Journal of Perinatology (2014) 34, 176–180

& 2014 Nature America, Inc. All rights reserved 0743-8346/14

www.nature.com/jp

ORIGINAL ARTICLE
The effect of obstetric practice change to reduce early term
delivery on perinatal outcome
SE Little1, JN Robinson1, KM Puopolo2, S Mukhopadhyay2, LE Wilkins-Haug1, DA Acker1 and CA Zera1

OBJECTIVE: To investigate whether the national emphasis on attaining X39 weeks gestation has altered obstetric practice,
and if so whether this has affected perinatal morbidity.
STUDY DESIGN: We examined trends in gestational age, neonatal morbidity, maternal complications and stillbirth for a
retrospective cohort of singleton, live births between 37 þ 0 and 39 þ 6 weeks of gestation over a 5-year period at a single
tertiary care center.
RESULT: There were 21 343 eligible deliveries. The proportion of deliveries in the early term (o39 weeks) decreased from 47.8 to
40.2% (Po0.01). The reduction was most pronounced for elective inductions (27.5 to 8.0%; Po0.01) and scheduled cesareans
(56.9 to 24.9%; Po0.01), although a similar trend was seen for nonelective inductions (51.2 to 47.9%; P ¼ 0.03). In multivariable
analysis, there was a 10% decreased odds of early term delivery per year (Po0.01). There were no changes in the rates of
neonatal intensive care unit (NICU) evaluation (29.8 to 28.1%; P ¼ 0.11), pre-eclampsia (7.6 to 8.5%; P ¼ 0.06) or stillbirth (11.5 to 14.4
per 10 000; P ¼ 0.55).
CONCLUSION: A 10% annual decline in the odds of early term delivery was not accompanied by significant changes in perinatal
morbidity.

Journal of Perinatology (2014) 34, 176–180; doi:10.1038/jp.2013.166; published online 9 January 2014
Keywords: 39 weeks; early term; elective delivery

INTRODUCTION period. We were not looking at one specific intervention, but

Neonatal outcomes improve with advancing gestation, even at
term. A growing body of literature demonstrates that neonates
born in the early term (37 to o39 weeks) have increased rates
of respiratory morbidity, hypoglycemia, temperature instability,
sepsis, intensive care utilization and death, as well as worse
neurodevelopmental outcomes compared with neonates born
after 39 weeks.1–6
Consequently, a number of different organizations have tried to
reduce the rate of elective delivery before 39 weeks. The American
College of Obstetricians and Gynecologists (ACOG) published
guidelines in 2009 clarifying the need for an appropriate
indication for labor induction before 39 weeks7 and in 2013
outlined what indications were considered medically appropriate.8
In 2010, the Joint Commission established new obstetric quality
measures that included elective deliveries in the early term, and
the March of Dimes launched a national campaign to reduce early
term elective deliveries.9,10
We hypothesized that obstetric providers are responding to this
national message by reducing the rate of scheduled delivery
before 39 weeks. However, we further hypothesized that 39 week
delivery policies may have less impact than hoped when
translated into real world practice. We anticipated that changes
in practice patterns would in reality only lead to a small increase in
the gestational age at delivery. If this were the case, consequent
changes in delivery gestational age would only be a small number
of days and subsequently may not affect neonatal outcome.
To investigate our hypotheses, we analyzed the trend in early
term delivery at a single, large tertiary care center over a 5-year
rather the overall trend over time. We determined whether there
was a decrease in the rate of early term delivery over time and
whether this decrease was accompanied by changes in either
neonatal or maternal morbidity, specifically focusing on the rates
of neonatal intensive care unit (NICU) evaluation, pre-eclampsia
and stillbirth. All deliveries were included in our analysis, rather
than just elective deliveries, as we hypothesized that the message
about 39 week elective deliveries may have lead to parallel
changes in nonelective deliveries and we wanted to assess the full
impact on the entire population.
METHODS
After obtaining approval from our institutional review board (IRB protocol
no. 2011-P-002264/1), we examined a retrospective cohort of all singleton
live births delivered at a single tertiary care center from 1 October 2006 to
30 September 2011. We included only those patients delivered between
37 þ 0 and 39 þ 6 weeks of gestation. Data were obtained from the
hospital’s electronic delivery information system, which captures obstetric
outcomes for all deliveries greater than 20 weeks gestational age, and from
the electronic NICU admission log.
We calculated the gestational age at delivery using the estimated date
of delivery recorded by the primary delivering provider based upon best
obstetrical dating. We compared the gestational age at delivery over time
for the entire cohort, as well as for specific subsets, including scheduled
cesarean deliveries and elective inductions. We defined scheduled
cesareans and elective inductions based on data entered electronically
by the delivering provider. Inductions were considered elective if the
indication was documented as elective or scheduled for social conve-
nience, an advanced cervical exam in the absence of labor, maternal

