PRETERM PREMATURE RUPTURE OF THE MEMBRANES

Preterm premature rupture of membranes (PROM) affects over 120,000

pregnancies annually in the United States and is associated with significant
maternal, fetal, and neonatal risk. Management of PROM requires an accurate
diagnosis as well as evaluation of the risks and benefits of continued pregnancy
or expeditious delivery. An understanding of gestational agedependent
neonatal morbidity and mortality is important in determining the potential
benefits of conservative management of preterm PROM at any gestation.
Where possible, the treatment of pregnancies complicated by PROM remote
from term should be directed towards conserving the pregnancy and reducing
perinatal morbidity due to prematurity while monitoring closely for evidence of
infection, placental abruption, labor, or fetal compromise due to umbilical cord
compression. Current evidence suggests aggressive adjunctive antibiotic
therapy to reduce gestational agedependent and infectious infant morbidity.
Similarly, review of evaluable data indicates that antenatal corticosteroid
administration in this setting enhances neonatal outcome without increasing
the risk of perinatal infection. It is not clear that tocolysis in the setting of
preterm PROM remote from term reduces infant morbidity. When preterm PROM
occurs near term, particularly if fetal pulmonary maturity is evident, the patient
is generally best served by expeditious delivery.
INTRODUCTION Incidence and Clinical Importance Preterm premature rupture
of membranes (PROM) occurs in 3% of pregnancies and is responsible for
approximately one third of all preterm births. Preterm PROM is an important
cause of perinatal morbidity and mortality, particularly because it is associated
with brief latency from membrane rupture to delivery, perinatal infection, and
umbilical cord compression due to oligohydramnios. Even with conservative
management, 5060% of women with preterm PROM remote from term will
deliver within 1 week of membrane rupture. Amnionitis (1360%) and clinical
abruptio placentae (412%) are commonly associated with preterm PROM. The
risk of these complications increases with decreasing gestational age at
membrane rupture. The frequency and severity of neonatal complications after
preterm PROM vary with the gestational age at which rupture and delivery
occur, and are increased with perinatal infection, abruptio placentae, and
umbilical cord compression. Respiratory distress syndrome (RDS) is the most
common serious complication after preterm PROM at any gestation. Other
serious acute morbidities including necrotizing enterocolitis, intraventricular
hemorrhage, and sepsis are common with early preterm birth but relatively
uncommon near term. Remote from term, serious perinatal morbidity that may
lead to long-term sequelae or death is common. Figures 1 through 3 present
recent gestational agedependent morbidity and mortality curves from a
prospective community-based evaluation of 8523 consecutive women
delivering at six hospitals in Shelby County, Tennessee between July 1997 and
March 1998. In this evaluation, we found that 33% of live-born and resuscitated
infants delivered at 23 weeks survived to discharge from hospital (Figure 1).

Figure 1. Survival by gestational age among live-born resuscitated

infants. Results of a community-based evaluation of 8523 deliveries,
19971998, Shelby County, Tennessee. Curves smoothed by 2-point
average.
One-week increments in gestational age were associated with impressive
improvements in survival when delivery occurred between 23 and 32
weeksgestation. Among infants surviving to discharge, RDS (more than 24
hours oxygen requirement or ventilation in the absence of other evident cause
of respiratory compromise) was the most common acute morbidity at any
gestational age (Figure 2).
Figure 2. Acute morbidity by gestational age among surviving infants. Results
of a community-based evaluation of 8523 deliveries, 19971998, Shelby
County, Tennessee. Curves smoothed by 2-point average.

Among surviving infants, intraventricular hemorrhage and necrotizing

enterocolitis were rare when delivery occurred after 32 weeks. Blood- or
cerebrospinal cultureproven sepsis declined rapidly among those delivering
between 27 and 30 weeks, with a modest decline in sepsis for each week
gained thereafter. Although this study did not have the opportunity to measure
long-term morbidity, retinopathy of prematurity and bronchopulmonary
dysplasia occurred almost uniformly among survivors born at 23 weeks, and
were almost never seen with delivery at or after 32 weeks (Figure 3).
Figure 3. Chronic morbidity by gestational age among surviving infants.
Results of a community-based evaluation of 8523 deliveries, 19971998,
Shelby County, Tennessee. Curves smoothed by 2-point average
These findings were similar to a recent multicenter observational study of
women delivering infants who weighed less than 1000 g.1 In that study, 15 of
40 live-born singletons delivered at 23 weeks gestation survived (37.5%).
However, two thirds of survivors suffered major morbidities potentially
associated with long-term morbidity (any of the following: grades 3 to 4
intraventricular hemorrhage, grades 3 to 4 retinopathy of prematurity,
necrotizing enterocolitis requiring surgery, oxygen dependence at 120 days or
at discharge, or seizures). Current data specific to infants delivering after
preterm PROM are not available. However, it has been found that perinatal
sepsis is two-fold more common in the setting of preterm PROM than preterm
birth after preterm labor with intact membranes.
Definitions Premature rupture of the membranes is defined as spontaneous
membrane rupture that occurs before the onset of labor. When spontaneous
membrane rupture occurs before 37 weeks gestation, it is referred to as
preterm PROM. The term latency refers to the time from membrane rupture
to delivery. Conservative management is defined as treatment directed at
continuing the pregnancy. Preterm PROM that occurs at or before 26 weeks
gestation complicates 0.60.7% of pregnancies, and has been defined as

