Prognosis-based management of unexplained infertility—why not?

Abstract Up to a half of couples seeking medical assistance for infertility are diagnosed with unexplained infertility, characterized by normal ovulation, tubal patency, and semen analysis results. This condition presents a challenge in determining the optimal treatment approach. Available treatments include IUI and IVF, but guidelines vary on when to offer each. Prognosis-based management is identified as a research priority, and various prediction models have been developed to guide treatment decisions. Prognostic factors include female age, duration of subfertility, and sperm parameters, among others. Prognosis-based strategies can enhance cost-effectiveness, safety, and patient outcomes, offering less invasive options to those with good prognoses and more aggressive interventions to those with poor prognoses. However, there is a gap between research evidence and its clinical application. In this article, we discuss the application of prognosis-based management in the context of unexplained infertility, highlighting its potential to improve clinical decision-making and patient outcomes.


Introduction
In up to half of all couples who seek medical assistance for infertility, a direct cause for their inability to conceive cannot be identified.They are therefore labelled as having unexplained infertility (Buckett and Sierra, 2019).Unexplained infertility is characterized by normal ovulation and at least one patent Fallopian tube in the female partner, as well as a normal (or slightly deviant) semen analysis in the male partner according to the World Health Organization criteria (Cooper et al., 2010).The causes of unexplained infertility remain, as determined by its definition, largely unknown (Quaas and Dokras, 2008).While couples with unexplained infertility even after 12 months of trying still have inherent chances of conception, they are typically considered to be less fertile.They may therefore, despite the lack of a clear understanding of the underlying physiological factors responsible for their reduced fertility, undergo a variety of diagnostic and therapeutic procedures (Wang et al., 2020;Raperport et al., 2023).On one hand, lack of conception might be put down to chance, whereas on the other hand, underlying unidentified and maybe even unknown causes for infertility prevent conception.
Available medical treatments for unexplained infertility include IUI with ovarian hyperstimulation and IVF.Treatments such as IUI and IVF might on one hand be effective because they increase the number of possible conceptions, for example by creating multiple follicles or embryos, but they also might overcome an unknown barrier to conception.
At present, the consensus on the management of unexplained infertility across continents is varied.For example, The National Institute for Health Care and Excellence (NICE) guideline (National Institute for Health and Care Excellence, 2013) for unexplained infertility recommends IVF after 2 years of expectant management.The American Society of Reproductive Medicine (ASRM) and the Canadian Fertility and Andrology Society recommend an initial approach of IUI with superovulation for up to three to four cycles followed by IVF, but the true challenge lies in determining the optimal timing for each treatment (Buckett and Sierra, 2019;Penzias et al., 2021).A recent ESHRE guideline also discusses the management of unexplained infertility and recommends IUI with ovarian stimulation as a first-line treatment for couples with unexplained infertility (Teede et al., 2018;Romualdi et al., 2023).
In couples diagnosed with unexplained infertility, natural conception is inherently a possibility.Therefore, assessing the prognostic profile of these couples becomes crucial in understanding the likelihood of natural conception.This opportunity for prognosis-based management of unexplained infertility has been identified as one of the top 10 priorities for infertility research (Duffy et al., 2020); however, it was not used in the above-mentioned guidelines to phrase the key research questions.The ESHRE guideline does indeed recognize the potential of a prognosis-centred strategy for couples facing unexplained infertility.However, it seems to adopt a more generalized stance, placing the onus on readers and clinicians to select their preferred methodologies (Romualdi et al., 2023).

Opinion
Here, we discuss the indication-setting for the treatment of couples with unexplained infertility.A first and fundamental step prior to the management of these couples is the exclusion of diagnoses linked to an underlying cause of infertility.We will then demonstrate how prognosis can be used to determine which couples require treatment and at what stage of their infertility.

