Routine prenatal ultrasonography as a screening tool

http://www.uptodate.com/contents/routine-prenatal-ultrasonography-as-a-screening-tool?
source=search_result&search=ultrasonographic+on+pregnancy&selectedTitle=15%7E150#H23
Routine prenatal ultrasonography as a screening tool
Authors
Anna K Sfakianaki, MD
Joshua Copel, MD
Section Editors
Charles J Lockwood, MD, MHCM
Deborah Levine, MD
Deputy Editor
Vanessa A Barss, MD, FACOG
Disclosures: Anna K Sfakianaki, MD Nothing to disclose. Joshua Copel, MD Nothing to disclose. Charles J Lockwood, MD,
MHCM Consultant/Advisory Boards: Celula [Aneuploidy screening (Prenatal and cancer DNA screening tests in development)]. Equity
Ownership/Stock Options: Celula [Aneuploidy screening (Prenatal and cancer DNA screening tests in development)]. Deborah Levine, MD Nothing
to disclose. Vanessa A Barss, MD, FACOG Nothing to disclose.
Contributor disclosures are reviewed for conflicts of interest by the editorial group. When found, these are addressed by vetting through a multi-level
review process, and through requirements for references to be provided to support the content. Appropriately referenced content is required of all
authors and must conform to UpToDate standards of evidence.
Conflict of interest policy
All topics are updated as new evidence becomes available and our peer review process is complete.
Literature review current through: Jun 2015. | This topic last updated: Jun 02, 2015.
INTRODUCTION Prenatal ultrasonography is a common procedure in many countries [1-4]. It is an accurate technique for determining
gestational age, number of fetuses, fetal cardiac activity, and placental location. In addition, many congenital structural anomalies and significant
abnormalities in fetal growth can be identified. However, whether all obstetrical patients should undergo ultrasound screening and whether such
screening improves pregnancy outcome is controversial.
Many countries have developed local guidelines for the practice of prenatal ultrasonography and most offer at least one mid-trimester ultrasound
examination as part of standard prenatal care, although worldwide obstetric practice varies [5]. In the United States, the American College of
Obstetricians and Gynecologists (ACOG) supports the use of ultrasound when there is a specific medical indication and advises against nonmedical
use of prenatal ultrasonography (eg, to accommodate parental curiosity about fetal sex or parental desire to view and/or obtain an image of the
fetus) [6,7]. ACOG also stated that the benefits and limitations of ultrasonography should be discussed with all patients, and performance of the
procedure is reasonable in patients who request it. In addition, a 2006 workshop organized by the Eunice Kennedy Shriver National Institute of
Child Health and Human Development (NICHD) reached a consensus that all fetuses should have a screening ultrasonogram for the detection of
fetal anomalies and pregnancy complications [8].
An evidence-based discussion of the application of ultrasound to an unselected obstetric population as a screening tool will be presented here. The
use of prenatal ultrasonography for specific obstetrical indications is reviewed separately. (See "Ultrasound examination in obstetrics and
gynecology".)
CRITERIA FOR A GOOD SCREENING TEST In general, a good screening test should be safe, have high sensitivity and specificity, ideally take
only a few minutes to perform, require minimum preparation by the patient, and be inexpensive. It should identify individuals with an important
disease or condition and be cost-effective, allowing for control of health care costs by identifying disease early and treating before the
consequences of the disease are overwhelming. (See "Evidence-based approach to prevention", section on 'Methodologic issues in evaluating
screening programs'.)
RATIONALE FOR ROUTINE SCREENING PRENATAL ULTRASOUND It is hypothesized that routine use of ultrasound in all pregnancies will
prove beneficial since adverse outcomes may occur in pregnancies without risk factors and clinical conditions that place the fetus at high risk may
not be detected by clinical examination. The primary objective is to obtain information that will enable delivery of optimal antenatal care and thus the
best possible outcomes for mother and fetus [5]. However, the benefit of such prenatal sonographic screening on neonatal outcomes remains
unproven.
The components of a standard prenatal ultrasound examination are described in the tables (table 1 and table 2) and reviewed in detail separately.
(See "Ultrasound examination in obstetrics and gynecology".)
BENEFITS OF ROUTINE PRENATAL ULTRASOUND EXAMINATION Compared to no prenatal ultrasound examination or ultrasound
examination in selected patients, routine ultrasound examination improves:
Estimation of gestational age
Identification of multiple gestation
Identification of congenital anomalies
Better estimate of gestational age/delivery date Determination of the expected date of delivery (EDD) is essential in obstetrics so that
misdiagnosis (and inappropriate intervention) of previable, preterm, and postterm pregnancy can be avoided. As more focus is placed on reducing
late preterm birth and early term birth, more accurate estimation of EDD takes on increasing importance. Ample evidence has accumulated that
routine ultrasound examination results in more accurate assessment of the EDD than last menstrual period (LMP) dating or physical examination,
even in women with regular and certain menstrual dates [9-11]. Prenatal assessment of gestational age and determination of EDD is discussed in
detail separately. (See "Prenatal assessment of gestational age and estimated date of delivery".)

The benefits of ensuring the most accurate estimation of gestational age/delivery date include a reduction in intervention for postterm pregnancy,
reduction in diagnosis of fetal growth restriction, and reduction in use of tocolysis. More accurate estimation of EDD may also reduce planned
cesarean delivery before 39 weeks of gestation resulting from misdiagnosis of gestational age.
Reduction in intervention for postterm pregnancy Ultrasound-based determination of EDD reduces intervention for postterm pregnancy. In a
2010 Cochrane review of 11 trials of routine/revealed ultrasound versus selective/concealed ultrasound before the 24th week of pregnancy, routine
use of early ultrasound and the subsequent adjustment of the EDD led to a significant reduction in induction of labor for postterm pregnancy (RR
0.59, 95% CI 0.42-0.83) [11]. (See "Postterm pregnancy".)
Although these trials showed that ultrasound determination of EDD as late as 24 weeks can reduce the number of pregnancies eventually
diagnosed as postterm, as well as scans performed at 10 to 18 weeks, a randomized trial looking specifically at the timing of the examination
demonstrated that first trimester ultrasound examination in a low-risk population was more effective than second trimester ultrasound examination in
decreasing postterm pregnancy [12].
Better detection of aneuploidy Serum screening protocols are dependent upon accurate estimation of gestational age. The majority of
abnormal tests recalculated based on corrected EDD are normal. Thus, routine ultrasound estimation of gestational age may reduce the number of
women who have anxiety caused by false positive results in serum screening. (See 'Better estimate of gestational age/delivery date' above
and "Laboratory issues related to maternal serum screening for Down syndrome", section on 'Method of gestational age determination'.)
In addition to gestational age determination, the combination of sonographic nuchal translucency measurement and maternal serum analyte
assessment in the first trimester can detect over 90 percent of fetuses with Down syndrome, as well as some fetuses with other aneuploidies.
(See "Sonographic findings associated with fetal aneuploidy" and "Down syndrome: Overview of prenatal screening" and "First trimester combined
test and integrated tests for screening for Down syndrome and trisomy 18".)
Risk assessment for fetal aneuploidy using cell-free DNA in maternal blood is not dependent on precise determination of gestational age.
(See "Noninvasive prenatal testing using cell-free nucleic acids in maternal blood".)
Better identification of multiple gestation A major benefit of routine ultrasound screening is early, reliable identification of twin pregnancies. In
women who do not undergo routine ultrasound examination, a significant number of twin pregnancies are not recognized until the third trimester or
delivery. (See "Twin pregnancy: Prenatal issues".)
In the 2010 Cochrane review of trials of routine ultrasound before the 24th week of pregnancy described above, multiple gestation was diagnosed
earlier in routinely scanned pregnancies and ultrasound in early pregnancy significantly reduced the failure to detect multiple pregnancy by 24
weeks of gestation (RR 0.07, 95% CI 0.03-0.17; 7 trials, 295 patients) [11]. In the Routine Antenatal Diagnostic Imaging with Ultrasound Study
(RADIUS), of over 15,000 gravidas (the largest trial in the Cochrane review), women who did not have a routine second trimester ultrasound
examination had 38 percent of twin pregnancies unrecognized until after 26 weeks of gestation and 13 percent of twins were not diagnosed until
delivery [13]. There were no twin pregnancies missed on ultrasound examination.
Similar findings were reported by the Helsinki trial, a prospective randomized trial of 9310 women, of which 4691 underwent screening ultrasound
between 16 and 20 weeks [14]. All twin pregnancies were detected before 21 weeks of gestation in the screened group, versus 76.3 percent in the
control group. Moreover, the perinatal mortality rate (PMR) of twins in the screened group was 27.8/1000 compared with 65.8/1000 in the control
group, although these were small numbers and did not reach statistical significance. Retrospective series have also suggested improved neonatal
outcomes with early diagnosis of twin pregnancy [15].
Better detection of congenital anomalies The ability of routine ultrasound screening to detect fetal anomalies in an unselected population is
highly controversial. Over 30 studies have evaluated this question, and numerous reviews have attempted to summarize them critically [16]. In the
2010 Cochrane review of trials of routine ultrasound before the 24th week of pregnancy described above, performance of routine/revealed early
pregnancy ultrasound significantly increased detection of fetal abnormalities before 24 weeks of gestation (RR 3.46, 95% CI 1.67-7.14; 2 trials, 387
patients) [11]. A number of factors (eg, gestational age at examination, type of malformation, number of ultrasounds performed, operator
experience, quality of equipment, population characteristics) affect detection rates, and were not addressed in the meta-analysis (see'Factors
affecting detection rates' below).
The largest trials and studies comparing second trimester screened and unscreened pregnancies are reviewed below (only RADIUS was included
in the Cochrane review).
Helsinki Trial The Helsinki Trial, performed from 1986 to 1987, randomly assigned 4691 women to screening ultrasound at 16 to 20
weeks to look for fetal anomalies and compared their outcome to 4619 controls who underwent ultrasound examination if obstetrically
indicated [17]. Ultrasounds were performed at one of two hospitals, 95 percent of all pregnant women in the Helsinki metropolitan area
agreed to participate, and only four women were lost to follow-up (but mortality data were complete). Seventy-seven percent of women in
the control group underwent ultrasound examination sometime during pregnancy. Routine second trimester ultrasound screening resulted
in:
Significantly increased detection of fetal anomalies. The rate of detection of malformations was 36 percent in the City Hospital and
77 percent at the University Hospital.
A significant reduction in the PMR (4.6/1000 versus 9.0/1000 in controls) because elective termination of anomalous fetuses
resulted in fewer fetal and neonatal deaths due to congenital anomalies.
RADIUS Trial The RADIUS trial was the first randomized trial of routine obstetrical ultrasound screening in the United States [13,18]. It
included over 15,000 women randomly assigned to either screening ultrasounds at both 15 to 22 weeks and 31 to 35 weeks or to ultrasound
for obstetrical indications only. Forty five percent of the control group had at least one ultrasound. Routine second trimester ultrasound
screening resulted in:
Significantly increased detection of fetal anomalies (34.8 versus 11 percent in controls). Of these anomalies, one-half were detected
prior to 24 weeks of gestation in the screened group (16.6 percent of the total cohort of anomalous fetuses). As noted in the Helsinki
trial, detection rates were significantly higher in tertiary facilities.

