Cephalad-caudad versus transverse blunt expansion of the low transverse uterine incision during cesarean delivery

Review

Cephalad–caudad

versus

transverse

blunt

expansion

of

the

low

transverse

uterine

incision

during

cesarean

delivery

Serena

Xodo

,

Gabriele

Saccone

,

Antonella

Cromi

,

Pinar

Ozcan

,

Emanuela

Spagnolo

,

Vincenzo

Berghella

_*

a

DepartmentofGynaecologyandObstetrics,SchoolofMedicine,UniversityofUdine,Udine,Italy

b

DepartmentofNeuroscience,ReproductiveSciencesandDentistry,SchoolofMedicine,UniversityofNaplesFedericoII,Naples,Italy

c

DepartmentofObstetricsandGynecology,UniversityofInsubria,Varese,Italy

d_{DepartmentofGynaecologyandObstetrics,BezmialemVakifUniversityHospital,Istanbul,Turkey} e

DepartmentofPerinatalMedicine,PoliclinicoAbanoTerme,Padova,Italy

f

DivisionofMaternal-FetalMedicine,DepartmentofObstetricsandGynecology,SidneyKimmelMedicalCollegeofThomasJeffersonUniversity, Philadelphia,PA,USA

Contents

Introduction... 76

Materialsandmethods... 76

Results... 77 Comment... 77 Funding ... 79 References... 79 ARTICLE INFO Articlehistory: Received11March2016

Receivedinrevisedform9April2016 Accepted23April2016 Keywords: Cesareansection Review Postpartumhemorrhage Metaanalysis Surgery ABSTRACT

Itisimperativetohaveevidence-basedguidelinesforcesareandelivery.Theaimofthismeta-analysis

wastoevaluatetheeffectivenessofacephalad–caudadcomparedtotransversebluntexpansionofthe

uterine incisionto reduceblood lossinwomenwhounderwentlow-segment transversecesarean

delivery.Wethereforeperformedasystematicsearchinelectronicdatabasesfromtheirinceptionuntil

March2016.Weincludedallrandomizedtrialscomparingcephalad–caudadversustransverse(control

group)bluntexpansionoftheuterineincisioninwomenwhounderwentalowtransversecesarean

delivery.Theprimaryoutcomewaspostpartumbloodloss,definedasthemeanamountofbloodloss

(mL).Twotrials(921women)wereanalyzed.Afterthetransverseuterineincisionintheloweruterine

segmentwiththescalpel,theuterineincisionwasthenbluntlyexpandedbythedesignatedmethod.

Bluntexpansionoftheprimaryincisionwas derivedbyplacing theindexfingersoftheoperating

surgeon into the incision and pulling the fingers apart laterally (transverse group) or cephalad

(cephalad–caudadgroup).Womenwhowererandomizedinthecephalad–caudadgrouphadlower:

meanofpostpartumbloodloss,hemoglobindropandhematocritdrop24haftercesarean,unintended

extension,uterinevesselsinjury,bloodloss>1500mLandneedforadditionalstitches.Therewasno

statisticallysignificantdifferenceintheincidenceofbloodloss>1000mL,intheoperatingtimeandin

post-operativepain.Inconclusion,expansionoftheuterineincisionwithfingersinacephalad–caudad

directionisassociatedwithbettermaternaloutcomesand,therefore,shouldbepreferredtotransverse

expansionduringacesareandelivery.

ß 2016ElsevierIrelandLtd.Allrightsreserved.

Abbreviations:CD,cesareandelivery;RCTs,randomizedclinicaltrials;Hgb,hemoglobin;Hct,hematocrit;RR,relativerisk;MD,meandifference;CI,confidenceinterval. * Correspondingauthorat:DivisionofMaternal-FetalMedicine,DepartmentofObstetricsandGynecology,ThomasJeffersonUniversity,833ChestnutStreet,FirstFloor, Philadelphia,PA19107,USA.Tel.:+12159557996;fax:+12159555041.

E-mailaddress:vincenzo.berghella@jefferson.edu(V.Berghella).

