Midtrimester, defined here as the 15th through the 24th week of gestation, is by far the most common time period for performing the amniocentesis procedure. Culture of amniotic fluid cells is optimal in this time period, both from the perspective of rapidity of cell growth (and therefore sample turnaround time) and because the culture failure rate is less than 0.5% in experienced laboratories [51, 52].

The risks associated with midtrimester amniocentesis include leakage of fluid, cramping, bleeding, infection, and miscarriage. The risk of miscarriage following midtrimester amniocentesis is related to practitioner experience, number of needle insertions, size of the needle, and other factors [53]. The appropriate risk figure to provide patients is still debated. In spite of the millions of amniocentesis procedures performed and the importance of an accurate risk figure to provide patients, there has been only one large prospective controlled study performed regarding the risks of amniocentesis. In this paper, known as “the Danish study,” 4,606 women comprised the final study population [54]. Of these, half were randomized to have amniocentesis, and the other half were randomly assigned to the control, non-amniocentesis group. At the conclusion of the study, it was found that the total rate of spontaneous abortion was 1.7% in the study group and 0.7% in the control group (p  <  0.05). When the women with a high maternal serum alpha-fetoprotein (AFP) were considered, it was found that they had a relative risk of spontaneous abortion after amniocentesis of 8.3 compared to women with a normal maternal serum alpha-fetoprotein level. This equated to an overall relative risk of 2.3. Other factors found to increase the risk of spontaneous abortion were transplacental passage of the needle (relative risk of 2.6) and discolored amniotic fluid (relative risk of 9.9).

Another study from Denmark in 2009 reported on results of a national registry-based cohort study, including all singleton pregnant women having an amniocentesis or CVS between 1996 and 2006 [55]. The fetal loss rate was defined as that occurring before 24 weeks’ gestation. The miscarriage rate after amniocentesis was 1.4% after amniocentesis and was not correlated with maternal age. The number of procedures performed at each center had a significant effect on the risk of fetal loss. In those performing fewer than 500 amniocenteses, the odds ratio for fetal loss was 2.2 (95% CI, 1.6–3.1) when compared to those performing more than 1,500 per year. There was no control group in this study.

In reviews of procedure-related risks from many publications, in which five included a control group, the authors concluded that the procedure-related miscarriage rate from amniocentesis is 0.5–1.0% [56, 57].

Odibo et al. reported on the fetal loss rate after second-trimester amniocentesis in a single center in a retrospective cohort study comparing the fetal loss rate in women having amniocentesis with those not having any procedure [58]. Of the 88% for whom complete outcome data were available, fetal loss in the amniocentesis group was 0.4% compared with 0.26% in those without amniocentesis (relative risk of 1.6, 95% CI, 1.1–2.2). Fetal loss less than 24 weeks occurred in 0.97% of the amniocentesis group and 0.84% of the no-procedure group, so the fetal loss rate less than 24 weeks attributable to amniocentesis was 0.13% (95% CI, 0.07–0.20%), or 1 in 769. The only subgroup with a significantly higher amniocentesis-attributable fetal loss rate was women with a normal serum screen (0.17%, p  =  0.03).

An important and often overlooked component of providing risk assessments to patients is the underlying incidence and timing of pregnancy losses. A prospective study of 220 ultrasonographically normal pregnancies in women recruited prior to conception (in order to avoid bias of selection) found a pregnancy loss rate after 8 weeks of 3.2% [59]. Other studies have shown a maternal age factor in the loss rate [38]. The prevalence of trisomies is about 50% higher at 16 weeks compared to term pregnancies (ibid.), so selection against chromosomally abnormal abortuses is still occurring at 16 weeks. The incidence of spontaneous pregnancy loss after 16 weeks is 1%.

Some genetic counselors and amniocentesis practitioners counsel patients regarding the risk of the amniocentesis relative to the risk of a fetal chromosome abnormality and in effect use this as a decision point. In this way, a woman with a risk of fetal chromosome abnormality of 1 in 200 might be inclined to decline amniocentesis if the risk of miscarriage as a result of the procedure was quoted as 1 in 100 and the risks compared during the counseling session. A maternal age of 35 has traditionally been used as a cutoff for the definition of advanced maternal age because the risk of a fetal chromosome abnormality at this age is roughly equivalent to the originally reported risk of miscarriage as a result of the amniocentesis. This is not sound reasoning because the burdens of the risks are quite different—one burden being the potential lifetime task of caring for an individual with mental retardation and physical/health problems and the other being miscarriage of a potentially healthy fetus [60].

Interest in early amniocentesis (EA) rose in the 1980s, due in large part to the continued desire to provide and receive prenatal diagnosis at an earlier gestation without some of the risks and limitations associated with chorionic villus sampling, which are outlined in the following paragraphs. An increase in sophistication in ultrasound technology has also made earlier imaging of fetuses more feasible and has added to the confidence level of the physicians performing the procedure. Adding to this is the opportunity to measure amniotic fluid alpha-fetoprotein and acetylcholinesterase, which is not possible with CVS. One center reported a rise in EA procedures from 3.2% of their 495 procedures in early 1985 to 6.5% of 980 procedures in late 1987 [61].

Early amniocentesis is usually described as one that occurs before 15 weeks’ gestation. It has been shown that the earlier a prenatal diagnosis procedure is performed, the higher the fetal loss rate is [62]. One should therefore further divide the periods at which amniocentesis is performed to provide better comparative data for a variety of procedures since, “…true risks …appear to be a function of gestational age and less related to the procedure performed” [62].

Although the procedure by which EA is performed is similar to that of midtrimester amniocentesis, practitioners report several challenges unique to EA. The placenta is more widely spread, the amniotic fluid volume is lower, and the amniotic membrane is not yet totally adherent to the uterine wall, leading to the “tenting” reported by some physicians [63].