1
Division of Maternal-Fetal Medicine, Brigham and Women’s Hospital, Boston, MA, USA and 2Department of Newborn Medicine, Brigham and Women’s Hospital, Boston, MA,
USA. Correspondence: Dr SE Little, Division of Maternal-Fetal Medicine, Brigham and Women’s Hospital, 75 Francis Street, Boston, MA 02115, USA.
E-mail: selittle@partners.org
Received 5 June 2013; revised 30 October 2013; accepted 12 November 2013; published online 9 January 2014
 Early term delivery and perinatal outcome
SE Little et al
177
Table 1. Patient characteristics by year

2006–2007 2007–2008 2008–2009 2009–2010 2010–2011 P-valuea

(n ¼ 4351) (n ¼ 4401) (n ¼ 4269) (n ¼ 4142) (n ¼ 6617)

Maternal age 32 32 32 32 32 0.70

Median (25–75%) (28–36) (28–36) (28–36) (28–36) (28–35)
Nulliparity 1736 1732 1725 1755 1773 o0.01
N (%) (39.9) (39.4) (40.4) (42.4) (42.5)
Public insurance 421 448 439 446 446 0.09
N (%)b (11.4) (12.1) (12.1) (12.7) (12.5)
Induced 941 1005 912 884 904 0.50
N (%) (21.6) (22.8) (21.4) (21.4) (21.7)
Cesarean delivery 1361 1514 1418 1351 1349 0.89
N (%) (31.3) (34.4) (33.3) (32.7) (32.3)

Provider group
N (%)
Faculty 1395 (32.2) 1465 (33.4) 1436 (33.8) 1368 (33.3) 1407 (34.0) 0.12
Private practice 2940 (67.8) 2927 (66.6) 2811 (66.2) 2736 (66.7) 2737 (66.1)
a
P-values for continuous variables based on the Kruskal–Wallis testing, P-values for categorical variables based on the Cochran–Armitage test for trend.
b
15% of the data for this category was missing, data based on sample size of 3698, 3713, 3635, 3511 and 3557, for each year, respectively.

discomfort or mood disorder, a history of a prior difficult or precipitous RESULTS

