Fetal karyotyping by chorionic villus sampling after the first trimester

Authors


Correspondence: Mr S. G. M. Carroll, Fetal Medicine Unit, St Michael's Hospital, Southwell St, Bristol BS2 8EG, UK.

Abstract

Objective To evaluate chorionic villus sampling (CVS) as a technique for karyotyping after the first trimester by examining the incidence of result failure, confined placental mosaicism, and false positive or negative results at different gestational ages.

Methods During a nine year period between 1989 and 1997, all results of CVS between 8 and 37 weeks of gestation provided by the Regional Cytogenetics Centre were analysed retrospectively by examining indications for CVS, weights of tissue received, gestational age at sampling and karyotype results.

Results There were 2424 chorionic villus samples analysed by the direct method and or cell culture. In 1548 cases CVS was performed before 14 weeks (Group 1), in 685 between 15 and 20 weeks (Group 2), in 160 between 21 and 28 weeks (Group 3) and in 31 cases after 29 weeks (Group 4). Although there was a trend for an increasing rate of failed direct preparation results from Groups 1 to 4 which were 3.8%, 4.7%, 5.6% and 6.6%, respectively; these results were not significantly different. There were 19 cases of confined placental mosaicism and the incidence was significantly greater in Group 3 compared with Group 1 (P < 0.05), and in Groups 3 and 4 combined compared with Group 1. There were six false positive and one false negative result following direct analysis with no significant differences between gestational ages.

Conclusions CVS is a useful test after the first trimester, especially when a fast result is clinically required. However, after 20 weeks, when cordocentesis is available, the higher rate of cytogenetic discordancy between the placenta and the fetus means that cordocentesis may be preferable.

INTRODUCTION

There are numerous publications on chorionic villus sampling (CVS) in the first trimester1–4. Because the mitotic index of the placenta decreases with advancing gestation5, CVS after the first trimester has gained less widespread acceptance. This may be based partly on the belief that insufficient metaphases in the cytotro-phoblast are available for analysis. There have been previous reports on the rates of failed results, confined placental mosaicism and false positive and negative results with CVS after the first trimester (Table 1)6–14. However, there are deficiencies in many of these studies including lack of analysis by gestational age, insufficient data on cell culture, relatively small patient numbers, and no examination of villus sample weights in relation to both gestational age and result. The purpose of our study was to address these deficiencies by evaluating cytogenetic aspects of CVS in the second and third trimester in relation to the first.

Table 1.  Studies on 2nd and 3rd trimester chorionic villus sampling showing gestational age (GA) in weeks (w), weights of samples (Wt), rates of failed results following direct preparation (FD) and cell culture (FC), confined placental mosaicism either in the cytotrophoblast preparation (CPM 1) or cell culture (CPM 2), and false positive and negative results with the direct preparation (FPD and FND,: respectively) and cell culture (FPC and FNC, respectively). Values are given as n/n (%), unless otherwise indicated. NA = not analysed.
StudyGA(w)(range)Wt (mg)(range)FDFCCPM 1CPM 2FPDFPCFNDFNC
Pijpers et al. 1988613–3810–504/127 (3.1)NA0/123 (0)NA0/123 (0)NA0/123 (0)NA
Hogdall et al. 1988715–1955/19 (26)NA0/19 (0)NA0/19 (0)NA0/19 (0)NA
Chieri&Aldini 1989814–25>210/200 (5)NA0/190 (0)NA0/190 (0)NA0/190 (0)NA
Wolstenholme et al 1989916–25>20/74 (0)NA0/74 (0)NA1/74 (1.3)NA1/74 (1.3)NA
Gatz et al. 19901016–322–3019/243 (7.8)NA2/224 (0.9)NA0/224 (0)NA1/224 (0.4)NA
Dalpra et al. 19931113–34>22/143 (1.3)0/68 (0)2/145 (1.4)NANANANANA
Cameron et al. 19941218 (mean)NA20/551 (3.6)15/551 (2.7)2/533 (3.7)3/539 (0.5)1/533 (0.2)0/538 (0)0/533 (0)0/538 (0)
Caspari et al. 19941313–38NA43/1405 (3.1)NA14/1306 (1.1)NA3/1306 (0.2)NA9/1306 (0.7)NA
Podobnik et al. 19971413–40NA10/1000 (1)NA0/1000 (0)NANANANANA
TOTAL  113/3762 (3)15/619 (2.4)20/3614 (0.5)3/539 (0.5)5/2469 (0.2)0/538 (0)11/2469 (0.4)0/538 (0)