In order to compare first-trimester prenatal diagnostic modalities, a number of investigators have published studies comparing CVS with early amniocentesis. Shulman et al. reported on 500 women, half of whom had transabdominal CVS (TA CVS) from 1986 to 1988, and half of whom had EA from 1987 to 1991 [79]. Of the EA specimens, all but 11 were obtained from weeks 12 to 14, and the rest were from weeks 9 to 11. Of the continuing pregnancies, loss rates of 3.8 and 2.1% for EA and TA CVS, respectively, were seen. This was not statistically significant. The culture failure rate for both procedures was 0.8%. This study has limited applicability inasmuch as the numbers were small and the patients not randomized, and the time intervals were different. Although all procedures were listed as initial cases, the relative degree of prior individual practitioner experience in the two procedures was not addressed.

In 1994, Nicolaides et al. reported on a prospective, partially randomized study comparing EA and TA CVS in 1,870 women [80]. The spontaneous loss rate was significantly higher after EA at 5.3% than with the CVS group (1.2%). The rate of successful sampling was the same at 97.5%. Culture failure occurred in 2.3% of the EA group, compared to 0.5% in the CVS group. Confined or true mosaicism was seen in 1.2% of the CVS group, compared to 0.1% of the EA group. The authors concluded that although EA and CVS are equally likely to produce valid cytogenetic results, CVS would probably become the “established technique” due to the 2–3% excess risk of fetal loss in the EA group.

In response to this study, Saura et al. stated that EA could be a “true alternative” to CVS after the 13th week, when the disadvantages of culture failure and fetal losses decrease [81]. Bombard et al. reported one loss in 121 procedures (0.83%) performed by one practitioner at 10–13 weeks using a 22-gauge needle [82]. They suggested that Nicolaides’ higher EA fetal loss rate could be related to the needle gauge and the multiple practitioners in his study, compared to one practitioner in Bombard’s center.

Similar results were reported by Vandenbussche et al., who, in a partially randomized study, reported eight fetal losses among 120 EA procedures, compared to none among the 64 CVS patients with a follow-up of 6 or more weeks [83].

Another response to these reports proposed the idea that the main drawback to the studies was the very small numbers of EA procedures performed and the evident greater practitioner experience with CVS than with EA. The authors reported a spontaneous abortion rate after EA of 1% up to week 24 on the basis of 1,800 pregnancies. The culture failure rate was 0.3% for gestations ranging to 10 weeks 4 days [84].

An important consideration raised by some investigators is that the banding quality of amniocentesis specimens of any gestation is generally superior to that of CVS specimens, which increases the informativeness of the cytogenetic analysis [78, 84]. The fact that amniotic fluid AFP levels and multiples of the median have been established in many laboratories down to 12 or 13 completed weeks of gestation adds another advantage to the diagnostic power of EA compared to CVS [85].

A 14-center study of 3,775 women randomized to having either CVS or EA was conducted to try to provide more answers to the questions as to the safety and accuracy of EA and transabdominal CVS at 11–14 weeks’ gestation [86]. Both types of procedures were performed by the physicians in each center. Early in the trial, reports of clubfoot at 11–12 weeks in EA patients caused procedures at these weeks to be discontinued.

Criteria for inclusion included advanced maternal age, serum marker screen positive, and prior trisomy [86]. The primary outcome was deemed to be preterm delivery or pregnancy loss of a cytogenetically normal fetus at less than 28 weeks’ gestation. Secondary outcomes included total fetal loss, including neonatal death; amniotic fluid loss; pregnancy outcome; limb and other congenital defects; and cytogenetic diagnostic success and accuracy. Multiple procedures were required for EA at 11–12 weeks (2.4% vs. 1.2% for CVS). Maternal cell contamination was seen in EA specimens at 11–12 and 13 weeks’ gestation (0.6% in both cases vs. 0% in CVS). Pseudomosaicism was seen in 1.2% of EA 11–12-week specimens versus 0.6% of CVS specimens. CVS specimens were harvested at 5.9–6.5 days across the sampling period, compared to 12.3–9.8 days for 11–12- or 14-week EA specimens, respectively. As for complications, the only difference that reached significance at the p  <  0.001 level was EA with a 9.6% amniotic fluid leakage rate. Gestational hypertension/preeclampsia was seen in 5.4% of the CVS patients compared to 3.5% of the EA patients, for a p value of 0.005. Of 1,914 CVS procedures, 34 had cytogenetic abnormalities, two were lost to follow-up, and 1,878 were cytogenetically normal. Thirty-nine, or 2.1%, were lost or delivered before 28 weeks. This compares to 1,861 EA procedures, of which 38 had cytogenetic abnormalities, three were lost to follow-up, and 1,820 were cytogenetically normal. Forty-two, or 2.3%, were lost or delivered before 28 weeks. Clubfoot was seen in 0.2% of CVS patients; in EA patients, it was seen in 1.2% of 11–12-week procedure offspring, 0.8% of 13-week offspring, and 0.2% of 14-week offspring for a relative risk of EA versus CVS of 4.1 (1.17–14.6). The authors concluded that, in general, CVS is the preferred prenatal diagnostic procedure between 12 and 14 weeks [86]. This conclusion was shared by Evans and Wapner and by Tabor and Alfirevic in their reviews of prenatal diagnostic procedures [56, 57]. Tabor and Alfirevevic stated, “Amniocentesis should…not be performed before 15  +  0 weeks’ gestation” due to increased fetal loss rates compared to conventional amniocentesis and the risk of talipes [57].