No improvement in any perinatal outcome under study, including mortality, preterm birth, birth weight, and neonatal morbidity
(retinopathy of prematurity, bronchopulmonary dysplasia, need for mechanical ventilation, necrotizing enterocolitis, intraventricular
hemorrhage).
No increase in the number of abortions performed for fetal anomalies, which was the same in both groups.
No improvement in survival among anomalous fetuses. Antenatal detection of fetal anomalies did not improve survival over that with
postnatal diagnosis.
Eurofetus study The Eurofetus study of 1999 is the largest study of routine ultrasonographic examination in unselected population [19].
Women were routinely scanned by trained sonologists at 18 to 22 weeks in 61 centers across Europe. Major findings were:
The sensitivity for detection of all anomalies was 56.2 percent.
The detection rate was higher for major (73.7 percent) than for minor (45.7 percent) anomalies, and higher for anomalies of the
central nervous system (CNS, 88.3 percent) and the urinary tract (88.5 percent) than for cardiac abnormalities (38.8 percent of major
cardiac and 20.8 percent of minor cardiac anomalies were detected).
Overall, 44 percent of anomalies and 55 percent of severe anomalies were detected before 24 weeks. Cardiac defects and
cleft lip/palate were diagnosed later in pregnancy than abnormalities of the CNS, urinary tract, or musculoskeletal systems.
The rate of live births for mothers carrying fetuses with anomalies was lower than that of mothers carrying fetuses with no detected
anomalies because many pregnancies with anomalous fetuses were electively terminated.
Factors affecting detection rates Several important factors need to be considered when analyzing these data: prevalence of anomalies,
characteristics of the population, the setting in which the ultrasound was performed, and study design.
Prevalence of anomalies Detection rates for anomalies are dependent upon the prevalence of each anomaly in the population studied,
and prevalence is highly dependent upon the quality of follow-up achieved. For example, how thoroughly were aborted fetuses, stillbirths,
and neonates examined for presence of anomalies? Were autopsies and x-rays performed when appropriate? Were neonates examined by
a physician with additional expertise in dysmorphology? How many neonates underwent noninvasive studies with incidental discovery of a
congenital anomaly? Some anomalies are not diagnosed until early childhood; was there examination of pediatric records to identify these
children?
The table shows that anomaly prevalence rates reported in the literature range from 0.3 to 3.2 percent (table 3). Studies with low prevalence
rates likely represent incomplete follow-up, thus detection rates and published sensitivities are not necessarily reliable. The last column in
the table corrects for the differing prevalence rates and standardizes sensitivity rates for each study. The adjusted overall cumulative
sensitivity becomes 40 percent, with the Eurofetus detection rate climbing to almost 70 percent.
Population characteristics Characteristics of the population studied also affect both detection rates and the applicability of data to
other populations. For example, two of the RADIUS trial's main inclusion criteria were that the patients have private obstetrical care and that
they be indifferent to the possibility of termination of pregnancy. In fact, after all exclusions, only 28 percent of the eligible women actually
participated in the study. One consequence of these selection criteria is that women who wished to undergo ultrasound examination,
including those who might consider termination in the event of fetal anomaly, did not elect to participate. Thus, the observation that 71
percent of women with ultrasound-detected anomalies prior to 24 weeks decided to continue their pregnancies is not generalizable to an
unselected population of pregnant women.
In this respect, data from the Helsinki trial, which included 95 percent of women in the area, more reliably describe the consequences of
detecting fetal anomalies antenatally. The Helsinki trial reported that birth rates of anomalous fetuses were low due to termination of many of
these pregnancies.
Another important aspect of the RADIUS trial is that approximately 85 percent of the population ultimately had an identifiable indication for
ultrasound. Seventy percent of women were excluded prior to randomization, primarily because potential indications for ultrasound
examination were present, and 50 percent of the remaining (enrolled) patients developed indications during the trial.
Setting The setting in which the ultrasound is performed significantly affects detection rates. Factors relating to the setting involve the
equipment available, as well as the skill and experience of the examiners. Both the RADIUS trial and the Helsinki trial showed that
examinations performed in hospital or tertiary care settings identified more anomalies than those performed in office-based or community
centers and ultrasound was more effective when performed by experienced operators. A review by the German Institute for Quality and
Efficiency in Health Care (IQWiG) also noted that higher qualifications or greater experience of examiners and superior device quality were
associated with higher detection rates for fetal abnormalities [20].
It is important to note that the RADIUS and Helsinki trials were both performed in the 1980s when ultrasound technology was still new. Since
that time, the technology has evolved dramatically, which could enhance the diagnostic efficacy of less experienced sonologists and
sonographers.
Study design Studies are designed and powered to detect differences in their main outcome variables. Investigators choose study
outcomes based upon their clinical significance and relevance to the question at hand. The endpoints chosen by the investigators of the
RADIUS trial were associated with prematurity, rather than with anomalies and their consequences. RADIUS was not powered to detect
differences in secondary outcome measures associated with the diagnosis of anomalies, such as impact on immediate survival, long-term
neonatal outcomes in anomalous fetuses, and economic and social implications. In fact, the survival rate for infants with acute lifethreatening anomalies was higher in the screened group (75 percent) than in the unscreened group (52 percent). Although this was not
statistically significant, this was likely due to the very small sample size.
Other investigators have found that patients with cardiac anomalies [21] and specifically hypoplastic left heart [22] and transposition of the
great arteries [23] have better outcomes when the diagnosis is made prenatally rather than postnatally. The data regarding long-term
outcome for prenatally diagnosed anomalies of the renal system are conflicting and limited [24].
First trimester screening Detection of fetal anomalies in the first trimester is limited by the small size of the fetus and ongoing fetal
development, which can result in later development of markers suggestive of an underlying abnormality (eg, hydramnios related to esophageal
atresia). Nevertheless, ultrasound technology is rapidly progressing and assessment of fetal anatomy in the first trimester is becoming more widely
available.

A 2013 systematic review of the accuracy of ultrasonography at 11 to 14 weeks of gestation for detection of fetal structural anomalies included 19
observational studies, approximately 78,000 fetal anatomy examinations, and 996 postnatally-confirmed malformed fetuses [25]. The overall
detection rate for fetal structural anomalies was 51 percent, but detection rates for specific abnormalities varied widely from 0 to 100 percent. For
example, no cases of renal agenesis were detected in this gestational age range, fewer than 50 percent of cases of spina bifida were detected,
more than 50 percent of cases of omphalocele, gastroschisis, and tetralogy of Fallot were detected, and 100 percent of anencephalic fetuses were
detected. Detection of anomalies was more likely if there were multiple anomalies, both abdominal and vaginal ultrasound examination was
performed, or the woman was at high risk. Although there was considerable heterogeneity among these studies, the findings affirm both the
potential benefits and limitations of the first trimester fetal anatomic survey. Most women will need a second trimester survey to provide a more
reliable assessment of fetal anatomy.
The use of nuchal translucency measurement in the first trimester as a marker for the detection of congenital heart disease is discussed elsewhere.
(See "First trimester cystic hygroma and increased nuchal translucency".)
POTENTIAL BENEFIT OF ROUTINE SCREENING
Prevention of preterm birth A 2013 Cochrane review did not find sufficient evidence to recommend routine cervical length screening of all
pregnant women [26]; however, the trials did not have a clear protocol for management of women based on cervical length and included
heterogeneous populations. Based on a meta-analysis of five trials of progesterone treatment of women with asymptomatic midtrimester cervical
shortening, cervical length screening for all singleton pregnancies in the midtrimester coupled with treatment of women found to have a short cervix
is a reasonable approach [27]. (See "Second trimester evaluation of cervical length for prediction of spontaneous preterm birth", section on
'Guidelines from national organizations'.)
In a 2012 practice bulletin, the American College of Obstetricians and Gynecologists (ACOG) neither mandated universal routine cervical length
screening in women without a prior preterm birth nor recommended against such screening [28]. However, in women undergoing obstetrical
ultrasound examination, ACOG has recommended that the cervix be examined when clinically appropriate and technically feasible [29]. In 2011, the
Society of Obstetricians and Gynaecologists of Canada (SOGC) concluded that routine transvaginal cervical length assessment was not indicated
in women at low risk [30].
UNPROVEN OR UNCLEAR BENEFIT OF ROUTINE SCREENING
Better diagnosis of deficient growth and improvement in perinatal outcome An early ultrasound examination serves as a key baseline
against which later examinations are compared for the evaluation of fetal growth. The assessment of appropriate fetal or neonatal size is based
upon the expected weight for gestational age. If gestational age is overestimated, then an appropriately grown fetus/neonate may be incorrectly
classified as growth restricted or small for gestational age (SGA). However, in the 2010 Cochrane review of trials of routine/revealed ultrasound
versus selective/concealed ultrasound before the 24th week of pregnancy described above, routine use of early ultrasound did not result in a
significant reduction in diagnosis of SGA (RR 1.05, 95% CI 0.81-1.35; 3 trials, 17,105 patients) [11]. (See "Fetal growth restriction: Diagnosis".)
A late pregnancy screening ultrasound (third trimester) for fetal growth disturbance in low-risk women also has not been effective for reliably
detecting these fetuses or improving outcome.
In a 2008 Cochrane review of trials of routine ultrasound >24 weeks versus no/concealed/selective ultrasound >24 weeks of gestation, routine
ultrasound did not improve detection of neonates with birthweight <10th percentile (RR 0.98, 95% CI 0.74-1.28; 4 trials, 20,293 patients) or decrease
stillbirth, neonatal morbidity, or neonatal mortality (RRs 0.94 to 1.04) [31].
Assessment of placental position, fetal presentation, amniotic fluid volume, macrosomia Although ultrasound examination late in
pregnancy can diagnose clinically important placental abnormalities (eg, placenta previa, placenta accreta), fetal malpresentation, disorders of
amniotic fluid volume (oligohydramnios, polyhydramnios), and excessive fetal growth, a 2008 Cochrane review found no evidence that a policy of
routine late pregnancy ultrasound screening for these conditions results in better maternal, fetal, or neonatal outcomes than performance of
ultrasound when indicated by clinical findings or high-risk factors [31].
Doppler ultrasound screening
Umbilical artery The application of Doppler to obstetrical ultrasound provides information on function in addition to form. Doppler velocimetry of
the fetal circulation is used to assess vascular impedance, especially in the setting of suspected growth restriction [32]. Multiple trials have
evaluated the use of Doppler ultrasound as a screening tool in low-risk women. These have consistently found neither maternal nor fetal benefit,
which is in contrast to the proven benefit in high-risk pregnancies. (See"Doppler ultrasound of the umbilical artery for fetal surveillance".)
In a 2015 Cochrane review including 5 trials involving 14,185 women, routine fetal and umbilical artery Doppler ultrasound examination in low-risk
or unselected populations did not result in increased antenatal, obstetric, and neonatal interventions, and no overall differences were detected for
substantive short-term clinical outcomes such as perinatal mortality, neonatal morbidity, stillbirth, birthweight, preterm birth, or cesarean delivery
[33].
Uterine artery Although meta-analyses show that uterine artery Doppler analysis can predict women at increased risk of preeclampsia [34,35],
we and most experts do not recommend these studies for screening purposes [36-40]. Close clinical monitoring for preeclampsia is already a major
component of prenatal care; improved identification of women at increased or decreased risk of a disease that cannot be prevented and has no
treatment other than delivery is unlikely to improve maternal or fetal outcome [41]. Furthermore, the false positive rate of this test is quite high
[39,40], leading to excessive patient anxiety and health care costs. Further research is needed before screening with uterine artery Doppler can be
recommended. (See "Prediction of preeclampsia", section on 'Uterine artery Doppler velocimetry'.)
OTHER CONSIDERATIONS
Safety issues The safety of obstetric ultrasound is well established and is reviewed in detail elsewhere. (See "Basic principles and safety of
diagnostic ultrasound in obstetrics and gynecology".)
Resource utilization The overall effect of routine screening on the health care system has yet to be determined. A small number of reports have
analyzed the cost-benefit ratio of screening low-risk women with obstetrical ultrasound. The value (ie, utility) that individual patients place on birth,

