OBJECTIVE: An observational study of a consecutive case series of pre-viable PPROM (16-24 gestational weeks) was performed between 2001 and 2007 in a single tertiary centre to identify factors that predict neonatal survival. METHODS: Detailed obstetric, ultrasound and neonatal data were abstracted from clinical records. Univariate, multivariate and receiver operator curve (ROC) analyses were performed to identify predictors of neonatal survival to discharge. RESULTS: A total of 143 cases of PPROM were identified. Survival to discharge was less with PPROM at 16-20 weeks than 20-24 weeks (17% versus 39%; p=0.042). GA at PPROM, latency, mode of delivery and electronic foetal monitoring (EFM) were all significant, independent, predictors of survival (p<0.05). Ultrasound assessed amniotic fluid index (AFI) was a poor predictor of survival (area under ROC=0.649, 95% CI=0.532-0.766). A multivariable predictive model, including GA at PPROM, latency, mode of delivery and EFM had an area under the ROC of 0.954 (95% CI=0.916-0.993, sensitivity 97%, specificity 89% and accuracy 92%). CONCLUSION: Pre-viable PPROM has a poor prognosis, though modern neonatal management techniques may improve survival in late pre-viable PPROM. The predictive model generated from this consecutive case series of this rare condition provides valuable data for counselling patients with this condition.
OBJECTIVE: An observational study of a consecutive case series of pre-viable PPROM (16-24 gestational weeks) was performed between 2001 and 2007 in a single tertiary centre to identify factors that predict neonatal survival. METHODS: Detailed obstetric, ultrasound and neonatal data were abstracted from clinical records. Univariate, multivariate and receiver operator curve (ROC) analyses were performed to identify predictors of neonatal survival to discharge. RESULTS: A total of 143 cases of PPROM were identified. Survival to discharge was less with PPROM at 16-20 weeks than 20-24 weeks (17% versus 39%; p=0.042). GA at PPROM, latency, mode of delivery and electronic foetal monitoring (EFM) were all significant, independent, predictors of survival (p<0.05). Ultrasound assessed amniotic fluid index (AFI) was a poor predictor of survival (area under ROC=0.649, 95% CI=0.532-0.766). A multivariable predictive model, including GA at PPROM, latency, mode of delivery and EFM had an area under the ROC of 0.954 (95% CI=0.916-0.993, sensitivity 97%, specificity 89% and accuracy 92%). CONCLUSION: Pre-viable PPROM has a poor prognosis, though modern neonatal management techniques may improve survival in late pre-viable PPROM. The predictive model generated from this consecutive case series of this rare condition provides valuable data for counselling patients with this condition.