Aerobic interval training vs. continuous moderate exercise as a treatment for the metabolic syndrome - “A Pilot Study”

Abstract

Background:

Individuals with the metabolic syndrome are three times more likely to die from heart disease than healthy counterparts. Exercise-training reduces several of the symptoms of the syndrome, but which exercise-intensity that yields maximal beneficial adaptations is controversial. We compared moderate to high exercise-intensity with regard to variables associated with cardiovascular function and prognosis in patients with the metabolic syndrome.

Methods and Results:

Thirty-two metabolic syndrome patients (52.3±3.7 yrs, maximal oxygen uptake (VO_2max) 34 ml/kg/min) were randomized to equal volumes of either moderate continuous-training (CME at 70% of highest measured heart rate; Hf_max), aerobic interval-training (AIT at 90% of Hf_max) 3-times/week for 16-weeks, or to a control group. VO_2max increased more after AIT vs. CME (35% vs. 16%, p<0.01) and was associated with removal of more risk factors that constitute the metabolic syndrome (Number of factors in AIT: 5.9 pre vs. 4.0 post, (p<0.01); CME: 5.7 pre vs. 5.0 post, group difference p<0.05). AIT was superior compared to CME in enhancing endothelial function (9% vs. 5%, p<0.001), insulin signaling in fat and skeletal muscle, skeletal muscle biogenesis and excitation-contraction coupling, and reducing blood glucose and lipogenesis in adipose tissue. The two exercise-programs were equally effective at lowering mean arterial blood pressure and reducing body weight (−2.3 and −3.6 kg in AIT and CME, respectively) and fat.

Conclusions:

Exercise-intensity was an important factor for improving aerobic capacity and reversing the risk factors of the metabolic syndrome. These findings may have important implications for exercise training in rehabilitation programs and future studies.

Introduction

The metabolic syndrome is a cluster of cardiovascular risk factors including elevated blood pressure, dyslipidemia, impaired glycemic control, and abdominal obesity.¹ This syndrome affects about 25% of the adult US population and is reaching an epidemic spread in parallel to obesity, which is estimated to affect about 312 million people worldwide. With at least 1.1 billion overweight people, the incidence of the metabolic syndrome is expected to continue to rise,² warranting thorough mechanistic understanding to enable optimal treatment at a socioeconomic scale.

The metabolic syndrome is associated with increased cardiovascular morbidity and mortality, as individuals with the metabolic syndrome are three times more likely to die from coronary heart disease than healthy counterparts, after adjusting for conventional cardiovascular risk factors.³ Aerobic fitness and endothelial function are reduced in individuals with the metabolic syndrome and have been proposed to be independent and strong predictors of mortality relative to other established risk factors.^{4, 5} Although a direct cause-effect relationship has not been established, impaired aerobic capacity is a common factor underlying cardiovascular and metabolic diseases. Indeed, it has been well established that exercise training at least partly, reverses the metabolic syndrome,³ but the optimal level of exercise needed to prevent and treat the metabolic syndrome and its associated cardiovascular abnormalities remains undefined.

Here we address this issue by testing the efficacy of two distinctly different modes of exercise to reverse features of the metabolic syndrome. That is, adaptations were measured in response to either continuous moderate exercise (CME) or aerobic interval training (AIT) at a high intensity in humans with clinically defined metabolic syndrome.

Methods

Patients

Thirty two patients with the metabolic syndrome defined according to the WHO-criteria⁶ were included. All subjects provided written, informed consent, and the regional committee for medical research ethics approved the protocol. Patients were randomized and stratified (age and gender) into aerobic interval training (AIT, n=12), continuous moderate exercise (CME, n=10) or a control group (n=10). Registration of diet behavior was performed using a questionnaire previously described.⁷

Maximal oxygen uptake (VO_2max)

Before measurements of VO_2max, subjects were informed about the test, and instructed to exercise to their maximum limit. Familiarization to the treadmill, warm-up, and the VO_2max ramp-procedure have been described in detail previously.⁸ A leveling-off of oxygen uptake despite increased work load and a respiratory exchange ratio >1.05 were used as criteria for reaching the true VO_2max, and this was achieved in all individuals in the present study.

