Preimplantation genetic testing for aneuploidies (abnormal number of chromosomes) in in vitro fertilisation

Simone Cornelisse; Miriam Zagers; Elena Kostova; Kathrin Fleischer; Madelon van Wely; Sebastiaan Mastenbroek

doi:10.1002/14651858.CD005291.pub3

Diagnósticos genéticos preimplantacionales para detectar aneuploidías (número anormal de cromosomas) en la fertilización in vitro

Authors' declarations of interest

Version published: 08 September 2020 Version history

https://doi.org/10.1002/14651858.CD005291.pub3

Collapse all Expand all

Resumen

available in

Antecedentes

En la fecundación in vitro (FIV) con o sin inyección intracitoplasmática de espermatozoides (ICSI), la selección del embrión o los embriones más competentes para la transferencia se basa en criterios morfológicos. Sin embargo, muchas mujeres no logran un embarazo incluso después de la transferencia de embriones de "buena calidad". Una de las posibles causas es que estos embriones morfológicamente normales muestran un número anormal de cromosomas (aneuploidías). Por consiguiente, los análisis genéticos preimplantacionales para detectar aneuploidías (PGT‐A ), antes conocidos como diagnóstico genético preimplantacional (DGP), se desarrollaron como un método alternativo para seleccionar embriones para su transferencia en la FIV. En el PGT‐A, el corpúsculo polar o una o unas pocas células del embrión se obtienen por biopsia y se analizan. Sólo se transfieren los corpúsculos polares y los embriones que muestran un número normal de cromosomas.

Se demostró que la primera generación de PGT‐A, mediante biopsia en estadio de división y la hibridación fluorescente in situ (FISH) para el análisis genético, no eran efectivos para mejorar las tasas de nacidos vivos. Desde entonces, se han desarrollado nuevas metodologías de PGT‐A que realizan el procedimiento de biopsia en otras fases de desarrollo y utilizan diferentes métodos de análisis genético.

El hecho de que el PGT‐A mejore los desenlaces de la FIV y sea beneficioso para los pacientes sigue siendo controvertido.

Objetivos

Evaluar la efectividad y la seguridad del PGT‐A en mujeres que han recibido FIV.

Métodos de búsqueda

Se realizaron búsquedas en el Registro de Ensayos del Grupo Cochrane de Ginecología y Fertilidad (Cochrane Gynaecology and Fertility Group), CENTRAL, MEDLINE, Embase, PsycINFO, CINAHL y dos registros de ensayos en septiembre de 2019, y se verificaron las listas de referencias de los artículos pertinentes.

Criterios de selección

Fueron aptos para su inclusión todos los ensayos controlados aleatorizados (ECA) que informaron de datos sobre desenlaces clínicos en participantes sometidas a FIV con PGT‐A versus FIV sin PGT‐A.

Obtención y análisis de los datos

Dos autores de la revisión de forma independiente seleccionaron los estudios para su inclusión, evaluaron el riesgo de sesgo y extrajeron los datos de los estudios. El desenlace principal fue la tasa acumulada de nacidos vivos (TaNV). Los desenlaces secundarios fueron la tasa de nacidos vivos (TNV) tras la primera transferencia de embriones, tasa de aborto espontáneo, tasa de embarazo en curso, tasa de embarazo clínico, tasa de embarazo múltiple, proporción de mujeres que llegan a recibir una transferencia de embriones y el número medio de embriones por transferencia.

Resultados principales

Se incluyeron 13 ensayos con 2794 mujeres. La calidad de la evidencia varió de baja a moderada. Las principales limitaciones fueron la imprecisión, inconsistencia y el riesgo de sesgo de publicación.

FIV con PGT‐A versus FIV sin PGT‐A mediante análisis del genoma completo

Biopsia del corpúsculo polar

Un ensayo utilizó la biopsia del corpúsculo polar con hibridación genómica comparada de chips (aCGH). No se sabe con certeza si la adición de PGT‐A mediante biopsia del corpúsculo polar aumenta la TaNV en comparación con la FIV sin PGT‐A (odds ratio (OR) 1,05, intervalo de confianza (IC) del 95%: 0,66 a 1,66, un ECA, N = 396, evidencia de calidad baja). La evidencia sugiere que para una TaNV observada del 24% en el grupo de control, la probabilidad de un nacido vivo tras los resultados de un ciclo de FIV con PGT‐A es de entre el 17% y el 34%. No se sabe si la TNV después de la primera transferencia de embriones mejora con el PGT‐A por biopsia del corpúsculo polar (OR 1,10, IC del 95%: 0,68 a 1,79, un ECA, N = 396, evidencia de calidad baja). El PGT‐A por biopsia del corpúsculo polar puede reducir la tasa de abortos espontáneos (OR 0,45, IC del 95%: 0,23 a 0,88, un ECA, N = 396, evidencia de calidad baja). No se disponía de datos de la tasa de embarazos en curso. El efecto de la PGT‐A por biopsia del corpúsculo polar en la mejora de la tasa de embarazos clínicos es incierto (OR 0,77, IC del 95% 0,50 a 1,16, un ECA, N = 396, evidencia de calidad baja).

Biopsia en la fase de blastocisto

Un ensayo utilizó la biopsia en fase de blastocisto con secuenciación de nueva generación. No se sabe con certeza si la FIV con PGT‐A por biopsia en fase de blastocisto aumenta la TaNV en comparación con la FIV sin PGT‐A, ya que no se disponía de datos. No se sabe si la TNV después de la primera transferencia de embriones mejora con el PGT‐A por biopsia en fase de blastocisto (OR 0,93, IC del 95%: 0,69 a 1,27, un ECA, N = 661, evidencia de calidad baja). No se sabe con certeza si el PGT‐A con biopsia en fase de blastocisto reduce la tasa de abortos espontáneos (OR 0,89, IC del 95%: 0,52 a 1,54, un ECA, N = 661, evidencia de calidad baja). No se disponía de datos sobre la tasa de embarazos en curso ni sobre la tasa de embarazos clínicos.

