675) | 27 (1795) | 31 (2082) |
| South | 34 (2315) | 34 (2300) | 34 (2287) | 33 (2198) | 32 (2173) |
| Midwest | 27 (1815) | 25 (1680) | 22 (1508) | 22 (1461) | 18 (1219) |
| West | 17 (1162) | 18 (1208) | 19 (1257) | 19 (1270) | 19 (1284) |
| Marital status, % (n) | |||||
| Married/living together | 41 (2753) | 43 (2905) | 44 (2968) | 45 (3053) | 43 (2893) |
| Separated/divorced/widowed | 28 (1912) | 29 (1946) | 27 (1795) | 26 (1743) | 27 (1809) |
| Single | 29 (1963) | 26 (1764) | 27 (1832) | 27 (1818) | 29 (1951) |
| BMI, % (n) | |||||
| <25 kg/m2 | 22 (1519) | 26 (1736) | 26 (1747) | 29 (1968) | 35 (2372) |
| 25–30 kg/m2 | 29 (1966) | 31 (2119) | 34 (2263) | 34 (2291) | 35 (2357) |
| 30–35 kg/m2 | 22 (1512) | 21 (1415) | 21 (1431) | 20 (1348) | 18 (1190) |
| ≥35 kg/m2 | 25 (1716) | 21 (1431) | 18 (1235) | 16 (1049) | 12 (793) |
| Smoking status, % (n) | |||||
| Never | 61 (4159) | 64 (4355) | 65 (4354) | 67 (4505) | 68 (4587) |
| Past | 20 (1322) | 21 (1416) | 23 (1532) | 23 (1539) | 24 (1644) |
| Current <15 cigarettes/d | 13 (902) | 11 (744) | 10 (670) | 8 (561) | 7 (469) |
| Current ≥15 cigarettes/d | 6 (397) | 4 (270) | 3 (187) | 2 (136) | 1 (77) |
| Alcohol consumption, % (n) | |||||
| <1 drink/wk | 69 (4686) | 69 (4658) | 68 (4600) | 68 (4594) | 71 (4783) |
| 1–6 drinks/wk | 24 (1622) | 25 (1708) | 25 (1708) | 26 (1749) | 24 (1646) |
| ≥7 drinks/wk | 6 (429) | 6 (394) | 6 (392) | 5 (359) | 5 (310) |
| Education level, % (n) | |||||
| ≤12 y | 19 (1312) | 16 (1065) | 14 (913) | 12 (793) | 9 (615) |
| 13–15 y | 38 (2580) | 37 (2494) | 37 (2480) | 34 (2295) | 30 (2054) |
| 16 y | 24 (1594) | 25 (1702) | 25 (1701) | 26 (1730) | 27 (1797) |
| ≥17 y | 19 (1261) | 22 (1488) | 24 (1607) | 28 (1871) | 33 (2260) |
| Physical activity, % (n) | |||||
| < 7 METs hour per week | 49 (3294) | 38 (2589) | 31 (2087) | 25 (1655) | 17 (1184) |
| 7-20 METs hour per week | 29 (1983) | 34 (2332) | 34 (2324) | 33 (2212) | 30 (2011) |
| ≥20 METs hour per week | 21 (1409) | 26 (1773) | 33 (2221) | 41 (2756) | 51 (3465) |
| Neighborhood SES quintiles, % (n) | |||||
| Quintile 1 (lowest) | 21 (1304) | 19 (1192) | 18 (1109) | 15 (960) | 15 (918) |
| Quintile 2 | 22 (1334) | 19 (1225) | 20 (1230) | 19 (1177) | 18 (1111) |
| Quintile 3 | 21 (1299) | 20 (1261) | 20 (1250) | 20 (1231) | 19 (1172) |
| Quintile 4 | 19 (1163) | 22 (1367) | 21 (1309) | 22 (1381) | 22 (1386) |
| Quintile 5 (highest) | 18 (1098) | 20 (1242) | 21 (1318) | 24 (1479) | 26 (1618) |
| Total energy intake (kcal), mean (SD) | 1607 (737) | 1446 (670) | 1441 (657) | 1443 (638) | 1541 (616) |
Abbreviations: MET, metabolic equivalents; SES, socioeconomic status.