delivery, suspected macrosomia or unstable lie. All other indications were
considered nonelective. Multiple reasons for induction could be provided
on the electronic delivery record; the induction was considered elective
only in the absence of any nonelective indication.
Maternal age, parity, race and delivery information were collected from
the electronic delivery record. Insurance status was collected from the
hospital billing records and available for 85% of our cohort.
Changes in covariates and delivery timing were analyzed using
nonparametric testing: Kruskal–Wallis tests for continuous variables and
the Cochran–Armitage test for linear trend for categorical variables. We
performed logistic regression to analyze the odds of early term delivery as
a function of year, controlling for potential confounders including maternal
age, self-reported race, parity, type of provider (faculty versus private
practice), spontaneous versus induced labor, mode of delivery and
birthweight. Given the missing data, we decided to include insurance
status as a potential confounders only if there was a significant changes in
this variables over time, which there was not, thus it was not included in
the final logistic regression. A P-value of 0.05 was considered statistically
significant.
Our primary neonatal outcome was need for NICU evaluation. We
included both admissions to the NICU for prolonged care and NICU triage
admissions (defined at our center as NICU admissions of p4 h duration,
with subsequent transfer to Newborn Nursery) combined together as ‘NICU
evaluation’. We had 80% power to detect a 3% absolute decrease in the
rate of NICU evaluation. Secondary outcomes included the indication for
NICU evaluation, birthweight, Apgar score and prolonged length of stay.
Neonatal length of stay was considered prolonged if 43 days for a vaginal
delivery or 45 days for a cesarean delivery. Meconium was not reliably
recorded, thus not included as an outcome in our analysis. All neonatal
deaths were identified in the neonatal records and individually reviewed.
We identified stillbirths using death certificates, all of which were
individually reviewed. We had 80% power to detect a fourfold increase in
the overall rate of stillbirth.
Our primary maternal morbidity of interest was hypertension, which
included any delivery where hypertension or pre-eclampsia was stated as
an indication for induction or was recorded as a complication. We analyzed
time trends with a Cochran–Armitage test. We had greater than 80%
power to detect a 2% absolute increase in the rate of hypertensive
disorders. Secondary outcomes included rates of severe pre-eclampsia,
(which we defined as any delivery where magnesium sulfate was
administered as it is standard practice at our institution to routinely
administer magnesium sulfate only for those with severe pre-eclampsia),
chorioamnionitis, (presumed if either coded as a complication or if
antibiotics consistent with coverage for chorioamnionitis were adminis-
tered), third or fourth degree perineal lacerations, postpartum hemorrhage
(4500 cc for a vaginal delivery and 41000 cc for a cesarean delivery),
shoulder dystocia and uterine rupture. We were underpowered to detect
small changes in rare events; for example, we had 80% power to detect
only a sixfold increase in the rate of uterine rupture.
There were 21 343 singleton live births at 37 þ 0 to 39 þ 6 weeks
of gestation delivered between 1 October 2006 and 30 September
2011. Of these, 90% were to mothers who had only one delivery
during the five-year period, 10% had two deliveries and o1% had
three or more deliveries. Patient demographics were similar across
years (Table 1), although there were more nulliparous women by
the end of the study period (39.9% in 2006 versus 42.5% in 2011,
Po0.01). There was no change in the overall rate of induction or
cesarean delivery during the 5-year period.
Median gestational age was unchanged at 39 þ 0 weeks
throughout the study period; however, the gestational age
distribution changed markedly over time (Figure 1). For example,
among all deliveries in 2006 to 2007, 6.6% occurred at 38 þ 6
weeks and 8.1% at 39 þ 0 weeks, whereas in 2010 to 2011, 4.8%
were delivered at 38 þ 6 weeks and 10.5% at 39 þ 0 weeks.
Figure 1 demonstrates the shift from 38 þ 6 weeks to 39 þ 0
weeks among elective inductions and scheduled cesareans.
Table 2 displays the percent of deliveries in the early term
(37 þ 0 to 38 þ 6) for the most common delivery indications. As
shown, for the entire cohort there were fewer deliveries in the
early term (37 þ 0 to 38 þ 6) by the end of the study period (47.8%
in 2006 versus 40.2% in 2011; Po0.01). The percent of deliveries in
the early term fell by the greatest magnitude for elective
inductions (27.5 to 8%; Po0.01) and scheduled cesareans (56.9
to 24.9%; Po0.01). There was also a significant decrease in early
term delivery for all nonelective inductions combined (51.2 to
47.9%; P ¼ 0.03) and a non-significant trend when individual
nonelective indications were examined, such as diabetes, hyper-
tension and growth restriction, the three most common none-
lective indications. Similarly, among cesarean deliveries there was
a non-significant trend toward decreased early term delivery for
women with a prior myomectomy or classical cesarean (Table 2).
Delivery year was a significant predictor of early term delivery.
In multivariable analysis, every year later delivery was associated
with a 10% decreased odds of an early term delivery (odds ratio
0.90, 95% confidence interval (CI) 0.88 to 0.92) independent of
maternal age, race, parity, birthweight, labor induction, provider
and mode of delivery.
There was no change in the rate of hypertensive disorders
(7.6 to 8.5%; P ¼ 0.06) nor the rate of severe pre-eclampsia
(1.3 to 0.9%) (Table 3). There was a decrease in the rate of
chorioamnionitis (6.6 to 5.2%; Po0.01) and severe perineal
lacerations (1.6 to 1.0%; P ¼ 0.03). Rates of other maternal
morbidities were unchanged.

& 2014 Nature America, Inc. Journal of Perinatology (2014), 176 – 180
 Early term delivery and perinatal outcome
SE Little et al
177
Table 1. Patient characteristics by year

2006–2007 2007–2008 2008–2009 2009–2010 2010–2011 P-valuea

(n ¼ 4351) (n ¼ 4401) (n ¼ 4269) (n ¼ 4142) (n ¼ 6617)

Maternal age 32 32 32 32 32 0.70

Median (25–75%) (28–36) (28–36) (28–36) (28–36) (28–35)
Nulliparity 1736 1732 1725 1755 1773 o0.01
N (%) (39.9) (39.4) (40.4) (42.4) (42.5)
Public insurance 421 448 439 446 446 0.09
N (%)b (11.4) (12.1) (12.1) (12.7) (12.5)
Induced 941 1005 912 884 904 0.50
N (%) (21.6) (22.8) (21.4) (21.4) (21.7)
Cesarean delivery 1361 1514 1418 1351 1349 0.89
N (%) (31.3) (34.4) (33.3) (32.7) (32.3)

