Human Reproduction Update Advance Access originally published online on June 10, 2004
Human Reproduction Update 2004 10(4):295-307; doi:10.1093/humupd/dmh024
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Diagnosis and management of the infertile couple: missing information
Correspondence should be addressed to: Prof. P.G. Crosignani, I. Clinica Ostetrica e Ginecologica, Università di Milano, Via Commenda 12, 20122, Milano, Italy. Email: piergiorgio.crosignani{at}unimi.it
 |
Abstract |
|---|
TOP
Abstract
IntroductionMissing information about reproductive processes is an important barrier to conception and live birth among infertile couples. In vitro and genetic techniques have been informative, but not enough is known about potential defects in fertilization, implantation and early embryo development to define explicit infertility diagnoses and direct specific treatment. Study of gamete quality is needed because endocrine and ultrasound measurements in the female cannot determine oocyte quality and semen analysis in the male is a limited predictor of fertilizing ability. Also, more study would help in understanding the conditions in the Fallopian tube and elsewhere that foster fertilization and embryo development. Research in related subjects is also needed: after many years of enquiry, there exists little or no evidence to determine how and if endometriosis causes subfertility. In contrast, the recently described impact of smoking in males is large, clinically important and possibly reversible. Missing information limits the choice of specific treatment, puts a ceiling on the overall prognosis, increases the likelihood of persistent infertility and forms a barrier to understanding the broader meanings of success in the management of infertility.
Key words: assisted reproductive technology / diagnosis / infertility / management
|
Introduction |
|---|
In addition to the recognized causes of infertility, numerous undetectable defects in the reproductive process might prevent conception among infertile couples. Deficiencies in oocyte or sperm quality and tubal function may contribute to fertilization failure and diminished embryo quality. Potential gaps in knowledge of gamete quality and the transport of gametes and embryos are discussed in the first three sections of this review. When gametes have combined to produce embryos of apparently high quality, the embryos may undergo implantation failure by means that are poorly understood. Thus, fertilization and post-fertilization processes are vulnerable to failure through mechanisms that are largely unknown. Such unknown mechanisms may underlie the impact of endometriosis on reproduction, a condition involving more theories than specific pathways. Smoking is an example of the many environmental factors that potentially may influence fertility, and the effects of female smoking on reproduction and early pregnancy are well-established (Laurent et al., 1992
). The information gap concerning the impact of male smoking on reproduction is now being filled with new data from clinical outcome studies. The extent to which lack of knowledge may jeopardize the probability of conception for infertile couples can best be judged by estimating the proportion of couples with persistent infertility after conventional treatment and the maximum success that can be expected with optimal use of conventional and assisted reproductive technology treatment. At present, however, many couples decline to complete what might be considered an optimal treatment regimen, even when the costs are covered by health care programmes. There remains much to be learned about the different levels of motivation that determine the demand for infertility treatment.
|
Quality of ovulation: myth or reality? |
|---|
Ovulation is of fundamental importance to survival of the species. Extrusion of an oocyte containing the maternal genetic material from the ovary starts the process which results eventually in a new individual. It is essential therefore that the appropriate number of oocytes are ovulated and that these oocytes should be capable of being fertilized and forming a normal embryo. Each species has developed complex systems involving endocrine and paracrine signals to ensure the production of developmentally competent oocytes (Baird and Mitchell, 2002
). The complexity of this process cannot be evaluted with sensitivity in the usual diagnostic assessment for infertility. Also, many of the inefficiencies associated with assisted reproductive technology arise from failure to reproduce these physiological control mechanisms. Normal ovulation
The events leading up to ovulation involve selection and subsequent maturation of the ovulatory follicle(s). During folliculogenesis the oocyte undergoes a series of maturational processes which involve both the cytoplasm and nucleus. Unless these events occur in the correct sequence, the oocyte will not acquire full developmental competence. For example, oocytes aspirated from small antral follicles (<10 mm) at IVF are developmentally incompetent and require maturation in vitro before fertilization can occur (Sirard and Trounson, 2003). Oocytes from large follicles have a greater developmental potential than those aspirated from follicles <12 mm (Bergh et al., 1998). The oocyte is arrested at the dictyate stage of the first meiotic division until a few hours before ovulation. The rising level of estradiol secreted by the ovulatory follicle(s) induces a surge of LH, with consequent changes in structure and function of the oocyte as well as the granulosa and theca cells. Blood vessels invade the follicular cavity as the basement membrane between the theca and granulosa cells breaks down. The cumulus cells surrounding the oocyte disperse due to secretion of mucinous substances. Inflammatory cells invade the theca and granulosa cell at the point of eventual rupture (stigma). Meanwhile the oocyte resumes meiosis so that by the time of ovulation it has reached metaphase II (Picton et al., 1998). The bivalent chromosomes separate on the spindle and segregate into the oocyte and the first polar body. These maturational changes in both nucleus and ooplasm are essential prior to ovulation if normal fertilization is to occur. The relatively low pregnancy rate following attempts to mature oocytes in vitro may be related to failure to reproduce these maturational changes (Hardy et al., 2000).
Abnormal ovulation
Little is known about the incidence or the causes of abnormal ovulation in spontaneous cycles (Baird, 2002). Almost all methods of assessing ovulation are indirect, involving measurement of ovarian and pituitary hormones and/or follicular growth by ultrasound. Pregnancy is the ultimate test of normal ovulation, but it only occurs in 
25% of cycles in apparently healthy women who have intercourse during the peri-ovulatory period (Wilcox et al., 1995). So errors in ovulation and/or the quality of the oocyte appear to be relatively common.
