Abstract
While intracytoplasmic sperm injection (ICSI) is the most significant advance in assisted reproductive technology (ART) for the alleviation of male factor subfertility, its use has become increasingly widespread and indiscriminate in ART clinics, extending well beyond the reasons for its necessary application. But ICSI is not “better” than IVF using any established outcome metric. Indeed, available evidence indicates that ICSI yields fewer embryos per treatment cycle, embryos which may have impaired developmental potential compared to IVF-derived embryos. This chapter investigates the basis for the over-use of ICSI, and identifies risks to which couples are exposed by the unjustified use of ICSI: a debate that has been raging for two decades, and is now also extending into considerations of “andrological ignorance”, how ICSI has effectively blocked scientific advances in andrology, and how obligate ICSI has effectively transferred the treatment burden for male factor infertility to the female partner, who is expected to undergo possibly unnecessary controlled ovarian hyperstimulation, oocyte retrieval and embryo transfer procedures.
11.1 Introduction
Unarguably, intracytoplasmic sperm injection (ICSI) is the most significant advance in assisted reproductive technology (ART) for the alleviation of male factor subfertility. However, its use has become increasingly widespread and indiscriminate in ART clinics, extending well beyond the reasons for its necessary application, which are:
cases of severe male factor subfertility where there is actual evidence for a serious risk of impaired sperm function that would lead to reduced or failed fertilization in vitro (also includes autoimmune infertility caused by antibodies directed against the sperm head, and many – although certainly not all – cases of retrograde ejaculation);
couples in whom conventional in vitro fertilization (IVF) has failed due solely to sperm dysfunction, such as failure to acrosome react;
use of spermatozoa recovered surgically from the male reproductive tract;
use of spermatozoa that were cryopreserved in finite limited quantities, e.g. prior to chemo- or radiotherapy or prior to vasectomy, or with known poor cryosurvival;
cases where pre-implantation genetic testing is to be performed for monogenic conditions, and there is a real risk of possible contamination with DNA from spermatozoa still attached to the zona pellucida; and
with cryopreserved or vitrified oocytes where sperm–oolemma fusion might be compromised.
The proposition for the widespread, even “universal,” use of ICSI is based on the commonly held perception among ART clinicians, as well as by numerous embryologists, that ICSI gives a higher fertilization rate than conventional IVF, and that it satisfies their desire to avoid unexpected IVF fertilization failure. In 2000, Fishel and colleagues proposed that ICSI should be used as a first option because it: (i) offered a higher incidence of fertilization; (ii) maximized the number of embryos; and (iii) minimized the risk of complete failure of fertilization for all cases requiring in vitro conception [1]. However, they did note that, among other concerns, the then current knowledge of ICSI birth outcomes did not provide the confidence to use it in all cases of IVF for the time being. But a year later, Ola and colleagues concluded that from safety, scientific and economic viewpoints, ICSI should only be used in cases where success of IVF was regarded as unlikely [2].
While wishing to avoid having to explain to a man why his sperm “didn’t work” with conventional IVF is perhaps understandable, the indiscriminate, even total, use of ICSI is increasingly being seen as unacceptable, and some jurisdictions have regulated against this approach. For example, an expert Panel Report for the Ontario (Canada) Ministry of Children and Youth Services stated that ICSI should be provided only for individuals where either severe male factor infertility is present, or there is demonstrated fertilization failure in a previous IVF cycle. Clearly, a “poor” semen analysis, even one with several characteristics below the World Health Organization reference values for recently fertile men [3], is not adequate justification for using ICSI.
This chapter investigates the basis for this reasoning, and identifies the risks to which couples are exposed by the unjustified use of ICSI, a debate that has now been raging for two decades [4–7] and is now also extending into considerations of “andrological ignorance,” how ICSI has effectively blocked scientific advances in andrology, and how the over-reliance on ICSI has effectively transferred the treatment burden for male factor infertility to the female partner, who is expected to undergo possibly unnecessary controlled ovarian hyperstimulation, oocyte retrieval and embryo transfer procedures [8].
11.2 Does ICSI Have Any Real Benefit in Non-Male Factor Infertility Cases?
11.2.1 Fertilization Rates Are Not Higher Using ICSI
While some studies have reported higher fertilization rates with ICSI than with IVF, a 2003 Cochrane Review concluded that IVF gave better fertilization results than ICSI in couples with male factor subfertility [9]. These authors also noted that pregnancy rates following IVF and ICSI were comparable for couples with non-male subfertility, and also that, if anything, ICSI did not improve on the IVF outcome in these couples.
