(254) A Model to Predict Latency in Preterm Prelabor Rupture of Membranes (PPROM)

Thursday, January 30, 2025

10:30 AM – 12:30 PM

Submitting Author and Presenting Author(s)

Kira Anne Bromwich, MD

University of Pennsylvania Perelman School of Medicine
Philadelphia, PA, United States

Coauthor(s)

Kelly Zafman, MD, MSCR (she/her/hers)

MFM Fellow
Hospital of the University of Pennsylvania
Philadelphia, Pennsylvania, United States
MF

Markolline Forkpa, MPH

Pennsylvania Medicine Women's Health Clinical Research Center
Philadelphia, Pennsylvania, United States
JC

Jesse Chittams, MS

University of Pennsylvania School of Nursing
Philadelphia, Pennsylvania, United States
MG

Miatta Goba, BS

University of Pennsylvania Perelman School of Medicine
Philadelphia, Pennsylvania, United States
Nadav Schwartz, MD

Professor, Maternal Fetal Medicine
University of Pennsylvania, Perelman School of Medicine
Philadelphia, Pennsylvania, United States
Rebecca F. Hamm, MD, MSCE (she/her/hers)

Assistant Professor
University of Pennsylvania Perelman School of Medicine
Philadelphia, Pennsylvania, United States

Objective: We sought to develop a model to predict length of latency in patients with PPROM.

Study Design: This was a retrospective cohort study of singleton gestations presenting with PPROM between 24w0d-32w6d from 2017-2022 at two large academic centers. Patients presenting with active labor, clinical chorioamnionitis, contraindications to expectant management, or known fetal anomalies were excluded. Primary outcomes were latency ≥2 days and ≥7 days. We collected data on maternal and obstetric history, prenatal course, and symptoms, labs, vitals, and fetal data at presentation via individual chart review. Multivariable logistic regression was used to produce prediction models.

Results: Of 304 patients who presented with PPROM between 24w0d and 32w6d, 192 (63.1%) met inclusion criteria, with most exclusions due to active labor or fetal anomalies. Gestational age (GA; p = 0.02), presence of contractions (p = 0.02), and cervical dilation (p = 0.03) were significantly associated with latency ≥2 days. (Table 1.) After adjusting for confounders, patients presenting with contractions had 62% lower odds of a ≥2 day latency. Each additional centimeter of dilation and week of gestation reduced the odds of ≥2 day latency by 27% and 17%, respectively. Together, these variables achieved moderate prediction of the outcome (AUC: 0.72. Figure 1). Examples of variables not determined to be significant in the model included parity, obstetric history, maternal vitals and basic lab markers, and fetal heart rate patterns. For latency ≥7 days, only GA (p = 0.01) and cervical dilation (p = 0.01) were found to be significant, yielding an adjusted model with an AUC of 0.68.

Conclusion: In patients presenting with PPROM, it is possible to generate a predictive model for odds of latency ≥2 days. This can be used to individualize counseling and help set patient expectations, as well as to assist in guiding resource utilization and necessity or timing of hospital transfer.