Values are means (SD) for continuous variables; percentages and ns for categorical variables and (except for age) are standardized to the age distribution of the study population. Values of categorical variables may not sum to 100% because of rounding.
Adherence to PHD was associated with a significantly lower risk of all-cause mortality and CVD-specific mortality but was not associated with cancer-specific mortality (Table 3). In multivariable analyses, the HR for females in the highest quintile of PHDI relative to the lowest quintile was 0.82 (95% CI: 0.71, 0.94) for all-cause mortality (Ptrend = 0.026) and 0.74 (95% CI: 0.55, 0.98) for CVD-specific mortality (Ptrend = 0.053). We also reported the HR for higher PHDI (quintiles 2–5 combined) compared with the lower PHDI (quintile 1). HRs for that comparison were 0.82 (95% CI: 0.74, 0.92) for all-cause mortality, 0.72 (95% CI: 0.57, 0.90) for CVD mortality, and 0.87 (95% CI: 0.74, 1.03) for cancer mortality.
TABLE 3.
Hazard ratios (95% confidence interval) for PHDI in relation to all-cause and cause-specific mortality.
| PHDI quintile1 | All-cause mortality |
CVD-specific mortality |
Cancer-specific mortality |
|||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| N death | Person-years | HR2 (95% CI) | HR3 (95% CI) | N death | Person-years | HR2 (95% CI) | HR3 (95% CI) | N death | Person-years | HR2 (95% CI) | HR3 (95% CI) | |
| Quintile 1 | 641 | 102,740 | 1.00 (ref.) | 1.00 (ref.) | 150 | 106,464 | 1.00 (ref.) | 1.00 (ref.) | 282 | 105,530 | 1.00 (ref.) | 1.00 (ref.) |
| Quintile 2 | 591 | 103,009 | 0.76 (0.67, 0.87) | 0.79 (0.69, 0.91) | 124 | 107,003 | 0.69 (0.52, 0.91) | 0.75 (0.56, 1.00) | 270 | 105,989 | 0.79 (0.64, 0.96) | 0.80 (0.65, 0.98) |
| Quintile 3 | 704 | 102,520 | 0.80 (0.71, 0.91) | 0.90 (0.79, 1.04) | 125 | 107,253 | 0.59 (0.44, 0.78) | 0.64 (0.48, 0.86) | 338 | 105,524 | 0.93 (0.77, 1.13) | 1.02 (0.83, 1.24) |
| Quintile 4 | 769 | 103,132 | 0.67 (0.59, 0.77) | 0.78 (0.68, 0.90) | 178 | 107,841 | 0.62 (0.47, 0.81) | 0.74 (0.55, 0.98) | 345 | 106,579 | 0.71 (0.58, 0.86) | 0.76 (0.62, 0.93) |
| Quintile 5 | 832 | 103,040 | 0.64 (0.56, 0.73) | 0.82 (0.71, 0.94) | 202 | 109,194 | 0.55 (0.42, 0.71) | 0.74 (0.55, 0.98) | 390 | 106,463 | 0.79 (0.65, 0.95) | 0.91 (0.74, 1.12) |
| Ptrend = 0.026 | Ptrend = 0.053 | Ptrend = 0.727 | ||||||||||
Abbreviations: CI, confidence interval; CVD, cardiovascular disease; HR, hazard ratio; PHDI, Planetary Health Diet Index.
The PHDI was 36.8–<71.2 for quintile 1, 71.2–<78.2 for quintile 2, 78.2–<84.4 for quintile 3, 84.4–<91.4 for quintile 4, and 91.4–126.6 for quintile 5.
HR adjusted for age and energy only.
HR adjusted for age, energy, education, BMI, smoking status, alcohol consumption, and physical activity.