Analyzing 486 couples without a diagnosis of male factor [10] reported that among the 99 cycles in which ICSI was used (there were no IVF/ICSI “splits”) the fertilization rate was lower than among the 598 IVF cycles (61.7% vs 72.9%, p < 0.001), and the fertilization failure rate was similar at 4% cf. 3%. Clinical outcomes as measured by positive ß-hCG and live birth rates were not different between the IVF and ICSI treatment cycles (40% cf. 32% and 22% cf. 17%, p = 0.13 and 0.24, respectively).
Nyboe Andersen and colleagues [11] and Carrell and colleagues [6] concluded that:
ICSI is invaluable in treating patients with male factor infertility, however, such patients should be evaluated by an andrologist to assure that ICSI is indeed necessary and that broader health concerns are addressed. ICSI may also be indicated in other situations, including the use of cryopreserved oocytes, and in conjunction with PGD, however, the data clearly show no benefit in routine use of ICSI for other patients.
Table 11.1 shows a model analysis of the number of zygotes that would be generated following IVF and ICSI from a “typical” cohort of 13 cumulus–oocyte complexes of which 85% contained MII oocytes by laboratories operating at the “competency” and “benchmark” performance levels, as defined by the Vienna consensus on ART laboratory key performance indicators [12]. In either situation, IVF would generate more zygotes than ICSI (7.8–9.8 vs 6.5–8.4). This represents, on average, between 1.3 and 3.3 fewer zygotes when ICSI is used in a case where IVF would have worked, or a 50% greater outcome for a benchmark IVF lab compared to a competent ICSI lab (9.8 vs 6.5 zygotes). Assuming both labs then had benchmark blastocyst development rates of 60%, the difference would result in two more blastocysts in the IVF lab compared to the ICSI-only lab.
Table 11.1 Model comparing IVF with ICSI in terms of the number of 2PN zygotes generated in a generic typical cycle considering “competency” and “benchmark” laboratory performance levels as per the Vienna Consensus [12].
| Parameter | IVF | ICSI | ||
|---|---|---|---|---|
| Number of COCs | 13 | 13 | ||
| Number of MII oocytes @ 85% | 11.05 | |||
| Damaged |
|
| ||
| Inseminated or injected | 13 | 9.95 | 10.50 | |
| Normal fertilization rate | Competency (60%) | Benchmark (75%) | Competency (65%) | Benchmark (80%) |
| Number of 2PN zygotes | 7.80 | 9.75 | 6.47 | 8.40 |
Note: Normal fertilization rate for IVF is defined as the proportion of 2PN zygotes per cumulus–oocyte complex (COC) inseminated, and for ICSI as the proportion of 2PN zygotes per MII oocyte injected.
Another situation where ICSI is often perceived to be better practice is in cycles where there are very few oocytes available. This was investigated by Borini and colleagues [13] who found no benefit in using ICSI, and concluded that “Performing ICSI in all cases of IVF is not advantageous, probably only more expensive and time consuming,” and commented that “Abandoning IVF appears to be questionable.” Considering ICSI for non-male factor infertility in older women, there was still no advantage of ICSI over IVF [14].
11.2.2 What is the Real Risk of Total IVF Fertilization Failure (TIFF)?
Many centres have published a prevalence of TIFF of 5–10%, or even higher, yet others report values of 2% [2]. This discrepancy, and the finding of high rates of TIFF, are the result of two main issues: (i) poor andrological (more correctly, spermatological) evaluation of the male partners; and (ii) poor sperm handling/preparation/capacitation systems that result in impaired sperm function in vitro. Centres with these issues create a self-fulfilling prophesy: sperm will show poor function, thereby reducing IVF fertilization rates and causing a high prevalence of TIFF. Although ICSI would eliminate this iatrogenic problem of TIFF due to poor sperm function, what is really needed is just better IVF.
