• Aucun résultat trouvé

Development of a nomogram for individual preterm birth risk evaluation

N/A
N/A
Protected

Academic year: 2021

Partager "Development of a nomogram for individual preterm birth risk evaluation"

Copied!
5
0
0

Texte intégral

(1)

HAL Id: hal-02006582

https://hal-amu.archives-ouvertes.fr/hal-02006582

Submitted on 11 Apr 2019

HAL is a multi-disciplinary open access

archive for the deposit and dissemination of

sci-entific research documents, whether they are

pub-lished or not. The documents may come from

teaching and research institutions in France or

abroad, or from public or private research centers.

L’archive ouverte pluridisciplinaire HAL, est

destinée au dépôt et à la diffusion de documents

scientifiques de niveau recherche, publiés ou non,

émanant des établissements d’enseignement et de

recherche français ou étrangers, des laboratoires

publics ou privés.

Development of a nomogram for individual preterm

birth risk evaluation

Marion Gioan, Florence Fenollar, Anderson Loundou, Jean-Pierre Menard,

Julie Blanc, Claude d’Ercole, Florence Bretelle

To cite this version:

Marion Gioan, Florence Fenollar, Anderson Loundou, Jean-Pierre Menard, Julie Blanc, et al..

Devel-opment of a nomogram for individual preterm birth risk evaluation. Journal of Gynecology Obstetrics

and Human Reproduction, Elsevier, 2018, 47 (10), pp.545-548. �10.1016/j.jogoh.2018.08.014�.

�hal-02006582�

(2)

Original

Article

Development

of

a

nomogram

for

individual

preterm

birth

risk

evaluation

Marion

Gioan

a

,

Florence

Fenollar

b

,

Anderson

Loundou

c

,

Jean-Pierre

Menard

d

,

Julie

Blanc

e

,

Claude

D’Ercole

e,f

,

Florence

Bretelle

e,f,

*

a

CHGSainte-Musse,54,rueHenri-Sainte-Claire-Deville,83100Toulon,France

b

Unite´ derecherchesurlesmaladiesinfectieusestropicalesetemergentes,UM63,CNRS7278,IRD198,INSERM1095,13000Marseille,France

c

MedicalEvaluation,DepartmentofPublicHealth,Assistancepublique–hoˆpitauxdeMarseille,AMU,Aix-MarseilleUniversite´,13000MarseilleFrance

dConseilde´partementalduVal-de-Marne,94000Cre´teil,France e

DepartmentofGynaecologyandObstetrics,Gynepole,AP–HM,Assistancepublique–hoˆpitauxdeMarseille,13000Marseille,France

f

AMU,Aix-MarseilleUniversite´,13000Marseille,France

Introduction

PretermBirth(PB)remainsthecauseofneonatalmorbidityand mortalityindevelopedcountrieswithupto9%ofpregnanciesword wile [1]. In EuropePB occursin 4.1 to 8.2%of birth leadingto maternalprolonghospitalisationsandtreatmentsespeciallyinhigh riskpregnancies[2,3].Basedonthe definition,anasymptomatic pregnant woman with a history of PB or late miscarriage is consideredathighriskforPB.Ultrasonographiccervical measure-mentisthestandardgoldtoevaluatethisrisk.Theriskdependson cervicallengthandgestationalageatmeasurement[4].Theaddition

ofriskfactorfortheevaluationofPBriskcouldbeacluetoimprove thepredictionofPBrisk.ForexampleBacterialvaginosis(BV)isarisk factorforobstetricalcomplications[5,6].Thedetectionbymolecular biology of microorganisms present in BV is a new diagnostic approach[7,8].Me´nardetalshowedthatthetimetodeliverywas shorterwhenhighatopobiumvaginalloadwasdetectedinahigh riskpopulation[8,9].Noreliablepredictivemethodexiststodayto definetheriskofPBinhigh-riskpregnancies[10].

Thenomogramisthegraphicrepresentationoftheprobability foreachpatientofanevent.Withthismodel,theriskcalculationis simple, reproducible and personalised. Two recent studies proposed an assessment of PB risk in high-risk populations [11,12].Unfortunately,theircalculationsdidnotincorporatemost ofthevariablesrecognisedasriskfactorsforPB,suchasvaginal swabresults,historyofadverseeventormaternalsmoking.Our studyaimwastodevelopanewtooltoevaluateindividualriskfor PBinahigh-riskpopulation.