In stratified analyses, HRs for associations of each of quintiles 2–5 compared with quintile 1 of PHDI with CVD mortality decreased with increasing quintile to 0.43 (95% CI: 0.21, 0.87) for quintile 5 (Ptrend = 0.004) among females <55 years of age, whereas there was no trend among females 55 and older. Similar trends for decreased mortality with increasing quintile were observed among females in other low-CVD-risk groups, including among never smokers (Ptrend = 0.004), individuals with education of 16+ years (Ptrend = 0.004), high neighborhood SES (Ptrend = 0.017) and BMI <25 (Ptrend = 0.055). There were marked inverse associations observed among never smokers, females with BMI < 30, females with a college education, and females who lived in higher SES neighborhoods (Table 4).
TABLE 4.
PHDI in relation to CVD-specific mortality, stratified by age, neighborhood SES, smoking status, education, and BMI.
| PHDI Quintiles1 | Age |
Smoking status |
||||||
|---|---|---|---|---|---|---|---|---|
| < 55 y |
≥ 55 y |
Never smokers |
Smokers |
|||||
| N2 | HR3 (95% CI) | N2 | HR3 (95% CI) | N2 | HR3 (95% CI) | N2 | HR3 (95% CI) | |
| Quintile 1 | 93 | 1.00 (ref.) | 57 | 1.00 (ref.) | 69 | 1.00 (ref.) | 81 | 1.00 (ref.) |
| Quintile 2 | 59 | 0.76 (0.48, 1.19) | 65 | 0.75 (0.53, 1.08) | 60 | 0.73 (0.49, 1.08) | 64 | 0.84 (0.49, 1.42) |
| Quintile 3 | 56 | 0.72 (0.45, 1.17) | 69 | 0.65 (0.46, 0.94) | 63 | 0.69 (0.46, 1.04) | 62 | 0.60 (0.34, 1.08) |
| Quintile 4 | 67 | 0.58 (0.33, 1.02) | 111 | 0.81 (0.58, 1.13) | 78 | 0.67 (0.46, 1.00) | 100 | 0.77 (0.43, 1.38) |
| Quintile 5 | 70 | 0.43 (0.21, 0.87) | 132 | 0.78 (0.56, 1.09) | 74 | 0.56 (0.37, 0.85) | 128 | 0.92 (0.50, 1.66) |
| Ptrend = 0.004 | Ptrend = 0.552 | Ptrend = 0.004 | Ptrend = 0.846 | |||||
| PHDI Quintiles1 |
Education |
Neighborhood SES4 |
||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| ≤12 y |
13–15 y |
≥16 y |
Low |
High |
||||||
| N2 | HR3 (95% CI) | N2 | HR3 (95% CI) | N2 | HR3 (95% CI) | N2 | HR3 (95% CI) | N2 | HR3 (95% CI) | |
| Quintile 1 | 45 | 1.00 (ref.) | 43 | 1.00 (ref.) | 61 | 1.00 (ref.) | 97 | 1.00 (ref.) | 53 | 1.00 (ref.) |
| Quintile 2 | 28 | 0.57 (0.31, 1.04) | 56 | 1.18 (0.74, 1.89) | 40 | 0.53 (0.34, 0.84) | 84 | 1.10 (0.78, 1.55) | 40 | 0.34 (0.21, 0.57) |
| Quintile 3 | 21 | 0.47 (0.25, 0.87) | 67 | 1.13 (0.71, 1.81) | 36 | 0.42 (0.26, 0.67) | 94 | 0.93 (0.65, 1.31) | 31 | 0.30 (0.20, 0.55) |
| Quintile 4 | 35 | 0.56 (0.30, 1.01) | 66 | 1.10 (0.68, 1.78) | 77 | 0.56 (0.37, 0.85) | 104 | 0.93 (0.65, 1.32) | 74 | 0.47 (0.30, 0.72) |