Unexplained infertility is a diagnosis per exclusion
The main diagnoses linked to a cause of infertility are anovulation, severe male factor, or tubal obstruction.There should be an egg, there should be sperm and the two should come together.If any of these three factors is missing, conception will not occur and infertility is a fact.
The diagnosis of anovulation, or its opposite, the confirmation of ovulation can be made through a variety of methods such as urinary LH tests, serum progesterone levels, basal body temperature charting, and ultrasound (Romualdi et al., 2023).The firstline and second-line treatments of an-or oligo-ovulation are medical ovulation induction (Teede et al., 2018).When this is not successful, IVF and/or laparoscopy with ovarian drilling can be applied.
Male infertility is diagnosed or excluded by conducting a simple semen analysis (The Sixth Edition of WHO Manual for Semen Analysis; Boitrelle et al., 2021).Severe male factor infertility involves severe oligozoospermia (<5 � 10 6 sperm per ml of ejaculate), asthenozoospermia, or even azoospermia (Mazzilli et al., 2022;Sharlip et al., 2002).Its treatment requires IVF and ICSI.In men with azoospermia, testicular sperm extraction or treatment with donor sperm can be considered.Finally, tubal blockage, or its opposite, namely patency of both or at least one Fallopian tube, can be diagnosed with hysterosalpingography and laparoscopy, and more recently hystero-salpingo contrast sonography (Penzias et al., 2021).Treatment could be tubal surgery, but is in most settings dominated by IVF.
Couples diagnosed with anovulation, severe oligozoospermia or azoospermia, and double-sided tubal blockage have one similarity: without treatment, their chances to conceive are-close to nil-treatment is therefore indicated in these couples (Thurston et al., 2019).
A special category is female ageing.Since ovarian reserve decreases with age, fertility declines too (Amanvermez and Tosun, 2016).In other words, even without any abnormalities, female fertility declines, with a turning point somewhere between 35 and 40 years.Above that turning point, infertility can be related to reduced ovarian competency and the potential benefits of ART are unlikely if infertility is attributed to the aging of the woman.
While there are many more diagnoses possible, the above category covers the majority of the cases where a diagnosis can be made.The remaining couples are diagnosed with unexplained infertility, and these couples do, by definition, not have a fundamental impediment to conception (Quaas and Dokras, 2008).As a consequence, natural conception is expected to occur in a substantial number of these women.As it is not useful to treat couples who have good prospects of natural conception in due time, before deciding on the start of treatment, a thorough understanding of the prognosis for natural conception is therefore needed.If the prognosis for a 'natural pregnancy' in due course is high enough, it might be justified to delay treatment in these couples for some time.