METHODS

During a nine year period between 1989 and 1997, samples between 8 and 37 weeks of gestation were forwarded to the Regional Cytogenetics Centre and analysed retrospectively by examining indications for CVS, weights of tissue received, gestational age at sampling and the karyotype results using the computer database. The weight of chorionic villus tissue was estimated using a photographic standard15. Samples with no villi were excluded. Maternal contamination was avoided by selection of villi microscopically in the laboratory. A short term incubation method for direct preparations was used and cell cultures were established with standard techniques16. The accuracy of the karyotypes from CVS was assured by local practice protocols, including karyotyping fetal tissue after termination of pregnancy in cases where a chromosomal abnormality had previously been diagnosed, follow up karyotyping on amniotic fluid cells or fetal blood in cases of discordancy between direct and culture preparations, and comparison of the results with paediatric cytogenetic evaluations, which are undertaken in the same Regional Cytogenetics Unit.

Confined placental mosaicism (CPM) was classified into three types17. CPM 1 was defined as the presence of two or more karyotypically different cell lines in the direct cytotrophoblast preparation but not in cultured chorionic cells, follow up amniotic fluid cells or fetal fibroblasts. CPM 2 refers to mosaicism in cultured cells but not in the direct preparation or follow up studies. CPM 3 is mosaicism present in both the direct preparation and cultured chorionic cells. Karyotype discrepancy was classified as a different direct preparation and chorionic cell culture result where both were abnormal, and did not include the cases with confined placental mosaicism and false positive or negative results. Uni-parental disomy in which both chromosomes of a pair originate from the same parent was examined for inappropriate cases by analysis of parental and neonatal blood using specific DNA probes18.

The maternal age and CVS weight distributions for different gestational ages were compared using the Mann-Whitney U test with interquartile ranges (IQR). Comparison of failure rates, confined placental mosaicism and false positive results in relation to gestational age was performed using the χ2 test with Yates continuity correction; Fisher's exact test was used when the number of observations was less than five. χ2 trend analysis was used to assess the effect of gestational age on result failure and confined placental mosaicism.

RESULTS

There were 2424 chorionic villus samples analysed by the direct method andor cell culture. The indications for CVS included maternal age greater than 35 years (41%), family history of chromosomal or genetic abnormality (20%), fetal abnormalities detected by ultrasound (19%), results of serum biochemistry screening (17%) and maternal anxiety (3%). The fetal karyotype was abnormal in 213 cases (8.8%). These abnormalities were trisomy 21 (n= 77), trisomy 18 (n= 43), trisomy 13 (n= 11), monosomy X (n= 27), triploidy (n= 6) and others (n= 49). The maternal age was significantly different (P < 0.0001) when Group 1 (median 38 years; IQR 34–40), Group 2 (median 33 years; IQR 28–37), Group 3 (median 31 years; IQR 25–37) and Group 4 (median 27 years; IQR 22–29) were compared. CVS was performed between 8 and 14 weeks of gestation in 1548 cases (Group 1), between 15 and 20 weeks in 685 (Group 2), between 21 and 28 weeks in 160 (Group 3) and in 31 cases after 29 weeks (Group 4). A direct analysis was not done in 24 cases in Group 1, four cases in Group 2, and one case in Group 4. Cell culture analysis was not performed in 69 cases in Group 1, 30 in Group 2, eight in Group 3 and three in Group 4.