Training protocols

Both exercise groups performed endurance training as walking/running “uphill” on a treadmill 3x/week for 16 weeks. AIT warmed-up for 10min at 70% of maximal heart frequency (Hf_max) before performing 4x4min intervals at 90% of Hf_max, with 3-min active recovery at 70% of Hf_max between intervals, and a 5min cool-down period, giving a total exercise time of 40min. To equalize training volumes to similar amounts of kcal/session in the two groups, CME had to perform 47-min at 70% of Hf_max each exercise session, as previously detailed.⁸ The control group followed advice from family physicians. Hearth frequency was continuously monitored during exercise to ensure the subjects trained at the intended intensity.

Endothelial function and blood pressure

Endothelial function was measured as flow-mediated dilatation (FMD) using high resolution vascular ultrasound (14 MHz echo Doppler probe, Vivid 7 System, GE Vingmed Ultrasound, Horten, Norway) according to the current guidelines.⁹ The procedures for measuring FMD and blood pressure were recently published by our group.¹⁰ Shear rate was calculated as blood flow velocity/diameter (cm·s⁻¹/cm) as previously published.¹¹ Since no group differences were found, data are not presented. All ultrasound images were analyzed in random order, using EchoPACtm (GE Vingmed Ultrasound AS, Horten, Norway) by a blinded investigator.

Biopsies

One week prior to the VO_2max-test, and at least four days after the last training session, muscle biopsies of the m. vastus lateralis were taken using a 5 mm diameter biopsy needle (Bergstrom, Sweden) under local anesthesia (2% Lidocaine). A small muscle sample was used immediately for sarcoplasmic reticulum (SR) Ca²⁺ ATPase (SERCA)-1 and 2 Ca²⁺ transport measurements using a fluorescence microscope as recently described.¹² The remaining biopsy was immediately frozen in liquid nitrogen and stored at −80°C. A fat biopsy was taken with the subjects lying in a prone position, by an aspiration from the gluteal fat pad using a 20 mL sterile plastic syringe and a 15 gauge needle. Fat samples were immediately frozen in liquid nitrogen and stored at −80°C.

Protein phosphorylation and expression levels

Frozen tissue samples were homogenized in ice-cold lysis buffer (1% Triton-X-100 in the presence of phosphatase and protease inhibitors) and the lysates treated with 100 nM insulin (Sigma) or buffer alone for 10min prior to phosphorylation in the presence of cold ATP. Proteins were immunoprecipitated with an anti-insulin receptor antibody (α-IR) (Santa Cruz Biotechnology Inc), analyzed on 7% SDS-PAGE and transferred to nitrocellulose membranes prior to detecting phosphorylated residues with a monoclonal anti-phosphotyrosine antibody (Cell Signalling Technology Inc). The blots were reprobed with an antibody against the β subunit of IR (Santa Cruz Biotechnology Inc) to account for the amount of IR in immunopellets. Moreover, 5 μg of total lysates were analyzed by 4-12% SDS-PAGE and immunoblotting with antibodies against fatty acid transporter protein 1 (FATP-1), fatty acid synthase (FAS) and the peroxisome proliferative activated receptor γ co-activator 1α (PGC-1α) (Santa Cruz Biotechnology Inc), before re-probing with antibodies against actin or tubulin to normalize for the amount of protein loaded. Following incubation with horseradish peroxidase-conjugated anti-IgG antibodies, proteins were detected by enhanced chemiluminescence (ECL; Amersham Pharmacia Biotech) and quantified by densitometry.

Blood analyses

If not otherwise stated, all blood analyses were performed using standard local procedures. Oxidized LDL and adiponectin were measured in plasma using specific Mercodia ELISA kits (Mercodia, Uppsala, Sweden), total nitrite (NO₂⁻) concentration was quantified using a commercially available assay for nitric oxide (NO)-detection (R&D systems, Inc., Minneapolis, MN, USA), and plasma insulin was analyzed by a radioimmunoassay (RIA) kit (Linco Research, St. Charles, MI). To estimate β-cell function and overall insulin sensitivity, the homeostasis assessment model (HOMA) was used.