Prognosis and unexplained infertility
The adoption of a prognosis-based approach has been steadily increasing in the field of medicine, proving to offer significant advantages for healthcare systems and patients alike.This approach is evident across various medical domains, such as oncology, where tools like the Nottingham Prognostic Index and Adjuvant Online have been developed to predict the prognosis of young breast cancer patients (<40 years old) (Hearne et al., 2015).Similarly, there is a growing demand for the implementation of a prognosis-based strategy in the management of couples facing unexplained infertility.This entails an initial assessment of the overall prognosis for untreated couples and an examination of the factors influencing this prognosis.Subsequently, these influential factors are integrated into a predictive model to distinguish between couples with a reasonable likelihood of natural conception and those who may require intervention to facilitate conception.Such a methodological framework aligns with the PROGRESS Framework, and we emphasize the importance of adhering to these recommendations when managing couples dealing with unexplained subfertility (Hemingway et al., 2013).
From the models above, the Hunault model integrated data from three previous cohorts into a unified prediction model, and found female age, duration of subfertility, primary or secondary subfertility, percentage of motile sperm, and referral status (being referred by a general practitioner or a fellow gynaecologist) to be independently predictive for natural conception (Hunault et al., 2004).
Before prognostic models can be used in clinical practice, they need to be validated in cohort studies or, even better, their impact should be assessed in randomized controlled trials (RCTs) (Leushuis et al., 2009).The Hunault model demonstrated good calibration during externally validated studies in the Netherlands, New Zealand, and Australia (van der Steeg et al., 2007;Farquhar et al., 2011;Song et al., 2021).Bensdorp et al. (2015) extended the model with additional variables such as cycle length, BMI of the woman, a basal FSH level, and a semen analysis.
More recently, updated prediction models (van Eekelen et al., 2017;McLernon et al., 2019) have been designed that provide predictions over time that allow adjustments at a repeated interval of time that could assist in offering chances of conception over multiple time intervals and has a distinct advantage of being able to provide prediction once the couple has returned after initial expectant management.At their return after the initial expectant management, the estimates of conception can be recalculated using one of the dynamic prediction models (van Eekelen et al., 2017;McLernon et al., 2019).While these 'dynamic' prediction models foresee a need for personalized predictions, they are not needed in a guideline, where prediction models If not specified, the unit is the number of couples.
Prognosis-based infertility management | 3 should guide not more than the first decision to start or delay treatment.
The prediction models can be used to determine if IUI and IVF are expected to add to natural fertility chances.It should be realized that in different situations, many predictive factors will work in the same direction and the same magnitude.For example, increased female age will reduce the fertility chances, independent of whether these are natural pregnancy chances or success after IUI or IVF.Predictors that strongly decrease natural fertility chances, but do not or to a lesser extent reduce fertility chances after treatment with IUI and IVF, are more suitable treatment selection markers.Examples of these are the duration of infertility, but maybe also the prognostic index derived from the Hunault model or other prediction models.
Prediction models can be used to triage couples with unexplained infertility as having a good, intermediate, or poor prognosis for natural conception.The power of using the prediction models lies in their capacity to guide treatments (Fig. 1).Couples with good prospects for natural conception-say >30% conception chances in 12 months-can initially be managed expectantly, while in couples with poor prognosis for natural conception-<30% conception chances in 12 months-but good prognosis after treatment, IUI, and later IVF/ICSI are justified (Steures et al., 2006).In couples with poor prognosis for natural conception where treatment is not expected to increase these chances, for example owing to a higher female age, IVF with oocyte donation has the best chance to overcome the low natural conception chances.
Importantly, the use of prediction models has not only been validated in cohort studies but also their impact has been assessed in RCTs.Steures et al. (2006) randomized 253 couples with an intermediate prognosis for natural conception (30-40% in 12 months) and found that IUI with hyperstimulation did not increase ongoing pregnancy rates after 6 months.In contrast, two other RCTs, the TUI (The Uterine Insemination) and the EXIUI (Expectant management vs IUI in unexplained subfertility and a poor pregnancy prognosis) trials, were carried out in couples with a poor prognosis for natural conception (<30% in 12 months) (Farquhar et al., 2018;Wessel et al., 2022).These trials, both with a large sample size, clearly demonstrated that IUI with ovarian stimulation is more effective than expectant management in these women.Bensdorp et al. (2015) showed in a large RCT that in these couples IVF does not bring a benefit over IUI with ovarian stimulation.
Thus, the prognostic index for natural conception, as established with the Hunault model, can be used as a treatment selection marker in couples with unexplained infertility (Hunault et al., 2004).In couples with a good prognosis, IUI with ovarian stimulation does not benefit over expectant management, while in poor prognosis couples, IUI with hyperstimulation should be recommended, with cancellation policies in the presence of multiple follicular development to reduce the risk of multiple pregnancy (Wang et al., 2019b;Wessel et al., 2022).
Several other ingredients can be added to the management plan.Lifestyle interventions should be considered where appropriate.Also, tubal patency testing with oil-based contrast is known to increase live birth rates in couples with unexplained infertility, in couples with a good prognosis (Wang et al., 2019a;Zhang et al., 2022).Prediction models and their adoption on a wider scale

