[Epiploic] Foramen Entrapment: a Retrospective Study of 145 Cases (2008-2016)

Horses undergoing surgery for omental [epiploic] foramen entrapment: a retrospective study of 145 cases (2008-2016)

Word count: 13.540

Michaël Swagemakers

Student number: 00906879

Supervisor: Prof. Dr. Ann Martens Supervisor: Dr. Thomas van Bergen

A dissertation submitted to Ghent University in partial fulfilment of the requirements for the degree of Master of Veterinary Medicine

Academic year: 2017 - 2018

Ghent University, its employees and/or students, give no warranty that the information provided in this thesis is accurate or exhaustive, nor that the content of this thesis will not constitute or result in any infringement of third-party rights. Ghent University, its employees and/or students do not accept any liability or responsibility for any use which may be made of the content or information given in the thesis, nor for any reliance which may be placed on any advice or information provided in this thesis.

Index

1. Abstract...... 4 Samenvatting ...... 4

2. Introduction ...... 5

3. Literature study...... 6 3.1 Anatomy of the omental foramen and the omental vestibule ...... 6 3.2 Occurrence ...... 6 3.3 Risk factors ...... 7 3.4 Pathophysiology ...... 8 3.5 Diagnosis ...... 9 3.6 Treatment ...... 11 3.7 Prognosis ...... 12

4. Retrospective study ...... 13 4.1 Materials and methods ...... 13 4.2 Results ...... 14

5. Discussion ...... 18

References ...... 21

Supplementary information items ...... 26

1. Abstract

The aim of this thesis is to document and analyze perioperative data, survival to and after discharge, and variables associated with survival of epiploic foramen entrapment (EFE) surgeries in a retrospective study of 145 horses that underwent surgery for EFE at the Faculty of Veterinary Medicine, Ghent University, Belgium between March 2008 and December 2016. Pre-, intra- and postoperative data were obtained. Post-operative process was assessed by follow-up telephone calls with the owner. Factors associated with postoperative ileus (POI), relaparotomy, hospital discharge, colic after hospital discharge and survival after discharge were assessed. The 145 surgeries were performed on 142 horses (recurrence-rate 3%). Warmblood horses represented 85% of the horses that underwent surgery and 62% of the patients were geldings. Windsucking/crib-biting was confirmed in 60% of these horses. In 100% of cases left-to-right entrapment occurred. In 74% of cases ileal involvement was recorded. Uncontrollable intraoperative hemorrhage was encountered in 6% of surgeries. One hundred and seven horses (74%) recovered from surgery. High morbidity and mortality associated with EFE were demonstrated. Survival to discharge of all horses undergoing surgery was 48% and survival to discharge of all horses recovering from surgery was 65%. Median survival of discharged horses was in excess of 3193 days. Statistical analysis on the 145 surgeries demonstrated that a resection performed predisposed horses for POI and that horses undergoing jejunoileostomy were more likely to develop POI compared to horses undergoing jejunojejunostomy. Horses that developed POI were less likely to be discharged and horses that underwent resection had shorter life expectations after hospital discharge.

Samenvatting

De bedoeling van deze thesis is om data omtrent de operatie, opvolging van de patiënt en de variabelen die geassocieerd zijn met het overleven van een “epiploic foramen entrapment” (EFE), te documenteren en te analyseren. Door middel van een retrospectieve studie is er onderzoek gedaan naar 145 paarden die tussen maart 2008 en december 2016 chirurgisch zijn behandeld voor een EFE aan de Faculteit Diergeneeskunde te Gent. Er zijn hierbij pre-, intra en postoperatieve data verzameld. Postoperatief bestond dit uit opvolging van de patiënten aan de hand van telefoongesprekken met de betreffende eigenaren. De nadruk van dit onderzoek lag op het verzamelen van de volgende gegevens: factoren die geassocieerd zijn met postoperatieve ileus (POI), re-laparotomie, ontslag uit de kliniek, koliek na ontslag en de overleving van de patiënt na het ontslag. Er zijn in totaal 145 operaties uitgevoerd op 142 paarden (3% recidieven). De patiënten werden voor 85% vertegenwoordigd door warmbloedpaarden en voor 62% door ruinen. Daarnaast werd in 60% van de gevallen bevestigd dat de patiënt een luchtzuiger of kribbebijter was. Bij 100% van de patiënten is een beknelling van links naar rechts vastgesteld en in 74% van de gedocumenteerde operaties was het ileum betrokken. In 6% van alle operaties ondervonden de chirurgen oncontroleerbare intra- operatieve bloedingen. 107 paarden (74%) hebben de operaties overleefd. Een hoge morbiditeit alsook een hoge mortaliteit, geassocieerd met EFE, is vastgesteld. 48% van de voor operatie aangeboden paarden konden succesvol worden ontslagen uit de kliniek, terwijl dit percentage voor de groep paarden die de operatie overleefden bij 65% lag. De mediaan van het overleven van de patiënten na ontslag uitgedrukt in aantal dagen, bedraagt 3193. Uit de statistische analyse van de 145 uitgevoerde operaties bleek dat darmresectie de kans op het ontwikkelen van POI vergroot, en dat de kans op POI bij een uitgevoerde jejuno-ileostomie hoger was dan bij een jejuno-jejunostomie. Ook kon worden geconcludeerd dat paarden waar POI is vastgesteld minder kans maakten om te worden ontslagen uit de kliniek en dat paarden, waarbij een darmresectie moest worden uitgevoerd, een kleinere overlevingskans hadden na ontslag.

2. Introduction

The omental [epiploic] foramen (Foramen omentale [epiploicum]) is a slit-like opening of about four to six cm between the abdominal cavity and the omental vestibule (vestibulum bursae omentalis) of the omental sac (Archer et al., 2004b). In human literature the omental foramen is called “Winslow foramen”, named after Jacques-Bénigne Winslow (1669-1769) who first described this inner structure (Bellary et al., 2012). Despite the preference for “omental foramen” in veterinary anatomical nomenclature, the term “epiploic foramen” is most commonly used amongst equine clinicians. Therefore, it was decided to use “epiploic foramen” throughout tis text. Strangulation of intestines through the epiploic foramen (epiploic foramen entrapment (EFE)) is a life- threatening condition in horses, necessitating surgical intervention (Wheat, 1975). It is found in about 1,7% to 6,9% of the surgical colic population (Kilcoyne et al., 2016; Steenhaut et al., 2001). Archer et al. (2004b, 2008a) identified an association between crib-biting/wind-sucking and EFE and mention this oral stereotypic behavior as an important factor increasing the risk to the development of EFE. Preoperative diagnosis of EFE can only be suggestive, based on ultrasonographic detection of distended small intestinal loops with a thickened wall at the level of the right cranial abdomen, whereas a definite diagnosis can only be made during surgery (Klohnen et al., 1996; Scharner et al., 2002; Freeman, 2002). Proudman et al. (2002a, 2002b, 2005) associated epiploic foramen entrapment with a high morbidity and mortality. In previous studies, short-term survival (survival to hospital discharge) of horses that successfully recovered after surgery for EFE, ranged between 77 and 95% (Freeman and Schaeffer, 2005; Archer et al., 2011). Ileal involvement consistently seems to occur in about two-thirds of the reported cases (Edwards, 1981; Vasey, 1988; Engelbert et al., 1993; Vachon and Fischer, 1995; Archer et al., 2004b; Freeman and Schaeffer, 2005; Kilcoyne et al., 2016) and in the vast majority of the cases, entrapment in a left-to- right fashion is described (Vachon and Fischer, 1995; Archer et al., 2004b; Steenhaut et al., 2004). In some case series, recurrence of EFE has been reported to be between 2 and 14% (Vachon and Fischer, 1995; Archer et al., 2004b; Freeman et al., 2014). Spontaneous closure of the epiploic foramen after surgical correction of EFE has recently been reported in 3/7 cases (43%) (van Bergen et al., 2016a) and could provide an explanation of the relatively low recurrence rates regarding the strong risk factors in the occurrence of EFE. Preventive laparoscopic closure of the epiploic foramen in the standing horse has been described in two recent studies (Munsterman et al., 2014; van Bergen et al., 2016b). The existing epidemiologic studies on survival after surgery for EFE were performed in the United Kingdom and North America and were mainly conducted in Thoroughbred dominated populations (Vachon and Fischer, 1995; Archer et al., 2004b, 2008a, 2011), whereas recent data from a larger European horse population regarding risk factors for EFE, survival rates, risk factors for postoperative complications and recurrence are lacking. Population specific and evidence-based information from a large patient group could allow surgeons to provide more accurate information to owners and referring veterinarians about likely prognosis following surgery to correct EFE. Therefore, the aim of this master thesis is to document and analyze perioperative variables, survival to and after discharge, and factors associated with survival of epiploic foramen entrapment in a retrospective study of a population of horses that underwent surgery for EFE in an equine colic referral institution on the European continent. In addition, a literature study will be done to summarize results that were obtained in earlier studies regarding EFE in horses.

3. Literature study

3.1. Anatomy of the omental foramen and the omental vestibule

To get an overview over the anatomy of the omental foramen and the vestibule of the omental bursa, it is important to know all the anatomical structures, which form the potential space. The lesser omentum is a serosal sheet between the lesser curvature of the stomach and the cranial part of the duodenum on the one hand, and the visceral surface of the liver on the other hand (Nickel et al., 1973). It consists of the hepatoduodenal and the hepatogastric ligament, and by this, helps forming the vestibule of the omental bursa as its ventral margin (Nickel et al. 1973; König and Liebich, 2007). The omental bursa has a vestibule, a dorsal, a caudal and also a splenic recess (Nickel et al. 1973). The Vestibulum bursae omentalis (omental vestibule) is defined as the “vestibule of the omental bursa, enclosed by Omentum minus, stomach and liver”, whereas the omental bursa (Bursa omentalis) is defined as the “lesser peritoneal sac, the potential space enclosed by the two omenta, the stomach and the liver, and it communicates with the greater peritoneal sac through the omental foramen” (Constantinescu and Schaller, 2012). Sisson (1975) describes the omental vestibule as the space between the hepatogastric ligament ventrally, the stomach on the left, the lesser omentum on the right and the gastropancreatic fold dorsally. Both, the vestibule as well as the epiploic foramen, are delineated by the plica gastropancreatica and the plica gastropencreaticoduodenalis (Freeman and Pearn, 2015). According to the Illustrated Veterinary Anatomical Nomenclature edited by Constantinescu and Schaller (2012), the Foramen omentale [epiploicum] (omental [epiploic] foramen) is defined as the “opening from the greater peritoneal sac to the Vestibulum bursae omentalis and it passes between the V. cava caudalis dorsally, the V. portae included into the hepatoduodenal ligament ventrally, the caudate lobe of the liver cranially, and the pancreas caudally”. It is a slit-like opening of about four to six centimeters located dorsally in the right cranial quadrant of the abdominal cavity (Archer et al., 2004b; Steenhaut et al., 2004). Nickel et al. (1973) define the borders of the epiploic foramen as the caudate process of the liver and the vena cava craniodorsally and the pancreas, the portal vein and the free border of the hepatoduodenal ligament caudoventrally. This stands in contrast to other publications where the hepatoduodenal ligament itself is thought to be the ventral boundary (Vasey, 1988; Archer et al., 2004b; Freeman and Pearn, 2015). A recent topographic anatomical study of the equine epiploic foramen defines the structures of the epiploic foramen as the hepatoduodenal and the hepatogastric ligament as well as the gastropancreatic and the hepatopancreatic folds (van Bergen et al., 2015).

