Beruflich Dokumente
Kultur Dokumente
ORIGINAL ARTICLE
Impact of duration of rupture of membranes on outcomes of
premature infants
MW Walker1, AH Picklesimer3, RH Clark1,2, AR Spitzer2 and TJ Garite2
OBJECTIVE: The primary aim of the study was to determine how the risk of adverse outcomes was related to the duration of
the latency period and gestational age at birth following preterm premature rupture of the fetal membranes (PPROM).
STUDY DESIGN: Retrospective review of infants discharged from 330 neonatal intensive care units. We defined four subgroups
based on gestational age: 23 to 25, 26 to 28, 29 to 31 and 32 to 34 weeks. Each gestational age group was evaluated by duration of
ROM: o24 h, 1 to 7 days, 8 to 14 days, 15 to 21 days, 21 to 28 days and >28 days and compared with a referent group (PPROM of
>24 h but o 7 days).
RESULT: In all, 239 808 non-anomalous infants 23 to 34 weeks’ gestational age were identified; 37 233 (15.5%) had rupture of
membranes (ROM) >24 h. Compared with a reference group (PPROM of >24 h but o 7 days), the risk of mortality for PPROM
of 8 to 14, 15 to 21 and 21 to 28 days varied depending on gestational age at birth. Only PPROM >28 days was consistently
associated with increased mortality and decreased likelihood of survival without morbidity in all gestational age subgroups.
CONCLUSION: PPROM for >28 days is associated with an increased risk of death and morbidity.
Journal of Perinatology (2014) 34, 669–672; doi:10.1038/jp.2014.73; published online 24 April 2014
1
Division of Neonatology, Children’s Hospital, Greenville Health System, Greenville, SC, USA; 2Departments of Education, Research and Quality Improvement, Pediatrix-Obstetrix
Center for Research and Education, Sunrise, FL, USA and 3Department of Maternal-Fetal Medicine, Greenville Health System, Greenville, SC, USA. Correspondence: Dr MW Walker,
Division of Neonatology, Children’s Hospital, Greenville Health System, Greenville, SC, USA.
E-mail: whit_walker@pediatrix.com
Received 2 December 2013; revised 7 March 2014; accepted 19 March 2014; published online 24 April 2014
Prolonged rupture of the membranes and outcome
MW Walker et al
670
Data selection mature (32 vs 33 weeks), minimally smaller (1.78 vs 1.79 kg), more
We included inborn infants with an estimated gestational age at birth often exposed to antenatal steroids (81% vs 62%), more often a
between 23 and 34 completed weeks that were discharged from the singleton birth (87% vs 64%) and less often delivered by cesarean
hospital of birth between 1 January 1997 and 1 July 2012. We excluded section (43% vs 67%). Overall mortality was decreased and survival
infants transferred in from another hospital (out born), infants transferred without morbidity was increased in infants whose membranes
out of the birth hospital and infants with major anomalies. were ruptured for >24 h but o 7 days compared with infants
whose membranes were ruptured o 24 h. The absolute differ-
Data analysis ences were small: mortality—3.5% vs 3.9%, AOR = 0.8, confidence
Comparisons of continuous variables (e.g., gestational age) were evaluated interval (CI) = 0.7 to 0.9; P o0.001; survival without morbidity—-
with analysis of variance. Categorical variables (e.g., race and sex) were 86% vs 85%, AOR = 1.13, CI = 1.07 to 1.19; P o0.01. Similar trends
evaluated with two-tailed χ2 tests or Fisher's exact test (2 by 2 tables). were seen within each gestational age group. The demographic
Nonparametric data were assessed with Kruskal–Wallis analysis of variance. and maternal treatment characteristics of the patients with
The linear trend test and the Cochran–Armitage trend test were performed rupture of membranes o 24 h were so distinctive that we did
to evaluate time-related changes. not use them as our reference group.
