Fetal Medicine Foundation Stillbirth risk score
Table 1 shows the maternal demographic data that synthesize the FMF Stillbirth Risk Calculator. In our study cohort, women after stillbirth compared to matched controls were found to have a significantly higher BMI (p= 0.002), were more frequently nicotine consumers (p <0.001) and suffered more frequently medical conditions, such as hypertension (p= 0.015), APS (p= 0.017) and diabetes (p <0.001) with higher need for insulin (p =0.006).
Figure 1 shows the distribution of the FMFStillbirth risk score in IUFDs and matched controls. The medianFMF Stillbirth risk score in the group of stillbirths was 0.45% (0.19-5.70%) [1:222 (1:526–1:17)], whilst the median risk score in the group of matched live births was 0.23% (0.18-1.30%) [1:435 (1:556–1:77); p <0.001]. Figure 2 and Figure 3 show the FMF Stillbirth risk scores in stillborn fetuses per cause of death with and without outliers, respectively.
To evaluate the discriminative power of the FMF Stillbirth Risk Calculator we performed a ROC analysis and calculated an area under the curve (AUC) of 0.72 (95% CI 0.67–0.78; p <0.001) to predict antepartum stillbirth in the total cohort (Figure 4 ). Also after exclusion of all stillborn fetuses with congenital anomalies, the AUC was 0.74 (95% CI 0.67-0.82;p <0.001). Assessing stillborn fetuses of unknown cause, AUC was 0.83 (95% CI 0.72-0.93; p <0.001). Assessing stillborn fetuses due to placental dysfunction, the AUC was 0.64 (95% CI 0.50-0.78; p =0.053).
Univariate binary logistic regression to examine the FMFStillbirth Risk Calculator´s predictive ability resulted in an OR of 6.22 (95% CI 3.91–9.89; p <0.001) at an optimalFMF Stillbirth risk score cut-off of ≥0.34% for predicting stillbirth with a specificity of 82% and a sensitivity of 58% in the total cohort of IUFDs.