Provider group
N (%)
Faculty 1395 (32.2) 1465 (33.4) 1436 (33.8) 1368 (33.3) 1407 (34.0) 0.12
Private practice 2940 (67.8) 2927 (66.6) 2811 (66.2) 2736 (66.7) 2737 (66.1)
a
P-values for continuous variables based on the Kruskal–Wallis testing, P-values for categorical variables based on the Cochran–Armitage test for trend.
b
15% of the data for this category was missing, data based on sample size of 3698, 3713, 3635, 3511 and 3557, for each year, respectively.

discomfort or mood disorder, a history of a prior difficult or precipitous RESULTS

delivery, suspected macrosomia or unstable lie. All other indications were
considered nonelective. Multiple reasons for induction could be provided
on the electronic delivery record; the induction was considered elective
only in the absence of any nonelective indication.
Maternal age, parity, race and delivery information were collected from
the electronic delivery record. Insurance status was collected from the
hospital billing records and available for 85% of our cohort.
Changes in covariates and delivery timing were analyzed using
nonparametric testing: Kruskal–Wallis tests for continuous variables and
the Cochran–Armitage test for linear trend for categorical variables. We
performed logistic regression to analyze the odds of early term delivery as
a function of year, controlling for potential confounders including maternal
age, self-reported race, parity, type of provider (faculty versus private
practice), spontaneous versus induced labor, mode of delivery and
birthweight. Given the missing data, we decided to include insurance
status as a potential confounders only if there was a significant changes in
this variables over time, which there was not, thus it was not included in
the final logistic regression. A P-value of 0.05 was considered statistically
significant.
Our primary neonatal outcome was need for NICU evaluation. We
included both admissions to the NICU for prolonged care and NICU triage
admissions (defined at our center as NICU admissions of p4 h duration,
with subsequent transfer to Newborn Nursery) combined together as ‘NICU
evaluation’. We had 80% power to detect a 3% absolute decrease in the
rate of NICU evaluation. Secondary outcomes included the indication for
NICU evaluation, birthweight, Apgar score and prolonged length of stay.
Neonatal length of stay was considered prolonged if 43 days for a vaginal
delivery or 45 days for a cesarean delivery. Meconium was not reliably
recorded, thus not included as an outcome in our analysis. All neonatal
deaths were identified in the neonatal records and individually reviewed.
We identified stillbirths using death certificates, all of which were
individually reviewed. We had 80% power to detect a fourfold increase in
the overall rate of stillbirth.
Our primary maternal morbidity of interest was hypertension, which
included any delivery where hypertension or pre-eclampsia was stated as
an indication for induction or was recorded as a complication. We analyzed
time trends with a Cochran–Armitage test. We had greater than 80%
power to detect a 2% absolute increase in the rate of hypertensive
disorders. Secondary outcomes included rates of severe pre-eclampsia,
(which we defined as any delivery where magnesium sulfate was
administered as it is standard practice at our institution to routinely
administer magnesium sulfate only for those with severe pre-eclampsia),
chorioamnionitis, (presumed if either coded as a complication or if
antibiotics consistent with coverage for chorioamnionitis were adminis-
tered), third or fourth degree perineal lacerations, postpartum hemorrhage
(4500 cc for a vaginal delivery and 41000 cc for a cesarean delivery),
shoulder dystocia and uterine rupture. We were underpowered to detect
small changes in rare events; for example, we had 80% power to detect
only a sixfold increase in the rate of uterine rupture.
There were 21 343 singleton live births at 37 þ 0 to 39 þ 6 weeks
of gestation delivered between 1 October 2006 and 30 September
2011. Of these, 90% were to mothers who had only one delivery
during the five-year period, 10% had two deliveries and o1% had
three or more deliveries. Patient demographics were similar across
years (Table 1), although there were more nulliparous women by
the end of the study period (39.9% in 2006 versus 42.5% in 2011,
Po0.01). There was no change in the overall rate of induction or
cesarean delivery during the 5-year period.
Median gestational age was unchanged at 39 þ 0 weeks
throughout the study period; however, the gestational age
distribution changed markedly over time (Figure 1). For example,
among all deliveries in 2006 to 2007, 6.6% occurred at 38 þ 6
weeks and 8.1% at 39 þ 0 weeks, whereas in 2010 to 2011, 4.8%
were delivered at 38 þ 6 weeks and 10.5% at 39 þ 0 weeks.
Figure 1 demonstrates the shift from 38 þ 6 weeks to 39 þ 0
weeks among elective inductions and scheduled cesareans.
Table 2 displays the percent of deliveries in the early term
(37 þ 0 to 38 þ 6) for the most common delivery indications. As
shown, for the entire cohort there were fewer deliveries in the
early term (37 þ 0 to 38 þ 6) by the end of the study period (47.8%
in 2006 versus 40.2% in 2011; Po0.01). The percent of deliveries in
the early term fell by the greatest magnitude for elective
inductions (27.5 to 8%; Po0.01) and scheduled cesareans (56.9
to 24.9%; Po0.01). There was also a significant decrease in early
term delivery for all nonelective inductions combined (51.2 to
47.9%; P ¼ 0.03) and a non-significant trend when individual
nonelective indications were examined, such as diabetes, hyper-
tension and growth restriction, the three most common none-
lective indications. Similarly, among cesarean deliveries there was
a non-significant trend toward decreased early term delivery for
women with a prior myomectomy or classical cesarean (Table 2).
Delivery year was a significant predictor of early term delivery.
In multivariable analysis, every year later delivery was associated
with a 10% decreased odds of an early term delivery (odds ratio
0.90, 95% confidence interval (CI) 0.88 to 0.92) independent of
maternal age, race, parity, birthweight, labor induction, provider
and mode of delivery.
There was no change in the rate of hypertensive disorders
(7.6 to 8.5%; P ¼ 0.06) nor the rate of severe pre-eclampsia
(1.3 to 0.9%) (Table 3). There was a decrease in the rate of
chorioamnionitis (6.6 to 5.2%; Po0.01) and severe perineal
lacerations (1.6 to 1.0%; P ¼ 0.03). Rates of other maternal
morbidities were unchanged.

