Independent Review of Assisted Reproductive Technologies

Transcription

1 Assisted Reproductive Technologies Review Committee Report of the Independent Review of Assisted Reproductive Technologies APPENDICES

2 Assisted Reproductive Technologies Review Committee APPENDIX A February 2006

3 NHMRC Clinical Trials Centre Assisted Reproductive Technologies Review February 2006 Suzanne Dyer, Alison Griffiths, Simon Eckermann and Sarah Lord NHMRC Clinical Trials Centre Sydney, Australia

4

5 Contents Glossary... ix Executive summary... xii Introduction...1 Background... 2 Assisted Reproductive Technologies... 2 Infertility... 3 ART procedures... 7 Artificial insemination procedures Clinical need...11 Safety of ART Methodological issues Current reimbursement arrangement International ART policies Methods Review of literature Data extraction Appraisal of the evidence Economic evaluation Results...31 Systematic literature review IVF IVF by maternal age IVF by cycle number IVF by duration of infertility IUI IUI by maternal age IUI by cycle number IUI vs IVF ICSI ICSI vs IVF effectiveness by male factor infertility ICSI vs IVF safety congenital malformation rate Policy of using frozen embryo transfer Cost-effectiveness Existing literature Model Structure Results Conclusions Appendix A ART Review Committee Appendix B HTA studies included in the review Assisted reproductive technologies iii

6 Appendix C Electronic databases and HTA websites Appendix D Literature searches Appendix E HTA reports, Appendix F MBS items for reproductive services Appendix G Economic scoping searches Appendix H Quality criteria Appendix I Economic inputs Appendix J Cost-effectiveness of IUI and ICSI IUI ICSI Abbreviations References iv Assisted reproductive technologies

7 Tables Table 1 MBS funding for assisted reproductive services Table 2 Australian Government expenditure on assisted reproductive services , in $million Table 3 PICO criteria for clinical questions in assisted reproductive technologies review Table 4 Clinical questions for assisted reproductive technologies review determined a priori Table 5 Electronic databases searched, limited to Table 6 Eligibility criteria for identification of relevant studies Table 7 Evidence dimensions Table 8 NHMRC designations of levels of evidence for interventions a Table 9 Interim NHMRC designations of levels of evidence for prognosis Table 10 Success rates of fresh, non-donor ART by maternal age, ANZARD data...33 Table 11 Success rates of all ART procedures (including donor and frozen cycles) by maternal age, ANZARD data Table 12 Live birth rates by maternal age for fresh, non-donor IVF cycle started, UK HFEA data 1995 to Table 13 Characteristics of studies reporting the safety or effectiveness of IVF by woman s age Table 14 Results of studies of IVF by maternal age Table 15 Success rates of fresh, non-donor IVF by maternal age and cycle number, ANZARD data Table 16 Characteristics of studies reporting the safety or effectiveness of IVF by cycle number Table 17 Results of studies of IVF by cycle number Table 18 Characteristics and results of study reporting the effectiveness of IVF by duration of infertility Table 19 Characteristics of studies reporting the safety or effectiveness of IUI by woman s age, cycle number or duration of infertility Table 20 Results of study of low-dose IUI by maternal age Table 21 Results of studies of safety or effectiveness of low-dose IUI versus IVF Table 22 Characteristics of studies reporting the effectiveness of ICSI versus IVF, by male factor infertility Table 23 Results of study of effectiveness of ICSI versus IVF by male factor infertility Table 24 Characteristics of studies ( ) reporting the safety of ICSI versus IVF Assisted reproductive technologies v

8 Table 25 Results of studies of safety of ICSI versus IVF Table 26 Characteristics of studies reporting the safety or effectiveness of a policy of using fresh and frozen embryo transfer Table 27 Results of studies of policy of using frozen embryos Table 28 MBS and PBS costs of fresh cycle IVF treatment delivery Table 29 Modelled cost estimates by maternal age for an average fresh cycle treatment from oocyte pick-up (Government perspective) Table 30 Modelled cost estimates by maternal age for an average fresh cycle treatment from oocyte pick-up (societal perspective) Table 31 Modelled cost estimates by maternal age for an average frozen cycle treatment (Government and societal perspective) Table 32 IVF outcome costs Table 33 Cumulative live births and costs by number of IVF attempts (treatment programs) per 1000 women aged years commencing IVF treatment Table 34 Cumulative live births and costs by number of IVF attempts (treatment programs) per 1000 women aged years commencing IVF treatment Table 35 Cumulative live births and costs by number of IVF attempts (treatment programs) per 1000 women aged years commencing IVF treatment Table 36 Cumulative live births and costs by number of IVF attempts (treatment programs) per 1000 women aged years commencing IVF treatment Table 37 Incremental cost per live birth (Government perspective) by maternal age and by IVF treatment program (fresh plus frozen cycles) Table 38 Incremental cost per live birth (societal perspective) by maternal age and by IVF treatment program (fresh plus frozen cycles) Table 39 Results of one way sensitivity analyses Table 40 One way sensitivity analyses pregnancy rates varied to lower 95 per cent confidence limit (societal perspective) Table 41 One way sensitivity analyses pregnancy rates varied to upper 95 per cent confidence limit (societal perspective) Table 42 Methods and results of previous Australian (AHTAC) report Table 43 Methods and results of HTA studies included in the review Table 44 Electronic databases and HTA websites searched in this review Table 45 Medline search strategy, 1966 to August week4, Table 46 PreMedline search strategy, 20 September Table 47 EMBASE search strategy, 1996 to 2005 Week Table 48 Scope of existing HTA reports included in the review Table 49 Scope of existing HTA reports excluded from the review vi Assisted reproductive technologies

9 Table 50 Medline search strategy, 1966 to November week3, Table 51 PreMedline search strategy, 8 December Table 52 EMBASE search strategy, 1980 to 2005 Week Table 53 Criteria for reviewing quality of included studies Table 54 Base case pregnancy rates by cycle and by maternal age Table 55 Base case live birth rates by cycle and maternal age Table 56 Base case discontinuation rates by cycle and maternal age Table 57 Proportion of fresh cycles modelled with failed fertilisation Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Treatment cycles started, pregnancies and live births, Australia and New Zealand, Clinical flowchart to illustrate current use of ART and IUI in Australia Number of transfer cycles by ART treatment type, Australia and New Zealand, Summary of the process used to identify and select studies for the review Live births per cycle started by woman s age for fresh, non-donor ART treatment, Figure 6 Success rates for ART by maternal age, US CDC data Figure 7 Modelled IVF treatment program Figure 8 Cumulative live births and costs by number of IVF treatment programs per couples commencing IVF treatment Government perspective (MBS plus Medicare Safety Net costs) Figure 9 Cumulative live births and costs by number of IVF treatment programs per couples commencing IVF treatment societal perspective (total costs) Assisted reproductive technologies vii

10

11 Glossary The following glossary is adapted for local usage from the assisted reproductive technology (ART) glossary produced by the World Health Organisation (WHO, 2002). Assisted hatching: an in-vitro procedure in which the zona pellucida of an embryo (usually at 8-cell stage or a blastocyst) is perforated by chemical, mechanical or laserassisted methods to assist separation of the blastocyst from the zona pellucida. Assisted reproductive technology (ART): for the purposes of this report, this is defined as all treatments or procedures for infertility for the purpose of achieving a live birth. This includes, but is not limited to, in-vitro fertilisation and transcervical embryo transfer, gamete intrafallopian transfer, zygote intrafallopian transfer gamete and embryo cryopreservation, oocyte and embryo donation, controlled ovarian stimulation and artificial insemination. Biochemical pregnancy (preclinical pregnancy): evidence of conception based only on biochemical data in the serum or urine before ultrasound evidence of a gestational sac. Blastocyst: an embryo with a fluid-filled blastocele cavity (usually developing by five or six days after fertilisation). Cancelled cycle: an ART cycle in which ovarian stimulation or monitoring has been carried out with the intent of undergoing ART but which did not proceed to follicular aspiration, or in the case of a thawed embryo, to transfer. Clinical pregnancy: evidence of pregnancy by clinical or ultrasound parameters (ultrasound visualization of a gestational sac). It includes ectopic pregnancy. Multiple gestational sacs in one patient are counted as one clinical pregnancy. Clinical pregnancy rate: number of clinical pregnancies expressed per 100 initiated cycles, Aspiration cycles or embryo transfer cycles. When clinical pregnancy rates are given, the denominator (initiated, aspirated or embryo transfer cycles) must be specified. Congenital abnormality: abnormality existing at birth, regardless of cause. Controlled ovarian hyperstimulation (COH): medical treatment to induce the development of ovarian follicles to produce oocytes. Cryopreservation: freezing and storage of gametes, zygotes or embryos. Delivery rate: number of deliveries expressed per 100 initiated cycles, aspiration cycles or embryo transfer cycles. When delivery rates are given, the denominator (initiated, aspirated or embryo transfer cycles) must be specified. It includes deliveries that resulted in a live birth and/or stillbirth. The delivery of a singleton, twin or other multiple pregnancy is registered as one delivery. Early neonatal death: death occurring within the first seven days after delivery. Ectopic pregnancy: a pregnancy in which implantation takes place outside the uterine cavity. Embryo: product of conception from the time of fertilisation to the end of the embryonic stage eight weeks after fertilisation (the term pre-embryo or dividing conceptus has been replaced by embryo). Embryo donation: the transfer of an embryo resulting from gametes that did not originate from the recipient and/or her partner. Embryo transfer (ET): procedure in which embryo(s) are placed in the uterus or fallopian tube. ix Assisted reproductive technologies

12 Embryo transfer cycle: ART cycle in which one or more embryos are transferred into the uterus or fallopian tube. Fertilisation: the penetration of the ovum by the spermatozoon and fusion of genetic materials resulting in the development of a zygote. Foetus: the product of conception starting from completion of embryonic development (at eight completed weeks after fertilisation) until birth or abortion. Full-term birth: a birth that takes place at 37 or more completed weeks of gestational age. This includes both live births and stillbirths. Gamete intrafallopian transfer (GIFT): ART procedure in which both gametes (oocytes and sperm) are transferred to the fallopian tubes. Gestational age: age of an embryo or foetus calculated by adding 14 days (2 weeks) to the number of completed weeks since fertilisation. Gestational sac: a fluid-filled structure containing an embryo that develops early in pregnancy usually within the uterus. Hatching: it is the process that precedes implantation by which an embryo at the blastocyst stage separates from the zona pellucida. Implantation: the attachment and subsequent penetration by the zona-free blastocyst (usually in the endometrium) which starts five to seven days following fertilisation. In-vitro fertilisation (IVF): an ART procedure which involves extracorporeal fertilisation. Infertility: failure to conceive after at least one year of unprotected sexual intercourse. Initiated cycles: ART treatment cycles in which the woman receives ovarian stimulation, or monitoring in the case of spontaneous (unstimulated) cycles, irrespective of whether or not follicular aspiration is attempted. Intracytoplasmatic (intracytoplasmic) sperm injection (ICSI): IVF procedure in which a single spermatozoon is injected through the zona pellucida into the oocyte. Karyotype: the chromosome set. Live birth: a birth in which a foetus is delivered with signs of life after complete expulsion or extraction from its mother beyond 20 completed weeks of gestational age. (Live births are counted as birth events, eg a twin or triplet live birth is counted as one birth event.) Live-birth delivery rate: number of live-birth deliveries expressed per 100 initiated cycles, aspiration cycles or embryo transfer cycles. When delivery rates are given, the denominator (initiated, aspirated or embryo transfer cycles) must be specified. It includes deliveries that resulted in at least one live birth. The delivery of a singleton, twin or other multiple birth is registered as one delivery. Malformation rate: includes all structural, functional, genetic and chromosomal abnormalities identified in aborted tissue or diagnosed before or subsequent to birth. Micromanipulation (also referred to as assisted fertilisation): the use of special micromanipulative technology that allows operative procedures to be performed on the oocyte, sperm or embryo. Neonatal death: death within 28 days of birth. Newborns or infants born: the number of live births plus stillbirths. Oocyte donation: an ART procedure performed with third-party oocytes. Oocyte pick-up (OPU): initiated ART cycle in which one or more oocytes are retrieved by aspiration. Assisted reproductive technologies x

13 Preimplantation genetic diagnosis (PGD): screening of cells from preimplantation embryos for the detection of genetic and/or chromosomal disorders before embryo transfer. Preterm birth: a birth which takes place after at least 20, but less than 37, completed weeks of gestation. This includes both live births and stillbirths. Births are counted as birth events (eg a twin or triplet live birth is counted as one birth event). Spontaneous abortion: spontaneous loss of a clinical pregnancy before 20 completed weeks of gestation or, if gestational age is unknown, a weight of 500 g or less. Stillbirth: a birth in which the foetus does not exhibit any signs of life when completely removed or expelled from the birth canal at or above 20 completed weeks of gestation. Stillbirths are counted as birth events (eg a twin or triplet stillbirth is counted as one birth event). Treatment cycle: a treatment episode commencing with the first patient intervention (usually controlled ovarian hyperstimulation). An ART cycle using unstimulated ovulation is considered as commencing with the provision of ovulation monitoring services. Zygote intrafallopian transfer (ZIFT): procedure in which the zygote, in its pronuclear stage of development, is transferred into the fallopian tube. Zygote: is the diploid cell, resulting from the fertilisation of an oocyte by a spermatozoon, which subsequently develops into an embryo. xi Assisted reproductive technologies

