Short-term high-salt consumption does not influence resting or exercising heart rate variability but increases MCP-1 concentration in healthy young adults

Abstract

High salt consumption increases blood pressure (BP) and cardiovascular disease risk by altering autonomic function and increasing inflammation. However, it is unclear whether salt manipulation alters resting and exercising heart rate variability (HRV), a noninvasive measure of autonomic function, in healthy young adults. The purpose of this investigation was to determine whether short-term high-salt intake 1) alters HRV at rest, during exercise, or exercise recovery and 2) increases the circulating concentration of the inflammatory biomarker monocyte chemoattractant protein 1 (MCP-1). With the use of a randomized, placebo-controlled, crossover study, 20 participants (8 females; 24 ± 4 yr old, 110 ± 10/64 ± 8 mmHg) consumed salt (3,900 mg sodium) or placebo capsules for 10 days each separated by ≥2 wk. We assessed HRV during 10 min of baseline rest, 50 min of cycling (60% V̇o_2peak), and recovery. We quantified HRV using the standard deviation of normal-to-normal RR intervals, the root mean square of successive differences (RMSSD), and additional time and frequency domain metrics of HRV. Plasma samples were collected to assess MCP-1 concentration. No main effect of high salt or condition × time interaction was observed for HRV metrics. However, acute exercise reduced HRV (e.g., RMSSD time: P < 0.001, condition: P = 0.877, interaction: P = 0.422). High salt elevated plasma MCP-1 (72.4 ± 12.5 vs. 78.14 ± 14.7 pg/mL; P = 0.010). Irrespective of condition, MCP-1 was moderately associated (P values < 0.05) with systolic (r = 0.32) and mean BP (r = 0.33). Short-term high-salt consumption does not affect HRV; however, it increases circulating MCP-1, which may influence BP in young adults.

NEW & NOTEWORTHY In healthy young adults, 10 days of dietary salt loading does not affect heart rate variability before, during, or after submaximal aerobic exercise. However, salt loading did increase the circulating concentration of the inflammatory cytokine monocyte chemoattractant protein 1 (MCP-1). Irrespective of the condition, MCP-1 was associated with resting blood pressures in our healthy young cohort.

INTRODUCTION

At the population level, excess dietary salt consumption increases blood pressure (BP) (1, 2) and cardiovascular disease (CVD) risk (3–5). Individuals who are salt-sensitive (i.e., increased BP with high-salt intake) are at increased risk for CVD events and mortality (6). However, even those who are salt-resistant may experience adverse cardiovascular consequences with high-salt loading. For example, short-term (10 days) salt loading augments systolic BP responses during dynamic submaximal exercise in healthy young adults despite no increases in resting BP (7). Elevated serum sodium, as a result of hypertonic saline infusion, also augments systolic BP during static exercise in healthy young adults (8). Furthermore, a salt-restricted diet decreases BP during acute static and dynamic submaximal exercise in people with hypertension (9). This is important because exaggerated cardiovascular reactivity during stressors, such as exercise, is prognostic of future hypertension (10, 11), coronary artery disease risk (12), and CVD mortality (11).

Another important cardiovascular variable that may be influenced by salt, irrespective of resting BP, is heart rate variability (HRV). HRV is the physiological variation in the time interval between heartbeats (13, 14). Specifically, HRV represents the ability to adjust to the demands placed by internal and external environments through the dynamic balancing of the sympathetic and parasympathetic branches of the autonomic system (15–18). In healthy people at rest, HRV is relatively high due to the normal dynamic balance between the autonomic branches (19). During acute exercise, HRV decreases due to an imbalance of autonomic activity favoring increased sympathetic outflow (20). Specifically, there is parasympathetic withdrawal and increased sympathetic outflow to increase cardiac output and systemic vascular tone to effectively meet the physiological demands of the exercise bout (20–22). In contrast, a sedentary lifestyle and hypertension decrease resting HRV, which is associated with higher risk for CVD and all-cause mortality (14, 15, 23–25).

A prior short-term salt manipulation study in middle-aged and older female participants demonstrated that compared with normal and high-salt diets, a low-salt intervention improved LF/HF (23), a measure of HRV thought to represent the balance between sympathetic and parasympathetic activity (26). In contrast, we previously demonstrated that short-term salt loading did not affect HRV at rest (27). Nonetheless, there is limited evidence regarding the influence of high dietary salt on HRV during or following (i.e., recovery from) acute exercise. Importantly, HRV may provide additional information regarding cardiovascular control during and following exercise (20, 28, 29). Therefore, the purpose of this study was to examine the effects of short-term high-salt consumption on HRV at rest, during submaximal exercise, and during recovery. Given our prior findings of augmented exercising BP, impaired postexercise hypotension, and vascular dysfunction (7, 27), we hypothesized that short-term salt loading would reduce exercising and recovery, but not resting HRV.