midtrimester PROM. Although the delineation of midtrimester PROM was

clinically relevant in the 1970s and 1980s, the limit of fetal viability has
progressively declined over the past 3 decades. As such it is currently more
clinically relevant to differentiate preterm PROM into previable PROM, which
occurs before the limit of viability (less than 23 weeks), preterm PROM remote
from term (from viability to about 32 weeks gestation), and preterm PROM
near term (approximately 3236 weeks gestation). When previable PROM
occurs, immediate delivery will lead to neonatal death. Conservative
management may lead to previable or periviable birth, but may also lead to
extended latency and delivery of a potentially viable infant. Immediate delivery
after preterm PROM remote from term is associated with a high risk of
significant perinatal morbidity and mortality that decreases with advancing
gestational age at delivery. Alternatively, with preterm PROM near term,
expeditious delivery of a noninfected and nonasphyxiated infant is associated
with a high likelihood of survival and a low risk of severe morbidity.
Pathophysiology
Premature rupture of membranes is multifactorial in nature. In any given
patient, one or more pathophysiologic processes may be evident.
Choriodecidual infection or in- flammation appears to play an important role in
etiology of preterm PROM, especially at early gestational ages.3 Decreased
membrane collagen content has been demonstrated in the setting of preterm
PROM and with increasing gestational age.4 In support of this, increases in
amniotic fluid matrix metalloproteases (1, 8, and 9) as well as decreases in
tissue inhibitors of matrix metalloproteases (1 and 2) have been identified
among women with preterm PROM.5,6 Other factors associated with preterm
PROM include lower socioeconomic status, cigarette smoking, sexually
transmitted infections, prior cervical conization, prior preterm delivery, prior
preterm labor in the current pregnancy, uterine distention (eg, twins,
hydramnios), cervical cerclage, amniocentesis, and vaginal bleeding in
pregnancy. Each of these may be associated with preterm PROM through
membrane stretch or degradation, local inflammation, or a weakening of
maternal resistance to ascending bacterial colonization. In many cases, the
ultimate cause of premature membrane rupture is unknown.
Prediction of Preterm PROM
Because the clinical course of preterm PROM is often one of brief latency and
increased infectious risk, it would be preferable to prevent it from occurring. In
a large prospective study,7 the National Institute of Child Health & Human
Development Maternal-Fetal Medicine Units Network found prior preterm birth
and preterm birth due to preterm PROM to be associated with subsequent
preterm birth due to preterm PROM. Those with a prior preterm birth at 2327
weeks had a 27.1% risk of preterm birth subsequently (P .001). Those with a
history of preterm PROM had a 13.5% risk of preterm birth due to preterm
PROM in a subsequent gestation (versus 4.1%, relative risk 3.3, P .01) and a
13.5-fold higher risk of preterm PROM at less than 28 weeks in the subsequent

gestation (1.8% versus 0.13%, P .01). In a separate publication from this study
the investigators found medical complications, work in pregnancy,
symptomatic contractions, bacterial vaginosis, and low body mass index to be
associated with preterm birth due to preterm PROM in nulliparas. Preterm
PROM in multiparas was associated with prior preterm PROM, prior preterm
birth due to preterm labor, and low body mass index. The presence of a short
cervix (less than 25 mm by transvaginal ultrasound) was associated with
preterm PROM in both nulliparas and multiparas. A positive fetal fibronectin
screen was also associated with preterm PROM in multiparas. Nulliparas with a
positive cervicovaginal fetal fibronectin and a short cervix had a 16.7% risk of
preterm birth due to preterm PROM, whereas multiparas with a prior history, a
short cervix, and a positive fetal fibronectin had a 25% risk of preterm PROM.
Multiparas with all three risk factors had a 31-fold increased risk of PROM with
delivery before 35 weeks relative to those without risk factors (25% versus
0.8%, P .001).8 Unfortunately, despite our developing ability to identify women
at increased risk for preterm PROM, such testing is expensive and inconvenient
to the patient, and will identify only a small fraction of those ultimately
delivering preterm. Because of this, our clinical efforts remain focused on
treatment of preterm PROM once it has occurred, rather than its prevention.
DIAGNOSTIC APPROACH
The approach to the diagnosis of membrane rupture is clinical, with over 90%
of cases being confirmed based on the presence of a suspicious history or
ultrasonographic finding followed by documentation of fluid passing from the
cervix or the presence of a Nitrazine (Bristol-Myers Squibb, Princeton,
NJ)/ferning positive vaginal pool of fluid. Other plausible causes of vaginal
discharge (eg, urinary incontinence, vaginitis, cervicitis, mucous show with
cervical effacement and dilatation, semen, vaginal douches) should be
excluded if the diagnosis is unclear. The Nitrazine test can be falsely positive
if the vaginal pH is increased by blood or semen contamination or alkaline
antiseptics, or if bacterial vaginosis is present. The ferning test should be
performed on a sample collected from the posterior fornix or lateral vaginal
sidewall to avoid cervical mucus, which may also yield a false positive result.
Prolonged leakage with minimal residual fluid can lead to a false negative
Nitrazine or ferning test. Should initial testing be negative but a clinical
suspicion of membrane rupture remain, the patient can be retested after
prolonged recumbency or alternate measures can be considered. Ultrasound
evaluation may prove useful if the diagnosis remains in doubt after speculum
examination. The diagnosis of membrane rupture can be confirmed
unequivocally with ultrasound-guided amnioinfusion of indigo carmine (1 mL in
9 mL of sterile normal saline), followed by observation for passage of blue fluid
per vaginum. Although oligohydramnios without evident fetal urinary tract
malformations or fetal growth restriction may be suggestive of membrane
rupture, ultrasound alone cannot diagnose or exclude membrane rupture with
certainty.

