Regular (ICSI) versus ultra‐high magnification (IMSI) sperm selection for assisted reproduction
Abstract
Background
Subfertility is a condition found in up to 15% of couples of reproductive age. Gamete micromanipulation, such as intracytoplasmic sperm injection (ICSI), is very useful for treating couples with compromised sperm parameters. An alternative method of sperm selection has been described; the spermatozoa are selected under high magnification (over 6000x) and used for ICSI. This technique, named intracytoplasmic morphologically selected sperm injection (IMSI), has a theoretical potential to improve reproductive outcomes among couples undergoing assisted reproduction techniques (ART). However, our previous version of this Cochrane Review was unable to find evidence that supported this possible beneficial effect. This is an update of Teixeira 2013.
Objectives
To identify, appraise, and summarise the available evidence regarding efficacy and safety of IMSI compared to ICSI in couples undergoing ART.
Search methods
We searched for randomised controlled trials (RCTs) in these electronic databases: the Cochrane Gynaecology and Fertility Group Specialised Register, CENTRAL, MEDLINE, Embase, PsycINFO, CINAHL, LILACS, and in these trial registers: ClinicalTrials.gov and the World Health Organization International Clinical Trials Registry Platform. We also handsearched the reference lists of included studies and similar reviews. We performed the last electronic search on 18 November 2019.
Selection criteria
We only considered RCTs that compared ICSI and IMSI; we did not include quasi‐randomised trials. We considered studies that permitted the inclusion of the same participant more than once (cross‐over or per cycle trials) only if data regarding the first treatment of each participant were available.
Data collection and analysis
Two review authors independently performed study selection, data extraction, and assessment of the risk of bias and quality of the evidence; we solved disagreements by consulting a third review author. We corresponded with study investigators to resolve any queries, as required.
Main results
The updated search retrieved 535 records; we included 13 parallel‐designed RCTs comparing IMSI and ICSI (four studies were added since the previous version), comprising 2775 couples (IMSI = 1256; ICSI = 1519).
We are uncertain if IMSI improves live birth rates (risk ratio (RR) 1.11, 95% confidence interval (CI) 0.89 to 1.39; 5 studies, 929 couples; I² = 1%), miscarriage rates per couple (RR 1.07, 95% CI 0.78 to 1.48; 10 studies, 2297 couples; I² = 0%, very‐low quality evidence), and miscarriage rate per pregnancy (RR 0.90, 95% CI 0.68 to 1.20; 10 studies, 783 couples; I² = 0%, very‐low quality evidence). We are uncertain if IMSI improves clinical pregnancy rates (RR 1.23, 95% CI 1.11 to 1.37; 13 studies, 2775 couples; I² = 47%, very‐low quality evidence). None of the included studies reported congenital abnormalities. We judged the evidence for all outcomes to be of very low‐quality. We downgraded the quality of the evidence due to limitations of the included studies (risk of bias), inconsistency of results, and a strong indication of publication bias.
Authors' conclusions
The current evidence from randomised controlled trials does not support or refute the clinical use of intracytoplasmic sperm injection (intracytoplasmic morphologically selected sperm injection (IMSI). We are very uncertain of the chances of having a live birth and of the risk of having a miscarriage. We found very low‐quality evidence that IMSI may increase chances of a clinical pregnancy, which means that we are still very uncertain about any real difference.
We did not find any trials reporting on the risk of congenital abnormalities. Well‐designed and sufficiently powered trials are still required.
PICOs
Plain language summary
Regular (ICSI) versus ultra‐high magnification (IMSI) sperm selection for assisted reproduction
Background: Sperm micromanipulation, such as intracytoplasmic sperm injection (ICSI), is very useful for treating couples in which the male partner has reduced sperm concentration, motility, or both. In the past decade, a different approach for sperm selection has been described, which analyses sperm under ultra‐high powered magnification (6000x). Initial studies showed that intracytoplasmic morphologically selected sperm injection (IMSI), using sperm selected under high magnification, was associated with higher pregnancy rates than those selected with conventional ICSI in couples with repeated implantation failures. However, the evidence from our previous Cochrane Review was uncertain of the real beneficial effects of this intervention.
Search date: We updated our search of the medical literature in November 2019, looking for studies that evaluated the effectiveness and safety of IMSI (using 6000x magnification) compared to conventional ICSI (using 200x to 400x magnification) procedures.
Study characteristics: We found 13 randomised controlled trials (four more than in the previous version), evaluating 2775 couples, that compared regular ICSI with IMSI for assisted reproduction. These studies were funded by fertility centres and universities.
Key results: Based on the very low‐quality evidence that we found, we are uncertain of the benefit of IMSI over ICSI. The chance of having a live birth with IMSI was between 20% and 32%, compared to 25% with ICSI. For women with a 7% risk of miscarriage with regular ICSI, the risk with IMSI was between 5% and 10%. The clinical pregnancy rate with IMSI was between 35% and 44%, compared with 32% with ICSI.