Luteinized unrupture follicle syndrome (LUF)
LUF occurs when, in spite of an LH surge and luteinization of the granulosa cells, the follicle wall does not rupture and the oocyte remains trapped within the follicle (Janssen-Caspers et al., 1986; ESHRE Capri Workshop Group, 2000). Because the granulosa cells are luteinized, the levels of progesterone rise normally in the luteal phase. However, ultrasound measurements demonstrate failure of collapse of the follicle, which persists as a cystic structure into the luteal phase. LUF occurs much more commonly in women who are taking non-steroidal anti-inflammatory agents such as indomethacin, which inhibit the synthesis of prostaglandins. The higher incidence of LUF in women treated with clomiphene may be due to an inappropriate LH surge and may account for the relatively low pregnancy rate per apparently ovulatory cycle (Randall and Templeton, 1991). Otherwise, LUF is not typically repetitive and is unlikely to be a cause of continuing infertility.
Abnormal oocyte
Ovulation of an abnormal or dysmature oocyte is a much commoner cause of relative infertility than failure of the follicle to rupture. The oocyte must be fertilized when it is ‘ripe’; immature or post-mature oocytes result in reproductive failure (Hassold and Hunt, 2001). Although the incidence of oocyte abnormalities in spontaneous cycles is uncertain, abnormal oocytes are more likely in assisted reproductive technology cycles where the aspiration of small follicles yields immature oocytes which require maturation in vitro before fertilization.
It is likely that a relatively large number of follicles in women during reproductive life contain aneuploid oocytes. The incidence of spontaneous miscarriage and live-born babies with certain congenital abnormalities rises strikingly with maternal age. It has been calculated that 20% of conceptions are aneuploid, the majority of which result in non-viable embryos (Eichenlaub-Ritter, 1998). Trisomy 21 (Down's), trisomy 18 (Edward's), trisomy 13 (Patau) and sex chromosomy monosomy XO (Turner's) may survive to birth. Aneuploidy is also seen more commonly in preimplantation genetic screening among women aged >37 years (Table I). The vast majority of aneuploidies are due to errors in chromosome segregation at the first meiotic division in the oocyte, that is they are maternal in origin. Non-disjunction can take place by way of two mechanisms: (i) non-disjunction of a whole chromosome at meiosis I or II; or (ii) pre-division of chromatids during methaphase II arrest (Angell, 1991). Thus, errors may arise at any time during meiosis from fetal life to adulthood, or in association with cytoplasmic deficiencies just prior to ovulation, due to spindle defects or nuclear aberrations. The incidence of chromosomally abnormal oocytes rises strikingly with age (Angell, 1997; Eichenlaub-Ritter, 1998; Sandalinas et al., 2002). The poor quality of oocytes in aged women is clearly illustrated by the improved pregnancy rates obtained with donated oocytes (Figure 1). About 80% of the abnormalties are due to non-dysjunction of homologous chromosomes at meiosis I (mainly 13, 15, 21 and 22). Pre-division of chromatids at meiosis I with precocious segregation at anaphase I can result in aneuploidy and is especially common in chromosomes with a distal chiasmata (Angell, 1991).
|
|
The reasons for these errors in ageing oocytes are uncertain. There may be cumulative lifetime damage to cytoskeleton and/or chromosomes due to intrinsic factors (reactive oxygen species, hypoxia, vascular endothelial growth factor) or extrinsic factors (radiation, environmental pollutants). Oocytes from older women and those failing to fertilize often contain abnormalities in the shape of the spindle and in chromosome alignment (Eichenlaub-Ritter et al., 1988
; Kovacic and Vlaisavijevic, 2000; Eichenlaub-Ritter, 2002). It has been suggested that primordial follicles are recruited for development in the order which they are formed in fetal life (‘first-in–first-out’) and that those formed last are more prone to errors (Henderson and Edwards, 1968). Alternatively, it may be the size of the oocyte pool rather than chronological age which determines risk, as women with a depleted oocyte pool have a higher risk of having a trisomy child (Brook et al., 1984; Eichenlaub-Ritter et al., 1988; Freeman et al., 2000). In summary, the selection of a competent oocyte in the ovulatory cycles that occur throughout reproductive life is crucial to the survival of the species. Nevertheless, there is now substantive evidence that a high proportion of the oocytes ovulated are developmentally incompetent. Even when folliculogenesis appears to proceed normally as detected by endocrine and ultrasound measurements, oocyte quality may not be adequate for pregnancy to occur. The percentage of defective oocytes increases strikingly with increasing age and is the main reason for the decline in fecundity in older women. Although there is extensive interaction between the oocyte and the somatic cells during folliculogenesis, there is little evidence that the occurrence of ovulation depends on the developmental competence of the oocyte. In the human, therefore, reproductive quality control is exerted at fertilization, implantation and fetal develoment rather than at ovulation. Unfortunately, while the number of oocytes developing in IVF cycles may reflect oocyte quality, direct diagnostic tests do not exist that can allow an assessment of oocyte quality.
|
Known and potential causes of oligozoospermia |
|---|
Epidemiological studies of infertility in the developed world indicate that semen abnormalities in the male partner are the commonest diagnostic abnormality (Irvine, 1998
). Our understanding of the causes of male subfertility remains limited, even though ICSI has revolutionized the management of these couples. Many of the difficulties arise from the problems of arriving at a meaningful diagnosis in the male. Although semen analysis methodology has been rigorously standardised by the World Health Organization (World Health Organization, 1998), it remains primarily a descriptive exercise, and the relationships between semen quality and biological fertility remain controversial. With respect to sperm concentration, for example, it has been suggested that an excess of ‘infertile’ over ‘fertile’ men is seen only at concentrations 
whereas from 13.5 to 
sperm concentrations are similar in fertile and infertile men (Guzick et al., 2001). In contrast, in studies which have examined the relationships between semen quality and the achievement of spontaneous pregnancy it has been shown that the probability of conception increased with increasing sperm concentration up to 
but any higher sperm density was not associated with additional likelihood of pregnancy (Bonde et al., 1998).
