Abstract
Unexplained infertility refers to the inability to conceive within 12 months of unprotected intercourse, not attributable to any known causes of infertility such as ovulatory dysfunction, reduced sperm quality, tubal pathology or other causes. Treatment for unexplained infertility can be done predominantly through intrauterine insemination with or without hyperstimulation or in vitro fertilisation. Given that these treatments are utilised to improve likelihood of conception in relation to the couple’s chances of spontaneous pregnancy, rather than targeting any specific pathology, a comparison should be drawn between these treatments and their natural conception prognosis. Utilisation of prognostic models can allow differentiation between those likely to benefit from immediate treatment from such individuals who have reasonable natural conception prognosis and thereby can delay treatment for 6 months in hopes of spontaneous pregnancy. This comparison is valuable given the aforementioned treatments have implications for both the woman and her future child, and the cost of such procedures also compromises care accessibility.
6.1 Introduction
Infertility, which refers to failure to conceive within one year of having regular unprotected intercourse, occurs in one of every six couples [1]. Infertility can broadly be categorised into tubal pathology, anovulation, reduced sperm quality and unexplained infertility.
From planning pregnancy to eventually seeking treatment for infertility, couples tend to follow a typical pattern of progression. In those desiring pregnancy, approximately 84% will have natural conception within the first 12 months of unprotected intercourse [2]. A proportion of unsuccessful couples then go on to seek fertility assistance. Generally, the typical fertility consultation involves conducting a comprehensive history and examination of the couple to determine if there is a medical cause for which the infertility can be attributed. Fertility investigations follow focusing on whether the gametes can meet each other, namely to identify issues with semen, ovulation and transport, that is, tubal patency or the pelvic cavity [3]. Following the work-up, only about 60% of cases have an identifiable cause and lead to a diagnosis of male and/or female subfertility, while the other 40% have no detected abnormalities, leading them to be labelled as so-called unexplained infertility [4].
There is consensus among the European Society of Reproductive Medicine (ESHRE) and National Institute for Health and Care Excellence (NICE) in their guidelines for infertility treatment that couples first be informed about their chances of spontaneous conception and consequently not be subjected to treatments that are ineffective or pose needless potential harm [3, 5]. Before a treatment is initiated, it must be determined whether it is appropriate for the specific couple and if potential benefits offset the risks of harm. Couples with regular cycles post-fertility work-up can have their probability for natural conception estimated using the Hunault synthesis prediction model [6, 7]. For couples with unexplained subfertility with low likelihood for spontaneous conception, medically assisted reproduction becomes warranted when treatment is expected to improve chances of achieving pregnancy above their chance naturally [3]. The rationale for which treatment should be approached is to view it in terms of prognosis rather than diagnosis; the question becomes ‘what will occur in the near future?’ as opposed to ‘what is causing the infertility?’.
6.2 Prediction Models
6.2.1 History of Prediction Models
To date, there are nine published prediction models pertaining to natural conception in infertile couples [6, 8, 9–13]. The first model was developed in 1993 by Bostofte et al. in a Danish cohort utilising three variables: subfertility duration; sperm penetration test and type of female subfertility, which could either be none, ovulatory or cervical condition, an anatomical issue or a mixture of conditions [8]. Most recently, a group in the Netherlands developed the synthesis model using combined data from the cohorts of the Snick, Collins and Eimers studies [9–11]. The synthesis model of Hunault et al. considers a number of different variables, namely female age, subfertility duration, whether subfertility is primary or secondary, percentage sperm motility, whether referral was by a specialist gynaecologist or general practitioner and optionally the post-coital test results [6].
At present, no current fertility treatment guidelines promote the use of prediction models to guide clinical practice, with perhaps one contributing factor being that validation within external populations of such prediction models has only recently occurred [14]. Table 6.1 summarises the different prognostic factors used in different research models to predict natural conception in couples with infertility.
