|Client directed care||Practitioner directed care|
|Focused on the person as a whole being||Focused on symptoms|
|Care provider is an expert in health||Care provider is an expert in pathology|
|Care provider guides||Care provider treats|
|Care is woman centered||Care is institution centered|
|Care is natural||Care is allopathic|
|Care is low tech||Care is high tech|
|Birth is safe||Birth has risk|
CHAPTER TWO: Review of the literature
Home birth with midwives is a safe, satisfying, cost effective option for women (ACNM, n.d.). Due to careful risk screening and judicious use of medical technology, the risk of experiencing complications is low (Anderson & Murphy, 1995). In approximately 5% to 10% of all pregnancies, however, premature rupture of membranes (PROM) does occur, and 60% to 80% of these take place at term (Duff, 1996; King, 1994). The length of time allowed between PROM and medical intervention is controversial due to inconsistencies within the literature. As a result, definitive clinical guidelines have not been determined. Since midwifery philosophy includes a belief in the use of intervention only with clear medical indication, home birth midwives are currently handicapped by overly conservative PROM management community standards (Keirse, Ottervanger & Smit, 1996). In order to ensure the safe, evidenced-based midwifery management of PROM for home birth, a thorough review of the literature is necessary.
The literature is being reviewed in order to develop an evidenced-based clinical practice guideline for PROM management in home birth. Relevant studies were analyzed following an on-line search of MEDLINE and CINAHL, using keywords home birth, home birth and safety, PROM, prelabor rupture of membranes, premature rupture of membranes, and PROM and infection. A total of thirty-seven full-text articles were obtained through the Iowa Methodist Hospital Health Sciences Library. For inclusion in this review, each article was required to be a report of actual research. Ten of the most pertinent studies are discussed.
% in spontaneous labor
Author / Year
|Grade of Evidence|
|Study Design and Validity|
|Anderson & Murphy (1995)|
|N=11,788 home births.|
29 states were represented.
Transfer rate = 8%.
Cesarean rate = 3%.
PROM = 13.3%.
PROM related maternal infection = 19 out of 905 intrapartum transfers (<0.02%).
Neonatal sepsis of unspecified etiology = 5 out of 905 intrapartum transfers (0.04%) with no related deaths.
Cord prolapse = 8 out of 905 intrapartum transfers (0.07%) with no related deaths.
Intrapartum neonatal mortality = 2/1000 (0.2 per 1000 after anomalies excluded).
No maternal deaths.
|Offers important descriptive information regarding CNM attended home birth outcomes. Although this study was not PROM specific, statistical calculations showing PROM related infections were included. Statistical findings show similar outcomes to those in other studies, clearly demonstrating a low risk of infection among the studied home birth population. Findings show relative safety of CNM attended home birth.|
|Retrospective cohort survey||N = 11, 788 intended home births.|
N = 11,592 hospital births.
Mean cost of hospital birth in 1991 = $5382.
Mean cost of home birth in 1991 = $1844.
Intrapartum home birth transfer rate = 8%.
Cesarean rate for home birth = 3%.
Cesarean rate for planned hospital birth = 8.3-26.9%).
Neonatal mortality for home birth = 2 / 1000.
Neonatal mortality for hospital birth = 2.2 /1000.
|The average financial cost of home birth is 68% less than birth in hospital. Home birth also offers significantly lower rates of cesarean section and neonatal mortality. For women transferred to hospital after intended home birth, total costs were similar to planned hospital births, due to less expensive prenatal midwifery care and fewer intrapartum / postpartum interventions.|
This study did not identify PROM and PROM related infections specifically. However, study results do contribute to the data showing home birth as a safe option. Infection rates of non-specified etiology in home birth are similar to rates of infection in hospital birth.
|Retrospective comparative study||N = 803 high risk subjects.|
PROM = 18.6% (<0.3% of the 5487 clinic patients).
Maternal fever = 3%.
Cesarean rate for the high risk CNM births = 17% compared to 21% for national data set.
5 minute Apgar of 7 = 1.7% for CNM high risk births.
5 minute Apgar of 7 = 2.5% for national data set.
|The reported incidence of PROM among the total number of clinic patients is similar to rates reported by other studies.|
CNM management of high-risk women results in higher numbers of spontaneous births, fewer instrumental births, fewer cesarean sections and higher 5-minute Apgar scores as compared to the national data set.
The authors recommend further research regarding the higher incidence of maternal fever in the midwifery sample, stating frequent overall underreporting within birth certificate data.
The study indicates that CNM care of higher-risk women result in favorable outcomes.
|Declercq (1995)||D||Retrospective longitudinal study||N = 82,210 home births in the United States from 1989-1992. CNMs attended 12% of the births, physicians attended 20%, and other midwives 30.5%. The attendants for the remaining 37.5% of home births were not identified. ||The average home birth client is Caucasian, married, older, para >3, lesser educated, and less likely to abuse substances. Home births are more common in the southwestern and western states.|
The health outcomes for babies born at home compared favorably to those born in hospitals.
|Retrospective comparative study||N = 1707 home births.|
N = 14,033 hospital births.
Intrapartum transfers from home group = 7.4%.
