Intervention Review

You have free access to this content

Regular (ICSI) versus ultra-high magnification (IMSI) sperm selection for assisted reproduction

  1. Danielle M Teixeira1,
  2. Mariana AP Barbosa1,
  3. Rui A Ferriani1,
  4. Paula A Navarro1,
  5. Nick Raine-Fenning2,
  6. Carolina O Nastri1,
  7. Wellington P Martins1,*

Editorial Group: Cochrane Gynaecology and Fertility Group

Published Online: 25 JUL 2013

DOI: 10.1002/14651858.CD010167.pub2


How to Cite

Teixeira DM, Barbosa MAP, Ferriani RA, Navarro PA, Raine-Fenning N, Nastri CO, Martins WP. Regular (ICSI) versus ultra-high magnification (IMSI) sperm selection for assisted reproduction. Cochrane Database of Systematic Reviews 2013, Issue 7. Art. No.: CD010167. DOI: 10.1002/14651858.CD010167.pub2.

Author Information

  1. 1

    University of Sao Paulo, Department of Obstetrics and Gynecology, Medical School of Ribeirao Preto, Ribeirao Preto, Sao Paulo, Brazil

  2. 2

    University of Nottingham, Division of Obstetrics and Gynaecology, School of Clinical Sciences, Nottingham, UK

*Wellington P Martins, Department of Obstetrics and Gynecology, Medical School of Ribeirao Preto, University of Sao Paulo, Hospital das Clínicas da FMRP-USP, 8 andar, Campus Universitário, Campus Universitario da USP, Ribeirao Preto, Sao Paulo, 14048-900, Brazil. wpmartins@gmail.com.

Publication History

  1. Publication Status: New
  2. Published Online: 25 JUL 2013

SEARCH

 
Characteristics of included studies [ordered by study ID]

MethodsRandomised controlled trial conducted in a private assisted reproduction centre (Italy). Period of enrolment not reported.


ParticipantsInclusion criteria: at least 2 previous diagnoses of severe oligoasthenospermia; 3 years of primary infertility; absence of female factor.


InterventionsCouples were randomised into 2 groups:

Intervention: oocytes injected with spermatozoon selected under ultra-high magnification (IMSI): on the basis of MSOME criteria, the examination and spermatozoa selection for IMSI procedure was performed in real time using an inverted light microscope equipped with high-power Nomarski optics, enhanced by digital imaging to achieve a magnification up to 6300x, and the Eppendorf Micromanipulation System (Transfer-Man NK2, Eppendorf, Germany). Only spermatozoa with normal head dimension (length 4.75 ± 0.28 μm; width 3.28 ± 0.20 μm) and shape, with no or maximum 1 vacuole (0.78 ± 0.18 μm) were microinjected; spermatozoa with abnormal head size were excluded (such spermatozoa were identified by superimposing a transparent celluloid form representing the correct spermatozoon size on the examined gametes). 2 spermatozoa for each oocyte were selected to be inseminated using the classical ICSI technique.

Control: oocytes injected with spermatozoon selected under regular magnification (ICSI): no further details.


OutcomesClinical pregnancy, miscarriage, and implantation rates.


NotesWe considered 2 publications to be related to the same study although the numbers of participants were not the same, because the only difference we observed was that the newer publication had evaluated more participants; the authors did not answer our e-mails to resolve these data queries.

Live birth and congenital abnormalities not reported.

Implantation rate: 23.0% (IMSI) vs. 16.6% (ICSI); P value not reported.


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Unclear riskNot reported.

Allocation concealment (selection bias)Low riskSealed, opaque envelopes prepared by a research nurse.

Blinding of participants and personnel (performance bias)
All outcomes
Low riskBlinding of participants and personnel was not considered as a potential source of bias.

Blinding of outcome assessment (detection bias)
All outcomes
Low riskBlinding of outcome assessment was not considered as a potential source of bias.

Incomplete outcome data (attrition bias)
All outcomes
Low riskNo loss to follow-up.

