Introduction

Sex chromosome abnormalities (SCAs), also known as X or Y chromosome abnormalities, can lead to dysplasia of sexual organs in both males and females1. SCAs are relatively common, with an estimated prevalence ranging from 250 to 400 per 100,000 live births, including 45,X (monosomy X or Turner syndrome) and various sex chromosome trisomies: 47,XXX (triple X syndrome), 47,XXY (Klinefelter syndrome), and 47,XYY (XYY syndrome)2. These conditions may result in a range of clinical symptoms such as short stature, infertility, behavioral problems, and abnormalites in sexual and cognitive development3,4,5. Currently, the majority of children with SCAs remain undiagnosed until puberty. Prenatal diagnose methods such as amniocentesis and chorionic villus sampling, can be used to detect SCAs before birth, but these procedures are invasive, costly, and carry a risk of miscarrige. Consequently, these procedures are typically reserved for high-risk pregnancies, such as those involving advanced maternal age or suspected trisomy 216,7.

For many years, the cornerstones of fetal aneuploidy screening were serum biochemical markers and ultrasound measurements. Current screening algorithms primarily focus on trisomies 21 and 18. Due to the nonspecific nature of serum biochemical screening combined with nuchal translucency (NT) measurements, many pregnancies that yield false-positive results for trisomy 21 or trisomy 18 may actually be affected by other chromosomal abnormalities, including deletions or duplications, sex chromosome aneuploidies, mosaicism, triploidy, and rare trisomies8,9,10,11. SCAs are associated with lower levels of pregnancy-associated plasma protein-A (PAPP-A) and increased NT in the first trimester, as well as elevated levels of free beta subunit of human chorionic gonadotropin (fβhCG), alpha-fetoprotein (AFP), progesterone, and inhibin A in the second trimester6,7,12. However, information regarding the tendencies of serum biochemical markers for SCAs remains limited. The efficacy of serum biochemical screening for SCAs is discordant in many countries. Some studies have indicated that serum biochemical screening detects only a minority of SCAs, whereas others have shown that it can detect most SCAs2,12. Additionally, the clinical performance of serum biochemical screening for various types of SCAs is unknown.

To explore these issues, we compared pregnancies with identified SCA with matched non-SCA controls to evaluate the tendency of serum biochemical markers for SCAs. We also analysed the clinical performance of serum biochemical screening for 45,X, 47,XXX, 47,XXY, 47,XYY, and mosaic sex chromosome abnormalities.

Materials and methods

Study design

This retrospective analysis included women with singleton pregnancies who underwent serum biochemical screening during their second trimester. We collected and analysed the results of pregnancy outcomes and serum biochemical screening tests, and then eveluated the performance of serum biochemical markers in each SCA group though a case–control study. For this purpose, 100 non-SCA pregnancies were matched with each SCA case. For trisomy 21, a risk value of ≥ 1 in 270 was classified as high risk; values ranging from 1 in 271 to 1 in 1000 indicated intermediate risk; and values< 1 in 1000 indicated low risk. Similarly, for trisomy 18, a value of ≥ 1 in 350 indicated high risk; values ranging from 1 in 351 to 1 in 1000 represented intermediate risk; while values < 1 in 1000 indicated low risk13,14. The high-risk group proceeded to ungergo invasive prenatal diagnosis, and the intermediate-risk group underwent non-invasive prenatal screening (NIPS). Consequently, we analysed and compared the efficacy of serum biochemical screening acrosis each SCA group.

