Previous studies have shown that acute increases in plasma NEFAs (non-esterified fatty acids) raise SVR (systemic vascular resistance) and BP (blood pressure). However, these studies have failed to distinguish between CNS (central nervous system) mechanisms that raise sympathetic activity and paracrine mechanisms that increase SVR directly, independent of CNS involvement. The aim of the present study was to directly determine whether the sympathetic nervous system contributes to the pressor response to NEFAs. On 2 days separated by at least 2 weeks, 17 lean healthy volunteers (ten male/seven female; age, 22±1 years; body mass index, 23±1 kg/m2; values are means±S.E.M.) received a 4-h intravenous infusion of 20% Intralipid® or placebo (in a single-blind randomized balanced order). MSNA (muscle sympathetic nerve activity), HR (heart rate), BP (oscillometric brachial measurement) and (cardiac output; acetylene rebreathing) were measured before and throughout infusion. The change in HR (+8.2±1.0 and +2.4±1.2 beats/min), systolic BP (+14.0±1.6 and +3.2±2.5 mmHg) and diastolic BP (+8.2±1.0 and −0.1±1.7 mmHg) were significantly greater after the 4-h infusion of Intralipid® compared with placebo (P<0.001). The change in BP with Intralipid® resulted from an increase in SVR (/mean arterial pressure; P<0.001) compared with baseline, without a change in . MSNA burst frequency increased during Intralipid® infusion compared with baseline (+4.9±1.3 bursts/min; P<0.05), and total MSNA (frequency×amplitude) was augmented 65% (P<0.001), with no change during placebo infusion. Lipid infusion increased insulin, aldosterone and F2-isoprostane, but not leptin, concentrations. On the basis of the concomitant increase in BP, MSNA and SVR, we conclude that central sympathetic activation contributes to the pressor response to NEFAs.

INTRODUCTION

Accumulating evidence links elevated plasma NEFA (non-esterified fatty acid) levels with insulin resistance and hypertension, two components of the metabolic syndrome. Epidemiological research indicates that elevated fasting NEFA concentrations are a strong independent predictor for the development of hypertension [1]. Levels of NEFAs in obese hypertensive patients are approximately twice that measured in lean normotensive individuals [2], and the ability of insulin to lower NEFAs is severely impaired in these patients. Moreover, NEFA concentrations correlate directly with BP (blood pressure) measured at rest and over 24 h, independently of insulin and insulin-mediated glucose disposal [3].

Studies over the past decade have examined whether acutely elevating the NEFA concentration increases BP, and the underlying mechanisms associated with the pressor response to NEFAs. For example, I/H (Intralipid®/heparin) infusion in minipigs significantly augmented BP and vascular resistance in most tissue beds [4]. Likewise, portal and femoral venous infusions of oleic acid induced a pressor response in rats [5,6], and I/H infusion in humans increased BP and resistance in most [79], but not all [10], investigations.

Despite the emerging literature linking NEFAs to metabolic abnormalities and hypertension, there is a paucity of results regarding the role of the SNS (sympathetic nervous system) in the pressor response to NEFAs. This is surprising given the evidence that some forms of hypertension with obesity and high-fat diets may be neurogenically mediated [11,12]. Studies of various lengths in normotensive rats and humans have shown that acutely elevating the NEFA concentration increases [5,8], decreases [13] or does not change [14] sympathetic activity. However, probably due to indirect measures of sympathetic activity [i.e. NA (noradrenaline) concentrations] [5,8,13] or insufficient duration of previous studies [14], it remains unclear whether NEFAs activate the SNS.

Therefore the main aim of the present study was to investigate the role of NEFAs on central sympathetic activation in humans by direct measurement of MSNA (muscle sympathetic nerve activity) for a sufficient duration to establish a complete temporal response. A secondary aim was to determine whether leptin and/or insulin are plausible mediators of the sympathetic response to NEFAs. Finally, we sought to fill the gaps of previous studies by measuring various hormonal [e.g. aldosterone and AngII (angiotensin II)] and haemodynamic variables, together with MSNA, in the framework of a randomized blinded placebo-controlled design. We hypothesized that NEFAs would augment MSNA and BP, and that this response would be associated with increases in insulin and leptin.

MATERIALS AND METHODS

Subjects

A total of 17 young (18–31 years of age) subjects participated in the present study (Table 1). All women were tested during days 2–7 of the follicular phase of their menstrual cycle [15]. Each subject underwent a complete medical screening, including DXA (dual energy X-ray absorptiometry) and a graded exercise test for determination of V̇O2max (maximal oxygen uptake). All subjects were healthy, active, but not athletically trained, normotensive, non-diabetic, non-obese and non-smokers who were not taking any medication, including oral contraceptives. Subjects abstained from alcohol for 2 days and caffeine for 1 day prior to the experiment. Approval was obtained from the Institutional Review Board of the Pennsylvania State University. Each subject gave written informed consent, and all procedures conformed to the Declaration of Helsinki.

Table 1
Characteristics of the study subjects

Values are means±S.E.M. BMI, body mass index; HDL, high-density lipoprotein; LDL, low-density lipoprotein; HOMA-IR, homoeostasis model assessment of insulin resistance; QUICKI, quantitative insulin sensitivity check.

CharacteristicValue
Gender (n) (male, female) 10, 7 
Age (years) 22±1 
Height (cm) 173±0 
Weight (kg) 69±3 
BMI (kg/m223±1 
Body fat (%) 20±2 
Waist/hip ratio 0.88±0.01 
V̇O2max (ml·kg−1 of body weight·min−145±2 
Total cholesterol (mg/dl) 155±6 
HDL (mg/dl) 53±3 
LDL (mg/dl) 86±5 
HOMA-IR 0.73±0.11 
QUICKI 0.43±0.01 
CharacteristicValue
Gender (n) (male, female) 10, 7 
Age (years) 22±1 
Height (cm) 173±0 
Weight (kg) 69±3 
BMI (kg/m223±1 
Body fat (%) 20±2 
Waist/hip ratio 0.88±0.01 
V̇O2max (ml·kg−1 of body weight·min−145±2 
Total cholesterol (mg/dl) 155±6 
HDL (mg/dl) 53±3 
LDL (mg/dl) 86±5 
HOMA-IR 0.73±0.11 
QUICKI 0.43±0.01 

Instrumentation

The study utilized a cross-over design involving two identical experimental days (differing only in the substance infused), separated by a period of at least 2 weeks. The infusion of I/H or S/G (saline/glycerol) was randomized, balanced and single-blinded. All subjects consumed a standardized meal of 54.4 kJ/kg of body weight, consisting of 55% carbohydrate, 30% fat and 15% protein, between 18.00 and 20.00 hours on the night prior to the experiment. After an overnight fast, subjects reported to the laboratory (approx. 07.00 hours) and lay in a supine position on a bed. A 20-gauge catheter was inserted into one antecubital vein for infusion, and an 18-gauge catheter was placed retrograde into the contralateral dorsal hand vein for blood sampling. This arm was kept on a heating pad to arterialize the venous blood. Subjects were then instrumented to measure HR (heart rate), BP, CBF (calf blood flow), (cardiac output) and MSNA.