Male infertility has been associated with hypogonadotrophic hypogonadism, undescended testes, structural abnormalities, genital infections, scrotal or inguinal surgery, varicocele, chronic illness, medication and exposure to chemicals. The nature of the association may be uncertain, as in varicocele, where operating on the lesion does not change pregnancy rates, thus identifying one further area of missing information (Evers and Collins, 2003). The WHO diagnostic classification for the infertile male partner is useful as a basis for standardization in multi-centre studies, but many of the diagnostic categories are of a descriptive nature (e.g. idiopathic oligozoospermia) or of controversial clinical relevance (e.g. male accessory gland infection) (Rowe et al., 1993). Also, in 
40% of men no cause of the infertility can be found (Crosignani et al., 1993). Moreover, recent advances in our understanding of the causes of male infertility mean that some elements of this classification need revision. Complex and detailed analyses show that both genetic and environmental factors may cause male infertility.
Genetic causes of male infertility
Traditionally, genetic causes of male infertility have been sought at the level of chromosomal abnormalities, which may be detected in 2.1–8.9% of men attending infertility clinics (Hargreave, 2000). In one study of 150 men with severe oligospermia 
or azoopsermia, 85 men (57%) had unexplained male infertility and they accounted for 27 (75%) of the 36 men with genetic abnormalities (Dohle et al., 2002) (Table II). Of the 38 abnormalities in 36 men, 16 were abnormal karyotypes (eight XXY), eight were deletions in the azoospermia factor region, and 14 were cystic fibrosis transmembrane conductance regulator (CFTR) gene mutations. It is now known that structural anomalies of the Y chromosome, which cause deletion of the distal fluorescent heterochromatin in the long arm, are associated with severe abnormalities of spermatogenesis. From recent studies which have defined a family of genes on the Y chromosome involved in spermatogenesis, it has become clear that a substantial proportion of non-obstructive azoospermia and severe oligozoospermia is associated with deletions in these genes (Krausz et al., 2003). Y deletions are found almost exclusively in men with 
in 0.4% of infertile males, 4% of ICSI candidates, 11% of azoopsermic men, and 35% of men with idiopathic Sertoli cell only syndrome (Krausz et al., 2003). The most frequently deleted region is AZFc (approx 60%) followed by AZFb/b + c/a + b + c (35%); AZFa deletions are rare (5%).
|
The recent sequencing of the human Y chromosome, and the discovery that the chromosome fights entropy with palindromes (Rozen et al., 2003
; Skaletsky et al., 2003) promises new insights into the molecular basis of male infertility. Oxidative stress and DNA damage
One defined cause of defective sperm function is oxidative stress created by excessive generation of reactive oxygen species (ROS) by the sperm and/or the disruption of antioxidant defence systems in the male reproductive tract. Excess free radical generation may involve defective spermiogenesis with high levels of cytoplasmic retention and consequent ROS generation. The consequences of such oxidative stress include a loss of motility and fertilizing potential and the induction of DNA damage in the sperm nucleus (Aitken and Krausz, 2001). The causes and consequences of oxidative damage to the DNA in the sperm nucleus are still not known with certainty. The available evidence suggests that, in addition to a reduced chance of spontaneous pregnancy (Loft et al., 2003), and a reduced chance of a live birth following IVF/ICSI (Evenson and Jost, 2000; Larson et al., 2000), early pregnancy loss and morbidity in the offspring, including childhood cancer, may also be associated with such damage.
Occupational and environmental causes of male infertility
Occupational and environmental exposure of the adult male to potential testicular toxicants is now clearly emerging as a significant cause of abnormal spermatogenesis (Sharpe, 2000; Sharpe and Franks, 2002), with recent evidence linking, for example, occupational heat exposure, pesticide exposure and smoking (Zitzmann et al., 2003a) to diminished semen quality, sperm DNA damage, and unfavourable outcomes from assisted conception.
Developmental causes of male infertility
During the past two decades, a number of reports have appeared which have raised concerns about the development of reproductive problems in animals and man. A body of evidence suggests that the incidence of testis cancer is increasing, both in Europe and North America, and that this increase is related to birth cohort. At the same time, there is evidence of an increasing incidence of congenital malformations of the male genital tract, particularly cryptorchidism and hypospadias, which are clearly present at birth. Linked to these observations, there have been controversial reports of adverse changes in human semen quality, and more recently, clear evidence of geographical variation in semen quality. Substantial controversy remains, however: are the reported changes in male reproductive health genuine? what are the potential causes? and what are the implications for human reproduction? It has been suggested that the increases in incidence may be due to factors acting during intrauterine and early neo-natal life, as well as to changes in adult exposures and lifestyle. Whether or not these observations have clinical consequences remains to be determined (Skakkebaek et al., 2001; Sharpe and Franks, 2002; Sharpe, 2003). Clearly, however, there is a potential for the discovery of new causes of male infertility before the existing causes are well understood.