Statistical analyses

The primary outcome variable was VO_2max. Prior experience suggests a standard deviation (SD) of about 2-3.¹³ No formal sample size calculation was done, but with 10 subjects in each group, a standardized within-group difference of 1.0 may be detected using a paired t-test with 80% power, at a significance level of 5%.¹⁴ Clinically, this corresponds to a detectable difference for VO_2max of 3 ml/kg/min. Continuous variables are presented as mean±SD for descriptive purposes, and as mean±SEM (standard error of the mean) when group differences are of main interest. Comparison of continuous variables was done according to design (number of groups, paired or unpaired), using the linear model with correction for baseline value.¹⁵ Wilcoxon's, Mann-Whitney's, or Kruskal-Wallis' non-parametric procedures were used if the assumption of normality and homogeneity of variance was in doubt. Within-group changes of individual risk factors constituting the metabolic syndrome was tested using McNemar's test, while the total number of factors changing across groups was tested using Mann-Whitney's test. The potential for spurious significance due to multiple testing and a large number of variables relative to subjects is recognized, but as this was a pilot study no correction for multiple comparisons was applied. Reported p-values are two-sided, and referred to within group comparison unless specified otherwise and values less than 0.05 are considered as statistically significant. SPSS® 14.0 (SPSS Inc. Chicago, IL.) was employed for all analyses.

Statement of responsibility

The authors had full access to and take full responsibility for the integrity of the data. All authors have read and agree to the manuscript as written.

Results

Baseline characteristics

At baseline patients had comparable body weights, body mass indexes (BMI), waist-to-hip circumference ratios (WHR), blood plasma parameters, and blood pressures (Tables 1-3). During the experimental period, four patients were unable to complete the training program; one relocated (AIT), one withdrew for personal reasons (CME), one had a minor hip injury (Control), and one refused to perform follow-up tests for personal reasons (CME) (Flow chart). Thus, 28 patients carried out 90±2% of the scheduled training sessions.

Table 1.

Patient characteristics

	CONTROL	MODERATE	INTERVAL
	Baseline	Baseline	Baseline
PHYSIOLOGICAL
Age, years	49.6±9.0	52.0±10.6	55.3±13.2
Gender, male/female	5/4	4/4	4/7
Weight, kg	96.4±12.1	91.2±19.5	91.8±17.5
Body mass index, kg/m2	32.1±3.3	29.4±4.9	29.8±5.5
Waist hip ratio	0.97±0.08	0.97±0.08	0.94±0.07
MEDICATIONS
Angiotensin II Blockers	4	1	2
β-Blockers	1	1	0
Calcium antagonists	1	1	1
α-Blockers	0	0	1
Statins	0	0	2
Acetylsalicylic acid	0	0	1
Metformin	1	1	1

Data are presented as mean ± SD or number of patients.

Table 3.

Overview over number of patients that had the different variables of the metabolic syndrome according to the definition (see methods)

	CONTROL		MODERATE		INTERVAL
	Baseline	Post	Baseline	Post	Baseline	Post
No. of metabolic syndrome factors, median (range)	5 (4-8)	6 (3-9)	5 (3-7)	4 (1-8)	6 (4-7)	4 (2-6)**$
No. of patients with metabolic syndrome	9/9	9/9	8/8	5/8	11/11	6/11*$
Separate variables
Fasting Glucose	8/9	9/9	6/8	7/8	11/11	7/11
2-h Post Glucose Load	6/9	5/9	7/8	3/8	8/11	4/11
High density lipoprotein	9/9	9/9	6/8	6/8	10/11	8/11
Triglycerides	3/9	6/9	2/8	3/8	4/11	5/11
Body mass index	5/9	6/9	4/8	3/8	4/11	4/11
Waist hip ratio	9/9	8/9	8/8	6/8	9/11	11/11
Systolic Blood Pressure	6/9	5/9	2/8	2/8	6/11	3/11
Diastolic Blood Pressure	4/9	6/9	5/8	3/8	8/11	4/11
Microalbuminuria	2/9	1/9	3/8	1/8	2/11	0/11

Data are presented as median (range), or number of patients.

^*Significantly different within each group from pre to post (p<0.05).

^**Significantly different within each group from pre to post (p<0.01).

^$Significant from the other groups (p<0.05).

Figure 3.

Flow chart of the study design.

Clinical follow-up

Despite no diet alterations during the 16 week-intervention period (data not shown), AIT and CME patients exhibited a slight reduction of 3% and 4%, respectively, in body weight (both p<0.05, with no difference when AIT and CME are compared, Table 1), and BMI (Table 1, both p<0.05). Similarly, the waist width was reduced by 5 and 6 cm in AIT and CME, respectively (p<0.05, Table 2). AIT reduced more of the other factors of the metabolic syndrome, including insulin sensitivity (HOMA) (group difference, p<0.05, Tables 2 and 3), when compared to CME. At the end of the training period, 46% (p<0.05) in AIT and 37% in CME (p=0.23; between-group difference p<0.05) were no longer diagnosed with the metabolic syndrome, as opposed to the control group in which the syndrome persisted in all subjects (Table 3).