Measurements

HR and arterial pressure

HR (three-lead ECG) was measured and recorded continuously. Oscillometric brachial BP (Colin 7000) was measured after a 30-min period of rest (twice at 1-min intervals) and every 10 min throughout infusion. MAP (mean arterial pressure) was estimated as diastolic pressure+1/3 pulse pressure.

CBF

CBF was determined by venous occlusion plethysmography on the calf [16]. During each data-recording period, blood flow was attained with two sets of measurements (separated by 4 min each) consisting of four cycles in succession. CVR (calf vascular resistance) was estimated as corresponding brachial MAP/CBF.

Q

graphic
was measured using the acetylene rebreathing technique that was first described by Triebwasser et al. [17]. A mass spectrometer (Perkin-Elmer MGA1100) was used to analyse gas concentrations during a 20-s rebreathing period. The accuracy and precision of this method has been validated against thermodilution and direct Fick. SVR (systemic vascular resistance) was calculated as
graphic
/MAP. SV (stroke volume) was calculated as
graphic
/HR.

MSNA

Peroneal MSNA was recorded as described previously [18]. Briefly, a tungsten electrode (FHC) was inserted into the peroneal nerve of the leg. Raw recordings were amplified, filtered and full-wave rectified to produce a recording of ‘integrated’ MSNA. Nerve traffic was expressed both as bursts/min and as bursts/min×mean burst amplitude, an index of total MSNA.

Protocol

Following instrumentation and a 30-min adaptation period, baseline haemodynamic and sympathetic data were obtained, and blood was sampled for metabolic and hormonal assays. Subjects then received a 4-h infusion of either I/H [20% Intralipid® (0.8 ml·m−2·min−1; Baxter) [7] and heparin (200 unit bolus, followed by 0.3 unit·kg−1 of body weight·min−1)] or S/G [sham infusion, 2.25% glycerol/saline (New England Compounding), infused at the same rate and volume as I/H]. Heparin was infused with Intralipid® to activate lipoprotein lipase and to accelerate the hydrolysis of fatty acids from TAG (triacylglycerol) [19]. Glycerol was infused with saline to match the glycerol contained in 20% Intralipid®. During the infusion,

graphic
and CBF were measured in duplicate every 30 min; MSNA and HR were recorded and averaged over a 5-min period every 30 min; and BP was taken every 10 min. Arterialized blood was sampled at baseline and for each hour of infusion thereafter for NEFAs, TAG, glycerol, glucose, insulin, leptin and catecholamines. Additional blood was sampled at baseline and at the end of infusion for aldosterone, AngII, and F2-isoprostanes. F2-isoprostanes are a biomarker of oxidative stress that may affect endothelial function.

Blood samples

Glucose was determined by the oxidase method immediately after blood collection. All other samples were collected on ice, centrifuged at 4°C, and the plasma was stored at −80°C until assay. Blood for NEFAs, TAG and glycerol was drawn into pre-chilled tubes containing EDTA and 0.275 g of paraoxon/l (Sigma) to inhibit lipoprotein lipase and to prevent hydrolysis of NEFAs from TAG in vitro [20]. Samples for catecholamines were collected in pre-chilled tubes containing EGTA and glutathione. Serum insulin and leptin concentrations were determined by ELISA. Plasma NEFA and TAG concentrations were determined by a colorimetric method. Glycerol was measured enzymatically, catecholamines by HPLC, and aldosterone and AngII by RIA. F2-isoprostanes were determined by immunoassay.

Statistical analysis

Recordings of MSNA and HR (5 min of data) were collected at baseline and every 30 min during infusion; each 5-min period was averaged to a single value. CBF was measured eight times, and was measured in duplicate at baseline and every 30 min during infusion; likewise, data from each period were averaged to a single value.

A repeated-measures ANOVA was used to detect treatment effects on neural, hormonal and haemodynamic variables. Least-squares means with Bonferroni correction were performed when appropriate to identify where differences between treatments occurred. Linear least-squares fit regression was conducted to examine the relationship between MSNA and CVR. The level of significance was set at P=0.05. Values are means±S.E.M. All calculations were made using SAS 9.1 software.

RESULTS

Anthropometric, metabolic and hormonal measurements

All subjects were young, healthy, normotensive, non-obese and were not insulin-resistant (Table 1). No gender differences in any responses were detected, so data from men and women were pooled for analysis. Infusions of I/H were well-tolerated by all of the subjects.

Plasma NEFA, TAG, glycerol and glucose concentrations during the infusions of S/G and I/H are shown in Table 2. Infusion of I/H was associated with a significant rise in plasma NEFAs and TAG compared with baseline and S/G infusion, whereas no changes occurred with S/G infusion. Plasma glycerol increased with S/G infusion and to a greater extent during I/H infusion. No significant changes in plasma glucose concentrations occurred with either day.

Table 2
Time course of plasma NEFA, TAG, glucose and glycerol concentrations during infusion

Values are means±S.E.M. of 17 subjects. *P<0.001 compared with baseline; †P<0.001 compared with S/G control day.