3.2. Occurrence

Colic can be defined as “acute abdominal pain” (White et al., 1990) and as “the manifestation of abdominal pain” (Hines, 2018). It contains a wide range of underlying causes, all leading to a comparable clinical presentation (Proudman et al., 2002b). The true incidence of pathologies causing colic in the global equine population is unknown (White, 1990). Estimations of the occurrence of enteralgia are reported only for limited populations. These limitations are originated in geographical disparity on one side, on the other side the vast majority of epidemiological studies concerning colic are performed on data derived from academic institutions, and thus merely reflect a portion of the equine colic population in the respective region. Tinker et al. (1997) studied 31 horse farms in the states of Virginia and Maryland (USA) and obtained a crude incidence density rate of 10,6 colic cases in 100 horse-years in 1427 horses, with a variation from 0 to 30 colic cases/100 horse-years for individual farms and a median of 7 cases/100 horse-years. Another author estimated the annual national incidence of colic in the United States to be 4,2 colic cases per 100 horses per year (Traub-Dargatz et al., 2001).

Small intestinal pathology seems to occur in about 19 to 25 percent of all colic cases, whereas strangulating lesions of the small intestines are responsible for 12,6% up to 15% of acute abdominal disease in horses (White and Lessard, 1986; Steenhaut et al., 2000, 2001; van der Linden et al., 2003) Colic surgery is performed in 19% (Johnson and Keller, 2005) to 65% (Sutton et al., 2009) of the reported cases in veterinary teaching hospitals. Nevertheless, Traub-Dargatz et al. (2001) describe only 1,4% of the colic events resulting in surgery in their epidemiologic survey. Between 19,9% (Sutton et al., 2009) and 41% (Johnson and Keller, 2005) of all colic surgeries are found to be indicated due to small intestinal pathologies. Incarceration of the jejunum and/or the ileum (jejunoileum) through the omental foramen leads to intestinal strangulation obstruction (Vachon and Fischer, 1995), necessitating surgical intervention (Wheat, 1975). It is believed to be the underlying cause in about 0,7% to 4% of horses with colic (Steenhaut et al., 2001; Kilcoyne et al., 2016). Omental foramen entrapment represents about 1,7% to 6,9% of the surgical colic population (Steenhaut et al., 2001; Kilcoyne et al., 2016) and between 8,9% and 24% of small intestinal lesions in colic surgery (Archer et al., 2004b; Müller et al., 2009).

3.3. Risk factors

Crib-biting/wind-sucking behavior

Crib-biting/wind-sucking (cribbing) is an oral stereotypic behavior, where horses draw air into the cranial esophagus by contracting their sternohyoid muscles, sternothyroid muscles, thyrohyoid muscles and omohyoid muscles, whether or not griping a horizontal fixed object with their incisor teeth (Waters et al., 2002; Albright et al., 2009). In the general horse population, the occurrence of cribbing behavior is reported to be around 5,5% (Albright et al., 2009; Malamed et al., 2010). Cribbing is significantly associated with an increased risk of herniation of bowel through the omental foramen, as about 50% (47-54,4%) of the cases were reported to show this stable vice. Horses with a history of cribbing are about 35 times more likely to develop EFE compared to non-cribbing horses (Archer et al., 2004a, 2004b, 2008a, 2008b). During and after an episode of cribbing, the intra-abdominal pressure is shown to be increased (Albanese et al., 2013), which could cause expansion of the omental vestibule, thus allowing small intestine to enter into it (Freeman and Pearn, 2015).

Age

Senile atrophy of the right lobe of the liver was believed to enlarge the omental foramen in horses of at least seven years of age, thus leading to a higher prevalence of intestines herniating through the foramen as the horses age increases (Wheat, 1975). This supposition however cannot be supported by more recent literature. In their study on age distributions, Freeman and Schaeffer (2001) noted that the age of 17 horses with EFE ranged from one to eighteen years with a mean of 9,6 years, which was not significantly different from horses with miscellaneous small intestinal lesions. Similar age distributions were seen by several other authors (Engelbert et al., 1993; Vachon and Fischer, 1995; Steenhaut et al., 2004; Archer et al., 2008a) and no predominant age group was found in fifteen horses with EFE (Vasey, 1988). The youngest horse in Vachon and Fischer’s (1995) study of 53 horses with EFE was eleven months, and in one case, incarceration of jejunum in the omental foramen in a four-month- old foal is reported (Murray et al., 1994). Despite the majority of publications seem to suggest the absence of an increased risk of small intestinal herniation through the omental foramen with increasing age, Kilcoyne et al. (2016) saw a significantly lower proportion of horses younger than three years of age (2,8%) and a significantly greater proportion of horses > 12-20 years (45,8%) in comparison to the colic population (24% and 23% respectively). In another study there was no significance for increasing age in its linear form, while the squared term for age (years2) was significantly correlated to an accrued risk of EFE (Archer et al., 2008a). ©2018 Michaël Swagemakers 7

Breed

In comparison to a control colic population, Vachon and Fischer (1995) noted an overrepresentation of Thoroughbreds in the EFE group. A significantly greater proportion of Thoroughbreds in horses with EFE was also seen by Kilcoyne et al. (2016), as were the proportions of warmbloods and Morgans, whereas the fraction of Quarter Horses and Paint Horses with EFE in the same study was significantly lower compared to the respective proportions in the colic population. In three studies that were performed on the European mainland, warmblood horses were considerably more likely to suffer from EFE than other breeds (van den Boom and van der Velden, 2001; Steenhaut et al., 2001, 2004).

Sex

It is suggested that the likelihood of geldings being affected with EFE is four times higher than in mares (Vachon and Fischer, 1995). Steenhaut et al. (2001) compared the frequency distribution for sex of horses with EFE to the sex ratio in the general colic population and found a significantly higher proportion of geldings and a significantly lower proportion of mares in the EFE group. In a similar study in the USA, mares as well as stallions were significantly underrepresented in the EFE group compared to the colic population, while geldings were overrepresented (Kilcoyne et al., 2016). Vasey (1988) on the other hand did not find a predominant gender in his study on incarceration of small intestine by the omental foramen, and no significant differences in sex distribution were found between horses with EFE and a control population in another publication (Archer et al., 2004b).

Season

EFE is seen throughout every month of the year, but it seems that the prevalence increases during winter period, with the greatest numbers in January (Steenhaut et al., 2001, 2004; Archer et al., 2008b). In another study by Archer et al. (2004b), 77,5% of the EFE cases were reported between October and March, and season therefore was identified to be a significant risk factor in the occurrence of omental foramen herniations.

Others

In two studies, Archer et al. (2008a, 2008b) also mention an episode of colic during the past twelve months, increased stabling in the previous twenty-eight days and a greater height as risk factors for horses to develop EFE.

3.4. Pathophysiology

Involved intestines and direction of herniation

In only very few case reports concerning intestinal herniation through the omental foramen, other parts than the jejunum and/or ileum are described to get entrapped. To the authors knowledge, herniation of the large colon is reported in five cases (Foerner et al., 1993; Steenhaut, et al., 1993; Scherzer, 1998; Mariën, 1999; Segura et al., 1999), herniation of the apex of the caecum is documented in another resent case (Grzeskowiak, et al., 2017) and in one horse the duodenum was found to be entrapped in the omental foramen (Steenhaut et al., 2004). In the vast majority of EFE however, it’s the jejunum and/or ileum that becomes strangulated. Ileal involvement seems to occur in about two-thirds of the reported cases: 7/7 cases (100%) (Edwards, ©2018 Michaël Swagemakers 8

1981), 13/15 cases (86,7%) (Vasey, 1988), 12/19 cases (63,2%) (Engelbert et al., 1993), 37/52 cases (71,2%) (Vachon and Fischer, 1995), 47/71 cases (66,2%) (Archer et al., 2004b), 17/21 cases (81%) (Freeman et al., 2005) and in 37/73 cases (50,7%) (Kilcoyne et al., 2016). The jejunum on the other hand gets entrapped in about eighty percent of the cases: 5/7 cases (71%) (Edwards, 1981), 11/15 cases (73,3%) (Vasey, 1988), 46/52 cases (88,5%) (Vachon and Fischer, 1995), 54/71 cases (76,1%) (Archer et al., 2004b) and in 61/73 cases (83,6%) (Kilcoyne et al., 2016). Entrapment of a small intestinal loop through the omental foramen can occur in two directions: displacement of bowel from the right abdominal cavity through the omental foramen into the omental bursa is described as right-to-left entrapment (also: latero-medial incarceration, retrograde displacement), whereas a left-to-right entrapment (also: medio-lateral incarceration, normograde displacement) is used to express intestinal displacement in the opposite direction. In a left-to-right entrapment, the migrating intestines push the lesser sac through the omental vestibule and the omental foramen into the greater sac, mostly resulting in rupture of the inverted omental bursa due to increasing pressure as the intestinal loop becomes strangulated (Turner et al., 1984; Steenhaut et al., 2004). Turner et al. (1984) reported the intestines to be herniated from the greater into the lesser sac (right- to-left) in 12/15 EFE cases (80%), while an entrapment in a left-to-right direction was found in the resting 3/15 cases (20%). In another 15 reported cases in 1988, all of the patients (15/15, 100%) were found to have a right-to-left entrapment of small intestines through the omental foramen (Vasey, 1988). More recent and larger studies however indicate that the vast majority of intestinal herniations through the omental foramen happen to occur in a left-to-right direction: 53/53 (100%) (Vachon and Fischer, 1995), 69/71 (97,2%) (Archer et al., 2004b), 46/47 (97,9%) (Steenhaut et al., 2004). The length of the incarcerated segment can range from a few centimeters up to almost twenty meters (Engelbert et al., 1993).

