The First Cesarean:

Role of “Fetal Distress” Diagnosis

Maged M. Costantine, MD, and George R. Saade, MD

The goal of this review is to synthesize the available information regarding factors leading
to the first cesarean, especially as it relates to the role of electronic fetal heart rate
monitoring (EFM) in the prevention of the first cesarean. The widespread use of EFM did
not reduce perinatal or neonatal adverse outcomes. Coupled with its inherent problems,
such as poor sensitivity and reliability, this has led to a dramatic increase in cesarean
delivery rates. To be realistic in any attempt to lower the current cesarean rate, we must
address the inherent limitations in our current EFM methods.
Semin Perinatol 36:379-383 © 2012 Elsevier Inc. All rights reserved.

KEYWORDS cesarean delivery, electronic fetal heart rate monitoring, perinatal and neonatal
adverse outcomes, fetal pulse oximetry, fetal blood sampling, fetal electrocardiogram

The rate of cesarean delivery (CD) has been progressively increasing over the past 3 decades owing to an increase in the rate of primary CD
coupled with disappointingly low
rates of vaginal birth after CD. Therefore, it is becoming clear that the best approach to decreasing the CD rates is to prevent the
first CD. For this reason, the Society for Maternal-Fetal Medicine (SMFM), the Eunice Kennedy Shriver National Institute of Child
Health and Human Development, and the American College of Obstetricians and Gynecologists (ACOG) cosponsored a workshop
during the SMFM annual meeting in February 2012 entitled “Prevention of the First Cesarean.” The goal of this review is to
synthesize the avail-able information regarding factors leading to the first CD, especially as it relates to the effect of electronic fetal
heart rate monitoring (EFM) on the rates of first CD.

The Unrealized Potential of EFM

1,2
EFM was introduced into obstetrical practice in the 60s, pri-marily to monitor complicated pregnancies. The original goal of EFM
1,2
was primarily to prevent intrapartum fetal death and significant perinatal morbidity, including neurologic damage. Although the
original purpose did not apply to all pregnancies, by the late 70s, it was used to monitor about

Division of Maternal-Fetal Medicine, Department of Obstetrics and Gyne-cology, The University of Texas Medical Branch, Galveston, TX.
Address reprint requests to Maged M. Costantine, MD, Division of Maternal-Fetal Medicine, Department of Obstetrics and Gynecology, The Univer-sity of
Texas Medical Branch, 301 University Boulevard, Galveston, TX 77555-0587. E-mail: mmcostan@utmb.edu
 3,4
two-thirds of laboring women in the United States, and this proportion rose to almost all women or 3.4 million women per year
5 4
by the new millennium, despite the lack of clinical trials to support its efficacy in reducing perinatal adverse outcomes. After the
introduction of the EFM in the United States and its widespread adoption, the CD rate rose from 5% to 6% in the 70s to 30%
5,6
currently. This trend in CD rates after the introduction of EFM was also noted in other devel-oped countries, such as the United
6-8
Kingdom and Sweden. However, despite the widespread use of EFM and the dra-matic increase in the rate of CD between 1960s
9
and 2000s in the United States and other developed countries, the rate of cerebral palsy (CP) remained static at 2‰-3‰.
The limitations of epidemiological studies are well known, but there are few randomized trials focusing on the benefits of EFM.
The most recent Cochrane review found 12 clinical trials (that included 37,000 women), only 2 of which were considered of high
10
quality. None of these trials included comparison with a nonmonitored or sham-monitored con-trol arm, and all were performed
20 years ago. When com-pared with intermittent auscultation (IA), continuous EFM had no significant effect on perinatal death,
neonatal inten-sive care admissions, or CP (Table 1). There was a 50% re-duction in neonatal seizure, but this benefit was only seen
in the trials where continuous EFM was used in conjunction with fetal blood sampling (FBS) (relative risk [RR]: 0.49, 95%
confidence interval [CI]: 0.29-0.84 for continuous EFM FBS vs RR: 0.51, 95% CI: 0.18-1.44 for continuous EFM only). By
contrast, continuous EFM was associated with 66% increase in CD and 16% increase in operative vaginal delivery

