Contingent versus routine third‐trimester screening for late fetal growth restriction

Objective To evaluate the use of third‐trimester ultrasound screening for late fetal growth restriction ( FGR ) on a contingent basis, according to risk accrued in the second trimester, in an unselected population. Methods Maternal characteristics, fetal biometry and second‐trimester uterine artery ( UtA ) Doppler were included in logistic regression analysis to estimate risk for late FGR (birth weight < 3 rd percentile, or 3 rd –10 th percentile plus abnormal cerebroplacental ratio or UtA Doppler, with delivery ≥ 34 weeks). Based on the second‐trimester risk, strategies for performing contingent third‐trimester ultrasound examinations in 10%, 25% or 50% of the cohort were tested against a strategy of routine ultrasound scanning in the entire population at 32 + 0 to 33 + 6 weeks. Results Models were constructed based on 1393 patients and validated in 1303 patients, including 73 (5.2%) and 82 late FGR (6.3%) cases, respectively. At the second‐trimester scan, the a‐posteriori second‐trimester risk ( a‐posteriori first‐trimester risk (baseline a‐priori risk and mean arterial blood pressure) combined with second‐trimester abdominal circumference and UtA Doppler) yielded an area under the receiver–operating characteristics curve ( AUC ) of 0.81 (95%  CI , 0.74–0.87) (detection rate ( DR ), 43.1% for a 10% false‐positive rate ( FPR )). The combination of a‐posteriori second‐trimester risk plus third‐trimester estimated fetal weight (full model) yielded an AUC of 0.92 (95%  CI , 0.88–0.96) ( DR , 74% for a 10% FPR ). Subjecting 10%, 25% or 50% of the study population to third‐trimester ultrasound, based on a‐posteriori second‐trimester risk, gave AUCs of 0.81 (95%  CI , 0.75–0.88), 0.84 (95%  CI , 0.78–0.91) and 0.89 (95%  CI , 0.84–0.94), respectively. Only the 50% contingent model proved statistically equivalent to performing routine third‐trimester ultrasound scans ( AUC , 0.92 (95%  CI , 0.88–0.96), P  = 0.11). Conclusion A strategy of selecting 50% of the study population to undergo third‐trimester ultrasound examination, based on accrued risk in the second trimester, proved equivalent to routine third‐trimester ultrasound scanning in predicting late FGR . Copyright © 2015 ISUOG. Published by John Wiley & Sons Ltd.