Cervicovaginal screening for fetal fibronectin has been suggested as a marker

for preterm PROM when the diagnosis remains in doubt after initial speculum
examination. However, the impact of prolonged membrane rupture on the
fibronectin result has not been elucidated. Further, a positive test may reflect
disruption of the decidua rather than membrane rupture. As such, fetal
fibronectin testing for the diagnosis of preterm PROM is not recommended for
routine practice. Until such time as the diagnosis of membrane rupture is
excluded, it is prudent to avoid digital cervical examinations, which have been
shown to decrease latency and increase infectious morbidity9,10 without
adding substantial information to that obtained by careful visualization during a
sterile speculum examination. The speculum examination also offers the
opportunity to perform appropriate cultures (eg, endocervical chlamydia and
gonorrhea), if these have not been recently performed or clinical suspicion of
new infection is present. Anovaginal cultures for group B streptococcus should
be obtained if these have not been performed within 5 weeks.
THERAPEUTIC APPROACH
General Considerations
Because the risk of perinatal complications changes dramatically with
gestational age at membrane rupture and delivery, a gestational agebased
approach to the management of preterm PROM is appropriate. Although there
is no apparent neonatal benefit to conservative management after membrane
rupture at term, there is a potential for neonatal benefit when conservative
management of preterm PROM is undertaken for the immature fetus. This
benefit may only be accrued if conservative measures lead to pregnancy
prolongation resulting in a reduction of gestational agedependent morbidity,
or through prevention of perinatal infection. Management should be based on
individual assessment of the estimated risk for maternal, fetal, and neonatal
complications should conservative management or expeditious delivery be
pursued (Figure 4).
Figure 4. An algorithm for evaluation and management of preterm premature
rupture of the membranes (pPROM). PgE2 prostaglandin E2; NIH National
Institutes of Health Maternal-Fetal Medicine Units Network.
Although practice varies and there is considerable controversy regarding the
optimal management of preterm PROM, there is general consensus regarding
some issues. First, gestational age should be established based on clinical
history and prior ultrasound assessment where available. Ultrasound should be
performed if feasible to estimate gestational age (if no prior ultrasound has
been performed) as well as fetal growth, position, and residual amniotic fluid
volume, and to evaluate for gross fetal abnormalities that may cause
polyhydramnios. The woman with preterm PROM should be evaluated clinically
for evidence of advanced labor, chorioamnionitis, abruptio placentae, and fetal
distress. Those with advanced labor, intrauterine infection, significant vaginal
bleeding, or nonreassuring fetal testing are best delivered regardless of
gestational age. If fetal malpresentation coexists with significant cervical

dilation, there is an increased risk of umbilical cord prolapse, which may

increase the risk of fetal loss. This circumstance may justify early delivery given
the increased fetal risk. Women with genital herpes simplex virus (HSV)
infection or human immunodeficiency virus (HIV) infection should generally not
be managed conservatively. In certain cases (eg, recurrent HSV or HIV infection
near or before the limit of viability), conservative management of preterm
PROM may be justified because of the extremely high risk of mortality and
chronic morbidity when immediate delivery is undertaken. These cases should
be managed individually. Although there are published case series supporting
conservative management in the setting of recurrent HSV infection, there are
no data in this regard for women infected with HIV. Intrapartum prophylaxis
against group B streptococcus is recommended for gravidas delivering preterm,
unless recent anovaginal group B streptococcus cultures are negative.11 If
conservative management of preterm PROM is to be pursued, the patient
should be admitted to a facility capable of providing emergent delivery for
abruptio placentae, fetal malpresentation in labor, and/or fetal distress due to
umbilical cord compression or in utero infection. The facility should also be
capable of providing 24-hour neonatal resuscitation and intensive care, as
conservative management should generally be performed only in pregnancies
where there is a signifi- cant risk of neonatal morbidity and mortality should
delivery occur. If adequate facilities for maternal and neonatal care do not
exist, patient transfer should be undertaken early in the course of management
to avoid emergent transfer once complications arise.
Preterm PROM Near Term (3236 Weeks)
When preterm PROM occurs at 3436 weeks gestation, the risk of severe acute
morbidity and mortality occurring is low when expeditious delivery is pursued.
Corticosteroids are generally not given to accelerate fetal pulmonary maturity.
Conversely, conservative management at 3436 weeks is associated with an
increased risk of amnionitis (16% versus 2%, P .001), prolonged maternal
hospitalization (5.2 versus 2.6 days, P .006), and a lower mean umbilical cord
pH at delivery (7.25 versus 7.35, P .009) without the benefit of a significant
reduction of perinatal complications related to prematurity.12 A mature
phosphatidyl glycerol, AmniostatFLM (Irvine Scientific, Santa Ana, CA),
fluorescence polarization TDX FLMII assay (surfactant/albumin ratio) (Abbott
Laboratories, Abbott Park, IL), lecithinsphingomyelin ratio, or a lamellar body
count result from fluid collected from either vaginal pool fluid or amniocentesis
is associated with a low risk of significant pulmonary complications after
preterm PROM near term, regardless of the presence of blood or meconium
contamination. The presence of blood or meconium in the setting of preterm
PROM should raise the level of suspicion for abruptio placentae or fetal
compromise and lead to reconsideration of the benefits of conservative
management. Should pulmonary maturity be evident based on vaginal pool
amniotic fluid or from amniocentesis at 3236 weeks gestation, it is evident
that the risk of major nonpulmonary perinatal complications are low and that
conservative management will prolong pregnancy only briefly (36 versus 14
hours, P .001), increase the risk of amnionitis (27.7% versus 10.9%, P .06), and

