The literature on salt intake and insulin sensitivity presents a mixed picture, as some studies have shown an increase, whereas others have shown a decrease, in insulin action as sodium intake is enhanced. In some cases, this may relate to the study of salt intake in patients with co-morbidities such as hypertension or diabetes. In the present study, we selected healthy normotensive lean volunteers who underwent a euglycaemic clamp following 6 days of a low-salt diet (20 mmol sodium daily) and, subsequently, 6 days of a high-salt diet (200 mmol sodium daily). Our results show an increase in insulin-mediated glucose disposal during euglycaemic clamp conditions that was significantly higher following the high-salt diet compared with the low-salt diet (7.41±0.41 compared with 6.11±0.40 mg·kg−1 of body weight·min−1 respectively; P=0.03). We measured calf blood flow before and during insulin infusion (no significant change after the two dietary salt interventions was detected) and plasma non-esterified fatty acids (also no significant differences were detected). We observed the expected increases in renin concentration and aldosterone activity in subjects on the low-salt diet, and also observed a significantly less increase in plasma noradrenaline concentration during euglycaemic insulin infusion following the high-salt compared with the low-salt diet. We propose that the 4–5-fold increase in serum aldosterone and the greater increase in plasma noradrenaline concentration following the low-salt intervention compared with the high-salt period may have contributed to the differences in insulin sensitivity following the adjustment in dietary sodium intake.

INTRODUCTION

Each meal ingestion initiates an integrated series of events, which facilitate digestion, delivery and storage of nutrients in target tissues. The haemodynamic and metabolic effects of insulin facilitate this process. The pivotal role of insulin in nutrient storage stems from its ability to both stimulate cardiac output, thereby increasing blood flow to [1] and transition through [2] the vasculature of insulin-sensitive target tissues such as skeletal muscle, and to direct nutrient uptake (principally glucose and fatty acids). When the responsiveness of insulin-sensitive target tissues such as skeletal muscle to insulin-mediated glucose uptake is reduced (i.e. insulin resistance), more insulin secretion is usually required.

The pathogenesis of insulin resistance is complex and only partially understood. We have shown previously [3] a relationship between plasma renin activity and insulin-stimulated glucose uptake measured during euglycaemic clamp conditions in normal volunteers. As plasma renin activity increased, insulin sensitivity fell [3], a finding also observed in hypertensive patients [4]. Alterations in salt intake directly influence plasma renin activity, with greater salt intake resulting in lower renin activity. Many of the studies over the last 2 decades, as summarized in Table 1, have pursued relationships between salt intake and insulin sensitivity, with some studies suggesting salt restriction increases insulin sensitivity [5], some finding a mixed result or no change in insulin sensitivity [68] and other studies suggesting salt restriction reduces insulin sensitivity in hypertension [9,10]. Moreover, public health initiatives that recommend reduced salt intake [11] often do not distinguish between those in whom clinical evidence implicates an increase in salt intake as injurious (such as those with hypertension or heart failure) and those without such co-morbidities where salt restriction could adversely affect parameters, such as insulin action, without clear cardiovascular benefit [12]. Thus information on the effects of salt restriction in healthy individuals are needed to help keep public health policy on this issue well-informed.

Table 1
Summary of studies on insulin sensitivity and sodium intake

HOMA, homoeostatic model assessment; ↓, decrease; ↑, increase; →, no change.

Author [reference]Sample size (n)Procedure used to assess insulin sensitivityInsulin sensitivity response to salt administrationComments
Donovan et al. [58* Euglycaemic clamp ↓ − 
Sharma et al. [623* OGTT → or ↑ Results differed by salt sensitivity of BP (→=salt-resistant) 
Foo et al. [718 Graded euglycaemic clamp → − 
Facchini et al. [819 3 h insulin-suppression test → − 
Iwaoka et al. [915 OGTT ↑ All hypertensive 
Raji et al. [10426 HOMA → or ↑ Depended on the nuances of renal blood flow responses to angiotensin II (modulators compared with non-modulators) 
Melander et al. [2928 Euglycaemic clamp ↓ or ↑ Results varied by gender and salt sensitivity; salt-sensitive subjects improved insulin sensitivity in higher salt intake; salt-resistant subjects worsened 
Perry et al. [3015* Euglycaemic clamp ↑ − 
Author [reference]Sample size (n)Procedure used to assess insulin sensitivityInsulin sensitivity response to salt administrationComments
Donovan et al. [58* Euglycaemic clamp ↓ − 
Sharma et al. [623* OGTT → or ↑ Results differed by salt sensitivity of BP (→=salt-resistant) 
Foo et al. [718 Graded euglycaemic clamp → − 
Facchini et al. [819 3 h insulin-suppression test → − 
Iwaoka et al. [915 OGTT ↑ All hypertensive 
Raji et al. [10426 HOMA → or ↑ Depended on the nuances of renal blood flow responses to angiotensin II (modulators compared with non-modulators) 
Melander et al. [2928 Euglycaemic clamp ↓ or ↑ Results varied by gender and salt sensitivity; salt-sensitive subjects improved insulin sensitivity in higher salt intake; salt-resistant subjects worsened 
Perry et al. [3015* Euglycaemic clamp ↑ − 
*

No women included.

