Arch Gynecol Obstet (2012) 285:15291534

DOI 10.1007/s00404-011-2179-0

MAT ERNAL-FE T A L M ED I C I N E

Previable preterm rupture of membranes: gestational

and neonatal outcomes
Marcela Furlan Margato Guilherme Lopes Pinheiro Martins
Renato Passini Jnior Marcelo Lus Nomura

Received: 26 July 2011 / Accepted: 12 December 2011 / Published online: 28 December 2011
Springer-Verlag 2011

Abstract
Introduction Previable (less than 24 weeks) premature
rupture of membranes complicates about 1 in every thousand
births and is responsible for substantial perinatal mortality.
Subjects and methods In this paper, we retrospectively
analyzed one twin and 35 singleton pregnancies.
Results Twenty cases occurred before and 16 after 20
weeks. Latency period ranged from 0 to 137 days, with an
average of 35 days. Amniotic Xuid index was reduced in 27
cases and normal in 6 cases. Expectant management was
adopted in 31 cases (86%), Wve patients declined and opted
for termination (14%) at admission or during the course of
pregnancy. Steroids were prescribed for 12 patients at or after
24 weeks (39%), leukocyte count at admission varied from
6,000 to 16,200/mm3, with an average of 11,310, in only 9%
it was greater than 15,000, immature forms were present in 10
cases (28%). Clinical chorioamnionitis occurred in 71%,
being three times more frequent in parous women. Bacteriuria
was present in 2 of 30 cases (6.6%). Two women developed
laboratorial and clinical signs of sepsis, none of them needed
hysterectomy. There were no maternal deaths. Mean gestational age at delivery was 24 weeks, ranging from 16 to 39
weeks. In the expectant group, preterm delivery rate was
68%. There was one case of abruption. Cesarean rate was
31%. Neonatal mortality was 42% (8 cases). Overall neonatal
survival was 35% (11 in 32 newborns).
Conclusion Perinatal mortality is high in pregnancies
complicated by previable rupture of membranes, however
gestational age at occurrence is a strong predictor of out-

M. F. Margato G. L. P. Martins R. Passini Jnior

M. L. Nomura (&)
Department of Obstetrics and Gynecology,
State University of Campinas, Campinas, Brazil
e-mail: mlnomura@unicamp.br

come. An individualized approach is the best management

option regarding maternal risks and fetal outcomes.
Keywords Preterm rupture of membranes
Preterm delivery Previable rupture of membranes
Chorioamnionitis

Introduction
Preterm rupture of membranes complicates about 4% of
pregnancies and previable (midtrimester), or less than
24 weeks, occurs in 3.7 of every 1000 births [1]. Neonatal
survival is poor in most case series, but with great variation
among reports, from 12 to 92% [2, 3]. This wide range of
reported survival rates depends on the gestational age at
which membranes ruptured, and is possibly inXuenced by
biased reports. Maternal morbidity is almost entirely due to
infection, and chorioamnionitis rates range from 5 to 77%
[4, 5]. Sepsis, abruption and maternal death have been
reported in pregnancies managed conservatively. Such
diVerent Wgures regarding neonatal survival and maternal
morbidity have led to uncertainties, and there is no consensus about optimal management.
In order to adequately counsel parents facing a rare
obstetrical situation, our objective was to evaluate neonatal
morbidity and mortality and maternal risks associated with
conservative management of previable preterm rupture of
membranes at our institution.

Methods
Patients were retrospectively identiWed at our electronic
database of diagnosis using ICD-9 and ICD-10 codes for