The volume of amniotic fluid obtained for prenatal diagnosis varies with the stage of gestation, with 15–20 mL conventionally removed by midtrimester amniocentesis practitioners. In one report, data from several small studies was pooled, and the volume of amniotic fluid for weeks 10–20 was calculated [68] (see Table 12.10). At gestations under 15 weeks, many practitioners have adopted the practice of removing 1 mL per week of gestation, and others have found excellent culture success rate and turnaround time with less fluid removed. For example, one group withdrew 4–12 mL in gestations of 9–14 weeks and obtained a 100% culture success rate in 222 specimens, while others withdrew 5–8 mL in pregnancies of 10 weeks and 4 days to 13 weeks and 6 days for an overall culture success rate of 99.7% [84, 87]. It has been observed that the total cell numbers rise exponentially from 8 to 18 weeks’ gestation, but the number of viable cells increases only slightly during that time [78]. This probably explains the comparable culture success rate of EA compared to midtrimester amniocentesis.

Risks associated with CVS have been extensively studied. Perhaps the issue receiving the most attention in the past few years was raised by Boyd et al. involving one case and then more extensively by Firth et al., who reported five babies with severe limb abnormalities out of 289 pregnancies in which TA CVS had been performed at 56–66 days’ gestation [88, 89]. Four of these had oromandibular-limb hypogenesis syndrome. They hypothesized that CVS undertaken up to 66 days’ gestation may be associated with an increase in the risk of oromandibular-limb hypogenesis syndrome and other limb-reduction defects. This report generated many others, with mixed conclusions.

A flurry of letters to the editor of Lancet in 1991 followed Firth’s report. Reporting evidence to support the association between CVS and limb-reduction defects were Mastroiacovo et al. and Hsieh et al. [90, 91]. Monni et al. suggested that the incidence and severity of limb defects was related to the gauge of the needle since they used a 20-gauge needle while Firth used an 18-gauge needle [92]. In a series of 525 CVS procedures done before 66 days’ gestation, no severe limb defects were seen, and only two mild defects were seen in 2,227 procedures that were done later [92]. Mahoney then reported on two multicenter studies that compared transcervical CVS with amniocentesis, and another comparing transabdominal CVS with transcervical CVS [93]. Of 9,588 pregnancies studied, 88% of the CVS procedures were performed after 66 days’ gestation. Significant limb-reduction defects were present in seven babies. Two of these defects were longitudinal, and five were transverse. Another baby had minor reduction defects of the toes. They compared these abnormalities to those reported to the British Columbia registry and found no significant increase in these birth defects. The timing of the CVS procedures that resulted in babies with abnormalities ranged from 62 to 77 days’ gestation.

Similar conclusions were reached in a study in which 12,863 consecutive CVS procedures were performed [94]. Five limb-reduction defects were seen, which were found not to be significantly different from the incidence observed in the British Columbia registry of birth defects. Of the 12,863 procedures, 2,367 were done at 56–66 days, and one of the limb defects was seen in this group. The authors observed no gestational time-sensitive interaction related to CVS and postulated that this was due to their larger experience base.

In 1993, Jahoda et al. reported on 4,300 consecutive transabdominal and transcervical CVS cases for which newborn follow-up information was obtained [95]. Of the 3,973 infants born in this group, three (0.075%) had a terminal transverse limb defect. Two of these occurred in the transcervical CVS group sampled before 11 weeks’ gestation (1,389 patients), and the other one was in the transabdominal CVS group, sampled after 11 weeks (2,584 patients). The authors found the latter figure to be comparable to the prevalence figure given in population studies. They concluded that postponement of CVS to the late first or early second trimester of pregnancy would contribute to the safety of the procedure.

In the same year, a report of the National Institute of Child Health and Human Development Workshop on Chorionic Villus Sampling and Limb and Other Defects was issued [96]. The conclusions, based on a review of the literature, were mixed; some concluded that exposure to CVS appeared to cause limb defects, while others did not. All agreed that the frequency of oromandibular-limb hypogenesis appeared to be more common among CVS-exposed infants. This seemed to correlate with CVS performed earlier than 7 weeks postfertilization (9 weeks post last menstrual period). Whether or not a distinctive type of limb defect was associated with CVS could not be determined, and it also was unclear whether the CVS-exposed infant had an increased frequency of other malformations, including cavernous hemangiomas.

A five-center retrospective cohort study was performed by the Gruppo Italiano Diagnosi Embrio-Fetali to examine this issue, with results published in 1993 [97]. Of 3,430 pregnancies in which CVS had been performed, outcome information was available for 2,759. Of these, three had transverse limb-reduction defects, two among 804 CVS procedures performed at 9 weeks and one among 1,204 CVS procedures performed at 10 weeks. There were no limb-reduction defects noted in 2,192 amniocenteses with completed follow-up performed during the same study period. The authors concluded that performing CVS at less than 10 weeks’ gestation, “should be discouraged until further evidence against this association can be obtained,” while noting that their follow-up rate was only 80%.

Hsieh et al. surveyed 165 obstetric units in Taiwan regarding the incidence of limb defects with and without CVS [98]. Of these, 67 hospitals responded, representing 78,742 deliveries. The incidence of limb defects was found to be 0.032% in the general population and 0.294% in the CVS population. The abnormalities seen in the CVS group included amelia, transverse reductions, adactylia, and digit hypoplasia, much like the abnormalities reported by Firth et al. [89]. The 25 limb abnormalities in the non-CVS group involved syndactyly or polydactyly. In addition, oromandibular-limb hypogenesis was seen in four of 29 CVS cases with limb abnormalities but in none of the non-CVS cases with limb abnormalities. The severity of the post-CVS limb abnormalities appeared to correlate with timing of the procedure, and the authors recommended performing CVS only after 10 full gestational weeks to minimize the risks.