ARTICLE INFO

Articlehistory: Received9May2018

Receivedinrevisedform16July2018 Accepted20August2018

Availableonline24August2018

Keywords: Prematurebirth Nomogram Highrisk Asymptomatic Bacterialvaginosis Shortcervix Latemiscarriage ABSTRACT

Objective. –Thisstudyaimedtodevelopanewtoolforpersonalisedpretermbirthriskevaluationin high-riskpopulation.

Study design. –813 high-risk asymptomatic pregnant women included in a French multicentric prospectivestudywereanalysed.Clinicalandparaclinicalvariables,includingscreeningforbacterial vaginosiswithmolecularbiology,cervicallength,havebeenusedtocreatethenomogram,basedonthe logisticregressionmodel.Thevaliditywascheckedbybootstrap.Adownloadablecalculatorwasbuild. Results. –Nineriskfactorswereincludedinthismodel:historyoflatemiscarriageand/orpreterm delivery,activesmoking,ultrasoundcervicallength,termofpregnancyatscreening,bacterialvaginosis, prematureruptureofmembranes,dailytravelmorethan30min.Discriminationandcalibrationofthe nomogramrevealedgoodpredictiveabilities.Theareaunderthereceiveroperatingcharacteristiccurve was0.77(95%CI;0.72–0.81).Themeanabsoluteerrorwas0.018,whichshowedpropercalibration.The optimalriskthresholdwas23.2%withasensitivityof74%,aspecificityof72.7%andapredictivenegative valueof90.6%.

Conclusion. –The nomogramcan help to better define individual preterm birth risk in high-risk pregnancies.

C 2018ElsevierMassonSAS.Allrightsreserved.

* Corresponding author. Unite´ de recherche sur les maladies infectieuses tropicales et emergentes,UM63, CNRS7278, IRD 198, INSERM1095, 13000 Marseille,France

E-mailaddresses:melkorsa@gmail.com(M.Gioan),florence.bretelle@ap-hm.fr

(F.Bretelle).

Available

online

at

ScienceDirect

www.sciencedirect.com

https://doi.org/10.1016/j.jogoh.2018.08.014

(3)

Materialsandmethods Targetpopulation

Theestablishednomogramhasbeencreatedfromthedatabase ofaprospectivemulticentricFrenchcohort[13].Pregnantwomen, whoare14to34weeks(wks)pregnant,admittedforprenatalcare ineightFrenchteachinghospitals(amongthemtwolevelIIandsix levelIII,allwerepublichospitalexceptone),wishingtoparticipate inthestudy,wereeligible.

Takingintoaccounta30%proportionofpretermbirthsbefore 37weeksinourhighriskpopulation,asamplesizeof690women should show a hazard ratio (HR) of approximately 3.3 for Atopobium vaginae at a power at 80. With lost to follow-up patients and missing data (estimated at 20%) into account, 820womenisrequired.

Inclusioncriteria

Thepatientsincludedwereolderthan18yearsandatriskforPB. Thisriskwasdefinedastheexistenceofashortcervix(acervical length<25mmmeasuredbytransvaginalultrasound)and/oran obstetrichistory:historyofpretermbirthand/orlatemiscarriage (spontaneousexpulsionofapregnancy14and<22wks).

The exclusion criteria were: multiple pregnancies, treated hypertension, foetal malformation, antiphospholipid syndrome, diabetes,pre-eclampsia,renaldisease,anyauto-immunedisease, oranantibiotictreatmentinthepast7days.

A completemedical examinationand interrogationcollected thedemographic data, medical history (age, parity, body mass index,smokingduringpregnancy,obstetrichistory, andcurrent pregnancydata) and theclinical characteristicsof each patient (uterinecontractions,clinicalsignsofBVorprematureruptureof membranes).