or avoidance of birth, of a child with an unsuspected anomaly needs to be incorporated into any cost-benefit equation of this issue, but is not
routinely assessed, nor are such values easily assigned in monetary terms.
The authors of the Routine Antenatal Diagnostic Imaging with Ultrasound Study (RADIUS) trial concluded that routine ultrasound was not costeffective and that offering the procedure would place a burden on health care systems. However, the authors neglected to take into consideration all
outcome measures in the study. As an example, although overall perinatal mortality was not changed, there was a significantly higher rate of
detection of anomalies in the screened group. The study was not powered to analyze what effect this detection rate had on secondary outcomes,
namely utilization of abortion services or reduction in the neonatal costs for pregnancies that were not continued. In addition, the trial found a
significant reduction in induction for postterm pregnancy and in the use of tocolysis, but these effects were not evaluated in the cost analysis. The
cost analysis also did not account for the societal costs of the ultrasounds performed on the patients who did not qualify for randomization.
Because of these criticisms, investigators not involved in the RADIUS study performed a complex cost-benefit analysis of routine second-trimester
ultrasound using data from the RADIUS trial and taking into consideration all of the findings of the trial [42]. Costs were estimated for each type of
fetal anomaly and projected for the follow-up each would need in neonatal life, and savings from decreased use of tocolysis and decreased
diagnosis of postterm pregnancy were calculated. This analysis concluded that routine ultrasound screening was associated with significant
savings, but only if the ultrasonography was performed in tertiary care centers, where experience and efficiency were greatest.
The Helsinki Trial also addressed the use of antenatal care services in its trial of routine ultrasound screening [43]. Ultrasound screening slightly, but
significantly, increased the number of routine antenatal visits (12.9 versus 12.6). However, it also reduced the need for specialist services, primarily
by improving the accuracy of gestational age estimation and removing questions regarding pregnancy dating. There were no differences in hospital
admissions.
A cost-effectiveness analysis on the same data set suggested that second trimester screening with ultrasound is cost-effective, at least in a society
with a publicly financed health system [44]. The absolute cost per ultrasound was $86, which included the cost of equipment, staff, travel, and lost
working time. When the costs of screening-induced examinations and procedures (repeat ultrasounds, amniocenteses, etc) were added, the cost
per ultrasound rose to $102. Cost savings were calculated based upon decreased utilization of health services and visits, and amounted to $182
per ultrasound, yielding a net saving of $80 per patient in the screened group. The gross cost of avoiding one perinatal death was $21,938, as
based upon the decreased perinatal mortality rate (PMR) reported in that trial. Stated differently, if the malformation rate was 2.95/1000, and the
cost per ultrasound was $102, the cost to identify a single fetus with a malformation was $34,620, a number which is much lower than the amount
spent during the hospitalization of a neonate with a congenital anomaly. The Helsinki data cannot be directly applied to the United States health
care system when one considers that costs in general are much greater in the United States. Importantly, this evaluation assumes the availability
and acceptability of termination of pregnancy.
A randomized trial from an urban center in South Africa also looked at the costs of routine ultrasound screening and perinatal outcomes [45]. The
study group consisted of patients without risk factors for congenital anomalies referred for ultrasound at 18 to 24 weeks of gestation, while the
control group had routine care only; both groups could be referred for additional scans as indicated. Women in the control group were more likely to
be diagnosed with postdate pregnancies and undergo amniocenteses for confirmation of lung maturity; however, the overall incidence of major
adverse perinatal outcomes was comparable in the two groups. Routine sonography was associated with increased costs that were not accounted
for by improved outcomes.
A systematic review of studies examining cost and cost-effectiveness in screening ultrasound concluded that the available data were of poor quality
and that data could not be summarized to draw useful conclusions [46]. The authors commented specifically that charges were examined instead of
costs, and maternal costs and long-term costs were not examined effectively. It is also important to emphasize that the data used in these cost
analyses come from countries outside of the United States, and also from a different time period, factors that limit their generalizability to the
contemporary United States population.
Ethical considerations The ethical concept of autonomy guides the physician in providing patients with as much information as they need to
make educated choices about their medical care. It implies a respect for patients' individual belief systems and patients' ability to choose among
appropriate options in health care. Engaging patients in a thoughtful discussion of the benefits and limitations of routine ultrasound respects the
concept of autonomy by providing patients with the information they need to make these choices [16,47].
However, moral dilemmas have emerged when healthy fetuses have been lost due to invasive diagnostic testing or pregnancy termination related to
ultrasound findings of uncertain clinical significance [48].
Women's views Ultrasound is very attractive to women because it provides early visual confirmation of pregnancy, contact with their babies, and
enhances feeling of bonding between the woman and her baby [1,49-56]. Although most examinations are reassuring about fetal well-being, anxiety
can be heightened while waiting for the scan and when the scan shows a problem, especially when findings of uncertain clinical significance are
noted.
INFORMATION FOR PATIENTS UpToDate offers two types of patient education materials, The Basics and Beyond the Basics. The Basics
patient education pieces are written in plain language, at the 5th to 6th grade reading level, and they answer the four or five key questions a patient
might have about a given condition. These articles are best for patients who want a general overview and who prefer short, easy-to-read materials.
Beyond the Basics patient education pieces are longer, more sophisticated, and more detailed. These articles are written at the 10 th to 12th grade
reading level and are best for patients who want in-depth information and are comfortable with some medical jargon.
Here are the patient education articles that are relevant to this topic. We encourage you to print or e-mail these topics to your patients. (You can
also locate patient education articles on a variety of subjects by searching on patient info and the keyword(s) of interest.)
Basics topics (see "Patient information: Prenatal care (The Basics)")
SUMMARY AND RECOMMENDATIONS
Routine early ultrasound is beneficial in an unselected population because of better estimation of gestational age resulting in significantly
reduced frequency of labor induction for postterm pregnancy. (See 'Better estimate of gestational age/delivery date' above.)

Routine early ultrasound examination can lead to earlier detection of clinically unsuspected fetal malformations (including aneuploidies)
and earlier detection of multiple pregnancies. The Eurofetus trial likely approaches the most accurate sensitivity rates for ultrasound in the
detection of anomalies (approximately 50 to 70 percent). One hundred percent of multiple gestations can be detected by ultrasound
examination. These findings have not been shown to improve ultimate fetal outcome, although studies have lacked power to assess for
secondary outcomes. (See 'Better detection of congenital anomalies' above and 'Better identification of multiple gestation' above.)
Improvements in technology, increasing experience, and refinements in visualization should make the detection of anomalies in the first
trimester more efficient at a time when termination of pregnancy is possible and private. (See 'First trimester screening' above.)
If one screening ultrasound examination is performed, we agree with the recommendation of the American College of Obstetricians and
Gynecologists (ACOG) that the optimal timing is at 18 to 20 weeks of gestation [29]. This allows for good visualization of anatomy and is
early enough to allow completion of prenatal diagnostic procedures (eg, fetal karyotype/microarray, additional imaging studies) while legal
termination of pregnancy is possible, if desired. The ultrasound examination should be carried out by an experienced practitioner.
(See 'Benefits of routine prenatal ultrasound examination' above.)
Routine use of ultrasound screening examination in the third trimester is not supported by available data. (See 'Unproven or unclear
benefit of routine screening' above.)
Use of UpToDate is subject to the Subscription and License Agreement.
Prenatal assessment of gestational age and estimated date of delivery
http://www.uptodate.com/contents/prenatal-assessment-of-gestational-age-and-estimated-date-of-delivery?
source=search_result&search=ultrasonographic+on+pregnancy&selectedTitle=8%7E150
Prenatal assessment of gestational age and estimated date of delivery
Authors
Andrew P MacKenzie, MD
Courtney D Stephenson, DO
Edmund F Funai, MD
Section Editor
Deborah Levine, MD
Deputy Editor
Vanessa A Barss, MD, FACOG
Disclosures: Andrew P MacKenzie, MD Nothing to disclose. Courtney D Stephenson, DO Nothing to disclose. Edmund F Funai, MD Nothing to
disclose. Deborah Levine, MD Nothing to disclose. Vanessa A Barss, MD, FACOG Nothing to disclose.
Contributor disclosures are reviewed for conflicts of interest by the editorial group. When found, these are addressed by vetting through a multi-level
review process, and through requirements for references to be provided to support the content. Appropriately referenced content is required of all
authors and must conform to UpToDate standards of evidence.
Conflict of interest policy
All topics are updated as new evidence becomes available and our peer review process is complete.
Literature review current through: Jun 2015. | This topic last updated: Jun 04, 2015.
INTRODUCTION Ultrasonography has advanced obstetric practice by enabling relatively detailed assessment of the fetus in utero, including an
accurate estimate of gestational age when performed in the first half of pregnancy [1]. This information is invaluable because most obstetrical
management decisions are strongly influenced by consideration of fetal development, which closely correlates with fetal age. A first trimester
ultrasound examination can decrease the rate of induction for postterm pregnancy [2] and improve estimation of viability when complications occur
at the limit of viability. (See "Limit of viability".)
Fetal biometric measurements used to calculate gestational age and estimated date of delivery ("due date" or EDD) will be reviewed here. The
discussion of pregnancy dating will be based on time since the last menstrual period (LMP, menstrual age), not time since conception (conceptional
or embryonic age).
ESTIMATED DATE OF DELIVERY The estimated date of delivery (EDD) is 280 days from the onset of the last menstrual period (LMP) and 266
days from date of conception. Although fertilization of an oocyte in vivo cannot be detected by any laboratory or imaging method, it occurs within 24
hours of ovulation, which can be accurately predicted by detecting the luteinizing hormone (LH) surge or by ultrasound examination of the
periovulatory follicle (see "Evaluation of the menstrual cycle and timing of ovulation") Only 4 percent of women deliver on their EDD, in part because
of the limitations of methods used to estimate gestational age, but also because of natural biologic variation in the pace of fetal maturation and the
timing of natural delivery [3].
CLINICAL ASSESSMENT The two clinical methods of gestational age assessment are history, using the date of the last menstrual period (LMP)
to calculate the estimated date of delivery (EDD), and physical examination. The clinical assessment of gestational age or duration of pregnancy
reflects the "menstrual age." In contrast, embryologists always date developmental events from the time of fertilization, which is the "embryonic
age."
Naegele's rule Naegele's rule is the most common method of pregnancy dating. The EDD is calculated by counting back three months from the
LMP and adding seven days. As an example, if the LMP is February 20, then the EDD will be November 27. If the LMP is May 28, then the EDD will
be March 4. This method assumes the patient has a 28-day menstrual cycle with fertilization occurring on day 14.
Several factors limit the accuracy of EDD based on the day of the LMP [4-6]:
Many women do not have regular 28-day cycles due to variability in the length of the follicular phase
The fertile window is more than one day, which may not be day 14
There are variations in the duration of time between fertilization and implantation

