Planetary Health Diet Index and risk of total and cause-specific mortality in three prospective cohorts

Abstract

Background: In 2019, the EAT-Lancet Commission proposed a healthy dietary pattern that, along with reductions in food waste and improved agricultural practices, could feed the increasing global population sustainably. We developed a Planetary Health Diet Index (PHDI) to quantify adherence to the EAT-Lancet reference diet.

Objectives: We aimed to assess associations between PHDI and total and cause-specific mortality in 3 prospective cohorts of males and females in the United States.

Methods: We followed 66,692 females from the Nurses' Health Study (1986-2019), 92,438 females from the Nurses' Health Study II (1989-2019), and 47,274 males from the Health Professionals Follow-up Study (1986-2018) who were free of cancer, diabetes, and major cardiovascular diseases at baseline. The PHDI was calculated every 4 y using a semiquantitative food frequency questionnaire. Hazard ratios (HRs) were calculated using multivariable proportional-hazards models.

Results: During follow-up, we documented 31,330 deaths among females and 23,206 among males. When comparing the highest with the lowest quintile of PHDI, the pooled multivariable-adjusted HRs were 0.77 [95% confidence interval (CI): 0.75, 0.80] for all-cause mortality (P-trend < 0.0001). The PHDI was associated with lower risk of deaths from cardiovascular diseases (HR: 0.86; 95% CI: 0.81, 0.91), cancer (HR: 0.90; 95% CI: 0.85, 0.95), respiratory diseases (HR: 0.53; 95% CI: 0.48, 0.59), and neurodegenerative diseases (HR: 0.72; 95% CI: 0.67, 0.78). In females, but not males, the PHDI was also significantly associated with a lower risk of deaths from infectious diseases (HR: 0.62; 95% CI: 0.51, 0.76). PHDI scores were also associated inversely with greenhouse gas emissions and other environmental impacts.

Conclusions: In 3 large United States-based prospective cohorts of males and females with up to 34 y of follow-up, a higher PHDI was associated with lower risk of total and cause-specific mortality and environment impacts.

Keywords: diet pattern; mortality; planetary health diet; prospective cohort; sustainable diet.

FIGURE 1

Pooled HRs (95% CIs) for deciles of the PHDI in relation to total mortality among 206,404 males and females (54,536 deaths). HRs of cumulative average deciles were stratified jointly by age in months and follow-up cycles and adjusted for race (White or not), marriage status (married or not), living status (alone or not), neighborhood socioeconomic status z-score, menopausal status (pre or postmenopausal [never, past, or current menopausal hormone use], for females only), multivitamin use (yes/no), aspirin use (yes/no), total energy intake (kcal/d), baseline BMI (kg/m²: <23, 23–24.9, 25–29.9, 30–34.9, or ≥35), smoking status (never smoker, former smoker, current smoker: 1–14, 15–24, or ≥25 cigarettes/d), alcohol drinking (g/d: 0, 0.1–4.9, 5.0–9.9, 10.0–14.9, 15.0–29.9, or ≥30), physical activity (MET-h/wk: <3, 3–8.9, 9–17.9, 18–26.9, or ≥27), history of hypertension (yes/no), history of hypercholesterolemia (yes/no), family history of myocardial infarction (yes/no), family history of diabetes (yes/no), family history of cancer (yes/no). Fixed effects meta-analysis of the 3 cohorts (NHS1, NHS2, and HPFS) based on Dersimonian–Laird approach with inverse-variance weight was used to pool HRs. Mean (SD) of each decile of the PHDI for all 3 cohorts are presented. CI, confidence interval; HPFS, Health Professionals Follow-up Study; HR, hazard ratio; MET, metabolic equivalent of task; NHS1, Nurses’ Health Study; NHS2, Nurses’ Health Study II; PHDI, Planetary Health Diet Index; SD, standard deviation.

