A racing start in life? The hurdles of equine feto-placental pathology

Summary

Significant progress has been made in understanding and monitoring the causes of equine abortion over past decades. However, not all in utero pathology results in abortion. It has long been recognised that some in utero pathology, such as twinning or chronic placentitis, can result in the birth of live but growth-retarded foals and there is historical evidence that birth weight may influence future athletic performance. Clinical experience (e.g. from twins) and experimental studies (pony-Thoroughbred embryo transfer) have highlighted the importance of reduced functional placental area in limiting growth in utero in horses. Many other nonfatal in utero pathologies (e.g. umbilical cord-related circulatory compromise) can potentially affect either placental function or other organ systems. Their influence on the short- and long-term health of the foal and its future athletic performance is in many cases poorly documented or understood. This review summarises the main causes of in utero pathology and reflects on how these may potentially affect the foal if born alive, highlighting the need for long-term studies on this important subject.

Introduction

Although life in utero occurs in a protected environment, the developing fetus is in no way immune from infectious and noninfectious insult. The outcome of in utero, parturient and post parturient pathology can be varied; some pathologies may be inconsequential for normal development of the foal, while many are fatal and lead to abortion or foal death. In between, there is a spectrum of less severe pathologies that can result in the birth of a live foal, but the disease process, or its effects, may continue into the post natal period and beyond. Data on the outcome of such pathology in live foals are surprisingly limited. The aim of this review is to provide some ideas on how individual pathologies may affect development of the foal and subsequent performance and to encourage further investigation in this field. Understanding the range of ante- and perinatal pathology, both fatal and nonfatal, should provide an evidence-base for optimisation of clinical diagnosis, decisions on whether to start or continue treatment, and the nature of treatments to provide live foals. The review will focus on mid and late gestational pathology rather than early pregnancy loss. Early pregnancy failure (Ricketts 2003) is an interesting and developing area of research, and some later pathology may have its origin in these stages; however, at present, in most cases it is not possible to harvest satisfactory material for pathological investigation.

Incidence of equine perinatal losses and importance of diagnostic pathology

Prompt and accurate diagnosis of perinatal disease has for many years played a vital role in the equine breeding industry. Thorough necropsy and histological examinations, combined with new developments in microbiology and molecular diagnostic techniques have advanced the prevention and control of a number of infectious diseases. Accurate laboratory data provide an evidence base to enable practitioners to manage disease outbreaks, often under guidelines such as those issued in the UK and some other European countries by the Horserace Betting Levy Board (Anon 2011).

Defining the incidence of equine pregnancy failure and perinatal losses is complex and many forms of analysis have been used. Analysis of Weatherbys' Annual ‘Returns’ of UK and Irish Thoroughbred mares suggests that there have been significant improvements in conception rates, live foal rates and barren mare rates during the last 30 years (Ricketts and Troedsson 2007) (Table 1). For 14,109 UK and Irish Thoroughbred mares mated by 626 Thoroughbred stallions (after having deducted ‘no returns’ and those mares who died or who were exported and for which accurate results are unknown), 2.17% were returned by their owners as having ‘aborted early’ (confirmed pregnant at 40–50 days and subsequently found not to be in foal), 2.11% as having ‘slipped’ (confirmed pregnant at 40–50 days and aborts after 5 months, i.e. mostly abortions) and 3.86% as having produced ‘dead’ (i.e. stillborn) foals. This suggests an overall pregnancy/neonatal foal loss rate of 8.14%. This compares with rates of 8.74% in 2000, 8.60% in 1990 and 8.88% in 1980.

Table 1. A review of Weatherbys' Annual Returns for UK and Irish Thoroughbred Mares for the last 30 years

Weatherbys' Annual Returns for UK and Irish Thoroughbred Mares	1980	1990	2000	2010
* Before the widespread use of ultrasound scanning for equine pregnancy diagnosis made accurate diagnosis of early pregnancy failure possible.
Mares covered by TB stallions minus no returns, dead and exported (n)	11,586	15,307	16,519	14,109
Conceptions (%)	79.73	86.26	89.59	93.64
Live foals (%)	70.84	77.66	80.86	85.50
Barren mares (%)	20.27	14.33	10.41	11.68
Pregnancy failures (%)	8.88	8.60	8.74	8.14
Early pregnancy failures (%)	1.94*	2.04	2.55	2.17
Abortions (%)	3.12	2.38	2.45	2.11
Stillbirths (%)	3.82	4.17	3.74	3.86

These data reflect each mare's fertility status at the end of each season, as defined by their owners. Significant numbers of mares will have suffered early pregnancy failures only to conceive again and produce a healthy foal next year and this will not be reflected in these returns, underestimating the true incidence of early pregnancy failures in the population. Morris and Allen (2002) analysed the veterinary examination and subsequent foaling records of 1393 Thoroughbred mares visiting stallions in the Newmarket area during the 1998 season in detail, and found that 17.38% had records of pregnancy failure, during that season.

