Association of adherence to various dietary patterns with mortality among survivors of myocardial infarction: a prospective cohort study

Abstract

Background:

Current dietary guidelines are largely based on evidence from the general population, and data on dietary patterns and survival after myocardial infarction (MI) remain limited. We evaluated post-MI adherence to multiple healthy dietary patterns, as well as changes in adherence from before to after MI diagnosis, in relation to mortality among MI survivors.

Methods:

We included 3,277 women from the Nurses’ Health Study and 2,618 men from the Health Professionals Follow-up Study who survived a non-fatal MI. The mean age at MI diagnosis was 68.2 (standard deviation, 10.2) years. Diet was assessed using validated food frequency questionnaires and updated every two to four years. We calculated post-MI adherence scores for eight healthy patterns: the Alternate Healthy Eating Index (AHEI), the Healthy Eating Index-2015 (HEI-2015), the Alternate Mediterranean Diet Score (AMED), the Dietary Approaches to Stop Hypertension (DASH), the healthful Plant-Based Diet Index (hPDI), the reversed empirical dietary inflammatory pattern (rEDIP), the reversed empirical dietary index for hyperinsulinemia (rEDIH), and the reversed empirical dietary index for insulin resistance (rEDIR). The main outcome was total mortality, with cardiovascular disease (CVD) mortality and recurrent non-fatal MI as secondary outcomes.

Results:

During 15.7 years of follow-up after MI, 3,858 deaths (65.4%) were documented, including 1,639 CVD deaths (42.5% of deaths). Comparing the highest with the lowest quintile, higher post-MI dietary scores were consistently associated with lower total mortality, with the hazard ratio (HR) of 0.67 (95% confidence interval [CI]: 0.60–0.74) for AHEI, 0.73 (0.65–0.81) for HEI-2015, 0.66 (0.59–0.75) for AMED, 0.77 (0.69–0.86) for DASH, 0.74 (0.66–0.83) for hPDI, 0.77 (0.68–0.86) for rEDIP, 0.73 (0.65–0.83) for rEDIR, and 0.71 (0.64–0.80) for rEDIH. Similar inverse associations were observed for CVD mortality and recurrent non-fatal MI. Improved AHEI adherence from before to after MI was associated with lower mortality (HR: 0.86; 95% CI: 0.76–0.97), whereas decreased adherence to most patterns was associated with higher mortality.

Conclusions:

Among individuals who survived MI, adherence to post-MI dietary patterns emphasizing higher diet quality was associated with better long-term survival. These findings support dietary guidance focusing on overall diet quality for the secondary prevention of deaths among MI survivors.

Introduction

Diet plays a fundamental role in the primary prevention of cardiovascular disease (CVD), the leading cause of death in the United States and other developed nations.^1,2 Numerous studies have evaluated the associations of individual nutrients or foods with CVD risk and mortality.^3,4 Rapidly-expanding literature has also clearly shown the importance of dietary patterns in CVD etiology. In comparison with single nutrients or foods, dietary patterns consider certain combinations of nutrients and foods, and as such may more closely reflect the diet consumed by individuals in the real-world setting.^5,6 In addition, dietary patterns might be easier for the public to interpret or translate into adherence to dietary recommendations. The 2020–2025 Dietary Guidelines for Americans highlights the importance of healthy dietary patterns at every life stage and encourages individuals to make shifts in their food and beverage choices to achieve a healthy pattern.⁷ Results of several prospective cohort studies have shown that adherence to healthy dietary pattern was consistently associated with a lower incidence of CVD and mortality, and related disease outcomes such as type 2 diabetes, primarily in the general population.^8–10 In contrast, evidence regarding the association of adherence to different dietary patterns with survival after myocardial infarction (MI) attack remains sparse.^11,12 Compared with the general population, individuals surviving MI attacks are at a particularly elevated risk for deaths, and consuming healthful diets may be especially critical for maintaining health and preserving lifespan after the MI. Effective dietary strategies may lead to further improvement of the extended life span of MI survivors observed in the last three decades.^13,14

To fill these knowledge gaps, using data collected from two large prospective cohort studies of U.S. women and men with up to 15.7 years of follow-up after MI, we derived post-MI dietary scores for five prevailing recommended healthy dietary patterns (i.e., Alternate Healthy Eating Index [AHEI], the Healthy Eating Index-2015 [HEI-2015], Alternate Mediterranean Diet Score [AMED], the Dietary Approaches to Stop Hypertension [DASH], healthful Plant-Based Diet Index [hPDI]), and also three empirically-derived dietary patterns (i.e., the reversed empirical dietary inflammatory pattern [rEDIP], the reversed empirical dietary index for hyperinsulinemia [rEDIH], and the reversed empirical dietary index for insulin resistance [rEDIR]), and evaluated their associations during the post-MI period, as well as changes from before to after MI diagnosis, with total and CVD mortality among MI survivors.

Methods

Data Availability Statement

The data that support the findings of this study are not publicly available because of participant confidentiality, informed consent restrictions, and cohort-specific data use policies. Further information on data access procedures for the Nurses’ Health Study and Health Professionals Follow-up Study is available from the corresponding author upon reasonable request and approval by the cohort leadership.

Study Population

The Nurses’ Health Study (NHS) included 121,700 female registered nurses aged 30 to 55 years who were enrolled in 1976.¹⁵ The Health Professionals Follow-up Study (HPFS) consisted of 51,529 male health professionals aged 40 to 75 years who were enrolled in 1986.¹⁶ Participants in both studies have been followed biennially through mailed questionnaires to collect and update information on lifestyles, health-related behaviors, and medical histories, with a cumulative response rate of over 90% per cycle.¹⁷ In the current analysis, we included men and women who survived a non-fatal MI diagnosis at and since baseline (n=8,122, 1984 for the NHS and 1986 for the HPFS, when a validated food frequency questionnaire [FFQ] with more than 110 items was first administered in both cohorts). Participants were excluded if they reported a diagnosis of diabetes, stroke, or cancer at baseline (n=507); reported stroke or cancer before MI diagnosis (n=95); had incomplete baseline dietary assessments, or daily energy intake out of a normal range (<2,510 or >14,644 kJ/day for women, and <3,347 or >17,573 kJ/day for men) (n=727); or had missing dietary data at time of MI diagnosis (n=898) (Figure S1). After exclusions, a total of 5,895 eligible survivors of MI were included in the present analysis, with an average of 15.7 years of follow-up after MI.

This study followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline for cohort studies. The completed STROBE checklist was provided in the Supplemental Material.

The study protocol was approved by the institutional review boards at the Brigham and Women's Hospital and the Harvard T.H. Chan School of Public Health. Completion of the self-administered questionnaire was considered to imply informed consent.

Assessment of dietary scores

Dietary information was collected using validated FFQs administered every two to four years. In each FFQ, participants were asked how often, on average, they consumed a standardized portion size of each food during the past year, with response options ranging from “never or less than once per month” to “at least six times per day”. Nutrient intake was computed by multiplying the frequency of intake for each food item by its nutrient content of the specified portion and then summing up nutrient contributions across all foods. Validation studies comparing intakes assessed by questionnaires with multiple diet records showed that these FFQs provided reasonably valid estimates for intakes of a wide range of foods and nutrients in both NHS and HPFS.^18,19 Using the nutrient and food components, we computed adherence scores for five established healthy dietary patterns (AHEI, HEI-2015, AMED, DASH, and hPDI) and three empirically-derived dietary patterns (EDIP, EDIR, and EDIH). The components and scoring criteria in detail for the five dietary prevailing healthy patterns defined a priori are summarized in the Supplemental Methods and Table S1. The AHEI included ten components, with a score possibly ranging from 0 to 100.²⁰ The HEI-2015 included 13 components, with a theoretical range of 0 to 100.²¹ The AMED included nine components, ranging from 9 to 45.^13,22 The DASH consisted of eight components, ranging from 8 to 40.²³ The hPDI included 18 components, ranging from 18 to 90.²⁴ For these five dietary patterns, higher scores represent greater adherence to the pattern. The three empirical dietary indices, i.e., EDIP, EDIR, and EDIH, were combinations of selected food items, of 39 predefined food groups, that maximize the prediction of biologically relevant cardiometabolic biomarkers.^25–27 Specifically, EDIP was derived to predict interleukin-6, C-reactive protein (CRP), and tumor necrosis factor α receptor 2.²⁵ Similarly, EDIR and EDIH were derived to predict insulin resistance (triglyceride to HDL cholesterol ratio) and hyperinsulinemia (C-peptide), respectively.^26,27 EDIP, EDIR and EDIH are weighted sums of 18 food groups, and higher (more positive) scores indicate higher inflammatory or insulinemic potential of diets and lower (more negative) scores indicate lower inflammatory or insulinemic potential of diets (Table S2). To facilitate comparison across dietary scores, we reversed the EDIP, EDIR, and EDIH scores so that higher reversed scores indicated lower inflammatory, insulin-resistant, or hyperinsulinemic potential of diet.

