Date of Award

January 2018

Document Type


Degree Name

Medical Doctor (MD)



First Advisor

Christian M. Pettker

Second Advisor

Seth Guller


Counsel Confidently: Reliability of Bedside Estimated Fetal Weight at Periviability

Krenitsky NM, Desai A, Johnson AW, Xu X, Pettker CM

Yale University Department of Obstetrics, Gynecology & Reproductive Sciences

Introduction: Counseling patients at periviability hinges on providing survival outcomes calculated using estimated fetal weight (EFW) as a proxy for birth weight (BW). Accurate assessment of EFW is a critical input but is subject to variability and error. Formal ultrasound by certified sonographers is often considered superior to bedside ultrasound conducted by housestaff. We hypothesize that bedside EFW is as accurate as formal EFW at periviability when compared to BW, and that there is no significant difference in neonatal survival estimates generated using bedside versus formal EFW.

Methods: We conducted a retrospective cohort study of liveborn deliveries 22 weeks 0 days to 25 weeks 6 days at a single institution from 2004 to 2014. Observations with documented BW and an EFW within 10 days of delivery were included. EFW error was calculated as (EFW-BW)/BW*100, and accuracy as |EFW-BW|/BW*100. Error and accuracy were analyzed using a linear regression model adjusted for a number of maternal, fetal, and ultrasound characteristics. Similarly, error and accuracy of NICHD Neonatal Research Network Outcomes Estimator survival statistics using either EFW or BW were computed, analyzed, and compared.

Results: A total of 307 ultrasounds (167 bedside and 140 formal) performed on 259 patients were analyzed. Unadjusted mean percent accuracy of EFW was 9.41% (SD=7.53%) for bedside ultrasound and 8.57% (SD=6.49%) for formal ultrasound. EFW was a strong predictor of BW (correlation coefficient = 0.867, p<0.001). Type of ultrasound (bedside versus formal) was not a significant predictor of ultrasound error or accuracy (p=0.099 and p=0.762, respectively) after adjusting for gestational age (GA), days between ultrasound and birth, fetal sex, number of fetuses, maternal age, body mass index, race, tobacco use, and year in which ultrasound was performed. Survival generated using EFW and GA at ultrasound was significantly correlated with survival using BW and GA (correlation coefficient = 0.900, p<0.001). There was no significant association between ultrasound type (bedside or formal) and error or accuracy of EFW predicted survival, survival without profound neurodevelopmental impairment, or survival without moderate to severe neurodevelopmental impairment (all p>0.05).

Conclusions: Bedside EFW performs as well as formal EFW at predicting BW at periviability. Providers should feel confident that patients are receiving accurate morbidity and mortality statistics based on either modality. We suggest that bedside EFW alone can be used to expeditiously and reliably counsel patients at risk of periviable birth, and that obtaining both bedside and formal scans within 10 days of delivery is unnecessary.


This thesis is restricted to Yale network users only. It will be made publicly available on 06/25/2100