Quality of the evidence: We downgraded the quality of the evidence because of limitations in the included studies (risk of bias), inconsistency of the observed effect across studies, and high risk of publication bias. There was no evidence concerning congenital abnormalities. We conclude that the current evidence is very limited for suggesting using IMSI instead of ICSI in clinical practice.
Authors' conclusions
Summary of findings
| Regular (ICSI) compared with ultra‐high magnification (IMSI) for assisted reproduction | |||||
| Patient or population: couples undergoing assisted reproduction treatment Setting: fertility clinics Intervention: sperm selection under ultra‐high magnification (IMSI) Comparison: sperm selection under regular magnification (ICSI) | |||||
| Outcomes | Illustrative comparative risks* (95% CI) | Relative effect | No of Participants | Quality of the evidence | |
| Assumed risk | Corresponding risk | ||||
| ICSI | IMSI | ||||
| Live birth per allocated couple | 243 per 1000 | 269 per 1000 (216 to 337) | RR 1.11 (0.89 to 1.39) | 929 | ⊕⊝⊝⊝ |
| Miscarriage per allocated couple | 70 per 1000 | 75 per 1000 (54 to 103) | RR 1.07 (0.78 to 1.48) | 2297 (10 studies) | ⊕⊝⊝⊝ |
| Miscarriage per clinical pregnancy | 230 per 1000 | 207 per 1000 (157 to 276) | RR 0.90 (0.68 to 1.20) | 783 | ⊕⊝⊝⊝ |
| Clinical pregnancy per allocated couple | 320 per 1000 | 394 per 1000 (355 to 438) | RR 1.23 (1.11 to 1.37) | 2775 | ⊕⊝⊝⊝ |
| Congenital abnormalities per live birth | No studies reported on this outcome | ||||
| The median control group risk across studies was used as the basis for the assumed risk. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | |||||
| GRADE Working Group grades of evidence High. We are very confident that the true effect lies close to that of the observed in this review Moderate. We are moderately confident in the effect estimate; although the true effect is likely to be close to the observed in this review, there is a possibility of being substantially different Low. Our confidence in the effect estimate is limited, since the true effect may be substantially different from that observed in this review Very low. We have very little confidence in the effect estimate, since the true effect is likely to be substantially different from that observed in this review | |||||
| aThe quality of the evidence was downgraded one level because of limitations of the included studies – all studies had unclear risk of selection bias | |||||
Background
Description of the condition
Subfertility is a condition found in up to 15% of couples of reproductive age; until the late 1970s, there were few treatment options for those couples. Compromised semen parameters account for 20% to 30% of infertility cases; at least 30 million men worldwide are considered subfertile (Agarwal 2015). Since the first successful in vitro fertilisation (IVF) was described, the efficacy of subfertility treatment has greatly improved. However, for those men, IVF was still insufficient. During the 1980s, other assisted reproductive technology (ART) techniques were developed, which focused on gamete micromanipulation; yet, for all these techniques, spermatozoa had to be progressively motile and have the potential for an acrosome reaction, leaving infertility due to severe male factors inadequately treated. After the introduction of the intracytoplasmic sperm injection (ICSI), couples in which the men had severe male factor infertility could now achieve pregnancy (Palermo 1992).
Description of the intervention
For ICSI, after sperm preparation, an optical magnification of 200x to 400x is used to examine the sample. The best 'normal looking' motile spermatozoa are selected, based on their major morphology, and then injected into oocytes retrieved after ovarian stimulation. However, live birth rates still remain low, which may justify the search for novel interventions to improve the results (Gleicher 2019). In the early 2000s, an alternative approach to sperm selection was described (Bartoov 2002). This technique requires the analysis of minor morphological criteria using ultra‐high magnification (≥ 6000x) microscopy. When using this technique, the motile sperm fraction is examined, based on six subcellular organelles: acrosome, postacrosomal lamina, neck, mitochondria, tail, and nucleus. When the ICSI used high magnification to select the sperm, it was named intracytoplasmic morphologically selected sperm injection (IMSI (Bartoov 2003)).
How the intervention might work
With high magnification, some organelle malformations that are not detectable using standard magnification may be seen, and it is thought that sperm selection based on these small details will improve reproductive outcomes (Berkovitz 2006).
Why it is important to do this review
Initial reports had shown that IMSI was associated with higher pregnancy rates in couples with repeated implantation failures (Bartoov 2002; Bartoov 2003). High magnification became available as a new technique, adding costs to the treatment, without any proven evidence of benefit or safety. Our previous Cochrane Review evaluating this comparison concluded that there was no evidence of benefit for live birth or miscarriage (Teixeira 2013). However, the uncertainty of our previous findings, and the fact that IMSI continues to be used in clinical practice justify the need for an updated review.