|
The assessment of tubal physiology |
|---|
Fallopian tube patency is necessary but far from the only tubal function necessary for successful conception. In addition to providing a conduit for gametes and embryos, the Fallopian tube controls sperm migration, regulates oocyte transport, promotes fertilization, nurtures gametes and embryos, and ensures the timely passage of the embryo to the uterus. No diagnostic test can assess these aspects of tubal physiology, and even the assessment of tubal patency is no more than descriptive. Until now the only indication that a Fallopian tube can function normally is the observation that an embryo has successfully implanted in the uterus. Current clinical tests can only give a picture of the internal or external anatomy, using radiology, ultrasound or visual techniques. Even laparoscopy, often described as a ‘gold standard’, is limited to an external description of normal uterine, tubal and ovarian relationships with an assessment as to whether there is any anatomical barrier to oocyte retrieval and transport.
Thus, evaluating tubal function has limitations that correspond to the limitations of assessing oocyte quality and sperm function. Semen analysis is a descriptive test that does not indicate that individual sperm can reach the site of fertilization, penetrate the oocyte investments, complete fertilization, or achieve formation of a pronucleus. Similarly, measurement of progesterone confirms the formation of a corpus luteum, but not the release from the ovary of a mature, fertilizable oocyte.
Until specific tests of tubal function become available, two questions about tubal patency tests are important for future clinical practice. First, is it likely that more sensitive and meaningful tests will be developed? Second, are current tests, limited as they are to the assessment of tubal patency, relevant to the assessment of every infertile couple?
Physiology of the Fallopian tube
Much is known about the physiology of the Fallopian tube, and how it achieves its key functions of regulating gamete transport and promoting fertilization, as well as nurturing gametes and embryos during their time in the tubal lumen (Croxatto, 2002). Although there is evidence that fertilization can take place in the uterine cavity, the Fallopian tube is not just a conduit between the ovarian follicle and the uterus, but is the site of a complex interaction between sperm, epithelium and luminal fluid. Sperm adhesion to oviductal epithelium can prolong the fertile life of sperm as well as delay capacitation such that even minimal damage to the luminal cells could affect sperm function. Transport of the cumulus–oocyte through the ampulla in women is rapid, at most a few minutes, and the oocyte is held at the ampullary–isthmic junction for several days, before the zygote proceeds through the isthmus to the uterus in a matter of hours. Ectopic pregnancy occurs never or rarely in species other than women, mainly because embryos which are retained in the oviduct lose their viability. Although the mechanisms of tubal transport are orchestrated by the ovarian hormones, their action modulates a complex response of muscular cells to myotrophic agents, such as neuropeptides, prostaglandins, endothelins and nitric oxide. Oviductal fluid also varies in composition throughout the ovarian cycle and plays an important role in both transport and nutrition. It remains unclear whether the embryo controls events such as entry into the uterus, and if it does, the extent of the control is unknown (Croxatto, 2002).
Current methods of tubal assessment
As noted above, clinical tests are better at assessing tubal patency than the physiology involved in tubal transport, fertilization or embryo nurturing functions. Tubal function is implied by recovery of sperm from the tubes and the peritoneal cavity and by observing the transport of labelled particles from the vagina through the uterus and tubes, but these tests have not been adopted in clinical practice (Templeton and Mortimer, 1982; Wildt et al., 1998).
Screening tests for tubal pathology include chlamydial antibody testing, hysterosalpingogram and hystero-contrast-sonography. Chlamydia antibody presence is associated with tubal pathology; although a positive test requires follow-up evaluation, if the test is negative, tubal pathology at laparoscopy is unlikely (7–12%) (Land et al., 2003).
Hysterosalpingogram has been used for decades, and the relationship to tubal pathology at laparoscopy is similar to that of chlamydia antibody testing, at least with respect to tubal patency: an abnormal result should be confirmed by laparoscopy, but with a normal result laparoscopy is likely to show tubal patency in 95% of cases (Swart et al., 1995). Recently attempts to avoid radiation have resulted in the development of hystero-contrast-sonography in which contrast enhancement enables tubal visualization by ultrasound. Reports initially indicated that the results were at least as good as hysterosalpingography, although pain is more frequent with the sonographic procedure (Dijkman et al., 2000; Stacey et al., 2000).
Laparoscopy and dye transit is the standard assessment for the evaluation of tubal pathology and provides the only practical means of assessing tubal and pelvic surfaces. Nevertheless, technical problems can cause failure of dye transit. The place of laparoscopy as a ‘gold standard’ for tubal pathology is compromised because the procedure is only a moderately good predictor of live birth, which is the desired infertility outcome for the infertile couple (Mol et al., 1999). The place of transvaginal hydrolaparoscopy, although the subject of an increasing number of studies, needs further assessment.
The possible therapeutic effect of diagnostic tubal patency tests is the subject of a systematic review (Johnson et al., 2003), which concludes that there is some benefit from tubal flushing with oil-soluble contrast media, although the effect is small and does not warrant widespread use if oil-soluble contrast material. In cases of apparent proximal tubal blockade, selective salpingography and tubal catheterization may clarify whether the proximal occlusion is due to cornual spasm, but whether these have any prognostic value remains unclear as there are no controlled trials and neither procedure is without risk.
