Relationship Between a Plant‐Based Dietary Portfolio and Risk of Cardiovascular Disease: Findings From the Women's Health Initiative Prospective Cohort Study

Background The plant‐based Dietary Portfolio combines established cholesterol‐lowering foods (plant protein, nuts, viscous fiber, and phytosterols), plus monounsaturated fat, and has been shown to improve low‐density lipoprotein cholesterol and other cardiovascular disease (CVD) risk factors. No studies have evaluated the relation of the Dietary Portfolio with incident CVD events. Methods and Results We followed 123 330 postmenopausal women initially free of CVD in the Women's Health Initiative from 1993 through 2017. We used Cox proportional‐hazard models to estimate adjusted hazard ratios (HRs) and 95% CI of the association of adherence to a Portfolio Diet score with CVD outcomes. Primary outcomes were total CVD, coronary heart disease, and stroke. Secondary outcomes were heart failure and atrial fibrillation. Over a mean follow‐up of 15.3 years, 13 365 total CVD, 5640 coronary heart disease, 4440 strokes, 1907 heart failure, and 929 atrial fibrillation events occurred. After multiple adjustments, adherence to the Portfolio Diet score was associated with lower risk of total CVD (HR, 0.89; 95% CI, 0.83–0.94), coronary heart disease (HR, 0.86; 95% CI, 0.78–0.95), and heart failure (HR, 0.83; 95% CI, 0.71–0.99), comparing the highest to lowest quartile of adherence. There was no association with stroke (HR, 0.97; 95% CI, 0.87–1.08) or atrial fibrillation (HR, 1.10; 95% CI, 0.87–1.38). These results remained statistically significant after several sensitivity analyses. Conclusions In this prospective cohort of postmenopausal women in the United States, higher adherence to the Portfolio Diet was associated with a reduction in incident cardiovascular and coronary events, as well as heart failure. These findings warrant further investigation in other populations.

enriched margarine). An extension of the diet includes adding monounsaturated fats (MUFAs; such as olive/ canola oil or avocado). 6 Early findings from a metabolically controlled randomized trial showed that the LDL-C lowering effect of the Portfolio Diet was similar to the control diet taken with 20mg lovastatin (−28.6% versus −30.9%). 3 Recently, a systematic review and meta-analysis of metabolically controlled and ad libitum trials showed that the Portfolio Diet significantly lowered LDL-C by 17% (27% in the intended combination with an National Cholesterol Education Program Step II diet). It also lowered other cardiovascular disease (CVD) risk factors, including the alternate blood lipid targets of non-high-density lipoprotein cholesterol by 14% and ApoB (apolipoprotein B) by 15%, and CRP (C-reactive protein) by 32%. 8 These benefits have been recognized by CVD and diabetes mellitus clinical practice guidelines internationally, including those of the Canadian Cardiovascular Society, 9 Diabetes Canada, 10 European Atherosclerosis Society, 11 and Heart UK. 12 Currently, it is not known if these beneficial effects of the diet translate into lower risk of clinical CVD events. The individual components of the Portfolio Diet have been found to be associated with lower incidence of CVD events in prospective cohorts, [13][14][15][16][17] and 2 components of the diet (nuts and extra virgin olive oil) were shown to reduce major vascular events in the landmark PREDIMED (Prevención con Dieta Mediterránea) trial compared to a low saturated fat 18 however, the additive/combined effects of the Portfolio Diet components have not been assessed with incident CVD. Although conducting a long-term randomized trial with CVD as the primary outcome would be preferable, this type of trial is not yet feasible. Analyses of established observational studies may be helpful in assessing the long-term effectiveness of the Portfolio Diet. We have therefore developed a scoring system to measure adherence to the Portfolio Diet for use in these study designs. Here, for the first time, we have evaluated the association of a Portfolio Diet score with CVD outcomes in the WHI (Women's Health Initiative).

