Effects of Diet and Sodium Reduction on Cardiac Injury, Strain and Inflammation: The DASH-Sodium Trial

Abstract

Background:

We recently determined that the DASH diet has beneficial effects on cardiac biomarkers. The effects of sodium reduction on cardiac biomarkers, alone or combined with the DASH diet, are unknown.

Objectives:

Our objective was to determine the effects of sodium reduction and the DASH diet, alone or combined, on biomarkers of cardiac injury, strain, and inflammation.

Methods:

DASH-Sodium was a controlled feeding study in adults with SBP 120–159 mm Hg and DBP 80–95 mm Hg, randomly assigned to the DASH diet or a control diet. On their assigned diet, participants consumed each of three sodium levels for 4 weeks. Body weight was kept constant. At the 2,100 kcal level, the three sodium levels were low (50 mmol/d), medium (100 mmol/d), and high (150 mmol/d). Outcomes were 3 cardiac biomarkers: high-sensitivity cardiac troponin I (hs-cTnI, measure of cardiac injury), N-terminal b-type pro natriuretic peptide (NT-proBNP, measure of strain), and high-sensitivity C-reactive protein (hs-CRP, measure of inflammation), collected at baseline and at the end of each feeding period.

Results:

Of the original 412 participants, mean age was 48 yr; 56% were women, and 56% black. Mean baseline SBP/DBP was 135/86 mmHg. DASH (versus control) reduced hs-cTnI by 18% (95%CI: −27%,−7%) and hs-CRP by 13% (95%CI: −24%,−1%), but not NT-proBNP. In contrast, lowering sodium from high to low levels reduced NT-proBNP independently of diet (19%; 95%CI: −24%,−14%), but did not alter hs-cTnI and mildly increased hs-CRP (9%; 95%CI: 0.4%, 18%). Combining both DASH with sodium reduction lowered hs-cTnI by 20% (95%CI: −31%,−7%) and NT-proBNP by 23% (95%CI: −32%,−12%), while hs-CRP was not significantly changed (−7%; 95%CI: −22%, 9%) compared to the high sodium-control diet.

Conclusions:

Combining a DASH dietary pattern with sodium reduction can lower two distinct mechanisms of subclinical cardiac damage: injury and strain, while DASH alone reduced inflammation.

Condensed Abstract:

The direct benefits of healthy diet on cardiac damage beyond cardiovascular disease (CVD) risk factors is fiercely debated. We measured 3 biomarkers implicated in the development of CVD in serum collected after 4-, 8-, and 12-weeks of feeding from the DASH-Sodium trial: high-sensitivity cardiac troponin I (hs-cTnI), N-terminal b-type pro natriuretic peptide (NT-proBNP), and high-sensitivity C-reactive protein (hs-CRP). Compared to a higher sodium, typical American diet, a lower sodium, DASH diet significantly reduced hs-cTnI and NT-proBNP, markers of subclinical cardiac injury and strain. These findings amplify the importance of healthy diet to prevent subclinical cardiac damage.

Cardiovascular disease (CVD) is the leading cause of death in the United States (1) and risk factors (e.g. optimal blood pressure) to prevent CVD events are worsening (2). Public health advocates point to low consumption of fruit and vegetables as well as high sodium consumption as potential causes of these concerning trends (3, 4), urging lawmakers to institute public policy interventions that promote healthier food choices (5). However, critics are quick to point out the lack of evidence linking poor diet or high sodium intake to CVD damage (6–8) or CVD events (9).

In two recent secondary analyses of controlled feeding studies, we demonstrated that diets rich in fruit and vegetables lowered subclinical cardiac injury and strain (10) and inflammation (11). However, these studies did not examine the effects of sodium reduction, alone or combined with healthy diet, on CVD damage, and one was limited by lack of a typical US control diet (11).

The Dietary Approaches to Stop Hypertension (DASH)-Sodium trial was a controlled-feeding study that examined the effects of 3 levels of sodium intake (low, medium, or high) in 2 distinct diets - a typical American diet (control diet) or a healthy diet (DASH Diet) on blood pressure (12). Ultimately, the trial demonstrated that both the DASH diet and sodium reduction independently reduced blood pressure in adults with elevated blood pressure or hypertension (12). However, the impacts of sodium reduction on subclinical cardiac injury, strain, or inflammation were not determined (12). Given the potent effects of sodium on BP, we hypothesized that sodium reduction would reduce subclinical cardiac injury, strain, and inflammation.

