Preterm prelabour rupture of the membranes before 28 weeks: Better than feared outcome of expectant management in Africa

European Journal of Obstetrics & Gynecology and Reproductive Biology 126 (2006) 186–192 www.elsevier.com/locate/ejogrb

Preterm prelabour rupture of the membranes before 28 weeks: Better than feared outcome of expectant management in Africa Chantal J.M. Stewart a,*, Shaun K. Tregoning a, G. Moller b, H. Wainwright c a

Department of Obstetrics and Gynaecology, University of Cape Town Medical School, Anzio Road, Observatory 7925, Cape Town, South Africa b Department of Paediatrics, New Somerset Hospital/University of Cape Town, Cape Town, South Africa c Department of Anatomical Pathology, University of Cape Town, Cape Town, South Africa Received 8 December 2003; received in revised form 5 July 2005; accepted 22 August 2005

Abstract Objective: To document the prognosis after conservative management of patients with membrane rupture at gestations less than 28 weeks. Study design: Prospective observational study of 78 women with confirmed membrane rupture at less than 28 weeks gestation, managed conservatively. Antibiotics were given from the time of membrane rupture till delivery. Patients were delivered if clinical infection supervened, there was fetal compromise, spontaneous labour ensued or if the pregnancy continued to 34 completed weeks gestation. Results: The mean gestational age at membrane rupture was 23.3  3.17 weeks (16.5–27.8) and the median 24 weeks. Mean latency period was 24.1  29.1 days (1.5–154) with a median of 12.5 days. Eight women (10%) delivered between 24 and 48 h, 25 (32%) within 7 days and 55 (70%) within 1 month. Of note is that 23 patients (30%) had latency periods of greater than 1 month. The mean gestational age at delivery was 26.7  3.92 weeks. Overall of the 78 women there were 81 fetuses delivered, of which 35 (43%) survived. Survival was related to latency period, birth weight and gestational age at delivery. Sixteen women (20%) developed chorioamnionitis. There was no increase in the incidence of clinical infection with increasing latency period. Compression limb abnormalities occurred in 17% of neonates and lung hypoplasia in 18%. Conclusion: Conservative management of patients with very preterm prelabour membrane rupture offers a survival rate of at least 40% with no serious complications in a study of 78 women. # 2005 Elsevier Ireland Ltd. All rights reserved. Keywords: Preterm prelabour membrane rupture; Survival; Infection rate; Latency period

1. Introduction Preterm prelabour rupture of the membranes (<37 weeks) occurs in about 1% of pregnancies [1,2] and in about 0.65% of pregnancies before viability [3]. The majority of patients presenting with this problem labour within 24 h. The remainder pose a significant obstetric dilemma. When rupture of membranes occurs in pregnancies before viability, controversy still abounds as to the optimal management. A traditional view involves immediate * Corresponding author. Tel.: +27 21 404 6032/671 8460/83 7188026; fax: +27 21 4486921. E-mail address: [email protected] (Chantal J.M. Stewart).

termination of the pregnancy to prevent potential maternal and neonatal morbidity and mortality. In addition to the maternal complications of chorioamnionitis and puerperal infectious morbidity, there are thought to be high risks to the fetus (abortion,cord prolapse, infectious sequelae, hyaline membrane disease) [4] If the rupture of membranes is prolonged, features of fetal compression such as limb abnormalities and lung hypoplasia can occur. There is no consensus regarding management of very preterm (less than 28 weeks) prelabour rupture of the membranes (VPPROM) as indicated by a survey of 52 obstetric units in Sweden [5]. Following this survey it was suggested that, if membrane rupture occurred before 20 weeks, the pregnancy should be terminated, whereas

0301-2115/$ – see front matter # 2005 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.ejogrb.2005.08.016

