MP

CCJ

*BS

BaanAgsmeuscaaspLpwtsr4a6.gtcjadiI

Ia

GASTROENTEROLOGY 2008;135:163–172

aladaptive Intestinal Epithelial Responses to Life Stress Mayredispose Healthy Women to Gut Mucosal Inflammation

ARMEN ALONSO,* MAR GUILARTE,*,‡ MARIA VICARIO,* LAURA RAMOS,* ZIAD RAMADAN,§ MARIA ANTOLÍN,*RISTINA MARTÍNEZ,* SERGE REZZI,§ ESTEBAN SAPERAS,* SUNIL KOCHHAR,§ JAVIER SANTOS,* andUAN RAMÓN MALAGELADA*

Digestive Diseases Research Unit, Institut de Reçerca, Department of Gastroenterology; ‡Allergy Unit, Hospital Universitari Vall d’Hebron, Universitat Autònoma dearcelona, Department of Medicine, Barcelona, Spain; and §Bioanalytical Science Department, Metabonomics & Biomarkers, Nestlé Research Center, Lausanne,

witzerland

rqmc

ybwsnemobopmntrpvis

boprdtlos

mvtMq

BA

SIC–

ALI

MEN

TARY

TRA

CT

See Shepherd SJ et al on page 765 andCamilleri M et al on page 772 in CGH.

ackground & Aims: Irritable bowel syndrome (IBS),highly prevalent disorder among women, has been

ssociated with life stress, but the peripheral mecha-isms involved remain largely unexplored. Methods:20-cm jejunal segment perfusion was performed in 2

roups of young healthy women, equilibrated by men-trual phase, experiencing either low (LS; n � 13) or

oderate background stress (MS; n � 11). Intestinalffluents were collected every 15 minutes, for 30 minutesnder basal conditions, and for 1 hour after cold paintress. Cardiovascular and psychological response,hanges in circulating stress and gonadal hormones,nd epithelial function (net water flux, albumin outputnd luminal release of tryptase and �-defensins) to coldtress were determined. Results: Cold pain induced asychological response stronger in the MS than in theS group, but similar increases in heart rate, bloodressure, adrenocorticotrophic hormone, and cortisol,hereas estradiol and progesterone remained unal-

ered. Notably, the jejunal epithelium of MS femaleshowed a chloride-related decrease in peak secretoryesponse (�[15-0 minutes]: LS, 97.5 [68.4–135.0]; MS,8.8 [36.6–65.0] �L/min/cm; P < .001) combined withmarked enhancement of albumin permeability (LSAUC,.35 [0.9–9.6]; MSAUC, 13.97 [8.3–23.1] mg/60 min; P �

008) after cold stress. Epithelial response in bothroups was associated with similar increases in luminalryptase and �-defensins release. Conclusions: In-reased exposure to life events determines a defectiveejunal epithelial response to incoming stimuli. Thisbnormal response may represent an initial step in theevelopment of prolonged mucosal dysfunction, a find-

ng that could be linked to enhanced susceptibility forBS.

rritable bowel syndrome (IBS) is a highly prevalentdisorder characterized by recurrent abdominal pain

nd altered defecation that is associated with a marked

eduction of quality of life.1 Although IBS remains as auite heterogeneous symptom-based disorder it shows, asany other chronic disorders,2 a clear but mechanisti-

ally unraveled female predominance (2–3:1).3

IBS origin and development are unclear and complex,et genetic, neurobiological, and psychosocial differencesetween sexes3 along with increased susceptibility amongomen to factors involved in IBS might help to under-

tand that predominance. In this sense, gastrointesti-al4 – 6 and extraintestinal infections7 represent the high-st risk factors for development of IBS and women areore likely to initiate IBS-like symptoms after an episode

f infectious gastroenteritis.8 Likewise, life stress haseen claimed as a risk factor associated with symptomnset and severity in postinfective9 and other diarrhea-redominant IBS groups.10 Although this associationay be attributable to differences in psychological and

euroendocrine reactivity to stress and increased sensi-ivity among women to life events,11,12 the ultimateeason remains unknown. Recent findings showing theresence of microscopic inflammation and immune acti-ation in the intestinal mucosa of IBS13,14 have renewednterest in linking, from a pathophysiologic perspective,tress, infections, and female predisposition.

One such potential link is the intestinal epithelialarrier which is faced with critical functions in regulationf mucosal homeostasis15,16: On one side, the barrier isoised at maintaining physical integrity to prevent dis-uptive host–microbial interactions. On the other side, itevelops an active immunologic surveillance and execu-ive program evolved to process enormous amounts ofuminal antigens. This program includes basic weepingff functions, such as water secretion, as well as moreophisticated tools, such as release of chemical mediators

Abbreviations used in this paper: ACTH, adrenocorticotrophic hor-one; AUC, area under the curve; Cl�, chloride; CV, coefficient of

ariation; ELISA, enzyme-linked immunosorbent assay; HPA, hypo-halamic–pituitary axis; IBS, irritable bowel syndrome; LS, low stress;S, moderate stress; PEG, polyethylene glycol; Q1, Q3, first and thirduartiles; Na�, sodium; SSRS, Subjective Stress Rating Scale.

© 2008 by the AGA Institute0016-5085/08/$34.00

doi:10.1053/j.gastro.2008.03.036

os

aeamlthitdboatip

ptscem

twde

pwfa(aa

ta2ocn(spip

smtpdalatbrdflavCdpcw

1pmtmN

tHylfw

bcpneepLoal

BA

SIC–

ALIM

ENTA

RY

TRA

CT

164 ALONSO ET AL GASTROENTEROLOGY Vol. 135, No. 1

r natural antibiotics like defensins to fight unwantedtimuli17 and keep epithelial permeability tight.

Growing experimental evidence reveals that both stressnd bacteria contribute to mucosal barrier function andpithelial integrity. Acute and chronic stress increase ionnd water secretion and intestinal permeability in animalodels18,19 and humans.20 Mast cells have been estab-

ished as key regulators of epithelial function, involved inhese responses to stress.21 Moreover, laboratory stressas also been shown to initiate and reactivate mucosal

nflammation and to disturb gut epithelial function22;hese changes have been also observed in diarrhea-pre-ominant IBS patients.23 These disturbances of intestinalarrier function might facilitate an excessive penetrationf food and bacterial antigens across the epithelial layernd result in inappropriate immune stimulation, leadingo chronic intestinal inflammation.24 However, no stud-es have assessed the ability of chronic life stress toromote intestinal barrier dysfunction in humans per se.Stress dysregulation of gut epithelial barrier could

rove determinant in understanding the association be-ween life stress and IBS epidemiology. Therefore, in thistudy we aimed to investigate whether and how life stressould shape intestinal epithelial function to account fornhanced risk to develop IBS and the participation ofast cells.