Selective reporting (reporting bias)Low riskNot suspected.

Other biasLow riskNone.


MethodsRandomised controlled trial conducted in a private assisted reproduction centre (Turkey) between February and July 2009.


ParticipantsEligibility criteria: unselected women undergoing assisted reproduction treatment.


InterventionsCouples were randomised into 2 groups:

Intervention: oocytes injected with spermatozoon selected under ultra-high magnification (IMSI). The procedure was performed in real time using an inverted microscope (Olympus IX-71; Japan) with actual digitally enhanced magnification, as determined by a 0.01 mm Olympus objective micrometer, at 6300x. Normal-shaped nuclei were defined as smooth, symmetric, having an oval configuration, with mean length and width limits of 4.75 ± 0.28 and 3.28 ± 0.20 μm, respectively, with a homogeneous nuclear chromatin mass with no regional nuclear disorders and containing no more than one small vacuole with a borderline diameter of 0.78 ± 0.18 μm. Mean time needed for IMSI = 21 minutes.

Control: oocytes injected with spermatozoon selected under regular magnification (ICSI): not specified; the mean time needed for regular ICSI = 14 minutes.


OutcomesDuration of ICSI procedure; 2-pronuclei fertilisation rate; embryos with 4 blastomeres on day 2 post fertilisation; embryos with 8 blastomeres on day 3 post fertilisation; grade 1 and 2 embryos on transfer day; clinical pregnancy; live birth; implantation; and multiple pregnancy rate.


NotesMiscarriage and congenital abnormalities not reported.

Implantation rate: 66/228 = 28.9% (IMSI) vs. 42/215 = 19.5% (ICSI); P value = 0.02 (however, study authors reported that P = NS [not significant]).


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Low riskComputer-generated randomisation list.

Allocation concealment (selection bias)Unclear riskNot reported.

Blinding of participants and personnel (performance bias)
All outcomes
Low riskBlinding of participants and personnel was not considered as a potential source of bias.

Blinding of outcome assessment (detection bias)
All outcomes
Low riskBlinding of outcome assessment was not considered as a potential source of bias.

Incomplete outcome data (attrition bias)
All outcomes
Low riskNo loss to follow-up.

Selective reporting (reporting bias)Low riskNot suspected.

Other biasLow riskNone.


MethodsRandomised controlled trial conducted in a private assisted reproduction centre (Brazil) between May and December 2009.


ParticipantsEligibility criteria: women undergoing assisted reproduction treatment in conjunction with pre-implantation genetic screening for aneuploidy, as a result of advanced maternal age; sperm concentration > 1,000,000/mL and sperm motility > 20%; at least 6 oocytes available on oocyte retrieval.


InterventionsCouples were randomised into 2 groups:

Intervention: oocytes injected with spermatozoon selected under ultra-high magnification (IMSI): sperm morphology selection was assessed using an inverted Nikon Diaphot microscope (Eclipse TE 300; Nikon, Tokyo) equipped with high-power DIC (Nomarski). The total calculated magnification was 6600x. The sperm cells exhibiting normally shaped nuclei (smooth, symmetric, and oval configuration) and normal nuclear chromatin content (if it contained no more than 1 vacuole, which occupies < 4% of the nuclear area) were selected for injection.

Control: oocytes injected with spermatozoon selected under regular magnification (ICSI): sperm morphology selection was assessed using an inverted Nikon Diaphot microscope (Eclipse TE 300; Nikon, Tokyo) with a Hoffmann modulation contrast system under 400x magnification.


OutcomesSperm nuclear morphology at high-magnification ICSI; incidence of aneuploidy in derived embryos; clinical pregnancy rate.


