Key words
preterm premature rupture of membrane - antibiotic susceptibility - bacterial and Candida colonization - prevalence
Schlüsselwörter
vorzeitiger Blasensprung - Antibiotikaresistenz - Bacteria- und Candida-Kolonisation - Prävalenz
Introduction
In Europe the rate of preterm births is about 5 to 9 % of all births. However in the
USA the preterm birth rate is even higher (12–13 %) and the rates are increasing,
despite a better knowledge of the risk factors involved [1]. This increase is partly due to rising numbers of medically indicated
preterm births (for maternal or fetal indications) and to the increase in assisted
reproductive technologies [1], [27]. Children born preterm have a higher risk for major disabilities,
such as cerebral palsy and respiratory morbidity, and are more likely to develop
behavioral and educational difficulties [2], [3]. The risk of subsequent morbidities and long-term
problems is increased in children born at a younger gestation age [2], [3], [26]. 25 % to 30 % of preterm births are associated with preterm premature
rupture of membranes (PPROM). Thus, optimal management of patients with PPROM is a
major priority in the prevention of preterm birth. Causes of and risk factors for
PPROM are multifactorial and include uterine ischemia or hemorrhage, tobacco
exposure, however the most important causes are infection and inflammation [1]. Di Giulio et al. demonstrated that microbial invasion
of the amniotic cavity affected half of all patients with PPROM (detected by
molecular methods), although invasion was frequently not detected using
culture-based methods, which suggests that molecular techniques (e.g. PCR) may be
more sensitive for the detection of amniotic infections [4], [5], [6]. The
same study group found PPROM to be more frequently associated with Candida spp.
infection than was previously known [4]. The
pathophysiology of infection causing PPROM is considered to be due to the production
of prostaglandins and matrix-degrading enzymes via microbial endotoxins and
pro-inflammatory cytokines (e.g. IL-8, IL1β, TNFα) which, in turn, are released
after the binding of microorganisms to pattern-recognition receptors (e.g. toll-like
receptors) [1]. Prostaglandins play a major role in
stimulating uterine contractions, while degradation of the extracellular matrix in
fetal membranes is implicated in PPROM [1], [7]. However, infection caused by ascending microorganisms
may also occur secondary to PPROM [8], [9]. Intrauterine infection has been demonstrated in up to
60 % of cases [8], [9]. More
recent data suggest that some microbes invade the amniotic cavity from the
bloodstream after dissemination from remote sites, e.g. from the gastrointestinal
tract [4].
The appropriate therapy for women with PPROM is still unclear. The ORACLE Children
Study found that the administration of antibiotics to patients with PPROM was
neither beneficial nor harmful for childrenʼs health at a follow-up of 7 years [10], [11], [12]. In women who had spontaneous labor with no preterm
rupture of membranes, there was some evidence of fetal harm due to the
administration of antibiotics. This highlights the importance of ascertaining that
the diagnosis of PPROM is correct before prescribing antibiotics [11], [12].
In patients with PPROM, the administration of antibiotics appears to offer some
benefit, especially at an earlier gestational age (< 32 weeks), and is associated
with significant benefits in short-term outcomes (prolongation of pregnancy,
prevention of chorioamnionitis and neonatal infection). Thus, antibiotic therapy is
still the standard of care [13], [14]. However, if antibiotics are prescribed, it is still unclear which
should be the antibiotic of choice. Penicillins and macrolide antibiotics can be
administered alone or in combination, parenterally or orally. The administration of
amoxicillin plus clavulanic acid to pregnant mothers was found to be associated with
a significantly higher risk of necrotizing enterocolitis in infants after delivery,
and it was recommended not to prescribe this combination to pregnant women [15], [16]. Thus, more
information on the prevalence, spectrum and antibiotic sensitivity of bacterial and
Candida colonization is important to determine the optimal treatment strategies for
patients with PPROM. The aim of our study was therefore to provide data on the
prevalence and antibiotic sensitivity of bacterial and Candida colonization in
patients with PPROM to improve the knowledge base for decisions on the use of
specific antibiotics. We also carried out a subgroup analysis comparing prevalence,
sensitivity and colonization in early PPROM (21st – 27th week of gestation) with
that in late PPROM (28th – 34th week of gestation).
Methods
Study population
All samples were collected at the perinatal center of Johann Wolfgang Goethe
University Frankfurt, Germany, between June 2006 and May 2011. A retrospective
cohort study was done in patients with PPROM who met the following inclusion
criteria: (i) gestational age between 21 and 34 weeks, and (ii) high vaginal
swabs taken on admission to hospital for microbiological diagnosis prior to
receiving antibiotic treatment. A subgroup analysis was done to compare findings
in the 21st to 27th week of gestation (group A) with those in the 28th to 33rd
week of gestation (group B).
