Research www.AJOG.org

OBSTETRICS

Umbilical venous flow rate in term fetuses: can variations inflow predict intrapartum compromise?Tomas Prior, BSc; Edward Mullins, BSc; Phillip Bennett, PhD; Sailesh Kumar, PhD

OBJECTIVE: The objective of the study was to investigate the distri- greater than the 80th centile), and the incidence of diagnoses of fetal

bution of umbilical venous flow rates, measured in early labor, in acohort of normal term pregnancies and to establish the relationshipbetween umbilical venous flow and subsequent intrapartum outcome.

STUDY DESIGN: Five hundred eighty-nine women with uncomplicated,term, singleton pregnancies were recruited to this prospectiveobservational study prior to active labor (dilation of 4 cm or less) atQueen Charlotte’s and Chelsea Hospital (London, UK). All participantsunderwent an ultrasound examination, during which fetal biometry,umbilical venous flow velocity, and umbilical vein diameter wererecorded. Umbilical venous flow rate was then calculated. Followingdelivery, intrapartum and neonatal outcomes were correlated with theultrasound findings. Cases were subdivided according to mode ofdelivery, and mean umbilical venous flow rates were compared be-tween the groups. Cases were also subdivided according to umbilicalvenous flow rate (less than the 20th centile, 20th-80th centile, and

From the Center for Fetal Care, Queen Charlotte’s and Chelsea Hospital, anReproductive and Developmental Biology, Imperial College London, Londonauthors), and Mater Research Institute/University of Queensland, South Bris(Dr Kumar).

Received June 4, 2013; revised Aug. 5, 2013; accepted Aug. 28, 2013.

T.P. was supported by Moonbeam Trust (Charity number 1110691). All auththe Imperial College Healthcare National Health Service Trust comprehensiveCentre scheme.

The authors report no conflict of interest.

Reprints: Sailesh Kumar, PhD, Mater Research Institute/University of QueensPlace, Raymond Terrace, SouthBrisbane, Queensland 4101, Australia. skuma

0002-9378/$36.00 � ª 2013 Mosby, Inc. All rights reserved. � http://dx.doi.org/10.1

compromise was compared.

RESULTS: Fetuses delivered by emergency cesarean for presumedfetal compromise had the lowest umbilical venous flow rates (bothcorrected for and uncorrected for birthweight) (P¼ .02 and P¼ .001,respectively). Fetuses with the lowest umbilical venous flow rates weresignificantly more likely to require emergency cesarean for presumedfetal compromise than those with the highest flow rates (15.7% vs5.6%, relative risk, 2.83; 95% confidence interval, 1.16e6.91).

CONCLUSION: Fetuses with the lowest umbilical venous flow rates areat increased risk of a subsequent diagnosis of intrapartum fetalcompromise. Measurement of umbilical venous flow could contributeto the risk stratification of pregnancies prior to labor.

Key words: fetal compromise, fetal Doppler, labor, umbilical venousflow

Cite this article as: Prior T, Mullins E, Bennett P, et al. Umbilical venous flow rate in term fetuses: can variations in flow predict intrapartum compromise? Am J ObstetGynecol 2013;209:xx-xx.

dequate umbilical venous blood

A flow delivering oxygen and nutri-ents is essential to support the fetuswithin the uterine environment. The rateof umbilical venous flow may be consid-ered a direct measure of the delivery ofnutrients to the fetus by the placenta, arelationship that is tested to the highestdegree in the intrapartum period becausefetoplacental blood flow is compromisedduring periods of uterine contraction.

Despite being first measured in the1980s,1 the clinical utility of umbilicalvenous flow measurement is yet to beestablished. Studies of umbilical venousflow have demonstrated reduced flowrates in growth-restricted fetuses.1,2 It issuggested that in cases of growth restric-tion, umbilical venous flow is reducedprior to changes in umbilical arteryresistance.3 Furthermore, abnormalitiesin umbilical venous flow have been

d the Institute for, England, UK (allbane, QLD, Australia

ors were supported byBiomedical Research

land, Level 3, Aubignyr@mmri.mater.org.au.

016/j.ajog.2013.08.042

MONTH 2013 Am

associated with an increased incidence ofoperative delivery for fetal compromise.4

The widespread use of electronic fetalheart rate monitoring (EFM) over the last3 decades has not led to a reduction inthe incidence of cerebral palsy5 but hasbeen associated with a rise in the numberof operative deliveries performed.6

This possibly is due to the poor positivepredictive value of EFM for true fetalcompromise. Currently, between 8% and25% of cases of neonatal encephalopathyare believed to be due to intrapartumhypoxia.7-9 Better identification of fetusesat risk of intrapartum compromise wouldallow a more targeted approach to intra-partum monitoring as well as allowingmore informed decisions to be maderegarding place and mode of delivery.