Related to physiological responses to salt loading that can influence cardiovascular regulation during exercise, excessive salt consumption increases circulating concentrations of the chemokine, monocyte chemoattractant protein 1 (MCP-1) (30), among other proinflammatory cytokines. MCP-1 has also been associated with BP in some (31), but not all (32), cohort studies, though underlying mechanisms remain inconclusive. MCP-1 is also negatively associated with metrics of HRV in mice. Specifically, compared with controls, MCP-1 knockout mice exhibit improved HRV metrics (33). From a clinical perspective, elevated MCP-1 is associated with CVD morbidity and mortality (34). Therefore, an additional aim of this study was to extend previously observed increases in circulating MCP-1 following short-term high-salt loading in an all-male cohort (30) to our cohort inclusive of male and female participants. Furthermore, we sought to examine whether there were associations between MCP-1 and BP or HRV metrics in healthy young adults.

METHODS

The data presented in this manuscript are derived from a secondary analysis of stored electrocardiogram (ECG) recordings for HRV data and a post hoc assessment of MCP-1 concentration from stored EDTA-treated plasma samples. These data were collected as part of a registered clinical trial (NCT03565653; conducted between 2017 and 2019). Importantly, this analysis is unique relative to three prior publications that addressed separate research questions and provide additional detail on the parent study (7, 27, 35). Plasma samples were obtained by investigators at Auburn University under a material transfer agreement from the University of Delaware.

Participants

All participants provided written and verbal consent before engaging in any study activities. The study protocol and procedures were approved by the Institutional Review Board of the University of Delaware and conform to the provisions of the Declaration of Helsinki. Twenty participants (eight females) were included in this study. Participants’ age ranged from 18 to 34 yr. Exclusion criteria included a physician’s diagnosis of hypertension, diagnosis of CVD, cancer, type 1 or type 2 diabetes, kidney disease, current pregnancy, obesity (body mass index > 30 kg/m²), and current or recent (within the preceding 6 mo) use of tobacco products.

Salt Intervention

We used a randomized double-masked, placebo-controlled, crossover study design. Participants were instructed on how to interpret nutrition labels by a member of the research team (M.C.B.) and instructed to consume a diet of the recommended sodium intake (2,300 mg/day) during both 10-day intervention periods. Participants consumed unmarked capsules each day containing either salt [Morton table salt (NaCl); 3,900 mg sodium/day] or a placebo (NOW Foods dextrose). Thus, total sodium intake was designed to be 6,200 mg/day during the high-salt condition and 2,300 mg/day during the dextrose condition. Prior publications indicate that this duration of salt manipulation is sufficient to elicit physiological effects, including studies utilizing this salt capsule model (7, 27, 35–37). Condition order was randomized with a washout period of at least 2 wk. Prior data demonstrate that randomized order salt manipulation diets with no washout period still result in demonstrable changes in vascular and autonomic function across diets (38, 39). Thus, we are confident in the duration of our washout period to avoid carryover effects. All female participants were using oral hormonal contraceptives, and experimental visits occurred 2–6 days into their inactive week (i.e., when an individual takes their placebo pills rather than the contraceptive). Participants recorded their diet during the first intervention. They were provided a copy of their initial diet log and asked to match their diet for macro- and micronutrient composition during the second intervention. We analyzed three days of diet records from a subset of participants (n = 13) including two weekdays and a weekend day using Nutrition Data System for Research to assess potential differences in background diet between conditions.

Twenty‐Four‐Hour Urine Collection

Urine was collected during the final 24 h of both interventions in a light‐protected, sterile 3,500-mL container. Participants returned the container on arriving at the laboratory for the experimental visit. We measured total urine volume, urine specific gravity (Goldberg Brix Refractometer, Reichert Technologies), urine electrolyte concentrations (EasyElectrolyte Analyzer, Medica), and urine osmolality (Advanced 3D3 Osmometer, Advanced Instruments) from a mixed aliquot from the 24‐h collection container. Urine flow rate was calculated (urine volume [mL]/urine collection time [min]) and used to determine 24‐h sodium excretion. Participants were instructed to abstain from alcohol, caffeine, and exercise for 24 h before and during the 24‐h urine collection.