Several surveys of equine abortion have been published (Bain 1969; Platt 1973a,b; Pospischil et al. 1992; Giles et al. 1993; Hong et al. 1993a; Tengelsen et al. 1997; Ricketts et al. 2001; Smith et al. 2003; Szeredi et al. 2008; Laugier et al. 2011). Direct comparisons between studies on the incidence of different causes of abortion is difficult due to use of different categories/classifications of disease, different levels of diagnostic investigation and in some studies inclusion of neonatal foal deaths as well as abortions. However, these studies do give some indication of general trends and geographical variation in the causes of abortion and foal death. Early, pre-ultrasound scan studies in Australia suggested an overall abortion rate of 19%, with a rate of 9.3% after 60 days gestation (Bain 1969), and abortion rates of up to 75% have been reported in prevaccination outbreaks of equine herpesvirus (EHV) in the UK (Mumford et al. 1987). Before ultrasound scanning for pregnancy became widespread, twin pregnancy was the most important cause of abortion diagnosed in UK, whereas twin pregnancy is now seldom seen and umbilical cord torsion, often associated with a long umbilical cord, is now the most important cause in the UK (Ricketts et al. 2001; Smith et al. 2003). In US surveys, bacterial fetoplacental infections are the most frequent cause of abortion (Fig 1) (Giles et al. 1993) and it is not clear if these global differences reflect differences in diagnostic procedures and interpretation, climatic/environmental or managemental differences influencing the risk of placentitis, or risk factors affecting cord length and a predisposition to torsion.

Figure 1. Relationship between foal birthweight and subsequent racing performance (Platt 1978). EHV = equine herpesvirus.

Effects of in utero disease on foal health and future performance: general principles

Thoroughbreds are bred for racing, and their performances are well documented. As discussed by Platt (1978), racehorses are therefore good subjects for studying the effects of perinatal disorders on physical development and achievement. A few studies have considered how post natal disease, particularly neonatal bacterial infections, may influence outcome and future performance (Baker et al. 1986; Morley and Townsend 1997; Smith et al. 2004; Sanchez et al. 2008; Hemberg et al. 2010). While it has long been recognised that twin foals rarely develop into successful racehorses, there are relatively few studies that examine the performance potential for single foals of low birth weight and, to the authors' knowledge, none that correlate the different causes of low birthweight (in utero pathology) with subsequent performance. In the 1970s, Platt (1975a, 1978) provided some preliminary data from foals that were undersized at birth. Of the 43 undersized single foals born during the survey, 36 survived to racing age but the proportion that actually raced was lower than in a control group of 997 foals of normal birthweight (Fig 2). About a quarter of the normal birthweight group achieved Timeform ratings of 80–109 (winning Thoroughbred racehorses achieve Timeform ratings >100 and elite racehorses >125), but only one animal in the low birthweight group reached this standard.

Figure 2. Incidence of abortion and perinatal pathology: Comparison of UK and USA surveys (Giles et al. 1993; Smith et al. 2003).

Since foal birth weight may affect future performance, it is relevant to consider how different pathologies lead to intrauterine growth retardation. In Platt's studies, twinning and fungal placentitis were documented as a cause of low birthweight, but many of the cases were unexplained. In an experimental setting, Allen et al. (2002) elegantly demonstrated, with the use of embryo transfer experiments between ponies and Thoroughbreds, how restricted uterine size leads to growth retardation. This is corroborated to some extent in the relationship between placental size/weight and foal weight (Whitwell and Jeffcott 1975; Cottrill et al. 1991; Elliott et al. 2009). Diseases that affect the functional placental area (twins, placentitis, villus atrophy, body pregnancy) are well recognised in causing growth retardation. Compensatory acceleration in post natal growth rate probably occurs in most cases (Allen et al. 2004), but growth is limited for some individuals and this particularly affects horses racing at a young age (Platt 1975a).

Experimentally, in utero growth retardation may also influence endocrine function in early post natal life (Ousey et al. 2004) and altered cardiovascular function (Giussani et al. 2003). The influence of inflammatory mediators, cardiovascular/cord compromise, hypoxia and other disease processes are poorly documented in horses, but may have interesting correlations with human perinatal disease and are discussed further in other reviews in this supplement Fowden et al. 2012; (Ousey and Fowden 2012).

With improving diagnostic techniques, it may be possible to detect and monitor dynamically more pathological conditions in utero to enable both preventative treatment and long-term follow-up studies. However, at present, we have to rely largely on placental examination to give a historical insight on life in utero and speculation as to how nonfatal pathologies may affect development of the foal from our knowledge of more severe fatal pathology (Table 2).

Table 2. Possible outcomes associated with some forms of perinatal pathology

	Abortion/stillbirth	Normal foal	Premature foal	Growth retardation	Infected at birth	Deformities/malformation
++ frequent; + sometimes; +/-rare; - never/very rare; ? uncertain. EHV = equine herpesvirus; EVA = equine viral arteritis.
EHV	++	?	+	?	++	-
EVA	+	+	+	?	+/-	-
Placentitis	++	+/-	+	++	+/-	-
Twinning	++	+/-	+	++	n/a	+/-
Cord torsion	++	?	+/-	?	n/a	-
Premature placental separation	++	+	+/-	+/-	n/a	-
Body pregnancy	++	+/-	+	+	n/a	+
Villus atrophy or hypoplasia	+	+	+/-	+	n/a	-
Hydrops	+/-	?	++	++	n/a	++