MethodsParticipantsFertile female healthy volunteers were prospec-

ively recruited by public advertising. The study protocolas approved by the Ethics Committee at Hospital Vall’Hebron. Written informed consent was obtained fromach participant.

Candidates gave a full medical history and underwenthysical examination. Previous gastrointestinal disordersere excluded by anamnesis and Rome II criteria for

unctional dyspepsia25 and IBS.26 Food and respiratoryllergy was ruled out using a battery of skin prick testsLeti SA, Barcelona, Spain) for 32 common foodstuffsnd 24 inhalants with histamine and saline as positivend negative controls, respectively.

Participants with a history of acute gastroenteritis inhe last 2 years, active smokers, and those having takenny drug, pharmaceutical compound, or herb in the lastweeks were excluded. No diet restrictions, other than

vernight fasting, alcohol, or caffeine intake, were indi-ated. Other exclusion criteria were age �18 years, ab-ormal baseline levels of adrenocorticotrophic hormone

ACTH) and/or cortisol, amenorrhea, abnormal men-trual cycle, pregnancy, hormonal contraceptive use,resent infection, psychiatric or organic diseases, inabil-

ty to give informed consent, and violation of the study

rotocol. p

Jejunal Perfusion MethodA modified double-lumen closed-segment perfu-

ion technique20 was used. After an overnight fast, aultichannel tube was introduced orally and placed in

he jejunum under fluoroscopic control. The infusionort was opened 5 cm distally to the angle of Treitz andrainage port 20 cm distally to the infusion port. Twodditional channels connected to inflatable latex bal-oons were placed just orad and caudal to the infusionnd drainage ports. When adequate placement of theube was achieved, we first and only inflated the proximalalloon to avoid biliopancreatic contamination, andinsed the distal gut for 5 minutes to wash out residualebris in the lumen. Thereafter, both balloons were in-ated under perception threshold, the segment occludednd perfused at a 5 mL/min rate using a calibratedolumetric pump (Compat enteral; Novartis Nutritionorporation, Minneapolis, MN). An additional gravityrainage port placed just orad to the proximal balloonrevented accumulation of gastroduodenal and biliopan-reatic secretions. Maintenance of constant pressureithin balloons was checked by barostat monitoring.The perfusate was a water-based solution containing

80 mmol/L mannitol, 100 mmol/L xylose, and 5 g/Lolyethylene glycol (PEG) 4000, as a nonabsorbablearker. To limit transmural water flow, it did not con-

ain glucose or electrolytes but was isosmotic (280Osm/kg). The pH was adjusted to 7.8 with NaOH 0.05and temperature was 37°C.

Baseline Stress and Depression EvaluationParticipants were asked to fill in 3 validated ques-

ionnaires: (1) The Modified Social Readjustment Scale ofolmes-Rahe, to evaluate significant life events in the last

ear27; (2) The Perceived Stress Scale of Cohen28 to assessevels of stress in the last month; and (3) Beck’s inventoryor depression29 to assess depression levels in the lasteek.

Menstrual CycleBecause several stress responses may be affected

y hormonal fluctuations, groups were equilibrated afterontrolling for menstrual phase by asking the partici-ants for their menstrual phase and by measuring go-adal hormones. Estradiol and progesterone plasma lev-ls were determined using automated chemiluminescentnzyme immunoassays (estradiol sensitivity, 10 pg/mL;rogesterone, 0.2 ng/mL; Immulite 2000; EUROIDPCtd, Glyn Rhonwy, UK). Intra- and interassay coefficientsf variation (CV) were �10% for both. Interpretation wass follows: Follicular phase, progesterone �1 ng/mL;uteal phase, progesterone, �1 ng/mL.

Cold Pain StressPhysical stress was induced by the cold water

ressor test.30 Briefly, participants immersed their non-

dltm

ff

ewO

(pn�s8A

tSsp

tnirmala

abs

g(

8Tpwawmss

tp1cmmms[[NmM

bsefitemc

tt

F

BA

SIC–

ALI

MEN

TARY

TRA

CT

July 2008 JEJUNAL EPITHELIAL DYSFUNCTION IN STRESSED WOMEN 165

ominant hand in iced water (4°C) for 45 seconds, fol-owed by withdrawal for 15 seconds to prevent adapta-ion to pain. This pattern was repeated for 15 consecutive

inutes.

Assessment of Stress ResponseHand pain perception. The level of hand discom-

ort was assessed using a simple analog scale rangingrom 0 (complete comfort) to 10 (intolerable pain).

Autonomic response. Autonomic response wasvaluated by measuring blood pressure and heart rateith an automated sphygmomanometer (Omron M4-I;mron Healthcare, West Sussex, UK).

Hormonal response. Hypothalamic–pituitary axisHPA) activation was assessed by ACTH and cortisollasma levels as determined by chemiluminescent immu-ometric assays (Immulite 2000; cortisol sensitivity, 0.02g/mL, intra- and interassay CV, 7.0 and 7.7%; ACTH

ensitivity, 5 pg/mL; intra- and interassay CV, 6.7 % and.2%). Baseline normal values were considered as follows:CTH �46 pg/mL; cortisol �25 �g/dL.

Psychological response. Psychological responseo cold pain stress was evaluated by the Subjectivetress Rating Scale (SSRS)31 to assess the level of acutetress experienced by participants during experimentalrocedures.

Gut epithelial response. Epithelial barrier func-ion was assessed by measuring (1) net water flux, lumi-al osmolality, and sodium (Na�) and chloride (Cl�),

ndicators of its absorptive/secretory ability; (2) luminalelease of albumin, an indicator of macromolecular per-

eability; (3) luminal release of �-defensins and fractalkines indicators of local immune innate response; and (4)uminal release of tryptase as an indicator of mast cellctivation.

Experimental DesignPhysical examination, allergy testing, and stress

nd depression assessments were performed the weekefore perfusion (Figure 1). Based on Holmes–Rahecores, participants were stratified into baseline stress

igure 1. Experimental design.

roups: low stress (LS, �150 points) or moderate stressMS, 151–300 points).