place the fetus at risk for occult cord compression while unmonitored, without
offering a significant reduction in neonatal morbidity.13 Thus, the woman who
suffers preterm PROM at 3436 weeks is generally best served by expeditious
delivery. When preterm PROM occurs at 32 to 33 weeks, it is recommended
that fetal pulmonary maturity be assessed by vaginal pool specimen if
available, and that amniocentesis by a skilled clinician be considered should
there be inadequate vaginal fluid for pulmonary maturity evaluation. When
fetal pulmonary maturity is evident after preterm PROM at 32 to 33 weeks,
there is little to be gained by continued pregnancy, and expeditious delivery is
recommended. There are no studies upon which to guide practice for women in
which pulmonary maturity testing is not performed or unavailable after preterm
PROM at 32 to 33 weeks gestation. At this gestation, the likelihood of survival
with delivery is high, but there remains signifi- cant risk of pulmonary
immaturity and other gestational agedependent morbidity should the fetus be
immature. Cox et al14 evaluated the practice of immediate delivery versus
conservative management in 129 unselected women with preterm PROM at 30
33 67 weeks gestation. Tocolytics and antenatal steroids were not
administered, and group B streptococcus prophylaxis was not given.
Conservative management of preterm PROM was associated with only a brief
increase in latency to delivery (59% versus 100% delivered within 48 hours, P .
001), but a significant increase in amnionitis (15% versus 2%, P .009), and no
evident reduction in gestational agedependent morbidity. There was a
significant risk of RDS in this population (35%). In addition, there was one
stillbirth due to suspected occult umbilical cord compression in the
conservatively managed group, and three neonatal deaths in the immediately
delivered group (two from sepsis and one from pulmonary hypoplasia).
Although this study does suggest that immediate delivery might reduce fetal
exposure to intrauterine infection and avoid fetal loss due to cord compression,
it confirms that the infant delivered at 3033 weeks remains at risk for neonatal
sepsis and other significant gestational agedependent morbidity in the
absence of documented pulmonary maturity. As such, should fetal pulmonary
immaturity be suspected at 32 to 33 weeks, or should fluid for testing be
unavailable, an option would be to treat conservatively with close fetal
monitoring, adjunctive antibiotic therapy, and antenatal corticosteroid
administration for fetal maturation (see below). Alternatively, if there is no plan
to either induce fetal maturation with corticosteroids or prolong pregnancy and
suppress infection with concurrent antibiotics, these patients may be better
served by expeditious delivery.
Preterm PROM Remote From Term (2331 Weeks)
Delivery before 32 weeks gestation is associated with a significant risk of
neonatal complications, including severe acute morbidity and death. Because
of this, the stable gravida with preterm PROM remote from term is generally
best served by conservative management in an attempt to prolong pregnancy
and reduce the risk of gestational agedependent morbidity in the newborn.
Despite conservative efforts, many women will ultimately deliver after a brief
latency. However, a subset of these women will remain pregnant for an

extended period of time, allowing the fetus to mature in utero. A low initial
amniotic fluid index has been shown to be associated with shorter latency and
an increased risk of amnionitis.15,16 However, amniotic fluid volume
assessment does not accurately predict who will ultimately deliver quickly
despite conservative measures, and thus should not be used in isolation to
make decisions regarding conservative management.
During conservative management, care should be directed toward the early
detection of labor, abruptio placentae, amnionitis, and umbilical cord
compression or occult fetal distress. This is generally best accomplished by
inpatient observation unless the membranes reseal. Conservative management
generally consists of initial prolonged continuous fetal and maternal monitoring
combined with subsequent modified bed rest to increase the opportunity for
amniotic fluid reaccumulation and spontaneous membrane sealing. The fetus
with preterm PROM remote from term is at significant risk (3276%) for
umbilical cord compression and fetal distress.17,18 Because abnormal fetal
heart rate patterns and contractions can occur in otherwise asymptomatic
women, fetal assessment should be performed at least daily on those with
initially reassuring testing. Although both the nonstress test and the biophysical
profile have the ability to confirm fetal well-being or identify fetal compromise
in the setting of preterm PROM, fetal heart rate monitoring offers the
opportunity to identify periodic heart rate changes such as variable and late
decelerations, in addition to allowing concurrent assessment of uterine activity.
Biophysical profile testing may be confounded by the presence of
oligohydramnios but can be helpful should the nonstress test be equivocal,
particularly remote from term, when the fetal heart rate pattern is less likely to
be reactive. Those with intermittent umbilical cord compression but
otherwise reassuring testing should undergo continuous fetal heart rate
monitoring, pending reassessment upon successful attainment of antenatal
corticosteroid benefit (2448 hours after initial dose). Complete pelvic rest and
avoidance of digital pelvic examinations should be undertaken to reduce the
risk of intrauterine infection and enhance latency. The combination of fever
(temperature exceeding 38.0C or 100.4F) with uterine tenderness and/or
maternal or fetal tachycardia, in the absence of another source of infection, are
suggestive of intrauterine infection and should lead to expeditious delivery.
Maternal white blood cell counts may be artificially elevated if antenatal
corticosteroids have been administered within 57 days. Alternatively, a rising
white blood cell count in the presence of suspicious clinical findings may be
useful in identifying those who require closer observation, including continuous
fetal monitoring or delivery. Maternal fever is not always present early in the
course of amnionitis. Should maternal symptoms, physical findings, and/or
hematological findings be equivocal, amniocentesis may provide further
suggestive information regarding the presence of intrauterine infection,
including a glucose concentration of less than 1620 mg/dL, a positive Gram
stain, or a positive amniotic fluid culture. Although positive amniotic fluid
cultures and elevated interleukin levels have been clearly associated with an
increased risk of early delivery and perinatal infectious morbidity, these test