0146-0005/12/$-see front matter © 2012 Elsevier Inc. All rights reserved. 379
http://dx.doi.org/10.1053/j.semperi.2012.04.023
380 M.M. Costantine and G.R. Saade

Table 1 Summary of Meta-analysis Results of Perinatal and Neonatal Outcomes When Continuous Cardiotocography was
Compared With Intermittent Auscultation as a Form of Electronic Fetal Monitoring for Fetal Assessment During Labor
Number
Number of
Outcome of Trials Patients RR (95% CI)
Perinatal death 11 33,513 0.85 (0.59-1.23)
NICU admission 10 33,067 1.01 (0.93-1.10)
Neonatal 9 32,386 0.50 (0.31-0.80)
seizures
Cerebral palsy 2 13,252 1.74 (0.97-3.11)
Apgar score <4 4 1919 1.43 (0.61-3.34)
at 5 minutes
Cord blood 2 2,494 0.92 (0.27-3.11)
acidosis
Cesarean 10 18,761 1.66 (1.30-2.13)
delivery
Operative 9 18,151 1.16 (1.01-1.32)
vaginal
delivery
10
Adapted from Alfirevic Z et al.
CI, confidence interval; NICU, neonatal intensive care unit; RR, relative risk.

10
(Table 1). The lack of beneficial effect on perinatal death was also confirmed in a Bayesian meta-analysis of the pub-lished
11
studies and led the United States preventive task force to give EFM a grade of D, which is the worst grade that can be given, and
12
to conclude that there was no evidence of benefit and there was evidence of harm in using EFM.

What Happened?
The unrealized potential of continuous EFM could have been predicted and potentially avoided. Continuous EFM is not a good
predictor of fetal acidosis. Even in the early phases of evaluation of EFM, it was well known that most fetal heart rate (FHR)
patterns that are deemed abnormal are associated with normal cord pH, and even with the most dramatic pat-terns, only a very small
13
proportion of neonates are born aci-dotic. Its ability to predict the most relevant outcome, CP, is even worse with a sensitivity of
14
only 26.9% and a positive predictive value of 0.14%. Another limitation of EFM is its subjective interpretation. Multiple studies
15,16
have reported on its poor inter- and intraobserver reliability, particularly when evaluating variability and decelerations. The
responses to EFM patterns, especially regarding timing and FHR charac-teristics that warrant further evaluation, also vary between
17
types of clinical providers, such as nurses and residents.
Because any management is dependent on a subjective interpretation, the variability in interpretation clearly leads to variability in
management. Management decisions then re-flect the comfort level of the decision maker rather than a clear algorithm. If one
desires to prevent any and all adverse outcomes, then the interpretation is more likely to have a higher false-positive rate compared
with another decision
maker who is willing to accept some false-negative interpre-tations. Because the interpretation is not based on immutable data,
practice in an environment where any decision leading to an adverse outcome is perceived as highly likely to be reviewed and
challenged, with resultant significant personal and financial repercussions, is likely to pressure and bias the decision maker toward
high false-positive interpretations. If the interpreter is under pressure not to miss any case, then he/she will have more false-positive
interpretations of the test, which will decrease its specificity. The lower the speci-ficity, the closer the likelihood ratio for a positive
test would be to 1, no matter how high the sensitivity, making the test not clinically useful as a positive test. This is not limited to
EFM, and the same can be said of any other screening test that depends on interpretation. Such tests cannot be compared with those
that depend on agreed-upon and well-accepted cutoffs, such as serum screening for aneuploidy or glucose tolerance test, and,
therefore, these tests are more objective and less amenable to interpretation. Although there may be variability in the management
once the latter tests are called positive, there is no subjectivity in calling them positive once they reached an agreed-upon threshold.
If there was no ac-cepted cutoff for these tests and if there was pressure not to miss any Down syndrome case or gestational
diabetes, like there is for not having a baby that can result in a lawsuit from not performing a cesarean, then the interpretation of
aneu-ploidy or diabetes screens, by providers, will slowly drift to more false-positive results and we would be doing more am-
niocenteses and treating more patients for diabetes.
Any test that depends on human interpretation will be subject to the pressures exerted on the individual making the
decision and the individual’s responses to the environment. This will lead to either higher false-negative or higher-false
positive test result depending on whether the decision maker fears more the implications of a mistaken diagnosis or the
implications of missing a diagnosis, respectively. In the case of EFM, the major implication of a false-positive interpreta-tion
is operative delivery, whereas the implication of a false-negative interpretation is an adverse outcome for the baby, along
with its associated consequences to the decision maker and hospital. It is clear that negative implications occur more
frequently with a false-positive interpretation (operative de-livery for a diagnosis of nonreassuring FHR) than with a false-
negative interpretation (adverse outcome for the baby). If this is the case, then one could argue that we should set our
threshold for calling a FHR nonreassuring (or calling it a positive test) at a more stringent level, where we would ac-cept
more false negative to decrease the false-positive results. However, decisions based on fear are not rational. Some have
argued that the threshold should be moved lower so as to decrease the false-negative rate because the consequences of a
false-negative test (adverse neonatal outcomes), while less frequent, are more dire than the consequences of a false-
positive test (CD). The evidence is overwhelming that con-tinuous EFM, even if one believes it may have prevented some
adverse outcomes, has overall caused more harm than good. The result of the high false-positive rate is a rise in CD rates,
placenta accreta, and associated maternal morbidities
The first cesarean 381