3.5. Diagnosis

A definitive diagnosis of EFE, based on physical examination of the patient, is beyond the bounds of possibility and can only be made during explorative laparotomy. However, a tentative diagnosis of small intestinal strangulating obstruction can be obtained in most cases (Vasey, 1988).

Clinical examination

In most cases, inconclusive and confusing clinical symptoms are observed in horses suffering from EFE (Steenhaut et al., 2004; Freeman, 2012). Symptoms of moderate to severe pain (anxiety, sweating, getting up and down, rolling, pawing, kicking and/or biting at the abdomen) are often observed and the response to analgesic treatment is only temporary (Vasey, 1988; Mair and Smith, 2005; Archer et al., 2004b; Sanchez, 2018). In a later stage of the disease however, the demeanor may progress towards depression as the incarcerated segment undergoes necrosis (Engelbert et al., 1993; Mair and Smith, 2005; Sanchez, 2018). The average heart rate is reported to be increased (Vasey, 1988; Engelbert et al., 1993; Vachon and Fischer, 1995; Archer et al., 2004b; Steenhaut et al., 2004), as is the mean respiratory rate (Vasey, 1988; Steenhaut et al, 2004). Due to a compromised circulatory system, abnormal oral mucous membranes are found in the majority of the cases and the average capillary refill time is reported to be delayed (Vasey, 1988; Steenhaut et al., 2004). Decrease or absence of borborygmi can be perceived in the vast majority of patients with EFE during adominal auscultation (Vasey, 1988; Steenhaut et al., 2004; Kilcoyne et al., 2016). A mean rectal temperature of 37,5°C (Vasey, 1988; Steenhaut et al., 2004) and 37,3°C (Kilcoyne et al., 2016) was recorded and does not differ from the average temperature in healthy horses (37,2-38,3°C). ©2018 Michaël Swagemakers 9

Rectal examination

Palpation of the small intestine per rectum is not common in healthy horses, in case of a strangulating obstruction however, small intestinal distention may be palpable as a result of ileus with gas and/or fluid retention (Sanchez, 2018). According to Vachon and Fischer (1995), determination of dehydrated content in the large colon might also signal strangulating lesions of the small intestine. Due to the cranial location of the omental foramen in the abdominal cavity, distended small intestinal loops as a result of entrapment are not always palpable, especially in an early stage of the disease (Vachon and Fischer, 1995). In about 60 to 85 percent of horses with EFE, distended loops of small intestine can be palpated during rectal examination (Vasey, 1988; Engelbert et al., 1993; Vachon and Fischer, 1995; Archer et al., 2004b; Steenhaut et al., 2004).

Ultrasonography

Ultrasonography is a handy diagnostic tool in horses with colic. The main ultrasonographic changes in horses with strangulation colic, like EFE, are progressive distention of the small intestines, reduction in motility up to amotile intestines, thickening and edema of the intestinal wall and an increase of the amount of free fluid in the abdominal cavity (Freeman, 2002; Steenhaut et al., 2004). In patients with strangulation of the small intestines, transabdominal ultrasonography has a sensitivity, specificity and positive and negative predictive values of 100%, if loss of motility and abnormal intestines were detected (Klohnen et al., 1996). Small intestinal strangulation is highly associated with distension of the affected bowel (Steenhaut et al., 2004). The location of the distended loops can indicate the origin, and in case of EFE, distended loops can be found in the cranial abdomen. As in those cases, small intestinal distention in the cranial region of the abdomen cannot always be palpated per rectum (Sanchez, 2018), thus leading to an added value of ultrasonography in the diagnosis of EFE (Freeman, 2002). In contrast to non- strangulating obstructions of the small intestine, a greater loss of motility and greater increase of wall thickness may be seen in cases of strangulating obstruction (Vachon and Fischer, 1995; Freeman, 2002).

Nasogastric intubation

Evacuation of reflux (defined as ˃2 L of fluid obtained through a nasogastric tube (Archer et al., 2011)) from the stomach via a nasogastric tube potentially has two functions. In terms of diagnostic reasons, reflux indicates an obstruction of the small intestine (Vasey, 1988). About 50% of the horses with EFE are recorded to have had nasogastric reflux before surgery (Vasey, 1988; Engelbert et al., 1993; Vachon and Fischer, 1995; Steenhaut et al., 2004). On the other hand, nasogastric decompression may prevent the stomach from rupturing (Sanchez, 2018).

Abdominal paracentesis

Abdominal paracentesis is often used in the diagnosis of colic in horses, in particular when small intestinal strangulation is suspected. Collection of free fluid from the peritoneal cavity may help to determine whether surgery is needed, or medical treatment is adequate (Matthews et al., 2002). Nevertheless, the use of abdominal paracentesis is controversial, as different studies concerning the predictive values, sensitivity and specificity premise conflicting data (Freden et al., 1998; Matthews et al., 2002). Peritoneal fluid is collected aseptically via the midline at the deepest point of the abdomen with a sterile bovine teat cannula.

Normal peritoneal fluid is clear and pale yellow. In the study of Brownlow et al. (1981), peritoneal free fluid had a mean specific gravity of 1.010 and an average total protein of 7,7 g/liter. The average cell count was 4,33 ± 2,5 times 109/liter, consisting of neutrophils (45,2%), monocytes (47%) and lymphocytes (7,8%) (Brownlow et al., 1981). Abnormal peritoneal fluid can be turbid, serosanguineous and/or containing fibrine. The color can vary from normal yellow color to orange or red, up to green or brown. A total protein content of >20g/liter is considered abnormal (Garma-Aviña, 1998; Matthews et al., 2002). Measurement of lactate in the peritoneal fluid helps to estimate the severity of intestinal lesions. In case of a strangulating obstruction of the small intestine, the blood supply towards and from the affected bowel is compromised, thus leading to anaerobic metabolism of the cells and production of lactate. According to Latson et al. (2005) the mean lactate concentration in healthy horses was 0,60 mmol/L, in colic patients however, the peritoneal fluid lactate concentration was 4,00 mmol/L. In the same study, lactate levels in blood and peritoneal fluids are shown to be highly correlated (Latson et al., 2005).

Bloodwork

In the majority of EFE cases, a certain degree of dehydration is displayed on blood analysis. The average packed cell volume (PCV) is reported to be between 40 and 50% on presentation (Vasey, 1988; Vachon and Fischer, 1995; Archer et al., 2004b; Steenhaut et al., 2004). The outcome of blood gas analysis shows great variations, with a base excess (BE) (mEq/L) ranging from -17 in horses with metabolic acidosis to +11 in horses with respiratory alkalosis. In the majority of the cases however, the BE remained between the normal ranges (Engelbert et al., 1993; Steenhaut et al., 2004). Kilcoyne et al. (2016) reported a mean blood lactate concentration of 6,2mmol/L, which lies above the upper limit of 2mmol/L.

3.6. Treatment

Epiploic foramen entrapment always necessitates surgical intervention. In order to avert massive necrosis of the strangulated intestine, explorative laparotomy should be performed as early as possible (Wheat, 1975). After opening the abdominal cavity, the ileocecal fold should be identified and followed to the ileum. From the ileum on, the intestine can be followed in cranial direction unto the omental foramen. Reduction of the incarcerated bowel can be achieved by moderate traction in a horizontal direction on the caudal segment with one hand. Simultaneously, the incarcerated segment may carefully be pushed through the omental foramen with the other hand. In some cases, it is necessary to decompress the strangulated intestine by gently milking the content into the empty distal segment. Whenever possible, it can be helpful to enlarge the strangulated loop by hauling non-incarcerated bowel through the omental foramen to diminish distension of the strangulated segment (Vachon and Fischer, 1995; Freeman, 2012). The boundaries of the omental foramen, including large vascular structures like the portal vein (Vena porta), the caudal vena cava (Vena cava caudalis) and the hepatic artery (Arteria hepatica), are of fragile nature (Livesey et al., 1991). Rupture of these vessels results in uncontrollable hemorrhage and may either occur spontaneously or provoked by an endeavor to reduce the incarcerated bowel from the omental foramen by traction as well as an attempt to enlarge the foramen manually (Vasey, 1988; Livesey et al., 1991; Steenhaut et al., 2004). Decease as a consequence of intraoperative hemorrhage can be expected in about five percent of the surgical EFE cases: 1/15 (6,7%) (Vasey, 1988), 3/53 (5,7%) (Vachon and Fischer, 1995), 4/71 (5,6%) (Archer et al., 2004b), 3/90 (3,3%) (Steenhaut et al., 2004), 2/126 (1,6%) (Archer et al., 2011) and 7/58 (12,1%) (Kilcoyne et al., 2016). Due to the increased risk of fatal hemorrhage, manual enlargement of the omental foramen is contraindicated (Livesey et al., 1991). ©2018 Michaël Swagemakers 11

Once the incarceration is successfully reduced, the small intestines are to be exteriorized in order to evaluate the viability of the strangulated segment. To gain a proper viability assessment, the color of the serosa, improvement of the color after reduction of the entrapment, intestinal motility, edema and hemorrhage of the intestinal wall and the quality of mesenteric arterial pulses should be taken into account (Vachon and Fischer, 1995; Freeman, 2012). In case the affected segment is considered viable, resection is not recommended and evacuation of gas and fluid from the small intestines into the caecum should be sufficient (Proudman et al., 2002b). If the degree of intestinal damage however implies insufficient viability, resection of the affected part is required (Freeman, 2012). Resection is performed in about 53 to 90% of the surgical cases that are not euthanized during surgery, with a tendency towards a lower portion of resections performed in more recent publications (Edwards, 1981; Turner et al., 1984; Vasey, 1988; Steenhaut et al., 2004; Freeman and Schaeffer, 2005; Kilcoyne et al., 2016). The length of the resected segment is reported to vary between fifteen centimeters and thirteen meters and a mean of approximately four meters (Engelbert et al., 1993; Archer et al., 2004b; Steenhaut et al., 2004). Depending on the location and the length of the resected intestine, an end-to-end or a side-to-side jejunojejunostomy, an end-to-end jejunoileostomie or an end-to-side or side-to-side jejunocaecostomie should be performed (Freeman, 2012).