NotesAll embryos were submitted to pre-implantation genetic diagnosis and aneuploidy screening. On the morning of day 3 of embryo development, 1 cell per embryo was biopsied by laser zona drilling using a 1.48 mm infrared diode laser (Octax Laser Shot System, MTG, Bruckberg, Germany). After the biopsies, the embryos were returned to the culture medium. The removed blastomere nuclei were spread using 0.1 N HCl and 0.01% Tween 20 (Sigma, Dorset, UK). Briefly, the individual nuclei were placed on a slide in 1 drop of HCl-Tween spreading solution and observed until the cell had lysed. The slides were then air dried and dehydrated before FISH analysis was performed. All embryos were analysed for chromosomes X, Y, 13, 16, 18, 21, and 22. For the purpose of this study, the blastomeres were classified as normal when 2 sexual and 2 of each tested autosomal chromosomes were present. Blastomeres with 2 or more chromosomal numerical abnormalities were classified as chaotic. Embryos with abnormal findings in biopsied nuclei were not submitted for re-analysis. Embryo transfer was performed on day 4 using a soft catheter with transabdominal ultrasound guidance. Only the embryos found to be chromosomally normal were considered for embryo transfer, and a maximum of 3 embryos were transferred. The cycle was cancelled if normal embryos were absent after FISH.

Live birth and congenital abnormalities not reported.

Implantation rate: 55.6% (IMSI) vs. 40.9% (ICSI); P value = 0.59.

Study authors were contacted to clarify information about the 4 different included studies, from the same groups of authors (Figueira 2011; Setti 2011; Setti 2012a; Setti 2012b). All questions on methods of randomisation, patient overlapping, and data "per woman" were clarified.


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Low riskComputer-generated balanced table in sets of 10.

Allocation concealment (selection bias)Unclear riskNot reported.

Blinding of participants and personnel (performance bias)
All outcomes
Low riskBlinding of participants and personnel was not considered as a potential source of bias.

Blinding of outcome assessment (detection bias)
All outcomes
Low riskBlinding of outcome assessment was not considered as a potential source of bias.

Incomplete outcome data (attrition bias)
All outcomes
Low riskNo loss to follow-up.

Selective reporting (reporting bias)Low riskNot suspected.

Other biasHigh riskThe number of oocytes retrieved and embryos transferred were significantly different between groups (P value < 0.01). In the article, study authors reported that differences were not significant (P value = 0.20 and P value = 0.17, oocytes retrieved and embryos transferred, respectively).


MethodsRandomised controlled trial conducted in an academic setting (Slovenia) between October 2009 and June 2010.


ParticipantsEligibility criteria: all embryos arrested after prolonged 5-day embryo culture to the blastocyst stage in their previous conventional ICSI attempts; poor semen quality characterised by the incidence of teratozoospermia by less than 14% of morphologically normal sperm according to the Strict Kruger Criteria, oligozoospermia by a sperm concentration of < 20 million/mL and asthenozoospermia by < 50% of motile sperm according the WHO criteria; women without PCOS or endometriosis and age < 42 years.


InterventionsCouples were randomised into 2 groups:

Intervention: oocytes injected with spermatozoon selected under ultra-high magnification (IMSI): sperm were selected in dishes with a glass bottom (GWSt 1000; Will Co., Wells BV, Amsterdam, The Netherlands) and monitored under an inverted microscope with a heated stage equipped with DIC (Nikon ECLIPSE TE2000-S, Japan). Approximately 5 elongated droplets of SpermSlow medium (Origio, Denmark) were placed on the bottom of a glass dish to immobilise the sperm. A smaller droplet of prepared sperm was placed near each SpermSlow droplet. Then the connections were made between the sperm and the SpermSlow droplets for sperm to swim into the SpermSlow droplets and to bind to the HA. All droplets were covered with paraffin oil (Origio, Denmark). For observation under 6000x magnification, a droplet of immersion oil was inserted underneath the glass dish (under the SpermSlow droplet). One droplet of SpermSlow with bound sperm was monitored by the immersion objective, DIC, and Nikon Digital Sight DS-Ri1 camera. The single (mature) sperm that was bound to the HA and had the best morphology was chosen, aspirated in the microinjection pipette, scored in 3-dimensions, and evaluated according to the morphology and head vacuoles at 6000x magnification.

