LEARNING OBJECTIVES:

1. To define and describe how to implement evidence-based medicine.
2. To list factors determining when patients should proceed to ART.
3. To describe the effectiveness of different procedures commonly used in ART.


INTRODUCTION

This manuscript will review the principles of evidence-based medicine, patient selection for ART and utilization of various ART technologies to assist the clinician in decision-making regarding optimal patient selection and application of selected assisted reproductive technologies.

EVIDENCE-BASED MEDICINE

Evidence-based medicine (EBM) involves the integration of our clinical expertise and our patients’ values with the best available research evidence.1 "Clinical expertise refers to the ability to use our skills and experience to identify each patient’s health state, individual risks and benefits of potential interventions, and values and expectations both about possible therapeutic maneuvers and about health. Without expertise, clinical practice risks becoming evidence-tyrannized because even excellent evidence may be inapplicable to a patient. Each patient brings unique preferences, concerns, and expectations to the clinical encounter, and we must incorporate these values into the decision-making process. The best evidence includes clinically relevant research, often from the basic sciences of medicine, but especially from patient-centered clinical research into the accuracy and precision of diagnostic tests, the power of prognostic markers, and the efficacy and safety of therapeutic, rehabilitative, and preventive interventions. Without this evidence, clinical practice risks becoming rapidly out of date."2

The practice of EBM is achieved by: (1) converting the need for information into answerable questions; (2) tracking down the best evidence with which to answer these questions; (3) critically appraising the evidence for validity, importance, and applicability; (4) integrating this appraisal with our clinical expertise and our patient’s unique biology, values and circumstances; and (5) evaluating our effectiveness and efficiency in executing steps 1 to 4 and finding ways to improve them for the future.

Recent developments that have made the practice of EBM more attainable include international groups reviewing the literature, development of evidence-based journals, availability of improved search strategies and electronic search services.

One of the biggest problems with EBM is simply the difficulty in quantifying clinical outcomes. All areas of clinical practice, including reproductive medicine, suffer from the lack of good outcomes data.3 Clinical research is often deficient and most physicians have a limited understanding of interpreting the literature and optimizing the use of the data. For example, it is necessary to know the different types of study designs and their limitations. The US Preventive Services Task Force rating scale for experimental design is either a randomized controlled trial (I) or a nonrandomized controlled trial (II-1). Observational studies are cohort or case-control studies (II-2), multiple time series before and after intervention (II-3) or descriptive studies, consensus panel, or expert opinion(III).4

Evidence-based medicine has its limitations.5 First, no studies have conclusively proven in a prospectively randomized trial that EBM has an effect on outcomes, because no one has overcome the problems of study design, size and follow-up that would be necessary. The principles of evidence-based medicine can be misapplied and the literature misinterpreted. Quantifying outcomes can be difficult, instruments can read incorrectly, measurements by the investigator are potentially biased, statistical significance can be confused with clinical relevance, data can be limited or not available when it is needed, multiple step algorithms are complex and publication bias can limit the reporting of negative outcomes.

But studies have shown that those patients who have evidence-based therapy have better outcomes than those who do not.6 Clinical protocols for standardizing care have also been shown to be an effective tool to improve quality and efficiency of care.7 Better methods to provide more informed consent to patients given the many variables and personal perspectives are needed, and more sophisticated interactive information systems are needed.

EFFECTIVENESS OF TREATMENT

Determining the effectiveness of treatment is difficult because so few good studies have been performed in reproductive medicine. This problem is confounded by the rapid pace of innovations and utilization of new technologies. Even studies which attempt to answer clinical questions scientifically can suffer from nonrepresentative sampling, poor selection of research subjects, improper allocation of treatment, improper stratification of confounding variables, use of inappropriate controls, bias of subjects or evaluators or non-objective assessment of outcome. Other flaws include inadequate sample size, failure to collect data on variables influencing interpretation of results, poor response rates in surveys, subjects lost to follow-up, extensive missing data and quality control problems. If a study has not been designed and conducted in a manner such that the investigator’s hypothesis can be tested by the selected methodological approach, then no amount of statistical revision or other manipulation will suffice to correct this basic flaw. Even good data sets, such as the SART Registry, need to be interpreted carefully.8

When evaluating the literature to determine effective treatment, clinicians need to use logic and common sense to distinguish between promising advances and self-promotion. They also need to be able to evaluate expressions of probability because P-values are frequently misunderstood. The P-value is the probability that the observed result could have occurred by chance alone. Many factors can affect the P-value and its importance in studies should not be over-estimated.9

The following sections in this manuscript will attempt to evaluate what works best in ART using evidence-based medicine, while recognizing that the available data often are inadequate to answer definitively our clinical questions. Importantly, issues of cost will not be addressed simply because of time constraints, but cost is clearly a critical element in medical decision-making.10,11

WHEN TO PROCEED TO ART

Infertility Evaluation

Formal evaluation of the woman is usually performed when she has one year of unprotected intercourse (6 months if age 35 or older) and has failed to conceive, or has two spontaneous pregnancy losses or has known infertility factors in her or her husband. Evaluation includes history and physical examination, testing for ovarian factors, uterine, tubal and peritoneal factors, cervical factor and male factor. These unique features of each couple and their psychosocial characteristics are then integrated with the best available clinical evidence to decide on the most appropriate management of their infertility.

