Nutrition, Metabolism & Cardiovascular Diseases (2007) 17, 427e435

www.elsevier.com/locate/nmcd

Development of abdominal fat and incipientmetabolic syndrome in young healthy menexposed to long-term stress

Stefan Branth a,*, Gunnar Ronquist b, Mats Stridsberg b,Leif Hambraeus a,c, Erik Kindgren a, Roger Olsson a,David Carlander b, Bengt Arnetz d

a Department of Medical Sciences, Section of Nutrition, University of Uppsala,University Hospital of Uppsala, SE-751 85 Uppsala, Swedenb Department of Medical Sciences, Clinical Chemistry, University Hospital, SE-751 85 Uppsala, Swedenc Unit for Preventive Nutrition, Department of Biosciences and Nutrition,Karolinska Institutet, SE-141 57 Huddinge, Swedend Section of Social Medicine, Department of Public Health and Caring Sciences,University of Uppsala, University Hospital of Uppsala, SE-751 85 Uppsala, Sweden

Received 1 September 2005; received in revised form 27 February 2006; accepted 14 March 2006

KEYWORDSBlood pressure;Metabolic syndrome;Obesity;Nutrition;Stress

Summary Background and aim: The sympathetic nervous system may be involvedin the pathophysiology of insulin resistance and metabolic cardiovascular syndromein young men. The aim was to study the effects of long-term stress on different fea-tures of the metabolic syndrome (MES) in formerly non-obese healthy young malesduring 5 months of defined conditions.Methods and results: Sixteen healthy male sailors (mean age 36.5 (SD) �7 years)participating in a sailing race around the world were recruited for the study. Inves-tigations were done before the start and at stopovers after finishing laps 1, 2 and 4(1, 2½ and 5 months, respectively). Anthropometric and blood pressure data as wellas biochemical data associated with MES were substantiated. Food intake and ex-ercise were chartered and largely controlled. A mean weight loss of 4.5 � 2 kg(P < 0.005), comprising both fat and lean body mass, was recorded during the firstlap. Subsequently after 5 months, a weight gain, mainly consisting of 1.2 � 1.1 kgbody fat (P < 0.05), took place, concomitantly with a protein mass drop of0.6 � 1.1 kg (P < 0.05). The body fat gain accumulated on the abdominal region.Elevated blood levels of HbA1c, insulin and the triglycerides/high-density lipoprotein

* Corresponding author. Tel.: þ46 18 611 0000; fax: þ46 18 552 562.E-mail address: stefan.branth@medsci.uu.se (S. Branth).

0939-4753/$ - see front matter ª 2006 Elsevier B.V. All rights reserved.doi:10.1016/j.numecd.2006.03.001

428 S. Branth et al.

ratio were also observed during the race. Likewise heart rate and systolic bloodpressure increased slightly but to a statistically significant extent.Conclusions: Non-obese healthy young men exposed to long-term stress developedabdominal obesity and signs of a metabolic syndrome in embryo, also emphasizedby biochemical and blood pressure alterations. It is suggested that long-term andsustained stress activation might be an additional risk factor for the developmentof MES, even after control of dietary and exercise habits.ª 2006 Elsevier B.V. All rights reserved.

Introduction

The metabolic syndrome (MES) is an importantpublic health problem worldwide, being one of themajor causes of cardiovascular disease. A cluster-ing of multiple risk factors associated to MES hasbeen recognized for cardiovascular disease andtype 2 diabetes also in young healthy non-obesesubjects [1,2]. Main risk factors of MES are upperbody obesity, insulin resistance, high levels of tri-glycerides (TG) and low levels of high-density lipo-protein (HDL) cholesterol and hypertension [3].The exact cause of MES is not known: geneticsplay a minor role [4], acquired in-utero factorsmay also play a role [5]. However, for most people,MES is primarily related to lifestyle factors, such asdietary habits and inadequate exercise causingobesity [6,7].

The sympathetic nervous system has been sug-gested to be involved in the pathophysiology of thecharacteristic insulin resistance; metabolic cardio-vascular syndrome in young men and chronic, or‘‘long-term’’ stress has recently been suggested tobe a contributing factor [8,9], and it results in neu-roendocrine changes in both the hypothalamic-pituitary-adrenal (HPA) axis and in gonads. Thesechanges can lead to deleterious metabolic effectsincluding abdominal obesity and contribute to thedevelopment of other risk factors of MES, such ashypertension [10]. There is also evidence that al-terations of the endocrine system, particularlythe HPA axis regulation by corticosteroids, playa pivotal role in the development of obesity, espe-cially the abdominal type of obesity [11]. Abdomi-nal obesity has been suggested to be decisive forthe development of MES, and insulin resistanceseems to be the pathogenic link underlying thedifferent metabolic abnormalities being part ofMES [12,13].

Although there exists an increasing amount ofconvincing animal experimental and epidemiolog-ical data, suggesting a link between chronic stressand the development of MES [13e15], there isa lack of human experimental data confirmingsuch findings. The main reason apparently is the

difficulty in finding the right experimental condi-tions and ethical approval of studies of long-termstress and its relationship to MES [16]. One majorchallenge has been to find real-life conditions elic-iting long-term stress of sufficient duration andintensity to initiate and sustain neuro-hormonalstress responses. Moreover, there has so far beenin sufficient control of important life style factorssuch as diet and exercise.