Assessment of nonfatal MI

On each biennial questionnaire, participants were asked to indicate whether they had physician-diagnosed non-fatal MI in the previous two years. For newly reported non-fatal MI, medical records were reviewed by study physicians blinded to the participants’ exposure status. Non-fatal MI was confirmed according to the World Health Organization criteria, which require typical clinical symptoms plus either diagnostic change on electrocardiogram or elevated cardiac enzymes.²⁸ For those whose medical records were unavailable, the diagnosis was considered probable if supported by telephone interview or additional corroborating information. We included all confirmed and probable cases in the analyses because the exclusion of probable cases did not materially change the results (data not shown).

Ascertainment of endpoint

The primary endpoint for this study was mortality, and recurrent non-fatal MI was a secondary endpoint. Deaths were reported by next of kin or the postal system or identified through searches of the National Death Index.²⁹ The follow-up for deaths in each cohort has been estimated to be more than 98% complete. The cause of deaths was identified from death certificates or review of medical records by physicians. For the present analysis, we evaluated all-cause mortality and deaths from CVD (International Classification of Diseases, ninth revision [ICD-9] codes: 390.0–459.9).

Data on recurrent non-fatal MI were ascertained in the NHS but were not collected in the HPFS; therefore, analyses of recurrent non-fatal MI were restricted to NHS participants. In the NHS, recurrent non-fatal MI was defined as a second non-fatal MI event occurring after participants had survived the first MI attack.

Assessment of covariates

Information on age, body weight, smoking status, physical activity, family history of MI, multivitamin use, use of antihypertensive or lipid lowering drug, aspirin use, diabetes drug use, and self-reported diagnosis of diseases, including hypertension, hypercholesterolemia, and diabetes, was assessed and updated in the biennial follow-up questionnaires. Body mass index (BMI) was calculated as self-reported weight in kilograms divided by height in meters squared. Physical activity was estimated by multiplying the energy expenditure in metabolic equivalent tasks (METs) measured in hours per week by hours spent on the activity, and the values of all activities were summed to derive total physical activity. Detailed descriptions on the validity and reproducibility of self-reported body weight, and physical activity have been published elsewhere.^30,31 For NHS participants, we also ascertained data on menopausal status and postmenopausal hormone use.

Statistical analysis

For each participant, this study calculated person-years from the date of MI diagnosis to last returned questionnaire, death, or end of follow-up (June 2021 for the NHS, and January 2020 for the HPFS), whichever came first. The distributions of continuous variables were assessed using the Kolmogorov-Smirnov normality test. Normally distributed variables were presented as means ± standard deviations (SDs), whereas skewed variables were presented as medians (interquartile ranges). Categorical variables were represented by frequency and percentage. To better assess long-term dietary intake and minimize the random within-person variation, the cumulative average of the dietary pattern scores from all FFQs were calculated by averaging the repeated measurements since MI diagnosis until outcome or end of follow-up. For example, the dietary scores in 1984 were used to predict mortality during the follow-up between 1984 and 1986, and the average score of 1984 and 1986 assessments was used for the follow-up between 1986 and 1990, and so forth. We stopped updating dietary information after a diagnosis of diabetes or cancer since these conditions may lead to changes in diet. Cox proportional hazards regression models with time-varying covariates were used to examine the associations between post-MI dietary scores and mortality. Follow-up was divided into two-year questionnaire cycles, and dietary scores and time-varying covariates were updated at the beginning of each cycle. To control finely for age, calendar time, and their potential interaction, models were stratified jointly by five-year age categories and two-year questionnaire cycles/calendar periods, allowing the baseline hazard to vary across age and calendar-time strata. Participants with missing dietary assessment at the time of MI diagnosis were excluded. For time-varying covariates with missing values during follow-up, the most recent previously reported value was carried forward. If participants had two or more consecutive missing FFQs, the most recent valid dietary information continued to be carried forward until a new valid FFQ became available. Hazard ratios (HRs) were used to estimate relative risks in each higher quintile of each dietary score in comparison with participants in the lowest quintile. In multivariable analysis, we adjusted for age, ancestry, MI duration, a family history of MI, BMI at MI diagnosis, as well as time-varying covariates, including smoking status, physical activity, multivitamin use, aspirin use, menopausal status and postmenopausal hormone use (women only), and total energy intake. Tests of linear trend across increasing quintiles of dietary score were conducted by assigning the median value for each quintile and fitting this continuous variable in the model. E values were calculated for the HRs to evaluate robustness subject to potential unmeasured confounding. A higher E value indicates that stronger unmeasured confounders would be required to account for the observed associations, whereas a small E value suggests that weak unmeasured confounders would be sufficient to explain the observed associations.³² Kaplan-Meier method was used to assess survival probability, and differences between curves were analyzed by the log-rank test. In addition, we used restricted cubic spline regressions to model the potential dose-response relationships between dietary scores after MI and mortality. Tests for nonlinearity were based on the likelihood ratio test, comparing the model with only the linear term to the model with the linear and the cubic spline terms. We also evaluated the association between the changes in dietary score from pre-MI to post-MI diagnosis (time-varying post-MI dietary score minus pre-MI dietary score) with mortality. In this analysis, dietary scores before MI diagnosis were also adjusted for in the multivariable model.

Several sensitivity analyses were performed to test the robustness of our findings. First, we restricted our analyses to adults with incident MI by excluding those with prevalent MI at baseline. Second, we excluded deaths occurring during the first four years after MI diagnosis to examine whether the results were impacted by reverse causation. Third, we examined potential confounding by partner’s education and self-rated socioeconomic status to the final model. Fourth, we further included the diagnoses of hypertension, hypercholesterolemia, and diabetes, which may be intermediate outcomes of diet, in our multivariable models to understand their impact on the associations of interest. Fifth, we continuously updated participants’ diet throughout follow-up regardless of diabetes or cancer diagnoses. Sixth, we also conducted a sensitivity analysis excluding current and former smokers to further reduce confounding by smoking status. Seventh, physical activity level could be an indicator of underlying severity of diseases, and MI survivors may avoid certain activity due to symptoms. We also adjusted for post-MI physical activity level with a two-year lag or excluding participants in the lowest quintile of physical activity. Eighth, wrist fracture was employed as a falsification endpoint for further evaluating the impact of uncontrolled confounding, as this outcome may not necessarily have a strong connection with the diet. Ninth, for hPDI, we also created alternative hPDIs by scoring certain healthy animal foods (fish, poultry, fermented dairy, low-fat dairy, and eggs) positively. Lastly, to account for competing risks, we further conducted sensitivity analyses using Fine-Gray sub-distribution Cox models. For CVD mortality, non-CVD deaths were treated as competing events. For recurrent non-fatal MI, deaths before recurrent non-fatal MI were treated as competing events. All statistical analyses were performed using SAS software, version 9.1 (SAS Institute Inc, Cary, NC), and a P value <0.05 was considered statistically significant.