In 1995, Olney et al. reported on a United States multistate case-controlled study comprising the years 1988–1992 [99]. The case population was 131 babies with nonsyndromic limb deficiency born to women 35 and older, and control subjects were 131 babies with other birth defects. These were drawn from a total of 421,489 births to women greater than 34 years of age. The odds ratio for all types of limb deficiency after CVS was 1.7, and for transverse digital deficiency, an odds ratio of 6.4 after CVS was observed. They estimated that the absolute risk for transverse digital deficiency in babies after CVS was one per 2,900 births (0.03%).

Froster and Jackson reported on outcome data in a World Health Organization (WHO) study on limb defects and CVS in 1996 [100]. From 1992 to 1994, 77 babies or fetuses with limb defects from 138,996 pregnancies exposed to CVS were reported to the WHO CVS registry. This group represented the entire experience of 63 European and American centers reporting to the registry. They found that the overall incidence of limb defects in the CVS cohort did not differ from that in the general population, and they did not see a different pattern of distribution of limb defects between the groups. No correlation between limb-reduction defects and gestational age was identified. They indicated that other studies finding an association between limb defects and CVS are confusing because of different methodologies and interpretations and that the numbers reported are too small to draw firm conclusions.

Larger numbers were collected by Kuliev et al., who summarized the accumulated experience of 138,996 cases of CVS from the same 63 centers that report cases to the World Health Organization CVS registry [24]. They reported an overall incidence of limb-reduction defects after CVS of 5.2–5.7 per 10,000, compared with 4.8–5.97 per 10,000 in the general population. They also found no difference in the pattern distribution of limb defects after CVS and similarly concluded that their data provide no evidence for any risk for congenital malformation caused by CVS.

The spontaneous decondensation of the constitutive heterochromatic regions of chromosomes 1, 9, 16, and Y has been observed in 46.6% of chorionic villus samples, per a study by Perez et al. [101]. This type of decondensation is occasionally observed in amniotic fluid cells (9%) and has never been found in fetal lymphocytes. This decondensation can lead to breaks, fragile sites, and loss of the chromosome, including, for example, the loss of 1q in culture.

Because CVS is usually performed on women 35 and older, the issue of whether the limb deficiencies seen after CVS were related to maternal age was raised by Halliday et al. in a study from Victoria, Australia [102]. A congenital malformations registry maintained there was reviewed by a medical geneticist, who classified all cases using the International Classification of Diseases, 9th revision [103]. All babies born with limb defects in 1990–1991 were identified, and the number of those whose mothers had amniocentesis, CVS, or no invasive study was known. Excluding babies with chromosome abnormalities, recognized inherited syndromes, or amniotic bands, the authors found a twofold relative risk of having a baby with a limb deficiency of any type among women at age 35 or older, compared to women under 35. They also discuss the difficulty in interpreting studies of limb defects and CVS, as others had, pointing out the importance of 100% follow-up, inclusion of all recognized cases of limb deficiencies (induced abortions as well as all other births), recognition of the heterogeneity of the condition, and the different risk estimates at different gestational ages [100].

A subsequent study found no maternal age confounding effect in interpretation of CVS/transverse limb deficiency data [104]. The authors analyzed the maternal age-specific rates of transverse limb deficiencies in the Italian Multicentric Birth Registry and used a case-control model for maternal age. No difference in the relative risk was seen between the 35-and-older group, whether or not CVS had been performed, and the under-35 group. The risk estimate for transverse limb defects associated with CVS was 12.63 and did not change after stratification for maternal age or for gestational age.

After 1991, the utilization of CVS dropped significantly, due in large part to the concern regarding limb deficiencies [105, 106]. Although national utilization numbers are not available, a large national prenatal genetic counseling company affiliated with a cytogenetics laboratory reported a decrease of CVS of 3% per year from 36 to 14% from the years 2007 to 2001. In their patient population of 55,019 women, 34% were offered CVS. The decline was statistically significant. The acceptance rate increased again in 2008 to 24%, thought due to American College of Obstetrics and Gynecology Practice Bulletin #77, 2007a, supporting first-trimester screening for fetal chromosome abnormalities, as well as likely being due to such factors as increased access to first-trimester screening [107].

In the first large controlled study of the safety of CVS, Rhoads et al. reported on seven centers’ experience with transcervical CVS in 2,235 women compared to that of 651 women who had amniocentesis at 16 weeks’ gestation [108]. They found an overall excess loss rate of 0.8% in the CVS group after statistical adjustments for gestational age and maternal age. CVS procedures in which more than one attempt was made were associated with a substantially higher loss rate, supporting the observation by Silver et al. and others that increased operator experience is a key factor in assessing the risks of CVS [109]. Silver’s group found that the number of placental passes and increased sample weight/aspiration attempt ratio may be more sensitive indicators of competence than the fetal loss rate.

Results of a randomized international multicenter comparison of transabdominal and transcervical CVS with second-trimester amniocentesis were reported in 1991 [110]. Outcome information was available for 1,609 singleton pregnancies in the CVS group and 1,592 in the amniocentesis group. Thirty-one centers participated, and the numbers of cases submitted ranged from 4 to 1,709. Significantly fewer surviving newborns were seen in the CVS group than in the amniocentesis group (4.6% difference, p  <  0.01). Most of the difference was in the significantly greater number of spontaneous fetal deaths before 28 weeks: 86/1,528 in the successfully sampled CVS group and 25/1,467 of the successfully sampled amniocentesis group (rate difference of 2.9%, p  <  0.02).

In a report from the Centers for Disease Control, an overall risk of spontaneous abortion attributed to CVS is reported from a literature survey as 0.5–1.0%, compared to 0.25–0.50% for amniocentesis procedures [111].

