angkatan 46

Upload: andika-siswanta

Post on 04-Apr-2018

216 views

Category:

Documents


0 download

TRANSCRIPT

  • 7/29/2019 ANGKATAN 46

    1/9

    ResearchOBSTETRICS

    First-trimestermetabolomicdetection oflate-

    onsetpreeclampsia

    Ray O. Bahado-Singh,

    MD, MBA; RanjitAkolekar, MD;

    Rupasri Mandal, PhD;

    Edison Dong, BSc;

    Jianguo Xia, PhD;

    Michael Kruger, MS;

    David S. Wishart,

    PhD; Kypros

    Nicolaides, MD

    www.

    AJOG .o

    r

    g

    OBJECTIVE: Wesoughttoidentifyfirst-

    trimestermaternalserumbio-

    markersforthepredictionoflate-

    onsetpreeclampsia(PE)using

    metabolomicanalysis.

    STUDYDESIGN: Inacase-

    controlstudy,nuclearmagneticresonance

    basedmetabolomicanalysiswasperformedonfi

    rst-trimestermaternal serumbetween110-

    136weeksofgestation.Therewere30cases

    oflate-

    onsetPE,ie,requiringdelivery37weeks,and59u

    naffected

    controls.Theconcentrationsof40metabolitesw

    erecomparedbe-tweenthe2groups.Wealsocompared30early-

    onsetcasestothe

    RESULTS:

    Atotalof14metabolitesweresignificantlyelevat

    edand3sig-

    nificantlyreducedinfirst-trimesterserumoflate-onsetPEpatients.Acom-plexmodelconsistingofmultiplemetabolitesandmaternaldemographicchar-acteristicshada76.6%sensitivityat100%specificityforPEdetection.Asimplifiedmodelusingfewerpredictorsyielded60%sensitivityat96.6%spec-ificity.Strongseparationoflate-vsearly-onsetPEgroupswasachieved.

    CONCLUSION:

    Significantdifferencesinthefirst-

    trimestermetabolites

    werenotedinwomenwhowentontodevelopedlate-onsetPEandbe- tweenearly-andlate-onsetPE.

  • 7/29/2019 ANGKATAN 46

    2/9

    late-onsetgroup.

    Keywords:metabolomics,preeclampsiaprediction

    Citethisarticleas:Bahado-SinghRO,AkolekarR,MandalR,etal.First-trimestermetabolomicdetectionoflate-onsetpreeclampsia.AmJObstetGynecol2013;208:58.e1-7.

    proves it is expected that thisnumber

    is associated with significant fetalmor-

    Metabolomics,arelat

    ivelyrecent

    addition totheomicsfamily

    ,could grow by a factor of 10.

    2

    Because

    biditie

    s.Thepathophysiologyisthought

    involvesthehigh-

    throughputcharacter- ization andinterpretation of the small- molecule

    metabolites (1500 d) pro- duced by

    cells, tissues, and organisms. To

    date, 8000 human metabolites from

    80 chemical classes have been

    identi- fied or catalogued.1 As

    technology im-

    FromtheDepartmentofObstetricsand Gynecology,WayneStateUniversity,Detroit,MI(DrBahado-SinghandMrKruger);Harris

    BirthrightResearchCentreforFetalMedicine,KingsCollegeHospital,London,United

    Kingdom(DrsAkolekarandNicolaides);and

    theDepartmentsofBiologicalSciences(Drs

    MandalandWishartandMrDong)and

    ComputingSciences(DrsXiaandWishart),

    UniversityofAlberta,Edmonton,Alberta, Canada.

    ReceivedJune18,2012;revisedNov.4,2012; acceptedNov.8,2012.Thisstudywaspartlysupportedbyagrant

    fromtheFetalMedicineFoundation,Charity

    Number1037116.

    Theauthorsreportnoconflictofin

    terest.

    Reprintsnotavailablefromtheauthors.

    0002-9378/$36.00

    2013PublishedbyMosby,Inc.

    http://dx.doi.org/10.1016/j.ajog.2012.11.0

    03

    SeeJournalClub,page8

    7

    of the wide chemical diversity of

    metab-

    olites,theirtightcouplingwithenviron-

    mental interactions (food, drugs, gut

    microbiota), and their huge phenoty-

    pic-dependent concentrationvariations

    6

    (10

    ),metabolomicsoffersapowerful,

    quantitative route to describe the

    actual phenotype of cells, tissues, or

    organisms in both normal and

    diseased states. Re- cently,

    significant advances have oc- curred

    both in metabolite identification

    techniques1,2 and computational

    tech- niques3foranalyzingthelargevolumeof data

    generated by metabolomic studies.

