Lethal variants of equine pregnancy: is it the placenta or foetus leading the conceptus in the wrong direction?

Keywords: abortion, aneuploidy, chromosome, CNV, early pregnancy loss, embryology, equine, fetus, genetics, horse, mare, miscarriage, monosomy, placenta, pregnancy complications, pregnancy loss, SNPs, translocation, trisomy, trophoblast.

It is believed that the ancestors of the modern horse were domesticated across the Western Eurasian steppe (modern day Northern Kazakhstan) around 4000 years ago (Outram et al. 2009). Unlike other prey animals that were domesticated for food purposes (cattle, sheep, pigs etc.), horses were more likely domesticated primarily for transportation, farming, and warfare purposes. Through the mating of related individuals with prized phenotypic traits, the process of domestication of horses has led to the modern horse possessing speed, agility and strength frequently on display on the arena, cross country course or racetrack. The majority of horse breeders therefore select mare and stallion combinations based on prior performance and musculoskeletal and cardiovascular traits. With reproductive characteristics playing second fiddle, it is unsurprising that the modern horse genome contains an excess of deleterious mutations compared with their wild ancestors [reviewed by (Raudsepp et al. 2019)].

Reproductive performance studies over the decades point to pregnancy loss as one of the greatest challenges to breeding operations. Due to the seasonal oestrous cycle and the programmed lifespan of endometrial cups in the mare, even pregnancies lost in the first 2 months of gestation can significantly disrupt breeding efficiency. A US study of Thoroughbred (TB) breeding showed the importance of mares producing a foal in six out of seven breeding seasons (Bosh et al. 2009a). Even relatively wealthy sports such as Thoroughbred racing, breeding is not devoid of economic challenges. Indeed, a more recent economic impact study in the UK reported that whilst the United Kingdom (UK) TB breeding industry alone contributes £427 million to the economy and over 19 000 jobs, the profitability of the sport to breeders has declined, with over 66% of operations now operating at a loss (PriceWaterhouseCoopers LLP 2018). For more endangered breeds, reproductive efficiency is critical to their survival (Orlando and Librado 2019).

Pregnancy can fail at any stage from conception to parturition. The underlying reasons for these losses are variable but ultimately most cases will be attributed to a primary pathology of either the mare (such as endometrial health, endocrine function, immunopathology and oocyte characteristics) or the embryo/foeto-placental unit (inherited via the germline from the stallion or mare or acquired during development). In both cases, defects could be intrinsic to the tissue or alternatively the response of that tissue to extrinsic factors such as pathogens, nutrients and environmental changes and/or contaminants. Whilst previous research has focused on both intrinsic and extrinsic factors that impact the environment in which the embryo develops, surprisingly little is known about defects intrinsic to the embryo itself. Both large and small genetic variants intrinsic in the conceptual tissues are commonly associated with miscarriage in women but to recently, little was known about their role in foetal viability in the horse. The aim of this review is therefore to focus on identifying all known genetic causes of pregnancy loss in the mare. We start by introducing the foundations of equine placental and foetal development essential for the delivery of a healthy offspring. We then move on to review the known causes and risk factors for pregnancy loss at the pregnancy, mare, stallion and environmental level with a focus on those that may relate to genetic conditions. In part three, we outline specific structural variants and deleterious mutations associated with embryonic or foetal lethality and discuss their origins and possible explanations for phenotypic heterogeneity.

Equine pregnancy lasts approximately 340 days (or 11 months) during which time, the foetus undergoes coordinated growth and development, allowing delivery of a new-born foal capable of running alongside its mother within just a few hours of birth (Allen and Wilsher 2009). There are two main phases of pregnancy: the early gestation period is approximately 2 months and encompasses the transformation of the simple single cell zygote into a foeto-placental unit comprised of fully formed foetus with a full cohort of organs and a chorioallantoic placenta that has firm contact with the endometrium. The rest of pregnancy thereafter is taken up with a period of continued, rapid and co-ordinated expansion and growth of the placenta and foetus.

Approximately 5.5–6 days following ovulation, the compact morula (or early blastocyst) is selectively transported through the uterine-tubal junction to the uterus, where it migrates between the two uterine horns and body for approximately 10 further days. This phase of mobility is unique in the mare and believed to contribute an anti-luteolytic signal to the endometrial surface, although its precise role remains undetermined (Swegen 2021). Conceptus mobility is facilitated by propulsive uterine contractions, probably resulting from the local secretion of prostaglandin, at least partly by the embryo itself (Stout and Allen 2002), and the spherical shape of the embryo and its glycoprotein capsule. The glycoprotein capsule forms underneath the pre-existing zona pellucida, growing to approximately one micron thick between days 6 and 8. The zona pellucida is shed by day 8, revealing the capsule (Flood et al. 1982), believed to protect the early conceptus from the powerful contractions of the maternal uterine environment (Allen and Wilsher 2009). Gradual disintegration of the capsule begins around day 21 (Oriol et al. 1993). The capsule is essential for equine embryo development as removal from day 6 embryos results in embryonic death (Stout et al. 2005). Puncturing of the capsule can also prove fatal, depending on the size of the hole (Wilsher et al. 2020): in vitro puncturing of day 7 or 8 embryos with an acupuncture needle of approximately 100 μm diameter led to no pregnancies from the 10 embryos transferred to recipient mares. Puncture of day-7 or day-8 embryos with far smaller needles – micromanipulator biopsy pipette (∼30 μm diameter) or microneedle (<1 μm tip) – resulted in pregnancy rates of 75% and 67% respectively, indicating a small allowance for repair mechanisms by the capsule. One potential protective function of the capsule is in aiding the expansion of the embryo (from morula stage to blastocyst) caused by an influx of fluid. This blastocoel fluid has recently been shown to contain several important proteins, including equine pregnancy-associated glycoprotein (PAG) and FK-506-binding-protein-4 (FKBP4) (Swegen et al. 2017).

The equine placenta is a remarkable organ, growing rapidly and constantly differentiating (Chavatte-Palmer et al. 2022). The entire conceptus is only 26 mm at day 20, yet by term, the placenta fills the entirety of the uterine body and the two horns and weighs around 3–5 kg. Equine placentae are epitheliochorial with six layers separating maternal and foetal blood streams (Allen and Wilsher 2009). Formation of extraembryonic membranes is critical to the survival of the conceptus. Prior to the development of the mature epitheliochorial placenta, the equine conceptus relies on a yolk sac placenta. This transitory, primitive structure is essential for supporting the early embryos reliance on histotrophic nutrients dynamically regulated and secreted from the endometrial glands (Bastos et al. 2019). The expanded blastocyst is composed of inner cell mass (ICM), which will develop into the embryo, and the trophectoderm, which forms the extra-embryonic cell linages, and gives rise to the feto-placental unit.