The day of the experiment, participants were cited at:00 AM, asked for menstrual phase, and then intubated.o equilibrate groups by menstrual cycle, the last 10otential participants were queried about the cycle theeek before perfusion. Once the segment was occludednd after 30 minutes of equilibration, intestinal effluentsere collected by siphonage every 15 minutes, for 30inutes under basal conditions, for 15 minutes under

tress conditions, and for 45 minutes after cold paintress.

Transit time for the 20-cm segment was calculated byhe time difference between visual appearance and disap-earance of phenol red after injecting a bolus of 1 mg inmL of 0.9% saline into the infusion tubing. Steady-state

onditions were considered to be achieved within 30inutes after segmental occlusion, as indicated by theean transit time for the 20-cm segment (4.7 � 1.2inutes; n � 8), and by osmotic, pH, and electrolyte

tability under basal conditions (Cl� output: LS, 3.41.7– 4.8]; MS, 2.7 [1.6 –3.4] mEq/15 min/20 cm; P � NS;Cl�]: LS, 35.5 [20 –53]; MS, 28.1 [17.0 – 42.4] mEq/L; P �

S; Na� output: LS, 3.4 [1.7– 4.4]; MS, 2.3 [1.4 –3.6]Eq/15 min/20 cm; P � NS; [Na�]: LS, 35.9 [20.7–50.4];S, 27.7 [16.2– 43.2] mEq/L; P � NS).Four 20-mL blood samples were obtained at the end of

asal and stress periods and 15 and 60 minutes aftertress. Hand pain perception was assessed before stress,very 5 minutes during the stress period, 5 minutes afternishing cold pain and up to total recovery of percep-ion. Autonomic and psychological responses werevaluated before intubation, before the stress period, 5inutes after stress initiation, 5 minutes after stress

essation, and at the end of the recovery period.

Calculations, Analysis, and Processingof SamplesIntestinal effluents were collected into chilled con-

ainers at 4°C and every 15 minutes transferred to plasticubes, snap frozen, and stored at �80°C until analyzed.

AwAyaRcdapoNtbUtMjflc3fitbtt

aticc

VNt

qlwmttFwacvnfacrr

mwsc(ls

sswic

pooliiomo

h�rug

T

NAM

F

H

HP

B

DAv

BA

SIC–

ALIM

ENTA

RY

TRA

CT

166 ALONSO ET AL GASTROENTEROLOGY Vol. 135, No. 1

protinin (Sigma-Aldrich, St Louis, MO) 0.01 mL/mLas added to intestinal samples to minimize proteolysis.ppearance of effluents was checked for the presence ofellow content and samples analyzed for lipase activity by

colorimetric method (Cobas Mira Plus CC Analyzer;oche Diagnostics Systems, Barcelona, Spain) to excludeontamination with biliopancreatic secretions. PEG wasetermined by turbidimetry32; albumin, Na� and Cl� byutomated spectrophotometry (Olympus AU5400; Olym-us, Barcelona, Spain). Osmolality was measured with ansmometer (Model 2020; Advanced Instruments, Inc,orwood, MA). Luminal levels of �-defensins and frac-

alkine were determined by enzyme-linked immunosor-ent assay (ELISA; �-defensins: Hycult Biotechnology,den, The Netherlands; detection limit, 50 pg/mL; Frac-

alkine: DuoSet ELISA Development kit, R&D Systems,inneapolis, MN; detection limit, 0.625 ng/mL). Tryptase

ejunal concentration was measured using a specificuoroenzyme-immunoassay (FEIA-UniCAP; Pharma-ia Diagnostics, Uppsala, Sweden) at baseline and for0 minutes after cold stress, according to our previousndings.21 Samples were lyophilized, to increase concen-ration, reconstituted in phosphate buffer saline � 1%ovine serum albumin � a multiprotease inhibitor cock-ail (Sigma-Aldrich) and centrifuged. Tryptase concentra-ion was assayed in supernatants.

Net intestinal water flux and luminal release of medi-tors were calculated after correction for PEG, accordingo standardized formulas.21 Samples with PEG recoveryndex outside 90%–110% range or detectable lipase wereonsidered unacceptable as were studies with �2 unac-eptable periods.

Blood samples were collected in plastic tubes (BDacutainer Plus Plastic K2-EDTA Tubes; Franklin Lakes,J), centrifuged and aliquoted for hormonal determina-

ions (ACTH, cortisol, estradiol, and progesterone).

Statistical AnalysisData are expressed as median with first and third

uartiles (Q1–Q3). Most of the dependent variables ana-yzed along time in both stress subgroups (LS and MS)ere normally distributed, but some variables, in deter-ined periods, failed to achieve normality (Shapiro–Wilk

est). Therefore, for single comparisons, nonparametricests (Mann–Whitney U test, Wilcoxon signed-rank test,isher exact test, and Spearman correlation coefficient)ere used to increase statistical assurance. Hormonal,utonomic, psychological, and epithelial changes wereompared using a 2-way repeated-measures analysis ofariance (ANOVA) because it is known to be robust whenormality is not severely violated and especially when

actor cells have similar sample size (balanced groups)nd variances are homogeneous (Levene test) as was thease. Impact of non-normality was negligible becauseeanalysis of transformed data (logarithmic or square

oot transformation) disclosed similar results (Supple- c

entary Table 1; see supplementary material online atww.gastrojournal.org). Baseline stress group was con-

idered as the between-subjects factor (stress group) andhanges along perfusion time as the within-subject factortime). Area under the curve over time (AUC) was calcu-ated via the trapezoidal method. P � .05 was consideredignificant.

Box plots were used for graphic representation of re-ults unless otherwise stated. The median for each data-et was indicated by the black center line and Q1 and Q3ere the edges of the rectangular area. Empty circles

ndicated minor outliers. All data were analyzed usingommercial software (SPSS 15.0; SPSS Inc, Chicago, IL).

ResultsDemographics, Baseline Stress, andDepression LevelsThirty-one women were initially recruited. Seven

articipants were excluded owing to inadequate placingf the tube (n � 2), unacceptable PEG recovery (n � 3),r because menstrual cycle did not match required stress

evels to equilibrate subgroups (n � 2). Twenty-four stud-es were considered evaluable, with no demographic, clin-cal or menstrual differences between LS and MS groups,ther than stress and depression levels (Table 1, Supple-entary Figures 1 and 2; see supplementary material

nline at www.gastrojournal.org).