Cesarean rate = 1.46% home vs16.46% for hospital group.
Assisted delivery = 2.11% home vs 26.60% hospital.
Perinatal death = 1.0% compared to 1.33% for hospital group.
|This study reflects statistics for a lay midwifery service that includes home birth care of higher risk populations (breech presentation, multiple gestation, higher parity, and fewer prenatal visits). Perhaps due to this reason, the perinatal mortality rate for the home birth population is higher in this study than in other studies. However, the perinatal mortality rate is comparative to that for the hospital group. Despite the inclusion of higher risk subjects, this outcomes for the home birth population are comparatively safe to the hospital birth group.|
|Jackson & Bailes (1995)|
|Retrospective Review of international research||N not defined.|
Looked at an undefined number of international studies on the safety of home birth with trained attendants, stringent screening criteria, and physician consulting opportunities. No statistical tables were provided.
|The studies reviewed, totally tens of thousands of birthing women, consistently demonstrate home birth outcomes to be "at least as good as, if not better than" hospital birth. No study definitively showed that hospitals are safer.|
|Janssen et al (1994)|
|Retrospective comparative study||N = 6944 licensed midwife (LM) out-of-hospital births.|
N = 23,596 physician (MD) attended low-risk hospital births.
N = 14,777 CNM attended low-risk hospital births.
N = 4054 CNM out-of-hospital births. Low birth weights: LM group = 10.9/1000, MD group = 16.5/1000 (RR=0.8, 95% CI 0.6-1.0). CNM in hospital low APARs = 10.1/1000. CNM low Apgars out-of-hospital = 9.9/1000. Neonatal deaths: LM group = 1.7/1000, MD group = 1.0/1000 (RR=1.0, 95% CI 0.7-3.0). Neonatal deaths for CNM in hospital group = 1.6/1000 and for CNM out-of-hospital = 1.7/1000.
|The risk for low birthweight was significantly lower for licensed out-of-hospital midwives than for the physician group, and similar to that for nurse-midwives in either setting.|
No differences were found in 5-minute Apgar scores, neonatal mortality, or postneonatal mortality between licensed midwives and the other groups.
This study suggests that out-of-hospital births attended by appropriately trained midwives (both LM and CNM) may be as safe as physician or CNM attended hospital birth.
|meta-analysis||N = 24,092 home birth women from 6 controlled observational studies.|
No significant differences found for perinatal mortality (OR=0.87, 95% CI 0.54-1.41).
Low Apgar scores were less prevalent for the home birth group (OR = 0.55; 0.41-0.74).
Fewer cesarean sections for the home birth group (0.05-0.31).
No maternal deaths.
|Home birth leads to a lower use of medical interventions. There is no evidence among these studies to support a claim that hospital birth is safer than planned home birth. Home birth is a viable option for low-risk women assisted by a skilled practitioner.|
|Pang et al (2002)|
|Retrospective analysis. Poor validity due to examiner bias (inclusion of statistics from preterm births) and conclusions based solely on birth certificate data.||N = 6133 home births. |
N = 10,543 hospital births.
Initial analysis included births >34 wks. Results: neonatal mortality = 3.5/1000 home births, 1.7/1000 hospital births (RR 1.99, 95% CI 1.06, 3.73).
Secondary analysis was performed for subjects > 37 wks gestation. Results: neonatal mortality =3.3/1000 home births, 1.7/1000 hospital births (RR 2.09, 95% CI 1.09, 3.97).
|The authors concluded that the risk of neonatal death was "almost twice as high for infants born of women intending to deliver at home as for infants born to women delivering in hospitals." |
This study is likely biased. The researchers relied only on birth certificate data which does not indicate planned vs unplanned locations of birth. Furthermore, study results included statistics from preterm births (>34 wks) that are not usually accepted in home birth practices. The possibility that unplanned, higher-risk home births were included cannot be ruled out.
|Grade of Evidence|
|Study design||Sample size and description||Research procedure||Measures used & their reliability & validity||Statistics reported|
(include type of statistic reported & p value/confidence intervals as indicated)
|Study results & midwifery perspective|
|Murphy & Fullerton|
|Prospective descriptive study||N = 1404 intended home births.|
One thousand, two hundred twenty-one (1221) remained eligible at onset of labor.
|All women accepted into one of 29 CNM home birth practices from 1994 through 1995 were enrolled in this study.||Individual CNM practice guidelines were used to determine eligibility for home birth. Uniform data collection forms were developed and pilot tested. Hospital records were reviewed for all women and newborns transferred. Intrarater and interrater reliability were calculated using simple percent agreement, ranging from 92%-96%.||Outcomes were compared using x2 and t test procedures. The probability value was set at .05%.|
PROM = 4/1404 (<1.0%).
Intrapartum transfer = 8.3%. Of these, 5.2% transferred for PROM and all infants remained well. Two neonates were evaluated for sepsis.