results are not rapidly available in most clinical laboratories, limiting their utility
in clinical practice. If amnionitis is suspected but the diagnosis is equivocal,
amniocentesis should be considered with Gram stain, differential count, and
glucose level used, for acute management. Should a positive amniotic fluid
culture become evident later, this information should be considered by the
obstetrician if the patient remains pregnant, and provided to the neonatologist
if the patient has delivered subsequent to the amniocentesis.
Optimal treatment of the woman reaching an advanced gestational age after
conservative management of preterm PROM remote from term has not been
studied. In the absence of labor, abruption, nonreassuring fetal testing, or
intrauterine infection, it is reasonable to consider labor induction when the
pregnancy reaches 34 weeks gestation. It may also be reasonable to assess
fetal pulmonary maturity at any time from 32 to 34 weeks and consider
delivery once maturity has been documented.
Adjunctive Treatment of PROM Remote From Term
Antibiotics. The benefits of narrow-spectrum intrapartum prophylaxis with
intravenous penicillin or ampicillin to prevent vertical transmission and earlyonset neonatal sepsis by group B streptococcus (Streptococcus agalactiae)
have been well established. Intrapartum group B streptococcus prophylaxis
with intravenous penicillin as a 5,000,000-U initial bolus followed by 2,500,000
U every 4 hours or ampicillin (2 g) followed by 1 g intravenously (IV) every 4
hours (erythromycin, 500 mg IV every 6 hours, or clindamycin, 900 mg IV every
8 hours, if penicillin allergic) should be administered during labor or before
cesarean delivery after preterm PROM unless there is inadequate time or there
is a negative anovaginal culture obtained within 5 weeks of delivery. Because
of
increasing resistance of group B streptococcus to erythromycin and
clindamycin, some investigators have suggested careful confirmation of
penicillin allergy before alternate treatment, and intravenous cefazolin for
those with documented penicillin allergy.19,20 Should the latter approach be
taken, the caregiver should closely observe the patient for allergy to
cephalosporins. It may be prudent to restrict such alternative treatment to
women with minor allergic reactions to penicillin therapy. The reader is
encouraged to watch for evolving recommendations regarding intrapartum
group B streptococcus prophylaxis.
The goal of adjunctive antibiotic therapy during conservative management of
preterm PROM remote from term is to treat or prevent ascending decidual
infection in order to prolong pregnancy and offer the opportunity for reduced
neonatal infectious and gestational agedependent morbidity. Over two dozen
randomized, prospective, controlled trials have been undertaken to determine
if there is value in the adjunctive administration of antibiotics during
conservative management of preterm PROM remote from term. These studies
are addressed in detail elsewhere (Svare J. Preterm delivery and subclinical urogenital infection [thesis]. Copenhagen, Denmark: University of Copenhagen,
Department of Obstetrics and Gynaecology Rigshospitalet, 1997).2124 The

broad range of antibiotics administered, differing routes and duration of

administration, inclusion of pregnancies at low risk for neonatal morbidity, and
the variability of concurrent tocolytic and corticosteroid administration make
interpretation of the findings difficult. However, two large multicenter clinical
trials with different approaches to this issue have adequate power to evaluate
the utility of adjunctive antibiotics after preterm PROM in reducing infant
morbidity. The combination of these studies, as well as specific details from
individual smaller trials, offer valuable insights regarding the potential role of
adjunctive antibiotics in this setting.
The National Institute of Child Health & Human Development Maternal-Fetal
Medicine Units Network elected to study women with preterm PROM from the
limit of 2432 07 weeks gestation.25 The primary goal was to see if adjunctive
antibiotic therapy during conservative management of preterm PROM would
reduce the number of infants who were acutely ill after birth; a secondary goal
was to see if such treatment reduced individual infectious or gestational age
dependent morbidities. The National Institute of Child Health & Human
Development Maternal-Fetal Medicine Units Network elected to use initial
aggressive IV therapy (48 hours) with ampicillin (2 g IV every 6 hours) and
erythromycin (250 mg IV every 6 hours), followed by limited duration oral
therapy (5 days) with amoxicillin (250 mg orally every 8 hours) and enteric
coated erythromycin base (333 mg orally every 8 hours). These agents were
chosen because of their broad-spectrum antimicrobial coverage and their
demonstrated safety when used in pregnancy. The duration of therapy was
limited because of concerns that prolonged therapy might lead to selective
survival of resistant bacteria that might be more difficult to treat should
neonatal infection occur. Once group B streptococcus carriage was identified
based on initial cultures, carriers were treated with ampicillin for 1 week and
then again in labor, regardless of whether they were assigned to study
antibiotics or placebo therapy. Similar to other studies, this trial found that
antibiotic treatment prolonged pregnancy, increasing two-fold the likelihood
that patients would not have delivered after 7 days of treatment. Despite
discontinuation of antibiotics at 7 days, treated women continued to be more
likely to remain pregnant at up to 3 weeks after randomization, suggesting that
the therapy actually successfully treated subclinical infection rather than just
suppressing it. Regarding infant morbidity, the National Institutes of Health
(NIH) trial found that antibiotics improved neonatal health by reducing the
number of infants with one or more major infant morbidity (composite
morbidity: death, RDS, early sepsis, severe intraventricular hemorrhage, severe
necrotizing enterocolitis) from 53% to 44% (P .05). Additionally, aggressive
antibiotic therapy was associated with significant reductions in individual
gestational agedependent morbidity including RDS (40.5% versus 48.7%),
stages 3 to 4 necrotizing enterocolitis (2.3% versus 5.8%), patent ductus
arteriosus (11.7% versus 20.2%), and chronic lung disease (bronchopulmonary
dysplasia: 13.0% versus 20.5%) (P
.05 for each). Regarding infectious
morbidity, the antibiotic study group had a lower incidence of neonatal group B
streptococcus sepsis (0% versus 1.5%, P .03). Amnionitis was also reduced with