Control: oocytes injected with spermatozoon selected under regular magnification (ICSI): the sperm selection for microinjection was performed at a magnification of 200x to 400x. Sperm with severe head-shape defects clearly seen under the magnification (pin, amorphous, tapered, round, and multinucleated head) were excluded from microinjection into the oocyte.


OutcomesFertilisation, blastocyst, implantation, and pregnancy rates.


NotesLive birth and congenital abnormalities not reported.

Implantation rate: 6/35 = 17.1% (IMSI) vs. 3/44 = 6.8% (ICSI); P = 0.17.

Study author was contacted to clarify information about the 2 different included studies (Knez 2011 and Knez 2012). All questions on methods of randomisation, patient overlapping, and data "per woman" were clarified.


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Low riskComputer-generated random numbers (unrestricted randomisation list).

Allocation concealment (selection bias)Low riskSealed, opaque envelopes prepared by a research nurse.

Blinding of participants and personnel (performance bias)
All outcomes
Low riskBlinding of participants and personnel was not considered as a potential source of bias.

Blinding of outcome assessment (detection bias)
All outcomes
Low riskBlinding of outcome assessment was not considered as a potential source of bias.

Incomplete outcome data (attrition bias)
All outcomes
Low riskNo loss to follow-up.

Selective reporting (reporting bias)Low riskNot suspected.

Other biasUnclear riskStudy authors did not report the number of oocytes retrieved in the groups. The number of transferred embryos per participant was not significantly different between groups.


MethodsRandomised controlled trial conducted in an academic setting (Slovenia) between January and October 2011.


ParticipantsEligibility criteria: at least 6 mature oocytes available upon oocyte retrieval; isolated teratozoospermia, which was determined as having < 14% of morphologically normal spermatozoa according to the Kruger strict criteria, > 15 million spermatozoa per millilitre and at least 40% motile spermatozoa; women without PCOS or endometriosis.


InterventionsCouples were randomised into 2 groups:

Intervention: oocytes injected with spermatozoon selected under ultra-high magnification (IMSI): a single spermatozoon bound to the HA and with the best morphology was chosen, aspirated in the microinjection pipette, scored in 3-dimensions and evaluated according to the morphology and head vacuoles at 6000x magnification.

Control: oocytes injected with spermatozoon selected under regular magnification (ICSI): spermatozoa without severe head shape defects clearly seen under the microscope (pin, amorphous, tapered, round, and multinucleated head) were selected at magnification 200x to 400x.


OutcomesFertilisation, blastocyst, implantation, and pregnancy rates.


NotesLive birth, miscarriage, and congenital abnormalities not reported.

Implantation rate not reported.

Study author was contacted to clarify information about the 2 different included studies (Knez 2011; Knez 2012). All questions on methods of randomisation, patient overlapping, and data "per woman" were clarified.


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Unclear riskNot reported.

Allocation concealment (selection bias)Unclear riskNot reported.

Blinding of participants and personnel (performance bias)
All outcomes
Low riskBlinding of participants and personnel was not considered as a potential source of bias.

Blinding of outcome assessment (detection bias)
All outcomes
Low riskBlinding of outcome assessment was not considered as a potential source of bias.

Incomplete outcome data (attrition bias)
All outcomes
Low riskNo loss to follow-up.

Selective reporting (reporting bias)Low riskNot suspected.

Other biasUnclear riskStudy authors did not provide the number of transferred embryos per participant neither the SD for the number of oocytes retrieved (mean 11.0 with IMSI vs. 9.8 with ICSI).


MethodsRandomised controlled trial conducted in a private assisted reproduction centre (Tunisia) between April 2009 and November 2010.


ParticipantsEligibility criteria: oligoasthenozoospermia based on WHO references values. Teratozoospermia evaluated by the strict criteria of Kruger sperm morphology; healthy woman aged < 37 years.