14 Executive summary Background Assisted reproductive technologies (ARTs) are treatments for infertility, performed to assist a couple to achieve a live birth. In Australia, ARTs reimbursed by the government include artificial insemination or stimulation of the ovaries to increase egg production, or procedures where the eggs and sperm are combined in a laboratory (in vitro) and the resulting embryo is transferred to a woman,. The most relevant measure of the success of ART is the number of live births produced expressed as a percentage of the number of cycles of ART treatment started. The Commonwealth Government set up an independent review of ART on 4 July 2005 to advise the Commonwealth Minister for Health and Ageing. As part of this review, the National Health and Medical Research Council (NHMRC) Clinical Trials Centre conducted a review of the current evidence of the clinical effectiveness and safety of ART, in addition to an economic analysis. Five specific questions were addressed: What are the effects of maternal age, treatment cycle number and the duration of infertility on the success rates of in-vitro fertilisation (IVF)? What are the effects of maternal age and treatment cycle number on the success rates of intrauterine insemination (IUI) with low-dose stimulation of egg production? How do the success rates of low-dose stimulated IUI and IVF with the transfer of a single embryo compare? How do the safety and the success rates of conventional IVF and IVF with injection of a sperm into the egg (intracytoplasmic sperm injection, ICSI) performed in couples with the same degree of male infertility compare? How do the safety and success rates of ART using fresh and frozen embryos compare to ART using only fresh embryos? A report was produced on the basis of the review and expert opinion from the ART Review Committee. Results Maternal age The review identified nine sources of evidence which were considered to provide information on the effect of a woman s age on the success rates of ARTs. The reviewed evidence provided a large quantity of data which consistently demonstrated declining success rates with increasing maternal age. There was evidence that this effect was maintained after the effects of a number of other factors that influence success rates were taken into account. This included, but was not limited to, whether or not the woman had had children previously, the number of previous IVF attempts, the number of years of smoking, the cause of infertility and the duration of infertility. The decrease in the success rates of ART for women over 40 years of age was marked. Two reports from the United Kingdom and the United States indicated that live birth rates for women over Assisted reproductive technologies xii

15 43 were less than 5 per cent. Australian data from 2002 indicated that the live birth rate for women aged 40 to 44 was approximately 6 per cent, compared to 25 per cent for women aged 30 to 34. Low-quality evidence from one study indicated that an increase in maternal age was also associated with a decrease in the success rates of IUI. Cycle number Six sources of information were reviewed to determine the success rates of ART after increasing numbers of treatment cycles. The type of evidence identified was considered poorer than that available for investigating the effect of age. There was evidence for a general trend of decreasing birth rates with increasing numbers of successive cycles. Only one study demonstrated that this effect was maintained after adjusting for other factors known to affect ART success rates (including age). However, the effects of the cause or duration of infertility were not taken into account. Australian data indicated that the effectiveness of IVF significantly decreased in cycles 2 and 3 by comparison with the first cycle. However, the reason for this relationship is not known. It may be due to other factors, such as maternal age. The cycle number at which success rates decreased was inconsistent between the studies reviewed. No studies provided evidence of the success rates of IUI in different cycle numbers. Duration of infertility The review identified two sources of evidence which indicated that the success of IVF is reduced in women with a longer duration of infertility. The size of this effect appeared to be smaller than that associated with increasing maternal age. However, the limited amount of evidence available means that this conclusion is uncertain. Intrauterine insemination (IUI) Only one study provided evidence of the comparative success rates of low-dose stimulated IUI and IVF. In this study up to two embryos were transferred to the woman during IVF. This study indicated that IVF was more successful than IUI, even though it was likely that the patients having IUI had worse male fertility than those in the IVF group. Since there is very little evidence for this comparison, whether or not this finding reflects what would be observed in a general population in Australian practice is uncertain. Intracytoplasmic sperm injection (ICSI) Intracytoplasmic sperm injection (ICSI) is an ART involving the injection of a single sperm into the cytoplasm of an egg to achieve fertilisation. ICSI is used in couples with male factor infertility and those with poor fertilisation with conventional IVF. For couples with severe male factor infertility ICSI is the only viable treatment option. The review identified only two studies reporting the success rates of ICSI and IVF in couples with a similar level of male factor infertility, although it appeared likely that the groups were not the same. These two studies reported conflicting results. High-quality research comparing the success rates of these treatments in couples with mild male infertility is required. Concern has been expressed about whether the non-natural selection of sperm used with the ICSI technique, or the physical damage to the egg results in poorer outcomes in the children conceived in this manner. Eleven reports provided evidence on the rates of xiii Assisted reproductive technologies

16 malformations present at birth between IVF and ICSI children. All of these studies have limitations in their design for determining the answer to this question. Higher quality studies did not indicate any significant differences in the congenital malformation rates in ICSI verses IVF children. Seven reports demonstrated no difference in the overall chance of malformations in IVF or ICSI children. One large Swedish study indicated a higher rate of malformations of the urogenital tract in the ICSI children. One other smaller study indicated a higher rate of stillbirths in ICSI twins, compared with IVF twins. Another report which combined the results of 10 studies indicated a possible increase in the major malformation rate in ICSI children, but this effect was not seen when only the four highest quality studies were analysed. Whether or not the differences seen in the rate of malformations with ICSI and IVF in some studies is real is uncertain at this time as the evidence is conflicting. A larger quantity of high-quality research is required to determine whether or not the use of ICSI is associated with different safety or effectiveness outcomes than IVF. ART cycles using fresh and frozen embryo transfer Four studies provided evidence on the comparison of ART using frozen embryo transfer versus ART with fresh embryos only. One study of high-quality indicated that a policy of elective single embryo transfer (SET), with transfer of a frozen embryo when necessary, will reduce the rate of multiple pregnancies. All of the reviewed studies indicated that this might occur without severely compromising the success rates of ART. Economic considerations An economic analysis based on information about IVF success rates for all women undergoing IVF in Australia and New Zealand in 2002 was conducted using data provided by the Australian Institute of Health and Welfare National Perinatal Statistics Unit (AIHW NPSU). This analysis estimated the incremental costs and effects of IVF treatment by age and by number of IVF attempts (treatment program number) using fresh and frozen embryo transfers. This analysis included costs of IVF treatment as well as the costs of pregnancy and all of the related pregnancy outcomes in line with standard international approaches to economic analysis. The results support previous findings that the cost of a live birth increases with maternal age. The findings also suggest that within younger (30 33 year) and older (42 45 year) age groups the cost per live birth increases with a greater number of treatment programs. Estimates based on Medicare costs and other clinic fees show that the cost for each live birth for women aged 30 to 33 years was an additional $27,373 in the first treatment program, $30,098 in the second treatment program and $31,836 in the third treatment program. The cost for each live birth for women aged 34 to 37 and 38 to 41 years was an additional $32,604 and $51,680 per treatment program, respectively. For women aged 42 to 45 years the cost per live birth increased from an additional $130,951 for the first treatment program to $187,515 for the second treatment program. The inclusion of costs associated with the pregnancy outcomes in the analysis narrowed the difference between the cost per live birth for the younger and older age groups compared to a study that did not include these costs (Chambers et al., 2005). This is due to the higher pregnancy rates from IVF treatment in the younger age groups. The differences between younger and older age groups would be expected to narrow further if the postnatal costs of multiple births were included in the analysis, as there is a higher Assisted reproductive technologies xiv

17 multiple birth rate in the younger women. Further research is required on the economic costs to the health system of multiple births to establish the true cost-effectiveness of IVF therapy by age and number of treatment programs. xv Assisted reproductive technologies

18

19 Introduction The Commonwealth Government set up an independent review of assisted reproductive technologies (ART) on 4 July 2005 to advise the Commonwealth Minister for Health and Ageing on the clinical and cost-effectiveness of ART for the purposes of public funding under the Medicare Benefits Schedule. As part of this review, the National Health and Medical Research Council (NHMRC) Clinical Trials Centre has been contracted to undertake a systematic review of the evidence to determine the clinical effectiveness of ART and an economic evaluation to determine the incremental cost-effectiveness of different ART techniques. This report updates information provided in the Australian Health Technology Advisory Committee s (AHTAC) 1998 report, Review of Assisted Reproductive Technology (AHTAC, 1998). The primary effectiveness outcome considered in this review is the live birth rate per ART treatment cycle started. An ART treatment cycle is considered as a treatment episode commencing with the first patient intervention (usually controlled ovarian hyperstimulation). An ART cycle using unstimulated ovulation is considered as commencing with the provision of ovulation monitoring services. Assisted reproductive technologies 1

20 Background Assisted reproductive technologies Assisted reproductive technology (ART) according to the WHO definition refers to infertility treatments where both eggs (oocytes) and sperm are handled to achieve a live birth. In conventional ART procedures the oocytes and sperm are combined in a laboratory (in vitro) and the resulting embryo is transferred to a woman. ARTs include in-vitro fertilisation (IVF) with or without injection of a sperm into the oocyte (intracytoplasmic sperm injection, ICSI). Using this definition, some treatments for infertility such as ovulation induction with timed intercourse and artificial insemination, while being a treatment for infertility, are not therefore classified as an ART. However, in this report, the terminology assisted reproductive technologies will be used to refer to ARTs as defined by the WHO, ovulation induction, and artificial insemination inclusively. ART involves a series of interventions usually commencing with the use of hormones in the female partner to stimulate the ovaries to produce mature oocytes. In ICSI and IVF these oocytes are collected by needle aspiration and combined with sperm from the male partner in vitro. IVF/ICSI conventionally involves transfer of the embryo to the women s uterus at the cleavage stage of embryo development (usually day 2 or 3). Embryo transfer may be performed with fresh embryos, or embryos may be frozen (cryopreserved) for thawing and transfer at a later time. Measurement of ART effectiveness Estimates of ART effectiveness may vary according to the definitions of treatment success and the treated population. Live birth rates provide the most clinically relevant measure of ART success. Other outcomes commonly reported in the literature include fertilisation, implantation and pregnancy rates. Pregnancy may be defined as biochemical pregnancy (elevated human chorionic gonadotrophin), clinical pregnancy (identification of a gestational sac in the uterus), on-going clinical pregnancy (foetal heart activity detected at ultrasonography at 12 weeks), or on-going pregnancy (pregnancy beyond 20 weeks). These intermediate outcomes will not necessarily reflect the desired outcome of a live birth. The European Society of Human Reproduction and Embryology consensus document (2003) recommends that the singleton live birth rate is a most appropriate measure of ART success, given the increased risks to mother and infant associated with multiple pregnancy (Land & Evers, 2003). In Australia, the Fertility Society of Australia s (FSA) Code of Practice states that the objective of ART must be the live birth of a single healthy child (FSA Reproductive Technology Accreditation Committee, RTAC, 2005). The outcome of a single, term gestation, live baby has been referred to as birth emphasizing a successful singleton at term (BESST) (Healy, 2004; Min et al., 2004). In Australia, overall live birth rates following ART are recorded as well as the singleton and multiple live birth rates (Bryant et al., 2004). The definition of the treated population has an impact on the calculation of the live birth rate. The inclusion of all women commencing treatment for the purpose of ART as the treated population is recommended (Bryant et al., 2004; Human Fertilisation and 2 Assisted reproductive technologies

21 Embryology Authority, HFEA, 2003; National Center for Chronic Disease Prevention and Health Promotion et al., 2002). Studies that only report ART live birth rates per oocyte recovery or per embryo transfer cycle exclude women with failed or cancelled controlled ovarian hyper-stimulation (COH) and thus may overestimate the true effects of ART. Defining an IVF/ICSI treatment as the initiation of treatment with COH allows inclusion of the results of the transfer of an initial fresh embryo and any subsequent frozen embryos stored after the initial oocyte retrieval within the same treatment cycle. The primary effectiveness outcome considered in this review is the live birth rate per ART treatment cycle started. An ART treatment cycle is considered as a treatment episode commencing with the first patient intervention (usually COH). An ART cycle using unstimulated ovulation is considered as commencing with the provision of ovulation monitoring services. Infertility Infertility is defined as the inability to conceive after 12 months of unprotected and regular sexual intercourse. Observational studies conducted in couples using natural methods to conceive have indicated that approximately 80 per cent of couples will conceive in the first six menstrual cycles and an additional 10 per cent will conceive spontaneously within the next six cycles (Gnoth et al., 2005). Of the 10 per cent of couples classified as infertile after one year of trying to conceive, approximately half will achieve a spontaneous conception over the next 36 months (Gnoth et al., 2005). The FSA estimates that one in six Australian couples suffer from infertility, which they define as the inability to conceive after 12 months or to carry pregnancies to a live birth (FSA, 2005). Aetiology A couple may be infertile due to impaired male or female fertility or a combination of factors from both partners. The dominant cause of infertility has been attributed to male factors in up to 30 per cent of couples, female factors in up to 37 per cent of couples and both partners in 20 to 35 per cent of couples (Boyle et al., 2004a). Standard practice in Australia is to undertake a full medical and sexual history and medical examination of both partners when couples present after failing to conceive after 12 months of trying. Investigation is then offered to both partners. The purpose of these initial routine investigations is to make a diagnosis, guide management, estimate prognosis or identify patients requiring further specific investigation (Evers, 2002). Many of the causes of infertility are reversible and are managed by non-art treatments. In 5 to 15 percent of couples no cause is identified and a diagnosis of unexplained infertility is made. Such couples are eligible for treatment using assisted reproductive technologies (ART). The common causes of infertility are described in the following. These can be classified under five headings based on their different aetiology, prognosis and non-art treatment options: male subfertility; ovulation dysfunction; tuboperitoneal disorders; defects in sperm-cervical mucus interaction; and unexplained (Evers, 2002). Assisted reproductive technologies 3