In the present study, we examined three biomarkers in stored specimens from participants of the DASH-Sodium trial to determine the independent and combined effects of the DASH diet and sodium reduction on subclinical (1) cardiac injury (high sensitivity cardiac troponin I, hs-cTnI), (2) cardiac strain (N-terminal pro-brain natriuretic peptide, NT-proBNP), and (3) inflammation (high sensitivity C-reactive protein, hs-CRP).

Methods

The DASH-Sodium trial was an investigator-initiated study conducted from September 1997 through November 1999 at four clinical centers (Baltimore, Maryland, Boston, Massachusetts, Durham, North Carolina, and Baton Rouge, Louisiana) and funded by the National Heart, Lung, and Blood Institute (12). In brief, DASH-Sodium compared the effects of three different sodium levels consumed in the setting of the DASH diet or a typical American diet (control) on blood pressure. The control diet was designed to reflect the 25^th percentile of several micronutrients and the average macronutrient profile of the US population. Institutional Review Boards at each institution approved the original study protocol. Furthermore, all participants provided written, informed consent to participate in the original study and for their biospecimens to be stored and used for subsequent analysis. The current study was determined by the Institutional Review Board of Beth Israel Deaconess Medical Center to be exempt research and was supported by the National Heart, Lung, and Blood Institute to measure biomarkers in stored specimens curated by the BioLINCC repository.

Participants

DASH-Sodium enrolled adults, aged 22 years and older with mean SBP of 120–159 mm Hg and DBP of 80–95 mm Hg; however, adults with a prior diagnosis of heart disease (i.e., myocardial infarction, coronary artery bypass graft, angioplasty, symptomatic ischemic heart disease, stroke, or congestive heart failure), renal insufficiency, poorly controlled dyslipidemia, diabetes mellitus, or heart failure were excluded (12). DASH-Sodium also excluded adults taking antihypertensive agents or insulin, and persons drinking more than 14 alcoholic drinks a week.

Dietary Interventions

The DASH-Sodium trial simultaneously tested two dietary patterns and three levels of sodium intake. The two dietary patterns were tested using a parallel design. Participants were randomized to either the DASH diet or a control diet. The three sodium levels were tested using a crossover design. In the context of their assigned dietary pattern, participants were assigned each of three sodium levels (low, medium, and high) in random order.

The DASH diet emphasized fruits, vegetables, and low-fat dairy products, included whole grains, poultry, fish, and nuts, and was reduced in saturated fat, total fat, cholesterol, red meat, sweets, and sugar-containing beverages (13). The control diet was designed to reflect a American diet typical of what many people in the U.S. eat (Supplemental Table 1) (14).

Each participant ate their assigned diet at 3 sodium levels: high (1.6 mg/kcal), medium (1.1 mg/kcal), and low (0.5 mg/kcal) based on their estimated energy intake. Estimated Calorie needs were monitored throughout the study and assigned Calories were adjusted as needed to maintain weight constant. The five energy levels prepared were 1,600, 2,100, 2,600, 3,100, and 3,600 kcal/day. At the 2,100 kcal level, the low level of sodium was 1,150 mg/d, the medium level was 2,300 mg/d, and the high level was 3,450 mg/d. The high level was consistent with average U.S. sodium intake. The medium level was based on sodium recommendations contemporaneous to the study. The low sodium intake, below current recommendations, was hypothesized to result in greater BP reduction. Differences in sodium level were achieved with unsalted or salted varieties of foods or by adding salt to entrees, ensuring similar nutrient profiles. Details of the dietary interventions are republished in Supplemental Table 1 (14).

All meals and snacks were provided to participants. The participants’ main meal (lunch or dinner) was consumed at the study centers under direct observation with close monitoring for waste or protocol deviations, while the remaining study meals were consumed off-site. Participants were weighed daily and, if needed, energy intake adjusted nearly weekly to ensure that weight remained constant.

Over a two-week run-in period, all participants ate the high sodium-control diet, and then were randomized to diet and sodium sequence. Each sodium level was consumed for a mean of 30 days, separated by a mean 5-day washout period during which participants could eat their usual diets. Over 98% of participants completed all intervention periods (12). With regards to attendance, participants adhered to the study protocol for 37,703 of 42,911 study days (87.9%).