    Thereiscurrentlytremendousinteresti

    n the use of metabolomics for the

    charac-

    terizationandearlydiagnosisofcomple

    x diseases.4

    Preeclampsia(PE)isacommonobstet-

    ricdisordercharacterizedbyhypertens

    ion and proteinuria during

    pregnancy. It is a cause of

    significant morbidities, affectingthehealthofboththemother5 andfetus.

    However, its causes and

    pathophysiol-

    ogylargelyremainamystery.Itnowap-

    pears that PE is at least 2 fairly

    distinct disorders,anearly-

    onsetandalate-onset form.6,7

    Theearly-onsetvarietytypically

    occurs 34-35 weeks of pregnancy,

    and

    to be failure of trophoblast invasion

    of thematernalspiralarteriole8

    resultingin maintenance of highmaternal vascular resistance. This is

    consistent with the high frequency

    of placental underperfu- sion

    reported9 in this disorder. Thelate-

    onsetformisconsideredtobe

    moreofamaternalconstitutionaldisor-

    der10

    duetounderlyingmaternalmicro-

    vascular disorders such as

    hypertension

    orageneticpredispositioninwhichpoo

    r

    trophoblastinvasionisthoughttoplaya

    lesssignificantrole.Late-

    onsetPEissig-

    nificantlymorecommonandwhileitof-

    ten has a mild course can be

    associated with significant clinical

    morbidities.11 It is therefore

    important to investigate its

    pathogenesis and if possible to

    develop biomarker predictors of this

    disorder.

    Studieshavenowconfirmedtheclinical

    feasibility of first-trimester

    screening for early-, late-, and

    intermediate-onset vari- eties of PE

    using demographic, clinical,

    biomarker,anduterinearteryDoppleri

    n- formation.12,13

    Recently,theNationalCol-

    laboratingCenterforWomensandChi

    l- drens Health in the United

    Kingdom issued clinical

    guidelines14 for routine early

    p

    re

    n

    a

    t

    a

    l

    s

    c

    r

    e

    e

    n

    i

    n

    g

    f

    o

    r

    P

    E

    b

    a

    s

    e

    d

    o

    n

    58.e1 AmericanJournalofObstetrics & Gynecology

  • 7/29/2019 ANGKATAN 46

    3/9

    JANUARY2013

  • 7/29/2019 ANGKATAN 46

    4/9

    www.AJOG.org

    ObstetricsResearchTABLE1Demographicandothercharacteristics:late-onsetpreeclampsiavscontrolgroup

    disorderofpregnancyandwhohadblood

    collectedwithin3daysofassessmentofthe

    late-onset PE case. There was no evident

    source of bias in the selection of cases orLate-onset controls. Thedefinition of PE used was

    Parameter preeclampsiaControl Pvalue

    thatproposedbytheInternationalSociety

    No.ofcases 30 59 for the Study of

    Hypertension in Preg-..............................................................................................................................................................................................................................................

    nancy,16 namely systolic pressure 140

    Maternalage,y,mean(SD) 31.2(6.4)30.8(5.6).81

    ..............................................................................................................................................................................................................................................

    mmHgordiastolicpressure 90mmHg

    Racialorigin,n(%) .02

    ..................................................................................................................................................................................................................................... on

    2occasions4hoursapart 20weeksWhite 14(46.7)44(74.6) of gestation,

    in women who were previ-.....................................................................................................................................................................................................................................

    Black 14(46.7)14(23.7) ously

    normotensive. Proteinuria was de-.....................................................................................................................................................................................................................................

    finedasatotalof300mgina24-hoururine

    Asian 0(0) 1(1.7)

    .....................................................................................................................................................................................................................................

    collectionor2readingsofatleast2 pro-

    Mixed 2(6.7) 0(0)

    ..............................................................................................................................................................................................................................................

    teinuria on a midstream or catheterizedNullipara,n(%) 12(40)31(52.5).37

    .............................................................................................................................................................................................................................................. urinespecimen in the absence of a 24-

    Weight,kg,mean(SD) 74.9(15.7)67.7(12.2).03 hour

    urine collection must also have..............................................................................................................................................................................................................................................

    Crown-rumplength,mm,mean(SD)62.0(9.1)62.7(7.6).69....................................................................................................................................................................