Study Population and Design
The design and methods of the WHI have been published elsewhere. [19][20][21] Briefly, between 1993 and 1998, postmenopausal women aged 50 to 79 years were recruited into clinical trials or an observational study (OS) (n=161 808). Recruitment and baseline data collection have been previously reported. 20 This analysis includes follow-up through February 28, 2017. We excluded participants who had a history of CVD at baseline (n=32 594), and missing information regarding diet and lifestyle covariates or implausible caloric intake (<600 kcal or >5000 kcal/day) (n=5884). The final analysis included 123 330 women ( Figure 1). The baseline characteristics of participants included or excluded due to missing data from the analysis are shown in Table S1. Written informed consent was obtained from all WHI participants and procedures were approved by institutional review boards at all participating institutions. The WHI data are accessible to qualified researchers trained in human subject confidentiality protocols and requests to access the data set may be sent to the WHI Publications and Presentations Committee.

Dietary Assessment
The exposure was diet as measured by a Portfolio Diet score. The foods and nutrients composing this score were self-reported using the food frequency questionnaire (FFQ) developed and validated for the WHI 22,23 at enrollment and again at year 3 for the OS participants. No further diet assessments were available for the WHI participants. We used a cumulative average score for those who completed the FFQ at baseline and year 3 (Data S1).

CLINICAL PERSPECTIVE
What Is New?
• Higher adherence to the Portfolio Diet was associated with a 11%, 14%, and 17% lower risk of total cardiovascular disease, coronary heart disease, and heart failure, respectively, but no association was seen with stroke or atrial fibrillation. • This study shows that the beneficial effects of the Portfolio Diet on cardiovascular risk factors from the clinical trials may translate into lower hard clinical cardiovascular disease events.
What Are the Clinical Implications?
• Given the increased interest in plant-based foods and diets around the world, and growing concerns related to ethical and environmental implications of diet, the Portfolio Diet warrants attention from healthcare professionals as another therapeutic dietary approach for cardiovascular disease risk reduction. Food items on the WHI FFQ that are characteristic of the Portfolio Diet were categorized into 6 components (plant protein, nuts, viscous fiber, phytosterols, MUFAs, and saturated fat/cholesterol sources). Intake was assessed as servings/day of targeted foods in all components except phytosterols, which used all FFQ food items to derive total daily intake (mg/day). For the 6 components, each was scored from 1 (unhealthy) to 5 (healthiest) according to participant's quintile of intake resulting in a score range between 6 and 30, with higher scores indicating higher adherence to the Portfolio Diet. Additional information on the Portfolio Diet score development is provided in Data S1 and Table S2.

Ascertainment of CVD Outcomes
Our primary outcomes included total CVD, coronary heart disease (CHD; defined as clinical myocardial infarction, definite silent myocardial infarction, or a death due to definite CHD or possible CHD), and stroke incidence and death as these CVDs are causally related to high LDL-C and the Portfolio Diet has an established cholesterol-lowering effect. 8 Total CVD was a composite of nonfatal myocardial infarction, CHD death, stroke, coronary revascularization and incident heart failure (HF). 24 Our secondary, or exploratory, outcomes included HF and atrial fibrillation (AF). The outcomes were ascertained in the WHI through self-reported medical questionnaires completed by participants every 6 to 12 months, depending on study assignment. Medical records and death certificates for all outcomes were reviewed by central physician adjudicators or trained local adjudicators. 25

