Effects of dietary patterns on the all‐cause mortality and cardiovascular disease mortality in patients with hypertension: A cohort study based on the NHANES database

Abstract Background Hypertension (HTN) patients have higher risk of all‐cause and cardiovascular disease (CVD)‐specific mortality. Dietary patterns have been reported related to the risk of mortality, but their roles in HTN patients is unclear. Hypothesis To explore the relationships between different dietary patterns and all‐cause/CVD‐specific mortality and provide dietary guidance for HTN patients' prognosis improvement. Methods Data of 27 618 HTN patients were extracted from the National Health and Nutrition Examination Survey (NHANES) database in this retrospective cohort study. The associations between Healthy Eating Index (HEI)‐2015, Alternate Healthy Eating Index (AHEI)‐2010, Dietary Approaches to Stop Hypertension (DASH), and Mediterranean (MED) diet and all‐cause and CVD‐specific mortality were explored using univariate and multivariate Cox regression analyses with hazard ratios (HRs) and 95% confidence intervals (CIs). Subgroup analyses of age, gender, body mass index, and comorbidity were also performed. Results The median follow‐up time was 83 months. A total of 3462 patients died for all‐cause and 1064 died due to CVD. After adjusting for covariates, we found that high adherence to AHEI‐2010 (HR = 0.84 for all‐cause; HR = 0.72 for CVD), and MED (HR = 0.84 for all‐cause; HR = 0.77 for CVD) diet were associated with decreased risks of both all‐cause and CVD‐specific mortality. In patients who aged ≥65 years old, were normal/overweight, without complications, the relationships between different dietary patterns and risk of mortality were different. Conclusion High scores of AHEI‐2010 and MED may be associated with decreased risks of all‐cause and CVD‐specific mortality in patients with HTN.


| INTRODUCTION
Hypertension (HTN) is a clinical syndrome characterized by elevated systolic blood pressure (SBP)/diastolic blood pressure (DBP) ≥ 140/ 90 mmHg, affecting 1.13 billion adults worldwide. 1HTN is a major modifiable risk factor for cardiovascular disease (CVD), 2 and patients with HTN have higher risk of both all-cause and CVD-specific mortality compared with the general population. 3It is paramount to identify effective strategies to improve the prognosis of HTN individuals.
Diet adjustment, a mean of nondrug treatment, is an important part of HTN management. 4Previous studies have showed that dietary patterns can significantly impact a HTN patient's health, with a noticeable effect on blood pressure (BP). 5,6A meta-analysis of 30 randomized controlled trials found that Dietary Approaches to Stop Hypertension (DASH) significantly decreased SBP and DBP in adults with HTN. 7 Adherence to the DASH in HTN patients has been associated with a lower risk of all-cause mortality. 8A systematic review and meta-analysis of 15 prospective cohort studies showed that higher diet quality characterized by the Alternate Healthy Eating Index (AHEI) was significantly related to decreasing all-cause mortality. 9Hu et al. 10 found that higher adherence to the Healthy Eating Index 2015 (HEI-2015) diet was linked to lower risks of CVDspecific and all-cause mortality among US adults.The adoption of the Mediterranean (MED) diet was also accompanied by a relatively small, but yet significant BP reduction in HTN patients. 11However, studies on the association between different diet patterns and prognosis in HTN patients have been scarce.
Herein, this study aims to explore the effects of different diet patterns on the all-cause and CVD-specific mortality in HTN patients, in order to provide some dietary guidance for the management and prognosis improvement of these patients.A total of 27 618 adults who diagnosed with HTN were absorbed in this study.The exclusion criteria were (1) aged <20 years old, (2)   having extremely low or high total energy intake (under 500 or over 8000 kcal/day), and (3) missing the information of dietary.After excluding those who lost to follow-up, finally, 21 757 of the adults were eligible.Since the NHANES protocol was approved by the Institutional Review Board (IRB) of the NCHS of CDC, and the data were publicly available, no ethical approval of our IRB was required.