    ..........................................................................

    Uterinepulsatilityindex,MoM,mean(SD)1.07(0.35)0.98(0.31).22....................................................................................................................................................................

    ..........................................................................

    MoM, multiplesofmedian.

    Bahado-Singh.Late-onsetpreeclampsia,metabolomics.AmJObstetGynecol2013.

    beenpresentinadditiontothehyperten-

    sion. Proteinuria must also have been

    present in addition to the hypertension

    for the diagnosis of PE. No HELLP

    syn- drome or gestational hypertension

    casesmaternal demographic, historical, and ics Committee.

    Briefly, women were re-

    were included.

    06

    Nuclear magnetic resonance (NMR)

    clinicalcharacteristics.Itispossiblethat,

    in the future, combining clinical with

    cruited at 1113

    wee

    ks gestation.spectrometry

    was used for metabolite

    biomarker predictors could further en-

    hance screening accuracy. Our primary

    objective was to evaluate the use of

    metabolomics to identify first-trimester

    biomarkers of late-onset PE. Second-

    arily, we evaluated the diagnostic accu-

    racy of these markers for late-onset PEprediction.Finally,weevaluatedtheca-

    pability of metabolomics for distin-

    guishinglate-fromearly-onsetPE.

    MATERIALS AND METHODS

    StudypopulationThis study is part of an ongoing

    prospec- tive study being conducted by

    the Fetal Medicine Foundation,

    London, United Kingdom, for the first-

    trimester predic-

    tionofimportantfetalandobstetricdisor-

    ders. Institutional review board project

    #02-03-033 approval was obtained on

    March14,2003.Thedetailsofpatienteval-

    uation and study methods have been

    ex- tensively described in a prior report

    of metabolomic prediction of early-

    onset PE.15 A routine population of

    Britishwomenwasprospectivelyscreenedfrom

    March 2003 through September 2009

    andtheyallgavewrittenconsenttopartic-

  • 7/29/2019 ANGKATAN 46

    5/9

    JANUARY2013

    AmericanJournalofObstetrics &

    Gynecology

    58.e2

  • 7/29/2019 ANGKATAN 46

    6/9

    Research Obstetrics

    www.AJOG.org

    are significant metabolite differences be-

    tweenthenormalandcontrolgroups.

    PLS-DAisusedtoenhancethesepara-TABLE2Serummetaboliteconcentrationsbynuclearmagneticresonancetionbetweenthegroupsbysummarizing Late-onsetPE,mean

    Controls,mean(SD)

    the data into a few latent variables that

    Metabolite (SD)

    (concentration

    (concentrationin Fold

    maximize covariance between the re-

    inmol/L) mol/L)

    P

    value

    change

    sponse and the predictors.18 To mini- mize the possibility that the observed..............................................................................................................................................................................

    Glycerol 800.7(541.7)312(296.8) .0012.4

    separation on PLS-DA is due to chance,

    permutation testing was performed...............................................................................................................................................................................................................................................

    1-methylhistidine70.3(40.0) 38.9(20.3) .0011.7

    ........................................................................................

    ........................................................................................

    ..............................................................

    Thisinvolvedrepeated(2000

    times)data sampling, with

    different random label-

    ing. A significant P valueindicates that

    .......................................................................................................................................................................

    .......................................................................

    Acetone

    22.1(11.4)

    14.9(8.5)

    .0031.6

    the separation observed between groups

    is very unlikely to be due to chance. The..............................................................................................................................................................................................................................................

    Trimethylamine6.03(2.0) 7.6(3.3) .0050.88........................................................................................

    ........................................................................................

    ..............................................................

    MetaboAnalyst computer

    used to perform (PCA and PLS-DA)

    analyses.19 Avariableimportanceinpro-

    Pyruvate..............................................................

    Hydroxyisovalerate_36.5(3.3)4.7(2.5)

    .0081.4

    jection (VIP) plot18 is a plot ranking the.....................................................................................

    .........................................................................................................................................................

    Acetamide

    11.9(7.8)

    16.1(6.4)

    .

    0080.7

    3

    metabolites based on their importance

    in discriminating study from control..............................................................................................................................................................................................................................................

    Glucose 4312.9(1783.0)3362.4(765.9).0081.2

    .......................................................................................

    .......................................................................................

    ................................................................

    groups. Metabolites with the

    Hydroxybutyrate_228.0(14.4)

    21.2(7.5)

    .021.3

    powerful group discriminators....................................................................