123

1530

rupture of membranes. We identiWed 36 patients in the

period ranging from January 1996 to September 2008.
Maternal and neonatal charts were reviewed for data collection. This study was previously approved by our Ethics
Board.
The diagnosis of PRM was established if one of the following criteria were met: typical maternal history with vaginal pH > 7, or ferning test positive, or visualization of
amniotic Xuid in vaginal speculum exam, or severe oligohydramnios at ultrasound scan.
Maternal demographics, clinical presentation, laboratory
exams (complete blood count and urine culture), and ultrasound assessment of amniotic Xuid index at admission were
collected. If no signs of maternal infection were detected at
admission, patients were counseled about choosing expectant or active management. All patients with amniotic
infection had their pregnancies terminated. During followup, serial blood counts, maternal hydration in the Wrst 48
72 h, temperature monitoring and ultrasound scans were
performed. Most patients have prolonged hospital stays at
the Wrst evaluation. Labor and delivery data were recorded.
Neonatal charts were reviewed for morbidity and mortality
data.
Our management protocol aimed to exclude infection at
presentation with an inpatient initial evaluation of all cases.
Chorioamnionitis was deWned by Gibbs criteria (fever, leukocytosis, uterine tenderness, maternal or fetal tachycardia,
malodorous amniotic Xuid). During hospital stay, management included maternal hydration, temperature monitoring,
fetal heart rate surveillance, serial blood counts and ultrasound scans after 4872 h to reassess amniotic Xuid index.
If chorioamnionitis was ruled out after this initial evaluation, patients were managed as an outpatient basis at our
high risk pregnancy unit. Steroids were prescribed only at

Fig. 1 Case outcome Xowchart

123

Arch Gynecol Obstet (2012) 285:15291534

or beyond 26 (in the early years) or 24 weeks. Antibiotics

were not routinely prescribed unless urinary or vaginal
infections were present at admission or during follow-up.
We do not administer therapeutic antibiotics for previable
rupture of membranes. Tocolysis was never prescribed for
any patient in any situation.
Pregnancies were allowed to proceed as long as possible
and medical interruption was indicated if labor or signs of
maternal infection ensured, or if fetal compromise was
detected after 26 weeks. Chorioamnionitis was treated with
gentamycin and metronidazole, and betalactam were added
if fever persisted after 48 h or maternal sepsis developed.

Results
During the study period (January 1996 to September 2008),
35,901 births occurred at our institution and 36 cases (35
singleton and 1 dichorionic twin) were identiWed, which
brings an incidence of 0.1% of previable rupture of membranes (PRM). See Fig. 1.
Five patients (13.9%) elected termination of pregnancy
(TOP) after counseling. These cases were excluded from
the analysis of outcomes, with remaining 31 mothers and
32 fetuses (one twin pregnancy). In this group, gestational
age at rupture ranged from 15 to 22 weeks, one patient had
sickle cell anemia, one had a large subchorionic hematoma
and the other had asthma. None of them had complications.
Mean maternal age was 26 years, ranging from 12 to
42 years, 19% had previous abortions, 25% were primigravid, 47% had at least one previous cesarean section and
none had prior preterm birth. Regarding ethnicity, 70% were
Caucasian, 25% were AfricanAmerican and 5% were black.

Arch Gynecol Obstet (2012) 285:15291534

Mean gestational age at rupture was 19 weeks, ranging

from 14 to 23 weeks, 20 cases (56%) occurred before
20 weeks. Mean latency was 35 days, ranging from 0 to
137.
Clinical history at admission was typical in 33 patients
(maternal report of loss of signiWcant amount of clear Xuid
from the vagina). One patient reported an uncertain complaint but ultrasound showed absence of amniotic Xuid,
another patient had physical examination very suggestive,
with vaginal pH > 7, ferning test positive, visualization of
amniotic Xuid in vaginal speculum exam, and severe oligohydramnios at ultrasound scan, the third patient had a
bleeding placenta previa and absence of amniotic Xuid at
ultrasound.
SigniWcant clinical comorbidities were present in Wve
patients: HIV disease, sickle cell anemia, asthma, portal
hypertension (secondary to schistosomiasis) and chronic
hypertension. Three patients had placenta previa. One
patient got pregnant with a copper IUD in situ that was
removed 3 weeks before PRM.
In only three patients ultrasound was not performed: two
patients elected TOP (both with conWrmatory history and
physical examination) and one delivered soon after admission. In the remaining 33 cases, there was a reduced amniotic Xuid index (AFI) at admission in 27 (81%). Of six
cases with initially normal AFI, four had oligohydramnios
in the follow-up. Interestingly, two patients had an initial
normal AFI and typical history and physical examination;
both developed chorioamnionitis and delivered at 25 weeks
gestation.
Anemia was present in six (17%) patients. Leukocyte
count at admission ranged from 6 to 16,000/mm3 (mean
11,310), with three patients with counts above 15,000/mm3.
Immature forms were observed in 10 patients. Twenty
patients had more than one count collected and 14 (70%)
had an increase in leukocyte count during pregnancy. Of
these 14 patients, 10 had chorioamnionitis diagnosed and 2
delivered spontaneously.
Bacteriuria was present in 2 of 30 collected urine samples at admission, Escherichia coli was isolated in these
two cases.
In 28 patients, maternal hydration was prescribed (23 l
of intravenous ringer solution in 24 h for 2 days). In nine
cases AFI increased, but in only four it reached the normal
percentile for correspondent gestational age. Steroids for
fetal maturation were prescribed at or beyond 24 weeks for
12 patients.
Chorioamnionitis was considered a probable diagnosis
in 71% of the patients based on Gibbs criteria [6]. Leukocytosis of unknown origin was the main Wnding, followed by
maternal tachycardia and fever (temperature >37.8C).
Maternal sepsis occurred in two patients, both admitted
at the ICU. None of the patients developed severe sepsis or