The American Society for Reproductive Medicine (ASRM) has stated that in routine ART for non-male factor infertility the risk of failed fertilization is low, and a similar frequency is found after IVF and ICSI [15]. With optimized IVF lab systems, TIFF rarely exceeds 2–3%, and the Vienna Consensus sets a limit of 5% [12] against a generally established background ICSI fertilization failure rate of 1–2%. Certainly, patients could choose to use ICSI if they found a 1–2% incremental risk of TIFF to be unacceptable, but it would be irresponsible to allow them to believe that the prevalence of TIFF is 10% or higher – although, if it were that high then the lab would not be considered to meet good practice standards [12].
From our own experience [16], after performing a careful assessment of the sperm prior to recommending IVF or ICSI treatment, low fertilization (<25% of eggs inseminated or injected) and failed fertilization at IVF can be no greater than levels seen in ICSI cases (1.9% and 1.7% for IVF vs 3.0% and 2.4% for ICSI; Table 11.2-A). More importantly, when analyzing such IVF cases, most abnormal IVF outcomes occurred in cycles with abnormal ovarian stimulation response, either very few mature oocytes or large numbers of oocytes (>15: Table 11.2-B). Key parameters used in the pre-treatment sperm assessment to differentiate these couples into IUI, IVF or ICSI treatment recommendations were: sperm morphology as assessed using the teratozoospermia index (TZI), and not the % normal forms; and the yield from a PureSperm density gradient “trial wash”; and sperm hyperactivation assessed using CASA (Hamilton Thorne IVOS, Beverly, MA, USA) analytical methods which can be found in [17]. Consequently, with adequate sperm pre-assessment, IVF can be recommended with good confidence.
Table 11.2 A: Prevalence of low and failed fertilization cases, and B: Stimulation responses in the IVF low and failed fertilization cases. Data from 2006–2014 at Atlantic Assisted Reproductive Therapies, Halifax, NS, Canada [16].
| A | B | ||||||
|---|---|---|---|---|---|---|---|
| Fertilization | IVF | ICSI | P | Stimulation response | Fertilization | ||
| 0% | 1–24% | ||||||
| (n = 14) | (n = 16) | ||||||
| n = 830 | n = 1129 | Normal | 5–15 COCs | 2 | 3 | ||
| Low | 16 (1.9%) | 34 (3.0%) | 0.148 | High | >15 COCs | 1 | 8 |
| Failed | 14 (1.7%) | 27 (2.4%) | 0.339 | Low | <5 COCs | 6 | – |
| <4 MIIs | 7 | 5 | |||||
11.2.3 Compromised Embryo Development Following ICSI
An older study reported a significant decrease in blastocyst development in ICSI-derived embryos compared to IVF-derived embryos, including in a limited sibling oocyte study [18], although the blastocyst development rates were low compared to modern-day expectations [12]. This finding complemented an earlier observational study on supernumerary embryo development [19].
In an experimental study on mouse blastocysts generated following IVF, ICSI and ICSI-A (ICSI with artificial oocyte activation using ionophore), Bridges and colleagues [20] found that expression of 197 genes differed between ICSI and IVF, while in blastocysts derived by ICSI-A versus IVF, and ICSI-A versus ICSI, the expression of 132 and 65 genes differed respectively. Classification of the differentially expressed genes into biological pathways revealed consistency to known treatment-induced adverse consequences, including the regulation of metabolic pathways (including cholesterol and lipid metabolism/catabolism), and structural and neural developmental pathways.
11.2.4 Pregnancy and Live Birth Rates are not Higher when Non-Male Factor Cases are Treated using ICSI
A randomized controlled trial showed that ICSI did not offer an advantage over IVF [21]. implantation and clinical pregnancy (fetal heart) rates were higher in the IVF group (30% vs 22%, relative risk 1.35 (95% CI 1.04–1.76) and 33% vs 26%, relative risk 1.17 (95% CI 0.97–1.35), respectively), and mean associated laboratory time was significantly shorter with IVF than with ICSI (22.9 ± 12.1 SD minutes vs 74.0 ± 38.1 minutes; 95% CI for difference 45.6–56.6). These authors concluded that “ICSI offers no advantage over IVF in terms of clinical outcome in cases of non-male factor infertility. Our results support the current practice of reserving ICSI only for severe male-factor problems.”
A large registry-based analysis from Australia recently reported that ICSI did not increase the cumulative live birth rate in non-male factor infertility, and that the fertilization rate per oocyte retrieval was lower by ICSI than IVF, 56.2% versus 59.8%, p < 0.001 [22] concluding that “These data suggest that ICSI offers no advantage over conventional IVF in terms of live birth rate for couples with non-male factor infertility.”