The decision to proceed to ART can still be difficult for many couples and physicians, but studies completed in the past few years do help in making this decision.12-20 Approximated pregnancy rates following standard treatments are shown in Table 1.21

Indications for ART

The major factor affecting the decision to proceed to ART is the diagnosis, but other factors can also be important. Absolute indications for IVF are shown in Table 2. Other factors that are relative indications for ART are shown in Table 3. Some factors that cause infertility cannot be treated by ART. These untreatable conditions can be managed by third party reproduction (donor oocytes, donor sperm, host uterus, surrogacy), adoption or "child-free living". Some of these conditions are listed in Table 4.

Historical Factors Affecting Pregnancy Rates

Age is the single most important factor in pregnancy rates with IVF. ART success rates decrease with age, from a baseline at age 31 about 3% per year to age 34, 8% per year to age 39 and 15% per year from baseline at age 39 to age 42, after which the prognosis is extremely limited. Therefore, as patients pass age 40 they need to give serious consideration to the use of donor oocytes or other options to start their families.22,23

A second factor is diagnosis. Some conditions such as severe tubal disease or severe male factor are almost always better managed immediately with IVF than other treatments. Patients who have severe or extensive endometriosis may have slightly lower pregnancy rates at IVF, possibly due to reduced ovarian response to stimulation as a result of disease and operations affecting the ovaries.24 Patients who have only one ovary may have a reduced prognosis. Sperm abnormalities can almost always be normalized through the use of ICSI.24 But patients with mild endometriosis, oligo-ovulation or minor sperm abnormalities often conceive without ART procedures, and indeed can have second or third pregnancies as a result of diagnosis and treatment of an underlying disease state which can be corrected. This avoids the costs of ART and limits the potential for high-order multiple pregnancies.11,12,18-20

A third factor is prior treatment. Patients who have failed standard treatments of surgery, ovarian stimulation and intrauterine stimulation, and male factor treatment should proceed to ART rather than persist longer than is appropriate with ineffective standard treatments. Second surgeries are not indicated very often, and ovarian stimulation with clomiphene citrate and gonadotropins for more than 3 to 6 cycles of each is almost never indicated.21

A fourth factor is duration of infertility. Patients who have not conceived after 3 to 4 years, regardless of age, diagnosis or treatment, are usually better treated by ART unless no attempt at diagnosis and treatment have been attempted in the past. It should be noted, however, that increasing duration of infertility is associated with reduced ART pregnancy rates, the odds ratio being 1.24 at one year, 0.92 at four years, and 0.66 at 13 years, regardless of other confounding factors.13,16

A fifth factor is prior pregnancy history. Patients who have been pregnant have a higher success rate with IVF, the odds ratios for a live birth in a patient having a prior IVF live birth being 2.14, IVF pregnancy not resulting in a live birth 1.35, live birth excluding IVF 1.26 and pregnancy not resulting in live birth excluding IVF pregnancy 1.12.13,16

Another factor to consider is the number of previous IVF attempts. A study by the SART Research Committee showed that delivery rates for the first IVF cycle was 23.4%, the second cycle 25.9 %, the third cycle 16.1%, the fourth cycle 21.0% and for greater than four cycles 15.4%. A summary of the literature shows that the second cycle has a success rate 92.7% that of the first, the third 91.5%, the fourth 87.0% and for more than four 83.6%. Therefore, for properly selected patients who can respond to ovarian stimulation medication, multiple IVF attempts can be justified on the basis of reasonable ongoing chances for success.25

Finally, importantly, the decision regarding ART must also take into account other factors that can affect success. These include medical conditions and lifestyle choices such as cigarette smoking, caffeine use, alcohol and illicit drugs, weight, diet and exercise that can impact the success rate.26-31 Frank discussions and appropriate responses by the patient are always indicated in deciding which patients should pursue ART.