C.J.M. Stewart et al. / European Journal of Obstetrics & Gynecology and Reproductive Biology 126 (2006) 186–192

between 20 and 25 weeks treatment should be individualised. More recently, several studies have looked at the outcome for mother and infant with expectant management [6– 12,3,13,14]. Survival rates ranged between 22% and 62% with maternal infectious morbidity between 37% and 77%. From these studies it is impossible to draw any firm conclusions about the incidence of pulmonary hypoplasia as many of the studies displayed a paucity of postmortem data, but rates of 9%–60% have been reported following expectant management of women with VPPROM [2,15–17]. A confounding factor in many of the studies is that they are retrospective with a variation in their use of antibiotics, steroids and tocolytics. Most of these studies are from First World countries. There are no studies from Africa. The purpose of this study was to report the experience of expectant management of cases with VPPROM in Africa. With a higher incidence of pregnancy related sepsis overall, as well as the limited neonatal facilities with a paucity of ventilators, little access to surfactant and overcrowded neonatal intensive care units, it was feared that expectant management would carry an unacceptably high risk with poor outcome. However, there was no data to support the assumption. The data from this study is important to help counsel patients in this location about the management options.

2. Study design We followed prospectively 78 women presenting with prelabour rupture of the membranes for more than 24 h, at gestations less than 28 weeks in a 212-year period. The women were managed in two of the perinatal referral centres in CapeTown-New Somerset Hospital and Groote Schuur Hospital. The exclusion criteria were: (1) Labour within 24 h of rupture of the membranes. Patients who laboured within 24 h of membrane rupture were not referred to the author and were not included in the study. The aim of the study was to assess long term complications and institute a management protocol for these patients with prolonged membrane rupture which included expectant management, close monitoring and antibiotic usage. For these reasons, only those with membrane rupture exceeding 24 h were included. (2) Cervical cerclage had been undertaken. No sutures were placed after the membranes had ruptured. (3) There was clinical evidence of chorioamnionitis. (4) There was significant vaginal bleeding immediately prior to or associated with the rupture of membranes. (5) Patient request for termination of the pregnancy. Gestational age was determined by menstrual history and/or ultrasonography. A sterile speculum was passed on admission to document the drainage of amniotic fluid and to obtain cervical cultures. Digital examinations were avoided.

187

Rupture of the membranes was confirmed by a combination of the pH test and the presence of ferning in fluid obtained from the vagina. If no fluid was obtained, an ultrasound amniotic fluid index of five or less in the presence of a suggestive history was taken as diagnostic. Initial assessment included temperature, pulse and fetal heart rate monitoring. A baseline leucocyte count and Creactive protein were obtained. Expectant management included admission to hospital and 8 hourly monitoring for any evidence of labour or infection. The women remained in hospital throughout the pregnancy. Oral amoxycillin 250 mg t.d.s. and Metronidazole 400 mg t.d.s. were administered to all women until delivery. If cervical cultures yielded Group B streptococcus, intravenous ampicillin was administered in labour. Prophylactic ampicillin was also administered prior to any caesarean section. Tocolytics were never used in these cases. Women were delivered if spontaneous labour occurred. Serial ultrasound examinations were performed every week in the cases which continued with the pregnancy sufficiently to allow this. Fetal growth was assessed and reaccumulation of amniotic fluid documented. Cardiotocographic monitoring was commenced at 28 weeks gestation, this being the gestation where the fetus is regarded as viable in our unit. Intramuscular dexamethasone injections 12 mg in two doses 24 h apart was given to patients who reached 28 weeks and repeated at 33 weeks. They were withheld if there was any suspicion of chorioamnionitis. In our neonatal unit, survival rates below 28 weeks gestation are less than 50% due to the lack of access to surfactant and a shortage of ventilators. Caesarean sections were therefore not performed in women of less than 28 weeks gestation or estimated fetal weight of less than 900 g. Labour was electively induced at 34 weeks if no other complications had occurred. A diagnosis of chorioamnionitis was made if: (1) temperature of greater than 37.5 8C on two or more occasions, (2) at least two of the following findings were presentmaternal tachycardia >100/min, fetal tachycardia >160/min, offensive vaginal discharge, persistent uterine tenderness or irritability, and (3) temperature of more than 37.2 8C plus one of the above findings. If chorioamnionitis developed, intravenous antibiotics were administered and labour was induced. The definition of postpartum endometritis was made using similar criteria, except that criterion three was replaced with the finding of a temperature of more than 38 8C on any occasion barring the first day after delivery. HIV screening was not available at the time of this study so the HIV status of the patients is not known. Neonatal morbidity was assessed by the paediatric staff. The presence of hyaline membrane disease, sepsis, compression abnormalities, a clinical course suggestive of lung hypoplasia and any complications of prematurity were documented. Neonatal infection was diagnosed by a positive urine test for Group B streptococcus, pus cells in the gastric aspirate, a positive blood culture or clinical features of infection. The criteria used were very broad in order to include all possibly infected neonates and so probably