22 Male factor infertility Routine investigation of male partners includes a detailed history, physical examination, semen analysis and hormone measurement. The semen analysis is fundamental to this assessment. A finding of no abnormality (normospermia) indicates that spermatogenesis (sperm production) is normal and the reproductive tract is patent. At the other extreme, a finding of no sperm (azoospermia) confirms male infertility. Findings between these extremes such as low sperm count (oligospermia); low sperm motility (asthenozoospermia); abnormal sperm forms (teratozoospermic) and combinations of these abnormalities are associated with male subfertility, but may only be one factor contributing to a couple s inability to conceive. WHO has defined reference values for sperm count, motility and morphology; however, males classified with abnormal parameters according to these guidelines may still be able to achieve conception naturally and thus the predictive value of semen analysis findings other than azoospermia are limited (WHO, 2000). In Australia, severe male infertility is commonly defined as the presence of a sperm density <5 x 10 6 per ml, or total motile sperm count (TMC) <10 x 10 6 (P. Illingworth, 2005, pers. comm., 16 September). Abnormal sperm parameters may be due to problems with the production of sperm due to abnormal hormone production or testicular failure, or to problems with sperm transport from the testes to ejaculation due to obstruction of the genital tract (see following list). In around one quarter of infertile men, no underlying cause is detected (National Collaborating Centre for Women's and Children's Health, NCCWCH, 2004). Among cases where an underlying diagnosis can be made, varicoceles (dilation of the veins that drain the testis thought to impair fertility due to increased intratesticular temperature) and genital tract obstruction are most common (38% and 13 per cent of diagnoses at male infertility clinics, respectively) (Boyle et al., 2004a). Both these problems can be treated by surgery. Less common causes such as sperm autoimmunity, gonadotrophin deficiency and other endocrine abnormalities are amenable to medical treatment. Aetiology of abnormal sperm parameters Problems with sperm production Abnormal testicular development or function - primary testicular failure due to cryptorchidism, chromosome disorders such as Klinefelter s syndrome - secondary testicular failure due to varicocele, testicular torsion, trauma, orchitis, systemic disease, radiotherapy, chemotherapy Hypothalamic/pituitary failure - congenital deficiencies of gonadotrophins, leutenising hormone (LH), follicle stimulating hormone (FSH), other syndromes - pituitary insufficiency, hyperprolactinaemia, haemochromatosis - other specific endocrine abnormalities Disorders of sperm motility or function - congenital defects of the sperm tail, maturation defects - anti-sperm antibodies - infection 4 Assisted reproductive technologies

23 Disorders of sperm transport vasectomy infection of epididymis, vas deferens neuropathic, eg spinal cord injury Sexual dysfunction. Female factors Routine investigation of female partners includes a detailed history, physical examination and investigation with hormone testing to assess ovulatory function; chlamydia screening to indicate potential tubal damage; and hysterosalpingography to assess tubal patency. Ovulatory dysfunction Identification of ovulatory failure or dysfunction is made by taking a menstrual cycle history and measuring the serum concentrations of two pituitary hormones involved in the regulation of ovulation: follicle stimulating hormone (FSH) and prolactin; and oestradiol produced by the ovary (Evers, 2002). The results of these tests are used to classify the aetiology and site of the underlying hormone disorder and determine appropriate treatment to correct the underlying cause. For example, women with a low concentration of FSH due to hypothalamic failure can be treated with gonadotropinreleasing hormones; and those with a high concentration of prolactin can be treated with dopamine antagonists and/or treatment of the underlying cause. Women with normal hormone levels are classified as having a disturbance of the hypothalamic-pituitaryovarian axis that determines hormone regulation. Treatment for this subgroup involves the use of ovarian stimulation drugs. No treatment is available to restore ovulation in women with premature ovarian failure (associated with a low concentration of oestradiol despite high levels of FSH). This may occur as a primary problem or secondary to chemotherapy, radiotherapy or other iatrogenic factors. Oocyte donation can be offered to achieve a pregnancy in these women. Tuboperitoneal factors Tubal obstruction and other distortions of the pelvic anatomy can prevent natural fertilisation of the egg and sperm and transport of the embryo to the uterus for implantation. Causes include post-infectious tissue damage, post-surgical adhesions, endometriosis (the growth of endometrial tissue in areas outside the uterus) and developmental abnormalities of the uterus. Diagnosis is suggested by a history of pelvic inflammatory disease or pelvic surgery, chronic pelvic pain and clinical examination findings. A blood test to detect past exposure to the sexually transmitted disease Chlamydia trachomatis is also undertaken to assess the possibility of tubal damage. Hysterosalpingography is routinely conducted to assess tubal patency, although its value as a standard investigation has been questioned (Evers, 2002). Treatment is surgical and is commonly performed laparoscopically. ART offers an alternative for patients who remain infertile after corrective surgery, or for those who require complex microsurgical procedures. Defects in sperm-mucus interaction A diagnosis of defects in sperm mucus interaction or cervical hostility is based on the finding of normospermia in the male partner with a well-timed post-coital test that shows Assisted reproductive technologies 5

24 sperm present in the vaginal vault but no moving sperm in the cervical mucus. The validity of the post-coital test has been questioned and this diagnosis is not universally accepted; however, it has been shown to be a useful prognostic indicator in couples without a long history of infertility where known causes of female infertility are absent (Evers, 2002). A negative post-coital test may also indicate the presence of anti-sperm antibodies produced by either the male or female partner which are known to interfere with sperm motility and egg fertilisation. Anti-sperm antibodies are generally detected in men by direct semen examination (prevalence up to 10% of infertile men, higher rates in those who have had surgery on their reproductive tract) and by blood test in females (prevalence less than 5% of women). Systematic reviews have indicated that couples with a negative postcoital test will benefit from IUI (Boyle et al., 2004a). Relative frequency of different underlying causes The relative frequency of each diagnosis varies across different studies. The findings of studies based on couples presenting to ART clinics does not reflect the distribution of infertility causes in all couples presenting with, or treated for, infertility. However, this review is only concerned with the value of ART for infertile couples opting for ART who are not eligible for non-art treatments. The principal causes of infertility in this subgroup of couples can be classified as male or female factors that are not amenable to treatment or where treatment has failed (eg non-obstructive azoospermia, severe tubal disease and severe endometriosis) and unexplained infertility. In 2002, the Australian and New Zealand ART database recorded 19,883 fresh, nondonor ART treatment cycles. The causes attributed to infertility in this population were multiple factors (31.3%), male factor only (26.4%), unexplained (16.7%), female factor only (16.0%: tubal disease 10.0%, endometriosis 6.0%) and other (9.7%). Risk factors The National Institute of Clinical Effectiveness (NICE) review provides detailed guidelines based upon the evidence available for known risk factors for infertility (NCCWCH, 2004). These factors are outlined briefly in the following. Female age is the major determinant of infertility. Natural female fertility falls gradually after age 30 years, with a rapid decline after age 35 years to cessation at menopause. The predictive value of female age on the success rates of reproductive services is a major focus of this review and is addressed in detail in the results section of this report (see page 31). Other factors associated with female infertility are obesity (body mass index greater than 29) and low body weight (body mass index less than 19 and irregular or absent menstruation) and smoking. Obesity has also been associated with male infertility. Excessive alcohol intake, smoking and elevated scrotal temperature due to sedentary work position, occupational heat exposure and wearing tight underwear has been associated with reduced semen quality in men, although the impact of this on male fertility is not known (NCCWCH, 2004). Prescription and recreational drugs and occupational hazards such as exposure to solvents have been associated with infertility in males and females. 6 Assisted reproductive technologies

25 Infertility treatments In most couples presenting with infertility, treatment is directed at reversing the underlying cause. As described previously in the aetiology section, this may involve surgical measures to treat genital tract obstruction, varicoceles or endometriosis, or hormone treatment to restore ovulatory function. ART procedures in Australia are reserved for couples where the underlying cause is not amenable to other treatments, or is unexplained. ART procedures IVF interventions IVF involves a series of interventions. The four fundamental steps are: COH and oocyte retrieval; sperm retrieval and preparation; in-vitro fertilisation; and embryo transfer. These are briefly described in the following based on a recent review by Boyle et al. (2004b). Ovarian stimulation and retrieval In the normal 28-day menstrual cycle the ovary develops around 10 oocyte-containing follicles, of which one develops to maturity under the complex physiological regulation of pituitary hormones (FSH and luteinizing hormone, H) and ovarian hormones (oestrogen and progesterone) with release of the oocyte (ovulation) around day 14 of the cycle. Controlled ovarian hyperstimulation refers to the use of drugs to induce the development of multiple dominant follicles in women with normal menstrual cycles undergoing ART (Fauser et al., 2005). Ovarian stimulation is also used to refer to the use of drugs to induce ovulation in anovulatory women (ovarian induction) who are not being treated with ART. Most ART includes COH to allow the harvest of multiple oocytes. The availability of multiple oocytes improves the efficiency of in-vitro fertilisation and consequently the number of embryos available for transfer. The development of multiple embryos also permits the selective transfer of high-quality embryos. Research into the most effective means of assessing embryo quality is ongoing (reviewed by Sakkas & Gardner, 2005). The drugs used for COH act by raising and sustaining high FSH concentrations to allow multiple follicles to mature rather than one single dominant follicle. These drugs include clomiphene citrate, an anti-oestrogen agent that increases the release of FSH from the pituitary, and exogenous hormone supplementation obtained from the urine of postmenopausal women (human menopausal gonadotropin, HMG) or recombinant human FSH produced in the laboratory (rfsh) (Fauser et al., 2005). Additional treatment with gonadotrophin-releasing hormone (GnRH) analogues to facilitate cycle control is also standard. Follicular development is monitored using serum oestradiol levels and transvaginal ultrasound. When the follicles reach a sufficient size, usually after 9 to 11 days, human chorionic gonadotropin (HCG) is administered to induce ovulation. Retrieval of oocytes is undertaken within 32 to 36 hours of administering HCG and before spontaneous ovulation. This can be performed transvaginally with sedation by passing a needle into the ovary under ultrasound guidance, or by laparoscopy under general anaesthetic. Assisted reproductive technologies 7

26 In women with unsuccessful oocyte retrieval, the use of donor oocytes can be offered; however, the question of the effectiveness of ART using donor oocytes, sperm or embryos is beyond the scope of this review. Sperm retrieval Fresh sperm are retrieved from ejaculate provided by the male partner on the same day as the oocyte retrieval, although frozen thawed sperm can also be used. If the male partner is azoospermic, percutaneous or open biopsy techniques may be used to retrieve sperm from the testes or epididymus. Percutaneous methods involve the needle aspiration of sperm from the testis, epididymus or vas. It can be performed in an outpatient setting under local or regional anaesthesia and takes approximately 30 minutes. Open biopsy of the testis is a more invasive technique performed in an operating theatre with a general anaesthetic. It is associated with a higher risk of surgical and anaesthetic complications, but provides a greater volume of tissue, which may be needed in men with severe sperm defects or may be desired to allow storage for later ART attempts. In-vitro fertilisation Sperm are washed, spun in a centrifuge and incubated in a specialised medium in preparation for fertilisation. Retrieved oocytes are separated from follicular fluid and classified to identify mature oocytes suitable for fertilisation. The selected samples of sperm and oocytes are combined in a culture medium in a Petri dish and inspected for fertilisation after 16 to 20 hours. Oocytes that are not fertilised or that have fertilised abnormally are discarded. At fertilisation, the resulting cell contains a pronucleus from both the oocyte and sperm (referred to as gametes) and is known as a pronuclear zygote. The two pronuclei then fuse and cell division commences, reaching a four-cell stage at approximately 40 hours. Embryos are inspected at this stage to assess viability. Embryo transfer Embryo transfer takes place as an outpatient procedure, conventionally two to three days after fertilisation when the embryo is at the cleavage (4 8 cell) stage (or after 4 6 days if transferred at the blastocyst stage). In Australia, a grading system based on embryo morphology is commonly used to select the healthiest embryos for injection into the uterus transvaginally. After embryo transfer, the female is treated with progesterone daily for up to the 10th week of pregnancy (or alternative regimens) to assist implantation and maintenance of pregnancy. The number of embryos transferred depends on the practice of the provider and the age and preferences of the treated couple. In Australia, 94 per cent of embryo transfers undertaken in 2002 involved single or double embryos (Bryant et al., 2004). Rates of single, double and triple/higher embryo transfer rates were 29, 68 and 3 per cent, respectively, for women less than 38 years; and 28, 60 and 11 per cent, respectively, for women aged 38 years or older (Bryant et al., 2004). Any remaining healthy embryos may be frozen for storage (cryopreservation) for thawing and transfer at a later date if needed. 8 Assisted reproductive technologies