Analysis of data reported to the US National Assisted Reproductive Technology Surveillance System during 1996–2012 revealed that in 317,996 cycles without male factor infertility, ICSI use was associated with lower rates of implantation (23.0% vs 25.2%; adjusted RR 0.93, 95%CI 0.91–0.95) and live birth (36.5% vs 39.2%; adjusted RR 0.95, 95%CI 0.93–0.97) compared to conventional IVF [23]. These authors concluded that “Compared with conventional IVF, ICSI use was not associated with improved post-fertilization reproductive outcomes, irrespective of male factor infertility diagnosis.”
Registry data from the USA reveals that in all female age groups younger than 43, cases without diagnosed male factor infertility achieved fewer live births following ICSI compared to IVF [24], confirming the ASRM official statement on ICSI for non-male factor infertility that “Routine use of ICSI for all oocytes does not appear to be justified in cases without male factor infertility or a history of prior fertilization failure” [15].
11.2.5 How Much ICSI is Needed?
From the US National Assisted Reproductive Technology Surveillance System data for 1996–2012, 65.1% of the 1,395,634 fresh IVF cycles used ICSI, although only 35.8% reported male factor infertility [23]. For cycles without male factor infertility, ICSI use increased from 15.4% in 1996 to 66.9% in 2012.
The latest report from the International Committee for Monitoring Assisted Reproductive Technologies (ICMART) [25] reported more than 455,000 ICSI treatment cycles were started in 2010 compared to only 220,000 IVF treatment cycles, with the proportion of ICSI cases ranging from 55% in Asia, through 65% in Europe and 73% in North America, to 86% in Latin America and almost 100% in the Middle East. In Europe, the prevalence of ICSI since 2007 has stabilized to between 65% and 70% [26].
Clearly the application of ICSI has little to do with the prevalence of male factor infertility throughout most of the world. Based on our experience, and that of many centres with which we have been associated over the past 25 years, even though 90% or more of couples might be defined as having a male factor according to current WHO semen analysis reference values, ICSI is typically necessary in no more than about 40% of cases due to a male factor that would be expected to impair sperm fertilizing ability [25,27]. The only exception to this would be for centres that specialize in treating men with spinal cord injuries or have a very specific focus on severe male factor subfertility.
11.2.6 Does using IVF/ICSI “Splits” have any Real Value?
A number of centres employ IVF/ICSI “splits,” where the available oocytes are assigned to two groups, one for conventional insemination and the other for ICSI. The rationale is that because these centres’ sperm assessments cannot identify those men whose spermatozoa have impaired fertilizing potential, this approach is “easier” than upgrading the diagnostic andrology testing to include sperm functional assessments.
We have even heard some clinicians say that doing “splits” is “easier on the lab,” a position that is hard to understand when the practice greatly increases the workload by requiring both IVF and ICSI forms of insemination in each such cycle, with the ICSI component taking far more time than a simple IVF insemination [21].
However, with “splits” it is not uncommon for the “better looking” cumulus-corona-oocyte complexes to be assigned to ICSI, rather than following a proper randomization – creating another self-fulfilling prophesy that (in these cases) ICSI achieves a higher fertilization rate.
11.2.7 Is a “Poor” Semen Analysis Justification for using ICSI?
The limited prognostic value of descriptive semen analysis characteristics has long been known, rendering simple assessments of sperm concentration, motility and even normal sperm morphology (even by properly trained, expert semen analysis technologists) of limited value in defining sperm fertilizing ability. Indeed, the Vienna Consensus recommended that any treatment selection decisions be based only on parameters derived from “trial wash” preparations and not on semen analysis characteristics [12]. Not even sperm morphology was considered sufficiently robust because the current visual evaluation of 200 or 400 spermatozoa used in the vast majority of laboratories to assess “percent normal forms” has such a large uncertainty of measurement that it cannot be considered a reliable predictor for IVF success/failure for individual men.
Moreover, there is no direct equivalence between morphological normality of a spermatozoon and its ability to fertilize an oocyte – or the converse. Many morphologically abnormal spermatozoa fertilize oocytes, and we have routinely employed IVF in men with less than 4% normal forms and achieved normal IVF fertilization rates (albeit so long as the Teratozoospermia Index (TZI) is below the threshold value of 1.80 [17].
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