Pathology of mid-late term pregnancy failure, stillbirth and perinatal foal death

Infectious – viral

Equine herpesvirus-1: EHV is the most important infectious cause of abortion and can be caused by EHV-1 and, less commonly, EHV-4. The EHV-1 circulates in horse populations via subclinical or clinical infection after primary infection or reactivation of a latent state (Brown et al. 2007). Pregnant mares can be infected without causing abortion but transplacental infection results in rapid placental separation and the fetus is aborted within its membranes. Abortion usually occurs between 7 months and full term, but foals can be born alive when infection occurs in late gestation. Typical gross lesions of EHV-1 infection in the fetus/foal are an excess of clear yellow body cavity effusion and multifocal hepatic necrosis, sometimes visible as white spots on the liver capsule and cut surface. There is often perirenal oedema. The lungs may be firm and mottled. The placenta may be oedematous or normal. Histopathological investigations will usually confirm hepatic and bronchiolar necrosis, as well as necrosis in the adrenal cortex, lymph nodes, spleen and thymus; detection of classical intranuclear viral inclusions is pathognomonic, but are not always present. Placental morphological changes are not specific and include endothelial necrosis and oedema. Using immunohistochemistry, polymerase chain reaction (PCR) or in situ hybridisation, viral antigen or deoxyribonucleic acid (DNA) can be detected (Gerst et al. 2003). The relative involvement of different organs varies between cases and sampling of multiple tissues is advisable to achieve a diagnosis. Examination of the placenta is important, since placental infection can occur without fetal infection (Smith et al. 1992), presumed to be related to rapid onset of placental separation at sites of maternal uterine micro-thrombosis prior to infection across the placenta.

Abortion rates as high as 75% can occur (Mumford et al. 1987) and investigation of all abortions and foal deaths is critical to enable a rapid diagnosis and appropriate preventative action. Unlike with equine viral arteritis (EVA) infection, gross pathological changes are often suggestive of disease and the diagnosis has been further improved with the availability of rapid molecular tests, such as real-time PCR.

Where mares are infected or develop recrudescent viraemia close to term, foals may be born alive but infected. Such foals are usually weak and develop viral pneumonia within the first few days of life. They often succumb to secondary infections and complications of hypoxia. Some foals may appear clinically normal for the first few days of life prior to succumbing to disease and, while most infected foals that are born alive die within the first few days of life from severe respiratory disease, there clearly is potential for some foals to survive if the pathological changes are milder. Whitwell et al. (1992) demonstrated that of 84 aborted fetuses/stillbirths, 9 had EHV-1 antibody titres, but only 3 were proven EHV abortions, suggesting perhaps that some fetuses survive long enough to generate an antibody response to infection (or that there was breakdown of the uteroplacental barrier). Since the placental separation in EHV infection occurs rapidly in most cases, in utero growth retardation is an unlikely consequence of infection for foals that survive infection close to term, but foals may be premature. If subclinical/nonabortigenic infection does occur, the effects on chorionic villi are not well documented, but perhaps could cause some loss of function.

Equine herpesvirus-4 causes abortion less commonly and, unlike EHV-1, cases are usually sporadic with less risk of an outbreak of abortion developing. An excess of fluid in the body cavities is usually a feature, but gross and histopathological lesions in the lung and liver may be very sparse or absent. The spleen is considered the most reliable organ from which to isolate virus (Whitwell et al. 1994). Since most of the pathological lesions observed in EHV-4 aborted foals are more suggestive of hypoxic injury than direct organ damage by the virus, infected foals that survive the birth process may be more likely to be viable than compared with EHV-1 infection.

Equine viral arteritis: This is caused by an arterivirus and the vascular system is the principal but not unique viral target. It has variable presentations in mature horses, including interstitial pneumonia, panvasculitis with oedema, thrombosis and haemorrhage, lymphoid necrosis, renal tubular necrosis, abortion and inflammation of male accessory genital glands. The disease has major financial implications on international trade in horses and equine semen (reviewed in Timoney 2011). In countries where EVA is endemic, the virus causes sporadic or occasionally epidemic outbreaks of abortion. In a survey in Hungary, 8.3% of abortions were EVA positive. Pathological lesions can be very mild or absent in the fetus. If present at all, there may only be mild perivascular lymphocytic infiltrates and mild interstitial pneumonia. Rarely, there can be more severe vasculitis involving the allantochorion, brain, liver, spleen and lung. The disease is most reliably detected by PCR, immunohistochemical or fetal serological investigations (Del Piero 2000; Szeredi et al. 2005).

Experimentally, infected mares can deliver a live fetus, and the fetus may or may not be infected with the virus (Coignoul and Cheville 1984). Although fetal death may occur in utero during acute EVA, abortion probably is due to lesions in the uterus of the mare rather than the fetus. As with EHV infection, significant in utero growth retardation is unlikely, but the effects of maternal illness could potentially affect development if abortion or premature delivery does not occur.

Equine infectious anaemia: This lentiviral infection can cause abortion. Similar to EVA, abortion is probably associated mainly with maternal infection and pyrexia, although transplacental infection of foals has been documented (Kemen and Coggins 1972).