Around the world offshore sailing races lastingseveral months include a number of legs (laps) oftough sailing lasting for 1 up to over 4 weeks, andstopovers (2e3 weeks) between these sailinglegs. This sailing competition has turned out tobe an ample model for studying metabolic reac-tions in humans under long-term stressful condi-tions and the arrangement of the race impliesrepeated tests at the stopovers [17]. In addition,crew members are healthy, fairly well trainedand lean subjects who are exposed to the samestressors under defined conditions during the race.

Hence, numerous studies suggest a link betweenexposure to chronic stress and the development ofMES [8,13,15,18]. However, to the best of ourknowledge, there is a lack of prospective studiesin humans actually demonstrating such a link. Inthe present study we investigated prospectivelythe effects of an intense and long-term stress,ocean sailing race, on metabolic and neuroendo-crine functions as well as on body composition con-comitantly with monitoring relevant factors,including food intake and exercise. The aim ofthis study was therefore to examine, whetheryoung, non-obese healthy males exposed to long-term stress can undergo said alterations leadingto the development of MES.

Methods

The study was carried out during 5 months includ-ing 4 legs of sailing and 3 stopovers of the largestaround the world-offshore sailing race, the VolvoOcean Race 2001e2002, in accordance with theschedule shown in Fig. 1. Twenty-four healthy

Long-term stress and metabolic syndrome 429

Sydn

ey

Auc

klan

d

Rest 13 days

Leg 2Sailing25 days

Leg 3Sailing10 days

Leg 4Sailing23 days

Leg 1Sailing34 days

Test 4 Test 3 Test 2

Rio

de

Jane

iro

Rest 19 days Rest 21 days

Cap

e T

own

Start

Test 1

Sout

hham

pton

Figure 1 Schematic representation of the offshore sailing race protocol including the 4 different tests and restoccasions.

male professional sailors selected to join twoteams (ASSA Abloy and Djuice Dragons) were askedto join the research project. All participants werefully advised of the protocol by oral and written in-formation and gave their informed consent. Thestudy protocol was carried out in accordancewith the Declaration of Helsinki II and was ap-proved by the Ethics Committee at the UniversityHospital of Uppsala, Sweden.

Diet

During sailingDuring sailing (3 legs) the diet-plans were set toreach a mean intake per subject between 4000 and5500 kcal per day, containing 55 energy per cent(E%) carbohydrates, 15 E% protein and 30 E% fat.Food on the boat consisted mainly of lyophilizedfood, cereals and powder soups. Olive oil wasused for cooking. Dried fruits, nuts, energy barsand sport-drinks were additives. The dietary in-take during sailing was assessed by recording allpacked food on the boat before start and theremaining food on the boat after finishing thesailing leg, employing a food record computer pro-gram (DietistTM, version 1996, Stockholm, Sweden).A mean dietary intake per sailor calculated asdescribed before [17].

StopoversThree, sometimes 4 meals per day were eaten bythe teams during all stopovers. Food similar toa Mediterranean diet, including high glycemicindex foods, was served [19] in an attempt toreach the following goals: carbohydrates, 55 E%;proteins, 20 E%; fat, 25 E%; saturated fat, lessthan 10 E%. The food record computer programcalculated the menus. A mean food intake of

each meal was approximately set to an averageportion size for these sailors judged by an experi-enced dietician, and the chefs registered eachmeal eaten by the sailors. The sailors’ food andalcohol habits were checked by interviews (dietaryhistory) and possible changes were recorded.

Physical training

All subjects had been on a stable physical traininglevel before start, which was above averagecompared to their respective age group. Duringstopovers there were compulsory training pro-grams, on an average of five 60-minute sessionsper week, consisting of both strength and endur-ance-like training. The subjects displayed a stableweight during 2 months preceding the study.

ExaminationsThe tests were done 2 weeks before start and thefirst morning after finishing legs 1, 2 and 4. Allexaminations were performed in the same order,under standardized conditions after overnight fast-ing (10e12 h) and after voiding. The same experi-enced investigator assessed all subjects and usedthe same technical equipment in all tests.

Anthropometric measurementsHeight was measured to the nearest 0.5 centimetre.An electronic scale (EKS� International AB, Sweden)with an accuracy of �0.1 kg was used for bodyweight measurements. Skinfold thickness was mea-sured on 5 locations (biceps, triceps, subscapular,suprailiac and anterior abdominal, 2 cm lateralfrom umbilicus) using a Harpenden skin caliper(John Bull, British Indicators, St. Albans, England)[20]. Whole body Bioimpedance spectroscopymeasurements were performed using XITRON 4000B

430 S. Branth et al.

(Xitron Technologies Inc. San Diego, CA, USA). Bodycomposition data including the proportion of water,body fat, lean body mass (LBM) and protein masswere determined applying the three-compartmentmethod based on skin fold and bioelectric imped-ance- measurements and calculations [17,21].

Blood pressure and heart rateTwo blood pressure and heart rate measurementswere undertaken 2 min apart, using a standardsphygmomanometer (MMM�) on the left arm tothe nearest 2 mmHg with the participants in supineposition after 10 min rest. Due to practical rea-sons, blood pressure and heart rate were onlymeasured at tests 1 and 4.