CalculaƟng probability of natural concepƟon in the next 12 months
Prediction models have been in existence for over three decades, yet their widespread integration into routine practice remains limited.Factors contributing to this underutilization include a lack of awareness about these models, concerns regarding their trustworthiness and appropriateness for practical use, and insufficient validation across diverse populations.In the context of fertility evaluation, couples often opt for immediate interventions over natural conception because of the perceived need for 'treatment' and impatience regarding the time required for natural conception.
The successful adoption of prediction models hinges on robust validation efforts in diverse populations, ensuring the models' generalizability and reliability.A critical aspect of achieving broader implementation involves raising awareness among clinicians about the clinical utility of prognostic models.Demonstrating how the incorporation of these models can lead to enhanced patient outcomes, improved decision-making, and resource optimization is essential.Moreover, the design of prognostic models should prioritize accessibility and userfriendliness for healthcare professionals, with careful consideration given to their seamless integration into existing healthcare systems (Collins et al., 2024).
While certain models exist, the Hunault model demonstrated good calibration during externally validated studies, making it a more generalizable and reliable model (Custers et al., 2007;van der Steeg et al., 2007).In spite of its validation on a wider scale, local application and validation are recommended before using in routine practice to improve the accuracy of the model performance.Alternative models, such as van Eekelen's and McLernon's (van Eekelen et al., 2017;McLernon et al., 2019), although existing, require further validation, especially in diverse populations.Practitioners in different geographical locations can effectively leverage prediction models by comprehensively understanding their purpose, clinical relevance, and underlying predictors.Regularly auditing model performance in conjunction with clinical judgment is essential for continued efficacy.Addressing the inclination towards immediate interventions requires patient education to empower informed decisionmaking between expectant routes and more intensive procedures such as IUI or IVF.A prospective survey exploring barriers and future adoption trends of these models would provide valuable insights for refining clinical practice.
In the endeavour to increase awareness and promote the adoption of prediction models, international societies can play a pivotal role.The active engagement of these models on international platforms, such as ESHRE or ASRM, can serve as a catalyst in illuminating their existence.By featuring these models prominently within the discourse of international societies, a broader audience of clinicians can be reached, transforming them into advocates for the incorporation of prognosis-based approaches to unexplained infertility.This concerted effort would contribute to the realization of these models as a practical and beneficial tool in the clinical management of infertility.

Advantages of prognosis-based management
Contemporary management of couples with unexplained infertility exhibits variation across different continents.In countries where reimbursement is available for assisted conception treatment, a pragmatic approach is adopted, wherein less invasive therapies are initially considered followed by more invasive interventions if the initial treatment proves unsuccessful.Conversely, in many countries where couples pay for infertility treatments entirely, or are largely out-of-pocket, IVF is offered as the first-line treatment despite its invasive nature, lower costeffectiveness, and associated safety concerns.
The adoption of a prognosis-based approach confers several advantages, such as increased cost-effectiveness, improved safety, and reduced invasiveness or medicalization.Pham et al. (2018) found that if 90% of couples with a good prognosis delayed IVF treatment by 6 months, there would be a substantial decrease in the cost without compromising pregnancy and live birth rates over an 18-month period.Other studies have confirmed the cost-effectiveness of this approach (van Eekelen et al., 2020;Tjon-Kon-Fat et al., 2015).On top of that, adopting a prediction-based approach offers advantages in terms of providing couples with less invasive and safer options at an earlier stage of their infertility, while those with a poor prognosis can proceed directly to treatments such as IUI with ovarian stimulation and IVF.Even IVF with single embryo transfer carries an increased risk of twins or of multiple pregnancies (Chaabane et al., 2015).On top of that, both IUI with ovarian stimulation and IVF bear an increased risk of perinatal complications such as preterm births, pre-eclampsia, and gestational diabetes (Pandey et al., 2012).
Owing to the aforementioned considerations, it is in our opinion crucial to evaluate couples with unexplained infertility to assess the individual prognosis as part of the clinical management.The validity of this approach has been confirmed in multiple RCTs in which the prognostic index is used as an inclusion criterion, the strongest type of validation available.

Table 1 .
Prediction models for pregnancy after expectant management, IUI, and IVF.

Sample size a End point Predictive factors Performance Model stage Natural conception
Note: After exclusion of other causes, couples are assessed by calculating a prognostic index for natural conception.In case of a favourable prognosis (natural conception chances with 12 months >30%) expectant management for 6 months is the preferred strategy.In other couples, intrauterine insemination with ovarian stimulation for three to six cycles should be the preferred strategy.
Figure 1.Management strategies for couples with unexplained infertility.