ContentslistsavailableatScienceDirect

European

Journal

of

Obstetrics

&

Gynecology

and

Reproductive

Biology

j o urn a l hom e pa ge : ww w. e l s e v i e r. c om/ l o ca t e / e j ogr b

http://dx.doi.org/10.1016/j.ejogrb.2016.04.035

Introduction

Cesarean delivery (CD)is one of themostcommon surgical procedures performed in the Western world and rates are increasingdespiteeffortstothecontrary[1].Itisimperative to haveevidence-based guidelinesfor each surgical step,in order tominimizemorbidityandmortality[2,3].

The most common complication of CD is hemorrhage [2]. Researchershaveidentifiedthefollowingsurgicalstepsascrucial momentsforreducingbloodlossduringtheoperativeabdominal delivery:useofuterotonics,spontaneousplacentalremovaland bluntexpansionoftheuterineincisionwithfingers,ratherthan scissors[3].Indeed,comparedtosharputerineincisionexpansion, bluntexpansion is associated with less unintended extensions andfavorablematernaloutcomes[3].However,whethertheblunt expansion of uterine incision should be performed cephalad– caudallyortransverselyisstillamatterofdebate.

The aimofthis studywastoevaluatetheeffectivenessof a cephalad–caudad compared to transverse blunt expansion to reducebloodlossinwomenundergoingalow-segmenttransverse CD.

Materialsandmethods

Thisreviewwasperformedaccordingtoaprotocoldesigned aprioriand recommendedforsystematicreview [4].Electronic databases (i.e. MEDLINE, PROSPERO, Scopus, ClinicalTrials.gov, EMBASE, Sciencedirect, the Cochrane Library, Scielo) were searched from their inception until March 2016. Search terms used were the following text words: ‘‘cesarean,’’ ‘‘caesarean’’, ‘‘cephalad–caudad blunt expansion’’, ‘‘transverse blunt expan-sion’’,‘‘expansion of uterine incision’’, ‘‘obstetrichaemorrhage’’, ‘‘randomized,’’ ‘‘randomized controlled trial’’ and ‘‘randomized clinicaltrial.’’Norestrictionsforlanguageorgeographiclocation wereapplied.Inaddition,thereferencelistsofallidentifiedarticles were examined to identify studies not captured by electronic searches.Theelectronicsearchandtheeligibility ofthestudies wereindependentlyassessedbytwoauthors(SX,VB).Differences werediscussedandconsensusreached.

We included allrandomizedclinical trials(RCTs) comparing cephalad–caudad(i.e.interventiongroup)versustransverse(i.e. controlgroup)bluntexpansioninwomenwhounderwenta low-segment transverse CD. Selection included women undergoing

alow-segmenttransverseCDafter30weeksofgestation,either planned or urgent.Quasi-randomized trials (i.e.trials in which allocationwasdone onthebasisofa pseudo-randomsequence, e.g.odd/evenhospitalnumberordateofbirth,alternation)were excluded.

After the transverse uterine incision in the lower uterine segmentwiththescalpel,theuterine incisionwasthenbluntly expanded by the designated method. Blunt expansion of the primaryincisionwasderivedbyplacingtheindexfingersofthe operatingsurgeonintotheincisionandpullingthefingersapart laterally(i.e.transversegroup)orcephalad–caudad(i.e.cephalad– caudad group). Women in thetransverse expansion grouphad theuterineincisionextendedbytheinsertionofbothindexfingers oftheoperatorintotheopeningwhothenpulledthefingerapart laterally.In thecephalad–caudadexpansion group, atransverse openingoftheloweruterinesegmentwascreatedbyseparation ofthefingersofthesurgeoninacephalad–caudaddirectionalong themidline(Fig.1).

Theriskofbiasineachincludedstudywasassessedbyusing the criteria outlined in the Cochrane Handbook for Systematic Reviews ofInterventions.Sevendomains relatedto riskof bias wereassessedineachincludedtrialsincethereisevidencethat these issues are associated withbiased estimates of treatment effect: (1) randomsequence generation; (2) allocation conceal-ment;(3)blindingofparticipantsandpersonnel;(4)blindingof outcomeassessment;(5)incompleteoutcomedata;(6)selective reporting; and (7) other bias.Review authors’judgments were categorizedas‘‘lowrisk’’,‘‘highrisk’’or‘‘unclearrisk’’ofbias[4]. Twoauthors(SX,GS)independentlyassessedinclusioncriteria, riskofbiasanddataextraction.Disagreementswereresolvedby consensus with a third reviewer (VB). Data from each eligible studywereextracted withoutmodificationoforiginal dataonto custom-made data collection forms. Differences were reviewed, andfurtherresolvedbycommonreviewoftheentireprocess.Data not presented in the original publicationswere requested from theprincipalinvestigators.