27 Blastocyst transfer Keeping embryos in in-vitro culture for a longer period of time before uterine transfer is now possible due to improvements in laboratory techniques. The transfer of embryos to the uterus at a later stage of development may mimic natural physiology more closely and also allow better selection of embryos for transfer. Much research on this approach has been conducted in the last decade, but with conflicting results. A recently updated Cochrane review has investigated the relative success of cleavage stage (day 2 3) versus blastocyst stage (day 5 6) embryo transfer (Blake et al., 2005). This high-quality review searched literature to May 2005 for high-quality trials (randomised controlled trials, RCTs) reporting live birth, clinical pregnancy or multiple pregnancy rates. The review identified 16 included trials published over the period 1998 to 2004 comparing cleavage to blastocyst stage transfer. Quasi-randomised studies were excluded in addition to a single study from 1993 using outdated culture techniques. Across seven RCTs, the live birth rates resulting from day 2/3 or day 5/6 embryo transfer did not differ (odds ratio, OR 1.16, 95% CI ). Data from 15 RCTs demonstrated no difference in clinical pregnancy rates (OR 1.05, 95% CI ). Multiple pregnancy rates did not differ between cleavage stage and blastocyst stage embryo transfer (12 RCTs, OR 0.85, 95% CI ). No differences were demonstrated for higher order multiple pregnancy rates or miscarriage rates (across 5 RCTs and 9 RCTs, respectively). No differences in any outcome were observed in subgroups of couples with good, poor or unselected prognosis. There were also no differences in outcomes observed in subgroups of studies with equal numbers of embryos transferred or in those with more cleavage stage than blastocyst embryos transferred. These results were unchanged in a sensitivity analysis excluding studies where different media were used for the two study arms. The authors of this review concluded that There is no difference in live birth or pregnancy outcomes between day 2 to 3 and day 5 to 6 transfer of embryos. Blastocyst transfer was associated with an increase in failure to transfer any embryos in a cycle and a decrease in embryo freezing rates. In the absence of data on cumulative live birth rates resulting from fresh and thawed cycles, it is not possible to determine if this represents an advantage or disadvantage. In this review, many of the studies identified were transferring two to three embryos. The only study which transferred a single blastocyst transferred two cleavage stage embryos. Further research comparing the outcomes from transfer of a single cleavage stage or blastocyst stage embryo is required. In addition, research into the best patient selection criteria for blastocyst culture is continuing as media composition and culture techniques improve. Cryopreservation and frozen embryo transfer The ability to preserve embryos by freezing (cryopreservation) for subsequent frozen thawed embryo transfer in unstimulated treatment cycles has the potential benefits of increasing the number of embryo replacement cycles without additional controlled ovarian hyperstimulation and oocyte retrieval. Although the live birth rate per embryo transfer is lower for frozen than fresh cycles (14.8% versus 23.5% in Australia in 2002, Bryant et al., 2004), this improves the overall pregnancy rate per treatment cycle, at the same time reducing the risk of ovarian hyperstimulation syndrome (OHSS). A recent cost-effectiveness analysis from Sweden compared a single embryo transfer (SET) strategy with transfer of one additional frozen thawed embryo if necessary to a double embryo transfer (DET) strategy (Kjellberg et al., 2006). This analysis indicated Assisted reproductive technologies 9

28 that, when maternal and paediatric complications were considered, SET was the more cost-effective approach. In Australia, 31,253 ART treatment cycles involving embryo transfer were recorded with non-donor oocytes or embryos in Of these 64 per cent involved fresh embryo transfers(bryant et al., 2004). Data from 1994 to 2002 show increasing numbers of treatment cycles resulting in the storage and thawing of frozen embryos; however, data about the average number of fresh and frozen embryo transfers resulting from one COH cycle with oocyte retrieval have not been published. Other technological advances to improve the safety and effectiveness of COH regimens, and new techniques for embryo culture, selection and transfer have contributed to the increased success rates of ART over recent times. These techniques include milder COH regimens (Pruksananonda et al., 2004; Vlaisavljevic et al., 2003), new culture media (Ben Yosef et al., 2004), more detailed assessment of embryo morphology prior to transfer (Rossi-Ferragut et al., 2003), and ultrasound-guided embryo transfer (Buckett, 2003; Sallam & Sadek, 2003). Other new techniques such as assisted hatching (a technique to improve embryo implantation) (Edi-Osagie et al., 2003; Sallam et al., 2003) and preimplantation genetic diagnosis for couples with known genetic diseases (Verlinsky et al., 2000) or for aneuploidy testing (Kahraman et al., 2004; Kuliev et al., 2003) offer advantages for specific indications. Research is ongoing to determine optimal therapies and techniques. Intracytoplasmic sperm injection Intracytoplasmic sperm injection (ICSI) is an ART involving the injection of a single sperm into the cytoplasm of an oocyte to achieve fertilisation. It is indicated for the treatment of couples with male factor infertility and those with poor fertilisation with conventional IVF, although some have recommended its broad use as first-line ART treatment (Abu-Hassan et al., 2003). ICSI is the only treatment option for couples with severe male factor infertility. It can be performed with ejaculated or surgically retrieved sperm. As described previously for IVF, oocytes are examined after 16 hours for fertilisation, and viable embryos are transferred 1 to 3 days later (Boyle, 2004). Concerns have been raised that ICSI may be associated with an increased risk of congenital malformations and long-term genetic consequences due to the ability to produce embryos using abnormal sperm that would not otherwise be able to achieve fertilisation, including sperm from males with genetic defects (Boyle, 2004). This issue is discussed on page 17 (Other antenatal complications). Gamete intrafallopian transfer Gamete intrafallopian transfer (GIFT) involves the laparoscopic aspiration of follicles and simultaneous transfer of oocytes and sperm into the fallopian tubes. Originally developed as a more physiological alternative to IVF in women with a high chance of success, the use of GIFT has declined in the last decade as the effectiveness of IVF techniques have improved. GIFT is now rarely performed in Australia (Bryant et al., 2004). 10 Assisted reproductive technologies

29 Artificial insemination procedures Artificial insemination (AI) refers to procedures involving the insertion of sperm from the male partner into the vaginal vault, cervix (intracervical insemination) or uterus (intrauterine insemination, IUI, also see following) of the female partner. It may be used as a less invasive alternative to IVF or ICSI for the initial treatment of mild male factor infertility or unexplained infertility with or without ovarian stimulation (NCCWCH, 2004). Donor insemination (DI) using donor sperm may be the only treatment option available to couples with azoospermia, male genetic disorders or infectious diseases, other co-morbidities or sexual factors preventing natural insemination and conception. Intrauterine insemination Intrauterine insemination (IUI) involves the placement of washed sperm into the uterus under ultrasound guidance to bypass the natural cervical mucus barrier. It is performed under sedation as an outpatient procedure with or without controlled ovarian hyperstimulation (COH). It is designed to bring a high concentration of sperm into close contact with one oocyte (after natural ovulation) or with multiple oocytes (after COH). The main indications for IUI with COH are in the treatment of unexplained infertility where investigations have excluded an obstructive cause (at least one open fallopian tube) and severe male factor infertility. Advantages are that the procedure is less invasive and better tolerated than IVF. Disadvantages are that the procedure has been associated with a lower success rate and higher multiple pregnancy rate than IVF. However, the use of low-dose COH regimens with abandonment of insemination when more than three dominant follicles develop may be expected to reduce the latter. In Australia, best practice involves ovarian stimulation with low-dose clomiphene citrate (usually mg daily) or FSH alone (usually IU daily), with IUI treatment only when one or two dominant follicles are present on the day of HCG administration. A recent report of 510 cycles of IUI with COH in an Australian IVF clinic described a live birth rate of 7 per cent and multiple birth rate of 1 per cent of IUI cycles started (Healy et al., 2003). In this study cycles were stopped if more than three follicles greater than or equal to 14mm diameter developed. One triplet pregnancy (in 0.2% of cycles) occurred. Clinical need In developed nations throughout the world, the total fertility rate has been declining and is currently below the population replacement value of 2.1 births per woman in many countries. In Australia, the total fertility rate in 2002 was 1.76, which is comparable to that in the United Kingdom, United States and Canada (Ford et al., 2005). It is estimated that 12 to 25 per cent of couples are affected by infertility, but the number of couples seeking medical advice is not accurately known (AHTAC, 1998). The FSA estimates that one in six Australian couples suffer from infertility, which they define as the inability to conceive after 12 months or to carry pregnancies to a live birth (FSA, 2005); however, there is no national data collection on infertility in Australia. A recent population-based telephone survey of Australian men (MATeS) found a self-reported failure to conceive of 7.6 per cent (Holden et al., 2005). In developed nations, as an increasing number of women delay having children until a later age, the prevalence of infertility and need for fertility assistance services is likely to Assisted reproductive technologies 11

30 increase. The median age at child-bearing in Australian women increased from 26.8 years in 1982 to 28.7 years in 1992, then to 30.2 years in 2002 (Australian Bureau of Statistics, ABS, 2005c). The average age of women bearing their first child is also increasing, from 26.2 years in 1993, to 27.6 years in 2003 (AIHW NPSU, 2005a; AIHW NPSU, 2005b). The age-specific fertility rates of women in the age groups 20 to 24 and 25 to 29 years have been decreasing, along with an increase in the age-specific fertility rates of women aged 35 to 39 and 40 to 44 years. Over the 10-year period from 1992 to 2002, agespecific fertility rates in women aged 20 to 24 and 25 to 29 fell from 75 to 57 births per 1000 and 132 to 105 births per 1000, respectively (ABS, 2005a). Over the same period, the fertility rates of women aged 35 to 39 and 40 to 44 years increased from 38 to 52 births per 1000 and from 6 to 10 births per 1000, respectively. Data from the Netherlands demonstrate that the age of women registering for their first infertility consultation is increasing alongside the increasing age at first birth (Snick et al., 2005). As shown in Figure 1, the use of ART services in Australia and New Zealand has also been steadily increasing over the same decade. The average age of women undergoing ART treatment in Australia and New Zealand in 2002 was 35.2 years (Bryant et al., 2004). Figure 1 Treatment cycles started, pregnancies and live births, Australia and New Zealand, Extracted from Bryant et al. (2004) In Australia, there is currently no national legislation covering the regulation of ART clinical practice. Legislation regarding ARTs is determined individually by each state and territory; therefore, laws and practices regarding ART services, including access, differ by region. Victoria, South Australia, and Western Australia have legislation that regulates access to ART services, including IVF. Victorian and South Australian laws restrict 12 Assisted reproductive technologies

31 access to IVF treatment to women who are married or in a de facto relationship with a man. Western Australia recently amended its Act to remove these restrictions. In July 2000, the Federal Court ruled that the Victorian Fertility Treatment Act 1995 was inconsistent with Section 22 of the Sex Discrimination Act, which prohibits discrimination in the provision of goods and services on the grounds of sex or marital status. The ruling was in response to a challenge to the Victorian Act lodged in the Federal Court by a doctor practising in Victoria on behalf of a single woman who wished to access IVF treatment. The Federal Court decision rendered invalid the restrictions in the Victorian Act and effectively any such restrictions in the legislation of other jurisdictions. There is little available information/statistics on same sex couples and single women accessing reproductive services. Reimbursement through the MBS is dependent upon the presence of a medical condition determining a clinical need and not dependent upon partner status (see page 20). Current use of ART in Australia As discussed, the main indications for ART are unexplained infertility, male factor infertility, and female ovulatory or tuboperitoneal disorders that are not amenable to treatment or where treatment has failed. Figure 2 illustrates the current use of ART in managing infertility in Australia. It has been used to identify appropriate comparators for this review. Infertility Failure to conceive after regular unprotected sexual intercourse Male factor Female factor Hypogonadotrophic hypogonadism: - gonadotrophins Obstructive azoospermia - surgery - sperm recovery Azoospermia Unknown cause (normal semen, no ovulation disorders, no tubal occlusion) Treatment Ovarian failure Severe Mild Donor insemination ICSI Stimulated IUI Unstimulated IUI IVF Oocyte donation No conception Fertilisation failed Figure 2 Clinical flowchart to illustrate current use of ART and IUI in Australia. Based on existing guidelines (NCCWCH, 2004) and information provided by the review committee. Assisted reproductive technologies 13

32 Profile of ART in Australia The first live birth resulting from IVF worldwide was in IVF is now a wellestablished treatment for infertility in developed countries. Advances in ART, including the use of donor gametes and ICSI, have increased the number of couples who have been treated with ART in recent years. ART accounts for 1 to 3 per cent of annual births in western developed countries (Gosden et al., 2003). In 2002, 2.3 per cent of babies born in Australia were conceived following the use of ART (Laws & Sullivan, 2004). Data from the Australian and New Zealand Assisted Reproduction Database (ANZARD) indicate that the use of ART has been steadily increasing over the period from 1993 to 2002 (Bryant et al., 2004) ( Figure 1). In this 10-year period the use of standard IVF has not increased markedly, but significant increases in the use of ICSI and frozen embryo transfer have occurred (Figure 3). Over the same period, the use of gamete intrafallopian transfer (GIFT) has virtually disappeared. There has also been a constant decline in the number of embryos transferred. In 1993, 55.3 per cent of treatment cycles transferred three or more embryos. In 2002 this had decreased to 5.8 per cent of cycles. It is now usual practice in Australia and New Zealand to transfer no more than two embryos. Figure 3 Number of transfer cycles by ART treatment type, Australia and New Zealand, Extracted from Bryant et al. (2004) 14 Assisted reproductive technologies

