Comparing the Hadlock fetal growth standard to the Eunice Kennedy Shriver National Institute of Child Health and Human Development racial/ethnic standard for the prediction of neonatal morbidity and small for gestational age.

BACKGROUND: The fetal growth standard in widest use was published by Hadlock >25 years ago and was derived from a small, homogeneous cohort. In 2015, The Eunice Kennedy Shriver National Institute of Child Health and Human Development Fetal Growth Study published updated standards that are specific to race/ethnicity. These do not allow for precise estimated fetal weight percentile calculation, however, and their effectiveness to predict neonatal morbidity and small for gestational age has not yet been compared to the long-standing Hadlock standard.

OBJECTIVE: We compared the ability of the Hadlock standard to predict neonatal morbidity and small for gestational age at birth with that of The Eunice Kennedy Shriver National Institute of Child Health and Human Development race-/ethnicity-specific standard. Our secondary objective was to compare their performance among our Native American population, which is not accounted for in the Eunice Kennedy Shriver National Institute of Child Health and Human Development standard.

STUDY DESIGN: For this retrospective study of diagnostic accuracy, we reviewed deliveries at the University of New Mexico Hospital from Jan. 1, 2013, through March 31, 2017. We included mothers with singleton, well-dated pregnancies and nonanomalous fetuses with an estimated fetal weight within 30 days of delivery. Cubic spline interpolation was performed on the Eunice Kennedy Shriver National Institute of Child Health and Human Development estimated fetal weight-percentile tables to calculate percentiles specific to the gestational day. Estimated fetal weight percentiles were then calculated using both the Hadlock and Eunice Kennedy Shriver National Institute of Child Health and Human Development race-/ethnicity-specific standards according to maternal self-identified race/ethnicity. We calculated the receiver operator area under the curve of each method to predict composite and severe composite neonatal morbidity and small for gestational age at birth (birthweight <10th percentile). As an additional measure of method accuracy, we calculated the mean ultrasound-birthweight percentile discrepancy. For Native Americans, percentiles were calculated using the Hadlock and Eunice Kennedy Shriver National Institute of Child Health and Human Development race/ethnicity standards (white, black, Hispanic, Asian), and test characteristics were calculated for each to predict neonatal morbidity and small for gestational age.

RESULTS: We included 1514 women, with a mean ultrasonography-to-delivery interval of 14.4 days (±8.8) and a small for gestational age rate of 13.6% (n = 206). For the prediction of both composite and severe composite neonatal morbidity, the Hadlock method had superior performance, with higher areas under the curve than the Eunice Kennedy Shriver National Institute of Child Health and Human Development method (P < .001 for both), though neither had good discriminatory value (all areas under the curve <0.8). For the prediction of small for gestational age at birth, the Hadlock standard had higher sensitivity (61.1%) than the Eunice Kennedy Shriver National Institute of Child Health and Human Development standard, both when using the interpolated Eunice Kennedy Shriver National Institute of Child Health and Human Development method (36.2%, P < .01) and the Eunice Kennedy Shriver National Institute of Child Health and Human Development whole-week 10th percentile cutoff (46.7%, P < .01). The Hadlock method also had a higher area under the curve than the Eunice Kennedy Shriver National Institute of Child Health and Human Development interpolated method to predict small for gestational age (0.89 vs 0.88, P < .01). The Hadlock method had a lower ultrasound-birthweight percentile discrepancy than the Eunice Kennedy Shriver National Institute of Child Health and Human Development method (6.1 vs 16.5 percentile points, P < .01). Fetuses classified as growth restricted by Hadlock but not Eunice Kennedy Shriver National Institute of Child Health and Human Development had significantly higher composite morbidity than normally grown fetuses. Among Native American women, the Hadlock method had the highest area under the curve to predict composite and severe composite morbidity, while the Hadlock and all Eunice Kennedy Shriver National Institute of Child Health and Human Development race-/ethnicity-specific methods performed comparably to predict small for gestational age.

CONCLUSION: Despite its publication >25 years ago, the Hadlock standard is superior to the Eunice Kennedy Shriver National Institute of Child Health and Human Development race-/ethnicity-specific standard for the prediction of both neonatal morbidity and small for gestational age.