812 Archives ofDisease in Childhood 1991; 66: 812-815

OBSTETRICS FOR PAEDIATRICIANS Arch Dis Child: first published as 10.1136/adc.66.7_Spec_No.812 on 1 July 1991. Downloaded from

Preterm prelabour rupture of membranes

David James

The problem due to oligohydramnios, placental abruption, Unfortunately there are too many examples in cord compression (either due to cord prolapse of very important conditions for or severe oligohydramnios), and mechanical which there is no consensus of professional difficulties at delivery. There is no information opinion about definition, diagnosis, risks, and to indicate whether these risks are any different above all, management. Preterm prelabour rup- for 'hindwater' as opposed to 'forewater' ture of the membranes (PPROM) is such a case. PPROM thus in practice no distinction is made The best, most widely used, and simplest between the two. definition of PPROM is membrane rupture occurring before the onset of regular uterine contractions prior to 37 weeks' gestation. Some The problem of diagnosis authors add caveats, for example, about the The diagnosis of membrane rupture is not diffi- interval between membrane rupture and the cult if there is liquor gushing from the or onset of labour or not including cases above 34 the on sterile speculum examination. weeks' gestation. In general, these are unhelpful However, in many instances this is not the case. modifications to the definition. However, while A careful history is important, with special care not helpful for definition, these two factors- being taken to distinguish PPROM from urin- duration of membrane rupture and gestational ary incontinence. The presence of white flecks age-critically influence the risks and manage- (vernix) or green coloration (meconium) of the ment of PPROM. the liquor may be a useful feature in the history. The reported incidence of prelabour mem- However, vernix and meconium are characteris- brane rupture in all varies from 1- tic of term liquor and seldom found before 32 10% depending on the population studied. weeks. If a sample of fluid is available for

However, if elective preterm deliveries and fetal examination then it may be possible to decide http://adc.bmj.com/ death before labour are excluded, the incidence whether it is liquor. contains of PPROM in women delivering preterm is protein and produces a typical fern like pattern between 40-60%. The main independent risk when allowed to dry on a microscope slide; factors for PPROM are antepartum vaginal urine does not, but in contrast contains creati- bleeding, maternal smoking, and previous pre- nine and uric acid. The pH of liquor, at 7 1- term delivery. 7 3, tends to be higher than that of urine.

The aetiology and pathogenesis of PPROM A variety of tests have been developed specifi- on September 29, 2021 by guest. Protected copyright. are poorly understood. For example, it is not cally to distinguish liquor not only from urine clear whether pregnancies complicated by spon- but also from vaginal secretions and other taneous preterm labour and delivery with fluids, but they all have limitations. Some, PPROM should be regarded as different in including the ferning test, are of little value at these respects from cases without preceding preterm gestations. For example, the Nile blue membrane rupture. Factors claimed to be poss- test, which produces a positive result by detect- ible causes of PPROM include infection, mem- ing the presence of orange staining anucleate brane collagen abnormalities, cervical incompe- cells in liquor, has a false negative rate of 90% at tence, raised intrauterine pressure, and low 36 weeks and 99% at 32 weeks. Detection of ct membrane calcium and magnesium content. fetoprotein using monoclonal antibodies should There have been many theories of the mechan- be more accurate at preterm gestations; isms whereby these factors produce PPROM however, it has not yet been adequately evalu- with much research currently involving platelet ated in clinical practice. The nitrazine test is activating factor and metabolism probably the most widely used. The underlying in the membranes. As with most aspects of principles are that nitrazine solution changes spontaneous preterm delivery, the explanation from yellow-green to dark blue at a pH of 6-5 remains as elusive as ever. or more and while vaginal pH during While the causes of PPROM are uncertain is normally between 4S5-6 0 that of liquor Bristol Maternity the risks are well known. The three serious is usually alkaline. Unfortunately nitrazine Hospital, complications, in order of importance, are pre- can also turn blue when to tap water, Southwell Street, exposed Bristol BS2 8EG term labour and delivery and its consequences, certain antiseptic solutions, cervical mucus, Correspondence to: infection, and pulmonary hypoplasia. Other semen, blood, and alkaline urine. Dr James. important complications are various deformities The transabdominal injection of various dyes Preterm prelabour rupture ofmembranes 813