Vaginal sampleand ultrasound measurementof thecervical lengthwereperformed.Gestationalagewasdeterminedfromthe date of the last menstrual period or on the first trimester ultrasoundincase ofaone-weeklag.Thedailytraveltimedata wascollected and divided into two groups: more or less than 30min.Obstetric and neonataloutcomes werecollected in the postpartumperiodthroughconsultationofmedicalfiles. Bacteriologicalanalysis

Thebacteriologicalanalysiswasperformedwithself-collected vaginalswabs.Menardetal.[14]demonstratedthevalidityand reliabilityof this methodversuspractitioner-collected swabfor molecularquantification.Eachvaginalsamplereceiveda molecu-larbiology analysisbasedonquantitativePCR.Theresultswere blinded for the medical team of obstetricians. The organisms targetedbyquantitativePCR andselectedin ourstudywereA. vaginaeandGardnerella vaginalis.Molecularquantificationof A. vaginae108 copies/ml and/orG. vaginalis 109 copies/ml has

beendescribedascommoninwomenwithBVflora[8].Thetrial wasregisteredatClinicalTrials.gov(identifierNCT00484653),and funded by the national hospital clinical research program (Programme de Recherche Clinique, number 2007-A00069-44). ‘‘Le Comite´ de Protection des Personnes Sud Me´diterrane´e V’’ approved the project (number 07.019). Analysis method was previouslyreported[13].

Statisticalanalysis

StatisticalanalysiswasperformedwiththeIBMSPSSStatistics version20software.Theassociationbetweenthevariablesselected in the preliminary study and the risk for PB was tested with

univariateand multivariatelogistic regression analyses. Logistic model calibration was assessed using the Hosmer–Lemeshow goodness-of-fittesttoevaluatethediscrepancybetweenobserved andexpectedvalues.Oddsratioswerereportedwith95%confidence interval(CI95%).Qualitativevariableswerepresentedintheformof enrollment countsand percentages. Quantitativevariableswere expressedasmeanstandarddeviation.Theretainedandintegrated variables were the term (expressed asweeks ofgestation (wks)), history of preterm and late miscarriage, premature rupture of membranes,smoking,dailytraveltime,BVdiagnosedwithmolecular biology, sonographic measurement of cervical length, and the combination of history of preterm birth and cervical length <25mm.Discrimination,calibration,andnomogramwereperformed usingthe‘‘rms’’libraryoftheRsoftware(http://www.R-project.org) [15].Thepredictivemodelwasinternallyvalidatedforcalibrationwith bootstrap resampling. This compares predicted PB and actual PB probabilities.The calibration wasstudiedwith a

x

2-test withtwo

degreesoffreedom.Discriminationwasexaminedusingtheareaunder the receiver operating characteristic (ROC) curve (AUC), graphic representationofthefalse-positiveratebasedonthesensitivityof eachmodelvalue.AUC>0.8representsanexcellentdiscriminating power,andisgoodwhencomprisedbetween0.7and0.8[16].AnAUC of0.5israndom.TheAUCisassociatedwith95%confidenceinterval (CI).TheoptimalthresholdvaluefortheriskofPBisthepointonthe ROCcurvethatisfurthestfromthediagonalandthatrepresentsthe zerocontributiontest[17].Allthetestsweretwo-sided.Thestatistical significancewasdefinedasP<.05.

Results Study

BetweenJuly2007andApril2012,813patientswereincluded. PCRscoreswereperformedon764vaginalsamples.Dataresults werereportedelsewhere[13].Inbrief,themeanmaternalagewas 29.4years(5.6)andthepercentageofmultiparous patientswas 17.5%.Meangestationalageofpregnanciesatinclusionwas26.3(5.1) wks.Atotalof220patients(28.8%)gavebirthbefore37wks,atotalof 142patientshadanobstetrichistoryofPBorlatemiscarriage.Among them,122women(86%)hadashortcervix.Acervixlength<25mm wasobservedin622patientswithoutadverseeventhistory.Basedon moleculardefinitionofBV,70(9.2%)patientswereBVcarriers.

Amongwomenwhodeliveredprematurely,24(10.9%)hadBV versus46(8.5%)inthegroupwherewomendeliveredatterm. Predictivemodelofpretermbirthriskinhigh-riskpopulation

Theresultsoftheunivariateandmultivariateanalysisarelisted inTable1.Amultivariatelogisticregressionanalysisdemonstrated significantandindependentassociationsbetweendeliverybefore

Table1

UnivariateandmultivariateanalysisofPretermbirthriskfactors.OR,oddsratio; aOR,adjustedOR;CI,confidenceinterval;PB,pretermbirth;PROM,premature ruptureofmembranes.

