Intracytoplasmic Sperm Injection (ICSI) – What are the Risks?

In vitro fertilization used in combination with intracytoplasmic sperm injection allows otherwise sterile couples to become parents. Despite recent studies on the safety of these technologies, there is still only an incomplete picture of the risks associated with the usage of these assisted reproductive techniques to offspring. The risk of multiple gestations continues to be of major concern because of its association with low birth weight, preterm delivery, and increased perinatal mortality. This article outlines the risks associated with in vitro fertilization/intracytoplasmic sperm injection as a well-defined treatment for couples with severe male factor infertility.

Over the past 30 years, the treatment of infertility has seen the development of revolutionary new assisted reproductive technologies (ARTs), first in 1978 with the birth of baby Louise Brown, who was conceived by in vitro fertilization (IVF) , then with microassisted reproduction using techniques such as intracytoplasmic sperm injection (ICSI) , and most recently with the development of preimplantation genetic diagnosis as a technique . These highly complex technologies are used with increasing frequency in the treatment of couples around the world; more than 1 million babies worldwide have been conceived in this manner . In 2003 nearly 3% of children born in Scandinavia and 1.7% of children born in France were conceived using ICSI. During this time, almost 500,000 cycles were performed in Europe (122,872 cycles in the United States), resulting in the birth of almost 300,000 infants . The number of IVF cycles in the United States has increased twofold since 1996, when Society for Assisted Reproductive Technology (SART) began monitoring the IVF programs .

Unlike most therapeutic procedures used in medicine, ARTs never underwent rigorous safety testing before clinical use. Treatments for infertility overcome natural barriers that prevent fertilization. Because these technologies are used to overcome infertility phenotypes that may have a genetic basis, the possibility exists that unwanted genetic traits may be transmitted to offspring. Although researchers believe that perhaps 75% or more cases of all infertility have a contributing genetic basis, our ability to diagnose these defects remains limited. For untreated couples, their infertility represents “lethality” within the gene pool, because their condition essentially blocks the transmission of undesirable genetic traits to any offspring. Put simply, large numbers of couples undergo fertility treatments without a complete understanding of the basis of their infertility or the potential long-term risks for their offspring.

No better example of this phenomenon exists than that of IVF/ICSI as a treatment for severe male factor infertility. Before the development of IVF/ICSI, these men simply could not have reproduced by any means. Currently, a growing percentage of the population of countries in the Western world is comprised of the offspring of these men. What health risks do these people face? What burden will they create from a public health standpoint? Questions regarding the safety of ART—and IVF/ICSI in particular—become even more important.

Retrospective data suggest that IVF and IVF/ICSI are safe. Health risks to mother and offspring that are significantly increased with assisted reproduction include multiple gestation, preterm delivery (even in singleton pregnancy), and congenital abnormalities in the offspring . Most IVF pregnancies proceed uneventfully and result in the birth of healthy babies; however, studies consistently identify an increased absolute risk of problems in IVF and IVF/ICSI pregnancies and deliveries .

It is challenging to establish the nature of these risks and dissect out whether they are related to the technology itself or the genetic defects of the parents. Other factors cloud this analysis. For example, in the United States, multiple embryos are routinely transferred to increase the ultimate likelihood of a live birth (while also enhancing the chances of a multiple gestation). Whether the altered endocrine milieu that results from hormonal induction of ovulation by ovarian hyperstimulation contributes further to this risk is unclear. Hormonal induction profoundly changes the normal uterine environment and body homeostasis. Because the oldest child conceived by IVF/ICSI is only approximately 15 years old, long-term studies of multiple cohorts of offspring conceived by IVF/ICSI from conception through to adulthood have not been possible. Incomplete follow-up caused by couples seeking treatment at large fertility clinics far from their homes is a problem. After a pregnancy is achieved, families are again followed by their community physicians. Because we are a mobile society, couples are frequently lost to follow-up for other reasons, which makes long-term follow-up of children conceived by ART difficult.

Unfortunately, epidemiologic studies regarding the safety of IVF and IVF/ICSI in general are faced with additional challenges, both in retrospective and prospective studies. These challenges include incomplete reporting and inconsistent definition of congenital abnormalities and other adverse outcomes. Importantly, the assessment of offspring conceived by IVF/ICSI is commonly performed by pediatricians as part of a routine neonatal health examination, yet a medical geneticist may have different criteria for disease. Alternatively, the physician may examine these children more closely than naturally conceived children, and inevitably the closer one looks, the greater the likelihood of finding an abnormality.

