Performance of non-invasive prenatal screening for fetal aneuploidy in twin pregnancies: a meta-analysis.

Abstract

The objectives of this study were to review the published studies of non-invasive prenatal screening (NIPS) in twin pregnancies and to evaluate the performance for screening fetal trisomies 21, 18, and 13 in twin pregnancies. Ten eligible studies were included in this meta-analysis. Data analysis, heterogeneity exploring, meta-regression, and subgroup analysis were all conducted with Meta-DiSc 1.4. Quality assessments were carried out with the Quality Assessment Tool for Diagnostic Accuracy Studies. Non-invasive prenatal screening for trisomy 21 screening in twin pregnancies showed a pooled sensitivity of 0.99 (95% CI, 0.92-1.00), a specificity of 1.00 (95% CI, 0.99-1.00), a positive likelihood ratio (LR) of 116.46 (95% CI, 53.84-251.92), a negative LR of 0.11 (95% CI, 0.06-0.22), and a diagnostic odds ratio of 1297.73 (95% CI, 438.06-3844.42). But the difference of detection rates between monozygotic and dichorionic twin pregnancies was uncertain for only a very small number of monochorionic twins available. For trisomy 18 screening, the pooled sensitivity, specificity, positive LR, negative LR, and diagnostic odds ratio were 0.85 (95% CI, 0.55-0.98), 1.00 (95% CI, 0.99-1.00), 86.11 (95% CI, 32.45-228.47), 0.25 (95% CI, 0.11-0.56), and 333.90 (95% CI, 35.15-3171.78), respectively. For trisomy 13 screening, three cases in three studies were all detected accurately. The meta-analysis results from 10 eligible studies demonstrated that NIPS has both high sensitivity and specificity for trisomy 21 screening in twin pregnancies. The results for trisomy 18 screening are less satisfactory than those for trisomy 21. The performance for trisomy 13 screening could not be determined as only three cases were identified. © 2017 John Wiley & Sons, Ltd.