In the present study, we measured insulin sensitivity following 1 week of isocaloric nutrient intake on a high-salt diet compared with a low-salt diet. We examined three mechanisms that could potentially be responsible for changes induced by varying salt intake on insulin sensitivity in healthy human subject volunteers, including measures of (i) skeletal muscle blood flow, (ii) plasma NEFA (non-esterified fatty acid) concentration, and (iii) plasma catecholamine metabolism before and during hyperinsulinaemic euglycaemia. In addition, the effects of dietary salt manipulation on seated BP (blood pressure) and heart rate (on day 6 of the diet) and fasting substrate oxidation, using indirect calorimetry during the clamp procedure, were determined. Given the growing importance of the metabolic syndrome, whose pivotal finding is insulin resistance, as a risk for future cardiovascular disease [13], we consider that the findings of studies such as this one would contribute to the body of knowledge from which dietary recommendations are drawn.

MATERIALS AND METHODS

All procedures were reviewed and approved by the Institutional Review Board of the University of Pennsylvania, and all subjects provided written informed consent. The protocol was also reviewed by the GCRC (General Clinical Research Center) Scientific Advisory Committee.

Selection of subjects

Normotensive men and women, within 20% of an ideal body weight as defined by data from the Metropolitan Life Insurance Company [14], of at least 18 years of age were recruited by local advertisement and word of mouth. There was no ethnic exclusion. At enrolment, each subject read and signed an informed consent form, and then underwent a standard medical history and physical examination, 12-lead ECG and routine measures of blood and urine chemistries, including urine toxicology for illegal substances. In addition, each subject also underwent a standard 3-h OGTT (oral glucose tolerance test) with 75 g of oral glucose [15] to exclude those with previously undiagnosed diabetes. Subjects with hypertension (>140/>90 mmHg) were excluded from the study. Subjects taking prescription medication for diabetes, thyroid disease, hypertension or sex-hormone replacement were excluded. Subjects having laboratory values greater than twice the upper limit of normal were also excluded. Subjects with illegal substances present in the urine would have been excluded; however, this did not occur in our screened subject population.

Study protocol

An outline of the main study design is shown in Figure 1. Three euglycaemic clamp procedures were performed on each subject in the GCRC. The first clamp was performed after a single overnight fast and was incorporated into the protocol to familiarize the subject with the complex procedures undertaken during the clamp, as done by others studying the effects of salt on insulin sensitivity [7]. Moreover, this normal condition clamp allowed the unmanipulated insulin sensitivity of the subject to be estimated. Women in this protocol underwent each euglycaemic clamp within 10 days of the onset of menses to minimize the effect cyclic hormonal changes have on insulin sensitivity. During the first admission to the GCRC, each subject was interviewed by the GCRC dietician to ascertain caloric needs and to record food preferences to craft a palatable diet of 20 mmol sodium/day. Subjects were re-admitted approx. 4 weeks later and randomized [using a pre-specified randomized blocked (block=ten subjects) table generated by the GCRC biostatistician and kept by the dietician] to receive the individualized 20 mmol sodium/day diet with either placebo tablets (ten per day; low-salt diet) or 1-g NaCl tablets (ten per day; high-salt diet). The tablets were divided across three meals for the six consecutive days. The diet was the same during each of the 1-week-long stays; the difference was limited to the NaCl component administered as a supplement. If the subject was randomized to the high-salt diet first, they underwent a euglycaemic clamp on the morning of day 7 in the GCRC (following six full days of the diet), and were re-admitted approx. 4 weeks later. At that time, they received the low-salt diet for 6 days, again undergoing the final euglycaemic clamp on day 7 of that GCRC stay. Subjects were weighed daily. Vital signs were recorded at 10.00, 14.00 and 16.00 hours as per the usual GCRC protocol for safety surveillance. One investigator measured seated BPs on day 6 at 14.00 hours (obtained in duplicate and averaged for a single systolic and diastolic value for each subject) and also obtained a single measurement of the heart rate of each subject. Urines samples (24 h) were collected on days 6–7 for sodium and urine urea nitrogen excretion.

Flow diagram detailing enrolment, study design, randomization and subject study completion

Figure 1
Flow diagram detailing enrolment, study design, randomization and subject study completion

Subj., subjects.

Figure 1
Flow diagram detailing enrolment, study design, randomization and subject study completion

Subj., subjects.

Euglycaemic clamp procedure

The euglycaemic clamp and indirect calorimetry studies were performed in the GCRC, as described previously [3,16]. For the first euglycaemic clamp, subjects were admitted to the GCRC between 17.00–18.00 hours on the evening prior to the study. A 12-h urine collection was obtained (18.00–06.00 hours) to estimate protein oxidation from urea nitrogen excretion. During the 1-week-long stay, 24-h urine samples were collected from 06.00 hours on the morning of day 6 to 06.00 hours on day 7 to determine sodium and urine urea nitrogen excretion.