In the WHO study, registry participants reported a ­spontaneous pregnancy loss rate after transabdominal or transcervical CVS of 2.5–3.0%, with several large-volume operators having loss figures of less than 2% [24]. This risk was deemed comparable to that of amniocentesis. Tabor et al. documented fetal loss rates in an 11-year national registry study in Denmark of the outcomes of 31,355 CVS procedures at 24 weeks’ gestation in a 2009 report [55]. The overall fetal loss rate was 1.9% (95% CI, 1.7–2.0) and was not correlated with maternal age. The number of procedures done in each center had a significant effect on the loss rate; in departments performing 1,500 or fewer during the 11 years, the risk was 40% greater than in those performing more than 1,500 procedures per year. There was no control group.

In reviews of procedure-related risks from many publications, in which five included a control group, the authors concluded that the procedure-related miscarriage rate from CVS is 0.5–1.0% [56, 57].

Efficacy and risks associated with transcervical CVS (TC CVS) and transabdominal CVS (TA CVS) have been studied at several centers [110, 112–114] (see Fig. 12.1). The majority of CVS had been performed transcervically until the late 1980s, when more centers began using TA CVS to avoid ­cervical microorganisms and to reach placentas more easily. In their pilot study in 1988, Smidt-Jensen and Hahnemann reported on 100 TA CVS cases at 8–12 weeks’ gestation followed to term, compared to 200 amniocentesis cases [114]. In all CVS cases, a sample was successfully obtained and cultured, and the fetomaternal complication rates were found not to be significantly different from those of previous TC CVS reports.

Transabdominal CVS has been increasingly used in recent years compared to TC CVS. Brambati et al. reported on efficiency and risk factors in 2,411 patients; 1,501 of whom had TC CVS and 910 of whom had TA CVS [112]. The two approaches had comparable success rates and complication rates, but TA CVS was considered easier to learn and less likely to be contraindicated by clinical and anatomical conditions. Subsequently, this group published results of a randomized clinical trial of TA and TC CVS [113]. All CVS procedures were performed by the same practitioner, who had prior similar experience in both techniques. The procedures were found to be equally effective, although TA CVS required significantly fewer insertions. The authors concluded, “…transabdominal and transcervical CVS appear equally effective, and by and large the choice may be based on the operator’s preferences.”

Chromosomal mosaicism is characterized by the presence of two or more karyotypically different cell lines within one individual. Confined placental mosaicism (CPM) is defined as a discrepancy between the chromosomal constitutions of placental and embryonic/fetal tissues. CPM results from viable mitotic mutations occurring in the progenitor cells of trophoblast or extraembryonic mesoderm during early embryonic development. In 1983, Kalousek and Dill reported on numerical discrepancies between the karyotypes of fetal and placental cells, either full trisomies or mosaic aneuploidies, and similar reports followed [115, 116]. Based on six cases in which placental/CVS cells had a different chromosome constitution from that of amniotic fluid cells, the authors concluded that the results of cytogenetic analysis from placental tissue may not be representative of the fetus. Their figures, though small, were similar to the 2% incidence of this phenomenon as previously reported [117]. Since then, others have found CPM to occur in 0.8–2% of viable pregnancies studied by CVS at 9–11 weeks’ gestation and in 0.1% or less in amniocentesis specimens [80, 118–125].

The outcomes of pregnancies in which CPM is diagnosed vary from apparently normal outcomes to severe intrauterine growth restriction (IUGR), although few follow-up reports are yet available in the literature. Kalousek et al. found six cases of IUGR among 17 gestations with CVS-detected CPM, 5 in liveborns, and one associated with intrauterine death [126]. They noted that others had found a 22% fetal loss rate among pregnancies with CPM. Wolstenhome et al. found 73 cases of CPM in 8,004 CVS specimens from women referred for advanced maternal age, previous child with aneuploidy, or family history thereof [125]. Comparison at delivery with the control population did not show a marked increase in adverse pregnancy outcome. In 108 other cases referred for ultrasound detection of isolated IUGR, seven were shown to have CPM involving the following chromosomes: 2 and 15 (1 case), 9 (1 case), 16 (3 cases), del(13) (1 case), and 22 (1 case).

Hahnemann and Vejerslev evaluated cytogenetic outcomes of 92,246 successfully karyotyped CVS specimens from 79 laboratories from 1986 to 1994 [127]. CVS mosaicism or nonmosaic fetoplacental discrepancy was found in 1,415 (1.5%) of the specimens. Table 12.11 shows the mosaic and nonmosaic chromosome findings seen. Updated CVS mosaicism reports are shown from other studies for specific chromosomes. Hahnemann and Vejerslev’s work on several cell lineages indicated that mosaic or nonmosaic trisomies found in cytotrophoblasts, with a normal karyotype in the villus mesenchyme, were not seen in fetal cells. However, if such trisomies were seen on cultured preparations, a risk of fetal mosaic or nonmosaic trisomy existed. They recommended amniocentesis in all pregnancies involving mosaic autosomal trisomy in villus mesenchyme.

A thoughtful study on this topic was published by Daniel et al., in which rare trisomies detected at the time of CVS and amniocentesis were analyzed [128]. The authors comment on the likelihood of cryptic fetal mosaicism as the cause of abnormal phenotypic findings as opposed to CPM, given the lack of phenotypic effect of maternal uniparental disomy (UPD) for chromosomes 1, 7, 9, 10, 13, 21, and 22 or for paternal UPD for chromosomes 1, 5, 6, 7, 8, 13, 17, and 21. The effect of maternal and paternal UPD 20 is still unclear, and for chromosomes 1 and 16, effects are less clear than previously believed. In the series, they reported there was some evidence for cryptic fetal mosaicism and none for UPD. They describe other similar findings in the literature and conclude that the finding of even very low-grade mosaicism in amniocytes should be regarded as significant.

