RESEARCHARTICLE Bioelectrical impedance vector analysis (BIVA) for measuring the hydration status in young elite synchronized swimmers MartaCarrasco-Marginet1,JorgeCastizo-Olier1,LaraRodr´ıguez-Zamora2,3, XavierIglesias1,FerranA.Rodr´ıguez1,DiegoChaverri1,DanielBrotons4,AlfredoIrurtia1* 1 INEFC-BarcelonaSportsSciencesResearchGroup,NationalInstituteofPhysicalEducationofCatalonia, UniversitatdeBarcelona,Barcelona,Spain,2 DepartmentofHealthSciences,MidSwedenUniversity, O¨stersund,Sweden,3 SwedishWinterSportsResearchCentre,MidSwedenUniversity,O¨stersund, a1111111111 Sweden,4 SportandHealthAreaoftheCatalanSportCouncil,GovernmentofCatalonia,Barcelona,Spain a1111111111 a1111111111 *[email protected] a1111111111 a1111111111 Abstract OPENACCESS Purpose Citation:Carrasco-MarginetM,Castizo-OlierJ, Theassessmentofbodyhydrationisacomplexprocess,andnomeasurementisvalidfor Rodr´ıguez-ZamoraL,IglesiasX,Rodr´ıguezFA, allsituations.Bioelectricalimpedancevectoranalysis(BIVA)hasemergedasarelatively ChaverriD,etal.(2017)Bioelectricalimpedance noveltechniqueforassessinghydrationstatusinsports.WeappliedBIVAa)todetermine vectoranalysis(BIVA)formeasuringthehydration statusinyoungelitesynchronizedswimmers. hydrationchangesevokedbyanintensesynchronizedswimming(SS)trainingsession;b) PLoSONE12(6):e0178819.https://doi.org/ tocharacterizethesampleofyoungeliteswimmersinrelationwithanonathleticreference 10.1371/journal.pone.0178819 population;andc)togenerateits50%,75%and95%percentilesofthebioelectrical Editor:TiagoM.Barbosa,NanyangTechnological variables. University,SINGAPORE Received:February14,2017 Methods Accepted:May21,2017 Forty-nineeliteSSfemaleswimmersoftwoagecategories,comen(C :13.9±0.9years, o Published:June7,2017 n=34)andjunior(J:16.3±0.6years,n=15),performedalong,highintensitytrainingses- r Copyright:©2017Carrasco-Marginetetal.Thisis sion.Bodymass(BM)andbioelectricalvariables(R,resistance;Xc,reactance;PA,phase anopenaccessarticledistributedundertheterms angle;andZ,impedancemodule)wereassessedpre-andpost-training.BIVAwasusedto oftheCreativeCommonsAttributionLicense, characterize1)thedistributionpatternofthebioelectricalvector(BIAvector)forbothage whichpermitsunrestricteduse,distribution,and reproductioninanymedium,providedtheoriginal groups,and2)pre-topost-trainingBIAvectormigration.Bioelectricalvariableswerealso authorandsourcearecredited. correlatedwithBMchangevalues. DataAvailabilityStatement:Allrelevantdataare withinthepaperanditsSupportingInformation Results files. Mostswimmersweremostlylocatedoutsidethe75%andsomebeyondthe95%percentile Funding:Thisworkwassupportedbyresearch grantsfromGeneralitatdeCatalunya,InstitutCatalà ofthebioelectricaltoleranceellipsesofthegeneralpopulation.TheBIAvectorshowedsta- delesDones(U-34/10)http://www20.gencat.cat/ tisticallysignificantdifferencesinbothC (T2=134.7,p=0.0001)andJ (T2=126.2,p< o r portal/site/icdones,andAgènciadeGestio´d’Ajuts 0.001).Bothgroupswerealsobioelectricallydifferent(T2=17.6,p<0.001).Afterthetrain- UniversitarisideRecerca(AGUAR),(VCP/3346/ ingsession,adecreaseinBM(p=0.0001)andanincreaseinBIAvariables(p=0.01)was 2009)http://agaur.gencat.cat/es/inici/,andConsejo SuperiordeDeportes(001/UPB10/11)http://www. observed.BIVAalsoshowedasignificantpre-postvectormigrationbothinC (T2=82.1; o PLOSONE|https://doi.org/10.1371/journal.pone.0178819 June7,2017 1/17 BIVAinsynchronizedswimming csd.gob.es/.MartaCarrasco-Marginetwasapre- p<0.001)andJ (T2=41.8;p<0.001).NocorrelationswereobservedbetweenBM r doctoralresearcheralsosupportedbyAGAUR changesandbioelectricalvariables. (VCP/3346/2009).Thefundershadnoroleinstudy design,datacollectionandanalysis,decisionto