| Quintile 5 | 44 | 0.73 (0.40, 1.33) | 69 | 1.16 (0.71, 1.88) | 86 | 0.41 (0.27, 0.64) | 106 | 0.81 (0.57, 1.17) | 96 | 0.50 (0.32, 0.77) |
| Ptrend = 0.366 | Ptrend = 0.957 | Ptrend = 0.004 | Ptrend = 0.081 | Ptrend = 0.017 | ||||||
| PHDI Quintiles1 |
BMI |
|||||||
|---|---|---|---|---|---|---|---|---|
| <25 kg/m2 |
25–<30 kg/m2 |
30–<35 kg/m2 |
≥ 35 kg/m2 |
|||||
| N2 | HR3 (95% CI) | N2 | HR3 (95% CI) | N2 | HR3 (95% CI) | N2 | HR3 (95% CI) | |
| Quintile 1 | 36 | 1.00 (ref.) | 33 | 1.00 (ref.) | 33 | 1.00 (ref.) | 51 | 1.00 (ref.) |
| Quintile 2 | 20 | 0.57 (0.29, 1.11) | 37 | 0.68 (0.38, 1.21) | 29 | 1.06 (0.55, 2.06) | 41 | 0.90 (0.56, 1.45) |
| Quintile 3 | 24 | 0.73 (0.39, 1.37) | 40 | 0.52 (0.29, 0.93) | 21 | 0.70 (0.35, 1.41) | 43 | 0.91 (0.56, 1.49) |
| Quintile 4 | 35 | 0.53 (0.28, 1.03) | 53 | 0.56 (0.32, 0.97) | 34 | 0.95 (0.50, 1.81) | 57 | 0.98 (0.61, 1.58) |
| Quintile 5 | 48 | 0.55 (0.28, 1.05) | 86 | 0.67 (0.39, 1.13) | 39 | 0.96 (0.50, 1.86) | 27 | 0.58 (0.33, 1.05) |
| Ptrend = 0.055 | Ptrend = 0.464 | Ptrend = 0.869 | Ptrend = 0.159 | |||||
Abbreviations: CI, confidence interval; CVD, cardiovascular disease; HR, hazard ratio; PHDI, Planetary Health Diet Index; SES, socioeconomic status.
The PHDI was 36.8–<71.2 for quintile 1, 71.2–<78.2 for quintile 2, 78.2–<84.4 for quintile 3, 84.4–<91.4 for quintile 4, and 91.4–126.6 for quintile 5.
Number of deaths from CVD.
HRs were adjusted for education, BMI, smoking status, alcohol consumption, physical activity, age, and energy, with the stratification variable removed from the model.
Low neighborhood SES was defined as neighborhood SES score < mean score in the population, and high neighborhood SES was defined as neighborhood SES score ≥ the mean score.
Discussion
In this large cohort of United States Black females, participants in the highest quintile of adherence to the PHD were estimated to have an 18% lower all-cause mortality and 26% lower mortality from CVD relative to participants whose diet was in the lowest quintile. Participants in the middle quintiles were estimated to have a similarly reduced mortality risk. Among participants who were at low baseline risk of dying from CVD because of younger age, never smoked, lower BMI, higher education, and residence in higher socioeconomic neighborhoods, there was a notably stronger association of PHDI with CVD mortality, with a further reduction with increasing quintile of PHDI. PHDI was not associated with cancer-specific mortality overall.
