The main procedures involved in IVF treatment are:
pituitary downregulation: switching off the natural ovulatory cycle to facilitate controlled ovarian stimulation
ovarian stimulation: administration of
gonadotrophins to encourage the development of several follicles followed by administration of
hCG to mature eggs ready for collection
egg collection followed by semen production or sperm recovery
transfer of resulting embryos to the uterus
luteal support: administration of hormones to aid implantation of the embryos.
The HFEA considers that a fresh IVF treatment cycle starts when drugs are administered for ovarian stimulation or, if no drugs are used, when an attempt is made to collect eggs.711 The HFEA also considers that a frozen IVF treatment cycle starts when a cryopreserved embryo is removed from storage in order to be thawed and then transferred.711
Immediate versus delayed in vitro fertilisation
A recent multicentred RCT (n = 139 couples) reported significantly higher live birth rates per IVF/ICI cycle when compared with no treatment for three months in women with fallopian tube patency (29% with IVF/ICSI versus 4% with no treatment).712 [Evidence level 1b]
Another RCT compared the effectiveness of immediate IVF with six-month delayed IVF in couples with all causes of infertility. Patients in the treatment group received up to four cycles of IVF treatment. Patients in the control group were permitted to have any form of fertility treatment other than IVF. Intention to treat analysis for this study showed significant differences in live birth rates per couple (12% with immediate IVF versus 5% with delayed IVF; RR 2.36, 95% CI 1.03 to 5.66) and pregnancy rates per couple (17% with immediate IVF versus 8% with delayed IVF; RR 2.43, 95% CI 1.18 to 5.07). No details of the fertility treatment received by the control group were presented.713 [Evidence level 1b]
A further RCT compared early IVF with late IVF (after six months) in couples with all causes of infertility. Patients in the treatment group received one cycle of IVF treatment. The control group received other fertility treatments, such as IUI with superovulation, donor insemination and tubal surgery during the six-month waiting period. Intention to treat analysis of all causes of infertility showed no significant differences in clinical pregnancy rates per couple (10% with immediate IVF versus 7% with delayed IVF; RR 1.51, 95% CI 0.65 to 3.51), nor in live birth rates per couple (9% with immediate IVF versus 5% with delayed IVF; RR 1.86, 95% CI 0.72 to 4.79).714 [Evidence level 1b]
The incidence of spontaneous pregnancy during IVF treatment has been examined in a retrospective study based on couples who had attempted one or more IVF procedures.715 However, the study was based on 484 subfertile couples, having excluded 110 truly infertile couples. Spontaneous pregnancies occurred in 11.2% of couples. The only characteristic that differed between couples with spontaneous and IVF pregnancy was duration of infertility; shorter duration of infertility was associated with spontaneous pregnancy. [Evidence level 3]
The decision to recommend IVF treatment should take into consideration the likelihood of spontaneous pregnancy without treatment, in particular in cases where significant spontaneous pregnancy rates may be expected, as in the case of mild endometriosis and unexplained infertility.716 [Evidence level 3]
In vitro fertilisation for management of fertility problems associated with tubal disease
We found no RCTs comparing IVF versus no treatment specifically in the management of tubal infertility, although two RCTs compared immediate or delayed referral for IVF (see above). In one of the RCTs, a subgroup of patients with infertility due to tubal factors (n = 45) reported a higher success rate with immediate IVF compared with delayed IVF; however, caution is needed in interpreting this result as the subgroup analysis was not conducted on an intention to treat basis.714 [Evidence level 1b]
In vitro fertilisation for management of fertility problems associated with endometriosis
One RCT (n = 245) compared immediate with delayed referral for IVF (see above). A subgroup analysis of 21 women with endometriosis did not detect a significant difference in pregnancy rates between immediate and delayed IVF (33.3% immediate IVF versus 0% delayed IVF). However, this result should be interpreted with caution because it is a subgroup analysis based on a small sample.714 [Evidence level 1b]
A systematic review of 22 observational studies of patients undergoing IVF treatment, suggested that those with endometriosis-associated infertility compared with couples with other causes of infertility had a lower pregnancy rate (OR 0.63, 95% CI 0.51 to 0.77).717 [Evidence level 2b] The overall chance of achieving a pregnancy with IVF in these 22 studies was about 25%. [Evidence level 2b] The effect of endometrioma on the outcome of IVF treatment is unclear.718–721 [Evidence level 3]
Duration of infertility has been shown to be an important factor in determining the chance of pregnancy, with or without treatment.722 Of those couples who have not conceived within one year 50% will do so spontaneously in the subsequent year. Couples who have not conceived after two years have only a 12% chance of conceiving in the following year (see Section 3.1). [Evidence level 3]
Analysis of the HFEA database showed a significant decrease in the IVF live birth rate with increasing duration of infertility from one to 12 years, which persisted after adjusting for the woman’s age.723 The cause of infertility did not have a significant effect on outcome but previous pregnancy and live birth increased the chance of treatment success. Another study found no significant differences in cumulative pregnancy rates between causes of infertility in women undergoing IVF treatment.724 [Evidence level 3]
Cumulative conception and live birth rates among women undergoing IVF treatment were reported to be lowest in patients with male infertility or multiple infertility factors. Cumulative pregnancy rates were significantly higher in couples with secondary infertility, when compared with couples with primary infertility. In cases of tubal, endocrinological and unexplained infertility the success rate of IVF was comparable with the probability of natural conception of young and fertile couples.725,726 [Evidence level 3]
With the exception of ovulatory disorders, the final treatment option for most categories of fertility problem is IVF and its related technologies. (With ovulatory disorders, the options centre on therapies to correct the specific disorders; see Chapter 7). The recognised indications for in vitro fertilisation treatment include:
male factor fertility problems where medical/surgical management and
intrauterine insemination have not resulted in a live birth or are judged to be inappropriate
tubal disease where tubal surgery has not resulted in a live birth or is judged to be inappropriate
endometriosis where surgery and
IUI have not resulted in a live birth or are judged to be inappropriate
unexplained fertility problems of three years’ duration where medical management and
IUI have not resulted in a live birth or are judged to be inappropriate
failure of spermatogenesis caused by prior treatment for cancer where cryopreserved semen is unsuitable for
IUIovarian failure caused by prior treatment for cancer where eggs or embryos have been cryopreserved
a requirement for egg donation.