1531

needed hysterectomy. There was one case of abruption and

one case of puerperal infection (endometritis). The true
incidence of endometritis and puerperal infection might be
underestimated, since most infections are mild and treated
in institutions of lower levels of care.
The overall incidence of preterm delivery was 90%.
Placental histology was performed in 25 of 31 specimens
(80%) and in 9 (36%) there was evidence of chorioamnionitis [6] and in 1 funiculitis.
Table 1 shows maternal characteristics and outcomes
compared between two groups according to gestational age
of membranes rupture (group 1: 1419 weeks and group 2:
2024 weeks). Only gestational age at rupture was signiWcantly diVerent between groups.
Table 2 shows fetal and neonatal outcomes for 32 newborns also according to gestational age of membranes rupture (group 1: 1419 weeks and group 2: 2024 weeks).
Group 2 had a non-signiWcant lower birth weight mainly
because there were no deliveries between 32 and 38 weeks.

Table 1 Maternal and obstetrical characteristics by gestational age at

rupture of membranes
Maternal
characteristics

Group 1 1419
weeks (n = 17)

Maternal age (years)a

27 5,7

25.5 6.2

0.34

Paritya

1.1 1.1

1.5 1.1

0.21

Previous abortion (%)

11.8%

28.6%

Gestational age
PRM (weeks)

16.9 1.67

22.1 1.46

Latency (days)a

39 40

40 34

Maternal ICU
admission

Fever (%)

3 (17.6%)

2 (14.2%)

0.59

Oligohydramnios
at admission

8 (50%)

3 (21.4%)

0.10

Placenta previa

Group 2 2024
weeks (n = 14)

0.23
<0.01
0.60

Mean standard deviation

Table 2 Perinatal outcome by gestational age at rupture of membranes

Fetal and neonatal
outcome

Group 1 1419
weeks (n = 17)

Group 2 2024
weeks (n = 15)

GA at delivery
Birthweighta

22 6

27 5.5

0.02

1,653 1,079

1,390 876

Preterm birth < 37 weeks

0.92

94.1%

85.7%

GA at delivery > 24 weeks 41.2%

73.3%

Stillbirth

10

Neonatal death

Perinatal survival

42.8% (3/7)

72.7% (8/11)

0.20

Overall survival rate

18% (3/17)

53% (8/15)

0.02

0.14

Mean standard deviation

123

1532

Gestational age at delivery was signiWcantly higher in

group 2, and this characteristic was also reXected in a much
higher perinatal survival (42.8 vs. 72.7%).
Neonatal morbidity was: admission in ICU (47%), neonatal sepsis (21%), pulmonary hypoplasia (10%), respiratory failure (42%), and intraventricular hemorrhage (10%).