Overall fetal/neonatal mortality for completed home births = 1.8/1000.
|Birth can be conducted safely at home, as long as care is provided by qualified CNMs within a system that allows for physician collaboration and hospital transfer, when indicated.|
|Study||Grade of Evidence|
|Study Design and Validity|
|Egarter et al (1996)|
|meta-analysis||N = 657 women from seven clinical trials with PPROM between 23 and 34 weeks gestation. Antibiotic treatment most frequently included IV ampicillin 2gm q6H X 24-48H, followed by oral ampicillin.|
Odds ratio and 95% confidence interval were calculated for each study separately, then for all studies combined. Statistics testing followed x2distribution with N =1 degree of freedom.
Results: Antibiotic therapy reduced the risk of neonatal sepsis by 68% (OR 0.32, CI 95% 0.16-0.65, p=0.001). There was no effect on neonatal mortality (OR 0.92, CI 95% 0.46-1.81). There was no effect on necrotizing enterocolitis (OR 1.27, CI 95% 0.61-2.62).
|The authors of this meta-analysis concluded that antibiotic prophylaxis for preterm PROM does improve neonatal morbidity.|
The use of antibiotics for preterm PROM can be beneficial, especially if the medication regimen allows the pregnancy to continue.
Alteration of intestinal flora can be a potentially serious side effect of broad spectrum antibiotic use. This study found no increase of such side effect.
The effect of prophylactic antibiotic administration in term PROM are not known. Due to the lack of impact on neonatal mortality in this metaanalysis, more study is indicated before prophylactic antibiotics in term PROM can be recommended.
|Ferguson et al (2002)|
|Ex-post facto, case-controlled study. |
Women were recruited for the study on admission to one of three tertiary care hospitals. Populations were matched for gestational age and previous vitamin supplementation. Dietary intake was determined through use of the Block Healthy Habits and History food frequency questionnaire.
This study demonstrates poor validity due to the very small sample size and lack of control over variables.
|N = 46 women with PPROM between 23-35.6 weeks gestation. This group was compared with similarly matched women having intact membranes. |
There were no differences in homocysteine, red blood cell folate, or vitamin B levels of either group. There were no differences in dietary intake. Lower hemoglobin levels were found in women with PPROM (P<0.001). Women with PPROM were also 3 times more likely to have family incomes < $25,000 (OR 3.1, CI 95%, 1.6-6).
|PPROM is associated with lower socioeconomic status. Women with PPROM have lower hemoglobin levels, but show no other differences in nutritional level.|
Midwives provide greater amounts of nutritional education, socioeconomic assessments, and counseling. These preventative midwifery activities may reduce some of the variables associated with PPROM by this study.
|Meta-analysis of 15 controlled trials||N = 2265 women in 15 controlled trials.|
Chorioamnionitis: Reduced with IV antibiotic therapy from summary risk of 26.1% to 15.2%. Oral antibiotic therapy prolonged pregnancy, but did not reduce overall maternal infection.
Neonatal sepsis: Intravenous antibiotics reduce neonatal sepsis (summary risk: 7.0 versus 10.5%, OR: 0.65, P = 0.01).
Oral therapy does not reduce maternal or neonatal morbidity.
|Intravenous antibiotic therapy can reduce the incidence of chorioamnionitis in PPROM.|
Oral antibiotic therapy was not shown to be effective. It may offer benefit to prolongation of pregnancy in PPROM, but does not reduce maternal or neonatal morbidity or mortality.
|Mozurkewich & Wolf (1997)|
|Meta-analysis of 23 controlled trials||N = 7493 women from 23 controlled trials reviewed three term PROM management options: immediate oxytocin induction, conservative management (delayed oxytocin), or vaginal PGE2 induction. |
Cesarean deliveries: Oxytocin vs conservative (OR 1.24, 95% CI 0.89, 1.73). PGE2 vs conservative (OR 0.67, 95% CI 0.34, 1.29).
Chorioamnionitis: Oxytocin vs conservative (OR 0.91, 95% CI, 0.51, 1.62). PGE2 vs conservative (OR 0.68, 95% CI 0.51, 0.91). PGE2 vs oxytocin (OR 1.55, 95% CI 1.09, 2.21).
Neonatal infection: oxytocin vs conservative (OR 0.73, 95% CI 0.47, 1.13). PGE2 vs conservative (OR 1.06, 95% CI 0.67, 1.66). PGE2 vs oxytocin (OR 1.50, 95% CI 0.91, 2.45).
|Results using the Mantel-Haenszel fixed-effects procedure favored oxytocin induction for reduction of chorioamnionitis and endometritis. Results using random-effects procedure showed no statistically significant differences.|
Immediate oxytocin induction does slightly increase cesarean deliveries. Although rates of maternal and fetal infections were higher with expectant management, the overall number of infections is too small to be statistically significant.
From a midwifery perspective, women should be informed of the small statistical differences in various risks, then empowered to choose their preferred option.
|Peleg et al. (1999)|
|Secondary analysis of TERM PROM study.|
Purpose: to identify predictors of cesarean section post PROM at term.
|N = 5028.|
90.1% delivered vaginally compared to 9.9% delivered by cesarean.
The association between the variables and mode of delivery was tested using x2 univariate analysis. Multivariate analysis was then done by step-wise logistic regression to determine the statistically significant (p<0.05) independent predictors of cesarean delivery.