Allanalysesweredoneusinganintention-to-treatapproach, evaluating women accordingto the treatment group towhich theywererandomlyallocatedintheoriginaltrials.Primaryand secondary outcomes were defined before data extraction. The primaryoutcomewaspostpartumbloodloss,definedasthemean amount of blood loss (mL)in case of CD. Secondary outcomes includedincidenceofunintendedextension,uterinevesselsinjury,

need for additional stitches, blood loss >1000 and >1500mL, hemoglobin(Hgb)andhematocrit(Hct)drop24hafterCD,mean ofoperatingtime(min)andpost-operativepain.

Thedataanalysiswascompletedindependentlybytwoauthors (SX,GS)usingReviewManagerv.5.3(TheNordicCochraneCentre, CochraneCollaboration,2014,Copenhagen,Denmark).The com-pleted analyses were then compared, and any difference was resolved by consensus with a third reviewer (VB). Statistical heterogeneityacrossstudieswasassessedusingtheHigginsI2_test.

In case of statistically significant heterogeneity (I2_{>0%) the}

randomeffects modelof DerSimonianand Lairdwasmanaged; otherwise,incaseofnoinconsistencyintheriskestimates(I2_=0),

afixedeffectmodelwasperformed[4].Thesummarymeasures werereportedasmeandifference(MD)orasrelativerisk(RR)with 95%ofconfidenceinterval(CI).Potentialpublicationbiaseswere statisticallyassessed byusing Begg’sand Egger’s tests. pValue <0.05wasconsideredstatisticallysignificant.

The meta-analysis was reported following the Preferred Reporting Item for Systematic Reviews and Meta-analyses (PRISMA)statement[5].Beforedataextraction,thereview was registeredwiththePROSPEROInternationalProspectiveRegister of Systematic Reviews (registration No.: CRD42015027791) followingthePRIMAguidelinesforprotocols(PRIMSA-P)[6].

Results

TheflowofstudyidentificationisshowninFig.2.ThreeRCTs wereassessedforeligibility[7–9].Onewasexcludedsinceblunt versussharputerineincisionexpansion wasevaluated[9].Two RCTs, including 921 women, were analyzed [7,8]. Of the 921 included women, 459 (50%) were randomized in the cephalad–caudadgroup,while462(50%)inthetransversegroup (i.e.controlgroup).Publicationbias,assessedstatisticallybyusing Begg’sandEgger’stests,showednosignificantbias(p=0.21and p=0.34,respectively).Thequalityofthestudiesincludedinour meta-analysiswasassessedbyusingthecriteriaoutlinedinthe Cochrane Handbook for Systematic Reviews of Interventions (Fig. 3) [4]. All the included studies had low risk of bias in ‘‘randomsequencegeneration.’’Adequatemethodsforallocation ofwomenwereusedinbothRCTs.Blindingwasconsiderednot feasible methodologicallygiven the intervention,and therefore noneoftheincludedstudieswasdoubleblind.Bothauthors(AC, PO)oftheoriginalRCTskindlyprovidedadditionalunpublished data.