3.7. Prognosis

Survival rates in horses and factors influencing the likelihood of successful discharge after EFE surgery are investigated and documented in several studies. As time went by, short-term survival (defined as survival to hospital discharge after surgery) seems to improve. Case reports from the 1980’s demonstrate rather low survival rates: 3/7 cases (43%) (Edwards, 1981), 2/14 cases (14,3%) (Turner et al., 1984) and 3/11 cases (27,3%) (Vasey, 1988) survived to hospital discharge. Since the 1990’s, the chance of successful hospital discharge increased. Documentation of fifty-three surgical EFE cases between 1987 and 1993 by Vachon and Fischer (1995) demonstrate a survival to discharge of 66% (35/53). Archer et al. (2004b) reported a short-term survival rate of 69% (49/71) in seventy-one horses that were admitted to surgery. In thirteen horses (18,3%) surgery was not finished and of the remaining fifty-eight horses that did survive surgery, 84,5% (49/58) left the hospital alive. Another large retrospective study containing 100 horses suffering from EFE was performed by Steenhaut et al. (2004). Ten of the horses (10/100; 10%) were euthanized prior to surgery. Of the 90 horses that underwent surgery, eleven (12,2%) deceased or were euthanized during surgery and 79 (87,8%) successfully recovered from anesthesia. Forty-four horses were discharged from the hospital, yielding a short-term survival of 49% (44/90) and 55,7% (44/79) in horses that underwent surgery and in horses that survived surgery, respectively. Of the thirty-five horses that did not necessitate resection, 60% (21/35) survived to discharge, whereas 23/44 (52%) of the patients that underwent intestinal resection were able to leave the clinic alive (Steenhaut et al., 2004). Kilcoyne et al. (2016) documented a survival rate of 50% (29/58) in fifty-eight surgically treated horses with EFE. Similar results have been published by van den Boom et al. (2001) (37/75 cases; 49%) and Morton and Blikslager (2002) (7/14 cases; 50%).

4. Retrospective study

4.1. Materials & methods

Study design

Retrospective case series (n = 145).

Case selection

Medical records at the Department of Surgery and Anaesthesiology of Domestic Animals, Faculty of Veterinary Medicine, Ghent University, Belgium, were reviewed to identify horses that were subjected to surgery for incarceration of the small intestine in the epiploic foramen between April 2008 and December 2016. All horses that underwent EFE surgery during this period were included in this study, whether they survived the surgery or not.

Medical records review

General data that were collected included the horses breed (warmblood [WB], pony, others), gender (mare, stallion, gelding), age (months) and weight (kg). The season (winter, spring, summer, autumn) in which the horse was presented for surgery was also mentioned. Specific preoperative clinical data were deprived from the database at the Department of Internal Medicine and Clinical Biology of Large Animals, Ghent University, Belgium. They included heartrate (HR) (Beats Per Minute (BPM)), respiratory rate (RR) (Respirations Per Minute (RPM)), Packed Cell Volume (PCV) (%), base excess (BE) (mEq/L), blood lactate (mmol/L) and abdominal lactate (mmol/L). The duration of colic before surgery (h) and whether the horse was a windsucker/crib-biter were also reported. Surgical reports were screened to identify the surgeon involved, the duration of surgery (min) (from incision to final closure) and the duration of general anesthesia (min). The direction of small intestinal entrapment through the epiploic foramen, whether or not the ileum was involved and the estimated length of the entrapment (cm) were noted, as was the performance of small intestinal resection. In case of intestinal resection, the length of the resected bowel (cm) as well as the type of anastomosis performed were described. Furthermore, intraoperative euthanasia, fatal hemorrhage during reposition or decease for other reasons were recorded, and whether the horse survived the recovery from anesthesia or not. Postoperative events reported are the number of days of hospitalization, the occurrence of Post- Operative Ileus (POI) (defined as ˃2 L reflux obtained through a nasogastric tube on at least 1 occasion (Archer et al., 2011)) and the number of days with POI, re-laparotomy in case it was performed during the period of hospitalization, whether the horse survived until discharge from the hospital and, in case the horse was not discharged, the reason for decease or euthanasia.

Telephone interview

In view of long-term follow-up, the owners of horses that were successfully discharged from the hospital were contacted by telephone and asked if the horse was still alive, whether the horse had suffered from one or more significant colic episodes that necessitated medical intervention after discharge and, in case of mortality, if it was related to a colic episode. To avoid overestimation of recurrence, cases were only recorded as recurrent EFE cases if the diagnosis was made during surgery in our hospital. ©2018 Michaël Swagemakers 13

Statistical analysis

Data were collected with commercially available software (Excel 2016, Microsoft Corporation, Redmond, WA, USA) for descriptive analysis and subsequently analyzed with a statistical program (SPSS, version 23, IBM Corporation, Armonk, NY, USA) to explore potential associations between relevant selected variables and outcome. Significance for all tests was set at P≤0.05. The effects of selected independent variables on POR occurrence, on relaparotomy during the hospitalization period after surgery for EFE, on successful hospital discharge and on recurrence of colic after hospital discharge were assessed. First, univariable logistic regressions were performed to estimate the general effect of the independent variables on each dependent variable. For continuous independent variables, the linearity of the relationship between the independent variable and the log odds of the model was assessed by examining scatter plots and by performing a Box-Tidwell procedure. Independent variables yielding an effect with P≤0.20 were then considered for inclusion in a multivariable logistic regression model. First, potential correlations between the independent variables were assessed using the Pearson’s correlation coefficient, with values >0.6 indicating significant correlation. When correlation was present, only the variable with the most significant effect was used in the multivariable analysis. Variables with >30% missing data were excluded from analysis. The multivariable models were then further refined by stepwise elimination of variables with a significance of P>0.05, until only independent variables with a significance of P≤0.05 remained in the model. At each elimination step, the model fit (assessed by a Hosmer-Lemeshow test) and how the elimination influenced the effect of other independent variables were evaluated. When an elimination decreased the model fit or drastically changed the effect of the remaining independent variables, the eliminated variable was forced back in the model, even when the effect was not significant. The variables remaining in the final model were then considered to have a significant effect on the respective dependent variable. Survival times for all horses that underwent surgery, for recovered horses and for discharged horses were measured as a continuous variable starting at the moment horses left the recovery box alive and ending the day of confirmed death or censoring (horses still alive at the end of follow-up). Survival analysis was performed to estimate which of the selected independent variables significantly influenced colic-related mortality, counting from the day the horses left the recovery box alive. The effect of independent variables on survival was first tested using a univariable Cox-regression analysis. For continuous independent variables, the linearity of the relationship with survival time was evaluated by calculating the martingale residuals and plotting them against the independent variable. For categorical variables, the assumption of proportional hazards was evaluated by examining cumulative hazard plots. Variables yielding an effect with a significance of P≤0.20 were then considered for a multivariable Cox-regression analysis. First, potential collinearity was assessed by calculating the Pearson’s correlation coefficients, with values of >0.6 indicating significant correlation. If there was correlation, only the variable with the most significant effect was included in the multivariable model. The model was then further refined by stepwise elimination of the variable with the least significant effect, until only significant variables (P≤0.05) remained.

4.2. Results

Included horses

During the study period of 8.75 years, one hundred and forty-five surgeries were performed on 142 different horses suffering from small intestinal herniation through the epiploic foramen. Two horses had a second surgery for recurrent EFE nine months and two years, respectively, after the first surgical intervention, whereas a second herniation of small intestines through the omental foramen developed in one horse within 24 hours after the first surgery. 85,9% (122/142) of the patients were warmblood horses, 5,6% (8/142) were pony’s and the remaining 8,5% (12/142) were horses of different breeds.

There were 88 geldings (62%), 18 stallions (12,7%) and 36 mares (25,4%). The horses mean age was 119,6 months (9,96 years) ± 63,1 months (median: 99 months), ranging from 12 months up to 278,5 25% months (23,2 years). The weight varied between 239kg and 777kg 34% with a mean of 536,7kg ± 88,4kg (median: 535kg). Information about windsucking/crib-biting was acquired in 70 patients. In 42 (60%) of these 70 cases this behavioral vice was 18% confirmed. 23% Seasonal distribution of EFE colic is illustrated in Figure 1. Winter Spring Summer Autumn Preoperative data Figure 1: Seasonal distribution of 145 The mean duration of colic before surgery was 7,72 hours (median: EFE surgeries 6,83 hours) with a minimum of 30 minutes and a maximum of 29 hours. A summary of the specific continuous clinical variables is shown in table 1.

Table 1: Continuous clinical variables of the horses that underwent EFE surgery. Variable Number of horses Mean Median Range

Heartrate (beats/min) 143 62,7 60 24-128

Respiratory rate (respirations/min) 105 34,6 28 16-120 Packed Cell Volume (%) 144 38,9 38 26-62

Base excess (mEq/L) 144 2,0 2,8 -12,5-23,5

Blood lactate (mmol/L) 33 5,0 3,3 0,0-18,9

Peritoneal fluid lactate (mmol/L) 46 6,1 4,6 0,0-18,0

Surgical data

All EFE surgeries were performed by a total of 10 surgeons, each with at least 6 interventions. In all of the 145 cases (100%) the small intestines were entrapped though the omental foramen from left-to-right. The mean length of the herniated bowel was 334,3 cm (median: 200 cm) and the ileum was involved in 107 of the cases (74%). Thirty-seven horses (26%) were euthanized during surgery. In eight cases (6%) euthanasia was due to incontrollable intraoperative hemorrhage at the level of the omental foramen during or after intestinal reduction, and twenty-nine horses (20%) were euthanized because of technical restrictions or owners unwilling to continue with intestinal resection for financial or prognostic constraints. One additional horse was put to sleep during recovery from anesthesia because of severe femoral nerve paralysis, resulting in 107 (74%) horses that survived recovery. After reduction of the intestines from the omental foramen, small intestinal massage without resection was performed in 57 (53%) cases. A small-intestinal resection was performed in 51 surgeries (47%), with an average length of the resected segment of 428,8 cm (median: 400 cm). End-to-end jejunoileostomy was performed in 24 cases (47%), end-to-end jejunojejunostomy was performed in 21 cases (41%) and 6 cases underwent a side-to-side jejunocaecostomy (12%). The average duration of surgery and anesthesia of horses that were successfully sent to recovery from anesthesia was 125,3 minutes (median: 120 minutes) and 160,2 minutes (median: 150 minutes) respectively.