188

C.J.M. Stewart et al. / European Journal of Obstetrics & Gynecology and Reproductive Biology 126 (2006) 186–192

resulted in overdiagnosis of this complication. Postmortems were performed for 20 (49%) of the stillbirths and neonates who died. Placental histology was obtained in all patients who consented to this. Histological chorioamnionitis was diagnosed by the presence of polymorphonuclear leucocytes in the chorion and amnion or the cord. The diagnosis of lung hypoplasia in babies who died was made histologically according to the Wiggelsworth criteria of lung:body weight ratio of less than 0.015 at gestations under 28 weeks and less than 0.012 above 28 weeks. In the survivors, clinical criteria used were small chest circumference, pulmonary hypertension, air-leak syndromes, bell-shaped chest on X-ray and ventilation at high pressures. Results were analysed by the x2 test to compare frequencies, the Mann–Whitney U-test for non-parametric data, and Spearman’s statistic for correlation. Differences were considered to be significant for p-value <0.05.

Table 2 Indications for delivery

3. Results

3.2. Infection

3.1. Pregnancy outcome

Of the 78 women, 16 (21%) developed clinical evidence of chorioamnionitis but there was no increase in the incidence of chorioamnionitis with increasing latency period as shown in Fig. 1. In fact, in the group of women whose pregnancies continued for more than 14 days, 30 (83%) never had clinical evidence of chorioamnionitis. Postpartum endometritis developed in 13 women (17%) and 6 (46%) of these had chorioamnionitis prior to delivery. The total rate of infection was thus 30%. Placental histology was performed in 54 cases of which 42 (78%) had histological chorioamnionitis. Twelve (22%) had no clinical signs thereof. Of the women with clinical evidence of infection, 19 had placental histology performed and 17 (89%) had histological chorioamnionitis. Eleven of the 16 women (69%) with clinical chorioamnionitis had induction of labour; the remainder laboured spontaneously. Five of the seven women (70%) who developed postpartum endometritis without preceding

The study group included 78 women aged 16–41 years with a median age of 27 years; 17 (22%) were primigravid. Thirty-five women (45%) had experienced at least one previous pregnancy loss. Of these, 30 women (38%) had had one or more miscarriages, 14 (18%) had had at least two miscarriages of which 14 occurred in the midtrimester, 1 an ectopic pregnancy and 1 a preterm delivery. There were five women with previous stillbirths or neonatal deaths. The gestational age at the time of rupture of membranes ranged from 16.5 to 27.8 weeks. The mean was 23.3  3.1 weeks and the median 24 weeks (Table 1). The length of time from rupture of membranes to delivery (latency period) ranged from 1.5 to 154 days. The mean was 24.1  29 days and the median 12.5 days. Eight women (10%) delivered between 24 and 48 h of membrane rupture, 25 (32%) within 7 days and 55 (70%) within 1 month of rupture of membranes. Of note is that 23 women (30%) had

Indication

No.