Table 6.1 A comparison of natural conception prediction models for infertile couples
| Authors | Prognostic factors |
|---|---|
| Bostofte et al. (1993) [8] |
|
| Eimers et al. (1994) [11] |
|
| Bahamondes et al. (1994) [12] |
|
| Wichmann et al. (1994) [13] |
|
| Collins et al. (1995) [10] |
|
| Snick et al. (1997) [9] |
|
| Hunault et al. (2004) – Synthesis model [6] |
6.2.2 Performance of the Synthesis Model for Natural Conception
The limitation in many predictive models is that conclusions drawn are largely specific to the population in which the model was developed, with overly optimistic outcomes being drawn when employed in populations which differ from the original [15]. As such, both internal and external validation of any prediction model should be conducted before considering use in clinical practice [16, 17].
Internal validation involves utilising the population on which the prediction model is based to determine what would constitute over-optimistic estimates and consequently adjust the model accordingly. Unlike with previous models, the more recent predictive models have had internal validation applied [8, 11–13]. When assessing prediction models, however, external validation is of greater practical value given that the model can be prospectively applied in a different population than that in which it was developed, albeit at great financial cost and with significant time consumption [16].
The process of external validation was applied to the Hunault natural conception prediction model, which demonstrated that the model’s calibration was nearly flawless [7]. The study population involved 3,000 subfertile couples, 537 (18%) of whom had spontaneous ongoing pregnancy, 55 (2%) had pregnancy loss, 1,316 (44%) commenced treatment, 820 (27%) were neither pregnant nor initiated treatment and 280 (9%) were lost to follow-up. Cumulative ongoing pregnancy rates in couples with prognosis ≥40% was 46% without treatment, which paralleled the discriminative capacity of the original sample population in which the model was based (c-statistic: 0.59).
Despite there being a common tendency to dwell on the principle of discrimination (differentiating between those who will and will not spontaneously conceive), the emphasis should be on determining the likelihood that a couple will conceive. Approached from this context, improvements can be made on the current Hunault prediction model by incorporating further prognostic factors such as a couple’s comprehensive fertility history, the woman’s body mass index (BMI), cycle duration, basal follicle-stimulating hormone (FSH) level and a semen analysis. Additional potential prognostic factors were explored in the analysis of an extended cohort [7, 18]. Factors which enhanced prediction of natural conception include history of pregnancy within current relationship, cycle duration, BMI, sperm volume, concentration and World Health Organization (WHO) morphology. Figure 6.1 depicts the additional prognostic indicators used in conjunction with those in the Hunault synthesis model for natural conception for infertile couples.
Figure 6.1 Prognostic factors for predicting natural conception in couples with subfertility. The likelihood of conceiving within one year can be predicted using various prognostic factors. Factors chosen in the Hunault synthesis model (boxes) have been externally validated and have shown good predictive capability for estimating the chance for natural spontaneous conception. Additional prognostic factors (circled), when combined with the Hunault synthesis model, have shown greater predictive capability than the synthesis model alone. BMI, body mass index; FSH, follicle-stimulating hormone.
6.3 Treatment of Unexplained Infertility
At present, the likelihood of conceiving can be naturally predicted for couples with unexplained infertility, which leads to the clinical question of which couples should be treated and when this should occur. The two predominant methods employed in the treatment of unexplained infertility are intrauterine insemination (IUI) with or without ovarian stimulation and in vitro fertilisation (IVF).
6.3.1 Intrauterine Insemination
The process of IUI begins with monitoring of the ovulatory cycle with ultrasound or urine testing for the luteinizing hormone (LH) surge. When ovulation is anticipated, the insemination is planned. This starts with processing of semen through laboratory ‘washing’, such that motile spermatozoa can be prepared and concentrated into a small volume for delivery into the uterus. Sperm is delivered directly into the uterine cavity via a small catheter, thereby bypassing the cervix entirely. The premise of IUI is to increase the likelihood of fertilisation by delivering the sperm into the uterus such that it is in close proximity to the released oocyte. The success of IUI is critically dependent on coinciding the procedure with ovulation; hence cycle monitoring is employed to determine timing precisely.