V̇o_2peak

As previously described (7), participants performed cycling exercise to volitional fatigue (Lode CPET, Lode, The Netherlands) using an established ramped protocol (40). Briefly, participants began cycling at a constant power of 30 W for 3 min. After the 3‐min warm-up, power increased by 1 W every 2 s until participants were unable to maintain a pedaling cadence of at least 60 revolutions per minute. Oxygen consumption and carbon dioxide production were measured and averaged in 15‐s intervals using indirect calorimetry via an automated open circuit system (Parvo Medics, Sandy, UT) throughout the exercise test. Gas analyzers were calibrated with standard gases (16% O₂, 4.05% CO₂) and the pneumotach was calibrated using a standard volume of air (3 L) before each test. Heart rate (HR) was monitored via a chest‐worn HR monitor (Polar H7, Polar). The highest V̇o₂ over a 30-s average was used as V̇o_2peak. A successful test included achieving at least three of the following: a plateau in V̇o₂ with increasing workload; a maximal achieved HR within 10 beats per minute of age-predicted maximal HR (220 – age); a rating of perceived exertion (Borg 6–20 scale) of ≥ 18; and a respiratory exchange ratio (CO₂:O₂) of ≥1.1 (41). After the V̇o_2peak test and a brief rest period (∼10 to 15 min), participants resumed cycling and the workload corresponding to 60% V̇o_2peak was determined via V̇o₂ measurement. This workload was used during the experimental visits for the 50‐min exercise bouts (described below).

Study Visits

On the tenth day of each diet, participants reported to the Cardiovascular Physiology Laboratory at the University of Delaware. On arrival, participants provided a spot urine sample for assessment of hydration status via urine-specific gravity (Goldberg Brix Refractometer, Reichert Technologies). Samples from female participants were tested using hormonal pregnancy tests (Moore Medical) to ensure that they were not pregnant. Body mass, body composition, and total body water were measured via bioelectrical impedance (Tanita Body Composition Analyzer, model TBF‐300A; Arlington Heights, IL). Participants were instrumented for single-lead ECG and oscillometric BP measurements on the upper right arm (Dash 2000, GE Medical Systems). Participants were scheduled at approximately the same time of day (±1 h) to account for circadian variation in cardiovascular measures. Because of the length of the visits and exercise component, participants were provided Nature Valley snack bars and instructed to eat them 1 to 2 h before study visits. Although the snack may have affected HRV metrics compared with consuming nothing or just drinking water (42), the within participant data should not be confounded as this practice was instructed for both visits.

Exercise Protocol

As previously reported (7, 27), we utilized a steady‐state submaximal (60% of V̇o_2peak) exercise protocol and HR was monitored continuously. Briefly, after instrumentation, participants rested quietly in a dimly lit room for 10 min before a blood draw and BP and vascular measures reported elsewhere (7, 27, 35). After this period of supine rest, participants moved to an upright cycle ergometer (Lode Excalibur, Lode). Participants began submaximal aerobic exercise (60% V̇o_2peak) with a 5‐min warm‐up period at self‐selected resistance. The resistance was then increased to the power output determined during the V̇o_2peak test (described in v̇o_2peak) for 50 min. After 50 min of exercise at 60% V̇o_2peak, participants completed 5 additional minutes of aerobic exercise at a self‐selected resistance as a cool down. Exercise workloads for warm-up, exercise, and cool down were matched between conditions. Participants pedaled at their preferred cadence above 60 revolutions per minute, and resistance was continuously adjusted to maintain a constant power output. In addition, ratings of perceived exertion (Borg 6–20 scale) were recorded every 5 min. Participants were permitted to drink water ad libitum during the first exercise trial and water intake during exercise and cool-down intensities were matched for the second exercise trial. Exercise was followed by 60 min of recorded supine recovery.

HRV Quantification

HRV was quantified from ECG readings during baseline, exercise, and recovery. All ECG readings were recorded into LabChart 8 Pro (43) and analyzed using Kubios HRV Scientific 4.0.1 software. HRV was initially analyzed (B.A.L.) and inspected (A.T.R.) by trained investigators. We used ECG from the following time points; all 10 min of baseline, minutes 30–40 of exercise, and the last 10 min of recovery (50 to 60 min postexercise). Automatic noise detection was set to “medium” within the program and analyses were done across all 10 min of each 10-min ECG section. Importantly, study visit start times, and thus ECG recordings were standardized to start at the same time of day for both visits. Our analysis included time and frequency domain HRV measures. Our time domain measures included average RR interval, HR, the standard deviation of RR intervals (SDRR), percent of detected RR intervals greater than 50 ms different from the immediately preceding RR interval (pRR50%), and root mean square of successive differences (RMSSD). Our frequency domain measures included very low-frequency power (VLF ms²) and contribution (VLF%), low-frequency power (LF ms²) and contribution (LF%), high-frequency power (HF ms²), and contribution (HF%), and the low-frequency to high-frequency contribution ratio (LF/HF). We defined the range of each frequency band as VLF range: <0.04 Hz; LF range 0.04–0.15 Hz; HF range 0.15–0.40 Hz; >HF range: >0.40 Hz (19, 20).