| Assessment of dietary patterns
The NHANES database uses two 24-hour recalls as the principal methodology to provide detailed quantitative food and nutrient intake assessment, and uses the food frequency questionnaires (FFQs) to provide typical or qualitative data. 13The FFQ was calculated in standardized portion sizes and nutrient intakes as grams per day.Participants without complete dietary data and those that reported excessively high or low values for total food or energy intake (<600 kcal/day in or more than 4200 kcal/day) were excluded. 14e DASH dietary pattern aims at lowering BP which focuses on high amounts of fruits, vegetables, and low-fat dairy products.DASH score was calculated based on protein, fiber, magnesium (Mg), calcium (Ca), potassium (K), total fat, saturated fat, cholesterol, and sodium (Na) to examine the degree of adherence to the DASH.
Provide one point when meeting the goal for each component, provide 0.5 points when meeting an intermediate goal of DASH and give zero points when the nutrient content of the DASH control diet and meeting neither goal.The control diet targets are from the previous DASH study. 15This pattern score ranges between 0 and 9 points, where higher scores indicate greater adherence to the DASH dietary pattern.
The AHEI-2010 includes information of 11 components of food groups and nutrients that are associated consistently with lower risk of chronic diseases in clinical and epidemiologic investigations, which is based on a comprehensive review of the relevant literature and discussions with other nutrition researchers. 16Briefly, higher intakes of vegetables, fruits, whole grains, nuts and legumes, long-chain ω-3 fatty acids (timnodonic acid and docosahexaenoic acid), polyunsaturated fatty acids, and moderate alcohol count favorably, whereas higher intakes of sugar-sweetened beverages, red/processed meats, trans fat, and Na count unfavorably.Each component of the score is rated on a scale of 0-10, and the total AHEI-2010 score ranged from 0 (nonadherence) to 110 (perfect adherence).
The HEI-2015 is a diet quality index to assess the adherence to the Dietary Guidelines in Americans, which were developed by the USDA's Center for Nutrition Policy and Promotion. 7HEI-2015 has replaced empty calories with saturated fat and added sugar basing on the HEI-2010.HEI-2015 focuses on the consumption of total fruits, whole fruits, total vegetables, greens and beans, whole grains, dairy foods, total protein foods, seafood and plant proteins, fatty acids, refined grains, Na, added sugars, and saturated fats.The score ranges from 0 to 100 points, and the higher score reflects the healthier eating.
The MED score was calculated according to the information in the Food Patterns Equivalents database (developed by the US Department of Agriculture): https://www.ars.usda.gov/northeast-areaeltsvillee-mdbhnrceltsvillee-human-nutrition-research-center/food-surveys-researchgroup/docs/fped-databases/.In summary, the calculation was based on assessed intakes of total fruits, vegetables (except potatoes), whole grains, legumes, nuts, fish, red and processed meat, ratio of monounsaturated to saturated fat, and alcohol.A single point was assigned to participants whose intakes were greater than the median of the study cohort, except for red and processed meat and alcohol.For alcohol and red/processed meat, one point was assigned to those who had moderate alcohol consumption (10-25 g/day for men and 5-15 g/day for women) or a meat intake that was less than the median for the cohort.If these criteria were not met, participants received point of zero.Hence, higher MED (total score = 18) score represents greater adherence to the MED diet. 17 the current study, the four dietary patterns were divided into three levels, respectively, according to their tertiles, including HEI-2015: ≤46.64, (46.64, 58.28), and >58.28,AHEI-2010: ≤30.5, (30.5,   39), and >39, DASH: ≤1.5, (1.5, 2.5), and >2.5, and MED score: ≤3, (3, 5), and >5.

