A Bayesian risk analysis for Trisomy 21 in isolated choroid plexus cyst: combining a prenatal database with a meta-analysis

Objective: The purpose of this study was to quantify the possible additional risk of a fetus with an isolated choroid plexus cyst (ICPC) for Trisomy 21 by combining a large controlled cohort study with data from existent studies.

Methods: We searched our prenatal database between 2000 and 2014 for all singleton pregnancies between 18 + 0 and 26 + 6 gestational weeks with either an isolated choroid plexus cyst (study group) or no abnormality found in the detailed ultrasound scan (control group). We assessed all prenatal karyotyping results if invasive testing was performed and attempted to collect the postnatal outcome reports of all patients. The prevalence of Down syndrome was calculated. By using previous studies that met our inclusion criteria, a meta-analysis following the Bayesian Independent Model was created. From this meta-analysis, we computed the posterior predictive distribution of the probability (Trisomy 21 | ICPC) = P1 including posterior means, standard deviations, quantiles (2.5, 50, and 97.5%). By calculating the posterior of the difference (Δ) between the probability (Trisomy 21 | ICPC) and the probability (Trisomy 21 | Normal Ultrasound) = P2, we investigated the additional risk of an ICPC (ΔB = P1-P2).

Results: Overall, we detected 1220 fetuses with an isolated plexus cyst at 19-27 weeks of gestational age (GA). In our study group, the prevalence of Trisomy 21 was 2/1220 (0.16, 95% CI: 0.1-0.6%). The median of the pooled probability of Trisomy 21 given isolated PC across the studies included in the meta-analysis was 0.2% (CI: 0.1-0.4%). In the given periods (GA and time), 66,606 (74.8%) out of 89,056 investigated fetuses met the inclusion criteria and had a normal ultrasound result without any abnormality. The Δ between our study group and the control group was 0.08% (CIΔA: 0-0.5%). Including the meta-analysis, the median of the posterior distribution of Δ between P1 and P2 was 0.08% (CIΔB: 0-0.4%) (ΔB = P1-P2).

Conclusion: The posterior distribution of Δ between P1 and P2 including the meta-analysis corresponds to showing no difference between the cases and controls (95% CIΔB: 0-0.4%). The additional risk of a fetus with an ICPC for Trisomy 21 is 97.5% likely to be lower than 0.4% (about 1/250). However, in our collective, the positive predictive value of ICPC for Down syndrome was 0.16% (about 1/625). In prenatal counseling, the additional risk should be added to the individual risk (based on maternal age, earlier screening test results, and sonographic markers) and the diagnostic options including fetal DNA and diagnostic procedures should be discussed according to the posterior individual risk.