Effect of Cold Stress on Hand PainPerceptionCold stress induced a similar enhancement of

and pain perception in both groups (�LS, 8.3 [5.3–9.4];MS, 8.2 [6.6 – 8.8]; P � .001 vs basal). Hand perception

eturned to basal levels a median of 8.0 (6.1–10.5) min-tes after the stress period with similar delays in bothroups.

able 1. Demographic and Clinical Characteristics ofEvaluated Participants

Low stress Moderate stress P

umber of subjects 13 11ge (y) 22 (22–26.5) 24.5 (22.5–28) .34enstrual phase:follicular/luteal

6/7 6/5 NS

ood and respiratoryallergy

0/13 0/11

istory ofgastroenteritis

0/13 0/11

olmes–Rahe’s 56 (25–121.5) 250.5 (203.5–271.3) .001erceived StressScale of Cohen

12 (10–17.5) 20 (16–23.5) .007

eck’s 1.5 (0–3) 3.5 (2.5–7) .007

ata are expressed as median with first and third quartiles (Q1–Q3).Mann–Whitney U test was used for comparisons of continuous

ariables between groups; the Fisher exact test was used to compare

ategorical variables.

besowfi

tHmaf

cT[pnq

r

3iga[s3

[a

bH[c14Mwbi4.

T

A

C

Ws(* ce (H

Fsb

BA

SIC–

ALI

MEN

TARY

TRA

CT

July 2008 JEJUNAL EPITHELIAL DYSFUNCTION IN STRESSED WOMEN 167

Effect of Cold Stress on Autonomic ResponseHeart rate and blood pressure were similar at

aseline. Cold stress significantly increased both param-ters (Supplementary Table 1; Supplementary Figure 3;ee supplementary material online at www.gastrojournal.rg) in both groups (P � NS, LS vs MS, for all 3 variables)ith all measurements returning to baseline 5 minutes afternishing stress.

Effect of Cold Stress on HPA and GonadalHormonal ResponseThere were no basal differences in ACTH or cor-

isol values. Cold stress induced a similar activation ofPA in LS and MS groups as indicated by a significantain effect for time in ACTH [F(3, 22) � 8.95; P � .001]

nd cortisol increases [F(3, 22) � 23.3; P � .001], but notor stress groups (Table 2).

Responses about menstrual cycle were accurate; allases since coincided with hormonal determinations.here were no differences in baseline estradiol (LS, 80.5

45.3– 82.3]; MS, 61.1 [43.5– 82.3] pg/mL; P � NS) androgesterone levels (LS, 1.2 [0.4 – 4.6]; MS, 0.45 [0.4 –2.6]g/mL; P � NS) between stress groups, confirming ade-uate equilibration by menstrual phase.

No changes in gonadal hormones were observed inesponse to cold stress (data not shown).

Effect of Cold Stress on PsychologicalResponseBoth groups scored equal in SSRS at baseline (LS,

.4 [1.7– 4.4]; MS, 5.1 [2.5– 6.9]; P � NS). Cold painnduced a marked raise in perceived stress levels in bothroups [F(5, 22) � 19.7; P � .001; Figure 2]. Althoughnalysis revealed no differences between stress groupsF(1, 22) � 2.9; P � .103], delta peak response wastronger in the MS group (�LS, 0.8 [�1.1 to 2.2]; �MS,.98 [1.6 – 4.9]; P � .002).

Effect of Cold Pain on Jejunal EpithelialResponseNet water flux was similar at baseline (LS, 38.5

28.45–48.6]; MS, 34.9 [25.6–42.3] �L/min/cm; P � NS)

able 2. Hypothalamic–Pituitary Axis Hormonal Response to

0 Minutes 15 Minu

CTH (pg/mL)LS 7.2 (6.1–10.6) 17.0 (13.6–MS 6.4 (5.3–9.0) 12.7 (7.1–1

ortisol (�g/dL)LS 5.25 (3.6–8.95) 9.1 (5.6–1MS 8.0 (5.3–8.9) 10.4 (7.1–1

omen with low (LS) or moderate (MS) background stress were submtress hormones measured before (0), and 15, 30, and 60 minutesQ1–Q3).P � .001 vs basal based on a 2-way ANOVA with post hoc differen

nd a significant increase was observed after cold stress in *

oth groups [F(5, 21) � 10.0; P � .001 for time; Figure 3).owever, this response was different between stress groups

F(1, 21) � 11.1; P � .003]. Indeed, the peak response wasonsistently lower in MS (�[15-0 minutes]: LS, 97.5 [68.4–35.0]; MS, 48.8 [36.6–65.0] �L/min/cm; P � .001; Figure) as was total response to stress (LSAUC, 4.4 [3.3–5.2];SAUC, 2.1 [1.5–2.8] mL/60 min; P � .003). AUC for netater flux showed a significant negative correlation withaseline stress group (r � �0.62; P � .002). Cold stress

ncreased Cl� output in the first 15 minutes only in LS (LS:.6 [1.7–7.1]; MS: 2.3 [1.6–3.6] mEq/15 min/20 cm; P �

041 vs LS [0 minutes]), although analysis revealed no sig-

Stress

30 Minutes 60 Minutes

)* 16.95 (11.7–24.5)* 7.9 (7.35–12.05))* 11.05 (9.3–23.9)* 7.9 (5.0–14.95)

7.5 (4.8–15.6)* 8.1 (6.1–13.9)*14.85 (11.2–16.15)* 10.15 (8.2–14.0)*

to the cold pressor test for 15 minutes and plasmatic variations inthe test. Data are expressed as median with first and third quartiles

olm–Sidak method).

igure 2. Psychological response to cold stress. Psychological re-ponse was evaluated by SSRS at �60 and 0 minutes, before (emptyoxes), and at 5, 20 and 60 minutes after (filled boxes) cold pain stress.

Cold

tes

22.44.25

1.2)*3.9)*

ittedafter

P � .002 for MS [5 min] vs LS [5 min] (Mann–Whitney U test).

ngoitg

1Ng23

tcces[msgfn

(ms[n2u

5Pb[NPa[ru

aaIicsa

Fmss

BA

SIC–

ALIM

ENTA

RY

TRA

CT

168 ALONSO ET AL GASTROENTEROLOGY Vol. 135, No. 1

ificant differences between stress groups, probably due toreater individual variation in the following periods. Na�

utput displayed similar dynamics as Cl�, yet stress-inducedncrease did not reach statistical significance (Supplemen-ary Figure 4; see supplementary material online at www.astrojournal.org).