InterventionsCouples were randomised into 2 groups:

Intervention: oocytes injected with spermatozoon selected under ultra-high magnification (IMSI): 6600x magnification, using Leica 6800 station.

Control: oocytes injected with spermatozoon selected under regular magnification (ICSI).


OutcomesFertilisation rate, percentage of good-quality embryos, and the rates of clinical pregnancy and implantation.


NotesLive birth, miscarriage, and congenital abnormalities not reported.

Implantation rate: 19.2% (IMSI) vs. 17.2% (ICSI); P value = not significant.

Study authors were not contacted because we were unable to obtain their contact details.


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Unclear riskNot reported.

Allocation concealment (selection bias)Unclear riskNot reported.

Blinding of participants and personnel (performance bias)
All outcomes
Low riskBlinding of participants and personnel was not considered as a potential source of bias.

Blinding of outcome assessment (detection bias)
All outcomes
Low riskBlinding of outcome assessment was not considered as a potential source of bias.

Incomplete outcome data (attrition bias)
All outcomes
Low riskNo loss to follow-up.

Selective reporting (reporting bias)Low riskNot suspected.

Other biasUnclear riskStudy authors did not report the number of oocytes retrieved and embryos transferred.


MethodsRandomised controlled trial conducted in a private assisted reproduction centre (Brazil). Period of enrolment not reported.


ParticipantsEligibility criteria: first IVF treatment; abnormal semen parameters according to WHO, except for azoospermia; use of fresh semen sample; absence of a known female factor infertility; and at least 6 oocytes available on retrieval.


InterventionsCouples were randomised into 2 groups:

Intervention: oocytes injected with spermatozoon selected under ultra-high magnification (IMSI): sperm selection was examined at high magnification with an inverted microscope (Eclipse TE 300; Nikon, Tokyo, Japan) equipped with high-power DIC optics (Nomarski). The total calculated magnification was 6600x.

Control: oocytes injected with spermatozoon selected under regular magnification (ICSI): sperm morphology selection was assessed using an inverted Nikon Diaphot microscope (Eclipse TE 300; Nikon, Tokyo) with a Hoffmann modulation contrast system under 400x magnification.


OutcomesClinical pregnancy, implantation rate, and miscarriage.


NotesLive birth and congenital abnormalities not reported.

Implantation rate: 158/664 = 23.8% (IMSI) vs. 28/156 = 25.4% (ICSI); P value = 0.60.

Study authors were contacted to clarify information about the 4 different included studies, from the same groups of authors (Figueira 2011; Setti 2011; Setti 2012a; Setti 2012b). All questions on methods of randomisation, patient overlapping, and data "per woman" were clarified.


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Low riskComputer-generated randomisation list.

Allocation concealment (selection bias)Unclear riskNot reported.

Blinding of participants and personnel (performance bias)
All outcomes
Low riskBlinding of participants and personnel was not considered as a potential source of bias.

Blinding of outcome assessment (detection bias)
All outcomes
Low riskBlinding of outcome assessment was not considered as a potential source of bias.

Incomplete outcome data (attrition bias)
All outcomes
Low riskNo loss to follow-up.

Selective reporting (reporting bias)Low riskNot suspected.

Other biasHigh riskThe number of oocytes retrieved and embryos transferred were significantly different between groups (P value < 0.01 for both). In the article, study authors reported that differences were non-significant (P value = 0.29 for oocytes retrieved and P = 0.27 for embryos transferred).


MethodsRandomised controlled trial conducted in a private assisted reproduction centre (Brazil) between May 2009 and December 2010.


ParticipantsEligibility criteria: women undergoing assisted reproduction treatment in conjunction with pre-implantation genetic screening for aneuploidy, as a result of advanced maternal age; no severe spermatogenic alteration; at least 6 oocytes available on oocyte retrieval.


