Obstetrics

Premature rupture of membranes at 34 to 37 weeksÕ gestation:

Aggressive versus conservative management
Robert W. Naef III, MD,c John R. Allbert, MD,b Elaine L. Ross, MD,a B. Michael Weber, MD,a
Rick W. Martin, MD,a and John C. Morrison, MDa
Jackson, Mississippi

OBJECTIVE: Our purpose was to compare induction of labor with preterm rupture of membranes between
34 and 37 weeks’ gestation with expectant management.
STUDY DESIGN: In this prospective investigation 120 gravid women at ³34 weeks 0 days and <36 weeks 6
days of gestation were randomized to receive oxytocin induction (n = 57) or observation (n = 63).
RESULTS: Estimated gestational age at rupture of membranes (34.3 ± 1.4 weeks vs 34.5 ± 1.4 weeks) and
ultrasonographically estimated fetal weight (2230 ± 321 gm vs 2297 ± 365 gm) were equivalent between
groups (not significant). Chorioamnionitis occurred more often (16% vs 2%, p = 0.007), and maternal hospi-
tal stay (5.2 ± 6.8 days vs 2.6 ± 1.6 days, p = 0.006) was significantly longer in the control group. Neonatal
sepsis was also more common in the observation group (n = 3) than among induction patients (n = 0), but
the difference was not statistically significant.
CONCLUSION: Aggressive management of preterm premature rupture of the membranes at ³34 weeks 0
days of gestation by induction of labor is safe for the infant in our population and avoids maternal-neonatal
infectious complications. (Am J Obstet Gynecol 1998;178:126-30.)

Key words: Premature rupture of membranes, induction, chorioamnionitis

The reported incidence of premature rupture of the
membranes averages from 6% to 10%,1 and about 20%
of these cases occur before 37 weeks’ gestation.1
Although preterm premature ruptured membranes com-
plicates only 1% to 2% of all pregnancies, it is associated
with 40% of preterm deliveries and can result in signifi-
cant perinatal morbidity and mortality.1-3 Balancing the
risk of ascending infection and cord prolapse against the
hazards of prematurity is a complex process about which
there is little consensus.1-4 Many studies have demon-
strated significant benefits in expectant or conservative
management for gestations of <34 weeks, whereas the
management of pregnancies complicated by preterm
premature rupture of membranes between 34 and 37
weeks’ gestation continues to be a controversial issue.3, 4
In contrast, for membrane rupture in gestations at >37
weeks most recommend aggressive management by in-
From the Departments of Obstetrics and Gynecology, University of
Mississippi Medical Center,a Carolinas Medical Center,b and Keesler
U.S. Air Force Medical Center.c
Supported in part by the Vicksburg Hospital Medical Foundation,
Vicksburg, Mississippi.
Received for publication April 22, 1997; revised June 30, 1997; ac-
cepted August 13, 1997.
Reprints not available from the authors.
6/1/85430
duction of labor to avoid maternal and neonatal morbid-
ity.1, 5, 6
The most favorable approach in dealing with preterm
premature rupture of the membranes at 34 weeks 0 days
to 36 weeks 6 days of gestation, however, remains the
most conflicting area. Recent studies suggest that pro-
longing gestation beyond 34 weeks 0 days results in little
or no reduction in neonatal morbidity because the inci-
dence of significant morbidity and death in these infants
is not different from those delivered at >36 weeks 6
days.7, 8 Alternatively, it has been suggested that induc-
tion of labor after rupture of the membranes, particu-
larly in gravid women with an unfavorable cervix, may be
associated with increased rates of cesarean delivery.5, 9
The purpose of this study was to investigate differences
in maternal and neonatal morbidity associated with ac-
tive versus expectant management of premature rupture
of membranes at 34 weeks 0 days to 36 weeks 6 days of
gestation.