Outcomes of Interest

Biomarkers used for the present analysis (hs-cTnI, NT-proBNP, and hs-CRP) were measured in 2019 from stored serum specimens. The biomarkers were selected based on their relationship with 3 distinct pathways of subclinical cardiac damage: subclinical myocyte injury (hs-cTnI), cardiac strain as a marker of wall stress (NT-proBNP), and inflammation (hs-CRP). Moreover, these markers have been shown to predict CVD events in adults without known CVD (15–21). Serum specimens were collected in participants after a 12-hour fast at baseline prior to feeding while participants ate their self-selected diets and at the conclusion of each of the three 4-week sodium feeding periods. All serum had been stored at −70°C and underwent at least 1 freeze-thaw cycle prior to the present measurements. Of a total of 1,648 specimens (412 participants with 4 visits), 1,547 (93.9%) were available for our study with 1,543 (93.6%) yielding results for all 3 biomarkers. Missing data were evenly distributed across dietary assignments (Supplemental Figure 1).

Measurement kits and assays were donated by Siemens (Siemens Healthineers, Malvern, Pennsylvania) for: (1) ADVIA Centaur High-Sensitivity Troponin I (reported within-run CV of 4.8% for a mean of 13.11 ng/L [or pg/mL]), (2) Dimension Vista N-terminal Pro-brain Natriuretic Peptide (reported within-run CV of 1.4% for a mean of 120 pg/mL), and (3) Dimension Vista high sensitivity C-reactive protein (reported within-run CV of 5.2% for a mean of 2.39 mg/L). Manufacturer estimates of assay accuracy and precision performance are found elsewhere (10). Biomarker assays had the following limits of detection: <1.60 ng/L (hs-cTnI), <5 pg/mL (NT-proBNP), and <0.160 mg/L (hs-CRP). Given the larger number of hs-cTnI values below the limit of detection, in sensitivity analyses we examined an alternate measurement cut point based on the assay’s the limit of blank (<0.5 ng/L). Of note, the limit of detection is defined by the manufacturer as the lowest concentration of hs-cTnI that can be detected with 95% probability, while the limit of blank is defined as the highest measurement that might be observed for a blank sample. For figures depicting geometric means, values below the limit of detection were imputed using a value two-thirds the distance to 0. Siemen’s hs-cTnI assay has been shown to be stable up to 15 years of storage and to freeze-thaw (22).

Covariates

Other participant characteristics were determined via self-reported questionnaire, laboratory specimens, and physical examination. Race was examined in categories of self-reported black and non-black. Seated SBP and DBP were measured at the same time with random-zero sphygmomanometers at three visits during the screening phase and at two visits during the 2-week run-in period. The average of these 5 measurements served as a baseline for this study. Hypertension, as defined in this trial, was SBP ≥140 mm Hg or DBP ≥90 mm Hg. High density lipoprotein cholesterol, triglycerides, and total cholesterol were measured using enzymatic colorimetry and used to estimate low density lipoprotein cholesterol (LDL-c) (23). Body mass index (BMI) was derived from measured height and weight, and obesity was defined as a BMI ≥30 kg/m².

Statistical analysis

We described baseline population characteristics by diet assignment using means (SD) and proportions. Given data skew, we determined the geometric mean of serum concentrations (SD) of biomarkers at baseline and after each of the three 4-week sodium periods and used both the absolute difference (the difference in geometric means) and the percent change (derived by exponentiating the difference in log-transformed baseline and end-of-period values) to compare change from baseline.

We compared log-transformed cardiac markers across dietary assignments, using the following contrasts: DASH versus control, high versus low sodium, and a combination of low sodium-DASH versus high sodium-control. We also examined trends across sodium levels (low, medium, and high) by treating sodium assignment as a continuous variable in models. We report both the difference in exponentiated values (geometric means) to estimate change on the original marker scale as well as exponentiated differences to present %-difference between the diets. The primary outcome was the %-difference. For sodium, the primary contrast was high versus low sodium.