Results

Participant characteristics

Among 5,895 MI survivors, 3,277 (55.6%) were women from the NHS and 2,618 (44.4%) were men from the HPFS; the mean age at MI diagnosis was 68.2 (SD, 10.2) years; and 5,643 participants (95.7%) were White. The age standardized characteristics at MI diagnosis of the study participants according to quintiles of the eight dietary scores are presented in Table 1 and Table S3. In both cohorts, compared with the MI survivors with lower dietary scores, those with higher dietary scores were more likely to be physically active, have lower BMI and use postmenopausal hormones (women), and were less likely to be current smokers. Higher dietary scores were also associated with higher consumption of vegetables, fruits, whole grains, and nuts/legumes, and lower consumption of refined grains and red/processed meats. AHEI, HEI-2015, AMED, DASH, and hPDI significantly correlated with each other (all P<0.001), with Spearman correlation coefficients (r) ranging from 0.41 to 0.75 (Figure S2). The rEDIP, rEDIH, and rEDIR had relatively lower correlations with other dietary patterns. Most individual components across different dietary patterns were concordantly distributed by dietary score quintiles (Figure S3). The participants were also classified largely consistently into the quintiles among the five established dietary scores and the three empirical scores, respectively (Figure S4).

Table 1.

Characteristics of survivors of myocardial infarction at diagnosis according to quintiles of well-established dietary scores.^*

