240

Labor induction is considered elective when it is un-

dertaken for the purpose of convenience and in the ab-
sence of any maternal or fetal condition that justifies
delivery.

1

Elective labor induction is commonly practiced

in Flanders (Northern Belgium). From 1996 to 1997,
30% of all deliveries were induced, of which 71% were
elective. Therefore, 1 of 5 pregnant woman in Flanders
had a labor induction for convenience. The outcome of
elective induction compared with labor of spontaneous
onset has not been extensively studied; to our knowledge,
only 1 small randomized controlled trial is on record.

2

Case-controlled studies or randomized trials

3

showed

that elective induction leads to more operative delivery,
more need for pain relief, but less meconium in labor.
However, these conclusions could not be generalized

merely because of the lack of a sizeable number of pa-
tients included in the different studies. The purpose of
this study was to determine whether the current practice
of elective labor induction in Flanders was associated with
differences in mode of delivery, demand for pain relief,
and need for neonatal care when compared with labor of
spontaneous onset.

Patients and methods

The region of Flanders covers 13,524 km

2

and has 5.9

million inhabitants. There are 80 maternity-obstetric
units: 4 in university hospitals, 15 in teaching hospitals,
and 61 in peripheral hospitals. Ninety-five percent of all
deliveries are performed or supervised by a gynecologist.
Most gynecologists practice in the private sector.

The following data were derived from the existing com-

puter files of the Flemish Centre for Perinatal Epidemiol-
ogy (SPE). For each newborn of at least 500 g, an official
perinatal form is completed (most often by the midwife)
and sent to the SPE, where all data are controlled by an
error detection program and feedback is provided. Dur-
ing the study period (January 1, 1996–December 31,
1997) 124,680 deliveries took place. Labor was induced
in 30% of the deliveries, and elective labor induction
took place in 21% of the deliveries. Of the deliveries, 46%

From the Department of Gynecology, Andrology, and Obstetrics,
Academic Hospital–Free University Brussels,

a

and the Department of

Medical Statistics, University Hospital, Ghent.

b

Received for publication April 27, 2001; revised August 7, 2001;
accepted August 21, 2001.
Reprint requests: H. Cammu, MD, Academic Hospital–Free Uni-
versity Brussels, Department of Gynecology, Andrology, and
Obstetrics, Laarbeeklaan, 101, B-1090, Brussels, Belgium. E-mail:
[e-mail].
Copyright 2002, Mosby, Inc. All rights reserved.
0002-9378/2002 $35.00 + 0

6/1/119643

doi:10.1067/mob.[phone]

Outcome after elective labor induction in nulliparous
women: A matched cohort study

Hendrik Cammu, MD,

a

Guy Martens, BS,

b

Gunther Ruyssinck, BS,

b

and

Jean-Jacques Amy, MD

a

Brussels and Ghent, Belgium

OBJECTIVE: To determine whether elective induction of labor in nulliparous women is associated with
changes in fetomaternal outcome when compared with labor of spontaneous onset.
STUDY DESIGN: All 80 labor wards in Flanders (Northern Belgium) comprised a matched cohort study.
From 1996 through 1997, 7683 women with elective induced labor and 7683 women with spontaneous labor
were selected according to the following criteria: nulliparity, singleton pregnancy, cephalic presentation, ges-
tational age at the time of delivery of 266 to 287 days, and birth weight between 3000 and 4000 g. Each
woman with induced labor and the corresponding woman with spontaneous labor came from the same labor
ward, and they had babies of the same sex. Both groups were compared with respect to the incidence of
cesarean delivery or instrument delivery and the incidence of transfer to the neonatal ward.
RESULTS: Cesarean delivery (9.9% vs 6.5%), instrumental delivery (31.6% vs 29.1%), epidural analgesia
(80% vs 58%), and transfer of the baby to the neonatal ward (10.7% vs 9.4%) were significantly more com-
mon (P < .01) when labor was induced electively. The difference in cesarean delivery was due to significantly
more first-stage dystocia in the induced group. The difference in neonatal admission could be attributed to a
higher admission rate for maternal convenience when the women had a cesarean delivery.
CONCLUSION: When compared with labor of spontaneous onset, elective labor induction in nulliparous
women is associated with significantly more operative deliveries. Nulliparous women should be informed
about this before they submit to elective induction. (Am J Obstet Gynecol 2002;186:240-4.)