In addition, female age should be considered when determining the timescale over which other treatments should be explored before proceeding to in vitro fertilisation treatment.
Cost effectiveness of in vitro fertilisation versus intrauterine insemination
A US study compared a protocol with clomifene citrate and hMG plus IUI with a protocol of only hMG and IUI.727 The study involved 99 subfertile couples undergoing a total of 225 cycles of IUI. The study design was a retrospective cohort and no explicit control group was identified. It was reported that the clomifene/hMG/IUI protocol was around a third as expensive (around $660) as the hMG plus IUI protocol (around $1,850). Cumulative pregnancy rates for clomifene/hMG plus IUI were similar to the more expensive regimen. The multiple pregnancy rate for clomifene/hMG plus IUI was reported to be 28% (all twin pregnancies).
A UK study has evaluated the efficacy and cost effectiveness of stimulated IUI (clomifene citrate and FSH) versus stimulated IVF using the same drug regimen.707 The study included 80 couples with unexplained fertility problems but with confirmed ovulation cycles who were randomised to a controlled trial (although this was compromised by treatment response and patient preference further on in the trial). There was no statistically significant difference in outcome per cycle completed (live birth rate) in a sample of 80 couples. The cost of treatment was £32,280 in the stimulated IVF group, compared with £15,384 in the stimulated IUI group. The cost of multiple birth was not included in the analysis. The authors calculated a cost per maternity of £4,611 for IVF and £1,923 for stimulated IUI. No statistical analysis or sensitivity analysis was performed to explore the robustness of these findings or the impact of small changes in outcome or in cost of treatment.
A retrospective cohort study undertaken in a Finnish fertility clinic considered the cost effectiveness of IUI with clomifene citrate/hMG/HCG stimulation protocol using partners’ sperm.728 The IUI cost-effectiveness data were compared with IVF. No control group was explicitly identified. Data on 924 cycles of IUI were included in the analysis. A pregnancy rate of 12.7% per cycle was reported; 70.6% of the pregnancies were viable, 23.5% resulted in spontaneous abortion and 5.9% resulted in ectopic pregnancy. A multiple pregnancy rate of 13.7% was reported. The cost per live birth was £1,670 for clomifene/hMG/IUI, which was less than half the reported cost of IVF over the same period (£4,450). The longer-term costs of multiple birth were not included in the analysis.
Another US study considered the cost effectiveness of three different assisted reproduction protocols: ovarian stimulation only (with clomifene citrate), IUI with hMG and IVF.729 The study was based on a nonsystematic review of the literature and ‘clinical experience’. This study was different from those discussed above because it considered protocols that used different combination of treatments, starting with the least expensive (clomifene citrate) and limiting the use of any type of treatment to three cycles. Using three cycles of clomifene citrate, plus three cycles of stimulated IUI and three cycles of IVF, the cost per delivery was $13,220 after the first cycle and $63,000 after completion of the whole protocol. When 50% of couples had conceived (between the sixth and seventh cycles of treatment in this case), the cost per couple was around $16,000. When clomifene citrate was dropped and only stimulated IUI and IVF were offered the cost per delivery was $22,380 after one cycle and $63,316 after the completed protocol. Around 50% of couples had conceived at a cost of $18,000 per couple. When IVF alone was used, the cost per delivery was $49,128 after one cycle and $71,825 after four cycles. It was estimated that 50% of couples would have had conceived after spending around $27,000. Thus the most cost-effective option turned out to be a protocol that began with the least expensive option.
11.1. Surgery for hydrosalpinges before in vitro fertilisation treatment
Hydrosalpinx is dilation of the fallopian tube in the presence of distal tubal obstruction, which may result from a number of causes.730 In women undergoing IVF, the presence of hydrosalpinx is associated with early pregnancy loss and poor implantation and pregnancy rates,730,731 probably due to alteration in endometrial receptivity.732,733 [Evidence level 2b]
A systematic review of three RCTs showed that tubal surgery such as laparoscopic salpingectomy significantly increased live birth rate (OR 2.13; 95% CI 1.24 to 3.65) and pregnancy rate (OR 1.75; 95% CI 1.07 to 2.86) in women with hydrosalpinges before IVF when compared with no treatment.734 [Evidence level 1a] There were no significant differences in the odds of ectopic pregnancy (OR 0.42; 95% CI 0.08 to 2.14), miscarriage (OR 0.49; 95% CI 0.16 to 1.52), treatment complication (OR 5.80; 95% CI 0.35 to 96.79) or implantation (OR 1.34; 95% CI 0.87 to 2.05).734
RECOMMENDATIONS
Women with hydrosalpinges should be offered salpingectomy, preferably by laparoscopy, before in vitro fertilisation treatment because this improves the chance of a live birth.
[A]
RESEARCH RECOMMENDATIONS
Further randomised controlled trials evaluating the effectiveness of in vitro fertilisation in comparison with no treatment are needed for different durations and causes of fertility problems.
Further research is needed to determine the relative effectiveness of intrauterine insemination and in vitro fertilisation in couples with unexplained fertility problems.
For women who have hydrosalpinges, the effectiveness of draining of hydrosalpinges or performing salpingostomy on improving live birth rate during in vitro fertilisation needs further evaluation.
11.2. Female age
Live birth rates
Fresh embryo treatment cycles
Analysis of HFEA data on all IVF cycles carried out in the UK between August 1991 and April 1994 showed that the overall live birth rate per cycle of treatment was 13.9%. The highest live birth rates were in the age group 25 to 30 years; younger women had lower rates and there was a decline in older women. At all ages over 30 years, use of donor eggs was associated with a significantly higher live birth rate than use of the woman’s own eggs, but there was also a downward trend in success rate with the recipient’s age.723 [Evidence level 3]
More recent data from the HFEA database (covering the period 1995 to 1999) were analysed by single year of age for this guideline (see , , , and below). The analyses were based on fresh and frozen IVF treatment cycles that were registered between January 1995 and March 1999 and involved use of the woman’s own eggs. Data collected after March 1999 have not been used in this guideline because they are self-reported data which have not been validated by the HFEA and are considered by the HFEA to be less reliable than data for the period January 1995 to March 1999.