33 Over this same time period, the success of ART in Australia and New Zealand has steadily increased (Bryant et al., 2004). The number of pregnancies per fresh embryo transfer cycle was 30.5 per cent in 2002, by comparison with 16.3 per cent in The pregnancy rate per frozen embryo transfer cycle increased from 13.5 to 19.9 per cent in the same 10-year period. In 2002, 18.3 per cent of all fresh, non-donor cycles started resulted in a live delivery. The success rate as determined by the BESST outcome for full-term, singleton babies was 12.9 per cent of all fresh, non-donor stimulated cycles. Safety of ART Research about the safety of ART compared to spontaneous pregnancies is based on observational studies that compare outcomes for couples achieving pregnancy with ART with those achieving pregnancy spontaneously. As outlined in the following, these studies show that ART procedures are associated with greater health risks for the mother and child than spontaneous pregnancies. These risks are largely associated with the use of controlled ovarian hyperstimulation regimens and multiple embryo transfers in ART, for example, OHSS, and complications due to multiple pregnancies. Other differences observed in antenatal and perinatal outcomes may reflect differences between fertile and subfertile couples, such as maternal age and paternal sperm abnormalities, rather than the independent effects of ART. However, research is ongoing to confirm or exclude an association between ART, in particular ICSI, and the risk of congenital malformations. A comparison of congenital malformation rates in ICSI and IVF offspring is considered in the results section of this report (see page 70). For infertile couples, ART may represent the only option to achieve a live birth without using donor gametes. Therefore, this current review focuses on an assessment of the evidence about the relative safety and effectiveness of different ART procedures rather than a comparison of ART versus spontaneous procedures. Safety of controlled ovarian hyperstimulation drugs Ovarian hyperstimulation syndrome OHSS is the most common and serious adverse event associated with controlled ovarian hyperstimulation. Other potential complications include pelvic pain and/or bleeding following oocyte retrieval; however, these events are usually self-limited. OHSS occurs in susceptible women who experience an excessive response to the drug treatment which results in increased oestrogen levels and ovarian size, increased capillary permeability and fluid imbalance. Onset occurs within nine days of oocyte retrieval, or later when it is usually associated with a pregnancy (Papanikolaou et al., 2005). In its mild form, symptoms include nausea, vomiting, diarrhoea and abdominal discomfort. More severe cases are associated with abdominal bloating. In its most severe form, severe fluid imbalance results in dehydration, breathlessness due to pleural effusions, and compromised cardiac, renal and liver function, which can be life threatening. Treatment is based on rehydration, removal of excess fluid collections and prevention of thromboembolisms. The incidence of all OHSS cases in Australia is not known as many cases do require hospitalisation. In Australia and New Zealand in 2002, 192 cases of OHSS requiring hospitalisation were recorded. These are likely to be the most severe cases. This represents 1.1 per cent of 18,186 oocyte pick-ups (OPU) cycles and 0.5 per cent of Assisted reproductive technologies 15

34 36,483 treatment cycles including cycles with unsuccessful OPUs and embryo thaws, and donor insemination (Bryant et al., 2004). Differences in the clinical criteria used to classify OHSS limit the indirect comparison of published rates; however, Australian and New Zealand rates appear to be broadly consistent with international experience. The ESHRE reported an overall risk of OHSS of 0.9 per cent in 204,147 cycles of IVF and ICSI undertaken in 20 European countries in 2001 (Andersen et al., 2005). Research is ongoing to determine the optimal regimens to prevent OHSS. A systematic review conducted by the NCCWCH for NICE (NCCWCH, 2004) concluded that there was no evidence to support the superiority of either HMG, rfsh or urinary preparations in preventing OHSS. This review identified one RCT that demonstrated that using lowdose and step-up FSH in women undergoing IUI is associated with a reduced incidence of OHSS compared with conventional FSH regimens with no difference in pregnancy or multiple pregnancy rates (27.1% versus 8.3%; RR 3.32, 95% CI 1.16 to 9.46) (Sengkou, 1999). There was also evidence that HCG for luteal support resulted in an increased risk of OHSS versus progesterone or no treatment meta-analysis: overall incidence of OHSS with HCG 5 per cent (n = 220) versus 0 per cent (n = 193) with progesterone or no treatment (Soliman et al., 1994). Cancer Systematic reviews of the literature have found no association between COH regimens and cancer of the breast, thyroid, endometrium, cervix, colorectum or melanoma (Brinton et al., 2005; Klip et al., 2000; NCCWCH, 2004). However, much of the evidence is limited to small studies with short follow-up and poor reporting of the type and indications of drug use. Other limitations include the difficulty in separating the effect of other known risk factors such as past reproductive history from the effect of the drugs used in ART (Brinton et al., 2005). Thus, further research with methodologically rigorous long-term prospective studies is required to fully investigate the potential links between COH regimens and cancer risk. Multiple pregnancy In Australia, no more than two embryos or oocytes can be transferred in one treatment in women younger than 40 years of age (FSA RTAC, 2005). RTAC recommends that no more than one fresh embryo or oocyte be transferred in the first attempt in women under 35 years of age. Multiple pregnancy is a common outcome of ART as a result of the practice of transferring more than one embryo in around 24.7 per cent of fresh cycles and 28.8 per cent of frozen IVF or ICSI cycles (Bryant et al., 2004). Multiple births occurred in 14.2 per cent of all ART cycles in 2002 (Bryant et al., 2004). Intrauterine insemination with COH also exposes women to the risk of multiple pregnancy due to the maturation of multiple follicles to optimise the chance of fertilisation. Rates of multiple pregnancy following IUI are not recorded by ANZARD and may vary according to the COH regimen and monitoring procedures used. Twin rates of around 20 per cent and triplet rates of less than 6 per cent following IUI with COH have been reported in Australia (Healy et al., 2003) and elsewhere (Guzick et al., 1999). Multiple pregnancy increases the risk of adverse events for both mother and foetus, which can largely be attributed to the increased risk of preterm birth (Rao et al., 2004; Umstad & Gronow, 2003). These include a higher risk of low birth weight and perinatal 16 Assisted reproductive technologies

35 mortality for infants. A retrospective cohort study of 304,466 twins and 17,696 higher order births delivered in the United States between 1995 and 1997 reported an increased risk of early death with each additional foetus (P <.001), with relative risk of 2.4 (95% CI: ) for triplets compared to twins, 3.3 (95% CI: ) for quadruplets, and 10.3 (95% CI: ) for quintuplets (Salihu et al., 2003). Maternal complications include anaemia, hypertension, polyhydramnios, gestational diabetes, antepartum and postpartum haemorrhage, and caesarean section (American Society for Reproductive Medicine, ASRM, 2004; Umstad & Gronow, 2003). Multiple births also increase parenting stress (ASRM, 2004). Despite these risks, surveys have suggested that multiple pregnancies may not be viewed as an adverse outcome by women with fertility problems (NCCWCH, 2004). The financial consequences of multiple pregnancy are high due to the increased need for prolonged antenatal hospitalisation and neonatal care (Cassell et al., 2004; Koivurova et al., 2004; Lukassen et al., 2004; Motohashi et al., 2004). In the United Kingdom, guidelines and health service policies exist to restrict the number of embryos transferred to the uterus in order to minimise multiple births resulting from ART. NICE guidelines recommend that prevention of iatrogenic multiple pregnancy involves the transfer of no more than two embryos for ART. The guidelines also recommend the judicious use of COH drugs and monitoring with ultrasound to chart follicular development by a specialist clinic (NCCWCH, 2004). The systematic review on which these guidelines were based did not find strong evidence that multiple pregnancy following ART resulted in poorer obstetric and neonatal outcomes than multiple pregnancy conceived spontaneously (NCCWCH, 2004). Another systematic review has suggested that the perinatal mortality rate for ART twins is lower than that for twins conceived spontaneously (Helmerhorst et al., 2004). This difference has been attributed to the higher rate of monozygotic twins conceived spontaneously and the increased rate of congenital abnormalities in monozygotic twins. A large registry study showed similar outcomes for dizygotic twins conceived by ART versus spontaneously (Pinborg et al., 2004b). Other antenatal complications The AHTAC review of ART (1998) and the Alberta Heritage Foundation for Medical Research (Corabian, 1998) both reported a higher risk of antenatal complications other than multiple pregnancies, such as pre-eclampsia (pregnancy hypertension) and placenta praevia (a placental abnormality which often leads to delivery by caesarean section) in women undergoing ART compared to women with spontaneous pregnancies. However, these risks may be associated with maternal age and obstetric history, including prior pregnancy loss, rather than an independent effect of ART. The AHTAC report identified two trials using age and parity matched controls to assess the link between ART and antenatal complications, only one of which reported an association. Another systematic review has suggested that singleton pregnancies from ART have a worse perinatal outcome than non-art singleton pregnancies (Helmerhorst et al., 2004). This question was not addressed by the recent NICE report. Assisted reproductive technologies 17

36 A recent retrospective cohort study, based upon Australian ANZARD data, found that the transfer of fresh embryos and female-factor infertility were independently associated with preterm birth and low birth weight for singletons and twins born following ART (Wang et al., 2005). This finding was based upon 18,429 infants conceived by ART and born between 1996 and 2000, using multivariate logistic regression. The effect persisted following adjustment for maternal age and parity, cause of infertility, number of embryos transferred, type of embryos and type of procedure. Congenital malformations As maternal age increases, the risk of congenital malformations (particularly chromosomal abnormalities) increases. The frequency of trisomies increases exponentially after a maternal age of 35 years. Couples presenting for ART show a higher rate of chromosomal change than the general population (Clementini et al., 2005; Foresta et al., 2005). This has raised concerns that ART may lead to higher rates of congenital malformations, in particular following ICSI, where fertilisation is not achieved through a process of natural selection. The use of culture media to support fertilisation and embryo development have also been raised as potential causes of congenital abnormalities. The NICE report found that, in general, the evidence about the relative safety of ART was broadly reassuring and did not confirm a higher rate of congenital malformations, developmental abnormalities or cancer in children born following ART versus spontaneous pregnancies (NCCWCH, 2004). However, the authors also cautioned that there are no adequate RCTs to assess these issues. The observational studies identified were limited by their method of surveillance, sample size, participant attrition, length of follow-up, and lack of standardised definitions. Furthermore, parental factors associated with infertility may confound the associations observed between the techniques used and congenital malformations. In light of these limitations, existing health technology assessment (HTA) reports have all recommended long-term follow-up studies to further assess the safety implications for children born as a result of ART, in particular ICSI. Although the NICE report concluded that there was insufficient evidence to determine whether ICSI is associated with chromosomal abnormalities in offspring of infertile couples with normal karyotypes, the authors cited a review of studies assessing foetal karyotypes (7 studies, n = 2139) which showed a significant increase in chromosomal abnormalities for ICSI offspring compared to the general neonatal population. These findings included an increase in the number of inherited structural abnormalities, most of which were inherited from infertile fathers. The NICE report also cited observational studies that have suggested that children born after ART have a higher risk of syndromes due to genomic imprinting defects (abnormalities due to the silencing of a gene inherited from one parent). These include Beckwith Wiedemann syndrome (BWS), which is characterised by gigantism (overgrowth), macroglossia (large tongue) and abdominal wall defects. At least one Australian study has investigated this possible link (Halliday et al., 2004). This case control study (37 cases of BWS in Victoria between 1983 and 2002, and 148 matched controls) indicated that the relative risk was significantly increased in children conceived through IVF; however, the uncertainty surrounding the estimate of effect was large (OR 17.8, 95% CI ) (Halliday et al., 2004). The authors of the NICE report suggested that, overall, the existing evidence was insufficient to confirm an association. 18 Assisted reproductive technologies

37 An Australian report produced for MSAC in 2003 that is not yet published drew similar conclusions based on the findings of 16 registry studies and case series. Since publication of the NICE report, several observational studies investigating various developmental outcomes with long-term follow-up of ARTs versus naturally conceived children have been published (Bonduelle et al., 2004; Leslie, 2004). A review of the congenital malformation rates in IVF versus ICSI offspring is presented as one of the research questions for this review (see page 70). Impact on quality of life Counselling is recommended for all couples accessing ART services in Australia (FSA RTAC, 2005). Infertility can be associated with feelings of stress, guilt, low self esteem, anxiety and depression (Chen et al., 2004; El Messidi et al., 2004; Ragni et al., 2005b; van den Akker, 2005). The investigation and treatment of fertility problems can also be a cause of stress for infertile couples (Ragni et al., 2005b; van den Akker, 2005). A prospective cohort study of educated and professional women undergoing IVF or GIFT in the United States found that women who were extremely concerned about the finances associated with the procedure had a high risk of not achieving a live birth (OR = 11.62, 95%CI 1.84, 73.59) (Klonoff-Cohen & Natarajan, 2004a). Despite these stressors, there is evidence from surveys that infertile couples undergoing ART (Connolly et al., 1992) and those who have failed ART (Sydsjo et al., 2005) sustain stable relationships. Some studies have indicated that there are equal or lower levels of parenting stress in parents of IVF offspring than parents conceiving spontaneously (Glazebrook et al., 2004), including Australian evidence of positive adjustment in parents at five years following an IVF conceived birth (McMahon et al., 2003), although it may be expected that higher order deliveries result in increased parental stress (Ellison et al., 2005). Methodological issues Issues relevant to interpreting evidence about the safety, effectiveness and costeffectiveness of ART to clinical practice include factors that have been shown to predict ART success, treatment discontinuation rates, and spontaneous pregnancy rates in couples who have been classified as infertile and received ART treatment. Predictors of ART success ART success rates fall with increasing maternal age (Bryant et al., 2004). Observational studies have also indicated that success rates may vary according to the primary diagnosis, past reproductive history, duration of infertility and number of previous ART cycles (Kupka et al., 2003a; NCCWCH, 2004; Templeton et al., 1996). ART success also varies according to the underlying diagnosis (Bryant et al., 2004), although estimates of relative effectiveness in different patient groups is not established. In couples with severe male factor infertility, IVF success rates may be so low that ICSI may be the only viable means for conception. Assisted reproductive technologies 19