Objectives
To identify, appraise, and summarise the available evidence regarding efficacy and safety of intracytoplasmic morphologically selected sperm injection (IMSI) compared to intracytoplasmic sperm injection (ICSI) in couples undergoing assisted reproductive technology (ART).
Methods
Criteria for considering studies for this review
Types of studies
We considered only truly randomised controlled trials (RCTs) to be eligible; we did not include quasi or pseudo‐randomised trials. We included cross‐over trials only if data regarding the first treatment of each couple were available.
Types of participants
Couples undergoing assisted reproductive technology (ART).
Types of interventions
Trials comparing intracytoplasmic morphologically selected sperm injection (IMSI), using high magnification (≥ 6000x), versus intracytoplasmic sperm injection (ICSI), using regular magnification (200x to 400x), for sperm selection.
Types of outcome measures
Primary outcomes
-
Live birth. However, we used ongoing pregnancy, defined as a clinical pregnancy that is less likely to result in a miscarriage (i.e. those with fetal heartbeat beyond 10 to 16 weeks of gestational age) as a surrogate in case live birth was not reported.
-
Miscarriage per allocated couple and per pregnancy
Secondary outcomes
-
Clinical pregnancy per allocated couple
-
Congenital abnormality per live birth
Although implantation rates were important outcomes for this review, we did not include them in the quantitative meta‐analysis, because of different denominators (transferred embryos). However, we reported them in the 'Characteristics of included studies' tables.
Search methods for identification of studies
We searched for all published and unpublished RCTs that compared IMSI to ICSI, in consultation with the Gynaecology and Fertility Group (CGF) Information Specialist. We did not limit searches by language or publication status.
Electronic searches
We searched the following electronic databases for relevant trials:
-
The CGF Specialised Register of Controlled Trials, PROCITE (searched 18 November 2019; Appendix 1);
-
CENTRAL; Ovid (searched, Issue October 2019, 18 November 2019; Appendix 2);
-
MEDLINE Ovid (1946 to 18 November 2019; Appendix 3);
-
Embase Ovid (1980 to 18 November 2019; Appendix 4);
-
PsycINFO Ovid (1806 to 18 November 2019; Appendix 5);
-
CINAHL EBSCO (Cumulative Index to Nursing and Allied Health Literature; 1961 to 18 November 2019; Appendix 6);
-
LILACS Web (Latin American and Caribbean Health Science Information database; 18 November 2019; Appendix 7)
We combined the MEDLINE search with the Cochrane highly sensitive search strategy for identifying randomised trials (Lefebvre 2011). We combined the Embase, PsycINFO, and CINAHL searches with trial filters developed by the Scottish Intercollegiate Guidelines Network (SIGN).
We also searched for relevant trial registrations in:
-
ClinicalTrials.gov (clinicaltrials.gov; searched 18 Noveber 2019; Appendix 8);
-
World Health Organization (WHO) International Clinical Trials Registry Platform search portal (apps.who.int/trialsearch/Default.aspx; searched 18 November 2019; Appendix 8).
Searching other resources
We handsearched reference lists of relevant trials and systematic reviews retrieved by the electronic searches, and contacted experts in the field to obtain additional trials. We also searched conference abstracts that were not covered in the CGF Specialised Register.
Data collection and analysis
Selection of studies
After DMT and AHM independently completed the initial screen of titles and abstracts retrieved by the search, we retrieved the full texts of all potentially eligible studies. Two review authors (DMT and AHM) independently examined these full‐text articles for compliance with the inclusion criteria, and select eligible studies. We corresponded with study investigators as required, to clarify study eligibility. We resolved disagreements by discussion. If any reports required translation, we described the process used for data collection. We documented the selection process with a PRISMA flow chart.
Data extraction and management
Two review authors (DMT and AHM) independently extracted data from eligible studies using a data extraction form designed and pilot‐tested by the authors. Any disagreements were resolved by discussion with all authors. Data extracted included study characteristics and outcome data. Where studies had multiple publications, the authors collated multiple reports of the same under a single study ID with multiple references.
We corresponded with study investigators for further data on methods or results, as required.
Assessment of risk of bias in included studies
Two review authors (DMT and AHM) independently assessed risk of bias in the included studies using the Cochrane 'Risk of bias' assessment tool to assess: selection (random sequence generation and allocation concealment); performance (blinding of participants and personnel); detection (blinding of outcome assessors); attrition (incomplete outcome data); reporting (selective reporting); and other bias (Higgins 2011). We assigned judgements as recommended in the Cochrane Handbook for Systematic Reviews of Interventions, Section 8.5 (Higgins 2011). We resolved disagreements through discussion. We described all judgements fully and present the conclusions in the 'Risk of Bias' table, which we incorporated into the interpretation of review findings by means of sensitivity analyses.