Additional means of improving timing of IUI with ovulation is to perform it together with mild ovarian hyperstimulation (MOH). MOH is typically achieved through use of clomiphene citrate (CC), letrozole or subcutaneous injection of gonadotropins. The premise of MOH is to improve accuracy of timing and to increase the number of oocytes available to be fertilised. Comparisons between IUI (with or without MOH) efficacy relative to expectant management have been drawn and it is a topic currently fiercely debated with no unanimous consensus. Concerns have been expressed regarding multiple pregnancy rates when IUI is augmented with MOH. A recent Cochrane review comprising 14 trials with a total population size of 1,867 women failed to demonstrate that women with unexplained infertility treated with IUI (irrespective of MOH use) compared to expectant management or timed intercourse differed significantly in terms of rates of live birth or multiple pregnancy [19]. The treatment effect is likely to be dependent on prognosis. While in couples with a poor prognosis for natural conception MOH IUI seems to have additional value, its additional effect in women with good prospects for natural conception is limited [20, 21].
6.3.2 In Vitro Fertilisation
IVF use has expanded over the years and it has been increasingly utilised in all causes of subfertility despite there being a distinct lack of data regarding efficacy in such cases. Together with the mounting evidence of IVF use being associated with adverse health outcomes in children born through such means, a frank discussion must be had regarding its use.
Conducting IVF involves a standard regimen, notably beginning with controlled ovarian stimulation (COS), followed by oocyte retrieval under ultrasound guidance, laboratory insemination, culturing of embryo/s and lastly transcervical replacement of embryos during either the cleavage or blastocyst stage. IVF is an invasive therapy that does carry risk for numerous complications, including multiple pregnancy and ovarian hyperstimulation syndrome (OHSS). Despite its risks, use of IVF in unexplained infertility may be warranted to overcome unknown biological shortcomings contributing to infertility by bypassing various in vivo steps which may be impairing the ability to conceive. Such factors may include ovarian dysfunction, cervical factors, issues with sperm and egg transport and sperm–egg interaction.
Hughes et al. conducted a randomised controlled trial (RCT) comparing IVF with expectant management [22]. The trial involved 139 couples and compared one cycle of IVF with 3 months expectant management, with results demonstrating higher live birth rates in the IVF-treated group. The evidence presented in the study, however, was considered to be low quality [23]. In particular, the average duration of subfertility among the patients was 4 years, which is a long duration relative to other studies drawing comparisons between treatment and expectant management in unexplained infertility. Furthermore, prospective studies have demonstrated that timed intercourse must be undertaken across a minimum of six cycles, not simply 3 months, to adequately cover the interval in which the majority of conceptions theoretically occur [24].
6.3.3 Treatments Preceding IUI and IVF
6.3.3.1 Tubal Flushing
Hysterosalpingography (HSG) refers to a diagnostic technique used to assess tubal patency that has been commonly used in the outpatient setting for this purpose since the establishment of the method in 1914. There has been speculation since the 1950s that potential fertility benefits are linked to HSGs, whereby the act of flushing tubes directly contributes to increased rates of pregnancy in the months post-procedure. Numerous studies have since demonstrated fertility benefits are derived specifically from HSG using oil-based contrast mediums; however, only three such studies were RCTs [26–28]. Analysis of pooled data from the RCTs suggested higher ongoing pregnancy rates in women who received an HSG with oil-based contract relative to no intervention (odds ratio [OR] 3.6; 95% confidence interval [CI] 2.1–6.3) [29]. A meta-analysis was conducted to compare how different types of HSG contrasts, notably oil- and water-based contrast, differ in terms of pregnancy rates [29]. The analysis included five RCTs and demonstrated that although there were higher ongoing pregnancy rates in the oil-based group compared to the water-based groups, the results were not statistically significant (OR 1.4; 95% CI 0.8–2.5). There should be caution with regards to interpretation of these results, given the studies included in the meta-analysis were deemed low-quality studies associated with a large degree of uncertainty regarding the result estimates.