All comparisons (both baseline and between diets) were performed via mixed effects tobit models (metobit command) that were left-truncated for the limits of detection or blank described above. A tobit model was used to address informative left-censoring that occurs below the limit of detection (or blank) for each assay, allowing us to fit a linear regression model in the detectable range, while designating undetectable biomarkers as below the detectable range (24). The fixed effects portion of the tobit model included diet assignment, visit (baseline and up to 3 follow-up visits) or sodium level, and the interaction of these terms (i.e., DASH -by-visit for DASH contrasts or DASH-by-sodium level for sodium contrasts). The random effects portion of the tobit model included participant id (a random intercept). In models of dietary effect (including those comparing the low sodium DASH diet to the high sodium control diet), the DASH-by-visit interaction term was used to determine the effect of diet compared to control. In models of sodium reduction, the DASH-by-sodium interaction terms were used to determine the effect of diet within sodium level. In order to determine the overall dietary effect (DASH versus control), all 3 follow-up visits were designated as follow-up (distinct from baseline). Measurements were not averaged together.

A two-tailed P-value of <0.05 was considered statistically significant without adjustment for multiple comparisons. All analyses were conducted using Stata version 15.1 (Stata Corporation, College Station, TX, USA).

Funding

The funding source of the present study played no role in the study’s design, conduct, analysis, and decision to submit the manuscript for publication.

Results

Baseline characteristics

Baseline characteristics of the participants of the DASH-Sodium trial by randomized dietary pattern are shown in Table 1. There were minimal differences by randomized dietary pattern.

Table 1.

Baseline characteristics according to diet assignment

	Control Diet		DASH Diet
	Total N	Mean (SD) or % or Median (25th to 75th percentile)	Total N	Mean (SD) or % or Median (25th to 75th percentile)
Age, yr	204	49.1 (10.4)	208	47.4 (9.6)
Women, %	204	54.4	208	59.1
Black, %	204	56.4	208	57.2
Blood pressure, mm Hg
Systolic	204	135.4 (9.4)	208	134.2 (9.6)
Diastolic	204	85.8 (4.1)	208	85.6 (4.8)
Baseline SBP≥140 or DBP≥90 mm Hg, %	204	40.7	208	40.9
High density lipoprotein cholesterol, mg/dL	204	48.1 (13.1)	208	48.5 (12.5)
Low density lipoprotein cholesterol, mg/dL	202	131.8 (31.6)	204	130.7 (29.8)
Triglycerides, mg/dL	204	94.0 (67.0 to 140.5)	208	92.0 (67.0 to 130.5)
Total cholesterol, mg/dL	204	202.7 (36.0)	208	202.0 (36.4)
Body mass index, kg/m2	204	29.5 (5.0)	208	28.8 (4.7)
Body mass index ≥30, %	204	40.2	208	37.5

Abbreviations: DASH, Dietary Approaches to Stop Hypertension; N, number; SD, standard deviation

Effects of the DASH Diet

The median (25^th to 75^th percentiles) of the cardiac biomarkers at baseline are in Supplemental Table 2. Mean baseline concentrations of hs-cTnI, NT-proBNP, and hs-CRP were similar between the control and DASH diet (Figure 1, Table 2). Compared with baseline, participants assigned the control diet had virtually no mean change in hs-cTnI or hs-CRP, but a significant reduction in NT-proBNP (−17.7; 95% CI: −23.8, −11.2) (Supplement Figure SF2). In contrast, compared to baseline, participants assigned the DASH diet had significant reductions in all three markers, reducing hs-cTnI by 16.1%, NT-proBNP by 21.6% and hs-CRP by 14.3.

Figure 1. Effects of sodium reduction by dietary assignment on cardiac biomarkers.

Geometric means (95% confidence intervals) of (A) high sensitivity cardiac troponin I (hs-cTnI, ng/L), a marker of cardiac injury, (B) N-terminal pro-brain natriuretic peptide (NT-proBNP, pg/mL), a marker of cardiac strain, and (C) high sensitivity C-reactive protein (hs-CRP, mg/L), a marker of inflammation according to diet assignment at baseline and after high, medium, and low sodium feeding periods. Values below the limit of detection were imputed as two-thirds the distance to zero.

Table 2.