The incidences of failed results following direct preparation and cell culture, confined placental mosaicism, false positive and negative results and karyotype discrepancies at different gestational ages are shown in Table 2. A comparison of the rates of failed results, confined placental mosaicism and false positive results between the four gestational ages is shown in Table 3. Although there was a trend for an increase in the number of cases with failed results and confined placental mosaicism with the direct preparation with advancing gestation, this was not statistically significant on analysis (Fig. 1 and 2). There was also no difference in the rate of result failure or false positive direct results when the groups were compared. However, there was a significant increase in the rate of confined placental mosaicism in Group 3 compared with Group 1 (P < 0.05); this was also found to be significant when Groups 1 and 2 were combined and compared with Groups 3 and 4 (P < 0.05), and when Group 1 was compared with combined Groups 3 and 4 (P < 0.05).

Table 2.  Study on chorionic villus sampling performed at four different gestational ages showing rates of failed results following direct preparation (D) and cell culture (C), confined placental mosaicism either in the cytotrophoblast preparation (CPM 1) or cell culture (CPM 2) or both (CPM 3), false positive (n?) and negative results (FW) with the direct preparation and cell culture, and karyotype discrepancy (DK) showing error in the direct or culture sample. Values are given as n/n (%).
 8–14 weeks15–20 weeks21–28 weeks≥ 29 weeks
Failed result (D)58/1524 (3.8)32/681 (4.7)9/160 (5.6)2/30 (6.6)
Failed result (C)35/1479 (2.4)12/655 (1.8)6/152 (3.9)1/28 (3.6)
Both D, C failed5/1479 (0.3)1/655 (0.1)6/152 (1.3)0
CPM 13/1479 (0.2)3/655 (0.4)00
CPM 24/1479 (0.2)2/655 (0.3)4/152 (2.6)0
CPM 31/1479 (0.1)2/655 (0.3)00
CPM total8/1479 (0.5)7/655 (1.1)4/152 (2.6)0
FP(D)2/1479 (0.1)3/655 (0.4)01/28 (3.6)
FP(C)0000
FN(D)001/152 (0.6)0
FN(C)0000
DK (D error)2/1479 (0.1)1/1655 (0.1)00
DK (C error)1/1479 (0.1)000
Table 3.  Study on chorionic villus sampling showing comparison of incidences (expressed as P values) in relation to gestational age of failed results on direct preparation (D) and/or cell culture (C), confined placental mosaicism (CPM total) in either the direct preparation, cell culture or both, and false positive results on cytotrophoblast (FPD). Group 1 = 8–14 weeks of gestation; Group 2 = 15–20 weeks; Group 3 = 20–28 weeks; Group 4 = at least 29 weeks.
 P
GroupsFailed DFailed CFailed D&CCPM totalFPD
  1. *Significant.

1 vs 20.390–540.410.140.17
1 vs 30.360.180.130.02*0.25
1 vs 40.320.500.05
2 vs 30.770.100.090.130.57
2 vs 40.430.420.15
3 vs 40.540.700.29
Figure 1.

Rates of failed results after CVS according to direct preparation (□), cell culture (▪) and both (▪) in relation to gestational age.

Figure 2.

Following CVS, rates of confined placental mosaicism in either the direct preparation, cell culture or both in relation to gestational age.

Of the 101 cases with failed results following direct preparation, there were 69 live births, one miscarriage, one intrauterine death at 26 weeks of gestation where the fetus was growth restricted, 12 terminations of pregnancy for abnormalities detected by ultrasound or by follow up cell culture or molecular genetic analysis, and no data recorded in 18. There were 38 live births among the 54 cases with failed cell culture results, eight terminations for abnormalities on scan or genetic analysis or for chromosomal defects detected by the direct preparation or amniocentesis; in another eight cases information was not available. Fourteen of the 19 cases of confined placental mosaicism resulted in live births, and there were five cases of termination of pregnancy. In the four cases in Group 1 where the patients opted for termination, one was for cystic hygroma detected by ultrasound, one was for trisomy 20 confined placental mosaicism type 3, one was for social reasons before the culture result was provided and the indication in the fourth case was not recorded in the database. The fifth case was in Group 2 for oligohydramnios seen on ultrasound. Details on the six cases and one case where the direct preparation gave false positive and negative results respectively are shown in Table 4.

