Materials and Methods
Case Series
In this retrospective study, we retrieved karyotyping and CMA results from POC specimens with clinical indications of “RPL,” “multiple miscarriages (>2),” “multiple spontaneous abortions (×2–5),” and “fetal demise or intrauterine fetal demise” referred to the Yale Clinical Cytogenetics Laboratory during a 16-year period of 2007 to 2023.[15 ] A small portion of cases was also noted with suspected fetal anomalies (n = 43), preterm premature rupture of the membranes (PPROM; n = 11), holoprosencephaly/anencephaly/acrania (n = 5), cystic hygroma (n = 4), and neural tube defect/spina bifida (n = 3). Women of RPL were divided into two groups: an SA group with cytogenomic analysis performed once or only one result, and a CA group with cytogenomic analyses performed on consecutive miscarriages. This project was categorized as a chart review retrospective case study and deemed exempt from institutional review board (IRB) approval and granted waiver of consent based on the policy of the Yale University IRB.
Karyotyping and Chromosome Microarray Analysis
Cell culture of chorionic villi and/or fetal skin tissue dissected from POC specimens and chromosome analysis were performed following the laboratory's standardized procedures.[16 ] For cases with cell culture failure or a normal karyotype, CMA was recommended and performed upon referring physician approval and insurance prior authorization. Genomic DNA was extracted from dissected POC tissues using the Gentra Puregene Kit (Qiagen, Valencia, CA). CMA was performed using the oligonucleotide array comparative genomic hybridization (aCGH, SurePrint G3 Human 8 × 60K, Agilent Inc, Santa Clara, CA) or the SNP-based OncoScan microarray assay (OMA, Affymetrix Inc. Santa Clara, CA) following manufacturers' protocols and laboratory's standardized procedures.[5 ]
[17 ] The aCGH data were processed through Agilent's CytoGenomics Software; the OMA data were analyzed using Affymetrix's Chromosome Analysis Suite version 3.3. The detected genomic imbalances and pCNVs were reported based on the February 2009 assembly of human genome (GRCh37/hg19).
Statistical Analysis
The abnormalities were categorized into numerical and structural chromosomal abnormalities. The numerical chromosomal abnormalities were further divided into simple aneuploidies (common/rare autosomal trisomies and sex chromosome aneuploidies), compound aneuploidies, polyploidies (triploidy and tetraploidy), and mosaic abnormalities. The structural chromosomal abnormalities were further divided into chromosomal rearrangements and pCNVs. To compare the spectrum of abnormalities between SA and CA groups, a chi-square test was performed to evaluate the equality of proportions of abnormalities at a 95% confidence level. The Pearson chi-square test was applied to determine the likelihood of the observed distribution fitting the expected distribution in the CA group. Categorical variables were presented as a percentage of total cases (%) and a chi-square test value p < 0.05 was accepted as statistically significant.
Results
Over a 16-year period of 2007 to 2023, cytogenetic analysis was performed on 4,674 POC specimens, which included 3,797 cases (81%) referred for SPL and 877 cases (19%) referred for RPL. For the 877 POC specimens by RPL, 113 specimens (13%) were not studied due to the absence of fetal tissue and the presence of only maternal decidua, 26 specimens (3%) yielded no result due to culture failure, 269 specimens (31%) from 266 women (including three twin pregnancies) with a single result from their miscarriages were collected in the SA group, and 469 specimens (53%) of consecutive miscarriages from 225 women were collected in the CA group ([Fig. 1A ]). The maternal age at the time of first cytogenetic testing for women in the SA and CA groups showed a similar distribution with 2 and 9% collected at 21 to 25 years, 16 and 17% at 26 to 30 years, 32 and 40% at 31 to 35 years, and 36 and 25% at 36 to 40 years, respectively ([Fig. 1B ]).
Fig. 1 Demographics of RPL cases in the SA and CA groups. (A ) Proportions of SPL and RPL cases from a total of 4,674 POC specimens. The 877 RPL cases were divided into no study (NS), single analysis (SA), and consecutive analysis (CA) groups. (B ) Maternal age distribution at the time of first cytogenetic analysis. (C ) Time intervals between first and second or third miscarriages from the CA group. (D ) Patterns of normal (white circle, number inside for observed cases ) versus abnormal (gray circle ) cytogenetic findings in the CA group with observed an expected distributions from the first and second miscarriages.