To our knowledge, no studies to datehave investigated the value of umbilicalvenous flow rate, in normal, appropri-ately grown term pregnancies, as apredictor of a subsequent diagnosis offetal compromise in labor. As in otherstudies of fetal hemodynamics, a

erican Journal of Obstetrics & Gynecology 1.e1

Research Obstetrics www.AJOG.org

significant proportion of studies ofumbilical venous flow have taken placein cohorts of fetuses known to begrowth restricted.

There is very little similar informationavailable from appropriately grown fe-tuses. In this prospective observationalstudy, we investigated the distribution ofumbilical venous flow rates in a cohortof normal, appropriately grown, fetusesat term and correlated the umbilicalvenous flow rate with intrapartumand neonatal outcomes.We hypothesizedthat appropriately grown fetuses, with thelowest umbilical venous flow rates, wouldbe at increased risk of a diagnosis ofintrapartum fetal compromise and sub-sequent emergency delivery.

MATERIALS AND METHODS

This was a prospective observationalstudy based at Queen Charlotte’s andChelsea hospitals (London, UK). Allwomen bookedwith low-risk pregnancies(those with no maternal or fetal concernsidentified in the antenatal period), pre-senting in early labor, and likely to deliverwithin 72 hours were considered eligiblefor inclusion. Early labor was defined asregular, painful contractions associatedwith cervical effacement and or dilatationup to and including 4 cm.

Exclusion criteria were multiple preg-nancy, known fetal growth restriction/preeclampsia, known fetal anomaly, cer-vical dilatation greater than 4 cm, andruptured membranes with meconiumstained liquor. Women were recruitedfrom the delivery suite and day assess-ment unit over a 2 year period. Each pa-tient gave written consent for inclusion inthe study. Patient demographics such asage, ethnicity, and parity were recorded.

Women then underwent an ultra-sound assessment inwhich fetal biometry(head circumference, biparietal diameter,abdominal circumference, femur length)and umbilical venous Doppler wereassessed. All ultrasound scans wereperformed by a single operator using aVoluson e machine (GE Healthcare,Buckinghamshire, UK), and a trans-abdominal curvilinear transducer (AB2-7-RS). Participants were asked to lie in asupine position, with a slight left lateraltilt to avoid caval compression.

1.e2 American Journal of Obstetrics & Gynecology

For Doppler assessment, the umbilicalveinwas imaged at a free loop, and pulsedwave Doppler was used to obtain um-bilical vein flow velocities. Time-averagedmaximum velocities were recorded,which are reported to show good corre-lation with actual mean flow velocities inexperimental models.10 The angle ofinsonation was maintained as close to0 degrees as possible, and always less than30 degrees, to ensure accurate assessmentof flow velocities. The angle correctionfunction was used when any flow veloc-ities weremeasured at an angle other than0 degrees. The umbilical vein was thenimaged at a free loop, in gray-scale and inlongitudinal section, with the umbilicalvein perpendicular to the ultrasoundbeam. The image was magnified and theinternal diameter of the vein measured.All readings were taken between

uterine contractions and were repeated 3times, with themean value being used fordata analysis. An estimated fetal weightwas calculated from fetal biometry usingthe formula of Hadlock et al.11

The umbilical vein flow velocity andumbilical vein diameter were then usedto calculate umbilical venous flow rate.For this calculation, the umbilical veinwas considered a cylindrical vessel andassumed to have laminar flow. Thefollowing formulawas used for calculation:umbilical venous flow rate (milliliters

per minute) ¼ velocity (centimeters persecond) � 0.3 � cross-sectional area(square millimeters)Derivation of the formula is presented

in the Appendix.Doppler assessment of umbilical ve-

nous flow rate has been demonstrated tobe accurate when compared with thegold standard methods of in vivo flowcalculation.12

Cases were then managed according tolocal protocols and guidelines. Staffmanaging the labor were blinded to theresults of the ultrasound scan to ensurethat the ultrasound findings did not affectobstetric management.Power calculations, based on data

from a pilot study of 100 cases, suggesteda power of 0.8 (to detect a difference inincidence of cesarean section for pre-sumed fetal compromise between fetuseswith the lowest and highest umbilical

MONTH 2013

venous flow rates of 0.15) could beachieved with a sample size of 500 cases.

Following delivery, patient notes andelectronic records were reviewed andintrapartum andneonatal outcome detailsrecorded. Cases were subcategorizedaccording to the mode of delivery andcorrected umbilical venous flow rate(milliliters per minute per kilogram).Intrapartum cardiotocograph (CTG) re-cordings were analyzed and classified byan obstetrician blinded to the ultrasoundscan results as normal, suspicious, orpathological, based on the National In-stitutes of Health and clinical excellenceguidelines.13 This is a nationally usedclassification of electronic fetal heart ratetracings. A diagnosis of intrapartum fetalcompromise, and the basis for a deliveryperformed for this indication, was apathological CTG or a suspicious CTGcombined with a pH less than 7.20 fromintrapartum fetal blood sampling.