Consistent with the current study, previous studies among Chinese [18], Swedish [19,21], and United States (predominately White) populations [1] reported an inverse association of adherence to PHD with all-cause and CVD-specific mortality [1,[18], [19], [20], [21]]. In the Singapore Chinese Health Study, which identified ∼23,000 deaths over a median follow-up of 23 y, adherence to PHD was associated with a 21% reduction in CVD-specific mortality risk comparing the highest and lowest quintiles of PHD Score [18]. The analysis of the Cohort of Swedish Men and the Swedish Mammography Cohort, which identified over 24,000 deaths over 20 y of follow-up, showed that adherence to PHD was associated with a 15% reduction in CVD-specific mortality risk and a 19% reduction in all-cause mortality risk among females comparing those with high and low adherence with PHD [19]. In the Malmö Diet and Cancer (MDC) Study, which observed 7030 deaths over 20 y of follow-up, adherence to PHD was associated with a 32% lower risk of CVD-specific mortality comparing those with the highest adherence to those with the lowest adherence [21]. A pooled analysis of Nurses’ Health Study (NHS), NHS Ⅱ, and Health Professionals Follow-up Study (HPFS), with over 54,000 deaths during 30 y of follow-up, found that adherence to PHD was associated with a 14% lower risk of CVD-specific mortality comparing the highest and lowest quintiles [1]. In contrast, the European Prospective Investigation into Cancer and Nutrition (EPIC) Oxford study reported a null association between adherence to PHD and all-cause mortality. It is possible that the lack of association observed in that study could be because of the relatively narrow range of PHD adherence score (0–14) and the large proportion of vegetarians in the study population [27].
The MDC Study and the pooled analysis of HPFS and NHS both reported an association of adherence to PHD with a lower risk of cancer-specific mortality [1,21], whereas our study did not. Of note, in the pooled study of HPFS and NHS, which used cumulative average PHDI as exposure, a sensitivity analysis showed that the association of PHDI with cancer-specific mortality was greatly attenuated when baseline PHDI was used instead of a cumulative average [1]. In the current study, we did not use a cumulative average because we had FFQ data on all 15 food groups from only 1 time point. Interestingly, the analysis of the Cohort of Swedish Men and the Swedish Mammography Cohort found that higher adherence to PHD was not only correlated with lower risk of all-cause and CVD-specific mortality, but also with increased predicted dietary exposure to food contaminants, which might attenuate a protective effect of PHD on cancer-specific mortality [19]; however, we were not able to assess the potential effects of dietary pesticide exposure in the BWHS.
Our findings regarding the PHDI are also consistent with previous studies on associations of Mediterranean diet and plant-based diet, 2 healthy dietary patterns similar to PHD, with all-cause and CVD-specific mortality [11,12,28]. A systematic review concluded that Mediterranean diet was associated with a significantly lower risk of CVD-specific mortality based on 6 prospective cohort studies involving over 100,000 participants and over 3000 deaths [11]. Regarding plant-based diet, based on a meta-analysis of 12 cohort studies with over 42,000 deaths among over 500,000 participants, there was a significant inverse association between adherence to plant-based diet and risk of coronary artery disease mortality (HR = 0.77, 95% CI: 0.70, 0.86) [12]. In the Multiethnic Cohort Study, which involved over 78,000 females and ∼9000 deaths in females over 21 years of follow-up, a healthful plant-based diet was associated with a 21% lower risk of CVD-specific mortality in females comparing the highest and lowest quintile of healthful plant-based diet index (HR = 0.79, 95% CI: 0.74, 0.85) [28].
The PHD was developed not only to promote healthy eating and improve health but also to support environmental sustainability by reducing food production-related environmental degradation, such as cropland use, biodiversity loss, water use, greenhouse gas emissions, and nitrogen and phosphorus pollution [3]. Previous environmental impact studies reported inconsistent estimations on the environmental impacts of PHD: some studies found positive environmental impacts of PHD such as lower greenhouse gas emissions, fertilizer needs, cropland use, and water needs [1,29], whereas some studies reported negative environmental impacts such as higher water use, cropland use, and greenhouse gas emission [18,30], potentially because of different populations and conversion tables used. Thus, the environmental impact of the PHD adoption remained inconclusive.
A limitation of our study is that dietary data were not collected again on the entire study population until 2021, so we were unable to capture changes in diet. This would likely have increased nondifferential measurement error, possibly obscuring or minimizing a true association. In addition, because the FFQ had only 85 questions, some food groups had limited inputs; for instance, the food group nuts could accept inputs on many nuts including walnuts, almonds, and pecans, but our FFQ asked only about peanuts. Nevertheless, the magnitude of associations observed in our study was very similar to that observed in the previous studies of 3 different populations with repeated measures of dietary intake.