The strongest predictor of cesarean was country of birth (Sweden < Denmark < United Kingdom < Canada < Australia).
Other strong predictors for cesarean delivery were nulliparity (OR 2.81, 95% CI, 1.95 to 4.05); active labor >12 hours (OR 2.78, 95% CI 2.01 - 3.85), previous cesarean (OR 2.75, 95% CI 1.46 - 5.16), and epidural (OR 2.75, 95% CI 1.94 - 3.65).
|Women with PROM were assessed by single sterile vaginal examination for initial diagnosis of PROM. Cervical dilation and effacement were also recorded.|
The duration of PROM and/or mode of PROM management were not associated with increased cesarean section rates, according to this secondary analysis.
However, nulliparity and active labor >12 hours were found to have higher cesarean section rates. Therefore, women with less ripe cervices may need longer durations of labor for vaginal delivery to occur.
Variations in the rates of cesarean deliveries were most affected by country. Different settings and environmental factors, then, were shown to impact this birth outcome.
Midwives offering lower intervention care in the home setting may provide a PROM management option that reduces cesarean section rates without increasing infection risk.
|Grade of Evidence|
|Study design||Sample size and description||Research procedure||Measures used & their reliability & validity||Statistics reported|
(include type of statistic reported & p value/confidence intervals as indicated)
|Study results & midwifery perspective|
|Alcalay et al (1996)|
|Prospective randomized trial||N = 154 women with PROM at term. |
Expectant management (EM) = 80
Oxytocin induction (OI) = 74
|Patients with term PROM > 36 wks gestation were randomly assigned to one of the two groups. Sterile speculum exam was used to diagnose PROM, then no further invasive procedures were performed. Preexisting infection was ruled out on admission to the study through amniotic fluid and blood cultures, and temperature monitoring. The study was conducted in the antepartum ward of an European hospital. |
Daily monitoring included: WBC, oral & rectal temps, and BPP. Full sepsis workup was performed on neonates with suspected infection, and for those with PROM >48 hours.
|Method used included two-tailed Student's t-test and x2testing. P-value of <0.05 was considered significant. Reliability & validity were met.||Mean duration of PROM to delivery: EM group = 27.9H, OI group =14.3 H, P <0.01.|
Spontaneous vaginal delivery:
EM = 93.7%, OI = 79.7%, P <0.05.
Fetal distress: EM = 2.5%, OI = 9.5%, P <0.07.
C-section: EM 2.5%, OI 4.1%, P NS.
Chorioamnionitis: EM 2.5%, OI 5.4%, P NS
(confirmed by culture): EM 1.25%, OI 1.35%, P NS.
|The mean period from PROM to delivery was shorter for the induced group. The mean duration of labor was significantly shorter for the expectantly managed group. One infant in each group experienced clinical evidence of infection, with both recovering fully.|
This study shows evidence that expectant management of term PROM is safe.
|Hannah et al (2000)|
|Secondary analysis of the TERM PROM study.||N = 653 managed partially or completely at home, after initial evaluation in hospital.|
N = 1017 managed in hospital.
|This study did not accurately investigate home management because all subjects received invasive digital examinations in the hospital prior to entry into the study. Some were admitted, then later discharged to home care. Single diagnostic sterile speculum examination in the home along with midwifery home management would prevent patient exposure to such invasive monitoring and potentially more virulent hospital microbes.||Multiple logistic regression analysis was used to determine whether adverse effects of expectant management for PROM at term and patient satisfaction were greater in those managed at home.|
Reliability and validity were met.
|Chorioamnionitis: 10.1% home compared to 6.4% hospital (P=0.006).|
Neonatal infection: 3.1% home vs 1.7% hospital (OR 1.97, CI 95% 1.00-3.90, P = 0.05).
|Authors conclude that home management of PROM may increase the likelihood of some adverse outcomes, while leading to reduced levels of patient satisfaction.|
|Hanna et al (1996)|
|Prospective, multi-center randomized controlled trial.||N = 5041 women with term PROM were assigned to one of four groups: Induction with IV oxytocin, induction with PGE2, expectant management up to 4 days followed by induction by one of the above.||Subjects were recruited from 72 centers internationally. Each was randomly assigned via centralized telephone randomization program. Labor management and monitoring varied according to site, practitioner, and subject allocation.||Clinical chorioamnionitis was defined as maternal temperature > 37.5C on two or more occasions, a single temp >38C, or a maternal WBC > 20,000 cells/mm3. |
Multiple logistic regression analysis was once again used. Reliability and validity were established.
|Outcomes for the groups were similar: |
Neonatal infection: Induction-Oxytocin group = 2%, induction-PGE2 = 3.0%, exp mgmt (oxytocin) = 2.8%, exp mgmt (PGE2) = 2.7%.
