The proportion of elderly pregnant women (the expected age of delivery $\ge$35 years) has increased significantly with the opening of the second- and third-child fertility policies, and older pregnant women account for more than half of the total number of those with prenatal diagnosis. With the increase in gestational age, the probability of giving birth to fetuses with abnormal chromosomes in elderly pregnant women is significantly higher than that in those of appropriate age [1, 2]. Currently, the technique of chromosome karyotype analysis is the “gold standard” of prenatal diagnosis. It can detect various chromosome abnormalities, such as aneuploidy and unbalanced rearrangement, although limited by its low resolution, time-consuming and tedious nature [3]. Fluorescence in situ hybridization (FISH) can quickly detect common fetal chromosome aneuploidy abnormalities without cell culture, with high specificity and ease of performing on a large scale. It can also detect small fragments of chromosome imbalance, but these require specific probes, and the type and number of probes are relatively limited. Therefore, known fragments belonging to specific regions are required for detecting some chromosome microdeletions and microrepeats [4].

With molecular biology and gene sequencing technology advances, chromosome microarray analysis (CMA), as a novel prenatal diagnosis method in molecular genetics, can detect and obtain numerous genome sequences at once within the entire genome and detect DNA copy number imbalance, that is, copy number variants (CNVs). It can detect submicroscopic chromosome deletions or duplications $>$50–100 kb that cannot be detected using conventional karyotype analysis without relying on cell culture and has various specimen types. It is an excellent complement to traditional karyotype analysis because of its rapidity, accuracy, high throughput, and the ability to detect small structural abnormalities that cannot be identified using “karyotype analysis” [5]. Therefore, we investigated the chromosome karyotype analysis combined with the CMA technique for the chromosomal examination of amniotic fluid specimens in elderly pregnant women to improve the detection rate of chromosomally abnormal fetuses and reduce the birth defect rate in elderly pregnancies.