Many women are not certain of the date of their last period
Early pregnancy bleeding or recent use of hormonal contraceptives may lead to an incorrect assumption of the date of ovulation
Women who rely on recall rather than written or electronic documentation of their LMP tend to show digit preference, most commonly the 15 th day of
the month [7].
Calculator Calculators that determine EDD and current gestational age are widely available (calculator 1 and calculator 2) and should be used
instead of traditional mechanical wheels. Electronic techniques, such as apps available for download to cell phones, appear to be more accurate
than these wheels. In a study of 31 mechanical gestational wheels from a variety of sources, the largest discrepancy was four days short of the
EDD predicted by Naegele's rule, two wheels yielded EDDs that differed from each other by seven days, and wheels from the same source did not
agree with each other [8]. In this study all 20 apps gave an EDD consistent with Naegele's rule and all corrected for a leap year, while none of the
mechanical gestational wheels made this adjustment. However, others have found that a high proportion of gestational age apps are inaccurate [9].
Clinicians and patients should be aware of this possibility when using a gestational age app, and clinicians should test the accuracy of the app they
use.
Uterine size On physical examination, the pregnant uterus is soft and globular. The size-gestational age correlation is learned by experience and
is often described in terms of fruit (eg, for singleton pregnancies: 6- to 8-week size = plum, 8- to 10-week size = orange, 10- to 12-week size =
grapefruit ), despite the imprecision of this terminology.
The uterus remains a pelvic organ until approximately 12 weeks of gestation, when it becomes sufficiently large to palpate abdominally just above
the symphysis pubis. At approximately 16 weeks, the uterine fundus is palpable midway between the symphysis pubis and umbilicus, and at
approximately 20 weeks it is palpable at the umbilicus. After 20 weeks, the symphysis to fundal height in centimeters should correlate with the week
of gestation. In the absence of other dating information, uterine enlargement two fingerbreadths above the umbilicus suggests the fetus is at a
gestational age at the limit of viability if neonatal intensive care is available. Leiomyoma, obesity, and other factors affecting uterine size or the ability
to palpate the uterus (eg, retroverted position) interfere with accurate assessment.
SONOGRAPHIC ASSESSMENT Ultrasound estimates of gestational age are based on the assumption that the size of the embryo/fetus or its
parts correlate with its age. First trimester sonographic gestational age assessment is more accurate than second trimester assessment for
estimating gestational age because biological variation of these measurements is less in the first trimester than later in gestation.
Several studies have demonstrated that ultrasound estimation of the expected date of delivery (EDD) in the first half of pregnancy is superior to
dating based on the last menstrual period (LMP) or physical examination [1,10,11]. These findings are illustrated by the following examples:
One study of 34,249 singleton pregnancies with both certain menstrual dates and ultrasonographic biometry reported that delivery
occurred within seven days of the EDD more often when sonography alone, rather than LMP alone, was used to calculate the delivery date
(55.2 versus 49.5 percent) [1]. "Certain menstrual dates" were defined as a patient certain of her LMP, with regular menstrual cycles, and no
use of oral contraceptives in the preceding three months. Sonography was performed transabdominally; 95 percent of patients were
scanned between 13 and 24 weeks.
A systematic review found reduced rates of induction of labor for postterm pregnancy (odds ratio [OR] 0.68, 95% CI 0.57-0.82) among
women who underwent ultrasound examination in early pregnancy (before 24 weeks) and concluded that routine ultrasound in early
pregnancy enabled better gestational age assessment [10].
In addition, a study of 44,623 women who had both a normal LMP and ultrasound examination before 20 weeks of gestation found that sonographic
dating resulted in a small downward estimation of gestational age more often than an upward estimation [11]. This finding is consistent with other
series showing that delayed ovulation (after day 14) occurs more often than early ovulation. Because routine ultrasound examination has the net
effect of reducing the estimated gestational age, populations assigned an EDD by routine sonographic examination have a lower rate of induction
for postterm birth than those assigned an EDD by menstrual dating (2 versus 3 percent); on the other hand, they have a higher rate of preterm
delivery (9.1 versus 7.6 percent) [10].
As with menstrual dating and physical examination, the performance of ultrasound for assessment of gestational age also has limitations. Factors
that affect accuracy include the quality of the images, multiple gestation, fetal position, fetal anomalies, and biologic variation.
Indications Because ultrasound estimation in the first half of pregnancy is superior to dating based on the LMP or physical examination and also
provides information about fetal development, it is a common practice to offer this examination routinely to all pregnant women (see "Routine
prenatal ultrasonography as a screening tool") Sonographic estimation is particularly important when menstrual cycles are irregular, the LMP is
unknown, and in patients who conceive while using hormonal contraception. Ultrasound is also used to establish a pregnancy's duration when the
uterine size estimated on physical examination differs from that predicted by menstrual dating. Some causes for a discrepancy between the
perceived uterine size and that expected by LMP include uterine leiomyomata, uterine malposition (eg, retroverted uterus), and multiple gestation.
First trimester Since gestational age calculations are based upon biometric measurements, the optimal time to obtain an estimate of gestational
age is during the first trimester when biologic variation in size from fetus to fetus is minimal [12]. First trimester sonograms can be performed via the
transvaginal (TVS) and/or the transabdominal (TAS) route. The preferred mode depends on both the gestational age at the time of the study and
the patient's pelvic and abdominal anatomy. In the earliest stages of pregnancy, TVS generally provides clear and accurate images while TAS may
be unable to even detect an intrauterine gestation [13]. Therefore, TVS is typically used for early first trimester evaluation of the gestational sac, yolk
sac, and developing embryo. Measurement of crown-rump length (CRL) in early gestation is easier with TVS than with TAS, but not more accurate
for determining gestational age [14]. As the uterus grows out of the pelvis into the mid and upper abdomen and the fetus gets larger, TAS usually
provides better visualization of the pregnancy.
Gestational sac Key points of gestational age assessment by observation of the gestational sac (using TVS) include:
The gestational sac is clearly visible at 4.5 to 5 weeks of gestation, with a double decidual sign appearing at 5.5 to 6 weeks (image 1)
Gestational age (days) = mean sac diameter (MSD) in mm + 30
MSD is accurate to 5 to 7 days
MSD should not be used to date pregnancies if a CRL can be obtained or when MSD is >14 mm

During the first three to five weeks of pregnancy (conceptus phase), the endometrium has a "trilaminar" appearance [15,16]; sonographic evaluation
does not show distinct evidence of an intrauterine pregnancy at this time. The gestational sac is the first sonographic sign of an intrauterine
pregnancy [12], and appears as a small, fluid-filled, sac-like structure eccentrically located within the endometrium. This structure and its echogenic
rim represent the chorionic cavity, implanting chorionic villi, and associated decidual tissue [17]. The "double sac sign" refers to an intrauterine fluid
collection surrounded by two concentric echogenic rings, and the "intradecidual sign" refers to a fluid collection with an echogenic rim located within
a markedly thickened decidua on one side of the uterine cavity [18,19]. These signs are primarily intended to help determine whether intrauterine
fluid in a woman with a positive pregnancy test represents a gestational sac or bleeding related to extrauterine pregnancy. They are not used for
calculating EDD and their absence does not exclude intrauterine pregnancy [20].
If there are no distinct structures within the gestational sac, then initial dating measurements are based on sac diameter. To correctly measure sac
diameter, the cursors should be placed on the sac itself and should not include the echogenic region surrounding the gestational sac [12]. A size of
2 to 3 mm is the smallest size that can be clearly visualized within the uterine cavity and correlates with a gestational age of approximately four
weeks and one to three days [21-23]. The gestational sac should be seen by 4.5 to 5 weeks of gestation [24]. As the gestation progresses, the sac
(chorionic cavity) enlarges. The appearance and thickness of the decidua basalis changes over time until it can no longer be identified at 10 weeks
of gestation [25]. The decidua capsularis becomes stretched and ultimately merges with the decidua parietalis.
Gestational age based upon gestational sac measurements can be calculated in several ways. One standardized formula that is frequently used is
the MSD, which is derived by calculating the mean of the three perpendicular sac diameter measurements [26-29]. The gestational age can then be
determined by consulting the table showing gestational age according to MSD (table 1). A simplified method of estimating gestational age is to
calculate the sum of 30 plus the sac size in millimeters. This number is equivalent to the gestational age in days. As an example, using a sac size of
5 mm, the calculated age would be 35 days (30 + 5) or five weeks [12,22,26,30].
The early gestational sac grows by approximately 1 mm per day, or 7 mm per week. Gestational sac measurement is less accurate later in
pregnancy when >14 mm [26] and when the embryonic/fetal pole can be identified [31]. Therefore, it is used for dating only very early pregnancies.
The shape of the sac may change from a circular to an irregular appearance on subsequent examinations. This may be indicative of various
physiologic and anatomic factors (ie, uterine contractions, enlarged bladder, fibroids, implantation bleeds) or may be an ominous sign of a failed
pregnancy [12,26]. (See "Spontaneous abortion: Risk factors, etiology, clinical manifestations, and diagnostic evaluation", section on 'Diagnostic
evaluation'.)
The CRL is a more accurate indicator of gestational age than mean gestational sac diameter, so once the embryonic pole is evident, CRL should be
used to determine an EDD. However, this MSD may be recorded when an embryo is not identified. Caution should be used in making the
presumptive diagnosis of a gestational sac in the absence of a definite embryo or yolk sac. Without these findings, an intrauterine fluid collection
could represent a pseudogestational sac associated with an ectopic pregnancy.
Yolk sac Key points of gestational age assessment by observation of the yolk sac (using TVS) include:
The yolk sac first becomes visible at 5 weeks of gestation and normally degrades between 10 to 12 weeks
Yolk sac diameter correlates poorly with gestational age, so it is not used for gestational dating
The yolk sac is the first anatomic structure to appear within the gestational sac and provides confirmation of an intrauterine pregnancy. The yolk sac
is spherical with a sonolucent center and echogenic periphery (image 2) [12]. It can be noted initially at the beginning of the 5th week of gestation
(MSD approximately 5 mm), although it may not appear until the MSD approaches 8 mm [24,32,33]. Yolk sac diameter is calculated by taking the
average of three measurements obtained with the calipers placed at the center of the yolk sac wall [34]. Tables are available correlating yolk sac
diameter and gestational age [35]. The yolk sac continues to grow to a maximum diameter of approximately 6 mm by 10 weeks of gestation
[34,36,37]. During this natural progression, the yolk sac migrates to the periphery of the chorionic cavity and becomes undetectable sonographically
by the end of the first trimester [34,36-38].
Reasonable criteria for an abnormal gestation are MSD of 8 mm or greater with an absent yolk sac [32]. However, because of patient and
sonographer variability, this sac size should not be taken as an absolute threshold for diagnosis of an abnormal pregnancy. Either correlation with
human chorionic gonadotropin (hCG) and/or sonographic follow-up should be obtained if an initial evaluation of pregnancy shows an 8 mm empty
sac without a yolk sac. Because of the wide normal variation in yolk sac size and appearance, a small, large, or irregular yolk sac is no more than
weakly predictive of spontaneous abortion [34,37,39-41].
By identifying the yolk sac, the embryonic disc (ie, the thickened region along the outer-most margin of the yolk sac) is also located [26,32,42,43].
The embryonic disc becomes visible at 1 to 2 mm in length, which correlates with a gestational age of five to six weeks [44-47]. As discussed
above, the gestational sac and yolk sac lose their accuracy for gestational age assessment as the pregnancy advances [31]. Direct measurement of
the CRL of the embryo provides the most accurate estimate of gestational age once the embryonic pole can be identified and should be used to
determine EDD [46-48].
Crown-rump length Key points of gestational age assessment by measurement of the CRL include the following:
CRL at 7 to 10 weeks of gestation is the most accurate biometric parameter for pregnancy dating (3 days). After 10 weeks, the accuracy
of the CRL falls slightly with a margin of error of 5 days at 10 to 14 weeks.
When CRL is <25 mm, gestational age (in days) = CRL (mm) + 42.
If CRL is >84 mm, biparietal diameter (BPD) should be used for assessment of gestational age.
The CRL is the standard biometric measurement of the embryo in the first trimester [46-48]. By definition, the CRL is the longest straight-line
measurement of the embryo measured from the outer margin of the cephalic pole to the rump (image 3) [48,49]. A length of at least 5 mm routinely
allows visualization of the embryo (image 4), but some embryos as small as 2 to 3 mm can be seen [31]. Tables have been formulated to estimate
gestational age for each numeric measurement of CRL up to 120 mm [48,50-52]. These tables vary because they are based on a variety of
equations. A 2014 systematic review [53] suggested use of one of four equations for calculating gestational age (GA) from CRL [48,54-56].
Subsequently, an international standard for ultrasound dating of pregnancy based on CRL measurement was developed by the International Fetal
and Newborn Growth Consortium for the 21st Century (INTERGROWTH-21) [57]. As a quick reference, when CRL is under 25 mm, the GA (in
days) is CRL + 42 [44].