Table 2 shows the timing of the infusions, samples and procedures during each euglycaemic clamp study. At 06.30 hours on the day of the clamp study, an intravenous catheter was inserted into each arm. One catheter was placed in the hand or wrist, and warmed to 55–60°C (to ‘arterialize’ the blood [17]) from which all blood sampling was performed. The other catheter was placed in the opposite forearm and was used for all infusions. Subjects rested for approx. 30 min before any studies were initiated.

Table 2
Time and events during the euglycaemic clamp study

The shading indicates an infusion during the time period. Each √ in a row indicates that a sample was drawn at that time. FFA, NEFAs; 15% DEX, 15% dextrose in water with 20 mmol/l KCl; ALDO, serum aldosterone concentration; NOREPI, plasma noradrenaline; Ind Cal, indirect calorimetry; QBf, calf blood flow measurement; *GLU, whole blood glucose (obtained at 10 min intervals from 07.00 to 09.00 hours, and sampled at 5 min intervals from 09.00 until 12.00 hours).

graphic
 
graphic
 

At 07.00 hours, a ‘blank’ blood sample for background deuterated glucose enrichment was obtained and was followed by a primed (4 mg/kg of body weight) and then continuous (50 μg·kg−1 of body weight·min−1) infusion of [2H6]glucose to determine the rate of appearance of glucose as an estimate of hepatic glucose output.

At 08.00 hours, samples for plasma renin and serum aldosterone concentrations were obtained to demonstrate any physiological effects of the dietary salt manipulation. Plasma for adiponectin and resistin concentrations were also collected at this time.

At 08.30 hours on the morning of clamp study, blood was sampled every 10 min to demonstrate stable whole-blood glucose concentrations, by using the glucose oxidase method on a glucose analyser (YSI 2300; Yellow Springs). At the same time, expired air was collected into an indirect calorimeter for 10 min via a canopy method to determine substrate oxidation. Expired air was collected by a canopy system connected to a DeltaTrac II calorimeter (Sensormedic) for determination of CO2 production and O2 usage 30 min before insulin administration and during the last hour of the euglycaemic clamp study. Expired air was analysed at 60 s intervals for a total of 9–10 min, and was averaged for each measurement.

At 08.45 hours, calf blood flow was determined (five measurements; averaged for a single value) using supine calf strain-gauge plethysmography (Hokansen) [18].

At 08.55 hours and again at 09.00 hours, blood was sampled to determine enrichment with [2H6]glucose and plasma NEFA concentrations.

At 09.00 hours, samples for plasma noradrenaline and insulin concentrations were drawn, which were repeated periodically as shown in Table 2. At 09.00 hours, a euglycaemic clamp was started [3,16]. A priming bolus of normal human insulin (approx. 0.5 unit adjusted based on body weight; Humulin-R; Eli Lilly) was followed by a constant infusion at a dose of 40 milli-units·m−2·min−1 for 3 h (from 09.00 to 12.00 hours). Whole-blood glucose concentrations were determined every 5 min, and a 15% dextrose in 20 mmol/l KCl was administered at a rate sufficient to maintain whole-blood glucose concentrations at the average level observed between 08.30 and 09.00 hours (typically between 75 and 85 mg/dl; 4.2 and 4.7 mmol/l).

At 11.00 hours, the indirect calorimetry data were repeated. At 11.15 hours, calf blood flow measurements were repeated. From approx. 11.20 hours until insulin infusion was completed, the subject was left as undisturbed as possible, so that the average glucose uptake during the final 30 min of the 3-h insulin infusion represented the target period.

Following 3 h of insulin infusion, all infusion solutions, except dextrose, were discontinued. Subjects ate a normal lunch, and the dextrose infusion tapered and discontinued between 12.00 and 13.00 hours.

Analyses and calculations

Urinary sodium was determined using a flame photometer (IL-943; Instrumentation Laboratory). Glucose enrichment was determined using GC/MS [19] by Metabolic Solutions, and the calculations for the rate of glucose appearance were performed using the Steele equation for steady state [20]. Active renin was determined using an immunoradiometric assay kit (Nichols Institute Diagnostic). Aldosterone was determined by an RIA kit (Diagnostic Products). The CVs (coefficients of variation) for both renin and aldosterone were <5 and 7% respectively. Plasma catecholamines were extracted using a Bio-Rad Laboratories kit and were determined by HPLC using an EC 1640 detector (Bio-Rad Laboratories). The CV for the catecholamines assay was <8%. All of the assays were performed in the GCRC Core Laboratory.

Plasma insulin levels were determined by RIA (INCSTAR), plasma NEFAs were determined using a standard kit (Wako Bioproducts), plasma adiponectin concentrations were determined by RIA, and plasma resistin concentrations were determined by ELISA. These assays were performed in the Diabetes and Endocrine Research Center (University of Pennsylvania, Philadelphia, PA, U.S.A.).