Given the ambiguity surrounding the effects of the two different HSG contrasts on pregnancy rates, a multicentre trial [30] randomised infertile women were randomised to HSG tubal flushing with either oil-based contrast medium (Lipiodol®) or water-based contrast medium (Telebrix Hystero®). The trial demonstrated both higher ongoing pregnancy rates (39.7% vs. 29.1%, respectively) and live birth rates (38.8% vs. 28.1%) in those receiving oil-based contrast.
6.3.3.2 Lifestyle Interventions
Obesity undoubtedly constitutes a large public health issue overall, especially if one considers that 20% of reproductive-aged women in developed nations are obese, but it also specifically impacts women desiring pregnancy given associations between obesity and anovulation, menstrual abnormalities and infertility in general [31–35]. Numerous guidelines on infertility treatment of obese women recommend the first step of management is to aim for a 5%–10% reduction of their weight [36, 37]. Despite such a recommendation being quite commonplace, there is a noticeable scarcity in studies investigating the impact on pregnancy outcomes with effective weight loss in accordance with recommended goals.
A multicentre RCT of comprising 577 obese infertile women aged between 18 and 39 [38] involved randomisation of 290 women to a 6-month lifestyle programme followed by 18 months of fertility treatment and the remaining 287 women to immediate fertility treatment over 24 months. Fertility treatment used in the study occurred as per Dutch infertility guidelines, with women being assigned treatment modalities based on presence of anovulation, the natural conception prognosis as per the Hunault synthesis model and by what treatments they had failed in the past. Women in this study thereby received treatment with ovulation induction, expectant management, IUI, and/or IVF or intracytoplasmic sperm injection (ICSI). With number of vaginal deliveries of healthy, term singletons at 24 months following randomisation used as the outcome, it was found that the intervention group had a significantly smaller proportion of births than the control groups (27.1% vs. 35.2%), with the rate ratio being 0.77 (95% CI 0.60–0.99). However, there was a significantly larger proportion of women with an ongoing pregnancy who conceived naturally in the intervention group (26.1%) than in the control group (16.2%), with the rate ratio being 1.61 (95% CI 1.16–2.24).
6.4 The Preferred Path: Comparative Assessment of First-Line Treatments
Four treatment domains come into play when considering the first-line therapy of choice to use in couples with unexplained subfertility: effectiveness, safety, burden and costs.
6.4.1 Effectiveness
When contrasting the effectiveness of both first-line treatments in unexplained infertility, the outcome is naturally measured by the ability for couples to have a live birth. Delivery rates with IUI (with MOH) is approximately 8% per cycle, compared to the 29% pregnancy rate which has been demonstrated with IVF [39]. While RCTs conducted on females aged between 18 and 38 with unexplained infertility have demonstrated higher per cycle pregnancy rates in IVF relative to IUI, no significant differences are observed in cumulative pregnancy rates over 12 months in treatment-naïve couples undergoing a maximum of six cycles of either IUI (with or without MOH) or IVF, or in one cycle of IVF with elective single embryo transfer compared to three cycles of IUI [40, 41].
6.4.2 Safety
Safety as it relates to treatment primarily has to do with potential for physical harm, which can manifest in the context of unexplained infertility as maternal or neonatal complications. The primary concern in IUI or IVF is the increased risk of multiple pregnancy, which has been estimated to be 7% with IUI and 19% in IVF [39]. Those multiple pregnancies which lead to two or more at or near full-term healthy babies do not technically present an issue. However, the potential for increased adverse maternal and neonatal complications and outcomes has prompted initiatives to reduce multiple pregnancy rates. Single embryo transfers (SETs) represent one such measure used to combat multiple pregnancy rates and have been adopted by a number of countries such as Australia, New Zealand, Scandinavia, Belgium and the Netherlands. Results from two RCTs demonstrated comparable multiple pregnancy rates in treatment-naïve couples receiving IUI and IVF with SET [40, 41]. IVF compares unfavourably to IUI, however, in terms of the former’s risk of OHSS, which is rarely an issue for IUI even with MOS.