Hydrosalpinges

There are now reasonable data to support salpingectomy, especially for large hydrosalpinges, prior to IVF because the live birth rate is reduced from approximately 25% to 15% per cycle in the presence of large hydrosalpinges.32-35 Some studies have suggested that salpingectomy might compromise ovarian function so it is imperative to perform the operation very carefully so as not to compromise the vascular supply to the ovary.36 A procedure to occlude the proximal tube appears to be as effective with less risk to the ovary and less surgical risk.37

Myomas

Myomas have long been considered a potential factor in infertility, but obtaining good data has been extremely problematic.38 It has long been recommended that intrauterine myomas or those distorting the uterine cavity should be removed, at hysteroscopy if possible.39-41 Another surgical decision which now must be considered is whether or not asymptomatic intrauterine myomas that do not distort the uterine cavity should be resected prior to IVF. A recent study has reported pregnancy rates of 16% per cycle with intramural myomas and 10% with submucosal myomas not distorting the uterine cavity compared to 30% or more in patients with no myomas or subserosal myomas.42 These results are consistent with the study by Bernard.40 More recently, Surrey et al. showed a significant decrease in implantation rates in patients less than 40 years of age with intramural myomas and structurally normal endometrial cavities, but an apparent somewhat lower pregnancy rate was not statistically significant.43 Another controlled study with a control group did not show any difference when intramural myomas less than 7 cm diameter did not encroach on the uterine cavity.44 It has been my practice to remove myomas before ART when the myoma distorted the uterine cavity, was larger that 8 centimeters, consisted of 3 or more myomas larger than 5 centimeters or was associated with menorrhagia. Additional prospective randomized studies are needed to elucidate this situation further.

Endometriosis

Patients with endometriosis generally have pregnancy rates that are similar to other diagnoses for IVF.23,24 However, some data suggest that results of IVF treatment are less in patients with severe endometriosis, possibly due to a reduction in ovarian response as a result of disease or prior surgeries compromising ovarian function, or as a result of existing endometriomas.45,46 Reasonably good data suggest that patients with severe endometriosis benefit from long-term down regulation with GnRH agonists.47 There is a great need for prospective randomized studies to evaluate the role of IVF for infertile patients with different stages of endometriosis.20 It is unclear whether endometriomas should be treated before IVF, and if so, at what size and by what surgical technique?

FSH Levels

Numerous studies have demonstrated the increase of FSH levels as ovarian reserve decreases.48 The use of different assays has confounded results of basal screening tests, so that each IVF program must be knowledgeable about their normal and abnormal levels. Only one elevated level is highly predictive of reduced chances at conception in an IVF cycle. Determination of concomitant estradiol level is common, but its interpretation is problematic because of difficulties in identifying threshold values in different studies. Estradiol levels over 80 pg/ml are probably abnormal. The clomiphene citrate challenge test is even more specific, and abnormal results are highly predictive of poor pregnancy outcomes. However, it is not very sensitive because many older patients have normal test results but poor ovarian reserve.48

Informed Consent and Guidelines for Care

Once all of the available clinical and laboratory data are collated, and attention given to the psychosocial issues with each couple, this enhanced information base will lead to a more informed consent. Due to the amount of information, sophisticated interactive information systems will probably become more common. The ASRM has developed informed consent guidelines that should be followed for every patient.49 Failure to inform patients adequately is increasingly a cause for litigation.
The ASRM also has published guidelines for the provision of infertility services as well as numerous practice guidelines.50,51 Physicians should follow these guidelines so that they can achieve and demonstrate proficiency in the care of their patients.

CLINICAL MANAGEMENT

IVF vs. GIFT vs. ZIFT

One of the first decisions is whether to perform IVF or GIFT or ZIFT. For many years, GIFT success rates have been higher than IVF success rates as published by SART and others.23,24 This seems to be particularly true for women who are in their forties, and numerous studies have confirmed these results. Supporters of GIFT have suggested that the intrafallopian environment is superior to that in the laboratory and that the timing of embryo implantation is superior in GIFT, leading to higher success rates. Many others have argued that the difference in results is due to patient population differences, with IVF patients having more severe pelvic disease, especially endometriosis, and male factor which are confounding variables leading to lower pregnancy rates. These differences in success rates seem to be narrowing. The 1997 SART success rates show that live birth rate per retrieval for IVF is 27.1%, for GIFT is 30.0% and for ZIFT is 28.0%. The general consensus is that IVF must be fairly equivalent because about 94.4% of the initial oocyte retrievals are for IVF, 3.6% for GIFT and 2.0% for ZIFT.23,24 Prospective randomized trials have not demonstrated differences between IVF and ZIFT or IVF and GIFT. 20,52-56 Patients sometimes undergo GIFT when they have other indications for a laparoscopy, have economic incentives to do GIFT compared to IVF or tend to be older.