Indirect calorimetry data were analysed for carbohydrate and lipid substrate oxidation using formulae described previously [21].

Statistical analysis

Values are expressed as means±S.E.M., and data were analysed using STATA v9.2. Differences in variables measured during euglycaemic clamps following the low-salt and high-salt diets were compared with a paired Student's t test. Glucose uptake for each of the last three 10-min periods during insulin infusion was averaged for each subject during each euglycaemic clamp and was used as their glucose uptake value for that dietary salt phase. A Mann–Whitney test was performed if the data failed an equal variance test. Two-tailed P values were used for the Student's t tests, with P values <0.05 considered to be statistically significant.

We proposed a 15% difference in glucose uptake as potentially clinically significant, and determined that a sample size of 20 subjects using an S.D. of 1.8 mg·kg−1 of body weight·min−1 (based on our previous clamp studies in normal volunteers [22]) would provide 80% power to detect a difference between the two interventions.

RESULTS

The demographics of the study subjects are shown in Table 3. There were eight women and 12 men evaluated in the protocol. Two subjects (both women) completed only one of the two dietary phases. In both cases, this was during the low-salt phase. One of the two subjects required multiple attempts to re-start the blood sampling line during the second euglycaemic clamp and declined to return for the last intervention. The other subject moved out of the area shortly after completing the second euglycaemic clamp.

Table 3
Demographics of the study subjects

Values are means±S.E.M.

African AmericanCaucasian
CharacteristicAll subjectsWomenMenWomenMenAsian menLatino women
Subjects (n20 
Age (years) 30±2 31±5 33±4 26±4 32±3 28 23 
BMI (kg/m223.1±0.6 24.1±1.4 24.5±0.6 23.1±0.2 22.7±1.6 19.1 20.2 
Glucose (mg/dl) 75±3 73±5 78±5 70±6 81±2 63 67 
African AmericanCaucasian
CharacteristicAll subjectsWomenMenWomenMenAsian menLatino women
Subjects (n20 
Age (years) 30±2 31±5 33±4 26±4 32±3 28 23 
BMI (kg/m223.1±0.6 24.1±1.4 24.5±0.6 23.1±0.2 22.7±1.6 19.1 20.2 
Glucose (mg/dl) 75±3 73±5 78±5 70±6 81±2 63 67 

BP, heart rate and weight changes on day 6 after each dietary intervention

Seated BPs at 14.00 hours on day 6 in subjects on the low-salt diet were 111±2/65±2 mmHg, and seated BPs in the afternoon on day 6 in subjects on the high-salt diet were 117±4/69±3 mmHg (P=0.16 and P=0.33 for changes in systolic and diastolic BPs respectively, in subjects on the low-salt diet compared with the high-salt diet). Seated heart rates in the afternoon on day 6 in subjects on the low-salt and high-salt diets were 72±2 and 67±2 beats/min respectively (P=0.10).

The average weight on the morning of day 7 after 6 days on the low-salt diet was 67.9±2.5 kg compared with an average weight following 6 days on the high-salt diet of 69.6±2.4 kg. The average weight on day 7, although higher in subjects on the high-salt diet, was not significantly different (P=0.64) when compared with the low-salt dietary intervention.

Calorimetry and regional (calf) blood flow

Fasting substrate oxidation and calf blood flow on the morning of the euglycaemic clamp before and following 2 h of euglycaemic insulin infusion are shown in Table 4. There were no significant differences in calf blood flow between the low-salt and high-salt diets before (P=0.89) or during (P=0.86) insulin infusion. Calf blood flow increased significantly during insulin infusion compared with the value obtained before insulin infusion began (P<0.03 for both low-salt and high-salt diets before compared with after insulin infusion on that particular diet).

Table 4
Blood flow and indirect calorimetry before and during insulin infusion

Values are means±S.E.M. 1 kcal≈4.184 J.

Low-salt dietHigh-salt diet
ParameterPre-insulin infusionDuring insulin infusionPre-insulin infusionDuring insulin infusion
Calf blood flow (cm3·100 g−1 of leg weight·min−12.3±0.2 2.8±0.3 2.4±0.2 2.8±0.2 
REE (kcal·kg−1 of body weight·h−122.7±1.2 24.2±0.9 22.1±0.8 23.1±0.8 
Glucose oxidation (mg·kg−1 of body weight·min−10.8±0.3 2.4±0.4 1.3±0.4 3.2±0.5 
Fat oxidation (mg·kg−1 of body weight·min−11.4±0.2 0.8±0.2 1.1±0.2 0.4±0.2 
Low-salt dietHigh-salt diet
ParameterPre-insulin infusionDuring insulin infusionPre-insulin infusionDuring insulin infusion
Calf blood flow (cm3·100 g−1 of leg weight·min−12.3±0.2 2.8±0.3 2.4±0.2 2.8±0.2 
REE (kcal·kg−1 of body weight·h−122.7±1.2 24.2±0.9 22.1±0.8 23.1±0.8 
Glucose oxidation (mg·kg−1 of body weight·min−10.8±0.3 2.4±0.4 1.3±0.4 3.2±0.5 
Fat oxidation (mg·kg−1 of body weight·min−11.4±0.2 0.8±0.2 1.1±0.2 0.4±0.2 