Discussion
Counseling in previable preterm rupture of membranes is
diYcult, mainly because of lack of data regarding maternal
and fetal safety. Our main objective was to provide local
data in order to adequate counsel mothers and fathers facing this situation. Our overall survival rate was a little
lower than the recent reports [7, 8]. However, when gestational age is analysed, ruptured membranes at or beyond
20 weeks have higher perinatal survival, despite having
similar median latency period. This may have occurred
because perinatal survival begins to rise sharply after
24 weeks. Increasing gestational age at delivery in 40 days
does not signiWcantly increases perinatal survival from 15
to 21 weeks, but dramatically increases from 21 to
27 weeks, for example. There are two large cohorts
reported [9, 10], one comprised 516 livebirths between 22
and 25 weeks with a perinatal survival rate of 35%, the
same rate observed in our sample [9]. The other report
included 553 newborns with PRM between 22 and
24 weeks, with survival rates of 6, 26 and 55% at 22, 23
and 24 weeks, respectively, but more than 90% survived
with major morbidities n the whole group [10]. Other
reports evaluated wider ranges of gestational ages and
therefore their results may not be comparable.
Maternal morbidity was signiWcant and sepsis developed
in two patients requiring intensive care support. Using
Gibbs criteria, chorioamnionitis was a common Wnding,
and fever developed only in Wve patients, but it was in this
group that sepsis was detected. A recent study found that
for each day a fetus stays in utero with premature rupture of
membranes, it is 1.01 times more likely to be exposed to
chorioamnionitis [11]. Although none of the septic patients
died, needed hysterectomy or suVered additional severe
morbidity, there was one report of maternal death [12] and
report bias regarding this outcome is possible. However,
maternal death is an outcome that is always reported by surveillance committees in most parts of the world.
There was one case of pregnancy with IUD, removed at
the 20th week. Although the outcome had been favorable,
retained IUD are also associated, although rarely, with fetal
and amniotic candida infections [13]. In a more recent
review of 30 cases, the rates of miscarriage, preterm rupture
of membranes, placental abruption, and preterm delivery
were signiWcantly higher in pregnant women with retained

123

Arch Gynecol Obstet (2012) 285:15291534

IUD compared with pregnancies after IUD removal in early

pregnancy [14]. So it is advisable to remove an IUD as
soon as possible if a woman becomes pregnant.
Placenta previa is a known risk factor for preterm delivery, mostly associated with heavy bleeding in third trimester. Only one of three patients had a good outcome (preterm
birth at 30 weeks with prolonged neonatal ICU stay).
Bleeding in placenta previa, particularly in early second trimester, with light or Xuid reddish discharge, must raise
PRM suspicion. Persistent uterine bleeding in early pregnancy is associated with PRM and perinatal mortality [15].
Of concern also is the occurrence of previable PRM in
women with complications, such as immune suppressive
diseases and severe clinical conditions. Our patient with
HIV infection was in good immune state (high CD4 count
and low viral load) with antiretroviral therapy, and she
opted for expectant management, allowing pregnancy to
proceed to term with excellent neonatal outcome. The
patient with sickle cell anemia had TOP performed, and the
patient with liver cirrhosis had a stillbirth. There is a paucity of data regarding management in such situations, but a
report bias must be possible, since TOP must be considered
when maternal preexisting conditions portray a very high,
unacceptable morbidity in face of a low probability of neonatal prolonged survival.
Steroids for fetal pulmonary maturation are prescribed
only after 24 weeks with signiWcant reductions in the incidence of neonatal respiratory distress syndrome, intraventricular hemorrhage and neonatal death. However, little is
known about their eVects at lower gestational ages.
Recently, a retrospective review in Japan showed that at
22 weeks there might be some beneWt, which is a 30%
reduction in mortality with antenatal corticosteroid use in
preterm infants born at 22 or 23 weeks gestation [16]. This
Wnding merits larger discussions, since there are disparate
data on neonatal survival at 22 and 23 weeks among several
centers, depending on local resources and level of care.
Overall Wgures are 2040% of survival beyond 28 days
[17].
Is isolated leukocytosis enough for clinical decision?
Maternal leukocytosis at admission has been associated
with a higher incidence of adverse neurodevelopmental
outcome at 2 years of age in infants born to women with
preterm rupture of membranes beyond 24 weeks [18]. Leukocytosis is one of the Gibbs criteria for chorioamnionitis,
but its sensitivity (2947%) [19] is relatively low for clinical decision as an isolated indicator for delivery. It is of
notice that nearly all studies were done in gestational ages
beyond 24 or 26 weeks, and there are no studies addressing
maternal laboratory parameters and chorioamnionitis in
previable rupture of membranes. In our study, 50% of
patients who presented increasing leukocyte counts during
follow-up had clinical chorioamnionitis. Therefore, an