Cytogenomic Findings in the Single Analysis Group
The results of karyotyping and CMA from the SA group are listed in [Table 1 ]. Of these 269 POC specimens (three twin miscarriages) from 266 women, normal karyotypes were noted in 98 cases (36%) and chromosomal abnormalities were noted in 171 cases (64%). Simple aneuploidies, compound aneuploidies, polyploidies, structural abnormalities/pCNVs, and mosaic abnormalities were detected in 109 cases (41%), five cases (2%), 35 cases (13%), 10 cases (4%), and 12 cases (4%) respectively. The most frequently observed numerical abnormalities were trisomy 16 (33/269, 12%), triploidy (29/269, 11%), trisomy 22 (19/269, 7%), monosomy X (14/269, 5%), trisomy 15 (13/269, 5%), trisomy 21 (10/269, 4%), and tetraploidy (6/269, 2%). Other autosomal aneuploidies included trisomies 9 and 14 in four cases each; trisomy 4 in three cases; trisomy 12 in two cases; and trisomies 2, 3, 8, 17, and 20 and monosomy 21 in one case each. Compound aneuploidy of two to three chromosomes was seen in five cases involving chromosomes X, 4, 8, 15, 16, 20, 21, and 22. Mosaic abnormalities in 12 cases were aneuploidy or polyploidy with a normal karyotype.
Table 1
Cytogenomic abnormalities detected in SA and CA groups of RPL
RPL-SA
%
RPL-CA
%
RPL[a ]
%
Normal cases
98
36%
227
48%
325
44%
Abnormal cases
171
64%[b ]
242
52%[b ]
413
56%[b ]
Numerical abnormalities
149
55%[b ]
207
44%[b ]
356
48%[b ]
Simple aneuploidy
109
41%
160
34%
269
36%
Common autosomal trisomies
76
28%
103
22%
179
24%
47, + 16
33
12%[b ]
27
6%[b ]
60
8%[b ]
47, + 22
19
7%
25
5%
44
6%
47, + 15
13
5%
14
3%
27
4%
47, + 21
10
4%
11
2%
21
3%
47, + 13
1
0
14
3%
15
2%
47, + 18
0
0
12
3%
12
2%
Rare autosomal trisomies
19
7%
33
7%
52
7%
Sex chromosomes
14
5%
24
5%
38
5%
45,X
14
5%
22
5%
36
5%
47,XXY
0
0
2
0
2
0
Compound aneuploidy
5
2%
11
2%
16
2%
Polyploidy
35
13%
36
8%
71
10%
Triploidy (3n)
29
11%
30
6%
59
8%
Tetraploidy (4n)
6
2%
6
1%
12
2%
Structural abnormalities
10
4%
19
4%
29
4%
Chromosomal rearrangements
7
3%
17
4%
24
3%
pCNVs
3
1%
2
<1%
5
1%
Mosaic abnormalities
12
4%
16
3%
28
4%
Total
269
469
738
Abbreviations: CA, consecutive analysis; RPL, recurrent pregnancy loss; SA, single analysis.
Note: Bold signifies major categories of cytogenomic findings.
a Data from SA + CA.
b Significant difference noted between the SA and CA groups, but no difference between the CA and RPL groups.