Ethical approval for this study wasgranted by the North London ResearchEthics Committee (reference numberREC 10/H0718/26).

RESULTS

Five hundred eighty-nine women wererecruited to the study over a 2 yearperiod from February 2011 to February2013. Patient demographics are shownin Table 1. The median interval betweenthe ultrasound scan and delivery was1 day (range, 0e14). Ninety-six percentof women delivered within 72 hours ofthe ultrasound scan taking place.

Umbilical venous flow was non-normally distributed in the study popu-lation (Kolmogorov-Smirnov, 0.02). Themedian umbilical venous flow rate in thestudy population was 213.6 mL/min andthe interquartile range was 74.19 mL/min.Umbilical venous flow rate was found tobe positively correlated with birthweight(R2 ¼ 0.19) but was not correlated withgestation (R2 ¼ 0.00) in our cohort.Therefore, umbilical venous flow rates inthis study are reported as raw values aswell as following correction for birth-weight. Corrected umbilical venous flowrate also had a nonnormal distribution inthe study population (Kolmogorov-Smir-nov, 0.01). The median corrected umbil-ical venous flow rate was 61.1mL/min per

TABLE 1Maternal demographics and intrapartum and neonatal outcome according to mode of delivery

Variable Overall

Emergencycesarean fetalcompromise SVD

Instrumentalfetal compromise

Instrumentalprolongedsecond stage

Emergencycesarean other

ANOVA/c2 P value

Number of cases 589 68 233 107 94 87

Primiparous, % 75.0 (442/589) 92.6 (63/68) 47.2 (110/233) 91.6 (98/107) 90.4 (85/94) 94.3 (82/97) < .001

Mean maternal age, y 32 (16e47) 32 (19e42) 32 (18e47) 32 (16e44) 32 (18e43) 34 (22e46) .14

Mean BMI 24.7 (16e42) 25.5 (19e42) 24.7 (17e38) 24.1 (18e35) 24.2 (16e41) 25.2 (18e40) .08

Ethnicity, %

White 67.4 (397/589) 54.4 (37/68) 68.2 (159/233) 71.0 (76/107) 69.1 (65/94) 69.0 (60/87) .73

Asian 17.7 (104/589) 26.5 (18/68) 13.3 (31/233) 19.6 (21/107) 18.1 (17/94) 19.5 (17/87) .21

Afro-Caribbean 10.0 (59/589) 13.2 (9/68) 13.3 (31/233) 4.7 (5/107) 6.4 (6/94) 9.2 (8/87) .11

Other 4.9 (29/589) 5.9 (4/68) 5.2 (12/233) 4.7 (5/107) 6.4 (6/94) 2.3 (2/87) .77

Ultrasound

Mean gestationof ultrasound scan

40.5 (37e42.3) 40.7 (37.7e42.0) 40.3 (37.0e42.0) 40.5 (37.0e42.1) 40.7 (37.6e42.1) 40.6 (37e42.3) .01

Umbilical venous flowrate, mL/min

213.6 (68.3e577.1) 195.4 (83.4e341.9) 216.1 (80.3e226.4) 205.8 (68.3e452.4) 227.5 (81.6e363.1) 228.0 (89.9e577.1) .001

Corrected umbilical venousflow rate, mL/min per kilogram

61.1 56.6 62.0 59.4 63.0 61.5 .02

Neonatal outcomes

Birthweight, g 3532 (1780e5026) 3423 (2478e4906) 3498 (2520e4480) 3474 (1780e4940) 3612 (2390e5026) 3696 (2690e4506) < .001

Birthweight centile 53.5 (1e100) 44.2 (1e100) 52.3 (2e99) 49.9 (1e100) 57.7 (2e100) 63.7 (5e99) < .001

Apgar <7 at 5 minutes 1.0% (6/589) 1.5% (1/68) 0.4% (1/233) 0.9% (1/107) 2.1% (2/94) 1.2% (1/87) .71

Cord arterial pH <7.20 31.6% (186/589) 26.5% (18/68) 35.6% (83/233) 48.6% (52/107) 23.4% (22/94) 12.6% (11/87) < .001

Neonatal unit admission 1.5% (9/589) 2.9% (2/68) 0.9% (2/233) 2.8% (3/107) 1.1% (1/94) 1.2% (1/87) .56

Neonatal encephalopathy 0% (0/589) 0% (0/68) 0% (0/233) 0% (0/107) 0% (0/94) 0% (0/87) n/a

Composite neonataloutcome score

0.76 (0e7) 0.63 (0e5) 0.83 (0e7) 1.21 (0e5) 0.62 (0e4) 0.31 (0e5) < .001

Composite neonatal outcome scored as follows: Apgar,>7 at 1 minute¼ 0,<7 at 1 minute¼ 1,<7 at 5 minutes¼ 2; cord arterial pH,>7.20¼ 0,<7.20¼ 1,<7.10¼ 2,<7.00¼ 3; base excess, greater thane8¼ 0, greater thane8 and less thane12¼1, greater than e12 ¼ 2; and NNU admission, no ¼ 0, yes ¼ 1.

