Residents’ Papers www. AJOG.

org

A retrospective review of ampicillin-sulbactam

and amoxicillin ⴙ clavulanate vs cefazolin/cephalexin
and erythromycin in the setting of preterm premature
rupture of membranes: maternal and neonatal outcomes
Robert M. Ehsanipoor, MD; Judith H. Chung, MD; Charlotte A. Clock, MD; Jennifer A. McNulty, MD; Deborah A. Wing, MD

OBJECTIVE: The purpose of this study was to compare the efficacy and
outcomes of 2 different antibiotic regimens that are used to prolong
latency in preterm premature rupture of membranes. The primary ob-
jective was to determine whether the use of ampicillin-sulbactam/
amoxicillin ⫹ clavulanate was associated with an increased risk of
necrotizing enterocolitis.
STUDY DESIGN: A retrospective review of pregnancies that were com-
plicated by preterm premature rupture of membranes from 1999-2006
at 2 institutions was performed. Outcomes were compared between
subjects who received parenteral ampicillin-sulbactam followed by oral
amoxicillin ⫹ clavulanate (protocol A) and subjects who received par-
enteral cefazolin and erythromycin followed by oral cephalexin and
erythromycin (protocol B).
RESULTS: There were 147 women who were evaluated; 88 women re-
ceived protocol A, and 59 women received protocol B. There were no
differences in latency period, gestational age at delivery, or route of
delivery. The incidence of necrotizing enterocolitis was 8.0% and
10.2% for protocol A and protocol B, respectively (P ⫽ .64).
CONCLUSION: Ampicillin-sulbactam/amoxicillin ⫹ clavulanate was
not associated with an increase in neonatal necrotizing enterocolitis.
Erythromycin in combination with cefazolin and cephalexin is an effec-
tive latency antibiotic regimen.
Key words: latency antibiotic regimen, necrotizing enterocolitis,
preterm premature rupture of membranes

A lthough the benefit of administering

antibiotics to patients in the setting of
preterm premature rupture of membranes
(PPROM) has been clearly demonstrated
in numerous randomized controlled trials,
the optimal regimen remains unclear.1
Mercer et al2 showed benefit with paren-
teral ampicillin and erythromycin fol-
lowed by oral amoxicillin and erythro-
mycin. However, increasing antibiotic
resistance, particularly to ampicillin, sug-
gests that this regimen may have decreased
efficacy in certain patient populations and
actually may cause harm.3 Antibiotic regi-
From the Department of Obstetrics and
Gynecology, Division of Maternal-Fetal
Medicine, University of California, Irvine,
Orange, CA (Drs Ehsanipoor, Chung,
Clock, and Wing); and Miller Children’s
Hopsital and Long Beach Memorial Medical
Center, Long Beach, CA (Dr McNulty).
Received July 31, 2007; accepted Dec. 21,
2007.
Reprints not available from the authors.
0002-9378/free
© 2008 Mosby, Inc. All rights reserved.
doi: 10.1016/j.ajog.2007.12.022
mens with a more broad spectrum of cov-
erage, such as ampicillin-sulbactam and
amoxicillin ⫹ clavulanate have also been
studied in the setting of PPROM; although
the use has been associated with improved
maternal and neonatal outcomes, the reg-
imens have not been widely adopted.4 This
perhaps is due to the concern that the reg-
imens may be associated with an increased
risk of necrotizing enterocolitis.5,6
At our institutions, 1 of 2 different anti-
biotic regimens is generally used in the set-
ting of PPROM: parenteral ampicillin-sul-
bactam followed by oral amoxicillin ⫹
clavulanate or parenteral cefazolin and
erythromycin followed by oral cephalexin
and erythromycin. This afforded the
unique opportunity to retrospectively
compare the 2 regimens. In particular, we
wished to investigate the risk of necrotizing
enterocolitis with the use of these 2 differ-
ent antibiotic regimens.
M ATERIALS AND M ETHODS
Institutional Review Board approval was
obtained from Long Beach Memorial
Medical Center and the University of
California, Irvine. Pregnancies that were
complicated by PPROM from 1999-
2006 were identified with the use of a
computerized obstetrics database and
delivery log books.
The choice of latency antibiotic regi-
men was at the discretion of the admit-
ting physician. Protocol A consisted of
48 hours of parenteral ampicillin-sul-
bactam (3 gm every 6 hours) followed by
5 days of oral amoxicillin ⫹ clavulanate
(500 mg every 8 hours), as described by
Lovett et al.4 Protocol B was a modifica-
tion of the protocol of Mercer et al2 in
which the subjects received 48 hours of
parenteral cefazolin (2 gm every 8 hours)
and parenteral erythromycin (250 mg
every 6 hours), followed by 5 days of oral
cephalexin (500 mg every 6 hours) and
oral erythromycin (250 mg every 6
hours). In general, subjects at the Uni-
versity of California, Irvine, received
protocol A, and subjects at Long Beach
Memorial Medical Center received pro-
tocol B, although some overlap did
occur.
Maternal and neonatal charts were re-
viewed for all singleton pregnancies that
were complicated by PPROM between
24-32 weeks of gestation that were