Serum biochemical screening tests were conducted at the Department of Medical Genetics / Prenatal Diagnosis Centre of the West China Second University Hospital from January 2009 to December 2018. Written informed consent was obtained from all participants, and approval was secured from the institutional review board. All methods were performed in accordance with the relevant guidelines and regulations. The inclusion criteria comprised singleton living fetuses, gestational age ranging from 15 weeks to 20 weeks and 6 days in the second trimester, and maternal age of 16 years or older. The exclusion criteria included cases involving vanishing twin syndrome, multiple pregnancies (twins or higher-order), as well as a family history of congenital deformities or chromosomal abnormalities. Amniocentesis was employed to assess the chromosomal status of the fetus. The indications for amniocentesis encompassed high-risk pregnant women identified through serum biochemical prenatal screening or NIPS, pregnant women with an expected delivery age of 35 years or older, and any other cases deemed necessary by healthcare providers for prenatal diagnosis. In situations of stillbirth, pregnancy termination, or spontaneous pregnancy loss, the fetal chromosomal status was ascertained through tissue sampling. The fetal chromosomal status of expectant mothers who did not undergo amniocentesis was assessed through telephone follow-up.

Serum biochemical screening

The procedures of serum biochemical indicator measurements and the assessment of risk of trisomy 21 and trisomy 18 in the second trimester were described in our previous study15.

Diagnosis of fetal chromosome results

Fetal chromosomes were validated using amniotic fluid or fetal tissue samples by copy number variation sequencing, chromosomal microarray analysis, or karyotyping. Meanwhile, the mosaicism was confirmed via fluorescence in situ hybridization (FISH). Each sample was independently analyzed by two experimenters. In FISH analysis, if the proportion of abnormal cells for a specific indicator is ≥ 10%, it indicates that the indicator is abnormal (if the percentage of abnormal cells is between 10 and 60%, this indicates the presence of mosaicism).

Statistical analyses

Descriptive data are presented as n (%) for categorical variables and median (range) for continuous variables. For each SCA group, differences in serum biochemical markers between patients with SCAs and matched controls were assessed using the Mann–Whitney U test. Statistical significance was set at P < 0.05. The data analysis was conducted using the statistical software program SPSS 23.0 (SPSS Inc., Chicago, IL, USA).

Results

We collected 159 identified SCA cases, including 23 cases of 45,X; 19 cases of 47,XXX; 36 cases of 47,XXY; 17 cases of 47,XYY; and 64 cases of mosaic sex chromosome abnormalities. Among the 159 patients, screening methods emplpoyed were as follows: double screening for 110 patients, triple screening using inhibin A for 22 patients, triple screening utilizing uE3 for 15 patients, and quadruple screening for the remaining 12 patients. Specifically, AFP and fβhCG detected 159 cases (Figs. 1 and 2), uE3 detected 27 cases (Fig. 3), and inhibin A detected 34 cases (Fig. 4). In addition, we assessed 100 randomly matched controls for each SCA based on their serum biochemical markers. Maternal age, gestational age, and maternal weight of the various study groups are shown in Table 1. The karyotype results for 159 cases and outcome of the pregnancies are detailed in Table 2.

Table 1 Median (range) for maternal age, gestational age and maternal weight in the various study groups.
Full size table
Table 2 Karyotyping results of 159 cases and pregnancy outcome.
Full size table
Table 3 Performance of serum biochemical screening.
Full size table

45,X

In the 45,X pregnancies, the distribution of AFP (mMoM = 0.71) was significantly lower than that in the control group (P < 0.005), and the AFP MoM was ≤ the 5th percentile of the controls in 5 (21.74%) cases and ≥ the 95th percentile of the controls in 1 (4.35%) case (Fig. 1A). The median MoM (0.62) of fβhCG was not significantly different from that of the control group (P = 0.998), while the fβhCG MoM was ≤ the 5th percentile of the controls in 6 (26.09%) cases and ≥ the 95th percentile of the controls in 1 (4.35%) case (Fig. 2A). The median MoM (0.70) of uE3 was not significantly different from that of the control group (P = 0.503), and uE3 was ≤ the 5th percentile of the controls in 1 (50%) case (Fig. 3). The median MoM (0.74) of inhibin A was not significantly different from that of the control group (P = 0.617), and inhibin A was ≤ the 5th percentile of the controls in 1 (25%) case (Fig. 4). Overall, 11 patients (47.83%) showed abnormally increased or decreased serum biochemical marker levels. Through serum biochemical screening, 3(13.04%), 7(30.43%), and 13 (56.52%) patients were classified as high-, intermediate-, and low-risk, respectively (Table 3).