Standard practice for determining gestational age is to take the mean of three CRL measurements [12]. When CRL is measured between 7 and 10
weeks, this method is accurate within three days [44,48,50,58]. However, accuracy wanes as the gestation progresses. Estimation of gestational
age by CRL between 10 and 14 weeks is accurate within 5 days [51,59,60] and expands to within 8.4 days at 15 weeks [51]. This variation has
been attributed to normal biological differences in embryologic development and variations in anatomical positioning of the fetal head and torso
[61,62]. Both first trimester growth delay and growth acceleration are active areas of investigation [63]. (See "Diagnosis and outcome of first
trimester growth delay".)
Because of these variations, some experts prefer that estimation of gestational age by CRL measurement be limited to gestations of 6 to 10 weeks
and suggest alternative biometric studies (BPD, head circumference [HC]) for gestational age assessment later in the first trimester [12]. The goal of
these additional measurements is to improve fetal dating after the 10th week of gestation when test performance of CRL measurement diminishes.
However, these measurements have not been proven to perform better than CRL in the late first trimester [64-67] and are more technically
challenging this early in pregnancy [12]. We agree with guidelines from national organizations that recommend using CRL up to 84 mm (14 weeks
of gestation) for gestational age assessment and BPD for gestational age assessment for CRL measurements >84 mm [68-71].
Cardiac activity If the embryo is visible but too small to measure adequately, detection of cardiac activity (movie 1) establishes a gestational age
of 5.5 to 6 weeks [33].
Second and third trimester The four standard biometric parameters commonly used to estimate gestational age and/or fetal weight in the
second and third trimesters are BPD, HC, abdominal circumference (AC), and femur length (FL) [64]. They are typically obtained by TAS
examination. (See "Prenatal sonographic assessment of fetal weight".)
Biparietal diameter The BPD is the best-studied biometric parameter [51,72-78] because it is highly reproducible and can predict gestational
age within 7 days when measured between 14 and 20 weeks of gestation [61,62,72-74]. As discussed above, we agree with guidelines from
national organizations that recommend using BPD for gestational age assessment when the CRL is >84 mm (14 weeks of gestation) and CRL for
gestational age assessment when the measurement is 84 mm [68-70]. BPD test performance diminishes as the gestation progresses beyond 14
to 20 weeks [74,75,79]. By the mid- to late-third trimester, the margin of error is three to four weeks. This significant variation is likely due to large
normal biological variation in fetal shape and size near term [74,80,81].
The BPD is measured on a plane of section that intersects both the third ventricle and thalami (image 5) [31,82]. This standardized method of
measurement helps to ensure reproducibility among examiners. To further enhance test performance, the calvarium should appear smooth and
symmetrical in the plane of section. The appropriate image can be obtained by positioning the abdominal transducer perpendicular to the fetal
parietal bones [31]. The cursors are then placed on the outer edge of the proximal skull and the inner edge of the distal skull. This length represents
the BPD [31,82,83].
HC may be more reliable than BPD in determining gestational age when there are variations in skull shape, such as dolichocephaly or
brachycephaly [84].
Cephalic index The fetal cranium may not always display a traditional shape, particularly with breech presentations, oligohydramnios, premature
rupture of the membranes, and neural tube abnormalities. Head compression or distortion from these conditions may result in an abnormal cranial
conformation (eg, dolichocephaly) that lowers test performance of the BPD for gestational age estimation [31,85,86]. In these cases, the cephalic
index (CI) should be measured.
CI refers to the ratio of the BPD and the occipitofrontal diameter (OFD) multiplied by 100 [31]. The standard CI range for normal-shaped craniums
approximates one standard deviation from the mean (>74 or <83) [84]. Therefore, if the CI measurement approaches the outer limits of the normal
range, the use of the BPD for estimation of gestational age is not accurate [84,87]. In these cases, HC (discussed below) is recommended for
cranial assessment because it provides a good estimate of gestational age despite the fetus' irregular cranial structure [84].
Head circumference Measurement of the fetal HC provides a good estimate of gestational age on routine sonograms, but also is useful in the
clinical setting of growth disorders or variations in skull shape when other measures may not perform well [72,80,88-92]. The accuracy is within one
week prior to 20 weeks of gestational age; some studies have even demonstrated superiority when compared with the BPD [72,81,88-91]. As with
other biometric measures, test performance falls in the second half of pregnancy, and by the late third trimester, variability can be within three to
four weeks [72,73,78,80].
The correct plane for the image passes through the thalami and third ventricle, and is similar to the plane for the BPD. However, additional
intracerebral landmarks that must be visualized to obtain the appropriate measurement include the cavum septum pellucidum (image 6) anteriorly
and the tentorial hiatus posteriorly [31]. This view depicts the greatest anterior-posterior length of the cranium and resembles an "arrow" with the
anterior portion appearing as tail feathers, the third ventricle and Sylvian aqueduct as the shaft, and the actual arrowhead comprised of the ambient
and quadrigeminal cisterns and the tentorial hiatus [31]. The standard view should not include the cerebellum or the lateral cerebral ventricles.
The calvarium should always appear symmetrical in the image. HC measurements are obtained by placing the cursors on the outer margins of the
calvarium bilaterally [31], in contrast to BPD measurements, which extend from the inner calvarium on one side to the outer calvarium on the other
side. By using the computerized ellipse function, the ultrasound machine will assist in the approximation of the outer perimeter of the calvarium. It is
important to avoid measuring the outer margin of the skin overlying the scalp since doing so will falsely increase the HC.
Femur length The FL can be measured as early as 10 weeks of gestation because of its size and echogenicity [31]. Correlation with true
gestational age is within one week prior to 20 weeks of gestation, but falls to within 2.1 to 3.5 weeks in the third trimester [81,89,91,93-96].
Even though FL is a simple "one-dimensional" image, errors in measurement often occur [97-99]. The transducer should be aligned along the long
axis of the femoral bone [31]. The proper view is obtained by visualizing either the femoral head or the greater trochanter at the proximal end of the
femur and the femoral condyle at the distal end [97]. The calipers should be placed at the junction of bone and cartilage (image 7) to measure only
ossified bone [97]. They should not contain the femoral head [60]. Including nonossified portions of the femur and not visualizing the full femur
(femoral head/greater trochanter to femoral condyle) are the major sources of error in gestational age assessment by FL. The former overestimates
gestational age and the latter underestimates gestational age.

Average FL appears to vary slightly among ethnic groups. Short femurs may be a normal finding or a marker of aneuploidy (trisomy 21). Severely
shortened (<5th percentile) or abnormal appearing femurs in the second trimester suggest a skeletal dysplasia or early-onset fetal growth restriction
[100,101].
Abdominal circumference The AC appears to have a slightly lower ability to predict gestational age early in the second trimester than the BPD,
HC, and FL [81,89-91]. Some of the variability may be due to error in ultrasound technique, along with natural biologic variations. In the second
trimester, gestational age assessment is accurate within two weeks [102]. In the late third trimester, variability increases, so accuracy falls to within
three to four weeks [81,102]. Due to the wide margin of error, AC is often used for estimations of fetal weight and interval growth evaluations rather
than gestational age assessment. However, the AC can be a valuable additional measure for dating during the second trimester, especially in
fetuses with cranial or limb abnormalities.
Measuring the AC is challenging since the abdomen is not symmetrical, not as echogenic as the cranium and femur, and changes during fetal
breathing and other movement [31]. Therefore, the AC is not as easily measured. The image is taken at the level of the largest diameter of the fetal
liver, denoted by the point of union of the right and left portal veins, which has a "hockey stick" appearance (image 8) [103]. This plane was chosen
because the size of the liver correlates well with overall fetal growth. The correct plane can be confirmed by visualizing the umbilical segment of the
left portal vein in its shortest length. If the cross-section through the fetal abdomen is skewed caudally, this vein will appear in a longer image [31].
Two additional guidelines for correct measurement are: (1) positioning the transducer perpendicularly to the fetal abdominal wall and (2) visualizing
the symmetrical appearance of the lower ribs [31]. The fetal stomach is typically visualized on the AC view. The measurement is taken by placing
four calibration points around the abdomen on the skin edge, not the rib cage (image 9). This will improve test performance while reducing the risk
of underestimation of the AC [31]. The ultrasound machine will approximate the AC. Another method is to use the electronic ellipse function on the
ultrasound machine.
Other biometric parameters Other less-commonly used biometric parameters include but are not limited to the intra- and interorbital diameters
[104], transverse cerebellar diameter (image 10) [105-107], clavicle length [108], foot length [109,110], and length of long bones of the extremities
[31,93,104,108,109,111-113]. These measures tend to have slightly wider ranges of variability in comparison with standard parameters during the
early second trimester, and once again, there is marked variability late in the gestation. Therefore, the value of these additional measurements is
questionable except when fetal anomalies (eg, omphalocele, anencephaly) preclude use of standard biometric parameters.
Use of multiple markers The estimation of gestational age based on one parameter may be significantly different from that calculated from the
others. A thorough evaluation is warranted to determine the etiology of the abnormal measurement. As an example, there may be a fetal
abnormality, such as hydrocephalus, distorting the HC or BPD. Various fetal body ratio indexes (eg, HC/AC, BPD/FL, FL/AC) have been established
across pregnancy to define how these measurements correlate in an average fetus and to determine if a parameter is potentially abnormal
[84,87,114-117]. After a careful assessment to determine potential clinical significance, it is usually appropriate to omit a single discordant parameter
from the gestational age calculation.
Human error and normal variation in fetal size can result in over- or underestimation of gestational age. Regression equations using the four
standard biometric parameters minimize, but do not eliminate these problems and yield the best estimation [89,90]. The ideal biometric assessment
incorporates the least number of images with the highest potential for accuracy. When multiple biometric parameters were compared, the best test
performance with the least variability was obtained with the four standard biometric measures (BPD, HC, FL, and AC) described above [89,91]. A
two biometric parameter (HC and FL) approach showed similar test performance as the four parameter approach for predicting gestational age in
the first half of pregnancy, but variability was slightly higher later in pregnancy [78,91,99]. The addition of a fifth parameter, such as length of the
radius, did not further improve test performance [91]. Therefore, the four measures of BPD, HC, FL, and AC has remained the standard formula for
estimating gestational age in the second and third trimesters.
Additional measurements may be useful when there are biometric discrepancies. If the four routine parameters do not correlate, choosing which
measurements are most accurate for dating purposes can be a difficult task. Therefore, incorporating extra parameters to assist in making an
educated estimate may be useful in this setting. As an example, the transverse cerebellar diameter (in millimeters) correlates with gestational age
up to 22 weeks of gestation [105]. When standard biometry is inconclusive, this additional measurement may help to clarify the situation. However,
as with all parameters discussed, the accuracy of additional measurements is age dependent with significant variability in late pregnancy.
A potential source of bias is when the ultrasound screen displays the calculated gestational age next to each biometric measurement. The
sonographer may adjust the position of the cursors to fit the expected fetal age as opposed to the actual anatomic reference points [31].
SIGNS OF FETAL MATURITY There are several signs suggestive of fetal maturity that can be observed sonographically and correlated with
gestational age when an early ultrasound examination has not been performed or menstrual dates are unknown or uncertain. As an example, the
femoral epiphyseal and proximal tibial ossification centers are well visualized by 32 and 35 weeks of gestation, respectively [118-122]. The proximal
humeral epiphysis also appears in the late third trimester and correlates with fetal lung maturity and gestational age [119,120]. Identification of such
landmarks of maturation of the bony skeleton cannot be incorporated into actual biometric evaluation of gestational age. They are merely visual
aides to help establish fetal maturity in late pregnancy.
ASSIGNING THE ESTIMATED DATE OF DELIVERY The standard deviations of biometric measurements taken from normal pregnancies
increase as pregnancy progresses. Thus, fetal biometric images are most useful for obtaining an accurate estimate of fetal age when performed in
the first trimester but at least by 20 weeks of pregnancy, with a reasonably accurate estimation as late as 28 weeks. The range of error in predicting
delivery dates by ultrasound examinations is shown in the table (table 2) [44,48,50,51,58].
A practical approach to assigning estimated date of delivery (EDD) is to use the sonographically derived gestational age if it differs from that
calculated using the last menstrual period (LMP) by more than 7 days before 16 weeks, by more than 10 days between 16 and 22 weeks, by more
than 14 days between 22 and 28 weeks, and by more than 21 days after 28 weeks (table 3) [71].
Alternatively, the 95% confidence interval (CI) for the gestational age estimate can be derived by using a constant error of 8 percent (two standard
deviations) of the estimate or by reporting variability in terms of specific days (2 to 11 days) [78]. Thus, if the sonographically-based EDD differs
from the menstrually-based EDD by more than 8 percent, then the sonographically-based EDD should be used.
Clinicians should be particularly cautious when changing the EDD based upon measurements obtained from an ultrasound examination performed
late in pregnancy. A single late examination cannot reliably distinguish between a pregnancy that is misdated and younger than expected and a