Arch Gynecol Obstet (2012) 285:15291534

increase in serial leukocyte counts may be more speciWc

and useful, although still limited in its sensitivity.
Is conservative management justiWable in light of the
current Wgures of perinatal survival? Two major reviews
were published. In a compilation of data from 11 articles,
an overall perinatal survival of 21%, despite a live birth rate
of 67%, was found [20]. In a recent review of 275 cases, a
livebirth rate of 68.4% and neonatal survival rate of 44%
were found [1] but diVerent and better outcomes were
reported when PRM occurred at or beyond 22 weeks, similar to the outcomes of our study. In conclusion, PRM after
20 weeks has diVerent perinatal outcomes and information
regarding these outcomes must be incorporated in counseling.
One weakness of our study is that we were not able to
collect data about long-term prognosis, which is important
information in the decision-making process involving termination. However, despite several reports on the longterm outcomes of children born at extremely preterm gestational ages, there are still diYculties in comparing outcomes in diVerent populations. Since preterm birth was
almost the rule in our population, this issue should be
addressed with the parents. Prematurity in the setting of
rupture of membranes has important implications in neonatal morbidity, especially at earlier gestational ages when
amniotic and fetal infection is the leading known cause of
preterm birth. There is a twofold increase in the risk of
cerebral palsy when clinical chorioamnionitis is present
[21], and survival without major morbidity (<10%) is a rare
event at such earlier gestational ages [9].
Lung hypoplasia is a known consequence from prolonged oligohydramnios during the canalicular phase of
fetal lung development (1628 weeks), regardless of its
cause. Its incidence is quite variable, mainly because of
lack of uniform diagnostic criteria, and mortality ranges
from 55 to 100% [22]. In our sample, pulmonary hypoplasia occurred in four fetuses or newborns, conWrmed by
pathological exam, which brings an incidence of 12.5%,
considering stillbirths. All cases occurred in women with
PROM before 20 weeks. Prediction of neonatal lung hypoplasia, particularly lethal lung hypoplasia, is a diYcult task,
and despite several studies on clinical and ultrasonographic
markers, gestational age at PROM seems to be the best predictor, with sensitivity of 96% and speciWcity of 48% for
PROM before 25 weeks [22]. Persistent oligohydramnios
and latency period are described by other authors as independent predictors [23], and both informations must be
incorporated into counseling.
Previable rupture of membranes is one of the most controversial and confusing issues in obstetrics. Most pregnancies complicated by PRM are terminated, based on
assumptions of poor perinatal survival and high maternal
risk of serious infections. With marked improvements in

1533

neonatal care, lowering limits of fetal viability, a new

approach is being developed. Counseling might change
during the course of pregnancy, especially in the presence
of persistent oligohydramnios and a long latency period.
There are several unanswered questions, and the evidence
so far is derived from case series, which are prone to important biases, such as underreporting of poorer outcomes and
small sample sizes. Whether if a prospective, randomized
clinical trial is ethically justiWable and feasible is an open
question, but previable rupture of membranes is a situation
where the level of evidence might be far from the ideal, and
an individualized approach is at present the best evidence.
ConXict of interest

None.