& 2014 Nature America, Inc. Journal of Perinatology (2014), 176 – 180
 Early term delivery and perinatal outcome
SE Little et al
178
Figure 1. Changes in gestational age distribution. This figure
displays the change in gestational age distribution between the
beginning of the study period (2006 to 2007) and the end of the
study period (2010 to 2011).
In terms of neonatal outcomes, there was no change in the rate
of NICU evaluation (29.8% to 28.1; P ¼ 0.11) (Table 3). There were
no in-hospital neonatal deaths to non-anomalous fetuses in our
study cohort. There was an increase in the rate of respiratory
distress (5.9 to 8.7%; Po0.01) but all other morbidities were
unchanged. In logistic regression, the odds of respiratory distress
increased by 11% for every year of delivery (odds ratio 1.11; 95%
CI 1.07 to 1.15) independent of gestational age but there was no
change in the rate of respiratory distress requiring a prolonged
length of stay (1.29 to 1.29%; P ¼ 0.55).
Journal of Perinatology (2014), 176 – 180
When considering elective deliveries alone, there was similarly
no significant improvement in neonatal outcome. For elective
inductions (n ¼ 1216), the rate of NICU evaluation and respiratory
distress had a nonstatistically significant decline from 14.3 to
10.4% (P ¼ 0.22) and from 3.5 to 2.5% (P ¼ 0.45), respectively.
However, for scheduled cesareans (n ¼ 3990) there was a
significant increase in both NICU evaluation (23.8 to 28.4%;
P ¼ 0.02) and respiratory distress (9.9 to 17.0%; Po0.01).
Though there were no changes in the overall rate of neonatal
morbidity over time, there was a decrease in neonatal morbidity
with advancing gestation (Figure 2). For the entire cohort, the
rates of NICU evaluation, respiratory distress, hypoglycemia/
hyperbilirubinemia and temperature instability and prolonged
length of stay all decreased significantly with increasing week of
gestation.
There were 29 term stillbirths during the study period. The rate
of stillbirth in 2006 was 11.5 per 10 000 (95% CI 3.7 to 26.8 per
10 000), compared with 14.4 per 10 000 in 2011 (95% CI 5.3 to 31.2
per 10 000) (P ¼ 0.89) (Table 3). When stillbirths were limited to
those in the early term, the rate increased from 4.6 per 10 000
(95% CI 0.6 to 16.6 per 10 000) to 12.0 per 10 000 (95% CI 3.9 to
27.9 per 10 000) (P ¼ 0.58), although in the lowest-risk subgroup of
non-anomalous fetuses born to women with no documented
comorbidities, the rate was unchanged (4.6 to 4.8 per 10 000)
(P ¼ 0.57). Of note, comorbidity exclusions included only pre-
gestational conditions (chronic hypertension and pre-gestational
diabetes).
DISCUSSION
Despite a 10% annual decrease in the odds of early term delivery,
there was no significant change in perinatal morbidity at a single,
high-volume tertiary care center. The rate of scheduled delivery in
the early term fell significantly over the 5-year period; there was a
71% reduction in early term elective inductions and a 56%
reduction in early term scheduled cesareans. Nevertheless, there
was no change in neonatal morbidity and the only change in
maternal morbidity that we found was a decrease in chorioam-
nionitis and severe perineal lacerations, both changes that are
more likely attributable to trends in labor floor practice during this
time period.
Consistent with prior research, we found that neonates born in
the early term had more morbidity, including increased rates of
NICU evaluation, respiratory distress and prolonged length of
stay.11–13 However, we did not see a decrease in neonatal morbidity
over time, despite the fact that there was a significant decline in
early term delivery. There was no change in the rate of NICU
evaluation or prolonged length of stay. We did find a slight increase
in NICU evaluations coded for respiratory distress, but this was
observed across gestational ages. Changes in our local algorithm for
neonatal sepsis evaluation resulted in a decline in the number of
sepsis evaluations during our study period and this may have
resulted in a reclassification of short-term NICU admission for
infants with both respiratory distress and indications for sepsis
evaluation as NICU admission for respiratory distress alone.14
Nevertheless, we observed no change in the incidence of
prolonged neonatal hospital stay for respiratory distress across
the study period, suggesting that serious respiratory morbidity was
unaffected by our observed increase in gestational age at birth.
We speculate that we did not see a decrease in neonatal
morbidity because the changes that we found in the gestational
age distribution were not particularly meaningful. Deliveries that
were previously in the late 38th week were pushed into the early
39th week. In fact, we found the most marked change in the
number of deliveries that occurred at exactly 38 þ 6 as compared
with exactly 39 þ 0 weeks over time. This shift produced a
significant decrease in the early term delivery rate but may not
translate into meaningful changes in neonatal morbidity.
& 2014 Nature America, Inc.
 Early term delivery and perinatal outcome
SE Little et al
179
Table 2. Trends in early term delivery