| Variables collection
We collected the variables including age, gender, race, education level, marital status, poverty-to-income ratio (PIR), smoking status, drinking status, height, weight, body mass index (BMI), waistline, physical activity, total energy intake, estimated glomerular filtration rate (eGFR), platelet, hemoglobin (HB), creatinine (Cr), C-reactive protein (CRP), CVD, diabetes mellitus (DM), dyslipidemia, cancer, asthma, anemia treatment, gout, chronic obstructive pulmonary disease (COPD), dialysis, hypothyroidism, human immunodeficiency virus (HIV) infection, and drug abuse from NHANES.or self-reported hypercholesterolemia, or use of a cholesterol medication. 10Medication data were compiled from the prescription drug questionnaire taken during the home interview. 11According to the multiple-choice question (MCQ) in NHANES: "Have you ever been told you had (congestive) heart failure, coronary heart disease, angina/angina pectoris, heart attack and stroke" or using of cardiovascular drugs in any one or more diseases or medications was diagnosed as CVD.Renal function was calculated using the Chronic Kidney Disease Epidemiology Collaboration equation 18 : eGFR = 141 × min (SCr/κ, 1) α × max (SCrκ,1) − 1.209 × 0.993Age × 1.018 (if female) × 1.159 (if Black) where SCr (mg/dL) is serum Cr collected as part of the standard biochemistry profile, κ is 0.7 if female or 0.9 if male, α is −0.329 if female or −.411 if male, min is the minimum of SCr/κ or 1, and max is the maximum of SCr/κ or 1.
Participants with SBP ≥ 130 mmHg or DBP ≥ 80 mmHg were defined as HTN according to the American Heart Association/ American College of Cardiology 2017 guideline.If the BP measurement data were missing, patients with antihypertensive medication prescription in the clinical history "Are you now taking prescribed medicine" were also considered hypertensive.Survey participants who had been told by a doctor or other health care provider to have emphysema (MCQ160G) and/or chronic bronchitis (MCQ160K) were considered COPD.Subjects who were asked "Has a doctor or other health professional ever told you that you had gout?" and had positive answers were categorized into gout subjects.Information of hypothyroidism, anemia, asthma, cancer, and HIV infection were all collected through the household interviews and MCQs.NHANES collected participants' blood samples in the mobile examination center (MEC) for further examinations of clinical indicators.

| Outcomes and follow-up
The study outcomes were all-cause mortality and CVD-specific mortality.The mortality information of patients was queried using the NHANES-linked social security account through the website: https:// www.cdc.gov/nchs/data-linkage/mortality.htm.CVD-specific mortality was defined and classified by the International Statistical Classification of Disease (ICD) codes, 10th revision (coded I00-I78). 19The follow-up ended until patients died or up to December 31, 2019.

| Statistical analysis
Normal distributed data were described by mean ± standard deviation (mean ± SD), and t test was used for comparation.Data  S1).

| Characteristics of participants
A total of 27 618 patients with HTN were initially included.We excluded those who aged less than 20 years old (n = 662), missing dietary data (n = 2255), and having extreme energy intakes (<500 or >8000 kcal) (n = 147).Then, individuals who lost to the follow-up were also excluded (n = 2795).Finally, 21 757 of the patients were eligible.
Table 1 showed the characteristics of patients with HTN.After a median 83 months' follow-up, 3 462 patients died for all-cause while that 1064 died due to CVD.The average age of eligible patients was 53.45 years old and 53.04% of them were male.We found the HEI-2015, AHEI-2010, DASH, and MED scores were all significantly different between survival group and all-cause mortality group, and between the survival group and CVD-specific mortality group (all p < .05).

| Association between different dietary patterns and all-cause mortality and CVD-specific mortality in patients with HTN
Covariates respectively related to all-cause mortality and CVDspecific mortality were showed in Supporting Information: Table S2.
We found that age, race, education level, marital status, PIR, physical activity, total energy intake, smoking, eGFR, drinking, DM, CVD, dyslipidemia, cancer, asthma, anemia treatment, gout, COPD, HB, and dialysis were both associated with all-cause mortality and CVDspecific mortality (all p < .05).
In addition, Figures 1 and 2 were the KM curves that reflected the effects of different level of dietary pattern scores on the all-cause mortality and CVD-specific mortality in patients with HTN.