Covariates
Covariates that were included in our models were based on information on the participants' lifestyle and risk factors for CVD assessed at baseline, including age, region in the United States, race/ethnicity, alcohol intake, physical activity, caloric intake, sodium intake, hysterectomy history, body mass index (BMI), hormone therapy use, personal history of hypertension and high cholesterol, family history of CVD and diabetes mellitus, diabetes mellitus or cancer diagnoses, smoking status, education, marital status, and clinical trial/study arm. Detailed descriptions of the validity and reproducibility of baseline measurements have been previously published. 21 Statistical Analysis Baseline characteristics were described by quartile of the Portfolio Diet score using means with SDs for continuous variables and frequencies with percentages for categorical variables. To compare baseline characteristics, χ 2 tests were used for categorical variables and analysis of variance for continuous variables.
Participants were categorized into quartiles of the Portfolio Diet score, with the lowest quartile serving as the reference group, as per our prespecified analysis plans. Cox proportional hazard models were used to estimate hazard ratios (HRs) and 95% CIs for the association between the Portfolio Diet score quartiles and CVD outcomes. Two multivariable models were used. Covariates commonly examined in studies of dietary pattern scores and CVD risk were included based on our a priori analysis plan. Model 1 was adjusted for age (continuous), region (Northeast, South, Midwest, West), smoking (never, past, current), and study arm (hormone replacement therapy arm, dietary modification arm, calcium and vitamin D) arm). Model 2 was adjusted for model 1+race/ethnicity (White, Black, Hispanic, Asian/Pacific Islander, Other [American Indian, Alaskan Native, other]), education (college or above, below college), marital status (presently married/other), hysterectomy history (yes/no), body mass index (continuous), physical activity (continuous), alcohol intake (>7 drinks/week, <7 drinks/ week), energy intake (continuous), cancer status (yes/no), hypertension status (yes/no), diabetes mellitus status (yes/no), sodium intake (continuous), family history of CVD (yes/no), family history of diabetes mellitus (yes/no), postmenopausal hormone use (never, past, current), and cholesterol-lowering medication use (yes/no). For all covariates, 5% or less of values were missing. When we checked the proportional hazard model assumptions using Schoenfeld residuals method, no violations of the assumption were found.
Tests for linear trend were conducted by assigning the median value to each quartile. Our main analysis (per our protocol) included all WHI participants (clinical trials+OS). We also conducted several sensitivity analyses to test the robustness of our main findings. First, we conducted analyses by restricting the data to the OS participants only as the clinical trials participants have received an intervention and may be different from the OS participants. We also then (1) restricted analyses to the baseline diet only, (2) excluded participants from the dietary modification trial (a low fat diet intervention), as their diet may have changed overtime, (3) excluded CVD events within the first 3 years of follow-up to address possible reverse causation, (4) excluded those with diabetes mellitus at baseline due to their higher CVD risk, and (5) completed multiple imputation for missing covariate data (using the multivariate imputation by chained equations method). 26 We also conducted post hoc sensitivity analyses where we created another Portfolio Diet score based on the recommendations from the Portfolio Diet randomized clinical trials (further details included in Table S3). We then also applied subgroup analyses according to several potential interactive factors (age, body mass index, family history of CVD, race/ethnicity, smoking status, and cholesterol-lowering medication) and conducted interaction tests via multiplicative interaction terms using model 2 to assess if the P for interactions were significant. Additional analyses we conducted included evaluating associations between the 6 individual components of the Portfolio Diet and risk of the CVD outcomes. Statistical tests were 2-sided and P<0.05 was considered statistically significant. The statistical analyses were conducted with Stata statistical software (Stata Statistical Software: Release 15., Stata Corp., College Station, TX). Further information on the methods can be found in Data S1.

Lifestyle Characteristics of the Participants
Baseline characteristics by quartiles of the Portfolio Diet score are shown in Table 1. Women with higher scores tended to be older, have a lower body mass index, engage in more physical activity, have a higher education, be less likely to smoke, as well as several other differences. All of these known risk factors at baseline were adjusted for in our analyses. Mean intake of the Portfolio Diet score components is shown in

Portfolio Diet Score and CVD Outcomes
During an average of 15.3 years of follow-up, we documented 13 365 incident CVD cases, including 5640 CHD cases, 4400 stroke cases, 1907 HF cases, and 929 AF cases. After adjusting for potential confounders, we observed that women in the top quartile (Q4) of the Portfolio Diet score, compared to those in the bottom quartile (Q1), had an HR of 0.89 (95% CI, 0.83-0.94; P<0.001 for trend) for risk of total CVD, 0.86 (95% CI, 0.78-0.95; P<0.001 for trend) for risk of CHD, and 0.97 (95% CI, 0.87-1.08; P=0.50 for trend) for stroke (Table 3 and Figure 2). For our exploratory outcomes, we observed that women in the top quartile compared   (Table 3 and Figure 2). Absolute incidence rates per 100 000 person-years among quartiles of adherence are shown in Tables 3 and 4.