Comparison of live birth rates per cycle started by age of woman based on fresh (not frozen) embryo transfer and excluding donor eggs, 1995 to 1999 (Source: Human Fertilisation and Embryology Authority).
Comparison of live birth rates per cycle started by age of woman based on frozen embryo transfer and excluding donor eggs, 1995 to 1999 (Source: Human Fertilisation and Embryology Authority).
Comparison of clinical pregnancy rates per cycle started by age of woman based on fresh (not frozen) embryo transfer and excluding donor eggs, 1995 to 1999 (Source: Human Fertilisation and Embryology Authority).
Comparison of ectopic pregnancy rates per cycle started by age of woman based on fresh (not frozen) embryo transfer and excluding donor eggs, 1995 to 1999 (Source: Human Fertilisation and Embryology Authority).
Comparison of miscarriage rates per cycle started by age of woman based on fresh (not frozen) embryo transfer and excluding donor eggs, 1995 to 1999 (Source: Human Fertilisation and Embryology Authority).
relates to live birth rates from fresh IVF cycles. This analysis was based on 110,538 IVF treatment cycles that were registered between January 1995 and March 1999 and involved use of the woman’s own eggs and fresh embryo transfer. The overall live birth rate per fresh treatment cycle in the period January 1995 to March 1999 was 17.6%. Between the ages of 23 years and 33 years the live birth rate per treatment cycle exceeded 20%. The live birth rates for women aged 18 years to 22 years are shown in but reliable conclusions cannot be drawn from the extremely small number of cases on which these rates are based (0.4% of all fresh IVF treatment cycles). Above the age of 33 years, live birth rates per treatment cycle declined, falling below 10% (i.e. less than half the rate in 23 to 33 year-olds) by the age of 40 years. Women of 40 and older have a declining chance, which reduces to 1% at the age of 45 years. [Evidence level 3]
Since the effectiveness of IVF treatment for women aged less than 23 years is uncertain, the use of IVF treatment can only be recommended where there is an absolute indication (for example, tubal blockage, very poor semen quality or prior treatment for cancer).
Frozen embryo treatment cycles
Embryo cryopreservation allows any supernumerary embryos arising from the initial egg collection and fertilisation to be stored for some time before a subsequent attempt at replacement either because the fresh embryo transfer has not resulted in a live birth or because further children are desired. The ability to preserve embryos routinely has the added benefits of increasing the number of potential embryo replacement cycles without additional egg retrievals thereby improving the overall pregnancy rate and decreasing the risk to the patient of OHSS by substituting frozen-thawed embryo transfer in unstimulated cycles. Embryo quality has the most significant impact on post-thaw survival.735 Freezing poor quality embryos will lead to poor cryosurvival and low implantation rates.736 As with fresh embryos, pregnancy rates are affected by factors such as patient age.736–739 A beneficial outcome is also more likely if a pregnancy resulted from the original stimulation cycle from which the frozen embryos were derived.738–740 The number of oocytes retrieved in the initial stimulation cycle and the number of embryos available for cryopreservation also affects outcome.741 [Evidence level 3] Methods of embryo freezing, protocols for post-thawing embryo selection and culture conditions may affect outcome.
HFEA data from the year 1997–98 reported a live birth rate (per attempted frozen embryo replacement) of 10.4% per treatment cycle in 4533 patients using their own gametes.742 The corresponding figure for 1999–2000 was 13.8% of 5131 treatment cycles.743 [Evidence level 3]
The most recent data on live birth rates with frozen IVF cycles obtained from the HFEA are shown in . This analysis was based on 22,546 IVF treatment cycles that were registered between January 1995 and March 1999 and involved use of the woman’s own eggs and frozen embryo transfer. The overall live birth rate per treatment cycle was 11.5%. Between the ages of 23 years and 38 years the live birth rate per treatment cycle varied between 10% and 16%. The live birth rates for women aged 18 years to 22 years are shown in , but reliable conclusions cannot be drawn from the extremely small number of cases on which these rates are based (0.3% of all frozen IVF treatment cycles). The live birth rate for women aged more than 38 years was less than 7%. [Evidence level 3]
Four further studies have shown decreasing live birth rates with increasing female age using fresh embryo transfer.744–747 [Evidence level 3] Two of these studies showed that live birth rates were positively associated with donor insemination,746 embryo quality,746,747 number of embryos transferred,747 and cause of infertility.747
A retrospective review of experience with embryo cryopreservation over an eight-year period (March 1984 to December 1991) reviewed freeze–thaw cycles (4898 frozen embryos, of which 3288 were thawed) excluding those following oocyte donation. Those that survived (n = 2002) were replaced in 897 cycles, resulting in an ongoing clinical pregnancy rate of 10.9%, comparable with an ongoing clinical pregnancy rate achieved with fresh IVF over the same time period of 13.3%. Overall, the cryopreservation of supernumerary embryos and subsequent thawing and transfer increased the overall pregnancy rate of their IVF/GIFT programme by 4%, increased the clinical pregnancy rate of women who had embryos cryopreserved by 7% and increased the cumulative pregnancy rate in those who returned for frozen-thawed embryo transfer cycles by 11%.741 [Evidence level 3] This study was conducted in the 1980s and early 1990s when it was usual practice to use all surviving embryos. The current practice of selecting embryos good quality embryos from a larger pool of surviving embryos could be expected to increase cumulative pregnancy rates. However, we found no recent studies that addressed this issue.
