科研之友
微信
新浪微博
Facebook
分享链接
摘要
The current gold standard for prenatal identification of heritable fetal anomalies is the use of invasive cytogenetic tests. Noninvasive prenatal determination of fetal sex using cell-free fetal DNA has been investigated as an alternative to invasive tests. In several European countries, this test has been used in routine clinical care despite the absence of clinically valid data evaluating its performance. This systematic review and meta-analysis was designed to assess the performance of cell-free fetal DNA testing for fetal sex determination and to identify covariables affecting performance. The investigators conducted a search of the PubMed database (1997-2011) to identify English-language studies reporting primary data involving humans. Reference lists in review articles were also searched. Studies with primary data suitable for analysis were identified by 2 authors who read the abstracts independently and extracted articles that met inclusion criteria. The covariables analyzed were publication year, sample type, amplification technique, Y chromosome sequence, and gestational age. The sensitivity, specificity, diagnostic odds ratio (OR), positive predictive value, negative predictive value, and area under curve (AUC) were determined to assess the performance of the test to detect Y chromosome sequences. The performances of 2 DNA amplification techniques, real-time quantitative polymerase chain reaction (RTQ-PCR) and conventional polymerase chain reaction (PCR), were compared. A total of 80 data sets extracted from 57 selected studies meeting inclusion criteria were analyzed. These data sets represented 3524 male and 3017 female fetuses. The overall performance of the test was as follows: sensitivity, 95.4% (95% confidence interval [CI], 94.7%-96.1%); specificity, 98.6% (95% CI, 98.1%-99.0%); diagnostic OR, 885; positive predictive value, 98.8%; and negative predictive value, 94.8%. The AUC was 0.993 (95% CI, 0.989-0.995), showing significant interstudy heterogeneity in test performance. Only gestational age and the DNA amplification technique were found to be significant predictors of test performance. Performance increased with gestational age: <7 weeks, AUC = 0.989 (95% CI, 0.965-0.998); 7 to 12 weeks, AUC = 0.994 (95% CI, 0.987-0.997); 13 to 20 weeks, AUC = 0.992 (95% CI, 0.983-0.996); and >20 weeks, AUC = 0.998 (95% CI, 0.990-0.999) (P = 0.02 for comparison of diagnostic ORs across age ranges). RTQ-PCR (AUC, 0.996 [95% CI, 0.993-0.998]) performed significantly better than conventional PCR (AUC, 0.988 [95% CI, 0.979-0.993]) for predicting fetal sex (P = 0.02). The sensitivity (96.0%) and specificity (99.0%) of RTQ-PCR were better than those of conventional PCR (sensitivity, 94.0%; specificity, 97.3%). The best test performance was demonstrated among women at >20 weeks' gestation (sensitivity, 99%; specificity, 99.6%). The worst test performance was in women <7 weeks' gestation (sensitivity, 74.5%; specificity, 99.1%). These findings show that the best overall performance of noninvasive fetal sex determination using RTQ-PCR is with a maternal blood sample taken at >20 weeks' gestation. Testing women before 7 weeks' gestation is unreliable.
- 出版日期2011-12