38 In studies comparing the outcomes of different ARTs where the distribution of the underlying predictive factors are not equivalent in the groups compared, results will be confounded by these differences. Discontinuation rate Observational studies have estimated that around 50 per cent of women discontinue ART treatment after an unsuccessful cycle, with higher discontinuation rates observed in older women (NCCWCH, 2004; Olivius et al., 2004). A prospective cohort study of 450 Swedish couples who started IVF treatment and did not achieve a live birth reported that 54 per cent discontinued the treatment before receiving the three cycles offered under subsidy. The investigators reported that a majority of these discontinuations were due to the psychological burden of treatment (26%) or a poor prognosis (25%). Other reasons for discontinuation included spontaneous pregnancy (19%), physical burden (6%) and serious disease (2%) (Olivius et al., 2004). Spontaneous pregnancies Despite a diagnosis of infertility, spontaneous pregnancies may still occur, with rates of between 7 and 21 per cent observed in couples treated with ART (Hennelly et al., 2000; Kupka et al., 2003b). Spontaneous pregnancies are more likely in younger women (Hennelly et al., 2000) and may vary according to the criteria used to select couples for ART and the underlying cause of infertility. Current reimbursement arrangement The Medicare Benefits Schedule provides reimbursement for reproductive services (including ART) in Australia. Reimbursement through the MBS is dependent upon the presence of a medical condition determining a clinical need and not dependent upon partner status. Medicare Australia advises that: Medicare pays benefits for clinically relevant services. A clinically relevant service means a service rendered by a medical practitioner that is generally accepted in the medical profession as being necessary for the appropriate treatment of the patient to whom it is rendered. The payment of Medicare benefits for ART services relies on the practitioner's decision that he/she is providing clinically relevant treatment. The practitioner is also required to take State/Territory laws into consideration when making that decision. A list of the relevant MBS items, a brief description and the associated MBS fees are provided in Table 1. Full details are provided in Appendix F. There is currently no limit to ART services under Medicare. Prior to 1 November 2000, patients were limited to a maximum of six stimulated interventions using item Following the lifting of the six intervention limit on item 13200, the impact on growth of services has been a minimal. Under the extended Medicare Safety Net, introduced in March 2004, 80 per cent of out-of-pocket costs for medical services provided outside hospital are met once an annual threshold is reached. 20 Assisted reproductive technologies

39 Table 1 MBS funding for assisted reproductive services MBS Item Item Description MBS fee, $ Number Planning and management of reproductive services treatment ART using ovarian stimulation. Includes hormone measurement, ultrasound, 1, counselling and embryology laboratory services ART without ovarian stimulation, or with ovarian stimulation by clomiphene citrate only. Includes hormone measurement, ultrasound, counselling and embryology laboratory services Ovulation monitoring for ovarian stimulation of less than 9 days duration and artificial insemination. Includes quantitative estimation of hormones and ultrasound examinations Oocyte retrieval Preparation for semen for ART or artificial insemination Embryo transfer Preparation and transfer of frozen or donated embryos or both ova and sperm For the five-year period from 1 January 2000 to 31 December 2004, the Australian Government, spent $428.5 million on ART, as shown in Table 2: Table 2 Australian Government expenditure on assisted reproductive services , in $million Calendar year Total MBS $39.3 $43.3 $46.0 $50.0 $78.6 $257.2 PBS $27.0 $31.6 $31.7 $37.1 $43.9 $171.3 Total $66.3 $74.9 $77.7 $87.1 $122.5 $428.5 For the 12-month period from 1 January 2004 to 31 December 2004: Medicare expenditure for ART services increased by 57 per cent from $50 million in calendar year 2003 to $78.6 million in calendar year The number of ART interventions undertaken increased by 9 per cent, from 145,517 in the calendar year 2003 to 159,181 in the calendar year International ART policies ART policies and practices vary among different countries reflecting differences in sociocultural values and health service resources. Variations in two key policies in developed countries restrictions to ART funding according to the age of the woman or prior number of failed cycles, and the maximum number of embryos transferred at one time are outlined in the following. Many countries have restrictions on public funding for ART based upon the woman s age. In the United Kingdom and Germany, women must be under 40 years of age to received reimbursement. In the United Kingdom, women must also be clinically assessed as having a reasonable chance of responding to treatment (ie not approaching menopause). In France, women must be under 43 years of age to receive government reimbursement (as at 2001). In the Netherlands, reimbursement is available for only up Assisted reproductive technologies 21

40 to three IVF/ICSI cycles (van Montfoort et al., 2005). Belgium restricts reimbursement to six IVF/ICSI cycles in a lifetime (Gerris, 2004; Koivurova et al., 2004). In Germany, public funding is available for up to four IVF/ICSI cycles, two GIFT cycles and six cycles of AI. The medical, social and economic burden of twins and higher order pregnancies have resulted in guidelines to limit the number of embryos transferred in ART treatments. Many ART centres in Australia and internationally limit the number of embryos transferred to two as a standard of care and some countries are introducing legislation regarding the number of embryos to be transferred. In Australia, no more than two embryos or oocytes can be transferred in one treatment in women less than 40 years of age (FSA RTAC, 2005). RTAC also recommends that no more than one fresh embryo or oocyte be transferred in the first attempt in women under 35 years of age. In the United Kingdom, the HFEA restricts embryo transfer to two embryos. In Germany a maximum of three embryos can be transferred. ESHRE recommends a policy of elective single embryo transfer with cryopreservation of any additional good-quality embryos for treatment of women 36 years of age or younger in a first or second IVF/ICSI cycle (Land & Evers, 2003). This approach is dependent upon selection of good-quality embryos and cannot be performed in countries such as Germany where embryo selection and cryopreservation is banned. In contrast, in Sweden, public reimbursement is provided for an unlimited number of single embryo transfer cycles, but only up to four cycles if more than one embryo is transferred. In the United States, the high national rate of multiple births has been attributed to the limited availability of insurance coverage for ART (ASRM, 2004). The high cost of the procedure is suggested to serve as an incentive to the transfer of a higher number of embryos to maximise pregnancy rates (Reynolds et al., 2003). In 2002, 35.4 per cent of live births from ART in the United States were multiple, with a 3.8 per cent rate of triplets or higher order births (Centers for Disease Control and Prevention, CDC, et al., 2004). The Society for Assisted Reproductive Medicine (SART) and ASRM have produced guidelines on the number of embryos to be transferred in ART cycles (SART & ASRM, 2004b). These guidelines recommend that no more than two embryos be transferred in women under 35. Transfer of only one embryo should be considered in women with a favourable prognosis (undergoing their first cycle with good morphological quality embryos and have excess embryos of a quality suitable for cryopreservation, or who have previously had successful IVF). For women 35 to 37 years with a favourable prognosis, no more than two embryos should be transferred, and no more than three embryos in other women in this age group. It is suggested that the maximum number of embryos transferred is increased by one for each of the age ranges 38 to 40 and over 40. These guidelines apply to transfer of cleavage stage embryos; fewer embryos should be transferred if they are transferred at a later stage of development (eg blastocyst). For patients with two or more previous failed IVF cycles and those with a less favourable prognosis, it is suggested that additional embryos may be transferred according to individual circumstances. It should be noted that these guidelines consider the possibility of transferring more embryos than guidelines produced by several other countries. Theoretically, double embryo transfer (DET) may be expected to offer a higher pregnancy and live birth rate than single embryo transfer, whereas single embryo transfer reduces the risk of multiple pregnancy and associated risks. A retrospective review of 2603 fresh embryo transfers conducted at a Victorian IVF centre between 2001 and 2003 reported comparable pregnancy rates with reduced multiple pregnancy rates for Assisted reproductive technologies

41 (6.5%) SETs versus 2436 (93.6%) DETs in couples with at least four good-quality embryos available for transfer (pregnancy rates 31.8 per cent for SET versus 33.5 per cent for DET with 28.4 per cent twin and 0.7 per cent triplet rate) (Kovacs et al., 2003). A recent health economic study conducted in Belgium found that transfer of a single top-quality embryo was as effective as, and substantially cheaper than, double embryo transfer in women younger than 38 years in their first IVF/ICSI treatment cycles. Costs to three months after delivery were considered in this analysis (Gerris et al., 2004). In Australia, the NHMRC and the FSA are funding the Australian Study of Single Embryo Transfer (ASSET). This is a multicentre double-blind RCT to compare the outcomes of pregnancy following the transfer of either a single embryo or two embryos in an optimal group of patients undergoing IVF, with or without ICSI. Subjects in the trial are women younger than 35 years with at least three high-quality embryos. Other policies regarding the use of donor oocytes and sperm, surrogacy, embryo selection, cryopreservation and pre-implantation genetic diagnosis also vary among countries. Assisted reproductive technologies 23

42 Methods The Commonwealth Government of Australia set up an independent committee to undertake a review of assisted reproductive technologies (ART). The committee formulated five clinical questions to be addressed in this review a priori by describing the relevant population, intervention, comparator, outcomes (PICO criteria, Table 3). The questions developed are given in Table 4. This report updates information provided in the Australian Health Technology Advisory Committee s 1998 report, Review of Assisted Reproductive Technology (see Appendix B). The questions were addressed by reviewing international HTA reports published since 1997, conducting a systematic literature search and undertaking a cost-effectiveness analysis. Table 3 PICO criteria for clinical questions in assisted reproductive technologies review Patients Intervention Comparator Outcomes Infertile couples a 1. stimulated treatment cycle b with 1. n versus n 1 cycles 4 versus 3 cycles subgroup analyses in vitro fertilisation 3 versus 2 cycles with a single embryo by maternal age transfer 2 versus 1 cycle by duration of infertility (1) by severity of male factor infertility (3) 2. stimulated (lowdose) treatment cycle with intrauterine insemination 3. stimulated treatment cycle with intracytoplasmic sperm injection 2a. n versus n 1 cycles 4 versus 3 cycles 3 versus 2 cycles 2 versus 1 cycle 2b. stimulated treatment cycle with IVF 3. stimulated treatment cycle with IVF Safety ovarian hyper-stimulation syndrome rate ectopic pregnancy rate multiple birth rate congenital malformation rate c neonatal mortality up to 28 days postpartum Effectiveness live birth pregnancy rate per treatment cycle/as percentage of total number of treatment cycles overall success rate for live births per patient maternal tolerance, quality of life Secondary outcomes clinical pregnancy rate miscarriage rate prematurity rate foetal abnormality rate patient satisfaction maternal anxiety/ depression Economic evaluation Resource use/costs a Infertile is defined as the inability to conceive after 12 months of unprotected intercourse. b A stimulated treatment cycle is a treatment episode leading to an egg harvest. c Studies reporting congenital malformation or other clinical diagnoses present at birth were included. Studies examining developmental outcomes were excluded. 24 Assisted reproductive technologies

43 Table 4 Clinical questions for assisted reproductive technologies review determined a priori Primary question: 1. What is the safety, effectiveness and cost-effectiveness of 4 stimulated treatment cycles with in-vitro fertilisation compared with 3 stimulated treatment cycles in women aged < 35 years, years and 40 years, or by duration of infertility, assuming all fresh cycles, (OR 1 fresh, 2 frozen embryo transfers per stimulated treatment cycle if rates available)? 3 versus 2 cycles 2 versus 1 cycle 6 versus 5 cycles 5 versus 4 cycles Secondary questions: 2. What is the safety, effectiveness and cost-effectiveness of 4 stimulated treatment cycles with intrauterine insemination compared with 3 stimulated treatment cycles in women aged < 35 years, years and 40 years? 3 versus 2 cycles 2 versus 1 cycle 6 versus 5 cycles 5 versus 4 cycles 3. What is the safety, effectiveness and cost-effectiveness of low-dose stimulated treatment cycles with intrauterine insemination compared with in-vitro fertilisation with a single fresh embryo transfer in women aged < 35 years, years and 40 years? 4. What is the safety, effectiveness and cost-effectiveness of a stimulated treatment cycle with intracytoplasmic sperm injection compared with in-vitro fertilisation in couples by severity of male factor infertility? a 5. What is the safety and effectiveness of a health management system recommending the transfer of frozen embryos (when available) compared with a system recommending the use of fresh embryo transfer only? a Severe male factor infertility = density <5x10 6 per ml, or total motile sperm count (TMC) <10x10 6. Review of literature Search strategy Existing health technology assessment reports A search of the electronic databases and websites of international HTA agencies listed in Appendix C was undertaken to identify HTA reports on ART published between 1997 and August Search terms used included: assisted reproduction, assisted reproductive technologies, in-vitro fertilisation, IVF, artificial insemination, intrauterine insemination and intracytoplasmic sperm injection. This search identified 11 HTA reports available in English or providing an English synopsis, including 2 unpublished Australian reports (Appendix E). Five reports investigated one or more of the clinical questions specified for this review. The methods and results of these reports which are relevant to the current review are summarised in Appendix B (page 113). A recent, comprehensive and high-quality international HTA report was identified which assessed all of the technologies addressed in the clinical questions posed for the current review. This systematic review was undertaken by the National Collaborating Centre for Women s and Children s Health for the National Institute for Clinical Effectiveness (NICE, 2004). Therefore the systematic review in the current report sought only to update the NICE literature search, identifying key studies published from 2003 to Assisted reproductive technologies 25