47,XXX

In the 47,XXX pregnancies, the median MoM of AFP (0.98) was not significantly different from that of the control group (P = 0.989). The AFP MoM was ≤ the 5th percentile of the controls in 4(21.05%) cases and ≥ the 95th percentile of the controls in 3(15.79%) cases (Fig. 1B). The distribution of fβhCG (mMoM = 2.12) was significantly different from that of the control group (P < 0.005) and the fβhCG MoM was ≥ the 95th percentile of the controls in 6 (31.58%) cases and ≤ the 5th percentile of the controls in 1 (5.26%) case (Fig. 2B). The distribution of uE3 (mMoM = 0.78) was significantly different from that of the control group (P < 0.05). The median MoM of inhibin A (1.43) was not significantly different from that of the control group (P = 0.182), and inhibin A was ≥ the 95th percentile of the controls in 2 (28.57%) cases (Fig. 4). Overall, 12 patients (63.16%) showed abnormally elevated or decreased serum biochemical marker levels. Through serum biochemical screening, 10 (52.63%), 5 (26.32%), and 4 (21.05%) patients were classified as high-, intermediate-, and low-risk, respectively (Table 3).

47,XXY

In 47,XXY pregnancies, the median MoM of AFP (1.05) was not significantly different from that of the control group (P = 0.534), and the AFP MoM was ≥ the 95th percentile of the controls in 4 (11.11%) cases (Fig. 1C). The distribution of fβhCG (mMoM = 1.97) was significantly different from that of the control group (P < 0.001), and the fβhCG MoM was ≥ the 95th percentile of the controls in 6 (16.67%) cases and ≤ the 5th percentile of the controls in 2 (5.56%) case (Fig. 2C). The distribution of uE3 (1.19mMoM) was significantly different from that of the control group (P < 0.05). The median MoM of inhibin A (1.53) was significantly different from that of the control group (P < 0.05), and inhibin A was ≥ the 95th percentile of the controls in 1 (12.50%) case (Fig. 4). Overall, 10 patients (27.78%) showed abnormally elevated or decreased serum biochemical marker levels. Through serum biochemical screening, 9 (25.00%), 14 (38.89%), and 13 (36.11%) patients were classified as high-, intermediate-, and low-risk, respectively (Table 3).

47,XYY

In the 47,XYY pregnancies, the median MoM of AFP (1.00) was not significantly different from that of the control group (P = 0.534). The AFP MoM was ≥ the 95th percentile of the controls in 3 (17.65%) cases and ≤ the 5th percentile of the controls in 2 (11.76%) cases (Fig. 1D). The median MoM of fβhCG (2.60) was also not significantly different from that of the control group (P = 0.332), and the fβhCG MoM was ≥ the 95th percentile of the controls in 4 (23.53%) cases and ≤ the 5th percentile of the controls in 4 (23.53%) cases (Fig. 2D). The distribution of uE3 (mMoM = 0.75) was significantly different from that of the control group (P < 0.05). The median MoM of inhibin A (1.64) was not significantly different from that of the control group (P = 1.61), and inhibin A was ≥ the 95th percentile of the controls in 2 (50.00%) patients (Fig. 4). Overall, 11 patients (64.71%) showed abnormally increased or decreased serum biochemical marker levels. Through serum biochemical screening, 4 (23.53%), 8 (47.06%), and 5 (29.41%) patients were classified as high-, intermediate-, and low-risk, respectively (Table 3).