Blood sampling and analysesAll blood samples were collected in the morning ina fasting state after 15 min rest in supine positionfrom a cubital vein using Vacutainer� tubes con-taining EDTA (for hematology and HbA1c) and hep-arin. For some analyses (insulin, c-peptide andcortisol) tubes without additives were used. Alllaboratory analyses were carried out at a certifiedlaboratory following standardized routines. Serumconcentrations of insulin, C-peptide, and cortisolwere measured by an automated system for immu-nological analyses (Auto-Delfia, Wallac OY, Turku,Finland). Plasma concentration of ACTH was mea-sured using a commercial RIA-kit (CIS bio int. OrisGroup, France).

StatisticsStudy variables were first reviewed for means anddispersion measures. Differences between twocontrasting sampling times were analyzed byStudent’s t-test for paired samples with regard tovariables with normal distributions and with non-parametric methods (Wilcoxon signed rank test)for all other variables. Changes over the entirestudy period were assessed using one-way analysisof variance (one-way ANOVA using Greenhouse-Geisser corrections). When warranted, post-hocadjustment, using Tukey’s procedures, was done.Statistical significance was set at P < 0.05, two-sided.

Results

Participants

Eight subjects dropped out from the initial groupof 24; the mean age of the remaining 16 sailors was36.5 � 7 (SD) years. Two sailors had to rest due to

injury and acute illness. Another two sailors didnot want to continue the blood sampling and theremaining 4 were exchanged due to competitivereasons. The dropouts did not differ regardingphysiological baseline data and age.

Dietary intake

During sailingThe mean energy intake was 4700 kcal per day.Except for the intake of saturated fat (15 E%), thedietary composition during sailing agreed wellwith the dietary goals (Table 1). Daily fiber contentwas 30 g due to a rather high intake of whole graincereals. During sailing a multi- vitamin-mineralsupplement was taken daily. According to thefood records, the dietary intake of essential min-erals and vitamins covered recommended dailyallowances. There was no alcohol intake duringsailing, during at least 2 days before start and be-fore tests.

StopoversMost of the food consumed was eaten jointly andestimated to an energy intake of 3100 kcal per dayand the consumption pattern reached our objec-tives of a Mediterranean-like diet rich in wholegrains, fruit and vegetables. The documentedfood composition was estimated to 55 E% caloriesfrom carbohydrates, 19 E% calories from proteinand 26 E% calories from fat, whereas saturatedfat contributed with 9 E%, monounsaturated fatwith 13 E% and polyunsaturated fat with 4 E% dis-playing a polyunsaturated to saturated fatty acid(P/S) ratio of 0.4, containing 240 mg cholesterol

Table 1 Total mean dietary intake per day andsubject during sailing

Nutrient Sailing

Mean intakeper subjectand day

% of energy

Energy (kcal/day) 4700Carbohydrates (g/day) 650 57Protein (g/day) 142 13Fat (g/day) 160 30Saturated fat (g/day) 78 15Monounsaturated fat (g) 78 15Polyunsaturated fat (g) 24 5Cholesterol (mg) 120P/S ratio 0.31Omega 6/3 ratio 6Fiber (g) 30

P/S, polyunsaturated- over saturated fatty acids. Omega6/3, omega 3- over omega 6 polyunsaturated fatty acids.

Long-term stress and metabolic syndrome 431

and 38 g fiber. Alcohol intake was low to moderate(equivalent to 10e30 g ethanol per day) consistingmainly of beer and to some extent wine beverage.

Physical training

High physical demands during the races due toheavy weather etc., were a reality. The crews hadjointly compulsory physical exercise sessions(around 60 min) 5 times a week, as a mean, duringthe stopovers. Additionally, there was periodicsailing training during stopovers.

Anthropometric findingsTable 2 shows the dynamics of anthropometricdata throughout the 5 month study. Only 15 sub-jects are included in specified body compositiondata due to a technical error. The subjects lostan average of 4.5 � 2 kg body weight (P < 0.001,Student’s t-test) during the first leg with no subse-quent body weight losses. Instead, the sailorsregained 2.1 � 2.3 kg body weight (P < 0.005,Student’s t-test) during the remaining time of thestudy (tests 2 to 4). One-way ANOVA, using allfour sampling periods, confirmed the time-relatedchanges in body weight (P for weight over time,Greenhouse-Geisser, P < 0.001) and the net bodyweight loss from start was accordingly2.4 � 2.8 kg (P < 0.005) after 4 legs. The regainedweight consisted mainly of body fat mass,1.2 � 1.1 kg (Student’s t-test, P < 0.05; one-wayANOVA over time; P < 0.01), with a concomitantloss of protein mass, 0.6 � 1.1 kg (Student’st-test, P < 0.05; one-way ANOVA over time;P < 0.01). Other alterations of body weight wererelated to an increase in TBW by 1.7 � 3.3 L(non-significant). The fat mass gain was confirmed

by the skin fold measurements, revealing a mean in-crease of 6.8 � 5.5 cm (Student’s t-test, P < 0.001;one-way ANOVA over time, P < 0.001). Most of thisincrease was due to an abdominal increment, com-prising a specific central anterior abdominal pointincrease of 4.6 � 3.1 cm during the 5 months-testperiod (Student’s t-test, P < 0.001; one-way ANOVAover time, P < 0.001), contrary to the generalbody fat mass reduction due to the first leg’s all-embracing loss of weight.