Table1showsthecharacteristicsofthetwoincludedstudies andofthe921includedwomen.Nodifferenceswerefoundinthe maternal characteristicsbetween thetwo groups. Both studies camefromEurope.InthesetwoRCTswomenwhounderwenta low-segmenttransverseCDwereassignedrandomlytohavethe bluntexpansionoftheuterineincisionbythephysicianseparating thefingerseitherinatransversedirectionorinacephalad–caudad direction.IntheItalianRCTtheprimaryoutcomewastheincidence of unintended extensions [7], while Ozcan et al. did not pre-specifiedtheprimaryoutcome[8].Alloperationswereperformed under spinal anesthesia and skin incisions were made with a classical Pfannenstiel incision. The two studies had different inclusioncriteria:Cromietal.enrolledallwomenwhounderwent alow-segmenttransverseCDafter30weeksofgestation,either plannedorurgent[7],whileOzcanetal.excludedwomenwith severemedicalconditions [8].One RCTincludedonlysingleton gestations[8].IntheItaliantrialbloodlosswasestimatedfromthe bloodthathadbeencollectedinthesuctiondevice,intheplastic pouchesofsteriledrapes,andinthesaturationofpadsandsponges

[7].InthestudybyOzcanetal.bloodlosswasestimatedusingthe weightofcompressesusingduringtheCD(theincreaseinweight

of compresses with absorbed blood) and the number of intraoperativecompresses[8].

Table 2 shows pooledresults of theprimary and secondary outcomes.The statistical heterogeneityacrossstudieswas low, withno inconsistency(I2=0)in risk estimates for theprimary outcomeandformostofthesecondaryoutcomes.Womenwho wererandomizedinthecephalad–caudadgrouphadlower:mean of postpartum blood loss (MD 67.64mL, 95% CI 102.85 to 32.43;Fig.4),Hgbdrop(MD 0.26g/dL,95%CI 0.37to 0.14) andHctdrop24hafterCD(MD 1.20g/dL,95%CI 1.87to 0.53), unintendedextension(4.8%vs.8.9%;RR0.51,95%CI0.30–0.88), uterinevesselsinjury(1.5%vs.2.8%;RR0.52,95%CI0.20–0.84), blood loss>1500mL(0.2%vs.1.7%;RR 0.12,95% CI0.02–0.99) andneedforadditionalstitches(20.3%vs.29.2%;RR0.60,95%CI 0.44–0.82).Therewasnostatisticallysignificantdifferenceinthe incidenceofbloodloss>1000mL(1.2%vs.3.0%;RR0.41,95%CI 0.14–1.18),intheoperatingtime(MD1.36min,95%CI 0.17to 2.89)and inpost-operativepain( 0.50points,95%CI 1.17to 0.17)(Table2).

Comment

Thismeta-analysisfromthetwohigh-qualitylowriskofbias RCTs,including921womenundergoingalow-segmenttransverse Fig.2.Flowdiagramofstudiesidentifiedinthesystematicreview.(Prismatemplate [Preferred Reporting Item for SystematicReviews and Meta-analyses]).RCTs, randomizedcontrolledtrials.

CD, showed that cephalad–caudad blunt expansion of the low transverse uterine incision decreased blood loss compared to transversebluntexpansion.Itwasalsoassociatedwithlowerdrop of Hgb and Hct 24h after CD and lower risk of unintended extension,uterinevesselsinjuryandneedforadditionalstitches. The comparison of blunt and sharp expansion of the low transverseuterineincisionhascommonlybeendiscussedinthe literature[9–17]. Results from a Cochrane review suggest that blunt expansions should be preferred over sharp expansions becausebluntexpansionsappeartobeassociatedwithlowerrisk

of blood loss, shorter operative time, and less unintended extension[17].Itremained,however,unclearwhichtypeofblunt technique–cephalocaudalversustransverse–shouldbeused.To our knowledge, ours is the first meta-analysis evaluating the effectivenessof cephalad–caudadcompared totransverseblunt expansiontoreducebloodlossinwomenwhounderwentalow transverseCD.

Ourstudyhasseveralstrengths.Thetwoincludedstudieshada low risk of allocation bias by Cochrane Collaboration tool assessment.Intent-to-treatanalysiswasused,and bothrandom Fig.3.Assessmentofriskofbias.(A)Summaryofriskofbiasforeachtrial;plussign:lowriskofbias;minussign:highriskofbias;questionmark:unclearriskofbias.(B)Risk ofbiasgraphabouteachriskofbiasitempresentedaspercentagesacrossallincludedstudies.

Table1

Characteristicsoftheincludedtrialsandoftheincludedwomen.