Current place of tubal assessment
Chlamydia testing can serve as an initial screening test: if negative, a visual test of tubal patency can be delayed for a few months while simple therapy is used. If chlamydia testing is abnormal, hysterosalpingogram or hysterosonography are indicated. Laparoscopy can be reserved for cases in which these screening tests for tubal patency are abnormal, although it is indicated as the initial investigation when the history suggests a high risk of tubal or peritoneal pathology (Table III).
|
The continuing high percentage of couples with unexplained infertility underlines the limitations of current investigations to the extent that certain patient characteristics (women's age, pregnancy history and duration of infertility) have become as important prognostic indicators as the presence or absence of apparent impediments to conception. In this light, the place of hysterosalpingogram and particularly laparoscopy, an invasive procedure requiring general anaesthetic, should be seriously considered in each individual case and the question asked: Is the result of this investigation going to alter future management? If future management is likely to be IVF or ICSI, then laparoscopy may not be relevant. Significant hydrosalpinges can be detected ultrasonically.
Summary of tubal assessment
Clinical tests of tubal function, as for other fertility investigations, are limited in what they can assess. They are mainly descriptive and do not begin to assess function. Normal appearances do not necessarily mean normal physiology. On the other hand the finding of gross tubal pathology can facilitate the decision to proceed to IVF where appropriate. Research is needed to improve understanding of the conditions that foster fertilization and embryo development. Recognizing the limitations of the current screening and reference tests, research is also needed to develop non-invasive investigations of tubal function.
|
The association between endometriosis and infertility |
|---|
Reasonable patients request well-founded, effective diagnostic tests and treatments, but in the case of endometriosis especially, hard evidence is rare since many of the diagnostic procedures are not based on reliable scientific data and many treatments lack objective evaluation.
When considering fertility in endometriosis patients, two clinical questions predominate: Do endometriosis patients suffer from impairment of fertility? If so, does treatment improve their pregnancy chances? From applying Koch's postulates to the endometriosis literature Wheeler and Malinak (1988) deduced that, in a clinical and epidemiological sense, there is insufficient scientific evidence for endometriosis and impaired fertility to be causally related.
Fertility in patients with endometriosis
Whether endometriosis decreases fertility is a question that can be studied in a prospective observational cohort study of two groups of patients, one with and one without documented endometriosis. Since this study would require a laparoscopy in perfectly healthy young women even before they would start attempting to achieve a pregnancy, a study like this will obviously never be performed.
The ideal study design being unattainable, several less robust approaches remain. One approach would compare spontaneous pregnancy rates in patients with unexplained subfertility to those in patients with unexplained subfertility who also have the finding of endometriosis at laparoscopy. Another way to assess spontaneous pregnancy rates in endometriosis patients is to study untreated control patients participating in randomized controlled trials (RCT). There have been ten studies reported from which figures like these can be derived (Telimaa et al., 1987; Thomas and Cooke, 1987; Bayer et al., 1988; Mahmood and Templeton, 1990; Fedele et al., 1992a,b; Overton et al., 1994; Marcoux et al., 1997; Tummon et al., 1997; Gruppo Italiano per lo Studio dell, 1999). These trials involved 462 untreated control patients, typically followed for 6–12 months, and the aggregate pregnancy rate was 27% [95% confidence interval (CI) 22–32] (Table IV). In six trials of unexplained infertility treatment involving 444 untreated control patients, typically followed for 3–6 months, the aggregate pregnancy rate was 13% (95% CI 10–17) (Harrison and O'Moore, 1983; Fisch et al., 1989; Deaton et al., 1990; Glazener et al., 1990; Fujii et al., 1997; Guzick et al., 1999). The aggregate spontaneous pregnancy rate in 20 studies involving 2026 patients with unexplained infertility, followed typically for >2 years, was 33% (95% CI 31–35) (Taylor and Collins, 1992). Thus, the likelihood of conception without treatment is not materially different between endometriosis and unexplained infertility.
|
Studying the results of artificial insemination in women with azoospermic husbands offers an alternative way to address the issue. Three studies may offer some insight in this respect (Portuondo et al., 1983
; Jansen, 1986; Rodriguez-Escudero et al., 1988). The crude spontaneous pregnancy rates during 12 months of follow-up ranged from a low figure of 29% in the few patients in one study (2/7; Jansen, 1986) to a ‘normal’ conception rate of 81% in another (17/21; Rodriguez-Escudero et al., 1988) in women with endometriosis, as compared to 51% in women without (Jansen, 1986); overall, pregnancies occurred in 38/59 cases (64%) and 46/91 controls (51%).
Profertility treatments in patients with endometriosis
Five RCT have shown that medical treatment does not increase the likelihood of pregnancy in subfertile endometriosis patients (Thomas and Cooke, 1987; Bayer et al., 1988; Telimaa et al., 1987; Telimaa, 1988; Fedele et al., 1992a). The issue of surgical treatment, however, is more delicate. Two studies, one from Canada involving 341 patients and one from Italy involving 101 patients, have been published reporting surgical removal of endometriosis and adhesions, if present (Marcoux et al., 1997; Gruppo Italiano per lo Studio dell, 1999) in subfertility patients with mild or minimal endometriosis. The Canadian study is by far the largest published RCT on any treatment of endometriosis in subfertility patients to date. It offers evidence that surgical treatment of minimal and mild endometriosis increases the likelihood of ongoing pregnancy, although fertility is not restored to normal. During the 36 weeks of follow-up, 63 (37%) of the 172 patients randomized to have surgery conceived and 50 pregnancies (29%) continued to 20 weeks. In the 169 randomized to no treatment, 37 (22%) conceived and 29 (17%) continued to 20 weeks (Table IV). The 22% pregnancy rate at 36 weeks in the untreated patients from this surgical RCT is the third lowest among the ten trials shown in Table IV. The Italian study included 101 patients. During the 12 months of follow-up, 12 (22%) of the 54 patients randomized to have surgery conceived and 10 (19%) had a delivery. Of the 47 randomized to no treatment, 13 (28%) conceived and 10 (21%) had a delivery. Results among patients who did not have adhesions or did not need lysis of adhesions performed would specifically address whether it is the removal of endometriosis lesions that improves fertility. In the Canadian study, 284 women did not have adhesions. In these, the destruction of implants also increased the 36 week cumulative probability of a pregnancy that lasts beyond 20 weeks (cumulative incidence ratio, 1.6; 95% CI 1.1–2.5; P < 0.05). Unfortunately the published results of this and other trials do not allow for calculation of the pregnancy rates in patients following surgical removal of endometriosis lesions only.