Material and methods
This prospective randomized study was performed
over a 2-year period (1992 to 1994) at the University of
Mississippi Medical Center in Jackson, Mississippi. The
clinical trial was approved by the University of Mississippi
Medical Center Institutional Review Board and written

126
Volume 178, Number 1, Part 1
Am J Obstet Gynecol
Table I. Demographic and clinical characteristics
Oxytocin Observation
(n = 57) (n = 63) Significance
Maternal age (yr) 22.8 ± 5.7 23.0 ± 5.4 p = 0.86
Race
Black 46 (81%) 48 (76%) p = 0.70
White 10 (18%) 13 (21%) p = 0.84
Native American 1 (2%) 1 (3%) p = 0.53
Gravidity 2.8 ± 1.6 2.7 ± 1.6 p = 0.84
Nulliparous 27 28 p = 0.63
EGA at PROM (wk) 34.3 ± 1.4 34.5 ± 1.4 p = 0.47
Ultrasonographic 2230 ± 321 2297 ± 350 p = 0.29
EFW (gm)
EGA, Estimated gestational age; PROM, premature rupture of
membranes; EFW, estimated fetal weight.
informed consent was obtained from all participants.
Candidates for this study included all patients with sin-
gleton gestations between 34 weeks 0 days and 36 weeks 6
days of gestation complicated by preterm premature rup-
ture of the membranes. Gestational age was confirmed
by a reliable last menstrual period, early sonogram, or
first-trimester pelvic examination. In the absence of se-
cure dates, an ultrasonographic estimate of fetal weight
³1800 gm and ²2500 gm on admission was used as the
threshold for entry in the study because this equates with
34 weeks 0 days to 36 weeks 6 days in our population8 and
in others like it.9
Amniorrhexis was confirmed by visualization of pool-
ing fluid in the posterior vaginal fornix through the
cervix during sterile speculum examination or fern ar-
borization specimen under microscopic review plus a
positive nitrazine. No digital examination was performed
in any patient. Exclusion factors included noncephalic
presentation, fetal distress, labor on admission, and med-
ical or obstetric complications (such as suspected
chorioamnionitis, hypertensive disorders, diabetes melli-
tus, active genital herpes, placenta previa, infection,
meconium-stained fluid, or severe fetal anomalies).
Eligible and consenting gravid women were randomly
assigned to either expectant management (observation)
or active treatment (oxytocin) group by use of blinded
computer-generated random number cards in opaque
sealed envelopes. Patients assigned to the obser vation
group were assessed in the labor and delivery suite for 2
to 4 hours with continuous external fetal heart rate
(FHR) monitoring and tocodynamometry. In the ab-
sence of nonreassuring fetal status, initiation of labor, or
infection, these women were transferred to an antepar-
tum room where maternal vital signs and the FHR were
monitored every 8 hours. Patients were restricted to bed
rest with bathroom privileges and remained hospitalized
until delivery. Treatment with 2 gm of ampicillin intra-
venously for group B streptococcal prophylaxis was car-
ried out in all patients. Tocolysis or corticosteroid treat-
Naef et al. 127
Table II. Maternal outcome
Oxytocin Observation
(n = 57) (n = 63) Significance
Admission-to-delivery 9.8 ± 7.8 119 ± 223 p = 0.001
interval (hr)
Chorioamnionitis 1 (2%) 10 (16%) p = 0.007
Total hospital stay 2.6 ± 1.6 5.2 ± 6.8 p = 0.006
(days)
Type of delivery
SVD 41 (72%) 48 (76%) p = 0.746
Forceps 12 (21%) 12 (19%) p = 0.964
Cesarean section 4 (7%) 3 (5%) p = 0.444
SVD, Spontaneous vaginal delivery.
ments were not used because we felt that efficacy for
these techniques from 34 weeks 0 days to 36 weeks 6 days
was not proved and because of possible bias by alternat-
ing gestational age or incidence of infant respiratory dis-
tress as outcome variables. Criteria for delivery in this
group included evidence of nonreassuring fetal status
(recurrent decelerations or persistent fetal tachycardia),
initiation of labor, or signs of clinical chorioamnionitis.