ANOVA, analysis of variance; BMI, body mass index; n/a, not applicable; NNU, neonatal unit; SVD, spontaneous vaginal delivery.

Prior. Umbilical venous flow and fetal compromise. Am J Obstet Gynecol 2013.

www.AJO

G.org

Obstetrics

Resea

rch

MONTH

2013American

JournalofObstetrics

&Gynecology

1.e3

TABLE 2Mode of deliveryVariable n (%)

Cases 589

Cesearean 155 (26.3)

Presumed fetalcompromise

68 (43.9)

Other indication 87 (56.1)

SVD 233 (39.6)

Instrumental deliveries 201 (34.1)

Presumed fetalcompromise

107 (53.2)

Prolongedsecond stage

94 (46.8)

Cesarean not caused by fetal compromise were per-formed for failure to progress in 86 of 87 cases, with 1performed because of unstable lie.

SVD, spontaneous vaginal delivery.

Prior. Umbilical venous flow and fetal compromise.Am J Obstet Gynecol 2013.

Research Obstetrics www.AJOG.org

kilogram and the interquartile range was18.75 mL/min per kilogram.

Cases were subclassified according tomode of delivery (mode of delivery de-tails for all participants are documentedin Table 2). The percentage of primiparouswomen was lower in the spontaneousvaginal delivery (SVD) group, but no otherdifferences in maternal demographicswere identified between the differentmode of delivery groups (Table 1).

Infants born by emergency cesarean forpresumed fetal compromise had thelowest umbilical venous flow rates, bothuncorrected and corrected for birth-weight. Infants requiring instrumentaldelivery for presumed fetal compromisehad umbilical venous flow rates higherthan those of the cesarean for fetalcompromise group but lower than thoseof infants born by SVD or instrumentaldelivery for a prolonged second stage.Infants born by SVD had higher flowrates, with only infants born by instru-mental delivery for a prolonged secondstage and cesarean delivery (other indica-tion, largely failure to progress in the firststage of labor) having higher uncorrectedand corrected umbilical venous flow rates.

Both uncorrected and corrected flowrates were significantly different betweenthe various modes of delivery groups

1.e4 American Journal of Obstetrics & Gynecology

when compared using 1 way analysis ofvariance (Table 1). Mean corrected um-bilical venous flow rates between eachmode of delivery groupwere also assessedusing independent sample Student t tests.Statistically significant differences wereobserved between the emergency cesar-ean for fetal compromise group and boththe SVD group (P ¼ .01) and instru-mental delivery for the prolongedsecond-stage group (P ¼ .005). Directcomparison between the other modes ofdelivery groups did not demonstrate sig-nificant variation. Statistically significantvariation in mean birthweight and meanbirthweight centile was also observedbetween the modes of delivery groups,with infants born by emergency cesareanfor presumed fetal compromise havingthe lowest mean birthweight (Table 1).Corrected umbilical venous flow rates

were also recalculated using ultrasound-derived estimated fetal weight asopposed to birthweight. The difference incorrected umbilical venous flow rate be-tween the mode of delivery groups re-mained statistically significant (P ¼ .03).Because the study group included 37

infants with a birthweight less than the10th centile for gestation, who may havebeen growth restricted, the analysis wasrepeated with these infants excluded.Both the umbilical venous flow rate andcorrected umbilical venous flow rateremained significantly different betweenthe mode of delivery groups (P ¼ .003and P ¼ .04, respectively).Receiver-operator characteristic (ROC)

curves were constructed to assess theability of umbilical venous flow ratesto predict emergency cesarean for fetalcompromise. This resulted in an areaunder curve of 0.62 (P ¼ .006, 95%confidence interval [CI], 0.54e0.69)and 0.61 (P ¼ .01, 95% CI, 0.53e0.68),respectively.Other clinical evidence of fetal com-

promise in labor was examined. Weobserved no significant relationship be-tween uncorrected or corrected flow ratesand the passage of meconium or require-ment for intrapartum fetal blood sampling(suggesting persistent CTG abnormalitiesclassified as suspicious but not patholog-ical). Amean umbilical venous flow rate of222.4mL/minwas observed in fetuseswith

MONTH 2013

normal CTG recordings, 210.1 mL/minin those with suspicious CTG record-ings, and 201.2 mL/min in those withCTG recordings defined as pathological(P ¼ .009).

