Bm. Mercer et al., The Preterm Prediction Study: Prediction of preterm premature rupture of membranes through clinical findings and ancillary testing, AM J OBST G, 183(3), 2000, pp. 738-745
OBJECTIVE: Our objective was to determine the relative importance of demogr
aphic characteristics, clinical risk factors, and ancillary screening tests
in the prediction of preterm birth as a result of premature rupture of mem
branes.
STUDY DESIGN: A total of 2929 women were evaluated in 10 centers at 23 to 2
4 weeks' gestation. Demographic and clinical characteristics were ascertain
ed. Cervicovaginal fetal fibronectin and bacterial vaginosis were evaluated
. Cervical length was measured by vaginal ultrasonography. Patients were ev
aluated for spontaneous preterm birth caused by preterm premature rupture o
f membranes at <37 and <35 weeks' gestation. Multivariate analyses were per
formed separately for nulliparous women and multiparous women.
RESULTS: Premature rupture of membranes at <37 weeks' gestation complicated
4.5% of pregnancies, accounting for 32.6% of preterm births. Univariate an
alysis revealed low body mass index, pulmonary disease, contractions within
2 weeks, short cervix (less than or equal to 25 mm), positive results of f
etal fibronectin screening, bacterial vaginosis, and a previous preterm bir
th caused by premature rupture of membranes tin multiparous women) to be si
gnificantly associated with preterm birth caused by premature rupture of me
mbranes in the current gestation. Short cervix, previous preterm birth caus
ed by premature rupture of membranes in multiparous women, and presence of
fetal fibronectin were the strongest predictors for both preterm birth caus
ed by premature rupture of membranes at <37 and <35 weeks' gestation. Women
with positive fetal fibronectin screening results and a short cervix had g
reater risks both of preterm birth caused by premature rupture of membranes
at <37 weeks' gestation (relative risk, 4.9) and of preterm birth caused b
y premature rupture of membranes at <35 weeks' gestation (relative risk, 13
.5) than did those without these risk factors. Multiparous women with all t
hree risk factors had a 31.3-fold increased risk of preterm birth caused by
premature rupture of membranes at <35 weeks' gestation.
CONCLUSION: The combination of short cervical length, previous preterm birt
h caused by preterm premature rupture of membranes, and positive fetal fibr
onectin screening results was highly associated with preterm delivery cause
d by preterm premature rupture of membranes in the current gestation.