Baseline albumin output was higher in the MS group (LS,.2 [0.3–2.15]; MS, 3.37 [0.8–4.1] mg/15 min/20 cm; P �S) and significantly enhanced after cold stress in both

roups as shown by a significant main effect for time [F(5,1) � 12.75; P � .001], although peak response was delayed0 minutes in MS (Supplementary Figure 5; see supplemen-

igure 3. Jejunal net water flux. Water flux was measured at 15-inute intervals before (empty boxes) and after (filled boxes) cold pain

tress. A 2-way ANOVA revealed significant differences for time and fortress groups.

ary material online at www.gastrojournal.org). However, aonsistently stronger (Figure 5) and longer sustained in-rease was observed in MS as shown by a significant mainffect for stress groups [F(1, 21) � 8.67; P � .008] and by aignificant larger AUC (LSAUC, 6.35 [0.9–9.6]; MSAUC, 13.978.3–23.1] mg/60 min; P � .008). In addition, this enhanced

acromolecular permeability (albumin output AUC)howed a significant positive correlation with baseline stressroup (r � 0.55; P � .005). Altogether, these striking dif-erences between groups indicate a distorted ability of jeju-al epithelium to respond to acute stress in the MS group.Baseline �-defensins output was similar in both groups

LS, 108.5 [49.8 –117.3]; MS, 68.9 [63.7–91.95] ng/15in/20 cm; P � NS) and cold pain induced a marked but

imilar increase in both groups (LS [15 minutes], 295.9124.0 –1170.8]; MS [15 minutes], 446.8 [127.6 – 886.6]g/15 min/20 cm; F(3, 20) � 4.6; P � .007 for time; F(1,0) � 0.1; P � .76 for stress groups). Fractalkine wasndetectable in intestinal effluents.Tryptase luminal output at baseline was similar (LS,

.65 [3.02–12.02]; MS, 7.98 [2.7–35.5] ng/15 min/20 cm;� NS). Cold pain induced comparable increases in

oth groups, in the first 15-minute period (LS, 24.68.9 – 41.6]; MS, 40.2 [18.2–55.3] ng/15 min/20 cm; P �

S; F(2, 14) � 29.65; P � .001 for time; F(1,14) � 2.42;� .142 for stress groups). Thereafter, tryptase decreased

long the next 15 minutes (LS, 8.75 [5.6 –19.6]; MS, 12.510.5–38.1] ng/15 min/20 cm; P � NS). AUC for tryptaseelease showed no differences between stress groups (Fig-re 6).

DiscussionIncreasing interest has emerged to understand

nd explore women’s excess risk to develop chronic painnd inflammatory and functional disorders.2 ConcerningBS, attention has been focused on mechanisms involvedn gender-based differences in pain modulation33 at theentral level with reports showing enhanced viscerosen-ory perception to colorectal distension34 and greaterctivation of limbic and paralimbic regions that might be

Figure 4. Peak jejunal net wa-ter flux response. Data plots rep-resenting increments in net wa-ter flux in LS and MS groups.Results are expressed as indi-vidual data values at 0 minutes(before cold stress) and at 15

minutes (after cold cessation).

ptteclimia

tass

tstIsmirt2c

sbmaoadtsmgdgassttttbrm

aw

Fbpagamac

F1Ws

BA

SIC–

ALI

MEN

TARY

TRA

CT

July 2008 JEJUNAL EPITHELIAL DYSFUNCTION IN STRESSED WOMEN 169

art of a pain-facilitating circuit.35 However, little atten-ion has been paid to elucidate putative mechanisms athe peripheral level. Here, we have shown that in appar-ntly healthy women experiencing significant life events,ompared to low-stressed women, intestinal and psycho-ogical, but not autonomic or hormonal responses toncoming stimuli were markedly disrupted. This abnor-

al epithelial response was characterized by a dimin-shed secretory ability paralleled by an enhanced perme-bility to antigenic macromolecules.

Stress, either physical or psychological, represents ahreat to the internal homeostasis. In response to stress,

coordinated autonomic, endocrine, and immune re-ponse is generated to maintain stability. However, exces-ive stress exposure, in susceptible individuals impairs

igure 5. Individual jejunal al-umin output responses. Datalots representing increments inlbumin output in LS and MSroups. Results are expresseds individual data values at 0inutes (before cold stress) and

t 15 minutes (after cold stressessation).

igure 6. Jejunal tryptase output. Tryptase output was measured at5-minute intervals for 30 minutes after cold pain stress (P � NS Mann–hitney U test). A 2-way ANOVA revealed no significant differences for

itress groups.

his adaptive response, eventually predisposing thoseubjects to development of new diseases or to exacerba-ion of previously existing ones.36 This may be the case ofBS and women; IBS is more prevalent in women, womeneem to be more vulnerable to life stress11 and to suffer

ore from anxiety and depression disorders,37 and grow-ng evidence, although disputed, indicates a prominentole of stress in IBS pathophysiology and clinical presen-ation.9,38,39 Therefore, to test our hypothesis, we selected

sample populations that varied in their basal levels ofhronic stress but were not depressed.

Surprisingly, evidence from animal and human studiesuggests that women’s vulnerability to life stress cannote easily attributed to neuroendocrine responses becauseen seem to have equal or greater autonomic and ACTH

nd cortisol responses to a variety of stresses.11,40 On thether hand, previous stress experience also influencesutonomic and HPA response and this may result inifferent clinical outcomes for men and women.41 Onhese grounds, women at familial risk of breast cancerhowed stronger cortisol response after brief speech and

ental arithmetic tasks42 and subjects with higher back-round stress also disclosed lower plasma cortisol levelsuring recovery from acute psychological stress,43 sug-esting a sensitizing effect of chronic stress. We observedrobust cardiovascular and HPA activation after cold

tress, but did not find differences either in basal ortress-induced autonomic or HPA recovery responses be-ween groups. Although several reasons may account forhis apparent disagreement, differences in the quality ofhe stressor (physiologic vs psychological), the inten-ional absence of participants showing high levels ofackground stress in our study, the duration of theecovery period, and variations in sample populations

ay explain our results.Hormonal fluctuations along menstrual cycle may be

lso implicated in enhanced stress and pain sensitivity inomen.44 Estrogens have been related to pain process-

ng45 with most, but not all, animal studies indicating an

eemistbe

seqtncrtawpi

ptcgobma

iasmrcttpscigewrhnteriscm

sasi

erclkamiurm

gmirdstnrmacorveAtapub

tlsmprwii

ct

BA

SIC–

ALIM

ENTA

RY

TRA

CT

170 ALONSO ET AL GASTROENTEROLOGY Vol. 135, No. 1

nhancement of pain sensitivity.46 Greater discrepanciesxist in human studies, some showing that women areore sensitive to pain and cardiovascular response dur-

ng luteal phase, whereas others found only small effectizes across the menstrual cycle.47 To prevent any poten-ial effect of gonadal fluctuations, we controlled groupsy menstrual phase. Moreover, we did not observe anyffect of cold stress on estradiol or progesterone levels.