With respect to within‐trial selective reporting, where identified studies failed to report the primary outcome of live birth, but did report interim outcomes, such as pregnancy, we assessed whether the interim values were similar to those reported in studies that also reported live birth.
Measures of treatment effect
For dichotomous data (e.g. live birth rates), we used the numbers of events in the control and intervention groups of each study to calculate the Mantel‐Haenszel risk ratio (RR). We preferred to use RR, because odds ratio (OR) is harder to understand and apply in practice. Misinterpretation of the OR, as if it is the same as the RR, will tend to overestimate the intervention effect, especially when events are common, and there is concern that this occurs quite frequently in published reports of individual studies and systematic reviews (Higgins 2011). However, had we observed a zero event count or prevalence less than 1%, we would have used the Peto fixed‐effect OR, because this method is found to be the least biased and most powerful, providing the best confidence interval (CI) coverage in these situations (Higgins 2011); the OR value in such situations is also very similar to RR, avoiding misinterpretations. We calculated the 95% CI to determine the precision of the estimates. Whenever estimates were statistically significant, we also determined the number needed to treat for an additional beneficial outcome (NNTB) or an additional harmful outcome (NNTH).
Unit of analysis issues
The primary analysis was per couple randomised. Exceptions were miscarriage, where we also considered the number of clinical pregnancies as the denominator, and congenital abnormality, which we will analyse per live birth when it is measured. We counted the delivery of a multiple pregnancy (e.g. twins or triplets) as one live birth event.
Dealing with missing data
We analysed the data on an intention‐to‐treat basis, when possible (i.e. including all randomised couples in analysis, in the groups to which they were randomised). We attempted to obtain missing data from the original trials. When these were unobtainable, we did not use imputation of individual values. We assumed that live birth did not occurred in couples with cycle cancellation.
Assessment of heterogeneity
We considered whether the clinical and methodological characteristics of the included studies were sufficiently similar for meta‐analysis to provide a clinically meaningful summary. We assessed statistical heterogeneity by the measure of the I². An I² value greater than 50% indicated substantial heterogeneity (Higgins 2011).
Assessment of reporting biases
In view of the difficulty of detecting and correcting for publication and other reporting biases, we aimed to minimise their potential impact by ensuring we conducted a comprehensive search for eligible studies, and by being alert for duplication of data. We used a funnel plot to explore the possibility of small study effects (a tendency for estimates of the intervention effect to be more beneficial in smaller studies).
Data synthesis
Using a fixed‐effect model, we combined the data from sufficiently similar studies that compared IMSI versus ICSI. An increased risk of a particular outcome associated with IMSI, which may be beneficial (e.g. live birth) or detrimental (e.g. miscarriage), was displayed graphically in the meta‐analysis to the right of the centre line, and a decreased risk to the left of the centre line.
Subgroup analysis and investigation of heterogeneity
When we detected a substantial heterogeneity (I² > 50%), we explored possible explanations in subgroup analyses (e.g. differing populations), sensitivity analyses (e.g. differing risk of bias), or both. We took statistical heterogeneity into account when interpreting the results, especially when there was any variation in the direction of effect.
We planned to perform the following subgroup analyses.
-
Sperm quality: studies including only couples where the male partner had poor sperm quality versus partners with good or unselected sperm quality.
-
Sperm source: ejaculate versus surgical.
-
Previously unsuccessful embryo transfers: studies only including women with repeated previously unsuccessful embryo transfers versus studies including any women (irrespective of the number of previous attempts)
Sensitivity analysis
We had planned to perform sensitivity analyses for the outcome, live birth or ongoing pregnancy, to determine whether the conclusions were robust, given our arbitrary decisions made regarding the eligibility and analysis. This analysis included consideration of whether the review conclusions would have differed if:
-
Eligibility had been restricted to studies with low risk of bias (defined as low risk of selection bias and no high risk of bias in any other domain);
-
Only studies reporting live birth without any imputation had been included.
Overall quality of the body of evidence: 'Summary of findings' table
We prepared a 'Summary of findings' table using GRADEpro and Cochrane methods (GRADEpro GDT; Higgins 2011). This table evaluates the overall quality of the body of evidence for the main review outcomes (live birth, clinical pregnancy, miscarriage, and congenital abnormality) for the comparison IMSI versus ICSI. We assessed the quality of the evidence using GRADE criteria: risk of bias, consistency of effect, imprecision, indirectness, and publication bias). Two review authors, working independently, made the judgements about evidence quality (high, moderate, low, or very low); they resolved disagreements discussion. They justified, documented, and incorporated judgments into the reporting of results for each outcome. The quality of the evidence was interpreted as follows (Balshem 2011): high: we are very confident that the true effect lies close to that of the observed in this review; moderate: we are moderately confident in the effect estimate, although the true effect is likely to be close to the observed in this review, there is a possibility of being substantially different; low: our confidence in the effect estimate is limited, since the true effect may be substantially different from that observed in this review; very low: we have very little confidence in the effect estimate, since the true effect is likely to be substantially different from the observed in this review.