study antibiotics (23% versus 32.5%, P .01). Neonatal sepsis (8.4% versus
15.6%, P .009) and pneumonia (2.9% versus 7.0%, P .04) were less frequent in
those who were not group B streptococcus carriers. In support of this studys
findings, Mantel-Haenszel 2 analysis of prospective randomized clinical trials
with initial broad-spectrum parenteral therapy for women with preterm PROM
at less than 34 weeks (Svare J. Preterm delivery and subclinical urogenital
infection [thesis]. Copenhagen, Denmark: University of Copenhagen,
Department of Obstetrics and Gynaecology Rigshospitalet, 1997)23-32 shows
that antibiotics increase the likelihood of women remaining pregnant for longer
than 1 week (odds ratio [OR] and 95% confidence interval [CI]: 2.52; 1.92,
3.30), reduce amnionitis (0.60; 0.47, 0.78), reduce neonatal sepsis (0.47; 0.33,
0.67), reduce RDS (0.76; 0.60, 0.96), and reduce intraventricular hemorrhage
(0.70; 0.52, 0.95), without an increase in the risk of cesarean delivery (1.00;
0.78, 1.28) or necrotizing enterocolitis (1.28; 0.84, 1.98). Because of these
findings, limited duration aggressive antibiotic therapy is recommended during
conservative management of preterm PROM remote from term. Kenyon et al33
performed a large multiarm, multicenter placebo-controlled trial of oral
antibiotic therapy of women with preterm PROM at less than 37 weeks. Women
were randomized to receive oral erythromycin, amoxicillinclavulonic acid,
both, or a placebo for up to 10 days. Over 400 presented and nonpresented
statistical analyses were performed. In summary, oral erythromycin was
associated with brief pregnancy prolongation (not significant at 7 days),
reduced need for supplemental oxygen (31.1% versus 35.6%, P .02), reduced
positive blood cultures (5.7% versus 8.2%, P .02), but no significant reduction
in the composite primary outcome (an infant suffering one or more of the
following: death, chronic lung disease, major cerebral abnormality on
ultrasonography) (12.7% versus 15.2%, P .08). Subanalysis of singleton
gestations revealed a reduction in oxygen dependence at 28 days (6.9% versus
8.9%, P .03), positive blood cultures (5.3% versus 7.4%, P .04), abnormal
cerebral ultrasonography (3.0% versus 4.6%, P .04), and composite morbidity
(11.2% versus 14.4%, P .02) with erythromycin. Oral amoxicillinclavulonic acid
prolonged pregnancy (43.3% versus 36.7% undelivered after 7 days, P .005)
and reduced the need for supplemental oxygen (30.1% versus 35.6%, P .05),
but was associated with an increased risk of necrotizing enterocolitis (1.9%
versus 0.5%, P .001) without reducing other neonatal morbidity or composite
morbidity.
The combination of oral amoxicillinclavulonic acid and erythromycin yielded
similar findings. Subanalysis of women with preterm PROM remote from term
was not presented but was reported to reveal the same pattern of findings.
There was no analysis of the benefits of oral antibiotics during conservative
management at 3236 weeks. The authors concluded that oral erythromycin,
when given to women with preterm PROM, has effects on the occurrence of
major neonatal disease, and might therefore have a substantial health benefit
on the long-term respiratory and neurological function of many children. They
raised concern regarding the potential negative effects of oral amoxicillin
clavulonic acid because of the increased incidence of necrotizing enterocolitis.

An alternate conclusion would be that oral antibiotic therapy with erythromycin

reduces perinatal morbidity when given to women with preterm PROM before
37 weeks gestation, but given the small differences in the actual incidences of
morbidity in the study groups, many women would need to be treated to
prevent one adverse outcome. Although the finding regarding increased
necrotizing enterocolitis with oral amoxicillinclavulonic acid is concerning, it is
at odds with the National Institute of Child Health & Human Development
Maternal-Fetal Medicine Units Network trial finding of reduced stages 2 to 3
necrotizing enterocolitis with aggressive antibiotic therapy in a higher risk
population. Further, there is no consistent trend towards a positive or negative
effect of antibiotics for necrotizing enterocolitis in the literature. In summary,
there appears to be a role for adjunctive aggressive antibiotic therapy with
erythromycin and amoxicillin or ampicillin during conservative management of
preterm PROM remote from term. If aggressive therapy is not given, oral
erythromycin may offer some benefit. The combination of erythromycin and
extended-spectrum ampicillinclavulonic acid in a lower risk population near
term does not appear to be beneficial and may be harmful. This latter regimen
is not recommended. Intrapartum group B streptococcus prophylaxis should be
given to carriers, regardless of prior therapy.
Corticosteroids. It is well established that administration of antenatal
corticosteroids to women at high risk of delivering an immature preterm infant
is one of the most effective obstetric interventions directed to reduce perinatal
morbidity and mortality. Antenatal corticosteroids reduce the risk of RDS,
intraventricular hemorrhage, and perinatal death and have been shown to have
longterm neurological benefits when given in a timely fashion before preterm
birth. The current NIH consensus conference recommendations regarding
corticosteroid administration in the setting of preterm PROM are that a single
course of betamethasone (12 mg intramuscularly [IM], two doses every 24
hours) or dexamethasone (6 mg IM, four doses every 12 hours) be given during
conservative management of preterm PROM before 3032 weeks gestation
because of the potential for reduction of intraventricular hemorrhage.34 There
continues to be discussion regarding the efficacy of corticosteroids in the
setting of preterm PROM in the prevention of RDS. It has been suggested that
antenatal corticosteroids may be inappropriate in this setting because 1)
women with preterm PROM will deliver too quickly to accrue the potential
benefits; 2) preterm PROM itself might induce fetal pulmonary maturation,
thereby making corticosteroids unnecessary; and 3) antenatal corticosteroids
might delay the diagnosis or increase the risk of perinatal infection through
their immunosuppressive effects. It is now clear that the majority of
conservatively managed women with preterm PROM remote from term will
remain pregnant for at least 2448 hours, particularly if adjunctive antibiotics
are administered. Although there have been conflicting reports regarding the
relative risk of RDS with preterm birth after preterm PROM, it is clear that RDS
remains the most common acute morbidity after preterm birth due to preterm
PROM. In the National Institute of Child Health & Human Development
Maternal-Fetal Medicine Units Network trial, the incidence of RDS after