To our knowledge, the present study is the first to examine the association between adherence to PHD and mortality in a Black population, which experiences a disproportionately high mortality burden. Study participants reside in diverse regions and neighborhoods in the United States. The current analysis involved over 30,000 Black females and 3527 deaths during 18 y of follow-up, which ensured high statistical power. Additionally, adherence to PHD was measured by PHDI, an index aligned with the EAT-Lancet diet. With a wide range of 0–140 and a sufficient range for food groups, the PHDI likely captures most variability in food group consumption and discriminates adherence levels to PHD [1]. Given varying approaches to dietary exposure assessment, our view is that finding associations across most FFQs and most populations adds weight to the evidence of a positive association of the PHDI with improved human health. Dietary data were collected using a validated FFQ [23] which ensured the validity of the dietary data used to calculate the PHDI. Moreover, we utilized comprehensive data on important potential confounders in multivariable analyses.
In conclusion, we observed that adherence to PHD was associated with a lower risk of all-cause and CVD-specific mortality among Black females. Although the magnitude of the associations was modest overall, females with a low baseline risk were estimated to have as much as half risk of dying from CVD if they had high compared with low adherence to the PHD. These findings strengthen evidence for the role of plant-based diets in reducing mortality across diverse populations. A plant-based diet has already been shown to be beneficial for the planet; evidence of its benefit to human health may accelerate the uptake of the diet.
Author contributions
The authors’ responsibilities were as follows – YS, KAB, JRP: contributed to the study conception and design; KAB, JRP: performed data collection; YS: performed statistical analyses and wrote the first draft of the manuscript; JLP, JRP, KAB, SS: provided critical feedback on the draft and read and approved the final manuscript; JRP, KAB: have primary responsibility for final content; and all authors: read and approved the final version.
Funding
This research was supported by the National Institutes of Health (CA164974). The content of this study is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.
Data availability
The data underlying this article cannot be shared publicly to protect the privacy of individuals who participated in the study. The data will be shared on reasonable request to the corresponding author.
Conflict of interest
The authors declare that they have no conflicts of interest.
Acknowledgments
We thank the participants and staff of the BWHS for their contributions.
References
- 1.Bui L.P., Pham T.T., Wang F., Chai B., Sun Q., Hu F.B., et al. Planetary Health Diet Index and risk of total and cause-specific mortality in three prospective cohorts. Am. J. Clin. Nutr. 2024;120(1):80–91. doi: 10.1016/j.ajcnut.2024.03.019. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Willett W., Rockström J., Loken B., Springmann M., Lang T., Vermeulen S., et al. Food in the Anthropocene: the EAT–Lancet Commission on healthy diets from sustainable food systems. Lancet. 2019;393(10170):447–492. doi: 10.1016/s0140-6736(18)31788-4. [DOI] [PubMed] [Google Scholar]