Table 2.

Parameters related to the metabolic syndrome before and after the experimental period

	CONTROL		MODERATE		INTERVAL
	Baseline	Post	Baseline	Post	Baseline	Post
BLOOD VARIABLES
Insulin, pmol/Ml	115.1±23.4	111.4±20.2	104.2±27.4	105.5±24.1	111.2±34.6	113.2±7.0
Fasting Glucose, mmol · L−1	6.1±0.2	6.8±0.3*	6.1±0.5	6.5±0.6	6.9±0.6	6.6±0.6$
Insulin sensitivity, (HOMA, %)	60.0±7.9	59.3±8.2	64.4±5.7	50.2±4.9	62.2±8.0	77.2±4.9*§
β-cell function (HOMA, %)	80.9±10.3	76.7±8.2	78.9±11.0	80.9±8.7	76.8±12.6	97.0±9.2*§
Microalbuminuria, Alb.excr, μg/min	42.2±32.3	42.8±32.2	21.4±10.0	13.5±5.8	23.0±11.3	7.7±1.3
High density lipoprotein, mmol · L−1	0.62±0.05	0.58±0.08	0.74±0.09	0.80±0.08	0.69±0.07	0.84±0.10*
Triglycerides, mmol · L−1	1.84±0.40	2.00±0.54	1.47±0.45	1.67±0.38	1.65±0.20	1.70±0.19
Adiponectin, μg/mL	6.4±2.0	7.0±1.1	6.7±1.5	8.2±1.4*#	7.8±2.3	9.4±3.0*#
BODY COMPOSITION
Weight, kg	96.4±4.0	96.2±4.9§	91.2±6.9	87.6±6.5*	91.8±5.3	89.5±4.9*
Body mass index, kg/m2	32.1 1.1	32.0±1.3	29.4±1.7	28.2±1.5*#	29.8±1.7	29.1±1.5*
Waist, cm	114.3±2.7	112.0±3.4	105.1±5.3	99.1±5.0*#	105.5±4.1	100.5±3.6*
Waist hip ratio	0.97±0.03	0.96±0.03	0.97±0.03	0.93±0.03	0.94±0.02	0.94±0.02
BLOOD PRESSURE
Systolic Blood Pressure, mmHg	146±6	141±5	131±6	121±5*	144±5	135±5*
Diastolic Blood Pressure, mmHg	95±5	96±4	88±4	82±5	95±3	89±3*
Mean arterial blood pressure, mmHg	111±5	108±5	102±4	95±5*	111±3	105±3*
EXERCISE PARAMETERS
VO2max, ml/kg/min	32.3±3.4	33.7±2.7	36.0±3.2	41.6±3.6*#	33.6±2.5	45.3±3.3**$
VO2max, ml/kg0.75/min	101.1±10	105.1±8	110.8±10	126.5±7*#	103.2±7	138.0±9**$
Peak heart rate, beats · min−1	175±9	184±3	189±4	192±4	176±5	178±5
Peak oxygen pulse, mL · beats−1	17.8±1.6	17.5±1.4	17.3±1.9	18.9±2.2	17.2±1.1	22.2±1.8*$

Data are presented as mean ± SEM.

^*Significantly different within each group from pre to post (p<0.05).

^**Significantly different within each group from pre to post (p<0.01).

^$Significantly different from the other groups (p<0.05).

^#Different from control (p<0.05).

VO_2max, maximal oxygen uptake.

Blood pressure

Both AIT and CME decreased systolic and diastolic blood pressures, by ∼10 mmHg (both, p<0.05), and ∼6 mmHg (AIT p<0.05 and CME p=0.24, Table 2), respectively.

Aerobic capacity

AIT and CME increased VO_2max by 35% and 16% (p<0.01), respectively (group difference p<0.01, Figure 1A). Only AIT had a significantly increased peak O₂ pulse (p<0.05, between-group difference p<0.001, Table 2).

Figure 1.