Key words:

Elective labor induction, cesarean delivery

Volume 186, Number 2

Cammu et al 241

Am J Obstet Gynecol

were nulliparous; 14.3% were cesarean deliveries; 13.2%
were instrumental deliveries; 53% required epidural
analgesia; 16.8% were transferred to the neonatal ward;
and 1.7% resulted in congenital malformations. The peri-
natal mortality rate was 0.7%.

The inclusion criteria for the women with electively in-

duced labor and those with a labor of spontaneous onset
were as follows: nulliparous women with a low-risk, sin-
gleton pregnancy in cephalic presentation, and delivery
of a live-born infant. Both study groups were matched for
maternal age, gestational age (between 266 and 287
days), birth weight (between 3000 and 4000 g), and sex of
the infant. These limitations were chosen to avoid
prematurity, postdating, small for gestational age, and
macrosomic babies. For each woman who was electively
induced, the control patient with spontaneous onset of
labor came from the same obstetric unit to minimize dif-
ferences in treatment. Fetal deaths, non–hospital deliver-
ies, and deliveries not performed or supervised by a
gynecologist were excluded from these subgroups. By
doing so, 7683 cases of electively induced labor and 7683
cases of spontaneous labor were selected. In both groups,
we determined the incidence of cesarean delivery, instru-
mental delivery, epidural analgesia use, transfer to the
neonatal ward, congenital malformations, and neonatal
death. Cesarean delivery for first- and second-stage dysto-
cia means that the main reason for performing a ce-
sarean delivery was a failure to progress to full dilatation
(first stage) or a failure of descent after full dilatation
(second stage). Fetal distress was diagnosed by the gyne-
cologist in charge, according to his/her own criteria. The
nonparametric Mann-Whitney U test was used for com-
paring the induction variable in the hospitals with a dif-
ferent level of specialization. Chi-square tests were used
to evaluate the association between discrete variables.
The relative risks were given together with their 95% CIs
to measure the strength of these associations. The level of
statistical significance was set at a probability value of .01.
Interrelations between the risk factors were assessed with
stepwise logistic regression analysis.

Results

During the study period, the incidence of elective

labor induction, which ranged from 10% to 48%, varied
greatly between the various Flemish labor wards. Elective
labor induction was most frequently performed in the 4
university hospitals (mean, 25%) followed by the teach-
ing hospitals (mean, 23%). The induction ratio in the
peripheral hospitals amounted to 20% (P < .001, Mann-
Whitney U test).

Table I shows that women with either induced or spon-

taneous labor were comparable. Maternal age of

≥35

years (3.1% vs 2.9%) and the birth weight of the baby
from 3750 to 4000 g (15% vs 14%) were almost equally
distributed.

Women whose labor had been induced more often re-

ceived an epidural analgesia (Table II). Induced labor was
associated with significantly more cesarean deliveries and
slightly, but significantly, more instrumental deliveries.
The increased frequency of cesarean delivery in induced
labor was predominantly the result of a higher incidence
of first-stage dystocia (Table II). Fetal distress as a reason
for cesarean delivery was more frequently encountered
when labor had been electively induced (Table II). We de-
termined that, when compared with spontaneous labor,
every 29 elective inductions accounted for 1 extra ce-
sarean delivery (95% CI, 23.4-39.4).

Women who had labor induced were more prone to

deliver during working hours and significantly less during
the night (Table III).

Babies who were born after induced labor were trans-

ferred more often to the neonatal ward (Table IV). This in-
creased rate in neonatal admission of the babies who were
born after induced labor was only seen during the first 48
hours (Table IV). Reasons for transfer to the neonatal
ward were comparable in both groups: asphyxia and respi-
ratory problems in 41% (induced labor) and 40%
(spontaneous labor), respectively; suspicion of perinatal
infection in 16% (induced labor) and 17% (spontaneous
labor) respectively; and hyperbilirubinemia in 17% (in-
duced labor) and 18% (spontaneous labor), respectively.