Characteristic	AHEI		HEI-2015		AMED		DASH		hPDI
	Q1	Q5	Q1	Q5	Q1	Q5	Q1	Q5	Q1	Q5
Overall
Dietary score	38.0±6.1	67.7±6.9	55.5±6.3	84.8±4.7	18.9±2.3	35.3±2.5	16.8±2.1	30.7±2.0	44.4±3.2	65.3±3.4
No. of participants	1299	1092	1296	1147	1307	1072	1154	1068	1169	1068
Age† (year)	67.7±10.1	68.6±9.7	67.7±10.3	68.5±9.8	67.8±10.3	68.8±9.7	67.3±10.3	69.3±9.7	68.3±10.2	68.3±9.7
Women	701(54.0)	621(56.9)	715(55.2)	665(58.0)	717(54.9)	565(52.7)	626(54.2)	596(55.8)	629(53.8)	629(58.9)
White	1250(96.2)	1034(94.7)	1239(95.6)	1101(96.0)	1247(95.4)	1027(95.8)	1101(95.4)	1022(95.7)	1128(96.5)	1016(95.1)
BMI (kg/m2)	27.1±5.2	26.3±4.6	27.0±5.3	26.3±4.5	27.1±5.2	26.1±4.4	27.0±5.3	26.0±4.6	27.2±5.3	26.4±4.7
Physical activity (MET-hours/week)	0.0(0.0,1.8)	1.5(0.0,5.0)	0.1(0.0,1.7)	1.5(0.0,4.8)	0.0(0.0,1.5)	1.7(0.0,5.0)	0.0(0.0,1.8)	1.5(0.0,5.0)	0.2(0.0,2.0)	1.2(0.0,4.0)
Current smoker	333(25.6)	99(9.1)	347(26.8)	102(8.9)	325(24.9)	100(9.3)	332(28.8)	78(7.3)	235(20.1)	137(12.8)
Menopausal hormone use‡	395(56.4)	387(62.3)	401(56.1)	426(64.0)	402(56.1)	375(66.4)	364(58.2)	370(62.1)	353(56.1)	399(63.5)
Multivitamin use	713(54.9)	470(43.0)	718(55.4)	485(42.3)	732(56.0)	460(42.9)	655(56.8)	436(40.8)	608(52.0)	515(48.2)
Family history of MI	516(39.7)	467(42.8)	505(39.0)	483(42.1)	499(38.2)	448(41.8)	448(38.8)	444(41.6)	441(37.7)	457(42.8)
Hypertension	778(59.9)	662(60.6)	741(57.2)	708(61.7)	795(60.8)	671(62.6)	697(60.4)	649(60.8)	718(61.4)	660(61.8)
Hypercholesterolemia	742(57.1)	660(60.4)	721(55.6)	687(59.9)	750(57.4)	659(61.5)	664(57.5)	640(59.9)	672(57.5)	666(62.4)
Regular aspirin use	655(50.4)	545(49.9)	658(50.8)	532(46.4)	664(50.8)	511(47.7)	584(50.6)	508(47.6)	586(50.1)	533(49.9)
Use of anti-hypertensive drug	629(48.4)	533(48.8)	590(45.5)	575(50.1)	638(48.8)	548(51.1)	556(48.2)	534(50.0)	566(48.4)	533(49.9)
Use of lipid-lowering medication	325(25.0)	322(29.5)	299(23.1)	328(28.6)	348(26.6)	318(29.7)	298(25.8)	294(27.5)	321(27.5)	328(30.7)
Dietary intake
Total energy, kcal/d	1897±601	1743±561	2038±651	1550±475	1544±556	2134±581	1803±614	1886±546	2141±610	1543±518
Alcohol intake, g/day	1.5(0.0,9.5)	1.5(0.0,7.6)	1.1(0.0,7.3)	1.5(0.0,8.4)	0.0(0.0,4.7)	3.9(0.0,11.9)	1.8(0.0,9.2)	1.2(0.0,7.3)	1.8(0.0,9.3)	1.5(0.0,8.9)
Vegetables, servings/d	2.4±1.5	4.4±2.3	2.6±1.6	3.7±2.0	1.9±1.2	5.0±2.3	2.2±1.2	4.6±2.4	2.8±1.7	3.9±2.4
Fruits, servings/d	0.9±0.7	2.5±1.4	1.0±0.9	2.1±1.2	0.9±0.8	2.6±1.6	0.8±0.6	2.7±1.5	1.2±1.0	2.1±1.6
Red and processed meats, servings/d	1.4±0.7	0.5±0.4	1.3±0.8	0.5±0.4	1.1±0.7	0.7±0.6	1.3±0.7	0.6±0.4	1.3±0.8	0.6±0.5
Whole grains, servings/d	1.0±1.0	1.9±1.5	0.9±1.1	1.8±1.4	0.7±0.8	2.2±1.5	0.7±0.7	2.2±1.5	1.0±1.0	1.8±1.5
Refined grains, servings/d	1.8±1.5	1.2±1.2	2.1±1.6	0.9±0.8	1.5±1.3	1.6±1.5	1.8±1.5	1.3±1.2	2.2±1.4	1.0±1.1
Nuts and legumes, servings/d	0.2±0.3	0.8±0.8	0.4±0.6	0.6±0.6	0.2±0.3	0.9±0.8	0.3±0.5	0.7±0.7	0.4±0.5	0.6±0.7
NHS
Dietary score	39.3±6.0	68.6±7.1	55.1±6.3	84.5±5.1	18.7±2.2	35.0±2.4	16.9±2.0	30.5±1.9	44.3±3.3	65.4±3.5
No. of participants	701	621	715	665	717	565	626	596	629	629
Age† (year)	68.2±9.8	68.3±10.4	68.3±9.9	68.3±10.7	68.4±10.1	68.2±10.2	68.2±9.8	68.4±10.4	68.4±10.0	67.9±10.4
White	687(98.0)	609(98.0)	704(98.4)	656(98.6)	701(97.8)	558(98.8)	612(97.8)	589(98.9)	622(98.9)	620(98.6)
BMI (kg/m2)	27.3±6.0	26.6±5.1	27.2±6.0	26.7±5.2	27.3±6.0	26.1±4.8	27.2±6.0	26.2±5.1	27.6±6.1	26.6±5.2
Physical activity (MET-hours/week)	0.0(0.0,1.0)	1.2(0.0,3.5)	0.0(0.0,1.0)	1.0(0.0,3.3)	0.0(0.0,1.0)	1.2(0.0,3.5)	0.0(0.0,1.2)	1.2(0.0,3.4)	0.0(0.0,1.2)	1.0(0.0,3.0)
Current smoker	225(32.1)	78(12.5)	237(33.2)	76(11.5)	226(31.5)	77(13.6)	218(34.9)	69(11.6)	162(25.7)	107(17.0)
Menopausal hormone use	395(56.4)	387(62.3)	401(56.1)	426(64.0)	402(56.1)	375(66.4)	364(58.2)	370(62.1)	353(56.1)	399(63.5)
Multivitamin use	367(52.4)	254(40.9)	380(53.2)	275(41.4)	374(52.1)	225(39.9)	344(54.9)	247(41.5)	291(46.3)	284(45.2)
Family history of MI	264(37.7)	253(40.7)	265(37.1)	278(41.8)	258(36.0)	234(41.4)	234(37.4)	233(39.1)	237(37.6)	259(41.2)
Hypertension	467(66.6)	404(65.0)	453(63.4)	436(65.6)	475(66.3)	374(66.2)	428(68.3)	379(63.6)	431(68.5)	402(63.9)
Hypercholesterolemia	432(61.6)	391(63.0)	429(60.0)	414(62.3)	447(62.4)	354(62.7)	388(62.0)	364(61.0)	392(62.3)	399(63.4)
Regular aspirin use	330(47.1)	276(44.4)	329(46.0)	286(43.0)	323(45.0)	236(41.8)	291(46.5)	246(41.2)	276(43.9)	292(46.5)
Use of anti-hypertensive drug	397(56.7)	323(52.0)	375(52.5)	352(52.9)	404(56.3)	306(54.2)	356(56.8)	313(52.5)	358(56.9)	323(51.4)
Use of lipid-lowering medication	191(27.2)	195(31.4)	182(25.5)	204(30.7)	211(29.4)	181(32.1)	179(28.6)	178(29.9)	191(30.3)	206(32.7)
Dietary intake
Total energy, kcal/d	1759±534	1627±519	1871±581	1437±452	1415±495	2011±522	1631±530	1759±483	1984±526	1400±455
Alcohol intake, g/day	0.0(0.0,2.8)	0.9(0.0,4.7)	0.0(0.0,2.7)	0.9(0.0,4.7)	0.0(0.0,1.8)	2.1(0.0,7.9)	0.9(0.0,3.6)	0.9(0.0,4.7)	0.0(0.0,3.4)	0.9(0.0,4.7)
Vegetables, servings/d	2.3±1.4	4.3±2.0	2.5±1.7	3.6±1.8	1.9±1.2	4.9±1.9	2.2±1.2	4.5±1.9	2.8±1.8	3.7±2.0
Fruits, servings/d	0.9±0.7	2.4±1.3	1.0±0.9	2.0±1.1	0.8±0.8	2.5±1.3	0.8±0.6	2.5±1.2	1.2±1.0	1.9±1.2
Red and processed meats, servings/d	1.2±0.6	0.5±0.4	1.1±0.7	0.5±0.4	0.9±0.6	0.7±0.6	1.1±0.6	0.6±0.4	1.2±0.6	0.6±0.4
Whole grains, servings/d	0.9±1.0	1.7±1.4	0.9±1.1	1.7±1.3	0.6±0.8	2.0±1.4	0.6±0.6	2.0±1.4	1.0±0.9	1.6±1.3
Refined grains, servings/d	1.8±1.4	1.1±1.0	2.1±1.6	0.9±0.8	1.5±1.3	1.4±1.2	1.8±1.4	1.1±1.0	2.2±1.4	0.9±0.8
Nuts and legumes, servings/d	0.2±0.2	0.7±0.7	0.3±0.4	0.4±0.5	0.2±0.3	0.7±0.7	0.2±0.4	0.6±0.6	0.3±0.3	0.5±0.6
HPFS
Dietary score	36.1±5.7	66.2±6.4	56.4±6.0	85.1±4.1	19.2±2.3	35.7±2.6	16.8±2.0	31.0±2.0	44.5±3.0	65.1±3.4
No. of participants	598	471	581	482	590	507	528	472	540	439
Age† (year)	68.3±9.9	68.4±9.3	68.4±9.9	68.4±9.3	68.3±10.3	68.4±9.4	68.2±10.0	68.7±9.3	68.4±9.9	68.2±9.6
White	560(93.6)	428(90.9)	532(91.5)	448(92.9)	542(91.9)	469(92.5)	484(91.7)	437(92.5)	504(93.3)	399(90.9)
BMI (kg/m2)	26.8±3.8	25.8±3.8	26.6±3.9	25.8±3.5	26.9±3.6	26.1±3.9	26.6±3.9	25.8±4.0	26.7±4.0	26.0±3.9
Physical activity (MET-hours/week)	0.6(0.0,3.0)	2.5(0.2,6.0)	0.5(0.0,2.5)	2.5(0.2,6.0)	0.2(0.0,2.5)	2.5(0.2,6.0)	0.4(0.0,2.5)	2.0(0.0,5.7)	0.7(0.0,3.0)	2.0(0.0,5.2)
Current smoker	90(15.0)	25(5.3)	89(15.4)	28(5.9)	83(14.1)	26(5.2)	87(16.4)	14(2.9)	68(12.6)	34(7.7)
Multivitamin use	342(57.2)	217(46.1)	331(56.9)	209(43.4)	355(60.1)	240(47.3)	304(57.5)	196(41.6)	318(58.8)	227(51.8)
Family history of MI	247(41.3)	215(45.6)	234(40.3)	207(42.9)	237(40.1)	215(42.5)	205(38.8)	215(45.5)	205(37.9)	200(45.5)
Hypertension	316(52.8)	258(54.7)	297(51.1)	272(56.4)	327(55.4)	286(56.5)	276(52.2)	260(55.0)	285(52.8)	259(58.9)
Hypercholesterolemia	313(52.4)	266(56.5)	299(51.5)	269(55.9)	309(52.3)	296(58.3)	282(53.5)	267(56.6)	278(51.5)	266(60.6)
Regular aspirin use	325(54.4)	268(56.8)	333(57.3)	247(51.2)	343(58.2)	275(54.3)	298(56.5)	259(54.9)	310(57.4)	237(54.0)
Use of anti-hypertensive drug	236(39.5)	206(43.8)	225(38.7)	221(45.8)	241(40.9)	229(45.1)	208(39.4)	208(44.0)	206(38.2)	207(47.1)
Use of lipid-lowering medication	140(23.4)	124(26.3)	125(21.5)	121(25.1)	144(24.4)	129(25.4)	128(24.2)	109(23.0)	130(24.0)	122(27.8)
Dietary intake
Total energy, kcal/d	2078±634	1879±580	2265±673	1687±467	1715±589	2280±610	2042±649	2030±576	2330±651	1714±534
Alcohol intake, g/day	5.5(0.0,16.1)	2.9(0.0,11.5)	3.7(0.0,13.5)	3.9(0.0,14.0)	2.0(0.0,11.5)	7.3(1.8,15.1)	4.5(0.0,15.3)	2.8(0.0,11.2)	4.5(0.0,14.7)	3.9(0.0,14.1)
Vegetables, servings/d	2.4±1.5	4.5±2.6	2.6±1.6	3.7±2.3	1.9±1.1	5.1±2.6	2.2±1.2	4.7±2.8	2.8±1.6	4.2±2.8
Fruits, servings/d	1.0±0.8	2.6±1.5	1.0±0.9	2.3±1.3	0.9±0.9	2.7±1.8	0.9±0.6	2.9±1.8	1.3±1.1	2.3±1.9
Red and processed meats, servings/d	1.6±0.8	0.5±0.5	1.6±0.8	0.5±0.4	1.3±0.7	0.8±0.6	1.6±0.8	0.6±0.5	1.6±0.8	0.7±0.6
Whole grains, servings/d	1.1±1.1	2.1±1.6	1.1±1.2	2.0±1.4	0.8±0.9	2.4±1.6	0.8±0.9	2.4±1.6	1.1±1.0	2.0±1.7
Refined grains, servings/d	1.8±1.6	1.3±1.4	2.1±1.7	1.0±0.8	1.5±1.3	1.7±1.6	1.9±1.6	1.4±1.4	2.2±1.6	1.1±1.3
Nuts and legumes, servings/d	0.3±0.3	0.9±0.9	0.5±0.8	0.7±0.7	0.2±0.2	1.0±0.9	0.4±0.7	0.8±0.8	0.5±0.6	0.7±0.8

AHEI: Alternate Healthy Eating Index; AMED: Alternate Mediterranean Diet; BMI: Body mass index; DASH: Dietary Approaches to Stop Hypertension; HEI: Healthy Eating Index; hPDI: healthful Plant-based Diet Index; HPFS: Health Professionals Follow-up Study; MET: Metabolic equivalents of task; MI: Myocardial infarction; NHS: Nurses’ Health Study.

^*Values are mean ± SD, median (interquartile range), or n (%). Percentages were calculated among participants with non-missing data. Missing values were <5% for all covariates unless otherwise indicated.

