Over the past three decades, multiple reports from individual centers have demonstrated the safety and low rates of pregnancy loss following CVS [22, 23]. There is a large amount of literature, which will not be extensively reviewed here, from the 1980s and 1990s that has detailed the development of CVS as a clinical procedure and led to the US Food and Drug Administration’s removal of the restrictions on the use of the CVS catheter in 1989. The large majority of these projects did not find any increased risk of CVS over amniocentesis, but overall loss rates of both procedures were considerably higher than claimed by many operators of both procedures [22, 23].

Three studies in the past decade have been very important in our current understanding. The first was a meta-analysis published in 2007 by Mujezinovic and Alfirevic [5]. Because of the higher background rate of loss in the first trimester, CVS would be expected to have a higher overall loss rate. Yet it is impossible precisely to parse which losses are procedure-related, as opposed to background. What they reported is that, in experienced centers, the loss rates within 2 weeks of the procedure, up to 24 weeks and at term were essentially identical (Table 20.1).

The second study is the recent ongoing experience of the Danish first-trimester screening program, which over the past several years has maintained a very high detection rate for Down syndrome (>90 %) while cutting their false positive rates in half. As a consequence of first-trimester screening, the utilization of CVS in Denmark has been three times that of amniocentesis. Their data show that the procedure risk of CVS is as low (and possibly lower) than amniocentesis [24]. Furthermore, the incidence of late complications such as late demise was significantly lower in CVS patients than amniocentesis patients.

Recently Akolekar published a meta-analysis that demonstrated that overall procedure loss rates for CVS and amniocentesis were much lower than previously shown, and not different from each other—approximately 1/500 [6]. The primary confusion is the difference in background loss rates, which actually explain the vast majority of any differences between the two procedures.

Overall, we counsel our patients as to a 1/400 procedural risk of either CVS or amniocentesis in very experienced hands. These risks may, in fact, err on the side of being too high in our hands, but we do not want patients taking these decisions frivolously without considering the possibility of loss. Importantly, both procedures have considerably higher risks in inexperienced physicians or those who cannot visualize the anatomy in a 3D manner as described above. As such we believe that both CVS and amniocentesis should be performed preferably in “centers of excellence” by physicians, with experienced sonographers, who perform the procedure regularly—and not as an occasional item.

In the early 1990s it was suggested that CVS may be associated with specific fetal malformations, particularly limb reduction defects (LRDs). Today, based on the published data, it appears safe to state that there is no increased risk for LRDs, or any other birth defect, when CVS is performed at >70 days of gestation [30–33].

The first suggestion of an increased risk for fetal abnormalities following CVS was reported by Firth et al. [7]. In a series of 539 CVS-exposed pregnancies, there were five infants with severe limb abnormalities in a cohort of 289 pregnancies sampled by TA-CVS at 55–66 days gestation. Four of these infants had the unusual and very rare oromandibular-limb hypogenesis (OLH) syndrome, and the fifth had a terminal transverse LRD. OLH syndrome occurs in 1 per 175,000 live births [34], and LRDs normally occur in 1 per 1690 births [35]. Thus, the occurrence of these abnormalities in more than 1 % of CVS-sampled cases raised a high level of suspicion. Subsequently, other groups reported the occurrence of LRDs and OLH following “early” CVS [36–41]. In 1992, a case–control study using the Italian Multi-Center Birth Defects Registry, reported an odds ratio of 11.3 (95 % CI 5.6–21.3) for transverse limb abnormalities following first-trimester CVS [36]. However, when stratified by gestational age at sampling, pregnancies sampled prior to 70 days had a 19.7 % increased risk of transverse limb reduction defects, while patients sampled later did not demonstrate significantly increased risk. Other case–control studies, however, have not seen any association of CVS with LRDs [31, 32].

The risk for fetal malformations is real when CVS is done at an earlier gestation age, i.e., during the period of limb genesis at 6–7 weeks [7, 8, 38]. Brambati et al., practicing in Milan, have a large population at risk for β thalassemia. The population is also predominantly Catholic, for whom abortion at any gestational age is religiously proscribed. However, to assuage the guilt of wanting diagnosis, they wanted it done as early as technically possible. They reported a 1.6 % incidence of severe LRDs in a group of patients sampled at 6 and 7 weeks gestation [38]. This rate decreased to 0.1 % at 8–9 weeks. These data support the assumption that early gestational sampling and excessive placental trauma may be etiologic in the reported clusters of post-CVS LRDs.