When the results of the Canadian and Italian studies are combined, assuming that on going pregnancy results correspond closely to live birth, the aggregate control birth rate is 18% (95% CI 13–25). The rate in treated patients is 8% higher (95% CI 1–16, P=0.03; P for heterogeneity = 0.13). With this 8% difference the number needed to treat (NNT) would be 12 (95% CI 6–122),indicating that for every 12 women undergoing laparoscopy that have ablation of minimal or mild endometriosis lesions there will be one additional pregnancy, compared with not doing ablation. Although small, the benefit seems important, given that it is relevant to decisions taken during a laparoscopy which is already under way, when the additional risks are small. In planning laparoscopy procedures, however, one cannot identify which women will have minimal or mild endometriosis before surgery. If the proportion of patients undergoing laparoscopy in a given clinical practice is 20%, then the NNT is 12/0.2, or 60 and if it is 50%, the NNT is 12/0.5, or 24 (Crosignani and Vercellini, 2000). This benefit is too small to be a deciding factor when considering whether to do a laparoscopy.
It therefore seems reasonable to conclude that there exists no or only very limited evidence to support the theory that endometriosis per se causes subfertility, that there is no support for the contention that medical treatment of minimal and mild endometriosis improves pregnancy chances in subfertile couples, and finally that there is statistical evidence for a slight beneficial effect of surgical removal of the lesions, but the size of this effect is small and may be short-lived.
|
Detrimental effects of smoking on fertility |
|---|
In women, smoking reduces conception rates in natural cycles. A meta-analysis of 12 studies on smokers found an odds ratio (OR) of 1.60 (95% CI, 1.34–1.91) for the risk of infertility compared to non-smoking women (Figure 2). There is a remarkable consistency of effect across the 12 studies.
|
Similarly studies of subfertile women undergoing IVF show a decreased fecundity among smoking women: OR 0.66 (95% CI 0.49–0.88) for pregnancies per number of cycles (Augood et al., 1998
). In IVF cycles the deleterious effect of smoking becomes apparent in older women (Zenzes et al., 1997). Smoking also increases the risk of miscarriage (Hughes and Brennan, 1996) and the risk of multiple pregnancy (Parazzini et al., 1996).
In men, smoking interferes with several aspects of fertility. By impairing the vascular endothelial function, male smoking causes erectile dysfunction and can be considered as an early sign of atherosclerosis (Feldman et al., 2000). The risk of developing erectile dysfunction is almost doubled for smokers compared to non-smokers (Mirone et al., 2002).
Male smoking may also affect semen variables, although despite several studies on semen variables in smokers and non-smokers, this issue is still controversial. Even recent studies carried out among subjects from similar cultural backgrounds found either no effects (in Austria) (Trummer et al., 2002) or significant decreases in sperm count and motility (in Switzerland) (Künzle et al., 2003) in smoking versus non-smoking infertile men.
With respect to conception, male smoking appears to impair the natural conception rate slightly (Curtis et al., 1997). Also, several studies report the negative effect of male smoking on IVF results (e.g. Klonoff-Cohen et al., 2001; Zitzmann et al., 2003a). Furthermore, recently it has been shown that male smoking almost halves the success rate in ICSI (Zitzmann et al., 2003a) (Table V). The reduction in success is correlated with the number of cigarettes consumed daily. Since the damaging effect disappears in ex-smokers after 2 years, there is an incentive to persuade would-be ICSI fathers to quit smoking (Zitzmann et al., 2003b).
|
Male smoking also damages sperm DNA, which may explain the lower birthweight of new-borns (Klonoff-Cohen et al., 2001
) and the higher malignancy rate in children of smoking fathers (Ji et al., 1997; Sorahan et al., 1997). In conclusion, male cigarette consumption is not only a risk for the smokers themselves but also for their paternity, their partners and their offspring.
|
The impact of missing information on management of infertility |
|---|
It remains uncertain whether missing information about fertility is merely a theoretical problem or whether it has an impact on the management of infertility. Given that there are large gaps in knowledge of oocyte quality, sperm function and tubal physiology, and that the correlation between fertility and a non-reproductive factor such as smoking status is stronger than the correlation with a disease like endometriosis, an effect on fertility might seem all too likely.