A cohort study (n = 485 couples, 1086 cycles) which assessed the efficiency and efficacy of an IVF programme between 1989 to 1991 found that embryo cryopreservation (n = 193) (within the limitations of Norwegian law, as frozen embryos can only be stored in Norway for 12 months) contributed a 5.2% increase in the live birth rate for women entering the IVF programme.748 Another case-series study (n = 364) reported a cumulative viable pregnancy rate of 40.7% following one fresh and two freeze-thaw embryo replacements (using two embryos only) in women requesting IVF.749 [Evidence level 3]
Available data on the effects of cryopreservation of embryos did not indicate any apparent negative impact on perinatal outcome, early infant development or congenital malformation rate.750 A retrospective study compared babies (n = 283) from births from cryopreserved embryos with babies (n = 961) after conventional IVF. There was no differernce in the incidence of twins, triplets, their mean gestational age, birth weight and perinatal mortality rates between the two groups. The incidence of major congenital malformations was significantly lower in the cryopreserved group (1%) than in the IVF group (3%).751 One study matched 255 children from cryopreserved embryos for maternal age, parity, single or twin pregnancy and date of delivery with 255 children born after standard IVF with fresh embryos and 252 children from spontaneous pregnancies. Growth, the incidence of major malformations and the prevalence of chronic diseases at 18 months were similar in all three groups.752 [Evidence level 3]
Pregnancy rates
relates to clinical pregnancy rates from IVF cycles. This analysis was based on 110,538 IVF treatment cycles that were registered by the HFEA between January 1995 and March 1999 and involved use of the woman’s own eggs and fresh embryo transfer. The overall pregnancy rate per treatment cycle was 21.0%. Between the ages of 22 years and 36 years the pregnancy rate per treatment cycle exceeded 20%. The pregnancy rates for women aged 18 years to 22 years are shown in , but reliable conclusions cannot be drawn from the extremely small number of cases on which these rates are based (0.4% of all fresh IVF treatment cycles). The pregnancy rate for women aged more than 36 years was less than 14%. [Evidence level 3]
A cohort study has shown that pregnancy rates decline significantly after the age of 40 years, and again after the age of 42 years.753 [Evidence level 2b]
Several other studies have shown that pregnancy rates following IVF treatment decline after the age of 35 years,754–757 37 years758 and 40 years.334,724–726,759–763 [Evidence level 3]
The decline in pregnancy rates with age may be related to declining embryo quality.746 Embryo quality is difficult to assess. For apparently equal embryo quality, maternal age does not significantly reduce pregnancy rates.764 In women with good ovarian response to controlled ovarian hyperstimulation, there was no significant difference in pregnancy rates between women aged more than 40 years and those who were younger.588 [Evidence level 3]
Clinical pregnancy rates and pregnancy loss rates are similar whether the frozen embryos are obtained from oocytes fertilised by conventional IVF or from oocytes fertilised by ICSI.765–767 [Evidence level 3]
A retrospective review on IVF outcomes of patients (n = 322) enrolled in a shared oocyte programme from 1997 to 1999 reported a significantly higher clinical pregnancy rate for recipients who had a fresh embryo transfer compared with recipients whose first embryo transfer consisted of frozen-thawed embryos (63.4% versus. 43.6%). However, no difference between the clinical pregnancy rates from fresh and frozen first embryo transfers were found (47.7% versus. 40.9%).768 [Evidence level 3]
Ectopic pregnancy rates
relates to ectopic pregnancy rates from IVF cycles. This analysis was based on 110,538 IVF treatment cycles that were registered by the HFEA between January 1995 and March 1999 and involved use of the woman’s own eggs and fresh embryo transfer. The overall ectopic pregnancy rate per treatment cycle was 0.5%. The ectopic pregnancy rate in women aged 18 years to 25 years was 0.9% and the rate in women aged more than 35 years was less than 0.3%. [Evidence level 3]
Another study has shown that there is no significant difference in ectopic pregnancy rates following IVF in women over 35 years compared with younger women.755 [Evidence level 3]
Miscarriage rates
relates to miscarriage rates from IVF cycles. These rates are presented as per treatment cycle and are therefore lower than if they were presented as per pregnancy. This analysis was based on 110,538 IVF treatment cycles that were registered by the HFEA between January 1995 and March 1999 and involved use of the woman’s own eggs and fresh embryo transfer. The overall miscarriage rate per treatment cycle was 2.7%. The miscarriage rate in women aged more than 35 years was 2.4%. [Evidence level 3] These data were based on numbers of pregnancies shown in and they give miscarriage rates per pregnancy of 10.5% at 30 years, 13.1% at 35 years, 22.7% at 40 years, and 40.7% at 43 years.
Several other studies have reported increased miscarriage rates following IVF in women aged more than 34 years,726 35 years755,757,769 and 40 years.758,759,762,770 [Evidence level 3]
Fertilisation rates
Several studies have reported decreased fertilisation rates following IVF in women aged more than 35 years,754 37 years771 and 40 years.770 Two other studies found significantly lower fertilisation rates in older women772,773 and after previous IVF failure.773 However, no significant decline in fertilisation rates with age was found in a further study.334 [Evidence level 3]
Implantation rates
Two studies have reported decreased implantation rates following IVF in women aged more than 35 years755 and 37 years.774 However, a third study showed no significant difference in implantation rates between women aged over 35 years and younger women.775 Although advancing maternal age predisposes to a reduced chance of success from IVF treatment, maternal age alone is not a useful predictor of embryo implantation or endometrial receptivity in completed IVF treatment cycles.775 [Evidence level 3]
Oocyte number and quality
The decline in success rates with age following IVF may be due to reduced oocyte production. In one study, the number of retrieved oocytes decreased with increasing age, without alteration of the cleavage rate.776 It has also been reported that the number of oocytes recovered and the number of embryos cleaved after two consecutive cycles of IVF treatment did not differ between women aged less than or over 35 years, although conception rates in older women were lower than the overall pregnancy rate in the IVF programme during the same time period.756 [Evidence level 3]
Older women with good ovarian response, producing more than three embryos suitable for transfer, may have a pregnancy rate similar to younger patients. Cycles yielding less than three embryos have a poor prognosis.777 [Evidence level 3]
Treatment discontinuation rates
A high percentage of women discontinue IVF treatment after unsuccessful cycles. An analysis of the French National In Vitro organisation (FIVNAT) database showed that 40–50% of women discontinued IVF treatment after unsuccessful treatment cycles.778,779 [Evidence level 3] One study found that 17.7% of women aged less than 30 years and 50% of women aged 38 to 40 years discontinued IVF treatment after unsuccessful cycles.780 [Evidence level 3] Another study found significant increases in discontinuation rates with age (38% for women aged 25 to 39 years, 50% for women aged 40 to 43 years and 70% for women aged 44 to 45 years).745 [Evidence level 3]
Although age alone may not be a deterrent to fertility treatment, older patients require thorough counselling regarding the decreased likelihood of success of IVF treatment as the woman’s age increases.