Haematological and biochemical variablesThe mean haemoglobin value increased slightlybut significantly from test 1 before the start to test4 after leg 4 from 145.5 � 8 to 149.8 � 8 g� L�1

(P < 0.01), while the haematocrit values werenot significantly augmented from test 1 to test 4,44.3 � 2 and 45.0 � 2%, respectively. A slight butsignificant leukocyte count decrease was notedfrom 5.8 � 1 to 5.5 � 1 � 109 cells � L�1

(P < 0.01) at tests 1 and 4, respectively.Important MES-associated blood, plasma and

serum variables of the sailors are presented inTable 3. A slight but significant increase was notedfor blood HbA1c levels (one-way ANOVA over time,P < 0.005), which appeared already after the firstleg despite the pronounced body weight loss.Plasma glucose was however unchanged. PlasmaLDL, HDL and total cholesterol concentrationsfell during leg 1, but returned to the initial levelsafter leg 4. A mean, statistically significant in-crease in TG concentration and TG/HDL ratiowere noted after leg 1 and remained significantlyelevated after leg 4 (one-way ANOVA over time,P < 0.05 respectively P < 0.001). Both serum insu-lin and C-peptide levels increased significantlyduring leg 1 and remained elevated throughout

Table 2 Anthropometric data of sailors collected on 4 different test occasions

Test 1 Test 2 Test 3 Test 4

Before start After leg 1 After leg 2 After leg 4

n P-value P-value P-value

Body weight (kg) 16 86.1 � 8.5 81.6 � 7.0 <0.001 82.3 � 7.4 <0.001 83.7 � 8.9 <0.001BMI (kg m�2) 16 26.1 � 2.3 24.7 � 1.8 <0.001 24.9 � 2.0 <0.001 25.4 � 2.2 <0.001LBM (kg) 15 73.1 � 6.8 71.8 � 6.2 ns 71.5 � 6.0 <0.05 73.0 � 6.7 nsProtein mass (kg) 15 18.3 � 1.7 17.2 � 1.6 <0.001 17.4 � 1.3 <0.001 16.6 � 1.7 <0.001TBW (L) 15 55.0 � 5.7 54.9 � 5.5 ns 54.6 � 5.2 ns 56.7 � 5.5 nsBody fat (%) 15 15.9 � 5.4 12.7 � 5.8 <0.001 14.0 � 4.2 <0.05 13.9 � 4.1 <0.05Fat mass (kg) 15 14.0 � 5.3 10.6 � 5.2 <0.001 11.7 � 3.9 <0.01 11.8 � 3.9 <0.01

Skinfold (cm)Sum (5 points) 16 38.9 � 9.0 34.3 � 7.0 <0.001 36.1 � 7.6 <0.01 41.1 � 9.8 <0.05Anterior abdominal 16 11.7 � 2.9 10.7 � 3.1 ns 13.4 � 4.1 <0.05 16.3 � 5.0 <0.001

Data are means � SD. P-value, two sided, Student’s t-test for paired samples comparing baseline (test 1) and tests 2, 3 and 4.BMI, body mass index. LBM, lean body mass. TBW, total body water.

432 S. Branth et al.

Table 3 Biochemical and endocrinological variables of sailors collected on 4 different test occasions

Test 1 Test 2 Test 3 Test 4

(before start) (after leg 1) (after leg 2) (after leg 4)

n P-value P-value P-value

Biochemical dataB-HbA1c (%) 16 4.26 � 0.2 4.37 � 0.2 <0.005 4.43 � 0.2 <0.005 4.40 � 0.2 <0.001P-glucose (mmol L�1) 16 5.6 � 0.3 5.5 � 0.8 ns 5.5 � 0.6 ns 6.1 � 1.2 nsP-triglycerides (mmol L�1) 16 1.03 � 0.5 1.48 � 0.7 <0.05 1.21 � 0.6 ns 1.41 � 0.9 <0.05P-HDL (mmol L�1) 16 1.53 � 0.2 1.29 � 0.2 <0.001 1.62 � 0.3 <0.01 1.48 � 0.2 nsP-TG/HDL-ratio 16 0.71 � 0.5 1.17 � 0.6 <0.005 0.77 � 0.4 ns 1.0 � 0.7 <0.05P-cholesterol (mmol L�1) 16 4.59 � 0.9 3.82 � 0.6 <0.001 4.51 � 0.6 ns 4.55 � 0.6 nsP-LDL (mmol L�1) 16 2.67 � 0.7 1.94 � 0.7 <0.001 2.37 � 0.6 <0.05 2.48 � 0.6 <0.05