Immediate postoperative data

Of the 107 horses that had successfully recovered from anesthesia, forty-one (38,3%) developed POI. Eleven (19,6%) of the fifty-six horses that only had been treated with small intestinal massage developed POI, whereas POI occurred in 19/24 (79,2%), 8/20 (40%) and 3/6 (50%) of the horses that respectively had an end-to-end jejunoileostomy, an end-to-end jejunojejunostomy or a side-to-side jejunocaecostomy. In comparison to jejunojejunostomy, the occurrence of POI was significantly more prevalent in horses after jejunoileostomy (P=0.03, OR=5.70, 95%CI 1.12-28.96), which was not the case compared to jejunocaecostomy. A total of seventy horses survived to discharge, resulting in a short-term survival rate of 48,3% in function of the 145 horses that underwent surgery and 65,4% in function of the 107 horses that successfully recovered from anesthesia. The mean duration of maintenance in hospital of successfully discharged horses was 13,1 days (median: 11 days). Thirty-six patients did not survive until hospital discharge. In twenty-five cases (69,4%) death or euthanasia was the consequence of POI, seven horses (8,3%) died because of peritonitis and/or surgical site infection, two horses (5%) were put to sleep with severe laminitis, one patient (2,8%) had an eventration after re-laparotomy and infection of the abdominal wound, and in one horse (2,8%) pelvic trauma was the reason for euthanasia. Re-laparotomy was performed in 10 horses (7%) with POI and ongoing colic symptoms during the postoperative hospitalization period. In three cases (30%) an obstipation at the level of the anastomosis was found, which, in two cases was resolved by massage, and in one case by performing a side-to-side jejunocaecostomy bypass. Three horses (30%) that had massage only during the initial surgery suffered from severe necrosis of the formerly incarcerated segment yet necessitating a resection to be performed during re-laparotomy (one end-to-end jejunojejunostomy, one end-to-end jejunoileostomy and one side-to-side jejunocaecostomy). In two other cases (20%) problems were caused by adhesions which manually were broken down during re-laparotomy, in one horse (10%) the small intestines had passed through a high mesenteric defect at the level of the anastomosis, and one horse (10%) developed a recurrent herniation of the ileum through the omental foramen within 24 hours after the initial EFE surgery, requiring jejuno-ileal resection of 470 cm and an end-to-end jejunoileostomy. All horses (100%) successfully recovered from this second intervention but only two of them (20%) survived to discharge (1 with adhesions and 1 with a resection performed during relaparotomy). Detailed results for the univariable and multivariable effects of potential explanatory independent variables on POI occurrence, on relaparotomy during the hospitalization period after surgery for EFE and on non-survival to hospital discharge are summarized in Supplementary Information 1, 2 and 3. Multivariable logistic regression identified 4 variables (PCV, being a stallion, resection and end-to-end jejunojejunostomy) that were associated with the occurrence of POI. The risk of POI development increased with an increasing pre-operative PCV (mean ± s.d. PCV horses without POI 36 ± 6%, mean ± s.d. PCV horses with POI 41 ± 7%; OR 1.13, 95% CI 1.04-1.23, P=0.01). Stallions were significantly less likely to develop POI (OR 0.07, 95% CI 0.01-0.63, P=0.02). Horses that underwent intestinal resection were more likely to develop POI (OR 4.81, 95% CI 1.86-12.45, P˂0.01) and horses that had a jejunojejunostomy performed were significantly less likely to develop POI compared to horses with a jejunoileostomy performed (OR 0.13, 95% CI 0.03-0.62, P=0.01). Multivariable analysis for the effects on relaparotomy during the hospitalization period after surgery for EFE yielded no significant results. Multivariable logistic regression identified 3 variables (POI, pre-operative heart rate and pre-operative PCV) that were associated with reduced likelihood of survival to hospital discharge. Horses with POI were significantly less likely to be discharged from the hospital (OR 129.83, 95% CI 9.03-1866.70, P˂0.01). The risk of death before discharge from the hospital significantly increased with an increasing pre-operative heart rate (mean ± s.d. heart rate discharged horses 54 ± 15 BPM, mean ± s.d. heart rate not discharged horses 70 ± 21 BPM; OR 1.07, 95% CI 1.00-1.15, P=0.04) and significantly decreased with an increasing pre-operative PCV (mean ± s.d. PCV discharged horses 40 ± 8%, mean ± s.d. PCV not discharged horses 37 ± 7%; OR 0.80, 95% CI 0.68-0.95, P=0.01). ©2018 Michaël Swagemakers 16

Post-hospital discharge events

Owners from 63/70 discharged horses could be reached (90%). One owner confirmed that the horse was still alive but was unwilling to provide further information resulting in 62/70 horses (89%) being available for analysis of post-discharge events. The maximum recorded follow-up time was 3193 days (8.75 years). After discharge, 19 horses (30%) had at least 1 colic episode. Thirteen horses (21%) died after hospital discharge resulting in 50 horses alive at follow-up, which represents 36% of the horses that underwent surgery and 79% of the horses that were discharged from the hospital. Nine horses (13%) died because of colic, of which 5 horses (7%) had a specific diagnosis (3 intestinal strangulation due to intra- abdominal adhesions (3%) and 2 recurrent EFE (3%)), and 4 horses (6%) died without a specific colic diagnosis. Four horses (6%) died because of reasons unrelated to colic. Detailed results for the univariable and multivariable effects of potential explanatory independent variables on occurrence of colic after hospital discharge are shown in Supplementary Information 4. Multivariable analysis demonstrated a significant effect of the length of hospitalization on the likelihood of colic after hospital discharge. The risk for colic after hospital discharge increased significantly with an increasing hospitalization time (mean ± s.d. hospitalization time for horses without colic after hospital discharge 11 ± 7 days, mean ± s.d. hospitalization time for horses with colic after hospital discharge 15 ± 11 days; OR 1.08, 95% CI 1.01-1.15, P=0.03). Survival information was missing for 7 horses. Median survival time for all 138 EFE surgeries for which survival data were available, including horses that underwent intraoperative euthanasia and horses that were not discharged was 9 days. Median survival time of horses recovered from anesthesia was 1749 days. Median survival of the 63 horses that were discharged from the hospital and for which survival data were available was in excess of 3193 days. Cumulative survival times for all affected horses that underwent surgery, for affected horses that recovered from anesthesia and for affected horses discharged from the hospital are illustrated in Kaplan-Meier plots in respectively Figure 1, 2 and 3. Detailed results for univariable Cox-regression analyses between potential explanatory variables and cumulative probability of survival of discharged horses are shown in Supplementary Information 5. Multivariable Cox-regession identified 2 variables (pre-operative PCV and whether a resection was performed) that were associated with survival after discharge. The cumulative probability of survival was significantly higher in horses with a higher pre-operative PCV (mean ± s.d. PCV in deceased horses was 34±7%, mean ± s.d. PCV in horses alive was 38±7%; HR 0.85, 95% CI 0.73-0.99, P=0.04), and lower in horses in which a resection was performed (HR 5.28, 95% CI 1.13-24.67, P=0.03).

Epiploic Foramen Entrapment recurrences

Recurrence of EFE was confirmed during 4 EFE surgeries; the initial surgery of one of these horses was performed before start of data collection, so within the described timeframe, 3 horses underwent 2 EFE surgeries. This represents 3% of the recovered horses. The time between the first and second surgery for EFE in the 4 recurrent cases ranged from 1 day to 3 years, 10 months and 10 days with a median of 2 years, 3 months and 4 days. One of these 4 horses survived to discharge following the second surgery.

5. Discussion

This study is the first to report a large case series from a single, large hospital population in a population consisting predominantly of European Warmblood horses. The present study enables the findings to be compared with similar studies performed in other equine populations to determine if the findings are consistent with other populations, and to enable evidence-based advice to be applicable to a broader population of horses. Consistent with other studies, EFE was associated with high rates of mortality as 26% of the horses underwent intraoperative euthanasia and 35% of recovered horses died prior to hospital discharge. Horses that underwent resection, especially those undergoing jejunoileostomy, were more likely to develop POI. Horses that developed POI were less likely to be discharged from the hospital and horses that underwent a resection had reduced post-operative survival after hospital discharge compared to horses without resection. Furthermore, beside these moderate results, the condition recurred in at least 3% of the survivors. Cribbing behavior was exhibited by most horses (60%) which is higher than in previous reports. A discharge rate of 48% for all horses that underwent surgery, and of 65% for horses that successfully recovered from anesthesia in our population of mainly Warmblood horses was substantially lower compared to 66-69% (all horses that underwent surgery) and 79-85% (horses that were recovered) rates of hospital discharge that were previously reported in other populations (Vachon and Fischer, 1995; Archer et al., 2004b, 2011). Long-term survival is difficult to compare between studies due to differences in follow-up period and in definition of survival (Vachon and Fischer, 1995; Archer et al., 2004b, 2011). In our study the median survival of the 63 discharged horses from which data were available was in excess of 3193 days, which is substantially longer compared to the 700 days reported by Archer et al. (2004b). This difference cannot be explained by the longer follow-up in our study compared to the study of Archer et al. (2004b) as their median survival time (700 days) will not increase with a longer follow-up. A more plausible explanation could be the more radical elimination of complicated cases or those with a poor prognosis during and immediately after surgery in our study, resulting in a selected group of horses leaving the hospital carrying a better long-term prognosis. This selection is further illustrated by the median survival time of all EFE cases undergoing surgery, which was only 9 days in our study compared to 397 days in the prospective study of Archer et al. (2004b). This rather low median survival of the whole surgery population in our case series can primarily be attributed to the high number of horses that were euthanized during surgery (26%) compared to previous studies (10-13%) (Vachon and Fischer, 1995; Archer et al., 2011). The most common reason for euthanasia in our study (20%) was a poor or hopeless prognosis combined with financial restrictions and/or unwillingness of the owners to perform a resection. Making the decision of intraoperative euthanasia is a complex and multifactorial process, impeding comparison between intraoperative euthanasia percentages. Uncontrollable intraoperative hemorrhage leading to intraoperative euthanasia occurred in the remaining 6% of the euthanized horses which is in line with previously reported numbers ranging from 1.6 to 12% (Vasey, 1988; Livesey et al., 1991; Vachon and Fischer, 1995; Archer et al., 2004b, 2011). Intraoperative hemorrhage can occur due to rupture of the portal vein, caudal vena cava or hepatic artery (van Bergen et al., 2015) and can be caused by difficult repositioning, a small EF, edematous small intestines or a lack of surgical skills. While the effect of PCV on POI, hospital discharge and survival, of heart rate on hospital discharge, and of duration of hospitalization on colic after hospital discharge were significant, their clinical importance is questionable. Indeed, the corresponding odds ratios (logistic regression analysis) or hazard ratios (survival analysis) are very close to 1 (Supplementary Information 1 to 5). Analysis on the effect of POI on hospital discharge, of intestinal resection on POI and on survival, and of the type of anastomosis on POI generated odds ratios or hazard ratios that merit further attention, as these ratios deviated substantially from 1 (Supplementary Information 1 to 5). Horses with POI after surgery were less likely to be discharged from the hospital, which is in line with previous results on surgery for EFE (Archer et al., 2011). Similarly, the negative influence of performing an intestinal resection on the development of POI has been shown in other studies (Holcombe et al., 2009).

