Implication of third-trimester screening accuracy for small-for-gestational age and additive value of third-trimester growth-trajectory indicators in predicting severe adverse perinatal outcome in low-risk population: pragmatic secondary analysis of IRIS study

Objectives: To examine the implications of third-trimester small-for-gestational-age (SGA) screening accuracy on severe adverse perinatal outcome (SAPO) and obstetric intervention in a low-risk population. Furthermore, we aimed to explore the additive value of third-trimester sonographic growth-trajectory measurements in predicting SAPO and obstetric intervention. Methods: This was a secondary analysis of a Dutch national multicenter stepped-wedge-cluster randomized trial among 11 820 low-risk pregnant women. Using multilevel multivariable logistic regression analysis, we compared SAPO and obstetric interventions in SGA neonates with and without SGA suspected prenatally (true positives and false negatives) and non-SGA neonates with and without SGA suspected prenatally (false positives and true negatives). In a subsample (n = 7989), we analyzed the associations of abdominal circumference (AC) and estimated fetal weight (EFW) < 10 ^th centile (p10) and third-trimester growth-trajectory indicators AC and EFW crossing > 20 and AC crossing > 50 centiles and the lowest decile of AC growth-velocity Z-scores (ACGV < 10%) with SAPO and obstetric interventions. Results: SGA infants, i.e. the true-positive and false-negative cases, had an increased risk of SAPO (adjusted odds ratio (aOR), 4.46 (95% CI, 2.28–8.75) and aOR 2.61 (95% CI, 1.74–3.89), respectively), and obstetric intervention (aOR for: induction of labor, 2.99 (95% CI, 2.15–4.17) and 1.38 (95% CI, 1.14–1.66); Cesarean section, 1.82 (95% CI, 1.25–2.66) and 1.27 (95% CI, 1.05–1.54); medically indicated preterm delivery, 2.67 (95% CI, 1.97–3.62) and 1.20 (95% CI, 1.03–1.40)). The false-positive cases did not differ from the true negatives for all outcomes, including obstetric intervention. Of the third-trimester growth-trajectory indicators, only ACGV < 10% was associated moderately with SAPO (aOR, 2.15 (95% CI, 1.17–3.97)), while AC and EFW crossing > 20 and AC crossing > 50 centiles were not. Both EFW < p10 alone (aOR, 1.95 (95% CI, 1.13–3.38)) and EFW < p10 combined with ACGV < 10% (aOR, 4.69 (95% CI, 1.99–11.07)) were associated with SAPO, and they performed equally well in predicting SAPO (area under the receiver-operating-characteristics curve, 0.71 (95% CI, 0.65–0.76) vs 0.72 (95% CI, 0.67–0.77), P = 0.51). Conclusion: Neonates who had been suspected falsely of being SGA during pregnancy had no higher rates of obstetric intervention than did those without suspicion of SGA prenatally. Our results do not support that third-trimester low fetal growth velocity (ACGV < 10%) may be of additive value for the identification of fetuses at risk of SAPO in populations remaining at low risk throughout pregnancy. AC and EFW crossing > 20 and AC crossing > 50 centiles performed poorly in identifying abnormal fetal growth.