Sensitivity Analyses
The associations between the Portfolio Diet score and CVD outcomes remained similar in all sensitivity analyses (in the OS participants only [ Table 4], and baseline diet only, excluding participants from the dietary modification trial, excluding CVD events within the first 3 years of follow-up, excluding those with diabetes mellitus, and completing multiple imputation for missing covariate data [ Table S4]). For HF, however, after excluding events diagnosed in the first 3 years, the association was slightly attenuated and no longer significant (Table S4). The association between the Portfolio Diet score based on the randomized clinical trials recommendations and CVD outcomes were attenuated and no longer significant for some outcomes; however, patterns were similar to our original a priori analysis (Table S3).

Subgroup Analyses
The results remained largely consistent in each of the subgroup analyses, apart from effect modification by smoking status and CHD (Figures S1 through S5).

Individual Component Analyses
When we individually assessed the 6 components of the Portfolio Diet score with the CVD outcomes,   Table S1 in the Supplementary Appendix. § Two points not given to any participants based on consumption on MUFAs (low in entire population). ‖ Higher quintiles represent higher intake; however, high intake and high quintiles of saturated fat/cholesterol received lower scores.
higher intakes of nuts, phytosterols, and MUFAs and lower intake of saturated fat sources had inverse associations with total CVD. Phytosterols and low saturated fat sources had inverse associations with CHD and phytosterols had inverse associations with stroke. Nuts also had an inverse association with HF (Table S5).

DISCUSSION
In this large prospective cohort study of US postmenopausal women, a higher Portfolio Diet score was associated with a 11% and 14% lower risk of total CVD and CHD, respectively, but no association was seen with stroke. These findings remained consistent across all Quartile 1 represents the least adherent to the Portfolio Diet, whereas quartile 4 represents the most adherence to the Portfolio Diet. Associations between Portfolio Diet and outcomes were determined by Cox proportional hazard models. Under/over energy reporters and those with baseline CVD were excluded from the analysis. Total CVD is a composite of incidence and death of CHD, stroke, heart failure, and coronary revascularization (coronary artery bypass grafting or percutaneous transluminal coronary angioplasty). CHD indicates coronary heart disease; CT, clinical trial; CVD, cardiovascular disease; HR, hazard ratio; OS, observational study; and Q, quartile. *Model 1 adjusted for age (continuous), region (Northeast, South, Midwest, West), smoking (never, past, current), and study arm (hormone replacement therapy arm, dietary modification arm, calcium and vitamin D arm). † Model 2 adjusted for model 1+race/ethnicity (White, Black, Hispanic, Asian/Pacific Islander, Other [American Indian, Alaskan Native, other]), education (college or above, below college), marital status (presently married/other), hysterectomy history (yes/no), body mass index (continuous), physical activity (continuous), alcohol intake (>7 drinks/week, <7 drinks/week), energy intake (continuous), cancer status (yes/no), hypertension status (yes/no), diabetes mellitus status (yes/no), sodium intake (continuous), family history of CVD (yes/no), family history of diabetes mellitus (yes/no), hormone therapy use (never, past, current), cholesterol-lowering medication use (yes/no). sensitivity analyses, including when we excluded the WHI clinical trial participants, highlighting the robustness of our results. There was also a strong linear trend for greater adherence to the Portfolio Diet with total CVD and CHD. For our secondary analysis, there was an association of a 17% lower risk of HF with a higher Portfolio Diet score, but no association was seen with AF. The true benefits of the Portfolio Diet on CVD risk reduction, however, are likely underestimated in the current study.