Study protocol
Membrane rupture was diagnosed by insulin-like growth factor binding protein-1
(IGFBP-1; Actim® PROM). After the diagnosis of PPROM, a high vaginal swab was
taken for microbiological culture of aerobic and anaerobic bacteria and fungi.
Microbiological culture techniques were performed in accordance with standard
laboratory techniques [17]. Bacterial susceptibility
testing was done according to the recommendations of the Clinical and Laboratory
Standards Institute, Wayne, PA, USA [18]. All
procedures were performed by the routine diagnostic laboratories of the
Institute of Medical Microbiology and Infection Control of the University
Hospital of Frankfurt (Germany). These laboratories are certified in accordance
with DIN EN ISO/IEC 17025 and 15189 standards (laboratory identification code:
D-ML-013102-01 and D-PL-13102-01-00).
Statistical analysis
All results are shown as mean values ± standard deviation. Wilcoxon signed rank
test and Fisherʼs exact test were used for statistical analysis. Analysis was
done using SPSS Statistics 17.0 software (Scientific Packages for the Social
Sciences, Inc., Chicago, IL, USA). A value of p < 0.05 for a two-tailed test
was considered statistically significant.
Results
Cervical swabs obtained from 245 subjects with PROM were analyzed. [Table 1] presents the baseline characteristics of the 245
enrolled patients. Aerobic bacteria were identified with culture testing ([Table 2]).
Table 1 Characteristics of the study population (n = 245;
n. s. = not significant).
|
Characteristic
|
Group A
21st to 27th week of gestation (n = 56)
|
Group B
28th to 33rd week of gestation (n = 189)
|
p-value
|
|
Maternal age (years)
|
32.6 ± 4.3
|
33.0 ± 5.4
|
n. s.
|
|
Nulliparity
|
62.3 % (33/53)
|
67.1 % (110/164)
|
n. s.
|
|
Candida colonization
|
11.1 % (6/54)
|
24.3 % (37/152)
|
0.04
|
|
Bacterial colonization
|
40.8 % (21/54)
|
41.4 % (63/152)
|
n. s.
|
|
Normal vaginal bacteria
|
55.6 % (30/54)
|
46.1 % (70/152)
|
n. s.
|
|
Overall resistance
|
|
|
|
|
|
71.4 % (15/21)
|
52.5 % (32/61)
|
n. s.
|
|
|
9.5 % (2/35)
|
11.7 % (7/60)
|
n. s.
|
|
|
28.6 % (6/35)
|
16.4 % (10/61)
|
n. s.
|
|
|
5.0 % (1/20)
|
5.4 % (3/56)
|
n. s.
|
Table 2 Bacteria identified on culture (n = 206; n. s. = not
significant).
|
Bacteria species
|
Group A
21st to 27th week of gestation (n = 54)
|
Group B
28th to 33rd week of gestation (n = 152)
|
p (Fisherʼs exact test)
|
|
E. coli
|
5.6 % (3/54)
|
13.2 % (20/152)
|
n. s.
|
|
Enterococcus faecalis
|
0 % (0/54)
|
0.7 % (1/152)
|
n. s.
|
|
Staphylococcus aureus
|
5.6 % (3/54)
|
2.6 % (4/152)
|
n. s.
|
|
Normal vaginal species
|
55.6 % (30/54)
|
46.1 % (70/152)
|
n. s.
|
|
Streptococcus agalactiae
|
5.6 % (3/54)
|
5.3 % (8/152)
|
n. s.
|
|
Streptococcus dysgalactiae
|
0 % (0/54)
|
0.7 % (1/152)
|
n. s.
|
|
Klebsiella pneumoniae
|
5.6 % (3/54)
|
3.3 % (5/152)
|
n. s.
|
|
Klebsiella oxytoca
|
3.7 % (2/54)
|
0 % (0/152)
|
n. s.
|
|
Salmonella enteritidis
|
0 % (0/54)
|
0.7 % (1/152)
|
n. s.
|
|
Enterobacter cloacae
|
0 % (0/54)
|
2.6 % (4/152)
|
n. s.
|
|
Pseudomonas aeruginosa
|
0 % (0/54)
|
0.7 % (1/152)
|
n. s.
|
|
Proteus mirabilis
|
0 % (0/54)
|
1.3 % (2/152)
|
n. s.
|
|
Proteus vulgaris
|
0 % (0/54)
|
0.7 % (1/152)
|
n. s.
|
|
Citrobacter koseri
|
0 % (0/54)
|
0.7 % (1/152)
|
n. s.
|
|
Acinetobacter Iwoffii
|
0 % (0/54)
|
1.3 % (2/152)
|
n. s.
|
|
No growth
|
18,5 % (10/54)
|
20.4 % (31/152)
|
n. s.
|
The cervical prevalence of bacteria which colonize normal vaginal flora
(Escherichia [E.] coli, Streptococcus agalactiae, Klebsiella spp. etc.)
and are potentially involved in amniotic infections was similar in both study groups
(40.8 vs. 41.4 %; p > 0.05), however a difference in the susceptibility to
antibiotics was noted, although this difference was not statistically significant
(overall resistance to ampicillin 71.4 vs. 52.5 %; resistance to cefuroxime 9.5 vs.