FIV con PGT‐A versus FIV sin PGT‐A mediante FISH para el análisis genético

En esta comparación se incluyeron 11 ensayos. No se sabe si la FIV con PGT‐A aumenta la TaNV (OR 0,59, IC del 95%: 0,35 a 1,01, un ECA, N = 408, evidencia de calidad baja). La evidencia sugiere que para una TaNV media observada del 29% en el grupo de control, la probabilidad de un nacido vivo después de los resultados de un ciclo de FIV con PGT‐A es de entre el 12% y el 29%. El PGT‐A por FISH probablemente reduce los nacidos vivos después de la primera transferencia en comparación con el grupo de control (OR 0,62, 95% CI 0,43 a 0,91, 10 ECA, N = 1680, I² = 54%, evidencia de calidad moderada). La evidencia sugiere que para una TNV media observada por primera transferencia del 31% en el grupo de control, la probabilidad de nacidos vivos después de la primera transferencia de embriones con PGT‐A es de entre el 16% y el 29%. Probablemente haya poca o ninguna diferencia en la tasa de abortos espontáneos entre el PGT‐A y el grupo de control (OR 1,03, 95%, IC 0,75 a 1,41; 10 ECA, N = 1680, I² = 16%; evidencia de calidad moderada). La adición de PGT‐A podría reducir la tasa de embarazos en curso (OR 0,68, IC del 95%: 0,51 a 0,90, cinco ECA, N = 1121, I² = 60%, evidencia de baja calidad) y probablemente reduce los embarazos clínicos (OR 0,60, IC del 95%: 0,45 a 0,81, cinco ECA, N = 1131; I² = 0%, evidencia de calidad moderada).

Conclusiones de los autores

No hay suficiente evidencia de buena calidad de una diferencia en la tasa acumulada de nacidos vivos, la tasa de nacidos vivos después de la primera transferencia de embriones o en la tasa de abortos espontáneos entre la FIV con y sin PGT‐A tal y como se realiza actualmente. No se disponía de datos sobre las tasas de embarazos en curso. El efecto del PGT‐A en la tasa de embarazos clínicos es incierto.

Las mujeres deben ser conscientes de que no se sabe con certeza si añadir a la FIV un PGT‐A con análisis del genoma completo es eficaz, especialmente en vista del carácter invasivo y los costes que entraña el PGT‐A. El PGT‐A con FISH para el análisis genético es probablemente perjudicial.

La evidencia disponible actual es insuficiente para apoyar el PGT‐A en la práctica clínica habitual.

PICOs

Population

Intervention

Comparison

Outcome

The PICO model is widely used and taught in evidence-based health care as a strategy for formulating questions and search strategies and for characterizing clinical studies or meta-analyses. PICO stands for four different potential components of a clinical question: Patient, Population or Problem; Intervention; Comparison; Outcome.

See more on using PICO in the Cochrane Handbook.

Resumen en términos sencillos

available in

Diagnósticos genéticos preimplantacionales para detectar cantidades anormales de cromosomas en parejas que se someten a una fertilización in vitro

Pregunta de la revisión

¿Mejoran los diagnósticos genéticos preimplantacionales para detectar cantidades anormales de cromosomas las posibilidades de un embarazo seguido de un bebé nacido vivo?

Antecedentes

En la fecundación in vitro (FIV) con o sin inyección intracitoplasmática de espermatozoides (ICSI), la selección del mejor o los mejores embriones para su transferencia se basa principalmente en la evaluación morfológica de los mismos, que incluye el número de células, la regularidad de las células y la presencia de fragmentos celulares. Desafortunadamente, casi dos tercios de las parejas no quedan embarazadas ni siquiera después de la transferencia de embriones de "buena calidad". Una de las posibles causas es que estos embriones tienen un número anormal de cromosomas (aneuploidía). El análisis genético preimplantacional para detectar aneuploidías (PGT‐A de sus siglas en inglés) es una técnica utilizada para analizar el número de cromosomas presentes en los embriones de FIV. En la PGT‐A se obtienen por biopsia para su análisis un corpúsculo polar (un producto de desecho de la meiosis materna) o una o unas pocas células del embrión. Sólo se transfieren al útero corpúsculos polares o embriones con un número normal de cromosomas en cada célula, los llamados "embriones euploides". La idea es que esto aumentará la tasa de nacidos vivos por cada ciclo de FIV iniciado. Estudios previos sobre PGT‐A que utilizaron una técnica de análisis genético llamada hibridación fluorescente in situ (FISH) encontraron que el PGT‐A no es efectivo para mejorar las tasas de nacidos vivos. Desde entonces, se han desarrollado nuevas metodologías y técnicas en el PGT‐A que realizan el procedimiento en los corpúsculos polares u otras etapas del desarrollo del embrión y utilizan diferentes métodos de análisis genético (hibridación genómica comparada de chips (del inglés aCGH) o secuenciación de nueva generación (del inglés NGS).

Se compararon los efectos beneficiosos y los riesgos de la FIV con y sin PGT‐A, realizada con diferentes técnicas en distintas etapas: corpúsculo polar u otra fase de desarrollo del embrión.

Características de los estudios

Se incluyeron 13 ensayos controlados aleatorizados (un tipo de estudio en que los participantes se asignan a uno de dos o más grupos de tratamiento mediante un método aleatorio) que incluyeron a 2794 mujeres. La evidencia está actualizada hasta septiembre de 2019.

Resultados clave

FIV con PGT‐A versus FIV sin PGT‐A mediante análisis del genoma completo

Biopsia del corpúsculo polar

No hubo suficiente evidencia para determinar si hay alguna diferencia en la tasa acumulada de nacidos vivos (TaNV) o en la tasa de nacidos vivos (TNV) después de la primera transferencia de embriones añadiendo un PGT‐A a la FIV mediante biopsia del corpúsculo polar. Puede haber una reducción en la tasa de abortos con la adición de PGT‐A. Ningún estudio informó sobre la tasa de embarazos en curso. También es incierto si la adición de PGT‐A a la FIV mediante biopsia del corpúsculo polar resulte en más embarazos clínicos.

