Results
PlGF was measured in addition to the other trisomy 21 screening
biomarkers in 11,518 women. 29 (0.25%) women were confirmed as having
trisomy 21 affected pregnancy. The maternal, pregnancy and screening
marker characteristics of these women are presented in Table 1. The
detection rate of the first trimester combined test was 89.7% (95%
confidence interval (CI) 72.6 to 97.8%) for a 5.7% (95%CI 5.3 to
6.2%) false positive rate.
The best fit regression model for PlGF was log-linear with gestational
age in days, weight (kg) and smoking being significant independent
predictors of its level. Expected median
log10PlGF in unaffected pregnancies was estimated as being = 2.50220583
−0.03872797 * gestational age + 0.00032803 * gestational
age2 − 0.00146749 * weight + 0.07319080 * smoker (1
for smoker or 0 for non-smoker) (R2=0.11).
The mean log10 PlGF MoM (standard deviation: SD) in
unaffected and trisomy 21 affected pregnancies were 0.0000 (0.1939) and
-0.2449 (0.2515), respectively. The correlations between
log10 PlGF MoM, log10 free β-hCG MoM and
log10 PAPP-A MoM in unaffected pregnancies were
respectively 0.0783 (95% CI 0.0622 to 0.0944) and 0.3184 (95% CI
0.3037 to 0.3329) and statistically significant (p<0.001). The
respective correlations in those with trisomy 21 were -0.2566 (95% CI
-0.5600 to 0.1076) and -0.1331 (95% CI -0.4655 to 0.2322) and not
statistically significant (p>0.16). The respective
correlations in the pooled data were 0.0754 (95% CI 0.0593 to 0.0916)
and 0.3205 (95% CI 0.3058 to 0.3349). The correlation between
log10 PlGF MoM and gestational age at screening was not
significant (r=0.13, p=0.48).
Figure 1 shows the ROC of PlGF based trisomy 21 screening tests compared
to the current combined first trimester test. Delong test indicated that
the difference in AUC between the combined test and after replacing
PAPP-A or by adding PlGF as an additional marker was not significant AUC
Combined vs Replacing PAPP-A: 0.984 (95% CI: 0.981 to 0.986) vs. 0.982
(95% CI: 0.980 to 0.985), p=0.58; Combined vs Adding PlGF: 0.984 (95%
CI: 0.981 to 0.986) vs. 0.979 (95% CI: 0.976 to 0.981); p=0.36).
For a term risk cut-off of 1:250, the detection rate and false positive
rate based on replacing PAPP-A by PlGF in the combined test were 89.7%
and 6.2%, respectively. The corresponding figures for the same
threshold after adding PlGF to combined screening were 89.7% and 5.5%,
respectively. Screening follow-up action remained unchanged in 96.0%
and 97.8% of women had we reported trisomy 21 risk based on replacing
PAPP-A by PlGF or adding PlGF to the existing combined test.
Post screening follow-up would have changed in 460 (4.0%) women by
replacing PAPP-A by PlGF, with 52 (0.45%) women previously having a
term risk of <1:250 now having a risk ≥1:250. When PlGF was
added to the combined test post screening, follow-up would have changed
in 257 (2.2%) women, with 29 (0.45%) women previously having a term
risk of ≥1:250 now having a term risk of <1:250.