pregnancy that is complicated by fetal growth restriction. Late pregnancy ultrasound measurements should never be used to change an EDD
established by an examination performed earlier in pregnancy.
After 28 weeks of gestation, it is inappropriate to rely solely on a single set of the usual biometric parameters for fetal dating [31]. When menstrual
dates are discordant with the gestational age suggested by a late pregnancy sonographic examination, serial measurements showing accelerated
or restricted growth over weeks may help to differentiate large for gestational age and growth-restricted fetuses from those with inaccurate
menstrual dating [123]. Thus, serial examinations may provide more accurate information and allow the clinician to make a more reliable clinical
assessment. Because of measurement variability, these serial examinations should be spaced apart by an interval of at least two weeks [68].
In some cases, early fetal growth restriction rather than inaccurate menstrual dating may be the reason for a discrepancy between ultrasound and
menstrual dating (or dating based on conception using assisted reproductive technologies) [78], although this is uncommon [124-126]. When an
EDD is revised because of smaller than expected size of the fetus on a second ultrasound examination, we suggest a follow-up ultrasound to
evaluate fetal growth over time and make sure this finding was not due to early growth restriction.
Assisted reproduction When the conception/fertilization date is known with certainty, as with assisted reproduction, the EDD should be
calculated based on this date. Online calculators and apps are available for calculating EDD and gestational age [127,128].
The patient/clinician enters one of the following dates, as appropriate: date of ovulation, date of egg retrieval, date of insemination, and date of
cleavage stage (day 3) or blastocyst (day 5) transfer.
Assuming the EDD is 280 days from the onset of the last menstrual period (LMP) and 266 days from date of conception in a natural cycle, then in a
fresh in vitro fertilization cycle, the EDD can be calculated by adding 266 days to the date of egg retrieval/fertilization (ie, 14 + 266 = 280). In a cycle
using a frozen cleavage stage embryo, the EDD can be calculated by adding 263 days to the date of embryo transfer to account for three days of
embryo culture (ie, 17 + 263 = 280).
SUMMARY AND RECOMMENDATIONS
The estimated date of delivery (EDD) can be calculated without tools by counting back three months from the last menstrual period (LMP)
and adding seven days. (See 'Naegele's rule' above.)
Ultrasound estimation of EDD in the first half of pregnancy is superior to dating based on LMP or physical examination, and is most
accurate in the first trimester. (See 'Sonographic assessment' above.)
Calculators that provide EDD and current gestational age based on ultrasound examination and/or LMP are widely available (calculator
1 and calculator 2) and are more accurate than traditional mechanical gestational wheels. (See 'Calculator' above.)
The sonographically-derived EDD should be used if it differs from that calculated using the LMP by more than 5 to 7 days in the first
trimester and by more than 10 to 14 days in the second trimester or by 8 percent. (See 'Assigning the estimated date of delivery' above.)
The gestational sac is the earliest sonographic landmark of pregnancy and can be visualized at 4.5 to 5 weeks by transvaginal (TVS)
ultrasound (table 4). Measurement of the gestational sac diameter (MSD) allows estimation of gestational age. The sum of 30 and the sac
size in millimeters is equivalent to the gestational age in days. (See 'Gestational sac' above.)
The embryonic/fetal pole can be visualized at about 6 weeks of gestation. Measurement of crown-rump length (CRL) at 7 to 10 weeks of
gestation is the most accurate biometric parameter for pregnancy dating (3 days). After 10 weeks, the accuracy of the CRL falls slightly
with a margin of error of 5 days at 10 to 14 weeks and 8.4 days by the 15th week. CRL should be used to estimate gestational age up to
84 mm (14 weeks of gestation), and the biparietal diameter (BPD) should be used for CRL measurements >84 mm. (See 'Crown-rump
length' above.)
After the first trimester, the preferred biometric measurements for assessment of gestational age are a combination of BPD, head
circumference (HC), femur length (FL), and abdominal circumference (AC). Prior to 16 weeks, the accuracy of BPD, HC, FL, and AC for
estimating gestational age is 7 days, but this gradually falls to three to four weeks near term (table 2). (See 'Use of multiple
markers' above.)
In pregnancies with an unknown or uncertain LMP and late first sonogram, ossification of the distal femoral epiphysis suggests a fetal age
of at least 32 weeks, and ossification of the proximal tibial and humeral epiphyses suggests a fetal age of at least 35 weeks. (See 'Signs of
fetal maturity' above.)
Sonographic examination initially performed in the third trimester does not reliably estimate gestational age because fetal biometry does
not correlate well with gestational age late in gestation. Serial measurements over two to three weeks may be helpful. Above average
growth between exams suggests a large for gestational age fetus who may be less far along in gestation than predicted by biometry,
whereas no or minimal growth between examinations suggests a growth restricted fetus who may be further along in gestation than
predicted by biometry. (See 'Assigning the estimated date of delivery' above.)

Twin pregnancy: Labor and delivery

http://www.uptodate.com/contents/twin-pregnancy-labor-and-delivery?
source=search_result&search=ultrasonographic+on+pregnancy&selectedTitle=21%7E150Twin pregnancy: Labor and delivery
Authors
Stephen T Chasen, MD
Frank A Chervenak, MD
Section Editor
Charles J Lockwood, MD, MHCM
Deputy Editor
Vanessa A Barss, MD, FACOG

Disclosures: Stephen T Chasen, MD Nothing to disclose. Frank A Chervenak, MD Nothing to disclose. Charles J Lockwood, MD,
MHCM Consultant/Advisory Boards: Celula [Aneuploidy screening (Prenatal and cancer DNA screening tests in development)]. Equity Ownership/Stock
Options: Celula [Aneuploidy screening (Prenatal and cancer DNA screening tests in development)].Vanessa A Barss, MD, FACOG Nothing to disclose.
Contributor disclosures are reviewed for conflicts of interest by the editorial group. When found, these are addressed by vetting through a multi-level review
process, and through requirements for references to be provided to support the content. Appropriately referenced content is required of all authors and must
conform to UpToDate standards of evidence.
Conflict of interest policy
All topics are updated as new evidence becomes available and our peer review process is complete.
Literature review current through: Jun 2015. | This topic last updated: Jun 26, 2015.
INTRODUCTION Twin pregnancies are at increased risk of intrapartum complications, such as fetal heart rate abnormalities and complications related to
malpresentation. This topic will review issues related to the delivery of twins, such as timing and route of delivery, fetal monitoring,
maternal analgesia/anesthesia, and management of delivery. Antepartum assessment and neonatal management of these pregnancies, and monoamniotic twin
pregnancy, are discussed separately. (See"Twin pregnancy: Prenatal issues" and "Monoamniotic twin pregnancy" and "Neonatal outcome, complications, and
management of multiple births".)
TIMING Spontaneous or medically indicated preterm birth complicates over 50 percent of twin pregnancies, thus scheduling the timing of delivery is not
subject to the discretion of the obstetrician in most cases [1]. However, in the absence of a spontaneous or medically indicated preterm delivery, when should
ongoing twin pregnancies be delivered?
In general, we agree with expert consensus guidelines for timing of delivery of twin gestations. A consensus opinion from a workshop held by the Eunice Kennedy
Shriver National Institute of Child Health and Human Development (NICHD) and the Society for Maternal-Fetal Medicine (SMFM) suggested delivery at 38
weeks for uncomplicated dichorionic twins and 34 to 37 weeks for uncomplicated monochorionic diamniotic twins [2]. The National Institute for Health and
Clinical Excellence (NICE) recommends elective delivery of dichorionic twin pregnancies from 37 weeks of gestation and monochorionic twin pregnancies from
36 weeks (after a course of corticosteroids) [3]. The American College of Obstetricians and Gynecologists (ACOG) recommends delivery of uncomplicated
dichorionic twins at 380/7ths to 386/7ths weeks and uncomplicated monochorionic twins at 340/7ths to 376/7ths weeks of gestation [4,5]. Twin pregnancies complicated by
fetal growth restriction are delivered earlier, with timing depending on the clinical scenario.
Dichorionic twin pregnancy For normal uncomplicated dichorionic/diamniotic twin pregnancies, we suggest elective delivery between 38 and 40 weeks of
gestation.
There are no high quality data from randomized trials on which to base a recommendation for the optimum timing of delivery of dichorionic/diamniotic twins.
Based on our experience and the data presented below, we and others avoid elective delivery (ie, in the absence of usual maternal or fetal indications for delivery)
of dichorionic/diamniotictwins prior to 38 weeks of gestation [5-8]. Assessment of fetal lung maturity should be performed if elective delivery is considered prior
to 38 weeks [9].
In the presence of complications associated with increased perinatal mortality, such as oligohydramnios, abnormal fetal growth, or maternal hypertension, delivery
prior to 38 weeks without confirmation of lung maturity is reasonable.
Optimal length of gestation The optimal length of gestation may be shorter in twin than in singleton pregnancies. Epidemiological evidence shows that the
lowest rate of perinatal mortality (PNM) occurs at an earlier gestational age and at lower birth weight in twins than in singletons [10-12]. This was illustrated in
two large analyses:
A retrospective study comparing PNM among 89,000 infants born of multifetal pregnancies to PNM in six million singleton infants reported that PNM
in twin gestations reached a nadir at 37 to 38 weeks of gestation and then increased, while the lowest PNM in singleton pregnancies was at 39 weeks [11].
Data from the United States National Center for Health Statistics, including over 11 million singleton and almost 300,000 twin births, showed that the
lowest PNM for singletons was at 39 to 41 weeks (approximately 1 fetal and 1 neonatal death per 1000 deliveries), but for twins lowest PNM was at 37 to
39 weeks (3 to 4 fetal and 2 neonatal deaths per 1000 deliveries) [13]. Moreover, as of 39 weeks gestation, the prospective risk of fetal death in an ongoing
twin pregnancy exceeded the risk of neonatal death.
Additional smaller series have confirmed these findings. However, the only two randomized trials of elective delivery at 37 weeks versus expectant management
of uncomplicated twin pregnancy had too few patients (n = 36, n = 235) to detect significant benefits or harms [14,15], even when the data were combined [16].
When all twin pregnancies are considered, the lowest perinatal mortality for twin pregnancy appears to be at about 37 to 38 weeks of gestation [17-21], while
delivery before 38 weeks is associated with higher neonatal morbidity than later delivery [18,22,23]. However, when only uncomplicated dichorionic twin
gestations are considered,morbidity/mortality falls through 39 weeks of gestation [8]. Therefore, in uncomplicated pregnancies and when the twins are continuing
to grow and mature, we and others feel it is reasonable to deliver at or after 38 completed weeks of gestation to minimize the frequency of neonatal complications
[5,8,22-25]. However, there is no consensus on the optimum time to deliver these pregnancies.
Assessment of pulmonary maturity Although there is widespread belief that fetuses of twin gestations experience more rapid pulmonary maturation than
singleton fetuses [26], this assertion has been challenged by conflicting data [27-29]. The potential for neonatal respiratory problems is particularly important when
considering the optimal timing of delivery of twins since many of these pregnancies are delivered by cesarean birth prior to the onset of labor [30,31].