Biopsia en la fase de blastocisto

Ningún estudio informó sobre la TaNV tras una biopsia en la fase de blastocisto. No está claro si la adición de PGT‐A mediante biopsia en fase de blastocisto mejora la TNV tras la primera transferencia de embriones o reduce la tasa de aborto. Ningún estudio informó sobre la tasa de embarazo en curso o la tasa de embarazo clínico.

FIV con PGT‐A versus FIV sin PGT‐A mediante FISH para el análisis genético

La adición de PGT‐A mediante FISH no aumenta la TaNV cuando se usa para el análisis genético. La tasa de nacidos vivos después de la primera transferencia de embriones se reduce probablemente con la adición de PGT‐A. Probablemente haya poca o ninguna diferencia en las tasas de aborto espontáneo entre la FIV con y sin PGT‐A mediante FISH. El PGT‐A mediante FISH puede reducir los embarazos en curso y probablemente disminuye los embarazos clínicos.

Calidad de la evidencia

La calidad de la evidencia varió de baja a moderada. Las principales limitaciones de la evidencia fueron el número reducido de estudios y episodios, la incoherencia de las estimaciones entre los estudios y los indicios de que los resultados pueden estar sesgados porque no se han publicado todos los estudios elegibles.

Authors' conclusions

Implications for practice

It is uncertain whether the addition of blastocyst stage preimplantation genetic testing for aneuploidies (PGT‐A) where genome‐wide analyses are used, that is the current standard in routine practice today, improves cumulative live birth rates or reduces miscarriage rate. Evidence based on one trial showed a decrease in miscarriage rate with the use of polar body biopsy. There were no data available on ongoing pregnancy rates. The effect of PGT‐A on the number of clinical pregnancies is also uncertain.

The currently available evidence does not support the routine use of PGT‐A with the use of fluorescence in situ hybridisation (FISH) for the genetic analysis. It is uncertain if PGT‐A with the use of FISH improves cumulative live birth rate. PGT‐A with the use of FISH probably decreases live birth rate after the first embryo transfer. There is probably little or no difference in the number of miscarriages between groups. PGT‐A with the use of FISH may reduce ongoing pregnancies and probably reduces the number of clinical pregnancies.

Women and partners who receive IVF should have access to up‐to‐date information on the uncertainty of the effectiveness of PGT‐A, especially considering the invasiveness and costs, in order to make informed choices regarding their treatment. The currently available evidence is insufficient to support PGT‐A in routine clinical practice.

Implications for research

Considering the outcomes of our review, new developments should be properly evaluated before their routine clinical application. This involves method assessment studies, pilot‐studies showing a potential benefit in terms of cumulative live birth per woman, followed by randomised controlled trials. New research on the effectiveness of PGT‐A should be specifically directed at retaining the highest cumulative live birth chances whilst preventing miscarriages. Another reason for the need for new research is the low number of new included studies in this update review. Future trials should ideally: randomise the participants before the ovum pick‐up so there is no withdrawal of participants without an embryo transfer; have a clear description of the indication(s) for PGT‐A in the study; include cumulative live birth rate and time to pregnancy as outcomes; and use the right unit of analysis (e.g. per woman, per ovum pick‐up).

New research on PGT‐A with the use of FISH for genetic analysis is no longer necessary.

Summary of findings

Open in table viewer

Summary of findings 1. Preimplantation genetic testing for aneuploidies with the use genome‐wide analyses in in vitro fertilisation

Outcomes per woman	Embryo stage of biopsy	*Illustrative comparative risks (95% CI)**		Relative effect (95% CI)	No. of participants (studies)	Quality of the evidence (GRADE)
Preimplantation genetic testing for aneuploidies with the use genome‐wide analyses in in vitro fertilisation
Patient or population: couples with an in vitro fertilisation indication Settings: fertility clinics Intervention: PGT‐A with the use of genome‐wide analyses Comparison: no PGT‐A
		Assumed risk	Corresponding risk
		Control	PGT‐A with the use of genome‐wide analyses
Cumulative live birth	Polar body biopsy	236 per 1000	245 per 1000 (169 to 338)	OR 1.05 (0.66 to 1.66)	396 (1 RCT)	⊕⊕⊝⊝ low^a
Live birth rate after the first embryo transfer	Polar body biopsy	199 per 1000	215 per 1000 (144 to 308)	OR 1.10 (0.68 to 1.79)	396 (1 RCT)	⊕⊕⊝⊝ low^a
Live birth rate after the first embryo transfer	Blastocyst stage biopsy	435 per 1000	417 per 1000 (347 to 494)	OR 0.93 (0.69 to 1.27)	661 (1 RCT)	⊕⊕⊝⊝ low^a
Miscarriage	Polar body biopsy	141 per 1000	69 per 1000 (36 to 127)	OR 0.45 (0.23 to 0.88)	396 (1 RCT)	⊕⊕⊝⊝ low^b
Miscarriage	Blastocyst stage biopsy	82 per 1000	73 per 1000 (44 to 121)	OR 0.89 (0.52 to 1.54)	661 (1 RCT)	⊕⊕⊝⊝ low^a
Ongoing pregnancy	No studies reported on ongoing pregnancy.
Clinical pregnancy	Polar body biopsy	366 per 1000	308 per 1000 (224 to 402)	OR 0.77 (0.50 to 1.16)	396 (1 RCT)	⊕⊕⊝⊝ low^a
The basis for the assumed risk* (e.g. the median control group risk across studies) is provided in footnotes. 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; OR: odds ratio; PGT‐A: preimplantation genetic testing for aneuploidies; RCT: randomised controlled trial
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.
^aDowngraded two levels for imprecision: results based on one study, small number of events; CI crosses the line of no effect. ^bDowngraded two levels for imprecision: results based on one study, small number of events.