e54 American Journal of Obstetrics & Gynecology MAY 2008

www.AJOG.org
treated with either protocol A or proto-
col B. Subjects who were in active labor,
who had an indication for immediate de-
livery, or who had an anomalous preg-
nancy were excluded. Definitions used
for the primary and secondary outcomes
are presented in Table 1.
Fisher exact or chi-square tests were
used for the analysis of categoric data.
The Student t test was used for continu-
ous variables. Univariate and multivari-
ate logistic regressions were then per-
formed for the primary outcome of
necrotizing enterocolitis as a function of
latency antibiotic. All analyses were per-
formed with Stata software (version 9.0;
Stata Corp, College Station, TX).
With the exception of latency antibi-
otic regimen, the management of
PPROM between the 2 institutions was
similar. In general, patients are admitted
and evaluated for evidence of infection,
placental abruption, and/or active labor.
Practices regarding fetal monitoring,
corticosteroid use, tocolytics, and gesta-
tional age of elective delivery are similar
at both institutions. Amniocentesis to
evaluate for infection or fetal lung matu-
rity was not used routinely in the treat-
ment of patients.
R ESULTS
During the study period, there were 147
subjects who met the inclusion criteria;
88 patients received protocol A, and 59
patients received protocol B. The groups
were well-matched with respect to base-
line demographics (data not shown).
The average gestational age at ruptured
membranes in each group was 28.5
weeks (P ⫽ .77). The rate of confirmed
necrotizing enterocolitis between the 2
groups was not statistically significant
with 7 patients (8.0%) who received pro-
tocol A and with 6 patients (10.2%) who
received protocol B (P ⫽ .64). This dif-
ference remained insignificant after ad-
justment for potential confounders that
included maternal transport, tocolytic
use, previous antibiotic exposure, and
hospital site of delivery (odds ratio, 0.76;
95% CI, 0.14-4.03). Outcomes between
the 2 groups were also comparable with
respect to latency, infectious morbidity,
and other adverse neonatal outcomes
Residents’ Papers
TABLE 1
Clinical criteria used for morbidities
Clinical outcome Diagnostic criteria
Necrotizing enterocolitis Modified Bell Staging criteria used
..............................................................................................................................................................................................................................................
Stage II-III: confirmed necrotizing enterocolitis);
confirmed surgically or radiograph finding of
pneumatosis intestinalis, pneumoperitoneum, portal
air, or intestinal dilatation in addition to compatible
symptoms
..............................................................................................................................................................................................................................................
Chorioamnionitis Any 2 of the following: Antepartum temperature of
⬎38°C, uterine tenderness, foul-smelling vaginal
discharge or amniotic fluid, maternal heart rate
⬎100 beats/min, fetal heart rate ⬎160 beats/min, or
white blood cell count ⬎20⫻105/L
..............................................................................................................................................................................................................................................
Endometritis Any 2 of the following: Persistent temperature
⬎38°C, foul lochia, uterine tenderness, white blood
cell count ⬎20⫻105/L, and no other identifiable
cause