^†Values were not age adjusted.

^‡Menopausal hormone use was based on NHS participants only.

Post-MI dietary scores and mortality

During 92,272 person-years of follow-up, we documented 3,858 deaths, of which 1,639 were due to CVD (Table S4). Each of the dietary scores was associated with a significantly lower all-cause mortality in multivariable analyses (P<0.001 for trend, Table 2 and Figure 1). The pooled HRs comparing extreme quintiles of individual dietary scores after MI were 0.67 (95% confidence interval [CI]: 0.60–0.74) for AHEI, 0.73 (95% CI: 0.65–0.81) for HEI-2015, 0.66 (95% CI: 0.59–0.75) for AMED, 0.77 (95% CI: 0.69–0.86) for DASH, 0.74 (95% CI: 0.66–0.83) for hPDI, 0.77 (95% CI: 0.68–0.86) for rEDIP, 0.73 (95% CI: 0.65–0.83) for rEDIR, and 0.71 (95% CI: 0.64–0.80) for rEDIH. All E values for the observed HRs for total mortality were over 1.92 (Figure 1). For CVD mortality, a significant inverse association was also observed for these dietary scores (Table 3 and Figure 1). Among NHS participants, in whom recurrent non-fatal MI was ascertained, higher post-MI dietary scores were also inversely associated with recurrent non-fatal MI (Table 3 and Figure 1). Kaplan-Meier curves showed differences in survival probabilities across quintiles of post-MI dietary scores (Figure 2 and Figure S5; all log-rank P<0.001).

Table 2.

Hazard ratios (95% confidence intervals) of total mortality among survivors of myocardial infarction during follow up according to quintiles of dietary scores.

Variable	Quintiles of dietary scores					P for trend	Per 1-SD increase†
	Q1	Q2	Q3	Q4	Q5
AHEI
Median score (interquartile range)	40.9 (36.4, 45.1)	49.7 (46.3, 52.9)	55.6 (52.0, 58.8)	61.3 (57.4, 65.0)	70.2 (65.5, 74.7)
No of participants/person years	990/18399	838/18713	736/18296	736/18458	558/18406
Age-adjusted	1.00	0.82 (0.75, 0.90)	0.75 (0.68, 0.82)	0.71 (0.65, 0.78)	0.56 (0.50, 0.62)	<0.001
Multivariable adjusted*	1.00	0.88 (0.80, 0.97)	0.83 (0.75, 0.91)	0.81 (0.73, 0.89)	0.67 (0.60, 0.74)	<0.001	0.85 (0.82, 0.89)
HEI-2015
Median score (interquartile range)	58.1 (53.0, 61.3)	67.5 (65.7, 69.6)	73.9 (72.2, 75.8)	79.5 (77.5, 81.4)	86.5 (84.1, 89.5)
No of participants/person years	940/17654	774/17433	769/17796	709/18519	666/20870
Age-adjusted	1.00	0.82 (0.74, 0.90)	0.83 (0.75, 0.91)	0.72 (0.65, 0.80)	0.61 (0.55, 0.67)	<0.001
Multivariable adjusted*	1.00	0.89 (0.80, 0.98)	0.91 (0.83, 1.01)	0.82 (0.74, 0.91)	0.73 (0.65, 0.81)	<0.001	0.89 (0.86, 0.93)
AMED
Median score (interquartile range)	19.0 (17.0, 21.0)	24.0 (23.0, 24.0)	27.0 (26.0, 28.0)	31.0 (30.0, 31.0)	35.0 (34.0, 37.0)
No of participants/person years	1026/19210	827/17525	854/20338	645/16663	506/18536
Age-adjusted	1.00	0.92 (0.84, 1.01)	0.83 (0.76, 0.91)	0.75 (0.67, 0.82)	0.55 (0.49, 0.61)	<0.001
Multivariable adjusted*	1.00	1.01 (0.92, 1.11)	0.91 (0.83, 1.01)	0.86 (0.77, 0.96)	0.66 (0.59, 0.75)	<0.001	0.86 (0.83, 0.90)
DASH
Median score (interquartile range)	17.0 (16.0, 19.0)	21.0 (20.0, 22.0)	24.0 (23.0, 25.0)	27.0 (26.0, 27.0)	30.0 (29.0, 32.0)
No of participants/person years	779/15938	824/18514	924/20832	682/17734	649/19254
Age-adjusted	1.00	0.90 (0.81, 0.99)	0.83 (0.75, 0.91)	0.73 (0.66, 0.81)	0.62 (0.56, 0.69)	<0.001
Multivariable adjusted*	1.00	0.97 (0.88, 1.07)	0.93 (0.84, 1.03)	0.88 (0.79, 0.98)	0.77 (0.69, 0.86)	<0.001	0.91 (0.87, 0.94)
hPDI
Median score (interquartile range)	45.0 (42.0, 47.0)	51.0 (49.0, 52.0)	55.0 (54.0, 56.0)	59.0 (58.0, 61.0)	65.0 (63.0, 67.0)
No of participants/person years	849/16958	902/19780	734/18267	842/20295	531/16972
Age-adjusted	1.00	0.93 (0.85, 1.03)	0.82 (0.74, 0.91)	0.86 (0.78, 0.94)	0.71 (0.64, 0.79)	<0.001
Multivariable adjusted*	1.00	0.92 (0.83, 1.01)	0.85 (0.76, 0.94)	0.84 (0.76, 0.93)	0.74 (0.66, 0.83)	<0.001	0.91 (0.88, 0.94)
rEDIP
Median score (interquartile range)	−0.34 (−0.48, −0.25)	−0.10 (−0.15, −0.06)	0.02 (0.00, 0.05)	0.17 (0.12, 0.21)	0.40 (0.32, 0.54)
No of participants/person years	954/20532	880/19885	874/18860	652/17745	498/15251
Age-adjusted	1.00	0.84 (0.77, 0.92)	0.87 (0.79, 0.95)	0.69 (0.62, 0.76)	0.67 (0.60, 0.75)	<0.001
Multivariable adjusted*	1.00	0.89 (0.81, 0.98)	0.86 (0.78, 0.95)	0.77 (0.69, 0.85)	0.77 (0.68, 0.86)	<0.001	0.91 (0.88, 0.95)
rEDIR
Median score (interquartile range)	−1.07 (−1.45, −0.80)	−0.61 (−0.79, −0.45)	−0.39 (−0.51, −0.26)	−0.18 (−0.28, −0.10)	0.09 (0.00, 0.23)
No of participants/person years	973/19984	838/19654	792/19961	733/17862	522/14811
Age-adjusted	1.00	0.89 (0.81, 0.98)	0.74 (0.68, 0.82)	0.77 (0.70, 0.85)	0.71 (0.63, 0.79)	<0.001
Multivariable adjusted*	1.00	0.96 (0.87, 1.06)	0.81 (0.73, 0.89)	0.82 (0.74, 0.91)	0.73 (0.65, 0.83)	<0.001	0.93 (0.90, 0.97)
rEDIH
Median score (interquartile range)	−0.73 (−0.90, −0.59)	−0.44 (−0.52, −0.35)	−0.29 (−0.36, −0.22)	−0.16 (−0.21, −0.11)	−0.01 (−0.05, 0.06)
No of participants/person years	759/15994	818/18527	771/19335	819/19687	691/18728
Age-adjusted	1.00	0.87 (0.79, 0.96)	0.75 (0.68, 0.83)	0.74 (0.67, 0.81)	0.65 (0.59, 0.73)	<0.001
Multivariable adjusted*	1.00	0.92 (0.83, 1.02)	0.79 (0.70, 0.87)	0.78 (0.70, 0.87)	0.71 (0.64, 0.80)	<0.001	0.92 (0.89, 0.95)

AHEI: Alternate Healthy Eating Index; AMED: Alternate Mediterranean Diet; CI: confidence interval; CVD: cardiovascular disease; DASH: Dietary Approaches to Stop Hypertension; HEI: Healthy Eating Index; hPDI: healthful Plant-based Diet Index; HR: hazard ratio; MI: myocardial infarction; rEDIH: the reversed empirical dietary index for hyperinsulinemia; rEDIP: the reversed empirical dietary inflammatory pattern; rEDIR: the reversed empirical dietary index for insulin resistance.