Maximal oxygen uptake (VO_2max) (A). Expression of peroxisome proliferactive activated receptor γ co-activator 1α (PGC-1α) in samples from m. vastus lateralis (B). Maximal rate of re-uptake of Ca²⁺ into sarcoplasmatic reticulum in m. vastus lateralis (C). Endothelial function measured as flow mediated dilatation (FMD) of the brachial artery in man (D). * Significantly different within each group from pre to post (p<0.05). **Significantly different within each group from pre to post (p<0.01). ^$ Significantly different from the other groups (p<0.05).^# different from control (p<0.05).

PGC-1α and SERCA

Decreased exercise capacity in individuals with the metabolic syndrome may be linked to skeletal muscle abnormalities related to mitochondrial biogenesis and excitation-contraction coupling.^{13, 16} Therefore, we examined the levels of PGC-1α, a critical coordinator for activation of metabolic genes controlling substrate utilization and mitochondrial biogenesis, and the re-uptake of Ca²⁺ into the SR. We found that PGC-1α levels in m. vastus lateralis increased by 138% by AIT, but were not significantly altered in the other groups (group difference p<0.01; Figure 1B). Furthermore, AIT, but not CME, increased the maximal rates of SR Ca²⁺ uptake by ∼50% (p<0.03, Figure 1C) with p<0.05 between group differences (Figure 1C).

Endothelial function

AIT and CME improved FMD by 9% (p<0.001) and 5% (p<0.001), respectively (group difference p<0.001; Figure 1D). Endothelium-independent dilatation by administrating nitroglycerine sublingually was equal in all groups (data not shown).

We observed increased availability of NO (36±3%, p<0.03) after AIT, but not CME (p=0.37, group difference p<0.05). Furthermore, oxidized LDL, which negatively regulates the bioavailability of NO, was reduced by 17%, (Pre: 102±8 vs. Post: 85±7 U/mL, p<0.001) in AIT, but not CME (not shown) (group differences p<0.01).

Insulin action in skeletal muscle and fat tissue

Because insulin receptor (IR) phosphorylation in skeletal muscle is an excellent surrogate measure of peripheral insulin sensitivity in obese and type 2 diabetic humans,^{17, 18} phosphorylation of the IR β-subunit (IR_β) in the presence of insulin was examined. Insulin induced a substantial increase (2.5-15 fold, p<0.001, Figure 2A) in receptor phosphorylation (and thus activation) in m. vastus lateralis after AIT as compared to the other groups (p<0.03). In fat tissue, AIT induced a larger increase in receptor phosphorylation compared to the other groups (p<0.01, Figure 2B).

Figure 2.

Insulin action in m. vastus lateralis (A), in fat tissue (B). pIR; phosphorylated insulin receptor. Relative expression of fatty acid transporter protein 1 (FATP-1) in fat tissue (C). Relative expression of fatty acid synthase (FAS) in fat tissue (D). * Significantly different within each group from pre to post (p<0.05). ** Significantly different within each group from pre to post (p<0.01). ^$ Significantly different from the other groups (p<0.05). ^# different from control (p<0.05).

FATP-1 and FAS

Endurance training did not affect the protein content of FATP-1, a fatty acid transporter, in the m. vastus lateralis (data not shown). In contrast, FATP-1 protein levels decreased 3-4 fold in fat tissue after AIT, but not CME (group difference p<0.05, Figure 2C). Consistently, AIT caused a more marked decrease in the protein content of FAS, a key lipogenic enzyme, in white adipose tissue compared to the other groups (p<0.01, Figure 2D), suggesting reduced lipogenesis in this tissue by AIT, but not CME.

Blood analysis

AIT caused a significant improvement in fasting blood glucose compared to the other groups (p<0.05, Table 2). Furthermore, insulin sensitivity and β-cell function appear to be mainly improved by AIT, as indicated by HOMA analysis (group difference p<0.05, Table 2). Moreover, HDL-cholesterol was increased by ∼25% after AIT (p<0.05, Table 2), but remained unaltered in the other groups (group difference n.s.). Consistent with improved insulin sensitivity by AIT, circulating adiponectin, an adipocyte-secreted hormone that is associated with improved insulin sensitivity and amelioration of the metabolic syndrome, was increased (p<0.05, Table 2). Similarly, Adiponectin levels were increased by CME (p<0.05, Table 2), perhaps due to a comparable decrease in visceral obesity, as suggested by reduced waist circumference in patients undergoing either exercise program (Table 2). Neither training program changed the levels of triglycerides, insulin, microalbumin (Table 2), total cholesterol, LDL, C-peptide or hemoglobin (data not shown).