publish,orpreparationofthemanuscript. Conclusions Competinginterests:Theauthorshavedeclared BIVAshowedspecificbioelectricalcharacteristicsinyoungeliteSSathletes.Considering thatnocompetinginterestsexist. thedecreaseinBMandthemigrationoftheBIAvector,weconcludethatthehomeostatic Abbreviations:BCM,Bodycellmass;BIA, hydrationstatusoftheseyoungelitefemaleswimmerswasaffectedbytheexecutionof Bioelectricalimpedanceanalysis;BMI,Bodymass intensetrainingsessions.Fromamethodologicalperspective,BIVAappearstobesensitive index;BIVA,Bioelectricalimpedancevector analysis;ECW,Extracellularwater;ECW:TBW, enoughtodetectsubtlehydrationchanges,butfurtherresearchisneededtoensureits Extracellularwater/totalbodywaterratio;Fc, validityandreliability.Moreover,thesefindingshighlighttheimportanceofensuringade- Characteristicfrequency;FFM,Fat-freemass;FM, quatefluidintakeduringtraininginyoungSSathletes. Fatmass;h,Bodyheight;Hotelling’sT2,Test comparingmeangroupvectors;ICW,Intracellular water;Cm,Membranecapacitance;PA,Phase angle;R,Bioelectricalresistance(R/hifadjustedby height);RXcgraph,R/hvs.Xc/hprobabilisticplot; SD,Standarddeviation;SF-BIA,Single-frequency Introduction bioelectricalimpedanceanalysis;s-RPE,Session ratingofpereceivedexertion;SS,Synchronized SincebecomingpartoftheOlympicprogramin1984,synchronizedswimminghasenjoyeda swimmming;TBW,Totalbodywater;Xc, growingworldwidepopularity.Thishighlytechnicalsportcombinesaerobicandanaerobic Bioelectricalreactance(Xc/hifadjustedbyheight); endurance,flexibility,strength,power,acrobaticsandperformanceskills,andchoreography Z,Bioelectricalimpedance;Zvector,Vectoryieldby theRXcgraph;˚C ,˚C ˚C ,Coreandskin [1]requiringlonghoursoftrainingtoattainsuchbroadattributes[2]. core hand, foot temperaturesoftherighthandandfoot. Mostsynchronizedswimmersenterthesportasyounggirlsattherecreationallevel,andby theageof13–15years,themoretalentedathletesstarttrainingandcompetingatamore intenselevel[3].Eliteswimmerstendtotrain6daysperweekwithonedayoff,andtraining sessionsusuallylastbetween3and5hours[2,3]andaredividedintwoworkoutsperdaywith differentcontent.Forexample,sport-specificskilltraininginthewatercouldfollowapoolses- sionofswimmingforaerobicfitness.Adrylandtrainingcouldoccurlaterinthesameday, consistingofflexibility,drylanddrills,orapsychologysession[3].Asaresult,trainingde- mandsattheeliteleveloftenresultinhigh-volume—averagingapproximately40hperweek— andhigh-intensitytrainingprograms[2,4]. Youngathletesmayexperiencefluidimbalancesifsomeconditionsaremet,withpossible consequencesontheirphysicalperformance,cognitiveperformanceandhealthmaintenance [5–8]. Despitethehighrequirementsatsuchayoungage,informationaboutfluidintakeand hydrationduringthestrenuousSStrainingisscarce.Femaleswimmersshowlowenergyavail- ability,especiallyinphasesofintensifiedtrainingperformedbeforecompetition[9].Findings highlighttheimportanceofensuringadequatefluidintakeduringsynchronizedswimming trainingtoenableoptimalperformance.Nevertheless,ithasbeensuggestedthatthereislower fluidreplacementduringpoolsessions,possiblyduetothelimiteddrinkbreaksorbecause athletestrytoavoidpotentialgastrointestinaldiscomfortiftheexerciserequiresthemtobe upsidedown[10]. Theassessmentofbodyhydrationisadynamicandcomplexprocess,andnomeasurement isvalidforallsituations[11].Inthiscontext,bioelectricalimpedancevectoranalysis(BIVA) emergesasarelativelynoveltechniqueforassessinghydrationstatuswithoutalgorithm-inher- enterrorsorrequiringassumptionssuchasconstanttissuehydration[12,13].BIVAusesraw