- 3.Dwyer-Lindgren L., Kendrick P., Kelly Y.O., Baumann M.M., Compton K., Blacker B.F. Cause-specific mortality by county, race, and ethnicity in the USA, 2000–19: a systematic analysis of health disparities. Lancet. 2023;402(10407):1065–1082. doi: 10.1016/s0140-6736(23)01088-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Cunningham T.J., Croft J.B., Liu Y., Lu H., Eke P.I., Giles W.H. Vital signs: racial disparities in age-specific mortality among Blacks or African Americans—United States, 1999–2015. MMWR Morb. Mortal. Wkly. Rep. 2017;66(17):444–456. doi: 10.15585/mmwr.mm6617e1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Kyalwazi A.N., Loccoh E.C., Brewer L.C., Ofili E.O., Xu J., Song Y., et al. Disparities in cardiovascular mortality between Black and White adults in the United States, 1999 to 2019. Circulation. 2022;146(3):211–228. doi: 10.1161/circulationaha.122.060199. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Micha R., Peñalvo J.L., Cudhea F., Imamura F., Rehm C.D., Mozaffarian D. Association between dietary factors and mortality from heart disease, stroke, and type 2 diabetes in the United States. JAMA. 2017;317(9):912. doi: 10.1001/jama.2017.0947. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Naghshi S., Sadeghi O., Willett W.C., Esmaillzadeh A. Dietary intake of total, animal, and plant proteins and risk of all cause, cardiovascular, and cancer mortality: systematic review and dose-response meta-analysis of prospective cohort studies. BMJ. 2020;370:m2412. doi: 10.1136/bmj.m2412. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Balakrishna R., Bjørnerud T., Bemanian M., Aune D., Fadnes L.T. Consumption of nuts and seeds and health outcomes including cardiovascular disease, diabetes and metabolic disease, cancer, and mortality: an umbrella review. Adv. Nutr. 2022;13(6):2136–2148. doi: 10.1093/advances/nmac077. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Aune D. Plant foods, antioxidant biomarkers, and the risk of cardiovascular disease, cancer, and mortality: a review of the evidence. Adv. Nutr. 2019;10(Suppl 4):S404–S421. doi: 10.1093/advances/nmz042. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Blanco Mejia S., Messina M., Li S.S., Viguiliouk E., Chiavaroli L., Khan T.A., et al. A meta-analysis of 46 studies identified by the FDA demonstrates that soy protein decreases circulating LDL and total cholesterol concentrations in adults. J. Nutr. 2019;149(6):968–981. doi: 10.1093/jn/nxz020. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Laffond A., Rivera-Picón C., Rodríguez-Muñoz P.M., Juárez-Vela R., Ruiz de Viñaspre-Hernández R., Navas-Echazarreta N., et al. Mediterranean diet for primary and secondary prevention of cardiovascular disease and mortality: an updated systematic review. Nutrients. 2023;15(15):3356. doi: 10.3390/nu15153356. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Jafari S., Hezaveh E., Jalilpiran Y., Jayedi A., Wong A., Safaiyan A., et al. Plant-based diets and risk of disease mortality: a systematic review and meta-analysis of cohort studies. Crit. Rev. Food Sci. Nutr. 2022;62(28):7760–7772. doi: 10.1080/10408398.2021.1918628. [DOI] [PubMed] [Google Scholar]
- 13.Chen W., Zhang S., Hu X., Chen F., Li D. A review of healthy dietary choices for cardiovascular disease: from individual nutrients and foods to dietary patterns. Nutrients. 2023;15(23):4898. doi: 10.3390/nu15234898. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Casas R., Castro-Barquero S., Estruch R., Sacanella E. Nutrition and cardiovascular health. Int. J. Mol. Sci. 2018;19(12):3988. doi: 10.3390/ijms19123988. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Mensah G.A., Brown A.G.M., Pratt C.A. Nutrition disparities and cardiovascular health. Curr. Atheroscler. Rep. 2020;22(4):15. doi: 10.1007/s11883-020-0833-3. [DOI] [PubMed] [Google Scholar]
- 16.Siegel R.L., Giaquinto A.N., Jemal A. Cancer statistics, 2024. CA Cancer J. Clin. 2024;74(1):12–49. doi: 10.3322/caac.21820. [DOI] [PubMed] [Google Scholar]