Mosaic sex chromosome abnormality

In the pregnancies with mosaic sex chromosome abnormalities, the distribution of AFP (mMoM = 0.90) was significantly different from that of the control group (P < 0.005), and the AFP MoM was ≤ the 5th percentile of the controls in 6 (9.38%) cases and ≥ the 95th percentile of the controls in 1 (1.56%) case (Fig. 1E). The median MoM (2.65) and distribution of fβhCG were both significantly different from that of the control group (P < 0.001); the fβhCG MoM was ≥ the 95th percentile of the controls in 27 (42.19%) cases and ≤ the 5th percentile of the controls in 10 (15.63%) cases (Fig. 2E). The median MoM (1.06) of uE3 was not significantly different from that of the control group (P = 0.814), and uE3 was ≤ the 5th percentile of the controls in 1 (10%) case (Fig. 3). The median MoM (0.71) of inhibin A was not significantly different from that of the control group (P = 0.182), and inhibin A was ≥ the 95th percentile of the controls in 1 (11.11%) case (Fig. 4). Overall, 33 patients (51.56%) showed abnormally increased or decreased serum biochemical marker levels. Through serum biochemical screening, 35 (54.69%), 15 (23.44%), and 14 (21.88%) patients were classified as high-, intermediate-, and low-risk, respectively (Table 3).

Fig. 1
figure 1

AFP and gestational age in SCAs. The abscissa represents gestational age (days) and the ordinate represents the AFP MoM value. The solid black line indicates the median AFP MoM in the control group and black dotted lines indicate the 5th and 95th percentiles of the AFP MoM in the control group. (A) 45,X; (B) 47,XXX; (C) 47,XXY; (D) 47,XYY; (E) mosaic sex chromosome abnormalities.

Full size image
Fig. 2
figure 2

fβhCG and gestational age in SCAs. The abscissa represents gestational age (days) and the ordinate represents the fβhCG MoM value. The solid black line illustrates the median fβhCG MoM in the control group and black dotted lines indicate the 5th and 95th percentiles of the fβhCG MoM in the control group. (A) 45,X; (B) 47,XXX; (C) 47,XXY; (D) 47,XYY; (E) mosaic sex chromosome abnormalities.

Full size image
Fig. 3
figure 3

uE3 and gestational age in SCAs. The abscissa represents gestational age (days) and the ordinate represents the uE3 MoM value. The solid black line indicates the median uE3 MoM in the control group and black dotted lines indicate the 5th and 95th percentiles of the uE3 MoM in the control group.

Full size image
Fig. 4
figure 4

Inhibin A and gestational age in SCAs. The abscissa represents gestational age (days) and the ordinate represents the inhibin A MoM value. The solid black line indicates the median inhibin A MoM in the control group and black dotted lines indicate the 5th and 95th percentiles of the inhibin A MoM in the control group.

Full size image

Performance of serum biochemical screening

The high-risk proportion of SCAs identified through serum biochemical screening for trisomy 21 or 18 was 38.36% (61 of 159), Table 3. The serum biochemical screening revealed that the high-risk proportions of 45,X, 47,XXX, 47,XXY, 47,XYY, and mosaic sex chromosomal abnormalities were 13.04%, 52.63%, 25.00%, 23.53%, and 54.69%, respectively. There was a significant difference in the high-risk proportion among the five groups (χ2 = 19.447, P < 0.001). The proportions of intermediate risk for 45,X, 47,XXX, 47,XXY, 47,XYY, and mosaic sex chromosomal abnormalities were 30.34%, 26.32%, 38.89%, 47.06%, and 23.44%, respectively. The combined high- and intermediate-risk proportions of 45,X, 47,XXX, 47,XXY, 47,XYY, and mosaic sex chromosomal abnormalities were reported at 43.48%, 78.95%, 63.89%, 70.59%, and 78.13%, respectively. There was a significant difference in the distribution of high- and intermediate-risk patients among the five groups (χ2 = 10.28, P < 0.05). Overall, the high- and intermediate-risk proportions of fetal SCAs in serum biochemical screening for trisomy 21 and trisomy 18 were 69.18%.