By day 10, the ICM has differentiated into two cell layers, forming the bilaminar embryonic disc. The inner layer, the endoderm, grows and expands to line the trophoblast layer to eventually enclose the yolk sac. On day 12, with the onset of formation of the primitive streak, the equine embryo measures approximately 12–14 mm in diameter. At day 14, now 17–22 mm, the pregnancy is usually diagnosed via transrectal ultrasound in stud medicine practice by the presence of an embryonic vesicle measuring 17–22 mm (Gaivão et al. 2014). Between days 30 and 40, the yolk sac ceases to expand and incorporates into the umbilical cord between days 40 and 50. Coinciding with the regression of the yolk sac is the rapid expansion of the allantochorial placenta that takes over nutrient, energy, and oxygen supply to the foetus, although histotroph secretion continues throughout gestation.

By day 20, the allantois (a fluid filled sac that collects waste from the embryo) appears from the embryonic hindgut, and begins to expand and surround the spherical conceptus (Allen and Wilsher 2009). During this time the capsule begins to thin and rupture, separating it from the conceptus (Oriol et al. 1993; Allen and Wilsher 2009). At day 24, the allantois meets the chorionic membrane and fuses to form the allantochorion (Ginther 1998). At day 28, the allantochorion occupies approximately one-third of the conceptus. By day 35, the allantochorion will surround almost the entirety of the conceptus, save for a small section at the embryonic pole called the bilaminar omphalopleure, the future site of umbilical cord attachment.

The primitive chorionic girdle, first identifiable microscopically at day 28, is a structure made of tall, columnar trophoblast cells, which becomes visible to the eye by day 34 reviewed extensively elsewhere (Allen and Wilsher 2009; Antczak et al. 2013). During the maturation process, the trophoblast layer of the primitive girdle undergoes rapid cell division resulting in finger-like folds. At day 35, it is mature and compromised of terminally differentiated binucleate, elongated columnar trophoblast cells (Allen and Wilsher 2009). Between days 35 and 37, chorionic girdle trophoblast cells invade into the maternal endometrium via the endometrial glands and begin endometrial cup formation (Allen and Wilsher 2009). Endometrial cups are the source of equine chorionic-gonadotrophin (eCG) which reaches peak serum concentrations between days 70 and 80 then declines as the cups regress until approximately day 120 (Wilsher and Allen 2011). The endometrial cups, and the secretion of eCG, are the primary hurdle preventing a mare from becoming pregnant again in the same season following a pregnancy loss (Crabtree et al. 2012).

The epitheliochorial placenta does not invade into the maternal endometrium to the same degree as a haemochorial placenta, therefore a far greater surface area of contact is required to support the developing foetus. The surface area of contact between the foetal allantochorion and maternal endometrium is estimated to be 38–50 m² in the TB mare, which is affected by numerous maternal factors reviewed by others (Derisoud et al. 2021). As a result, primiparous mares produce lighter and smaller foals and placentae with a decrease foal to placenta surface ratio when compared with multiparous mares (Robles et al. 2018). In the mare, the surface area is maximised by both having contact across the entire exterior of the conceptus (termed diffuse) and also through the development of microcotyledons, which are small highly folded areas of contact. From day 40, chorionic villi begin to extend into the maternal endometrium, forming a microvillous attachment with capillary beds rapidly developing on both the maternal and foetal sides (Allen and Wilsher 2009). Microvilli continue to fold, branch, and elongate to give rise to the mature microcotyledons by day 150, with the villi within continuing to lengthen and branch throughout the remainder of gestation (Macdonald et al. 2000).

The first signs of an embryo proper are visible by days 12–14 post ovulation with the visualisation of the primitive streak, the earliest stage of neural tube formation, with rudimentary optical vesicles visible on day 15 (Franciolli et al. 2011) (Fig. 1). The embryonic circulation is established around day 18 (Gaivão et al. 2014), with the cardiac prominence first visible at day 19 and a primitive heartbeat detectable by ultrasonography around or shortly after that time (Franciolli et al. 2011). By day 25, the embryonic skin begins to lose its transparency and limb buds are visible (Barreto et al. 2016), the genital tubercle (primitive reproductive tract) can be distinguished, and the tail begins developing (Franciolli et al. 2011). Hoof formation is evident by day 27 (Franciolli et al. 2011), while the primitive skull (cartilaginous chondocranium) is first visualised at day 30, then calcifies into bone (osteocranium) by day 35 (Barreto et al. 2016). Over the period of 30–65 days post ovulation, the muscles of the fore and hind limbs develop through the differentiation of mesenchyme into muscle fibres (Barreto et al. 2016). During this time, the tongue, oesophagus, and intestine develop (Rodrigues et al. 2014). Development of the lungs, trachea, and thoracic oesophagus is first visible by day 34 (Franciolli et al. 2011) and by day 36 the trachea and lungs have developed and the diaphragm is identifiable. Pigmentation of the retina is first observed at day 30 post ovulation (Franciolli et al. 2011) at the same time the heart has completely formed. The embryo proper transitions to being termed a foetus at day 36, reflecting the development of these key organs (Terminology Working Party of the Third International Equine Reproduction Symposium held in Sydney Australia 1982; Allen and Wilsher 2009). Therefore throughout this review, we will refer to an embryo proper up to 36 days and thereafter the foetus, including if encompassing both periods.

Fig. 1.  Early foetal development of the horse. In utero development of the embryo/foetus during the first 144 days post ovulation, emphasising the rapid growth and development during the first 4 months. Summary of key developmental events in the transition from embryo to fetus (a).  Demonstrated in images of fetuses obtained from pregnancies clinically normal at the time of termination or death of the mare showing limb buds (day 27, b), retinal pigment (day 30, 31, c–d), curving and early limb segmentation (day 34, 40, e–f), fetus identified as horse with hoof formation clearly evident (day 64–144, g–i). Images Equine Pregnancy Laboratory.