Cromi2008[7] Ozcan2015[8]

Studylocation Italy Turkey

Numberofpatients 811(405vs.406) 110(54vs.56) Inclusioncriteria Womenwhounderwentalowtransverse

cesareandeliveryafter30weeksofgestation, eitherplannedorurgent

Singletonswithoutseveremedicalconditionsb whounderwentalowtransversecesareandelivery

Maternalage 32.64.9vs.32.74.8 30.44.6vs.29.75.6 Parity 0.40.6vs.0.50.7b 1.31.4vs.1.21.0 Nulliparous 344(84.9%)vs.351(86.5%)a 3(5.6%)vs.6(10.7%) Primigravida 240(59.3%)vs.261(64.3%)a N/R PreviousCD 104(25.7%)vs.90(22.2%) 42(77.8%)vs.46(82.1%)a BMI 26.74.0vs.27.34.2 28.12.3vs.28.71.8 GAatbirth(weeks) 38.32.4vs.38.52.6 38.51.1vs.38.71.1 Laborstage Notinlabor 274(67.7%)vs.296(72.9%) 49(90.7%)vs.52(92.9%)a Firststage 64(15.8%)vs.46(11.3%) 5(9.3%)vs.4(7.1%)a Secondstage 67(16.5%)vs.64(15.8%) 0vs.0a

Indicationforcesarean

Priorcesarean 101(24.9%)vs.89(21.9%) 42(77.8%)vs.46(82.1%)a Dystocia 88(21.7%)vs.87(21.4%) 5(9.3%)vs.4(7.1%)a Fetaldistress 87(21.5%)vs.91(22.4%) 0vs.0a

Malpresentation 57(14.1%)vs.62(15.3%) 7(13.0%)vs.6(10.7%)a

Other 72(17.8%)vs.77(18.9%) 0vs.0a

Fetuses’birthweight 3112588vs.3150554 3328517vs.3470518 Fetuseswithmacrosomia 17(4.2%)vs.15(3.7%) 0vs.0a

Primaryoutcome Incidenceofunintendedextensions N/R

Dataarepresentedastotalnumber(numberinthecephalad–caudadgroupvsnumberinthetransversegroup)withpercentageorasmeanstandarddeviation.Nulliparous, nopreviousvaginaldelivery;Primigravida,nopreviouspregnancy;BMI,bodymassindex;GA,gestationalage;CD,cesareandelivery;N/R,datanotrecorded.

a

Additionalunpublisheddatakindlyobtainedbytheauthorsoftheoriginaltrials.

b

Severemedicalconditions:diabetesmellitus,moderate-severehypertension,anybloodorthrombophiliadisorders,presenceofuterineoverdistension(multiple pregnancies,suspectedmacrosomia,polyhydramnios),emergencysurgery(placentaabruption,placentaprevia),anti-coagulationtherapyorahistoryofothermajor abdominalsurgeries.

andmixedeffectsmodelswereusedwhenappropriate.Inaddition, publicationbiaswasnotapparentbystatisticalanalysis. Hetero-geneity between studies was variable but generally was not significant.Thesearekey elementsthat areneededtoevaluate thereliabilityofa meta-analysis[4]. Wepooled datafromdue different RCTs in order to draw conclusions to this clinical dilemma.

Limitationsofourstudyaremostlyinherenttothelimitations oftheincludedstudies.Thetwotrialsincludedhad somewhat differentinclusion criteria.Giventheintervention,noneofthe includedRCTswasdouble-blind.Theoutcomeparameterswere notmeasuredwithavalidscore.Thelownumberoftheincluded trialsandtheheterogeneityintermsofrateofpriorCDareother shortcomingsofthismeta-analysis.Anothersignificantconcern wasthepre-specifiedprimaryoutcome(i.e.estimatedbloodloss), whichis difficultto measure accuratelyandconsistently, and wheretheassessorisnotblindedtotheallocation.

Thereareclinicalscenarioswhichwerenotaddressedinthe includedRCTs.Theseinclude, asexamples, a smalllow uterine segment(e.g.inpretermgestations),thepresenceofmyomataat thelevel of theuterine incision,placenta previaor accreta,or uterinedehiscence.Ingeneral,wewouldspeculatethatcephalad– caudadexpansion wouldstill befeasibleand preferredin most (if not all) of these scenarios, but more data is needed for a definiteanswer.Thelong-termrisksofpreviaoraccretaorother complicationsinasubsequentpregnancyaccordingtothetypeof uterineexpansionwerealsonotreportedintheincludedRCTs.