There is now solid evidence demonstrating that womenuffer more distress and tendency to experience negativemotions after acute laboratory stress at a greater fre-uency and intensity than men.11 This increased suscep-ibility is not explained by gender differences in auto-omic or cortisol reactivity during the challenge andannot be attributed to gonadal hormones,12 but may beelated to genetic predisposition, social roles, and emo-ion regulation strategies. In addition, psychological re-ctivity to acute stress has been reported stronger inomen with higher background stress.48 Consistent withrior research, we found enhanced psychological reactiv-

ty to cold stress in women with higher chronic stress.The intestinal epithelium acts as a physical barrier,

reventing most luminal antigens from crossing acrosso reach the mucosal-associated immune system, a pro-ess that prevents overactivation of resident cells andeneration of inadequate proinflammatory signals. Toptimize processing of antigens and luminal content,eyond its physical ability, intestinal epithelial cellserge basic properties, such as water secretion, with

dvanced ones like immune defense and permeability.14

Both acute and chronic stress have been shown tonfluence intestinal barrier. Most animal models of acutend chronic stress have shown increases in water and ionecretion in jejunum and colon.24,49 Studies using seg-

ental perfusion techniques in human jejunum haveevealed that acute stress, either psychological or physi-al, reduced net water absorption20,50 or increased secre-ion.21 These effects were mediated by the parasympa-hetic nervous system and mast cell mediators. Ourresent observations confirm and extend previous oneshowing an increase in intestinal water secretion duringold pain stress in healthy volunteers. Interestingly, novelnformation came up when analyzed according to back-round stress: those women with low background stressxhibited a marked secretory response to cold stresshereas those with MS displayed a significantly reduced

esponse. A similar secretory blunted response patternas been described after neural stimulation in the jeju-um of chronically stressed rats compared with con-rols.51 Chronic stress also seems to promote an impairedpithelial responsiveness to acute noxious stimuli in theat intestine.17 Main driving forces for fluid movementnto the gut lumen are Na�, for absorption, and Cl�, forecretion.52 We have observed higher Cl� output afterold stress only in the LS group, as well as a parallel

ovement of Na�. Therefore, we postulate that our re- j

ults in the MS group are best explained by decreasedctive secretion of Cl�. In our support, other studies havehown that acute and chronic stress induce Cl� secretionn the animal and human jejunum.20,49 –51

Acute and chronic stress have been reported to increasepithelial permeability to small and large molecules inodent jejunum and colon24,49,53 involving mast cells,orticotropin-releasing hormone54 and several cytokinesike interferon-�, interleukin-4, and myosin light chaininase.55 Physical stressors such as surgery and traumand gastrointestinal infections56 increase intestinal per-eability in humans, and, recently, corticotropin-releas-

ng hormone has been shown to enhance transcellularptake of macromolecules in human colonic mucosa viaeceptor subtypes R1 and R2 located on subepithelial

ast cells.57

Cold stress enhanced epithelial permeability in bothroups, but response was greater among women withoderate background stress. Although cold stress also

nduced mast cell activation in both, the differentialesponse in permeability was not related to tryptase; weid not find differences between groups. These findingsuggest that either other mast cell mediators, rather thanryptase, are involved and/or stress-induced intrinsic ab-ormalities in epithelial cells in the MS group may beesponsible for the defective behavior. Enhanced per-

eability could lead to excessive uptake of luminalntigens and bacterial products that may initiate a mu-osal inflammatory response.19 In this sense, our findingf augmented release of �-defensin, a natural antibioticeleased by Paneth cells in response to cholinergic acti-ation, a mechanism induced by cold stress, or bacterialxposure indicates activation of local innate immunity.58

lthough we did not find differences between groups,his measurement was made after single exposure tocute stress. It remains to be determined whether re-eated exposure to incoming stressful stimuli could havencovered a different response in women with moderateackground stress.

In conclusion, healthy women with excessive exposureo stressful life events show an abnormal jejunal epithe-ial response to incoming stimuli. This abnormal re-ponse indicates loss of regulatory mechanisms and

ight represent an initial step in the development ofersistent mucosal dysfunction. Although our intriguingesults may have several implications for understandingomen’s excess risk to develop IBS, clearly, similar stud-

es in males are needed to determine whether our find-ngs are gender specific.

Supplementary Data

Note: To access the supplementary material ac-ompanying this article, visit the online version of Gas-roenterology at www.gastrojournal.org, and at doi: 10.1053/

.gastro.2008.03.036.

1

1

1

1

1

1

1

1

1

1

2

2

2

2

2

2

2

2

2

2

3

3

3

3

3

3

3

3

3

3

4

4

4

4

4

4

BA

SIC–

ALI

MEN

TARY

TRA

CT

July 2008 JEJUNAL EPITHELIAL DYSFUNCTION IN STRESSED WOMEN 171

References

1. Gralnek IM, Hays RD, Kilbourne A, et al. The impact of irritablebowel syndrome on health-related quality of life. Gastroenterology2000;119:655–660.

2. Federman DD. The biology of human sex differences. N EnglJ Med 2006;354:1507–1514.

3. Chang L, Heitkemper MM. Gender differences in irritable bowelsyndrome. Gastroenterology 2002;123:1686–1701.

4. Garcia Rodriguez LA, Ruigomez A. Increased risk of irritable bowelsyndrome after bacterial gastroenteritis cohort study. BMJ 1999;318:565–566.

5. Marshall JK, Thabane M, Garg AX, et al. Walkerton Health StudyInvestigators. Incidence and epidemiology of irritable bowel syn-drome after a large waterborne outbreak of bacterial dysentery.Gastroenterology 2006;131:445–450.

6. Mearin F, Perez-Oliveras M, Perello A, et al. Dyspepsia and irrita-ble bowel syndrome after a Salmonella gastroenteritis outbreak:one-year follow-up cohort study. Gastroenterology 2005;129:98–104.