conservatively managed preterm PROM was 41% when corticosteroids were not
administered, despite adjunctive antibiotic administration. Regarding perinatal
infection, studies performed before the era of adjunctive antibiotic
administration revealed con- flicting results as to the risk of neonatal infection
when antenatal corticosteroids were administered. There are two recent
randomized clinical trials that have assessed the utility of antenatal
corticosteroid administration concurrent to adjunctive antibiotic administration.
In the first, Lewis et al35 found a reduced incidence of RDS (18.4% versus
43.6%, P .03) with no obvious increase in perinatal infection (3% versus 5%, P
NS) when antenatal corticosteroids were administered after initiation of
ampicillin-sulbactam for preterm PROM at 2434 weeks. In the Dexiprom
multicenter trial,36 women with preterm PROM were evaluated after being
allocated to either a single course of dexamethasone (12 mg IM, two doses
every 24 hours) or placebo at 2834 weeks gestation. All patients received
amoxicillin and metronidazole for 5 days. Although there was no significant
reduction in morbidity in the overall population, no increase in maternal or
neonatal morbidity was evident in the corticosteroid group. Among those
remaining pregnant at least 24 hours after study entry, newborns exposed to
antenatal steroids had a lower incidence of perinatal death (1.3% versus 8.3%,
P .05). The most recent meta-analysis regarding this issue37 includes data
from the original trial of corticosteroid administration published by Liggins and
Howie in 1972. These authors found corticosteroid administration in the setting
of preterm PROM to substantially reduce the risks of RDS (20% versus 35.4%;
OR 0.56; CI 0.46, 0.70), intraventricular hemorrhage (7.5% versus 15.9%; OR
0.47; CI 0.31, 0.70), and necrotizing enterocolitis (0.8% versus 4.6%; OR 0.21;
CI 0.05, 0.82), without significantly increasing the risks of maternal infection
(9.2% versus 5.1%; OR 1.95; CI 0.83, 4.59) or neonatal infection (7.0% versus
6.6%; OR 1.05; CI 0.66, 1.68). The authors suggested that there was no need
for additional study of this specific issue. Thus, based on current information, it
is recommended that a single course of antenatal corticosteroids be
administered concurrently to initial adjunctive antibiotic therapy during
conservative management of preterm PROM before 32 weeks. Similar
treatment should also be considered if conservative management is pursued
for suspected fetal pulmonary immaturity at 3234 weeks gestation. The risks
and benefits of repeated corticosteroid administration after an initial course of
antenatal corticosteroids near the limit of viability remain to be determined.
Currently routine repeated administration of antenatal corticosteroid is not
recommended. Tocolysis. There are inadequate data upon which to make firm
recommendations regarding tocolytic therapy in the setting of preterm PROM.
Two small studies of prophylactic intravenous or oral betamimetic therapy and
a study of prophylactic betamimetic or magnesium sulfate tocolysis after
preterm PROM3840 have suggested brief pregnancy prolongation in this
setting. Alternatively, a practice of expectant management with therapeutic
tocolysis initiated only after the onset of contractions did not show that
tocolysis prolonged latency.41 Similarly, improved neonatal outcome has not
been conferred with a practice of tocolysis and serial assessment of fetal
pulmonary maturity followed by delivery. Although no study has demonstrated

tocolytics to improve or worsen neonatal outcome, studies of conservative

management have not evaluated tocolysis when corticosteroids and antibiotics
are given concurrently. It is plausible that short-term pregnancy prolongation
with prophylactic tocolysis could enhance the potential for corticosteroid effect
and allow more time for antibiotics to act against subclinical decidual infection.
Given evidence of short-term pregnancy prolongation without evident risk in
some studies, it is not unreasonable to initiate tocolysis in women at high risk
for infant morbidity should there be concurrent attempts to prevent infection,
prolong pregnancy, and induce fetal pulmonary maturity. However, this
approach should not be considered an expected practice as further research is
needed.
Previable Preterm PROM (Less Than 23 Weeks)
Approximately half of women with second-trimester PROM will deliver within 1
week of membrane rupture. Alternatively, up to 22% will reamin pregnant for at
least 1 month. In addition to the high risk of chorioamnionitis (39%), maternal
morbidities associated with conservative management of second-trimester
PROM include endometritis (14%), abruptio placentae (3%), retained placenta,
and postpartum hemorrhage necessitating dilation and curettage (12%).
Maternal sepsis is a rare but serious complication, complicating 0.8% of cases.
One maternal death from sepsis has been reported in 731 pregnancies
complicated by second-trimester PROM (0.14%).43 The incidence of stillbirth
subsequent to second-trimester PROM (15%) is higher than that seen with
preterm PROM later in pregnancy (1%). This may reflect increased fetal
susceptibility to umbilical cord compression, hypoxia, and intrauterine
infection, but may also reflect a practice of nonintervention for fetal distress in
the pre- or periviable fetus. Previable delivery is immediately lethal. Periviable
delivery after preterm PROM may be associated with more frequent morbidity
and mortality than that seen with delivery due to other causes because preand periviable PROM are associated with perinatal infection and pulmonary
hypoplasia. Women with early previable PROM before 20 weeks, and those with
persistent oligohydramnios, are at particular risk for pulmonary hypoplasia.
Lethal pulmonary hypoplasia rarely occurs with membrane rupture subsequent
to 24 weeks gestation, presumably because alveolar growth adequate to
support postnatal development has occurred. Ultrasound estimates of interval
fetal lung growth including lung length, chest circumference, chest/abdominal
circumference ratio, or chest circumference/femur length ratio carry a high
positive predictive value for lethal pulmonary hypoplasia.4447 The pattern of
fetal restriction deformities occurring after prolonged intrauterine crowding is
similar to that seen with Potter syndrome. The nonpulmonary features include
abnormal facies and limb positioning abnormalities. The ears may be low set
and epicanthal folds present. The extremities may be flattened and
malpositioned.
It is estimated that 5684% of survivors will be neurologically intact after
midtrimester PROM.43,4852 However, developmental delay (24%), delayed
motor development (23%), and other less frequent complications including