Infectious – bacterial and fungal

Placentitis: The incidence and aetiology of placentitis shows quite marked geographical variation, with higher incidence usually reported in the USA than Europe. From a diagnostic point of view, it is important to be aware that some cases of placentitis, particularly acute cases, are not evident macroscopically and require histologically confirmation (Whitwell 1988). Histology of normal and an obviously affected chorion are illustrated in Figures 4. In most countries, Streptococcus equi ssp. zooepidemicus and Escherichia coli infections are commonly encountered (Whitwell 1988; Hong et al. 1993a; Smith et al. 2003), but many other bacteria can be involved, including Pseudomonas aeruginosa, Klebsiella pneumonia, Actinobacillus spp., Staphylococcus spp. and Leptospira spp. Nocardiform actinomyctes have been a significant cause of focal placentitis in parts of the USA since the 1990s. Aspergillus and other mucor species are the most common cause of mycotic placentitis (Hong et al. 1993a).

Figure 4. a) Normal chorionic villi. b) Bacterial placentitis. The villi are expanded by dense infiltrates of neutrophils, sometimes surrounding colonies of bacteria (arrow, inset). c) Ischaemic necrosis of the cervical pole: full-thickness necrosis of the chorioallantois (N) with some mineralisation of the chorionic villi and neovascularisation at the junction between the viable and necrotic zones (arrow). d) Umbilical cord torsion: mineralisation of blood vessels and stroma (arrows) is often observed within the chorionic villi. H&E staining; a)-c) 20 × magnification; d) 200 × magnification.

Placental infection often results from ascending spread through the cervical canal (Platt 1975b) and is most common in later pregnancy. Focal placentitis is more likely to be related to haematogenous spread from the mare; it is less common in the UK, but the incidence of focal nocardioform placentitis has increased in the USA. This placentitis is usually associated with infection by the Gram-positive, filamentous, branching bacterium, Crossiella equi sp. nov. (Giles et al. 1993; Donahue et al. 2002; Cattoli et al. 2004), but other species of nocardioform bacteria can be involved (Bolin et al. 2004). The often focal distribution of the exudative mucopurulent and necrotising placentitis, most commonly centred on the ventral aspect of the junction of the placental horns and body, suggests a haematogenous route of infection rather than an ascending infection, but the route of entry and reason for the rise in incidence is currently uncertain.

Environmental bacteria that are not normally pathogenic may become opportunistic pathogens, particularly if the host is compromised. Less commonly reported causes of placentitis and abortion include Rhodococcus equi (pyogranulomatous placentitis and fetal pneumonia) (Patterson-Kane et al. 2002), Encephalitozoon cuniculi (necrotising placentitis) (Patterson-Kane et al. 2003; Szeredi et al. 2007), Dermatophilus congolensis (placentitis, funisitis and fetal bronchopneumonia) (Sebastian et al. 2008a) and Corynebacterium pseudotuberculosis (Poonacha and Donahue 1995).

Leptospirosis has been reported as a significant cause of fetal loss in horses in Kentucky and Northern Ireland, but is less commonly documented in other countries (Donahue et al. 1995; Donahue and Williams 2000; Whitwell et al. 2009). Most abortions result from infection by Leptospira interrogans serovars kennewicki or bratislava. When present, gross placental lesions include funisitis, nodular adenomatous hyperplasia of the allantois, oedema, areas of necrosis of the chorion, and necrotic mucoid exudate coating the chorion (Poonacha et al. 1993; Sebastian et al. 2005; Szeredi and Haake 2006). Histological lesions may include thrombosis, vasculitis, mixed inflammatory cell infiltration of the stroma and villi (Fig. 4b), cystic adenomatous hyperplasia of allantoic epithelium, and villous necrosis and calcification. Fetal lesions are variable, but may include hepatocellular dissociation, mixed leucocytic infiltration of the portal triads, giant cell hepatopathy, interstital nephritis, pulmonary haemorrhages, pneumonia, myocarditis and adrenocortical inflammatory infiltrates (Hodgin et al. 1989; Whitwell et al. 2009).

In general terms, the consequences of most types of chronic placentitis will be similar. The fetus/foal is usually growth retarded, probably as a combination of loss of functioning placental area, direct effects of inflammatory mediators on the developing fetus and prematurity. Not all foals born alive will be infected, but the risk of congenital sepsis may be increased. In 14 out of 17 cases of abortion with acute placentitis there was evidence of septicaemia in the foal (Whitwell 1988). It is important to remember that inflammation is not confined to the placenta, but also affects the uterus, stimulating production of prostaglandins (PGE2 and PFG2α) resulting in myometrial contraction and premature lactation (LeBlanc et al. 2002). In man, proinflammatory cytokines may have beneficial as well as negative effects, indirectly stimulating the fetal hypothalamic-pituitary-adrenal axis resulting in precocious maturation of the fetus and increasing the chance of survival if the fetus survives long enough to benefit from these influences (Gravett et al. 2000).