We extracted study data, formatted our comparisons in data tables, and prepared a 'Summary of findings' table before writing the results and conclusions of our review.
Results
Description of studies
See: Characteristics of included studies; Characteristics of excluded studies.
Results of the search
The updated search retrieved 535 records. We considered 27 to be potentially eligible, and examined them for eligibility. Combining results of the updated search and the previous review, 13 trials (in 17 records) met our inclusion criteria, and we excluded 15 studies (in 17 records). Five studies are awaiting classification. There is one ongoing study. The study flow diagram is shown in Figure 1.
Study flow diagram
Included studies
Study design and setting
We included 13 RCTs (2775 couples) in this review. All were single‐centre studies, conducted in private assisted reproduction centres or academic centres, from Italy (Antinori 2008; Marci 2013), Turkey (Balaban 2011), Brazil (Figueira 2011; Setti 2011; Setti 2012a; Setti 2012b), Slovenia (Knez 2011; Knez 2012), Tunisia (Mahmoud 2011), France (Leandri 2013), USA (Check 2013), and Iran (Mangoli 2019).
Participants
The studies included 1256 couples in the intervention groups (intracytoplasmic morphologically selected sperm injection (IMSI)) and 1519 couples in the control groups (intracytoplasmic sperm injection (ICSI)). Seven studies included only couples in which the male partner had poor sperm quality (Antinori 2008; Knez 2011; Knez 2012; Leandri 2013; Mahmoud 2011; Mangoli 2019; Setti 2011); three included women with advanced maternal age (Figueira 2011; Setti 2012a; Setti 2012b); one study included only women with repeated implantation failure (Check 2013); and one study included couples who underwent assisted reproductive technology (ART), without specifying further details (Balaban 2011). One trial excluded couples with female factor infertility (Antinori 2008), and three excluded women with polycystic ovaries syndrome (PCOS) or endometriosis (Knez 2011; Knez 2012; Setti 2012b).
Interventions
All studies compared IMSI versus ICSI .
Outcomes
-
Five studies reported live birth (Balaban 2011; Check 2013; Leandri 2013; Mangoli 2019; Marci 2013);
-
All thirteen studies reported clinical pregnancy.
-
Nine studies reported miscarriage (Antinori 2008; Check 2013; Figueira 2011; Knez 2011; Leandri 2013; Marci 2013; Setti 2011; Setti 2012a; Setti 2012b).
-
No studies reported congenital abnormalities.
Excluded studies
We excluded 15 studies from the review, for the following reasons:
-
ten were not RCTs: nine were observational studies (Apryshko 2010; Bartoov 2003; Berkovitz 2005; Berkovitz 2006; Cassuto 2011; Ghazali 2017; Hazout 2005; Oliveira 2011; Wilding 2011), and one was a retrospective study (Gatimel 2016);
-
three randomly allocated the oocytes, not the couples (Braga 2011; De Vos 2013; Mauri 2011);
-
one was pseudo‐randomised (La Sala 2015);
-
in one, we could not determine if the study was adequately randomised. In addition, the number of women or couples enrolled was higher than the number of cycles reported, and we found several inconsistencies in numbers reported and statistical analysis performed (Karabulut 2019).
Risk of bias in included studies
See table Characteristics of included studies, Figure 2, and Figure 3 for detailed information.
Risk of bias graph: review authors' judgements about each risk of bias item presented as percentages across all included studies.
'Risk of bias' summary: review authors' judgements about each risk of bias item for each included study
Allocation
Eight of the thirteen studies were at low risk of selection bias related to sequence generation, as they used computer randomisation or a random numbers table (Balaban 2011; Figueira 2011; Knez 2011; Leandri 2013; Mangoli 2019; Marci 2013; Setti 2011; Setti 2012b). The other five studies did not describe the method used, and were at unclear risk of this bias.
Two studies were at low risk of selection bias related to allocation concealment, as they used sealed, opaque envelopes prepared by research nurses (Antinori 2008; Knez 2011). The other eleven studies did not describe the method used for allocation concealment, and we classified them to be at unclear risk of bias.
Blinding
Four studies were at low risk of performance bias as the couples and personnel were blinded to the intervention performed (Antinori 2008; Knez 2011; Leandri 2013;Setti 2012b); the other nine studies did not report blinding and we judged them to be at unclear risk of bias in this domain.
One study was at high risk of detection bias, as primary study team was not blinded to the intervention performed (Setti 2012b). All other twelve studies did not report blinding of outcome assessment and we judged them to be at unclear risk of detection bias.