The ideal study design to answer the question of whether any ART is safe is nearly impossible to achieve. Fertility cannot be compared directly using standard statistics, because the combined fertility potential of the couple ultimately determines whether a couple is fertile or infertile. The identification of an appropriate control population for infertile couples can present a nearly impossible quest. Selection criteria may bias findings because variability between the two groups exists at baseline. Birth defects and adverse ART events are relatively rare compared with the overall total number of pregnancies and live births each year. The design and implementation of adequately powered research studies are difficult.

Despite these significant limitations, numerous investigations of IVF/ICSI safety have been performed and are reviewed in this article. Although ARTs include various methods to process oocytes and sperm in vitro to enhance the likelihood of fertilization and pregnancy, this article focuses specifically on IVF with ICSI regardless of whether the sperm are retrieved from an ejaculate, the epididymis, or the testis. Importantly, no discussion of the risks associated with IVF/ICSI can be conducted outside of the context of the existing IVF safety data. As such, the safety of IVF and IVF/ICSI are considered here.

Congenital disorders and hormonal abnormalities in children conceived by in vitro fertilization/intracytoplasmic sperm injection

Multiple large studies ( Table 1 ) consistently show a higher risk of genitourinary, cardiovascular, musculoskeletal, and gastrointestinal defects in offspring conceived by IVF and IVF/ICSI . Other older studies with less methodologic consistency also report similar findings . Comparison of the birth defect studies is confounded by differences in the definitions used to classify the birth defects, reporting of congenital abnormalities, and methods used for statistical analysis. These caveats present challenges for interpretation of the data. Given these issues, it may be difficult to conclude that the increased risk of birth defects seen in the IVF and IVF/ICSI cohorts results from ART.

Table 1

Congenital malformation in children conceived through assisted reproductive technologies

Author	Country	Year	Study type	Control	Outcomes	Sample size	Findings
Hansen	Australia	1993–1997	Registry	Yes	Congenital malformations	1138 ART children (301 IVF/ICSI, 837 IVF alone, versus 4000 SC children)	Significantly higher likelihoods of birth defects with IVF and ICSI even after correction (OR 2.0 for both)
Westergaard	Denmark	1994–97	Registry	Yes	Congenital malformations, pregnancy outcomes	2245 ART children children versus 2245 naturally conceived children (cohorts matched for maternal age, parity, and multiplicity)	No difference in risk of congenital malformation (poorer pregnancy outcomes observed in ART cohort)
Loft	Denmark	1994–1997	Registry/questionnaire	No	Congenital malformations, genetic abnormality, and pregnancy outcome in ICSI conceptions	665 questionnaires returned	No statistically significant differences identified, >90% responder rate to questionnaire
Zhu	Denmark	1997–2003	Registry/questionnaire	Yes	Time to pregnancy, treatments for infertility, congenital malformations	64,405 children (50,897 singletons and 1366 twins from fertile couples, 5764 singletons and 100 twins naturally conceived by subfertile couples, 4588 singletons and 1690 twins by ART)	A higher likelihood of congenital malformation in ART group and subfertile group with delay to spontaneous conception (> 12 mo); as delay increased so did likelihood of malformation
Koivurova	Finland	1990–1995	Registry	Yes	Congenital malformations and pregnancy outcomes	304 IVF children versus 569 SC children	All adverse pregnancy outcomes were significantly higher in the IVF group but corrected largely after consideration of multiplicity; cardiac malformations higher in IVF cohort regardless of multiplicity
Ludwig	Germany	1998–2002	Registry	Yes	Congenital malformations	3372 IVF/ICSI children versus 30,940 spontaneous conceptions	RR 1.25 (95% CI 1.11–1.40); used separate population (Mainz Birth Registry) for control cohort
Katalinic	Germany	1998–2002	Registry	Yes	Congenital malformations and pregnancy outcomes	3372 IVF/ICSI children versus 8016 naturally conceived children	Adjusted OR 1.24 (95% CI 1.02–1.50), higher incidence of birth defects in ICSI cohort after correcting for multiplicity, true control cohort
Zádori	Hungary	1995–2002	Registry	Yes	Pregnancy outcomes, neonatal complications (including birth defects)	221 ART pregnancies (185 singleton and 36 twin versus identical SC cohort)	A minimal difference between cohorts, except for a significantly higher prematurity rate in IVF singletons; IVF triplets had much higher risks of adverse outcomes in a separate analysis
Anthony	The Netherlands	1995–1996	Registry	Yes	Congenital malformations	4224 ART children versus 314,605 SC children	Risk of any congenital malformation only slightly higher, corrects after accounting for confounders [Crude OR 1.20 (95% CI 1.01–1.43), corrected OR 1.03 (95% CI 0.86–1.23)]; cardiovascular malformation higher in ART cohort regardless
Wennerholm	Sweden	1993–1998	Registry	Yes	Congenital malformations	1008 IVF/ICSI children versus all SC in Sweden over comparable time period (no number given)	Adjusted OR (after correcting for multiplicity) 1.19 (95% CI 0.79–1.81); only hypospadias higher in ICSI cohort (RR 3.0)
Ericson	Sweden	1982–1997	Registry	Yes	Congenital malformations	9111 IVF children versus 1,690,577 SC children	Unadjusted OR 1.47, adjusted 0.89; higher risk of alimentary atresia, neural tube defects, and hypospadias in ICSI cohort even after correction
Sutcliffe	UK	1989–1994	Prospective	Yes	Congenital malformation and neurologic development in cryopreserved embryos	91 ART children (cryopreserved embryos) versus 83 SC children	Incidences were higher in the ART cohort, although no numbers reached statistical significance
Olson	US	1989–2002	Registry	Yes	Congenital malformation and mortality	1462 ART children versus 8422 SC controls	Adjusted OR 1.30 (95% CI 1.00–1.67), higher incidence of birth defects in IVF cohort even after correcting for multiplicity
Rimm	NA	2004	Meta-analysis	Yes	Compiled 16 IVF studies, 7 ICSI studies	28,524 IVF versus 2,520,988 SC children; 7234 IVF/ICSI children versus 978,078 SC children	Pooled OR of 1.29 (95% CI 1.01–1.67) (statistically significant); all studies examined had design flaws
Lie	NA	2005	Meta-analysis	Yes	Compiled 4 prospective, well-designed studies (out of 22)	5395 IVF/ICSI children versus 13,086 IVF children (pooled from all four studies)	Pooled RR 1.12 (95% CI 0.97–1.28), no significant increase for any single category of defect with ICSI
Hansen	NA	2005	Meta-analysis	Yes	Compiled 25 studies (only 7 well-designed per author’s criteria)	28,638 ART children	On both analyses (all 25 studies or limited just to the 7 appropriate studies) pooled OR of malformation was significantly higher; 1.40 (95% CI 1.28–1.53) on analysis of 7 well-designed studies