Chorioamnionitis induction-oxytocin to exp mgmt (oxytocin): 4.0% vs 8.6%, P<0.001.
|Women with PROM in this study appeared to have lower risk for infection when induced, as compared to expectant management. The author recommends induction. Women were also reported to have reduced satisfaction when managed at home, possibly due to the requirement for frequent trips to the hospital for testing.|
Expectant management with midwives may reduce patient risk and dissatisfaction in a home birth population due to the opportunity for 100% home care.
|Hannah et al (1997)|
|Secondary analysis or TERM PROM study.||N = 516 total GBS positive women.|
N = 121 induced with oxytocin
N = 133 induced with PGE2
N = 149 expectant, then oxytocin
N = 115 expectant, then PGE2.
|The rates of maternal antibiotic use, clinical chorioamnionitis, neonatal infection, neonatal antibiotic use, and NICU admission were compared for each of 4 groups.||Multiple logistic regression analysis was used to determine the effect of induction of labor on neonatal infection in GBS + women with PROM.||When treatment groups were combined to assess for neonatal infection, the odds ratio was 3.08 (p <0.001) for the induction-oxytocin and expectant-oxytocin groups, 3.31 (p <0.001 for induction-oxytocin and induction PGE2 groups, and 4.45 (P < 0.001) for the induction-PGE2 and expectant-PGE2 groups.||GBS status was predictive of neonatal infection for those induced with PGE2 and for those in the expectant management groups. No increase in GBS infection was seen with the oxytocin induced group.|
The duration of membrane rupture was not a significant predictor of neonatal infection.
|Kappy et al (1982)|
|Prospective controlled trial|
|N = 150 hospitalized women with PROM >36 wks gestation.|
N = 112 expectant management group
N = 28 early induction group
|Diagnosis made through sterile speculum examination. No SVE allowed. Women with Bishops > 8 were offered induction by oxytocin. All others were managed expectantly until labor ensued or signs of infection developed.||All clients managed expectantly had a daily CBC with diff, temp a4H, daily evaluation of uterine tenderness and daily NST. Clients with signs of chorioamnionitis were treated with IV antibiotics, followed by oxytocin induction.At delivery, all neonates received appropriate bacterial cultures. Microscopic evaluation of placentas were done for those with suspected chorioamnionitis.||Duration of expectant management was not time limited. Spontaneous labor: 87% within 48 hours, 3.6% went between 7 and 18 days.|
Cesarean rate: 39% for induction group vs 12% for expectant mgmt group (P = < 0.01).
No cases of culture positive sepsis, even for those with 7-18 day PROM.
No maternal or neonatal deaths.
|Extended time periods for the expectant management of term PROM with an unfavorable cervix is safe and may reduce the incidence of cesarean section without an increase in infection.|
|Keirse et al (1996)|
|Meta-analysis of controlled trials.|
Inclusion criteria: Trials were limited to random controlled studies that took place over the same period of time.
|N = 2441 subjects from a total of 15 studies.||Meta-analysis||Reliability and validity were established.||Data show a 30% increase in cesarean section rate for PROM with active induction compared with expectant mgmt (OR 1.3, 95% CI).|
Data for infectious morbidity within the expectant management groups ranged from 0-28% for maternal and from 0-16% for neonatal.
Neonatal infectious morbidity did occur less frequently in the active group (N=3/843) than in the expectant groups (N=11/805) (OR 0.37, 95 CI: 0.19 to 0.73).
|The studies examined reveal a lack of consistent methodology, wide variations in diagnostic criteria, and large differences in the allotted time limits for expectant management.|
There is a high potential for bias within the studies examined because clinical diagnosis of infection was usually made by persons who knew the type of management option applied. Therefore, more cultures may have been collected on neonates with longer durations of PROM, resulting in higher rates of diagnosis.
Overall, the data shows little difference in maternal morbidity between active and expectant management policies. An expectant policy may increase the risk of neonatal morbidity, but the effect (if any) is small and may be impacted by observer bias. PROM by itself should not be an indication for induction. Mothers should be offered a choice among the options.
|Kenyon et al (2001)|
|Randomized, blinded, multicenter trial||N = 4826 women with PPROM <37 weeks gestation.|
Women were randomly assigned to one of four treatment groups:
|Participants took the trial medicine orally, QID x 10 days or until delivery.|
Group assignment was randomly selected by computer, and blinded by sequential numbering system.
Data were collected three times using entry forms and outcome forms: At entry into study, at discharge of the mother after delivery, and at death or discharge of the baby. Endpoint for data collection was hospital discharge.
|All data were entered and verified by two people on two separate occasions. Random samples of pediatric data (10%) were checked against source documentation for accuracy.|
Statistical analysis was done by use of the Ztest; two-sided p-values were cited throughout. Unpaired ttest was used for normal distributions; Mann-Whitney test otherwise.
Reliability and validity were well established.
|There were statistically significant differences between the erythromycin only group & placebo group in regard to positive blood cultures (5.3% vs 7.4%, p = 0.03). No differences in necrotising enterocolitis (NEC) were found.|
For Augmentin group: positive blood cultures = 6.8%. NEC = 4.1% vs 2.7% placebo only (p=0.08).
Combined antibiotics: positive cultures = 7.0%, p =0.27. NEC = 3.5%, p=0.23.
There were no appreciable differences for maternal outcomes.
|Oral erythromycin alone may reduce the incidence of neonatal infection in PPROM without an increase in necrotizing enterocolitis.|
Use of oral Augmentin increases necrotizing enterocolitis without a significant decrease in rates of PROM neonatal sepsis.