Open in table viewer

Summary of findings 2. Preimplantation genetic testing for aneuploidies with the use of fluorescence in situ hybridisation (FISH) for the genetic analysis in in vitro fertilisation

Outcomes per woman	*Illustrative comparative risks (95% CI)**		Relative effect (95% CI)	No. of participants (studies)	Quality of the evidence (GRADE)
Preimplantation genetic testing for aneuploidies with the use of fluorescence in situ hybridisation (FISH) for the genetic analysis in in vitro fertilisation
Patient or population: couples with an in vitro fertilisation indication Settings: fertility clinics Intervention: PGT‐A with the use of FISH for the genetic analysis Comparison: no PGT‐A
	Assumed risk	Corresponding risk
	Control	PGT‐A with the use of FISH for the genetic analysis
Cumulative live birth	287 per 1000	192 per 1000 (124 to 289)	OR 0.59 (0.35 to 1.01)	408 (1 RCT)	⊕⊕⊝⊝ low^a
Live birth rate after the first embryo transfer	307 per 1000	215 per 1000 (160 to 287)	OR 0.62 (0.43 to 0.91)	1680 (10 RCTs)	⊕⊕⊕⊝ moderate^b
Miscarriage	105 per 1000	108 per 1000 (81 to 143)	OR 1.03 (0.75 to 1.41)	1680 (10 RCTs)	⊕⊕⊕⊝ moderate^c
Ongoing pregnancy	274 per 1000	204 per 1000 (161 to 253)	OR 0.68 (0.51 to 0.90)	1121 (5 RCTs)	⊕⊕⊝⊝ low^b,c
Clinical pregnancy	333 per 1000	219 per 1000 (174 to 267)	OR 0.56 (0.42 to 0.73)	1131 (5 RCTs)	⊕⊕⊕⊝ moderate^c
The basis for the assumed risk* (e.g. the median control group risk across studies) is provided in footnotes. 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; OR: odds ratio; PGT‐A: preimplantation genetic testing for aneuploidies; RCT: randomised controlled trial
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.
^aDowngraded two levels for imprecision: results based on one study, small number of events. ^bDowngraded one level for inconsistency: I² > 50% in results across studies. ^cDowngraded one level for imprecision: the CI for most studies crosses the line of no effect, or small number of events.

Background

Description of the condition

One in six couples experience subfertility, defined as the failure to conceive after one year of unprotected intercourse, at least once during their reproductive lifetime (Dyer 2016). In vitro fertilisation (IVF) with and without intracytoplasmic sperm injection (ICSI) has evolved as the intervention of choice to help these couples conceive.

In IVF, selection of the most competent embryo(s) for transfer is mainly based on morphological criteria, such as the number of pronuclei, the number and regularity of blastomeres, and the percentage of fragmentation in cleavage stage embryos, and assessment of trophectoderm, inner cell mass, and expansion in the blastocyst stage of embryo development. Even so, two‐thirds of women do not reach pregnancy, whilst some others might result in a miscarriage, after transfer of embryos that are morphologically of good quality (De Geyter 2018).

One of the suggested causes is that such morphologically normal embryos contain an abnormal number of chromosomes (aneuploidies). Those embryos have been regarded as the main reason for implantation failure, miscarriage, or recurrent miscarriages and prolonged time to pregnancy in IVF.

Since the early 1980s, many reports have been published showing numerical chromosome abnormalities in morphologically normal human cleavage stage embryos (Angell 1983; Munné 1993; Benadiva 1996; Delhanty 1997). Since aneuploid embryos are not expected to develop to term, preimplantation genetic testing for aneuploidies (PGT‐A) was introduced in 1995 to improve live birth rates (Verlinsky 1995; Sermon 2016).

Description of the intervention

The previous terminology of 'preimplantation genetic screening' (PGS) has been replaced by 'preimplantation genetic testing', or PGT, following a revision of terminology used in infertility care (Zegers‐Hochschild 2017). PGT is defined as a test performed to analyse the DNA from mature oocytes (polar bodies) or embryos (cleavage or blastocyst stage) for determining genetic abnormalities. This includes PGT for chromosomal structural rearrangements (PGT‐SR), PGT for monogenic/single‐gene defects (PGT‐M), and PGT for aneuploidies (PGT‐A). Although the technology used in PGT‐A and PGT‐SR is nearly identical, PGT‐A and PGT‐SR have completely different indications. PGT‐A aims to improve treatment outcome in subfertile couples undergoing an IVF treatment, whereas PGT‐M and PGT‐SR aim to prevent the birth of affected children in fertile couples with a high risk of transmitting genetic disorders.

In PGT‐A, embryos created in vitro are analysed for aneuploidies, and only those that show a normal number of chromosomes, that is those that are euploid for the chromosomes tested, are transferred into the uterine cavity. There are different approaches to obtain nuclear material for this genetic analysis.

One approach is aspiration of the first and second polar body from the unfertilised oocyte or the zygote (Verlinsky 1995; Montag 2009; Geraedts 2016; Verpoest 2018). Polar bodies are a waste product of maternal meiosis. Mitotic errors and paternally derived meiotic errors and mutations cannot be detected from polar bodies. Polar body biopsy can be an alternative to embryo biopsy due to regulations that prohibit embryo biopsy in specific regions. Another approach is removal of one or two blastomeres from embryos at the early cleavage stage (Handyside 1989). A third approach is removal of trophectoderm cells at the blastocyst stage (Dokras 1990). The amount of DNA is higher since multiple cells are analysed. Finally, new sources of embryonic genetic material can be obtained by blastocyst fluid aspiration, Gianaroli 2014; Capalbo 2018, or spent embryo culture medium analysis, where cell‐free genomic DNA is obtained (in a non‐invasive way) from the embryo culture medium, may potentially be used for genetic testing (Hammond 2017; Vera‐Rodriquez 2018; Belandres 2019; Leaver 2020).

The method of genetic analysis can be limited to a certain number of chromosomes, mostly when using fluorescence in situ hybridisation (FISH) for the analysis, or it can encompass almost all or all chromosomes in genome‐wide technologies. FISH technology is based on the use of specific DNA probes that are labelled with distinctive fluorochromes; following hybridisation, results are visualised via fluorescence microscopy. Multiple genome‐wide technologies are described; array‐based comparative genomic hybridisation (aCGH) (Wells 2008; Geraedts 2011), single nucleotide polymorphism (SNP) array (van Uum 2012), or next generation sequencing (NGS) (Treff 2013). The principle of aCGH is a molecular cytogenetic technique for the detection of chromosomal copy number changes. aCGH compares the genome against a reference DNA sample, utilising the same principles as FISH. The application of NGS for the detection of copy number changes differs from aCGH by using direct reads of genomic sequencing fragments and their quantification according to sequence read numbers instead of a signal intensity comparison between fluorescently labelled test and reference DNA samples.