Mean baseline, follow-up visits, and difference in concentrations of cardiac biomarkers by diet assignment, N = 411 unique persons

			Change from Baseline (95% CI)		Between Diet Comparison (95% CI)
	Baseline*	Follow-up*	Absolute†	%-Difference‡	Absolute†	%-Difference‡
High sensitivity cardiac troponin I, ng/L
Control	1.7 (0.2)	1.6 (0.2)	−0.04 (−0.2, 0.1)	−2.5 (−10.6, 6.3)	Ref	Ref
DASH	1.4 (0.1)	1.1 (0.1)	−0.2 (−0.3, −0.1)	−16.1 (−23.3, −8.3)	−0.2 (−0.4, −0.0)	−14.0 (−24.0, −2.6)
N-terminal pro-brain natriuretic peptide, pg/mL
Control	28.7 (1.9)	23.6 (1.4)	−5.1 (−7.3, −2.9)	−17.7 (−23.8, −11.2)	Ref	Ref
DASH	29.0 (1.9)	22.7 (1.4)	−6.3 (−8.5, −4.1)	−21.6 (−27.4, −15.4)	−1.1 (−3.7, 1.4)	−4.8 (−14.5, 6.1)
High sensitivity C-reactive protein, mg/L
Control	2.1 (0.2)	2.0 (0.2)	−0.03 (−0.2, 0.2)	−1.2 (−10.0, 8.5)	Ref	Ref
DASH	1.9 (0.2)	1.7 (0.1)	−0.3 (−0.5, −0.1)	−14.3 (−21.9, −5.9)	−0.3 (−0.5, −0.0)	−13.2 (−24.0, −1.0)

There were 1,545 high sensitivity troponin I and C-reactive protein measurements and 1,547 N-terminal pro-brain natriuretic peptide measurements.

^*Exponentiated log-transformed markers or the geometric mean (SE). In this table there were up to three follow-up visits contributing to the Follow-up geometric mean.

^†Difference in exponentiated log-transformed markers (i.e. difference in geometric means). This may be derived by taking the follow-up minus baseline for DASH and subtracting the follow-up minus baseline for control.

^‡Exponentiated differences of the change on the log-scale

Abbreviations: CI, confidence interval; DASH, Dietary Approaches to Stop Hypertension; N, number; Ref, reference; SE, standard error

Compared with the control diet, the DASH diet significantly reduced hs-cTnI by 14.0% and hs-CRP by 13.2%, but not NT-proBNP.

Effects of Sodium Reduction

While there were significant effects from diet assignment on hs-cTnI and hs-CRP within each strata of sodium intake (Table 3; Supplemental Tables 3 & 4), there was no evidence that diet modified the effects of a low versus high sodium level on any of the three biomarkers examined (P-interactions of 0.66, 0.17, and 0.76, respectively). Low versus high sodium did not significantly reduce hs-cTnI. There was no evidence of a trend in hs-cTnI across high, medium, and low sodium levels (P-trend = 0.22). In contrast, low versus high sodium intake reduced NT-proBNP (−19.2%). There was evidence of a significant decreasing trend across high, medium, and low sodium levels (P-trend < 0.001). Meanwhile, we observed an increase in hs-CRP with low versus high sodium (8.8%) with a significant increasing trend across high, medium, and low sodium levels (P-trend = 0.03).

Table 3.

Effects of sodium reduction on percent change in cardiac biomarkers from baseline and between diets, N = 411 unique persons

		%-Change from Baseline (95% CI)†			Low versus High (95% CI)†
	Baseline*	High Sodium	Medium Sodium	Low Sodium	%-Difference	P-trend
High sensitivity cardiac troponin I, ng/L	1.5 (0.1)	−7.7 (−14.5, −0.4)	−7.8 (−14.6, −0.4)	−12.4 (−18.8, −5.4)	−5.0 (−12.0, 2.6)	0.22
N-terminal pro-brain natriuretic peptide, pg/mL	28.8 (1.3)	−10.5 (−16.1, −4.5)	−20.3 (−25.3, −14.9)	−27.6 (−32.2, −22.7)	−19.2 (−24.2, −13.7)	<0.001
High sensitivity C-reactive protein, mg/L	2.0 (0.1)	−13.5 (−20.2, −6.3)	−4.3 (−11.7, 3.7)	−5.9 (−13.2, 2.0)	8.8 (0.4, 17.9)	0.03

^*Exponentiated log-transformed markers or the geometric mean (SE)

^†Exponentiated differences of the change on the log-scale

There were 1,545 high sensitivity troponin I and C-reactive protein measurements and 1,547 N-terminal pro-brain natriuretic peptide measurements. There were no significant interactions between diet and sodium. Sodium effects by diet are in Supplement Table ST3 & ST4.