Discussion

The current study emphasizes that exercise-intensity was an important factor for improving aerobic capacity and reversing the risk factors of the metabolic syndrome including endothelial function, insulin action and lipogenesis in individuals with the metabolic syndrome.

Aerobic capacity

Of all established risk factors, low aerobic exercise capacity appears to be the strongest predictor of mortality.⁵ We herein demonstrated that high-intensity exercise increased VO_2max to a higher degree than moderate-intensity exercise in patients with the metabolic syndrome. Central O₂ delivery and peripheral O₂ utilization in addition to skeletal muscle capacity contribute to training-induced changes in VO_2max.¹³ Accordingly, AIT improved maximal stroke volume (measured as peak O₂ pulse), as well as Ca²⁺ cycling and mitochondrial capacity in skeletal muscle (as assessed by improved SERCA-capacity and PGC-1α levels in m. vastus lateralis) to a larger extent than CME.

Similar weight loss between exercise intensities, but a larger total reduction of the cardiovascular risk factors that constitute the metabolic syndrome after AIT, are in line with epidemiological observations suggesting that it may be more important to become fit than to loose weight per se.^{19, 20} Moreover, although obesity and aerobic capacity are strong and independent prognostics markers for cardiovascular mortality, the link between aerobic capacity and mortality seems to be stronger.^{19, 20} This suggests that it is more beneficial to target aerobic capacity before weight loss, although improvement of both risk factors probably would be ideal. However, larger prospective studies are needed to address this issue.

The main goal of the present study was to evaluate the effect of exercise intensity on the metabolic syndrome by subjecting patients to two different exercise regimens, which differ solely by their level of intensity. The high intensity exercise was set to 90% of Hf_max and performed as aerobic interval exercise because this training method has previously yielded the greatest improvement of aerobic capacity over a relatively short time in healthy individuals,²¹ and in patients with coronary artery disease,⁸ intermittent claudi-catio²² and post-infarction heart failure.¹³ The rationale behind interval training is that most evidence suggests that it is the pumping capacity of the heart (i.e. stroke volume) that limits VO_2max, and the interval design allows for rest periods that enable the patients to complete short work periods at higher intensities, which thereby challenge the heart's pumping ability more than what would be possible by lower intensities.

There is a general agreement that we need better and more affordable prevention and treatment strategies to improve wide-scale health outcome and slow down the current epidemic of overweight to prevent the epidemic of metabolic syndrome from reaching global proportions and straining public health and economy.²³ Furthermore, weight regulation has a complex nature, such that the likelihood of a pharmacological solution is small.²⁴ In fact, there is no ideal drug at the present time for the majority of patients with metabolic syndrome.²⁵ The current study suggests that exercise in general and AIT in particular is partly or fully able to reverse the metabolic syndrome, suggesting that this may be a promising treatment strategy. However, despite the results presented here, the interval protocol may not be readily acceptable to the general population of patients with the metabolic syndrome. On the other hand, an exercise intensity of 90% of Hf_max corresponds to ∼20 beats/min below maximum. This implies that the patients would have tolerated higher exercise intensities (but with shorter durations), and that most patients walk “up-hill” on the treadmill to avoid running and thereby reduce the risk of musculoskeletal injuries caused by the mechanical load at high running speeds. Interestingly, informal comments from the patients in the exercise groups indicate that patients in AIT found it motivating to have a varied procedure to follow during each training session, whilst those in CME found it “quite boring” to walk continuously during the whole exercise period. To extend the clinical usefulness of the exercise training for those not having access to heart rate monitors, we recently asked patients with post-infarction heart failure to indicate the level of effort on a Borg scale after each interval or during continuous walking, i.e. identical training protocols as in the present study, and found Borg scores of 17± and 12± to correspond to AIT and CME, respectively.¹³ Unpublished data from our laboratory confirm these numbers in both healthy and overweight individuals. To obtain the intended exercise intensity during home-based interval training, we normally instruct patients to exercise such that they are breathing heavily and talking becomes uncomfortable during each interval, without developing severe leg stiffness, and that it is an absolute requirement that they should be able to perform 4-minute intervals in succession. These instructions were satisfactory in old patients with post-infarction heart failure,¹³ and may therefore also be feasible for patients with the metabolic syndrome. These observations suggest that it may be possible to instruct patients to perform such training without monitoring heart rate.