Incomplete outcome data
We considered all thirteen studies to be at low risk of bias in this domain, as they stated that all allocated women or couples were analysed.
Selective reporting
We considered all thirteen studies to be at low risk of selective reporting bias, because clinical pregnancy was reported by all included studies. Although eight studies did not report live birth, and none reported congenital malformations, we believe these studies might not have been designed to evaluate these outcomes, as they require longer follow‐up to be properly assessed.
Other potential sources of bias
We deemed eight studies to be at high risk of other bias: four because there were substantial differences for the mean number of oocytes retrieved, or embryos transferred between groups, or both (Figueira 2011; Setti 2011; Setti 2012a; Setti 2012b); one because reported miscarriage rates were quite different than the calculated rates (Marci 2013); two because different interventions were performed for the study and control groups (Knez 2011; Knez 2012); and one because the number of embryos transferred for each group was not reported, and demographic characteristics between groups were not compared (Mangoli 2019). We considered one study to be at unclear risk of bias, because there was insufficient information to compare the number of oocytes retrieved or embryos transferred, or both, per woman (Mahmoud 2011). We judged four studies to be at low risk of other potential sources of bias (Antinori 2008; Balaban 2011; Check 2013; Leandri 2013).
Effects of interventions
See: Summary of findings for the main comparison
1. IMSI versus ICSI for assisted reproduction
Primary outcomes
1.1 Live birth (effectiveness)
We are uncertain whether IMSI improves live birth (risk ratio (RR) 1.11, 95% confidence interval (CI) 0.89 to 1.39; 5 studies, 929 couples; I² = 1%; very low‐quality evidence; Analysis 1.1; Figure 4). This suggests that if the chance of having a live birth with ICSI is assumed to be 24%, the chance following IMSI, would be between 21% and 33%.
Forest plot of comparison: Ultra‐high (IMSI) versus regular magnification (ICSI), outcome: 1.1 Live birth per allocated couple.
There were no studies at low risk of selection bias, and therefore, we did not perform a sensitivity analysis.
1.2 Miscarriage per allocated couple (adverse events)
We are uncertain whether IMSI reduces miscarriage rate per couple (RR 1.07, 95% CI 0.78 to 1.48; 10 studies, 2297 couples; I² = 0%, very low‐quality evidence; Analysis 1.2). This suggests that if the chance of having a miscarriage with ICSI is assumed to be 7%, the chance following IMSI, would be between 5% and 10%. Sensitivity analysis restricting the eligibility to studies with low risk of bias would make the estimates even more uncertain and imprecise (RR 0.93, 95% CI 0.09 to 9.58; 1 study, 57 couples; I² = not applicable).
1.3 Miscarriage per clinical pregnancy (adverse events)
We are uncertain whether IMSI reduces miscarriage rate per clinical pregnancy (RR 0.90, 95% CI 0.68 to 1.20; 10 studies, 783 couples; I² = 0%, very low‐quality evidence; Analysis 1.3; Figure 5). This suggests that if the chance of having a miscarriage with ICSI is assumed to be 23%, the chance following IMSI, would be between 16% and 28%. Sensitivity analysis restricting the eligibility to studies with low risk of bias would make the estimates even more uncertain and imprecise (RR 0.30, 95% CI 0.04 to 2.06; 1 study, 8 couples; I² = not applicable).
Forest plot of comparison: Ultra‐high (IMSI) versus regular magnification (ICSI), outcome: 1.3 Miscarriage per clinical pregnancy
Secondary outcomes
1.4 Clinical pregnancy (effectiveness)
Due to the very low quality evidence, we are uncertain whether IMSI was associated with an increase in clinical pregnancy rates (RR 1.23, 95% CI 1.11 to 1.37; 13 studies, 2775 couples; I² = 47%; very low‐quality evidence; Analysis 1.4; Figure 6). This suggests that if the chance of having a clinical pregnancy with ICSI is assumed to be 32%, the chance following IMSI would be between 35% and 44%. The resulting number needed to treat for an additional beneficial outcome (NNTB) was 12 (95% CI 7 to 45).
Forest plot of comparison: Ultra‐high (IMSI) versus regular magnification (ICSI), outcome: 1.4 Clinical pregnancy per allocated couple
Subgroup analysis (separating the studies by those that included only couples with poor sperm quality and those that included couples with good or unselected sperm quality) did not reduce the observed heterogeneity (Analysis 2.2). Sensitivity analysis restricting the eligibility to studies with low risk of bias changed this estimate, making the evidence of benefit highly uncertain and imprecise (RR 3.08, 95% CI 0.82 to 11.59; 1 study, 57 couples; I² = not applicable). The asymmetric funnel plot suggested a small studies effect, and therefore, we strongly suspected publication bias (Figure 7).