The proportion of elderly pregnant women has increased with the opening of the second- and third-child policies. The older the pregnant woman is, the higher the risk of birth defects. Some studies show that the incidence of fetal aneuploidy and birth defects in elderly pregnant women are significantly higher than those in young women, and the rate of chromosome abnormality increases gradually with the increasing age [6]. Advanced age is an important indicator of prenatal diagnosis and also a high-risk factor for fetal chromosome abnormalities [7]. This study uses chromosome karyotype analysis combined with the CMA technique to examine the chromosomes of amniotic fluid samples of elderly pregnant women. Currently, chromosome karyotype analysis is the gold standard in the clinic for detecting fetal chromosome abnormalities, which can detect chromosome number and structural abnormalities ˃10 Mb. CMA currently has two detection techniques as follows: comparative genomic hybridization (CGH) and single nucleotide polymorphism (SNP). CGH is the main clinical CMA technology. Most CGH chips used in prenatal diagnosis are targeted microarrays designed for chromosome aneuploidy, typical microdeletions or microrepeats, and subtelomere or other chromosome structural rearrangements of obvious clinical significance. Compared with CGH, SNP uses arrays based on high-density oligonucleotides and can extract other clinically useful information from the genotype map. This includes single parent diploid, chimera, maternal cell contamination, and blood relationship, and can also identify triploids CGH cannot detect [8]. Here, amniocentesis, chromosome karyotype analysis, and CMA detection were performed in 210 elderly pregnant women. Overall, 26 cases of abnormal karyotypes were detected using chromosome karyotype analysis, with a detection rate of 12.38% (26/210). The total detection rate of CMA chromosome abnormalities was 24.76% (52/210), which was 12.38% higher than that of karyotype analysis. It is significantly higher than that reported by Shaffer et al. [9]. that CMA can increase the abnormal detection rate by 2.9% compared with chromosome karyotype analysis. However, 3.6% more chromosomal abnormalities can be detected than in traditional karyotype techniques; this is higher than that in the CMA technique reported by Hillman et al. [10]. The reason for the analysis may be that the cases in this study are elderly pregnant women and some elderly pregnant women also combine some high-risk factors, and old age will also increase the rate of chromosome abnormality. Second, there are more thalassemia gene carriers because Guangxi belongs to the high incidence area of thalassemia, resulting in a higher detection rate of pathogenic CNV. For chromosome number abnormalities, 16 cases of pathogenic chromosome aneuploidy were detected using CMA and karyotype analysis, including 6, 4, and 6 cases of trisomy 21, sex chromosome aneuploidy, and chimerism, respectively. CMA also detected an additional case of chimerism. The karyotype analysis result was 46 Magi Xnjue 1qhmage, whereas that of the CMA test was 46 Magi XXB and 4Q XY chimerism (the proportion of 46 Magi XX was ~15%); however, the two results were inconsistent. Based on the karyotype analysis, 22 additional cases of pathogenic CNV were detected using CMA, with an additional detection rate for pathogenic CNV of 10.48%, similar to the results of Srebniak et al. [11] and Van den Veyver et al. [12]. An Indian study indicated [13] that compared with karyotype analysis, Cmas detected an additional 3.78% copy number variation in pathogenicity and detected 3 pCNVs (13.04%) among 23 maternal age high schools. It is lower than the result of our study, which may be attributed to the fact that our study mainly targeted at older pregnant women and did not subdivide other indicators of prenatal diagnosis, leading to bias. In pregnant women who received amniotic fluid puncture with no ultrasound abnormalities, simple old age or positive aneuploidy screening, chromosome microarray analysis detected pathogenic CNVs in only 1.7% of cases [8, 14]. Wapner et al. [15] also reported 20 cases of genetic abnormalities that were not found using karyotype analysis. Of the 20 cases, 9.3% and 11.3% of pregnant women were found with chromosome abnormalities of clinical significance and uncertain clinical significance, respectively. However, CMA can not detect some balanced chromosome structural abnormalities such as balanced translocation, inversion, insertion, and gene point mutation [16]. Some studies have shown that the traditional G-banding karyotype analysis can detect abnormalities undetected using 3% CMA. Here, 5 cases were found with normal CMA results but abnormal using chromosome karyotype, including 1, 1, and 3 cases of chromosome translocation, deletion, and inversion, respectively, accounting for 2.4%, which was close to the foreign-related studies [17, 18]. Studies have shown that balanced translocation and inversion of chromosomes are important causes of reproductive abnormalities, and the rate of chromosomal abnormalities in full-term fetuses of couples carrying balanced translocation is about 10%, higher than that of the general population [19]. Balanced translocation is also associated with repeated abortion [20]. In our study, the abnormal rate of karyotype analysis combined with CMA detection was 27.14% (57/210), which was higher than that of CMA or chromosome karyotype analysis alone. The results of chromosome karyotype analysis combined with CMA were statistically significant compared with those of chromosome karyotype analysis alone (p 0.001). The results of chromosome karyotype analysis combined with CMA detection had no statistical significance compared with those of CMA detection alone (p $>$ 0.05). The results of CMA alone were statistically significant compared with those of chromosome karyotype alone (p 0.001). Kappa consistency test was used to analyze the consistency of the two methods (Kappa = 0.447). CMA has more advantages in detecting chromosome microdeletion or microduplication. Karyotype analysis can detect some chromosome structural abnormalities such as chromosomal translocation, inversion and insertion that cannot be identified by CMA. Therefore, combining karyotype analysis and CMA for prenatal diagnosis of elderly pregnant women, both of which complement each other, is crucial.

The results of 22 cases with no abnormalities in chromosomal testing and pathogenic CNV results in CMA, case A202246, were analyzed, and duplication of ~2.02 Mb was detected in sample 22q11.21, a pathogenic CNV. This duplication overlaps with the 22q11.2 duplication region, and the ClinGen database was queried to find a clear triple dose effect for the 22q11.2recurrent DiGeorge/palatine facial syndrome (DGS/VCFS) region (proximalA-B, chr22:18,912,231-20,287,208), and the individuals carrying this duplication have varying clinical symptoms. The 22q11 repeat is inherited from phenotypically normal or near-normal parents in ~70% of cases. Rosenfeld et al. [21] performed a Bayesian analysis of data from a large sample of people and showed that the 22q11.2 microrepeat has ectopic incompleteness with an ectopic rate of ~21.9%. This study’s patient had a history of adverse pregnancy and delivery with two embryonic stoppages; we performed CMA on the peripheral blood of the parents, but the microarray analysis did not reveal any clinically significant chromosomal abnormalities, and the pregnant fetus was born with pathological jaundice without significant abnormalities in appearance and is currently growing well.