When a conceptus is trisomic, this aneuploidy is said to be “corrected” if by chance there is early loss of one of the trisomic chromosomes. Depending upon the parental origin of the trisomy and of the chromosome that is lost, this can lead to an apparently normal diploid cell line with uniparental disomy (both chromosomes in a pair from one parent) for that chromosome. Because most trisomies are maternally derived, the disomy seen is often maternal, as was the case in two previously reported cases of trisomy 15 mosaicism seen at CVS in which the neonates subsequently manifested Prader-Willi syndrome due to maternal disomy 15 [125]. The authors also note the reports of several cases of chromosome 16 CPM-associated IUGR in which maternal disomy 16 was seen in most of the cases. The presence of mosaic trisomy 16 itself may be of most significance in such cases, however. In this regard, Daniel et al. presented data questioning the clinical significance of UPD 16 and CPM versus that of true cryptic fetal mosaic trisomy 16 [128].

The evaluation of parental disomy in all CPM cases involving chromosome 15 should be offered, and this recommendation has extended to other chromosomes as more information has become available.

For a thorough discussion of UPD, refer to Chap. 20.

Interphase fluorescence in situ hybridization (FISH, see Chap. 17) can be useful for the diagnosis of CPM, given that interphase FISH is rapid and has the great advantage of not requiring growing, dividing cells to obtain results. Harrison et al. examined the placentas of 12 pregnancies in which nonmosaic trisomy 18 had been diagnosed and found significant levels of mosaicism, confined to the cytotrophoblast, in 7 of the 12 [118]. Based on their observation that most of the mosaic results were seen in stillborn or newborn trisomy 18 babies, and on the fact that the great majority of trisomy 18 conceptuses spontaneously abort, they suggested that a normal diploid trophoblast component in placental tissue may be necessary to facilitate the prolonged survival of trisomy 18 conceptuses.

Schuring-Blom et al. used FISH to document CPM in three pregnancies in which mosaic trisomy 8, mosaic trisomy 10, and nonmosaic monosomy X were observed following CVS but which were found to be chromosomally normal at amniocentesis [129]. In all three cases, FISH showed the presence of the mosaic cell line confined to one part of the placenta.

Henderson et al. performed a cytogenetic analysis using a “mapping” technique of nine term placentas after CPM had been diagnosed and found tissue-specific and site-specific patterns of mosaicism [130]. In addition to metaphase chromosome analysis, they employed interphase FISH to ­examine several areas of the placentas. Noting that the outcomes of pregnancies are highly variable after CPM is diagnosed, they proposed a wider study involving extensive analysis of term placentas when this occurs in order to obtain more information regarding the outcome of such pregnancies.

Direct CVS preparations involve the rapid metaphase analysis of villous cytotrophoblastic tissue. Cultured preparations involve the mesenchymal cells in the villi. Some laboratories use only cultured cell preparations, and others utilize both methods. Investigations into the outcomes of pregnancy after CVS support the use of both techniques to maximize the accuracy of the test [121, 123, 124]. These studies documented false-negative and false-positive results using direct and cultured preparations, and the first two groups concluded that results from both direct and cultured techniques were necessary in a substantial number of cases to accurately predict the fetal karyotype. In one study, long-term culture was advocated as having higher diagnostic accuracy, and the direct method was said to be a useful adjunct to the culture method [121]. In a study by Los et al. of 1,829 consecutive CVS procedures with direct and long-term cultures, one conclusion was that using both modalities decreased the necessity for follow-up amniocentesis by 35% compared to that of long-term culture alone [131]. In part at least, the finding that both techniques add to the diagnostic accuracy appears to be related to the nonrandom findings of some trisomies in direct versus long-term cultured tissues. Trisomy 2 is seen more in cultured cells, and trisomy 3 is more often seen in direct preparations [124, 125]. False-positive trisomy 7 or 18 can occur with either technique. To add to the complexity, it should be kept in mind that true trisomy 2 and trisomy 7 mosaicisms have been documented in liveborn children after having been diagnosed prenatally by amniocentesis [132, 133].

Maternal cell contamination (MCC) in CVS is generally due to the lack of complete separation of chorionic villi from maternal decidua, and it is reported in an estimated 1.0–1.8% of cases [121, 123, 124]. The MCC reported in these studies is about half of the figures above, reflecting the XX/XY admixtures, and is doubled to account for the likely equal incidence of MCC in female fetuses. MCC occurs more often in cultured cells than in direct preparations, thus underscoring the importance of using both methods in a full CVS cytogenetic analysis. In one report, the rate of MCC was significantly higher in specimens obtained by the transcervical method (2.16%) than in samples obtained by the transabdominal method (0.79%) [121].

A note of caution is prudent here. Generally, when there is a discrepancy between the direct and the cultured preparations, a subsequent amniocentesis is considered to provide the “true” result. However, a case of mosaic trisomy 8 reported by Klein et al. illustrates the fact that a true ­low-level tissue-specific mosaicism can exist [134]. In this case, the CVS showed a normal direct preparation and mosaic trisomy 8 in culture. Subsequent amniocentesis showed normal chromosomes, but peripheral blood cultures of the newborn showed trisomy 8 mosaicism. Therefore, when considering amniocentesis or PUBS as follow-up studies because of possible CPM observed in CVS, factors such as the specific aneuploidy involved, the likelihood of detecting it using a given sampling technique, and the risks of the additional invasive procedure need to be weighed.

The minimum amount of chorionic villus material necessary to obtain diagnostic results and the transport medium should be established in advance with the laboratory. In general, a minimum of 10 mg of tissue is needed to obtain both a direct and a cultured cell result, and 20 mg is ideal. If possible, the specimen should be viewed through a dissecting microscope to ensure that villi are present. The specimen should be transported at ambient temperature to the cytogenetics laboratory as soon as possible.