Time (h)
VariableInfusion01234
NEFAs (μmol/l) I/H 502±52 1419±51*† 2104±142*† 2301±93*† 2266±110*† 
 S/G 502±52 518±50 556±50 617±36 657±38 
TAG (mmol/l) I/H 0.87±0.07 3.03±0.13*† 4.17±0.30*† 4.99±0.43*† 5.76±0.54*† 
 S/G 0.79±0.07 0.97±0.07 0.94±0.06 0.97±0.07 1.00±0.07 
Glycerol (μmol/) I/H 72±9 774±44*† 1072±61*† 1239±81*† 1288±101*† 
 S/G 79±7 239±10* 244±12* 243±16* 262±16* 
Glucose (mmol/l) I/H 4.45±0.10 4.69±0.07 4.50±0.09 4.35±0.09 4.46±0.08 
 S/G 4.52±0.10 4.52±0.12 4.38±0.13 4.57±0.13 4.55±0.12 
Time (h)
VariableInfusion01234
NEFAs (μmol/l) I/H 502±52 1419±51*† 2104±142*† 2301±93*† 2266±110*† 
 S/G 502±52 518±50 556±50 617±36 657±38 
TAG (mmol/l) I/H 0.87±0.07 3.03±0.13*† 4.17±0.30*† 4.99±0.43*† 5.76±0.54*† 
 S/G 0.79±0.07 0.97±0.07 0.94±0.06 0.97±0.07 1.00±0.07 
Glycerol (μmol/) I/H 72±9 774±44*† 1072±61*† 1239±81*† 1288±101*† 
 S/G 79±7 239±10* 244±12* 243±16* 262±16* 
Glucose (mmol/l) I/H 4.45±0.10 4.69±0.07 4.50±0.09 4.35±0.09 4.46±0.08 
 S/G 4.52±0.10 4.52±0.12 4.38±0.13 4.57±0.13 4.55±0.12 

Serum leptin and insulin, and plasma adrenaline concentrations are shown in Figure 1, and plasma aldosterone, AngII and F2-isoprostane concentrations are shown in Table 3. Baseline hormone concentrations were similar in both experimental conditions. The leptin concentration decreased during both infusions; however, the treatment×time interaction (P=0.053) suggested a tendency for NEFAs to maintain leptin levels compared with the decline in leptin with placebo. The insulin concentration rose significantly by 3 and 4 h of I/H infusion, but did not increase in response to S/G infusion. Infusion of I/H significantly lowered plasma NA concentrations by 3 and 4 h, whereas NA increased by the final hour of S/G infusion. When expressed as a change from baseline (Figure 2), the NA concentration was unchanged with I/H infusion; however, a significant treatment interaction (P=0.001) was still present. Both F2-isoprostanes and aldosterone increased during I/H infusion, whereas aldosterone significantly decreased during S/G administration. No changes in AngII and adrenaline concentrations were observed with either I/H or S/G infusion (Table 3).

Hormonal responses to the infusion of I/H (fat) and S/G (saline)

Figure 1
Hormonal responses to the infusion of I/H (fat) and S/G (saline)

Values are means±S.E.M. of 17 subjects. *P<0.05 compared with baseline; †P<0.05 compared with saline control day.

Figure 1
Hormonal responses to the infusion of I/H (fat) and S/G (saline)

Values are means±S.E.M. of 17 subjects. *P<0.05 compared with baseline; †P<0.05 compared with saline control day.

Table 3
Baseline (0 h) and final (4 h) values for neural, cardiovascular and hormonal variables during infusion

Values are means±S.E.M. for subjects infused with I/H or S/G solutions at baseline and at 4 h. *P<0.05 compared with baseline; †P<0.05 compared with S/G control day. See Figures for the time course of changes. Interaction, P value of a two-way (treatment×time) repeated-measures ANOVA interaction.

Infusion
I/HS/G
Variable0 h4 h0 h4 hInteraction
HR (beats/min) 58.7±2.4 69.0±3.1*† 60.9±2.3 63.3±2.5 <0.001 
SBP (mmHg) 109.2±2.1 123.2±2.7*† 112.4±2.5 116.0±3.3 <0.001 
DBP (mmHg) 59.7±1.9 68.5±1.8*† 62.1±1.4 61.7±2.4 <0.001 
MAP (mmHg) 76.2±1.7 86.7±2.0*† 78.8±1.6 79.8±2.5 <0.001 
graphic
(litres/min) 
6.5±0.4 6.4±0.4 6.5±0.4 6.4±0.4 0.983 
SV (ml) 86.7±6.1 79.1±6.9 85.3±5.6 79.8±6.4 0.904 
SVR (dyne·s·cm−5991±80 1151±80* 1007±56 1063±64 0.070 
MSNA (bursts/min) 16.3±1.7 21.2±1.8* 20.5±2.2 20.8±2.3 0.005 
Total MSNA (% baseline) 100±0 165±24*† 100±0 112±16 0.002 
CBF (ml·100 ml−1·min−11.65±0.23 2.60±0.29*† 1.95±0.14 2.36±0.17* 0.032 
CVR (units) 54.4±4.6 38.7±3.2 45.1±4.5 37.2±3.5 0.080 
NA (nmol/l) 0.93±0.06 0.74±0.06*† 0.76±0.07 0.99±0.08* 0.001 
Leptin (pg/ml) 5.43±1.24 5.11±1.16* 5.37±1.19 4.55±1.05* 0.053 
Insulin (pmol/l) 23.5±4.2 30.0±4.3*† 23.3±2.3 18.8±2.0 0.004 
Adrenaline (pmol/l) 479.9±34.0 491.6±58.3 461.0±35.3 456.2±29.3 0.937 
Aldosterone (nmol/l) 0.08±0.01 0.11±0.01*† 0.09±0.01 0.05±0.01* <0.001 
AngII (pmol/l) 3.84±0.58 3.02±0.51 4.0±0.62 3.5±0.54 0.624 
F2-isoprostanes (pg/ml) 12.4±1.2 24.2±1.1*† 12.9±1.1 14.8±0.7 0.015 
Infusion
I/HS/G
Variable0 h4 h0 h4 hInteraction
HR (beats/min) 58.7±2.4 69.0±3.1*† 60.9±2.3 63.3±2.5 <0.001 
SBP (mmHg) 109.2±2.1 123.2±2.7*† 112.4±2.5 116.0±3.3 <0.001 
DBP (mmHg) 59.7±1.9 68.5±1.8*† 62.1±1.4 61.7±2.4 <0.001 
MAP (mmHg) 76.2±1.7 86.7±2.0*† 78.8±1.6 79.8±2.5 <0.001 
graphic
(litres/min) 
6.5±0.4 6.4±0.4 6.5±0.4 6.4±0.4 0.983 
SV (ml) 86.7±6.1 79.1±6.9 85.3±5.6 79.8±6.4 0.904 
SVR (dyne·s·cm−5991±80 1151±80* 1007±56 1063±64 0.070 
MSNA (bursts/min) 16.3±1.7 21.2±1.8* 20.5±2.2 20.8±2.3 0.005 
Total MSNA (% baseline) 100±0 165±24*† 100±0 112±16 0.002 
CBF (ml·100 ml−1·min−11.65±0.23 2.60±0.29*† 1.95±0.14 2.36±0.17* 0.032 
CVR (units) 54.4±4.6 38.7±3.2 45.1±4.5 37.2±3.5 0.080 
NA (nmol/l) 0.93±0.06 0.74±0.06*† 0.76±0.07 0.99±0.08* 0.001 
Leptin (pg/ml) 5.43±1.24 5.11±1.16* 5.37±1.19 4.55±1.05* 0.053 
Insulin (pmol/l) 23.5±4.2 30.0±4.3*† 23.3±2.3 18.8±2.0 0.004 
Adrenaline (pmol/l) 479.9±34.0 491.6±58.3 461.0±35.3 456.2±29.3 0.937 
Aldosterone (nmol/l) 0.08±0.01 0.11±0.01*† 0.09±0.01 0.05±0.01* <0.001 
AngII (pmol/l) 3.84±0.58 3.02±0.51 4.0±0.62 3.5±0.54 0.624 
F2-isoprostanes (pg/ml) 12.4±1.2 24.2±1.1*† 12.9±1.1 14.8±0.7 0.015 