Biomarker Assessments

Venous blood samples were collected in silicone-coated vacuum blood collection tubes (for serum analyses) and K⁺ EDTA-treated vacuum blood collection tubes (for plasma and whole blood analyses). The serum and plasma samples were centrifuged at 750 g for 15 min at 4°C (Allegra X-22R, Beckman Coulter). Serum electrolytes (EasyElectrolyte Analyzer, Medica), plasma osmolality (Advanced 3D3 Osmometer, Advanced Instruments), hemoglobin (Hb 201+, HemoCue), and hematocrit (Sure prep capillary tubes, Clay Adams spun in a microcentrifuge at 1,950 g for 5 min, Legend Micro 17, Thermo Sorvall) were measured. The estimated change in plasma volume between conditions was calculated using the following equation, which is based on changes in hemoglobin (Hb) concentration and hematocrit (Hct) between conditions (44):

$$\text{Plasma\hspace{0.17em}volume}(\%)=\left\{\right[100\times ({\text{Hb}}_{\text{dextrose}}/{\text{Hb}}_{\text{salt}})]\times \hspace{0.17em}[1-({\text{Hct}}_{\text{salt}}/100)]/(1-{\text{Hct}}_{\text{dextrose}}/100\left)\right\}-100$$

The remaining serum and plasma samples were stored at −80°C for future analysis. Plasma MCP-1 concentration was quantified using the Human CCL2/MCP-1 Quantikine Enzyme Linked Immunosorbent Assay Kit (R&D Systems, Minneapolis MN, Catalog Number: DCP00) at Auburn University by a single investigator (B.A.L). Samples were run in triplicate across two 96-well plates. Average intra-assay sample CV was 3.82%, and quality controls indicated adequate precision (Control group DCP00, RD6Q [expected range] actual values obtained across two kits); Control 1 [137–291], 138.9 pg/mL, 223.0 pg/mL; Control 2 [539–879], 773.2 pg/mL, 815.6 pg/mL; Control 3 [1,091–1,781], 1434.3 pg/mL, 1645.5 pg/mL. The absorbance from 96-well plates from the ELISA kits was read on a SpectraMax iD3 plate reader (Molecular Devices, San Jose, CA) at 450 nm.

Statistical Analysis

Outliers were identified when outcome variables were greater than 1.5 times the interquartile range outside of the third and first quartiles of each variable and subsequently removed from all analyses and are not presented in any figures (45–47). We inspected all variables for normality using the Shapiro-Wilk test and QQ plots to compare the shapes of distributions. Urine sodium excretion, urine potassium excretion, urine chloride excretion, serum potassium, SDRR, pRR50%, LF/HF, VLF (ms²), LF (ms²), and HF (ms²) were not normally distributed. Two-tailed, paired sample t tests were conducted to assess the effect of salt loading on all baseline characteristics and resting BP measures, HRV metrics, and MCP-1 concentrations when data were normally distributed. Wilcoxon-matched tests were used when paired data were not normally distributed. Cohen’s d was used to assess effect size. Effects sizes were interpreted as small (d = 0.2), medium (d = 0.5), and large (d = 0.8) (48). Nonparametric effect sizes are represented as rank biserial correlation Mixed effects models were conducted using conditions (dextrose and salt) and time points (baseline, exercise, and recovery) as fixed factors to determine the effects of salt loading and activity levels on metrics of HRV. All data are presented as raw data, irrespective of normality. Sidak’s multiple comparisons were used to determine pairwise comparisons on detecting significant main or interaction effects. Finally, Pearson’s correlations were completed to investigate the relation between MCP-1, BP measures, and HRV metrics. An a priori power analysis was not conducted as this was a secondary data analysis. Statistical analyses were completed using jamovi 2.0.0 (49), GraphPad 9.3.1 (GraphPad Prism Software, San Diego, CA) with a priori significance set to P < 0.05.

RESULTS

Demographics

Participant characteristics from their screening visit are presented in Table 1. Participants were young adults free of clinical hypertension, CVD, cancer, metabolic syndrome, diabetes, and kidney disease diagnoses. Most participants fell within the healthy range of body mass index. On average, participants had normal BP values. Regarding baseline sex differences in variables of interest, we conducted exploratory sex comparisons on data collected during the placebo visit. There were sex differences observed at rest, with males exhibiting higher LF (%), females exhibiting higher HF (%), and thus males exhibiting higher LF/HF.

Table 1.

Participant descriptive statistics at screening

Variable	Means ± SD	Range
n (female/male)	20 (8/12)
Race and ethnicity	13 NHW, 1 BR, 4 A, 2 L
Age, yr	24 ± 4	18–34
Height, cm	173 ± 10	152–185
Body mass, kg	69.4 ± 12.5	48.9–96.8
Body mass index, kg/m2	23.1 ± 2.6	18.4–28.9
Brachial systolic BP, mmHg	112 ± 10	86–125
Brachial diastolic BP, mmHg	64 ± 8	64–92
Brachial mean BP, mmHg	80 ± 8	53–76
V̇o2peak, mL/kg/min	40.3 ± 9.9	23.9–59.9

All data are presented as means ± SD. A, Asian; BR, biracial; L, Latinx; NHW, non-Hispanic white.