Interpretation of Results and Implications
These findings are consistent with the Portfolio Diet trial evidence assessing effects on intermediate risk factors for CVD. The Portfolio Diet has been shown to result in clinically meaningful reductions in the lipid targets for CVD prevention (LDL-C, non-high-density lipoprotein cholesterol, ApoB), as well as CRP, with smaller reductions in blood pressure. 8 In particular, LDL-C, the primary risk factor that the Portfolio Diet was designed to reduce, is considered causal in the pathogenesis of atherosclerotic CVD based on evidence from cardiovascular outcomes trials involving 3 different classes of drugs (statins, ezetimibe, and PCSK9 inhibitors), Mendelian randomization studies and prospective cohorts. 27 Our strongest finding of a 14% inverse association with CHD is consistent with these lines of evidence and closely reflects the predicted 10-year CHD risk reduction of 13% estimated in our systematic review and meta-analysis of the Portfolio Diet trials. 8 The 0.73 mmol/L reduction in LDL-C that corresponds to this 13% reduction in the Portfolio Diet trials Figure 2. Summary of findings of incident cardiovascular disease, coronary heart disease, stroke, heart failure, and atrial fibrillation comparing low to high adherence to the Portfolio Diet. Hazard ratios (HRs) and for comparing participants in Q1 (low adherence [reference category]) to Q4 (high adherence) of the Portfolio Diet with CVD outcomes in the Women's Health Initiative (Clinical Trials+Observational Study). Multivariate-adjusted models were adjusted for the following: age, region, smoking, clinical trial study arm, ethnicity, education, marital status, hysterectomy history, body mass index, physical activity, alcohol intake, energy intake, cancer status, hypertension status, diabetes mellitus status, sodium intake, family history of CVD, family history of diabetes mellitus, hormone therapy use, and cholesterol-lowering medication use. P trend was determined by assigning a median value to each quartile. Horizontal lines represent 95% CIs. CHD indicates coronary heart disease; and CVD, cardiovascular disease.
is predicted by the regression line for the observed risk reduction per mmol/L of LDL-C seen within the updated analyses of the CTT (Cholesterol Treatment Trialists) collaboration. 28 We are unaware of other studies examining the association of a Portfolio Diet with CVD events. The individual components of the Portfolio Diet, however, have been associated with lower rates of CVD events in prospective cohorts. Systematic reviews and metaanalyses have shown that consumption of legumes, 13 dietary fiber including viscous fiber sources, 14 nuts, 15 and MUFAs 16 are associated with reductions in CVD events, and consumption of foods high in saturated fat (such as red and processed meats) are associated with an increased risk of CVD. 17 The inverse association of increasing phytosterol intake from natural sources with Quartile 1 represents the least adherent to the Portfolio Diet, whereas quartile 4 represents the most adherence to the Portfolio Diet. Associations between Portfolio Diet and outcomes were determined by Cox proportional hazard models. Under/over energy reporters and those with baseline CVD were excluded from the analysis. Total CVD is a composite of incidence and death of CHD, stroke, heart failure, and coronary revascularization (coronary artery bypass grafting or percutaneous transluminal coronary angioplasty). CHD indicates coronary heart disease; CVD, cardiovascular disease; HR, hazard ratio; and Q, quartile. *Model 1 adjusted for age (continuous), region (Northeast, South, Midwest, West), and smoking (never, past, current). † Model 2 adjusted for model 1+ethnicity (White, Black, Hispanic, Asian/Pacific Islander, Other [American Indian, Alaskan Native, other]), education (college or above, below college), marital status (presently married/other), hysterectomy history (yes/no), boyd mass index (continuous), physical activity (continuous), alcohol intake (>7 drinks/week, <7 drinks/week), energy intake (continuous), cancer status (yes/no), hypertension status (yes/no), diabetes mellitus status (yes/no), sodium intake (continuous), family history of CVD (yes/no), family history of diabetes mellitus (yes/no), hormone therapy use (never, past, current), cholesterol-lowering medication use (yes/no).