Multivariable logistic regression was used to calculate the adjusted odds Compared with the infants with PPROM 1 to 7 days, infants with
ratios (AOR) and 95% confidence intervals (CIs) for the variables associated
PPROM >7days were smaller (1.68 vs 1.82 kg), more often exposed
with an increased risk of death or morbidity. Because mortality is a
competing variable with morbidities (i.e., if you die you cannot develop a to antenatal steroids (89% vs 77%) and less often delivered by
late morbidity), ORs for each individual morbidity were not calculated, but cesarean section (40% vs 49%). Gestational age (32 vs 32) and
instead a variable was created for survival without major morbidity (stage 3 singleton birth (89% vs 86%) were similar for infants with PPROM
or 4 retinopathy of prematurity, grade 3 or 4 intraventricular hemorrhage, >7 days and those with PPROM 1 to 7 days. Within each
necrotizing enterocolitis that required medical or surgical treatment, gestational age group the demographic and maternal treatment
oxygen use at 36 weeks postmenstrual age for infants born at 32 weeks or characteristics of the patients with PPROM 1 to 7 days were similar
less and at 28 days for infants >32 weeks) and the AORs for survival and the differences were small compared with the other PPROM
without morbidity were determined. The variables in the model for groups (Supplementary Tables). For this reason, we used the
calculating the AORs included estimated gestational age, birth weight,
PPROM 1 to 7 days as our reference group.
gender, multiples/singleton fetal status, reported exposure to antenatal
steroids, year of discharge and duration of latency in each PPROM group.
For the data in the Supplementary Tables, the logistic model evaluating Medications and support. Compared with the infants with PPROM
ORs for death and survival without morbidity in each gestational age 1 to 7 days, and when corrected for gestational age, birth weight,
group was considered an independent sample population.
gender, multiples, antenatal steroids and discharge year using
In our original design, we used two reference groups for calculation of
the odds of death and the odds of survival without morbidity: PPROM logistic modeling, infants with PPROM >7 days were more often
o24 h and the PPROM 1 to 7 days group (Supplementary Tables 1–4). The treated with vasopressors (11% vs 6%, AOR = 1.8, 95%tile CI
reason for using two reference groups was that infants born to mothers = 1.7 to 2.0), more often required ventilator support (37% vs
with ruptured membranes o24 h were significantly different from infants 24%, AOR = 1.7, 95%tile CI = 1.6 to 1.8) and were more often
with rupture of membranes 24 h or more. Premature infants born within treated with surfactants (35% vs 23%, AOR = 1.6, 95%tile CI
24 h of rupture of membranes may represent a group that was delivered = 1.5 to 1.7). The Supplementary Tables show that the treatment
for fetal distress or emergent medical indications like pre-eclampsia. differences are most pronounced in infants in the more mature
However, our database cannot always define the exact rationale behind gestational age groups (>26 weeks) and in the infants exposed to
each delivery. Because of this potential selection bias, we have not
the rupture of membranes >28 days.
included the data on ORs for this reference group compared with the other
PPROM groups.
Mortality and survival without morbidity. Compared with the
infants with PPROM 1 to 7 days, and when corrected for
RESULTS gestational age, birth weight, gender, multiples, antenatal steroids
Study population and discharge year, infants with more prolonged rupture of
Between January 1997 and July 2012, 239 808 non-anomalous, membranes were more likely to die: 15 to 21 days (4.7% vs
inborn, infants between 23 and 34 weeks’ gestational age were 3.0%, AOR = 1.5, CI = 1.2 to 1.8); 22 to 28 days (5.5% vs
reported to the Clinical Data Warehouse; 37 233 (15.5%) were 3.0%, AOR = 1.6, CI = 1.2 to 2.2) and >28 days (12.6% vs 3.0%,
reported to have rupture of membranes >24 h before delivery. To AOR = 5.7, CI = 4.7 to 6.9). There was no significant difference in
understand the impact of duration of ruptured membranes within mortality for infants with PPROM 1 to 7 days compared with those
specific estimated gestational ages, the infant populations were with PPROM 8 to 14 days (3.8% vs 3.0%). The largest mortality
broken down into four subsets based on gestational age at birth: difference was observed when comparing infants with PPROM of
23 to 25 weeks (n = 12 595; Supplementary Table 1), 26 to 28 weeks 1 to 7 days with those with PPROM >28 days and this was true
(n = 24 459; Supplementary Table 2), 29 to 31 weeks (n = 4949; across all gestational age groups (Supplementary Tables and
Supplementary Table 3) and 32 to 34 weeks (n = 153 263; Figure 1)