While foetal sex identification is possible by day 47 when the external reproductive organs are first visible (Franciolli et al. 2011), clinically it is usually performed shortly after, between days 55 and 75, or even later gestational stages (Resende et al. 2014). Differentiation of the skull begins around day 55, with ossification completed by day 65 (Barreto et al. 2016). The foetus undergoes rapid growth during mid–late gestation, along with further cardiac, pulmonary and gonad development (Fowden et al. 2020). Foetal size and growth rate may be abnormal in failed or failing pregnancies, with intrauterine growth restriction described in foetuses obtained from clinical cases of pregnancy loss in mares as early as 28–46 days (Kahler et al. 2021). Morphologically these foetuses had specific (i.e. failure of neural tube closure) and non-specific abnormalities of the neural tissues and presence of distinct bilateral subcutaneous haemorrhage (Fig. 2). Although this does not reveal an aetiology for the lethality, genetic causes of pregnancy loss may be suspected, similar to associations between chromosomal and foetal morphological anomalies in humans (Byrne et al. 1985, 1986; Ursell et al. 1985; Kasarskis et al. 1998).

Fig. 2.  Characteristics of clinical early pregnancy loss (EPL) abortus material submitted to the Equine Pregnancy Laboratory between 2013 and 2022 (a) and mapped based on maternal age at conception and gestational age (days) when pathology was first noted by the veterinary surgeon (b).

Pregnancy loss can occur anytime from fertilisation to parturition, although very little is reported on losses that occur in the 2 weeks immediately prior to clinical detection in the mare. Whilst miscarriage is the term most commonly used to describe embryonic or foetal death in women, phrases employed in the mare are less standardised. Early pregnancy loss (EPL) is commonly used to encompass losses during the combined embryonic and early foetal stages up to approximately 2 months. Overlapping with this period, early embryonic death (EED) is used to refer to losses during the embryonic stage (prior to 36 days gestation), and early foetal death/loss (EFD/EFL) refers to losses between 40 and 150 days (Morehead et al. 2002). After this point, universally any losses up to 300 days are considered to be abortions and thereafter stillbirths (Roach et al. 2021). In practice, research studies often find it difficult to distinguish between stillbirth and early neonatal death through records and will therefore use perinatal death as a phenotype to reflect an inability to accurately pinpoint the time of death. This variation in terms and timing can make comparison of studies challenging and would benefit from standardisation.

The incidence of pregnancy loss has been regularly reported globally over decades (Table 1). While the exact figures vary by population and methodology, collectively it is clear that the early gestation period is the most likely stage when pregnancies fail, mirroring that seen in other species. Further, these figures likely underestimate the true scale of early pregnancy loss as they do not account for losses prior to clinical detection. One study that carefully monitored pregnancies prior to 145 days found that approximately 19% of mare pregnancies were lost between 11 and 15 days gestation and then a further 6% were lost between 15 and 40 days (Ginther 1985). Recently, power doppler has been trialled to allow pregnancy detection in mares 7–8 days post ovulation (Nieto-Olmedo et al. 2020) providing hope that we can study pre-clinical embryonic losses more closely in the future. Over the last decade the Equine Pregnancy Laboratory has received conceptus material from clinical cases of early pregnancy loss ranging from 12 to 67 days of gestation (Figs 2, 3), reflecting the gestational age range of EPLs that can practically be isolated to study. There is clear bias in numbers towards EPLs that fail after 28 days compared to what might be expected from clinical monitoring of mares (Fig. 2), showing just how hard it is to perform research on EPLs that fail in the first 4 weeks of gestation.

Table 1.  Incidences of global equine pregnancy loss reported (1980–2020) from natural mating.

Fig. 3.  Representative samples of equine EPLs, submitted to the Equine Pregnancy Laboratory. Top: Conceptuses during their initial assessment and processing on the day of arrival in the laboratory. Please note the appearance of the early foetal membranes, e.g. ‘translucent’ (a) or ‘thickened and cloudy’ (b, d); and the absence (b) or presence [* in (c)] of blood vessels. Diameter of the allantochorion (ALC) in mm. Bottom: Microscopic images of foetuses from conceptus (b), (c) and (d), respectively. An intact amnion (AMN) with amniotic fluid was appreciable at day 62 (e). Crown-rump-length (CRL) in mm. Gross morphological abnormalities include an irregular pattern the CNS surface [+ in (b) and (d)] and a distinct bilateral dark-red discolouration. This ‘red patching’ is mostly located within the area of neck, shoulder and cranial thorax of a foetus and presents as localised severe subcutaneous haemorrhage in histology [* in (b) and (d)]. Please note a severely altered development of the foetal head and face and delayed limb development of the foetus in (c), appreciable despite autolytic changes of this sample. The visible right front limb (o) is not subdivided and ‘bent towards the midline’ as expected for this gestational age and development of foot pads has not occurred.

Whilst risk factors alone cannot inform us of the underlying cause of pregnancy loss, epidemiological studies still play an important role when collectively building evidence of cause of foetal death particularly for those associated cases with congenital foetal and genetic abnormalities (Reddy et al. 2009). In order to investigate risk factors for pregnancy loss, studies have used varying phenotypes including clinical confirmation of pregnancy (conception and pregnancy losses in the first 2 weeks of gestation) (Allen and Wilsher 2012; Hanlon et al. 2012; Lane et al. 2016), EPL (including losses from 15 days to 42, or 65 days) (Nath et al. 2010; de Mestre et al. 2019) and foetal losses from day 35 to foaling (early pregnancy, abortion and stillbirths) (Miyakoshi et al. 2012). They also include a mixture of univariable and multivariable analysis (Allen et al. 2007; Bosh et al. 2009b). Very few studies have explored risk factors specific to pregnancy loss after the first 2 months, and none, to the authors’ knowledge, have looked at equine abortion or stillbirth as overt phenotypes using multivariable models. The pathophysiology of pregnancy loss varies significantly depending on the stage of gestation. Whilst there may be general risk factors for pregnancy loss throughout gestation, given the differing underlying causes, it is important to consider early and later term losses separately if we are to use the evidence to better understand causes of pregnancy loss. Most work has focused on identifying risk factors for losses in the first 2 months of gestation, the period with the greatest contribution to reproductive inefficiency and the poorest rate of diagnosed cause. It is also the period when genetic variants are most likely to contribute to lethality, so will be the focus on discussion below. Variables found to modify the risk of EPL fall under pregnancy, mare, stallion, or environmental/external factors.