bioelectricalimpedanceparameters,i.e.,resistance(R,theoppositiontoflowthroughintra- andextracellularionicsolutions)andreactance(Xc,additionaloppositionfromthecapaci- tanceeffectofcellmembranesandtissueinterfaces),standardizedbyheight(h)toremovethe PLOSONE|https://doi.org/10.1371/journal.pone.0178819 June7,2017 2/17 BIVAinsynchronizedswimming effectofconductorlength,whichyieldsavectorthatisplottedinanRXcgraph[14].Overall, BIVApropertiesareespeciallyinterestingforhydrationassessmentinsports,duringboth competitionsandtraining[15,16]. Theaimofthisstudywas,first,todeterminethehydrationchangesevokedduringasyn- chronizedswimmingtrainingsessionbyfocusingonchangesofthewhole-bodyimpedance vector.Secondly,wecomparedtheSSyoungelitesamplewithareferencenonathleticpopula- tionandgeneratedits50%,75%and95%percentilesofthebioelectricalvariablesdistribution, alsoknownastoleranceellipses.Wehypothesizedthatthehydrationstatusoftheyoungswim- merswouldbealteredbythelong,intensetrainingsessionsandthebarriersforanadequate fluidintake.Inthisline,theseswimmerswouldbecharacterizedbyaspecificdistributionof BIVAvariableswhencomparedtothereferencepopulation. Materialsandmethods Participants Eighty-fourfemaleSSathletesoftwocompetitivecategories,comen(C ,n=53)andjunior o (J,n=31)swimmers,includingtheentireSpanishnationaljuniorteam,wererecruitedfor r thestudyinMarch2012.Thirty-five(C ,n=19;J,n=16)didnotmeetinclusioncriteria. o r Inclusioncriteriawereasfollows:(1)tohavecompetedatnationaland/orinternationallevel atleastintheprevioustwoyears;(2)tonotpresentinjuriesoranyclinicalconditionatthe timeofthestudy;(3)tobeinapostmenarchealstatewiththeovariancyclebetweendays5thto 11th[17];(4)tonotbeundercontraceptivesormenstrualcyclepharmacologicalregulators treatment.Samplesizewascalculatedtodetectaneffectsize(ES)=0.5,withanestimatedsam- plestandarddeviation(SD)=7.0,andaSDforchanges=0.7,requiringaminimumof15sub- jectspergroup.Power(P=1–β)wassetat0.80,andtheconfidenceintervalwasα=0.05. Forty-nineathleteswereselected(C ,n=34;J,n=15).Allsubjectsvoluntarilyparticipatedin o r thestudyanddeliveredwritteninformedconsent,withparentalpermissionwhenneeded.The studywasconductedfollowingtheWMAHelsinkiDeclarationStatement[18]andapproved bytheEthicsCommitteeforClinicalSportResearchofCatalonia.Thecharacteristicsofthe participantsareshowninTable1. Studydesign Thispre-postquasi-experimentalstudywasbothdescriptiveandcorrelationalandaimedto approachthetopicfromanecologicalperspective.Thestudyanalyzedtheacuteadaptations inducedbysynchronizedswimmingtrainingsessiononbodymass–BM(kg),bioelectricalvec- torvariables[resistance(R,Ω),resistanceadjustedbyheight(R/h,Ω/m),reactance(Xc,Ω), reactanceadjustedbyheight(Xc/h,Ω/m),impedancemodule(Z,Ω),andphaseangle(PA,˚)] andtheextracellularwater/totalbodywaterratio(ECW:TBW,%).Inadditiontotheseinde- pendentvariables,severalotherswererecordedtocharacterizethesample(Table1)andthe training(Table2). Procedures ThestudywasconductedtwoweeksbeforetheSpanishNationalSynchronizedSwimming Championship,withinthe4-weekprecompetitivemesocycle.Onetrainingsessionwasper- formedbyeachgrouponthesameday.TheprotocolischronologicallysummarizedinFig1. ToattainastateofeuhydrationpriortoBIAmeasurements[6],swimmerswererequiredto abstainfromcaffeine,alcoholandexercisethedaybeforetheinvestigation[19].Theywere alsoinstructedtodrink3.0Loffluidover24h(2.0Ltobeconsumedbetween6:00p.m.and PLOSONE|https://doi.org/10.1371/journal.pone.0178819 June7,2017 3/17 BIVAinsynchronizedswimming Table1. Characteristicsofparticipants. Allswimmers Comen Junior Unpairedt-test (95%CI) (95%CI) (95%CI) n=49 n=34 n=15 t p General Age(years) 14.6±1.4 13.9±0.9 16.3±0.6 -10.851 0.0001* (14.2–15.0) (13.6–14.2) (16.0–16.7) Training(h/week) 