RECOMMENDATION
Women should be informed that the chance of a live birth following in vitro fertilisation treatment varies with female age and that the optimal female age range for in vitro fertilisation treatment is 23–39 years. Chances of a live birth per treatment cycle are:
[C]
greater than 20% for women aged 23–35 years
15% for women aged 36–38 years
10% for women aged 39 years
6% for women aged 40 years or older.
The effectiveness of in vitro fertilisation treatment in woman younger than 23 years is uncertain because very few women in this age range have in vitro fertilisation treatment.
[C]
11.3. Number of embryos to be transferred and multiple pregnancy
Multiple gestations are associated with more complications during pregnancy, increased perinatal, neonatal and infant morbidity and mortality,592 as well as significant financial593,594 and psychological595 consequences for the parents. Surveys have suggested that the prospect of multiple pregnancies may not be viewed as an adverse outcome by prospective patients.598–602 [Evidence level 3]
Much of the increased risk for multiple births is due to the increased risk of preterm birth. The care required for these infants also has resource implications for the health services. However, in assisted reproduction, multiple pregnancies do not appear to be at any more risk of poor obstetric and neonatal outcomes than those conceived spontaneously.596,597 [Evidence level 3] The increase in incidence of multiple births in most countries is reported to be almost entirely the result of the use of gonadotrophins and other agents for ovulation induction or assisted reproduction.781 [Evidence level 2b]
The increase in triplet deliveries following assisted reproduction has been linked to the increased sale and use of ovulation induction agents.782 [Evidence level 3] A report by the FIVNAT showed that 7.3% of all IVF conceptions between 1986 and 1993 related to triplets or higher-order multiple gestation.783 [Evidence level 3]
In IVF, the number of embryos transferred to the uterus is the main determinant of the maximum number of babies that might result. In the UK and before the regulation of IVF by the HFEA, the maximum number of embryos transferred was four, with many clinics restricting the number to three. Under the regulation provided by the HFEA since 1991, the maximum number of embryos transferred has been three. In August 2001, the HFEA announced its decision to reduce the maximum number of embryos transferred from three to two, except in exceptional circumstances, where three might be transferred.784 The HFEA 6th Code of Practice, 2004, 218 states that in a single treatment cycle, a maximum of two eggs or embryos can be transferred to a woman of less than 40 years of age, regardless of the procedure used. Women aged 40 years and over may receive a maximum of three eggs or embryos, regardless of the procedure used. [Evidence level 4]
It has been suggested that the concept of an elective single embryo transfer may warrant serious consideration in future to reduce the overall incidence of multiple pregnancy.785 [Evidence level 4]
An RCT (n = 932) comparing superovulation versus no superovulation and intracervical insemination versus intrauterine insemination found that 23.6% of superovulation live births were twins, 5.6% were triplets and 4.2% were quadruplets.688 [Evidence level 1b] There were no multiple pregnancies in the no superovulation group. In the UK, analysis of data from the HFEA (1991 to 1995) showed that among 29,262 transfers of three embryos, 1755 of 6091 deliveries (28.9%) were twins and 5.8% were triplets or more.786 [Evidence level 3]
Analysis of data from 7170 IVF and 530 ICSI cycles reaching fresh embryo transfer at one fertility centre in the UK between 1984 and 1997 showed that 1889 cycles (25%) resulted in pregnancy. A total of 1256 of these pregnancies continued to delivery (16% per transfer) and 355 (28%) of the resulting births were multiple: 292 (23%) twins, 58 (5%) triplets and 5 (0.4%) quadruplets. The probability of birth has increased but the probability of multiple births has remained unchanged, despite HFEA legislation limiting the number of embryos transferred to three in 1991.787 [Evidence level 3]
Provisional data from the HFEA showed birth rates for twins and triplets per started cycle of IVF (using fresh and frozen embryos) to be 6.2% and 0.52%, respectively, in 1999 to 2000, as compared with 6.2% and 0.43% in 2000 to 2001.743 [Evidence level 3] The corresponding birth rates for twins and triplets per live birth were 30% and 2.5% in 1999 to 2000 and 28.6% and 1.9%, in 2000 to 2001, respectively.743 [Evidence level 3]
The most recent validated data from the HFEA database (covering the period 1995 to 1999) were analysed for this guideline. This analysis was based on 110,538 IVF treatment cycles that were registered by the HFEA between January 1995 and March 1999 and involved use of the woman’s own eggs and fresh embryo transfer. The overall multiple live birth rate per treatment cycle was 5%.
A systematic review of the literature reported results from two completed and one ongoing RCTs which compared transfers of one versus two embryos.788–790 [Evidence level 1b] All three RCTs had excluded women who had a poor prognosis (i.e. increased age, history of failed treatment and poor embryo numbers or quality). Sample sizes were small in all three RCTs. A meta-analysis of results from the first (fresh) treatment cycle in each of the RCTs showed that the combined odds ratio for pregnancy rate per cycle with single embryo transfer was 0.54 (95% CI 0.32 to 0.91). The combined OR for live birth was 0.48 (95% CI 0.27 to 0.86). These results indicate that pregnancy rate per cycle is significantly lower following single embryo transfer. However, the multiple pregnancy rate associated with single embryo transfer was significantly lower (combined OR 0.17, 95% CI 0.07 to 0.40). [Evidence level 1a]
Cumulative pregnancy rates were reported in two of the RCTs.789,790 [Evidence level 1b] In the first RCT, 47.3% of women who received a single embryo transfer achieved a clinical pregnancy, whereas 58.6% of women who received a double embryo transfer achieved a clinical pregnancy.790 In the second RCT, 36.4% of women who received two single embryo transfers (in separate treatment cycles) achieved a clinical pregnancy, whereas 28.6% of women who received a double embryo transfer (in a single treatment cycle) achieved a clinical pregnancy.789
These data suggest that in selected groups of women, while single embryo transfer significantly reduces the risk of multiple pregnancies, it is associated with lower pregnancy and live birth rates per cycle of treatment. Cryopreservation of surplus embryos and replacement in subsequent cycles may be associated with higher cumulative pregnancy rates. Larger, definitive RCTs are required with cumulative live birth as the end point.