As well as the more common bacterial causes of placentitis, Leptospira spp. (Hodgin et al. 1989), Salmonella abortusequi (Madićet al. 1997), Campylobacter spp. (Hong and Donahue 1989) and Listeria monocytogenes (Welsh 1983) are examples of organisms that can cause systemic pathology in the fetus/foal as well as the placenta. Distinguishing infection acquired in utero from those acquired neonatally (usually associated with failure of passive transfer of immunity or other immunocompromise) may be difficult. The longer-term consequences for the foal will clearly depend on the extent of disease/pathology, but may include damage to respiratory function following pneumonia, neurological damage secondary to hypoxia, or septic arthritis. Interestingly, Whitwell noted an increased incidence of carpal flexion in cases of chronic placentitis (Whitwell 1988). In most cases, if a compromised foal is born alive in association with placentitis and survives neonatal critical care to thrive in an apparently satisfactory manner, lasting effects on its organ systems can only be speculated upon.

Mare reproductive loss syndrome/equine amnionitis and fetal loss: An unusual epidemic of early fetal loss (EFL), late fetal loss (LFL), fibrinous pericarditis, and unilateral uveitis occurred in the USA during the spring of 2001 – now referred to as mare reproductive loss syndrome (MRLS) (Cohen et al. 2003; Sebastian et al. 2008b). The same syndrome with lesser intensity recurred in 2002. The estimated economic loss from this syndrome in 2001 and 2002 together was estimated at approximately $500 million. Both EFL and LFL were characterised by the absence of specific clinical signs in aborting mares. Nonhaemolytic Streptococcus spp. and Actinobacillus spp. accounted for 65% of the organisms isolated from fetuses submitted for a post mortem examination during the MRLS outbreaks of 2001 and 2002. Reported gross lesions in LFL include pale brown, thick, oedematous placenta, a thick, oedematous, yellowish umbilical cord, and, in some fetuses, pulmonary oedema. Some foals born alive have hyphaema present at the time of delivery. Histopathological changes associated with LFL included funisitis, mostly affecting the amniotic portion, placentitis centred on the extra-embryonic coelom and pneumonia. In the early fetal loss cases, the fetus is often autolysed, but a degree of placentitis may be detected and turbid amniotic fluid was seen during transrectal ultrasound scan examinations of mares undergoing EFL (Sebastian et al. 2008a). Epidemiological studies suggest an association between the presence of eastern tent caterpillars (ETC) in pastures with MRLS. Experimental studies in pregnant mares by exposure to ETC, or their administration by stomach tube or with feed material, reproduced EFL and LFL. Currently, 2 hypotheses are proposed for MRLS. One proposes that an ETC-related toxin with secondary opportunistic bacterial invasion of the fetus leads to MRLS. The second hypothesis suggests that a breach of gastrointestinal mucosal integrity by barbed hairs (setae) of ETC leads to an opportunist bacteraemia and results in MRLS. It is interesting that the chorionic villi are not infiltrated by inflammatory cells as observed in classical ascending placentitis, and the route of infection from the mare to fetus and amnion requires further elucidation in the proposed hypotheses. In 2004, a similar equine abortion storm, named equine amnionitis and fetal loss, was reported from Australia and caterpillar exposure was again identified as a risk factor for the abortion (Cawdell-Smith et al. 2011; Todhunter et al. 2009)

Noninfectious

Twinning: Historically, twinning was the most common cause of abortion in Thoroughbreds (Jeffcott and Whitwell 1973), but the incidence has dramatically decreased since the advent of ultrasonographic examinations in routine studfarm practice. The presence of 2, rather than one diffuse microcotyledonary placenta in the equine uterus, with juxtaposition of the 2 placentae resulting in avillous areas of the chorion, reduces the functional placental surface area available to support both fetuses. Death of the smaller fetus is usually attributable to placental insufficiency and, if the larger fetus survives to term, it will usually be impoverished and growth retarded, highlighting the critical importance of functional placental area in restricting growth. Abortion or stillbirth of both foals occurs in 64.5% of mares that carry twins, 21% delivering one live foal and 14.5% both foals live. There are different patterns of contact between the 2 placentae, and this influences the size of each fetus and odds of survival (Jeffcott and Whitwell 1973). Although twins are nearly always dizygotic, vascular anastomoses can occur between the 2 placentae, resulting in blood chimerism. However, this is thought to occur late in development after gonadal differentiation, so ‘freemartinism’ is not encountered, unlike the situation in cattle. Pseudo-anastamoses at the site of apposition of the twin placentae can remain functional after the death of one twin.

Umbilical cord pathology: Excessive length or shortness of the umbilical cord are risk factors for significant placental pathology and fetal death. The total length of the cord in normal Thoroughbred foals at term is between 36–83 cm (Whitwell and Jeffcott 1975). Factors governing cord length are not known, although they are statistically longer in male foals and foals from older mares, and there is a weak sire effect (Whitwell and Wood 1992). Increased length has been associated with a risk of excessive torsion of the cord or entrapment around the fetus. Most abortions due to cord torsion occur in Months 6–8 of gestation (Whitwell 1975). High fetal mobility in Months 4–7 of gestation provides the greatest opportunity for torsion, with mobility diminishing greatly after 7 months (Ginther 1992).