The study populationwas then dividedinto 3 groups based on corrected umbil-ical venous flow rates of less than the 20thcentile, 20th-80th centile, and greaterthan the 80th centile. The 20th centilevalue for corrected umbilical venous flowwas 48.5 mL/min per kilogram, whereasthe 80th centile value was 73.4 mL/minper kilogram. These cutoff values wereidentified from our ROC curve analyses.

Maternal demographics and neonataloutcomes for the 3 groups are docu-mented in Table 3. Those fetuses with thelowest corrected umbilical venous flowrates had the highest incidence of emer-gency cesarean or instrumental deliveryfor presumed fetal compromise as well asthe lowest rates of SVD or vaginal de-livery of any kind. Conversely, those fe-tuses with the highest corrected umbilicalvenous flow rates had the lowest inci-dence of emergency cesarean delivery aswell as the highest incidence of a vaginaldelivery of any kind (Tables 4 and 5). Anumbilical venous flow rate less than the20th centile was found to have a positivepredictive value for cesarean delivery of16.4%, whereas an umbilical venous flowrate greater than the 80th centile had anegative predictive value of 94.4%.

Neonatal outcomes including theincidence of an Apgar score less than 7 at5minutes, cord arterial pH less than 7.20at delivery, neonatal unit admission, andneonatal encephalopathy were alsocompared (Table 3). No significant dif-ference in Apgar scores or cord arterialpH was noted between the correctedumbilical venous flow rate groups. Acomposite neonatal outcome score wasthen calculated (Apgar score at 1 and5 minutes, cord arterial pH and baseexcess, and neonatal unit admission).This demonstrated no significant dif-ference in neonatal outcomes betweenthe 3 corrected umbilical venous flowrate groups.

COMMENT

The results from this study dem-onstrate that among appropriately

TABLE 3Maternal demographics and intrapartum and neonatal outcome according to corrected umbilical venous flowrate centile group

Demographic OverallUV flow rate<20th centile

UV flow rate20the80th centile

UV flow rate>80th centile

ANOVA/c2 P value

Primiparous, % 75.0 (442/589) 80.6 (87/108) 73.7 (275/373) 74.1 (80/108) .76

Mean maternal age, y 32 (16e47) 31.2 (16e44) 32.3 (18e47) 33.8 (21e46) .003

Mean BMI 24.7 (16e42) 24.8 (18e40) 26.5 (16e42) 24.5 (17e36) .07

Ethnicity, %

Caucasian 67.4 (397/589) 71.3 (77/108) 65.4 (244/373) 70.4 (76/108) .74

Asian 17.7 (104/589) 17.6 (19/108) 17.4 (65/373) 18.5 (20/108) .97

Afro-Caribbean 10.0 (59/589) 8.3 (9/108) 11.3 (42/373) 7.4 (8/108) .45

Other 4.9 (29/589) 2.8 (3/108) 5.9 (22/373) 3.7 (4/108) .36

Mean gestation at time of U/S 40.5 (37.0e42.0) 40.8 (37.4e42.0) 40.5 (37.0e42.0) 40.1 (37.0e42.0) < .001

Evidence of fetal compromise

Rate of meconium stainedliquor, %

12.2 (72/589) 12.0 (13/108) 11.3 (42/373) 15.7 (17/108) .50

Incidence of CTG abnormalitiesdefined as pathological, %

19.9 (117/589) 26.9 (29/108) 17.8 (67/373) 19.4 (21/108) .19

Neonatal outcome

Birthweight, g 3532 (1780e5026) 3550 (2590e5026) 3538 (1780e4858) 3498 (2390e4900) .70

Birthweight centile 53.5 (1e100) 52.4 (2e100) 53.7 (1e100) 53.7 (1e100) .83

Apgar <7 at 5 minutes, % 1.0 (6/589) 0.9 (1/108) 0.5 (2/373) 2.8 (3/108) .12

Cord arterial pH <7.20, % 31.6 (186/589) 31.5 (34/108) 30.3 (113/373) 36.1 (39/108) .64

Neonatal unit admission, % 1.5 (9/589) 3.7 (4/108) 0.3 (1/373) 3.7 (4/108) .004

Neonatal encephalopathy, % 0 (0/589) 0 (0/108) 0 (0/373) 0 (0/108) n/a

Composite neonatal outcome score 0.76 (0e7) 0.79 (0e7) 0.71 (0e4) 0.92 (0e5) .25

Composite neonatal outcome scored as follows: Apgar,>7 at 1 minute¼ 0,<7 at 1 minute¼ 1,<7 at 5 minutes¼ 2; cord arterial pH,>7.20¼ 0,<7.20¼ 1,<7.10¼ 2,<7.00¼ 3; baseexcess, less than e8 ¼ 0, greater than e8 and less than e12 ¼ 1, greater than e12 ¼ 2 ; NNU admission, no ¼ 0, yes ¼ 1.