Effect of Salt on Resting Cardiovascular Outcomes

The effects of salt loading on resting cardiovascular measures, blood and urine biomarkers, and self-reported nutrient intake are described in Table 2. As expected, there was an increase in 24-h urine osmolality and sodium excretion following the 10-day high-salt intervention compared with the placebo intervention. There were no differences in plasma osmolality and serum sodium between conditions. However, we did observe a significant increase in estimated plasma volume following the 10-day high-salt intervention relative to the placebo intervention. No differences were detected between the high-salt and placebo conditions in systolic BP, diastolic BP, or mean BP.

Table 2.

Comparison of resting BP measures and fluid/salt regulatory biomarkers between the placebo and high-salt intervention conditions

Variable	Placebo	High Salt	P Value	Effect Size
Blood pressure
Brachial systolic BP, mmHg	108 ± 11	108 ± 8	0.969	0.009
Brachial diastolic BP, mmHg	61 ± 7	61 ± 6	0.831	0.048
Brachial mean BP, mmHg	77 ± 7	77 ± 6	0.897	0.029
Biomarkers
Urine osmolality, mosmol/kg H2O	476 (213)	568 (310)	0.006	0.686
Urine Na+, mmol/24 h	117.7 (88.9)	264.0 (81.7)	<0.001	0.905
Urine K+, mmol/24 h	62.0 (42.9)	72.1 (29.6)	0.004	0.367
Urine Cl−, mmol/24 h	61.15 (30.8)	137.5 (95.6)	<0.001	0.962
Plasma osmolality, mosmol/kgH2O	294 ± 5	294 ± 5	0.691	0.090
Serum Na+, mmol/L	141.0 ± 1.8	141.4 ± 2.2	0.320	0.228
Serum K+, mmol/L	4.0 (0.4)	3.9 (0.5)	0.462	0.209
Serum Cl−, mmol/L	104.8 ± 2.0	106.2 ± 1.9	0.016	0.589
Plasma volume Δ, %		7.2 ± 11.3	0.012
Self-reported nutrients	Placebo	High salt
Calories	1,875 ± 459	1,921 ± 315	0.827	0.049
Dietary Na+, mg/day	2,506 ± 779	2,752 ± 554	0.493	0.155
Dietary K+, mg/day	2,279 ± 710	2,220 ± 691	0.871	0.036

Parametric data are presented as means ± SD. Nonparametric data are presented as median (interquartile range). Parametric effect sizes are represented as Cohen’s d. Nonparametric effect sizes are represented as rank biserial correlation. Bold text indicates significance at P < 0.05. BP, blood pressure; Cl⁻, chloride; K⁺, potassium; Na⁺, sodium.

Time domain HRV.

Time domain HRV data are presented in Fig. 1. There was not a significant main effect of the condition (placebo vs. high salt) or an interaction effect on any of the time domain HRV metrics. However, there was a significant main effect of time point on all metrics of time domain HRV. On an exploratory basis, we completed factorial ANOVAs (time × condition) within sexes to examine sex-specific differences in HRV measures and MCP-1 with salt loading. There were no sex-specific differences for time domain HRV measures.

Figure 1.

Visual representation of a two-way repeated-measures ANOVA on time domain HRV measures between time points (baseline vs. exercise vs. recovery) and condition (placebo and high-salt intervention). Data displayed as individual data points superimposed on means ± SD. Orange bars and circles represent placebo, and blue bars and squares represent the high-salt condition (n = 20; 11 males, 8 females). HRV, heart rate variability.

Frequency domain HRV.

Frequency domain measures of HRV are presented in Table 3. We split frequency domain HRV measures into relative power (ms²) and the relative contribution of each band to the total power as percentages (%). Regarding power measures (ms²) all HRV metrics were significantly affected by time point but not the experimental condition. These data are also visually depicted in Supplemental Figs. S1 and S2. Regarding the percent of the contribution of power measures (%), VLF% and LF% were, and HF% and LF/HF were not, affected by time point (Table 3). Across all frequency domain HRV metrics, VLF% was the only measure affected by condition and LF (ms²) was the only measure where there was an interaction for time × condition (Table 3). These data are also presented in Supplemental Figs. S3 and S4. Although the data are not shown, there were sex-specific time × condition interactions whereby females experienced an interaction for LF (%) and LF/HF, but males did not.

Table 3.