A degree of cord twisting is common in apparently normal foals. Obstruction to urine flow from the bladder to the allantoic cavity (e.g. by its compression at a site of cord twisting, or entrapment) leads to dilatation of the thin-walled urachus proximal to a compression site. Multiple twists can give rise to multiple dilatations. If close to the navel, bladder distension can occur and, in extreme cases contribute to rupture of the bladder or urachus, or to pervious urachus after birth. Excessive and/or tight cord twisting, however, is life threatening as it compromises or obstructs feto-placental bloodflow in the local cord vessels, leading to fetal death (Fig. 3b). This is evidenced by compression of the umbilical vessels at sites of torsion, and other gross changes locally such as cord swelling with allantoic ‘stretch marks’, oedema and haemorrhage into the wall of the cord vessels, tearing of the intima of the vessels, small aneurysms and thrombosis within vessels (Whitwell 1975). Macroscopically the fetus is usually autolysed, since its death precedes abortion. There is usually an excess of haemorrhagic fluid in the body cavities and sometimes oedema around the umbilical area. Fetuses with long and twisted cords often have apparent elongation or stretching of the navel stalk, possibly associated with tension. The most consistent, but not specific, histopathological features of cord torsion are mineralisation of blood vessels and stroma in the chorioallantois (Fig. 4d) and signs of villus core degeneration, with chorionic surface epithelium appearing better preserved. Microthrombi or their fragments are often noted.

Figure 3. a) An emaciated and growth retarded foal close to term highlights the consequences of growth retardation. In this case, the cause was secondary to chronic placentitis but other pathology can have similar consequences. b) Cord torsion: appearance of the amniotic region of an umbilical cord after untwisting of a tightly twisted sector. Blanched bands reflect sites of maximum constriction (arrows). A line of congestion and probably local haemorrhage borders the left pale band. c) Chronic placentitis at the cervical pole: the chorion is thickened and roughened, with cream surface exudates. d) Ischaemic necrosis of the cervical pole of the chorion. The image shows the allantoic surface of the body of the chorion and a well-demarcated line separating necrotic from normal chorion (white arrows). e) Body pregnancy: very wide chorionic body dimensions, which contained the whole fetus and fetal fluids. The horns of the chorion have expanded into the uterine horns but at abortion the horns are not distended by fetal fluids or any part of the fetus. f) Extensive villus atrophy (pale zones, higher magnification inset) over the body and pregnant horn of this placenta was considered secondary to hydrops amnion in this case.

The consequences of mild or intermittent nonfatal cord twisting is not known in horses but is recognised in human pregnancies (Baergen 2005). In man, approximately 50% of stillbirths are unexplained after fetopsy and placental examination, but fatal hypoxic injury due to restriction of umbilical flow (‘cord accident') may be causal in a subset of these stillbirths. Histopathological features seen in the placentae of suspected cord accident cases are similar to those that we observe in cases of equine cord torsion (Parast et al. 2008). Human infants that survive cord compromise may, in the short term, develop ischaemic renal damage and may have longer-term cognitive deficits. Possible influences of nonfatal cord compromise have not been investigated in horses.

Long cords have also been linked to localised ischaemic necrosis of the allantochorion, usually at the cervical pole (Fig. 3d & 4c) or sometimes other loci, e.g. horn tips. Thrombosis and hypoxia are probably factors in the aetiology of such lesions. A short cord has also sometimes been associated with ischaemic necrosis (K.E. Whitwell, personal observation) and with premature intrapartum cord rupture.

Three variants of chorioallantoic vascular patterns have been described in the mare pregnancy, centred on cord attachment sites at the base of the horns (Whitwell KE, unpublished observations, cited in Rossdale and Ricketts 2002). Abnormal cord attachment results from an inappropriate fixation position of the conceptus in the uterus: the potential for this to result in fetal and neonatal compromise has been considered (Wilsher et al. 2009). Fixation of early equine pregnancy in the uterine body rather than one of the uterine horns can result in abortion.

Placentopathies: The causes of premature placental separation are multifactorial, and in many cases the cause remains unidentified. Rapid premature separation is a feature of EHV infection, secondary to uteroplacental vascular pathology. Separation may also occur with fescue toxicosis, MRLS, placentitis and ischaemic necrosis of the cervical pole (Fig. 4c). Extensive placental separation will result in death of the fetus/foal and abortion, but partial separation may be compatible with continuation of the pregnancy to term, with fetal compromise a likely outcome of reduced functional placental area.

Body pregnancy (Fig 3e) is a rare condition. Implantation usually occurs in the correct location and the chorioallantois does extend into the uterine horns, but the fetus and fetal fluids do not. When abortion occurs, the horns are markedly narrowed and the uterine body wider than usual. The cause for this utero-chorial horn shrinkage is not known.

Areas of the placenta will sometimes have small sparse villi or appear avillous. It is often difficult to determine if this is atrophy or hypoplasia of the villi Fig 3f. It can be a remarkable feature of several causes of abortion, including hydrops allantois, or may relate to focal or diffuse chronic degenerative endometrial disease. It has obvious potential consequences for placental function and retarded growth, depending on the extent. Variable degrees of villus atrophy are not uncommon alongside the cervical folds in otherwise normal placentae, usually considered an incidental finding.