References
1. Waters TP, Mercer BM (2009) The management of preterm premature rupture of the membranes near the limit of fetal viability.
Am J Obstet Gynecol 201(3):230240
2. Chen KC, Hsu JJ, Hsieh CC, Hsieh TT (1992) Pregnancy outcome
of preterm premature rupture of the membranes before 28 weeks.
Changgeng Yi Xue Za Zhi 15(4):176182
3. Verma U, Goharkhay N, Beydoun S (2006) Conservative management of preterm premature rupture of membranes between 18 and
23 weeks of gestationmaternal and neonatal outcome. Eur J Obstet Gynecol Reprod Biol 128(12):119124
4. Hibbard JU, Hibbard MC, Ismail M, Arendt E (1993) Pregnancy
outcome after expectant management of premature rupture of the
membranes in the second trimester. J Reprod Med 38(12):945951
5. Muris C, Girard B, Creveuil C, Durin L, Herlicoviez M, Dreyfus
M (2007) Management of premature rupture of membranes before
25 weeks. Eur J Obstet Gynecol Reprod Biol 131:163168
6. Gibbs RS, Castelo MS, Rodgers PG (1980) Management of acute
chorioamnionitis. Am J Obstet Gynecol 136:709713
7. Farooqi A, Holmgren PA, Engberg S, Serenius F (1998) Survival
and 2-year outcome with expectant management of second-trimester rupture of membranes. Obstet Gynecol 92(6):895901
8. Dinsmoor MJ, Bachman R, Haney EI, Goldstein M, MacKendrick
W (2004) Outcomes after expectant management of extremely preterm rupture of the membranes. Am J Obstet Gynecol 190:183187
9. Fischer N, Steurer MA, Adams M, Berger TM, Swiss Neonatal
Network (2009) Survival rates of extremely preterm infants (gestational age < 26 weeks) in Switzerland: impact of the Swiss
guidelines for the care of infants born at the limit of viability. Arch
Dis Child Fetal Neonatal Ed 94(6):407413
10. Stoll BJ, Hansen NI, Bell EF, Shankaran S, Laptook AR, Walsh
MC, Hale EC et al (2010) Neonatal oucomes of extremely preterm
infants from the NICHD neonatal research network. Pediatrics
126(3):443456
11. Newman DE, Paamoni-Keren O, Press F, Wiznitzer A, Mazor M,
Sheiner E (2009) Neonatal outcome in preterm deliveries between
23 and 27 weeks gestation with and without preterm premature
rupture of membranes. Arch Gynecol Obstet 280(1):711.
doi:10.1007/s00404-008-0836-8
12. Moretti M, Sibai BM (1988) Maternal and perinatal outcome of
expectant management of premature rupture of membranes in the
midtrimester. Am J Obstet Gynecol 159(2):390396
13. Delprado WJ, Baird PJ, Russell P (1992) Placental candidiasis: report of three cases with a review of the literature. Pathology
14(2):191195

123

1534
14. Deveer R, Engin-Ustun Y, Sarikaya E, Aydogan P, Dofanay M,
Mollamahmutoflu L (2011) Comparison of C-reactive protein levels in pregnancies with retained and removed intrauterine device.
J Matern Fetal Neonatal Med 24(9):11521154. doi:10.3109/
14767058.2010.545925
15. Saraswat L, Bhattacharya S, Maheshwari A, Bhattacharya S (2010)
Maternal and perinatal outcome in women with threatened miscarriage in the Wrst trimester: a systematic review. BJOG 117(3):245257
16. Mori R, Kusuda S, Fujimura M (2011) Antenatal corticosteroids
promote survival of extremely preterm infants born at 22 to
23 weeks of gestation. J Pediatr 159(1):110114
17. Itabashi K, Horiuchi T, Kusuda S, Kabe K, Itani Y, Nakamura T,
Fujimura M, Matsuo M (2009) Mortality rates for extremely low
birth weight infants born in Japan in. Pediatrics 123(2):445450
18. Pasquier JC, Bujold E, Rabilloud M, Picaud JC, Ecochard R, Claris
O, Moret S, Mellier G (2007) EVect of latency period after premature rupture of membranes on 2 years infant mortality (DOMINOS
study). Eur J Obstet Gynecol Reprod Biol 135(1):2127

123

Arch Gynecol Obstet (2012) 285:15291534

19. Ismail MA, Zinaman MJ, Lowensohn RI, Moawad AH (1985) The
signiWcance of C-reactive protein levels in women with premature
rupture of membranes. Am J Obstet Gynecol 151(4):541544
20. Dewan H, Morris JM (2001) A systematic review of pregnancy
outcome following preterm premature rupture of membranes at a
previable gestacional age. Aust N Z J Obstet Gynaecol 41(4):389
21. Shatrov JG, Birch SC, Lam LT, Quinlivan JA, McIntyre S, Mendz
GL (2010) Chorioamnionitis and cerebral palsy: a meta-analysis.
Obstet Gynecol 116(2 Pt 1):387392
22. van TeeVelen AS, van der Ham DP, Oei SG, Porath MM, Willekes
C, Mol BW (2010) The accuracy of clinical parameters in the prediction of perinatal pulmonary hypoplasia secondary to midtrimester prelabour rupture of fetal membranes: a meta-analysis. Eur J
Obstet Gynecol Reprod Biol 148(1):312
23. Winn HN, Chen M, Amon E, Leet TL, Shumway JB, Mostello D
(2000) Neonatal pulmonary hypoplasia and perinatal mortality in
patients with midtrimester rupture of amniotic membranesa critical analysis. Am J Obstet Gynecol 182(6):16381644