Patients assigned to active management underwent in-
duction of labor with intravenous oxytocin with use of a
controlled infusion pump. Oxytocin was administered by
continuous intravenous infusion beginning at 0.5
mU/min, doubling the dose every 30 minutes to 2
mU/min, and then increasing by 2 mU/min every 30
minutes thereafter until a satisfactory labor pattern was
achieved. In both groups digital examinations were pro-
hibited (in the absence of progressive labor on the basis
of speculum examination) and cesarean delivery was per-
formed for standard obstetric indications. Clinical
chorioamnionitis was defined, in the absence of other
causes of hyperpyrexia, by a temperature of >100.4° F
with either uterine tenderness (or contractions), leuko-
cytosis, maternal or fetal tachycardia, or a foul-smelling
vaginal discharge. All patients with chorioamnionitis re-
ceived intravenous ampicillin and gentamicin, regardless
of group assignment, and the antibiotic therapy was con-
tinued until the patient was afebrile for 48 hours post
partum. In the observation group, if chorioamnionitis
was diagnosed, patients also received oxytocin induction
and antibiotic therapy. Postpartum endometritis was de-
fined as a temperature of >100.4° F after the first 24 post-
partum hours with associated uterine tenderness.
The infants were managed by neonatologists who were
not blinded to the perinatal clinical course. Intrapartum
maternal hyperpyrexia or clinical findings suggestive of
neonatal infection resulted in admission to the neonatal
intensive care unit for a sepsis evaluation. The diagnosis
of neonatal sepsis was made only in neonates with posi-
tive blood cultures. However, all babies with suspected
128 Naef et al. January 1998
Am J Obstet Gynecol

Table III. Neonatal outcome ment (n = 63). The demographic and clinical character-
istics of each group are shown in Table I. Maternal age
Oxytocin Observation
(n = 57) (n = 63) Significance
and gravidity were similar and there was no difference in
the racial distribution within each group. The duration
Birth weight (gm) 2314 ± 347 2416 ± 365 p = 0.121 of pregnancy at preterm premature rupture of mem-
Apgar score, 1 min 8.3 ± 1.6 8.3 ± 1.3 p = 0.943 branes and the estimated fetal weight were also equiva-
Apgar score, 5 min 9.1 ± 0.9 9.1 ± 0.7 p = 0.613
Cord pH 7.34 ± 0.07 7.25 ± 0.06 p = 0.009 lent. More than 95% of the women in each group were
NICU admission 11 (19%) 15 (24%) p = 0.706 admitted to the hospital within 24 hours of membrane
RDS 3 (5%) 3 (5%) p = 0.611 rupture.
Mechanical 2 (3%) 3 (5%) p = 0.548
ventilation Maternal pregnancy outcome and maternal morbidity
Sepsis 0 3 (5%) p = 0.151 are summarized in Table II. As expected, the latency
Total hospital stay 4.5 ± 4.9 4.8 ± 5.1 p = 0.693 from admission to delivery was significantly prolonged in
(days)
the observation group (p < 0.001). The overall incidence
NICU, Neonatal intensive care unit. of clinical chorioamnionitis was eight times higher
among women managed expectantly (2% vs 16%, p =
0.007). All but one of these 10 patients had signs and
sepsis received empiric broad-spectrum antibiotic ther- symptoms of clinical chorioamnionitis before the onset
apy until cultures returned and were negative (usually 3 of active labor.
days). Antibiotic therapy was continued for 7 to 10 days Similarly, maternal hospitalization was significantly
in those neonates with culture-proved sepsis. prolonged for women in the expectant observation
Other neonatal outcome variables of interest included group, with these women having on the average 3 addi-
respiratory distress syndrome (RDS), bronchopul- tional days in the hospital compared with patients who
monary dysplasia, patent ductus arteriosus, intraventricu- underwent induction of labor (5.2 ± 6.8 days vs 2.6 ± 1.6
lar hemorrhage, and necrotizing enterocolitis. RDS was days, p = 0.006).