Theloweruterinesegmentconsistsofcircularandtransversely runningmuscularbundles.Thisanatomicalfeatureexplainswhy the uterine incision can be easy widened. When the surgeon expands laterally the uterine incision, he actually applies a separating force to the cleavage planes between the muscular bundles.Inordertoavoidtheconsequencederivingfromthelack

of control of magnitude’s force and the following unintended extensionwithpossibleuterinevesselinjury,thecephalad–caudad expansion has been proposed. Our meta-analysis showed that cephalad–caudadbluntexpansionof thelow transverseuterine incision decreased blood loss compared to transverse blunt expansion. The biological plausibility to explainour findings is not completely clear. However, thesupposed advantage of the cephalad–caudadapproachmightbethecontroloftheexpansion, thusavoidingthedamageoflateraluterineandparametrialblood vessels.

Insummary,expansionoftheuterineincisioninacephalad– caudad direction is associated with lower risks of postpartum bloodloss,unintendedextension,uterinevesselsinjuryandneed for additional stitches, and should therefore be preferred to transverseexpansionwhenaCDisperformed.Furtherstudiesare required concerning subsequent pregnancies, e.g. risk of scar dehiscenceduringavaginalbirthaftercesarean,riskofplacenta previa/accretainthenextpregnancy.

Funding

Nofinancialsupportwasreceivedforthisstudy. References

[1]MartinJA,HamiltonBE,OstermanMJ,CurtinSC,MatthewsTJ.Births:finaldata for2013.NatlVitalStatRep2015;64:1–68.

[2]BerghellaV,BaxterJK,ChauhanSP.Evidence-basedlaboranddelivery man-agement.AmJObstetGynecol2008;199:445–54.

[3]Berghella V,BaxterJK,ChauhanSP. Evidence-basedsurgeryforcesarean delivery.AmJObstetGynecol2005;193:1607–17.

[4]HigginsJPT,GreenS.Cochranehandbookforsystematicreviewsof interven-tions,version5.1.0(updateMarch2011).TheCochraneCollaboration;2011, Availableat:www.cochrane-handbook.org[accessed20.02.15].

Fig.4.Forestplotofthemeanofestimatedbloodloss(mL)incaseofcesareandelivery.IV,independentvariable;CI,confidenceinterval. Table2

Primaryandsecondaryoutcomes.

Cromi2008[7] Ozcan2015[8] Total I2

(%) RRorMD(95%CI) Estimatedbloodloss(mL)b

398242vs.440341 407196vs.551179a – 0 S_{67.64mL(S102.85to}S_32.43) Unintendedextension 15/405(3.7%)vs.30/406(7.4%) 7/54(12.9%)vs.11/56 (19.6%) 22/459(4.8%)vs.41/462 (8.9%) 0 0.51(0.30–0.88)

Uterinevesselsinjury 0/405vs.2/406(0.5%) 7/54(12.9%)vs.11/56 (19.6%)

7/459(1.5%)vs.13/462 (2.8%)

0 0.52(0.20–0.84)

Needforadditionalstitches 93/405(23.0%)vs.135/406(33.3%) 0/54vs.0/56 93/459(20.3%)vs.135/462 (29.2%) 0 0.60(0.44–0.82) Bloodloss>1000mL 5/405(1.2%)vs.12/406(3.0%)a _{N/R 5/405(1.2%)vs.12/406} (3.0%) 0 0.41(0.14–1.18) Bloodloss>1500mL 1/405(0.2%)vs.8/406(2.0%) 0/54vs.0/56 1/459(0.2%)vs.8/462 (1.7%) 0 0.12(0.02–0.99) Hbdrop24hafterCD(g/dL) 10.8vs.1.21.0 0.90.7vs.1.40.7 – 7 S0.26g/dL(S0.37toS0.14) Hctdrop24hafterCD(g/dL) N/R 2.91.8vs.4.11.8 – 0 S1.20g/dL(S1.87toS0.53) Operatingtime(min) 40.411.8vs.38.911.9 42.311.6vs.4212.1 – 0 1.36min(S0.17to2.89) Post-operativepain N/R 4.61.8vs.5.11.8 – 0 S0.50points(S1.17to0.17) Boldfacedata,statisticallysignificant.Dataarepresentedastotalnumber(numberinthecephalad–caudadgroupvs.numberinthetransversegroup)withpercentageoras meanstandarddeviation.RR,relativerisk;MD,meandifference;CI,confidenceinterval;Hb,hemoglobin;Hct,hematocrit;post-operativepain,evaluatedbythefacespainrating scale24haftertheoperation;CD,cesareandelivery;h,hours;N/R,datanotrecorded.