Variables OR[CI95%] P aOR[CI95%]P HistoryofPB 1.5[1–2] 0.05*

3.6[1.2–11]0.019*

Historyoflatemiscarriage 1.2[0.7–1.9] 0.5 1.6[0.5–5] 0.41 Gestationalage 0.9[0.9–0.96]<.001*0.9[0.8–0.9] 0.002*

Bacterialvaginosis 1.3[0.8–2.2] 0.3 1.6[0.6–4] 0.33 Smokingduringpregnancy 1.2[0.8–1.7] 0.3 1.5[0.8–2.9] 0.23 Sonographiccervicallength 0.9[0.9–0.96]<.001*

0.9[0.8–0.9] 0.001*

HistoryofPB+cervicallength<25mm1.9[1.2–2.8] 0.002*

0.4[0.1–1.4] 0.16 Daytraveltime>30min 1.4[1.00–1.9]0.05*

2.6[1.4–4.9]0.002*

PROM 2.3[1.3–4.1] 0.002*3.6[1.7–7.5]0.001*

*

Significantvariables. M.Gioanetal./JGynecolObstetHumReprod47(2018)545–548 546

(4)

37wksandsonographiccervicallength,historyofpretermbirth, early gestational age at inclusion, premature rupture of mem-branes(PROM),andsuperiordailytraveltimeof30min.Historyof late miscarriage, combination of history of preterm birth and cervicallength<25mm,smokingduringpregnancyandmolecular

diagnosisofBVwerenotassociatedwithPBinourcohort,although these variables were significantly associated with a PB in a univariate analysis in the preliminary study. The literature recognizesthesevariablesasriskfactorsforPB.Wehavechosen to integrate these elements into the calculator. The AUC after bootstrap for the validation set was 0.77 (95% CI; 0.72–0.81) (Fig.1).

We used500 bootstrap resamples for internal validation of accuracyestimatesandtoreduceoverfitbias.Themeanabsolute error was 0.018. The maximum error was 0.0682. The mean squarederrorwas0.00051.Itrepresentsthequalityofmeasureof anestimator,asaresultclosetozeroshowsgoodcalibration.We have therefore developed a nomogram to predict individual pretermbirthriskinasymptomatichigh-riskpopulation(Fig.2). The optimal predictive value of the PB risk in high-risk populationshasbeendeterminedfromtheROCcurve.Theoptimal probability thresholdis 23.2% which offers differentthresholds dependingonstatisticalvariables(Table2).Prioritizingsensitivity todeterminethethresholdtouseseemsessential.Itsuseinclinical practicewillallowminimisingtherateoffalse-negativeresults. Discussion

The present study developed a new predictive model for assessingtheindividualriskofPBinahigh-riskpopulation.Our nomogramevaluatesaprobabilityscorebasedonthemainwell known riskfactorsofspontaneousPBandincludesnewlyother risk factors. In our study according to literature [18], cervical lengthwasoneofthestrongestriskfactorforPBwiththehistoryof

Fig.1.ReceiverOperatingCharacteristiccurve:predictivemodeldiscrimination.

Fig.2.Nomogram.

Table2

Predictionoftheriskofindividualpretermbirthinhigh-riskpopulation.Se,sensitivity;Sp,specificity;LHR+,positivelikelihoodratio;LHR ,negativelikelihoodratio;PPV, positivepredictivevalue;NPV,negativepredictivevalue;95%CI,95%confidenceinterval.

PBrate(%) [95%CI] 8 15 23,2 30 55 74 Se 100(95.4–100) 82(71.7–89.8) 74(63.2–83.6) 56(44.7–67.6) 23(14.3–34) 6.4(2–14) Sp 0(0.0–1.4) 52.3(46–58.4) 72.7(67–78) 80.7(75.4–85.3) 97.3(94.6–99) 100(98.6–100) LHR+ 1(1.0–1.0) 1.7(1.5–2) 2.7(2.2–3.5) 2.92(2.1–4.0) 8.7(3.8–20.1) NA LHR NA 0.34(0.2–0.6) 0.35(0.2–0.5) 0.54(0.4–0.7) 0.79(0.7–0.9) 0.94(0.9–1.0) PPV 22.8(18.7–27.5) 33.5(27.2–40.5) 44.6(36.3–53.12) 45.8(36.2–55.8) 72(52.4–85.7) 100(56.5–100) NPV NA 90.7(85–94.4) 90.6(85.8–93.8) 86.2(81.31–89.9) 81(76.4–85) 78.3(73.6–82.4)