FIGURE 2

Pooled HRs (95% CIs) of the PHDI for 20-point increase in total and cause-specific mortality in the 3 prospective cohorts. HRs of the continuous cumulative average PHDI scaled by 20 points (i.e., approximate to the difference between 10th and 90th percentiles of the PHDI) were stratified jointly by age in months and follow-up cycles and adjusted for race (White or not), marriage status (married or not), living status (alone or not), neighborhood socioeconomic status z-score, menopausal status (pre or postmenopausal [never, past, or current menopausal hormone use], for females only), multivitamin use (yes/no), aspirin use (yes/no), total energy intake (kcal/d), baseline BMI (kg/m²: <23, 23–24.9, 25–29.9, 30–34.9, or ≥35), smoking status (never smoker, former smoker, current smoker: 1–14, 15–24, or ≥25 cigarettes/d), alcohol drinking (g/d: 0, 0.1–4.9, 5.0–9.9, 10.0–14.9, 15.0–29.9, or ≥30), physical activity (MET-h/wk: <3, 3–8.9, 9–17.9, 18–26.9, or ≥27), history of hypertension (yes/no), history of hypercholesterolemia (yes/no), family history of myocardial infarction (yes/no), family history of diabetes (yes/no), family history of cancer (yes/no). Fixed effects meta-analysis of the 3 cohorts (NHS1, NHS2, and HPFS) based on Dersimonian–Laird approach with inverse-variance weight was used to pool HRs. P value was estimated from Wald test of the PHDI variable. ∗NHS2 data were not included due to low number of cases. CI, confidence interval; HPFS, Health Professionals Follow-up Study; HR, hazard ratio; MET, metabolic equivalent of task; NHS1, Nurses’ Health Study; NHS2, Nurses’ Health Study II; PHDI, Planetary Health Diet Index.

FIGURE 3

Pooled HRs (95% CIs) of PHDI for 20-point increase in total mortality across subgroups in the 3 prospective cohorts. HRs of continuous cumulative average PHDI scaled to 20 points were stratified jointly by age in months and follow-up cycles and adjusted for race (White or not), marriage status (married or not), living status (alone or not), neighborhood socioeconomic status z-score, menopausal status (pre or postmenopausal [never, past, or current menopausal hormone use], for females only), multivitamin use (yes/no), aspirin use (yes/no), total energy intake (kcal/d), baseline BMI (kg/m²: <23, 23–24.9, 25–29.9, 30–34.9, ≥35), smoking status (never smoker, former smoker, current smoker: 1–14, 15–24 or ≥25 cigarettes/d), alcohol drinking (g/d: 0, 0.1–4.9, 5.0–9.9, 10.0–14.9, 15.0–29.9, or ≥30), physical activity (MET-h/wk: <3, 3–8.9, 9–17.9, 18–26.9, or ≥27), history of hypertension (yes/no), history of hypercholesterolemia (yes/no), family history of myocardial infarction (yes/no), family history of diabetes (yes/no), family history of cancer (yes/no), except the corresponding subgroup variables. Time-fixed effect modifiers included race, baseline hypercholesterolemia, baseline hypertension, family history of diseases (diabetes, myocardial infarction, or cancer), baseline BMI; other effect modifiers were time-varying. Fixed effects meta-analysis of the 3 cohorts (NHS1, NHS2, and HPFS) based on Dersimonian–Laird approach with inverse-variance weight was used to pool HRs. P value for interaction was estimated from Wald test of the interaction terms between continuous PHDI variables and the corresponding subgroup variables in the pooled data set. CI, confidence interval; HPFS, Health Professionals Follow-up Study; HR, hazard ratio; MET, metabolic equivalent of task; NHS1, Nurses’ Health Study; NHS2, Nurses’ Health Study II; PHDI, Planetary Health Diet Index.

FIGURE 4

Pooled HRs (95% CIs) of Q5 vs. Q1 of each food group index and intake for total mortality in the 3 prospective cohorts. Hazard ratios (HRs) of each food group intake and index were stratified jointly by age in months and follow-up cycles and adjusted for race (White or not), marriage status (married or not), living status (alone or not), neighborhood socioeconomic status z-score, menopausal status (pre or postmenopausal [never, past, or current menopausal hormone use], for females only), multivitamin use (yes/no), aspirin use (yes/no), total energy intake (kcal/d), baseline BMI (kg/m²: <23, 23–24.9, 25–29.9, 30–34.9, ≥35), smoking status (never smoker, former smoker, current smoker: 1–14, 15–24, or ≥25 cigarettes/d), alcohol drinking (g/d: 0, 0.1–4.9, 5.0–9.9, 10.0–14.9, 15.0–29.9, or ≥30), physical activity (MET-h/wk: <3, 3–8.9, 9–17.9, 18–26.9, or ≥27), history of hypertension (yes/no), history of hypercholesterolemia (yes/no), family history of myocardial infarction (yes/no), family history of diabetes (yes/no), family history of cancer (yes/no). Fixed effects meta-analysis of the 3 cohorts (NHS1, NHS2, and HPFS) based on Dersimonian–Laird approach with inverse-variance weight. ∗The scores for these unhealthy food groups were reverse-coded to show that higher scores correspond to lower consumption. ∗∗Due to low variation of this food consumption, only 2 quantiles were made; HRs of the highest quantile compared with lowest quantile were reported. CI, confidence interval; HPFS, Health Professionals Follow-up Study; HR, hazard ratio; MET, metabolic equivalent of task; NHS1, Nurses’ Health Study; NHS2, Nurses’ Health Study II; Q, quantile.