Responses to the infusion of I/H (fat) and S/G (saline)

Figure 2
Responses to the infusion of I/H (fat) and S/G (saline)

Values are means±S.E.M. of 17 subjects, except n=13 in ΔMNSA and n=8 in Δtotal MNSA. *P<0.05 compared with baseline; †P<0.05 compared with saline control day.

Figure 2
Responses to the infusion of I/H (fat) and S/G (saline)

Values are means±S.E.M. of 17 subjects, except n=13 in ΔMNSA and n=8 in Δtotal MNSA. *P<0.05 compared with baseline; †P<0.05 compared with saline control day.

Haemodynamic measurements

As shown in Table 3, I/H infusion was associated with a significant increase in SBP (systolic BP, +14.0±1.6 compared with +3.2±2.5 mmHg), DBP (diastolic BP, +8.2±1.0 compared with −0.1±1.7 mmHg), and HR (+8.2±1.0 compared with +2.4±1.2 beats/min) compared with S/G infusion (P<0.001), especially during the last 2 h of infusion. Despite the increase in HR with I/H infusion (Figure 3),

graphic
remained unchanged due to a decrease in SV, which was not different compared with the S/G infusion day. Paralleling the increase in MSNA, with I/H infusion SVR was significantly elevated compared with baseline (P<0.001) at 4 h, but did not change during S/G infusion. Although CBF (Figure 2) increased significantly by 2 h and remained elevated for the duration of the I/H infusion, the increase during S/G infusion was not significant until the final time point. The change in CBF was greater with I/H compared with S/G over the last 2 h of infusion.

Systemic haemodynamic responses to the infusion of I/H (fat) and S/G (saline)

Figure 3
Systemic haemodynamic responses to the infusion of I/H (fat) and S/G (saline)

Values are means±S.E.M. of 17 subjects. *P<0.05 compared with baseline; †P<0.05 compared with saline control day.

Figure 3
Systemic haemodynamic responses to the infusion of I/H (fat) and S/G (saline)

Values are means±S.E.M. of 17 subjects. *P<0.05 compared with baseline; †P<0.05 compared with saline control day.

MSNA and vascular resistance

Changes in MSNA burst frequency and total MSNA in response to I/H and S/G infusions, along with CVR, are shown in Figure 2. MSNA recordings could not be obtained in four subjects because of shifts in electrode position or failure to meet signal-to-noise criteria. Furthermore, data for total MSNA were excluded if there was a shift in electrode position at any point during the experiment, leaving data from eight subjects for the total MSNA quantification. Owing to the challenges in maintaining the microneurography recording over a period of 4–5 h, we recruited more subjects than necessary to attain sufficient statistical power to identify differences in MSNA in the case of missing data. Thus the more limited size of the MSNA sample should be regarded as adequate and representative of the overall sample. Of interest, haemodynamic responses were similar in individuals without and with suitable MSNA recordings.

The baseline burst frequency was not different between the two days, and Bland–Altman analysis revealed that 85% of the baseline MSNA values for both days were repeatable within 25%. This compares favourably with previous studies [21]. Both the change in burst frequency (+29.9% compared with +0.0%) and total MSNA (+65.1% compared with +12.3%) in comparison with baseline and S/G infusion were augmented after 2 h with I/H infusion and continued to increase until the end of infusion. CVR tended to decrease over the 4-h I/H infusion compared with S/G infusion (P=0.08). Finally, the change in SVR during I/H infusion was directly related to the change in burst frequency (Figure 4).

Direct relationship between the changes in MSNA and SVR during lipid infusion as determined using a linear least-squares fit regression

Figure 4
Direct relationship between the changes in MSNA and SVR during lipid infusion as determined using a linear least-squares fit regression
Figure 4
Direct relationship between the changes in MSNA and SVR during lipid infusion as determined using a linear least-squares fit regression

DISCUSSION

The primary new findings of the present study were that (i) acute elevation of the plasma NEFA concentration with infusion of I/H increases central sympathetic activation measured from sympathetic post-ganglionic neurons, (ii) the augmentation of MSNA was associated with a rise in insulin, but not leptin, concentrations, and (iii) NEFAs increased aldosterone and F2-isoprostanes, but not AngII, to contribute to the haemodynamic response.

Our present results confirm earlier reports that increased NEFA and/or TAG concentrations, as might be observed in insulin-resistant individuals following a meal, elevate BP and HR [79]. Our findings are similar to one study that reported a rise in SBP/DBP and HR by 13.5/8.0±2.1/1.5 mmHg and 9.4±1.4 beats/min respectively, following 4-h infusion of I/H, resulting from both an increase in SVR and

graphic
[7]. However, the increase in BP in the present study resulted from augmented SVR, despite a rise in CBF and no change in
graphic
. The observation that SVR was elevated while
graphic
was unaltered and CBF was increased with lipid infusion suggests a redistribution of blood flow and vasoconstriction in vascular beds other than muscle [4]. Most studies are in agreement that NEFAs promote insulin resistance and impair endothelium-dependent vasodilation [2224]; however, the endothelial dysfunction documented in these investigations is typically associated with a paradoxical increase in basal limb blood flow. Although few studies have examined the underlying mechanisms of increased basal limb blood flow from elevated NEFA levels [2527], possible explanations include insulin-induced NO release or vasodilating prostaglandin production [26].