Glucose oxidation was higher following the high-salt diet, but the results were not statistically significant (P=0.23). Similarly, lipid oxidation was lower following the high-salt diet compared with the low-salt diet, but not significantly so (P=0.27). REE (resting energy expenditure) was similar between the groups (P=0.66), and the changes in REE and substrate oxidation during the euglycaemic clamp were similar following the high-salt and low-salt diets. Difficulties with instrument calibration resulted in oxidation measurements only being obtained in 15 out of the 20 subjects.

Urine sodium, hormonal markers of sodium balance, rate of glucose appearance, cytokines and plasma NEFAs

Urine sodium excretion, as per the protocol, averaged 23±4 and 194±12 mmol/day in subjects on the low-salt and high-salt diets respectively. The low-salt diet was associated with significant increases in renin (33±5 compared with 16±2 μ-units/l in the high-salt diet; P<0.001) and aldosterone (15.9±2.6 compared with 3.5±0.7 ng/dl in the high-salt diet; P<0.03) concentrations. Simple pairwise correlations were performed using serum aldosterone concentrations in the low-salt and high-salt diets to compare the levels with the glucose uptake value for each subject. In the low-salt group, this pairwise correlation gave an r value of 0.36 (P=0.17), whereas, in the high-salt group, an r value of −0.02 (P=0.94) was obtained. Plasma adiponectin concentrations were 7.5±0.6 and 6.7±0.8 μg/ml in subjects on the low-salt and high-salt diets respectively (P=0.42), and plasma resistin concentrations were 8.3±0.7 and 6.8±0.7 ng/ml in subjects on the low-salt and high-salt diets respectively (P=0.14). The rate of glucose appearance was 1.9±0.01 and 2.1±0.9 mg·kg−1 of body weight·min−1 respectively, in subjects on the low-salt and high-salt diets (P=0.13). Plasma NEFA concentrations were 0.51±0.22 and 0.43±0.15 mmol/l in subjects on the low-salt and high-salt diets respectively (P=0.23).

Euglycaemic clamp

Blood glucose (Figure 2A) and insulin concentrations were similar during the clamps performed following the low-salt and high-salt diets. There were no differences in the insulin concentrations averaged over the last 30 min in the high-salt diet compared with the low-salt diet (P=0.82; Figure 2B). In the case of one subject (subject 3), the samples stored for the determination of the insulin concentrations (from all three clamp studies) were discarded accidentally before being assayed.

Whole-blood glucose and insulin levels, and glucose uptake during the euglycaemic clamp, in the low-salt (○) and high-salt (■) diets

Figure 2
Whole-blood glucose and insulin levels, and glucose uptake during the euglycaemic clamp, in the low-salt (○) and high-salt (■) diets

(A) Whole-blood glucose levels during the euglycaemic clamp before and after insulin infusion during the diet interventions. (B) Plasma insulin levels during euglycaemic clamp conditions on the two diet interventions. (C) Glucose uptake during euglycaemic clamp conditions following the two diet interventions. Glucose uptake on the first GCRC admission (random salt intake) is shown by a broken line. M, glucose uptake.

Figure 2
Whole-blood glucose and insulin levels, and glucose uptake during the euglycaemic clamp, in the low-salt (○) and high-salt (■) diets

(A) Whole-blood glucose levels during the euglycaemic clamp before and after insulin infusion during the diet interventions. (B) Plasma insulin levels during euglycaemic clamp conditions on the two diet interventions. (C) Glucose uptake during euglycaemic clamp conditions following the two diet interventions. Glucose uptake on the first GCRC admission (random salt intake) is shown by a broken line. M, glucose uptake.

Glucose uptake was significantly greater following the high-salt diet compared with the low-salt diet (average from 150–180 min was 7.41±0.41 compared with 6.11±0.40 mg·kg−1 of body weight·min−1 respectively; P=0.03; Figure 2C). When the subjects were divided according to their BP response to the dietary intervention on day 6, half of the subjects were salt-sensitive (‘salt-sensitive’ was defined as an increase in BP of at least 3 mmHg). There were no significant differences in the increase in the glucose uptake value in salt-resistant subjects compared with salt-sensitive subjects following the high-salt and low-salt diets (results not shown).