Discussion

In pregnancies with trisomy 21, second trimester serum levels of fβhCG and inhibin A are significantly elevated, while those of AFP and uE3 are decreased. In pregnancies affected by trisomy 18, the serum levels of hCG, AFP, uE3, and inhibin A are significantly reduced7. Previous studies have indicated that changes in serum biochemical markers during the second trimester for both 45,X pregnancies and trisomy 21 pregnancies exhibit similarities; for example, 45,X was associated with lower levels of uE37. Possibly due to the limited number of cases analyzed in our study, we observed that the median MoM of the four serum biochemical markers in 45,X were consistently lower than those in the control group, and two out of three (66.67%) high-risk patients were identified as being at a high risk for trisomy 18. Changes in serum biochemical markers observed in 45,X pregnancies were consistent with those in trisomy 18 pregnancies which manifested as growth retardation. For 47,XXX, we found that the median MoM of fβhCG was increased and that of uE3 was decreased. Changes in serum biochemical markers in 47,XXX pregnancies were comparable to those observed in trisomy 21 pregnancies. Serum biochemical screening indicated that the high-risk proportion for trisomy 21 among 47,XXX pregnancies was 47.37%(9/19). Although results of serum biochemical screening of 47,XXY showed that the high-risk proportion was only 25%, we found that 4 (11.11%) and 6 (16.67%) cases were > the 95th percentile of AFP and fβhCG, respectively. In addition, the median MoM of fβhCG, uE3, and inhibin A were all significantly elevated compared to those of controls. Previous research has demonstrated that elevated levels of AFP, fβhCG, and inhibin A in structurally normal pregnancies are associated with an increased risk of preeclampsia, fetal demise, and intrauterine growth restriction7,16,17. We hypothesize that the increase in these serum biochemical markers is related to 47,XXY pregnancies. For 47,XYY, the high-risk proportion was 23.53% and high-risk pregnancies were all at a high risk for trisomy 21; meanwhile, 11 cases (64.71%) were > the 95th percentile or < the 5th percentile of AFP, fβhCG, or inhibin A. Mosaicism occurs more frequently in SCAs than in autosomal aneuploidies, and mosaicism involving more than two cell lines is also more prevalent within SCAs18. We found that the median MoM of AFP was decreased and that of fβhCG was increased in mosaic sex chromosome defects, while those of uE3 and inhinbin A remained unchanged. Changes in serum biochemical markers in the double screening for trisomy 21 and mosaic sex chromosomes were consistent19. Therefore, the high-risk proportion (54.69%) of mosaic sex chromosomal abnormalities was higher than that of the other four groups.

The changes in serum biochemical markers were different in each SCA group. For instance, while prenatal fβhCG levels were significantly elevated in 47,XXX, 47,XXY, and mosaic chromosome abnormalities, AFP levels showed no correlation with 47,XXX and 47,XXY. In addition, the levels of uE3 in the 45,X and mosaic sex chromosomes remained unchanged; 47,XXY had higher levels, and the other two groups had lower levels. While the level of inhibin A was higher in 47,XXY, it remained unchanged across the other groups. This also takes into account the fact that, even though there were numerous statistically significant differences in serum biochemical markers between matched controls and fetal SCAs, these differences were not always clinically significant for the majority of SCA cases, which remained well within normal ranges for these markers during pregnancy. Determining discriminating thresholds proved to be challenging due to the variability in serum biochemical markers among pregnant women with SCAs fetuses usually ranged from normal to either elevated or decreased levels. It is challenging to determine discriminating criteria since serum biochemical marker variance across pregnancies with SCAs usually ranged from normal to either increased or decreased levels. However, serum biochemical markers of fetal sex chromosomal abnormalities were usually > the 95th percentile or < the 5th percentile, and the proportion ranged from 27.78 to 64.71%. These results suggest that fetal sex chromosome abnormalities are correlated with abnormally increased or decreased serum biochemical markers.