For midwives, use of oral erythromycin for home management of PROM may offer some benefit. But, additional study is first needed in order to verify results of this ORACLE trial.
|Ladfors et al (1996)|
|Prospective, randomized controlled trial||N = 1385 women with PROM between 34 and 42 weeks.|
N = 502 to early induction group.
N = 510 to late induction group.
|Women were randomized to early induction at 24 hours or late induction at 72 hours after PROM. Digital examinations were avoided until onset of active labor. All labors experienced internal fetal scalp monitoring.|
Labor was induced with oxytocin if no spontaneous contractions occurred at the end of the allotted time period, or if signs of chorioamnionitis or fetal distress were detected.
|Outcome data tested for included: frequency of spontaneous deliveries, operative deliveries, maternal and neonatal infection, and Apgar scores.|
Data were analyzed using Duncan's multiple range test or Tukey's studentised range test HSD. A P-value of < 0.05 was considered significant.
Data were further analyzed for nullipara and multipara sub groups.
|80% had spontaneous contractions within 26 hours of PROM. After 50 hours, only 10% were still awaiting labor.|
Chorioamnionitis was diagnosed in 2 women with early induction and 4 women with late induction (0.8% overall). There were no differences in the rates of neonatal infection among the groups.
|Expectant management for up to 72 hours offers benefit of reduced cesarean section rates without a statistically significant rise in maternal or neonatal infection, especially for nulliparous women.|
|McCaul et al, 1997)|
|Prospective, randomized, controlled trial||N = 96 women with PROM and an unfavorable cervix between 36-42 weeks gestation.|
Three treatment groups:
N = 25 PGE2 induction
N = 35 oxytocin induction
N = 31 expectant management
|On admission to hospital, PROM was diagnosed by microscopic examination of pooling of amniotic fluid. Fetal monitoring was used to exclude persistent uterine activity and/or fetal distress. |
Inclusion criteria: women aged 16-35 years, cervical dilation < 3cm, effacement < 75% determined by SVE, cephalic singleton fetus, PROM < 24 hours duration.
Random selection was via computer. Daily NST and AFI were used to monitor the expectant group.
|Chi-square, Fisher's Exact Test, and ANOVA were used in statistical analysis A P value of < 0.05 was considered significant.|
Sample size was small. Reliability was well established, but validity is questionable.
|Study results showed no differences in neonatal infection, maternal febrile morbidity, or in cesarean section. There was a 13% incidence of variable decelerations in the expectant management group, but no associated adverse outcomes.||This study failed to establish PGE2 as superior to oxytocin induction. The authors demonstrated bias through via their interpretation of isolated EFM data. The authors felt that evidence of variable decelerations signifies increased risk of cord accident for the expectant management group. Longer hospital stays during the latency period were also identified as potential client deterrents, without evidence of such.|
Expectant management was not shown to increase morbidity or mortality rates. Concern about client satisfaction provides rationale for home expectant management with midwives.
|Naef et al (1998)|
|Prospective randomized controlled trial||N = 120 women between 34 and 36 6/chorioamnionitis.|
N = 57 immediate induction with oxytocin
N = 63 expectant management
|Sterile speculum examination was used to confirm pooled amniotic fluid, along with microscopic confirmation. No SVE were performed.|
Blinded computer-generated random assignment was utilized. Patients were monitored by continuous EFM for 2-4 hours, then those in the expectant management group were transferred to the antepartum unit and monitored every 8 hours. GBS prophylaxis was done for all GBS positive patients.
The active management group received oxytocin IV until active labor was established.
Neonatologists were not blinded to perinatal clinical course.
|Outcomes measured included: chorioamnionitis, type of delivery, birth weight, Apgar scores, RDS, NICU admission, and sepsis.|
Statistical analysis of group differences was done through use of the x2 test and Fisher's exact test for discrete data; and the Student t test for continuous data. P < 0.05 was accepted as significant.
Reliability was established. The small sample size threatens applicability of findings.
|Chorioamnionitis was higher for those managed expectantly (16% vs 2%, p - 0.007).|
Neonatal sepsis was slightly more common in the expectant management group (0% vs 5%, p = 0.151). All recovered completely.
Cesarean section was slightly lower in the expectant management group (7% vs 5%, p = 0.444)).
There were no other significant differences in outcome measures.
|The authors recommend active induction of PROM labor due to this study's finding of higher chorioamnionitis in the expectant management group. However, the study's percentage of difference does not concur with the rates found in previous studies, perhaps due to overdiagnosis and small sample size used. This study cannot be used as justification for immediate induction.|
|Seaward et al. (1997)|
|Secondary analysis of TERM PROM study to determine predictors of clinical chorioamnionitis.||N = 5028.|
Spontaneous labor occurred in 2225 (44%).