6.4.3 Burden
The concept of burden pertains to how receiving a treatment will affect a couple’s overall wellbeing and functional capacity. The fertility treatments utilised in unexplained infertility have an accompanying physical burden, which is compounded given there is often a need for repeated visits to the fertility clinic. IUI and ovarian stimulation imposes a physical burden, as does IVF, which is generally more painful than the former given adverse effects from medication used in the stimulation phase, need for follicle aspiration and ovarian enlargement within the luteal phase. Burden is also present as psychological distress, commonly anxiety and depression, which is typically prominent during or after failed or cancelled treatments [42].
6.4.4 Costs
Costs represent an important consideration in treatment and are defined as the financial expenditure required to successfully achieve the outcome of interest. The cost of IVF is far greater than that of IUI in a comparison per cycle [43].
6.4.5 The Preferred Path?
Bearing in mind each of the four criteria which dictate first-line treatment, IUI with mild ovarian stimulation (MOH) is seemingly better given it carries less burden, has a greater safety profile, is less costly and appears to have approximately equal efficacy compared to IVF. Figure 6.2 summarises the characteristics of IUI and IVF across the four domains as they pertain to first-line treatment of unexplained infertility.
Figure 6.2 Comparison of first-line treatments in unexplained infertility. First-line treatments for unexplained infertility include intrauterine insemination (IUI) with or without controlled ovarian stimulation (COS) and in vitro fertilisation (IVF). When comparing IUI ± COS with IVF, effectiveness as measured by cumulative live birth rates is comparable among treatment-naïve couples across six cycles for each therapy, with the former also being safer, associated with less burden and having lower costs.
A recent trial sought to investigate if different methods of IVF – IVF with SET or IVF in a modified natural cycle – could have comparable live birth rates to IUI–COS while reducing multiple pregnancy rates in couples with unexplained infertility and poor prognosis for natural conception [44]. The findings demonstrated that, across all three groups, the number of couples who delivered healthy children after one year were comparable and that all groups had low multiple pregnancy rates. The main distinction between all these modalities is that both forms of IVF are substantially more expensive than IUI [45]. The INeS trial results have been widely used to support the effectiveness of IUI and justify it as appropriate first-line therapy [46]. Drawing such a conclusion may be unwarranted at this stage, however, given there are implicit suggestions that, first, IUI has proven superiority over no treatment, and second that IVF similarly has an advantage over sexual intercourse in couples with unexplained subfertility.
6.5 IUI and IVF Use in Unexplained Infertility: What’s the Evidence?
6.5.1 IUI
6.5.1.1 IUI versus Sexual Intercourse
A recent review by the Cochrane Collaboration examined live birth rates in couples with unexplained infertility undergoing IUI, with or without controlled ovarian hyperstimulation (COS), compared to unprotected sexual intercourse in the presence or absence of cycle monitoring [19]. Three RCTs were identified by authors, which in total included a population of 690 couples with unexplained infertility of average duration 2–4 years, with females being on average 33 years old [47–49]. Among the three studies, only one compared IUI and timed sexual intercourse using cycle monitoring [47], while the other two studies contrasted IUI with sexual intercourse in the absence of medical co-interventions [48, 49]. Using the outcome of clinical pregnancy, a comparison between IUI without MOH relative to sexual intercourse demonstrated an OR of 1.53 (95% CI: 0.88–2.64), while IUI with COS in comparison to sexual intercourse was 1.00 (95% CI: 0.59–1.67). The OR for multiple pregnancy in IUI without COS compared to sexual intercourse was 0.50 (95% CI: 0.04–5.53), while for IUI with COS relative to sexual intercourse was 2.00 (95% CI: 0.18–22.34).