(5)

previousPB,PROMandlongdailytraveltime.Someofthelatest variableshaveneverbeenusedinanomogrambefore.Forexample ourresultsshowanindependentassociationbetweendailytravel timeand therisk ofPB and wasincludedin themodel.Active smokingwasalsoincludedinthemodel,evenifnon-significantin ourresultsbecauseofitspreviouslywasreportedanassociation withPB(OR:1.27,95%CI:1.21–1.33)[19].PreviousPBwasastrong and independent risk factor for PB in our study as previously reportedin ameta-analysiswithanabsoluterecurrence riskof 20.2% (95% CI: 19.9–20.6) [2]. PROM was an independent and strongriskfactorforPBinourstudyandbuthadpreviouslybeen includedinarisk-calculatingnomogram[11,12].

Inourmodelbacterialvaginosiswasincludedevenifitdidnot increasetheriskforPBaftermultivariateanalysis.Theassociation betweenBVandPBhasbeenpreviouslyreported,withashorter time to delivery, reason why it was included in the model. A pregnantwoman carryingBV hastwicethe risk of givingbirth prematurely (OR: 2.16, 95% CI: 1.56–3.00) [20], especially if screening was performed early during pregnancy [6]. In this situationtheriskoflatemiscarriageissixtimeshigher(OR:6.32; 95%CI:3.65–10.94)[20].Neverthelessthescreeningandtreatment ofasymptomaticBVinlow-riskpatientsarenotrecommendedand remain controversial for high-risk patients [21,22]. Because of heterogeneousBVdefinition,treatment,gestationalageat screen-ing.While waitingforthe resultsofan on-goingstudy[23],BV screeningandtreatmentarerecommendedforthesub-population ofasymptomatic patientswith a historyof adverseevents in a mother–foetalinfectiouscontext[24].

Areliablescreeningtoolshouldhelpforoptimalmanagement.The betteridentificationofhigh-riskPBpatientsbyusingapredictivetool couldlimitmedicalcosts.Thiscouldhelpalsotodecidethetimeof corticosteroidsinordertoavoidtooearlyorrepeatedtreatmentsthat couldbeinefficientordangerousforchildren[25].

Ourpredictivetoolhasbeencreatedbasedonamulticentricstudy performedwithalargeFrenchcohortofpatients.Itcouldthereforebe reproducibleinhighriskpopulation.Untilnow,theonly recommen-ded monitoring tool was the length of the cervix or the foetal fibronectinbetween16and22wks[26,27].Ourmodelproposesto personalisethecalculationofPBriskwithadditionalriskfactors.

Our study has some limits. To begin the use of molecular biologycannotbesetupeverywhereatthemoment.Ourmodel hasnotbeenvalidatedonanindependentpopulation. Wehave usedthere-samplingbybootstraptechniquetocounterthelackof external validity. The internal validity of our nomogram is improvedbyahighnumberofrepetitions.Unfortunately,atool isnotapplicabletopatientsatriskofpretermdeliveryformultiple pregnancyoruterinemalformation.Thelow-riskpopulationand inducedlabourarealsonotconcerned.Otherriskfactorsasthe numberofpreviousPBs,particularlyincaseofsuccessiveevents, andthegestationalofdeliveryofthepreviousPBcouldhavebeen alsoincludedinordertoimproveaccuracyofthenomogram[28]. Conclusion

Aninnovativecalculatortoolwasdevelopedinordertobetter define the individual risk of PB in high risk pregnancies. The accuracyofourpredictivetoolfortheriskofPBshouldbefurther evaluatedinaprospectivestudy.

References

[1]GoldenbergRL,CulhaneJF,IamsJD,RomeroR.Epidemiologyandcausesof pretermbirth.Lancet2008;371:75–84.