- 17.Clinton S.K., Giovannucci E.L., Hursting S.D. The World Cancer Research Fund/American Institute for Cancer Research Third Expert Report on Diet, Nutrition, Physical Activity, and Cancer: impact and future directions. J Nutr. 2020;150(4):663–671. doi: 10.1093/jn/nxz268. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Ye Y.-X., Geng T.-T., Zhou Y.-F., He P., Zhang J.-J., Liu G., et al. Adherence to a planetary health diet, environmental impacts, and mortality in Chinese adults. JAMA Netw. Open. 2023;6(10) doi: 10.1001/jamanetworkopen.2023.39468. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Pitt S., Kałuża J., Widenfalk A., Åkesson A., Wolk A. Adherence to the EAT-Lancet diet in relation to mortality and exposure to food contaminants in population-based cohorts of Swedish men and women. Environ. Int. 2024;184 doi: 10.1016/j.envint.2024.108495. [DOI] [PubMed] [Google Scholar]
- 20.Karavasiloglou N., Thompson A.S., Pestoni G., Knuppel A., Papier K., Cassidy A., et al. Adherence to the EAT-Lancet reference diet is associated with a reduced risk of incident cancer and all-cause mortality in UK adults. One Earth. 2023;6(12):1726–1734. doi: 10.1016/j.oneear.2023.11.002. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Stubbendorff A., Sonestedt E., Ramne S., Drake I., Hallström E., Ericson U. Development of an EAT-Lancet index and its relation to mortality in a Swedish population. Am. J. Clin. Nutr. 2022;115(3):705–716. doi: 10.1093/ajcn/nqab369. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Rosenberg L., Adams-Campbell L., Palmer J.R. The Black Women's Health Study: a follow-up study for causes and preventions of illness. J. Am. Med. Womens Assoc. 1995;50(2):56–58. 1972. [PubMed] [Google Scholar]
- 23.Block G., Hartman A.M., Naughton D. A reduced dietary questionnaire: development and validation. Epidemiology. 1990;1(1):58–64. doi: 10.1097/00001648-199001000-00013. [DOI] [PubMed] [Google Scholar]
- 24.Kumanyika S.K., Mauger D., Mitchell D.C., Phillips B., Smiciklas-Wright H., Palmer J.R. Relative validity of food frequency questionnaire nutrient estimates in the Black Women's Health Study. Ann. Epidemiol. 2003;13(2):111–118. doi: 10.1016/s1047-2797(02)00253-3. [DOI] [PubMed] [Google Scholar]
- 25.Coogan P.F., Cozier Y.C., Krishnan S., Wise L.A., Adams-Campbell L.L., Rosenberg L., et al. Neighborhood socioeconomic status in relation to 10-year weight gain in the Black Women's Health Study. Obesity. 2010;18(10):2064–2065. doi: 10.1038/oby.2010.69. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Boggs D.A., Palmer J.R., Spiegelman D., Stampfer M.J., Adams-Campbell L.L., Rosenberg L. Dietary patterns and 14-y weight gain in African American women. Am. J. Clin. Nutr. 2011;94(1):86–94. doi: 10.3945/ajcn.111.013482. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Knuppel A., Papier K., Key T.J., Travis R.C. EAT-Lancet score and major health outcomes: the EPIC-Oxford study. Lancet. 2019;394(10194):213–214. doi: 10.1016/s0140-6736(19)31236-x. [DOI] [PubMed] [Google Scholar]
- 28.Kim J., Wilkens L.R., Haiman C.A., Le Marchand L., Park S.-Y. Plant-based dietary patterns and mortality from all causes, cardiovascular disease, and cancer: the Multiethnic Cohort Study. Clin. Nutr. 2024;43(6):1447–1453. doi: 10.1016/j.clnu.2024.04.035. [DOI] [PubMed] [Google Scholar]
- 29.Tuninetti M., Ridolfi L., Laio F. Compliance with EAT–Lancet dietary guidelines would reduce global water footprint but increase it for 40% of the world population. Nat. Food. 2022;3(2):143–151. doi: 10.1038/s43016-021-00452-0. [DOI] [PubMed] [Google Scholar]
- 30.Guo Y., He P., Searchinger T.D., Chen Y., Springmann M., Zhou M., et al. Environmental and human health trade-offs in potential Chinese dietary shifts. One Earth. 2022;5(3):268–282. doi: 10.1016/j.oneear.2022.02.002. [DOI] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
The data underlying this article cannot be shared publicly to protect the privacy of individuals who participated in the study. The data will be shared on reasonable request to the corresponding author.