Supplementary Table 4). Compared with the infants with PPROM 1 to 7 days and when
Each subset of premature infants within their respective gesta- corrected for gestational age, birth weight, gender, multiples,
tional age groups was evaluated by duration of rupture of mem- antenatal steroids and discharge year, infants with more
branes: o24 h (n = 202 585, 84.5%), 1 to 7 days (n = 25 194, 10.5%), prolonged rupture of membranes were less likely to survive
8 to 14 days (n = 6507, 2.7%), 15 to 21 days (n = 2501, 1.0%), 21 to without morbidity: 15 to 21 days (83% vs 88%, AOR = 0.9,
28 days (n = 1067, 0.4%) and >28 days (n = 1954, 0.8%). In multi- CI = 0.7 to 1, P = 0.03); 22 to 28 days (78% vs 88%, AOR = 0.6,
variate analysis, the 1 to 7 days group was used as the reference CI = 0.5 to 0.7, P o 0.001) and >28 days (71% vs 88%, AOR = 0.3,
group to which the other PPROM groups were compared. CI = 0.3 to 0.4, Po 0.001). There was no significant difference in
survival without morbidity for infants with PPROM 1 to 7 days
compared with those with PPROM 8 to 14 days (85% vs 88%). The
Comparison of infants by gestational age and duration of rupture largest difference was observed when comparing infants with
of membranes PPROM of 1 to 7 days with those with PPROM >28 days and this
Demographics. Compared with infants with rupture of mem- was true across all gestational age groups (Supplementary Tables
branes o24 h, infants with PROM (⩾24 h) were slightly less and Figure 1)
23 to 25
26 to 28
29 to 31
32 to 34
All Patients
23 to 25
26 to 28
29 to 31
32 to 34
All Patients
23 to 25
26 to 28
29 to 31
32 to 34
All Patients
EGA Group
2
PPROM Groups; Reference Group is PPROM >1 and <=7 days
Adjusted Odds Ratio For Survival
1.8
8 to14 days 15 to 21 days 22 to 28 days > 28 days
1.6
without Major Morbidity
1.4
1.2
1
0.8
0.6
0.4
0.2
0
23 to 25
26 to 28
29 to 31
32 to 34
All Patients
23 to 25
26 to 28
29 to 31
32 to 34
All Patients
23 to 25
26 to 28
29 to 31
32 to 34
All Patients
23 to 25
26 to 28
29 to 31
32 to 34
All Patients
EGA Group
Figure 1. Impact of PPROM on the odds of mortality (a) and survival without morbidity (b) within each gestational age group. Within each
gestational age group, the impact of each PPROM group on the outcome is compared with the reference group PPROM 1 to 7 days. (a) AOR
for mortality by duration of rupture of membranes within each gestational age group. (b) AOR for survival without morbidity (stage 3 or 4
retinopathy of prematurity, grade 3 or 4 intraventricular hemorrhage, necrotizing enterocolitis that required medical or surgical treatment,
oxygen use at 36 weeks postmenstrual age for infants born at 32 weeks or less and at 28 days for infants >32 weeks) by duration of rupture of
membranes within each gestational age group. In the logistic model, each gestational age group was considered an independent sample
population. The variables in the model for calculating the AORs for each outcome included gestational age, birth weight, gender, multiples,
antenatal steroids, discharge year and duration of PPROM group.
DISCUSSION Our findings support the existing literature and suggest that the
In premature infants admitted to our NICUs over the past 15 years, ability to prolong latency for at least 1 to 7 days will allow time for
approximately 15% had prolonged rupture of the membranes, but treatment of the fetus with antenatal steroids and antibiotics (or
only 33% of these infants had rupture of the membranes for longer other additional therapies) and will result in improveds outcomes
than 1 week (3.1% of the total sample). Within all gestational groups, (especially crucial in pre- and periviable cases of PPROM).11,20 Most
the best outcomes were in the patients with PPROM of 1 to 7 days. importantly, the risk of bronchopulmonary dysplasia, vertically
Latency of >24 h and less than a week appeared to provide the transmitted neonatal infection and hospital length of stay did not
greatest benefit in the most immature GA group (Supplementary appear to increase as PPROM duration extended to 2 weeks. Our
Tables). As prolonged rupture of the membranes continued beyond data did not allow us to determine if the improved outcomes of
21 days, however, these benefits began to disappear and mortality these infants is due to a treatment effect or a difference in
and death or morbidity began to increase significantly. In all maternal health and reason for delivery. There likely is some 'self-
gestational age groups, survival and survival without mortality was selection' of pregnancies that are able to progress to greater
lowest in those infants where PPROM was >28 days. gestational age that partially explains the differences in outcomes.
Supplementary Information accompanies the paper on the Journal of Perinatology website (http://www.nature.com/jp)