2006–2007 2007–2008 2008–2009 2009–2010 2010–2011 P-valuea

% of deliveries for that indication that took place in the early term (o39 weeks)
All deliveries (n ¼ 21 343) 47.8 47.0 42.4 40.7 40.2 o0.01

Inductions
All (n ¼ 4646) 44.0 45.1 42.0 38.5 40.7 o0.01
Elective (n ¼ 1216) 27.5 29.0 18.1 13.1 8.0 o0.01
Nonelective (n ¼ 3430) 51.2 52.2 51.5 45.8 47.9 0.03
Diabetes (444) 34.8 25.0 25.3 18.6 24.5 0.06
Hypertension (1363) 62.8 64.9 63.8 54.4 60.3 0.10
Growth restriction (n ¼ 333) 68.8 62.8 76.8 65.5 66.2 0.93

Scheduled cesarean deliveriesb

All (n ¼ 3990) 56.9 52.7 38.1 26.8 24.9 o0.01
Repeat (n ¼ 2539) 54.7 54.0 38.7 24.0 19.8 o0.01
Breech (n ¼ 592) 53.8 42.1 27.6 27.5 24.6 o0.01
Prior myomectomy/classical (n ¼ 240) 73.5 71.1 58.5 57.4 60.8 0.07
a
P-values are based on a two-sided Cochran–Armitage trend test.
b
Does not include unplanned deliveries (such as those patients who presented in labor).

Table 3. Morbidities by year

2006–2007 2007–2008 2008–2009 2009–2010 2010–2011 P-valuea

Maternal
Hypertensive d/o any 331 (7.6%) 340 (7.7%) 336 (7.9%) 346 (8.4%) 356 (8.5%) 0.06
Severe pre-eclampsia 56 (1.3%) 46 (1.0%) 53 (1.2%) 41 (1.0%) 36 (0.9%) 0.07
Uterine rupture 2 (0.05%) 3 (0.07%) 0 (0%) 0 (0%) 3 (0.07%) 0.85
Shoulder dystocia 13 (0.3%) 9 (0.2%) 10 (0.2%) 10 (0.2%) 14 (0.3%) 0.66
Chorioamnionitis 287 (6.6%) 274 (6.2%) 237 (5.6%) 221 (5.3%) 215 (5.2%) o0.01
Severe perineal laceration 69 (1.6%) 40 (0.9%) 64 (1.5%) 40 (1.0%) 41 (1.0%) 0.03
Postpartum hemorrhage 172 (4.0%) 164 (3.7%) 156 (3.7%) 142 (3.4%) 159 (3.8%) 0.52