Plasma noradrenaline concentrations were similar at the start of the euglycaemic clamp on day 7 of the low-salt and high-salt diets (198±18 and 173±13 pg/ml respectively; P=0.29). After 180 min of euglycaemic hyperinsulinaemia, plasma noradrenaline levels were significantly higher in the low-salt diet (238±25 pg/ml) compared with the high-salt diet (181±19 pg/ml; P=0.02). Assay problems were encountered when analysing plasma catecholamines, such that there were only 17 paired samples available (0 and 180 min) for the low-salt group and 12 samples for the high-salt group. In addition, we performed simple pairwise correlations of noradrenaline and glucose uptake during the euglycaemic clamps as follows. First, we calculated the slopes of the plots of hourly plasma noradrenaline concentrations over time during the two dietary salt conditions. Subsequently, we calculated both the difference in slopes of the plasma noradrenaline concentration change over time after the two interventions and compared these slope differences with the differences in glucose uptake after the two dietary salt conditions. This pairwise correlation gave an r value of 0.46 (P=0.13).

DISCUSSION

The results of the present study have shown the anticipated stimulation of the renin–aldosterone axis by salt restriction, as well as a 21% increase in insulin-stimulated glucose uptake during euglycaemic clamp conditions after 6 days of a high-salt diet compared with a low-salt diet, which supports some findings of others [6,9,10] in suggesting that restriction of salt intake reduces insulin sensitivity.

Although we can only speculate on the mechanism(s) involved in this observation, we failed to show any significant changes in skeletal muscle blood flow (before or during insulin infusion) or significant changes in plasma NEFA concentrations in subjects on the low-salt diet compared with the high-salt diet. In addition, there were no significant alterations in plasma adiponectin or resistin concentrations to assist in determining the mechanism behind the changes in insulin sensitivity. Our results are consistent, however, with a reduction in sympathetic neural activity in conjunction with a markedly lower aldosterone concentration in the high-salt intervention, which may have contributed to our observations of insulin sensitivity in subjects on the two salt diets. These data are consistent with our previous observations of an important role of the involuntary nervous system in insulin sensitivity [23,24] and the known effects of aldosterone on insulin action [25,26].

Previous studies have shown that increases in salt intake suppress sympathetic nervous system activity [27]. In the present study, we observed that the plasma noradrenaline concentration during insulin infusion increased more during the low-salt diet compared with the high-salt diet. Euglycaemic insulin infusions increase plasma noradrenaline concentrations in a time-dependent fashion [28]. Moreover, although not achieving statistical significance (P=0.10), the average heart rate in subjects on day 6 of the high-salt diet (66±2 beats/min) was lower than the average heart rate in subjects on day 6 of the low-salt diet (72±2 beats/min), consistent with reduced adrenergic activity during the high-salt intervention.

Relatively recent work has shown that increased aldosterone concentrations down-regulate insulin action in target cells [26]. Interestingly, this in vitro study found a 23% reduction in glucose uptake when aldosterone was increased in the medium, similar to our 21% difference in glucose uptake during the high-salt diet compared with the low-salt diet [26].

Our results do differ though from those of Donovan et al. [5], who found that a high-salt diet decreased insulin-mediated glucose uptake. Their study enrolled eight men, but was otherwise quite similar in design to the present study. We studied 20 subjects, both men and women, but, aside from the gender and numbers studied, we have no additional explanation for the discordance in the findings between the two investigations.

Our results are also different from the findings of Foo et al. [7]. Our biochemical (catecholamines, renin and aldosterone) changes were similar to those reported by Foo et al. [7]; however, in their study, they observed (unlike in the present study) an increase in calf blood flow during the high-salt intervention, which did not translate into greater glucose uptake. Foo et al. [7] assessed insulin sensitivity using two doses of insulin for 120 min at each dose, whereas we studied a single insulin dose for 180 min. When we re-analysed our data at 120 min (see Figure 2C), we found that the differences in salt interventions were smaller at that point with a loss of statistical significance (P=0.10). Moreover, the differences in the serum aldosterone concentrations on the low-salt diet (mean, 15.9 ng/dl) compared with the high-salt diet (mean, 3.5 ng/dl) in the present study were approx. 4–5-fold higher in the low-salt diet compared with approx. 2-fold in the study by Foo et al. [7]. Differences in the duration of insulin infusion and the biochemical responses to salt deprivation may explain some of the differences between these two studies.

The results of the present study are, to some extent, in agreement with those of Melander et al. [29]. In their study, a strong correlation between stimulation of the renin system (increased plasma renin activity and serum aldosterone concentrations) and decreased insulin sensitivity was observed in the low-salt diet compared with the high-salt diet. No plasma catecholamine measurements (only urine) were reported, so the present study differs in this respect. In their summary table, the difference in glucose disposal, although higher in the high-salt intake, was not statistically significant (P=0.10). It was only in the correlation of biochemical parameter changes with insulin sensitivity that significant differences between the low-salt and high-salt diets were detected.

The results of the present study are also similar in magnitude and direction to those reported by Perry et al. [30]. Their studies in normal volunteers demonstrated a 15% reduction in insulin sensitivity on a low-salt diet (<80 mmol/day) compared with an NaCl-supplemented diet. In their study, a dose of insulin was used that achieved plasma concentrations approximately twice those of the present study. They did not propose a mechanism to explain their results, since they were pursuing a role of angiotensin II on adipocyte metabolism (lipolysis and glucose transport), but found no support for their hypothesis that increases in angiotensin II were causative in subjects on the low-salt diet.