defined as the early onset of tachypnea, retractions, and The incidence of cesarean delivery was similar in each
oxygen requirement for >24 hours or mechanical ventila- group and was lower than the 18% incidence noted in
tion with radiographic confirmation. Bronchopul- our overall obstetric population during the study period
monary dysplasia was diagnosed when the need for sup- (7% vs 5%, p = 0.444). All cesarean deliveries were per-
plemental oxygen or mechanical ventilation persisted formed for either nonreassuring fetal status or secondary
after the twenty-eighth day of life. Patent ductus arterio- arrest of labor. No abdominal delivery was performed for
sus was defined as a systolic murmur accompanied by failed induction of labor. Abnormal FHR patterns (repet-
bounding peripheral pulses with either development of itive variable decelerations) during labor were common
pulmonary edema or echocardiographic confirmation of and of similar incidence in both groups (32% vs 27%, p =
a left-to-right shunt. Intraventricular hemorrhage was di- 0.725). Umbilical cord prolapse occurred in only one pa-
agnosed with ultrasonography or computerized tomogra- tient in the study. Additionally, similar numbers of pa-
phy. Necrotizing enterocolitis was diagnosed by a combi- tients required operative intervention during vaginal de-
nation of abdominal distention and evidence of livery (p = 0.96) or had postpartum endometritis (p =
intramural air or perforation. 0.18) in each group.
Sample size calculation indicates that a change in the Neonatal outcome variables are depicted in Table III.
chorioamnionitis rate from 1% to 10% would require 55 Birth weight was similar in the two groups and correlated
patients in each group to attain significance at 0.05 with a well with the ultrasonographic estimated fetal weights ob-
power of 80%. More than 500 patients in each group tained before delivery (see Table I). Apgar scores at 1
would be necessary to obtain significance in RDS rates and 5 minutes were nearly identical; however, the arterial
(3% to 6%). All randomized patients remained in their cord pH was higher in the active management group
assigned group and were included in the statistical evalu- (7.34 ± 0.07 vs 7.25 ± 0.06, p = 0.009). This difference is
ation at the completion of the study. Statistical analysis of probably not clinically significant because no infant had
group differences was accomplished with use of the c2 a pH <7.0 and only two infants <7.2 in the observation
test and Fisher’s exact test as appropriate for discrete group (both had normal 5-minute Apgar scores). More
data and the Student t test for continuous data. babies in the observation group were admitted to the
Significant differences were accepted at p ² 0.05. neonatal intensive care unit, but the difference did not
reach statistical significance (24% vs 19%, p = 0.706).
Results Neonatal complications, such as hyaline membrane
During the 2-year study period 120 women were ran- disease and the need for mechanical ventilation or sup-
domized to either active (n = 57) or expectant manage- plemental oxygen >24 hours were similar between the
Volume 178, Number 1, Part 1
Am J Obstet Gynecol
two groups. Culture-proved sepsis occurred in three
neonates in the expectant management group (5%); all
recovered without sequela (staphylococcal epidermitis,
group D streptococci), whereas no infant in the active
management group had sepsis but this difference was
not statistically significant. No cases of intraventricular
hemorrhage, patent ductus arteriosus, or bronchopul-
monary dysplasia were noted in either group. The aver-
age neonatal hospital stay was similar between groups
(4.5 ± 4.9 days vs 4.8 ± 5.1 days, p = 0.693). There were no
stillbirths or neonatal deaths in the 120 patients.