To date, limited findings suggest [6,8], but have not definitively proven [13,14], a link between the rise in BP and sympathetic activation. In normotensive rats, the pressor response induced by oleic acid infusion was inhibited by the α1-adrenoceptor antagonist prazosin, suggesting sympathetic activation. Paolisso et al. [8] reported an increase in plasma NA with I/H infusion in humans, whereas another study [13] showed a small, but significant, decline in NA and a trend towards decreased renal venous NA spillover. It is difficult to reconcile these disparate findings, particularly in the absence of a control group and haemodynamic results in the latter study. More recently, a study assessing baroreflex function did not show a change in sympathetic activation during a 2-h lipid infusion [14]; however, the limited duration of that study may have precluded a complete sympathetic response to elevated NEFA concentrations. The increase in burst frequency (approx. 30%) and total MSNA (approx. 65%) recorded over the 4-h period in the present study clearly supports a role for central sympathetic activation in the pressor response to NEFAs. An additional finding in the present study is that the increase in SVR was directly associated with central sympathetic activation. Further studies blocking NEFA-induced sympathetic activation are necessary to determine a causal relationship.

Notably, NA concentrations decreased modestly during I/H infusion despite an increase in MSNA. This finding may at first appear counterintuitive; however, several studies, including one with lipid infusion [13], have documented a dissociation of the MSNA–plasma NA relationship with a modulation of blood flow, neuronal re-uptake, NA degradation or NA clearance [2832].

Possible mechanisms of central sympathetic activation

We considered four potential mechanisms by which NEFAs increased MSNA: (i) activation of vagal afferents, (ii) direct sensing in the hypothalamus, (iii) mediation by nutrient-sensing hormones, such as leptin and/or insulin, and (iv) modulation of the baroreflex. First, chemosensitive afferents arising from the abdominal viscera respond to lipids to influence discharge of vasomotor neurons of the rostral ventrolateral medulla [33]. Secondly, NEFAs cross the blood–brain barrier in proportion to circulating plasma NEFA concentrations and activate the POMC/CART (pro-opiomelanocortin/cocaine- and amphetamine-regulated transcript) neural pathway, resulting in decreased energy intake and activation of the SNS [34]. Although we cannot rule out the contribution of either of these mechanisms, the latter two can be examined in greater detail.

Hormonal mediation

Leptin and insulin have emerged as possible mediators of central sympathetic activation from NEFAs due to their characteristic responses to changes in adiposity and food intake. Both leptin and insulin stimulate POMC/CART neurons in the hypothalamus, which project to other brain centres and, ultimately, stimulate sympathetic pre-ganglionic neurons to modulate energy expenditure and BP [35]. Further support is shown by the observations that intravenous and intracerebroventricular administration of leptin increase both sympathetic activity and BP in conscious animals [36].

Our present results support a modest role for insulin as a mediator of sympathetic activation; however, leptin does not appear to be regulated extensively by elevated NEFA concentrations, at least in the acute setting. The approx. 30% increase in plasma insulin reported in the present study is in agreement with some [8,24,37], but not all [10,38], previous studies.

Previous studies have reported divergent findings regarding a leptin response to changes in NEFA concentrations. That peak leptin is positively correlated with a nocturnal peak in NEFA levels [39] suggests a link between NEFAs and leptin; however, results from lipid infusion in rats and humans are equivocal [4043]. As the nadir for leptin levels occurs in the early afternoon, it is possible that the decline in leptin observed in the present study on the control day simply reflected the diurnal rhythm, whereas elevated NEFA levels supported maintenance of leptin concentration. Nevertheless, its null response to NEFAs precludes leptin as a sympathetic activator in this study.

Baroreflex modulation

Two studies have presented conflicting evidence regarding acute baroreflex resetting and reduced baroreflex sensitivity in response to 1-h lipid infusion [14,44]. Although our present study was not designed to assess baroreflex function, the concomitant increase in BP, HR and MSNA supports the concept of resetting of the arterial baroreflexes to defend higher pressures. Alternatively, increases in MSNA may be a reflex response to muscle vasodilation with lipid infusion; however, two observations refute this notion. First, the BP response to I/H infusion far exceeds the expected baroreflex compensation, although CBF increased 20% within 1 h and 60% by 4 h of I/H infusion. The expected final MAP if the increase in CBF is representative of all skeletal muscle and if

graphic
and resistance in non-skeletal muscle beds remained at control levels would be 72.8 mmHg compared with the basal MAP of 76.2 mmHg, considerably less than the measured BP after I/H infusion. Secondly, as MSNA did not increase until after 2 h of infusion with I/H, whereas CBF and BP increased by 1 h, it is unlikely that MSNA augmentation resulted from a pure baroreflex response.

Could peripheral mechanisms contribute to the pressor response?

Although central sympathetic activation contributes to the pressor response to NEFAs, results from the present study also suggest other peripheral mechanisms are involved. As in previous studies [9], F2-isoprostanes, a biomarker of oxidative stress that has been linked to hypertension, doubled during I/H infusion in the present study, supporting its role in the BP response to NEFAs. Although no changes were observed in AngII concentrations on either experimental day, aldosterone increased approx. 38% by the end of I/H infusion, in contrast with the approx. 44% fall on the control day. The increase in aldosterone is consistent with a previous study that demonstrated the adrenal-stimulating effects of oxidized linoleic acid derivatives [45]. The reason for the decline in aldosterone on the control day is unclear; however, it is likely that the postural shift in blood volume and central venous pressure is a controlling factor [46]. Other mediators of a pressor response to NEFAs not measured in the present study include direct stimulation of protein kinases that cause constriction and increased sensitization of vessels to α-adrenergic stimulation [19]. From our present findings it appears likely that both peripheral and central mechanisms play a role in the haemodynamic response to elevated NEFA concentrations. Interestingly, in agreement with a previous study [7], a biphasic response was observed such that BP increased within the first 30 min and remained relatively stable, followed by an even greater increase during the final 2 h of lipid infusion. Additionally, in our present study, HR and SVR did not begin to rise until the final 2 h, coinciding with the increase in MSNA. These observations may reflect an initial direct vascular effect of fat followed by central activation (after approx. 2 h of infusion) and the corresponding haemodynamic responses.