Glenn et al Dietary Portfolio and Cardiovascular Disease
CVD risk in our study, however, was not shown in an earlier study. 29 The Portfolio Diet also shows similar results to other recognized dietary patterns for CVD prevention, such as the Dietary Approaches to Stop Hypertension, vegetarian, Nordic, and Mediterranean diets, which share important overlap in core foods (nuts, legumes, whole grains, fruit/vegetable sources, and/or monounsaturated fat). 18,30-33 Systematic reviews and metaanalyses of prospective cohort studies and large individual cohort studies have shown the Dietary Approaches to Stop Hypertension diet is associated with a 20% (95% CI, 0.76-0.85 HRs) reduction in CVD and a 21% reduction (0.71-0.88) in CHD incidence, 30 whereas Nordic and vegetarian diets are associated with 29% (0.65-0.78) and 22% (0.69-0.88) reductions in CVD and CHD mortality, respectively. 31,33 Similarly, the PREDIMED trial, a large randomized cardiovascular outcomes trial of the effect of a Mediterranean diet supplemented with either extra virgin olive oil or nuts compared with a low-fat diet, found reductions in major vascular events of 31% (0.53-0.91) and 28% (0.54-0.95), respectively. 18 Dietary patterns have also shown similar results specifically within the WHI. Higher adherence to the Healthy Eating Index 2010, Alternative Healthy Eating Index 2010, Alternate Mediterranean and Dietary Approaches to Stop Hypertension diets have been associated with 18% to 26% lower CVD mortality risk in the OS participants, 34 which falls within the 95% CIs (HR, 0.85; 95% CI, 0.78-0.93) of our findings for total CVD comparing lowest to highest adherence of the Portfolio Diet score in these participants. The 30% reduction in HF associated with higher adherence to the Alternative Healthy Eating Index 24 also falls within the 95% CIs (HR, 0.80; 95% CI, 0.64-0.99) of our findings for the Portfolio Diet in the OS participants.
Unlike some other dietary patterns, adherence to the Portfolio Diet was not associated with a reduction in stroke in our study. Both the Mediterranean and Dietary Approaches to Stop Hypertension diets have shown inverse associations with stroke. 30,32 Although the Portfolio Diet resulted in a reduction in blood pressure in the randomized trials, 8 this effect may not be strong enough to translate into an association with lower stroke risk, given that the reductions were small and hypertension is the most important risk factor for stroke. 35 The larger reductions in LDL-C and other lipid targets, as well as CRP, may be more relevant for the inverse associations seen with CHD and total CVD than with stroke. AF is also a major risk factor for stroke, 36 and we did not observe a significant association with lower AF risk in our study.
These findings highlight the plant-based Portfolio Diet as another dietary therapeutic approach for CVD prevention, alongside other dietary patterns recommended for CVD prevention. As adherence is one of the most critical determinants for attaining the benefits of any diet, as recognized by cardiovascular clinical practice guidelines, 9 the Portfolio Diet may best fit with the values and preferences of some patients and allow them to achieve the greatest adherence long term. The Portfolio Diet also has a small ecological footprint, emphasizing plant-based components with low environmental impact (eg, legumes, oats, barley, temperate fruit, etc). 37,38 Given increasing public concerns regarding ethical and environmental impact of food, 39,40 healthcare professionals will likely have more patients interested in this dietary pattern.

Strengths and Limitations
Strengths of our study include the prospective cohort design, large sample size, and long follow-up for incident CVD events. Nevertheless, this study does have limitations. First, our study included only 1 or 2 assessments of diet, and diet was self-reported. Second, the population included health-conscious postmenopausal women and therefore the results may not be generalizable to men or other populations; however, the Portfolio Diet trials were conducted in both men and postmenopausal women and benefits were seen in both sexes. Third, causation cannot be established because of the observational design, and residual confounding also cannot be ruled out. Lastly, consumption of many of the Portfolio Diet components remained low, particularly plant protein and MUFAs, even in the top quintiles. A few of the Portfolio Diet foods, such as some viscous fiber sources (eg, barley), were also not included on the FFQ. This finding was further highlighted in our post hoc sensitivity analysis where we created a Portfolio Diet score based on the recommendations of the Portfolio Diet trials. No participants in the WHI received the maximum amount of points possible, and maximum points suggested ≈50% adherence to the trial recommendations, with an average estimated adherence of ≈22%. These adherence estimations are, however, likely underestimated given the nature of FFQs and their inability to determine absolute intake of diets. Taken together, we expect that the associations are likely underestimated, and a stronger association with CVD events may be seen with greater consumption of the Portfolio Diet components. This low adherence reflects an important opportunity for individuals to achieve cardiovascular benefits of the Portfolio Diet. Typical dietary patterns in North America and Europe do not meet the targets for plant protein, viscous fiber, nuts, phytosterols, and MUFAs of the Portfolio Diet, [41][42][43][44][45][46] and therefore public health initiatives that focus on the components of the diet may improve cardiovascular outcomes globally.
It will be of great interest to apply this Portfolio Diet score to other populations, particularly in those that consume greater amounts of the diet components, to assess if similar or stronger associations with incident CVD events are found.