19.4±7.6 15.0±2.7 30.0±3.8 -15.911 0.0001* (17.4–21.8) (14.0–15.9) (28.0–32.1) Practice(years) 6.9±1.8 5.9±1.1 9.1±1.0 -9.980 0.0001* (6.4–7.4) (5.6–6.3) (8.6–9.7) Anthropometric Height(cm) 163.3±7.6 161.9±8.2 166.3±4.8 -1.943 0.058 (161.1–165.4) (159.0–164.8) (163.7–169.0) BM(kg) 49.1±7.0 47.2±7.0 53.5±5.2 -3.103 0.003* (47.1–51.2) (44.8–49.7) (50.6–56.3) BMI(kg/m2) 18.4±1.8 18.0±1.9 19.3±1.3 -2.514 0.015* (17.9–18.9) (17.3–18.6) (18.6–20.0) Fatmass(%) 16.5±4.4 15.6±4.7 18.6±2.6 -2.382 0.021* (15.2–17.8) (13.9–17.2) (17.2–20.1) Musclemass(%) 38.0±4.7 37.7±5.4 38.8±2.6 -0.722 0.474 (36.7–39.4) (35.8–39.6) (37.3–40.2) Bioelectrical R/h(Ω/m) 319.7±36.7 328.4±38.8 299.9±21.6 3.286 0.002* (309.1–330.2) (314.9–341.9) (287.9–311.9) Xc/h(Ω/m) 39.9±3.9 40.0±4.5 39.6±2.2 0.395 0.695 (38.7–41.0) (38.4–41.5) (38.4–40.8) PA(˚) 7.1±0.5 7.0±0.5 7.5±0.4 -4.166 0.0001* (7.0–7.3) (6.8–7.1) (7.3–7.7) Valuesaremean±SD;BM,bodymass;BMI,bodymassindex;R,resistance;Xc,reactance;PA,phaseangle;h,height;CI,95%confidenceinterval *significantdifferencesbetweencomenandjuniorswimmers(p<0.05). https://doi.org/10.1371/journal.pone.0178819.t001 10:00p.m.)inadditiontotheirhabitualwesterndietarypractices.From10:00p.m.untilthe startofthepre-testnextmorning,nofurtherfluidorfoodintakewasallowed[6].From7:00a. m.to8:00a.m.thefollowingday,aftermonitoringbodyandskintemperatures,pre-training measurementswereperformedinathermoneutralroom(25˚C)toobtainanthropometric (BM)andbioelectricaldata(R,R/h,Xc,Xc/h,PA,Z,andECW:TBW).Immediatelyafter,par- ticipantsconsumedastandardizedbreakfastconsistingof1cheeseandhamsandwich,1plain yogurt,1banana,and220mLofnaturalorangejuice[20].At10:00a.m.,allswimmersper- formedacategory-specifictrainingsessionina50-mindoorpoolwith30mavailableforuse (watertemperature:25–26˚C).ThecharacteristicsofbothtrainingareshowninTable2. Fluidintake(H2O)duringthetrainingwasmonitoredbyacertifieddietician.Swimmers wereinstructedtodrinkasimilaramountofwaterinthemiddleandattheendoftraining. Pre-andpost-training,BIAmeasurementswereconductedafterurinationanddefecation[21] tominimizetheinfluenceoffood/fluidingestion[22]andexercise[23];pre-trainingmeasure- mentswereconductedwhilefasting,andthepost-trainingdatawereobtainedwithinthefirst hourofrecovery.Justaftercompletingthetrainingsession,theratingofperceivedexertion Table2. Characteristicsofthetrainingsessions. Allswimmers Comen Junior Unpairedt-test (95%CI) (95%CI) (95%CI) n=49 n=34 n=15 t p Duration(min) 167.6±28.0 149.6±3.3 208.4±10.3 -21.695 0.001* (159.6–175.7) (148.5–150.8) (202.7–214.1) Internaltrainingload RPE(a.u) 6.6±0.5 6.4±0.5 6.8±0.6 -2.220 0.03* (6.4–6.7) (6.3–6.6) (6.5–7.1) Session–RPE 1102.4±231.3 963.9±78.5 1416±129 -12.572 0.001* (1036.0–1168.9) (963.5–991.3) (1344.8–1488.0) Waterintake(L) 0.6±0.2 0.5±0.2 0.7±0.3 -2.177 0.04* (0.5–0.6) (0.4–0.6) (0.5–0.8) Valuesaremean±SD;RPE,ratingofperceivedexertion(CR-10scale);a.u,arbitraryunits;CI,95%confidenceinterval *significantdifferencesbetweencomenandjuniorswimmers(p<0.05). https://doi.org/10.1371/journal.pone.0178819.t002 PLOSONE|https://doi.org/10.1371/journal.pone.0178819 June7,2017 4/17 BIVAinsynchronizedswimming Fig1.Studyprotocol.