Funnel plot of comparison: Ultra‐high (IMSI) versus regular magnification (ICSI), outcome: 1.4 Clinical pregnancy per allocated couple
1.5 Congenital abnormalities (adverse events)
None of the included studies reported congenital abnormalities.
Discussion
Summary of main results
High magnification led to inconclusive results for live birth or miscarriage and was associated with a small improvement in clinical pregnancy rates. However, all evidence was of very low‐quality, and thus, we have very little confidence in these results. The true effects are likely to be substantially different from those observed in this review. There is no available evidence on the impact on congenital abnormalities. See summary of findings Table for the main comparison for further details.
Overall completeness and applicability of evidence
The objectives of this review were addressed by the included studies. Six studies included only couples with poor sperm quality, and six included couples with good or unselected sperm quality; however, these subgroup analyses did not add information to the main analysis (Analysis 2.2). No study sorted the couples according to the sperm source ‐ ejaculate or surgical, and only one evaluated women with previously unsuccessful embryo transfers. However, the quality of the pooled evidence does not allow robust conclusions, and we are uncertain about the true effect of intracytoplasmic morphologically selected sperm injection (IMSI) on the studied reproductive outcomes. Because of this, the review findings still do not support the use of IMSI in clinical practice.
Quality of the evidence
We considered the evidence to be very low‐quality (see summary of findings Table for the main comparison). Issues, such as risk of bias in the included studies, inconsistency, and strong suspicion of publication bias contributed to our decision to downgrade the quality of the evidence.
For live birth, we deemed the evidence to be very low‐quality because we assessed all studies as unclear risk of selection bias. Evidence was downgraded another two levels due to serious imprecision (95% confidence interval (CI) was compatible with both appreciable harm and no effect; and small number of events).
For miscarriage, we deemed the evidence to be very low‐quality because we considered seven of the ten studies that reported this outcome at high risk of bias: five related to differences in the mean number of oocytes retrieved or embryos transferred between groups, one reported inconsistent numbers for miscarriage rates, and one was due to the impossibility of ensuring that groups were comparable. There was also very serious imprecision in the estimate: there were only 158 miscarriages across both groups, and the 95% CI was compatible with both appreciable harm and no effect.
There was very low‐quality evidence for clinical pregnancy because all studies were at risk of bias (see Assessment of risk of bias in included studies; Figure 3), and publication bias was strongly suspected; funnel plot analysis suggested a small studies effect (Figure 7). We also found substantial heterogeneity, which added inconsistency to our estimates.
There was no evidence from randomised controlled trials (RCT) on the effect of IMSI on congenital abnormalities.
Potential biases in the review process
We did not identify any potential bias in the review process. However, we acknowledge a potential risk of bias regarding the source of the included studies. Eight of the thirteen included studies were performed in private fertility centres; although not often declared, this might be interpreted as a possible conflict of interest, since researchers involved might be interested in proving the success of new techniques. In addition to all the other limitations of the studies that built this body of evidence, this is something to consider with even more caution when interpreting the findings and applying them to clinical practice and future research.
Agreements and disagreements with other studies or reviews
Over the past five years, we have observed a lack of scientific interest on the topic; only a few studies comparing IMSI and intracytoplasmic sperm injection (ICSI) were published; none of them supports the use of IMSI in clinical practice.
Three other reviews evaluated the effect of IMSI (Duran‐Retamal 2019; He 2018; Setti 2010). Setti 2010 included studies recovered from a single database (MEDLINE) and supplemented evidence from a single RCT with non‐RCT studies. They reported a significant improvement in clinical pregnancy rates and a decrease in the risk of miscarriage with IMSI, while we did not observe this effect. The findings of the two more recent reviews are similar to ours: inconclusive results for live birth and miscarriage rates with IMSI; these authors did not support the routine use of IMSI either (Duran‐Retamal 2019; He 2018).
The benefit of IMSI on live birth and miscarriage is only seen when data from observational studies are pooled for meta‐analysis. However, supplementing data from RCTs with non‐RCTs must be cautiously interpreted: it may represent an exchange of undesirable uncertainty for unacceptable error (Higgins 2011).
The updated results are also similar to our previous review (Teixeira 2013). We included four new studies, but the results are similar to our previous findings; we found inconclusive results for live birth and miscarriage, and we are still very uncertain about the beneficial effects of IMSI on clinical pregnancy.
Risk of bias graph: review authors' judgements about each risk of bias item presented as percentages across all included studies.
'Risk of bias' summary: review authors' judgements about each risk of bias item for each included study
Forest plot of comparison: Ultra‐high (IMSI) versus regular magnification (ICSI), outcome: 1.1 Live birth per allocated couple.