A direct comparison of breeding methods has found natural mating, artificial insemination, as well as artificial insemination, embryo flush followed by transfer all have a similarly low incidence of early pregnancy loss (4.5–6%) when compared to pregnancies established with cryopreserved in vitro generated embryos following ovum pick up, intracytoplasmic sperm injection (ISCI) and embryo transfer (13%) (Cuervo-Arango et al. 2019). Further, 26% of in vitro produced (IVP) pregnancy losses appear to be associated with a higher proportion of anembryonic vesicles (26% of losses) (Cuervo-Arango et al. 2019), compared with just 7/55 (13%) in non-IVP EPLs (Shilton et al. 2020). This is even more pronounced if embryo diameter at the time of transfer is considered: with 45% of embryos less than 300 μm at the time of transfer developing into an anembryonic vesicle (Claes et al. 2019; Cuervo-Arango et al. 2019). The exact underlying reason for anembryonic vesicles is not fully understood, with just one associated with autosomal aneuploidy (Shilton et al. 2020). The size of the embryo is also associated with modified risk of EPL for both IVP (Cuervo-Arango et al. 2019) and in vivo produced embryos (de Mestre et al. 2019). The odds of EPL is reduced by 0.26 for each 10 mm increase in embryo size at first scan in TB mares (de Mestre et al. 2019).

The presence of two or more conceptuses is associated with a high risk of early and late pregnancy loss (Jeffcott and Whitwell 1973). In the mare, these multiples are nearly always dizygotic, with monozygotic multiple pregnancies reported in the literature in just a handful of in vivo generated embryos (Meadows et al. 1995; Govaere et al. 2009) and 1.6% of IVP embryos (Dijkstra et al. 2020). Although conception of multiple pregnancies may be less relevant to EPL in modern veterinary practice: multiple conceptions are routinely reduced at around day 16 of gestation so only one conceptus is carried to term (Ginther and Griffin 1994). Such a reduction is shown across multiple studies to not increase the risk of EPL for the remaining conceptus (Nath et al. 2010). Its impact on abortion and stillbirth is not reported, but current investigations in the Equine Pregnancy Laboratory suggest abortion has distinct risk factors to EPL. Whilst twin reduction has not been associated with an increased risk of EPL (de Mestre et al. 2019), flunixin meglumine at time of twin reduction was found to be protective (de Mestre et al. 2019).

As in many other mammalian species studied to date (Quenby et al. 2021), advancing maternal age is a well-known risk factor for early pregnancy loss in horses (Rose et al. 2018; de Mestre et al. 2019; Fawcett et al. 2021). The pathologies of older mares are extensively reviewed by others (Derisoud et al. 2021) and in brief include endometrial pathologies such angiosclerotic changes (Grüninger et al. 1998) and lower quality oocytes and subsequently embryos (Rizzo et al. 2019, 2020). In an attempt to differentiate between oocyte related issues and uterine related issues of older mares, morphologically normal embryos from cycling donor mares were transferred to either normal or sub-fertile mares (Ball et al. 1987). There was no difference in the loss rate between the embryos transferred to the normal (9%) or sub-fertile (11%) mares by day 28 of gestation. The same group undertook another experiment in which embryos from normal mares (aged 5.7 and primi/multiparous) and sub-fertile mares (aged 19.4 with a history of reproductive failure) were transferred into normal fertile recipient mares and monitored until day 14 (Ball et al. 1989). A significant difference in the rate of embryonic loss between the normal donors (9%) and the sub-fertile donors (62%) was found, although only in the very early stages of gestation. These two publications highlight the importance of oocyte quality for early pregnancy outcome and are highly relevant when considering genetic causes of pregnancy loss.

The quality of the endometrial environment has also been associated with increased risk of EPL. Uterine cysts are fluid-filled structures and can be separated into two categories: glandular (located in the connective tissue layer, up to 1 cm diameter) or lymphatic (bulging into the uterine lumen, up to 20 cm diameter) (Stanton et al. 2004). Two recent studies using multivariable analysis showed an increased risk of EPL with the presence of cysts (Miyakoshi et al. 2012; de Mestre et al. 2019) even after maternal age and status had been accounted for in the models. It remains unclear whether the cysts themselves cause EPL and/or rather indicate underlying endometrial pathologies that could cumulate in EPL as reviewed here (Katila and Ferreira-Dias 2022). Mares can be bred on their foal heat, however, there are concerns that this practice may lead to an increased incidence of EPL. Multivariable analysis has shown both no modified risk (de Mestre et al. 2019) or an increased risk of EPL with foal heat cover (Miyakoshi et al. 2012; Lane et al. 2016).

Body condition score (BCS) of the mare is not a factor that has been explored in multivariable modelling of pregnancy losses however univariable studies indicate that it is possibly the change in BCS which is important as opposed to the score itself (Miyakoshi et al. 2012). Increasing BCS was found to be associated with a significantly lower PL rates between day 17 and 35 of gestation compared to a constant or decreasing BCS, highlighting the benefits of an increasing plane of nutrition during the early pregnancy period. Conversely, a small study found high-energy diets may increase the risk of EPL in Shetland Ponies (D’Fonseca et al. 2021). Nutrition is clearly important for foetal morbidity and mortality (Robles et al. 2022) and levels and constituents should be controlled if pregnancy outcome is to be optimised.

Advancing stallion age is associated with decreased fertility due to declining testicular function, affecting both gametogenesis and hormone production (Turner 2019). Stallion level risk factors though have been more difficult to pinpoint when considering pregnancy loss of populations, with most studies suggesting it is mare not stallion that has the greatest contribution to variability (Hanlon et al. 2012; Lane et al. 2016; de Mestre et al. 2019). When stallion factors do contribute, the impact is more commonly felt on day 14 pregnancy rates and this conception/very early pregnancy loss than on pregnancy losses beyond 15 days. Greater than 21 covers per week is associated with decreased pregnancy rates but not loss of clinically confirmed pregnancies (Lane et al. 2016). Age of stallion or individual stallion when included as a random effect have not been found to be significant contributors to EPL risk in multivariable analysis (de Mestre et al. 2019). Therefore, whilst genetic variants inherited through the germline via the stallion may still yet contribute to pregnancy loss in the mare, these risk factor studies suggest their contribution is small perhaps as such stallions are quickly removed from the breeding pool. Indeed, when 22 stallions were directly compared just 1/22 had significantly higher odds of EPL compared with a reference stallion (de Mestre et al. 2019) consistently with work of others (Allen et al. 2007).