ANOVA, analysis of variance; BMI, body mass index; CTG, cardiotocograph; n/a, not applicable; NNU, neonatal unit; U/S, ultrasound; UV, umbilical venous.

Prior. Umbilical venous flow and fetal compromise. Am J Obstet Gynecol 2013.

www.AJOG.org Obstetrics Research

grown fetuses at term, umbilicalvenous flow shows considerable varia-tion. Fetuses with umbilical venousflow rates less than the 20th cen-tile (both uncorrected and correctedfor birthweight) are at significantly in-creased risk of a subsequent diagnosis ofintrapartum fetal compromise requiringemergency delivery. These data demon-strate that even among a cohort ofapparently low-risk pregnancies (withappropriately grown fetuses), significantvariations exist in the delivery of oxygenand nutrients to the fetus and that thesevariations can be correlated to the needfor subsequent emergency delivery inlabor.

To our knowledge, this is the first timethat quantitative variations in umbilicalvenous flow have been demonstrated inan appropriately grown cohort of fetusesand correlated to intrapartum outcomes.The mean value for corrected umbilical

venous flow rate in our study was61.1 mL/min per kilogram. Variousstudies have examined corrected umbilicalvenous flow rate at differing gestations,2,14

but few have examined umbilical venousflow rates in appropriately grown fetusesat term. Acharya et al15 reported a 50thcentile value for corrected umbilicalvenous flow rate at 40 weeks’ gestationof 65.86 mL/min per kilogram, withmeasurements of vessel diameter and

MONTH 2013 Am

umbilical flow velocity measured at theintraabdominal portion of the vessel.

In our study, we chose tomeasure boththe umbilical vein diameter and flow ve-locity at a free loop of cord. Althoughumbilical vein diameter has been ob-served to decrease from the fetus to theplacenta, acquiring accurate ultrasoundimages of the umbilical vein at both a0 degree and 90 degree angle to deter-mine diameter and velocity, respectively,are most easily achieved at a free loop ofcord.12 When calculating flow volumes,small errors in the measurement of vesselsize or flow velocity can lead to large er-rors in volume flow. Despite this, previ-ous authors have demonstrated that

erican Journal of Obstetrics & Gynecology 1.e5

TABLE 4Delivery details according to corrected umbilical venous flow rate centile groups

Variable Overall <20th centile 20the80th centile >80th centile P value<20th vs >80th centile,RR (95% CI)

Emergency cesareanfetal compromise

11.5% (68/589) 15.7% (17/108) 12.1% (45/373) 5.6% (6/108) .07 2.83 (1.16e6.91)

Instrumental fetalcompromise

18.2% (107/589) 25.0% (27/108) 16.1% (60/373) 18.5% (20/108) .16 1.35 (0.81e2.26)

SVD 39.6% (233/589) 35.2% (38/108) 41.0% (153/373) 38.9% (42/108) .69 0.90 (0.64e1.28)

Instrumental failureto progress

16.0% (94/589) 8.3% (9/108) 16.9% (63/373) 20.4% (22/108) .06 0.41 (0.20e0.85)

Fetal compromisediagnosed at anytime during labor

29.7% (175/589) 40.7% (44/108) 28.2% (105/373) 24.1% (26/108) .05 1.69 (1.29e2.54)

Vaginal deliveryof any kind

73.7% (434/589) 68.5% (74/108) 74.0% (276/373) 77.8% (84/108) .73 0.88 (0.75e1.04)

CI, confidence interval; RR, relative risk; SVD, spontaneous vaginal delivery.

Prior. Umbilical venous flow and fetal compromise. Am J Obstet Gynecol 2013.

Research Obstetrics www.AJOG.org

umbilical venous flow rates measuredusing Doppler ultrasound demonstrategood correlation with in vivo models.16

The methodological differences be-tween this study and that of Acharya et al15

may explain the slightly lower mean um-bilical venous flow rate reported in ourstudy. Tchirikov et al17 correlated umbili-cal venous flow rate to a neonatal outcomescore based onApgar at 1minute, cord pH,birthweight, gestation, and the need for

TABLE 5Delivery details according to correcteinfants excluded from analysis)