Limitations

Results from previous studies have suggested that an infusion duration of at least 3–4 h was necessary to evoke the hypothesized hormonal, neural and haemodynamic responses [8,43]. Owing to the technical challenges of maintaining sympathetic nerve recordings, particularly due to the length of the study, two issues were of concern. First, any leg movement by the subject could alter the nerve recording or disrupt the signal altogether. Consequently, results could not be analysed for burst frequency in approx. 20% and for total MSNA in approx. 50% of subjects. Despite the reduction in MSNA data, the statistical power was large enough to identify differences for lipid infusion compared with control. Secondly, it is possible that stress or bladder distension contributed to the sympathetic and haemodynamic response [47] over the approximate 5-h basal and infusion period. To mitigate bladder distension, subjects voided using a condom catheter or urinal while remaining supine and stationary. Finally, the findings result from an acute increase in NEFA levels similar to that in previous studies [7,13,14]; any conclusion about the long-term effects of chronic augmentation of NEFA concentrations on BP or cardiovascular risk should be considered speculative.

Conclusions

The present study provides direct evidence that in humans the pressor response to NEFAs is mediated, at least in part, by sympathetic activation. Moreover, insulin, F2-isoprostanes and aldosterone, but not leptin, may contribute to the haemodynamic response during lipid infusion. The neurogenic response documented in the present study implicates central sympathetic activation as a possible mechanism linking conditions characterized by high NEFA concentrations (e.g. obesity and dyslipidaemia) with hypertension. With the age-adjusted prevalence rate of hypertension in the U.S.A. now over 28% [48], due in large part to the increasing obesity epidemic, more investigations of the mechanistic underpinnings of the NEFA–SNS link are warranted. In order to more precisely elucidate the relationship between NEFAs and sympathetic activation in some forms of hypertension, future studies should examine the effect of more chronic elevations of NEFA levels on autonomic control in humans and whether reducing the NEFA concentration results in a corresponding decrease in sympathetic activity.

FUNDING

This work was supported by an American College of Sports Medicine Foundation Research Grant; and the National Institutes of Health [grant numbers T32 AG00048, M01 RR10732].

We thank those who assisted in this research, the GCRC (General Clinical Research Center) at The Pennsylvania State University, University Park, PA 16802, U.S.A. for outstanding clinical support, the Core Labs at The Pennsylvania State University, University Park, PA 16802, U.S.A. for hormone analyses, and the research subjects for their amiable participation.

Abbreviations

     
  • AngII

    angiotensin II

  •  
  • BP

    blood pressure

  •  
  • CBF

    calf blood flow

  •  
  • CVR

    calf vascular resistance

  •  
  • DBP

    diastolic BP

  •  
  • HR

    heart rate

  •  
  • I/H

    Intralipid®/heparin

  •  
  • MAP

    mean arterial pressure

  •  
  • MSNA

    muscle sympathetic nerve activity

  •  
  • NA

    noradrenaline

  •  
  • NEFA

    non-esterified fatty acid

  •  
  • POMC/CART

    pro-opiomelanocortin/cocaine- and amphetamine-regulated transcript

  •  
  • Q

    cardiac output

  •  
  • S/G

    saline/glycerol

  •  
  • SBP

    systolic BP

  •  
  • SNS

    sympathetic nervous system

  •  
  • SV

    stroke volume

  •  
  • SVR

    systemic vascular resistance

  •  
  • TAG

    triacylglycerol

  •  
  • V̇O2max

    maximal oxygen uptake

References

References
1
Fagot-Campagna
 
A.
Balkau
 
B.
Simon
 
D.
Warnet
 
J. M.
Claude
 
J. R.
Ducimetiere
 
P.
Eschwege
 
E.
 
High free fatty acid concentration: an independent risk factor for hypertension in the Paris Prospective Study
Int. J. Epidemiol.
1998
, vol. 
27
 (pg. 
808
-
813
)
2
Davda
 
R. K.
Stepniakowski
 
K. T.
Lu
 
G.
Ullian
 
M. E.
Goodfriend
 
T. L.
Egan
 
B. M.
 
Oleic acid inhibits endothelial nitric oxide synthase by a protein kinase C-independent mechanism
Hypertension
1995
, vol. 
26
 (pg. 
764
-
770
)
3
Egan
 
B. M.
Hennes
 
M. M.
Stepniakowski
 
K. T.
O'shaughnessy
 
I. M.
Kissebah
 
A. H.
Goodfriend
 
T. L.
 
Obesity hypertension is related more to insulin's fatty acid than glucose action
Hypertension
1996
, vol. 
27
 (pg. 
723
-
728
)
4
Bulow
 
J.
Madsen
 
J.
Hojgaard
 
L.
 
Reversibility of the effects on local circulation of high lipid concentrations in blood
Scand. J. Clin. Lab. Invest.
1990
, vol. 
50
 (pg. 
291
-
296
)
5
Grekin
 
R. J.
Vollmer
 
A. P.
Sider
 
R. S.
 
Pressor effects of portal venous oleate infusion. A proposed mechanism for obesity hypertension
Hypertension
1995
, vol. 
26
 (pg. 
193
-
198
)
6
Grekin
 
R. J.
Dumont
 
C. J.
Vollmer
 
A. P.
Watts
 
S. W.
Webb
 
R. C.
 
Mechanisms in the pressor effects of hepatic portal venous fatty acid infusion
Am. J. Physiol.
1997
, vol. 
273
 (pg. 
R324
-
R330
)
7
Stojiljkovic
 
M. P.
Zhang
 
D.
Lopes
 
H. F.
Lee
 
C. G.
Goodfriend
 
T. L.
Egan
 
B. M.
 
Hemodynamic effects of lipids in humans
Am. J. Physiol. Renal Physiol.
2001
, vol. 
280
 (pg. 
R1674
-
R1679
)
8
Paolisso
 
G.
Manzella
 
D.
Rizzo
 
M. R.
Ragno
 
E.
Barbieri
 
M.
Varricchio
 
G.
Varricchio
 
M.
 