Ileal involvement in the majority of our EFE cases (74%) could possibly contribute to the observed high morbidity and mortality. Even though we were not able to demonstrate a direct significant effect, it is possible that ileal involvement influenced the prognosis of the group of horses needing intestinal resection. Indeed, performing an end-to-end jejunojejunostomy compared to an end-to-end jejunoileostomy significantly reduced the risk of POI development, and logically end-to-end jejunojejunostomy is only performed in cases without ileum involvement. These findings are in line with other researchers who also obtained better results without involvement of the ileum performing end-to-end jejunujejunostomy compared to side-to-side jejunocaecostomy when ileum was involved (Proudman et al., 2007). In case the ileum is involved and a resection is needed, either and end-to-end jejunoileostomy, or an side-to-side jejunocaecostomy can be performed. When the ileum is damaged until the base of the caecum, the surgeon is forced to the latter technique, but when a smaller portion is involved, the choice between techniques is up to the surgeon. Elimination of the ileocaecal valve by performing a side-to-side jejunocaecostomy is not always without consequences as demonstrated in a study where postoperative ileus (POI) was more likely to occur after side-to-side jejunocaecostomy than after end-to-end jejunojejunostomy (Proudman et al., 2007). In the comparison of side-to-side with end-to-end anastomotic techniques, the disparity in wall thickness between the jejunum and ileum should also be taken in account. Considerable cuff formation can result from an inverting suture pattern on the thick ileal wall, that can be responsible for problematic short-term outcomes with end- to-end jejunoileostomy (Proudman et al., 2007). More horses undergoing end-to-end jejunoileostomy underwent repeat laparotomy during the hospitalization period compared to horses with end-to-end jejunujejunostomy or side-to-side jejunocaecostomy (Stewart et al., 2014). However, there was no difference in short-term outcome between groups, but horses with a side-to-side jejunocaecostomy were more likely to have long-term complications with colic (Stewart et al., 2014). In another study where a large number of horses were treated by side-to-side jejunocaecostomy instead of end-to-end jejunoileostomy when the ileum was involved, high hospital discharge percentages were obtained suggesting that avoidance of end-to-end jejunoileostomy could improve results (Freeman and Schaeffer, 2005). In our study we were not able to demonstrate any benefit from side-to-side jejunocaecostomy compared to end-to-end jejunoileostomy in regard of POI development, but only 6 side-to-side jejunocaecostomies were performed. Final recommendations about the influence of the involved part of the intestine and the anastomosis technique on postsurgical complications and prognosis are difficult to provide based on literature review as another study could not demonstrate different outcomes between end-to-end jejunojejunostomy and end-to-end jejunoileostomy (Rendle et al., 2005). Stallions seemed to be less prone to POI development in the present case series, but the number of stallions was small, and the prognostic value of this finding should probably not be overestimated. In an earlier study reporting 53 EFE cases, 12/44 (27%) recovered horses underwent repeat laparotomy, mainly due to POI (Vachon and Fischer, 1995). Survival after relaparotomy was 50% (6/12) and a mechanical factor (adhesion, impaction, stricture) at the anastomotic side was the most common cause for relaparotomy (Vachon and Fischer, 1995). Another study conducted amongst horses that underwent surgery for all types of colic reports hospital discharge of 53% and 37% after relaparotomy due to persistent colic and POI, respectively (Findley et al., 2016). In our study relaparotomy was conducted in 10/145 (7%) surgeries and hospital discharge after relaparotomy was only achieved in 2/10 horses (20%). The majority of the non-discharged horses from the present case series were suffering from POI (70%). Specific causes for this POI were found and resolved in all horses that underwent relaparotomy. Nevertheless, only 2 out of 10 horses that underwent relaparotomy survived to discharge. The cause of POI most probably is multifactorial, with functional ileus as only one component, while many other factors related to the surgery probably contribute individually or more collectively to the problem (Freeman, 2017). The findings at relaparotomy in the present study illustrate that mechanical problems can contribute to POI in a large amount of cases, and that the horses with post-operative reflux should not be considered to simply have a primary physiological cause of small intestinal dysmotility (Salem et al., 2016). A more proactive approach with more frequent relaparotomy in horses suffering POI compared to the approach used in our study, might be beneficial. ©2018 Michaël Swagemakers 19

Recurrence of EFE was reported in 3% of the cases and was in line with previously reported recurrence rates ranging from 2% to 14% (Vachon and Fischer, 1995; Archer et al., 2004b; Freeman et al., 2014). Recurrence of EFE in the present study was probably underestimated due to the stringent classification of this event taking place (diagnosis made in our hospital). Four horses died because of colic without any specific diagnosis during the follow-up period of the study. Several colic related deaths amongst these 4 horses might be related to EFE recurrence. Spontaneous closure of the EF has been reported in 43% of horses undergoing laparoscopic inspection after surgery for EFE (van Bergen et al., 2016a). It has been suggested that horses that do not have spontaneous EF closure after surgery for EFE are likely to be at a substantially higher risk of developing EFE recurrence (van Bergen et al., 2016a). Standing right flank laparoscopy about 1 month after the initial colic surgery has been proposed (van Bergen et al., 2016a) and subsequent mesh closure can be performed if the EF is still open (van Bergen et al., 2016b). Alternatively, mesh closure could be performed during the initial surgery for EFE (van Bergen et al., 2018). These approaches could be of particular interest for windsucking/crib-biting horses who are known to be at increased risk of EFE (van Bergen et al., 2016a). The time of death after discharge was not directly recorded during this study, but was gathered from telephone interviews. As a consequence, the outcome variable “time to death” was biased by the best guess the owner could provide (recall bias). Other limitations of the study consisted of missing data in the clinical files, which is inherent to retrospective studies, and the EFE recurrence rate which might be underestimated due to the stringent classification of this event. The analysis on relaparotomy is underpowered, as there were only 10 horses that underwent relaparotomy. Retrospective analysis of a large dataset of mainly Warmblood horses on the European continent undergoing surgery for EFE demonstrated high morbidity and mortality, as 26% of the horses were euthanized during surgery and 35% of recovered horses died prior to hospital discharge. Horses that underwent resection were more likely to develop POI and horses undergoing jejunoileostomy were more prone to POI development compared to horses undergoing jejunojejunostomy. Horses that developed POI were less likely to be discharged from the hospital and horses that underwent a resection had reduced postoperative survival after hospital discharge compared to horses without resection. Postoperative reflux in a high percentage of cases could have contributed to the disappointing outcome after surgery. A combination of good intestinal anastomosis technique, more frequent relaparotomy and prevention of EFE recurrence might be beneficial to improve results presented in the present case series.

References

Albanese, V., Munsterman, A.S., DeGraves, F.J., Hanson, R.R., 2013. Evaluation of intra-abdominal pressure in horses that crib. Veterinary Surgery 42, 658-662. Albright, J.D., Mohammed, H.O., Heleski, C.R., Wickens, C.L., Houpt, K.A., 2009. Crib-biting in US horses: breed predispositions and owner perceptions of aetiology. Equine Veterinary Journal 41, 455-458. Archer, D.C., Freeman, D.E., Doyle, A.J., Proudman, C.J., Edwards, G.B., 2004a. Association between cribbing and entrapment of the small intestine in the epiploic foramen in horses: 68 cases (1991-2002). Journal of the American Veterinary Medical Association 224, 562-564. Archer, D.C., Proudman, C.J., Pinchbeck, G., Smith, J.E., French, N.P., Edwards, G.B., 2004b. Entrapment of the small intestine in the epiploic foramen in horses: a retrospective analysis of 71 cases recorded between 1991 and 2001. Veterinary Record 155, 793-797. Archer, D.C., Pinchbeck, G.K., French, N.P., Proudman, C.J., 2008a. Risk factors for epiploic foramen entrapment colic: an international study. Equine Veterinary Journal 40, 224-230. Archer, D.C., Pinchbeck, G.L., French, N.P., Proudman, C.J., 2008b. Risk factors for epiploic foramen entrapment colic in a UK horse population: a prospective case-control study. Equine Veterinary Journal 40, 405-410. Archer, D.C., Pinchbeck, G.L., Proudman, C.J., 2011. Factors associated with survival of epiploic foramen entrapment colic: a multicentre, international study. Equine Veterinary Journal 43, 56-62. Bellary, S.S., Walters, A., Gielecki, J., Shoja, M.M., Tubbs, R.S., Loukas, M., 2012. Jacob B. Winslow (1669-1760). Clinical Anatomy 547, 545-547. Brownlow, M.A., Hutchins, D.R., Johnston, K.G., 1981. Reference values for equine peritoneal fluid. Equine Veterinary Journal 13, 127-130. Constantinescu, G.M., Schaller, O., 2012. Splanchlogia, in: Illustrated Veterinary Anatomical Nomenclature. Enke Verlag, Stuttgart, Germany, pp. 140-233. Edwards, G.B., 1981. Obstruction of the ileum in the horse: A report of 27 clinical cases. Equine Veterinary Journal 13, 158-166. Engelbert, T.A., Tate, L.P., Bowman, K.F., Bristol, D.G., 1993. Incarceration of the small intestine in the epiploic foramen: report of 19 cases (1983-1992). Veterinary Surgery 22, 57-61. Findley, J.A., Salem, S., Burgess, R., Archer, D.C., 2016. Factors associated with survival of horses following relaparotomy. Equine Veterinary Journal 49, 448–453. Freden, G.O., Provost, P.J., Rand, W.M., 1998. Reliability of using results of abdominal fluid analysis to determine treatment and predict lesion type and outcome for horses with colic: 218 cases (1991- 1994). Journal of the American Veterinary Medical Association 213, 1012-1015. Freeman, D.E., Schaeffer, D.J., 2001. Age distributions of horses with strangulation of the small intestine by a lipoma or in the epiploic foramen: 46 cases (1994-2000). Journal of the American Veterinary Medical Association 219, 87-89. Freeman, D.E., Schaeffer, D.J., 2005. Short-term survival after surgery for epiploic foramen entrapment compared with other strangulating diseases of the equine small intestine in horses. Equine Veterinary Journal 37, 292-295.