Comment
Several studies have evaluated women with premature
rupture of membranes at term and have suggested that
increased maternal infection and abdominal delivery
rates occur with induction of labor.10, 11 Alternatively,
other investigators have shown that expectant manage-
ment is associated with significantly increased neonatal
infectious morbidity.6, 12 Three studies5, 13, 14 have been
undertaken on patients with preterm premature rupture
of the membranes at <37 weeks’ gestation. Spinnato et
al.5 found significantly increased maternal infectious
morbidity without identifiable neonatal benefit after ex-
pectant management. Nelson et al.13 showed a similar in-
cidence of cesarean delivery and neonatal infection with
a 50% increase in maternal infectious morbidity with ac-
tive management and induction of labor (18% vs 12%,
not significant). In contrast, Mercer et al.14 found a 52%
increase in maternal infectious morbidity (29.8% vs
19.6%, not significant) with expectant management but
a similar incidence of abdominal delivery and neonatal
infection. Therefore it appears that the relative benefits
and risks of active versus expectant management after
premature rupture of membranes in near-term gestation
have not been clearly elucidated.
Delaying delivery appears to be associated with a po-
tential increase in infectious morbidity for both the
mother and neonate. Prolonged latency offers time for
the development of chorioamnionitis, which is either
subclinical at the time of rupture of the membranes or
arises as a result of ascent of bacteria subsequent to rup-
ture. Thus it was our contention that the patient with
preterm premature rupture of membranes at 34 weeks 0
days to 36 weeks 6 days of gestation may benefit from ac-
tive intervention leading to delivery compared with ex-
pectant observation. The data from this study confirmed
a significantly increased risk of clinical chorioamnionitis
in the expectantly managed patient (16% vs 2%, p =
0.007). There was also an increased incidence of culture-
proved neonatal sepsis (4.8% vs 0%) associated with ex-
pectant observation, although this difference did not
reach statistical significance. These findings are consis-
tent with other reports that have suggested an association
Naef et al. 129
between expectant management and increased maternal
and neonatal infectious morbidity.5, 13, 14
We did not find an increase in abdominal delivery
after active management with induction of labor, as sug-
gested by studies of premature rupture of the mem-
branes at term10, 11 and none of our abdominal deliveries
in the active management group were a result of failed
induction. The incidence of cesarean delivery in both
groups was lower than that observed in our general ob-
stetric population during the study period (18%), and we
feel that this finding is the result of exclusion of patients
admitted with suspicious fetal heart tracings or meco-
nium-stained amniotic fluid.
The morbidity from preterm delivery was low in this
group of near-term neonates, as shown by our data and
those of others.7-9 Other than sepsis as noted above, the
only other neonatal complication observed was RDS,
which occurred with similar frequency in both groups
(5% vs 5%, p = 0.611). Furthermore, the observed inci-
dence of RDS in each group was lower than that pre-
dicted by Robertson et al.7 of 13.5% at 34 weeks’ gestation
and 6% at 35 weeks’ gestation. This discrepancy may be
the result of a different patient population: poor black
women in our study versus affluent white patients in in-
vestigation of Robertson et al. based in northern
California. None of the neonates in our study with hyaline
membrane disease required prolonged mechanical venti-
lation or other aggressive supportive measures. Thus no
neonatal benefit was observed from delayed delivery in
the expectant management group, although the numbers
are too small to prevent effective statistical analysis.
As expected, the total maternal hospital stay was signif-
icantly prolonged among patients in the observation
group. Most of this increase is accounted for by the in-
creased latency-to-delivery interval among these women.
Importantly, an additional 164 days of hospitalization
were required in the observation group, leading to in-
creased cost in health care dollars without evident neona-
tal benefit.
In conclusion, in a limited population induction of
labor in patients with preterm premature rupture of the
membranes at 34 weeks 0 days to 36 weeks 6 days of ges-
tation is associated with less chorioamnionitis and neona-
tal infectious morbidity. Active management of these pa-
tients did not result in an increase in cesarean delivery.
In contrast, expectant management was associated with
increased maternal infectious morbidity and a signifi-
cantly longer duration of maternal hospitalization. In the
observation group there were no significant neonatal
benefits among these patients who are already at low risk
for complications of prematurity. Patients with preterm
premature rupture of the membranes at 34 to 37 weeks’
gestation should be considered as candidates for induc-
tion of labor.