^*Multivariable adjusted for age (continuous), sex (men or women), white ethnicity (yes or no), duration of myocardial infarction (years), smoking status (never, former, current 1–14 cigarettes/day, current ≥15 cigarettes/day), physical activity (<3.0, 3.0–8.9, 9.0–17.9, 18.0–26.9, ≥27.0 metabolic equivalents-hours/week), multivitamin use, aspirin use (yes or no), menopausal status and post-menopausal hormone use (pre-menopause, post-menopause [never, former, or current hormone use], or missing; Nurses’ Health Study only), family history of myocardial infarction (yes or no), body mass index at myocardial infarction diagnosis (kg/m²: <21, 21–24.9, 25–29.9, 30–35, ≥35), and intake of total energy (continuous). Quintiles were defined within each cohort and then pooled for analysis.

^†1-SD increment in eight dietary scores: 11.6 points for AHEI, 11.2 points for HEI-2015, 6.0 points for AMED, 4.8 points for DASH, 7.1 points for hPDI, 0.33 points for rEDIP, 0.31 points for rEDIH, and 0.57 points for rEDIR.

Figure 1. Associations of post-MI dietary scores with total mortality, cardiovascular mortality, and recurrent non-fatal myocardial infarction.

AHEI: Alternate Healthy Eating Index; AMED: Alternate Mediterranean Diet; CVD: cardiovascular disease; DASH: Dietary Approaches to Stop Hypertension; HEI: Healthy Eating Index; hPDI: healthful Plant-based Diet Index; MI: myocardial infarction; rEDIH: reversed empirical dietary index for hyperinsulinemia; rEDIP: reversed empirical dietary inflammatory pattern; rEDIR: reversed empirical dietary index for insulin resistance.

The forest plots show the hazard ratios comparing extreme categories of dietary scores (visually represented by the centers of the error bars), the 95% confidence intervals (visually represented by the error bars) and E values and their lower bound. Hazard ratios and 95% confidence intervals are displayed on a logarithmic scale. Hazard ratios (95% confidence intervals) of recurrent non-fatal MI were calculated only in the NHS. Hazard ratios were adjusted for age (continuous), sex (men or women), white ethnicity (yes or no), duration of myocardial infarction (years), smoking status (never, former, current 1–14 cigarettes/day, current ≥15 cigarettes/day), physical activity (<3.0, 3.0–8.9, 9.0–17.9, 18.0–26.9, ≥27.0 metabolic equivalents-hours/week), multivitamin use, aspirin use (yes or no), menopausal status and post-menopausal hormone use (pre-menopause, post-menopause [never, former, or current hormone use], or missing; Nurses’ Health Study only), family history of myocardial infarction (yes or no), body mass index at myocardial infarction diagnosis (kg/m²: <21, 21–24.9, 25–29.9, 30–35, ≥35), and intake of total energy (continuous). Analyses of recurrent non-fatal MI were restricted to Nurses’ Health Study participants because recurrent non-fatal MI was ascertained in the Nurses’ Health Study but was not collected in the Health Professionals Follow-up Study.

Table 3.

HRs (95% CIs) of CVD mortality and recurrent non-fatal MI among survivors of MI during follow up according to quintiles of dietary scores.^*

Variable	Quintiles of dietary scores					P for trend	Per 1-SD increase†
	Q1	Q2	Q3	Q4	Q5
CVD mortality
AHEI	1.00	0.93 (0.80, 1.08)	0.95 (0.81, 1.10)	0.86 (0.74, 1.01)	0.70 (0.59, 0.83)	<0.001	0.90 (0.85, 0.96)
HEI-2015	1.00	0.91 (0.78, 1.06)	0.96 (0.83, 1.13)	0.93 (0.79, 1.08)	0.76 (0.64, 0.90)	0.007	0.93 (0.88, 0.98)
AMED	1.00	0.97 (0.84, 1.12)	0.99 (0.86, 1.14)	0.84 (0.71, 0.99)	0.66 (0.55, 0.80)	<0.001	0.90 (0.84, 0.96)
DASH	1.00	1.01 (0.86, 1.19)	1.00 (0.85, 1.16)	0.97 (0.82, 1.14)	0.76 (0.64, 0.91)	0.003	0.92 (0.87, 0.98)
hPDI	1.00	1.13 (0.97, 1.32)	1.00 (0.85, 1.18)	1.02 (0.87, 1.19)	0.83 (0.69, 0.99)	0.04	0.95 (0.90, 1.01)
rEDIP	1.00	0.88 (0.76, 1.02)	0.89 (0.77, 1.04)	0.75 (0.64, 0.88)	0.76 (0.63, 0.91)	<0.001	0.90 (0.85, 0.95)
rEDIR	1.00	1.01 (0.87, 1.17)	0.88 (0.75, 1.02)	0.84 (0.72, 0.99)	0.69 (0.57, 0.84)	<0.001	0.93 (0.89, 0.98)
rEDIH	1.00	1.01 (0.86, 1.19)	0.84 (0.71, 0.99)	0.84 (0.71, 1.00)	0.72 (0.60, 0.86)	<0.001	0.91 (0.87, 0.95)
Recurrent non-fatal MI ‡
AHEI	1.00	0.90 (0.69, 1.16)	0.83 (0.64, 1.09)	0.85 (0.65, 1.12)	0.58 (0.43, 0.79)	0.001	0.81 (0.73, 0.89)
HEI-2015	1.00	0.69 (0.52, 0.91)	0.87 (0.67, 1.14)	0.70 (0.53, 0.92)	0.59 (0.44, 0.79)	0.001	0.84 (0.76, 0.92)
AMED	1.00	0.88 (0.68, 1.14)	1.00 (0.77, 1.30)	0.86 (0.65, 1.15)	0.56 (0.39, 0.79)	0.008	0.85 (0.77, 0.95)
DASH	1.00	0.87 (0.66, 1.14)	0.95 (0.73, 1.23)	0.55 (0.40, 0.76)	0.60 (0.44, 0.81)	<0.001	0.82 (0.75, 0.90)
hPDI	1.00	0.80 (0.61, 1.04)	0.76 (0.57, 1.02)	0.69 (0.52, 0.92)	0.61 (0.45, 0.83)	0.001	0.83 (0.76, 0.92)
rEDIP	1.00	0.88 (0.69, 1.13)	0.95 (0.73, 1.22)	0.62 (0.46, 0.84)	0.62 (0.44, 0.87)	<0.001	0.86 (0.77, 0.95)
rEDIR	1.00	0.70 (0.54, 0.92)	0.77 (0.59, 1.01)	0.67 (0.50, 0.89)	0.37 (0.25, 0.55)	<0.001	0.79 (0.68, 0.90)
rEDIH	1.00	1.04 (0.78, 1.38)	0.78 (0.58, 1.06)	0.85 (0.63, 1.16)	0.51 (0.35, 0.72)	<0.001	0.87 (0.81, 0.93)

AHEI: Alternate Healthy Eating Index; AMED: Alternate Mediterranean Diet; CI: confidence interval; CVD: cardiovascular disease; DASH: Dietary Approaches to Stop Hypertension; HEI: Healthy Eating Index; hPDI: healthful Plant-based Diet Index; HR: hazard ratio; MI: myocardial infarction; rEDIH: the reversed empirical dietary index for hyperinsulinemia; rEDIP: the reversed empirical dietary inflammatory pattern; rEDIR: the reversed empirical dietary index for insulin resistance. SD: standard deviation.

^*Multivariable adjusted for age (continuous), sex (men or women), white ethnicity (yes or no), duration of myocardial infarction (years), smoking status (never, former, current 1–14 cigarettes/day, current ≥15 cigarettes/day), physical activity (<3.0, 3.0–8.9, 9.0–17.9, 18.0–26.9, ≥27.0 metabolic equivalents-hours/week), multivitamin use, aspirin use (yes or no), menopausal status and post-menopausal hormone use (pre-menopause, post-menopause [never, former, or current hormone use], or missing; Nurses’ Health Study only), family history of myocardial infarction (yes or no), body mass index at myocardial infarction diagnosis (kg/m²: <21, 21–24.9, 25–29.9, 30–35, ≥35), and intake of total energy (continuous).