in order to ascertain whether they appear The problem of infection Arch Dis Child: first published as 10.1136/adc.66.7_Spec_No.812 on 1 July 1991. Downloaded from through the cervix as an invasive method of The early diagnosis of intrauterine infection in diagnosing membrane rupture is not without PPROM is important yet difficult. The classical risk for the fetus and is not widely practised. signs of maternal fever, maternal and fetal Ultrasound evaluation of liquor volume has tachycardia, a tender , and foul smelling merit in theory. The presence of oligohydram- liquor are probably relatively late features. nios on scan with a history suggestive of mem- Furthermore intrauterine infection is not the brane rupture would probably correctly confirm only explanation of maternal fever or maternal the diagnosis. However, in practice ultrasound or fetal tachycardia. Another confounding fac- is of limited value. To produce appreciable oli- tor is that the discharge from fulminating aero- gohydramnios on ultrasound, the clinical fea- bic infection may be inoffensive. tures are usually clear cut with liquor leak con- A number of laboratory methods have been tinuing and oligohydramnios being obvious on introduced in an attempt to achieve earlier abdominal palpation. In contrast, the com- diagnosis ofintrauterine infection with PPROM. moner clinical dilemma is where there has poss- In general they have proved unreliable. A ibly been a small leak which has stopped. In summary of the tests is given in the table. Gram such cases ultrasound will not detect the small stain, white cell count, and culture of the liquor reduction of liquor volume. is probably the best of those available. But from where should the liquor be obtained for this bacterial examination? Collecting it vaginally is The problem of preterm delivery easier and identifying certain fetal pathogens Preterm labour and delivery and its complica- such as group B streptococci, Escherichia coli, or tions (respiratory distress syndrome, intraven- bacteriodes, especially if single isolates, may be tricular haemorrhage, and chronic lung disease) important. However, the usefulness of finding are the most common and important consequ- such organisms in planning subsequent ences of PPROM. Approximately 60% of cases management is not known. Negative cultures, will go into labour within one week of PPROM on the other hand, are reassuring. Because of at 24 weeks' gestation or less; this figure rises to this dilemma of possible vaginal commensals 80% after 34 weeks. contaminating liquor collected via that route, Because of the difficulties in accurate early obtaining liquor by has been diagnosis of intrauterine infection (see below) advocated to allow a more specific diagnosis of clinical management has concentrated on trying intrauterine infection. Unfortunately, there is to assess the consequences of preterm delivery no evidence to show that this approach is more especially when PPROM occurs before 34 accurate at diagnosing intrauterine infection. weeks. Almost exclusively the evaluation has Furthermore, liquor is not obtained in 30-50% centred on fetal pulmonary maturity and the of cases and the procedure carries a risk to the likelihood of respiratory distress syndrome. In fetus of trauma and the introduction of infec- this regard the traditional lecithin:sphing- tion. omyelin ratio in amniotic fluid has proved unre- Of more recent interest are the observations liable, especially when the liquor has been col- that early intrauterine infection appears to be http://adc.bmj.com/ lected vaginally. Fortunately, phosphatidylgly- associated with disturbances in fetal behavioural cerol is not found in any other body fluids than physiology. Abnormalities of fetal heart rate mature lung surfactant and thus its estimation is monitoring () and the fetal the test of choice in such circumstances. Of biophysical profile (concurrent recording of course, demonstration of fetal lung maturity fetal heart rate, breathing, movements, tone, does not necessarily mean that other consequ- and liquor volume) correlate closely with

ences of preterm delivery will not occur and that proved intrauterine sepsis. These non-invasive on September 29, 2021 by guest. Protected copyright. early elective delivery should be automatically tests have obvious advantages over amniocente- undertaken. Arguably the main value of demon- sis. However, at present it is not known strating fetal surfactant production is that whether use of these tests will prove to be help- maternal administration of steroids can be ful in determining the appropriate management avoided (see below). At present the timing of and improving fetal outcome. elective delivery in PPROM even with fetal lung There is a widely held view that vaginal maturity is a matter of clinical judgment rather examinations in PPROM introduces or causes than on the basis of properly conducted rando- infection. Although there is a reported associa- mised trials. In practice, most clinicians tend to tion of vaginal examinations and the develop- be more conservative the more preterm the pre- ment of sepsis, there are no controlled data to gnancy and opt for elective delivery only after confirm or refute the view that this relationship 34 weeks. is causal. There is no evidence as to whether