..............................................................................................................................................................................................................................................
Respiratory distress syndrome Compatible symptoms and radiographic findings of
hyaline membrane disease or respiratory insufficiency
that requires ventilatory support for at least 24 hours
..............................................................................................................................................................................................................................................
Pneumonia Compatible symptoms and diagnostic radiographic
findings
..............................................................................................................................................................................................................................................
Bronchopulmonary dysplasia Oxygen requirement at 36 weeks after menstrual age
Ehsanipoor. PPROM and latency antibiotics. Am J Obstet Gynecol 2008.
(Tables 2 and 3). Specifically, rates of that then become invasive is thought to
neonatal sepsis were similar, and there be of importance.7
were no cases of infection because of am- An increased risk of necrotizing en-
picillin resistant Escherichia coli or other terocolitis with the use of amoxicillin ⫹
highly resistant organisms. clavulanate in PPROM has been re-
ported in 2 studies.5,6 Although there
C OMMENT may be a plausible mechanism by which
Necrotizing enterocolitis is a neonatal this association may occur, the strength
complication that is seen primarily in of the association is questionable. The
premature infants and is characterized first report had a relatively small sample
by ischemic injury to the bowel. The size (n ⫽ 62 patients), and necrotizing
cause is unknown; however, intestinal enterocolitis was 1 of multiple secondary
colonization of pathogenic organisms outcomes that were evaluated.5 The
TABLE 2
Maternal outcomes
Amoxicillin & Cefazolin/cephalexin
clavulanic acid & erythromycin
Outcome (n ⴝ 88) (n ⴝ 59) P value
Chorioamnionitis (n) 16 (18.2%) 9 (15.3%) .64
..............................................................................................................................................................................................................................................
Endometritis (n) 7 (7.9%) 2 (3.4%) .32
..............................................................................................................................................................................................................................................
Latency period (d) a
11.1 ⫾ 12.7 10.5 ⫾ 12.7 .79
..............................................................................................................................................................................................................................................
Latency ⱖ 2 d (n) 72 (81.8%) 41 (69.5%) .08
..............................................................................................................................................................................................................................................
Latency ⱖ 7 d (n) 44 (50%) 25 (42.4%) .36
..............................................................................................................................................................................................................................................
Data were analyzed by ␹ test or Fisher’s exact test for categoric data and the Student t test for continuous data.
2
a
Data are presented as mean ⫾ SD.
Ehsanipoor. PPROM and latency antibiotics. Am J Obstet Gynecol 2008.
MAY 2008 American Journal of Obstetrics & Gynecology e55
Residents’ Papers www.AJOG.org