Vaginal

Caesarean

Spontaneous labour Infection Fetal distress Cord prolapse Elective induction at 34 weeks IUD Abruptio placentae Other

50 16 3 2 2 1 2 2

46 14 0 1 1 1 1 1

4 2 3 1 1 0 1 1

latency periods of longer than 1 month. There was an inverse correlation between gestational age at membrane rupture and latency period (r = 0.4; p = 0.0003). The mean gestational age at delivery was 26.7  3.9 weeks with a median of 26.9 weeks. The reasons for delivery are tabulated in Table 2. In total, 13 women (16%) were delivered by caesarean section and the rest vaginally.

Table 1 Distribution of gestational age at onset of premature rupture of membranes Completed week of gestation

No. of patients

16 17 18 19 20 21 22 23 24 25 26 27

2 4 2 5 4 1 7 10 15 7 12 9

Fig. 1. Chorioamnionitis and latency period.

C.J.M. Stewart et al. / European Journal of Obstetrics & Gynecology and Reproductive Biology 126 (2006) 186–192 Table 3 Organism

No. of patients

Group B streptococcus Gardnerella sp. Microaerophilic streptococcus Trichomonas vaginalis Candida sp. Other streptococci Escerichia coli Staphylococcus aureus Bacteroides sp. Proteus mirabilis Bacillus fragilis

6 3 2 2 2 1 2 2 1 2 1

189

Twenty-one women (35.6%) had positive cultures on high vaginal swab. Organisms cultured are listed in Table 3. Of the women with clinical chorioamnionitis, 46% had positive cultures compared with 32.6% who had no evidence of infection, which was not significantly different ( p = 0.39). 3.3. Perinatal outcome The study group included three sets of twins for a total of 81 deliveries. There were 21 stillbirths and 25 neonatal deaths with an overall perinatal mortality rate of 568 per 1000. All stillbirths had birthweights of less than 800 g. Of the neonatal deaths, 12 neonates died in the delivery room or within 24 h of admission to the neonatal intensive care unit from extreme prematurity. Four infants died of complications of prematurity (respiratory distress syndrome, intraventricular haemorrhage), three died with a diagnosis of pulmonary hypoplasia, three had asphyxia (two secondary to cord prolapse and one following vaginal breech delivery), one from sepsis, one from unexplained pulmonary hypertension and one infant had multiple congenital abnormalities. No babies born at less than 25 weeks gestation survived. Only four of the infants born between 25 and 26 weeks survived, and seven out of 15 babies (47%) born between 26 and 28 weeks gestation survived. Table 4 shows the survival rate in relation to gestational age at membrane rupture. The neonatal morbidity in the survivors is shown in Table 5. Forty-six percent of the neonates had evidence of sepsis, although these signs were often mild with only 25% having clinical features of infection.

chorioamnionitis had delivered vaginally and only one by caesarean section. Two of the 13 women (15%) who had caesarean sections had postpartum endometritis compared with 11 of the 65 women (17%) delivered vaginally ( p = 0.93). Operative intervention, therefore, did not appear to increase the risk of maternal infective morbidity. There was no serious infectious morbidity in any of the women in terms of septicemia, surgery for pelvic infection or septic thrombophlebitis. Three patients did, however, need readmission for evacuation of septic retained products of conception. Dexamethasone was administered to 33 women prior to delivery using the regime previously described. These were only given at gestational ages of 28 weeks or more, as the remainder were considered pre-viable. There was no increase in the incidence of clinical chorioamnionitis in the steroid group compared with the group who were not given steroids (18% in both groups). Table 4 Perinatal outcome

Gestational age at membrane rupture (weeks)

Mean gestation at rupture of membranes Stillbirth Neonatal death Perinatal survival

<20 (n = 14)

20–24 (n = 22)

24–26 (n = 24)

26–28 (n = 21)

All (n = 81)

17.8 7 4 3 (21.4%)

22.2 8 6 8 (36.4%)

