Results
After excluding data that met exclusive criteria, a total of 2340 cases were included for final analysis. Among whom, the number of FGR cases defined by SMFM definition, ACOG definition and ISUOG definition were 115 (4.9%), 63 (2.7%) and 48 (2.1%) separately. Due to the overlapping of these three definitions, those FGR cases identified by ACOG and ISUOG were included in the cases defined by SMFM (Figure 2). Therefore, there were 2225 cases in the non-FGR group.
As shown in Table 1, the incidences of preeclampsia in the FGR groups according to different definitions were significantly higher than those in the non-FGR group (13.91%, 17.46%, 29.17% vs 1.39%), so was the preterm birth rate (26.96%, 36.51%, 50% vs 6.2%) .
Regarding the neonatal outcomes, there were 147 cases (6.28%) of SGA newborns. The incidence of SGA in the different FGR groups was significantly higher than that in the non-FGR group (52.17%, 57.14%, 62.5% vs 3.91%). There were 127 foetus (5.43%) who were complicated with ANO. The incidence of ANO in the FGR group was higher compared to the non-FGR group (19.13%, 25.4%, 31.25% vs 4.72%). Likewise, the incidence of each neonatal complications in the different FGR group were also higher than that in the non-FGR group (Table 2).
When comparing the discriminatory capacities of the three different definitions for predicting SGA (Table 3), the SMFM criteria had the highest sensitivity (40.82% vs 24.49%, 20.41%), while the ISUOG criteria had the highest specificity (99.18% vs 97.49%, 98.77. As for the predictive value for a composite ANO, it was displayed that all three definitions had low sensitivity (17.32%, 12.6%, 11.81%) and high specificity of 95.8%, 97.88% and 98.51%. The AUCs of SMFM for predicting SGA (0.692) and ANO (0.566) were slightly higher than those of ACOG (0.616, 0.552) and ISUOG (0.598, 0.552)