Elevated plasma fatty acid concentrations stimulate the cardiac autonomic nervous system in healthy subjects
Am. J. Clin. Nutr.
2000
, vol. 
72
 (pg. 
723
-
730
)
9
Lopes
 
H. F.
Martin
 
K. L.
Nashar
 
K.
Morrow
 
J. D.
Goodfriend
 
T. L.
Egan
 
B. M.
 
DASH diet lowers blood pressure and lipid-induced oxidative stress in obesity
Hypertension
2003
, vol. 
41
 (pg. 
422
-
430
)
10
Polak
 
K.
Schmetterer
 
L.
Luksch
 
A.
Gruber
 
S.
Polska
 
E.
Peternell
 
V.
Bayerle-Eder
 
M.
Wolzt
 
M.
Krebs
 
M.
Roden
 
M.
 
Free fatty acids/triglycerides increase ocular and subcutaneous blood flow
Am. J. Physiol. Renal Physiol.
2001
, vol. 
280
 (pg. 
R56
-
R61
)
11
Hall
 
J. E.
 
The kidney, hypertension, and obesity
Hypertension
2003
, vol. 
41
 (pg. 
625
-
633
)
12
Straznicky
 
N. E.
Louis
 
W. J.
McGrade
 
P.
Howes
 
L. G.
 
The effects of dietary lipid modification on blood pressure, cardiovascular reactivity and sympathetic activity in man
J. Hypertens.
1993
, vol. 
11
 (pg. 
427
-
437
)
13
Grekin
 
R. J.
Ngarmukos
 
C. O.
Williams
 
D. M.
Supiano
 
M. A.
 
Renal norepinephrine spillover during infusion of nonesterified fatty acids
Am. J. Hypertens.
2005
, vol. 
18
 (pg. 
422
-
426
)
14
Monahan
 
K. D.
Dyckman
 
D. J.
Ray
 
C. A.
 
Effect of acute hyperlipidemia on autonomic and cardiovascular control in humans
J. Appl. Physiol.
2007
, vol. 
103
 (pg. 
162
-
169
)
15
Minson
 
C. T.
Halliwill
 
J. R.
Young
 
T. M.
Joyner
 
M. J.
 
Influence of the menstrual cycle on sympathetic activity, baroreflex sensitivity, and vascular transduction in young women
Circulation
2000
, vol. 
101
 (pg. 
862
-
868
)
16
Hokanson
 
D. E.
Sumner
 
D. S.
Strandness
 
D. E.
 
An electrically calibrated plethysmograph for direct measurement of limb blood flow
IEEE Trans. Biomed. Eng.
1975
, vol. 
22
 (pg. 
25
-
29
)
17
Triebwasser
 
J. H.
Johnson
 
R. L.
Burpo
 
R. P.
Campbell
 
J. C.
Reardon
 
W. C.
Blomqvist
 
C. G.
 
Noninvasive determination of cardiac output by a modified acetylene rebreathing procedure utilizing mass spectrometer measurements
Aviat. Space Environ. Med.
1977
, vol. 
48
 (pg. 
203
-
209
)
18
Wallin
 
B. G.
Eckberg
 
D. L.
 
Sympathetic transients caused by abrupt alterations of carotid baroreceptor activity in humans
Am. J. Physiol.
1982
, vol. 
242
 (pg. 
H185
-
H190
)
19
Stepniakowski
 
K. T.
Goodfriend
 
T. L.
Egan
 
B. M.
 
Fatty acids enhance vascular alpha-adrenergic sensitivity
Hypertension
1995
, vol. 
25
 (pg. 
774
-
778
)
20
Zambon
 
A.
Hashimoto
 
S. I.
Brunzell
 
J. D.
 
Analysis of techniques to obtain plasma for measurement of levels of free fatty acids
J. Lipid Res.
1993
, vol. 
34
 (pg. 
1021
-
1028
)
21
Kimmerly
 
D. S.
O'Leary
 
D. D.
Shoemaker
 
J. K.
 
Test-retest repeatability of muscle sympathetic nerve activity: influence of data analysis and head-up tilt
Auton. Neurosci.
2004
, vol. 
114
 (pg. 
61
-
71
)
22
de Kreutzenberg
 
S. V.
Crepaldi
 
C.
Marchetto
 
S.
Calo
 
L.
Tiengo
 
A.
Del Prato
 
S.
Avogaro
 
A.
 
Plasma free fatty acids and endothelium-dependent vasodilation: effect of chain-length and cyclooxygenase inhibition
J. Clin. Endocrinol. Metab.
2000
, vol. 
85
 (pg. 
793
-
798
)
23
Steinberg
 
H. O.
Tarshoby
 
M.
Monestel
 
R.
Hook
 
G.
Cronin
 
J.
Johnson
 
A.
Bayazeed
 
B.
Baron
 
A. D.
 
Elevated circulating free fatty acid levels impair endothelium-dependent vasodilation
J. Clin. Invest.
1997
, vol. 
100
 (pg. 
1230
-
1239
)
24
Steinberg
 
H. O.
Paradisi
 
G.
Hook
 
G.
Crowder
 
K.
Cronin
 
J.
Baron
 
A. D.
 
Free fatty acid elevation impairs insulin-mediated vasodilation and nitric oxide production
Diabetes
2000
, vol. 
49
 (pg. 
1231
-
1238
)
25
Kearney
 
M. T.
Chowienczyk
 
P. J.
Brett
 
S. E.
Sutcliffe
 
A.
Ritter
 
J. M.
Shah
 
A. M.
 
Acute haemodynamic effects of lipolysis-induced increase of free fatty acids in healthy men
Clin. Sci.
2002
, vol. 
102
 (pg. 
495
-
500
)
26
Stepniakowski
 
K. T.
Lu
 
G.
Davda
 
R. K.
Egan
 
B. M.
 
Fatty acids augment endothelium-dependent dilation in hand veins by a cyclooxygenase-dependent mechanism
Hypertension
1997
, vol. 
30
 (pg. 
1634
-
1639
)
27
Riksen
 
N. P.
Bosselaar
 
M.
Bakker
 
S. J. L.
Heine
 
R. J.
Rongen
 
G. A.
Tack
 
C. J.
Smits
 
P.
 
Acute elevation of plasma non-esterified fatty acids increases pulse wave velocity and induces peripheral vasodilation in humans in vivo
Clin. Sci.
2007
, vol. 
113
 (pg. 
33
-
40
)
28
Esler
 
M.
Jennings
 
G.
Lambert
 
G.
Meredith
 
I.
Horne
 
M.
Eisenhofer
 
G.
 