Structural chromosomal abnormalities detected in seven cases and pCNVs in three cases are shown in [Table 2 ]. The most frequently seen recurrent structural abnormalities were Robertsonian translocations resulting in gains of 14q in three cases. One case had an extra Robertsonian translocation of 14q/14q for four copies of 14q, and two cases showed three copies of 14q from Robertsonian translocations of 14q/14q and 14q/22q. Four cases were detected with unbalanced rearrangements involving chromosomes 1q/11q, 2q, 8q/21p, and 18p/q. CMA defined a 53.352 Mb duplication of 8p23.3q12.1 and a 22.684 Mb deletion of 21p11.2q22.11 in a male complement, likely the result of a derivative chromosome from an 8q/21q translocation ([Fig. 2A ]). CMA detected pCNVs in three cases. One case had a 1.064 Mb duplication at 16p11.2 which encompasses 27 OMIM genes including the KIF22 (OMIM#603213), PRRT2 (OMIM#614386), TLCD3B (OMIM#615175), ALDOA (OMIM#103850), TBX6 (OMIM#602427), and CORO1A (OMIM#605000) morbid genes ([Fig. 2B ]). This duplication overlaps with the dosage-sensitive 16p11.2 region (proximal, BP4–BP5) (includes TBX6 ) and is diagnostic for chromosome 16p11.2 duplication syndrome (OMIM#614671). The second case had a 1.411 Mb deletion at 17q12 which encompasses 14 OMIM genes including the ZNHIT3 (OMIM#604500), PIGW (OMIM#610275), ACACA (OMIM# 200350), and HNF1B (OMIM#189907) morbid genes. This deletion includes the dosage-sensitive 17q12 region (includes HNF1B ) and is diagnostic for chromosome 17q12 deletion syndrome (OMIM#614527). The third case had a 7.6 Mb deletion at Xp22.33p22.31 in a male complement which encompasses 27 OMIM genes, including the SHOX (OMIM#312865), CSF2RA (OMIM#306250), XG (OMIM#314700), ARSL (OMIM#302950), NLGN4X (OMIM#300427), and STS (OMIM#300747) morbid genes. This deletion overlaps with the dosage-sensitive Xp22.31 recurrent region (includes STS), and deletion of this region is associated with X-linked recessive ichthyosis (OMIM#308100).
Table 2
Structural abnormalities and pCNVs detected in SA and CA groups
Case No.
Karyotype/CMA (GRCh37/hg19)
SA
SA143
46,XY,der(1)t(1;11)(q43;q14.1)mat
SA173
46,XY,add(2)(q37.3)
SA139
46,XX,+14,rob(14;14)(q10;q10)
SA183
47,XX,+rob(14;14)(q10;q10)
SA51
46,XY,+14,rob(14;22)(q10;q10)
SA231
mos 46,XX,del(18)(p10)[11]/46,XX,i(18)(q10)[4]
SA26
arr 8p23.3q12.1(172416_56524696)x3,21p11.2q22.11(9648314_32332613)x1
SA35
46,XX.arr 16p11.2(29133476_30197490)x3
SA209
46,XX.arr 17q12(34832202_36243169)x1
SA215
arr Xp22.33p22.31(177941_7792383)x0
CA
CA37
46,X,-X,+der(6)t(X;6)(q13;q16.2)mat
CA11
46,XX,t(3;5)(q25;q22).arr(X,1–22)x2
CA136
47,XX,t(6;9)(p21.1;q34),+14
CA149
46,XY,+der(7)t(7;15)(q31;q11.2),−15
CA149
46,XX,+der(7)t(7;15)(q31;q11.2),−15
CA166
mos 46,XY,add(8)(p11.2)[7]dn/46,XY[13]
CA128
46,XY,i(8)(q10),der(9)t(8;9)(p11;p24)
CA163
48,XX,+9,rob(14;14)(q10;q10),+15
CA213
47,XX,t(11;22)(q23;q11.2),+16
CA213
47,XX,+der(22)t(11;22)(q23;q11.2).arr 11q23.3q25(116690577_134938847)x3,22q11.1q11.21(16054712_20747763)x3
CA157
46,XY,del(13)(q13q31).arr 13q13.1q31.1(32913501_82562672)x1
CA207
46,XY,del(13)(q11)
CA190
46,XX,der(18)t(18;21)(q23;q22.2)mat
CA151
46,XY,der(20)t(11;20)(q21;p13)pat
CA196
46,XY,i(22)(q10)
CA50
47,XX,+mar.arr 22q11.1q11.21(16054712_21968221)x4
CA26
mos 46,X,r(X)[45]/45,X[4].arr Xq27.1q28(139012778–155260560)x1
CA210
arr 7q11.23(72726378_74139531)x1
CA134
46,XY.arr Xp22.31(6458165_8116400)x0
Abbreviations: CA, consecutive analysis; CMA, chromosome microarray analysis; pCNV, pathogenic copy number variants; SA, single analysis.
Fig. 2 Chromosome microarray results on three POC specimens. (A ) A 53.352 Mb duplication of 8p23.3q12.1 and a 22.684 Mb deletion of 21p11.2q22.11 resulted from a derivative chromosome 8. (B ) A 1.064 Mb duplication at 16p11.2. (C ) A 1.66 Mb deletion at Xp22.31. Top panel shows log2 ratio, and lower panel shows biallelic frequency, and bottom panel (in B/C) shows genomic coordinates and genes.
