In this study, we confirmed that NIPS had a high sensitivity, specificity and NPV for detecting fetal trisomies 21, 18 and 13 in AMA pregnant women. In addition, the incidence of fetal trisomy 21 increased with maternal age. These findings point to the clinical significance of NIPS to detect fetal trisomies 21, 18 and 13 in AMA pregnant women and may help doctors and pregnant women to choose a suitable prenatal screening and diagnosis way.
Serological screening is widely used. At present, as for the high-risk pregnant women, fetal karyotyping with amniotic fluid cells or cord blood cells is used as the diagnosis for fetal chromosomal abnormalities. In China, all AMA pregnant women are advised to undergo prenatal diagnosis. However, with the implementation of the two-child policy, the number of AMA pregnant women has increased significantly, which has greatly increased the demand for prenatal diagnosis. However, amniotic fluid sampling or umbilical cord blood collection are invasive procedures, with the risk of miscarriage which was estimated at 0.5 to 1.0%[6,19]. There is also a risk of infection in such procedures. As a result, the overall utilization rate of both methods is low. Moreover, some pregnant women may have contraindications for invasive prenatal diagnosis, such as the high risk of inducing abortion, fever, increased tendency for bleeding, and infection. Therefore, it is needed to find prenatal screening methods that better meet the clinical needs.
NIPS is a noninvasive prenatal screening technique for fetal aneuploidies. NIPS is based on high-throughput sequencing to detect cell-free fetal DNA (cffDNA) in maternal blood. In 1997, Lo et al. found cffDNA in maternal blood and revealed that cffDNA was suitable for prenatal examination. However, it was not widely applied in clinic until the emergence of high-throughput sequencing . Bianchi et al  compared NIPS and serological screening in general population, which recruited 1,914 women with singleton pregnancies from 21 centers in USA. Each sample was tested by both methods. The positive predictive values for NIPS and standard screening were 45.5% and 4.2% for trisomy 21, and 40.0% and 8.3% for trisomy 18, respectively. NIPS showed significantly better performances than serological screening. Meanwhile, Bianchi et al  also found that the false negative rates were 0.3% and 0.2% for trisomies 21 and 18 as detected by NIPS, respectively, which were much lower than those of serological screening (3.6% and 0.6%, respectively). Similarly, in a study of 146,958 women , it was revealed that the sensitivity was 99.17%, 98.24% and 100%, that the specificity was 99.95% ,99.95% and 99.96%, that the PPV was 92.19%, 76.61% and 32.84%, and that the NPV was 99.99%, 100% and 100%, for trisomies 21,18 and 13, respectively. Using expanded non-invasive prenatal screening, Liang  demonstrated that the PPVs were 95%, 82% and 46%, for trisomies 21, 18 and 13, respectively. Those findings obtained from large size of general populations were consistent with the results of ours, indicating NIPS has similar performance and is suitable for pregnant women at AMA.
Lots of investigations demonstrated that NIPS is superior to serological screening and suitable for the detection of trisomies 21, 18 and 13 in all high risk or low risk populations, AMA or not[24-26]. Thus, the International Society for Prenatal Diagnosis (ISPD), the American College of Obstetricians and Gynecologists (ACOG), the Royal College of Obstetricians and Gynecologists (RCOG), and the American College of Medical Genetics and Genomics (ACMG), have recommended NIPS as the preferred screening method for all pregnant women. Additionally, NIPS has been included in a national policy or national program in 14 European countries . Considering the excellent efficiency of NIPS, NIPS could be promoted as the preferred screening method for AMA pregnant women. However, invasive screening methods such as amniotic fluid analysis and cord blood collection are still needed to carry out karyotype analysis for high-risk women identified by NIPS.
Since maternal age is closely associated with the incidence of fetal chromosomal abnormalities , we also studied the correlation between maternal age and the incidence of trisomy 21. Generally, the incidence increased with maternal age. This is consistent with previous report. In the present study, among the 29,343 AMA pregnant women, 37 is the optimal cut-off point for identifying the fetal trisomy 21 with an AUC of 0.638 indicating that the prevalence of trisomy 21 was significantly higher in pregnant women aged 37 or older. Therefore, AMA pregnant women less than 37 years old can choose NIPS as a priority. As for those women aged 37 or older, invasive prenatal diagnosis should be advised first.
Although NIPS is confirmed as a test with a high sensitivity and specificity in common fetal aneuploidy, the false-positive results and false-negative results can still occur. It was previously reported that NIPS had a false positive rate of 0.09%, 0.13% and 0.13% for trisomies 21, 18 and 13, respectively. Several factors might cause false-positive and false-negative results, like confined placental mosaicism (CPM)[32,33], fetal mosaicism, vanishing twin[34,35], maternal malignancy, low fetal concentration of low fetal fraction and technical or human errors. In the present investigation, the false-positive rate were 0.04%, 0.06%, 0.04% for trisomy 21 , trisomy 18 and trisomy 13, respectively. Additionally, the false-negative rate of trisomy 21, trisomy 18 and trisomy 13 were 0.89%, 0, 0, respectively. The false negative case in our study is a 39-year-old pregnant woman, whose NIPS results were low-risk at gestational age of 13 weeks. However, when she carried out regular prenatal organ screening at gestational age of 21 weeks, ultrasound revealed fetal edema and complete endocardial cushion defect. She was advised to accept invasive prenatal diagnosis and fetal trisomy 21 was confirmed. Therefore, there are still limitations in NIPS, which is a screening not a diagnostic method. Especially when the pregnant women meet any of the exclusion criteria, she should be advised to accept invasive prenatal diagnosis rather than NIPS.
We noted some shortcomings of this study. On the one hand, there were some cases without diagnosis in the high-risk population detected by NIPS. Many of these cases might have had fetal aneuploidies, especially those with ultrasound abnormalities or fetal death. Therefore, the PPV of fetal trisomies 21, 18, and 13 were likely to be higher than what were described here. On the other hand, the low incidence of trisomy 18 and trisomy 13 made it impossible to carry out an age stratification study as was done for the trisomy 21. Multicenter studies with larger sample sizes are expected in the future and that should provide additional data in support of optimizing prenatal screening and diagnosis strategies for AMA pregnant women.