As with singletons, there appears to be a significant risk of respiratory problems in late preterm twins:
One study evaluated the prevalence of neonatal respiratory disorders in the infants of 126 women with twins who underwent elective cesarean delivery
before labor (at 36 to 40 weeks of gestation) [6]. The majority (65 percent) of cesarean deliveries were performed due to malpresentation in one or both
fetuses; there were no maternal or fetal complications (eg, severe preeclampsia, fetal growth restriction) requiring medically indicated delivery. Neonatal
respiratory disorders occurred frequently, and more often among infants born at 36 to 37 compared to 38 or more weeks of gestation (13 versus 2 percent).
In another series of twins delivered by the vaginal or abdominal route, the prevalence of respiratory morbidity at 36 to 37 weeks was 23 percent,
compared to 7 percent at 37 to 38 weeks [32].
Therefore, fetal pulmonary maturity should be evaluated if elective early delivery is planned. Some authors, including ourselves, feel that amniocentesis of only
one twin is adequate if the gestation is 36 weeks [26,33], but others test both twins in all cases because pulmonary maturity can be asynchronous [34].
Based on review of these data [6,33,34], the American College of Obstetricians and Gynecologists suggested amniocentesis before elective delivery of twins less
than 38 and 0/7ths weeks of gestation [9]. They state that amniocentesis of only one twin is generally sufficient at a gestational age 33 0/7ths weeks since
discordancy is uncommon at this time [33,34], but they suggest amniocentesis of both twins when the procedure is performed earlier in pregnancy. When only one
sac is sampled, it would be reasonable to sample the sac of the fetus less likely to be mature. As an example, a non-presenting, larger, male fetus would be less
likely to have achieved lung maturity than a smaller, presenting female fetus. (See "Assessment of fetal lung maturity".)
Upper gestational age limit In the rare twin pregnancy in which labor has not occurred by 40 weeks of gestation and no fetal or maternal indications for
delivery are present, we electively intervene to deliver the twins. There are few studies evaluating the upper gestational age limit for allowing a twin pregnancy to
remain undelivered. One such series included over 60,000 twin pairs delivered at 37 or more weeks of gestation [24]. The odds of neonatal death (unrelated to
congenital anomalies) significantly increased at 40 weeks compared with 37 weeks of gestation (OR 4.24, 95% CI 2.65-7.00). NICE guidelines state continuing
uncomplicated twin pregnancies beyond 38 weeks 0 days increases the risk of fetal death [3].
Monochorionic twin pregnancy
Diamniotic There are no randomized trials that have evaluated the optimum time for delivery of monochorionic/diamniotic twin gestations. Based on the
higher risk of intrauterine mortality in these pregnancies [35], some experts have argued for delivery as early as 32 weeks of gestation [2,36-39]. While it is not
clear that elective delivery prior to 37 weeks of gestation is warranted in apparently uncomplicated monochorionic/diamniotic twin pregnancies [40,41], it is our
practice, and that of other experts [2,8,25,42-45], to discuss the increased risk of intrauterine demise and offer early delivery of monochorionic/diamniotic twins;
we suggest delivery at 36 to 37 weeks of gestation. Delivery at about this time provides a reasonable trade-off between the morbidity of preterm birth, which falls
with advancing gestation age, and the risk of unanticipated subsequent fetal demise. In one study, the perinatal morbidity rate and the prospective risk of fetal
death at 36 to 37 weeks were 9 and 1 percent, respectively, in closely monitored pregnancies [8]. We do not perform amniocentesis to document lung maturity
prior to delivery at 36 to 37 weeks inmonochorionic/diamniotic twins.
The American College of Obstetricians and Gynecologists suggests delivery of monochorionic twins at 34 0/7 to 376/7 weeks of gestation [5] and the North American
Fetal Therapy Network suggests delivery at 360/7 to 376/7 weeks of gestation [45].
Monochorionic/diamniotic twin pregnancies with twin-twin transfusion syndrome are delivered earlier. (See "Twin-twin transfusion syndrome: Management".)
Monoamniotic Additional issues must be addressed in timing the delivery of a monoamniotic twin pregnancy. Delivery earlier in the third trimester may be
indicated because of the high rate of perinatal mortality described in these pregnancies despite intensive fetal surveillance, (30 to 70 percent, which is likely due to
cord entanglement) [46,47]. Management of monoamniotic twin pregnancy is discussed in detail separately. (See "Monoamniotic twin pregnancy".)
DELIVERY ROUTE
Dichorionic twin pregnancy Whether route of delivery affects morbidity and mortality is controversial. The presentation and, in some situations, the
gestational age of twins can influence the selection of vaginal versus cesarean delivery. However, in most cases, women with twin pregnancies with the presenting
twin in cephalic presentation should be counseled to attempt vaginal delivery [48].
Should routine cesarean delivery be offered? Cesarean delivery is common in twin pregnancies: over 60 percent of twin births are by cesarean [49]. Some
investigators have proposed that neonatal outcome could be improved by a policy of routine cesarean delivery for all twin pregnancies. The basis for this proposal
is concern that the relative risk of anoxic death of the second twin is increased as a result of mechanical problems (eg, compound presentation, cord prolapse,
abruptio placentae) after vaginal birth of the first twin [50]; this is most significant in gestations 36 weeks [51]. However, others have pointed out that even
though the relative risk of neonatal mortality may be increased, the absolute risk remains low and, therefore, a large number of cesareans would have to be done to
prevent one death of a second twin [52]. Cesarean delivery also poses risks to the mother. (See "Cesarean delivery: Postoperative issues", section on
'Complications' and "Cesarean delivery: Postoperative issues", section on 'Long-term risks'.)
We do not favor a policy of planned cesarean delivery of all twin pregnancies. With appropriate intrapartum monitoring and management (see below), the second
twin is not at high risk of neonatal mortality or morbidity with trial of labor, even when remote from term [53-58]. (See 'Should VLBW fetuses be delivered by
cesarean?' below.)
We choose a delivery route based upon presentation and amnionicity (see below), as well as the presence/absence of standard obstetrical indications for cesarean
delivery (eg, placenta previa). We favor a trial of labor for concordant diamniotic twins with the first twin in vertex presentation.
The following evidence supports this approach:

In a 2011 systematic review, for twins with vertex/vertex presentation, vaginal delivery was safer than cesarean for the first twin, and was as safe as
cesarean for the second twin [54]. In pregnancies with vertex/nonvertex presentation, a trial of labor was a safe option in the absence of risk factors that
increased the risk of a cesarean delivery of the second twin after vaginal delivery of the first twin, although predicting combined delivery was essentially
impossible antepartum. The morbidity rates for planned vaginal and planned cesarean delivery were 15.1 and 7.4 percent, respectively (OR 1.11, 95% CI
0.65-1.88); mortality rates were 0.7 and 0.1 percent, respectively (OR 3.04; 95% CI 0.37-25.2). Given the wide confidence intervals, this analysis is
inadequate for conclusively determining which route of delivery should be offered.
Subsequently, the Twin Birth Study randomly assigned 1398 women between 320/7ths and 386/7ths weeks of gestation with twin pregnancy with the first
twin in cephalic presentation to planned cesarean or planned vaginal delivery (cesarean only if indicated) [59]. Elective delivery was planned between
375/7ths and 386/7thsweeks of gestation. The rate of cesarean delivery was 90.7 percent in the planned cesarean delivery group and 43.8 percent in the planned
vaginal delivery group, illustrating the high frequency of cesarean even when vaginal delivery is planned. The study found no significant difference in the
composite outcome (fetal or neonatal death or serious neonatal morbidity) between the planned cesarean and the planned vaginal delivery groups (2.2
versus 1.9 percent, respectively; odds ratio 1.16; 95% CI 0.77-1.74). This trial represents the best available evidence that planned cesarean delivery does
not significantly improve neonatal outcome as compared with planned vaginal delivery.
Should VLBW fetuses be delivered by cesarean? Some investigators have suggested that cesarean delivery may decrease the risk of intracranial hemorrhage
in very low birth weight (VLBW) preterm twin fetuses, regardless of presentation [60,61]. We and others have not found adequate evidence to recommend a policy
of elective cesarean delivery of VLBW babies [58,62]. (See "Delivery of the preterm low birth weight singleton fetus".)
Effect of fetal presentation
Vertex-vertex twins This presentation accounts for approximately 42 percent of twins [63]. The general consensus is that a trial of labor with the goal of a
vaginal delivery of vertex-vertex twins is appropriate at any gestational age [7,64].
Nonvertex presenting twin A nonvertex first twin comprises approximately 20 percent of twin presentations. Breech presenting twins are paired with vertex
twins (7 percent), breech twins (6 percent), and transverse twins (5 percent); breech-oblique, transverse-vertex, and transversetransverse presentations each
comprise fewer than 1 percent of twins. A unique, potential complication of breech-vertex twin delivery, as opposed to breech singleton delivery, is the possibility
of interlocking chins (ie, locked twins), but this is rare.
A retrospective case-control analysis of data from 13 centers that allowed vaginal birth for breech first twins reported no difference in Apgar scores and neonatal
mortality related to route of delivery among infants weighing more than 1500 grams [65]. However, nonvertex presenting fetuses weighing less than 1500 grams
that were delivered vaginally had significantly higher rates of low Apgar scores and neonatal mortality compared to those delivered by cesarean (37 versus 20
percent and 45 versus 8 percent, respectively).
In contrast to this report, we and others suggest cesarean delivery when the first twin is not in the vertex presentation because the safety of vaginal delivery in
these cases has not been confirmed by randomized trials and the general consensus in the obstetric community is against vaginal delivery of the breech presenting
fetus, even in singleton pregnancies [64,66]. (See "Delivery of the fetus in breech presentation".) An exception is the breech presenting second twin (see below).
Vertex-nonvertex twins Vertex-nonvertex twins comprise 38 percent of twins; the nonvertex twin may be breech (26 percent), transverse (11 percent), or
oblique (1 percent). Options for delivery include cesarean delivery of both twins, vaginal delivery with breech extraction of the second twin, or vaginal delivery
with cephalic version of the second twin. We suggest a trial of labor, offering the patient an attempt at breech extraction of the second twin and proceeding to
cesarean delivery if unsuccessful.
The only randomized trial of planned vaginal versus abdominal birth was performed in 60 women with vertex-nonvertex twins at 35 to 42 weeks of gestation and
otherwise uncomplicated pregnancies [67]. Maternal morbidity was higher with planned cesarean delivery, with no difference in neonatal outcome. However, this
trial was too small to detect clinically important differences in outcome between the two groups.
Many studies have reported successful vaginal delivery of both twins using internal or external version or breech extraction of the second twin [68-84]. The second
twins of vertex-nonvertex and vertex-vertex pairs generally had similar neonatal outcomes irrespective of mode of delivery or procedures performed during
delivery, but these series were small and observational. Successful vaginal delivery appeared to be less likely when external version was attempted than when
breech extraction was performed immediately after delivery of the presenting twin [70,71,78,82]; external version was completed in 40 to 50 percent of cases (the
remainder were delivered by cesarean delivery), while breech extraction followed by vaginal birth succeeded in 96 to 100 percent of patients [70,71]. Of note, the
mean gestational age was approximately 34 to 35 weeks and the mean birthweight was 2100 to 2200 grams in these studies; 7 to 15 percent of the neonates
weighed less than 1500 grams.
In contrast, a population based cohort study of twin deliveries in the United States reported infants weighing 1500 to 4000 grams had a significantly higher
frequency of neonatal death (adjusted for anomalies), asphyxia, and injury when both twins of a vertex-nonvertex pair were delivered vaginally than when both
twins were delivered by cesarean [85]. Interpretation of these data is limited because it was obtained from birth certificates.
In the absence of good data favoring one approach over another, we suggest offering the patient an attempt at breech extraction of the second twin and proceeding
to cesarean delivery if unsuccessful. If the patient does not wish to attempt breech extraction of the second twin, we give her the option of attempting external
cephalic version of the second twin or undergoing cesarean delivery of both twins. When discussing the options of breech extraction or external cephalic version
with patients, the obstetrician should include information about his or her experience and comfort with these procedures. Many obstetricians, based on training and
experience, may feel more comfortable performing cesarean delivery. Under these circumstances, cesarean delivery of both twins is recommended.
Contraindications to breech extraction We generally do not offer the option of breech extraction when:
The estimated fetal weight of the second twin is 20 percent more than that of the presenting twin.