and mortality. When discussing EFM, the fetus that was saved is frequently held as proof of benefit of this technology, but the many
more women who underwent unnecessary pro-cedures, had significant morbidity, or even died as a result of a false-positive
interpretation are rarely mentioned. More-over, the fact that the risks of a CD are cumulative over future pregnancies is frequently
overlooked. It would be interesting to determine whether, in the aggregate, preventing one ad-verse neonatal outcome in the index
pregnancy by perform-ing a CD actually leads to more adverse outcomes in subse-quent pregnancies, either from uterine rupture or
from an increased risk of stillbirth in women with previous cesarean.
Another common misconception about EFM is that it is a reliable screening test for absence of injury. The fact that when the
FHR is “reassuring,” the vast majority of the neo-nates have normal outcome is not related to the test itself, but to the extremely low
rate of abnormal outcomes. EFM is no better than tossing a coin. If we pick “head” as a “reassuring” test, then the vast majority of
time we get “head” we will have a normal neonatal outcome just because an adverse outcome is so rare. Therefore, using this logic,
we can claim that coin tossing is a reliable screening test for absence of injury. The correct way to determine if a test is a reliable
screen for absence of an outcome is to calculate the likelihood ratio for a negative test. This is calculated as (1 sensitivity)/specific-
ity. Because sensitivity of EFM for prediction of CP is so low, the numerator is only slightly less than the denominator. So the
likelihood ratio is only slightly 1 and far from being below 0.1, the accepted cutoff for calling a test clinically useful to predict
absence of an outcome. This common mis-conception regarding screening test is because of the focus on the test’s negative
predictive value, which is dependent on the prevalence of the outcome rather than the test itself, rather than the likelihood ratio,
which is dependent on the test itself. Although this may seem like semantics, it is criti-cally important to correct this misconception,
18
as it has con-tinued to somewhat inappropriately justify the use of contin-uous EFM.