Donor Oocytes vs. Patients’ Own Oocytes

The use of donor oocytes in the older patient, the woman who has not responded well to ovarian stimulation or the woman with multiple failed ART cycles and poor quality embryos, can be an effective alternative to achieve a pregnancy.22,23 However, donor oocytes must be considered an alternative rather than treatment. Before donor oocytes are recommended to a patient, she should have appropriate testing of FSH/estradiol, possibly with a clomiphene stimulation test, or preferably have undergone ovarian stimulation with gonadotropins.48 The combination of increasing age over 38 with borderline or elevated FSH and/or estradiol is highly suggestive of a very poor prognosis with the patient using her own eggs. The use of donor oocytes is very successful with the 1997 SART report showing a live birth rate of 40.0% per transfer.22,23

Previous guidelines on therapeutic donor insemination and oocyte donation, published by the American Society for Reproductive Medicine, have advised an arbitrary limit of no more than 25 pregnancies per sperm or oocyte donor in a population of 800,000, in order to minimize risks of consanguinity.57 This suggestion may require modification if the population using donor gametes represents an isolated subgroup or the specimens are distributed over a wide geographic area.

Recently published data have not demonstrated an association between the use of ovulation-inducing agents and ovarian cancer, although definitive conclusions await further follow-up.20,58-62 The only study that specifically suggested that the repetitive use of fertility medications presented greater risk than short-term use addressed the administration of clomiphene citrate and not exogenous gonadotropins. In that study, the risk of ovarian cancer was significantly increased only when treatment exceeded 12 cycles.63 Limitation of the participation of a given oocyte donor to approximately six stimulated cycles would appear reasonable, particularly as the donor might herself eventually require fertility therapy.

Host Uterus (Gestational Carrier) vs. Patient’s Own Uterus

The use of a host uterus or gestational carrier can be effective in situations in which the intended mother has no uterus, or there exists severe congenital anomalies, myomas, adenomyosis, intrauterine adhesions, incompetent cervix or medical problems which preclude or make very unlikely a successful outcome using her own uterus. In 1997 a total of 600 cycles were initiated and the delivery rate was 31.2% per initiated cycle.22,23 Careful attention to informed consent, counseling, and legal issues is mandatory.

Immunotherapy vs. No Immunotherapy

Recent years have seen the promotion of immunological therapy in conjunction with IVF, especially those who have had prior failed cycles. However, careful evaluation of the literature documents that indications for treatment are repeated pregnancy loss with anticardiolipin antibody syndrome based on abnormal anticardiolipin antibody and abnormal activated partial thromboplastin time or similar test.64,65 Treatment of confirmed disease consists of aspirin and heparin therapy. The use of empiric aspirin and heparin for IVF is not supported by data.66,67 The ASRM Practice Committee states, "The assessment of APA is not indicated among couples undergoing IVF. Therapy is not indicated on the basis of existing data."68 The CDC states, "Because the potential for bleeding exists with heparin and aspirin, the risks for and benefits of anticoagulation therapy to improve success rates in IVF patients require vigorous scientific investigation before being accepted as routine practice."69 Aspirin 81 mg daily has been used by some clinicians for treatment of all infertility patients, those undergoing IVF, those who have prior failed cycles, patients with repeated pregnancy loss or other reasons, and at least one prospective randomized trial supports its use.70 Such empiric use of aspirin alone may have benefits that justify the risks in some patients, but further prospective randomized studies are needed.

The use of paternal leukocyte immunization is experimental for repeated pregnancy loss and not indicated for IVF. A recent large, randomized trial has shown no benefit from paternal leukocyte immunization even for repeated pregnancy loss.71

The use of immunoglobulin therapy is not indicated for repeated pregnancy loss and there are no data to support its routine use in IVF patients.72 Antithyroid antibodies may occur more often in patients with recurrent pregnancy loss, but not in women undergoing IVF.73

The ASRM/SART and CDC through the Maternal Mortality Weekly Report, as well as others, have made strong statements concerning the experimental nature of immunotherapy in ART and the need for well designed prospective randomized trials before such therapy is provided and the patient charged for it.64,68,69 There are no convincing data that routine immunologic screening for antiphospholipid antibodies, natural killer cells, anti-thyroid antibodies, anti-ovarian antibodies or paternal leukocyte antibodies, or immunologic treatment alters outcomes for infertile patients. Prospective randomized trials have yet to show that any benefit exists for selected patients.

Ovulation Induction

The first IVF success, Louise Brown, occurred as a result of an unstimulated cycle. A few clinicians have recommended unstimulated cycles.74 Advantages include easier monitoring, lower cost, no increase in multiple pregnancy rate, no ovarian hyperstimulation and no need for ovarian stimulating drugs. Disadvantages include monovular cycles, missed LH surges, rigid patient selection criteria and lower success rates.75

Stimulated IVF is now the preferred protocol for almost all IVF cycles.22,23,76 Advantages are primarily the ability to obtain multiple embryos to select for implantation or cryopreservation, both of which increase the pregnancy rate. The cycle cancellation rate is lower and oocytes are almost always obtained at retrieval. Disadvantages include an increased multiple pregnancy rate, more frequent monitoring, use of stimulating drugs, ovarian hyperstimulation syndrome, higher cost and altered endometrial receptivity.