Overflow of catecholamine neurotransmitters to the circulation: source, fate, and functions
Physiol. Rev.
1990
, vol. 
70
 (pg. 
963
-
985
)
29
Floras
 
J. S.
 
Sympathetic activation in human heart failure: diverse mechanisms, therapeutic opportunities
Acta Physiol. Scand.
2003
, vol. 
177
 (pg. 
391
-
398
)
30
Khan
 
M. H.
Sinoway
 
L. I.
MacLean
 
D. A.
 
Effects of graded LBNP on MSNA and interstitial norepinephrine
Am. J. Physiol. Heart Circ. Physiol.
2002
, vol. 
283
 (pg. 
H2038
-
H2044
)
31
Thompson
 
J. M.
O'Callaghan
 
C. J.
Kingwell
 
B. A.
Lambert
 
G. W.
Jennings
 
G. L.
Esler
 
M. D.
 
Total norepinephrine spillover, muscle sympathetic nerve activity and heart-rate spectral analysis in a patient with dopamine β-hydroxylase deficiency
J. Auton. Nerv. Syst.
1995
, vol. 
55
 (pg. 
198
-
206
)
32
Leuenberger
 
U.
Gleeson
 
K.
Wroblewski
 
K.
Prophet
 
S.
Zelis
 
R.
Zwillich
 
C.
Sinoway
 
L.
 
Norepinephrine clearance is increased during acute hypoxemia in humans
Am.J. Physiol.
1991
, vol. 
261
 (pg. 
H1659
-
H1664
)
33
Ueta
 
Y.
Kannan
 
H.
Higuchi
 
T.
Negoro
 
H.
Yamaguchi
 
K.
Yamashita
 
H.
 
Activation of gastric afferents increases noradrenaline release in the paraventricular nucleus and plasma oxytocin level
J. Auton. Nerv. Syst.
2000
, vol. 
78
 (pg. 
69
-
76
)
34
Lam
 
T. K.
Schwartz
 
G. J.
Rossetti
 
L.
 
Hypothalamic sensing of fatty acids
Nat. Neurosci.
2005
, vol. 
8
 (pg. 
579
-
584
)
35
Elmquist
 
J. K.
 
Hypothalamic pathways underlying the endocrine, autonomic, and behavioral effects of leptin
Int. J. Obes. Relat. Metab. Disord.
2001
, vol. 
25
 
Suppl. 5
(pg. 
S78
-
S82
)
36
Dunbar
 
J. C.
Hu
 
Y.
Lu
 
H.
 
Intracerebroventricular leptin increases lumbar and renal sympathetic nerve activity and blood pressure in normal rats
Diabetes
1997
, vol. 
46
 (pg. 
2040
-
2043
)
37
Vollenweider
 
L.
Tappy
 
L.
Owlya
 
R.
Jequier
 
E.
Nicod
 
P.
Scherrer
 
U.
 
Insulin-induced sympathetic activation and vasodilation in skeletal muscle. Effects of insulin resistance in lean subjects
Diabetes
1995
, vol. 
44
 (pg. 
641
-
645
)
38
Fugmann
 
A.
Millgard
 
J.
Sarabi
 
M.
Berne
 
C.
Lind
 
L.
 
Central and peripheral haemodynamic effects of hyperglycaemia, hyperinsulinaemia, hyperlipidaemia or a mixed meal
Clin. Sci.
2003
, vol. 
105
 (pg. 
715
-
721
)
39
Heptulla
 
R.
Smitten
 
A.
Teague
 
B.
Tamborlane
 
W. V.
Ma
 
Y. Z.
Caprio
 
S.
 
Temporal patterns of circulating leptin levels in lean and obese adolescents: relationships to insulin, growth hormone, and free fatty acids rhythmicity
J. Clin. Endocrinol. Metab.
2001
, vol. 
86
 (pg. 
90
-
96
)
40
Chen
 
M. D.
Song
 
Y. M.
 
Effect of lipid infusion on plasma leptin and neuropeptide Y levels in women
Mayo Clin. Proc.
2002
, vol. 
77
 (pg. 
1391
-
1392
)
41
Garcia-Lorda
 
P.
Nash
 
W.
Roche
 
A.
Pi-Sunyer
 
F. X.
Laferrere
 
B.
 
Intralipid/heparin infusion suppresses serum leptin in humans
Eur. J. Endocrinol.
2003
, vol. 
148
 (pg. 
669
-
676
)
42
Nisoli
 
E.
Carruba
 
M. O.
Tonello
 
C.
Macor
 
C.
Federspil
 
G.
Vettor
 
R.
 
Induction of fatty acid translocase/CD36, peroxisome proliferator-activated receptor-γ2, leptin, uncoupling proteins 2 and 3, and tumor necrosis factor-α gene expression in human subcutaneous fat by lipid infusion
Diabetes
2000
, vol. 
49
 (pg. 
319
-
324
)
43
Wang
 
J.
Liu
 
R.
Hawkins
 
M.
Barzilai
 
N.
Rossetti
 
L.
 
A nutrient-sensing pathway regulates leptin gene expression in muscle and fat
Nature
1998
, vol. 
393
 (pg. 
684
-
688
)
44
Gadegbeku
 
C. A.
Shrayyef
 
M. Z.
Taylor
 
T. P.
Egan
 
B. M.
 
Mechanism of lipid enhancement of α1-adrenoceptor pressor sensitivity in hypertension
J. Hypertens.
2006
, vol. 
24
 (pg. 
1383
-
1389
)
45
Goodfriend
 
T. L.
Ball
 
D. L.
Egan
 
B. M.
Campbell
 
W. B.
Nithipatikom
 
K.
 
Epoxy-keto derivative of linoleic acid stimulates aldosterone secretion
Hypertension
2004
, vol. 
43
 (pg. 
358
-
363
)
46
Gauer
 
O. H.
 
Recent advances in the physiology of whole body immersion
Acta Astronaut.
1975
, vol. 
2
 (pg. 
31
-
39
)
47
Fagius
 
J.
Karhuvaara
 
S.
 
Sympathetic activity and blood pressure increases with bladder distension in humans
Hypertension
1989
, vol. 
14
 (pg. 
511
-
517
)
48
Cutler
 
J. A.
Sorlie
 
P. D.
Wolz
 
M.
Thom
 
T.
Fields
 
L. E.
Roccella
 
E. J.
 
Trends in hypertension prevalence, awareness, treatment, and control rates in United States adults between 1988–1994 and 1999–2004
Hypertension
2008
, vol. 
52
 (pg. 
818
-
827
)