Comparison of frequency domain table data compared between condition and time points

	Baseline	Exercise	Recovery	Time Point	Condition	Interaction
VLF, ms2
Placebo	80.28 ± 50.73	2.32 ± 1.85	204.6 ± 197.6	P < 0.001	P = 0.366	P = 0.170
High salt	162.4 ± 143.8	1.27 ± 0.78	186.1 ± 131.0
LF, ms2
Placebo	872.0 ± 574.8	11.1 ± 8.58	2,733 ± 2,341	P < 0.001	P = 0.813	P < 0.001
High salt	2,248 ± 2,129	6.03 ± 5.71	1,209 ± 395.1
HF, ms2
Placebo	1,568 ± 795.2	18.5 ± 17.9	2,275 ± 1,946	P < 0.001	P = 0.632	P = 0.182
High salt	2,333 ± 1,520	5.05 ± 2.76	1,951 ± 1,923
VLF, %
Placebo	4.68 ± 3.52	7.87 ± 5.26	4.46 ± 2.36	P < 0.001	P = 0.024	P = 0.409
High salt	5.60 ± 4.60	12.0 ± 7.26	6.49 ± 4.68
LF, %
Placebo	36.5 ± 14.3	36.8 ± 10.6	48.0 ± 14.0	P = 0.039	P = 0.793	P = 0.554
High salt	40.5 ± 16.3	39.0 ± 15.6	45.4 ± 17.6
HF, %
Placebo	58.2 ± 16.9	51.0 ± 18.9	44.7 ± 8.61	P = 0.162	P = 0.668	P = 0.620
High salt	52.3 ± 18.7	48.7 ± 19.2	48.1 ± 19.6
LF/HF
Placebo	0.70 ± 0.49	0.69 ± 0.40	1.06 ± 0.52	P = 0.439	P = 0.173	P = 0.289
High salt	0.96 ± 0.60	1.12 ± 0.93	1.01 ± 0.71

Parametric data are presented as means ± SD. Nonparametric data are presented as median (interquartile range). Bold text indicates significance at P < 0.05. HF, high frequency; LF, low frequency; VLF, very low frequency.

Monocyte Chemoattractant Protein-1

MCP-1 data are presented in Fig. 2. We detected a significant increase in plasma MCP-1 concentration following 10 days of high salt compared with placebo (Fig. 2A). Irrespective of condition, there were significant correlations between MCP-1 concentration and resting systolic BP (Fig. 2B) and mean BP (Fig. 2D), but not diastolic BP (Fig. 2C). Regarding Fig. 2, we completed factorial ANOVAs (time × condition) within sex to examine sex-specific differences in HRV measures and MCP-1 with salt loading. Interestingly, it appears that males drove the difference observed in MCP-1 from placebo to high salt (within male comparison P = 0.005, within females P = 0.688), hence we have used separate symbols for male and female participants. We also investigated whether there were any relations between MCP-1 and our HRV measures. Because both measures exhibiting high interindividual variability and having no clinical cut points, primary outcome variables were converted to relative change scores (i.e., percentages) using the equation: [(High Salt – Placebo)/Placebo] × 100. There were no significant relations between MCP-1 and any baseline measures of HRV (Ps > 0.126). For baseline time domain metrics, we assessed relations with plasma MCP-1 and average RR (r = −0.26, P = 0.320), SDRR (r = −0.14, P = 0.595), average HR (r = 0.22, P = 0.407), RMSSD (r = −0.13, P = 0.613), and pRR50% (r = −0.10, P = 0.709). For baseline frequency domain metrics, we assessed relations with plasma MCP-1 including VLF (ms²) (r = 0.05, P = 0.851), LF (ms²) (r = 0.06, P = 0.831), HF (ms²) (r = 0.11, P = 0.687), VLF (%) (r = 0.04, P = 0.876), LF (%) (r = −0.15, P = 0.571), HF (%) (r = 0.05, P = 0.844), and LF/HF (r = −0.15, P = 0.972).

Figure 2.

Visual representation of a paired t test between MCP-1 concentrations between the placebo condition short-term high-salt loading and placebo-controlled conditions (A). For A, data are displayed as individual data points superimposed on means ± SD. For B–D, Pearson’s correlations are presented between MCP-1 and systolic BP (B), diastolic BP (C), and mean BP (D). Orange bars (A) and circles (B–D) represent placebo and blue bars (A) and squares (B–D) represent the high-salt condition (n = 20; 12 males, 8 females). BP, blood pressure; MCP-1, monocyte chemoattractant protein-1.

DISCUSSION

In the present study, we investigated the influence of short-term high-salt consumption on HRV metrics at baseline, during submaximal aerobic exercise, and during recovery from acute exercise. Our main findings are that acute exercise reduced HRV, but that short-term high-salt consumption for the most part did not influence HRV metrics at baseline, during aerobic exercise, or after aerobic exercise in healthy young adults. We also found that short-term high-salt consumption increases the circulating concentration of the chemokine MCP-1. Furthermore, MCP-1 was correlated with resting brachial systolic and mean BP, irrespective of condition (i.e., salt capsules vs. placebo). Notably, the increase in MCP-1 with high-salt intake occurred in a cohort that on average exhibited salt-resistant BP (see Table 2). This highlights a potentially novel deleterious effect of high dietary salt on human physiology, even in adults who do not exhibit an increase in resting BP.