Endocrinological dataS-insulin (mU L�1) 16 4.8 � 3.4 13.0 � 7.6 <0.001 8.9 � 6.9 <0.05 19.0 � 11.8 <0.001S-C-Peptide (nmol L�1) 16 0.50 � 0.2 0.92 � 0.2 <0.001 0.75 � 0.3 <0.05 1.08 � 0.5 <0.001S-cortisol (nmol L�1) 16 377.3 � 111 317.0 � 125 ns 395.8 � 112 ns 333.9 � 109 nsP-ACTH (ng L�1) 16 35.6 � 22.6 16.5 � 5.4 ns 34.7 � 15.7 ns 22.8 � 10.0 <0.05

Data are means � SD. P-value, two sided, Student’s t-test for paired samples comparing baseline (test 1) and tests 2, 3 and 4.B, whole blood. P, plasma. S, serum. HDL, high density lipoprotein. TG, triglycerides. LDL, low density lipoprotein.

the race (one-way ANOVA over time, P < 0.01).Serum cortisol concentrations displayed a ratherinconsistent profile and no statistical change overtime was recorded, although a significant dip wasnoted between tests 3 and 4 (P < 0.05, Student’st-test). A significant reduction was apparent inthe plasma ACTH concentration from the start tothe end of the study after leg 4 (one-way ANOVAover time, P < 0.005).

Blood pressure and heart rateSystolic blood pressure increased significantly fromleg 1 to leg 4, from 120.1 � 8 to 124.6 � 9(P < 0.05, Student’s t-test). Resting heart rate in-creased from 54 � 6 to 63 � 10 beats � min�1

(P < 0.05, Student’s t-test) from the start to theend of the study after 4 legs.

Discussion

The current study is to our knowledge the firstprospective study in real-life of the possible role oflong-term stress of significant intensity in theaetiology of the metabolic syndrome-MES. Long-term stress elicits MES in animals [14] and has alsobeen implied to cause MES in epidemiological stud-ies of humans [11,13,15]. We herewith show signif-icant alterations in principal MES variables asserum insulin, blood HbA1c, abdominal obesityand blood pressure. Prior to discussing the actualfindings it is important to consider some of theweaknesses of the study.

There are many pitfalls in a field study like this.All test procedures outside strictly laboratoryconditions will be burdened by standardisationproblems. However, strong efforts were made toachieve as far as possible standardised conditionsduring each test situation. Disturbances in circa-dian rhythm and sleeping pattern affects a numberof the variables studied [22]. So are the various en-ergy intakes, especially marked during the first legwith weight losses similar to short term ‘‘beneficialdieting’’ [23]. However, these limitations couldalso be considered as this study’s strength, sinceit resembled very well a ‘‘normal living’’ in thegrowing risk population that this study aimed to in-vestigate, fairly young men exposed to chronic,i.e., long-term stress, bad sleeping habits, variabledietary intakes as well as short periods of unsuc-cessfully continued dieting and training periodswith oscillating weight as an unfavourable result.

Dietary factors are linked to individual meta-bolic features of MES [24] and therefore importantto consider in this context. Except intake of satu-rated and polyunsaturated fatty acids which weresomewhat deviating from recommendations duringsailing, the diet was well in agreement with knownrecommendations preventing MES [19,24]. For nat-ural reasons the intake of fresh vegetables andfruits was restricted during sailing but compen-sated for during stopovers. Also, intake of wholegrains and fiber was high both during sailing andstopovers. It is interesting in this context to notethat a high intake of whole grains and fiber hasbeen suggested to be a protective factor against

Long-term stress and metabolic syndrome 433

MES [25]. In this study it was not possible to per-form any detailed control of the dietary intakeduring stopovers (2 of 5 months of the study),which is a limitation. Moreover, the sailors’ usualfood habits, judged as being in accordance withNorth European life style, before joining the racewere not possible to control in any detail. How-ever, a switch occurred during the race to a Medi-terranean type of diet, meaning that their foodintake approached known recommendations thatare claimed to prevent development of a metabolicsyndrome [19,24].

Mild to moderate alcohol consumption, espe-cially of beer and wine, has been suggested to beassociated with a lower prevalence of the meta-bolic syndrome and with a favourable influence onserum lipids, waist circumference, and fastingserum insulin in some studies [26]. Since therewas no alcohol intake during sailing and a moderateconsumption during stopovers the overall alcoholconsumption decreased during the study periodand could be considered mild. Therefore, it is rea-sonable to believe that this small change of alcoholhabits was not influential on our results in a deci-sive way. We also noted stable and normally lowvalues for serum ALAT, ASAT and g-GT throughoutthe race in favour of no or minimal impact of alco-hol on the liver.

Exercise training also in the absence of bodyweight reduction or abdominal fat loss is associ-ated with improvements of peripheral insulinsensitivity and markers of glucose and fat metab-olism [27]. This is in harmony with the high degreeof physical activity concomitant with a high energyexpenditure typical of this kind of sailing race [17].Furthermore, stopovers included a physical exer-cise program on a regular basis meaning that theexercise load was reasonably high throughout thestudy. Accordingly, it is justified to conclude thatneither diet and insufficient physical activity noralcohol could be the underlying cause of our find-ings. Rather, the continuous stress factor standsout as the main cause of MES development.