Endothelial function

High intensity training improved endothelial function (FMD) more than continuous exercise. In keeping with the notion that exercise-induced improvement in vessel relaxation is mainly mediated by NO,²⁶ we observed no group-differences in endothelium-independent relaxation in the brachial artery after administrating nitroglycerine sublingually. Recently, Hambrecht et al²⁷ demonstrated 2-fold higher endothelial NO synthase (eNOS) and a 4-fold higher eNOS Ser¹¹⁷⁷-phosphorylation levels in the left internal mammary artery after 4 weeks of exercise training in patients with stable coronary artery disease compared to sedentary counterparts. This suggests that the effect of AIT is mediated by improved NO-bioavailability. Consistent with this hypothesis, AIT, but not CME, normalized the level of blood glucose, insulin sensitivity and the amount of oxidized LDL, which directly influence NO bioavailability.²⁸ The reason for the differences in FMD between groups is not fully understood, but it seems reasonable to suggest that the low- and high-intensity training exercise programs differently affect shear stress on the walls of blood vessels during exercise training, and that this yields differences in molecular responses.

Blood pressure

CME and AIT reduced systolic and diastolic blood pressure by ∼10 and 6 mmHg, respectively. Based on a meta-analysis of one million adults, a blood pressure lowering of this magnitude would in the long term be associated with about 40% and 30% decreases in the risks of premature deaths due to stroke and ischemic heart disease.²⁹

Insulin action

Physical activity and a healthy lifestyle promote insulin sensitivity, while a sedentary lifestyle and a western diet are associated with insulin resistance. In our study, AIT promoted insulin sensitivity and β-cell function more than CME. Increased insulin sensitivity and β-cell function after exercise have been proposed to result from peripherally enhanced insulin response and signalling.³⁰ Consistently, exercise, and in particular high-intensity training, improved IR phosphorylation (and activation) by insulin in muscle and fat tissue. The mechanism for improved insulin action in muscle by AIT is not clear. However, exercise is generally known to promote insulin action in muscle by decreasing the intracellular accumulation of triglycerides and promoting fatty acid oxidation.³¹ Consistently, AIT increased mitochondrial biogenesis in skeletal muscle of patients with the metabolic syndrome. In agreement with recent reports,³² AIT did not alter fatty acid transport into muscle, as indicated by intact FATP-1 levels. In contrast, the larger improvement of IR activation by high-intensity training in fat tissue was associated with a stronger reduction of fatty acid uptake and lipogenesis in this tissue, as suggested by reduced FATP-1 and FAS levels, respectively. Although the effect of exercise on visceral adiposity per se was not measured, decreased lipogenesis together with reduced waist circumference and body weight, in parallel to increased circulating adiponectin levels in exercise suggest decreased intra-abdominal visceral obesity. Taken together, the data suggest that both exercise programs decreased body weight, but AIT was particularly beneficial for decreasing fatty acid transport into the adipose tissue and promoting the suppressive effect of insulin on lipogenesis in this tissue. Together with reduction in plasma oxidized LDL levels, decreased lipogenesis in fat tissue by AIT demonstrates larger improvement in the overall fat metabolism in patients with the metabolic syndrome by high-intensity training.

AIT's positive effect on fasting glucose levels and insulin sensitivity may explain some of the group differences in these parameters. The slight worsening in CME with respect to these variables is not fully understood; especially since there were no diet alterations during the 16 week-intervention period. However, this suggests that CME, in contrast to AIT, did not produce enough stimulus to revert the phenotype.

Limitations to the study

The number of patients in our study was small and with limited inference about the safety and injury risk of this training protocol in the general population. However, in a number of studies in our laboratory,^{8, 13, 21, 22} we have observed only two knee injuries in extremely overweight patients. These data are provocative, but should encourage larger multi-center studies.

Conclusion

This study demonstrates that high-intensity exercise training is superior to moderate-intensity training in reversing risk factors of the metabolic syndrome. The closely supervised training regimens and the comparable training volumes between the two exercise groups are strong features of our study and demonstrate the importance of exercise training in reducing the risks of metabolic syndrome. Exercise training, especially high-intensity, appears to be highly beneficial in preventing the metabolic syndrome relative to any other currently known interventions. These findings may have important implications for exercise training in rehabilitation programs. While multi-center prospective studies using exercise with high relative intensity to treat patients with the metabolic syndrome are needed to advance our conclusions, we propose that high-intensity exercise training programs may yield more favorable results than those with low-to-moderate intensities.