|Internal fetal monitoring was used in 1541 (30%) of patients.||Univariate analysis and multivariate analysis were used. X2and Fisher's exact test were used to test categoric data. Student's ttest was used for data not normally distributed. Logical regression analysis was performed. All p values were based on two-tailed tests with p<0.05 identified as statistically significant.||Variables found to be associated with chorioamnionitis include: number of digital vaginal examinations >8, 7 to 8, 5 to 6, 3 to 4, (OR 5.07, 3.80, 2.62, 2.06); duration of labor >12, 9 to < 12, 6 to < 9, (OR 4.12,2.94,1.97); meconium-stained fluid (OR 2.28); parity of 0 (OR 1.80); time frame from PROM to active labor >48, 24 to <48 (OR 1.76, 1.77); and GBS colonization (OR 1.71).||This study demonstrates that vaginal examinations lead to significant increases in the rates of chorioamnionitis.|
The authors claimed the IFM was not associated with chorioamnionitis, but no specific data showing multiple logistic regression analysis for this variable were included in the report. With 30% of women exposed to IFM, the impact of this invasive procedure needs to be further addressed.
Avoidance of vaginally invasive iatrogenic procedures seems prudent in light of these findings. Midwives providing expectant management of PROM in the home setting may be the best option, since home midwifery care is known for reduced rates of intervention.
|Ottervanger et al. (1996)|
|Prospective, randomized controlled trial||N = 123 hospitalized women with PROM between 37 and 42 weeks gestation.|
N = 62 expectant management up to 48 hours
N = 61 immediate oxytocin induction
|Women presenting to the hospital with PROM were randomly assigned to one of the two study groups. PROM was diagnosed by history, loss of fluid and occasionally sterile speculum exam. Cervical swabs were taken at first vaginal examination. Cultures were taken from the baby at delivery, and a cord section was sent for pathological examination.|
Those in expectant care were monitored with EFM, daily WBC, temp qid. Ambulation was not restricted. After 48 hours with no labor, IV oxytocin induction was offered.
|Randomization was conducted in blocks of 20 with an interim analysis after 120 patients had been entered. Outcome measures reviewed included: cesarean section, operative delivery, use of analgesia, infectious morbidity of mother and/or newborn, and infant hospitalization after discharge.|
Women were excluded if they presented with obstetrical problems such as infection, abnormal FHT, or HTN. Random assignment was performed by means of sealed envelopes.
Reliability and validity were met.
|Fifty-eight (one-sided) to 71 (two-sided) women were required per group to demonstrate statistical significance (p<0.05) with a power of 0.08.|
55% went into spontaneous labor within 24 hours, an additional 16% within the next 24 hours, for a total of 81% by study endpoint.
Cesarean sections: 15.0% induction group vs 5.2% expectant management group.
Instrumental delivery: 16.4% induced vs 6.4% exp mgmt (OR 9.9 95% CI -1.2 to 21.1).
Positive initial cervical cultures (timing unspecified): 15 induced vs 16 exp mgmt.
Maternal clinical infection: 1.6% induced vs 3.2% exp mgmt (OR -1.6, 95% CI -7.0 to 3.8).
Positive culture of gastric aspirate = 37, but none showed clinical sepsis.
|No differences were found in the rate of infectious morbidity, use of antibiotics, or positive cervical cultures.|
Expectant management for up to 48 hours can decrease cesarean rates without increasing infectious morbidity.
Further study is needed using longer durations of expectant management to determine maximal duration of benefit.
|Shalev et al (1995)|
|Prospective, controlled trial||N = 566 women with PROM between 37 and 42 weeks.||PROM was confirmed by a single sterile speculum examination and nitrazine testing. Existing infection was ruled out through blood and amniotic fluid cultures. No vaginal examinations were permitted.|
Women were assigned to either a 12-hour or 72-hour expectant management plan. All women were placed on bedrest during this time. After the designated time period, women not in labor were induced with IV oxytocin. At birth, cultures were taken from the fetal membranes, placenta, and newborn oral cavity.
|Results were analyzed using x2 and Yates correction and Student t test methods. |
Reliability and validity were met.
|83% in the 72-hour group entered spontaneous labor, compared to 43% in the 12-hour group.|
There were no significant differences in the number of uterine infections or in the number of cesarean sections.
There were no differences in neonatal Apgar scoring or in rates of neonatal sepsis.
|72-hour periods of expectant management for PROM can produce higher rates of spontaneous labor without increasing infectious morbidity or mortality. Women may be more comfortable during this time in their own homes.|
|Sperling et al. (1993)|
|Prospective randomized controlled trial||N = 362 women with PROM > 36 weeks were invited into the study. |
N = 62 induced at 6 hours after PROM
N = 62 induced after 24 hours of PROM
238 declined participation due to desire for autonomy over management option and/or desire for "natural birth."
|PROM was confined by history and by midwife performed SVE. WBC and mid-stream urine culture were taken at admission to study.|
Women were randomized to one of two IV oxytocin induction groups: Early induction at 6 H after PROM or later induction at 24 H after PROM. Pulse, temp and FHT were recorded every 6 hours. Labor monitoring was performed by IFM. Vaginal examinations were allowed, but minimized until active labor. After delivery, the placenta and membranes were sent to pathology. Newborns skin cultures were taken for PROM > 24 H.
|Statistical analysis used the Mann-Whitney U-test for variables in labor, and the Chi-square-test (x2, two-tailed) was used for frequency of obstetrical interventions.|
Jonckheere-Terpstra Test was used to evaluate rates of chorioamnionitis in relation to duration of PROM and length of labor.