Abbreviations: CI, confidence interval; DASH, Dietary Approaches to Stop Hypertension; N, number; SE, standard error

Combined effects of Sodium Reduction and DASH

Compared to baseline, the high sodium-control diet did not change hs-cTnI, NT-proBNP, or hs-CRP (Table 4). In contrast, the low sodium-DASH diet reduced hs-cTnI (−18.9), NT-proBNP (−28.4), and hs-CRP (−13.5). Compared to a high sodium-control diet, a low sodium-DASH diet, reduced hs-cTnI by 20.1% and NT-proBNP by 22.9%, but had a non-significant effect on hs-CRP.

Table 4.

Mean baseline, follow-up, and difference in concentrations of cardiac biomarkers comparing the low sodium-DASH diet to the high sodium-control diet, N = 410 unique persons

			Change from Baseline (95% CI)		Between Diet Comparison (95% CI)
	Baseline*	Follow-up*	Absolute†	%-Difference‡	Absolute†	%-Difference‡
High sensitivity cardiac troponin I, ng/L
High Sodium - Control	1.9 (0.2)	1.9 (0.2)	0.03 (−0.2, 0.2)	1.5 (−8.1, 12.2)	Ref	Ref
Low Sodium – DASH	1.6 (0.2)	1.3 (0.1)	−0.3 (−0.5, −0.1)	−18.9 (−27.0, −9.9)	−0.3 (−0.6, −0.1)	−20.1 (−30.9, −7.7)
N-terminal pro-brain natriuretic peptide, pg/mL
High Sodium - Control	28.7 (1.9)	26.7 (1.8)	−2.0 (−4.7, 0.6)	−7.1 (−15.4, 2.0)	Ref	Ref
Low Sodium – DASH	29.3 (1.9)	21.0 (1.4)	−8.3 (−10.9, −5.8)	−28.4 (−34.8, −21.4)	−6.2 (−9.7, −2.8)	−22.9 (−32.4, −12.0)
High sensitivity C-reactive protein, mg/L
High Sodium - Control	2.0 (0.2)	1.9 (0.2)	−0.1 (−0.4, 0.1)	−6.6 (−17.0, 5.1)	Ref	Ref
Low Sodium - DASH	1.9 (0.2)	1.7 (0.1)	−0.3 (−0.5, −0.0)	−13.5 (−23.0, −2.8)	−0.1 (−0.4, 0.2)	−7.4 (−21.6, 9.3)

There were 781 specimens contributing data for each of the biomarker comparisons.

^*Exponentiated log-transformed markers or the geometric mean (SE)

^†Difference in exponentiated log-transformed markers (i.e. difference in geometric means)

^‡Exponentiated differences of the change on the log-scale

Abbreviations: CI, confidence interval; DASH, Dietary Approaches to Stop Hypertension; N, number; Ref, reference; SE, standard error

Sensitivity Analyses

All the dietary effects on hs-cTnI were virtually identical when examined using the limit of blank rather than the limit of detection (Supplemental Tables 5–7).

Discussion

In this population of adults with above normal blood pressure, the DASH diet alone reduced hs-cTnI and hs-CRP compared to a typical American diet, while reduced sodium intake lowered NT-proBNP and mildly increased hs-CRP (Central Illustration). Compared to a typical American diet with high sodium, the combined DASH-low sodium diet reduced hs-cTnI and NT-proBNP, but not hs-CRP. Together, these findings imply that two distinct dietary strategies might improve two key pathways of subclinical cardiac damage: injury and strain.

Central Illustration. DASH Diet, Sodium Reduction, and Cardiac Biomarkers: Results from DASH-Sodium.

(Top left) Population description and study outcomes. (Top right) Study design. (Bottom left) Effects of the DASH diet (compared to control), reduced sodium (compared to higher sodium), and the combined effects of DASH and reduced sodium (compared to control and higher sodium) on cardiac biomarkers. (Bottom right) Study conclusions.