are needed to determine the safety and

TABLE 3 efficacy of the cephalosporin/erythro-
Neonatal outcomes mycin regimen and to further evaluate a
Amoxicillin & Cefazolin/cephalexin potential association with necrotizing
clavulanic acid & erythromycin enterocolitis with these and other antibi-
Outcome (n ⴝ 88) (n ⴝ 59) P value
otic regimens that are used in the setting
Necrotizing enterocolitis 7 (8.0%) 6 (10.2%) .64 of PPROM. f
stage II/III (n)
..............................................................................................................................................................................................................................................
Neonatal sepsis (n) 17 (19.1%) 13 (22.0%) .24
..............................................................................................................................................................................................................................................
REFERENCES
Respiratory distress 31 (35.6%) 25 (42.4%) .41
1. Canavan TP, Simhan HN, Caritis S. An evi-
syndrome (n)
.............................................................................................................................................................................................................................................. dence-based approach to the evaluation and
Bronchopulmonary 13 (14.9%) 10 (16.9%) .74 treatment of premature rupture of membranes:
dysplasia (n) part I. Obstet Gynecol Surv 2004;59:669-77.
..............................................................................................................................................................................................................................................
Pneumonia (n) 6 (6.8%) 1 (1.7%) .24 2. Mercer BM, Miodovnik M, Thurnau GR, et al.
.............................................................................................................................................................................................................................................. Antibiotic therapy for reduction of infant morbid-
Intraventricular 4 (4.6%) 2 (3.4%) 1.0 ity after preterm premature rupture of the mem-
hemorrhage grade III/IV (n) branes; a randomized controlled trial: National
..............................................................................................................................................................................................................................................
Neonatal length of stay 45.3 ⫾ 35.7 53.4 ⫾ 44.8 .23 Institute of Child Health and Human Develop-
(d)a ment Maternal-Fetal Medicine Units Network.
.............................................................................................................................................................................................................................................. JAMA 1997;278:989-95.
Median 35 42 3. Terrone DA, Rinehart BK, Einstein MH, Britt
..............................................................................................................................................................................................................................................
Interquartile range 22.5-59.5 25-69 LB, Martin JN, Perry KG. Neonatal sepsis and
.............................................................................................................................................................................................................................................. death caused by resistant Escherichia coli: pos-
Death before discharge (n) 1 (1.1%) 3 (5.1%) .30 sible consequences of extended maternal am-
..............................................................................................................................................................................................................................................
Data were analyzed by ␹2 test or Fisher’s exact test for categoric data and the Student t test for continuous data. picillin administration. Am J Obstet Gynecol
a
Data are presented as mean ⫾ SD. 1999;180:1345-8.
4. Lovett SM, Weiss JD, Diogo MJ, Williams PT,
Ehsanipoor. PPROM and latency antibiotics. Am J Obstet Gynecol 2008.
Garite TJ. A prospective, double-blind, ran-
domized, controlled clinical trial of ampicillin-
ORACLE I trial also evaluated necrotiz- findings also suggest that ampicillin-sul- sulbactam for preterm premature rupture of
ing enterocolitis as 1 of many secondary bactam/ amoxicillin ⫹ clavulanate and a membranes in women receiving antenatal cor-
outcomes in a sample of 4826 subjects; cephalosporin combined with erythro- ticosteroid therapy. Am J Obstet Gynecol
1997;176:1030-8.
although this study had a notably larger mycin are both safe and effective antibi- 5. Cox SM, Leveno KJ, Sherman ML, Travis L,
sample size, the clinical relevance of the otic regimens for patients with PPROM. DePalma R. Ruptured membranes at 24 to 29
different necrotizing enterocolitis rates Thus, in hospitals with high levels of am- weeks: a randomized double blind trial of anti-
is unclear. This is because their reported picillin-resistance, the substitution of a microbials versus placebo. Am J Obstet Gy-
rates of 1.8% and 0.7% in the treated and cephalosporin, which is to be adminis- necol 1995;172:412.
6. Kenyon SL, Taylor DJ, Tarnow-Mordi W, for
untreated subjects, respectively, were tered in conjunction with erythromycin,
the ORACLE Collaborative Group. Broad-spec-
well below those that would be expected may be a reasonable alternative for the trum antibiotics for preterm, prelabour rupture
in a preterm population, which may be treatment of PPROM. of fetal membranes: the ORACLE I randomised
explained by their inclusion of subjects One limitation of our study is the trial. Lancet 2001;357:979-88.
up to 37 weeks of gestation.6 small sample size. A post hoc sample size 7. Fanaroff AA, Martin RJ, Rodriguez RJ. Iden-
tification and management of problems in the
In contrast, our study results do not calculation revealed that our study had
high-risk neonate. In: Creasy RK, Resnick R,
demonstrate a higher risk of necrotizing 50% power to detect a 10% difference in editors. Maternal fetal medicine: principles and
enterocolitis in the infants who are ex- necrotizing enterocolitis between the 2 practice. 5th ed. Philadelphia: Elsevier; 2004:
posed to amoxicillin ⫹ clavulanate. Our groups. Larger prospective clinical trials 1263-301.

e56 American Journal of Obstetrics & Gynecology MAY 2008