Cesarean delivery as the sole reason for being trans-

ferred to the neonatal ward was mentioned more fre-
quently for the babies who were born after induced labor
(16%) than after spontaneous labor (12%; P < .01; relative
risk, 1.08 < 1.39 < 1.79). Thirty percent of the babies who
were born after induced labor that ended in a cesarean
delivery were admitted to the neonatal ward (222/762 ba-
bies), as compared with 33% of the babies who were born
after spontaneous labor that ended in a cesarean delivery
(166/500 babies;

χ

2

, 2.16; P = .14, not significant). In both

the induced labor group (316/2429 women) and the
spontaneous labor group (297/2236 women), 13% of the
babies who were delivered instrumentally were trans-
ferred to the neonatal ward (

χ

2

, 0.06; P = .81, not signifi-

cant). A stepwise logistic regression was performed to
elucidate independent variables that were associated with
induced or spontaneous labor (Table V). This analysis
confirmed the association of cesarean delivery and
epidural analgesia with the type of labor but showed that
babies who were born after induced labor were not trans-

Table I. Patient characteristics

Induced Spontaneous

labor

labor

Characteristic

(n = 7683)

(n = 7683)

Maternal age (y ± SD)

27.0 ± 4.1

27.2 ± 4

Birth weight (g ± SD)

3461 ± 264

3456 ± 264

Gestational age (wk ± SD)

39.8 ± 0.9

39.8 ± 0.9

Male baby (%)

51.2

51.2

242 Cammu et al

February 2002

Am J Obstet Gynecol

ferred significantly more often to the neonatal ward be-
cause of an impaired condition of the neonate. There
were 82 congenital malformations (1%) in the induced
labor group and 77 congenital malformations (1%) in the
spontaneous labor group. Neonatal death within the first
week occurred 6 times in the induced labor group and 6
times in the spontaneous labor group.

Comment

The wide variation in frequency of elective induction

between labor wards in the same region was also found by
others.

4-6

A Finnish team

4

found that the level of special-

ization of the hospital was a determinant in the incidence
of elective induction: the least frequent at the highest
level of specialization, the most frequent at the lowest
level. Another finding was that the induction practice was
not consistent in the hospitals of the same level. There-
fore, it was concluded

4

that the opinions of individual

practitioners and staff routines had influenced the induc-
tion policy to an important degree. A Dutch study

5

showed large differences in labor induction rates among
hospitals, even after an adjustment was made for popula-
tion differences. An American study

6

came to similar con-

clusions: large differences were seen in induction rates
across levels of nonclinical factors, such as ethnicity and
hospital type.

We also found different induction rates between hospi-

tals, but unlike the investigators in the Finnish

4

and the

American

6

studies, we observed that elective induction of

labor was significantly more frequent in university and
teaching hospitals than in peripheral hospitals. No con-
clusive explanation can be given for that finding. Possibly,

the availability of epidural analgesia played a role. We
have previously reported

7

that, for low-risk labor, this

form of analgesia was significantly more often applied in
university and teaching hospitals (higher availability)
than in peripheral hospitals (lower availability). Unfortu-
nately we have no data on the social class and marital sta-
tus of the parturients, but Järvelin et al

4

found that

neither the socioeconomic level nor the marital status
were significant determinants. The use of epidural anal-
gesia is more commonly associated with electively in-
duced labor than with labor of spontaneous onset. This
has been reported consistently in the literature since
1975.

8,9

It remains unknown whether epidural analgesia

influences induction, or whether it is the other way
around. A reason for electively inducing labor might be
that the women want to make the best possible use of the
anesthetic facilities. But, induced labor can be slower and
more arduous than spontaneous labor and therefore re-
quire greater doses of oxytocin. This can lead to more
epidural analgesia.

10

Elective induction leads to more

daytime deliveries (Table III). This is probably 1 of the
main reasons for resorting to elective induction.

3,11

A

day-time delivery can have administrative and organiza-
tional advantages. It allows better planning for the pa-
tient, her family, and the medical staff.

Elective induction is associated with more cesarean

deliveries. Authors who evaluated nulliparous women
separately, as we did, found an increase in the rate of ce-
sarean deliveries in the group that was electively in-
duced.

9,11-14

However, this increase did not always reach

statistical significance.