Cytogenomic Findings in the Consecutive Analysis Group
Cytogenomic results on 469 POC specimens from consecutive miscarriages of 225 women in the CA group are listed in [Table 1 ]. Normal karyotypes were noted in 227 cases (48%), and chromosome abnormalities were detected in 242 cases (52%). Simple aneuploidies, compound aneuploidies, polyploidies, structural rearrangements/pCNVs, and mosaic abnormalities were detected in 160 cases (34%), 11 cases (2%), 36 cases (8%), 19 cases (4%), and 16 cases (3%), respectively. The most frequently observed numerical abnormalities were triploidy (30/469, 6%), trisomy 16 (27/469, 6%), trisomy 22 (25/469, 5%), monosomy X (22/469, 5%), trisomy 13 (14/469, 3%), trisomy 15 (14/469, 3%), trisomy 18 (12/469, 3%), trisomy 21 (11/469, 2%), and tetraploidy (6/469, 1%). Other autosomal aneuploidies included trisomies 14 and 20 in five cases each, trisomy 9 in four cases, trisomies 2, 4, and 8 in three cases each, trisomies 3, 12, and 17 in two cases each, and trisomies 6, 7, 10, and 11 in one case each. Compound aneuploidy of two to three chromosomes was seen in 11 cases involving chromosomes X, Y, 8, 9, 10, 13, 14, 15, 16, 18, 20, 21, and 22. Mosaic abnormalities in 16 cases were aneuploidy or polyploidy with a normal karyotype.
Structural chromosomal abnormalities detected in 17 cases and pCNVs in two cases are shown in [Table 2 ]. Recurrence of structural abnormalities derived from parental translocations t (7;15) and t (11;22) were noted in two women. CMA further defined the genomic imbalances of a large 49.65 Mb deletion of 13q13.1q31.1, a marker chromosome for a 5.91 Mb triplication of 22q11.1q11.21, and a 16.43 Mb deletion in a ring chromosome X. CMA detected a pCNV of a 1.41 Mb deletion at 7q11.23 with a female complement in one POC, which encompasses 25 OMIM genes including the DNAJC30 (OMIM#618202), ELN (OMIM#130160), and FKBP6 (OMIM#604839) morbid genes. This deletion overlaps with the dosage-sensitive 7q11.23 region and is diagnostic of Williams–Beuren syndrome (WBS, OMIM#194050). Another POC with a 1.66 Mb deletion at Xp22.31 in a male complement encompasses the dosage-sensitive STS gene (OMIM#300747) for X-linked recessive ichthyosis (OMIM#308100) ([Fig. 2C ]).
Spectrum and Patterns of Cytogenomic Findings in the Single Analysis and Consecutive Analysis Groups
The spectrum of cytogenomic abnormalities from the SA and CA groups and combined as RPL is shown in [Table 1 ]. The abnormality detection rate of 64% in the SA group, 52% in the CA group, and 56% in RPL showed significant differences among the three groups (SA vs. CA vs RPL), between the SA and CA groups and the SA and RPL groups (p < 0.05), but no difference between the CA and RPL groups (p = 0.16). There were no significant differences in the detection of various types of numerical and structural abnormalities and pCNVs among the three groups (p > 0.05), except for a significant difference in the detection of trisomy 16 in the above group comparisons (p < 0.05) and no difference between the CA and RPL groups (p = 0.12). The noted differences could be induced by the small sample size of the SA group. The lower detection rate of some specific chromosomal abnormalities such as trisomy 13, trisomy 18, and XXY in the SA group could also be explained by the small sampling size. The abnormality detection rate of 56% from 738 cases of RPL was comparable to the 49% from 1,001 POC cases of SPL in a previous study.[5 ] In summary, a similar spectrum of numerical and structural chromosomal abnormalities was noted in the SA and CA groups. Common autosomal trisomies of chromosomes 13, 15, 16, 18, 21, and 22, rare autosomal trisomies of other autosomes, sex chromosome aneuploidies, compound aneuploidy, and polyploidy were detected in approximately 24, 7, 5, 2, and 10% of RPL, respectively. Chromosomal rearrangements and pCNVs were detected in 3 and 1% of RPL, respectively. Mosaic abnormalities were detected in approximately 4% of RPL.