44 Electronic databases The medical literature was searched to identify relevant studies and reviews for the period between 2003 and Searches were conducted via Medline, PreMedline, EMBASE and Cochrane databases (Table 5). Table 5 Electronic databases searched, limited to Database Period covered Medline 1966 to August week PreMedline To 20 September 2005 EMBASE 1996 to 2005 Week 38 The Cochrane Controlled Trial Register (CENTRAL) Issue a a Cochrane Database of Systematic Reviews searched from 1998 in HTA search. The search strategy used for each database and the number of citations identified is provided in Appendix D. The search conducted in all databases other than CENTRAL was filtered by indexing and text terms used to identify clinical trials, registry and comparative studies. The scope of this search does not allow detailed evaluation of the evidence for focused research questions regarding individual technologies or prognostic factors. An additional targeted review may be required to comprehensively address any individual comparisons of interest. The search terms used included: assisted reproductive techniques, in-vitro fertilisation, IVF, artificial insemination, intrauterine insemination, intracytoplasmic insemination, ICSI, embryo transfer clinical trial, comparative study, registry, retrospective study, prospective study. 26 Assisted reproductive technologies

45 Selection criteria Table 6 Eligibility criteria for identification of relevant studies Characteristic Criteria Publication type Clinical studies included. Non-systematic reviews, letters, editorials, animal and laboratory studies will be excluded. Reports published as abstracts only will be excluded. Study design Comparative studies (randomised and non-randomised) that report outcomes for two or more assisted reproductive technologies, or different patient groups for the same technology (as indicated below under comparator) will be included. Patient Infertile couples where the primary cause of infertility is unknown, due to male factors, or mixed patient groups where the primary cause has been treated or is not amenable to other treatment will be included. Studies investigating one specific patient subgroup (other than unexplained or male factor infertility) will be excluded. Intervention/test Stimulated treatment cycles with IVF Stimulated treatment cycles with low-dose IUI a Stimulated treatment cycles with ICSI Comparator For stimulated treatment cycles with IVF: intrauterine insemination or intracytoplasmic sperm injection; OR comparisons of outcomes by age, duration of infertility, cycle number For stimulated treatment cycles with IUI: in-vitro fertilisation; OR comparisons of outcomes by age, duration of infertility, cycle number For stimulated treatment cycles with ICSI: in-vitro fertilisation Outcomes Primary outcomes: b Safety: ovarian hyperstimulation syndrome rate, ectopic pregnancy rate, multiple pregnancy rate, congenital malformation rate, neonatal mortality rate. Effectiveness: live birth rate per treatment cycle / as percentage of total number of treatment cycles, overall success rate for live births per woman, maternal tolerance/ quality of life, anxiety/depression, patient satisfaction Economic evaluation: Resource use/costs Language Non-English language articles will be excluded unless they are deemed necessary to the review. a Low-dose IUI was defined as insemination following an ovarian stimulation protocol leading to maturation of a maximum of two follicles. b The following secondary outcomes will be extracted from studies meeting the above inclusion criteria: preterm delivery rate, foetal abnormality rate, clinical pregnancy rate, miscarriage rate. Where large numbers of studies were identified for a particular research question, systematic reviews, RCTs and prospective cohort studies, as well as retrospective cohort studies which were based on national registry data or were multicentre were included for full review. Single-centre non-australian retrospective cohort studies were not included for full review. In a hierarchy of evidence from observational studies, single institution studies are considered to provide a lower quality of evidence (Delaney et al., 2005). Search results A total of 2733 unique citations were identified in the literature and HTA search. One reviewer evaluated these citations to determine whether or not they met the inclusion and exclusion criteria outlined in Table 6. A second reviewer evaluated a subset of 500 citations. No additional included studies were identified by the second reviewer. Assisted reproductive technologies 27

46 QUORUM Flowchart Potentially relevant studies identified in the literature search and screened for retrieval (n = 2733) Studies retrieved for more detailed evaluation (n = 217) Potentially appropriate studies to be Included in the systematic review (n = 71) Studies included in the systematic review (n = 38) Studies with usable information by outcome (n = 36) a IVF by maternal age (n = 8) IVF by cycle number (n = 7) IVF by duration of infertility (n = 2) Low-dose IUI (n = 4) ICSI vs IVF effectiveness (n = 7) ICSI vs IVF safety (n =14) Policy using frozen embryos (n = 6) Studies excluded, with reasons (n = 2516) Editorial, non-systematic review, animal/in-vitro study (n = 766) Not an ART study (n = 274) Not comparative/wrong comparator/prognostic factor not of interest (n = 1019) Wrong patient group/donor gametes (n = 163) Did not report outcomes of interest (n = 293) Non-English language (n = 1) Studies excluded, with reasons (n = 146) Editorial, non-systematic review, animal/in-vitro study (n = 36) Not an ART study (n = 1) Not comparative/wrong comparator/prognostic factor not of interest (n = 66) Wrong patient group/donor gametes (n = 4) Did not report outcomes of interest (n = 38) Non-English language (n = 1) Studies excluded from systematic review with reasons (n = 33) Studies reporting duplicate data (n = 3) IVF by maternal age single-centre studies (n = 12) Studies reporting non-current CDC data (n = 3) IUI not low-dose or not reported (n = 11) ICSI versus IVF not equivalent male factor (n = 3) ICSI single-centre studies reporting congenital malformation rates only (n = 1) Studies withdrawn by outcome with reasons (n = 2) 3 IVF cycles versus average all cycles (n = 1) Inadequate data reporting (n = 1) Figure 4 Summary of the process used to identify and select studies for the review Adapted from Moher et al. (1999). a Some studies included in more than one category. 28 Assisted reproductive technologies

47 Data extraction Data was extracted from included studies using a standardised data extraction form. Data extraction was performed by one reviewer and checked by a second. Appraisal of the evidence The evidence presented in the selected studies was assessed and classified using the dimensions and levels of evidence defined by the NHMRC (2000a). These dimensions (Table 7) consider important aspects of the evidence supporting a particular intervention and include three main domains: strength of the evidence, size of the effect and relevance of the evidence. The first domain is derived directly from the literature identified as informing a particular intervention. The last two require expert clinical input as part of its determination. Table 7 Evidence dimensions Type of evidence Strength of the evidence Level Quality Statistical precision Size of effect Relevance of evidence a See Table 8 and Table 9 Definition The study design used, as an indicator of the degree to which bias has been eliminated by design. a The methods used by investigators to minimise bias within a study design. The p-value or, alternatively, the precision of the estimate of the effect. It reflects the degree of certainty about the existence of a true effect. The distance of the study estimate from the null value and the inclusion of only clinically important effects in the confidence interval. The usefulness of the evidence in clinical practice, particularly the appropriateness of the outcome measures used. The three subdomains (level, quality and statistical precision) are collectively a measure of the strength of the evidence. Evidence regarding the relative effectiveness of different ART interventions (research questions 3 to 5) was assessed according to the NHMRC levels of evidence shown in Table 8. However, it should be noted that evidence regarding rare adverse events is unlikely to be adequately captured in RCTs and observational studies will be more appropriate for such outcomes. National multicentre registry studies are considered to provide the most accurate estimation of the rates of such events (Tyldesley et al., 2001). Evidence regarding the relative effectiveness of a single ART intervention in women of different ages, undergoing a different number of cycles of treatment, or with different duration of infertility (research questions 1 and 2) were assessed according to the interim NHMRC levels of evidence for prognosis, shown in Table 9. Assisted reproductive technologies 29

48 Table 8 NHMRC designations of levels of evidence for interventions a Level of evidence I II III-1 III-2 III-3 IV Study design A systematic review of level II studies A randomised controlled trial A pseudorandomised controlled trial (ie alternate allocation or some other method) A comparative study with concurrent controls: non-randomised, experimental trial cohort study case-control study interrupted time series with a control group A comparative study without concurrent controls: historical control study two or more single arm studies interrupted time series without a parallel control group Case series with either post-test or pre-test/post-test outcomes a NHMRC (2005). Table 9 Interim NHMRC designations of levels of evidence for prognosis 1 Level of evidence I II III-1 III-2 III-3 IV Study design A systematic review of level II studies A prospective cohort study All or none Analysis of prognostic factors among untreated control patients in a randomised controlled trial. A retrospective cohort study Case series, or cohort study of patients at different stages of disease 1 NHMRC (2005). The quality of included studies was assessed within the appropriate level of evidence according to the criteria listed in Appendix H (page 132). Economic evaluation This report uses data from the Australian Institute of Health and Welfare, National Perinatal Statistics Unit, to perform a cost-effectiveness analysis of ART in Australia. Where Australian data were not available, additional data were sourced from the literature. 30 Assisted reproductive technologies

49 Results Systematic literature review HTA reports Five international HTA reports published since the AHTAC report in 1998 investigated one or more of the clinical questions specified for this review. The methods and results of these reports, in addition to the AHTAC report, are discussed in the following in relation to the relevant clinical questions about ART effectiveness or safety (also summarised in Appendix B, Table 43). Three reports were classified as high-quality systematic reviews according to the criteria listed in Table 53 (Corabian, 1998; Medical Services Advisory Committee, 2003; National Collaborating Centre for Women's and Children's Health, 2004). The guidelines document produced by the National Collaborating Centre for Women s and Children s Health for the National Institute for Clinical Effectiveness (NCCWCH, 2004) provides the most comprehensive and recent review of the literature. This report cites or updates the earlier reports identified and the key results are presented under the relevant research question headings below. In-vitro fertilisation IVF by maternal age ANZARD data An AIHW NPSU report on ART in Australia comprehensively captures the success rates of more than 99 per cent of all ART cycles in Australia and New Zealand for 2002 (Bryant et al., 2004). These data from the Australian and New Zealand Assisted Reproduction database (ANZARD) provide the highest quality of observational epidemiological data (Tyldesley et al., 2001). Limitations in the accuracy of the data are related to patient self-reporting of pregnancy complications and neonatal morbidity and mortality. Key biases in interpreting the data on the effectiveness of IVF by maternal age are likely to be due to differences in characteristics of the women due to factors other than age. Different proportions of causes of infertility, duration of infertility or practices in different age categories may confound the results. For example, the proportion of cycles with three embryos transferred increases by maternal age from 1.7 per cent for women aged 25 to 29 years to 12.7 per cent for women aged 40 to 44. Information on the comparability of age groups for other characteristics is not available. The clinical pregnancy and live birth rates of fresh non-donor ART cycles by maternal age are shown in Table 10 and Figure 5. The outcome rates for all ART procedures combined are provided in Table 11. The latter data include outcomes for donor oocytes/embryos and frozen cycles. ART success rates are highest in the maternal age group 25 to 29 years, with a live birth rate of approximately 26 per cent for fresh, non-donor cycles (Table 10, Table 11). Live birth rates decline significantly in the age categories over 35 years, to approximately 6 per cent in women aged 40 to 44. This is a 76 per cent reduction in the chance of a live birth Assisted reproductive technologies 31

50 from fresh, non-donor cycles, by comparison with women aged 25 to 29 (rate ratio reduction = 0.76, 95% CI ). Across all ART procedures, miscarriage rates per cycle increased in women aged 30 to 34, 35 to 39 and 40 to 44, by comparison with women 25 to 29 (see Table 11). Miscarriages for women aged 25 to 59 occurred in 3 per cent of cycles started, and 4 per cent of cycles in women aged 40 to 44. This equated to miscarriage rates per pregnancy of approximately 12 and 35 per cent, respectively. Ectopic pregnancy rates per cycle and stillbirths per delivery did not differ significantly across the age groups. Neonatal death rates per delivery did not differ significantly across the age categories between 25 and 40 years of age. However, a significantly higher neonatal death rate was observed in the age group younger than or equal to 25 years than women aged 25 to 29. The cause of this association cannot be determined as differences in the populations for factors other than age are not known. Importantly, multiple birth rates, when expressed per live birth, also decreased in women 35 to 39 and 40 to 44 years of age, by comparison with women 25 to 59 (see Table 11). This effect was observed despite a tendency to transfer more embryos in older women (P<0.0001, χ2 test for interaction between age and the transfer of 3 embryos). Three embryos were transferred in 1.7 per cent of (fresh and frozen) transfer cycles in women aged 25 to 29, and 2.4, 5.6 and 12.7 per cent of cycles in women in the age ranges 30 to 34, 35 to 39 and 40 to 44, respectively. The multiple birth rate per live birth was similar in women aged 45 years or older, by comparison with women aged 25 to 29 (Table 11). However, the smaller number of cycles in this age group limits the power of this analysis to detect and estimate a true effect should one exist. Figure 5 Live births per cycle started by woman s age for fresh, non-donor ART treatment, 2002 Source: Bryant et al., 2004 Note: values for women aged 45 years were combined due to small numbers. 32 Assisted reproductive technologies

51 Table 10 Success rates of fresh, non-donor ART by maternal age, ANZARD data Maternal age, years Cycles Started, N Live birth rate, % per cycle started Clinical pregancy rate, % per cycle started Rate RR 95% CI a Rate RR 95% CI a b b b b > np b Source: (Bryant et al., 2004) a versus year age group; b P< Abbreviations: CI = confidence interval; RR = rate ratio; np = data not published by ANZARD due to small cell size. Note: The few data available for cycles in women younger than 25 years and older than 45 years make estimates of success rates in these age groups unreliable. Assisted reproductive technologies 33

52 Table 11 Success rates of all ART procedures (including donor and frozen cycles) by maternal age, ANZARD data 34 Assisted reproductive technologies Maternal age, years Cycles Started, N Live birth rate, % per cycle started Singleton birth rate, % per cycle started Miscarriage rate, % per cycle started Ectopic or heterotopic PR % per cycle started Rate RR a 95% CI Rate RR a (95% CI) Rate RR a 95% CI Rate RR a 95% CI b e e d e e b np > e e np Live births, N Multiple birth rate, % per live birth Neonatal death rate, % per live birth Rate RR a 95% CI Rate RR a 95% CI Deliveries, N Still birth rate, % per delivery Rate RR a 95% CI c e e np > np Source: (Bryant et al., 2004) a versus year age group; b P<0.05; c P<0.01; d P<0.001; e P< Abbreviations: CI = confidence interval; RR = rate ratio; np = data not published by ANZARD due to small cell size.