11,14

Two studies

11,14

were case-con-

trolled and involved a limited number of patients. Ce-
sarean delivery rates were, respectively, 12%

11

and 6.2%

14

higher after induction of labor than after labor of spon-
taneous onset. The differences were not significant be-
cause of the lack of power. Two other studies

9,12

compared a relatively small group of women who were
electively induced with a large group of women in spon-
taneous labor. Differences in cesarean delivery rates were
significant: 11.5%

9

and 9.7%,

12

respectively. Our study is

a large matched cohort study that clearly shows that elec-
tive induction is associated with significantly more ce-
sarean delivery (relative risk, 1.52), which is a somewhat

Table II. Labor and delivery characteristics

Characteristic

Induced labor (n)

Spontaneous labor (n)

Relative risk

(n = 7683)

(n = 7683)

Epidural analgesia*

6128 (79.8%)

4428 (57.6%)

1.35

≤ 1.38 ≤ 1.42

Cesarean delivery*

762 (9.9%)

500 (6.5%)

1.37

≤ 1.52 ≤ 1.70

Dystocia, 1st stage

454 (5.9%)

250 (3.3%)

1.56

≤ 1.82 ≤ 2.11

Dystocia, 2nd stage

96 (1.2%)

97 (1.3%)

0.75

≤ 0.99 ≤ 1.31

Fetal distress

197 (2.6%)

139 (1.8%)

1.14

≤ 1.42 ≤ 1.76

Other reasons

15 (0.2%)

14 (0.2%)

Instrument delivery*

2429 (31.6%)

2236 (29.1%)

1.04

≤ 1.09 ≤ 1.14

*P = .001.

Table III. Time of delivery

Induced labor

Spontaneous labor

Time span

(n)

(n)

8:00

AM

-4:59

PM

(A)

4985 (64.9%)

4.190 (54.5%)

6:00

PM

-11:59

PM

(B)

2088 (27.1%)

1.638 (21.3%)

12:00

AM

-7:59

AM

(C)

610 (7.9%)

1855 (24.1%)

A versus B: Relative risk, 0.94

≤ 0.97 ≤ 1.00.

A versus C: Relative risk, 2.04

≤ 2.20 ≤ 2.36.

B versus C: Relative risk, 2.10

≤ 2.26 ≤ 2.44.

Volume 186, Number 2

Cammu et al 243

Am J Obstet Gynecol

lower risk than that found by Maslow and Sweeny

9

(rela-

tive risk, 2.4) and by Seyb et al

12

(relative risk, 2.49). It

comes close to the relative risk of 1.38 found by Coonrod
et al.

6

However, this latter epidemiologic study combined

elective and nonelective inductions.

The main reason for the higher frequency of cesarean

delivery is the more common failure to progress in the
first stage of labor (Table II), which was also noted by
other researchers.

11,15

First-stage dystocia in nulliparous

women is more likely when the cervix is unfavorable at
the moment of induction.

3,11,12,14

In Flanders, two thirds

of all induced labors are preceded by a cervical ripen-
ing.

16

Unfortunately, information concerning the cervi-

cal status before induction was not available for this study.
Maslow and Sweeny

9

found that, in induced labor in nul-

liparous women, both an increased maternal age and an
increased weight of the baby accounted for an increase in
the percentage of cesarean deliveries. In our study, birth
weight and maternal age were kept comparable in both
groups. The small increase in cesarean delivery for fetal
distress that we found in induced labor may be indirectly
the result of dystocia. A longer labor means a longer feto-
maternal exposure to stress. The diagnosis of a subopti-
mal cardiotocogram is more likely to be put forward
when labor is prolonged, and all parties want labor to
come to an end.

There was only a slight increase in instrumental deliv-

eries when labor had been induced (Table II); this could
be related possibly to the more frequent application of
epidural analgesia in this group.

7

Macer et al,

11

in their case-controlled study, noticed

that babies who were born after spontaneous labor were
more likely to be admitted to the neonatal intensive care
unit. We found the opposite to be true (Table IV), al-

though the relative risk of 1.14 of admission of the baby
to the neonatal ward after induced labor is very modest.
Moreover, a logistic regression analysis failed to confirm
this association. Having had a cesarean delivery was, on
its own, an important reason for neonatal transfer. This
could be due to the fact that women who underwent this
procedure were given some relief in the immediate post-
operative phase.

Our study is the reflection of current obstetric practice

in the whole of a given region and not necessarily of the
best treatment of either induced or spontaneous labor.
We made use of a routinely collated database and
recorded what happened in reality.

We are aware of the fact that the randomized con-

trolled trial is the optimal method to measure differ-
ences between electively induced and spontaneous
labor. However, this is not easy to achieve. On the basis
of the results of this study, we would need a sample of
1017 subjects per arm to have enough power (

α, 5%;

power, 80%) to prove that a cesarean delivery rate of
9.9% after induced labor is statistically, significantly
higher than the 6.5% that was found when labor was of
spontaneous onset.