Our finding that short-term high-salt consumption largely did not influence HRV metrics is somewhat in contrast to previous data indicating that salt loading leads to deleterious cardiovascular responses during and after acute exercise. For example, we previously demonstrated that short-term salt loading leads to augmented systolic BP responses during dynamic submaximal exercise in healthy young adults despite no increase in resting BP (7). In addition, a salt-restricted diet decreases BP during acute static and dynamic submaximal exercise in people with hypertension (9). We also recently reported that short-term salt loading suppresses postexercise hypotension in healthy young adults (27). However, in the same cohort, we demonstrated that short-term salt loading did not affect baseline HRV (27). In the current investigation, we have extended these findings with additional measures of HRV. We also assessed HRV during acute exercise and recovery from exercise. Although salt loading did not influence HRV when compared with placebo conditions, our finding that acute exercise altered HRV is consistent with the literature (20). Specifically, all time domain measures of HRV were influenced by acute exercise, which aligns with the current literature (22). We also observed no changes between conditions for RMSSD. RMSSD has been strongly correlated with one aspect of nonlinear HRV measures (19), so although we did not assess nonlinear HRV, the lack of change observed in our analysis provides prospective insight into how salt and exercise may not strongly affect nonlinear metrics of HRV, at least in healthy young adults.

Regarding spectral metrics, the VLF power band is influenced by acute physical activity (50) and also may be affected by physiological responses to exercise such as alterations in hormones, shear stress, and neural control (19). The LF power band is influenced by both branches of the autonomic nervous system and potentially by changes in ventilation (19). We did not detect any pairwise differences between conditions in the contribution of LF and HF in response to exercise in our analysis. However, nearly all measures of frequency domain HRV were substantially reduced by acute exercise (Table 3), which is consistent with previous findings (20, 22). Specific to power band contribution, VLF% and LF% contribution were, and HF% and LF/HF were not, affected by time point. Interestingly, we demonstrated that HRV responses in VLF% (condition effect) and LF (ms²) (interaction effect) and may also be influenced by high-salt consumption, despite our largely null findings. Nonetheless, these data do contribute to the growing body of literature characterizing the effects of high dietary salt on cardiovascular responses during acute exercise.

A prior cross-sectional study demonstrated that participants with normal BP or prehypertensive BP who had a preference for diets containing more salt had increased HR, systolic BP, and diastolic BP and had lower HRV (51). These findings and prior data demonstrate that high dietary salt augments BP responses during exercise and indicate that excessive salt consumption can induce cardiovascular dysregulation, even in those at lower risk for CVD (4, 7, 8, 36, 38). The reasons we observed no differences in baseline, exercise, or recovery HRV metrics between placebo and high-salt conditions in the present investigation may include the age and relatively good health of our participants. For example, high-salt loading may alter HRV metrics at baseline, during exercise, or recovery in older adults or patient populations. Though BP reactivity is prognostic for future cardiovascular event risk (38), HRV reflects autonomic inflections specifically on the heart rhythm. In addition to these analyses assessing the autonomic influence on HRV during sustained exercise, ultrashort-term [< 5-min ECG recordings (52)] HRV measures have also been used to assess cardiac-autonomic control in endurance, power, and tactical athletes (53). Time-domain ultrashort-term HRV metrics have been proposed to be a valid and more convenient method of assessing autonomic function and could also be utilized as a means of investigating whether high-salt diets impact health and performance on a day-by-day basis.

Another novel component of our study was the investigation of the relation between the cytokine MCP-1 with BP and HRV. Importantly, proinflammatory cytokines, such as MCP-1, have been implicated in the pathogenesis of atherosclerosis (54–56). MCP-1 [also known as chemokine (C-C motif) ligand 2 [CCL2]] is responsible for the initial transmigration of monocytes across the arterial wall (57, 58). Experimental studies suggest that targeting MCP-1 signaling attenuates atherosclerosis progression (59–61). Our data expand on previous findings that short-term high-salt consumption increased circulating MCP-1 concentration in apparently healthy males to also include females (30), and that increased MCP-1 is associated with resting BP (31). These data may serve to bridge our understanding of the role high-salt diets may play in inducing proinflammatory responses through chemokine expression, like MCP-1, in endothelial and immune cells. For example, our group and others have reported that short-term high-salt consumption leads to endothelial dysfunction even in healthy young and salt-resistant cohorts (4, 7, 62–64). Specifically, MCP-1 strongly binds to C-C chemokine receptor 2 (CCR2), which has two different isoforms. CCR2A is predominantly expressed by immune and vascular smooth muscle cells, whereas CCR2B is predominantly expressed specifically by monocytes and activated natural killer cells (57). MCP-1 has approximately five times greater affinity to the CCR2B isoform. In addition, antagonistic treatment of the CCR2B isoform delays the development of hypertension in rats (65). Although MCP-1 is upregulated in response to high salt, future studies are needed to examine how upregulated MCP-1 may differentially interact with the two isoforms of CCR2 (receptors on immune cells and vascular smooth muscle) and the resulting inflammatory responses. Furthermore, a recent pilot study (66) observed that the dietary approaches to stop hypertension (i.e., DASH) diet significantly and substantially decreased circulating levels of MCP-1 in patients with stable coronary artery disease and alleviated hallmarks of atherosclerotic pathogenesis. The authors also observed significant effects of fiber and vitamin C consumption on MCP-1 concentration, both of which may have provided a synergistic effect in addition to low-salt consumption.