PGT‐A was first recommended and carried out for the following indications: advanced maternal age (Gianaroli 1999; Munné 1999; Kahraman 2000; Obasaju 2001; Munné 2003; Montag 2004; Platteau 2005), repeated IVF failure (Gianaroli 1999; Kahraman 2000; Munné 2003; Pehlivan 2003; Wilding 2004), recurrent miscarriage (Pellicer 1999; Rubio 2003; Munné 2005; Rubio 2005), and severe male factor (Silber 2003; Platteau 2004). Later PGT‐A was also offered to younger women with a good prognosis for a pregnancy, as high aneuploidy rates were found in their embryos as well (Baart 2006; Goossens 2009).

How the intervention might work

Initially PGT‐A was introduced based on the hypothesis that selection of an aneuploid embryo from a cohort of embryos available for transfer in an IVF treatment would lead to better clinical outcomes (Verlinsky 1995). Part of the embryos available in IVF were reported to be aneuploid, and aneuploidies were expected to result in arrested development, implantation failure, or miscarriage (Wilton 2002). Nowadays, the main reason for choosing to use PGT‐A is the hypothesis that PGT‐A could reduce the chance of miscarriage and shorten time to pregnancy leading to live birth (Sermon 2016; Cimadomo 2020). PGT‐A can never increase the cumulative live birth rate, as it selects out the aneuploid embryos that would likely not have resulted in a pregnancy (Paulson 2017).

First‐generation PGT‐A was characterised by biopsy at the cleavage stage of embryo development using fluorescence in situ hybridisation (FISH) for the genetic analysis; however, it was demonstrated to be ineffective in improving IVF live birth rates and reducing miscarriage rates (Twisk 2006; Mastenbroek 2011). The mosaic nature of human preimplantation embryos, which means that not all cells of an embryo have the same chromosomal constitution, and the limited capability and accuracy of the FISH analysis, are considered the main reasons for this failure (Scriven 2010; van Echten‐Arends 2011; Mastenbroek 2014; Taylor 2014). Consequently, PGT‐A at the blastocyst stage with the use of genome‐wide analyses became common practice, as mosaicism was claimed to be less of a problem at later stages of embryo development, or could at least be noticed as multiple cells are available for the analysis at the blastocyst stage, and genome‐wide analyses were claimed to be more accurate.

PGT‐A and IVF is a controversial subject, as evidence has appeared undermining the rationale of PGT‐A (Rosenwaks 2018; Mochizuki 2020). It was demonstrated that mosaic embryos could also lead to healthy babies (Greco 2015). Subsequently it was proposed that mosaic embryos, which till then were in most cases discarded, could now be transferred to the uterus, although perhaps with a lower priority than euploid embryos (Cram 2019). An assumed reason is that there is a self‐correction process of aneuploidies in preimplantation embryos, allowing mosaic embryos to result in healthy live births, although perhaps with lower efficiency than fully euploid embryos (Bolton 2016; Singla 2020). But next to mosaicism, there is still a debate surrounding the accuracy of analysis methods, not only since healthy live births have been reported after transfer of aneuploid embryos (Patrizio 2019), but also since there is increasing evidence that the genome‐wide analyses currently being used have a high false‐positive rate (Popovic 2018; Lawrenz 2019; Popovic 2020).

Why it is important to do this review

It has been demonstrated that the former approach for PGT‐A, that is cleavage stage biopsy with the use of FISH for genetic analysis, is ineffective in improving IVF live birth rates and in reducing miscarriage rates (Twisk 2006). New approaches of PGT‐A have been developed that perform the procedure on polar bodies or other stages of embryo development and use different methods of genetic analysis.

These new approaches of PGT‐A have been introduced into routine clinical practice and are often offered; however, their effectiveness remains unclear (Mastenbroek 2014). We therefore undertook and updated this systematic review to investigate whether there is a difference in cumulative live birth rate, live birth per first embryo transfer, miscarriage rate, ongoing pregnancy rate, or clinical pregnancy rate. This involved comparing IVF with PGT‐A versus IVF without PGT‐A.

Objectives

To evaluate the effectiveness and safety of PGT‐A in women undergoing an IVF treatment.

Methods

Criteria for considering studies for this review

Types of studies

Randomised controlled trials (RCT) were eligible for inclusion in the review. We excluded non‐randomised studies and quasi‐randomised trials. We also excluded cross‐over trials, which are not appropriate in this context.

Due to bias by design in favour of PGT‐A, we excluded RCTs that only reported results on women reaching a euploid embryo transfer, studies where PGT‐A was not performed on the full cohort of embryos, and studies where multiple oocyte retrievals were allowed before a transfer was attempted without the possibility to extract data per ovum pick‐up.

Types of participants

Women undergoing IVF treatment with or without PGT‐A offered for all suggested indications or a combination of indications, that is:

advanced maternal age (AMA);
repeated IVF failure (RIF);
recurrent miscarriage;
testicular sperm extraction (TESE‐ICSI);
good‐prognosis patients.

Types of interventions

We compared IVF with PGT‐A versus IVF without PGT‐A.

We included the following two randomised comparisons.

IVF with PGT‐A versus IVF without PGT‐A with the use of genome‐wide analyses
IVF with PGT‐A versus IVF without PGT‐A with the use of FISH for the genetic analysis

In these two randomised comparisons, different subgroups exist between the moment of biopsy in embryo development, as follows.

Polar body biopsy
Cleavage stage biopsy
Blastocyst stage biopsy

Types of outcome measures

Primary outcomes

Cumulative live birth rate per woman, defined as the birth of a living child after 20 weeks of gestation, after one IVF cycle, including the results of cryopreserved thawed embryo transfers.