InterventionsCouples were randomised into 2 groups:

Intervention: oocytes injected with spermatozoon selected under ultra-high magnification (IMSI): sperm selection was examined at high magnification using a similar inverted microscope equipped with high-power DIC optics (Nomarski). The total calculated magnification was 6600x. The sperm cells exhibiting normally shaped nuclei (smooth, symmetric, and oval configuration) and normal nuclear chromatin content (if it contained no more than 1 vacuole, which occupied < 4% of the nuclear area) were selected for injection.

Control: oocytes injected with spermatozoon selected under regular magnification (ICSI): sperm morphology selection was assessed using an inverted Nikon Diaphot microscope (Eclipse TE 300; Nikon, Tokyo, Japan) with a Hoffmann modulation contrast system under 400x magnification.


OutcomesClinical pregnancy, implantation rate, miscarriage, and gender incidence.


NotesAll embryos were submitted to pre-implantation genetic diagnosis and aneuploidy screening. On the morning of day 3 of embryo development, 1 cell per embryo was biopsied by laser zona drilling using a 1.48 mm infrared diode laser (Octax Laser Shot System, MTG, Bruckberg, Germany). Following the biopsies, the embryos were returned to the culture medium. The removed blastomere nuclei were spread using 0.1 mol/L HCl and 0.01% Tween 20 (Sigma, Dorset, UK). Briefly, the blastomeres were placed on a slide in a drop of HCl-Tween spreading solution and observed until the cell had lysed. The slides were then air-dried and dehydrated before FISH analysis was performed. All embryos were analysed for chromosomes X, Y, 13, 16, 18, 21, and 22. For the purpose of this study, the blastomeres were classified as normal when 2 sexual and 2 of each tested autosomal chromosomes were present. Embryo transfer was performed on day 5 using a soft catheter with transabdominal ultrasound guidance. Only the embryos found to be chromosomally normal were considered for embryo transfer, and up to a maximum of 3 embryos were transferred.

Live birth and congenital abnormalities not reported.

Implantation rate: 46.1% (IMSI) vs. 41.6% (ICSI); P value = 0.59.

Study authors were contacted to clarify information about the 4 different included studies, from the same groups of authors (Figueira 2011; Setti 2011; Setti 2012a; Setti 2012b). All questions on methods of randomisation, patient overlapping, and data "per woman" were clarified.


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Unclear riskNot reported.

Allocation concealment (selection bias)Unclear riskNot reported.

Blinding of participants and personnel (performance bias)
All outcomes
Low riskBlinding of participants and personnel was not considered as a potential source of bias.

Blinding of outcome assessment (detection bias)
All outcomes
Low riskBlinding of outcome assessment was not considered as a potential source of bias.

Incomplete outcome data (attrition bias)
All outcomes
Low riskNo loss to follow-up.

Selective reporting (reporting bias)Low riskNot suspected.

Other biasHigh riskThe number of oocytes retrieved and embryos transferred were significantly different between groups (P value = 0.01). In the article, study authors reported these differences as not significant.


MethodsProspective randomised clinical trial performed in a private fertility centre (Brazil). Period of enrolment not reported.


ParticipantsEligibility criteria: women of good physical and mental health; undergoing ICSI as a result of advanced maternal age (≥ 37 years old); with regular menstrual cycles of 25-35 days; normal basal FSH and LH concentrations; body mass index < 30 kg/m2; presence of both ovaries and intact uterus; absence of PCOS, endometriosis and gynaecological/medical disorders; and a negative result in a screening for sexually transmitted diseases. All male partners were normozoospermic, according to the WHO reference values (2010). No woman had received any hormone therapy for at least 60 days preceding the study.