Variable Overall <2

Emergency cesareanfetal compromise

10.9% (60/552) 16

Instrumental fetalcompromise

17.8% (98/552) 24

SVD 39.5% (218/552) 32

Instrumental failureto progress

16.5% (91/552) 8

Fetal compromisediagnosed at anytime during labor

28.6% (158/552) 41

Vaginal deliveryof any kind

73.7% (407/552) 66

CI, confidence interval; RR, relative risk; SVD, spontaneous vagi

Prior. Umbilical venous flow and fetal compromise. Am J O

1.e6 American Journal of Obstetrics & Gynecology

ventilation/neonatal unit admission. In thestudy by Tchirikov et al,17 the compro-mised infants had a mean birthweight of1790 g, suggesting a significant proportionof these babies may have been growthrestricted, a cohort specifically excluded inour study. Our results, however, suggestthat umbilical venous flow may also bepredictive of a diagnosis of intrapartumcompromise and subsequent emergencydelivery in appropriately grown fetuses.

d umbilical venous flow rate centile grou

0th centile 20the80th centile >80th cent

.8% (17/101) 10.9% (38/350) 5.0% (5/1

.8% (25/101) 16.3% (57/350) 15.8% (16/

.7% (33/101) 41.4% (145/350) 39.6% (40/

.9% (9/101) 17.4% (61/350) 20.8% (21/

.6% (42/101) 27.1% (95/350) 20.8% (21/

.3% (67/101) 75.1% (263/350) 76.2% (77/

nal delivery.

bstet Gynecol 2013.

MONTH 2013

When compared with those with aflow rate greater than the 80th centile,fetuses with an umbilical venous flow rateless than the 20th centile were noted tohave a significantly increased risk of beingdiagnosed with fetal compromise duringlabor (relative risk [RR], 1.69, 95% CI,1.29e2.54). These fetuses were alsonoted to have the highest incidenceof cesarean section for presumed fetalcompromise (15.7%) and were almost

ps (with small-for-gestational-age

ile P value<20th vs >80th centile,RR (95% CI)

01) .04 3.40 (1.30e8.86)

101) .18 1.56 (0.89e2.75)

101) .47 0.83 (0.57e1.19)

101) .09 0.43 (0.21e0.89)

101) .02 2.00 (1.28e3.12)

101) .63 0.87 (0.73e1.04)

www.AJOG.org Obstetrics Research

3 times more likely to be delivered in thisway than fetuses with the highest cor-rected umbilical venous flow rates (RR,2.83, 95% CI, 1.16e6.91). These differ-ences persisted when the analysis wasrepeated with small-for-gestational-agefetuses excluded (Table 5). Although theabsolute difference in the incidence ofcesarean delivery for fetal compromisebetween the different umbilical venousflow centile groups was small, the relativerisk between the groups does suggest thatassessment of umbilical venous flow ratemay have a clinical application.

The data from this study did notdemonstrate a significant difference inApgar scores or cord arterial pH at de-livery between infants with differingumbilical venous flow rates. Although adeterioration of these values is associat-ed with fetal compromise in labor, theyreflect the condition of the fetus atthe time of delivery and not the time adecision is made to deliver. There istherefore potential for these values toboth deteriorate, such as following aprolonged, difficult instrumental de-livery, and improve, such as with cessa-tion of Syntocinon augmentation andmaternal oxygen treatment during theinterval between the decision to deliverand the delivery itself. The incidenceof adverse neonatal outcomes (Apgar<7at 5 minutes, neonatal unit admission)were minimal in this study, limiting thevalidity of any association with umbilicalvenous flow rate.

Our findings suggest that quantitativeassessment of umbilical venous flow ratemay be of value in developing a model toidentify those fetuses at risk of a diag-nosis of fetal compromise during labor.The results presented in this studydemonstrates that among an appropri-ately grown cohort of low-risk fetuses,reduced umbilical venous flow rates canbe correlated to a subsequent diagnosisof intrapartum fetal compromise.

Our ROC curve analyses suggest thatalthough umbilical venous flow rate canbe correlated with intrapartum outcome,it may not have adequate sensitivity andspecificity to be used in isolation. Ourgroup has recently shown that a cere-broumbilical ratio greater than the 90thcentile,measured prior to active labor, hasa negative predictive value of 100% forthe development of fetal distress in laborrequiring emergency cesarean delivery.18

We suggest that umbilical venous flow,in conjunction with other fetal cardiovas-cular Doppler indices such as the cere-broumbilical ratio, umbilical artery, andmiddle cerebral artery pulsatility index,may have a potential clinical application inthe risk stratification of normal pregnanciesprior to labor. Such stratification wouldallow more informed decisions to be maderegarding themode and place of delivery aswell as a more targeted approach tointrapartum monitoring. Further work isnecessary to develop such a model/algo-rithm to accurately predict, prior to labor,fetuses at risk of intrapartum compromise.