Adenomatous hyperplasia/dysplasia of the equine allantois is usually considered secondary to chronic placental pathology of other causes (Shivaprasad et al. 1984; McEntee et al. 1988; Hong et al. 1993b). The lesions can be focal or multifocal and probably do not directly affect placental function significantly in themselves. The nodular masses are formed on the allantoic surface by hypertrophy and hyperplasia of allantoic epithelium, with metaplastic formation of epithelial glandular structures, often associated with luminal infiltrates of neutrophils.

The aetiology of hydrops amnion and hydrops allantois remains obscure. The production and regulation of both allantoic and amniotic fluids is complicated (Waelchli 2011). Both forms of hydrops may be associated with congenital defects (Vandeplassche et al. 1976; Allen 1986; Waelchli and Ehrensperger 1988). Altered fetal swallowing (e.g. secondary to a cleft palate or neurological dysfunction) may lead to an accumulation of amniotic fluid, but experimental obstruction of the oesophagus in other species does not support this theory (Wintour et al. 1977). Hydrops can precipitate prepubic tendon or body wall rupture in the mare and sometimes results in uterine inertia during delivery, probably as a result of overstretching of the abdominal and uterine musculature. The prognosis for survival of the foal is usually poor; not only may the foal have congenital defects, but the placentae have poor villous development and placental function is usually severely compromised.

Endocrine abnormalities: Endocrine pathology is relatively rare as a primary cause of abortion. A syndrome of neonatal foals characterised by hyperplasia of the thyroid gland and concurrent musculoskeletal deformities (TH-MSD) has been described in western Canada (Allen et al. 1994), but has not been widely reported elsewhere, perhaps supporting an environmental/feed related cause. Although primary endocrine dysfunction is rare, abnormalities in maturation of the hypothalamic-pituitary-adrenal axis are a feature of premature or dysmature foals (Rossdale et al. 1991).

Neoplasia: Rarely, fetuses or neonatal foals may develop neoplastic lesions. Benign cutaneous papillomas are occasionally observed. Lymphoma has been documented, confined to the fetus and not the mare (Haley and Spraker 1983). Other fetal tumours include hepatoblastoma (Neu 1993), pleuropulmonary blastema (Woolford et al. 2010) and rhabdomyoma (Meyerholz et al. 2004), as well as a range of congenital hamartomas. Placental neoplasia is also extremely rare, with reported cases of placental teratoma (Gurfield and Benirschke 2003), teratocarcinoma (Allison et al. 2004) and choriocarcinoma (Scase and Whitwell 2004). Metastasis of the teratocarcinoma was suspected in the foal post natally (Allison et al. 2004).

Maternal injuries, illness and toxicity: Maternal illness can lead directly to abortion via transplacental infection. Other illness, such as colic, pyrexia, starvation and neoplasia, may affect fetal growth indirectly (Wilsher and Allen 2006). In a study of pregnant mares with equine grass sickness (EGS), pregnancy failure occurred in gestations of <10 weeks. Later stage fetuses in mares with EGS did not show indication of ganglionic pathology (Whitwell 1992). Rarely, trauma to the mare's abdomen may sometimes cause in utero fractures to the foal, which may heal inappropriately during gestation and result in the birth of a live foal with a deformed limb (S.W. Ricketts and K.E. Whitwell, unpublished data).

Ingestion of tall fescue (Lolium arundianceum) infected with the endophytic fungus Neotyphodium coenophialum results in exposure to ergot alkaloids. This syndrome has been associated with several problems of pregnancy in the USA – increased gestational length, agalactia, stillborn foals, placental thickening/oedema, placental separation, dysmature foals and reduced serum progesterone and prolactin levels in the affected mares (Cross 2011). Foals that survive the prolonged gestation period are reported to be large-framed dysmature and emaciated, with overgrown hooves, sometimes with ‘dummy-like’ behaviour. They are prone to succumb to sepsis, probably from a failure to obtain adequate passive immunity from the colostrum. Lung maturation is reported to be impaired, with reduced pulmonary phospholipids in the amniotic fluid and impaired thyroid function (reviewed in Cross 2011).

Other conditions, which do not necessarily prevent conception, can compromise normal fetal development and/or the birth process. These include extreme forms of uterine pathology, e.g. extensive endometrial scarring, transuterine adhesions, congenital uterine malformation, and healed pelvic fractures, which may critically reduce pelvic dimensions.

Intrapartum and neonatal disease

Prematurity/dysmaturity: These syndromes are mostly caused by other conditions discussed in this review and have their origins in prenatal disturbances of fetal maturation and physiology (Rossdale 1993). Prematurity is a term ascribed to foals delivered at <320 days gestation whereas dysmaturity describes foals born in the full-term period showing premature-like signs. Dysmature foals are generally associated with placental pathology. However, the distinction between the 2 groups is tenuous and placental pathology is often present in premature foals (Cottrill et al. 1991). Premature/dysmature foals fall into 2 groups; those with a favourable clinical outcome and those that make progress during the first 24 h post partum only to deteriorate thereafter with the development of neurological, metabolic and respiratory deficits (‘second day syndrome’) (Rossdale 1993).