REFERENCES

1. Solomon JE, D’Alton ME. Induction of labor. In: Creasy RK, ed-

itor. Management of labor and delivery. Oxford: Blackwell Sci-
ence; 1997. p. 293-315.

2. Amano K, Saito K, Shoda T, Tani A, Yoshihara H, Nishijima M.

Elective induction of labor at 39 weeks of gestation: a prospec-
tive randomized trial. J Obstet Gynaecol Res 1999;25:33-7.

3. Jackson M, Regan C. Elective induction of labor. Clin Obstet Gy-

necol 1997;40:496-509.

4. Järvelin MR, Hartikainen-Sorri AL, Rantakallio P. Labour induc-

tion policy in hospitals of different levels of specialisation. Br J
Obstet Gynaecol 1993;100:310-15.

Table IV. Neonatal outcome

Outcome

Induced labor (n)

Spontaneous labor (n)

Relative risk*

Admitted to neonatal ward

824 (10.7%)

725 (9.4%)

1.03

≤ 1.14 ≤ 1.25

Duration of stay

0-48 Hours

437 (53%)

345 (47%)

48 Hours-7 days

180 (22%)

198 (27%)

>7 Days

90 (11%)

92 (13%)

Unknown duration

117 (14%)

90 (12%)

*P = .009,

χ

2

.

Table V. Outcome predictors of induced versus spontaneous labor: stepwise logistic regression

Independent predictor

B

P value

Odds radio

95% CI

Mode of delivery

0.271

<.001

1.311

1.162-1.478

Epidural analgesia

1.046

<.001

2.846

2.648-3.059

Admitted to neonatal ward

0.012

.911

NS

NS

Neonatal deaths

0.619

.431

NS

NS

NS, Not significant.

244 Cammu et al

February 2002

Am J Obstet Gynecol

5. Elferink-Stinkens PM, Brand R, le Cessie S, Van Hemel OJ. Large

differences in obstetrical intervention rates among Dutch hospi-
tals, even after adjustment for population differences. Eur J Ob-
stet Gynecol Reprod Biol 1996;68:97-103.

6. Coonrod DV, Bay RC, Kishi GY. The epidemiology of labor in-

duction: Arizona 1997. Am J Obstet Gynecol 2000;182:1355-62.

7. Cammu H, Martens G, Van Maele G. Epidural analgesia for low

risk labour determines the rate of instrumental deliveries but not
that of caesarean sections. J Obstet Gynaecol 1998; 18:25-9.

8. Cole RA, Howie PW, MacNaughton MC. Elective induction of

labour: a randomized prospective trial. Lancet 1975;1:767-70.

9. Maslow AS, Sweeny AL. Elective induction of labor as a risk fac-

tor for cesarean delivery among low-risk women at term. Obstet
Gynecol 2000;95:917-22.

10. Cammu H, Verlaenen H, Amy JJ, De Koster K, Derde MP,

Buekens P. Epidural analgesia in active management of labor.
Acta Obstet Gynecol Scand 1994;73:235-9.

11. Macer JA, Macer CL, Chan LS. Elective induction versus sponta-

neous labor: a retrospective study of complications and out-
come. Am J Obstet Gynecol 1992;166:1690-7.

12. Seyb ST, Berka RJ, Socol ML, Dooley SL. Risk of cesarean deliv-

ery with elective induction of labor at term in nulliparous
women. Obstet Gynecol 1999;94:600-7.

13. Yeast JD, Jones A, Poskin M. Induction of labor and the relation-

ship to cesarean delivery: a review of 7001 consecutive induc-
tions. Am J Obstet Gynecol 1999;180:628-33.

14. Prysak M, Castronova FC. Elective induction versus spontaneous

labor: a case-control analysis of safety and efficacy. Obstet Gy-
necol 1998;92:47-52.

15. Smith LP, Nagourney BA, McLean FH, Usher RH. Hazards and

benefits of elective induction of labor. Am J Obstet Gynecol
1984;148:579-85.

16. Defoort P, Spitz B, Huyghebaert G. Inductie in Vlaanderen.

Brussels: Jaarboek SPE; 1995. p. 62-73.