Despite the fact that it is very difficult tomonitor and quantify stress [28], it was quite obvi-ous that the crew members in this study were ex-posed to a homogenous and continuous multitudeof demands (stressors) during the period theywere studied, both during the legs and on shore.During sailing there is a competitive pressure ofmaking rapid decisions, the watch system at seadisturbs sleep and circadian rhythm, the weatherconditions with repeated periods of markedlyheavy sea and sometimes even risk of hitting ice-bergs cause different degrees of fear. In addition,the sailors have to share cramped quarters during

all stressful situations causing typical social prob-lems [17]. The mental stress factors during thelegs more or less continue on land due to very tightschedules preparing the boats for the next leg andby different social demands like pressure fromsponsors and attempts to compensate families,etc. Although stress itself could not be quantifiedand our data just represented indirect evidenceof continuously high stress level during the study,we can state, by definition, that all of the subjectsdefinitely have been exposed to more or less con-tinuous stress during the study which very well ful-fils the two principal criteria of stress, loss ofcontrol over one’s environment and alteration inlifestyle patterns. These two features have beenregarded to be of importance in the developmentof stress-induced MES, comprising abdominal obe-sity [28e31].

Our study demonstrates development of con-spicuous abdominal subcutaneous obesity, hyper-insulinemia and a slight but significant elevation ofHbA1c indicating a growing peripheral insulin re-sistance. Increased serum triglyceride concentra-tions were other signs of a metabolic syndrome inembryo. The elevated TG/HDL ratio is an impor-tant risk factor for cardiovascular disease [32]. Ad-ditionally, small but significantly increased meanvalues of the systolic blood pressure and heartrate were noted, possibly via an activation of thesympathetic nervous system [33]. Elevated cortisollevels are believed to be the main contributor tothe development of abdominal obesity and the en-tire spectrum of the MES [9,11,34]. It should benoted though, that no increased morning serumcortisol levels were observed in the present study,neither were the morning plasma concentrationsof ACTH increased. However, morning values ofserum cortisol and plasma ACTH are probably notenough to give a close-fitted and correct pictureof the levels over day and night of these two hor-mones, e.g. studies rather indicate that higherlevels of evening cortisol are the best sign ofstress-disturbed HPA axis [34,35]. Moreover, de-spite overall higher basal cortisol levels in obesesubjects, their plasma morning cortisol concentra-tions are lower than in non-obese subjects [34].Unfortunately, it was impossible, due to practicalreasons, to get evening cortisol samples separatelyand/or to administer salivary cortisol collectionsthroughout the day as would have been the pre-ferred mode of action.

There was a net body weight loss includingmainly fat mass during the first leg, althoughthere were signs of insulin resistance, as indicatedby higher levels of insulin, c-peptide and HbA1c.Most probably, the reason for the body mass loss

434 S. Branth et al.

during the first leg was the difficulties of thesailors to follow the dietary advice. During thesubsequent legs the sailors followed their individ-ual food programs better. Consequently, no majorbody weight losses occurred during the subsequentpart of the race but the sailors actually regaineda substantial part of their body weight. However,this body weight recovery consisted, surprisingly,mostly of an accumulation of body fat whilea continuous loss of protein mass proceeded.

Weight loss is well known to counteract not onlyinsulin resistance but also the metabolic risk factorsassociated with the MES [36,37]. However, despitethe pronounced weight loss during the first leg,both serum insulin and C-peptide concentrationsincreased concomitantly with an augmentation ofthe HbA1c and TG-levels. These seemingly contra-dictory results could most likely be explained bythe stressful conditions exerting adverse effectson glucose homeostasis as previously suggested[28,38].

The marked increase in the subcutaneous skin-fold at the anterior abdominal point and the fatmass increase during the period after leg 1, wereunexpected findings, since these subjects werehighly physically active with beneficial effects onabdominal adiposity [39], and they had mostlyeaten according to best known dietary advices[6,19,24]. However, it could be well in agreementwith the theory by Bjorntorp [9e12], that re-peated continuous stress could cause abdominalobesity even in previously lean young men. Thecontinued loss of protein mass, most likely fromthe muscles, corroborated this theory. Centralfat rather than peripheral fat is certainly associ-ated with insulin resistance [40], and visceral fatis more related to insulin resistance than subcuta-neous abdominal fat in obese individuals [41].However, also subcutaneous abdominal adiposityhas been associated with MES related distur-bances, especially valid in young lean subjects[42e44].

We conclude that subjects under continuousstress develop metabolic alterations including dis-tinctive abdominal obesity, biochemical changesand a slight blood pressure elevation towards theMES despite regular exercise and a proper diet. Wedo not believe that the kind of mental stress thesesailors were exposed to was uniquely extreme.There are, for instance, highly charged type Apersonalities and/or individuals under high jobdemands, financial strain or psychosocial pressure,who typically live under the same burden ofstressors [28]. Since exercise and adequate dietdid not counteract the development of these signsof a metabolic syndrome in embryo, there might be

a need to consider the relevance of expanding pre-ventive efforts to also include stress managementin addition to exercise and nutritional advice.