One way of estimating the impact of missing information is to compare actual success of current management with ideal success. Assuming optimal treatment, the gap between actual success and 100% success could be said to indicate the extent of missing information about the causes of infertility, the establishment of a correct diagnosis, and the selection of specific treatment. Another approach would be to estimate the frequency of persistent infertility after the use of appropriate, specific treatment. Of course, the value of assessments about the impact of treatment cannot be judged without understanding why couples sometimes fail to complete what clinicians might view as optimal treatment. These three topics comprise the remaining sections of this review.
|
The maximum prognosis with current infertility treatment |
|---|
The maximum prognosis with current treatment involves the likelihood of conception among couples with a broad range of diagnoses and treatments. This is reported infrequently, but even in the most recent report from Shanghai in 2002, the live birth rate was only 35% among 394 couples followed for
2 years (Che and Cleland, 2002).
A recently published model which estimated the overall live birth rate with the use of an optimal treatment protocol may be useful to assess what might be the maximum prognosis with infertility treatment (Collins and van Steirteghem, 2004). In the model the diagnostic groupings and appropriate treatments were based on comprehensive reviews (ESHRE Capri Workshop Group, 1996, 1997, 2000; Philips et al., 2000). The duration of each treatment and the proportion of couples receiving treatment were derived from prognostic studies in untreated couples (Eimers et al., 1994; Collins et al., 1995; Snick et al., 1997). Assisted reproductive technology utilization was set arbitrarily at three levels: 3, 10 or 50% of the couples in each diagnostic category that had no live birth after conventional treatment. The assisted reproductive technology utilization rates represented North American (3%), Northern European (10%) and an optimal (50%) level of utilization, have been suggested in a recent review (ESHRE Capri Workshop Group, 2000). The model estimated total live births, singleton live births and multiple live births for each diagnostic category and treatment per 10 000 couples.
The results showed that typically 59% of infertile couples seeking medical care would have treatment. The overall live birth rate during a 3 year period with non-assisted reproductive technology treatment only would be 37%. If assisted reproductive technology were utilized by 3, 10 and 50% of couples with persistent infertility in each diagnostic category, live birth rates would be 39, 43 and 47% respectively (Table VI).
|
The maximum prognosis was estimated in supplemental analyses hypothesizing that all couples would have treatment and utilize assisted reproductive technology at even higher rates than the optimal 50% rate. When all couples receive treatment, the live birth rate rises to 59% with 50% assisted reproductive technology utilization and to 76% with 100% utilization, in each case with a three cycle assisted reproductive technology protocol. With 100% assisted reproductive technology utilization and a six cycle assisted reproductive technology protocol, 61% of 10 000 couples would utilize 7778 cycles per 106 population per annum and the hypothetical overall live birth rate would be 94%. However, 23% of the births would be multiple, and assisted reproductive technology would account for 76% of the multiple births.
An optimal protocol of current treatment for infertility would fall short of complete success even if it were offered to all couples. Overall success rates can be raised >50% only by excluding untreated couples and hypothesizing extensive deployment of empiric assisted reproductive technology treatments. The models show that despite the use of an intense treatment protocol, many infertile couples would not conceive within a reasonable 3 year interval and we assume that some would never conceive. Thus, specific treatment and powerful empiric treatments cannot overcome the underlying unknown and therefore untreated factors that remain as barriers to greater overall success in the treatment of infertility.
Missing information about the true causes of infertility severely limits the prognosis for successful conception and live birth, hinders the selection of specific treatment, and contributes to the relatively high prevalence of persistent infertility.
|
Persistent infertility after conventional treatment |
|---|
A frustrating problem in the treatment of infertility associated with ovulation defects, male problems, tubal disease or endometriosis and with unexplained infertility is the persistent failure to conceive after multiple attempts with the conventional treatments such as ovarian stimulation or intrauterine insemination. Theoretically, specific treatments such as induction of ovulation for anovulation should restore normal fertility, but that does not happen, as many couples fail to conceive. Treatment failure might be a consequence of an inaccurate or incomplete diagnostic assessment or it might be due to defects at three different levels, fertilization, embryo quality and implantation. Such abnormalities cannot be detected with the conventional diagnostic tests, although the existence of problems with fertilization and embryo quality may become clear during IVF procedures.
Diagnostic failure
Unexplained infertility is a diagnosis made by exclusion of known aetiologies in couples who have not conceived and in whom standard investigations have not detected any abnormality. It accounts for 8–28% of couples with infertility. The proportion of couples with unexplained infertility is related to the extent of diagnostic tests performed to uncover putative causes of the delay in conception. In couples with unexplained infertility, the chances of spontaneous conception relates to the duration of infertility and to the age of the woman. Even in these couples with no abnormal diagnostic tests, the cumulative pregnancy rates are 
33–60% at 3 years and 64% at 9 years (Templeton and Penney, 1982; Collins et al., 1983).
Similarly abnormal results from semen analysis often fail to explain infertility. For example, among 120 men treated by ICSI for severe oligozoospermia, four (3%) had a subsequent pregnancy without treatment (Silber, 2001). Among 200 couples interviewed 4 years after one to four failed ICSI cycles, there were 23 deliveries (11.5%) without further treatment. All conceiving female partners were <32 years of age. Thus, even the most severe semen analysis results do not exclude the possibility of conception.
Fertilization failure
Successful fertilization is generally assumed to depend on gamete quality and normal tubal function. Tubal function is only partly assessed by tubal patency tests and tubal appearance. Oocyte quality cannot be measured and semen parameters, although they correlate with fertility, do not actually measure sperm quality. Fertilization failure may be due to oocyte deficiencies, impaired sperm or defects in the sperm–oocyte interaction.