In the USA, the population group most interested in early CVS has been the Orthodox Jewish community—at high risk for Tay Sachs and other Ashkenazi diseases. In the observant Jewish community, abortion is permitted, by religion, until 40 days post-conception (7 weeks 5 days from LMP). Wapner and Evans have shown that, in very experienced centers, CVS can be safely and reliably performed, even in very early gestation [30]. In a study they conducted, CVS was performed at less than 8 weeks gestation in a population of Orthodox Jews. Of the 82 cases of early CVS, there was only a single case of severe LRDs, a rate of 1.6 %. While this risk is considerably higher than when done at the usual time, we believe that in comparison to a 25 % risk of lethal disorder and with proper genetic counseling, it could be a reasonable decision for a couple to make. However, in the one case, despite the patient having had three previous early CVSs, they chose to sue the doctor stating that they did not “know.” As a result of the legal exposure, virtually all centers that were willing to do the procedure, stopped performing it.

The question whether CVS sampling after 10 weeks has the potential of causing more subtle defects, such as shortening of the distal phalanx or nail hypoplasia, was a major concern, debated thoroughly in the literature [40, 42]. The overall incidence of LRDs after CVS is estimated to be 1 in 1881 (ranging from 5.2 to 5.7 per 10,000), compared with 1 in 1642 (ranging from 4.8 to 5.97 per 10,000) in the general population [33, 42]; hence, there are no data to substantiate this concern. As noted, in most experienced centers performing CVS after 10 weeks, no increase in limb defects of any type was observed [31, 33, 41, 42].

Early amniocentesis is a procedure that has come and gone. It is a first-trimester procedure, i.e., performed before 14 weeks of gestation (usually from 11 + 0 to 13 + 6) [59, 60]. Some series have included procedures as early as 9 + 0 weeks. Traditional amniocentesis is usually performed after 15 + 0 weeks of gestation; invasive procedures between 14 + 0 and 14 + 6 are usually considered early and have been included in some series but not others [61]. Since 1987, various sized observational studies on EA reported rates of procedure-related fetal loss from 1.4 to 8.1 % (see Table 20.1). Early series concluded that EA is an appropriate technique for early diagnosis but is associated with an increased fetal loss rate [62, 63]. Assel et al. [64] compared EA with mid-trimester amniocentesis and found a significant increased post procedure fetal loss rate (1.8 % vs. 0.4 %) [65].

A prospective partially randomized study by Nicolaides et al. [66, 67] showed a higher rate of fetal loss after EA compared with CVS (4.9 % vs. 2.1 %), which was significant for pregnancies at 10–11 weeks but not significant for the 12–13 weeks gestation period.

The CEMAT study compared EA between 11 + 0 and 12 + 6 weeks with standard amniocentesis (15 + 0–16 + 6). This multicenter randomized trial reporting on 1916 EA procedures showed an increased total pregnancy loss (pre-procedure and post procedure losses including intrauterine and neonatal deaths) with the EA procedure (7.6 % vs. 5.9 %; P = 0.012) [68].

Fetal reduction (FR) has changed considerably over the last 25 years since we first published on the subject [15, 73]. These changes have taken place in medical technology outcomes, patient choices, and the larger demographic and cultural shifts that are driving the pace and direction of change.

At its core, FR started out as a way of managing pregnancies in which the risks to both mother and fetuses from carrying multiple embryos were extreme. Selective termination (as it was called then) of some of the embryos to increase the viability of the remaining ones and reduce the risk of morbidity and mortality for the mother was a desperate approach to salvage the situation. As with numerous other technological changes, what began as a dominant concern with matters of life and death has eventually become accepted. Then indications transform from crisis “life and death” into issues of quality of life [74, 75].

FR was developed as a clinical procedure in the 1980s, when a small number of clinicians in both the USA and Europe attempted to reduce the usual and high adverse sequelae of multifetal pregnancies, by selectively terminating or reducing the number of fetuses to a more manageable number. The first European reports by Dumez and Oury [20], and the first American report by Evans et al. [73], followed by a further report by Berkowitz et al. [76], and later Wapner et al. [77] described a surgical approach to improve the outcome in such cases.

Multiple papers in the 1990s demonstrated that with triplets or more, there was clear improvement in reducing to twins. Numerous papers argued whether triplets had better outcomes “reduced” or not. Yaron et al. [78] compared triplets-to-twins data to unreduced triplets with two large cohorts of twins. The data showed substantial improvement of reduced twins as compared to triplets. The data from the 2001 collaborative series and others suggested that pregnancy outcomes for cases starting at triplets or even quadruplets reduced to twins at about 12 weeks do fundamentally as well as starting as twins.

Overall, statistics on reductions have improved noticeably over these past 25 years [14, 79, 80]. In the early 1990s, when half the cases were quadruplets or more, loss rates (up to 24 weeks) were 13 %. Early premature deliveries were an additional 10 %. Now, overall with decreasing starting numbers, better ultrasound, better understanding of zygosity, and a limited number of practitioners with extensive experience accounting for a high percentage of reductions, losses are overall down to about 4 %. Counseling must be tailored to specific starting and finishing numbers and the experience of the operator.