For the 225 women in the CA series, the time interval for the second or third miscarriages within 1, 2, 3, 4, and 5+ years were seen in 64, 17, 9, 4, and 6% of RPL, respectively, indicating approximately 80% of consecutive miscarriages occurred within 2 years ([Fig. 1C ]). Based on the cytogenetic findings from the first and second miscarriages, patterns of recurrent normal karyotypes, alternating normal and abnormal karyotypes, and recurrent abnormal karyotypes were observed in 74 (33%), 71 (32%), and 80 (35%) of consecutive miscarriages. Thus, a skewed likelihood of 68% for recurring normal (74/108) or abnormal karyotypes (80/117) and 32% for alternating karyotypes in subsequent miscarriages was noted ([Fig. 1D ]). The third miscarriages included results from only 17 POC specimens and were insufficient to evaluate this skewed likelihood but a continuous tendency was suggested. For the 80 women with abnormal karyotypes in consecutive miscarriages, 86% (69/80) had recurrent numerical abnormalities involving different chromosomes, 11% (9/80) had repeat numerical abnormalities, and approximately 3% (2/80) were caused by parental carriers of a balanced translocation ([Tables 2 ]). As shown in [Table 3 ], repeat numerical abnormality in consecutive miscarriages included monosomy X in two women, trisomy 16 in one woman, triploidy in five women, and tetraploidy in one woman. Based on previous data that approximately 50% of pregnancy losses are caused by chromosomal abnormalities,[5 ]
[18 ] an expected distribution for the above patterns should be 56 (25%), 113 (50%), and 56 (25%), respectively. The observed distribution was significantly different from the expected distribution with a chi-square statistic of 16.31 and p -value <0.05.
Table 3
Repeat numerical abnormalities in the CA group
Patients/RPL
1
2
3
CA19
45,X
45,X
CA59
45,X
45,X
CA219
47,XY,+16
47,XX,+16
CA35
69,XXY
68,XXY,−3
CA93
70,XXY,+6
69,XXY
CA119
69,XXY
69,XXY
CA177
69,XXX
69,XXX
CA186
69,XXY
69,XXY
CA44
92,XXXX
92,XXXX
92,XXXX
Abbreviations: CA, consecutive analysis; RPL, recurrent pregnancy loss.
Discussion
Current cytogenetic testing on POC is rarely performed on the first pregnancy loss due to the low risk (3–5%) of RPL and high success rate in following pregnancies. The cytogenetic results from so-called “SPL” may represent the first cytogenetic evaluation on unspecified RPL events. A retrospective analysis of 12,096 POC specimens over a 20-year period revealed similar rates and patterns of aneuploidies, sex chromosome abnormalities, triploidies, tetraploidies, complex abnormalities, and structural rearrangements between patients with a single spontaneous miscarriage and RPL.[12 ] A recent meta-analysis on a total of 8,320 POC in 19 studies suggested that the incidence of chromosomal abnormalities in SPL was significantly higher than that in RPL, but the distribution of types of abnormalities showed no differences between SPL and RPL. However, the prevalence and distribution of chromosomal abnormalities could be affected by the size of sampling, the year of publication, regional practice standards, variability in patient selection and inclusion, and methods of analysis.[14 ] Our analysis revealed a similar spectrum of cytogenomic abnormalities between the SA and CA groups of RPL, which was also comparable to that observed from previous studies on SPL and RPL.[4 ]
[5 ]
A recent study on a large case series of 1,081 PRL specimens detected abnormal karyotypes in 36% of them, including approximately 30% of patients with two or more POC specimens analyzed; recurrent abnormalities related to parental chromosomal rearrangements, recurrent polyploidy, repeated and recurring aneuploidy were detected in 8, 7.5, and 4% of patients, respectively.[12 ] Our results from the CA group indicated an abnormality detection rate of 51% (114/225) in the second miscarriage and 59% (10/17) in the third miscarriage ([Fig. 1D ]). Results from previous studies and this case series supported a recommendation of cytogenetic analysis on consecutive miscarriages.[11 ]
[12 ]
[13 ]
[14 ]