REFERENCES

1. Gill RW, Kossoff G, Warren PS, Garrett WJ.Umbilical venous flow in normal and complicatedpregnancy. Ultrasound Med Biol 1984;10:349-63.2. Boito S, Struijk PC, Ursem NTC, Stijnen T,Wladimiroff JW. Umbilical venous volume flow inthe normally developing and growth-restrictedhuman fetus. Ultrasound Obstet Gynecol2002;19:344-9.3. Rigano S, Bozzo M, Ferrazzi E, Bellotti M,Battaglia FC, Galan HL. Early and persistentreduction in umbilical vein blood flow in thegrowth-restricted fetus: a longitudinal study. AmJ Obstet Gynecol 2001;185:834-8.4. Ghosh GS, Fu J, Olofsson P,Gudmundsson S. Pulsations in the umbilical veinduring labor are associatedwith increased risk ofoperative delivery for fetal distress. UltrasoundObstet Gynecol 2009;34:177-81.5. Clark SL, Hankins GDV. Temporal and de-mographic trends in cerebral palsy—fact andfiction. Am J Obstet Gynecol 2003;188:628-33.

MONTH 2013 Am

6. Alfirevic Z, Devane D, Gyte GM. Continuouscardiotocography (CTG) as a form of electronicfetal monitoring (EFM) for fetal assessment dur-ing labour. Cochrane Database Syst Rev 2006:CD006066.7. Hagberg B, Hagberg G, Beckung E,Uvebrant P. Changing panorama of cerebralpalsy in Sweden. VIII. Prevalence and origin inthe birth year period 1991-94. Acta Pædiatr2001;90:271-7.8. Blair E, Stanley FJ. Intrapartum asphyxia: a rarecauseof cerebral palsy. JPediatr 1988;112:515-9.9. Kumar S, Paterson-Brown S. Obstetric as-pects of hypoxic ischemic encephalopathy.Early Hum Dev 2010;86:339-44.10. Li S,HoskinsPR,AndersonT,McDickenWN.Measurement of mean velocity during pulsatileflow using time-averaged maximum frequency ofDoppler ultrasound waveforms. Ultrasound MedBiol 1993;19:105-13.11. Hadlock FP, Harris RB, Sharman RS,Deter RL, Park SK. Estimation of fetal weightwith the use of head, body, and femur meas-urementsea prospective study. Am J ObstetGynecol 1985;151:333-7.12. Figueras F, Fernández S, Hernández-Andrade E, Gratacós E. Umbilical venous bloodflowmeasurement: accuracy and reproducibility.Ultrasound Obstet Gynecol 2008;32:587-91.13. Clinical guideline 55: intrapartum care. Na-tional Institute For Health and Clinical Excellenceguideline. National Institute For Health andClinical Excellence; 2007 (Sept. 26).14. Barbera A, Galan HL, Ferrazzi E, et al.Relationship of umbilical vein blood flow togrowth parameters in the human fetus. Am JObstet Gynecol 1999;181:174-9.15. Acharya G, Wilsgaard T, RosvoldBerntsen GK, Maltau JM, Kiserud T. Referenceranges for umbilical vein blood flow in the sec-ond half of pregnancy based on longitudinaldata. Prenat Diagn 2005;25:99-111.16. GalanH, JozwikM,RiganoS, et al. Umbilicalvein blood flow determination in the ovine fetus:comparison of Doppler ultrasonographic andsteady-state diffusion techniques. Am J ObstetGynecol 1999;181(5 Pt 1):1149-53.17. Tchirikov M, Strohner M, Förster D,Hüneke B. A combination of umbilical artery PIand normalized blood flow volume in the umbil-ical vein: venous-arterial index for the predictionof fetal outcome. Eur J Obstet Gynecol ReprodBiol 2009;142:129-33.18. Prior T, Mullins E, Bennett P, Kumar S.Prediction of intrapartum fetal compromise us-ing the cerebroumbilical ratio: a prospectiveobservational study. Am J Obstet Gynecol2013;208:124.e1-6.

erican Journal of Obstetrics & Gynecology 1.e7

Research Obstetrics www.AJOG.org

APPENDIX

Cross-sectional area was calculatedusing the measured diameter of theumbilical vein and the formulafor the area of a circle (pr2):1 mL/min ¼ 1 cm/s per squarecentimeter � 60.

1.e8 American Journal of Obstetrics & Gynecology

A multiplication factor of 0.5 isrequired to correct for laminar flow in acylindrical vessel (because of slower flowat the perimeter of the vessel caused byfriction with the vessel walls).1 mL/min ¼ 1 cm/s per square

centimeter � 60 � 0.5.

MONTH 2013

1 mL/min ¼ 1 cm/s per squarecentimeter � 30.

Because 1 cm2 ¼ 100 mm2,1 mL/min ¼ 1 cm/s per square

millimeter � 30/100.1 mL/min ¼ 1 cm/s per square

millimeter � 0.3.