Several limitations of our present study are important to highlight. We studied only relatively young healthy volunteers with a normal BMI (body mass index) range, so our results are not generalizable to those with co-morbidities, such as diabetes or high BP, those who are older than 45 years of age or obese subjects (BMI >27.5 kg/m2).

In summary, we have observed a decrease in insulin-mediated glucose uptake during euglycaemic clamp conditions in healthy volunteers given an isocaloric low-salt diet compared with a high-salt diet for 6 days. We did not observe any significant changes in calf muscle blood flow or plasma NEFA concentrations, which might have explained this result. However, we did observe lower sympathetic neural responses, as reflected by reduced changes in plasma noradrenaline concentrations during euglycaemic insulin infusion after the high-salt diet compared with the low-salt diet. We also observed the expected changes in serum aldosterone concentrations, and propose that reductions in sympathetic neural activation when insulin is infused after 1 week of salt loading in conjunction with a reduction in serum aldosterone concentration are plausible mechanisms to explain at least part of this increase in insulin responsiveness.

Abbreviations

     
  • BMI

    body mass index

  •  
  • BP

    blood pressure

  •  
  • CV

    coefficient of variation

  •  
  • GCRC

    General Clinical Research Center

  •  
  • NEFA

    non-esterified fatty acid

  •  
  • OGTT

    oral glucose tolerance test

  •  
  • REE

    resting energy expenditure

This work was supported by NIH (National Institutes of Health) grants M01-RR00040 and K24-DK-02684.

References

References
1
Clark
 
M. G.
Wallis
 
M. G.
Barrett
 
E. J.
, et al 
Blood flow and muscle metabolism: a focus on insulin action
Am. J. Physiol. Endocrinol. Metab.
2003
, vol. 
284
 (pg. 
E241
-
E258
)
2
Vincent
 
M. A.
Clerk
 
L. H.
Rattigan
 
S.
Clark
 
M. G.
Barrett
 
E. J.
 
Active role for the vasculature in the delivery of insulin to skeletal muscle
Clin. Exp. Pharmacol. Physiol.
2005
, vol. 
32
 (pg. 
302
-
307
)
3
Townsend
 
R. R.
Zhao
 
H.
 
Plasma renin activity and insulin sensitivity in normotensive subjects
Am. J. Hypertens.
1994
, vol. 
7
 (pg. 
894
-
898
)
4
Lind
 
L.
Reneland
 
R.
Andersson
 
P. E.
Haenni
 
A.
Lithell
 
H.
 
Insulin resistance in essential hypertension is related to plasma renin activity
J. Hum. Hypertens.
1988
, vol. 
12
 (pg. 
379
-
382
)
5
Donovan
 
D. S.
Solomon
 
C. G.
Seely
 
E. W.
Williams
 
G. H.
Simonson
 
D. C.
 
Effect of sodium intake on insulin sensitivity
Am. J. Physiol.
1993
, vol. 
264
 (pg. 
E730
-
E734
)
6
Sharma
 
A. M.
Ruland
 
K.
Spies
 
K.-P.
Distler
 
A.
 
Salt sensitivity in young normotensive subjects is associated with a hyperinsulinemic response to oral glucose
J. Hypertens.
1991
, vol. 
9
 (pg. 
329
-
335
)
7
Foo
 
M.
Denver
 
A. E.
Coppack
 
S. W.
Yudkin
 
J. S.
 
Effect of salt-loading on blood pressure, insulin sensitivity and limb blood flow in normal subjects
Clin. Sci.
1998
, vol. 
95
 (pg. 
157
-
164
)
8
Facchini
 
F. S.
DoNascimento
 
C.
Reaven
 
G. M.
Yip
 
J. W.
Ni
 
X. P.
Humphreys
 
M. H.
 
Blood pressure, sodium intake, insulin resistance, and urinary nitrate excretion
Hypertension
1999
, vol. 
33
 (pg. 
1008
-
1012
)
9
Iwaoka
 
T.
Umeda
 
T.
Ohno
 
M.
, et al 
The effect of low and high NaCl diets on oral glucose tolerance
Klin. Wochenschr.
1988
, vol. 
66
 (pg. 
724
-
728
)
10
Raji
 
A.
Williams
 
G. H.
Jeunemaitre
 
X.
, et al 
Insulin resistance in hypertensives: effect of salt sensitivity, renin status and sodium intake
J. Hypertens.
2001
, vol. 
19
 (pg. 
99
-
105
)
11
Food and Drug Administration
 
 
Food labeling: declaration of sodium content of foods and label claims for food on the basis of sodium content, final rule
Fed. Register
1984
, vol. 
49
 (pg. 
15.510
-
515.535
)
12
Controy
 
M. M.
 
AMA targets sodium reduction as part of a plan to reduce cardiovascular disease
2006
 
13
Reilly
 
M. P.
Rader
 
D. J.
 
The metabolic syndrome: more than the sum of its parts?
Circulation
2003
, vol. 
108
 (pg. 
1546
-
1551
)
14
Metropolitan Life Foundation
 
Metropolitan Height and Weight Tables
1983
Stat. Bull. Metropol. Life Insur. Co.
 