24.4 6 10 8 (33.3%)

26.7 0 5 16 (76.2%)

23.3 21 25 35 (43.2%)

Table 5 Neonatal morbidity Gestational age at membrane rupture (weeks)

Admission to neonatal ICU Mean birthweight (g) Respiratory distress syndrome Sepsis Bronchopulmonary dysplasia Intraventricular haemorrhage Neonatal jaundice

<20 (n = 3)

20–24 (n = 8)

24–26 (n = 8)

26–28 (n = 16)

All (n = 35)

0 2200 0 1 0 0 0

5 1622 2 2 1 1 2

6 1358 3 4 2 0 3

12 1335 7 11 1 0 2

23 1829 12 16 4 1 7

190

C.J.M. Stewart et al. / European Journal of Obstetrics & Gynecology and Reproductive Biology 126 (2006) 186–192

Table 6 Characteristics of survivors and non-survivors

Gestational age at rupture of membranes (weeks) Gestational age at delivery(weeks) Latency period (days) Chorioamnionitis Steroid administration Birth weight (g) AFI at rupture of membranes AFI prior to delivery * **

Survivors (n = 35)

Non-survivors (n = 37)

24.4  3

22.5  3.1 *

29.6  2.8 35.7  36.4 7 25 1484  619 3.7 4.5

25.9  2.3 * 16.4  19.6* 8 11 811  257** 2.3 2.1

p < 0.05. p < 0.001.

Fig. 2. Survival rate with gestational age at membrane rupture.

Of infants with a birth weight of greater than 500 g, survivors had a higher mean birth weight, more advanced gestational age at delivery and a longer latency period than non-survivors (Table 6). Survival rates in relation to gestation at membrane rupture are illustrated in Fig. 2. The incidence of compression abnormalities was 17%. It was not possible to comment on lung hypoplasia in 23 of the 81 fetuses delivered because postmortems were not done. In the remaining 58, the incidence was 18%.

4. Discussion The clinical course of women who rupture their membranes in the midtrimester is often complicated by increased maternal and fetal morbidity, mainly related to sepsis and fetal loss. With improving perinatal care, there has been a tendency to manage these women expectantly. Our study evaluates the outcome of 78 pregnancies with preterm prelabour rupture of the membranes between 16 and 28 weeks gestation, managed expectantly. The age distribution of the patients in the study was in keeping with the general pregnant population cared for by the University of Cape Town Perinatal service. They were mainly indigent women who elected to continue with their pregnancies. Forty-five percent had had previous pregnancy

losses, either as miscarriages, stillbirths or neonatal deaths. Tocolytics were not used at all due to the conflicting evidence of its benefits and risks [18–22]. There is now considerable evidence to suggest that prophylactic antibiotics are beneficial in patients with preterm prelabour membrane rupture who are conservatively managed. The cochrane metanalysis [23] shows a longer latency period and a decrease in infective complications in patients to whom antibiotics are administered. Further metanalyses by Mercer and co-workers [24,26] and Egarter et al. [25] show similar results. More recently, the ORACLE study has shown an improvement in neonatal outcome [27]. The motivation for our use of prophylactic antibiotics routinely in this study was based on these data. Previous studies of expectant management of very preterm prelabour rupture of membranes have either not used antibiotics at all, or have used them sporadically only. The perinatal outcome in patients rupturing membranes before 24 weeks was improved in our series as compared with other studies. While one might speculate that this is due to routine antibiotic use, the closer follow-up, tighter guidelines for inclusion and more frequent use of steroids might also be contributory. Patients were all hospitalized as, with limited resources, it was not possible for a program of home visits, and patients had difficulty making regular trips to the hospital. Carlan et al. [28] showed that there was better survival with inpatient rather than outpatient monitoring. In an indigent population, this may be of value. The mean latency period in this study was 24 days with 30% of pregnancies continuing for longer than 1 month. Thirty-two patients (39%) reached 28 weeks gestation and 45 (56%) reached 26 weeks gestation. As survival was shown to be related to gestational age at delivery, prolongation of the latency period is important in this regard. A table is provided for comparison with other studies (Table 7). Patients were electively delivered at 34 weeks gestation, this being the hospital protocol in the case of preterm rupture of the membranes. The rationale for this is that, in our population, more than 95% of fetuses delivered at 34 weeks gestation will have lung maturity. This had to be weighed against the risk of infection related to the membrane rupture. With the evidence from this study that infection rates do not increase with increasing latency period, this policy needs review. Prolongation of the pregnancy to 37 weeks may have had even better outcomes in those cases. The overall survival rate in this study was 43%. Women who ruptured membranes at 26 weeks gestation or more had a 76% survival rate, compared with the group who ruptured membranes between 20 and 26 weeks where survival rate was 35% ( p = 0.003). We found no difference in survival rates in those women who ruptured their membranes between 20 and 23 weeks, and those who ruptured membranes between 23 and 26 weeks ( p = 0.83). The main factor determining survival is the gestational age at delivery. Survival is dependent on the neonatal facilities available. It