Whilst therapeutics targeted at the endometrium have increased significantly in use over the last decade, neither the administration, nor type of uterine treatment have been found to be associated with a modified risk of EPL across multiple studies (Allen et al. 2007; Lane et al. 2016; de Mestre et al. 2019). This suggests that endometrial conditions, including endometritis, are mostly being managed effectively at the population level to enable maintenance of the pregnancy to at least 65 days and possibly beyond. Induction of ovulation has been found to reduce the risk of EPL (de Mestre et al. 2019), possibly acting via the quality of the ovulated oocyte or early luteal function (Köhne et al. 2014). Whilst not epidemiological studies are lacking, it is still clear that the presence or absence of pathogens is key to pregnancy loss (Macleay et al. 2022) as demonstrated by the marked geographic variation in reporting of certain pathogens.

Pregnancy loss is a condition that can arise from a vast array of possible causes which vary by gestational age at the time of loss, geographic region, breed and management intensity. Broadly, losses fall into two categories, infectious and non-infectious. A recent scoping review of mid and late gestational losses indicates the intensive interest of the research community in infectious causes of pregnancy loss with studies of placentitis, EHV-1, Chlamydia, Leptospirosis and Nocardioform placentitis dominating the literature of the last three decades (Macleay et al. 2022). Equally, once we drill down to individual populations, the picture can look quite different. For example, a cohort study of 3516 pregnancies in a subpopulation of Thoroughbred mares in the UK and Ireland found a very different distribution of causes (Roach et al. 2021). None that underwent post-mortem examination were attributed to infectious agents such as Chlamydia, Leptospirosis or Nocardioform placentitis and just 5% of pregnancy losses found to be positive for EHV-1 infection and 5% diagnosed with infectious placentitis, a clear reminder that geography and management practices are key when it comes to understanding pregnancy loss in mares.

As well as dividing causes of pregnancy loss by presence of a pathogen, one could also consider the tissue involved in the pathology. The endometrium is one such tissue that has garnered great attention over the decades due to its key role in nurturing the developing embryo and foetus. Endometritis (Morris et al. 2020) and endometrial degeneration (Miyakoshi et al. 2012; de Mestre et al. 2019; Katila and Ferreira-Dias 2022) have been linked with pregnancy loss in the mare both in early and late gestation. Even in mare populations routinely swabbed as a prerequisite to breed, endometrial bacterial infections can be found around the time of the loss in 5–16% of EPLs (Ricketts et al. 2003; Hamstead et al. 2012; Rose et al. 2018). Similarly, the quality of the endocrinological environment has been a topic of hot debate, in particular when it comes to the importance of progesterone levels for pregnancy maintenance. Progesterone production across equine gestation is complex (Conley and Ball 2019) and its role in pregnancy loss has been controversial for decades (Allen 2001). Luteal insufficiency in early pregnancy is categorically a cause of pregnancy loss but an old but well cited study suggested this was uncommon (Irvine et al. 1990). Work over the last decade has questioned this early evidence (Betteridge et al. 2018) suggesting a precipitous drop in progesterone occurs more frequently preceding early pregnancy loss than previously thought. Further, more nuanced effects of early luteal progesterone levels on a number of pregnancy related parameters, including day 14 pregnancy rates (Hollinshead et al. 2022), endometrial function (Beyer et al. 2019), as well as early embryonic and foetal growth and organ development (Willmann et al. 2011; Okada et al. 2020) suggest much needs to be done to understand the role of progesterone levels in pregnancy loss.

Whilst the endometrial and endocrine environment are clearly important, it is estimated that 50–80% of early pregnancy losses (Hamstead et al. 2012) and approximately half of mid to late gestation losses (Roach et al. 2021; Macleay et al. 2022) have no attributed cause. There are a number of well described genetic variants of human foetuses associated with a phenotype of foetal lethality that account for well over 50% of miscarriages (Chen et al. 2021). Whilst congenital abnormalities of the developing equine foetus have been long identified (Giles et al. 1993; Hong et al. 1993), it has only been relatively recently that specific genetic variants have been linked to lethal congenital abnormalities (Ducro et al. 2015; Monthoux et al. 2015; Shilton et al. 2020). One of the challenges when it comes to congenital defects with or without a genetic cause is proving that the congenital abnormality constitutes a cause of death. Reddy et al. (2009) provide a valuable classification system developed for human stillbirth. They state in order for it to be a cause of death, ‘there are epidemiologic data demonstrating an excess of intrauterine mortality, the process is rarely seen in liveborn neonates and when the process is seen in liveborn neonates, if frequency results in death or there is biologic plausibility that it can result in death’. Direct application of this definition to equine pregnancy loss is not possible and will inevitably be clouded by euthanasia of individuals for financial reasons or perceived implications on athletic potential. Nevertheless, with refinement it is likely to be useful and worth considering here when looking at genetic variants.

One of the challenges of studying causes of pregnancy loss, in particular in the first 2 months, is access to the abortus material. We have created a biobank of this tissue (Fig. 2), working closely with clinicians in practice who have been instrumental to its success (Rose et al. 2016). The biobank is, though, not without its limitations. For example, there are just 12 EPLs that failed prior to 28 days of gestation, suggesting mares suffering EPL within 4 weeks of cover generally expel the tissue prior to or very soon after examination by the veterinary surgeon. EPLs have a mixed presentation, with translucent (Fig. 3a) or thickened and cloudy extraembryonic membranes (Fig. 3b, d), and absence (Fig. 3b) or presence of intact vasculature (Fig. 3c, d) sometimes with exsanguination (Fig. 3d). The foetuses associated with these membranes show subcutaneous haemorrhage (Fig. 3b), mismatch in developmental features such as limb development when compared gestation age of 41 days and CNS abnormalities (Fig. 3c), similar to that reported for earlier foetuses from the bank (Kahler et al. 2021). Surprisingly, despite the autolytic nature of these tissues, trophoblast cells and quality DNA have been successfully isolated (Rose et al. 2016) from the majority of EPL cases. Attempts to karyotype this material has been challenging (Blue 1981; Haynes and Reisner 1982) as the growth rate of the cells is compromised. DNA based assays are therefore likely to be the most rewarding when assessing the genomes of EPLs.