64:21983
15
National Diabetes Data Group
 
Classification and diagnosis of diabetes mellitus and other categories of glucose intolerance
Diabetes
1979
, vol. 
28
 (pg. 
1039
-
1057
)
16
Townsend
 
R. R.
DiPette
 
D. J.
 
Pressor doses of angiotensin-II increase insulin-mediated glucose uptake in normotensive man
Am. J. Physiol.
1993
, vol. 
265
 (pg. 
E362
-
E366
)
17
McGuire
 
E. A. H.
Helderman
 
J. H.
Tobin
 
J. D.
Andres
 
R.
Berman
 
M.
 
Effects of arterial versus venous sampling on analysis of glucose kinetics in man
J. Appl. Physiol.
1976
, vol. 
41
 (pg. 
565
-
573
)
18
Hokanson
 
D. E.
Sumner
 
D. S.
Strandness
 
D. E.
 
An electrically calibrated plethysmography for direct measurement of limb blood flow
IEEE Trans. Biomed. Eng.
1975
, vol. 
22
 (pg. 
25
-
29
)
19
Guo
 
Z. K.
Lee
 
W. N.
Katz
 
J.
Bergner
 
A. E.
 
Quantitation of positional isomers of deuterium-labeled glucose by gas chromatography/mass spectrometry
Anal. Biochem.
1992
, vol. 
204
 (pg. 
273
-
282
)
20
Steele
 
R.
Wall
 
J.
DeBodo
 
R.
Altszuler
 
N.
 
Measurement of size and turnover rate of body glucose pool by the isotope dilution method
Am. J. Physiol.
1956
, vol. 
187
 (pg. 
15
-
24
)
21
Jequier
 
E.
Acheson
 
K.
Schutz
 
Y.
 
Assessment of energy expenditure and fuel utilization in man
Ann. Rev. Nutr.
1987
, vol. 
7
 (pg. 
187
-
208
)
22
Townsend
 
R. R.
Zhao
 
H.
 
Differential effect of insulin on saturated and unsaturated fatty acids
Metab. Clin. Exp.
2002
, vol. 
51
 (pg. 
779
-
782
)
23
Teff
 
K. L.
Townsend
 
R. R.
 
Prolonged mild hyperglycemia induces vagally mediated compensatory increase in C-peptide secretion in humans
J. Clin. Endocrinol. Metab.
2004
, vol. 
89
 (pg. 
5606
-
5613
)
24
Petrova
 
M.
Townsend
 
R.
Teff
 
K. L.
 
Prolonged (48-hour) modest hyperinsulinemia decreases nocturnal heart rate variability and attenuates the nocturnal decrease in blood pressure in lean, normotensive humans
J. Clin. Endocrinol. Metab.
2006
, vol. 
91
 (pg. 
851
-
859
)
25
Corry
 
D. B.
Tuck
 
M. L.
 
The effect of aldosterone on glucose metabolism
Curr. Hypertens. Rep.
2003
, vol. 
5
 (pg. 
106
-
109
)
26
Campion
 
J.
Maestro
 
B.
Molero
 
S.
, et al 
Aldosterone impairs insulin responsiveness in U-937 human promonocytic cells via the down regulation of its own receptor
Cell Biochem. Funct.
2002
, vol. 
20
 (pg. 
237
-
245
)
27
Watson
 
R. D.
Esler
 
M. D.
Leonard
 
P.
Komer
 
P. I.
 
Influence of variation in dietary sodium intake on biochemical indices of sympathetic activity in normal man
Clin. Exp. Pharmacol. Physiol.
1984
, vol. 
11
 (pg. 
163
-
170
)
28
Rowe
 
J. W.
Young
 
J. B.
Minaker
 
K. L.
Stevens
 
A. L.
Pallotta
 
J.
Landsberg
 
L.
 
Effect of insulin and glucose infusion on sympathetic nervous system activity in normal man
Diabetes
1981
, vol. 
30
 (pg. 
219
-
225
)
29
Melander
 
O.
Groop
 
L.
Hulthen
 
U. L.
 
Effect of salt on insulin sensitivity differs according to gender and degree of salt sensitivity
Hypertension
2000
, vol. 
35
 (pg. 
827
-
831
)
30
Perry
 
C. G.
Palmer
 
T.
Cleland
 
S. J.
, et al 
Decreased insulin sensitivity during dietary sodium restriction is not mediated by effects of angiotensin II on insulin action
Clin. Sci.
2003
, vol. 
105
 (pg. 
187
-
194
)