The delivery of the presenting twin suggests that the pelvis may not be adequate for a breech delivery, such as when there is a prolonged second stage or
marked molding of the head.
The gestational age is less than 28 weeks or the estimated fetal weight of the second twin is less than 1500 grams. Under these circumstances, we
recommend performing cesarean delivery of both twins, rather than attempting cephalic version of the nonvertex second twin.
Trial of labor after previous cesarean delivery There are insufficient data to establish the safety of attempting vaginal birth of twins after a previous low
transverse cesarean delivery. Available data are reassuring that outcomes are similar to those in women with singletons undergoing a trial of labor. One of the
largest series reported uterine rupture in 16 of 1850 women with twins (0.9 percent) undergoing a trial of labor after previous cesarean delivery; this rate was
comparable to that in singleton pregnancies undergoing trial of labor (0.8 percent) [86]. Successful vaginal delivery was achieved in 45 percent of the twin
gestations and 62 percent of the singletons. Smaller series have reported similar findings [87-93]. (See "Choosing the route of delivery after cesarean birth".)
It is our practice to offer a trial of labor to women with twin pregnancies and one prior cesarean delivery, provided they go into spontaneous labor. Because the
most common initial sign of uterine rupture is fetal heart rate changes, we continually monitor both fetuses. If this is not technically possible, then cesarean
delivery is performed.
Monochorionic twin pregnancy
Diamniotic Monochorionic placentation is not a contraindication to a trial of labor and vaginal birth [42,94-96]. Decision-making about route of delivery is the
same as that described above for dichorionic twins: Cesarean delivery is performed when the presenting twin is not in vertex presentation and for standard
obstetric indications.
Acute peripartum twin-twin transfusion syndrome has been reported in 1.8 to 5.5 percent of monochorionic twin deliveries [97,98]. Although labor and vaginal
delivery may be risk factors, we believe this observation should not influence choice of delivery route until more data are available.
Monoamniotic Monoamniotic twins are usually delivered by cesarean. The rationale for this approach and other aspects of monoamniotic twin pregnancy are
reviewed separately. (See "Monoamniotic twin pregnancy".)
MANAGEMENT Management of labor and delivery is the same for dichorionic and monochorionic diamniotic twins.
Labor It is not clear whether multiple gestation has an effect on the progress of labor; studies have reported conflicting results (ie, both faster and slower
progression than singleton labor) [99-101]. Oxytocin for augmentation or induction of labor appears to be effective; there are inadequate data to allow evaluation
of safety [102-105].
Electronic fetal heart rate monitoring Multiple gestations are at increased risk of intrapartum complications; therefore, we monitor both twins continuously
during labor. Intermittent auscultation is not practical and may not reliably distinguish one twin from the other.
The fetal heart rate of each twin can be monitored using a single machine (waveform 1). These rates are often synchronous, thus requiring frequent careful review
of the tracing to make sure each fetus' heart rate is being monitored. Ultrasound can assist in ensuring that both fetal heart rates are traced. If separate monitors are
used, internal clocks must be synchronized, paper speeds must be identical, and contractions must be displayed on both tracings.
Electronic fetal heart rate monitoring is particularly useful for assessing the well-being of the second twin during the high risk period after delivery of the first twin
(see'Interval between delivery of the two twins' below).
Analgesia and anesthesia Epidural analgesia/anesthesia is generally recommended because it provides good pain relief, does not cause neonatal depression,
and is a suitable anesthetic if uterine manipulation (eg, version) or operative delivery (eg, forceps, cesarean) becomes necessary.
First twin For a variety of reasons, morbidity and mortality tend to be lower in first born than second born twins, regardless of route of delivery. A systematic
review of observational studies reported the overall neonatal morbidity of first and second twins was 3.0 and 4.6 percent, respectively (OR 0.53, 95% CI 0.390.70), and overall neonatal mortality was 0.3 and 0.6 percent, respectively (OR 0.55, 95% CI 0.38-0.81) [54].
When vaginal birth is attempted, the capacity for immediate cesarean delivery is important in the event that complications necessitating urgent delivery arise (eg,
prolapsed umbilical cord, nonreassuring fetal heart rate, or failed breech extraction/cephalic version). Emergency cesarean delivery has been reported in 10 to 30
percent of pregnancies in which vaginal births of twins had been planned, and may be necessary for delivery of the second twin after vaginal birth of the first twin
(see below). It is our practice to deliver all twin pregnancies in an operating room where cesarean delivery can be performed. In our hospital, patients are moved to
the operating room in the second stage of labor. However, if a patient can be transported to an operating room rapidly, delivery of vertex-vertex twins in a labor
room is not unreasonable.
Delivery of the first twin of a diamniotic pair is similar to delivery of a singleton except the umbilical cords should be marked with progressive numbers of clamps
(eg, one for the first twin birth, two for the second twin birth). If surgical clamps are used initially, they should be replaced with a like number of plastic umbilical
cord clamps prior to sending the placenta for formal examination. Recall that 'twin A' on ultrasound may not be first born at delivery (especially if the delivery is
by cesarean), and this infant is typically called 'baby A' by delivery room and nursery personnel [106-109].
In monochorionic twins, signs of acute peripartum twin-twin transfusion syndrome include bradycardia or a sinusoidal fetal heart rate pattern, and may necessitate
urgent delivery.

In all monochorionic twins, it is important to promptly clamp the umbilical cord after delivery of the first twin as the in utero co-twin could develop hypovolemic
shock from exsanguinating into the placenta and out the unclamped cord of the first twin [45]. Monochorionic twins are not good candidates for delayed cord
clamping.
Second twin Second born twins have been reported to have a higher incidence of adverse outcome (morbidity and mortality) due to lower birth weight; higher
frequency of malpresentation, cord prolapse, and abruptio placentae; and more deliveries involving internal podalic version [50,110-113].
After delivery of the first twin, the heart rate and position of the second twin should be evaluated using ultrasound and electronic fetal monitoring. If the second
twin is in a nonvertex presentation, ultrasound can be used to assist external cephalic version, breech extraction, or internal podalic version of the second twin, if
necessary.
As discussed above, our preference is breech extraction if the second twin is not in a vertex presentation and there are no contraindications to breech extraction
(see'Vertex-nonvertex twins' above). Intrauterine manipulation is aided by ultrasonographic visualization of the orientation between the physician's hands and fetal
parts (figure 1A-B) [114] and can be facilitated by administering intravenous nitroglycerin or inhalational anesthesia, which relax uterine muscle [115]. Effective
maternal analgesia is also crucial. When the fetus is in the desired presentation for delivery, oxytocin is given if labor has not resumed. Amniotomy is avoided until
after the presenting part is engaged [55,116].
Others have taken a somewhat different approach. One historic cohort study of 130 planned vaginal twin deliveries reported no patient who had a vaginal delivery
of the first twin required cesarean delivery of the second twin [84]. In this group's practice, all patients who were undelivered at 38 weeks of gestation underwent
induction if they met strict criteria (estimated weight of the second twin 1500 g and no more than 20 percent greater than the weight of the presenting twin,
absence of usual contraindications to vaginal delivery). After vaginal birth of the first twin, the second twin was delivered as a vertex presentation if the vertex was
engaged immediately after delivery of the first twin, by breech extraction if in breech presentation after delivery of the first twin, and by breech extraction after
internal podalic version if in vertex presentation but unengaged after delivery of the first twin. The authors attributed their success to active management of the
second stage of labor by obstetricians experienced in breech delivery and internal podalic version, and to their criteria for selecting candidates for vaginal delivery.
Interval between delivery of the two twins Historically, a prolonged interval between delivery of the first and second twins was thought to be associated with
poorer outcomes. Intervals of less than 25 to 30 minutes have been advocated, and maneuvers such as internal podalic version or breech extraction of the second
twin have often recommended to hasten delivery [117-119].
Subsequent studies undertaken after the universal routine use of electronic fetal monitoring during labor suggest that there does not have to be a finite interval
between delivery of the first and second twin, as long as the fetal heart rate tracing is reassuring [120-122]. Electronic fetal monitoring and the availability of realtime ultrasound have enabled obstetricians to identify those second twins who would benefit from expedited delivery, while allowing most cases to be managed
expectantly [121]. Thus, spontaneous delivery of both twins is more likely to be achieved.
Oxytocin augmentation of labor after delivery of the first twin is reasonable and sometimes necessary due to a temporary reduction in contraction frequency after
the first birth [7].
Delayed-interval delivery in previable gestations is discussed separately. (See "Delayed-interval delivery in multifetal pregnancy".)
Unplanned cesarean delivery An unplanned cesarean for delivery of the second twin is not uncommon, occurring in about 4 to 10 percent of planned vaginal
births [123-125]. As an example, a population based cohort study of twin deliveries in the United States reported 9.5 percent of second twins were delivered by
cesarean after vaginal birth of the first twin [123]. This rate fell to 6.3 percent if the second twin was vertex [126], but increased to 24.8 percent if only vertexnonvertex live births were considered [85].
A systematic review and meta-analysis including over 39,000 twins sets found that for the second twin, morbidity following combined delivery (19.8 percent) was
higher than after vaginal delivery (9.5 percent; OR 0.55, 95% CI 0.41-0.74) or cesarean delivery (9.8 percent; OR 0.47, 95% CI 0.43-0.53) [54]. Factors related to
unplanned cesarean for delivery of the second twin likely account for the increased morbidity. In combined deliveries, there may be an increased risk of maternal
or neonatal infection due to exposure to labor and ruptured membranes.
EXAMINATION OF THE PLACENTA Examination of the placenta can help to determine zygosity and the pathogenesis of neonatal findings (eg, discordant
growth, structural anomalies, or infection). (See "Gross examination of the placenta" and "The placental pathology report".)
Monozygotic or dizygotic? Approximately two-thirds of naturally-conceived twins are dizygotic. It is of importance to parents and twins to know whether
same sex twins are monozygotic (in lay terms, "identical"). Based upon a genotype and placental study of 668 consecutive twin pairs in Birmingham, England,
parents of naturally-conceived twins can be informed in the delivery room that 37 percent of all same sex twins are "identical" [127]. Same sex twins are virtually
always "identical" if monochorionic, while 18 percent of same sex twins with dichorionic membranes are "identical" (figure 2).
These proportions do not apply to pregnancies conceived by in vitro fertilization. Interestingly, in these pregnancies the rate of monozygotic twinning appears to be
higher than in spontaneously conceived pregnancies (2.3 versus 0.4 percent of pregnancies) [128]. In addition, there are several reported cases of dizygotic twins
with monochorionic placentation [129-134]. Blood studies in these twins demonstrate chimerism. The pathogenesis of monochorionic dizygotic twinning has not
been explained, but may be related to assisted reproductive technology, although the occurrence has also been reported in naturally-conceived pregnancies [133].
Many cases of monochorionic dizygous twins are probably unrecognized because the newborns have the same sex. When counseling parents of monochorionic
twins of discordant sex, both the possibility of dizygosity and a disorder of sexual differentiation in one of the twins should be addressed.
Zygosity can be determined conclusively using blood type or DNA markers [135].
Vascular anastomoses In monochorionic twins, vascular anastomoses are evaluated by placental injection. This study takes about an hour to perform and
involves catheterizing the arteries and vein of each umbilical cord and injecting the vessels with a substance, such as dyed undiluted barium sulfate, until the

peripheral branches are filled and backpressure prevents further injection [136]. The amniotic membranes are then removed from the chorionic surface and the
placenta is rinsed with cold tap water to improve visualization of the number and type of anastomoses. Several arteriovenous and venoarterial anastomoses in
combination with an arterioarterial and/orvenovenous anastomosis are seen in 90 percent of placentas. The remaining 10 percent are equally divided between those
with only arteriovenous anastomoses and those with no anastomoses [137,138]. (See "Twin-twin transfusion syndrome: Pathogenesis and diagnosis", section on
'Pathophysiology'.)
INFORMATION FOR PATIENTS UpToDate offers two types of patient education materials, The Basics and Beyond the Basics. The Basics patient
education pieces are written in plain language, at the 5th to 6th grade reading level, and they answer the four or five key questions a patient might have about a given
condition. These articles are best for patients who want a general overview and who prefer short, easy-to-read materials. Beyond the Basics patient education
pieces are longer, more sophisticated, and more detailed. These articles are written at the 10 th to 12th grade reading level and are best for patients who want in-depth
information and are comfortable with some medical jargon.
Here are the patient education articles that are relevant to this topic. We encourage you to print or e-mail these topics to your patients. (You can also locate patient
education articles on a variety of subjects by searching on patient info and the keyword(s) of interest.)
Basics topics (see "Patient information: Having twins (The Basics)")
SUMMARY AND RECOMMENDATIONS
We suggest avoiding elective delivery of dichorionic/diamniotic twins prior to 38 weeks or after 40 weeks of gestation (Grade 2C). We suggest delivery
ofmonochorionic/diamniotic twins at 36 to 37 weeks of gestation (Grade 2C). (See 'Timing' above.)
For vertex-vertex twins, we suggest vaginal delivery in the absence of standard indications for cesarean delivery (Grade 2B). (See 'Vertex-vertex
twins' above.)
When the first twin is not in vertex presentation, we suggest cesarean delivery (Grade 2C). (See 'Nonvertex presenting twin' above.)
For vertex-nonvertex twins, we suggest breech extraction of the second twin only if the obstetrician has the requisite experience and if the patient
provides informed consent (Grade 2C). (See 'Vertex-nonvertex twins' above.)
Available data are reassuring that outcomes in women with twins attempting vaginal birth after a previous cesarean delivery are similar to those with
singletons undergoing a trial of labor. However, these data are insufficient to definitively establish that uterine rupture rates are comparable. (See 'Trial of
labor after previous cesarean delivery' above.)
Oxytocin for augmentation or induction of labor appears to be effective in twin gestations; there are inadequate data to establish the safety of this
intervention. (See'Labor' above.)
We perform continuous electronic fetal monitoring of both fetuses during labor (See 'Electronic fetal heart rate monitoring' above.)
We suggest epidural analgesia/anesthesia during labor (Grade 2C). (See 'Analgesia and anesthesia' above.)
After delivery of the first twin, the heart rate and position of the second twin should be evaluated using ultrasound and electronic fetal monitoring. As
long as the fetal heart rate tracing is reassuring, there is no duration of elapsed time from delivery of the first twin that necessitates intervention to deliver
the second twin. Six to 25 percent of second twins will be delivered by cesarean after vaginal delivery of the first twin. (See 'Second twin' above.)
Approximately 18 percent of naturally-conceived same sex twins with dichorionic membranes are "identical." (see 'Monozygotic or dizygotic?' above).
Use of UpToDate is subject to the Subscription and License Agreement.