53 HTA reports The authors of the NICE report presented data from the UK HFEA database from 1995 to 1999 to conclude that an increase in maternal age reduces IVF success (NCCWCH, 2004). Data were presented per year of age, but no statistical analysis, or measures of variance were reported. The overall live birth rate for fresh treatment cycles in this data set was 17.6 per cent. The rates were greater than 20 per cent for women between the ages of 23 and 33 years. Above 33 years the live birth rates declined to less than 10 per cent by the age of 40. Women older than 40 years of age had a further decrease in the chance of a live birth. Live births rates were below 5 percent for women 42 years or older, reducing to 1 per cent at the age of 45. The report guidelines recommend that the optimal female age range for IVF is 23 to 39 years. The live birth rates per treatment cycle started for fresh non-donor IVF cycles in women aged over 33 years of age are shown in Table 12. Table 12 Live birth rates by maternal age for fresh, non-donor IVF cycle started, UK HFEA data 1995 to 1999 Maternal age, years Treatment cycles, N Live birth rate, % Source: (National Collaborating Centre for Women's and Children's Health, 2004) The HFEA data reviewed in the NICE report also showed a decline in pregnancy rates by maternal age with a distribution similar to that of the live birth rates. Ectopic pregnancy rates per treatment cycle were 0.9 per cent for women aged 18 to 25 years and less than 0.3 per cent in women aged more than 35 years. However, the authors pointed out that another study showed no significant difference in ectopic pregnancy rates following IVF in women over 35 compared with younger women. Miscarriage rates per pregnancy were 10.5 per cent at 30 years of age, 13.1 per cent at 35 years, 22.7 per cent at 40, and 40.7 per cent at 43 years. No tests for statistical significance or confidence intervals were provided for the ectopic pregnancy or miscarriage rate data. The NICE report also indicated that the effectiveness of IVF in women younger than 23 years is uncertain as very few women in that age range undergo IVF treatment. Primary studies ( ) Twenty-four studies were identified in the period from 2003 to 2005 that reported on outcomes of IVF by maternal age. Due to the large number of studies identified, systematic reviews, RCTs, prospective cohort studies, and retrospective cohort studies Assisted reproductive technologies 35

54 which were based on national registry data or were multicentre were included for full review. Of these, four studies reported data from the annual US CDC SART report (Jain et al., 2004; Kissin et al., 2005; SART & ASRM, 2004a; Vahratian et al., 2003). Three of these studies were excluded from review (Jain et al., 2004; Kissin et al., 2005; SART & ASRM, 2004a); instead, updated data from the most recent published report was included for review (CDC et al., 2004). That report provided IVF success rates for all cycles conducted in the United States in The fourth study was retained for review as it provided novel information in the form of adjusted odds ratios (OR) for live birth by maternal age (Vahratian et al., 2003). One single-centre retrospective study conducted in a Sydney IVF clinic was included for review (Jansen, 2003). The remaining 13 singlecentre studies were excluded from review. Characteristics of the seven studies reporting the prognostic value of maternal age on IVF outcome which were reviewed in full are summarised in Table 13. Three reviewed studies were prospective cohort studies, providing high-level evidence for prognosis (level II by NHMRC criteria). Two of these were multicentre studies conducted in the United States (Keye et al., 2004; Klonoff-Cohen & Natarajan, 2004b), one of which was considered a high-quality study (Klonoff-Cohen et al., 2004b). The applicability of the US studies to the Australian setting may be limited as practice differs between the countries. The third study was a single-centre study conducted in Japan, which reported success rates expressed per embryo transfer and thus may overestimate the true effects of IVF. A further three reviewed studies were registry or multicentre retrospective cohort studies, providing a lower level of evidence (NHMRC level III-3 evidence for prognosis). Two of these were reports using ART data from the US national registry. The SART report provided success rates of both fresh and frozen thawed non-donor cycles in 2002 by maternal age (CDC et al., 2004). A second US registry study reported outcomes from frozen thawed embryos in , and statistically adjusted for differences in characteristics of the women other than age (adjusted ORs Vahratian et al., 2003). The evidence from this study was therefore considered high quality. One further retrospective study reported outcomes by maternal age from first IVF cycles from all IVF clinics in the Netherlands. The final study reviewed was a retrospective study conducted in a single Sydney private IVF clinic (Jansen, 2003). The results from these studies are summarised in Table 14. Three studies provided high level evidence of the influence of maternal age on success of ART. Two of these studies were conducted in the United States. Both of these studies demonstrated a significant decrease in live birth rates in older women. One study was conducted over the period 1993 to 1998 in seven centres with high reported success rates (Klonoff-Cohen & Natarajan, 2004b). This study provided high-quality evidence of the 36 Assisted reproductive technologies

55 Table 13 Characteristics of studies reporting the safety or effectiveness of IVF by woman s age Assisted reproductive technologies 37 Study author/s Study design Setting year, region, site Prospective cohort studies Hara et al. (2005) Prospective cohort, single centre study Keye et al. (2004) EMBRACE Study Group Klonoff-Cohen et al. (2004b) Prospective cohort study (2 cohorts are total populations of 2 RCTs) Prospective cohort, multicentre study Dec 1999 Dec 2003, Japan USA 21 centres USA 7 centres with high success rates Study population N 64 transfers 32 <35 years years 239 women 111 <34 years years Selection criteria All couples undergoing conventional, fresh IVF, transfer 2 day 5 embryos. Tubal factor (34%), male factor (8%), endometriosis (14%), unexplained (44%). Sperm count > 1 x 10(6)/ml, or motility >20%. Women undergoing IVF-ET. Infertility >1 year, normal male factor, nonsmokers, regular ovulatory cycles Exclusion: Failed pituitary suppression, positive pregnancy test in last 3 months, BMI >34, history abnormal uterine bleeding, chemotherapy, intolerance or allergy to gonadotropin, active history of substance abuse, not breast feeding, no oral contraceptives during prior cycle, or participated in experimental drug study. 221 couples Fresh, non-donor IVF (62%), GIFT (36%), or ZIFT (3%). Single cycle per couple. >20 years age. Primary or secondary infertility due to tubal disease, endometriosis, immunological causes, male factor or unexplained or failing conventional therapy. Excluded preexisting medical conditions. Outcomes assessed Multiple pregnancy rate Clinical pregnancy rate Miscarriage rate LBR Clinical pregnancy rate Ectopic pregnancy rate OHSS rate pain score (injection site) LBR Clinical pregnancy rate Level of evidence (prognosis) & quality assessment Level II evidence Quality: fair Single centre. Unclear if consecutive. Embryo transfer comparable. Comparability of age groups for other factors unknown. Level II evidence Quality: fair Multicentre Baseline characteristics of age groups reported. Differences in cause inferitlity. Primary question not comparison of ages. Level II evidence Quality: high Multicentre. Clear inclusion criteria. Adjustment for many known confounders.

56 38 Assisted reproductive technologies Retrospective cohort studies CDC, SART & ASRM (2004) Jansen (2003) Linsten et al. (2005) OMEGA study Vahratian et al. (2003) CDC Retrospective cohort, national registry data Retrospective cohort, single academic centre study Retrospective cohort, multicentre (national) study Retrospective cohort, national registry study 2002 USA 391 clinics 1998 Australia Single private centre The Netherlands USA 85,826 cycles Fresh, non-donor ART cycles. IVF 52.5%, ICSI 46.3%, GIFT 0.2%, ZIFT 0.5%. 648 cycles Non-donor fresh + frozen IVF/ICSI cycles with oocyte pick-up. Excludes 54 cycles stopped before OPU women First cycles fresh, non-donor conventional IVF. Subfertility >1 year, completing questionnaires. 21,555 cycles ART procedures using non-donor thawed embryos, aged years, excludes cycles without ET or where embryos transferred to uterus & fallopian tubes. LBR Multiple birth rate Singleton LBR Clinical pregnancy rate LBR Multiple birth rate Clinical pregnancy rate Ectopic rate Miscarriage rate LBR LBR, Multiple birth rate Clinical pregnancy rate Level III-3 evidence Quality: fair National registry data. 100% follow-up. Comparability of age groups unknown. Level III-3 evidence Quality: fair Single centre. Comparability of age groups for other factors unknown. Level III-3 evidence Quality: fair Possible selection bias questionnaire. Adjustment for known confounders. Level III-3 evidence Quality: high National registry data. 100% follow-up. Adjustment for many known confounders. Abbreviations: ART = assisted reproductive technology; ASRM = American Society for Reproductive Medicine; CDC= Centers for Disease Control and Prevention; ET = embryo transfer; GIFT= gamete intrafallopian transfer; ICSI = intracytoplasmic sperm injection; IVF= in-vitro fertilisation; LBR = live birth rate; OPU = oocyte pick-up; SART = Society for Assisted Reproductive Technology; ZIFT = zygote intrafallopian transfer.

57 impact of maternal age on ART success rates by adjusting for many factors other than age that may influence the outcome. The range of the number of embryos transferred (average 4) as well as the time period of the study limited the applicability of the findings to current practice in Australia. In this study, each additional year of maternal age was associated with a 19 per cent increase in the odds of not achieving a live birth (OR 1.19, 95% CI , P=0.003), after adjustment for the number of embryos transferred, parity, number of previous IVF attempts, number of years of smoking, cause of infertility, type of procedure and clinic site. Duration of infertility was not adjusted for in the analysis. In an analysis of different age ranges of women, the odds of women not achieving a pregnancy or a live birth was significantly higher for women over 40 than for women under 35 (OR for no pregnancy = 4.74, 95% CI , P=0.01; OR for no live birth= 20.31, 95% CI , P=0.01). This study also stated that no significant association was found between maternal age and miscarriage (however, rates or odds ratios were not reported, n=71). Similarly, when analysed per live birth (n=41) multiple birth rates were not significantly associated with maternal age. The absolute rates of these outcomes were not reported and the findings are limited by the small sample size. The second study, conducted in 21 US centres, provided fair-quality evidence on IVF outcomes in women younger than 34 years by comparison with women 34 to 40 years of age (Keye et al., 2004). The inclusion criteria for this study were quite strict (eg excluded smokers, see Table 13), so the generalisability of the findings may be limited. The women in the two age groups did not differ significantly in their mean body mass index (BMI) or the proportion of women with tubal factor infertility. However, there were significantly more women with endometriosis (18.3% vs 9.2%, P=0.04, respectively) and significantly fewer women with unexplained infertility (28.8% vs 42.4%, P=0.03, respectively) in the younger than 34 years group. Day 1 serum LH was significantly lower and oestradiol levels significantly higher in the women in the older age group. This may be as a result of ageing in the women, or may be a sign of other underlying differences between the two groups. This study demonstrated a lower live birth rate, but similar ectopic pregnancy and clinical pregnancy rates, between women 34 to 40 years (34.2, 44.2 and 0.8%, respectively) compared to women younger than 34 years of age (41.7, 45.2 and 0.9%, respectively) overall. No variability measures or statistical testing of these overall event rates in the two age groups of women were reported. Statistical adjustment for the imbalance in the infertility diagnoses in the two groups of women to the different success rates of IVF was not performed. The prospective study based on data from a single Japanese centre reported lower clinical pregnancy and multiple pregnancy rates and a higher miscarriage rate in women 35 years of age or older (31.3, 20.0 and 12.5% per ET, respectively) by comparison with women younger than 35 (46.9%, 26.7% and 6.3% per ET, respectively), although these differences were not statistically significant (variability estimates not reported). The authors reported that women in the two age groups had the same number of embryos at the same stage of development transferred. The similarity of the groups for the cause or duration of infertility, or other factors predictive of IVF success rates is unknown. Four additional studies provided lower level (level III-3) evidence for the association between maternal age and success of ARTs. Two studies reported success rates of ART by maternal age from US registry data. Variations in practice between the United States and Australia may limit the generablisability of the findings in these studies. For example, approximately 62 per cent of fresh, non-donor ART cycles in the United States in 2002 involved the transfer of Assisted reproductive technologies 39

58 three or more embryos (CDC et al., 2004). Unadjusted live births and clinical pregnancy rates from fresh ART cycles in 2002 declined with increasing maternal age; however, statistical testing was not reported (CDC et al., 2004, see Figure 6 & Table 13). Live birth rates decreased sharply between the ages of 40 and 43, from approximately 16 per cent at age 40 to 2 per cent for women aged over 43 (Table 14). The data also indicated decreasing rates of multiple pregnancy and multiple births with increasing maternal age. These findings were supported by a high-quality study of ART success rates from frozen embryo transfers in (Vahratian et al., 2003). This study found that other patient and procedure characteristics (including the number of embryos transferred) varied according to age. Nevertheless, a significant trend of declining live birth and multiple birth rates was demonstrated with increasing age, after adjustment for the number of embryos transferred, prior births, prior ART cycles, and use of assisted hatching. These data are likely to underestimate the true effect, as the rates are reported per embryo transfer. The number of embryos transferred was shown to be significantly associated with success rates. The comparability of the age groups for the cause of infertility was not reported. No data or adjustment was reported for the duration of infertility. Figure 6 Success rates for ART by maternal age, US CDC data 2002 Extracted from CDC report (CDC et al., 2004) Assisted reproductive technologies 40