a

Additionalunpublisheddatakindlyobtainedbytheauthorsoftheoriginaltrials.

b

[5]MoherD,LiberatiA,Tetzlaff J,AltmanDG.Preferredreportingitemsfor systematicreviewsandmeta-analyses:thePRISMAstatement.JClin Epide-miol2009;62:1006–12.

[6]ShamseerL,MoherD,ClarkeM,etal.Preferredreportingitemsforsystematic reviewandmeta-analysisprotocols(PRISMA-P)2015:elaborationand expla-nation.BMJ2015;349:g7647.

[7]CromiA,GhezziF,DiNaroE,SiestoG,LoverroG,BolisP.Bluntexpansionofthe lowtransverseuterineincisionatcesareandelivery:arandomizedcomparison of2techniques.AmJObstetGynecol2008;199:292.e1–292.e16.

[8]OzcanP,AtesS,GunerCanM,YardımcıAS,BatmazG,KilicG.Iscephalad–caudad bluntexpansionofthelowtransverseuterineincisionreallyassociatedwithless uncontrolledextensionstodecreaseintra-operativebloodloss?Aprospective randomised-controlledtrial.JMaternFetalNeonatalMed2015;25:1–5.

[9]RodriguezAI,PorterKB,O’BrienWF.Bluntversussharpexpansionofthe uterineincisioninlow-segmenttransversecesareansection.AmJObstet Gynecol1994;171:1022–5.

[10]MagannEF,ChauhanSP,BufkinL,FieldK,RobertsWE,MartinJrJN. Intra-operativehemorrhagebybluntversussharpexpansionoftheuterineincision atcesareandelivery:arandomizedclinicaltrial.BJOG2002;109:448–52.

[11]HidarS,JerbiM,HafsaA,SlamaA,BibiM,KhairiH.Theeffectofuterine incision expansion at cesarean delivery on perioperative hemorrhage:

aprospectiverandomizedclinicaltrial[inFrench].RevMedLiege2007;62: 235–238.

[12]SekhavatL,DehghaniFirouzabadiR,MojiriP.Effectofexpansiontechniqueof uterineincisiononmaternalbloodlossincesareansection.ArchGynecol Obstet2010;282:475–9.

[13]AsıcıogluO,GungordukK,AsıcıogluBB,YıldırımG,GungordukOC,ArkC. Unintendedextensionofthelower segmentuterine incisionat cesarean delivery;arandomizedcomparisonofsharpversusblunttechniques.AmJ Perinatol2014;31:837–44.

[14]XuLL,ChauAM,ZuschmannA.Bluntvs.sharputerineexpansionatlower segmentcesareansectiondelivery:asystematicreviewwithmetaanalysis. AmJObstetGynecol2013;208:62.e1–e8.

[15]SaadAF,RahmanM,CostantineMM,SaadeGR.Bluntversussharputerine incisionexpansionduringlowtransversecesareandelivery:ametaanalysis. AmJObstetGynecol2014;211:684–8.

[16]DahlkeJD,Mendez-FigueroaH,RouseDJ,BerghellaV,BaxterJK,ChauhanSP. Evidence-basedsurgeryforcesareandelivery:anupdatedsystematicreview. AmJObstetGynecol2013;209:294–306.

[17]DoddJM,AndersonER,GatesS,GrivellRM.Surgicaltechniquesforuterine incisionand uterine closure atthe time ofcaesarean section. Cochrane DatabaseSystRev2014;7.CD004732.