C.J.M. Stewart et al. / European Journal of Obstetrics & Gynecology and Reproductive Biology 126 (2006) 186–192

191

Table 7 Pregnancy outcomes with conservative management latency period

Taylor and Garite Beydoun and Yasin Moretti and Sibaia,b Bengston et al. Major and Kitzmillera,b,c Morales and Talley Chen et al. [29] Hibbard et al. This studyb,d a b c d

N

Mean

>7 days

>14 days

>28 days

Amnionitis

53 70 118 59 70 94 47 44 78

16.8 19.1 13 21.5 12 10.5 – 13.1 24.1

38% 50% 33% 51% 37% – 9% 41% 68%

28% 29% 17% – – 25% – – 47%

23% 19% 12% – 43% – – – 30%

42% 59% 39% 46% 24% 34% 77% 21%

Tocolysis in some patients. Steroids in some patients. Antibiotics in some patients. Routine antibiotics.

may be noted that no babies born at less than 25 weeks gestation survived. Only four of the infants born between 25 and 26 weeks survived, and seven out of 15 babies (47%) born between 26 and 28 weeks gestation survived. Thus, perinatal mortality in the group of women who ruptured membranes between 26 and 28 weeks was significantly higher than in other studies. Major and Kitzmiller [11] and Morales and Talley [14] both showed a 100% survival rate in patients delivering after 26 completed weeks. Similarly, the mean gestational age at delivery in surviving neonates was 26.5 weeks versus 23.8 weeks in non-survivors in Morales’s study, and 26.5 weeks compared with 25.4 weeks in that of Hibbard et al. [8]. In our study, a higher mean gestational age was associated with improved survival—29.5 weeks versus 25.9 weeks. The mean birthweight in the surviving group was 938 g in Morales’s study and 1007 g in that of Hibbard. In comparison, the mean birthweight in our study was 1484 g. This reflects the limited neonatal facilities in our setting. The question of whether continuation of the pregnancy is warranted if membrane rupture occurs before 23 weeks is an important one. While overall survival rates vary from 22% [3] to 63% [11], authors uniformly report poorer survival rates when membrane rupture occurs earlier than 23 weeks. Moretti and Sibai [9] reported a 13% perinatal survival with VPPROM before 23 weeks; Beydoun and Yasin [12] 28% and Morales 23%. Our study showed a survival rate of 30% in women rupturing their membranes less than 24 weeks. This reflects the longer latency period and is of even greater significance in view of the reduced neonatal facilities in comparison with the above reports. Clinical chorioamnionitis developed in 21% of the women in this study. There was no difference in the incidence of chorioamnionitis related to duration of rupture of membranes (Fig. 1). The use of steroids did not alter the incidence of maternal infection. Postpartum endometritis occurred in 17% of women. There were no serious maternal nfectious sequelae, although hysterectomy, septic thrombophlebitis and maternal death have been reported in other studies. Thus the overall incidence of maternal infection in this study was lower than in previous studies.