Accuracy ofprediction of intrauterine infection by laboratory tests in published studies (extracted from Ohlsson and Wang). Results are shown as range Test Sensitivitv Specificitl Positive Negative predictive predictive value value (1) White cell count, band count () 23-80 61-98 40-75 40-91 (2) C reactive protein (%) 37-100 44-100 10-100 50-100 (3) Gram stain, white cells, and culture of liquor (%) 36-80 83-98 78-87 56-92 (4) Leucocyte esterase (%) 8-83 84-100 4-100 68-92 (5) Gas chromatography (%) 29-93 56-91 20-88 67-96 Decimal places areomitted forclarity. Thedefinition of'intrauterine infection'varied. 814 James

digital or speculum examination is better. In dence from controlled studies is complicated by Arch Dis Child: first published as 10.1136/adc.66.7_Spec_No.812 on 1 July 1991. Downloaded from general there are two alternative practices. factors such as the heterogeneous nature of Firstly, a sterile examination is performed on all study groups, possible selection bias, small cases of PPROM at presentation, though there numbers, uncontrolled and variable con- are then differences of opinion as to how fre- comitant treatment. Some generalisations are quently examinations should be subsequently possible, however. repeated. In contrast, other obstetricians main- There is no consistent evidence that the use of tain that vaginal examinations should be carried prophylactic antibiotics in PPROM reduces out with PPROM only in active labour. There is maternal infection risk before delivery, in- no evidence to suggest which approach is creases neonatal sepsis risk after delivery, or superior. What is clear, however, is that vaginal significantly prolongs pregnancy. The only examination should only be undertaken to proved benefit is that maternal infection rates obtain information that is unavailable by other after delivery are significantly reduced by anti- means. biotic prophylaxis. These observations together with the largely theoretical rather than proved consequences of the development of resistant The problem of pulmonary hypoplasia strains and masking of infection in the newborn Preterm delivery is numerically the most impor- are probably the reasons for many authors not tant complication of PPROM but pulmonary supporting the routine use of prophylactic hypoplasia due to oligohydramnios is in many antibiotics in PPROM. ways more distressing. The prognosis for respir- There is no proved benefit from the use of atory distress syndrome in the preterm baby oral tocolytics in PPROM. The few comparative delivering after PPROM and after 28 weeks' studies that have been published suggest a non- gestation is very good with over 90% survival significant lower incidence of delivery within 48 rates due to measures such as ventilatory hours with their use but any such suggestion of support and exogenous surfactant treatment. benefit is lost when delivery rates at 10 days Pulmonary morphological development can be after PPROM are compared. impaired with very preterm membrane rupture There have been several trials of the use of especially before 26 weeks. Thus a mother maternally administered corticosteroids, such might spend weeks in hospital, the membranes as dexamethasone, in PPROM and all have having ruptured before 26 weeks, only for the demonstrated a clear advantage of the treat- baby to be delivered, perhaps after 30 weeks, ment. Overall, treatment with such agents sig- and dying from pulmonary hypoplasia. Theories nificantly reduced the risk of development of of the pathogenesis of the condition include respiratory distress syndrome by approximately fetal compression reducing lung growth, 50%. However, in addition all the studies did decreased fetal breathing, and excessive loss of demonstrate a trend to slightly higher neonatal lung fluid. sepsis rates, though this was not statistically sig- The problem before is that there is no nificant. The conclusion from such work is that reliable way of diagnosing pulmonary hypopla- the benefits of corticosteroids would seem to sia. Fetal breathing movements, chest wall outweigh the risks. http://adc.bmj.com/ growth, and liquor volume measurements have Comparative studies have been made of all been advocated but all have limitations. In 'active approach' to delivery, namely inducing practice, an expectant approach is followed even labour after corticosteroid treatment or as soon if there is a strong suspicion that the condition as there is evidence of fetal lung maturity, with may be likely to develop. For example, with the 'conservative approach' of waiting to a much membrane rupture early in the second tri- later gestation such as 34 weeks before inducing mester, continued fluid loss, and severe oli- labour or until spontaneous labour occurs. on September 29, 2021 by guest. Protected copyright. gohydramnios then pulmonary hypoplasia is There are consistent trends in the published highly probable but not every case will result in reports of higher rates of maternal sepsis, the problem. caesarean section, neonatal sepsis, respiratory Amnioinfusion (restoration of the amniotic distress syndrome, and neonatal mortality with fluid volume with, tor example, N-saline) is an the 'active' policy of early induction of labour alternative approach to the problem currently and it would seem that it confers more harm under evaluation. Although there is evidence than benefit. that such an approach may reduce caesarean When spontaneous labour occurs after section rates, fetal heart abnormalities, and fetal PPROM the correct approach is to allow it to acid-base status in the presence of PPROM, it is proceed. There is circumstantial evidence to not clear whether the overall outlook for the suggest that the event itself may be a manifesta- fetus is any better. The outcome of randomised tion of 'subclinical' intrauterine infection. controlled trials is awaited. There is no evidence of benefit to mother or baby from the use of tocolytics in this circum- stance. The only exception to this view might The problem of management be where uterine activity commences at a non- Since preterm labour and delivery are the main viable gestation such as below 24 weeks. risks of PPROM, referral to a centre with facili- However, there are no studies that have addres- ties for neonatal intensive care, if these are not sed this issue and if this approach is followed available locally, is generally accepted as wise because 'there is nothing to lose', great vigilance management. There is, however, less agreement must be maintained for infection. When sponta- over other aspects of management of PPROM. neous labour occurs after PPROM and there is Furthermore, a review of the published evi- evidence of infection delivery must be allowed Preterm prelabour rupture ofmembranes 815