| Subgroup analyses of age, BMI, gender, CVD, DM, and dyslipidemia
Table 3 shows the results of subgroup analyses.In patients aged ≥65 years old, high HEI-2015, AHEI-2010, and MED scores were all associated with decreased risks of both all-cause mortality and CVDspecific mortality (all p < .05).Additionally, high DASH score was only linked to a decreased risk of all-cause mortality (HR = 0.84, 95% CI: [0.72-0.99]).
Greater adherence to HEI-2015 has been reported to be inversely correlated with the risk of incident HTN, and among its components, a higher score of total fruits, fatty acids, and Na was associated with a reduced risk of incident HTN. 22The AHEI, modeled on the original HEI, was constructed according to extensive epidemiological evidence for noncommunicable diseases prevention. 23Although HEI and AHEI diet scores were developed for slightly different purposes, they both emphasize on high intake of fruits, vegetables, monounsaturated fat, legumes, whole grains, and low consumption of saturated fatty acids and recommends lowers intakes for added sugars, refined grains, and Na, which associated with the increasing risk of HTN. 24,25Diets rich in fiber, low energy density, and low glycemic load are protective against HTN. 26The favorable impacts of vegetables/fruits on HTN could be attributed to potential healthy constituents such as phytochemicals, 27 vitamins, 28 Mg, and K 29 and antioxidants, 30 which have been independently related to a reduction in BP.The consumption of nuts and soy and moderate alcohol intake appeared to be the most important independent contributors to decreased mortality risk. 31The relatively high consumption of nuts, olive oil and moderate intake of red wine during meals (which can be considered a primarily plant-based diet), makes the MED diet unique and different from the other healthy diet patterns. 32Moreover, it has also been reported that BP can be decreased when individuals had high adherence to MED diet. 33,34ese may explain our finding that high AHEI-2010 and MED scores were associated with low risks of all-cause mortality and CVDspecific mortality.
Due to HTN may be a result of and can be influenced by multiple factors, we further performed subgroup analyses.Age, low educational attainment, overweight, obesity, abdominal obesity, and high cholesterol and blood glucose were strongly associated with HTN. 35 age subgroups, the highest tertiles of HEI-  dysfunction, and salt retention. 37However, we did not find any difference between genders in the relationships of different dietary patterns and mortality.The potential explanations for the gender difference have been reported, which could be the higher carbohydrate intakes typically observed among women, and evidence that females have increased sensitivity to insulin in skeletal muscle than men, leading to higher levels of intramuscular fat and subsequently increased HTN risk. 38Moreover, we also observed that different dietary patterns were linked to the mortality in HTN patients with different complications including CVD, DM, and dyslipidemia.High BP not only results in deleterious mechanical stress on blood vessels but also on the myocardium, leading to the development of hypertensive heart disease and congestive heart failure. 39The potential pathways of action may be dietary fiber can reduce LDL cholesterol and TG uptake to improve the elasticity of blood vessel walls to decrease vascular resistance and maintain adequate tissue perfusion. 40High-fiber foods such as vegetables also contain beneficial nutrients metabolized into compounds which may improve BP through greater bioavailability for use in vasodilation. 41Furthermore, recent studies have found some supports for reductions in body weight with higher fiber intakes, as shown in evidence synthesis of the general population and those with DM, with weight loss beneficial in the treatment and prevention of HTN. 42,43Herein, we considered that our study may partly provide some new ideas for the public policies related to dietary recommendation to improve health in patients with HTN, because the traditional DASH dietary may not consider the difference among all interindividual health conditions.
T A B L E 3 (Continued)