The most common type of genetic variant associated with pregnancy loss in the mare is aneuploidy (Shilton et al. 2020), mirroring that reported in women (Jia et al. 2015; Chen et al. 2017). Aneuploidy is the gain or loss of an entire chromosome compared to the normal diploid number for that species (Torres et al. 2008). There are three categories of aneuploidy type: monosomy (loss of a single chromosome), trisomy (gain of a single chromosome) and nullisomy (loss of two homologous chromosomes). As chromosomes are the carriers of the nuclear genome required for normal cell function, the genomic imbalance caused by aneuploidy are rarely tolerated to term. The best-known exception to this is Down’s Syndrome (trisomy 21) in humans, although this variant may also result in pregnancy loss (Hwang et al. 2021). Autosomal monosomy (2n − 1) has never been reported in a live animal of any species, strong evidence of a foetal lethal phenotype (Bugno-Poniewierska and Raudsepp 2021). Historically, research on aneuploidy was focused on human pregnancy loss and conditions, with surprisingly little reported in any veterinary species beyond the rare report in cattle (Schmutz et al. 1996), pigs (Hornak et al. 2012) and horse blastocysts (Rambags et al. 2005). The last decade has resulted in increasing uptake in use of veterinary species to study underlying mechanisms of aneuploidy (Rizzo et al. 2020). In the horse, both lethal and non-lethal aneuploid types have been reported (Shilton et al. 2020; Bugno-Poniewierska and Raudsepp 2021).

Non-lethal autosomal trisomies in horses are rare (Bugno-Poniewierska and Raudsepp 2021). There are just 16 individuals reported in the literature surviving to term (three of which were mosaic) (Power et al. 1992). Those identified are often euthanised at a young age (Kubien and Tischner 2002; Bugno et al. 2007; Holl et al. 2013). The exact phenotypes displayed for non-lethal autosomal aneuploidy are heterogenous. Trisomy of chromosome 28 (65XY,+28) resulted in a colt that displayed reduced growth and cryptorchidism. Trisomy of chromosome 23 (65,XY,+23) in a male Standardbred (Klunder et al. 1989) was associated with limb deformities, facial asymmetry, along with other developmental defects. Many of the other reported individuals diagnosed with aneuploidy displayed limb deformities including: angular limb deformities (Bowling and Millon 1990), bilateral contracted tendons of the forelimbs (Buoen et al. 1997; Lear et al. 1999), bilateral carpal flexural deformity (Zhang et al. 1992; Buoen et al. 1997), stiff hindlimb gait (Brito et al. 2008), and laxity of rear limb flexor tendons (Lear et al. 1999). Developmental Orthopaedic Disease is the most common congenital defect reported as a cause of death in thoroughbred foals (Mouncey et al. 2022), which begs the question: could some of the cases of neonatal and foal lethality be attributed to underlying trisomies?

Aneuploidy in equine conceptus tissue has been reported in just two studies. An early study by (Rambags et al. 2005) identified copy number abnormalities of chromosomes 2 and 4, in 22 in vivo and seven in vitro produced embryos. As these blastocysts were sacrificed for the study, their viability was not able to be assessed. Nevertheless, this early observation was important in building evidence that aneuploidy was not just a peculiarity of human gametes and embryos. More recently, we reported autosomal aneuploidy in 11/55 (20%) EPLs assessed making it by far the most common pathology of early pregnancy reported to date in well managed breeding mares (Shilton et al. 2020). With the exception of trisomy 30, the trisomies and monosomies reported were unique and therefore likely to represent a lethal phenotype, although more time and data is needed to state this with absolute confidence (Shilton et al. 2020; Bugno-Poniewierska and Raudsepp 2021).

Aneuploidy types identified to be associated with equine early pregnancy loss are not only unique, but they tend to involve the larger chromosomes and include trisomy 1, 3, 15, 20 and combined chromosomes 23/24. In contrast, aneuploidies of smaller autosomes and sex chromosomes have been reported in live equine births (chromosomes 23, 26, 27, 28, 30, 31, and X). Biology always reveals its nuances with time, but for the moment lethal and non-lethal phenotypes separate out quite clearly, with chromosome sizes of 65 Mb (or 2.8% of the Genome) associated with lethality and those 55 Mb or less (or 2.4% of the Genome) associated with liveborns or a mixed phenotype (Shilton et al. 2020; Bugno-Poniewierska and Raudsepp 2021). Though, interestingly, there have been no reports of liveborn equine trisomy 29 (34 Mb), suggesting that it is not only size, but also genetic composition of a chromosome, that matters when it comes to viability (Bugno-Poniewierska and Raudsepp 2021). Monosomies of small (26, 27, 31) chromosomes all lead to foetal demise by 65 days of gestation. It is worth noting that monosomies of some of the large chromosomes are yet to be identified, perhaps suggesting these are lethal prior to clinical detection and may well be revealed once routine testing of equine blastocysts is established.

Aneuploidies of allosomes, the most frequent of which is X monosomy, are more commonly reported in adult horses. Pure X monosomy is usually associated with infertility, and mosaic X monosomy subfertility, extensively reviewed by Bugno-Poniewierska and Raudsepp (2021). A study of 500 randomly selected horses (across 11 breeds) found that of the 10 chromosome anomalies detected, nine involved the sex chromosomes, eight of which were mosaic (Bugno et al. 2007). This is not always the case, with reports from other cytogenetic laboratories failing to identify any examples of mosaic X monosomy (Bugno-Poniewierska and Raudsepp 2021). Whether allosome aneuploides also play a role in pregnancy loss is not known. X trisomy has been identified in a case of EPL (Shilton et al. 2020). This could be a coincidence and the pregnancy failed due to another cause. Alternatively, as X trisomy is rare in liveborn horses presenting with gonadal dysgenesis, this aneuploidy type may have a mixed phenotypic outcome as has been shown in humans (Skuse et al. 2018).

Mosaicism is a feature of the term human placenta (Coorens et al. 2021) and with large scale screening of human blastocysts now common place, we know an aneuploidy trophectoderm cell identified in a human blastocyst does not always equate to foetal aneuploidy (West and Everett 2022). Whilst mosaicism is, as discussed above, a feature of adult horses (Bugno et al. 2007), examples of it are as yet to be identified in the equine foeto-placental unit. Only a limited number of EPL tissues have been assessed across both compartments with all assessed placental-foetal duos having a matched diploid status in foetal and placental or multiple placental samples (Shilton et al. 2020). We have identified additional aneuploid pregnancies in unpublished work in our laboratory but are still yet to find examples of aneuploid mosiacism. Still, this is likely a reflection of sample size and they may well be revealed in due course, although perhaps not at the frequency as seen in human placentae (West and Everett 2022).