No randomised trials that compared transfers of two versus three embryos could be identified. A single controlled observational study791 compared two embryo transfers (n = 80) in ‘good prognosis’ women with three embryo transfers (n = 130) in a similar nonrandomised group. The clinical pregnancy rates were similar (OR 1.26, 95% CI 0.70 to 2.26). Multiple pregnancy rates were higher in the three-embryo-transfer group but the difference did not reach statistical significance (OR 2.17. 95% CI 0.98 to 4.82). [Evidence level 2b]
A single randomised trial that compared transfers of two versus four embryos was identified.792 The RCT did not detect a difference in either clinical pregnancy rates (OR 1.34 95% CI 0.46 to 3.87), live birth rates (OR 2.88, 95% CI 0.95 to 8.72) or multiple pregnancy rates per cycle (OR 2.27, 95% CI 0.51 to 10.18). The wide confidence levels reflect the imprecision of the results due to the small sample size. [Evidence level 1b]
An increase in the number of embryos transferred invariably results in higher likelihood of multiple birth but without necessarily improving the overall success rate of IVF.786 [Evidence level 3] This observational study suggests that when more than four eggs are fertilised and available for transfer, the woman’s chance of a birth is not diminished by transferring only two embryos.786 [Evidence level 3]
Economic consequences
An American study based on a single retrospective cohort study in one IVF centre followed 413 treatment cycles.793 This study reported cost differences of about $39,000 for single and twin pregnancies, and $342,788 for triplet and quadruplet pregnancies.
A Scottish study examined the costs associated with IVF before and after the introduction of a policy to restrict the number of embryos transferred. There were 92 women in the ‘before’ group (historical cohort) and 93 women in the ‘after’ group (later cohort).794 There was no significant difference in clinical pregnancy rates between the two groups. A higher rate of multiple births in the historical cohort was associated with higher rates of preterm birth and low birth weight. The cost analysis included cost of intensive care, midwifery, drugs and equipment. In the historical cohort, 50 intensive care days and 115 special care cost days were recorded at a cost of over £500,000. In the later cohort, the costs of these additional services associated with multiple births were £56,000.
A Swedish study examined the transfer of one embryo compared with two in a single institution setting.795 A decision tree was used to model 1488 transferred embryos. The final outcomes were based on case series and opinion and not on robust research evidence. The model assumed that for IVF with one embryo transfer the chance of having a child was 21% and the chance of a twin pregnancy was 0.0021%. The transfer of two embryos was associated with a 24.8% chance of a singleton child and a 7.8% chance of twin children, with a 64% chance of no baby. The total costs of IVF with one embryo were reported to be about SEK11,000 (£822) and SEK43,286 (£3,320) for two embryos. These costs included sick leave, hospital care during pregnancy, cost of delivery, neonatal care and disability care.
These studies suggest that there may be significant resource savings from adopting a policy of limiting embryo transfer after IVF. The cost effectiveness of alternative embryo transfer policies in assisted reproduction is the subject of a study being undertaken at the National Perinatal Epidemiology Unit in Oxford. The results of the study are not yet available, but are due to be disseminated in 2004.
RECOMMENDATION
Couples should be informed that the chance of multiple pregnancy following in vitro fertilisation treatment depends on the number of embryos transferred per cycle of treatment. To balance the chance of a live birth and the risk of multiple pregnancy and its consequences, no more than two embryos should be transferred during any one cycle of in vitro fertilisation treatment.
[C]
RESEARCH RECOMMENDATION
Further research is needed to improve embryo selection to facilitate single embryo transfers.
11.4. Number of previous treatment cycles
The largest study to address the success of IVF treatment according to the number of previous unsuccessful cycles used the HFEA database of all IVF cycles carried out in the UK between 1991 and 1994 (n = 33,701 cycles).723 [Evidence level 3] This study reported that the probability of success decreased with each IVF treatment cycle from 14.0% (95% CI 13.5 to 14.5) at the first attempt, to 13.0% (95% CI 12.2 to 13.7) at the second attempt, 11.4% (95% CI 10.4 to 12.5) at the third attempt, 11.5% (95% CI 10.1 to 13.2) at the fourth attempt, 8.9% (95% CI 7.2 to 11.2) at the fifth attempt, 9.3% (95% CI 6.7 to 12.9) at the fifth attempt and 10.2% (95% CI 7.7 to 13.7) at the sixth to ninth attempts.
In addressing the effectiveness of IVF treatment in the context of the number of previous unsuccessful cycles, the HFEA was unable to provide these data for all 110,538 fresh IVF cycles registered in the period January 1995 to March 1999 that involved use of the woman’s own eggs. However, the HFEA was able to provide these data for a subset of 2247 of these cycles (see ). The data show that the live birth rate per treatment cycle is largely unchanged over the first four attempts, but the sample sizes for the fifth, sixth and seventh attempts are too small to make valid conclusions. [Evidence level 3]
Comparison of live birth rates per cycle started by number of previous unsuccessful treatment cycles based on fresh (not frozen) embryo transfer and excluding donor eggs, subset of 1995 to 1999 data (Source: Human Fertilisation and Embryology Authority). (more...)