Table 4.  Cytogenetic results of false positive and negative cases on direct chorionic villus sampling preparation. w = weeks; LB = live birth; UPD = uniparental disomy; CV = chorionic villi; TOP = termination of pregnancy; SGA = small for gestational age; IUD = intrauterine death.
 No. of analysed metaphasesOutcome
False positives   
Case 1 (8–14 w)46, XX47, XX, +7Normal LB; no UPD
  CV direct07 
  CV culture300 
  Neonatal blood1000 
Case 2 (8–14 w)46, XX47, XX, +3Normal LB
  CV direct012 
  CV culture300 
Case 3 (15–20 w)46, XY47, XY, +13Hydrops; TOP
  CV direct024 
  CV culture300 
  Skin biopsy400 
Case 4 (15–20 w)46, XY46, XY, add 4pNormal LB
  CV direct07 
  CV culture300 
  Amniocentesis300 
  Neonatal blood300 
Case 5 (15–20 w)46, XY46, XY, del 8Oligohydramnios; TOP
  CV direct017 
  CV cultureNot doneNot done 
  Skin biopsy300 
Case 6 (≥ 29 w)46, XY47, XY, +16LB; SGA, No UPD
  CV direct012 
  CV cultureNo growthNo growth 
  Neonatal blood300 
False negatives   
Case 1 (21–28 w)46, XX47, XX, +18IUD
  CV direct110 
  CV culture08 
  Cordocentesis015 

The median weight of samples in all four groups was 5 mg. However there was a variation in the weight distributions and the interquartile ranges in Groups 1 to 4 were 3–10, 3–8, 4–8 and 4–10, respectively. The only significant difference was that larger samples were obtained from Group 1 compared with Group 2 (P < 0.05). For both the direct and cell culture preparations, the sample weights were significantly higher in the group with successful results (median 5; IQR 3–8 and median 5; IQR 4–8, respectively) compared with those with failed results (median 4; IQR 2–6 and median 3; IQR 2–5, respectively; P < 0.0001).

DISCUSSION

These data confirm that CVS is an acceptable technique in the second and third trimester. There is no significant difference in failure rate or cytogenetic discordancy between the placenta and fetus when CVS is performed at 15–20 weeks of gestation compared with before 14 weeks. After 20 weeks, the discordancy between placen-tal and fetal karyotype is significantly higher than in the first trimester.

Between 15 and 20 weeks fetal karyotyping may be performed by amniocentesis or CVS Direct analysis of cytotrophoblast may provide a fetal karyotype within one to three days. Karyotyping by this technique between 15 and 20 weeks is associated with a direct preparation failure rate of 4.7%, compared with 3.8% before 14 weeks. However, when cell culture is also performed, a result is provided in 99.9% of cases. These results show that CVS is an acceptable alternative to amniocenteis between 15 and 20 weeks of gestation and has advantages if rapid karyotyping is required, such as in the presence of fetal anomalies. Rapid karyotyping after 20 weeks can be provided by CVS or cordocentesis. Although the mitotic index of cytotrophoblast cells is high between six and ten weeks and then decreases towards term5, this decrease does not appear to be large enough to have had a significant effect, despite the rising trend of failed results with gestation in our study. Nevertheless, in view of the 6% failure rate with the direct preparation after 20 weeks of gestation in our study, and data from another nine studies involving 3762 patients undergoing second and third trimester CVS demonstrating a total failure rate of 3% (Table 1)6–14, the option of cordocentesis may be preferable.