[2]KazemierB,BuijsP,MigniniL,LimpensJ,deGrootC,MolB,etal.Impactof obstetrichistoryontheriskofspontaneouspretermbirthinsingletonand multiple pregnancies: a systematic review.BJOG Int J Obstet Gynaecol 2014;121:1197–208.

[3]DelnordM,MortensenL,Hindori-MohangooAD,BlondelB,GisslerM,Kramer MR,etal.Euro-PeristatScientificCommittee.Internationalvariationsinthe gestationalagedistributionofbirths:anecologicalstudyin34high-income countries.EurJPublicHealth2017.

[4]HauteAutorite´ deSante´.Mesuredelalongueurducanalcervicalducolde l’ute´ruspare´chographieparvoievaginale.Inte´reˆtdanslepronostic d’ac-couchementpre´mature´.Cadrage.E´valuationdesactesprofessionnels.Paris: HAS;2009.

[5]MenardJP,Bretelle F. Vaginosebacte´rienne etaccouchement pre´mature´. Gyne´colObste´trFertil2012;40:48–54.

[6]LeitichH,Bodner-AdlerB, BrunbauerM,KaiderA,Egarter C,HussleinP. Bacterialvaginosis asariskfactorforpretermdelivery:ameta-analysis. AmJObstetGynecol2003;189:139–47.

[7]MenardJ-P,MazouniC,FenollarF,RaoultD,BoubliL,BretelleF.Diagnostic accuracyofquantitativereal-timePCRassayversusclinicalandGramstain identification of bacterial vaginosis. Eur J Clin Microbiol Infect Dis 2010;29:1547–52.

[8]MenardJ,FenollarF,HenryM,BretelleF,RaoultD.Molecularquantificationof GardnerellavaginalisandAtopobiumvaginaeloadstopredictbacterial vagino-sis.ClinInfectDis2008;47:33–43.

[9]MenardJP,MazouniC,Salem-CherifI,FenollarF,RaoultD,BoubliL,etal.High vaginal concentrations ofAtopobium vaginae andGardnerella vaginalis in womenundergoingpretermlabor.ObstetGynecol2010;115:134–40.

[10]CompanC,RossiA,Piquier-PerretG,DelabaereA,VendittelliF,LemeryD,etal. Pre´dictiondelapre´maturite´ encasdemenaced’accouchementpre´mature´ : revuedelalitte´rature.JGyne´colObste´trBiolReprod2015;44:740–51.

[11]Mailath-PokornyM,PolterauerS,KohlM,KueronyaiV,WordaK,HeinzeG, etal.Individualizedassessmentofpretermbirthriskusingtwomodified predictionmodels.EurJObstetGynecolReprodBiol2015;186:42–8.

[12]AlloucheM,HuissoudC,Guyard-BoileauB,RouzierR,ParantO.Development andvalidationofnomogramsforpredictingpretermdelivery.AmJObstet Gynecol2011;204.242.e1–8.

[13]BretelleF,RozenbergP,PascalA,FavreR,BohecC,LoundouA,etal.High AtopobiumvaginaeandGardnerellavaginalisvaginalloadsareassociatedwith pretermbirth.ClinInfectDis2015;60:860–7.

[14]MenardJ-P,FenollarF,RaoultD,BoubliL,BretelleF.Self-collectedvaginal swabs forthequantitative real-timepolymerasechain reaction assayof AtopobiumvaginaeandGardnerellavaginalisandthediagnosisofbacterial vaginosis.EurJClinMicrobiolInfectDis2012;31:513–8.

[15]R:TheRProjectforStatisticalComputing;n.d.https://www.r-project.org/

[accessed19.02.17].

[16]Swets JA. Measuring the accuracy of diagnostic systems. Science 1988;240:1285.

[17]DelacourH,ServonnetA,PerrotA,VigezziJF,RamirezJM.LacourbeROC (receiveroperatingcharacteristic):principesetprincipalesapplicationsen biologieclinique.AnnBiolClin(Paris)2005;63:145–54.

[18]CraneJMG,HutchensD.Transvaginalsonographicmeasurementofcervical lengthtopredictpretermbirthinasymptomaticwomenatincreasedrisk:a systematicreview.UltrasoundObstetGynecol2008;31:579–87.