    ...................................................................

    ........................................................................................................

    Creatinine

  • 7/29/2019 ANGKATAN 46

    7/9

    63.2(16.5)

    55.1(14.7)

    .0211.1

    In comparing the concentrations of.....................................................................................

    .........................................................................................................................................................

    Creatine

    metabolitesbetweengroups,outliertest- ....................................................................................................................................................................................................

    ingwasperformedusingDixonQtest.20 The

    Dixon Q test is used for identifica-

    Citrate..........................................................

    Hydroxybutyrate_349.9(46.7)

    29.7(19.1)

    .0381.4

    tion of outliers in the dataset and re-.....................................................................................

    .........................................................................................................................................................

    Leucine

    114.5(98.5)

    87.1(61.9)

    .1121.2

    places that value with the one closest to.....................................................................................

    .........................................................................................................................................................

    Acetate

    it.Replacementofoutliershelpstomeet

    the assumption of normal distribution..............................................................................................................................................................................................................................................

    Betaine80.6(101.5)49.1(52.5)33.3(23.6) 21.6(9.4).121.6.141.5

    andequalvariancebetweengroups.Onlya

    singlevalue(forvaline)wasadjustedinthis..............................................................................................................................................................................................................................................

    Glutamine 253.1(131.1)218.5(66.9) .1821.2

    ...................

    ...................

    ...................

    .........................................................................

    .........................................................................

    ...................................

    67.7(42.

    6)

    56.1(37.

    2)

    .191.2

    fashion, however. Kolmogorov-Smirnov.....................................................................................

    .........................................................................................................................................................

    Ornithine

    andShapiro-Wilktestsofnormaldistribu-

    tion were performed. Metabolite concen-..............................................................................................................................................................................................................................................

    Acetoacetate36.8(17.4)18.9(9.8)42.3(22.5)16.5(9.6).240.87

    .271.1trations in late-onset PE vs controlswere

    comparedusingthe2-tailedttest.Mann-WhitneyUtestwasusedincomparingme-..............................................................................................................................................................................................................................................

    Alanine 366.8(204.8)323.8(151.2).271.1..............................................................................................................................................................................................................................................

    Lactate 1213.1(564.7)1100.9(689.3).441.1........................................................................................

    ........................................................................................

    ..............................................................

    tabolite concentrations between

  • 7/29/2019 ANGKATAN 46

    8/9

    thatwerenotnormallydistributed.Other

    independentvariablesincludingfetalCRL,

    Threonine................................................

    ........................................................................

    ........................................................................

    ..............................................

    66.2(62.5).50.94

    Propyleneglycol11.1(5.0)

    11.8(4.9)

    .510.93

    uterine artery Doppler, PI, and maternal.....................................................................................

    .........................................................................................................................................................

    Formate

    27.0(13.8)

    29.0(17.8)

    .61.02

    age,parity,weight,ethnicity,smoking,and.....................................................................................

    .........................................................................................................................................................

    Tyrosine

    medical disorders were included in the...............................................................................................................................................................................

    ..............................

    geneticcomputinganalysesalongwiththe

    metabolite concentrations for PE

    Proline.......................................................................................

    .......................................................................................

    ................................................................

    Serine

    172.2(57.7)165.7(56.4)

    148.4(103.4)158.6(92.2)

    .611.04 .6350.93

  • 7/29/2019 ANGKATAN 46

    9/9

    prediction.

    Genetic programming is a branch of..............................................................................................................................................................................................................................................

    Arginine 136.3(55.5)131.2(35.9) .651.04.......................................................................................

    .......................................................................................

    ..........................................................

    ......evolutionary computing

    computingisabranchofgeneticprogram-

    ming.Theadvantageofgeneticcomputing

    Phenyl

    alanine

    78.0(4

    5.9)........

    ..................

    ..................

    ..................

    ..................

    ..................

    ..................

    ........................

    ........................

    ........................

    ........................

    ........................

    ..80.9(45.

    4).780.96

    Glycine

    238.4(129.3)244.0(115.7).840.97

    lies in its ability to handle nonnormally

    distributed outcome measures and the

    largevolumeofdatageneratedfromom-..............................................................................................................................................................................................................................................

    Bahado-Singh.Late-onsetpreeclampsia,metabolomics.AmJObstetGynecol2013. (continued)58.e3

    AmericanJournalofObstetrics & Gynecology JANUARY2013