cerebral palsy, chronic lung disease, hydrocephalus, and mental retardation are
serious reported sequelae. The frequency of these outcomes may be biased
because of the lack of follow-up in over 40% of cases. There are no available
recent data regarding the outcomes of survivors delivering after conservatively
managed previable preterm PROM. Women presenting with PROM before
viability should be counseled regarding the impact of immediate delivery and
the potential risks and benefits of conservative management. Counseling
should include a realistic appraisal of neonatal outcomes, including the
availability of obstetric monitoring and neonatal intensive care facilities.
Because of the significant risk of adverse maternal and neonatal sequelae after
previable PROM, some women will desire expeditious delivery. Depending on
the experience of the caregiver and the patients wishes, this can be
accomplished by dilation and evacuation, or by labor induction with high-dose
oxytocin or vaginal prostaglandin therapy in the absence of contraindications.
There are currently no data upon which to make recommendations regarding
initial conservative management of the woman with previable PROM. The
benefits of an initial period of inpatient observation may include strict bed and
pelvic rest to enhance the opportunity for resealing, and for early identification
of infection and abruption. Ultrasound should be performed to exclude the
presence of associated fetal anomalies. Serial ultrasound should be offered
every 1 to 2 weeks initially to evaluate for reaccumulation of amniotic fluid and
to evaluate interval pulmonary growth. Some women will elect not to continue
the pregnancy if interval ultrasound suggests a high likelihood of lethal
pulmonary hypoplasia based on the presence of persistent severe
oligohydramnios and/or lagging pulmonary growth. A number of novel
treatments including amnioinfusion and fibrin/ platelet/cryoprecipitate or gel
foam sealing of the membranes have been preliminarily investigated.5355 In
some approaches cervical cerclage placement concurrent to attempted
membrane sealing is used. The maternal risks and fetal benefits of these
aggressive interventions have not been adequately demonstrated, so currently
they should not be incorporated into routine clinical practice. Many clinicians
will discharge the patient to bed rest at home after an initial observation
period, with readmission to hospital bed rest once the patient reaches a
gestation at which they would intervene for fetal benefit should labor,
amnionitis, nonreassuring testing, or abruption occur. There are no published
data evaluating the risks and benefits of this strategy versus continued
hospitalization in this population.
CERVICAL CERCLAGE
Cervical cerclage is a widely described risk factor for PROM and other adverse
pregnancy outcomes. Preterm PROM complicates about one in four pregnancies
with a cerclage, and about half of pregnancies after emergent cerclage.
Although women undergoing cerclage placement, whether elective or
emergent, are at high risk for preterm birth, a thorough discussion of the
potential adverse sequelae despite cerclage placement should be undertaken
before the procedure. There are no prospective studies regarding the treatment
of preterm PROM subsequent to cervical cerclage in situ. A number of

retrospective studies have suggested that when cerclage is removed on

admission, the risk of adverse perinatal outcomes is not higher than those seen
after preterm PROM without a cerclage.56,57 Alternatively, studies comparing
pregnancies in which the cerclage was retained or removed have been small
and have yielded conflicting results.5860 Each study found a statistically
insignificant trend towards increased maternal infection with retained cerclage.
One study found increased infant mortality and death from sepsis with retained
cerclage, with the only evident benefit being an decreased likelihood of
delivery within 24 hours.58 One study, which largely reflected differing
practices at two institutions, found significant pregnancy prolongation with
cerclage retention.59 Among those with preterm PROM before 28 weeks,
cerclage retention was associated with improved birth weight (942 versus 758
g, P .04). However, no controlled study has found a significant reduction in
infant morbidity with cerclage retention after preterm PROM. Given the
potential risk without evident neonatal benefit, the general approach to
management should include early cerclage removal after preterm PROM. The
role for short-term cerclage retention while attempting to enhance fetal
maturation with antenatal corticosteroids in the periviable gestation has not
been determined.
CEREBRAL PALSY/ADVERSE NEUROLOGICAL OUTCOME
It is established that preterm birth is a significant risk factor for long-term
sequelae such as chronic lung disease, neurosensory impairment, cerebral
palsy, and developmental delay. There are accumulating data linking perinatal
infection to neurological complications. Because preterm PROM is associated
with early delivery and perinatal infection, it is a potential risk factor for longterm neurological morbidity. Cerebral palsy and cystic periventricular
leukomalacia have been linked to the presence of amnionitis, which is
commonly seen after preterm PROM.61 Similarly, elevated amniotic fluid
cytokines and fetal systemic inflammation, which may accompany or reflect
maternal or fetal infection, have been associated with preterm PROM as well as
brain lesions such as periventricular leukomalacia and/or the subsequent
development of cerebral palsy.6264 Currently there are no data to suggest
that immediate delivery of the candidate for conservative management after
preterm PROM will prevent these sequelae. Therefore, conservative
management, with adjunctive antibiotics to reduce the risk of intrapartum
infection, is recommended. However, at more advanced gestational ages,
particularly if pulmonary maturity can be documented, expeditious delivery
may reduce the risk of exposure to intrauterine infection and subsequent
morbidity.
SUMMARY
Premature rupture of the membranes affects over 120,000 pregnancies
annually in the United States and is associated with significant maternal, fetal,
and neonatal risk. Management of PROM requires an accurate diagnosis and
evaluation of the risks and benefits of continued pregnancy or expeditious

delivery. An understanding of gestational agedependent neonatal morbidity

and mortality is important in determining the potential benefits of conservative
management of preterm PROM at any gestation. Where possible, the treatment
of pregnancies complicated by PROM remote from term should be directed
towards conserving the pregnancy and reducing perinatal morbidity due to
prematurity while monitoring closely for evidence of infection, abruptio
placentae, labor, or fetal compromise due to umbilical cord compression.
Alternatively, when preterm PROM occurs near term, the patient is generally
best served by expeditious delivery, particularly if fetal pulmonary maturity is
evident. It is important that the patient be well informed regarding the
potential for subsequent maternal, fetal, and neonatal complications regardless
of the management approach.