Patterns of recurrent normal karyotypes, alternating normal and abnormal karyotypes, and recurrent chromosomal abnormalities were observed in 33, 32, and 35% of 225 RPL women in the CA group. A similar distribution of 37, 27, and 35% was also observed in a study of 108 primary and secondary RPL in 51 women.[11 ] These observations supported a skewed likelihood of 2/3 chance to have recurrent normal and abnormal patterns and 1/3 chance for an alternating karyotype in subsequent miscarriages to maintain an overall distribution of 1/3 for each pattern with chromosomal abnormality in approximately 50% of RPL. Further genomic analysis for RPL cases with recurrent normal karyotypes could lead to the identification of embryonic lethal and detrimental variants affecting early fetal development. Exome sequencing on a cohort of SPL with normal karyotypes detected pathogenic and likely pathogenic variants in 22% of cases.[19 ] Embryonic lethal variants in chromosomally normal pregnancy losses were detected from a case-control study and putatively detrimental variants in a panel of genes known to be associated with miscarriages were detected from euploid miscarriages.[20 ]
[21 ] Exome sequencing in cases of RPL identified novel pathogenic variants in several candidate genes and their functional implications were further analyzed by gene expression knockdown using targeted siRNA in cell lines and by reduced reproductive performance in variant knock-in mice.[22 ]
[23 ] A new enrichment approach for candidate gene detection in unexplained RPL and implantation failure was proposed by exome sequencing, detailed bioinformatic analyses, and in silico protein structural analyses.[24 ]
RPL with recurrent aneuploidy, compound aneuploidy, and polyploidy could be caused by pathogenic variants affecting meiotic chromosome segregation. A rare occurrence of three consecutive aneuploid pregnancies of trisomy 21, trisomy 9, and trisomy 18 suggested a high risk for nondisjunction.[25 ] A healthy woman ascertained by 18 consecutive miscarriages with triploidy lead to the identification of candidate variants in the PLCD4 and OSBL5 genes with functional implications in oocyte activation and completion of meiosis II.[26 ] The detection of repeated chromosomal aneuploidies and pCNVs could also suggest possible gonadal mosaicism, which could be a technical challenge for a firm diagnosis.[11 ]
[27 ] The repeated trisomy 16, monosomy X, and triploidy by CA are the most frequently seen numerical chromosomal abnormalities in pregnancy loss, the possibility of a coincidence event cannot be ruled out.
Structural rearrangements detected in POC could be de novo events or inherited from one of the parents carrying a balanced rearrangement. The detection rate for structural abnormalities from SA and CA cases is approximately 3%. Despite recommendations for follow-up parental studies upon the detection of structural abnormalities, only 5 out of 24 cases in this series confirmed a parental carrier of a balanced translocation. Consistent with our result, cytogenetic analysis on peripheral blood specimens from 570 and 224 couples with RPL showed a detection rate of 3% (18/57) and 7% (16/224) for parental carriers of a balanced translocation and 0.3% (2/570) and 2% (4/224) for carriers of a Robertsonian translocation, respectively, which provides cytogenetic etiology of consecutive two to nine abortions in these couples.[28 ]
[29 ] A recent meta-analysis focused on CMA results from large cases series of POC indicated a detection rate of approximately 3% for pCNVs and 1% for genomic disorders; the risk of pregnancy loss by pCNVs of 16p11.2 duplication, 17q12 deletion, Xp22.3 deletion, and 7q11.23 deletion was estimated in the range of 21 to 50%.[18 ] The detection rate of 1% for pCNVs in this case series was most likely an underestimation due to the incomplete implementation of CMA on POC specimens. For mosaic abnormalities detected in POC, confined placenta mosaicism, pseudomosaicism versus true mosaicism, and maternal cell contamination should be taken into considerations. A six-attribute classification of genetic mosaicism based on the location, pattern, and disease-causing mechanisms has been recently proposed and could be applied for mosaic abnormalities in POC.[30 ] The introduction of genomic sequencing on POC could detect all numerical and unbalanced structural chromosomal abnormalities, pCNVs, and single gene variants and thus provide evidence for diagnostic interpretation and functional implication on genetic causes of SPL and RPL.[24 ]
[31 ] However, there is a limitation on the detection of balanced chromosome rearrangements such as the Robertsonian translocations in parental carriers of SPL and RPL by CMA and genomic sequencing.