Freeman, D.E., Schaeffer, D.J., Cleary, O.B., 2014. Long-term survival in horses with strangulating obstruction of the small intestine managed without resection. Equine Veterinary Journal 46, 711-717. Freeman, D.E., Pearn, A.R., 2015. Anatomy of the vestibule of the omental bursa and epiploic foramen in the horse. Equine Veterinary Journal 47, 83-90. Freeman, D.E., 2017. Post-operative reflux: a surgeon’s perspective. Equine Veterinary Education, 10.1111/eve.12735. Freeman, S., 2002. Ultrasonography of the equine abdomen: findings in the colic patient. In Practice 24, 262-273. Foerner, J.J., Ringle, M.J., Junkins, D.S., Fischer, A.T., MacHarg, M.A., Phillips, T.N., 1993. Transection of the pelvic flexure to reduce incarceration of the large colon through the epiploic foramen in a horse. Journal of the American Veterinary Medical Association 203, 1312-1313. Garma-Aviña, A., 2010. Cytology of 100 samples of abdominal fluid from 100 horses with abdominal disease. Equine Veterinary Journal 30, 435-344. Grzeskowiak, R.M., Barret, E.J., Rodgerson, D.H., 2017. Cecal entrapment within the epiploic foramen in a mare. Canadian Vetrinary Journal 58, 842-844. Hines, M.T., 2018. Clinical Approach to Commonly Encountered Problems, in: Reed, S.M., Bayly, W.M., Sellon, D.C. (Eds.), Equine Internal Medicine. Fourth edition. Elsevier, St. Louis, Missouri, USA, pp. 232- 310. Holcombe, S.J., Rodriguez, K.M., Haupt, J.L., Campbell, J.O., Chaney, K.P., Sparks, H.D., Hauptman, J.G., 2009. Prevalence of and risk factors for postoperative ileus after small intestinal surgery in two hundred and thirty-three horses. Veterinary Surgery 38, 368–372. Johnson, A., Keller, H., 2005. Ergebnisse von 1.431 stationär behandelten Kolikpatienten, unter besonderer Berücksichtigung der 285 chirurgisch versorgten Pferde, in den Jahren 1990 bis 1997 in der Klinik für Pferde, Allgemeine Chirurgie und Radiologie der Freien Universität Berlin. Pferdeheilkunde 21, 427-438 Kilcoyne, I., Dechant, J.E., Nieto, J.E., 2016. Comparison of clinical findings and short-term survival between horses with intestinal entrapment in the gastrosplenic ligament and horses with intestinal entrapment in the epiploic foramen. Journal of the American Veterinary Medical Association 249, 660- 667. Klohnen, A., Vachon, A.M., Fischer, A.T., 1996. Use of diagnostic ultrasonography in horses with signs of acute abdominal pain. Journal of the American Veterinary Medical Association 209, 1597-1601. König, H.E., Liebich, H.-G., 2007.Veterinary anatomy of domestic animals: textbook and colour atlas. Schattauer Verlag, Stuttgart, Germany. Latson, K.M., Nieto, J.E., Beldomenico, P.M., Snyder, J.R., 2010. Evaluation of peritoneal fluid lactate as a marker of intestinal ischaemia in equine colic. Equine Veterinary Journal 37, 342-346. Livesey, M.A., Little, C.B., Boyd, C., 1991. Fatal hemorrhage associated with incarceration of small intestine by the epiploic foramen in three horses. Canadian Veterinary Journal 32, 434-436. Mair, T.S., Smith, L.J., 2005. Survival and complication rates in 300 horses undergoing surgical treatment of colic. Part 1: short-term survival following a single laparotomy. Equine Veterinary Journal 37, 296-302.

Malamed, R., Berger, J., Bain, M.J., Kass, P., Spier, S.J., 2010. Retrospektive evaluation of crib-biting and windsucking behaviours and owner-perceived behavioural traits as risk factors for colic in horses. Equine Vetrinary Journal 42, 686-692. Mariën, T., 1999. Lateromedial incarceration of the large colon through the foramen omentale (epiploic foramen) in a horse. Vlaams Diergeneeskundig Tijdschrift 68, 188-190. Matthews, S., Dart, A.J., Reid, S.W.J., Dowling, B.A., Hodgson, D.R., 2008. Predictive values, sensitivity and specificity of abdominal fluid variables in determining the need for surgery in horses with an acute abdominal crisis. Australian Veterinary Journal 80, 132-136. Morton, A.J., Blikslager, A.T., 2002. Surgical and postoperative factors influencing short-term survival of horses following small intestinal resection: 92 cases (1994–2001). Equine Veterinary Journal 34, 450- 454. Müller, J.-M.V., Wehrli-Eser, M., Waldmeier, P., Rohn, K., Feige, K., 2009. Kurz- und Langzeitprognose von Dünndarmkolikern nach Kolikoperation. Tierärztliche Praxis Großtiere 37, 247-254. Munsterman, A.S., Hanson, R.R., Cattley, R.C., Barrett, E.J., Albanese, V., 2014. Surgical technique and short-term outcome for experimental laparoscopic closure of the epiploic foramen in 6 horses. Veterinary Surgery 43, 105-113. Murray, R.C., Gaughan, E.M., Debowes, R.M., Huston, L.C., Cooper, V.L., Welsh, T., 1994. Incarceration of the jejunum in the epiploic foramen of a four moth old foal. Cornell Veterinarian 84, 47-51. Nickel, R., Schummer, A., Seiferle, E., 1973. The Viscera of the Domestic Mammals. Verlag Paul Parey, Berlin-Hamburg, Germany. Proudman, C.J., Smith, J.E., Edwards, G.B., French, N.P., 2002a. Long-term survival of equine surgical colic cases. Part 1: Patterns of mortality and morbidity. Equine Veterinary Journal 34, 432-437. Proudman, C.J., Smith, J.E., Edwards, G.B., French, N.P., 2002b. Long-term survival of equine surgical colic cases. Part 2: Modelling postoperative survival. Equine Veterinary Journal 34, 438-443. Proudman, C.J., Edwards, G.B., Barnes, J., French, N.P., 2005. Factors affecting long-term survival of horses recovering from surgery of the small intestine. Equine Veterinary Journal 37, 360-365. Proudman, C.J., Edwards, G.B., Barnes, J., 2007. Differential survival in horses requiring end-to-end jejunojejunal anastomosis compared to those requiring side-to-side jejunocaecal anastomosis. Equine Veterinary Journal 39, 181–185. Rendle, D.I., Wood, J.L.N., Summerhays, G.E.S., Walmsley, J.P., Boswell, J.C., Phillips, T.J., 2005. End- to-end jejuno-ileal anastomosis following resection of strangulated small intestine in horses: a comparative study. Equine Veterinary Journal 37, 356–359. Salem, S.E., Proudman, C.J., Archer, D.C., 2016. Prevention of post operative complications following surgical treatment of equine colic: Current evidence. Equine Veterinary Journal 48, 143-151. Sanchez, C.L., 2018. Disorders of the Gastrointestinal System, in: Reed, S.M., Bayly, W.M., Sellon, D.C. (Eds.), Equine Internal Medicine. Fourth edition. Elsevier, St. Louis, Missouri, USA, pp. 709-842. Scharner, D., Rötting, A., Gerlach, K., Rasch, K., Freeman, D.E., 2002. Ultrasonography of the abdomen in the horse with colic. Clinical Techniques in Equine Practice 1, 118–124. Scherzer, S., 1998. Inkarzeration des Colon ascendens durch das Foramen omentale (epiloicum) bei einem Warmblutwallach. Wiener Tierärztliche Monatsschrift 85, 273-275