Higher doses of gonadotropins are frequently used for IVF than non-IVF cycles because the risk of multiple birth can be more easily managed by choosing the number of embryos to transfer. The data with respect to live birth rate following treatment are mixed on the relative merits of human menopausal gonadotropins versus recombinant gonadotropins and the different gonadotropin preparations available, with some, but not all, studies showing a slight benefit with recombinant gonadotropins. 77-85 Supplemental LH is needed for patients with hypogonadotropic hypogonadism, and possibly for others. Luteal phase support with progesterone or hCG is needed when down-regulation with GnRH agonists has occurred or pregnancy rates will be lower.86-88 The benefit of adjunctive growth hormone therapy for women with poor responses to exogenous gonadotropins has not yet been demonstrated.77 There is no consensus regarding the impact of an early elevation of progesterone before LH surge on pregnancy rates nor of the impact of high estradiol levels on endometrial receptivity, oocyte and embryo quality, and cycle fecundity.77 Many studies now support the individualization of stimulation protocols, but much controversy still exists regarding the comparative efficacy of different protocols and their application to individual patients.

Gonadotropin releasing hormone agonists (GnRH agonists) have become a standard part of most ovarian stimulation protocols for IVF, largely because they prevent premature luteinization and cycle cancellation.89 However, it is not clear that these medications are beneficial for every patient or that pregnancy rates are higher because of their use.89 GnRH antagonists are an exciting new class of drugs that have different clinical properties than the GnRH agonists. However, their ultimate role in IVF is not clear at this time.89

Embryo Transfer

Numerous studies have been reported regarding embryo transfer. Topics include patient position, rest following transfer, cleansing of the cervix, flushing of the endocervix, aspiration of mucus, cervical manipulation and dilation, trial or mock transfers, anesthesia, uterine relaxants, prophylactic antibiotics, catheter choice, ultrasound guided transfer, depth of embryo placement, transfer technique, blood/mucus residues in the catheter, use of fibrin sealants, speed of insertion, variability of success with different practitioners and sequential embryo transfers.90

However, very few well-designed prospective randomized studies have been performed and the available data are primarily anecdotal and highly susceptible to bias. About the only conclusions that can be drawn from the literature is that the embryo transfer procedure is a major determinant of success rates, gentle entry into the uterine cavity is important, endometrial thickness of 6 to 8 mm is associated with higher success rates and that individual experience appears to affect pregnancy rates.90-95

LABORATORY PROCEDURES

Intracytoplasmic Sperm Injection vs. Donor Insemination

ICSI has revolutionized treatment for male factor infertility since its inception in 1992.96 When male factor is present ICSI should be discussed with the couple prior to IVF. The decision to utilize ICSI will usually be made by the gamete biologist at the time of the IVF retrieval. A few programs perform ICSI on all patients.97 Different laboratories use different criteria, but usually ICSI will be considered if the male produces fewer than 100,000 total motile sperm or morphology by strict morphology is less that 5% normal. ICSI has been very successful with 1997 results showing a live birth rate (LBR) per retrieval of 27.0% for ICSI patients and 28.4% for non-ICSI patients. ICSI was used in 35.7% of IVF cycles in 1997.22,23

ICSI has now been successfully used to treat all types of poor sperm parameters secondary to impaired spermatogenesis, agglutination, other immunological factors, failure of fertilization with IVF, ejaculatory dysfunction, cryopreserved sperm subsequent to cancer treatment, for diagnosis of fertilization, to avoid polyspermia, and with PGD. ICSI can be used with epididymal spermatozoa in cases of congenital bilateral absence of vas deferens (CBAVD), Young’s syndrome, failed vaso-epididymo- or vasovasostomy, bilateral herniorraphy or obstruction of both ejaculatory ducts. ICSI can be used with testicular spermatozoa for the same indications as epididymal spermatozoa as well as with extensive epididymal scarring, in azoospermia caused by testicular failure and in necrozoospermia. ICSI can be successful in most cases in which there is at least one spermatozoan with a functional genome and centrosome for the fertilization of each oocyte.98 The situations in which no injected oocytes have fertilized normally were associated with only a very few metaphase II oocytes, only totally immotile spermatozoa, gross abnormalities in the oocytes, round-headed sperm, or with all the oocytes damaged in the procedure.99,100

In cases in which ICSI is not able to be successful or the couple prefers not to utilize ICSI for financial, psychological, ethical or genetic considerations, donor sperm can be an alternative. Couples need to be fully informed about this option and its potential advantages, especially before the male undergoes any surgical procedures for sperm retrieval. Counseling with a mental health professional is always recommended before donor sperm can be used. This is simply to ensure that the couple have considered and resolved all the potential issues before proceeding with this option.