Potential Solutions
Given the lack of benefit of continuous EFM over IA and the potential for harm, the argument in favor of IA is gaining more
19
adherents. The American College of Nurse Midwives recently published a clinical bulletin endorsing IA in low risk patients. A
number of professional organizations in the United States and elsewhere (Association of Women’s Health, Obstetric and Neonatal
Nurses, ACOG, Society of Obstetri-cians and Gynecologists of Canada, Royal College of Obste-tricians and Gynecologists,
American College of Nurse Mid-wives) have published guidelines on intermittent FHR auscultation in women who are at low risk
during labor and recommend frequency of FHR auscultation approximately every hour in the latent phase, between 15 and 30
19
minutes in the active phase and every 5 minutes (for the majority) in the second stage of labor. However, none of the organizations
clearly and unambiguously define “low risk” patients. Al-though IA seems like a reasonable choice, there are some
issues that remain to be addressed. In addition to the defini-tion of “low risk,” it is imperative to address the staffing,
documentation, and criteria to transition to continuous EFM.
Another opportunity to improve on fetal monitoring dur-ing labor revolves around improving the reliability of inter-
20
pretation of the FHR. Toward that end, a number of groups have developed computerized systems to analyze the FHR.
Although some studies are promising, it is critical to have evidence from randomized trials comparing computer anal-ysis
21
with visual assessment, lest we repeat the mistakes of the past. One such trial is in progress.
In an effort to improve the consistency of interpretation of FHR, the Eunice Kennedy Shriver National Institute of Child Health
and Human Development, SMFM, and ACOG devel-oped a 3-tier system, along with specific definitions for the various patterns
22
and components of the FHR, such as base-line, variability, and periodic changes. Although the system had clear category I
(normal) and category III (abnormal) FHR and straightforward management for these (expectant for category I and intervention for
category III), the interme-diate category (category II) comprised all those FHR that did not fit into the other 2. Category II FHR is
23
very common, occurring in up to 84% of FHRs during labors, and its pro-portion increases in the second stage of labor. The
22
recom-mendation for this category was to use a back up test because the fetal status was indeterminate. Because of its indetermi-
nate characteristic, it is likely that this category of FHR is a major contributor to the variability in practice, especially that its
24
interobserver agreement is barely moderate (kappa 0.44). Given that a sizable proportion of FHRs is in category II, and that this
category is more open to variability in inter-pretation than the other categories, it seems logical that fo-cusing on this category
would have the best potential to im-pact the unnecessary CDs. To accomplish this, we would have to find an effective back up test.
There was a hope that fetal pulse oximetry would be this backup test. Fetal pulse oximetry allows measurement of the fetal
oxygen saturation (reassuring if $30%) using sensors placed against the fetal face. However, the trials did not bear this out. The
Cochrane review on fetal pulse oximetry for fetal assessment in labor found 5 trials (which included 7424 women) and found a 35%
reduction (4 trials, 4008 patients;
RR: 0.65, 95% CI: 0.46-0.90) in CD for nonreassuring fetal status in women randomized to fetal pulse oximetry and regular
cardiotocography compared with cardiotocography alone. However, the overall rate of CD did not differ signifi-
25
cantly (RR: 0.99, 95% CI: 0.86-1.13).
FBS allows measurement of the pH of fetal blood obtained from scalp sampling and is another backup test that is used in some
places outside the United States. The UK National In-stitute for Health and Clinical Excellence clinical guidelines include FBS in
its algorithm of intrapartum continuous
26
EFM. An FBS #7.2 is considered abnormal, requiring ur-gent birth, and between 7.22 and 7.24 is considered border-line. A meta-
analysis of 24 studies for FBS found a summary positive likelihood ratio (LR ) for pH 7.20 in predicting low Apgar score in the
potentially clinically useful range 4.51 (95% CI: 3.66-5.56), but a summary negative likelihood ra-
382 M.M. Costantine and G.R. Saade