^†1-SD increment in eight dietary scores: 11.6 points for AHEI, 11.2 points for HEI-2015, 6.0 points for AMED, 4.8 points for DASH, 7.1 points for hPDI, 0.33 points for rEDIP, 0.31 points for rEDIH, and 0.57 points for rEDIR.

^‡Analyses of recurrent non-fatal MI were restricted to Nurses’ Health Study participants because recurrent non-fatal MI was ascertained in the Nurses’ Health Study but was not collected in the Health Professionals Follow-up Study.

Figure 2. Survival probability according to quintiles of well-established dietary scores after myocardial infarction.

AHEI: Alternate Healthy Eating Index; AMED: Alternate Mediterranean Diet; DASH: Dietary Approaches to Stop Hypertension; HEI: Healthy Eating Index; hPDI: healthful Plant-based Diet Index

Kaplan-Meier curves were used to estimate the association between the dietary scores and survival probability among survivors of myocardial infarction. Adjusted for age (continuous), sex (men or women), white ethnicity (yes or no), duration of myocardial infarction (years), smoking status (never, former, current 1–14 cigarettes/day, current ≥15 cigarettes/day), physical activity (<3.0, 3.0–8.9, 9.0–17.9, 18.0–26.9, ≥27.0 metabolic equivalents-hours/week), multivitamin use, aspirin use (yes or no), menopausal status and post-menopausal hormone use (pre-menopause, post-menopause [never, former, or current hormone use], or missing; Nurses’ Health Study only), family history of myocardial infarction (yes or no), body mass index at myocardial infarction diagnosis (kg/m²: <21, 21–24.9, 25–29.9, 30–35, ≥35), and intake of total energy (continuous).

Restricted cubic spline analyses showed that the relationship was linear for the dietary scores after MI (Figure 3 and Figure S6). Per one-standard deviation (SD) increment in each dietary score after MI was associated with a reduction of 7% to 15% in total mortality (Table 2). Significant inverse associations with the CVD mortality and recurrent non-fatal MI were also observed for the increase in most dietary scores after MI (Table 3).

Figure 3. Dose-response associations between well-established dietary scores after myocardial infarction and total mortality.

AHEI: Alternate Healthy Eating Index; AMED: Alternate Mediterranean Diet; DASH: Dietary Approaches to Stop Hypertension; HEI: Healthy Eating Index; hPDI: healthful Plant-based Diet Index.

Restricted cubic spline models were used to evaluate the association between the dietary scores and mortality among survivors of myocardial infarction. Adjusted for age (continuous), sex (men or women), white ethnicity (yes or no), duration of myocardial infarction (years), smoking status (never, former, current 1–14 cigarettes/day, current ≥15 cigarettes/day), physical activity (<3.0, 3.0–8.9, 9.0–17.9, 18.0–26.9, ≥27.0 metabolic equivalents-hours/week), multivitamin use, aspirin use (yes or no), menopausal status and post-menopausal hormone use (pre-menopause, post-menopause [never, former, or current hormone use], or missing; Nurses’ Health Study only), family history of myocardial infarction (yes or no), body mass index at myocardial infarction diagnosis (kg/m²: <21, 21–24.9, 25–29.9, 30–35, ≥35), and intake of total energy (continuous).

Changes in dietary scores and mortality

The dietary scores before and after MI were modestly correlated (Spearman correlation coefficients from 0.57 to 0.71). As compared with participants who had relatively stable AHEI over time, those with the improvement in AHEI after MI had a significant 14% (95% CI, 3%−24%) lower risk of total mortality (Figure 4 and Table S5). Increased adherence to other scores was non-significantly associated with lower mortality. In contrast, a decrease in all these dietary scores, as compared with no change in this dietary score, was associated with increased total mortality. For AHEI, HEI-2015, AMED, DASH, hPDI, and rEDIH, the associations achieved statistical significance with HRs ranging from 1.18 for rEDIH to 1.45 for AHEI. A similar pattern of associations was observed for CVD mortality and recurrent non-fatal MI.

Figure 4. Associations of changes in well-established dietary scores from before to after myocardial infarction diagnosis with total mortality, cardiovascular mortality, and recurrent non-fatal myocardial infarction.

AHEI: Alternate Healthy Eating Index; AMED: Alternate Mediterranean Diet; CVD: cardiovascular disease; DASH: Dietary Approaches to Stop Hypertension; HEI: Healthy Eating Index; hPDI: healthful Plant-based Diet Index; MI: myocardial infarction.

Hazard ratios for total mortality and CVD mortality comparing extreme categories of changes in dietary scores from before to after myocardial infarction diagnosis were adjusted for age (continuous), sex (men or women), white ethnicity (yes or no), duration of myocardial infarction (years), smoking status (never, former, current 1–14 cigarettes/day, current ≥15 cigarettes/day), physical activity (<3.0, 3.0–8.9, 9.0–17.9, 18.0–26.9, ≥27.0 metabolic equivalents-hours/week), multivitamin use, aspirin use (yes or no), menopausal status and post-menopausal hormone use (pre-menopause, post-menopause [never, former, or current hormone use], or missing; Nurses’ Health Study only), family history of myocardial infarction (yes or no), body mass index at myocardial infarction diagnosis (kg/m²: <21, 21–24.9, 25–29.9, 30–35, ≥35), and intake of total energy (continuous). Decreased adherence, no substantial change, and increased adherence were defined using score-specific cutoffs based on the distribution of changes in dietary scores from before to after MI diagnosis. Analyses of recurrent non-fatal MI were restricted to Nurses’ Health Study participants because recurrent non-fatal MI was ascertained in the Nurses’ Health Study but was not collected in the Health Professionals Follow-up Study.

Sensitivity analyses

In sensitivity analyses, the significant inverse associations between post-MI dietary scores and total mortality were not materially changed when we excluded participants with a history of MI at baseline, or when we excluded deaths occurring during the first four years after MI diagnosis (Table S6). Further adjustment for socioeconomic status or the diagnoses of hypertension, hypercholesterolemia, and diabetes in the multivariable analyses did not materially alter the associations. The results also remained virtually unchanged when we continuously updated dietary information even after diagnosis of chronic diseases. Also, the associations were largely similar to the results from primary analyses when we restricted our analyses to never smokers. Results were also similar after excluding participants in the lowest fifth of physical activity or using a two-year lag for physical activity. None of the dietary scores after MI was significantly associated with the risk of wrist fracture. In addition, assigning positive scores individually to certain healthy animal foods in hPDI did not appreciably alter the HR (0.72; 95% CI: 0.64–0.80). In competing-risk sensitivity analyses, the inverse associations of post-MI dietary scores with CVD mortality were generally similar after treating non-CVD deaths as competing events. Among NHS participants, the associations with recurrent non-fatal MI were also largely consistent after treating deaths before recurrent non-fatal MI as competing events (Table S7).

Discussion

In two large prospective cohorts, greater adherence to various healthy dietary patterns after MI was consistently associated with better survival among individuals who had experienced a non-fatal MI. Each one-SD increase in these dietary scores after MI was significantly associated with a 7% to 15% reduction in mortality. In addition, deteriorated adherence to these healthy patterns from before to after MI diagnosis was particularly associated with elevated mortality.