The incidence of fetal compression abnormalities in our study was 17%. This is in keeping with the 12% incidence quoted by Nimrod. Other studies have reported far fewer orthopaedic problems. Fetal compression in our study was not related to the duration of rupture of membranes (range 6–133 days) as suggested by Blott and Greenough [30]. One baby had an amniotic band which had severed two toes. All of the other abnormalities were minor and did not require corrective surgery. Pulmonary hypoplasia was present in 18% of cases. Many of these were delivered very preterm without any chance of survival. This figure may be falsely low as not all non-surviving babies had postmortems. The incidence of pulmonary hypoplasia in other studies varies from 0.8% [9] to 26% [27]. The lower incidence is probably due to a lack of postmortem data. In conclusion, our study supports the view that conservative management in the setting of very preterm rupture of the membranes is warranted. A prolonged latency period results in improved neonatal survival. The low infection rate may be due to the routine use of antibiotics. Before embarking on conservative management in the current time, HIV counselling and testing would have to be discussed with the patient, as the risks of this approach would be influenced by this. This study provides information which enables us to provide appropriate counseling for patients with early membrane rupture in a developing country. References [1] Gibbs R, Blanco J. Premature rupture of the membranes. Obstet Gynaecol 1982;60:671–9. [2] Nimrod C, Varela-Gittings F, Machin G, Campbell D, Wesenberg R. The effect of very prolonged membrane rupture on fetal development. Am J Obstet Gynaecol 1984;148:540–3. [3] Taylor J, Garite T. Premature rupture of membranes before fetal viability. Obstet Gynaecol 1984;64:615–20. [4] Mercer BM. Management of premature rupture of membranes before 26 weeks gestation. Obstet Gynaecol Clin North Am 1992;19(2): 339–51.

192

C.J.M. Stewart et al. / European Journal of Obstetrics & Gynecology and Reproductive Biology 126 (2006) 186–192

[5] Olofsson P, Rydhstrom H, Sjoberg N. How Swedish obstetricians manage premature rupture of the membranes in preterm gestations. Am J Obstet Gynaecol 1988;159:1028–34. [6] Dowd J, Permezel M. Pregnancy outcome following preterm premature rupture of the membranes at less than 26 weeks gestation. Aust NZ J Obstet Gynaecol 1992;32(2):120–4. [7] Hadi H, Hodson C, Strickland D. Premature rupture of membranes between 20 and 25 weeks gestation: role of amniotic fluid volume in perinatal outcome. Am J Obstet Gynaecol 1994;170:1139–44. [8] Hibbard J, Hibbard M, Ismail M, Arendt E. Pregnancy outcome after expectant management of premature rupture of membranes in the second trimester. J Reprod Med 1993;38(12):945–51. [9] Moretti M, Sibai B. Maternal and perinatal outcome of expectant management of premature rupture of membranes in the midtrimester. Am J Obstet Gynaecol 1988;159:390–6. [10] Nelson L, Meis P, Hatjis C, Ernest J, Dillard R, Schey H. Premature rupture of membranes: a prospective, randomised evaluation of steroids, latent phase, and expectant management. Obstet Gynaecol 1985;66:55–8. [11] Major C, Kitzmiller J. Perinatal survival with expectant management of midtrimester rupture of membranes. Am J Obstet Gynaecol 1990;163:838–44. [12] Beydoun S, Yasin S. Premature rupture of membranes before 28 weeks: conservative management. Obstet Gynecol 1986;155:471–9. [13] Bengtson JM, VanMarter LJ, Barss VA, Greene MF, Tuomala RE, Epstein MF. Pregnancy outcome after premature rupture of the membranes at or before 26 weeks’ gestation. Obstet Gynecol 1989;73:921–6. [14] Morales W, Talley P. Premature rupture of membranes at <25 weeks: a management dilemma. Am J Obstet Gynaecol 1993;168:503–7. [15] Roberts A, Mitchell J. Direct ultrasonographic measurement of fetal lung length in normal pregnancies and pregnancies complicated by prolonged rupture of membranes. Am J Obstet Gynaecol 1990;163:1560–6. [16] D’Alton M, Mercer B, Riddick E, Dudley D. Serial thoracic versus abdominal circumference ratios for the prediction of pulmonary hypoplasia in premature rupture of the membranes remote from term. Am J Obstet Gynaecol 1992;166:658–63. [17] Fong K, Ohlsson A, Zalev A. Fetal thoracic circumference: a prospective cross-sectional study with real-time ultrasound. Am J Obstet Gynaecol 1988;158:1154–60. [18] Garite T, Keegan K, Freeman R, Nageotte M. A randomised trial of ritodrine tocolysis versus expectant management in patients with