Acknowledgements

We are grateful to the sailors of the Volvo OceanRace teams ASSA Abloy and Djuice Dragons fortheir kind willingness to join the study and theircooperation and patience during the investigation.We also want to thank both teams’ on shoreorganisations for all practical help, and the Atlan-tic racing team, Stockholm, Sweden and DjuiceDragons A/S for economic support for travel costs.Special thanks to Maja Storeide MD, Norway, forher positive and professional assistance through-out the investigations. Dr Lara Keytel, SportsScience Institute of South Africa, Cape Town,South Africa and Dr Paul Fitzgerald, Sydney,Australia, for their kind hospitality at local labo-ratories and their skilled laboratory help andmedical assistance. This study was supported bygrants from the University Hospital of Uppsala.

References

[1] Reaven GM. Banting Lecture 1988: role of insulin resistancein human disease. Diabetes 1988;37:1595e607.

[2] Alberti KG, Zimmet PZ. Definition, diagnosis and classifica-tion of diabetes mellitus and its complications. Part 1.Diagnosis and classification of diabetes mellitus, provi-sional report of a WHO consultation. Diabet Med 1998;15:539e53.

[3] Obesity: preventing and managing the global epidemic. Re-port of a WHO consultation. World Health Organ TechnicalReport Series 892: iexii. 2000;1e253.

[4] Arya R, Blangero J, Williams K, Almasy L, Dyer TD,Leach RJ, et al. Factors of insulin resistance syndrome-related phenotypes are linked to genetic locations onchromosomes 6 and 7 in non-diabetic Mexican-Americans.Diabetes 2002;51:841e7.

[5] Jaquet D, Leger J, Czernichow P, Levy-Marchal C. Theeffect of in-utero undernutrition on the insulin resistancesyndrome. Curr Diab Rep 2002;2:77e82.

[6] Riccardi G, Rivellese AA. Dietary treatment of the meta-bolic syndromedthe optimal diet. Br J Nutr 2000;83:S143e8.

[7] Mayer-Davis EJ, D’Agostino Jr R, Karter AJ, Haffner SM,Rewers MJ, Saad M, et al. Intensity and amount of physicalactivity in relation to insulin sensitivity: the Insulin Resis-tance Atherosclerosis Study. JAMA 1998;279:669e74.

[8] Raikkonen K, Keltikangas-Jarvinen L, Adlercreutz H,Hautanen A. Psychosocial stress and the insulin resistancesyndrome. Metabolism 1996;45:1533e8.

[9] Rosmond R, Bjorntorp P. The hypothalamic-pituitary-adre-nal axis activity as a predictor of cardiovascular disease,type 2 diabetes and stroke. J Internat Med 2000;247:447e54.

[10] Dallman MF, Pecoraro N, Akana SF, La Fleur SE, Gomez F,Houshyar H, et al. Chronic stress and obesity: a new view

Long-term stress and metabolic syndrome 435

of ‘‘comfort food’’. Proc Natl Acad Sci USA 2003;100:11696e701.

[11] Rosmond R, Dallman MF, Bjorntorp P. Stress-related corti-sol secretion in men: relationships with abdominal obesityand endocrine, metabolic and hemodynamic abnormali-ties. J Clin Endocrinol Metab 1998;83:1853e9.

[12] Bjorntorp P. Metabolic implacations of body fat distribu-tion. Diabet Care 1991;14:1132e43.

[13] Brunner EJ, Marmot MG, Nanchahal K, Shipley MJ,Stansfeld SA, Juneja M, et al. Social inequality in coronaryrisk: central obesity and the metabolic syndrome.Evidence from the Whitehall II study. Diabetologia 1997;40:1341e9.

[14] Jayo JM, Shively CA, Kaplan JR, Manuck SB. Effects of ex-ercise and stress on body fat distribution in male cynomol-gus monkeys. Int J Obes Relat Metab Disord 1993;17:597e604.

[15] Wamala SP, Lynch J, Horsten M, Mittlemann MA,Schenck-Gustafsson K, Orth-Gomer K. Education andthe metabolic syndrome in women. Diabet Care 1999;22:1999e2003.

[16] Chrousos GP, Gold PW. A healthy body in a healthyminddand vice versadthe damaging power of ‘‘uncon-trollable’’ stress. J Clin Endocrinol Metab 1998;83:1842e5.

[17] Branth S, Hambraeus L, Westerterp K, Andersson A,Edsgren R, Mustelin M, et al. Energy turnover in a sailingcrew during offshore racing around the world. Med SciSports Exerc 1996;28:1272e6.

[18] Seematter G, Binnert C, Martin JL, Tappy L. Relationshipbetween stress, inflammation and metabolism. Curr OpinClin Nutr Metab Care 2004;7:169e73.

[19] de Lorgeril M, Salen P, Martin JL, Mamelle N, Monjaud I,Touboul P, et al. Effect of a Mediterranean-type of dieton the rate of cardiovascular complications in coronary pa-tients: insights into the cardioprotective effect of certainnutrients. J Am Coll Cardiol 1996;28:1103e8.

[20] Lohman TG, Roche AF, Martorell R. Anthropometric Stan-dardization Reference Manual. IL: Human Kinetics Cham-pagne; 1988.

[21] Siconolfi SF, Gretebeck RJ, Wong WW. Assessing total bodyprotein, mineral, and bone mineral content from totalbody water and body density. J Appl Physiol 1995;79:1837e43.

[22] Vgontzas AN, Bixler EO, Chrousos GP. Metabolic distur-bances in obesity versus sleep apnoea: the importance ofvisceral obesity and insulin resistance. J Intern Med 2003;254:32e44.