Aneuploidy most often arises during gametogenesis, as a sporadic event of a single gamete. Maternal meiotic errors account for approximately 84% of human trisomy pregnancies, while paternal meiotic errors account for 11% and post-zygotic errors account for the remaining 5% of cases (Mikwar et al. 2020). These figures are not available for the mare. The failure to identify mosaicism in the samples screened to date, together with reported high incidence of aneuploidy of equine MII oocytes (56%) from old mares compared with young mares (16%) (Rizzo et al. 2020) strongly supports maternal gametogenesis as the primary contributor to aneuploid pregnancies in the mare. Origins in maternal meiotic errors is further supported by other abnormal features of aged equine oocytes, including enhanced chromosome misalignment (Rizzo et al. 2019) and weaker centromeric cohesion of in vitro matured oocytes (Rizzo et al. 2020). A study investigating the effect of in vitro maturation process on the rate of aneuploidy (Franciosi et al. 2017) found that in vitro maturation IVM oocytes were significantly more affected by aneuploidy than in vivo matured oocytes (45.5% vs 0%, respectively). Whether this explains the higher rates of early pregnancy loss of IVP embryos (Cuervo-Arango et al. 2019) is not known although worthy of consideration.

It has been suggested that the prolonged arrest of equine oocytes from in utero foetal development to ovulation contribute to development of aneuploidy in oocytes. In equines, oocytes from older mares (aged ≥14 years) were found to have both a thicker metaphase plate and a reduction in the correct alignment of chromosomes across it during metaphase II (Rizzo et al. 2019). Displacement of chromosomes on the metaphase plate has also been shown to be markedly increased in oocytes during the second meiotic division in older women compared with younger women (40–45 years and 20–25 years, respectively (Battaglia et al. 1996). Interestingly, the major spindle axis of equine oocytes is significantly increased in those with chromosome misalignment, regardless of mare age (Rizzo et al. 2019).

Aneuploidy can also arise through errors in spermatogenesis (Hassold et al. 2007). In analysis of sperm from healthy human donors, 1–2% of spermatozoa were found to be aneuploid (Martin 2007). In stallion sperm, advancing paternal age increases aneuploidy of the allosomes but not autosomes (Bugno-Poniewierska et al. 2011). This would suggest aneuploidy of stallion gametes in the horse is more likely to contribute to subfertility of the resulting offspring as opposed to lethality. Risk factor studies of equine pregnancy loss support this conclusion, with neither stallion age nor individual stallion found to be risk factors for early pregnancy loss (Hanlon et al. 2012; Lane et al. 2016; de Mestre et al. 2019) or pregnancy loss at any stage of gestation (Lane et al. 2016). Whilst genomic structural errors introduced in early embryonic cell division are well described for human and rodent embryos, they are as yet to be described in the horse.

Translocations are the breaking and fusing of chromosome segments between non-homologous chromosomes and have been a cornerstone of genetic diversity throughout evolution. Translocations can be broadly separated into two categories: balanced (the entire complement of genetic material is retained in all cells) and unbalanced (an unequal split between daughter cells, resulting in loss of genetic material in some cells). Of the 15 cases of equine translocations reported to date (Table 2), only a single male had an unbalanced autosomal translocation (64,XY,t(4;30),+4p) (Ghosh et al. 2020), with the remaining being balanced autosomal or unbalanced allosomal translocations. Thirteen mares all presented with subfertility with six of 13 mares presenting with a specific history of recurrent pregnancy loss. This is because translocations disturb meiosis and gametogenesis, resulting in formation of both genetically balanced (1/3) and unbalanced (2/3) gametes (Raudsepp 2020). The latter, if involved in fertilisation, may result in pregnancy loss, whereas gametes with the balanced form of translocation will pass the defect to liveborn offspring.

Table 2.  Summary of 15 equine translocations reported to date and their associated phenotype.

Single nucleotide polymorphisms (SNPs) are one of the most well-known and commonly studied genetic variant types. A change in one of the bases of the codon that results in a change of amino acid is termed a missense mutation which can result in compromised protein function and an altered phenotype dependent on exactly how the protein is impacted (Alberts et al. 2015). There are surprisingly few SNPs reported in the equine literature that are associated with lethality in utero or immediately following parturition (Table 3). This is likely more a reflection on the limited research in the area and limited availability of abortus tissue, as opposed to an absence of lethal haplotypes in equine foetuses which are more comprehensively described in women.

Table 3.  Equine SNPs identified to be associated with pregnancy loss.

Congenital hydrocephalus, excessive accumulation of fluid in the brain, has been identified in several horse breeds. Affected foals are either stillborn or euthanised immediately following birth (Ojala and Ala-Huikku 1992; Ferris et al. 2011; Pannu and Singh 2014). Next generation sequencing identified a nonsense mutation in β-1,3-N-acetylgalactosaminyltransferase 2 (B3GALNT2) associated with congenital hydrocephalus in Friesian foals (Ducro et al. 2015). While capable of leading to stillbirth and perinatal death, whether this SNP also contributes to EPL or abortion is unexplored. Pathologies of the central nervous system in foetuses from pregnancies lost in the first 6 weeks of gestation are common (Kahler et al. 2021), so a heterogenous phenotype can’t be ruled out.

Aberrations in the TP53 gene have also been associated with negative pregnancy outcomes in both horses (de Leon et al. 2012) and women (Su et al. 2011). A study of 105 mares on a single stud farm in Brazil (de Leon et al. 2012), found a heterozygote genotype in exon 4 of p53 (nonsense mutation of non-polar proline to positively charged arginine) was associated with increased risk of abortions (OR = 14.5). It should be noted this was a SNP identified in mares not abortus material, so likely acts on the embryonic environment via the endometrium or immune function although this remains to be explored. Altered expression of p53 has been demonstrated in equine endometrial progenitor cells in obsese mares who in turn present with suboptimal endometrial function (Smieszek et al. 2022).

Fragile Foal Syndrome (FFS) is a monogenic disorder with Mendelian inheritance of a recessive allele originally described in a Warmblood (Monthoux et al. 2015) but more recently also a Thoroughbred (Grillos et al. 2022). Foals homozygous for a single missense mutation (c.2032G>A) in procollagen-lysine, 2-oxoglutarate 5-dioxygenase 1 (PLOD1) either die in utero and present as late gestation abortion or are euthanised shortly following delivery (Aurich et al. 2019). Post mortem examinations describe dermal lesions and reduction of collagen fibres (Monthoux et al. 2015), keratosis-like thickening of the skin, abnormal flexibility of joints, and spinal cord deformities (Aurich et al. 2019). The largest cross breed analysis of PLOD1 recessive allele frequency genotyped 3365 novel individuals and found the mutation was present in 20 of the 38 breeds (Reiter et al. 2020) suggesting FFS is likely to be diagnosed in further breeds in the future.