˚C,bodyandskintemperaturemeasurements;Pre-T,pre-trainingmeasurements;Post-T,post-trainingmeasurements;RPE, ratingofperceivedexertion;C ,comen;J,junior. o r https://doi.org/10.1371/journal.pone.0178819.g001 (RPE)wasassessedusingtheBorgCR-10scale[24].Finally,aftercheckingthatbodyandskin temperaturesweresimilartothoseregisteredinthepre-trainingmeasurements,thepost- trainingassessmentwasperformed. Anthropometricassessment. BMwasmeasuredtothenearest0.05kgusingacalibrated weighingscale(Seca710,Hamburg,Germany).Height(h)wasmeasuredtothenearest1mm usingatelescopicstadiometer(Seca220,Hamburg,Germany).Bodymassindex(BMI)was calculatedasbodymass/height2(kg/m2).Anthropometricmeasurementsweretakenaccord- ingtothestandardcriteriaofTheInternationalSocietyfortheAdvancementofKinanthropo- metry[25]. Whole-bodybioimpedanceassessment. RandXcweremeasuredusingapreviouslycali- bratedplethysmograph(Z-Metrix,BioparHom,LeBourget-du-Lac,France)thatemitteda 77μAalternatingsinusoidalcurrentatsevenoperatingfrequencies(1,5,50,150,200,250,and 325kHz).The50-kHzsinglefrequencywasselectedforBIVA[26];meanwhile,multi-fre- quencycapabilitieswereusedtoestimatebodycomposition–fatmass(FM)andmusclemass (MM),andtheECW:TBWwascalculatedbyECW/TBW•100.Thedeviceprovidesimpedance valueswithanaccuracyaverageerrorof0.95±1.58%andaveragerepeatabilityerrorsof 0.55±0.38%forallthefrequencyrange(1to1000kHz)[27].Bioelectricalmeasurementswere conductedundercontrolledconditions[14]throughthestandardwhole-body,tetrapolar,dis- talBIAtechnique[28].Theanatomicalsitesforelectrodes(RedDot2660–5,3MCorporate Headquarters,St.Paul,MN,USA)weremarkedwithawaterproofpen[29].Bioelectricalmea- surementswererepeateduntiltheywerestabletowithin1Ω(usuallyuptothreetimeswithin anintervalof20–30s).Theaveragevaluewasusedincalculations[21]. RegardingtheBIVAmethod,thecorrelationbetweenRandXcdeterminestheellipsoidal formofthebivariateprobabilitydistributions(confidenceintervalsforaveragevectorsandtol- eranceforindividualvectors).Thevectordirectionisdefinedasthephaseangle(PA)andis thegeometricrelationshipbetweenRandXc.PAhasbeenvalidatedasanindicatorofcellular health[12,28]andhasbeeninterpretedasanindexoffluiddistributionbetweentheintracel- lularandextracellularcompartments[30],showinganinversecorrelationwiththeECW:TBW [31].Ontheotherhand,thelengthofthevectorindicateshydrationstatusfromfluidoverload (decreasedresistance,shortvector)toexsiccosis(increasedresistance,longervector),anda sidewaysmigrationofthevectorduetoloworhighreactanceindicatesadecreaseorincrease inthedielectricmass(membranesandtissueinterfaces)ofsofttissues[32].Theindividual vectorcanberankedontheRXcpointgraphwithregardtotoleranceellipsesrepresenting 50%,75%and95%accordingtothevaluesofareferencepopulation[14].Acomparison PLOSONE|https://doi.org/10.1371/journal.pone.0178819 June7,2017 5/17 BIVAinsynchronizedswimming betweenthemeanvectorsofdifferentsampleswiththe95%confidenceellipsescanbeper- formedontheRXcmeangraph.Furthermore,themeanvectordisplacementofagroupwith the95%confidenceellipsepre-topost-interventionwasplottedontheRXcpairedgraph[33]. Temperatureassessment. Core(˚C )andskintemperaturesoftherighthand(˚C ) core hand andfoot(˚C )weremeasuredusingthermistorsconnectedtoadatalogger(Squirrel2010, foot GrantInstrumentsLtd,Cambridge,UK).Allswimmerswereinstructedtotakeacoldshower (ascoldastolerable)for10–15minutespost-training,inordertoreducecutaneousbloodflow andtemperatureandremoveaccumulatedelectrolytes[34].Skintemperature,asasurrogate forcutaneousbloodflow[35],wasmeasuredjustbeforeBIAmeasurements;thisverifiedthe returntotemperaturesclosetothepre-trainingvalues(p<0.05):Pre-˚C :36.8±0.2˚Cvs. core Post-˚C :37.2±0.3˚C;Pre-˚C :29.6±0.8˚Cvs.Post-˚C :29.2±1.1˚C;Pre-˚C : core hand hand foot 29.0±1.2˚Cvs.Post-˚C :28.6±1.0˚C.Ambientairtemperatureandrelativehumidityin foot theindoorpoolareawerealsocontrolled(27.5±0.5˚Cand64.5±1.5%,respectively). Internaltrainingloadassessment. Theindividualsession-RPE(s-RPE)waschosenfor ratingtheperceivedexertionduringtraining[36].TheCR-10RPEscale[24]wasshowntothe