As ICSI is an invasive procedure that has become very widely used, especially for men who otherwise would not be able to conceive, concerns have arisen about its safety, especially the potential for genetic abnormalities in offspring. Reports of follow-up studies of pregnancies following ICSI are limited in number. All suffer from significant methodological problems that are always associated with ascertainment bias in congenital abnormality studies.23,98 The largest single-center series concluded that there is a statistically significant increase in sex chromosomal aberrations and structural de novo aberrations compared to a control neonatal population, but other studies have had different results.98 Therefore, patients who have ICSI are generally counseled regarding chorionic villus sampling (CVS), amniocentesis, serum screening tests and ultrasonography. Overall, the data suggest that there is not a significant increase in congenital anomaly rates. However, it is definite that selected males, especially some of those with severe infertility treated with ICSI, have an increased probability of passing on genetic abnormalities including both aneuploidy and gene abnormalities to their male offspring. Approximately 2-7% of men in an infertility program have genetic factors, while about 15% of azoospermic men have chromosomal abnormalities.101-103 Studies have shown child development to be both normal and also abnormal.104-107 The obstetric outcome after prenatal diagnosis in pregnancies created by ICSI appear to be no different than usual.108 Clearly, there is a need for further well-designed long term follow-up studies of babies born as a result of ICSI and to determine which cases are best treated with ICSI.109,110

Assisted Hatching vs. No Assisted Hatching

Another laboratory procedure to be considered is assisted hatching. The literature data are mixed on the efficacy of this technique, with different populations of patients being treated with different techniques and reported with different study designs. Three prospective randomized studies confirm its effectiveness in specific populations with multiple IVF failures.111-113 Three others show no benefit of assisted hatching in a broad population of good prognosis patients.114-116 The assisted hatching procedure has been implicated in an increased rate of monozygotic twinning and possibly conjoined twins.117-120

The United Kingdom Royal College of Obstetricians and Gynaecologists Guidelines for Management of Infertility in Tertiary Care states, "Assisted hatching has not yet been shown to be an effective treatment….it should not be offered outside the context of a clinical trial."121 The SART Practice Committee is developing guidelines for the use of assisted hatching. Many clinicians currently select patients who are older than 38, have a thickened zona pellucida or have prior failed IVF cycles. The potential benefit of the procedure needs to be balanced against the cost and potential damage to the embryo undergoing hatching. Review of these reports suggests that assisted hatching may be clinically useful, and that individual ART programs should evaluate their own patient populations in order to determine which subgroups may benefit from the procedure. However, the routine or universal performance of assisted hatching in the treatment of all IVF patients appears, at this point, to be unwarranted and its utility in any patient population is yet to be conclusively proven.

Blastocyst Culture/Transfer vs. Embryo Transfer
High multiple births are a serious problem facing the ART industry today. Blastocyst culture has become an important adjuvant to IVF therapy in the past few years since it offers the opportunity to reduce the number of high order multiple births by transferring at day 5 or 6 following oocyte retrieval a smaller number of blastocysts, each of which has a higher probability of implanting and resulting in a live birth, than embryos at day 3.122-128

However, the promise of this technology to markedly reduce the number of high order multiple births may not be as realizable as initially hoped, for a number of reasons.

First, the contribution of ART to the overall triplet and higher-order multiple birth ratio was estimated to be 38.7% in 1996 and 43.3% in 1997. The balance are approximately 20% from spontaneously occurring triplets and higher-order multiples and the rest were attributable to ovulation-inducing drugs without ART.129

Second, the numbers of patients who produce sufficient oocytes and good quality embryos to develop further to blastocysts represents a minor percentage of all the patients seen in most IVF programs. While a few programs recommend culturing all patients to blastocysts, most programs are not comfortable doing this because of the possibility of not having any blastocysts to transfer. Numerous studies have been performed in an attempt to identify factors that will allow viable embryos to be selected, or to determine if embryos should be allowed to grow to blastocyst stage. At this time, there are no definitive answers to this issue, although many feel that the number of cells on day 3 is the most important factor and the morphology the second most important. Other parameters have not been shown to be useful.130-133 It is possible this situation will change as programs become more familiar with the culture systems and/or data conclusively demonstrate that patients benefit from culturing to blastocyst stage.