There is substantial observational evidence that the DASH diet decreases risk of CVD events in population studies (25–27). We recently showed that three DASH-pattern diets lowered hs-cTnI compared to baseline after a 6-week feeding period (11). In a subsequent analysis of the DASH trial, we further showed that 8-weeks of the DASH diet versus a controlled diet similarly reduced hs-cTnI (10). The present study is further evidence in yet a third population of adults with elevated BP and hypertension, that adoption of the DASH diet, in this case over a 12-week period, significantly reduced subclinical cardiac injury. In contrast, sodium reduction alone had no affect on hs-cTnI, which is consistent with at least one other trial of large mammals (28). Of note one community-based study of nearly 3,000 adults in the Netherlands found that lower 24-hour urine potassium excretion (a central DASH micronutrient) was significantly associated with higher high sensitivity cardiac troponin T and non-significantly associated with higher hs-cTnI, while lower urine sodium excretion was not associated with high sensitivity cardiac troponin T or hs-cTnI, consistent with the effects observed in our study (29).

Our prior work, using data from the DASH trial, demonstrated that the DASH diet lowered NT-proBNP (10). However, in that trial, the estimated nutrient content of the DASH diet appeared to be somewhat lower in sodium by ~200mg/d compared with the control diet, and thus we were unable to fully distinguish DASH effects from sodium reduction. The present analysis clarified these mechanisms, showing that while sodium reduction significantly lowered NT-proBNP, the DASH diet had no effect on NT-proBNP. This is consistent with other studies of sodium reduction and NT-proBNP in the context of reducing cardiac strain in patients with congestive heart failure (30). Our present study extends this literature by showing the potential for sodium reduction to reduce subclinical cardiac strain in adults without established CVD (31).

While many studies have shown that DASH-like diets are inversely associated with CRP (32–34), our prior work demonstrated mixed effects of the DASH diet on hs-CRP. In the OmniHeart trial, we found that healthy diets reduced inflammation from baseline (11). However, we did not observe this effect in the DASH trial (10). Notably, in a prior analysis of these data, using a less sensitive CRP assay, there was no effect from DASH on CRP (35). In the present analysis of the DASH-sodium trial, the DASH diet reduced hs-CRP whereas there was a mild inverse relationship between sodium reduction and hs-CRP. We speculate that this might explain the null findings in the original DASH trial in that the estimated sodium intake based on menus was ~200 mg/d less in the DASH diet compared to the control diet. A similar inverse relationship was observed in a study of 41 non-obese, hypertensive adults in which a 3-week period of 1380 mg/day of sodium increased hs-CRP compared to a 3-week period of 3680 mg/d of sodium (36). However, similar findings were not observed in a randomized crossover study of 32 uncontrolled hypertensive patients, where sodium intake of 460 versus 4600 mg/d did not affect hs-CRP (37). Mechanisms behind why sodium reduction would increase hs-CRP are unclear, but are thought to reflect the proinflammatory effects of increased aldosterone stimulated by low sodium intake (38).

Our study has limitations. First, the durations of the two interventions were relatively short with 12 weeks for the dietary pattern contrast and 4 weeks for each of the three sodium feeding periods. Thus, longer-term effects of the dietary interventions could not be evaluated in our study. Second, DASH was a dietary pattern, involving multiple food groups and micronutrients. This limits our ability to isolate specific dietary factors driving effects on cardiac injury and inflammation. Third, the range of sodium intake in our study was limited as the average sodium intake of the U.S. population corresponds roughly to the high sodium level (150 mmol/day or 3450 mg/day) in the DASH-sodium trial. Thus, we are unable to determine the potential effects of very high sodium intake on cardiac biomarkers consumed by a large proportion of the population in the US and throughout the world. Fourth, while prior community-based population studies have demonstrated strong positive associations between troponin I and CVD events, similar associations using the Siemens assay have yet to be established. Fifth, the mean hs-cTnI levels observed in our study were lower than levels often observed in patient populations. Thus, performance of the assay may be less precise than suggested by the assay insert. Nevertheless, we are reassured that even using the substantially less precise limit of blank yielded virtually identical findings. Finally, while protocol compliance was high, intake of non-study foods could bias findings (if differential) or attenuate the impact of the intervention.

Our study also has strengths. First, both dietary interventions were tightly controlled and administered independently of each other in randomized fashion. Second, the isocaloric design minimized the effects of weight loss on the markers. Third, we used highly sensitive markers of subclinical cardiovascular disease, thought to reflect the effects of CVD risk factors on the mechanisms of cardiac damage. These markers are strongly related to CVD events and elucidate distinct mechanisms of cardiac injury.