[23] Field AE, Colditz GA. Frequent dieting and the develop-ment of obesity among children and adolescents. Nutrition2001;17:355e6.

[24] Hung T, Sievenpiper LJ, Marchie A, Kendall CWC,Jenkins DJA. Fat versus carbohydrate in insulin resistance,obesity, diabetes and cardiovascular disease. Curr OpinClin Nutr Metab Care 2003;6:165e76.

[25] Fung TT, Hu FB, Pereira MA, Liu S, Stampfer MJ,Colditz GA, et al. Whole-grain intake and the risk of type2 diabetes: a prospective study in men. Am J Clin Nutr2002;76:535e40.

[26] Temelkova-Kurktschiev T, Henkel E, Koehler C, Hanefeld M.Light-to-moderate alcohol consumption, risk factor profileand early atherosclerosis: the RIAD study. Atherosclerosis2001;156:239e40.

[27] Duncan GE, Perri MG, Theriaque DW, Hutson AD, Eckel RH,Stacpoole PW. Exercise training, without weight loss, in-creases insulin sensitivity and postheparin plasma lipaseactivity in previously sedentary adults. Diabet Care 2003;26:557e62.

[28] Steptoe A, Marmot M. Burden of psychosocial adversity andvulnerability in middle age: association with biobehavioralrisk factors and quality of life. Psychosom Med 2003;65:1029e37.

[29] Ravaja N, Keltikangas-Jarvinen L, Viikari J. Life changes,locus of control and metabolic syndrome precursors in ad-olescents and young adults: a three-year follow-up. Soc SciMed 1996;43:51e61.

[30] Marniemi J, Kronholm E, Aunola S, Toikka T, Mattlar CE,Koskenvuo M, et al. Visceral fat and psychosocial stressin identical twins discordant for obesity. J Intern Med2002;251:35e43.

[31] Feldman PJ, Steptoe A. Psychosocial and socioeconomicfactors associated with glycated hemoglobin in non-dia-betic middle-aged men and women. Health Psychol 2003;22:398e405.

[32] Maruyama C, Imamura K, Teramoto T. Assessment of LDLparticle size by triglyceride/HDL-cholesterol ratio in non-diabetic, healthy subjects without prominent hyperlipid-emia. J Atheroscler Thromb 2003;10:186e91.

[33] Palatini P. Sympathetic overactivity in hypertension: a riskfactor for cardiovascular disease. Curr Hypertens Rep2001;3:S3e9.

[34] Walker BR, Soderberg S, Lindahl B, Olsson T. Independenteffects of obesity and cortisol in predicting cardiovascularrisk factors in men and women. J Intern Med 2000;247:198e204.

[35] Plat L, Leproult R, L’Hermite-Baleriaux M, Fery F,Mockel J, Polonsky KS, et al. Metabolic effects of short-term elevations of plasma cortisol are more pronouncedin the evening than in the morning. J Clin Endocrinol Metab1999;84:3082e92.

[36] Wing RR, Vendetti E, Jakicic JM, Polley BA, Lang W. Life-style intervention in overweight individuals with a familyhistory of diabetes. Diabet Care 1998;21:350e9.

[37] Laaksonen DE, Kainulainen S, Rissanen A, Niskanen L. Rela-tionships between changes in abdominal fat distributionand insulin sensitivity during a very low calorie diet in ab-dominally obese men and women. Nutr Metab CardiovascDis 2003;13:349e56.

[38] Arnetz B. The potential role of psychosocial stress onlevels of Hemoglobin A1c (HbA1c) and fasting plasma glu-cose in elderly people. J Gerontol 1984;39:424e9.

[39] Wong SL, Katzmarzyk P, Nichaman MZ, Church TS, Blair SN,Ross R. Cardiorespiratory fitness is associated with lowerabdominal fat independent of body mass index. Med SciSport Exerc 2004;36:286e91.

[40] Wagenknecht LE, Langefeld CD, Scherzinger AL, Norris JM,Haffner SM, Saad MF, et al. Insulin sensitivity, insulin secre-tion, and abdominal fat: the Insulin Resistance Atheroscle-rosis Study (IRAS) Family Study. Diabetes 2003;52:2490e6.

[41] Bonora E. Relationship between regional fat distributionand insulin resistance. Int J Obes Relat Metab Disord2000;24:S32e5.

[42] Weyer C, Foley JE, Bogardus C, Tataranni PA, Pratley RE,Weyer C, et al. Enlarged subcutaneous abdominal adipocytesize, but not obesity itself, predicts type II diabetes indepen-dent of insulin resistance. Diabetologia 2000;43:1498e506.

[43] You T, Murphy KM, Lyles MF, Demons JL, Lenchik L,Nicklas BJ. Addition of aerobic exercise to dietary weightloss preferentially reduces abdominal adipocyte size. IntJ Obes 2006:1e6 [Lond].

[44] Tai ES, Lau TN, Ho SC, Fok AC, Tan CE. Body fatdistribution and cardiovascular risk in normal weightwomen. Associations with insulin resistance, lipids andplasma leptin. Int J Obes Relat Metab Disord 2000;24:751e7.