CONCLUSIONS
Greater adherence to the plant-based Portfolio Diet score was significantly associated with lower risk of total CVD, CHD, and HF in postmenopausal women. These findings provide the strongest evidence to date on the long-term benefits of a Portfolio Diet in the primary prevention of CVD, although our Portfolio Diet score needs to be assessed in other cohorts/populations to confirm these findings. Evidence from randomized trials with clinical CVD events is also needed. In this regard, we await the results of the PortfolioEX trial (Clini calTr ials.gov Identifier: NCT02481466) of the effect of the Portfolio Diet plus exercise on a surrogate marker of atherosclerotic CVD risk (magnetic resonance imaging of atherosclerosis [plaque volume]). In the meantime, our results support the Portfolio Diet as another therapeutic dietary approach for managing CVD risk that fits with current guidelines emphasizing plant-based diets.

ARTICLE INFORMATION
Author contributions: A.J.G., K.L., S.L., and J.L.S. had full access to the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. A.J.G., K.L., D.J.A.J., B.A.B., A.J.H., C.W.C.K., S.L., J.L.S. designed the study. A.J.G., K.L., and S.L. performed the statistical analyses. A.J.G. drafted the article. S.L. and J.L.S. supervised the project. All authors contributed to the data interpretation and revised each draft for important intellectual content. All authors read and approved the final article.

Sources of Funding
The Women's Health Initiative (WHI) was funded by the National Heart, Lung, and Blood Institute, National Institutes of Health, and U.S. Department of Health and Human Services through contracts HHSN268201600018C, HHSN268201600002C, HHSN268201600003C, HHSN268201600004C, and R01DK125403 (SL). Glenn was supported by the Nora Martin
The hazard ratios (HRs) are for assessing the Portfolio Diet score as a continuous exposure (0-7 points) and for comparing participants in Q1 (low adherence) to Q3 (high adherence) to the Portfolio Diet components based on recommendations from the Portfolio Diet RCTs (50g plant protein, 45g nuts, 20g viscous fiber, 2g plant sterols, 45g MUFAs, <7% energy from saturated fat and <200mg cholesterol per day). The scoring system to assess adherence to the Portfolio Diet score was based on 0-14 points (2 points for meeting previously listed 7 targets, 1 point for meeting half, and 0 points for less than half). Maximum score obtained was 7 points, therefore, 50% maximum adherence to the Portfolio Diet recommendations from the RCTs, with an average 22% adherence. Adherence to plant protein, viscous fiber and MUFAs was based on estimations of grams/servings for most commonly consumed foods in this category (plant protein was lentils/beans & green peas, viscous fiber was apples & oranges, and MUFAs were avocados). The remaining components were based on grams/day from the entire day as estimated from the FFQ.
The Cox regression models were adjusted for age, ethnicity, education, marital status, hysterectomy history, BMI, physical activity, smoking, alcohol intake, region in the U.S, study arm, energy intake, cancer status, hypertension status, diabetes status, sodium intake, family history of CVD, family history of diabetes, hormone therapy use, cholesterol lowering medication use. Under/over energy reporters and those with baseline CVD were excluded from the analysis. The hazard ratios (HRs) are for comparing participants in Q1 (low adherence) to Q5 (high adherence) to the Portfolio Diet components at baseline and year 3 (cumulative average). The Cox regression model was adjusted for age, ethnicity, education, marital status, hysterectomy history, BMI, physical activity, smoking, alcohol intake, region in the U.S, study arm, energy intake, cancer status, hypertension status, diabetes status, sodium intake, family history of CVD, family history of diabetes, hormone therapy use, cholesterol lowering medication use.