11.7 %; resistance to gentamicin 28.6 vs. 16.4 %; resistance to ciprofloxacin 5.0
vs. 5.4 %). [Table 3] shows the individual susceptibility
to antibiotics for the majority of pathogenic bacteria. Group A had a significantly
lower rate of Candida colonization (11.1 vs. 24.3 %; p = 0.04) and the difference
was statistically significant.
Table 3 Resistance of E. coli, Streptococcus
agalactiae and Klebsiella spp. (n = 245; n. s. = not
significant).
|
Characteristic
|
Group A
21st to 27th week of gestation (n = 56)
|
Group B
28th to 33rd week of gestation (n = 189)
|
p-value
|
|
E. coli resistance
|
|
|
|
|
|
66.7 % (8/12)
|
55.0 % (22/40)
|
n. s.
|
|
|
16.7 % (2/12)
|
2.5 % (1/40)
|
n. s.
|
|
|
25.0 % (3/12)
|
5.0 %(2/40)
|
n. s.
|
|
|
8.3 % (1/12)
|
7.5 % (3/40)
|
n. s.
|
|
Streptococcus agalactiae resistance
|
|
|
|
|
|
0 % (0/2)
|
0 % (0/6)
|
n. s.
|
|
|
0 % (0/2)
|
0 % (0/6)
|
n. s.
|
|
|
100 % (2/2)
|
100 %(6/6)
|
n. s.
|
|
|
0 % (0/2)
|
0 % (0/6)
|
n. s.
|
|
Klebsiella spp. resistance
|
|
|
|
|
|
100 % (4/4)
|
100 % (1/1)
|
n. s.
|
|
|
0 % (0/4)
|
0 % (0/1)
|
n. s.
|
|
|
0 % (0/4)
|
0 %(0/1)
|
n. s.
|
|
|
0 % (0/4)
|
0 %(0/1)
|
n. s.
|
Conclusion
The study showed that the prevalence of bacteria colonizing the normal vaginal flora
and potentially involved in amniotic infection in PPROM was around 41 % for both
subgroups. These results correspond to previously published data [4], [19], [20]. It is interesting to note that 10 and 71 %, respectively, of
detected bacterial species were resistant to the commonly used broad-spectrum
antibiotics cefuroxime and ampicillin. The major pathogenic bacteria showed a varied
susceptibility to antibiotics ([Table 3]). Different
antibiotic susceptibilities were noted for group A (21st to 27th week of gestation)
compared to group B (28th to 33rd week of gestation). Group A showed a higher
resistance to broad-spectrum ampicillin (71.4 vs. 52.5 %) while resistance to
cefuroxime was higher in group B, although at a markedly lower level (9.5 vs.
11.7 %). These findings are in accordance with data showing increasing numbers of
antibiotic-resistant E. coli infections among preterm infants.
Ampicillin-resistant E. coli infections would seem to be linked to prolonged
antenatal administration of antibiotics in cases with PPROM and appear to
preferentially affect preterm infants [21]. Furthermore,
the intrapartum administration of ampicillin has been shown to be an independent
risk factor for ampicillin-resistant E. coli early-onset sepsis and also
increases E. coli late-onset sepsis [21]. Thus,
there is a need for further and larger studies evaluating the bacterial spectrum and
antibiotic sensitivity in patients with PPROM, with a special focus on certain
subgroups of patients, as the data suggests that different treatment approaches
should be used for patients in the 21st to 27th week of gestation and for patients
in the 28th to 33rd week of gestation. We found some aerobic bacteria, but most
vaginal bacteria are anaerobic, especially in women with bacterial vaginosis which
commonly affects around 20 % of the study population [22].
Special emphasis should be laid on treating Candida colonization, especially later in
pregnancy, as our subgroup B had a significantly higher rate of Candida colonization
(11.1 vs. 24.3 %; p = 0.04). The finding that Candida colonization is significantly
higher later on in pregnancy is well known [23]; two
studies noted the impact of vaginal Candida colonization on preterm birth but it is
still not fully understood and has not been investigated in prospective studies
[24], [25]. As mentioned
above, infection with Candida spp. was found to be more frequently associated with
PPROM than was previously thought [4] and that effective
therapy should be administered to prevent preterm birth.
Future studies should include outcome measures (e.g. intra-amniotic inflammation at
presentation, histopathologic inflammation of the placenta and chorioamniotic
membranes at delivery) as well as pregnancy and perinatal outcomes.
Clinical Practice
Around 41 % of women with PPROM have pathological bacterial contamination of the
vagina resistant to the most commonly administered broad-spectrum antibiotics
ampicillin and cefuroxime.