A large scale study of 2556 horses (using both internal samples and publically available genotypes) found two SNPs in Killer cell lectin-like receptor subfamily 49 B (LY49B), one completely absent in the homozygous state and one with reduced homozygous frequency in the adult populations, suggestive of an inherited lethal phenotype (Todd et al. 2020). As LY49B is expressed in the trophoblast cells during equine placentation (Read et al. 2018) aberrations in this gene, could theoretically result in pregnancy loss but the variant is yet to be identified in abortus material from any stage of gestation.

Copy number variations (CNV) are large deletions or duplications of the genome, over 1 kb in length, naturally occurring during the cell cycle. It has been estimated that 4.8–9.5% of the human genome comprises CNVs (Kloosterman et al. 2015; Zarrei et al. 2015) compared with approximately 1.3% of the genome in the horse (Ghosh et al. 2014). Variation in copy number from wild type can, however, be associated with disease susceptibility, notably with the ageing process and increased risk of tumour (Hanahan and Weinberg 2011) and neurological disease (Girirajan et al. 2011). The size, location and the CNV type (duplication or deletion), as well as the genes located within the CNV, are all factors that contribute to the arising phenotype. It has been estimated that 14.6% of CNVs are found in exons compared with approximately 1.3% of SNPs (Kloosterman et al. 2015), suggesting that CNVs may contribute to disease risk at higher rates than previously thought.

Copy number variations are a normal feature of human placenta (Kasak et al. 2015; Coorens et al. 2021) which increase in number across gestation and also exceed numbers expected in human blood. Numerous studies have added to the evidence for CNV involvement in human pregnancy loss across gestation (Rajcan-Separovic et al. 2010a, 2010b; Levy et al. 2014; Nagirnaja et al. 2014; Karim et al. 2017; Kasak et al. 2017; Li et al. 2018; Fan et al. 2020; Wang et al. 2020) and unexplained stillbirth (Harris et al. 2011; Ernst et al. 2015). A single study to date has profiled the CNVs in normal human placental development (Kasak et al. 2015). A three-fold increase of N_CNV in placentae was noted compared with the parental blood, and a significant increase in N_CNV was found through healthy placental development. Using trios (maternal–paternal–placental), Kasak et al. (2015) found 11.1% of the CNVs in the placental tissue were inherited, and were equally inherited from each parent. Research into CNV influence on both normal pregnancy and pregnancy complications in the horse is lacking, although unpublished data in our laboratory suggests this feature will also be important to equine placental function and pathologies.

The first 2 months of equine pregnancy involve transformation of the simple zygote to a sophisticated foeto-placental unit through a series of spatially and temporally co-ordinated developmental events. Perhaps it is not surprising then that this period is frequently found as the most common for a pregnancy to fail, although foetal lethality is possible right up to parturition. Identifying causes of pregnancy loss throughout gestation, remains an area of intense interest of clinicians and researchers alike. As in vitro produced embryos become increasingly popular with breeders (Lazzari et al. 2020), excitement reignites around application of IVF in the horse (Felix et al. 2022), and the need to meet the challenge of managing the negative impact of deleterious mutations in breeds becomes more urgent (Orlando and Librado 2019), it seems no better time to expand our understanding of lethal genetic variants of the equine foetus.

In this review, we identified three types of variants directly associated with pregnancy loss in the mare: aneuploidy (Shilton et al. 2020), SNPs (Ducro et al. 2015; Monthoux et al. 2015) and translocations (Bugno-Poniewierska and Raudsepp 2021), with varying levels of evidence for lethality. Structural variants and SNPs of the mare’s genome have been found to increase the risk of that individual suffering an abortion or recurrent early embryonic loss. Presumably these variants act on the environment that mare provides for the developing foetus or lead to an imbalanced gamete genome which is inherited by the embryo through the germline. Duos have not been assessed to confirm either of these assumptions.

Variants intrinsic to the foeto-placental unit (aneuploidy, SNPs) have been identified that are thought to directly impact the development of the foetus and/or placenta through either (1) modulating the expression and/or function of an individual protein (Monthoux et al. 2015) or (2) leading to widespread genome instability (Shilton et al. 2020). These foetal variants have stronger evidence for lethality. The National Institute of Child Health and Human Development classifies chromosomal abnormalities as meeting the criteria for causality (Reddy et al. 2009). The evidence that equine Trisomy 1, 3, 15, 20, 23/34 and monosomy 27 and 31 are foetal lethal lies the fact that none of these variants have ever been reported in viable horses, despite both large scale and targeted screening by cytogenetic laboratories globally. Using this argument, trisomy 30 in EPL tissue may not be causal for pregnancy loss as it is described in a number of liveborn horses, although heterogenous phenotypes associated with human Trisomy 21 suggests this remains plausible (Hwang et al. 2021). It is curious that to date neither confined placental mosaicism nor foetal–placental mosaicism is yet to be identified, but we must not be complacent and assume it does not exist. It may well not be until blastocyst screening takes off that we really will know the true extent, plausible now with expanding numbers of IVP embryos (Lazzari et al. 2020; Felix et al. 2022). Evidence available up to this point suggests equine aneuploidy likely originates in maternal meiosis, responsible for 84% of human aneuploid conceptuses (Mikwar et al. 2020). Based on the criteria of Reddy et al. 2009, SNPs in PLOD1 and B3GALNT2 also have strong evidence for causality with few or no examples of viable offspring, and clear biological plausibility linking the affected protein and resultant foetal pathology.

In conclusion, research in this field in the horse is very much in its infancy but shows great promise for future diagnostic tests and to inform breeding decisions that avoid lethality. Research on techniques such as non-invasive prenatal testing through the detection of foetally derived DNA in pregnant mare serum has been initiated (Tonekaboni et al. 2020; Kadivar et al. 2021) and based on its success in women (Guy et al. 2021), is likely to be valuable to make this research relevant to clinical practice. To date, there are no foetal only or placental only variants, although it is reasonable to assume with further research these will be revealed in due course.

All data is available in the original manuscripts discussed in this review.

The authors declare no conflicts of interest.

This research did not receive any specific funding. Authors were funded for related work by Horserace Betting Levy Board and Thoroughbred Breeders Association.