Secondary outcomes

Live birth rate per woman after the first embryo transfer, defined as the birth of a living child after 20 weeks of gestation after the first embryo transfer in an IVF cycle.
Miscarriage rate per woman and per pregnancy.
Ongoing pregnancy per woman, defined as ultrasound‐confirmed evidence of a gestation sac with fetal heart motion at 12 weeks.
Clinical pregnancy per woman, defined by the presence of an intrauterine gestational sac or the presence of intrauterine gestational sac with fetal heartbeat.
Multiple pregnancy per woman and per live birth.
Proportion of women reaching embryo transfer.
Mean number of embryos per transfer.

Search methods for identification of studies

We searched for all published and unpublished RCTs on PGT‐A versus non‐PGT‐A, with no language or date restriction and in consultation with the Cochrane Gynaecology and Fertility Group (CGF) Information Specialist.

Electronic searches

We searched the following electronic databases for relevant trials:

the Cochrane Gynaecology and Fertility Group (CGF) Specialised Register of Controlled Trials; ProCite platform, searched on 9 September 2019 (Appendix 1);
Cochrane Central Register of Controlled Trials (CENTRAL); Ovid platform, searched on 9 September 2019 (Issue August 2019) (Appendix 2);
MEDLINE Epub Ahead of Print, In‐Process & Other Non‐Indexed Citations; Ovid platform, searched from 1946 to 9 September 2019 (Appendix 3);
Embase; Ovid platform, searched from 1980 to 9 September 2019 (Appendix 4);
PsycINFO; Ovid platform, searched from 1806 to 9 September 2019 (Appendix 5);
CINAHL (Cumulative Index to Nursing and Allied Health Literature); EBSCO platform, searched from 1961 to 9 September 2019 (Appendix 6).

We combined the MEDLINE search with the Cochrane Highly Sensitive Search Strategy for identifying randomised trials described in the Cochrane Handbook for Systematic Reviews of Interventions (Chapter 6, 6.4.11; Lefebvre 2011). The Embase, PsycINFO, and CINAHL searches are combined with trial filters developed by the Scottish Intercollegiate Guidelines Network (SIGN) (www.sign.ac.uk/what-we-do/methodology/search-filters/).

We also searched the following other electronic sources of trials.

Trial registers for ongoing and registered trials, searched on 9 September 2019:

- US National Institutes of Health Ongoing Trials Register ClinicalTrials.gov (www.clinicaltrials.gov);
- World Health Organization International Clinical Trials Registry Platform (www.who.int/trialsearch/Default.aspx).

Note: it is now mandatory for Cochrane Reviews to include searches of trial registers.

LILACS (Latin American and Caribbean Health Science Information database; from 1982 to 9 September 2019) and other Spanish and Portuguese language databases found in the Virtual Health Library Regional Portal (VHL) (bvsalud.org/portal/?lang=en) (the right‐hand drop down box allows you to filter out MEDLINE records).

Searching other resources

We searched the following conference abstracts:

American Society for Reproductive Medicine and Canadian Fertility and Andrology Society (ASRM/CFAS) Conjoint Annual Meeting (2018), Abstracts of the Scientific Oral and Poster Sessions, Program Supplement;
European Society of Human Reproduction and Embryology (ESHRE) Annual Meeting (2018), Abstracts of the Scientific Oral and Poster Sessions, Program Supplement.

We handsearched the references cited in all obtained studies. We searched PubMed and Google for any recent trials that had not yet been indexed in MEDLINE.

We contacted experts in the field to obtain additional data. We contacted original authors for clarification and further data if the trial report was unclear.

We did not perform a separate search for adverse effects of PGT‐A. We considered adverse effects described in the studies only.

Data collection and analysis

Selection of studies

After an initial screen of titles and abstracts retrieved by the search, conducted by the CGF Information Specialist (Marian Showell), two review authors (SC and MZ) retrieved the full texts of all potentially eligible studies. We used the software program Covidence to manage the screening of titles and abstracts and to generate the PRISMA flow diagram (Covidence). Two review authors (SC and MZ) independently examined the full texts against the inclusion criteria to determine their eligibility. In the case of disagreement between review authors, a third review author (SM) was consulted to establish consensus on whether to include the trial or not. We documented the selection process with a PRISMA flow chart (Figure 1).

Figure 1

Study flow diagram.

Data extraction and management

Two review authors (SC and MZ) independently extracted the outcome data and information on location, clinical and design details, funding, and participants. In the case of disagreement between review authors, a third review author (SM) was consulted to establish consensus. Any differences were resolved by discussion. Details of the studies are provided in the Characteristics of included studies tables. Studies that appeared to meet the inclusion criteria but were excluded from the review are presented in the Characteristics of excluded studies tables along with the reasons for their exclusion in brief.

Assessment of risk of bias in included studies

Two review authors (SC and MZ) independently assessed the included studies for risk of bias using the Cochrane 'Risk of bias' tool (Higgins 2011). The tool addresses the following domains:

selection bias (random sequence generation, allocation concealment);
performance bias (blinding of participants and personnel);
detection bias (blinding of outcome assessors);
attrition bias (incomplete outcome data);
reporting bias (selective reporting);
other forms of bias such as selective reporting of subgroups, or potential influence from funders.

In case of disagreement between review authors, a third review author (EK) was consulted to establish consensus. We assigned judgements to each of these domains, as recommended in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011). Special attention was given to selective reporting, as it affects the internal validity of an individual study. Studies might not present data per woman starting an IVF cycle. This information is described in the Characteristics of included studies table and Figure 2 and Figure 3 and provides a context for discussing the reliability of the results. Our conclusions are presented in the 'Risk of bias' table.

Figure 2

Risk of bias graph: review authors' judgements about each risk of bias item presented as percentages across all included studies.

Figure 3

Risk of bias summary: review authors' judgements about each risk of bias item for each included study.

Measures of treatment effect

All of our outcomes represented dichotomous data (e.g. live birth rates). We used the numbers of events in the control and intervention groups of each study to calculate Mantel‐Haenszel odds ratios (ORs). We reversed the direction of effect of individual studies if required to ensure consistency across trials. We used mean difference (MD) for continuous outcomes. We presented 95% confidence intervals (CIs) for all outcomes. We assessed whether the estimates calculated in the review for individual studies were compatible in each case with the estimates reported in the study publications.