A202677 detected a deletion of ~1.77 Mb in the 17q12 region of the sample, a pathogenic copy number variant. This deletion region is involved in 17q12 deletion syndrome, also known as renal cysts and diabetes syndrome; the key gene is HNF1B, which is expressed in all renal tubular epithelial cells that constitute the renal units and collecting ducts, controls and participates in membrane transport, cellular differentiation, and expression of metabolic genes, and has a role in the regulation of important renal genes such as PKHD1 and PKD2 [22]. The clinical presentation of patients with this syndrome varies widely among individuals, with the main clinical features including abnormal kidney or urogenital development (polycystic kidney, renal insufficiency), diabetes in late adolescence (MODY5), neurodevelopmental/neuropsychiatric disorders (autism, schizophrenia, anxiety, and epilepsy), developmental delay, language backwardness, and mental retardation. 17q12 deletion syndrome is dominantly inherited, with 70% and 30% of deletions being de novo variants and inherited from the parents, respectively, and the epistatic rate of 34.4%. A study [23] showed a significant correlation between enhanced renal parenchymal echo and 17q12 microdeletion syndrome. Quintero-Rivera et al. [24] reported the first case of duodenal atresia associated with 17q12 microdeletion revealing for the first time that the phenotypic spectrum of 17q12 microdeletion syndrome should include duodenal atresia. Here, the case patient had fetal ultrasound findings of bilateral renal pelvis separation, and the fetal outcome at gestational follow-up was a healthy boy delivered alive without significant abnormalities.

A204280 detected an ~910.84 kb duplication in sample Xq28, a pathogenic CNV. This duplication is involved in Xq28 duplication syndrome with the key gene MECP2, which is an X-linked neurodevelopmental disorder where patients exhibit severe psychomotor retardation, hypotonia, language development disorders, progressive muscle spasms, various seizures, and in some cases, developmental regression, ataxia, choreiform movements, sleep disorders, and recurrent respiratory infections. Males are 100% ectopic, whereas female carriers are usually unaffected or show only psychoneurological abnormalities. Subsequent lineage analysis suggested maternal inheritance, and the fetal sample Xq28 with ~910.84 kb of repeat inheritance from the mother was confirmed through CMA analysis, followed by selection for labor induction.

A210791 detected a deletion of ~8.07 Mb in sample 5p15.33p15.31, a pathogenic CNV. The deletion overlaps most of the region with the cri-du-chat syndrome (also known as catcalling syndrome), which has multiple pathogenic and possibly pathogenic reports in DECIPHER and ClinVar databases. The clinical phenotypes of these patients include intrauterine growth retardation, prematurity, feeding difficulties, short stature, peculiar facial features, mental retardation, and voice abnormalities; the deletion region contains 41 proteins, including TERT. The gene encoding TERT was mentioned in the ClinGen database as possibly having a single-dose effect (Haploinsufficiency Score: 1), but the evidence was insufficient. A duplication of 4.68 Mb was also detected in sample 10p15.3p15.1, of unknown clinical significance. This duplication was queried in the patient databases, DECIPHER and ClinVar, for 3 cases of (probable) pathogenicity reported and unqueried in the general population database, DGV; this region contains 12 protein-coding genes, including IDI2, and the ClinGen database was unqueried for the above gene/region duplication dose effect. The pregnancy outcome in this pregnancy was direct induction of labor.

A203673 detected ~2.5 Mb deletion in the sample 22q11.21, which was a pathogenic CNV. This deletion involves 22q11.2 deletion syndrome, which is mainly classified into three subtypes as follows: DGS syndrome (#188400), VCFS(#192430), and vertebral shaft abnormal facial syndrome (CAFS). Among them, DGS is mainly characterized by congenital heart disease, immunodeficiency, and hypocalcemia and is common in newborns; VCFS mainly shows cleft palate, congenital heart disease, special face, slender fingers, and mental and behavioral abnormalities, among others; CAFS mainly shows special face and heart deformities. This female patient had a history of adverse pregnancy with billet abortion, and a genealogical analysis suggested that this mutation was a de novo mutation, and the fetus was induced.