Our study has important clinical implications. Due to the expense and challenges of long-term, controlled feeding trials on CVD events, our study represents some of the strongest evidence that diet directly impacts cardiac damage. In particular, we show how two dietary strategies can improve distinct mechanisms of subclinical cardiac injury in a relatively short time period, suggesting that the improvements in CVD risk factors observed from a reduced sodium, DASH diet may also reduce concurrent cardiac damage. These findings reinforce the importance of a lifestyle that includes a reduced sodium, DASH diet to minimize cardiac damage over time.

In conclusion, the DASH diet reduced hs-cTnI and hs-CRP, while sodium reduction reduced NT-proBNP, but mildly increased hs-CRP. Combined together, the DASH diet and reduced sodium lowered hs-cTnI and NT-proBNP, biomarkers of subclinical cardiac injury and strain. These findings should strengthen public resolve for public policies that promote the DASH dietary pattern and lower sodium intake in the United States and globally.

Supplementary Material

1

Supplement Figure SF1. CONSORT Figure of Specimen Availability

Supplement Figure SF2. Effects of the DASH diet or sodium reduction on percent change in cardiac biomarkers

Supplement Table ST1. Nutrient composition by diet

Supplement Table ST2. Baseline median, 25th, and 75th percentiles for the cardiac biomarkers

Supplement Table ST3. Effects of sodium reduction on absolute changes in cardiac biomarkers.

Supplement Table ST4. Effects of sodium reduction on percent changes in cardiac biomarkers.

Supplement Table ST5. Effects of the DASH diet on high sensitivity cardiac troponin I (limit of blank and limit of detection)

Supplement Table ST6. Effects of sodium reduction on high sensitivity cardiac troponin I (limit of blank and limit of detection)

Supplement Table ST7. Effects of the low sodium-DASH diet versus the high sodium-control diet on high sensitivity cardiac troponin I (limit of blank and limit of detection)

Supplement Figure ST1.

Supplement Figure SF2. Percent changes (95% confidence intervals) from baseline for (A) high sensitivity cardiac troponin I (hscTnI), a marker of cardiac injury, (B) N-terminal pro-brain natriuretic peptide (NT-proBNP), a marker of cardiac strain, and (C) high sensitivity C-reactive protein (hs-CRP), a marker of inflammation according to diet assignment after high, medium, and low sodium feeding periods. All changes reflect the limit of detection.

Supplement Table ST1. Nutrient composition of the two diets used in DASH-Sodium

Supplement Table ST2. Baseline median, 25th, and 75th percentiles for the cardiac biomarkers, N= 390

Supplement Table ST3. Effects of sodium reduction on absolute change in cardiac biomarkers from baseline and between diets, N = 411 unique persons

Supplement Table ST4. Effects of sodium reduction on percent change in cardiac biomarkers from baseline and between diets, N = 411 unique persons

Supplement Table ST5. Mean baseline, follow-up, and difference in concentrations of high sensitivity troponin I by diet assignment, comparing the limit of blank (sensitivity) with the limit of detection (primary), N = 411 unique persons

Supplement Table ST6. Effects of sodium reduction on percent change in high sensitivity cardiac troponin I from baseline and between diets, comparing the limit of blank (sensitivity) with the limit of detection (primary), N = 411

Supplement Table ST7. Mean baseline, follow-up, and difference in concentrations of high sensitivity cardiac troponin I, comparing the effects of the low sodium-DASH diet to the high sodium-control diet on the limit of blank (sensitivity) or the limit of detection (primary), N = 410 unique persons

Perspectives.

Competency in Patient Care:

The DASH diet and dietary sodium reduction lower high sensitivity cardiac troponin I and N-terminal B-type pro-natriuretic peptide, which are associated with subclinical cardiac damage and long-term cardiovascular risk.

Translational Outlook:

More pervasive public policies are needed to promote reduction in dietary sodium intake and broadly encourage the DASH eating pattern.

Abbreviations:

BMI

body mass index

CI

confidence interval

CVD

cardiovascular disease

DASH

Dietary Approaches to Stop Hypertension

DBP

diastolic blood pressure

hs-CRP

high-sensitivity C-reactive protein

hs-cTnI

high-sensitivity cardiac troponin

LDL-c

low density lipoprotein cholesterol

NT-proBNP

N-terminal b-type pro natriuretic peptide

SBP

systolic blood pressure