130 Naef et al. January 1998
Am J Obstet Gynecol

REFERENCES
1. Maxwell GL. Preterm premature rupture of membranes. Obstet
Gynecol Surv 1993;48:576-83.
2. Merenstein GB, Weisman LE. Premature rupture of the mem-
branes: neonatal consequences. Semin Perinatol 1996;20:375-
80.
3. Douvas SG, Brewer JM, McKay ML, Rhodes PG, Kahlstorf JH,
Morrison JC. Treatment of preterm rupture of the membranes.
J Reprod Med 1984;29:741-4.
4. Ghidini A, Romero R. PROM: induction versus expectant man-
agement. Contemp Ob/Gyn 1993;38:79-85.
5. Spinnato JA, Shaver DC, Bray EM, Lipshitz J. Preterm prema-
ture rupture of the membranes with fetal pulmonary maturity
present: a prospective study. Obstet Gynecol 1987;69:196-201.
6. Rydhström H, Ingemarsson I. No benefit from conservative
management in nulliparous women with premature rupture of
the membranes (PROM) at term: a randomized study. Acta
Obstet Gynecol Scand 1991;70:543-7.
7. Robertson PA, Sniderman SH, Laros RK Jr, Cowan R, Heilbron
D, Goldenberg RL, et al. Neonatal morbidity according to gesta-
tional age and birth weight from five tertiary care centers in the
United States, 1983 through 1986. Am J Obstet Gynecol
1992;166:1629-45.
8. Fox MD, Allbert JR, McCaul JF, Martin RW, Roberts WE,
McLaughlin BN, et al. Neonatal morbidity between 34 and 37
weeks’ gestation. J Perinatol 1993;13:349-53.
9. Copper RL, Goldenberg RL, Creasy RK, et al. A multicenter
study of preterm birth weight and gestational age–specific
neonatal mortality. Am J Obstet Gynecol 1993;168:78-84.
10. Kappy KA, Cetrulo CL, Knuppel RA, Ingardia CJ, Sbarra AJ,
Scerbo JC, et al. Premature rupture of the membranes at term:
a comparison of induced and spontaneous labors. J Reprod
Med 1982;27:29-33.
11. Conway DI, Prendiville WJ, Morris A, Speller CE, Stirrat GM.
Management of spontaneous rupture of the membranes in the
absence of labor in primigravid women at term. Am J Obstet
Gynecol 1984;150:947-51.
12. Van Reempts P, Kegelaers B, Van Dam K, Van Overmeire B.
Neonatal outcome after very prolonged and premature rupture
of membranes. Am J Perinatol 1993;10:288-91.
13. Nelson LH, Meis PJ, Hatjis CG, Ernest JM, Dillard R, Schey HM.
Premature rupture of membranes: a prospective randomized
evaluation of steroids, latent phase, and expectant manage-
ment. Obstet Gynecol 1985;66:55-9.
14. Mercer BM, Crocker LG, Boe NM, Sibai BM. Induction versus
expectant management in premature rupture of the mem-
branes with mature amniotic fluid at 32 to 36 weeks: a random-
ized trial. Am J Obstet Gynecol 1993;169:775-82.

Availability of Journal back issues

As a service to our subscribers, copies of back issues of the American Journal of Obstetrics
and Gynecology for the preceding 5 years are maintained and are available for pur-
chase from Mosby until inventory is depleted at a cost of $15.00 per issue. The following
quantity discounts are available: 25% off on quantities of 12 to 23 and one third off on quan-
tities of 24 or more. Please write to Mosby, Inc., Subscription Services, 11830 Westline
Industrial Drive, St. Louis, MO 63146-3318, or call (800)453-4351 or (314)453-4351 for in-
formation on availability of particular issues. If back issues are unavailable from the pub-
lisher, photocopies of complete issues may be purchased from UMI, 300 N. Zeeb Road, Ann
Arbor, MI 48106-1346. Tel.: (313)761-4700.