Another 17 cases detected a heterozygous deletion of ~17.19 kb in sample 16p13.3, a pathogenic CNV. This deletion contained at least 4 OMIM genes, including HBM, HBA2, HBA1, and HBQ1, of which HBA1 and HBA2 are key genes for $\alpha$-thalassemia (referred to as $\alpha$-thalassemia). Individuals with heterozygous deletion of HBA1 and HBA2 genes (–/$\alpha$$\alpha$) are those with mild $\alpha$-thalassemia, also known as $\alpha$-thalassemia gene carriers, who are usually asymptomatic or may have mild anemia manifestations. This is because Guangxi, Guangdong, is a high-prevalence area for thalassemia, and more carriers of the gene for thalassemia exist among pregnant women. In these 17 cases, one fetus was induced after diagnosing thalassemia major in the elderly pregnant women; the remaining 16 cases were delivered alive without any significant abnormalities at 6 months of follow-up.

Here, seven cases of CNV of unknown clinical significance were detected, and all seven fetuses with CNVs of unknown clinical significance were delivered alive. In our study, 7 cases with unknown clinical significance accounted for 3.33% of CNVs, higher than the result of a foreign study of 1.89% [13]. This is consistent with the report of another foreign study that VOUS accounts for no more than 5% [25]. In the case of CNVs of unknown significance, it has been suggested that this may cause significant stress and even panic among pregnant women and their families and may lead to unnecessary labor induction in some cases. Therefore, the clinical indications for prenatal diagnosis are strictly defined before performing CMA, the pregnant women and their families are fully informed of the possible outcome, and consent is obtained. Furthermore, genetic counseling is adequately conducted before the prenatal diagnosis is performed [26]. In addition, most of the older pregnant women are the second child or more, so the proportion of cesarean section is larger, further increasing the risk of pregnancy. Recently, Vimercati et al. [27] used an innovative ultrasound parameter to measure the level of scar-vesicovaginal fold distance to determine the risk of uterine rupture in pregnant women. Combined with prenatal diagnosis, this method will better predict the occurrence of adverse pregnancy outcomes.

Summarily, the detection rate of chromosome abnormalities in CMA was significantly higher than that in the routine karyotype analysis if the CMA technique was used alone in this study. However, CMA will miss some chromosomes with structural abnormalities, such as chromosome translocation and inversion, which may cause the risk of infertility and miscarriage [28]. Here, the effect of CMA in detecting chromosome aneuploidy variation is similar to that of chromosome karyotype analysis. CMA has more advantages in detecting chromosome microdeletions and microrepeats; therefore, CMA detection is recommended for elderly pregnant women, regardless of whether they are combined with other indications, but CMA can not identify some chromosome structural abnormalities such as chromosome translocation, inversion, and insertion, among other. Since CMA cannot replace karyotype analysis, it is suggested that CMA should be combined with various detection methods.

For the prenatal diagnosis of fetal amniotic fluid in elderly pregnant women, the combined application of chromosome karyotype analysis and CMA detection technology can further improve the detection rate of abnormal chromosomes and reduce the rate of missed diagnosis, thus, reducing the birth defects rate and achieving the goal of eugenics.

The data used to support the findings of this study are included within the article.

GTL, CO and WWL designed the research study. CO and WWL provided the idea for the research. GTL and WWL analyzed the data and performed the literature search. GTL, CO and WWL provided oversight and were responsible for the study organization and implementation, and writing of the manuscript. GTL, CO and WWL revised the manuscript’s intellectual content. All authors contributed to editorial changes in the manuscript. All authors read and approved the final manuscript.

All those who underwent invasive prenatal diagnosis signed an informed consent form before the procedure and were approved by the ethics committee of Medical Ethics Committee of Yulin Maternal and Child Health Hospital (Approval number: YLSFYLL2021-04-29-04). The sample-related data analyzed for this manuscript was entirely retrospective with no patient or patient-related identifiers included in the analysis.

We thank the staff in the Yulin Maternal and Child Health Hospital and Perkin Elmer Co., Ltd for their assistance in conducting this clinical research study.

This research received no external funding.

The authors declare no conflict of interest. We declare there is no conflict of interest between this research and Perkin Elmer Co., Ltd.