Neonatal
NICU evaluation 1297 (29.8%) 1319 (30.0%) 1336 (31.3%) 1233 (29.8%) 1174 (28.1) 0.11
Respiratory distress 255 (5.9%) 344 (7.8%) 351 (8.2%) 357 (8.6%) 363 (8.7%) o0.01
Hypoglycemia, hyperbilirubinemia or 115 (2.6%) 37 (0.8%) 46 (1.1%) 41 (1.0%) 84 (2.0%) 0.05
temperature instability
5 min Apgar o7 29 (0.7%) 31 (0.7%) 26 (0.6%) 24 (0.6%) 22 (0.5%) 0.29
Birthweight 44000 g 268 (6.2%) 327 (7.4%) 307 (7.2%) 257 (6.2%) 270 (6.5%) 0.63
Extended LOS 129 (3.0%) 118 (2.7%) 144 (3.4%) 136 (3.3%) 146 (3.5%) 0.05

Stillbirth
Total (per 10 000 ) 5 (11.5) 9 (20.4) 6 (14.1) 5 (12.7) 6 (14.4) 0.89
Early term (per 10 000) 2 (4.6) 7 (15.9) 3 (7.0) 4 (9.7) 5 (12.0) 0.58
Early term without fetal anomalies or maternal 2 (4.6) 4 (9.1) 2 (4.7) 1 (2.4) 2 (4.8) 0.57
comorbidities (per 10 000)
Abbreviation: NICU, neonatal intensive care unit.
a
P-values are based on a two-sided Cochran–Armitage trend test.

We were reassured that fewer early term deliveries were not
accompanied by a detectable increase in adverse events. There
was no statistically significant change in the rate of maternal
hypertension, although there was a trend toward increased
hypertensive disorders from 7.6 to 8.5%. We would have needed
3.4 times the sample size for the change from the first year to the
last to reach statistical significance. We were similarly under-
powered to detect changes in rare outcomes, such as uterine
rupture or stillbirth, even with our large sample size. We would
have needed 5.6 times the sample size for the increase in early
term stillbirth from 4.6 to 12 per 10 000 to have reached statistical
significance. Prior studies have found conflicting results on
stillbirth; one found a significant increase in early term stillbirth
accompanying a decline in elective early term delivery12 and
another found a significant decline.13 This highlights the need for
continued surveillance on a national level to detect any changes
in rare adverse events that may accompany large scale changes in
obstetric practice.
Our study has a number of strengths. The large sample size
allowed us to see changes over the short period of time during
which decreasing elective deliveries before 39 weeks had become
a national priority. Unlike prior studies, we were not attempting to
study the effect of a specific policy intervention. The only
intervention at an institutional level was a minor change in our
induction scheduling system to discourage elective induction at
less than 39 weeks that occurred near the end of the study period,
in December 2010. Rather, we sought to look at trends over time
to capture any influence of the growing national awareness of the
neonatal morbidity associated with early term delivery.
We must, however, consider our results within the context of the
study design. To facilitate review of large numbers of mother/infant
pairs, we did not review individual patient medical records (other