Further data relating to the success of IVF treatment according to the number of previous unsuccessful cycles were provided by the Oxford Fertility Unit for this guideline (see ). This analysis was based on 5028 IVF treatment cycles started between January 1995 and December 2001 and involved use of the woman’s own eggs and fresh embryo transfer. These data show that for women aged less than 39 years and those aged 39 years and over, the live birth rate per treatment cycle is largely unchanged over the first three attempts (the live birth rate for women aged 39 years and over with two previous unsuccessful cycles is less reliable than the other rates because it is based on fewer cycles). [Evidence level 3]
Comparison of live birth rates per cycle started by age and number of previous unsuccessful treatment cycles based on fresh (not frozen) embryo transfer and excluding donor eggs, 1995 to 2001 (Source: Oxford Fertility Unit).
Data from 8362 patients who underwent a first cycle of IVF treatment between 1988 and 1989 have been analysed using the FIVNAT database.778 This study found a decline in pregnancy rate with rank of attempt, although the transfer rate and the number of transferred embryos increased with successive attempts. A more recent analysis of the FIVNAT database using data on 35,714 couples who underwent IVF treatment between 1990 and 1996 showed that the clinical pregnancy rate per oocyte recovery decreased from 20.2% on the first attempt to 17.4% on the second attempt, 16.0% on the third attempt, 13.3% on the fourth attempt, 13.4% on the fifth attempt, 12.7% on the sixth attempt, 7.3% on the seventh attempt, and 11.9% on the eighth attempt.779 This relationship was independent of the woman’s age and the cause of infertility. However, the woman’s age remained the most important factor: the cumulative pregnancy rate decreased from 60% for women aged less than 35 years to 17% for those aged more than 41 years.779 [Evidence level 3]
Another study reported data from 4225 women (8207 IVF cycles) who underwent IVF treatment in Australia between 1993 and 1997.796 [Evidence level 3] This study showed that clinical pregnancy rate per oocyte recovery using fresh or frozen embryo transfer decreased from 20.7% on the first attempt to 20.1% on the second attempt, 17.5% on the third attempt, 6.2% on the fourth attempt, 15.0% on the fifth attempt, 14.8% on the sixth attempt, and 11.7% on the seventh to tenth attempts.796 [Evidence level 3]
A multicentre retrospective study conducted in the USA reported pregnancy rates per cycle for cycles 1, 2, 3, 4 and over 4 to be 33.7%, 33.9%, 28.9%, 25.9% and 21.0%, respectively; the corresponding delivery rates were 27.0%, 27.4%, 23.4%, 16.1% and 15.4%, respectively.797 [Evidence level 3] The pregnancy and delivery rates decreased significantly after the fourth cycle and third cycles, respectively. A smaller study found that pregnancy and live birth rates declined with successive treatment cycles.726 [Evidence level 3] Another small study found that implantation rate was significantly associated with rank of attempt.774 [Evidence level 3] Another study reported similar clinical pregnancy rates for up to seven treatment cycles (25%, 29%, 28%, 33%, 35%, 30%, and 40%, respectively).758 [Evidence level 3]
RECOMMENDATION
Couples should be informed that the chance of a live birth following in vitro fertilisation treatment is consistent for the first three cycles of treatment, but that the effectiveness after three cycles is less certain.
[C]
11.5. Pregnancy history
Analysis of the HFEA database showed that previous pregnancy and live birth were associated with increased treatment success.723 [Evidence level 3] However, rates of secondary infertility are higher in the general population than in IVF clinic referrals.798 Another study based on the FIVNAT register showed that women with primary infertility were significantly younger than women with secondary infertility; they also had significantly more oocytes and fewer embryos, and significantly decreased fertilisation and pregnancy rates.799 [Evidence level 3] A further study that examined the relationship between the first cycle of IVF and subsequent cycles found that a previous pregnancy significantly improved a couple’s probability of conception in a later IVF cycle.763 [Evidence level 3]
RECOMMENDATION
Women should be informed that in vitro fertilisation treatment is more effective in women who have previously been pregnant and/or had a live birth.
[C]
11.6. Alcohol, smoking and caffeine consumption
Maternal and paternal alcohol consumption in excess of 12 g (one unit) per day up to one year before assisted reproduction have been associated with a significant decrease in the success rates of IVF and GIFT.800 [Evidence level 3]
Maternal and paternal smoking before assisted reproduction have been associated with significant decreases in the success rates of IVF and GIFT.801–804 Smoking by males is also associated with a decrease in the success rates of IVF and ICSI (OR 2.95; 95% CI 1.32 to 6.59).805 [Evidence level 3]
In an observational study, caffeine consumption (over 2–50 mg/day versus 0–2 mg/day; 100 mg caffeine in one cup of coffee) during a lifetime (i.e., usual intake) and during the week of initial visit for infertility were strong risk factors for not achieving a live birth in women undergoing IVF or GIFT, after adjusting for smoking, alcohol, age, race, education, parity, types of infertility, types of procedure, number of assisted reproduction attempts and number of embryos transferred.806 [Evidence level 3] This study also reported an association between maternal coffee consumption and decreased infant gestational age.806 [Evidence level 3]
11.7. Body weight
It has been reported that a weight loss programme may improve ovulation and pregnancy outcomes in obese infertile women for all forms of fertility treatment, including ovulation induction, IUI and IVF treatment (see Sections 3.6 and 7.1).497,498 [Evidence level 2b]
Obesity (BMI 25.8 to 30.8 kg/m2) has been shown to be a risk factor for spontaneous abortion in women after IVF or ICSI.807 [Evidence level 2b] Obesity is also associated with lower pregnancy rates after IVF when compared with women with a BMI of 25 kg/m2 or under.808 [Evidence level 2b]
Extremes of BMI (over 25–28 kg/m2 or under 20 kg/m2) have been associated with negative effects on IVF parameters leading to decreased chances of pregnancy.809,810 [Evidence level 2b]
RECOMMENDATION
Women should be informed that female body mass index should ideally be in the range 19–30 before commencing assisted reproduction, and that a female body mass index outside this range is likely to reduce the success of assisted reproduction procedures.
[B]
RESEARCH RECOMMENDATION
Further randomised controlled trials are needed to evaluate the effectiveness of assisted reproduction procedures in relation to female body mass index.