Existing evidence from prospective observational studies and clinical trials has shown that adherence to different healthy dietary patterns is associated with lower risks of CVD and mortality for general populations.^9,13,33–35 Shan et al. reported that a significant inverse association was observed between higher adherence to four healthy eating patterns (HEI-2015, AHEI, AMED, and hPDI) and risk of CVD in comparison with poorer adherence.⁹ Results from a meta-analysis of nine prospective cohort studies also indicated that a higher overall plant-based diet index, especially hPDI scores, was associated with a reduced CVD risk.³³ In an analysis in the Physicians' Health Study, greater adherence to AHEI, MED, and DASH scores was each inversely associated with mortality from all causes.³⁴ Sotos-Prieto et al. found that improvement in the scores of AHEI, MED, and DASH over time was consistently associated with a decreased risk of CVD and deaths.^36,37

In contrast, existing studies of dietary patterns in relation to secondary prevention of MI were based mostly on the Mediterranean diet.^38–40 The Lyon Diet Heart Study examined the effect of a Mediterranean diet compared to usual diet in 605 CHD patients over 46 months and found that assignment to a Mediterranean diet reduced the risk of deaths caused by CVD and all causes by 65% and 56%, respectively.³⁸ In The CORDIOPREV study evaluating the effects of a Mediterranean diet versus a low-fat diet over seven years of follow-up in 1,002 patients with CHD, the Mediterranean diet was superior to the low-fat diet in preventing a major cardiovascular event, with a decrease of HR of 26%.⁴¹ These findings were also consistent with results from a recent meta-analysis of seven prospective cohort studies in which the risk of all-cause and CVD mortality decreased by 15% and 9% for each 2-unit increment in a score of adherence to Mediterranean diet among people with a history of CVD.⁴¹ In a previous analysis in the NHS and HPFS, Li et al. found that the AHEI score was inversely associated with 24% lower all-cause mortality and 26% lower CVD mortality.¹² In the present study, we generated novel evidence through considering other dietary patterns with larger sample size and longer disease follow-up in the two large cohorts. Collectively, evidence from the current and previous studies clearly suggested that adherence to healthy dietary patterns is not only important for the primary prevention of CHD but also the secondary prevention of deaths among MI survivors. In addition, our findings also demonstrated that increased adherence to the healthy diet patterns may be associated with more favorable long-term prognosis among MI survivors.

The consistency among associations of various dietary patterns with mortality may be ascribed to the fact that the five prevailing recommended healthy dietary patterns all emphasize a common core group of healthful foods, such as high intake of fruits and vegetables, whole grains, and nuts, moderate intake of alcohol, and low intake of refined grains and red/processed meats, which have long been suggested as healthy choices that may be effective for CVD secondary prevention.¹⁴ Meanwhile, the dietary patterns were developed for different purposes and vary by different emphases of foods or nutrients, which are worth discussing. AHEI specifically emphasizes higher long-chain omega-3 fatty acids, polyunsaturated fat intake, lower trans fat and sugar-sweetened beverages/added sugars. Previous evidence has indicated that the type of fat consumed appeared to be more important than the total fat intake to achieve positive health outcomes.⁴² Dietary guidelines generally recommend increasing omega-3 polyunsaturated fatty acids and monounsaturated fatty acids intake and reducing the intake of saturated fat and trans fat to facilitate primary and secondary prevention of chronic diseases.^14,43 The HEI-2015 prioritizes legumes, total dairy, the unsaturated/saturated fat ratio and de-emphasizes refined grains.⁴⁴ The AMED emphasizes consumption of olive oil, fish, legumes and nuts, while the DASH diet focuses on intake of fruits, vegetables, low fat dairy foods, and reduced saturated and total fat.^40,45 In contrast to above dietary patterns, hPDI mainly emphasizes the quality of foods derived from plant sources and de-emphasizes any animal foods. Although each dietary pattern represents a unique combination of dietary components, our findings suggest that the differences among these dietary patterns are unlikely major determinants of the associations, and rather the common food groups are the cornerstones of healthy diets.

Unlike the five established dietary patterns, the three empirical indices were not based on the consideration of the quality of individual food groups. Instead, they were derived based on their predictivity of causal cardiometabolic disease risk markers. Of note, the three empirically-derived dietary patterns were developed based on finer classification of food groups and thus consisted of food groups that are not particularly emphasized or de-emphasized by the five established dietary patterns.²⁶ Nevertheless, these patterns emphasize vegetables and coffee among other healthy choices of foods and universally de-emphasize red and processed meats and beverages and foods rich in refined carbohydrates and added sugar.⁴⁶ Meanwhile, future studies are warranted to understand how these particular combinations of food items, despite some deviations from the established dietary patterns, are still robustly associated with the better health outcomes.^47,48

The biological mechanisms underlying the inverse association between healthy dietary patterns and mortality among MI survivors may involve improvement of glucose and insulin metabolism or dyslipidemia and alleviation of chronic inflammation. Previous studies have shown that adherence to healthy dietary patterns, such as the AHEI, AMED, and DASH, was inversely associated with biomarkers of inflammation (e.g., CRP, IL-6) and insulin response (e.g., fasting insulin).^49,50 The rEDIP, EDIR and rEDIH scores represent dietary patterns empirically constructed, based on foods that are either positively or inversely associated with biomarkers of inflammation, hyperinsulinemia, and insulin resistance.^26,48 Our findings on these dietary patterns and survival among MI survivors were consistent with our initial hypothesis that pro-inflammatory and insulin-modulating diets increase the risk of mortality in MI survivors by promoting a physiologic milieu favorable to disease progression. The weighting of foods by a measured biological response may also account for differential measurement errors among specific foods.

The strengths of this study include the prospective design, large sample size of the cohort, long-term follow-up with a high retention rate, repeated assessments of diet and lifestyle both before MI and after MI, and use of multiple diet-quality scores. However, potential limitations also need to be considered. First, given the observational nature of the present study, we cannot completely exclude the possibility that unmeasured or residual confounding that might affect the observed association, although we have carefully adjusted for several potential confounding variables, and E value analyses indicated that our results were robust to potential unmeasured confounders. Nonetheless, our findings should be interpreted with caution and warrant replication in intervention studies to establish cause-effect relationships. Second, because dietary information was self-reported and assessed by FFQs, measurement error and misclassification were inevitable. However, the FFQs used in our cohorts have been extensively validated against diet records with reasonable reproducibility and validity. Furthermore, the use of the repeated measures of diet during the follow-up to calculate cumulative averages not only reflected long-term habitual intake but also reduced the influences of measurement errors caused by within-person variation. In addition, due to the prospective design, any measurement errors and random misclassifications was more likely to be non-differential and therefore would attenuate true associations toward the null. Third, the differences in food components and scoring algorithms among dietary patterns may introduce some heterogeneity influencing the strength of associations. The wider range and more variabilities of AHEI and hPDI may lead to better characterization of dose-response relationships than DASH. Lastly, the study population consisted predominantly of White health professionals, which may limit the generalizability of the findings to more racially, ethnically, socioeconomically, and culturally diverse populations.

In conclusion, our results from these two prospective cohort studies consistently showed inverse associations of post-MI adherence to healthy dietary patterns with total and CVD mortality among MI survivors. In addition, improvement in adherence to these dietary patterns over time also tended to lead to improved long-term survival among the MI survivors. Overall, our findings provide evidence that adopting healthy dietary patterns explicitly emphasizing diet quality confers benefits for secondary prevention of deaths among MI survivors. The data also support the current American Heart Association dietary recommendations that focus on key healthy foods/nutrients as the cornerstone for a possibly much wider spectrum of dietary practices that incorporate personal preferences for improving cardiovascular health.

Supplementary Material

Supplemental Methods

Tables S1–S6

Figures S1–S6

Clinical Perspective.

What Is New?

• Adherence to various healthy dietary patterns after myocardial infarction (MI) was consistently associated with lower mortality among MI survivors.

• Deteriorated adherence to these healthy patterns from before to after MI diagnosis was associated with an elevated mortality.

What Are the Clinical Implications?

• Adopting healthy dietary patterns explicitly emphasizing diet quality confers benefits for the secondary prevention of deaths among MI survivors.

• The data imply that MI survivors can practice any dietary patterns that adequately emphasize the diet quality to achieve health benefits based on personal preferences, cultural traditions, and budgetary considerations.

Non-standard Abbreviations and Acronym

AHEI

Alternate Healthy Eating Index

AMED

Alternate Mediterranean Diet Score

BMI

Body mass index

CI

confidence interval

CRP

C-reactive protein

CVD

cardiovascular disease

DASH

Dietary Approaches to Stop Hypertension

FFQ

food frequency questionnaire

HEI-2015

the Healthy Eating Index-2015

hPDI

healthful Plant-Based Diet Index

HPFS

Health Professionals Follow-up Study

HR

Hazard ratios

MI

myocardial infarction

NHS

Nurses’ Health Study

rEDIH

the reversed empirical dietary index for hyperinsulinemia

rEDIP

the reversed empirical dietary inflammatory pattern

rEDIR

the reversed empirical dietary index for insulin resistance

SD

standard deviation