The conclusions inferred by the authors based on these results was that there is inconclusive evidence to support that there are differences in pregnancy outcomes between IUI, with or without COS, compared to sexual intercourse. Notably, the authors outlined the need to determine if multiple pregnancy rates after IUI could be lessened to an appropriate degree without compromising live birth rates, which could be achieved through contrasting IUI without COS to IUI with COS with low-dose gonadotropins.
6.5.1.2 The Trial on Intrauterine Insemination (TUI Study): Is IUI Better in Good or Bad Prognosis Couples?
A recent trial sought to investigate how outcomes differ between IUI and expectant management in a two-centre, open-label, RCT [20]. A total of 201 couples with unfavourable prognosis for natural conception were randomised for three cycles to either IUI with ovarian stimulation (using oral CC or oral letrozole) or expectant management. Of the 101 women who received IUI, there were 31 live births (31%), compared to 9 (9%) live births in the 100 women receiving expectant management. The risk ratio for the outcome of live birth in IUI compared to expectant management was 3.41 (95% CI: 1.71–6.79), implying that the couples receiving IUI when they have poor prognosis in unexplained infertility have higher cumulative live birth rates and that the treatment is effective in this clinical context. Findings from a similar study, which differed from the above trial primarily in that couples had a favourable prognosis for natural conception (less than 2 years attempting to conceive), demonstrated that use of immediate IUI did not improve upon cumulative live birth rates from 6 months of expectant management [48].
6.5.2 IVF
6.5.2.1 IVF versus Sexual Intercourse
The Cochrane Collaboration similarly published a review comparing the efficacy and safety of IVF in treatment of unexplained infertility to sexual intercourse [17]. Only two relevant studies, both of which were RCTs, were identified by authors for this analysis [22]. In both original studies, only a subset of the population could be used in the Cochrane review, as not all the couples had a diagnosis of unexplained infertility. As such, the first study contributed 35 out of 245 couples while the second study had 51 of 139 couples who could be included in this analysis. The Cochrane review thereby had a total study size of 86 couples, in whom the average duration for trying to conceive was 5 years and the average female age was 33 years. In the first study, a comparison was drawn between one IVF cycle and 6 months of sexual intercourse with or without reproductive treatment excluding IVF. The second study similarly compared one IVF cycle, albeit to 90 days of sexual intercourse alone. Other outcome measures, such as rate of multiple pregnancy, OHSS, miscarriage or costs, were not reported on in these studies.
The two RCTs had results demonstrating opposing treatment effect directions, which suggests that a high degree of heterogeneity was present. For the outcome of clinical pregnancy in IVF compared to sexual intercourse, the first study had an OR of 0.30 (95% CI: 0.02–3.67), while the second study had an OR of 8.00 (95% CI: 1.89–33.85), leading to a pooled OR 3.24 (95% CI: 1.07–9.80). Considering the second study, given its apparent positive treatment effect, a considerable difference in conception rates (29% vs. 1%) seemingly implies IVF is an effective tool in subfertility treatment. Notably, the study population had average duration of subfertility exceeding 4 years, which is far longer than seen in many couples who seek IVF. Additionally, it is likely that there was a proportion of couples within this study who were included despite having conditions such as severe oligospermia, anovulation or severe endometriosis (American Fertility Society [AFS] class III or IV), which would make them ineligible for expectant management given the absent chance for conceiving spontaneously.
The authors’ conclusion based on the pooled results of both studies is that there is inadequate evidence to suggest that IVF significantly provides better clinical pregnancy outcomes than sexual intercourse in unexplained infertility. The authors underscored that future directions should involve subsequent subfertility trials having concordant study designs, methods and result presentation such that data can be better pooled, in addition to having studies that investigate the timing at which the transition from sexual intercourse to invasive first-line treatment options should occur. Table 6.2 summarises the current literature comparing the efficacy of sexual intercourse to IUI and IVF in couples with unexplained infertility.
Table 6.2 Comparison between sexual intercourse and first-line treatments in treatment of unexplained infertility
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