InterventionsCouples were randomised into 2 groups:

Intervention: sperm selection in the IMSI group was analysed under high magnification using an inverted Nikon Diaphot microscope equipped with high-power DIC optics. The total calculated magnification was 6600x. An aliquot of the sperm cell suspension was transferred to a microdroplet of modified human tubal fluid medium containing 8% polyvinyl pyrrolidone (Irvine Scientific, Santa Ana, CA, USA) in a sterile glass dish (FluoroDish; World Precision Instrument, Sarasota, FL, USA). The dish was placed on a microscope stage above an Uplan Apo 100 oil/1.35 objective lens previously covered by a droplet of immersion oil. The sperm cells that were selected for injection exhibited normally shaped nuclei (smooth, symmetric and oval configuration) and normal nuclear chromatin content (if it contained no more than 1 vacuole that occupied < 4% of the nuclear area).

Control: sperm selection in the ICSI group was analysed under 400x magnification using an inverted Nikon Eclipse TE 300 microscope.


OutcomesFertilisation rate, high-quality embryo rate on day 3, blastocyst formation rate, number of transferred embryos, implantation rate, miscarriage rate and pregnancy rate.


NotesLive birth and congenital abnormalities not reported.

Implantation rate: 38.3% (IMSI) vs. 12.1% (ICSI). P value = 0.03.

This article has been withdrawn. Authors were contacted for more information regarding the reasons why this happened. Authors had also been contacted to clarify information about the 4 different included studies, from the same groups of authors (Figueira 2011; Setti 2011; Setti 2012a; Setti 2012b). All questions on methods of randomisation, patient overlapping and data "per woman" were clarified.


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Low riskComputer-generated randomisation table.

Allocation concealment (selection bias)Unclear riskNot reported.

Blinding of participants and personnel (performance bias)
All outcomes
Low riskBlinding of participants and personnel was not considered as a potential source of bias.

Blinding of outcome assessment (detection bias)
All outcomes
Low riskBlinding of outcome assessment was not considered as a potential source of bias.

Incomplete outcome data (attrition bias)
All outcomes
Low riskNo loss of follow-up.

Selective reporting (reporting bias)Low riskNot suspected.

Other biasHigh riskAlthough there was no significant difference in the number of oocytes retrieved, and despite the total number of embryos being higher in the ICSI group, more embryos per woman were transferred in the IMSI group.

This article has been withdrawn.

 
Characteristics of excluded studies [ordered by study ID]

StudyReason for exclusion

Apryshko 2010Not an RCT (observational study).

Bartoov 2003Not an RCT (observational study).

Berkovitz 2005Not an RCT (observational study).

Berkovitz 2006Not an RCT (observational study).

Braga 2011Oocytes (not participants) were allocated.

Cassuto 2011Not an RCT (observational study).

De Vos 2013Oocytes (not participants) were randomly allocated.

Hazout 2005Not an RCT (observational study).

Mauri 2011Oocytes (not participants) were randomly allocated.

Oliveira 2011Not an RCT (observational study).

Wilding 2011Not an RCT (observational study).

 
Characteristics of studies awaiting assessment [ordered by study ID]

MethodsProspective randomised clinical trial performed in a fertility centre (USA). Period of enrolment not reported.

ParticipantsEligibility criteria: women aged ≤ 39, undergoing ART, with failure to have a successful conception after 3 consecutive embryo transfers and whose male partner had a DNA fragmentation index > 30%.

InterventionsCouples were randomised into 2 groups:

Intervention: sperm selection in the ICSI group was analysed under high magnification.

Control: sperm selection in the ICSI group was analysed under normal magnification.

OutcomesLive birth and pregnancy rate.

NotesAuthors were contacted for further information regarding period of enrolment, eligibility criteria and results.


MethodsRandomised controlled trial conducted in 9 different assisted reproduction centres (France). Period of enrolment not reported.

ParticipantsEligibility criteria: couples undergoing ICSI due to male infertility, with < 1 million motile spermatozoa recovered after gradient preparation and at least 3 million sperm cells in the ejaculate. Couples in which the female age was > 38 years or FSH level was > 9.0 mL/mlU were excluded from the study.

InterventionsCouples were randomised into 2 groups:

Intervention: sperm selection in the ICSI group was analysed under 6000x magnification.