2 | METHODS 2 . 1 |
Study design and participantsData of participants in this retrospective cohort study were extracted from the National Health and Nutrition Examination Survey (NHANES) database in 2005-2018.NHANES includes publicly available 2-year cross-sectional surveys that focus on the health and nutrition of noninstitutionalized US residents, which is administered by the Centers for Disease Control and Prevention (CDC) and the National Center for Health Statistics (NCHS). 12Details including questionnaires, data sets, and related documentation of study implementation are available for online access: https://wwwn.cdc.gov/nchs/nhanes/Default.aspx.
Physical activity was calculated by the physical activity questionnaire in NHANES, using the following formula: energy expenditure (metabolic equivalent [MET]•min) = recommended MET × exercise time of corresponding activity (min).Participants who had laboratory diagnosis of fasting plasma glucose ≥7 mmol/L, HbA1c ≥ 6.5% (48 mmol/L), self-reported previous diagnosis of DM by medical professionals, or taking glucose-lowering agents or insulin (survey or prescription medication check) were considered to have DM.Dyslipidemia was defined as having total cholesterol ≥ 200 mg/dL (5.2 mmol/L), or triglyceride (TG) ≥ 150 mg/dL (1.7 mmol/L), or low-density lipoprotein cholesterol ≥ 130 mg/dL (3.4 mmol/L), or high-density lipoprotein cholesterol ≤ 40 mg/dL (1.0 mmol/L), were described by median and quartiles [M (Q1, Q3)], and Mann-Whitney U rank sum test was used for comparation.The composition ratio (N [%]) was used to describe the distribution of measurement data, and we used chi-square test for comparation.The MEC exam weight (wtmec2yr) was used for weighting since data sets in the current study incorporated seven cycles of NHANES 2-year data collection.Weighted univariate Cox regression analysis was used to screen covariates related to allcause mortality and CVD-specific mortality.Univariate and multivariate Cox regression analyses were used to explore the relationship of different dietary patterns and all-cause mortality and CVD-specific mortality, respectively.Model 1 was the crude model.Model 2 adjusted for demographic data including age, gender, education level, and marital status.Model 3 of all-cause mortality adjusted for age, race, education level, marital status, PIR, physical activity, smoking, total energy intake, eGFR, drinking, BMI, DM, CVD, dyslipidemia, cancer, anemia treatment, gout, COPD, HB, dialysis, CRP, and hypothyroidism.Model 3 of CVD-specific mortality adjusted for age, race, education level, marital status, PIR, physical activity, smoking, eGFR, drinking, DM, dyslipidemia, cancer, anemia treatment, gout, COPD, HB, dialysis, CRP, and hypothyroidism.We further drew the Kaplan-Meier (KM) curves to reflect the influences of different dietary patterns on all-cause mortality and CVD-specific mortality, respectively.Subgroup analyses of age, BMI, CVD, DM, and dyslipidemia were also performed.The evaluation indexes were hazard ratios (HRs) and 95% confidence intervals (CIs).Two-sided p < .05 was considered as significant.Statistical analyses were completed by SAS 9.4 (SAS Institute.),and R version 4.2.0 (Institute for Statistics and Mathematics).Missing data were interpolated.We further performed the sensitivity analysis on characteristics of participants before and after the interpolation of missing data (Supporting Information: Table The current study explored the association between different dietary patterns and all-cause mortality and CVD-specific mortality in patients with HTN.Our results showed that high AHEI-2010 and MED scores were linked to decreased risks of both all-cause mortality and CVD-specific mortality.Also, these relationships in patients with different age, BMI, CVD, DM, and dyslipidemia were different.Although there are few reports on the relationship between different diet quality scores and the prognosis in patients with HTN, previous research can still be instructive.Zhong et al.20 analyzed data from 6 prospective cohorts collected in 1985-2016 to assess the role of diet quality in the long-term absolute risks for CVD and mortality in adults.The results showed that participants who consumed a diet with the highest quintile of AHEI-2010 score or DASH score had lower long-term absolute risks for incident of CVD and all-cause mortality.20In a large-sample prospective study in the Middle East, researchers found that individuals with the highest quintile of dietary scores including AHEI-2010, MED, and DASH, compared with that of the lowest quintile dietary scores, had a significantly decreased allcause mortality, and a reduced CVD-specific mortality was found for those who had high AHEI-2010 and DASH scores.21Another prospective cohort study of participants from the Atherosclerosis Risk in Communities Study showed that compared with participants in the lowest quintile of dietary scores including HEI-2015, AHEI-2010, MED, and DASH, participants in the highest quintile had lower risks of incident CVD, CVD-specific mortality, and all-cause mortality.10In the current study, we found patients with HTN in the highest tertile of AHEI-2010 and MED score had lower risks of both all-cause mortality and CVD-specific mortality.Differently, we could not find significant association between other dietary scores (including HEI-2015 and DASH) and all-cause mortality or CVD-specific mortality in this study.The possible reasons may be the study population was different from the previous study, or patients with HTN usually received appropriate treatment, or they focused more on keeping a F G U R E 1 Kaplan-Meier curve of the effects of different dietary patterns on all-cause mortality in patients with hypertension.LI ET AL.

|
1367       difference between survival and mortality groups which the difference was over 10 years.The potential mechanisms of the roles of these dietary patterns in elderly persons with HTN to be more significant compared with that in youngers are needed further exploration.HTN patients who were normal or overweight other than obese had high AHEI-2010 were seemed to have decreased risks of all-cause mortality and CVD-specific mortality.Our study was in line with previous studies.Mohebbi et al.36 considered BMI (normal and overweight) can be a predictor of adherence to diet.High sugar and/or refined grains foods can alter HTN risk through mechanisms related to obesity, metabolic dysregulation, vascular T A B L E 1 Characteristics of patients with HTN.Association between different dietary patterns and all-cause mortality and CVD-specific mortality.
Association between different dietary patterns and all-cause and CVD-specific mortality in age, gender, BMI, and comorbidity subgroups.