[19]ShahNR,BrackenMB.Asystematicreviewandmeta-analysisofprospective studiesontheassociationbetweenmaternalcigarettesmokingandpreterm delivery.AmJObstetGynecol2000;182:465–72.http://dx.doi.org/10.1016/ S0002-9378(00)70240-7.

[20]LeitichH,KissH.Asymptomaticbacterialvaginosisandintermediatefloraas riskfactorsforadversepregnancyoutcome.BestPractResClinObstet Gynae-col2007;21:375–90.

[21]MoralesWJ,SchorrS,AlbrittonJ.Effectofmetronidazoleinpatientswith pretermbirthinprecedingpregnancyandbacterialvaginosis:a placebo-controlled,double-blindstudy.AmJObstetGynecol1994;171:345–7. discus-sion348-349.

[22]BrocklehurstP,GordonA,HeatleyE,MilanSJ.Antibioticsfortreatingbacterial vaginosisinpregnancy. In:TheCochraneCollaboration,editor.Cochrane DatabaseSyst.Rev..Chichester,UK:JohnWiley&Sons,Ltd;2013.

[23]BretelleF,FenollarF,BaumstarckK,FortanierC,CocallemenJF,SerazinV,etal. Screen-and-treatprogrambypoint-of-careofAtopobiumvaginaeand Gardne-rellavaginalisinpreventingpretermbirth(AuToptrial):studyprotocolfora randomizedcontrolledtrial.Trials2015;16:470.

[24]BrabantG.Vaginosebacte´rienneetpre´maturite´ spontane´e.Recommandations pourlapratiqueclinique:pre´ventiondelapre´maturite´ spontane´eetdeses conse´quences(horsrupturedesmembranes).JGyne´colObste´trBiolReprod 2016;45:1247–60.

[25]MurphyKE,HannahME,WillanAR,HewsonSA,OhlssonA,KellyEN,etal. Multiplecoursesofantenatalcorticosteroidsforpretermbirth(MACS):a randomisedcontrolledtrial.Lancet2008;372:2143–51.

[26]SentilhesL,Se´natM-V,AncelP-Y,AzriaE,BenoistG,BlancJ,etal. Recom-mandationspourlapratiqueclinique:pre´ventiondelapre´maturite´ sponta-ne´e et de ses conse´quences (hors rupture des membranes)—Texte des recommandations (texte court). J Gyne´col Obste´tr Biol Reprod 2016;45:1446–56.

[27]DeFrancoEA, LewisDF,Odibo AO.Improvingthescreeningaccuracy for pretermlabor:isthecombinationoffetalfibronectinandcervicallengthin symptomatic patientsa useful predictorof preterm birth?Asystematic review. AmJObstetGynecol2013;208.233.e1–6.

[28]BlancJ,BretelleF.Outilspre´dictifsdel’accouchementpre´mature´ dansune populationasymptomatiquea` hautrisque.JGyne´colObste´trBiolReprod 2016;45:1261–7.

M.Gioanetal./JGynecolObstetHumReprod47(2018)545–548 548

Figure

Fig. 2. Nomogram.

Références

Documents relatifs

Conditional recommendation (only in health-care facilities where intubation, ventilator care, blood gas analysis, newborn nursing care and monitoring are available) based

Je file vers Liam pour lui faire un gros câlin et des tas de bisous : -Tu seras sage, mon cœur, tu sais comme maman t’aime.. -Oui, tu m’aimes comme ca, dit il en faisant un cœur

This research utilises classification techniques and patient data consisting of known risk factors such as age, the number of pregnancies, STD’s, and smoking habits with the intent

In this European multi-cohort study including almost 200,000 pregnancies we found that compared to abstaining, drinking a maximum of three alcohol drinks per week during pregnancy

Data from the Timoun Mother – Child Cohort Study conducted in Guadeloupe between 2004 and 2007 were used to examine the associations of chlordecone concentrations in maternal

This observation combined with specific social risk factors (older maternal age, single living) less frequent on the French mainland probably explains a large

Troxel The objective of this project is the development of a Process Flow Representation (PFR) for the inte- gration of technology (process and device) CAD (TCAD)

Given the proofs of correctness of our mediated reordering channel and our key-value store implemented on top of reliable channels, we invoke our composition theorem to produce