Serum biochemical screening can identify approximately 60–80% of trisomy 21 pregnancies with a false positive rate of 5% and approximately 80–85% of trisomy 18 pregnancies with a false positive rate ranging from 1% to 5% during the second trimester6,7,13,19. Previous studies have suggested that this screening test accidentally detected a portion of sex chromosomal abnormalities3,6,9. Our study indicated that fetal SCAs exhibited a high-risk proportion of serum biochemical screening for both trisomy 21 and trisomy 18, ranging from 13.04% to 54.69%, with an overall high-risk proportion of 38.36%. In addition to identifying fetal trisomy 21 and trisomy 18, serum biochemical screening procedures can accidentally identify some fetal SCAs. These pregnant women with sex chromosome abnormalities fetuses chose to continue the pregnancy or induce labor according to their personal preferences after genetic counseling.

The most reliable screening method for common autosomal aneuploidies is currently well-established to be non-invasive prenatal screening (NIPS), which is based on cell-free DNA in maternal plasma and the target diseases include trisomy 21, trisomy 18, and trisomy 13, as well as the sex chromosomal aneuploidies, surpassing the performance of serum biochemical screening6,18,20,21. Nevertheless, because to its high cost, NIPS is unlikely to be employed for population-wide regular screening. Recent studies have suggested that the best screening model for trisomy 21 and trisomy 18 was to offer NIPS contingent on the results of first-line screening using serum biochemical screening9,22,23,24. Currently, this contingent screening model is recommended in China and other countries. According to this concept, all cases in the high-risk group should undergo invasive testing, whereas NIPS should be recommended for those in the intermediate-risk group, and invasive testing should then be performed for those who receive a screen-positive result13,14. Our study showed that the intermediate risk proportion of 45,X, 47,XXX, 47,XXY, 47,XYY, and mosaic sex chromosomal abnormalities were 30.43%, 26.32%, 38.89%, 47.06%, and 23.44%, respectively. On this basis, we assume that the proportion of intermediate-risk cases receiving NIPS is 100% and the proportion of positive detection of SCAs by NIPS is 100%. In other words, through contingent screening, the prenatal diagnosis proportion of 45,X, 47,XXX, 47,XXY, 47,XYY, and mosaic sex chromosomal abnormalities were 43.48%, 78.95%, 63.89%, 70.59%, and 78.13%, respectively. Approximately 69.18% of overall SCAs were diagnosed.

Theoretically, contingent screening did not increase the false-positive rate through serum biochemical screening for invasive tests, but NIPS for fetal SCAs had its own false-positive rate. According to a prior study, NIPS, which uses enormous parallel genomic sequencing, has a low false-positive rate (< 1%) when screening for SCA in fetuses25. Consequently, the total false-positive rate of contingent screening was comparable to serum biochemical screening; nevertheless, different study groups had variable detection rates in detecting sex chromosomal aneuploidy utilizing NIPS. In a meta-analysis, the false-positive and detection rates of NIPS in singleton pregnancies were 0.23% (95% CI, 0.14–0.34%) and 90.3% (95% CI, 85.7–94.2%), respectively, for monosomy X; and 0.14% (95% CI, 0.06–0.24%) and 93.0% (95% CI, 85.8–97.8%) for sex chromosome aneuploidies other than monosomy X26. An additional investigation found that with a 0.3% false-positive rate, the detection rate of SCAs ranged from 82 to 90%18. In other words, while we speculated good performance of contingent screening against the backdrop of a 100% detection rate of SCAs through NIPS, this good performance may be lower than in practice.

In view of the varying testing capabilities and economic conditions in different regions of China, the current contingent screening strategy still relies on traditional serum biochemical screening as the first-line screening. Our study explores the trends of serum biochemical markers for SCAs and evaluating the clinical performance of SCAs in serum biochemical screening during the second trimester, thereby providing additional insights for interpreting clinical report.

Conclusions

Fetal sex chromosome abnormalities correlate with significantly elevated or decreased levels of serum biochemical markers, which also indicate an increased risk for trisomy 21 or trisomy 18. Typically, serum biochemical markers indicative of fetal sex chromosomal abnormalities fall above the 95th percentile or below the 5th percentile. The majority of fetal SCAs were identified as high- and intermediate-risk through serum biochemical screening, suggesting that fetal SCAs may be incidental findings within the framework of the current contingent screening strategy.