Personal or nonessential information may be redacted at the editor's discretion.

INTRODUCTION: Fetal growth abnormalities are among the main drivers of perinatal mortality and morbidity and a cornerstone of prenatal care. However they are defined based on arbitrary definitions. Our objective was to define fetal growth abnormalities based on increase risk of stillbirth. METHODS: In a prospective cohort of 35,251 singleton pregnancies, stillbirth was defined as fetal demise at > 20 weeks without chromosomal or structural abnormalities. Fetal growth was percentile of birth weight for gestational age at delivery. The risk of stillbirth as a function of percentile of fetal growth and gestational age was determined using logistic regression. The threshold in the relationship adjusted for gestational age at delivery was determined using the Bayesian information criterion to identify the point where the relationship’s strength changes. RESULTS: There were 61 (0.2%) non-anomalous stillbirths. At the 8.0th percentile of fetal growth the risk of stillbirth increased from 1.5 to 6.5 per 1,000 pregnancies. Risk of stillbirth at or below 8th percentile was 0.0065 (95% CI 0.0060-0.0070) and above the 8.0th percentile the risk of stillbirth was 0.0015 (95% CI 0.0013-0.0017); P<.0001 for both associations. CONCLUSION: Definition of fetal growth impairment based on the risk of stillbirth will result in fewer small but normal pregnancies to be misclassified as abnormally grown and thus decreasing unnecessary testing and interventions. Additionally, lower threshold 8th vs 10th percentile for defining fetal growth restriction would decrease by 2% of all pregnancies (80,000 per year in the US) the number of pregnancies requiring further testing and intervention.