Abbreviations: CI, confidence interval; OR, odds ratio; RR, relative risk.

In this regard, several studies are noteworthy. Hansen and colleagues reported a higher likelihood of congenital abnormalities in children conceived by IVF and IVF/ICSI based on a study of the birth registry of Western Australia, which focused on 1138 children conceived using ART (301 ICSI children, 837 standard IVF children). This report was followed by a meta-analysis of 25 existing studies of birth defects in children conceived by ART . There was a significantly increased incidence of birth defects in children conceived by ART, including cardiovascular, urogenital, musculoskeletal, and chromosomal abnormalities in the IVF cohort. This trend was not statistically significant in the IVF/ICSI cohort, with the exception of musculoskeletal and chromosomal abnormalities, perhaps because of inadequate sample size in the IVF/ICSI cohort. Overall, this meta-analysis demonstrated a higher likelihood of birth defects in children conceived by IVF and IVF/ICSI.

A cohort of offspring conceived by IVF/ICSI has been followed over time by the Bonduelle group . This group of 8-year-old children conceived by ICSI was compared with a group of naturally conceived children . Although the children conceived by IVF/ICSI were generally healthy and had no higher likelihood of requiring surgery, hospitalization, or rehabilitation, there was an increased likelihood of congenital malformation in the IVF/ICSI cohort. Approximately 10% of these children had major birth defects, compared with 3.3% of the control group. Minor birth defects were similar in the children conceived by IVF/ICSI and the control group. Using the Western Australian birth registry system referenced in the Hansen studies, 6 of 150 children conceived by IVF/ICSI had a major malformation compared with 1 of 147 of the control children.

Again, some controversy remains. Analysis of the Danish National Birth Registry suggested no greater likelihood of congenital abnormalities for children born to women with significant delay to natural conception (>12 months) compared with children conceived by IVF/ICSI, with one exception . The incidence of genitourinary tract abnormalities is statistically significantly increased in children conceived by IVF/ICSI. These findings are consistent with other studies, which suggest that patients with subfertility are at higher risk of having a child born with congenital abnormalities. This finding argues that the ART procedures in and of themselves do not contribute to this risk.