swimmersimmediatelyafterthetrainingwascompleted.Scoreswerecomputedbymultiply- ingthedurationofthetrainingbytherelativeRPEvalues.Oneweekbeforethestudy,allpar- ticipantswereassessedrepeatedlyduringatleast3trainingtodiscloselearningeffectsandto improvetheconsistencyofthemeasurements[37]. Statisticalanalysis Descriptivestatistics(mean,SD)werecalculatedforeachindependentvariableandagecategory. Oncethedataweretestedfornormality(Shapiro-Wilkstest),differencesinanthropometric (BM)andbioelectricalvariables(R,Xc,R/h,Xc/h,PAandZ)betweenpre-andpost-training wereanalyzedbytheStudent’spairedt-test.TheStudent’sunpairedt-testwasusedtoanalyze groupdifferencesbetweenagecategories.Whole-bodybioimpedancevectorswereanalyzedby theRXcgraphmethod[14]usingtheBIVAsoftware[38].Eachswimmerwasplottedinthetol- eranceellipses(50%,75%and95%)ofthe14-to15-year-oldhealthyfemaleItalianreference population[39]asthiswasthereferencepopulationclosestinagetooursample.TheBIVA meangraphwasperformedtocomparewhole-bodyvectorsofC vs.J groups,andeachSS o r groupvs.thereferencepopulation.TheBIVApairedgraphwasusedtoanalyzepre-topost- trainingchangesinthevectorsofC andJ.Toexaminethemagnitudeofpre-postratiochanges o r inanthropometricandbioelectricalvariables,deltavalues(Δ,%ofpre)werecalculated.Toesti- matetherelevanceofthesechanges,relativeESwerecalculatedusingCohen’sd.Accordingto Cohen[40],ESwasdefinedassmall,d(cid:20)0.2;medium,d(cid:20)0.5;andlarge,d(cid:20)0.8.Pearson’scor- relationcoefficientwasusedtodeterminepossiblestatisticalassociationsbetweena)PAvs.chro- nologicalageandPAvs.theECW:TBW;andb)ΔBMvs.BIAvectorvariables(ΔR/h,ΔXc/h, ΔPA,ΔZ).Apairedone-sampleHotelling’sT2testwasusedtoanalyzepre-topost-training changesinthevectorthroughthe95%confidenceellipses.Atwo-sampleHotelling’sT2testwas usedtodeterminetheBIAvectordifferencesbetweenC andJ andbetweeneachSSgroupvs. o r thereferencepopulation.P<0.05wasconsideredsignificant. Results DeterminantsofBIAvectordistributionpatterninsynchronized swimmers TheBIVApointgraph(Fig2)indicatedthatswimmersfellmostlyoutsidethe75%tolerance ellipseregardlessofageorcompetitionlevel;inmanycases,theywereoutsidethe95% PLOSONE|https://doi.org/10.1371/journal.pone.0178819 June7,2017 6/17 BIVAinsynchronizedswimming Fig2.BIVApatternsbeforeandaftertraining.Ontheleftside,scattergramsoftheC andJ individual o r (aswellasthemean)impedancevectors,plottedonthe50%,75%,and95%toleranceellipsesofthe correspondinghealthyfemalereferencepopulation[39]aredisplayedbothforpre-andpost-training(Pre-T andPost-T,respectively).Ontherightside,meanvectordisplacementsofC andJ frompre-topost-training o r areshown.R/h,height-adjustedresistance;Xc/h,height-adjustedreactance;T2,Hotelling’sT2test;p-value (significanceatp<0.05). https://doi.org/10.1371/journal.pone.0178819.g002 toleranceellipse,denotingahigherdensityofbodycellmass(BCM)thanthereferencepopula- tion.DifferencesintheBIAvectorincomparisonwiththereferencepopulationwerefound forC (T2=134.7,p=0.0001)andJ (T2=126.2,p<0.001),aswellasbetweenbothgroupsof o r SSswimmers(T2=17.6,p<0.001)(Fig3). Fig4showsthe50%,75%and95%toleranceellipsescorrespondingtothewholeSSsample (C andJ together):R/h=319.7±36.7Ω/m;Xc/h=39.9±3.9Ω/m;r=0.78. o r Pre-postdifferences TheBIAvectormigration(Fig2)wascharacterizedbyanincreaseinR/handXc/h,indicating milddehydrationaftertrainingbothinC (T2=82.1)andJ (T2=41.8)(p<0.001).Thiswas o r paralleledbyadecreaseinBMinbothgroupsofswimmers(p=0.0001)asshowninTable3. Incontrast,allbioelectricalvariablessignificantlyincreased(Table3). BIVAcorrelations Apositivecorrelation(r=0.45,p=0.001)wasfoundbetweenPAandchronologicalageinthe wholeSSsample.Additionally,PAwasnegativelyrelated(r=-0.91;p<0.001)totheECW: TBW.NocorrelationswereobservedbetweenbioelectricalpretopostchangesinrelationtoBM. Discussion Thisstudyshowedthatsynchronizedswimmersexperiencedamodestlevelofdehydration afteranintensetrainingsession(BMloss~0.6–0.8%BM)thatwasdetectedbyBIVA.In addition,wereportaspecificBIAvectordistributionintheseyoungeliteSSswimmersincom- parisonwithahealthy,nonathleticreferencepopulationofsimilarage.Infact,thisisthefirst PLOSONE|https://doi.org/10.1371/journal.pone.0178819 June7,2017 7/17 BIVAinsynchronizedswimming Fig3.RXcmeangraph.The95%confidenceellipsesforthemeanimpedancevectorsofC (dottedline o ellipse),J (darkdashedlineellipse)andthehealthyfemalereferencepopulation(solidlineellipsewithvector) r [39]areshown.R/h,height-adjustedresistance;Xc/h,height-adjustedreactance;C ,comen;J,junior;SS, o r synchronizedswimmers. https://doi.org/10.1371/journal.pone.0178819.g003 timethatspecificreferencedistributionellipsesinafemalesportinggroupisbeingreported (Fig4). BIVAallowsforananalysisofboththehomeostaticstateandpossibleBIAvectormigra- tion,arisingfromanyvariationinbodyfluid[12,26].Nowadays,BIVAisawidelyusedtech- niqueinmedicineasavalidtoolintheassessmentofdifferentphysiologicalstatesandclinical conditionsinwhicheuhydrationisfrequentlyaltered,suchasrenaldisease[41],criticallyill patients[42],pulmonarydisease[43],heartfailure[44],gastrointestinaldisease[45],andpreg- nancyandpostpartum[46].Itspropertiesareespeciallyinterestingforhydrationassessments inboththetrainingprocessandcompetitivesportingevents[16,47]. Nevertheless,inprotocolsmeasuringparametersbeforeandafterexercisetoanalyzeacute vectorshifts,certainfactorsthatmaygenerateerrorsinthebioelectricalsignalshouldbecon- trolledinordertoprovideaccurateandreliableresults,including:skinpreparation[48];previ- oushydrationstatus[49];previousconsumptionoffoodorbeverage[35,50,51];body positionandpostureduringmeasurements[21,51,52];electrodeimpedance[53],position andplacementmodification[51,52];timeofbodyfluidstabilization[54,55];variationsin cutaneousbloodflowandtemperature[35,52];skinelectrolyteaccumulationproducedby physicalexercise[35];reproducibilityofbioelectricalmeasurementsinfluencedbybiological intra-day[56]andinter-dayvariations[50];environmentalconditions[21,52];menstrual cycle[17,57];andinjurycondition[58]. PLOSONE|https://doi.org/10.1371/journal.pone.0178819 June7,2017 8/17 BIVAinsynchronizedswimming Fig4.Toleranceellipses.50%,75%,and95%toleranceellipsesgeneratedoftheentiregroupof synchronizedswimmers.R/h,height-adjustedresistance;Xc/h,height-adjustedreactance. https://doi.org/10.1371/journal.pone.0178819.g004 Despitetheecologicaldesignofthisresearch,thestudyprotocolattemptedtocontrolthese factors.Asmentionedabove,ingestionofamealorbeveragehasaninfluenceonZ,which maydecreaseovera2-to4-hperiodafterameal,generallyrepresentingachangeof<3%inZ values[22].Therefore,inourstudy,post-exerciseBIVAmeasurementscouldhavebeeninflu- encedbybreakfastandwaterintakeinthemiddleofthetrainingsession,possiblyunderesti- matingZvaluesby~9–10Ω.Theamountofwaterintakeattheendofthetrainingshouldnot haveaffectedtheBIVAmeasurementsbecausetherecentingestionofamealorbeverage(<1 hfromtheingestiontoBIAmeasurements)appearstobe"electricallysilent"andtohavea minimaleffectonwhole-bodyZ[59].Withregardtothetemperaturecontrol,itisknownthat every1.0˚CincreaseintheskincanleadtoadecreaseinRofupto~11%[60].Possiblepre- postBIAdifferencesrelatedtoenvironmentalandcutaneoustemperaturesoftheswimmers werecontrolled,acceptingincreasesordecreaseslowerthan1˚Cas,inthisrangeofvalues,dif- ferencesinZappearnottobesignificant[61].Finally,itshouldbenotedthatthestudysample