P < 0.05 was identified as statistically significance.
|N participants = 124.|
40% entered spontaneous labor within 12 hours. 68% entered spontaneous labor within 24 hours.
Two primiparas developed chorioamnionitis at 14.7 hours and 18.9 hours respectively. One of the newborns was treated with antibiotics for fever and tachypnea, but cultures were negative.
Spontaneous vaginal delivery occurred for 71% in the early induction group and 69% in the late induction group.
Cesarean sections rates: 10% in the early induction group (all primiparous) and 13% in the late induction group (OR 0.72. 95% CI 0.33 to 1.60, p < 0.05).
Intrapartum infection: 0% early induction, 3% late induction (all primiparous).
None of the newborns had positive cultures.
|The authors concluded that increasing time spans from PROM to delivery lead to increased maternal infection. However, the study sample and time frame are too small to justify such a conclusion. Furthermore, invasive vaginal examinations and internal fetal monitoring were universally applied. These two practices have been shown in later studies to increase infection rates due to ascending bacteria. The increase found in this study is likely iatrogenic.|
This study shows that invasive monitoring procedures may increase the rates of PROM related infection, and are thus preventable.
|Theunissen & Van Lierde (1989)|
|Prospective quasi-experimental design||N = 215 women with PROM between 24 and 36 weeks gestation.|
Each neonate born after PROM was compared with a premature neonate matched for the same birth age and year of birth.
Statistical analysis was done using Chi-square tests of independence.
|PROM was diagnosed via history and sterile speculum examination with microscopic examination of pooled fluid for ferning. GBS cultures, US, and EFM were done on admission. Women were confined to bed rest in hospital. The treatment group received IV ampicillin or erythromycin and tocolysis.|
Maternal temperature, eval of uterine tenderness, EFM bid, CBC with diff qod, and weekly cervical cultures were utilized.
|Outcome measurements included: gestational age at birth, birth weight, Apgar scores, neonatal infection. and neonatal death by age 1 week.|
Neonatal infection was diagnosed by central culture (blood, urine, CSF).
|Neonatal infection after maternal chorioamnionitis: 23.7%; after absence of chorioamnionitis: 5.6% (p < 0.0001).||This study found no relationship between the latency period and rates of chorioamnionitis or neonatal infection. Even prolonged latency > 7 days did not increase the risk of chorioamnionitis or neonatal sepsis.|
Clinical suspicion of chorioamnionitis does increase the risk of neonatal infection.
Therefore, midwives finding no clinical evidence of chorioamnionitis can safely continue expectant management, even > 7 days after PROM.
Five to ten percent of all term pregnancies will experience premature rupture of membranes. Fewer than 3% of these will develop infectious morbidity as long as invasive monitoring procedures, such as vaginal examinations, are avoided. Rushing to immediate labor induction for PROM unnecessarily interrupts the processes that serve to prepare the body for labor and carries risks that include protracted labor, instrumental delivery, and cesarean section. In the home setting, women are better able to avoid iatrogenic risks and may find comfort in opportunities to continue their normal routines. Watchful waiting in an absence of clinical signs of infection or fetal distress is a safe option for low risk women under certain conditions. Both immediate induction and expectant management options carry potential risks and benefits. Therefore, each woman should be fully informed of the choices available so that she can be free to choose the option that best coincides with her individual health concerns, personal value system and goals. This evidenced-based clinical practical practice guideline is a tool that can guide informed consent and standardize the practice of the expectant management of PROM in home birth.
C. Outline of EB-CPG/clinical algorithm
Does client seek home birth with a midwife?
|Duration of PROM||Evidence of abnormal fetal heart patterns|
|Type of delivery||Apgar scores at 1, 5, and 10 minutes|
|CBC with diff at 24 hours after delivery||T-P-R birth to 48 hours|
|T-P-R during labor to 48 hours after delivery||Sepsis workup, if performed|
|Other clinical signs of infection||Other clinical signs of infection|
|Risks include: Failed induction if the cervix does not respond. This can lead to increased risks of vacuum or forceps delivery OR increased cesarean section.||Risks include: No statistically significant increase in infection for the first 3 days, as long as no vaginal penetration (examinations, intercourse, etc) is done. There is also a very small risk for cord problems, prolapsed cord, or poor positioning of the fetus.||Risks include: Castor oil causes diarrhea and can be uncomfortable. Herbs and homeopathic medicines may stimulate contraction similar that feel similar to natural labor. There are no known risks to the baby as long as proper dosing is followed.|
|Benefits include: Lowest risk (1%) for infection as long as minimal or no vaginal examinations are performed.||Benefits include: Less intervention means less side effects from drugs or surgery. Home birth is still possible as long as no signs of infection or other problems occur.||Benefits include: Labor may start in a more natural way, especially if your body is ready to respond. You avoid the more intense side effects of drug induced labor. If natural encouragement does not work within 4 hours, the option can be stopped to give you adequate rest.|