Throughout the world, the criteria used to define birth defects differ, which presents unique challenges to analysis of ART safety data. This is best illustrated in a study from Kurinczuk and Bower , who analyzed an earlier study published by the Bonduelle group. Four hundred twenty-three children were followed prospectively after IVF/ICSI in the Bonduelle study with no evidence of increased major congenital malformations. Kurinczuk correctly observed that the control population used in this series included spontaneous conceptions pooled from worldwide registries, including the Western Australian birth registry. Again, the different definitions of birth defects in Australia and Belgium (the location of the Bonduelle group) were not considered in the study; when Kurinczuk recategorized the IVF/ICSI cohort based on Western Australian birth registry standards, approximately a twofold higher risk of major congenital malformation in the IVF/ICSI cohort was observed. Clearly, standardization of reporting of birth defects should be mandatory for any study of ART safety given the fact that these two analyses reached different conclusions, albeit with the same data.

An increased incidence of genitourinary tract abnormalities (specifically hypospadias in boys) is consistently found in offspring conceived by ART. Although impaired or abnormal hormonal function is one possible cause of hypospadias, the causes of these defects in some offspring conceived by IVF/ICSI remain unknown. Genital tract abnormalities in some male parents of offspring conceived by IVF/ICSI (eg, hypogonadism or poor testis function) raise the question of whether the congenital genitourinary defect in offspring conceived by IVF/ICSI is the consequence of a genetic abnormality inherited from their fathers. A recent Danish series did observe statistically significantly lower testosterone levels in 125 male offspring conceived by ICSI when compared with testosterone levels in age-matched boys who were naturally conceived .

Detrimental effects of oocyte handling and in vitro maturation of immature oocytes were considered possible contributors to the hormonal and developmental effects found in offspring . Studies to date show no risk with in vitro maturation; however, this area remains one of active investigation.

Multiple gestations, preterm labor, and other perinatal complications are more common in pregnancies that result from in vitro fertilization/intracytoplasmic sperm injection

In the United States, most pregnancies conceived by ART result in a multiple-birth delivery. In contrast, only 1.5% of naturally conceived pregnancies result in a multiple birth . Multiple gestation is associated with an increased risk of preterm delivery, low birth weight, and increased perinatal mortality. Theoretically, multiple gestation could even account for higher risks of seemingly unrelated conditions in children conceived by IVF/ICSI, such as the risk of cerebral palsy. As such, it is important to consider the reported differences observed in IVF and IVF/ICSI cohorts .

When considering this line of reasoning, at first glance it is difficult to account for the fact that several meta-analyses demonstrate an increase in perinatal complications even in singleton pregnancies that result from IVF . The IVF/ICSI cohorts in these series, regardless of multiplicity, had higher likelihoods of low birth weight, preterm labor (< 36 weeks), and usage of hospital resources (ie, the need for surgery or neonatal intensive care unit admission).

The Danish National Patient Registry provided a large body of IVF/ICSI data; in this population, a higher likelihood of adverse outcomes was observed in multiple gestational pregnancies. Importantly, no such trend existed in IVF/ICSI singleton pregnancies . These authors addressed the inconsistencies between their own findings and the meta-analyses and hypothesized that vanishing twin syndrome in these pregnancies plays a causal role in the poorer outcome observed in IVF/ICSI singleton pregnancies . This risk relates to the number of embryos transferred. Importantly, a multicenter, prospective, randomized controlled trial compared outcomes after single-embryo transfer (SET) with transfer of multiple embryos. Pregnancy rates were not substantially lower in the SET cohort, but the number of multiple births was dramatically reduced .

SET represents an important intervention because it allows for an adequate success rate to be maintained with IVF and IVF/ICSI while it significantly decreases the risks in offspring that derive from multiple gestation . Remarkably, SET is not routinely performed as standard of care in the United States, and it is still only offered to few women in Europe . The idea that routine performance of SET can mitigate risks such as preterm labor in offspring conceived by IVF/ICSI might seem obvious to most readers, but additional benefits such as a decreased rate of cerebral palsy in offspring conceived by IVF/ICSI would help to minimize the risk of conditions, such as neurologic and developmental delay, in offspring conceived by ICSI cohorts. Prospective research with the goal of confirming this hypothesis is vitally important.

Only gold members can continue reading. Log In or Register to continue