to proceed. The chosen mode of delivery would toring techniques as for sepsis; and (c) uterine Arch Dis Child: first published as 10.1136/adc.66.7_Spec_No.812 on 1 July 1991. Downloaded from be that appropriate for any equivalent preterm activity. fetus without PPROM. That said, the issue of (5) Corticosteroid administration confers which is the best method for delivery of the pre- more benefit than harm, but the routine use of term fetus presenting by the breech is not tocolytics or prophylactic antibiotics is not resolved. Furthermore, there is some evidence recommended. that in the presence of infection caesarean sec- (6) Vaginal examinations should be kept to a tion may carry a higher morbidity than vaginal minimum and only be used where the informa- delivery. Though there are no trials of the use of tion to be gleaned cannot be obtained by any antibiotics before delivery in the presence of other means. infection commonsense argues for their use as (7) Where infection occurs intravenous anti- the objections are largely theoretical. biotics should be used and delivery by the nor- mal route for that stage of pregnancy should be undertaken. Conclusions Finally, further research is needed in the fol- Although there are more divergent opinions lowing areas: over the management of PPROM than the sci- (1) The relative risks of hindwater versus entific evidence can support, the following forewater membrane rupture. general guidelines are suggested: (2) More accurate methods of diagnosing (1) Accurate diagnosis and intrauterine infection. estimation are important at the outset. (3) The use of prophylactic antibiotics in a (2) Transfer to a tertiary referral centre selective manner (for example, with concurrent should be considered if local facilities for neona- administration ofcorticosteroids). tal intensive care are not adequate. (4) Amnioinfusion. (3) Prolongation of the pregnancy to 32-34 weeks (rather than active delivery after steroid Suggestions for further reading 1 Keirse MJNC, Ohisson A, Treffers PE, Kanhai HHH. treatment or evidence of fetal lung maturity) Prelabour rupture of the membranes preterm. In: Chalmers should be the-aim. I, Enkin M, Keirse MJNC eds. Effective care in pregnancy and . Oxford: Oxford University Press, 1989: (4) Surveillance should be maintained for: (a) 666-93. intrauterine sepsis-with regular maternal pulse 2 Kilbride HW, Yeast JD, Thibeault DW. Intrapartum and delivery room management of premature rupture of and temperature recordings, fetal heart rate, membranes complicated by oligohydramnios. Clin Pen- and behaviour monitoring and regular examina- natol 1989;16:863-88. 3 Ohlsson A. Treatments of preterm premature rupture of the tion of liquor collected vaginally (Gram stain, membranes: a meta-analysis. Am J Obstet Gynecol 1990; leucocyte count and culture); (b) fetal com- 162:890-906. 4 Ohlsson A, Wang E. An analysis of antenatal tests to detect promise-not just from infection but from infection in preterm premature rupture of the membranes. cord compression using the same regular moni- AmJ Obstet Gynecol 1989;160:809-18. http://adc.bmj.com/ on September 29, 2021 by guest. Protected copyright.