7. McKeown ES, Parry SD, Stansfield R, et al. Postinfectious irrita-ble bowel syndrome may occur after non-gastrointestinal andintestinal infection. Neurogastroenterol Motil 2006;18:839–843.

8. Gwee KA, Leong YL, Graham C, et al. The role of psychologicaland biological factors in postinfective gut dysfunction. Gut 1999;44:400–406.

9. Spence MJ, Moss-Morris R. The cognitive behavioural model ofirritable bowel syndrome: a prospective investigation of patientswith gastroenteritis. Gut 2007;56:1066–1071.

0. Levy RL, Cain KC, Jarrett M, et al. The relationship between dailylife stress and gastrointestinal symptoms in women with irritablebowel syndrome. J Behav Med 1997;20:177–193.

1. Kudielka BM, Kirschbaum C. Sex differences in HPA axis re-sponses to stress: a review. Biological Psychology 2005;69:113–132.

2. Kelly MM, Tyrka AR, Anderson GM, et al. Sex differences inemotional and physiological responses to the Trier social stresstest. J Behav Ther Exp Psychiatry 2007:39:87–98.

3. Chadwick VS, Chen W, Shu D, et al. Activation of the mucosalimmune system in irritable bowel syndrome. Gastroenterology2002;122:1778–1783.

4. Guilarte M, Santos J, de Torres I, et al. Diarrhoea-predominantIBS patients show mast cell activation and hyperplasia in thejejunum. Gut 2007;56:203–209.

5. Fiocchi C. Intestinal inflammation: a complex interplay of immuneand nonimmune cell interactions. Am J Physiol 1997;273:G769–G775.

6. Ismail AS, Hooper LV. Epithelial cells and their neighbors. IV.Bacterial contributions to intestinal epithelial barrier integrity.Am J Physiol Gastrointest Liver Physiol 2005;289:G779–G784.

7. Shao L, Serrano D, Mayer L. The role of epithelial cells in immuneregulation in the gut. Semin Immunol 2001;13:163–176.

8. Santos J, Saunders PR, Hanssen NP, et al. Corticotropin-releas-ing hormone mimics stress-induced colonic epithelial pathophys-iology in the rat. Am J Physiol 1999;277:G391–G399.

9. Söderholm JD, Yang PC, Ceponis P, et al. Chronic stress inducesmast cell-dependent bacterial adherence and initiates mucosalinflammation in rat intestine. Gastroenterology 2002;123:1099–1108.

0. Barclay GR, Turnberg LA. Effect of psychological stress on saltand water transport in the human jejunum. Gastroenterology1987;93:91–97.

1. Santos J, Saperas E, Nogueiras C, et al. Release of mast cellmediators into the jejunum by cold pain stress in humans. Gas-

troenterology 1998;114:640–648.

2. Qiu BS, Vallance BA, Blennerhassett PA, et al. The role of CD4�lymphocytes in the susceptibility of mice to stress-induced reac-tivation of experimental colitis. Nat Med 1999;5:1178–1182.

3. Alonso C, Santos J, Guilarte M, et al. Corticotropin-releasinghormone promotes jejunal proinflammatory responses in IBSpatients. Gastroenterology 2004;126(Suppl 2):A703.

4. Söderholm JD, Perdue MH. Stress and gastrointestinal tract. II.Stress and intestinal barrier function. Am J Physiol GastrointestLiver Physiol 2001;280:G7–G13.

5. Talley NJ, Stanghellini V, Heading RC, et al. Functional gastrodu-odenal disorders. Gut 1999;45:II37–42.

6. Thompson WG, Longstreth GF, Drossman DA, et al. Functionalbowel disorders and functional abdominal pain. Gut 1999;45:II43–47.

7. Holmes TH, Rahe RH. The social readjustment rating scale.J Psychosom Med 1967;11:213–218.

8. Cohen S, Kamarck T, Mermelstein R. A global measure of per-ceived stress. J Health Soc Behav 1983;24:385–396.

9. Beck AT, Ward CH, Mendelson M, et al. An inventory for measur-ing depression. Arch Gen Psychiatr 1961;4:53–63.

0. Lovallo W. The cold pressor test and autonomic function: a reviewand integration. Psychophysiology 1975;12:268–282.

1. Naliboff BD, Benton D, Solomon GF, et al. Immunologicalchanges in young and old adults during brief laboratory stress.Psychosom Med 1991;53:121–132.

2. Hyden S. A turbidimetric method for the determination of higherpolyethylene glycols in biological materials. LantbrukshoegskAnn 1955;22:139–145.

3. Payne S. Sex, gender, and irritable bowel syndrome: making theconnections. Gend Med 2004;1:18–28.

4. Chang L, Mayer EA, Labus JS, et al. Effect of sex on perceptionof rectosigmoid stimuli in irritable bowel syndrome. Am J PhysiolRegul Integr Comp Physiol 2006;291:R277–R284.

5. Naliboff BD, Berman S, Chang L, et al. Sex-related differences inIBS patients: central processing of visceral stimuli. Gastroenter-ology 2003;124:1738–4177.

6. Gunnar M, Quevedo K. The neurobiology of stress and develop-ment. Ann Rev Psychol 2007;58:145–173.

7. Altemus M. Sex differences in depression and anxiety disorders:potential biological determinants. Horm Behav 2006;50:534–538.

8. Mayer EA, Naliboff BD, Chang L, et al. Stress and the gastroin-testinal tract, V. Stress and irritable bowel syndrome. Am JPhysiol Gastrointest Liver Physiol 2001;280:G519–524.

9. Bennet EJ, Tennant CC, Piesse C, et al. Level of chronic lifestress predicts clinical outcome in irritable bowel syndrome. Gut1998;43:256–261.

0. Kajantie E, Phillips DIW. The effects of sex and hormonal statuson the physiological response to acute psychosocial stress. Psy-choneuroendocrinology 2006;31:151–178.

1. Edwards KM, Burns VE, Reynolds T, et al. Acute stress exposureprior to influenza vaccination enhances antibody response inwomen. Brain, Behav Immun 2006;20:159–168.

2. Gold SM, Zakowski SG, Valdimarsdottir HB, et al. Stronger endo-crine responses after brief psychological stress in women atfamilial risk of breast cancer. Psychoneuroendocrinology 2003;28:584–593.

3. Matthews KA, Gump BB, Owens JF. Chronic stress influencescardiovascular and neuroendocrine responses during acutestress and recovery, especially in men. Health Psychol 2000;20:403–410.