[19]

[20]

[21]

[22]

[23]

[24] [25]

[26]

[27]

[28]

[29]

[30]

premature rupture of membranes at 25 to 30 weeks of gestation. Am J Obstet Gynaecol 1987;157:388–93. Weiner C, Renk K, Klugman M. The therapeutic efficacy and costeffectiveness of aggressive tocolysis for premature labour associated with premature rupture of the membranes. Am J Obstet Gynaecol 1988;159:216–22. Matsuda Y, Ikenoue T, Ibara S. The efficacy of prophylactic antibiotic and tocolytic therapy for premature rupture of the membranes-a prospective randomised study (Japanese). Acta Obstet Gynaecol Jpn 1993;45(10):1109–14. Decavalas G, Mastrogiannis D, Papadopoulos V, Tzingounis V. Shortterm versus long-term prophylactic tocolysis in patients with preterm premature rupture of membranes. Eur J Obstet Gynaecol Reprod Biol 1995;59(2):143–7. Fortunato S, Welt S, Eggelston M, Bryant E. Active expectant management in very early gestations complicated by premature rupture of the fetal membranes. J Reprod Med 1994;39(1):13–6. Kenyon S, Boulvain M, Neilson J. Antibiotics for preterm premature rupture of membranes. Cochrane Database Syst Rev 2001;4:CD001058 [review]. Update in: Cochrane Database Syst Rev 2003;2:CD001058. PMID: 11687090 [PubMed-indexed for MEDLINE]. Mercer BM. Antibiotic therapy for preterm premature rupture of membranes. Clin Obstet Gynaecol 1998;41(2):461–8. Egarter C, Leitich H, Karas H, et al. Antibiotic treatment in preterm premature rupture of membranes and neonatal morbidity: a metanalysis. Am J Obstet Gynaecol 1996;174(2):589–97. Mercer BM, Miodovnik M, Thurnau GR, et al. Antibiotic therapy for reduction of infant morbidity after preterm premature rupture of the membranes, a randomized controlled trial. JAMA 1997;278(12):989– 95. Kenyon SL, Taylor DJ, Tarnow-Mordi W. ORACLE Collaborative Group. Broad spectrum antibiotics for preterm prelabour rupture of fetal membranes: the ORACLE 1 randomised trial. Lancet 2001;357(9261):979–88. Carlan SJ, O’Brien WF, Parsons MT, Lense JJ. Preterm premature rupture of membranes: a randomized study of home versus hospital management. Obstet Gynaecol 1993;81(1):61–4. Chen KC, Hsu JJ, Hsieh CC, Hsieh TT. Pregnancy outcome of preterm premature rupture of the membranes before 28 weeks. Chang Keng I Hsueh-Chang Gung Med J 1992;15(4):176–82. Blott M, Greenough A. Neonatal outcome after prolonged rupture of the membranes starting in the second trimester. Arch Dis Child 1988;63:1146–50.