IVF offers the opportunity to test the quality, or fertilizing potential, of the gametes. In 5–11% of all IVF cycles there is no fertilization of any oocytes (Ruiz et al., 1997; Bhattacharya et al., 2001; Repping et al., 2002). In 30% of cases the fertilization failure recurs in a subsequent IVF attempt (Barlow et al., 1990; Hamberger et al., 1995). Failed fertilization may be due to impaired sperm, oocyte deficiencies or defects in sperm–oocyte interaction (Mahutte and Arici, 2003). Standard semen analysis has limited ability to predict fertilization failure as judged by receiver-operating characteristics analysis (ROC: area under the curve 0.75). Including additional parameters such as pre- and post-wash total motile sperm count, the male partner's age and the number of oocytes only slightly improved the prediction (area under the curve 0.80) of fertilization failure (Repping et al., 2002).
ICSI, which bypasses sperm-related factors, may differentiate between sperm and oocyte defects. When failed fertilization occurs in 3% of ICSI cycles, it is due mainly to failure of oocyte activation (Mahutte and Arici, 2003). Several studies have compared IVF with ICSI in couples with non-male factor (Aboulghar et al., 1996a; Moreno et al., 1998; Bukulmez et al., 2000; Bhattacharya et al., 2001) or unexplained infertility (Aboulghar et al., 1996b). No significant differences were found in the fertilization or pregnancy rates between IVF and ICSI, except in cases with moderate oligoteratoasthenozoospermia, in which ICSI was associated with significantly better results (Tournaye et al., 2002). Nevertheless, ICSI could prevent the occurrence of complete fertilization failure, which was observed in 5/22 couples with unexplained infertility (Aboulghar et al., 1996b; Ruiz et al., 1997). It has been proposed in these cases to split the oocytes between IVF and ICSI (Hamberger et al., 1998), which would serve therapeutic but also diagnostic purposes. Since then, higher quality evidence has emerged in the form of a large trial in non-male infertility (Bhattacharya et al., 2001). Failed fertilization rates were 5 and 2% respectively in IVF and ICSI cycles. The difference was neither significant nor large enough to be clinically important: it would require 33 cycles to have one less failed fertilization with ICSI than with IVF.
Thus, failed fertilization is not common and ICSI can reduce the sperm contribution to failure, but the oocyte and sperm defects that might cause failed fertilization have yet to be clarified.
Poor embryo quality
Another potential cause of persistent infertility that can possibly be unveiled by IVF is embryo quality. In preimplantation genetic diagnosis (PGD) for aneuploidy screening with fluorescence in situ hybridization (FISH), 41–54% of the normal-looking embryos from couples with multiple (>3) IVF failures were found to be aneuploid (Gianaroli et al., 1999; Kahraman et al., 2000), whereas, using comparative genomic hybridization, 60% of the embryos had chromosomal abnormalities (Voullaire et al., 2002) (Table VII). Higher rates of chromosomal abnormality might be expected in the embryos from translocation carriers, who represent 0.6% of infertile couples and 3.2% of multiple IVF failures compared to 0.2% in the general population (Hook and Hamerton, 1977; Stern et al., 1999). The overall incidence of aneuploidies in the preimplantation embryos from translocation carriers of Robertsonian translocation was 77%, only somewhat higher than in the above studies of repeat IVF failure, while from those with reciprocal translocation was 89%. (Gianaroli et al., 2002). Interestingly, presence of chromosomal anomalies in preimplantation embryos correlates with the occurrence of embryo fragmentation (Alikani et al., 1999).
|
Implantation failure
The third reproductive process that may account for persistent infertility is implantation failure, which remains poorly understood. Several groups of genes and gene families are involved in the development of receptive endometrium as well as in the implantation process (Kao et al., 2002; Riesewijk et al., 2003). Moreover, during the normal window of implantation, some genes are up- and others down-regulated. It is therefore tempting to speculate that dysregulation of certain genes may be associated with multiple implantation failures.
In conclusion, several possible mechanisms could explain persistent infertility after conventional treatment: inadequate diagnostic assessment, defective gametes or gamete interaction leading to fertilization failure; poor embryo quality; and failure of good embryos to implant. IVF and ICSI might identify which patients have failures, but not the cause of those failures, and some processes such as endometrial receptivity and implantation remain beyond assessment. Against this background, age is potentially the most understandable, if not the most common, reason for persistent infertility.
|
Why do couples fail to take up assisted reproductive technology when the costs are covered? |
|---|
Currently, birth rates after assisted reproductive technology are
25% per cycle, which is comparable to in vivo fertilization. Despite this level of success, assisted reproductive technology is not accepted by many infertile couples, although this might be their only means of conception (Schlaff, 2003). Different factors may account for primary rejection of assisted reproduction treatment and secondary rejection after one or more assisted reproduction treatments (Hughes and Giacomini, 2001; Collins, 2001). Primary rejection of assisted reproduction treatment in cases of infertility
Medical factors
These may include aversion to needles and pain (hormonal injections, ovum aspiration), fear of side-effects (hyperstimulation, multiple pregnancies, malformations and premature birth) and lack of confidence in the outcome due to poor prognostic factors (female age, duration of infertility and history of multiple unsuccessful infertility treatments).
Psychosocial factors
These may include conflict between work and time-consuming treatment, mental stress during treatment and the anticipation of disappointment at the end of the cycle (because 70% of all patients fail to become pregnant in a single cycle). Family and social attitudes against the use of in vitro methods may influence some couples.
Ethnic and religious factors
These can also be the cause of the rejection of assisted reproductive technology, for example by some strict religions.
Secondary rejection of assisted reproduction treatment
For couples who have already had one or more assisted reproduction treatment cycles without success, additional factors may play an i