4. Houghton LA, Lea R, Jackson N. The menstrual cycle affectsrectal sensitivity in patients with irritable bowel syndrome but nothealthy volunteers. Gut 2002;50:471–474.

5. Aloisi AM. Gonadal hormones and sex differences in pain reac-

tivity. Clin J Pain 2003;19:168–174.

4

4

4

4

5

5

5

5

5

5

5

5

5

eG1�

SI((t

AtCtf

Rpi

BA

SIC–

ALIM

ENTA

RY

TRA

CT

172 ALONSO ET AL GASTROENTEROLOGY Vol. 135, No. 1

6. Fillingim RB, Ness TJ. Sex-related hormonal influences on painand analgesic responses. Neurosci Biobehav Rev 2000;24:485–501.

7. Kowalczyk WJ, Evans SM. Sex differences and hormonal influ-ences on response to cold pressor pain in humans. J Pain2006;7:151–160.

8. Valdimarsdottir HB, Zakowski SG, Gerin W, et al. Heightenedpsychobiological reactivity to laboratory stressors in healthywomen at familial risk for breast cancer. J Behav Med 2002;25:51–65.

9. Santos J, Benjamin M, Yang PC, et al. Chronic stress impairs ratgrowth and jejunal epithelial barrier function: role of mast cells.Am J Physiol Gastrointest Liver Physiol 2000;278:G847–G854.

0. Barclay GR, Turnberg LA. Effect of cold-induced pain on salt andwater transport in the human jejunum. Gastroenterology 1988;94:994–998.

1. Saunders PR, Kosecka U, McKay DM, et al. Acute stressorsstimulate ion secretion and increase epithelial permeability in ratintestine. Am J Physiol 1994;267:G794–G799.

2. Wood JD. Enteric neuroimmunophysiology and pathophysiology.Gastroenterology 2004;127:635–657.

3. Meddings JB, Swain MG. Environmental stress-induced gastroin-testinal permeability is mediated by endogenous glucocorticoidsin the rat. Gastroenterology 2000;119:1019–1028.

4. Santos J, Perdue MH. Stress and neuroimmune regulation of gutmucosal function. Gut 2000;47:49–52.

5. Ferrier L, Mazelin L, Cenac N, et al. Stress-induced disruption ofcolonic epithelial barrier: role of interferon-gamma and myosinlight chain kinase in mice. Gastroenterology 2003;125:795–804.

6. Dunlop SP, Hebden J, Campbell E, et al. Abnormal intestinalpermeability in subgroups of diarrhea-predominant irritable bowel

syndromes. Am J Gastroenterol 2006;101:1288–1294. w

7. Wallon C, Yang P, Keita AV, et al. Corticotropin releasing hor-mone (CRH) regulates macromolecular permeability via mastcells in normal human colonic biopsies in vitro. Gut 2007;57:7–9.

8. Selsted ME, Ouellette AJ. Mammalian defensins in the antimi-crobial immune response. Nat Immunol 2005;6:551–557.

Received April 23, 2007. Accepted March 13, 2008.Address requests for reprints to: Javier Santos, MD, Digestive Dis-

ases Research Unit, Institut de Recerca Vall d’Hebron, Department ofastroenterology, Hospital Universitari Vall d’Hebron, Paseo Vall d’Hebron19-129, 08035 Barcelona, Spain. e-mail: jsantos@ir.vhebron.net; fax:34 93 4894032/93 4894456.Supported in part by Nestec S. A. and the Spanish Ministry of

anidad y Consumo, Subdirección General de Investigación Sanitaria,nstituto Carlos III, Fondo de Investigación Sanitaria. Dr C. AlonsoCM04/00019), Dr Laura Ramos (CM05/00055), Dr Maria VicarioCD05/00060) and Dr J. Santos (F.I.S. 05/1423) were the recipients ofhese grants.

The authors thank Milagros Gallart, Montse Casellas, and Carmenlastrue for their collaboration in the preparation of perfusion solu-

ions and sample analysis. We also thank Anna Aparici, Maria Teresaasaus, and Purificación Rodríguez for their invaluable assistance inhe performance of jejunal perfusion and Rosa Galard and Olga Illánor their assistance in hormonal determinations.

This paper belongs to a broader group of studies developed within aesearch Agreement with the present study sponsor’s. The sponsor’sarticipation in this study included from collaboration in the process-

ng of samples to analysis of data and review of the manuscript and

as not limited to funding the research project.

S

C

A

N

D

T

A

C

H

S

S

D

H

S

H

S

D

Hab

July 2008 JEJUNAL EPITHELIAL DYSFUNCTION IN STRESSED WOMEN 172.e1

upplementary Table 1. Significance Levels and StatisticalPower

Nontransformeddata

Transformeddata

P P

hloride outputWithin subjects .004 .004Between subjects .68 .77

lbumin outputWithin subjects .001 .001Between subjects .008 .03

et water fluxWithin subjects .001 .001Between subjects .003 .001

efensin outputWithin subjects .007 .001Between subjects .76 .52

ryptase outputWithin subjects .001 .001Between subjects .14 .2

drenocorticotrophichormone

Within subjects .001 .001Between subjects .32 .45

ortisolWithin subjects .001 .001Between subjects .21 .2

and pain perceptionWithin subjects .001 .001Between subjects .25 .35

ubjective Stress RatingScale

Within subjects .001 .001Between subjects .103 .15

ystolic blood pressureWithin subjects .001 .001Between subjects .67 .71

iastolic blood pressureWithin subjects .001 .001Between subjects .87 .89

eart rateWithin subjects .001 .001Between subjects .89 .84

upplementary Table 2. Autonomic Response to Cold Stress

Low stress

eart rate (bpm)Baseline 68.5 (65–72)Stress 75 (68.75–85)

ystolic blood pressure (mmHg)Baseline 109 (106–119.5Stress 124.5 (116–132.5

iastolic blood pressure (mmHg)Baseline 70 (66.5–74)Stress 81 (73–86)

eart rate and systolic and diastolic blood pressures were measured as median with first and third quartiles (Q1–Q3). Data were analyzed

Moderate stress P

70 (61–79) NS78 (73–83) � .01 vs basal

NS vs stress groups

) 109 (108–119) NS) 124 (113–135) � .001 vs basal

NS vs stress groups

70 (69–74) NS87 (75–89) � .001 vs basal

NS vs stress groups

t baseline and 5 minutes after initiation of cold stress. Data are expressedby 2-way ANOVA and Mann-Whitney U test.

pm, beats per minute; NS, not significant.