OBJECTIVE: Our purpose was to determine risk factors predictive of preterm premature rupture of the membranes in women treated for preterm labor with intact membranes. STUDY DESIGN: Women with intact membranes participating in a National Institute of Child Health and Human Development multicenter randomized trial of adjunctive antibiotic therapy for preterm labor (24 to 34 weeks) were studied (n = 275). After randomization, 22 women continued to have contractions and were delivered of their infants. The remaining 253 women whose contractions had ceased composed our study population. Preterm premature rupture of the membranes was diagnosed if ruptured membranes occurred ≥1 hour before the onset of recurrent preterm labor. As part of the study protocol, most women underwent amniocentesis on admission. RESULTS: Preterm premature rupture of the membranes developed in 44% women (17.4%). Women who had preterm premature rupture of the membranes were more likely to be black (p = 0.004), to be multiparous (p = 0.014), to have a history of abortion(s) (p = 0.001), to have had a preterm birth(s) (p = 0.036), to have early onset preterm labor (p = 0.04), to have more advanced cervical dilatation (p = 0.0001), to have one or more amniotic fluid markers suggestive of infection (p = 0.01, odds ratio 4.2), and to have positive amniotic fluid cultures (p = 0.0007, odds ratio 27). Assignment to antibiotic therapy did not prevent preterm premature rupture of the membranes in the 253 women randomized or in the 16 women with a positive amniotic fluid marker(s) of infection. CONCLUSION: Black race, multiparity, a history of abortion or preterm birth, advanced dilatation, and a positive amniotic fluid marker(s) are associated with preterm premature rupture of the membranes in women with preterm labor. Antibiotic treatment did not prevent preterm premature rupture of the membranes. © 1995.