11.8. Clinical effectiveness and referral for in vitro fertilisation treatment
The cost-effectiveness models for IVF treatment are described in detail in Appendix B. These show cost-effectiveness by age and by the number of treatment cycles.
Age-specific costs per live birth using three cost estimates (baseline, lower and upper) for IVF treatment and an OHSS incidence rate of 0.2% were calculated. The costs per live birth were very similar for ages 24 years to 33 years, after which they rose steeply with increasing age. For example, using the baseline cost of IVF treatment (£2,771), the costs per live birth were £11,917 at 24 years, £12,931 at 35 years and £20,056 at 39 years. Sensitivity analyses using lower and higher costs for IVF treatment (£1,771 and £3,500, respectively) resulted in costs per live birth of £8,103 and £14,697 at 24 years, £8,800 and £15,943 at 35 years, and £13,723 and £24,673 at 39 years.
Cycle-specific costs where the live birth rate varied by cycle were also calculated using the baseline cost estimate for IVF treatment and the HFEA live birth rates by number of previous unsuccessful IVF cycles shown in . The cost per live birth in the first cycle of IVF treatment was £15,281. The corresponding costs for the second, third and fourth cycles of IVF treatment were £16,169, £14,793, and £14,336. These costs reflect the varying live birth rates by number of previous unsuccessful IVF cycles. Sensitivity analyses using the lower and higher costs for IVF treatment are presented in Appendix B.
Cycle-specific costs were also calculated using the baseline cost estimate for IVF treatment and the Oxford Fertility Unit live birth rates by number of previous unsuccessful IVF cycles shown in . For women aged less than 39 years, the cost per live birth in the first cycle of IVF treatment was £11,694. The corresponding costs for the second and third cycles of IVF treatment were £11,548 and £12,758. For women aged 39 years and over, the costs per live birth were £27,611 for the first cycle of treatment, £28,938 for the second cycle of treatment, and £12,835 for the third cycle of IVF treatment. These costs reflect the varying live birth rates by number of previous unsuccessful IVF cycles (see ) and the cost per live birth for the third cycle of treatment is not very reliable because of the small number of cycles on which the live birth rate was based.
The cost-effectiveness ratios (cost per live birth) presented here can be compared with cost-effectiveness ratios reported for other countries using RCT clinical effectiveness evidence. A review of this evidence shows far higher cost-effectiveness ratios (cost of IVF per delivery) in the USA (as might be expected) but similar results in Scandinavian countries.811 The data reported in are for the year 1994.
Cost of in vitro fertilisation per delivery (1994).
RECOMMENDATIONS
Couples in which the woman is aged 23–39 years at the time of treatment and who have an identified cause for their fertility problems (such as azoospermia or bilateral tubal occlusion) or who have infertility of at least 3 years’ duration should be offered up to three stimulated cycles of in vitro fertilisation treatment.
[GPP]
Embryos not transferred during a stimulated in vitro fertilisation treatment cycle may be suitable for freezing. If two or more embryos are frozen then they should be transferred before the next stimulated treatment cycle because this will minimise ovulation induction and egg collection, both of which carry risks for the woman and use more resources.
[GPP]
11.9. Gamete intrafallopian transfer and zygote intrafallopian transfer
Gamete intrafallopian transfer
GIFT is a technique which has been developed alongside IVF using much of the same technology, but where eggs, once collected, are transferred laparoscopically to the fallopian tube with prepared motile sperm to allow fertilisation to occur in vivo. GIFT is not now widely used because of the need for a laparoscopy. It has been most commonly used in the management of people with unexplained male factor fertility problems, and where transcervical embryo transfer is impossible.
We did not find any RCTs that compared GIFT with no treatment in couples with unexplained infertility.
One RCT compared GIFT with stimulated and unstimulated IUI in woman with unexplained infertility. It found higher pregnancy rates with GIFT (OR 0.12, 95% CI 0.02 to 0.20 with GIFT versus OR 0.018, 95% CI 0 to 0.05 with IUI plus OS; versus OR 0.018, 95% CI 0 to 0.05 with IUI in spontaneous cycle).812 [Evidence level 1b]
Another RCT compared GIFT and conventional infertility treatments in couples with female infertility excluding tubal factors. Overall, it showed higher pregnancy rates in the group receiving GIFT but in the subgroup of woman with unexplained infertility (number of women not specified) there was no significant difference in pregnancy rates per cycle (23.6% with GIFT versus 36.8% with conventional treatments).813 [Evidence level 1b]
The third RCT (n = 39) compared GIFT with ovarian stimulation in couples with unexplained infertility or failure of donor insemination. It found no significant difference in pregnancy rates between the two interventions in those women with unexplained infertility (8% with GIFT versus 13% with ovarian stimulation; RR 0.63, 95% CI 0.10 to 3.98).814 [Evidence level 1b]
A small RCT (n = 13) found no significant difference between GIFT and IVF in terms of pregnancy rates (33% with GIFT versus 28.5% with IVF) in couples with male factor fertility problems.815 [Evidence level 1b]
Zygote intrafallopian transfer
ZIFT is a technique that is not widely practised; it has been developed alongside IVF using much of the same technology. When transcervical embryo transfer is impossible, laparoscopic transfer of embryos to the fallopian tube after fertilisation in vitro offers an alternative route.
A meta-analysis of six RCTs (458 women, 548 cycles) found no significant difference in pregnancy rates between women undergoing ZIFT and IVF and embryo transfer for all causes of infertility exluding tubal factors (OR 0.99; 95% CI 0.62 to 1.57). There was a trend towards a two-fold greater chance of having an ectopic pregnancy in ZIFT than in IVF (OR 2.05; 95% CI 0.21 to 20.22)816 [Evidence level 1a]
The dominant adverse effect of female age on the success of IVF, GIFT and ZIFT has been highlighted in two cross-sectional studies, with a higher cycle cancellation rate and pregnancy loss rate associated with older women with unexplained infertility undergoing assisted reproduction.817,818 [Evidence level 3]