Perinatal asphyxia syndrome: The neonatal maladjustment syndrome (NMS)/perinatal asphyxia syndrome (PAS/hypoxic-ischaemic-encephalopathy) is a consequence of hypoxia leading to cerebral oedema, haemorrhage and/or ischaemic necrosis. Perinatal asphyxia is mostly associated with large foals, maiden mares, unattended deliveries and malpresentations leading to dystocia (Hong et al. 1993a). In addition to the neurological injuries, there may be associated atelectasis, meconium aspiration and/or rib fractures. Renal tubular necrosis, glomerular damage, ischaemic muscosal necrosis and ulceration of the intestinal tract, hepatocellular necrosis, myocardial ischaemia/infarction and adrenal necrosis have been associated with this condition (Vaala 1999).

Malformation/deformation of the foal may be a factor in dystocia, as can previous pelvic fractures in the mare narrowing the birth canal. As it is the foal's responsibility to achieve the correct posture in the birth canal, fetal malformation, deformation or illness, e.g. sepsis leading to weakness, may result in malposition of limbs or head leading to dystocia. It is not uncommon for clinicians dealing with dystocia cases to find such abnormalities in the foal during or following correction of dystocia and delivery. Degrees of neurological injury following asphyxia are variable and the long-term consequence of this injury on future performance of the horse is not well established. Many foals that recover appear to do well clinically but in such cases it is often impossible to exclude lasting neurological deficits.

Congenital malformations/deformations: A wide range of congenital defects occur in foals. While some have a genetic or toxic aetiology (true malformations), some of the more common conditions are believed to be a consequence of restricted positional life in utero (deformations) (Vandeplassche et al. 1984). Wry nose is a congenital deformation, more common in Arabian horses, but probably related to malposition in utero at critical stages of development. Flexural limb deformities are also a common deformation (Crowe and Swerczek 1985). Malformations include microphthalmia, craniofacial malformations, cleft palate, brachygnathia, heart defects, aganglionosis, intestinal atresia, umbilical defects, microencephaly, cerebellar hypoplasia, hydrocephalus hyperelastosis cutis and epitheliogenesis imperfecta. Although usually considered separately, immunodeficiencies, such as severe combined immunodeficiency (SCID) of Arab foals are a ‘malformation’ of the immune system as a consequence of a genetic defect (Wiler et al. 1995). Aetiological studies on most of these malformations are lacking, and some defects can have more than one cause – for example, torticollis/scoliosis may be a consequence of abnormal position during vertebral development or true vertebral malformation.

Neonatal sepsis and failure of passive transfer of immunity: Many cases of sepsis, manifesting during the first week of life, may be acquired post partum with the most important portals of entry for infection being the umbilicus, gastrointestinal and respiratory tracts. Other infections may have been acquired during gestation, most commonly in association with placentitis (see above) or perhaps parturition. While foals can recover if treated aggressively, success as an athlete may be compromised (Smith et al. 2004; Sanchez et al. 2008). Failure of transfer of colostral immunity remains an important predisposing factor for neonatal sepsis. Indirectly, achieving passive immunity may be affected by in utero pathology if that pathology affects the vitality of the foal and its ability to suckle or if it is associated with premature lactation in the mare, resulting in poor quality colostrum.

Conclusions and need for future studies

Rapid and accurate diagnostic pathology remains vital for the benefit of the attending veterinary surgeon, the individual owner and the wider equine industries. With the ever increasing risk of global movement of equine diseases, continuing surveillance for and awareness of new and emerging disease conditions remain as important as ever.

Standardisation of diagnosis and classifications would be useful to enable better global monitoring of changes in disease incidence (for example categories used in some surveys such as ‘placental oedema’, ‘villus atrophy’ or ‘fetal stress syndrome’ are relatively nonspecific features and probably have several different aetiologies, as discussed above.

Since the 1960s, by detailed post mortem examination of fetuses, placentae and foals, much progress has been made in identifying and defining details of the feto-placental pathology underlying equine pregnancy and neonatal losses. However the patho-physiological mechanisms leading to many of these failings is still poorly understood and requires much more research in order to control them. Using improved methodologies for imaging and clinico-pathological monitoring, a more holistic approach to assessment of a pregnancy's viability is needed to screen not only the health of the feto-placental unit but also the effects of external factors with the potential to influence the uterine environment and the mare herself (exercise-related, environmental, dietary and behavioural). In that way, the time frame allowing study of feto-placental disease can be extended retrogradely, so that the causative patho-physiology can be identified and studied when it first occurs and while the fetus is still alive. Improved pregnancy screening modalities may lead, for example, to better evaluation of the causes of cord-related problems, to how infections enter the pregnant uterus, the causes of foal deformations and stillbirth fatalities.

Hugh Platt's 1975 correlation of birth weight and athletic performance highlighted that life in utero can affect athletic ability, but there remain plenty of unanswered questions and, to progress, further long-term research studies are now required. In particular, nonfatal defects that develop during pregnancy or at birth need to be defined, and the extent to which such pathologies affect the foal's subsequent development, competence for extrauterine life and its performance as an athlete.

Conflicts of interest

The authors did not declare any conflict of interest.

Source of funding

None.

Acknowledgements

Thanks are due to the equine diagnostic laboratories at Rossdale & Partners and the Animal Health Trust, and their staff.