Percutaneous umbilical blood sampling (PUBS) is also known as periumbilical blood sampling, fetal blood sampling, or cordocentesis. The largest series in the literature regarding risks of PUBS included outcomes of 1,260 diagnostic cordocenteses among three fetal diagnosis centers and 25 practitioners[135]. A fixed needle guide was used in this study, and prospective data was compared to the published experience of large centers that use a freehand technique, where a 1–7% fetal loss rate has been reported. The procedure-related loss rate at a mean gestation of 29.1 +/−5 weeks at the time of sampling was 0.9%, leading to the conclusion that technique is a variable in the loss rate for cordocentesis.

PUBS experience at an earlier gestation was described by Orlandi et al. in 1990, who pointed out that, while cordocentesis was a technique largely confined to the middle of the second trimester to term, in their experience it could be performed as early as the 12th week with acceptable results [136]. They evaluated the outcomes of 500 procedures performed between 12 and 21 weeks for thalassemia study (386), chromosome analysis (97), fetomaternal alloimmunization (10), and infectious disease diagnosis (7). One practitioner performed the procedures, and the volume of blood obtained ranged from 0.2 to 2.0 mL, depending on the gestational age. Of the 370 pregnancies not electively terminated and for which outcome information was available, the fetal loss rate was 5.2% for fetuses of 12–18 weeks’ gestation and 2.5% between 19 and 21 weeks. Indicators of adverse ­outcome included cord bleeding, fetal bradycardia, prolonged procedure time, and anterior insertion of the placenta. Fetal bradycardia is a commonly reported complication after PUBS and is associated with a higher likelihood of fetal loss. In a review of 1,400 pregnancy outcomes after PUBS, the overall incidence of recognizable fetal bradycardia was estimated at 5% [137]. It was significantly more likely to occur when the umbilical artery was punctured. Boulot et al. performed 322 PUBS and noted fetal bradycardia, usually transitory, in 7.52% of their cases [138]. Fetal bradycardia occurred in 2.5% of cases with normal outcome and in 12.5% of cases of fetal loss in one study, while in another, 11 of 12 fetal losses were associated with prolonged fetal bradycardia [136, 137].

The underlying fetal pathology is a significant factor in fetal loss rate. Of these 12 losses, 10 were fetuses with a chromosome abnormality or severe fetal growth restriction. In gestations from 17 to 38 weeks, Maxwell et al. compared the loss rates within 2 weeks of the procedure with the indications [139]. Of 94 patients having prenatal diagnosis with normal ultrasound findings, one pregnancy of the 76 that were not electively terminated was lost. Of the group with structural fetal abnormalities, 5 in 76 were lost; and in the group of 35 with nonimmune hydrops, 9 were lost. It is important to take this factor into account when counseling patients before the procedure.

It has been said that no other fetal tissue “…can yield such a broad spectrum of diagnostic information (cytogenetic, biochemical, hematological) as fetal blood” [136]. As a means of fetal karyotyping, it has the advantage of generating results in 2–4 days, compared to 6–14 or more for ­amniotic fluid and CVS cells. When pseudomosaicism or mosaicism is seen in amniotic cell cultures, PUBS can provide valuable additional information regarding the likelihood of true mosaicism and thereby assist the couple in their decision making [140–143].

Although pseudomosaicism in amniotic fluid cell cultures is usually associated with normal chromosome analysis after PUBS, the absence of trisomic cells in fetal blood does not guarantee that mosaicism has been definitely excluded [144]. For example, fetal blood karyotyping is not useful for the evaluation of mosaic or pseudomosaic trisomy 20. For further discussion of mosaicism, see “Special Issues” later, and see also Chap. 8.

Because PUBS is associated with a significantly higher fetal loss rate than other prenatal diagnostic procedures, use of this technique should be recommended and provided with great care and only in certain high-risk situations such as those mentioned previously.

Ideally, 1–2 mL of blood should be obtained and put into a small sterile tube containing sodium heparin. Results can usually be obtained from 0.5 mL, and in some cases 0.2 mL, so even small amounts obtained should not be discarded. A Kleihauer-Betke test may be useful in evaluating the possibility of maternal cell admixture, particularly when a 46,XX karyotype results.

Six causes have been proposed for nuchal thickening/folds:

The fluid collects in the posterior neck fold, causing the appearance of a nuchal membrane separation on ultrasound examination (Fig. 12.3). With resolution of the fluid collection, a nuchal fold or thickening develops.

Nuchal membranes have been recorded as early as 9 weeks’ gestation. Measurement of the nuchal thickness, with or without first-trimester serum screening, has become the most sensitive first-trimester ultrasound finding used for Down syndrome detection [146–148]. Nicolaides, a pioneer of first-trimester nuchal thickness ultrasound scans, cites a detection rate of 90% for chromosome abnormalities when performed in conjunction with pregnancy-associated plasma protein A (PAPP-A) and free ß-hCG at 11–14 weeks of pregnancy, with an invasive pregnancy testing rate of 5% [146].

Nuchal folds and cystic hygromas have been known to be associated with chromosome abnormalities since 1966, with an incidence of chromosome abnormalities ranging from 22% to more than 70% in various series [149]. Based on 22 other studies, plus their own data, Landwehr et al. found that 32% of 1,649 karyotyped fetuses with nuchal folds or membranes and/or cystic hygromas had a chromosome abnormality. These included 207 cases of trisomy 21; 108 cases of trisomy 18; 30 cases of trisomy 13; 131 cases of 45,X; and 48 other chromosome abnormalities. This study included first- and second-trimester ultrasound scans, which employ different criteria for nuchal thickness.

In a 12-center study designed to determine the sensitivity and specificity of second-trimester soft-tissue nuchal fold measurement for the detection of trisomy 21, 3,308 fetuses of 14–24 weeks’ gestation were evaluated [150]. Using 6 mm as a cutoff, a nuchal skin fold was seen in 8.5% of chromosomally normal fetuses and in 38% of those with trisomy 21. A false-positive rate below 5% was obtained by 81% of the investigators. The authors concluded that this sign is useful in skilled hands in the second trimester, but it does not appear suitable for population screening because of the high variability in the results among the investigators.