tio (LR ) in the “not clinically useful” range 0.58 (95% CI 0.46-0.73). There are no studies showing that the use of FBS
26
improves long-term neonatal outcome.
Finally, a recent test that has been used as a backup is fetal electrocardiogram (ECG) monitoring or STAN (ST segment and T
wave analysis). This technology is widely used in Eu-rope and was approved by the Food and Drug Administra-tion in 2005 for use
in the United States as an adjunct to FHR monitoring to determine whether obstetrical intervention is warranted when there is an
indeterminate FHR pattern. This test is based on the established principles of ECG analysis and the development of myocardial
ischemia. With the develop-ment of metabolic acidosis, there is a switch from aerobic to anaerobic metabolism, increase
glycogenolysis, release of in-tracellular potassium ions, and change in membrane poten-tial, which leads to ST segment elevation
and T wave changes. These changes in the ST segment and T wave characteristics in relation to the QRS complex amplitude provide
the basis for the STAN technology. The most recent Cochrane review of randomized trials found no difference in neonatal enceph-
alopathy (RR: 0.54, 95% CI: 0.24-1.25), metabolic acidosis defined as cord arterial pH 7.05 and base deficit 12 mmol/L (RR: 0.78;
27
95% CI: 0.44-1.37), or overall CD (RR: 0.99; 95% CI: 0.91-1.08) when compared with continuous EFM. However, most trials
used FBS in the management, and the meta-analysis found a 40% reduction in the need for FBS when STAN was used as an adjunct
to continuous EFM. This has lead some to hypothesize that if used in the United States where FBS is not available, ST segment
anal-ysis may decrease the CD rate. A large multicenter trial to evaluate the use of STAN as an adjunct to EFM is currently under
way in the United States (ClinicalTrials.gov identi-fier: NCT01131260).

Conclusions
The widespread use of EFM was not associated with reduc-tion in perinatal or neonatal adverse outcomes. This, coupled with
the inherent problems of EFM, particularly as it relates to its poor sensitivity and reliability, has led to a dramatic increase in
CD rates. To be successful, any attempt to lower the current CD rate must address the inherent limitations in our EFM
methods. To impact the CD rate through EFM, we recommend the following:

1. Improve the reliability and reduce the subjectivity of FHR analysis by simplifying the current FHR guide-lines, introducing
more objective methods of interpre-tations, such as computerized analysis of FHR tracings.
2. Use of backup tests in the setting of category II tracing. These may include FBS with or without fetal ECG anal-ysis.
3. Promote the use of IA of FHR in low-risk population for which continuous FHR was not originally designed. This
requires clear definition of low-risk candidates, standardization of frequency and documentation, and guidelines on
when to transition to continuous EFM during labor.
 4. Decrease the false-positive interpretation rate by in-creasing the threshold required to perform a CD for FHR indications. As
an example, CD should not be performed in the presence of moderate FHR variability (determined by variability between
decelerations) re-gardless of the presence of decelerations. This means that the health care system will accept higher false
neg-ative and accept a very small increase in immediate adverse outcomes to decrease the cumulative maternal and perinatal
risks in the future. This can only be ac-complished by decreasing the pressures on the decision maker to maintain the high
false-positive rate and by changing the institutional environment. These can be accomplished through tort reform and peer
review, respectively.