Third, there are no prospective randomized trials to evaluate whether the eventual pregnancy rate is as good with blastocyst transfer as it is with day 3 embryo transfer in all populations. Such studies are difficult to perform in the United States because of limitations on human embryo research. While it is clear that transferring fewer blastocysts can be an effective technique to lower high multiple rates, it is not clear that the pregnancy rate is as high in matched populations. Some studies that have been widely reported show high pregnancy rates in populations that would be expected to have high pregnancy rates, and have still shown high multiple rates as well. It has not yet been demonstrated in large multi-center prospective randomized trials that the pregnancy rate is as high in all populations undergoing IVF, or whether any specific identifiable population, such as younger women, will benefit and by how much.

Fourth, many programs are not having much success with frozen blastocyst transfer, so that any potential benefits of blastocyst transfer may be reduced by the loss of pregnancies usually obtained by transfer of frozen embryos following the fresh cycle.

Fifth, some data suggest that blastocyst culture may expose the conceptus to greater risk, such as that of increased monozygotic twinning, possibly induced genetic alterations, and altered sex ratios.134

Finally, given the insurance coverage situation in most areas of the United States, it is likely that many patients will choose not to take the chance to grow a few embryos out to 5 days when they have healthy appearing embryos at day 3. The blastocyst development rate is approximately 30 to 60% and some patients who elect to culture the embryos to day 5 will end up having no blastocysts to transfer and will be very unhappy with their choice. In our clinic, we counsel all patients about blastocyst culture, with the advantages and disadvantages. We generally recommend blastocyst culture if the couple have 6 to 8 normally developing embryos at day 3.

Number of Embryos or Blastocysts to Transfer

If the couple choose to culture to blastocyst stage, usually two blastocysts are replaced, although arguments can be made in selected populations of young patients, especially those with a congenital anomaly of the uterus or other medical conditions making a multiple pregnancy especially risky, to replace only one.135,136 Any extra blastocysts are cryopreserved.
If the couple chooses not to culture the embryos to blastocyst, the SART guidelines state that in patients with the most favorable prognosis usually no more than two good quality embryos should be transferred.137 This could include women under age 35 who have embryos of sufficient number and quality for cryopreservation, and improved embryo quality as judged by morphologic features. In patients with above average prognosis, for example a woman under 35 without cryopreserved embryos, usually no more than three good quality embryos should be transferred. I patients with an average prognosis, such as a woman age 35-40, usually no more than four good quality embryos should be transferred. In patients with below average prognosis, such as a woman age greater than 40 or multiple failed cycles, usually no more than five good quality embryos should be transferred.137 The 1996 SART pregnancy rates per number of embryos transferred are shown in Table 5.138

These results show that the pregnancy rate increases as up to three embryos are transferred and that the singleton and multiple pregnancy rate remain relatively constant following transfer of anywhere from three to seven embryos. This is likely a reflection of physicians and patients choosing to replace more embryos as their embryo quality decreases. Studies have shown that two good-quality embryos or blastocysts would be sufficient to get acceptable pregnancy rates while limiting the multiple birth rate.135-140 The results show that with two embryos the multiple birth rate is much lower, being only about 25% of that which occurs when 3 or more embryos are replaced (approximately 4% vs. 16%). However, the pregnancy rate is only approximately 60% as high (approximately 20% vs. 35%).138 This difference between the replacement of 2 embryos in Europe compared to 3 embryos in the United States likely is the reason for the higher United States pregnancy rate and associated high multiple pregnancy rate. Newer SART guidelines will hopefully maintain the pregnancy rate while reducing the multiple pregnancy rate.

The key to solving the problem of high multiples lies in being able to determine the quality of the embryo, being able to grow to blastocyst which will ensure higher quality and higher pregnancy rates per blastocyst transferred, and having adequate insurance coverage so that patients and physicians are not motivated to take unnecessary risks by transferring high numbers of good-quality embryos.141,142 Cryopreservation also enables the patient to take advantage of all of the embryos or blastocysts which develop. It must be emphasized, however, that these numbers refer to good quality embryos, and therefore the limitation of absolute numbers of embryos for transfer in all clinical situations is not acceptable to most American physicians. The results above show that some patients must have higher numbers of embryos transferred to have a reasonable chance for pregnancy. This is especially true for the older patient.

The issue of high multiple births is a serious one threatening our industry at this time. Highly publicized high-order multiple births result in a public perception of an unregulated industry needing regulation, of patients who are acting irrationally, and of irresponsible physicians creating costly problems for families and society. We have a collective responsibility to work towards better solutions for this problem.