A nuchal thickness cutoff of 4 mm was chosen by Nadel et al. in a study of 71 fetuses of 10–15 weeks’ gestation, of which 63 were karyotyped [151]. Abnormal karyotypes were found in 31 of 37 hydropic fetuses and in 12 of 26 nonhydropic fetuses. The nonhydropic fetuses also had no septations in the hygromas. Twenty-two of the fetuses with septated hygromas had chromosome analysis, and 19 had abnormal chromosomes. Of fetuses with hydrops and no septations, 11 of the 14 had abnormal chromosomes.

There have been several first-trimester ultrasound studies of nuchal thickening. Van Vugt et al. karyotyped 102 first-trimester fetuses with a nuchal translucency of 3 mm or more and found that 46% had an abnormal karyotype: 19 had trisomy 21; 9 had trisomy 18; 13 had 45,X; 1 had 47,XXX; and 5 had other chromosome abnormalities [152]. Multiple logistic regression analysis was used to take into account data modifiers such as gestational age and maternal age. The authors examined the septated versus the nonseptated nuchal translucencies. Septa were seen in 45 (44%) of the fetuses, of whom 36 (80%) had chromosome abnormalities. Of 57 fetuses with no septation, 11 (19%) had abnormal chromosomes. This compared to a 56% incidence of chromosome abnormalities in first-trimester fetuses with septation and 23% incidence of chromosome abnormalities in first-trimester fetuses without septation in Landwehr’s study [149].

In 1,015 fetuses of 10–14 weeks’ gestation with nuchal fold thicknesses of 3, 4, 5, and >5 mm, Pandya et al. found incidences of trisomies 21, 18, and 13 to be approximately 3 times, 18 times, 28 times, and 36 times higher than the respective numbers expected on the basis of maternal age alone [153]. This corresponded to risks of one of these chromosome abnormalities to be 5, 24, 51, and about 60%, respectively.

Using a 4-mm cutoff in fetuses of 9–13 weeks, Comas et al. detected 57.1% of aneuploidies with a false-positive rate of 0.7% and a positive predictive value of 72.7% [154].

Szabó et al. evaluated 2,100 women under 35 years of age by ultrasound at 9–12 weeks’ gestation [155]. Women were offered CVS if the nuchal fold was 3 mm or greater. The authors found an incidence of first-trimester nuchal fold to be 1.28% in women under 35, with a corresponding percentage of chromosome abnormalities being 0.43%. This indicated a 1 in 3 risk for chromosome aneuploidy in this age group when a thickened nuchal fold was seen.

Given that nuchal thickening is clearly associated with chromosome abnormalities, most commonly trisomy 21, and that it is the most common abnormal ultrasound finding in the first trimester, ultrasound evaluation of nuchal thickness in the first trimester in combination with maternal serum markers has proven to be one of the most important early screening tools to evaluate an increased risk of aneuploidy [156]. In a review of ultrasound diagnosis of fetal abnormalities in the first trimester, Dugoff cites the work of Hyett et al., who reported on an association between increased nuchal translucency and heart abnormalities. In that study, the prevalence of major cardiac defects increased with nuchal thickness from 5.4 per 1,000 for translucency of 2.5–3.4 mm to 233 per 1,000 for translucency  ≥  5.5 mm. The authors recommended that when fetuses have a thickened nuchal fold and normal chromosomes, fetal echocardiography at 18–22 weeks’ gestation is merited, besides close scrutiny of cardiac anatomy in the first trimester [157] (Fig. 12.4). This topic was reviewed by Clur et al. in 2009 [158]. They concluded that an increased NT is associated with an increased risk for congenital heart disease with no bias for one form or another. The risk increases with increasing NT measurement. In combination with tricuspid regurgitation and an abnormal ductus venosus Doppler flow profile, however, it is a strong marker for congenital heart disease. The authors recommended a fetal echo at 18–22 weeks’ gestation in fetuses with a nuchal translucency of ≥95th percentile but less than the 99th percentile. In fetuses with a nuchal translucency measurement ≥99th percentile or in which tricuspid regurgitation and/or an abnormal ductus venosus flow pattern had been found, an earlier fetal echo was recommended.

Sau et al. evaluated the significance of a positive second-trimester serum screen in women who were screen negative after a first-trimester nuchal translucency scan. Of 2,683 women screened, eight cases of trisomy 21 were detected, all of which had a positive nuchal screen result. Serum screening of 1,057 women who screened negative by nuchal translucency showed 46 high-risk results, all of which proved to be false-positive. The authors concluded that second-trimester biochemistry screening following a negative nuchal translucency screen did not increase the detection of trisomy 21 [159].

Cuckle and Maymon reported on a method whereby fetus-specific Down syndrome risks in twins could be assessed taking the other fetal nuchal fold into account [160]. This was based upon the previous report by Wøjdemann et al. that there is a correlation coefficient of 0.34 between the pairs of NTs, expressed in log multiples of the median for crown-rump length [161]. This was seen in both monochorionic and dichorionic twins. Cuckle and Maymon found a correlation coefficient in unaffected pregnancies of 0.45 (P  <  0.0001) and estimated to be 0.12 and 0.04 in discordant and concordant twins, respectively [160].

Women whose second- or third-trimester fetuses have large cystic hygromas may not have an easily accessible fluid pocket in which to perform an amniocentesis. In such cases, paracentesis of the hygroma may yield a cytogenetic result, and at fetal demise or delivery, chorionic villus or placental cell cultures may prove beneficial in obtaining chromosomal diagnosis. The yield from amniocentesis is still the greatest, so if it can be accomplished, this is still the procedure of choice for cytogenetic diagnosis in such cases [162].