Outcome-based comparative analysis of five fetal growth velocity models to define slow growth

Fetal growth surveillance includes assessment of size as well as rate of growth, and various definitions for slow growth have been adopted into clinical use. The aim of this study was to evaluate the effectiveness of different models to identify stillbirth risk, in addition to risk represented by the fetus being small-for-gestational age (SGA).

This was a retrospective analysis of a routinely collected and anonymized dataset of pregnancies that had two or more third-trimester ultrasound measurements of estimated fetal weight (EFW). SGA was defined as EFW < 10^th customized centile, and slow growth was defined according to five published models in clinical use: (1) a fixed velocity limit of 20 g per day (FVL₂₀ ); (2) a fixed > 50 centile drop, regardless of scan-measurement interval (FCD₅₀ ); (3) a fixed > 30 centile drop, regardless of scan interval (FCD₃₀ ); (4) growth trajectory slower than the third customized growth-centile limit (GCL₃ ); and (5) EFW at second scan below the projected optimal weight range (POWR), based on partial receiver-operating-characteristics-curve-derived  cut-offs specific to the scan interval.

The study cohort consisted of 164 718 pregnancies with 480 592 third-trimester ultrasound scans (mean ± SD, 2.9 ± 0.9). The last two scans in each pregnancy were performed at an average gestational age of 33 + 5 and 37 + 1 weeks. At the last scan, 12 858 (7.8%) EFWs were SGA, and of these, 9359 were also SGA at birth (positive predictive value, 72.8%). The rate at which slow growth was defined varied considerably (FVL₂₀ , 12.7%; FCD₅₀ , 0.7%; FCD₃₀ , 4.6%; GCL₃ , 19.8%; POWR, 10.1%), and there was varying overlap between cases identified as having slow growth and those identified as SGA at the last scan. Only the POWR method identified additional non-SGA pregnancies with slow growth (11 237/16 671 (67.4%)) that had significant stillbirth risk (relative risk, 1.58 (95% CI, 1.04-2.39)). These non-SGA cases resulting in stillbirth had a median EFW centile of 52.6 at the last scan and a median weight centile of 27.3 at birth. Subgroup analysis identified methodological problems with the fixed-velocity model because it assumes linear growth throughout gestation, and with the centile-based methods because the non-parametric distribution of centiles at the extremes does not reflect actual difference in weight gain.

Comparative analysis of five clinically used methods to define slow fetal growth has shown that only the measurement-interval-specific POWR model can identify non-SGA fetuses with slow growth that are at increased risk of stillbirth. © 2023 The Authors. Ultrasound in Obstetrics & Gynecology published by John Wiley & Sons Ltd on behalf of International Society of Ultrasound in Obstetrics and Gynecology.