One limitation of our analysis was that during exercise, we faced technical difficulties in keeping clear ECG readings (e.g., increased movement and muscular electrical interference, reduced adherence of electrodes to the skin, etc.), which is why we focused on steady-state exercise as opposed to breaking the exercise bout into smaller epochs and/or focusing on periods of changing hemodynamics (e.g., beginning and end of exercise). Another consideration for our study was that salt consumption was not normalized to total caloric intake. Normalizing salt consumption to individual caloric needs may better elucidate the role of high-salt consumption on physiological parameters (4). Furthermore, HRV is highly sensitive and short-term HRV exhibits relatively poor interday reliability (67). Therefore, future research should capture multiday HRV profiles, which would sufficiently represent an individual’s cardiac-autonomic profile by providing means and coefficients of variance across several days (68). These measures should be considered to ensure accurate insight into the effects of training status and nutrition interventions. Finally, as we recently discussed (35), participants did not adequately limit salt intake during either experimental arm of this study. This is indicated in Table 2 by ∼30% higher sodium excretion rate than expected based on the prescribed 2,300 mg sodium diet. One could argue that this means our data have higher ecological validity in that this represents the effects between “nearly habitual” salt intake in America versus higher salt consumption. Nonetheless, it could be that compared with the actual recommended salt intake (i.e., 2,300 mg/day), high-salt intake may indeed influence HRV. Additional future directions include investigating potential sex differences in a larger adequately powered sample, and how salt loading may influence HRV in other populations such as predominately salt-sensitive cohorts, older adults, and patient populations. Finally, future investigations of salt manipulation on directly assessed sympathetic outflow via microneurography and investigating nonlinear measures of HRV during cycling exercise and other physical stressors could provide complementary important information.

Perspectives and Significance

High-salt consumption does not influence time domain or frequency domain measures of HRV at baseline, during aerobic exercise, and during recovery from exercise in healthy young adults. High-salt consumption did increase the circulating concentration of MCP-1, which was significantly correlated to systolic BP and mean BP, irrespective of condition. Future studies are needed to determine whether high-salt consumption influences HRV in at-risk cohorts such as older adults or certain patient populations. In addition, studies that examine strategies to mitigate high-salt-induced elevations in inflammation (e.g., MCP-1) are warranted.

DATA AVAILABILITY

Data will be made available upon reasonable request.

SUPPLEMENTAL DATA

Supplemental Figure S1: https://doi.org/10.6084/m9.figshare.21191311.v2;

Supplemental Figure S2: https://doi.org/10.6084/m9.figshare.21191326.v3;

Supplemental Figure S3: https://doi.org/10.6084/m9.figshare.21191329.v3;

Supplemental Figure S4: https://doi.org/10.6084/m9.figshare.21191362.v2.

GRANTS

This work was supported by National Institutes of Health (NIH) Grants K01HL147998 (to A.T.R.), K01HL160772 (to J.C.W.) and UL1TR003096 (TL-1 Fellowship to B.A.L.), an American College of Sports Medicine Foundation Doctoral Student Research Grant 17-00521 (to M.C.B.), an American Heart Association Grant 18POST34060020 (to A.T.R.), and an Auburn University Presidential Graduate Research Fellowship (to B.A.L.).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

M.C.B., K.U.P., J.C.W., and A.T.R. conceived and designed research; B.A.L., M.C.B., K.U.P., J.C.W., and A.T.R. performed experiments; B.A.L., M.C.B., and A.T.R. analyzed data; B.A.L., M.C.B., K.U.P., J.C.W., and A.T.R. interpreted results of experiments; B.A.L., and A.T.R. prepared figures; B.A.L., J.C.W., and A.T.R. drafted manuscript; B.A.L., M.C.B., J.C.W., and A.T.R. edited and revised manuscript; B.A.L., M.C.B., K.U.P., J.C.W., and A.T.R. approved final version of manuscript.