Unit of analysis issues

We analysed the data per woman randomised. We also analysed per clinical pregnancy for miscarriage. Data that did not permit valid analysis (e.g. per woman or per ovum pick‐up) were excluded. We counted multiple pregnancy as one event; this outcome was also analysed per live birth.

Dealing with missing data

We analysed the data on an intention‐to‐treat (ITT) basis as far as possible (i.e. including all randomised participants in the analysis in the groups to which they had been randomised). We anticipated that trials conducted over 10 years ago might not have data on cumulative live birth rates. When there was insufficient information in the published report, we attempted to contact the authors of the included studies for clarification. If missing data became available, we included these in the analysis. When we were unable to obtain these data, we undertook imputation of individual values for cumulative live birth rate and live birth rate per first embryo transfer only. Live birth was assumed not to have occurred in participants without a reported outcome. For other outcomes, we analysed only the available data.

Assessment of heterogeneity

We considered whether the clinical and methodological characteristics of the included studies were sufficiently similar for meta‐analysis to provide a meaningful summary. We assessed statistical heterogeneity by measuring the I² statistic. We assumed that there was substantial heterogeneity when I² was calculated as greater than 50% (Higgins 2011).

Assessment of reporting biases

We planned that if more than 10 studies were identified, we would produce a funnel plot to evaluate the risk of reporting bias, to explore the possibility of small‐study effects (a tendency for estimates of the intervention effect to be more beneficial in smaller studies) (Chaimani 2013). In view of the difficulty of detecting and correcting for publication bias and other reporting biases, we aimed to minimise their potential impact by ensuring a comprehensive search for eligible studies and by being alert to duplication of data.

Data synthesis

When multiple studies were available for a similar comparison, we used Review Manager 2014 software to perform the meta‐analyses, employing the Mantel‐Haenszel method with a fixed‐effect model; otherwise, results from trials that could not be combined were presented in data tables in a narrative format.

Subgroup analysis and investigation of heterogeneity

We planned that when sufficient data were available, we would perform subgroup analyses of the following subgroups to determine the potential causes of heterogeneity.

Polar body biopsy
Cleavage stage biopsy
Blastocyst stage biopsy

If we detected substantial heterogeneity, we would explore it by employing the random‐effects model. We aimed to take any statistical heterogeneity into account when interpreting the results, especially if there was any variation in the direction of effect.

We did not pool the results per subgroup for the comparison ‘IVF with PGT‐A versus IVF without PGT‐A with the use of genome‐wide analyses’, since the stages of biopsy are different.

Sensitivity analysis

We conducted sensitivity analyses for the primary outcome. These analyses included consideration of whether the review conclusions would have differed if:

the summary effect measure had been risk ratio rather than odds ratio;
eligibility had been restricted to studies with low risk of bias for randomisation and allocation concealment.

Overall quality of the body of evidence: 'Summary of findings' tables

We prepared 'Summary of findings' tables using GRADEpro GDT and Cochrane methods. These tables evaluate the overall quality of the body of evidence for the main review outcomes (cumulative live birth, live birth rate after the first embryo transfer, miscarriage rate per woman, ongoing pregnancy rate, and clinical pregnancy rate) for the two main review comparisons (IVF with PGT‐A versus IVF without PGT‐A with use of genome‐wide analyses; and IVF with PGT‐A versus IVF without PGT‐A with use of FISH for genetic analysis). We assessed the quality of the evidence using the GRADE criteria: risk of bias, consistency of effect, imprecision, indirectness, and publication bias. Two review authors (SC and MZ) independently made judgements about evidence quality (high, moderate, low, or very low); in the case of disagreement between review authors, a third review author (EK) was consulted to establish consensus. Judgements were justified, documented, and incorporated into the reporting of results for each outcome.

Results

Description of studies

Results of the search

For this update, we screened 1899 titles and identified 22 articles from searching electronic databases and other resources. From these 22 studies, we included three new studies, nine studies were earlier included. One study was excluded in the previous version of the review, but is included in this updated version (Werlin 2003).12 articles were excluded because they did not meet the inclusion criteria of the review; the reasons for their exclusion are shown in Characteristics of excluded studies. One of these trials was stopped early due to recruitment problems and is not published (NCT02265614). Ten other studies are ongoing and documented in the Characteristics of ongoing studies section of the review. One study is awaiting classification (see Characteristics of studies awaiting classification). Consequently, we included 13 studies with a total of 2794 participants in the review. Details of the screening and selection process are shown in the PRISMA study flow diagram (Figure 1).

Included studies

Study design and setting

We included a total of 13 studies in this update review, nine studies from the previous version and three new studies. One earlier excluded study was included in this version of the review (Werlin 2003). This study was previously excluded for the reason that only biochemical pregnancy rates were reported, but it was possible to extract data for one secondary outcome, therefore this study is now included.

A couple was offered one treatment cycle in 10 trials, Werlin 2003; Staessen 2004; Blockeel 2008; Hardarson 2008; Jansen 2008; Staessen 2008; Meyer 2009; Schoolcraft 2009; Rubio 2013; Munné 2019, and a maximum of three cycles in one trial (Mastenbroek 2007). In one trial a non‐randomised treatment cycle was offered to participants in the PGT‐A group if in the first treatment cycle all oocytes were aneuploid (Verpoest 2018). Eleven participants started a second cycle, 10 of which had ICSI in a second cycle, and PGT‐A was performed for nine of them. Of the nine evaluated participants, six did not have any euploid embryos in this second cycle. The authors kindly provided us data about those outcomes, which permitted an analysis per first treatment cycle (Verpoest 2018 [pers comm]). In one trial a couple could participate in the study several times with independent randomisation for each cycle (Debrock 2010). The authors of the Twisk 2006 Cochrane Review provided data for this trial, in which each participant was only included in one treatment group, by personal communication, and we used them again in this update (DeBrock 2010 [pers comm]).