Control: sperm selection in the ICSI group was analysed under 400x magnification.

OutcomesNumber of deliveries and implantation rate.

NotesThe study was registered at ClinicalTrials.gov and has already been completed. Authors were contacted for further information, since study has not been published yet.


MethodsProspective randomised clinical trial performed in a private fertility centre (Brazil). Period of enrolment not reported.

ParticipantsEligibility criteria: women undergoing ICSI as a result of unexplained fertility.

InterventionsCouples were randomised into 2 groups:

Intervention: sperm selection in the ICSI group was analysed under 6000x magnification.

Control: sperm selection in the ICSI group was analysed under 400x magnification.

OutcomesFertilisation rate, high-quality embryo rate, number of transferred embryos, implantation rate and pregnancy rate.

NotesMiscarriage, live birth and congenital abnormalities not reported.

Implantation rate: 39.6% (IMSI) vs. 19.4% (ICSI). P value = 0.019.

Authors were contacted for further information regarding period of enrolment and eligibility criteria.

 
Comparison 1. Ultra high (IMSI) versus regular magnification (ICSI)

Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size

 1 Live birth per allocated couple1168Risk Ratio (M-H, Fixed, 95% CI)1.14 [0.79, 1.64]

 2 Clinical pregnancy per allocated couple92014Risk Ratio (M-H, Random, 95% CI)1.29 [1.07, 1.56]

 3 Miscarriage per clinical pregnancy6552Risk Ratio (M-H, Fixed, 95% CI)0.82 [0.59, 1.14]

 
Comparison 2. IMSI versus ICSI: subgroup analysis by sperm quality

Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size

 1 Clinical pregnancy92014Risk Ratio (M-H, Random, 95% CI)1.29 [1.07, 1.56]

   1.1 Only poor sperm quality
51500Risk Ratio (M-H, Random, 95% CI)1.29 [0.98, 1.70]

   1.2 Good or unselected sperm quality
4514Risk Ratio (M-H, Random, 95% CI)1.33 [0.97, 1.82]

 
Summary of findings for the main comparison.

Regular (ICSI) compared with ultra-high magnification (IMSI) for assisted reproduction

Patient or population: couples undergoing assisted reproduction techniques

Intervention: sperm selection under ultra-high magnification (IMSI)

Comparison: sperm selection under regular magnification (ICSI)

OutcomesIllustrative comparative risks* (95% CI)Relative effect
(95% CI)
No of Participants
(studies)
Quality of the evidence
(GRADE)

Assumed riskCorresponding risk

ICSIIMSINNTB

Live birth per allocated couple38 per 10044 per 100
(30 to 63)
-RR 1.14 (0.79 to 1.64)168
(1 study)
⊕⊕⊝⊝
low1

Clinical pregnancy per allocated couple33 per 10043 per 100
(36 to 52)
10

(5 to 33)
RR 1.29 (1.06 to 1.55)2014
(9 studies)
⊕⊝⊝⊝
very low2

Miscarriage per clinical pregnancy22 per 10018 per 100
(13 to 25)
-RR 0.82 (0.59 to 1.14)552
(6 studies)
⊕⊝⊝⊝
very low3

Congenital abnormalities per clinical pregnancyNo evidence.


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).
CI: confidence interval; ICSI: intracytoplasmic sperm injection; IMSI: intracytoplasmic morphologically selected sperm injection; NNTB: number needed to treat for an additional beneficial outcome; RR: risk ratio.

GRADE Working Group grades of evidence
High quality: Further research is very unlikely to change our confidence in the estimate of effect.
Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate.
Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate.
Very low quality: We are very uncertain about the estimate.

 1. The quality of the evidence was downgraded two levels due to very serious imprecision.
2. The quality of the evidence was downgraded one level due to the high risk of bias in the included studies; another level due to inconsistency across studies; and one further level because publication bias was strongly suspected.
3. The quality of the evidence was downgraded two levels due to very serious imprecision; and another level due to the high risk of bias in the included studies.