wascomposedoffemaleathletes.Thus,tominimizethebodyfluidfluctuationscausedbythe effectoffemalehormonalkinetics[17]andconsequentchangesinperformanceability[62], internationalrecommendationswerefollowed[28,57].Thus,thoseswimmerswhowereina premenstrualphase(lutealorsecretoryphase)orwhoweretakingcontraceptivesand/ormen- strualcyclepharmacologicregulatorswereexcluded. PLOSONE|https://doi.org/10.1371/journal.pone.0178819 June7,2017 9/17 BIVAinsynchronizedswimming Table3. Anthropometricandbioelectricalparametersbefore(Pre)andafter(Post)training. Pre Post Δ-value Pairedt-test p-value Cohen’sd (95%CI) (95%CI) % t p d Comen(n=34) Anthropometric BM(kg) 47.2±7.0 46.9±7.0 -0.8±0.6 8.081 0.0001* 0.20a (44.8–49.7) (44.5–49.3) Bioelectrical R(Ω) 529.5±46.1 548.8±48.5 3.7±3.0 -7.251 0.0001* 0.39b (513.4–545.6) (531.9–565.7) Xc(Ω) 64.4±5.1 67.7±5.0 5.2±3.3 -9.193 0.0001* 0.67c (62.7–66.2) (66.0–69.5) R/h(Ω/m) 328.4±38.8 340.5±41.0 3.7±3.0 -7.104 0.0001* 0.30b (314.9–341.9) (326.1–354.8) Xc/h(Ω/m) 40.0±4.5 42.0±4.6 5.2±3.3 -8.905 0.0001* 0.43b (38.4–41.5) (40.4–43.6) PA(Ω) 7.0±0.5 7.1±0.5 1.5±2.5 -2.863 0.007* 0.20a (6.8–7.1) (6.9–7.2) Z(Ω/m) 330.9±38.9 343.1±41.2 3.7±3.0 -7.178 0.0001* 0.20a (317.2–344.4) (328.7–357.4) r(R/h,Xc/h) 0.84 0.84 — — — Junior(n=15) Anthropometric BM(kg) 53.5±5.2 53.2±5.1 -0.6±0.4 4.634 0.0001* 0.17a (50.6–56.3) (50.3–56.0) Bioelectrical R(Ω) 498.5±35.1 518.5±38.9 4.0±3.3 -4.870 0.0001* 0.53c (479.1–518.0) (497.0–540.1) Xc(Ω) 65.8±2.9 70.2±4.8 6.6±3.9 -6.447 0.0001* 0.82c (64.2–67.4) (67.5–72.8) R/h(Ω/m) 299.9±21.6 311.9±23.4 4.0±3.3 -4.864 0.0001* 0.53c (287.9–311.9) (298.9–324.9) Xc/h(Ω/m) 39.6±2.2 42.2±3.4 6.6±3.9 -6.352 0.0001* 0.62c (38.4–40.8) (40.3–44.1) PA(Ω) 7.5±0.4 7.7±0.4 2.4±3.3 -2.909 0.011* 0.45b (7.3–7.8) (7.5–7.9) Z(Ω/m) 302.5±21.7 314.8±23.5 4.1±3.3 -4.928 0.0001* 0.51c (290.5–314.5) (301.7–327.8) r(R/h,Xc/h) 0.66 0.76 — — — Valuesarethemean±standarddeviation;BM,bodymass;R,resistance;Xc,reactance;h,height;PA,phaseangle;Z,impedancevectormodule;r, PearsoncorrelationcoefficientbetweenR/handXc/h;%Δ,percentdifferencesPretoPost;CI,95%confidenceinterval *significantdifferencesbetweenPreandPost,p-value<0.05(pairedt-testPrevs.Post);a,smalleffectsize((cid:20)0.2);b,mediumeffectsize(d(cid:20)0.5);c, largeeffectsize(d(cid:20)0.8). https://doi.org/10.1371/journal.pone.0178819.t003 BIAvectorchangesevokedbytraining ThisstudyisthefirsttouseBIVAtocharacterizevariationsinhydrationstatusinyoungSS athletesevokedbytraining.RXcpairedgraphsshowedsignificantvectorchangesafterexercise inbothgroups(Fig2),whichwereinterpretedasmilddehydration(averageloss<1%BM) [63].Nevertheless,nocorrelationwasobservedbetweenchangesinBMandBIAvectormigra- tion.Thiscouldbeduetothefluidintakeoftheathletesduringtraining,whichwasmain- tainedduetotheecologicalstudydesign.Arecentinvestigationinwhichnofood/fluidintake wasallowedfoundsimilarresultswithexercise-induceddehydration[29].Theresearchers notedthatthiscouldbeinfluencedbyinadequatecriteriaforstablebioelectricalimpedance measurementsorbyexercise-relatedfactors,suchassweatrate,respiratorywaterlossandoxi- dativewaterproduction,thatmayleadtoBMlosswithoutaneffectivenetnegativefluidbal- ance[64].However,theseresultsdifferfromotherstudies[65,66]thatfoundasignificant relationshipbetweenchangesinbioelectricalvaluesandBMinducedpassivelyand/orchroni- cally.Itispossiblethatgreaterchangescouldhavebeenobservediftheswimmershadnot ingestedfluidsduringthelongandintensetraining(Table2).Nevertheless,nosignificantrela- tionshipswerefoundwithBMinthepresentstudy. PLOSONE|https://doi.org/10.1371/journal.pone.0178819 June7,2017 10/17