Rubio 2013 published one paper where two studies are described. In the first study patients with AMA were included, and in the second study patients with RIF could participate. We only used the data for women with RIF in this review. In the study group of participants with AMA, a second IVF treatment cycle was offered to participants before randomisation, therefore no data per started cycle could be extracted.

In three trials one embryo was transferred (Jansen 2008; Staessen 2008; Munné 2019); in one trial two embryos were transferred if there were two embryos available (Mastenbroek 2007); in two trials one or two embryos were transferred according to the policy per centre, participant wishes, and the availability of normal embryos (Hardarson 2008; Verpoest 2018); and in one trial up to three embryos were transferred (Blockeel 2008). In one trial the number of embryos transferred depended on the age of the woman, namely up to three blastocysts when the woman was between 37 and 39 years old and up to a maximum of six blastocysts if the woman was 40 years of age or older (Staessen 2004). In one trial a maximum of two to three embryos were transferred before 1 July 2003, and only one embryo after 1 July 2003 in the first IVF attempt in women younger than 36 years (Debrock 2010). In two trials the transfer policy was not described (Meyer 2009; Schoolcraft 2009).

All trials performed fresh embryo transfers, except for one recent trial where a freeze‐all strategy was used (Munné 2019). Data on cryopreservation and pregnancies originating from frozen‐thawed embryos were available in four trials (Mastenbroek 2007; Debrock 2010; Rubio 2013; Verpoest 2018). In one of these trials only one embryo transfer was performed in the PGT‐A group and two in the control group, although many more embryos were cryopreserved (Debrock 2010). In this study, one pregnancy was obtained in the PGT‐A group after a mixed transfer of both a non‐biopsied embryo (from a previous IVF cycle) and a biopsied embryo, so no conclusion regarding the origin of the embryo could be made. In this study no pregnancies were obtained in the control group.

In all of the included studies, in the control group the morphologically best embryos were transferred, and in the intervention group embryos that were found to be chromosomally normal were transferred. In two studies undetermined embryos with good morphologic features were transferred if no chromosomally normal embryos with good morphologic features were available (Mastenbroek 2007; Verpoest 2018); in the other studies undetermined embryos were not transferred. Although in one study, one embryo was transferred in which no result was obtained for chromosomes 16 and 18 due to technical difficulties, and one embryo was transferred in which the chromosomal pattern was only evaluated in one nucleus, whilst this study normally removed and investigated two embryos (Debrock 2010).

The sample size was based on a power calculation in 10 trials (Staessen 2004; Mastenbroek 2007; Blockeel 2008; Hardarson 2008; Jansen 2008; Staessen 2008; Meyer 2009; Debrock 2010; Verpoest 2018; Munné 2019). In four trials the calculated number of inclusions was not reached. In one trial the study was ended prematurely because an interim analysis showed such a lower implantation rate for the PGT‐A group that it was considered unethical to continue (Meyer 2009); in one trial the study was ended prematurely because an interim analysis showed futility (Staessen 2008); and a further trial was ended prematurely because the trend was opposite to that required to disprove the null hypothesis and because in the control group many more cryo‐stored blastocysts were accumulating (Jansen 2008). In one trial the study was ended prematurely because an interim analysis showed a very low conditional power of superiority for the primary outcome (Hardarson 2008). In one trial the targeted sample size was not reached because of suboptimal recruitment; however, the included sample allowed a 90% power to detect the targeted increase (Verpoest 2018). In three studies the power calculation was based on embryos instead of women (Staessen 2004; Blockeel 2008; Debrock 2010).

In the Staessen 2004 study, 400 women were randomised, with 200 assigned to the PGT‐A group and 200 to the control group. In the PGT‐A group one woman did not fulfil the inclusion criteria, whilst in the control group 10 women did not fulfil the inclusion criteria, therefore a total of 199 versus 190 women were correctly assigned to the treatment and control group respectively. In 51 women in the treatment group and 49 women in the control group ovum pick‐up was not performed. In the original article these women were excluded from the analysis, but we included these women in an ITT analysis, therefore we included from this article a total of 199 women in the PGT‐A group and 190 women in the control group.

In the Blockeel 2008 study, 95 women were randomised to the PGT‐A group and 105 women were randomised to the control group. In the PGT‐A group eight women did not fulfil the inclusion criteria, whilst in the control group 10 women did not fulfil the inclusion criteria; we excluded these from the meta‐analysis. The authors excluded nine women in the PGT‐A group and 11 women in the control group because of wrong allocation and a spontaneous pregnancy, therefore a total of 87 versus 95 women were correctly assigned to the treatment and the control group respectively. Nine women in the PGT‐A group and 10 women in the control group were wrongly allocated. These women should have been included in an ITT analysis; however, this was not possible because the results of their treatment are unknown to us. The study authors further excluded six women in the PGT‐A group and 18 women in the control group because of insufficient ovarian response, stop further fertility treatment, and spontaneous pregnancy. These women should have been included in an ITT analysis, but this was not possible for the same reason described above. Consequently, from this article we included a total of 72 women in the PGT‐A group and 67 women in the control group.

In the Meyer 2009 study, four women were excluded from the analysis because no embryo transfer was performed for personal reasons. In the original article these women were excluded from the analysis, but we included these women in an ITT analysis, therefore from this article we included 23 women in the PGT‐A group and 24 women in the control group.

In the Debrock 2010 study, 52 women were randomised to the PGT‐A group and 52 women to the control group. However, women could be included several times with independent randomisation for each cycle, which introduces a bias since these cycles are not independent. After excluding these cycles, that is when each woman could be included only once, 44 women could be included in the PGT‐A group and 50 women in the control group (DeBrock 2010 [pers comm]). Fifteen women in the PGT‐A group and 28 women in the control group did not receive the intended treatment; we included these women in an ITT analysis, therefore from this study we included 44 women in the PGT‐A group and 50 women in the control group.

One study reported only percentages and used transfers and pregnancies as units of analysis, therefore we recalculated the numbers per participant for the various outcomes (Schoolcraft 2009). From this article we included 32 women in the PGT‐A group and 30 women in the control group.