References
1. Hon EH: The electronic evaluation of the fetal heart rate; preliminary report. Am J Obstet Gynecol 75:1215-1230, 1958
2. Hon EH, Hess OW: The clinical value of fetal electrocardiography. Am J Obstet Gynecol 79:1012-1023, 1960
3. Banta HD, Thacker SB: Assessing the costs and benefits of electronic fetal monitoring. Obstet Gynecol Surv 34:627-642, 1979
4. Banta HD, Thacker SB: Fetal monitoring. Obstet Gynecol 54:667-670, 1979
5. Martin JA, Hamilton BE, Sutton PD, et al: Births: Final data for 2002. National Vital Statistics Reports; Vol 52. Hyattsville, MD, National Cen-ter for
Health Statistics, 2003
6. Hamilton BE, Martin JA, Ventura SJ: Births: Preliminary data for 2010. National Vital Statistics Reports; Vol 60. Hyattsville, MD, National Cen-ter for
Health Statistics, 2011
7. Notzon FC, Placek PJ, Taffel SM: Comparisons of national cesarean-section rates. N Engl J Med 316:386-389, 1987
8. Hospital Episode Statistics (UK) Maternity Data 2010-11. Available at: http://www.hesonline.nhs.uk/Ease/servlet/ContentServer?siteID
1937&categoryID 1815. Accessed April 1, 2012
9. Clark SL, Hankins GD: Temporal and demographic trends in cerebral palsy—Fact and fiction. Am J Obstet Gynecol 188:628-633, 2003
10. Alfirevic Z, Devane D, Gyte GM: Continuous cardiotocography (CTG) as a form of electronic fetal monitoring (EFM) for fetal assessment during labour.
Cochrane Database Syst Rev 19:CD006066, 2006
11. Sutton AJ, Abrams KR: Bayesian methods in meta-analysis and evi-dence synthesis. Stat Methods Med Res 10:277-303, 2001
12. U.S. Preventive Services Task Force: Screening for Intrapartum Electronic
Fetal Monitoring. Available at: http://www.uspreventiveservicestaskforce. org/uspstf/uspsiefm.htm. Accessed April 1, 2012
13. Beard RW, Filshie GM, Knight CA, et al: The significance of the changes in the continuous fetal heart rate in the first stage of labour. J Obstet Gynaecol
Br Commonw 78:865-881, 1971
14. Nelson KB, Dambrosia JM, Ting TY, et al: Uncertain value of electronic fetal monitoring in predicting cerebral palsy. N Engl J Med 334:613-618, 1996
15. Chauhan SP, Klauser CK, Woodring TC, et al: Intrapartum nonreas-suring fetal heart rate tracing and prediction of adverse outcomes: Interobserver
variability. Am J Obstet Gynecol 199:623.e1-623.e5, 2008
16. Westerhuis ME, van Horen E, Kwee A, et al: Inter- and intra-observer agreement of intrapartum ST analysis of the fetal electrocardiogram in women monitored
by STAN. BJOG 116:545-551, 2009
17. Chiossi G, Costantine MM, Pfannstiel JM, et al: Intervention for fetal
distress among obstetricians, registered nurses, and residents: Similar-ities, differences, and determining factors. Obstet Gynecol 118:809-817, 2011
18. Grimes DA, Peipert JF: Electronic fetal monitoring as a public health
screening program: The arithmetic of failure. Obstet Gynecol 116: 1397-1400, 2010
The first cesarean 383

19. American College of Nurse-Midwives: Intermittent auscultation for intrapartum fetal heart rate surveillance (replaces ACNM clinical bulletin #9,
March 2007). J Midwifery Womens Health 55:397-403, 2010
20. Parer JT, Hamilton EF: Comparison of 5 experts and computer analysis in rule-based fetal heart rate interpretation. Am J Obstet Gynecol 203: 451.e1-
451.e7, 2010
21. Ayres-de-Campos D, Ugwumadu A, Banfield P, et al: A randomised clinical trial of intrapartum fetal monitoring with computer analysis and alerts
versus previously available monitoring. BMC Pregnancy Childbirth 10:71, 2010
22. Macones GA, Hankins GD, Spong CY, et al: The 2008 National Institute of Child Health and Human Development workshop report on elec-tronic fetal
monitoring. Update on definitions, interpretation, and re-search guidelines. Obstet Gynecol 112:661-666, 2008
23. Jackson M, Holmgren CM, Esplin MS, et al: Frequency of fetal heart rate categories and short-term neonatal outcomes. Obstet Gynecol 118: 803-808,
2011
24. Blackwell SC, Grobman WA, Antoniewicz L, et al: Interobserver and intraobserver reliability of the NICHD 3-tier fetal heart rate interpreta-tion system. Am J
Obstet Gynecol 205:378.e1-378.e5, 2011
25. East CE, Chan FY, Colditz PB, et al: Fetal pulse oximetry for fetal assess-ment in labour. Cochrane Database Syst Rev 18:CD004075, 2007
26. National Institute for Clinical Excellence: Intrapartum Care: Care of Healthy Women and Their Babies During Childbirth. London, NICE, 2007.
Clinical Guidance No 55. Available at: http://www.nice.org.uk/ guidance/CG55/NICEGuidance. Accessed April 2, 2012
27. Neilson JP: Fetal electrocardiogram (ECG) for fetal monitoring during labour. Cochrane Database Syst Rev 19:CD000116, 2006