The incidence of cytogenetic discordancy between the direct and cell culture preparations and the fetus increased with gestation. There was a significantly higher number of cases with confined placental mosaicism after 20 (2.2%) weeks, compared with cases less than 14 weeks (0.5%). This finding differed from another study where confined placental mosaicism was shown to occur in 1% of samples both before and after 14 weeks13. In a third report the incidence of confined placental mosaicism in the first and second trimester was 0.2% and 1%, respectively, but half of the second trimester cases were CVS carried out between 13 and 14 weeks11. Other studies have shown that confined placental mosaicism occurs in 1%-2% of pregnancies studied by CVS analysis of cytotrophoblast or cell culture at 9–12 weeks of gestation1–3. The increasing incidence of confined placental mosaicism with gestation is difficult to explain and may be regarded as being evidence for somatic mutation. The alternative approach of carrying out cordocentesis after 20 weeks, apart from obviating the risk of confined placental mosaicism, provides fetal lymphocytes with superior chromosomal banding patterns with greater sensitivity in the detection of microdeletions10. Indeed, in the report of the UK Quality Assessment Scheme in Clinical Cytogenetics, the overall quality rating for chromosomal banding from fetal blood samples was higher than that for CVS19. Additionally, when both tests are performed by experienced personnel the post-procedure fetal loss rates are similar. In seven studies of second and third trimester CVS involving 2834 patients, there were 27 fetal losses (0.9%)6–8,10,12,14,20, which is comparable to data from a large review of 13 studies involving 3517 patients having cordocentesis, where 46 fetal losses (1.3%) occurred21. However, where cordocentesis may be technically difficult, such as in cases with oligohydramnios, karyotyping by CVS using the direct preparation is an appropriate option.

There were no false positive or known false negative results on cell culture, unlike the direct preparation where seven such cases were found. This is not surprising because the cell culture is derived from the inner cell mass and is therefore more likely than the direct preparation to be reflective of the fetal karyotype22. After 14 weeks the false positive and negative rates for direct preparation were 0.5% and 0.1%, respectively. These figures are consistent with the results of seven other studies on CVS after the first trimester involving 2469 patients, with rates of 0.2% and 0.4%, respectively (Table 1)6–14. Similar figures are also quoted for the first trimester12,23.

Because of the possibility of a false positive direct result, unless there is a high suspicion of abnormal karyotype (such as in cases with significant fetal anomalies detected on ultrasound), it is reasonable to defer termination of pregnancy until the result of the cell culture is available. Indeed, of the six false positive direct results in our study, in the two cases undergoing termination of pregnancy, abnormalities were seen on ultrasound. The false negative case was trisomy 18; this karyotype was prominent among false negative results in another study on CVS after the first trimester, where five of the nine false negatives had trisomy 188. A recent publication describes three false negative results of trisomy 21 in 9000 first trimester direct tests24. From these data, it appears that it is necessary to use both the direct and culture preparation to obtain the most accurate prediction of the chromosome analysis. Additionally, when discrepant findings in chorionic villi are found, amniocentesis or cordocentesis should be considered to determine the presence or absence of a cyto-genetic abnormality in the fetus.

The finding that there was an inverse correlation between maternal age and gestation reflects indications for CVS. The maternal age is likely to be higher in the first and early second trimester, stages in pregnancy when patients present for testing either because of advanced age alone or in combination with abnormal serum biochemistry. This study shows that it is possible to get adequate villus samples from all gestations. The finding that there were heavier samples in the group with successful results compared with the group where no result was obtained is most likely due to laboratory techniques being facilitated by samples of greater density.

CONCLUSION

CVS should not be restricted to the first trimester because the technique can provide a sufficient number of analysable metaphases in the second and third trimester. There was no significant difference in failure rate or cytogenetic discordancy between the placenta and fetus when CVS was performed at 15–20 weeks of gestation, compared with the first trimester. After 20 weeks, because of the result failure rate and the risk of cytogenetic discordancy between the placenta and the fetus, a fetal blood sample is preferable.

Ancillary