Forest plot of comparison: Ultra‐high (IMSI) versus regular magnification (ICSI), outcome: 1.3 Miscarriage per clinical pregnancy
Forest plot of comparison: Ultra‐high (IMSI) versus regular magnification (ICSI), outcome: 1.4 Clinical pregnancy per allocated couple
Funnel plot of comparison: Ultra‐high (IMSI) versus regular magnification (ICSI), outcome: 1.4 Clinical pregnancy per allocated couple
Comparison 1 Ultra‐high (IMSI) versus regular magnification (ICSI), Outcome 1 Live birth per allocated couple.
Comparison 1 Ultra‐high (IMSI) versus regular magnification (ICSI), Outcome 2 Miscarriage per allocated couple.
Comparison 1 Ultra‐high (IMSI) versus regular magnification (ICSI), Outcome 3 Miscarriage per clinical pregnancy.
Comparison 1 Ultra‐high (IMSI) versus regular magnification (ICSI), Outcome 4 Clinical pregnancy per allocated couple.
Comparison 2 IMSI versus ICSI: subgroup analysis by sperm quality, Outcome 1 Live birth per allocated couple.
Comparison 2 IMSI versus ICSI: subgroup analysis by sperm quality, Outcome 2 Clinical pregnancy per allocated couple.
| Regular (ICSI) compared with ultra‐high magnification (IMSI) for assisted reproduction | |||||
| Patient or population: couples undergoing assisted reproduction treatment Setting: fertility clinics Intervention: sperm selection under ultra‐high magnification (IMSI) Comparison: sperm selection under regular magnification (ICSI) | |||||
| Outcomes | Illustrative comparative risks* (95% CI) | Relative effect | No of Participants | Quality of the evidence | |
| Assumed risk | Corresponding risk | ||||
| ICSI | IMSI | ||||
| Live birth per allocated couple | 243 per 1000 | 269 per 1000 (216 to 337) | RR 1.11 (0.89 to 1.39) | 929 | ⊕⊝⊝⊝ |
| Miscarriage per allocated couple | 70 per 1000 | 75 per 1000 (54 to 103) | RR 1.07 (0.78 to 1.48) | 2297 (10 studies) | ⊕⊝⊝⊝ |
| Miscarriage per clinical pregnancy | 230 per 1000 | 207 per 1000 (157 to 276) | RR 0.90 (0.68 to 1.20) | 783 | ⊕⊝⊝⊝ |
| Clinical pregnancy per allocated couple | 320 per 1000 | 394 per 1000 (355 to 438) | RR 1.23 (1.11 to 1.37) | 2775 | ⊕⊝⊝⊝ |
| Congenital abnormalities per live birth | No studies reported on this outcome | ||||
| The median control group risk across studies was used as the basis for the assumed risk. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | |||||
| GRADE Working Group grades of evidence High. We are very confident that the true effect lies close to that of the observed in this review Moderate. We are moderately confident in the effect estimate; although the true effect is likely to be close to the observed in this review, there is a possibility of being substantially different Low. Our confidence in the effect estimate is limited, since the true effect may be substantially different from that observed in this review Very low. We have very little confidence in the effect estimate, since the true effect is likely to be substantially different from that observed in this review | |||||
| aThe quality of the evidence was downgraded one level because of limitations of the included studies – all studies had unclear risk of selection bias | |||||
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1 Live birth per allocated couple Show forest plot | 5 | 929 | Risk Ratio (M‐H, Fixed, 95% CI) | 1.11 [0.89, 1.39] |
| 2 Miscarriage per allocated couple Show forest plot | 10 | 2297 | Risk Ratio (M‐H, Fixed, 95% CI) | 1.10 [0.81, 1.51] |
| 3 Miscarriage per clinical pregnancy Show forest plot | 10 | 783 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.90 [0.68, 1.20] |
| 4 Clinical pregnancy per allocated couple Show forest plot | 13 | 2775 | Risk Ratio (M‐H, Fixed, 95% CI) | 1.23 [1.11, 1.37] |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1 Live birth per allocated couple Show forest plot | 5 | 929 | Risk Ratio (M‐H, Random, 95% CI) | 1.11 [0.89, 1.39] |
| 1.1 Only poor sperm quality | 2 | 405 | Risk Ratio (M‐H, Random, 95% CI) | 1.22 [0.61, 2.42] |
| 1.2 Good or unselected sperm quality | 3 | 524 | Risk Ratio (M‐H, Random, 95% CI) | 1.11 [0.82, 1.50] |
| 2 Clinical pregnancy per allocated couple Show forest plot | 13 | 2775 | Risk Ratio (M‐H, Random, 95% CI) | 1.27 [1.09, 1.47] |
| 2.1 Only poor sperm quality | 7 | 1905 | Risk Ratio (M‐H, Random, 95% CI) | 1.30 [1.04, 1.62] |
| 2.2 Good or unselected sperm quality | 6 | 870 | Risk Ratio (M‐H, Random, 95% CI) | 1.24 [0.99, 1.56] |