Cryopreservation vs. No Cryopreservation

Cryopreservation is an additional procedure which can benefit some ART patients. In 1997 approximately 20% of patients had sufficient embryos to cryopreserve.23 The delivery rate per transfer was 16.9% which compares to 29.8% for fresh embryos. This is within the expected range given the loss of implantation potential that occurs with cryopreservation.143 However, the cost savings and markedly decreased medical care and medication make cryopreservation a very useful adjunct to ART procedures. Cryopreservation can be especially useful for the younger patient who has a large number of embryos and is at significant risk for high multiple birth if more than two or three high-quality embryos are replaced. Newer techniques should continue to improve cryopreservation results.144

Multifetal Reduction vs. High Order Multiple Pregnancy

Multifetal reduction (MFR), or induced reduction, is a difficult issue for most infertile patients to consider because many feel that they would be lucky to get pregnant and can hardly fathom multiple births.145,146 Yet this topic should be discussed with all patients early in the consideration of ART procedures. MFR is generally only performed for triplet or higher pregnancies and results in an improved outcome relative to the higher order multiples but does not reach the same satisfactory outcome rate of pregnancies which start out only as twins. The chance that the entire pregnancy will be lost as a result of the procedure is less than 5%.147,148 In our program we will not transfer more than 3 embryos in any patient who has not agreed in advance to consider MFR a satisfactory procedure if they have triplets or higher. Some patients do agree to undergo the procedure and later change their minds, but at least they have had the opportunity to consider all the ramifications of their decision over an extended period of time and with fully informed consent.

Preimplantation Genetic Diagnosis

It has been almost a decade since the first reported pregnancies using preimplantation genetic diagnosis (PGD) have been achieved and over several hundred babies have been born world-wide using these techniques.149-151

However, the clinical experience of preimplantation genetic diagnosis is limited to the few centers reporting their experiences. There has not been any systematic approach to presenting errors. Therefore, the literature is confined to those who choose to report and does not represent complete experience of the technique.152 There have been enough informal reports to estimate an error rate in the 1-10% range depending on the particular disease and assay being evaluated.

The initial experience involved gender determination of embryos as an indirect method to avoid X-linked genetic diseases such as hemophilia, muscular dystrophy, X-linked mental retardation, Lesch-Nyhan syndrome, adrenoleukodystrophy and others. Currently, it is possible to say that the error rate is quite small with these two techniques. Limited data thus far suggest a less than 1% error rate for gender determination.153,154

There are still issues regarding the relative advantages and disadvantages of biopsying the polar body or the blastomere and the concept of allelic dropout, where one of the two alleles selectively amplifies, thus contributing to diagnostic errors.155-163 There have been reports of this phenomenon in as much as 30% of blastomeres, although most have reported much lower numbers (0.3-5.6%). Perhaps, fluorescent PCR, which is more sensitive, may reduce this even further.

There have been more than 200 babies born by this procedure worldwide thus far. To date there are no reports of increased fetal malformation rates or other measurable, identifiable problems. However, a systematic follow-up needs to be done.164,165 The International Working Group is initiating such an effort but confidentiality issues and selective reporting will remain a continuing problem.

PGD appears to be a viable alternative to post-conception diagnosis and pregnancy termination. The procedure is limited to certain genetic diseases and at centers where expertise in molecular genetics and embryology coexist. It is imperative that patients be aware of the uncertainty in diagnostic error and the unknown long-term consequences of this procedure on the fetus.

SUMMARY

ART is a powerful new technology that has had a profound impact on infertility care in the past two decades. The technological advances continue at an almost unbelievable rate. These scientific advances create opportunities for patients but also difficult clinical situations because the most appropriate utilization of these new technologies is not always obvious, and the chance for harming the patient is always present. It is incumbent on all physicians and embryologists practicing ART to practice evidence-based medicine, know when to proceed to ART, use sound clinical judgment in choosing ART procedures and understand which laboratory technologies are appropriate for each patient.

REFERENCES

TABLE 1

TREATMENT	MONTHLY FECUNDITY (%)
No treatment	3
IUI	4
Clomiphene	6
Clomiphene plus IUI	10
Gonadotropin	8
Gonadotropin plus IUI	18
IVF (Clinical pregnancy per retrieval)	34
GIFT (Clinical pregnancy per retrieval)	38

Modified from Guzick et al 21,166

TABLE 3

RELATIVE INDICATIONS FOR ART

TABLE 4

CONDITIONS NOT TREATABLE BY ART

# Embryos transferred	Live Birth Rate Per Transfer (Fresh, <35)	Singleton Live Birth Rate Per Transfer	Multiple Live Birth Rate Per Transfer
One	0.091	0.091	0
Two	0.203	0.162	0.041
Three	0.358	0.212	0.146
Four	0.367	0.205	0.162
Five	0.344	0.192	0.152
Six	0.369	0.191	0.178
Seven	0.318	0.162	0.156