& 2014 Nature America, Inc. Journal of Perinatology (2014), 176 – 180
 Early term delivery and perinatal outcome
SE Little et al
180
Figure 2. Neonatal morbidity by gestational age. This figure displays
the rate of neonatal morbidities at each gestational week of delivery
for the entire study period combined. There was a statistically
significant difference for each morbidity by gestational week
(Po0.01 for all).
than for stillbirths and neonatal deaths). The decision to deliver a
patient is complex, and there are clinical factors driving scheduled
delivery that we were not able to capture. Moreover, we were
reliant on the delivering provider to record the gestational age and
delivery indication. Inaccuracies in coding, however, would not
hinder our overall findings as our goal was to assess real world
translation. In addition, even by the last year in our study, the rate of
elective delivery was still outside the recommended guideline of
5%, as per the Joint Commission standards. As such, our results may
not be generalizable to all populations.
We found that a reduction in elective early term delivery was
accompanied by a reduction in early term delivery for nonelective
indications. This may well be a positive finding; prolonging stable
high-risk pregnancies until 39 weeks may be an appropriate
strategy in many situations. Our study was not designed to
evaluate the appropriateness of delivery timing decision making. It
is possible, however, that the national emphasis on the
importance of achieving 39 weeks before elective delivery may
have the unintended consequence of encouraging providers to
prolong pregnancies for which attaining 39 weeks may pose more
risk than benefit. This emphasizes the continued need to study not
only the benefits of prolonging pregnancies to 39 weeks or
beyond but also the risks of delaying pregnancies where early
term delivery may be an appropriate strategy.
Overall, our study highlights that 39 week delivery policies may
have minimal impact on neonatal morbidity when they are
translated into the real world. At a single, large tertiary care center
the rate of early term delivery fell steadily, but there was only a
minor prolongation of pregnancy by just a few days to cross the
39 weeks threshold; this did not translate into any improvement in
neonatal morbidity. Continued research and surveillance efforts
should address both the true impact of the 39 week delivery
policies as well as the potential for unintended consequences as
we continue to pursue the national goal of eliminating
unnecessary early term delivery.
Journal of Perinatology (2014), 176 – 180
CONFLICT OF INTEREST
The authors declare no conflict of interest.
ACKNOWLEDGEMENTS
This research was presented orally (abstract no. 27) at the 33rd Annual Meeting
of the Society for Maternal-Fetal Medicine, San Francisco, CA, 11 to 16 February
2013.
REFERENCES
1 Tita AT, Landon MB, Spong CY, Lai Y, Leveno KJ, Varner MW et al. Timing of
elective repeat cesarean delivery at term and neonatal outcomes. N Engl J Med
2009; 360: 111–120.
2 Clark SL, Miller DD, Belfort MA, Dildy GA, Frye DK, Meyers JA. Neonatal and
maternal outcomes associated with elective term delivery. Am J Obstet Gynecol
2009; 200: 156. e1–4.
3 Ghartey K, Coletta J, Lizarraga L, Murphy E, Ananth CV, Gyamfi-Bannerman C.
Neonatal respiratory morbidity in the early term delivery. Am J Obstet Gynecol
2012; 207(4): 292. e1-4.
4 Zhang X, Kramer MS. Variations in mortality and morbidity by gestational age
among infants born at term. J Pediatr 2009; 154: 358–362.
5 Yang S, Platt RW, Kramer MS. Variation in child cognitive ability by week of
gestation among healthy term births. Am J Epidemiol 2010; 171: 399–406.
6 Chiossi G, Lai Y, Landon MB, Spong CY, Rouse DJ, Varner MW et al. Timing of
delivery and adverse outcomes in term singleton repeat cesarean deliveries.
Obstet Gynecol 2013; 121: 561–569.
7 Induction of labor. ACOG Practice Bulletin No. 107. American College of Obste-
tricians and Gynecologists. Obstet Gynecol 2009; 114: 386–397.
8 Medically indicated late-preterm and early-term deliveries. ACOG committee
opinion no. 560. American College of Obstetricians and Gynecologists. Obstet
Gynecol 2013; 121(4): 908–910.
9 Specifications Manual for Joint Commission National Quality Core Measures
[Version 2011A]. Oakbrook Terrace (IL): The Joint Commission; 2010 Dec. PC-1 to
PC-242 p.
10 Main E, Oshiro B, Chagolla B, Bingham D, Dang-Kilduff L, Kowalewski L.
Elimination of Non-medically Indicated (Elective) Deliveries Before 39 Weeks
Gestational Age. (California Maternal Quality Care Collaborative Toolkit to
Transform Maternity Care) Developed under contract #08-85012 with the Cali-
fornia Department of Public Health; Maternal, Child and Adolescent Health
Division; First edition published by March of Dimes 2010.
11 Donovan EF, Lannon C, Bailit J, Rose B, Iams JD, Byczkowski T. Ohio Perinatal
Quality Collaborative Writing Committee. A statewide initiative to reduce inap-
propriate scheduled births at 36(0/7)-38(6/7) weeks’ gestation. Am J Obstet
Gynecol 2010; 202(3): 243. e1–e8.
12 Ehrenthal DB, Hoffman MK, Jiang X, Ostrum G. Neonatal outcomes after
implementation of guidelines limiting elective delivery before 39 weeks of
gestation. Neonatal outcomes after implementation of guidelines limiting
elective delivery before 39 weeks of gestation. Obstet Gynecol 2011; 118(5):
1047–1055.
13 Oshiro BT, Henry E, Wilson J, Branch DW, Varner MW. Women and
Newborn Clinical Integration Program. Decreasing elective deliveries before 39
weeks of gestation in an integrated health care system. Obstet Gynecol 2009;
113(4): 804–811.
14 Mukhopadhyay S, Eichenwald EC, Puopolo KM. Neonatal early-onset sepsis eva-
luations among well-appearing infants: projected impact of changes in CDC GBS
guidelines. J Perinatol 2013; 33(3): 198–205.
& 2014 Nature America, Inc.