Segura, D., Garzón, N., Nomen, C., Prades, M., Valls, S., 1999. Entrapment of large colon through the epiploic foramen in a horse. Equine Vetrinary Education 11, 227-228. Sisson, S., 1975. Digestive system, in: Getty, R. (Ed.), Sisson and Grossman’s The Anatomy of the Domestic Animals. W.B. Saunders Company, London, UK, pp. 454-497. Steenhaut, M., Vandenreyt, I., Van Roy, M., 1993. Incarceration of the large colon through the epiploic foramen in a horse. Equine Veterinary Journal 25, 550-551. Steenhaut, M., Martens, A., Vlaminck, L., Gasthuys, F., Desmet, P., De Moor, A., Mariën, T., Deprez, P., 2000. Koliek bij het paard: een retrospectieve studie. Prevalentie, resultaten van conservatieve en operatieve behandeling. Vlaams Diergeneeskundig Tijdschrift 69, 24-30. Steenhaut, M., Martens, A., Vlaminck, L., Desmet, P., Gasthuys, F., De Moor, A., Mariën, T., Deprez, P., 2001. Een retrospectieve studie van chirurgisch behandelde specifieke koliekaandoeningen bij het paard. Vlaams Diergeneeskundig Tijdschrift 70, 389-404. Steenhaut, M., Martens, A., Vlaminck, L., Vertenten, G., 2004. Incarceratie van dunne darm doorheen het foramen omentale (epiploïcum): een retrospectieve studie bij 100 paarden. Vlaams Diergeneeskundig Tijdschrift 73, 17-30. Stewart, S., Southwood, L.L., Aceto, H.W., 2014. Comparison of short- and long-term complications and survival following jejunojejunostomy, jejunoileostomy and jejunocaecostomy in 112 horses: 2005- 2010. Equine Veterinary Journal 46, 333–338. Sutton, G.A., Ertzman-Ginsburg, R., Steinman, A., Milgram, J., 2009. Initial investigation of mortality rates and prognostic indicators in horses with colic in Israel: A retrospective study. Equine Veterinary Journal 41, 482-486. Tinker, M., White, N., Lessard, P., Thatcher, C., Pelzer, K., Davis, B. and Carmel, D., 1997. Prospective study of equine colic incidence and mortality. Equine Veterinary Journal 29, 448-453. Traub-Dargatz, J., Kopral, C., Hillberg Seitzinger, A., Garber, L., Forde, K., White, N., 2001. Estimate of the national incidence of operation-level risc factors for colic among horses in the United States, spring 1998 to spring 1999. Journal of the American Veterinary Medical Association 219, 67-71. Turner, T.A., Adams, S.B., White, N.A., 1984. Small intestine incarceration through the epiploic foramen of the horse. Journal of the American Veterinary Medical Association 184, 731-734 Vachon, A.M., Fischer, A.T., 1995. Small intestinal herniation through the epiploic foramen: 53 cases (1987-1993). Equine Veterinary Journal 27, 373-380. van Bergen, T., Doom, M., van den Broeck, W., Wiemer, P., Clegg, P.D., Cornillie, P. and Martens, A., 2015. A topographic anatomical study of the equine epiploic foramen and comparison with laparoscopic visualisation. Equine Veterinary Journal 47, 313-318. van Bergen, T., Wiemer, P., Schauvliege, S., Paulussen, E., Ugahary, F. and Martens, A., 2016a. Laparoscopic Evaluation of the Epiploic Foramen after Celiotomy for Epiploic Foramen Entrapment in the Horse. Veterinary Surgery 45, 596-601. van Bergen, T., Wiemer, P., Bosseler, L., Ugahary, F., 2016b. Development of a new laparoscopic Foramen Epiploicum Mesh Closure (FEMC) technique in 6 horses. Equine Veterinary Journal 48, 331- 337. van Bergen, T., Rötting, A., Wiemer, P., Schauvliege, S., Vanderperren, K., Ugahary, F., Martens, A., 2018. Foramen Epiploicum Mesh Closure (FEMC) through a ventral midline laparotomy. Equine Veterinary Journal 50, 235-240.

Van den Boom, R., van der Velden, M.A., 2001. Short- and long-term evaluation of surgical treatment of strangulating obstructions of the small intestine in horses: A review of 224 cases. Veterinary Quarterly 23, 109-115. van der Linden, M.A., Laffont, C.M., Sloet van Oldruitenborgh-Oosterbaan, M.M., 2003. Prognosis in equine medical and surgical colic. Journal of Veterinary Internal Medicine 17, 343-348. Vasey, J.R., 1988. Incarceration of the small intestine by the epiploic foramen in fifteen horses. Canadian Veterinary Journal 29, 378-382. Waters, A.J., Nicol, C.J., French, N.P., 2002. Factors influencing the development of stereotypic and redirected behaviours in young horses: findings of a four year prospective epidemiological study. Equine veterinary Journal 34, 572-579. Wheat, J.D., 1975. Causes of colic and types requiring surgical intervention. Journal of the South African Veterinary Association 46, 95-99. White, N.A., Lessard, P., 1986. Risk factors and clinical signs associated with cases of equine colic. Proceedings of the annual convention of the American Association of Equine Practitioners 32, 637-643. White, N.A., 1990. The Equine Acute Abdomen. Lea & Febiger, London, UK.

Supplementary information items

S1: Univariable and multivariable associations between potential explanatory variables and postoperative reflux (POR). Horses available for analysis: 107. The odds ratios (OR) represent the effect of the specific variable in relation to POR with an OR between 0 and 1 as protective factor and an OR ˃ 1 determining a risk-factor for POR.

Univariable analysis Multivariable analysis Variable Number Missing OR 95% CI P-value OR 95% CI P-value available data (%) Continuous variables Age (years) 107 0 1.001 0.10-1.01 0.73 Weight (kg) 107 0 1.002 0.10-1.01 0.48 Duration colic signs (min) 105 2 1.02 0.94-1.11 0.59 Heart rate (beats/min) 106 1 1.03 1.00-1.05 0.03*,§ Respiratory rate (respirations/min) 79 26 0.97 0.94-1.01 0.11§ Packed cell volume (%) 106 1 1.14 1.06-1.22 ˂0.01*,§ 1.13 1.04-1.23 0.01* Base excess (mEq/L) 106 1 0.99 0.91-1.08 0.78 Blood lactate (mmol/L) 18 83 Peritoneal fluid lactate (mmol/L) 32 70 ; Duration anaesthesia (min) 107 0 1.02 1.01-1.03 0.00*,§ Length involved intestine (cm) 101 6 1.00 1.00-1.00 0.14§ Length resection (cm) 50 0 1.00 1.00-1.00 0.28 Categorical variables Breed 107 0 Warmblood 91 Ref Others 16 0.88 0.31-2.74 0.37 Sex 107 0 Mare 25 1.51 0.60-3.80 0.38 1.21 0.40-3.65 0.74 Stallion 15 0.10 0.1-0.80 0.03*,§ 0.07 0.01-0.63 0.02* Gelding 67 Ref Ref Ileum involved 107 0 NO 25 Ref YES 82 0.96 0.39-2.40 0.93 Resection 107 0 NO 57 Ref Ref YES 50 5.63 2.40-13.19 ˂0.01*,§ 4.81 1.86-12.45 ˂0.01* Type anastomosis 50 0 End-to-end jejunoileostomy 24 Ref Ref End-to-end jejunojejunostomy 20 0.18 0.05-0.66 0.01*,§ 0.13 0.03-0.62 0.01* Side-to-side jejunocaecostomy 6 0.26 0.40-1.72 0.16§ 0.20 0.02-2.32 0.20 *: P value ≤ 0.05 is considered significant; §: P value ≤ 0.2 selected for the multivariable binary logistic regression model; Ref: referent

S2: Univariable and multivariable associations between potential explanatory variables and relaparotomy during the post epiploic foramen entrapment (EFE) surgery hospitalization period. Horses available for analysis: 107. The odds ratios (OR) represent the effect of the specific variable in relation to re-laparotomy with an OR between 0 and 1 as protective factor and an OR ˃ 1 determining a risk- factor for re-laparotomy during the post EFE surgery hospitalization period. This analysis is underpowered, as there were only 10 horses that underwent re-laparotomy.

Univariable analysis Multivariable analysis Variable Number Missing OR 95% CI P-value OR 95% CI P-value available data (%) Continuous variables Duration colic signs (min) 105 2 0.82 0.64-1.04 0.096§ 0.82 0.64-1.06 0.13 Heart rate (beats/min) 106 1 1.03 0.99-1.07 0.13§ 1.02 0.99-1.06 0.24 Respiratory rate (respirations/min) 79 26 1.04 0.99-1.09 0.14 Packed cell volume (%) 106 1 1.03 0.93-1.13 0.58 Base excess (mEq/L) 106 1 1.07 0.90-1.26 0.44 Blood lactate (mmol/L) 18 83 Peritoneal fluid lactate (mmol/L) 32 70 Duration anaesthesia (min) 107 0 1.01 1.00-1.02 0.17§ 1.01 0.99-1.02 0.35 Length involved intestine (cm) 101 6 1.00 1.00-1.00 0.85 Length resected intestine (cm) 39 24 1.00 0.99-1.01 0.85 Post-operative reflux (days) 107 0 1.09 0.86-1.38 0.47 Categorical variables Resection 107 0 NO 56 Ref YES 51 1.80 0.48-6.81 0.38 Type anastomosis 50 0 End-to-end jejunoileostomy 24 Ref End-to-end jejunojejunostomy 20 0.26 0.027-2.57 0.25 Side-to-side jejunocaecostomy 6 1.00 0.091-11.028 1.00 Post-operative reflux 107 0 NO 66 Ref YES 41 2.54 0.67-9.62 0.17§ §: P value ≤ 0.2 selected for the multivariable binary logistic regression model; Ref: referent

S3: Univariable and multivariable associations between potential explanatory variables and non- survival to hospital discharge. Horses available for analysis: 145.The odds ratios (OR) represent the effect of the specific variable in relation to hospital discharge with an OR between 0 and 1 as protective factor and an OR ˃ 1 determining a risk-factor for death during hospitalization.

Univariable analysis Multivariable analysis Variable Number Missing OR 95% CI P-value OR 95% CI P-value available data (%) Continuous variables Age (years) 145 0 1.00 0.10-1.01 0.65 Weight (kg) 145 0 1.00 0.10-1.01 0.21 Duration colic signs (min) 143 1 0.97 0.89-1.04 0.37 Heart rate (beats/min) 141 3 1.05 1.03-1.08 ˂0.01*,§ 1.07 1.00-1.15 0.04* Respiratory rate (respirations/min) 105 28 1.00 0.98-1.03 0.77 Packed cell volume (%) 144 1 1.06 1.01-1.11 0.02§ 0.80 0.68-0.95 0.01* Base excess (mEq/L) 144 1 0.99 0.92-1.06 0.68 Blood lactate (mmol/L) 33 77 Peritoneal fluid lactate (mmol/L) 46 68 Duration anaesthesia (min) 145 0 0.10 0.99-1.00 0.30 Length involved intestine (cm) 141 3 1.00 1.00-1.00 ˂0.01*,§ Length resection (cm) 51 0 1.00 0.10-1.00 0.80 Categorical variables Recurrence EFE 145 0 NO 141 0 Ref YES 4 0 2.86 0.29-28.31 0.37 Ileum involved 141 3 NO 34 Ref YES 107 1.19 0.55-2.58 0.66 Resection 108 0 NO 57 Ref YES 51 3.25 1.42-7.44 ˂0.01*,§ Type anastomosis 51 0 End-to-end jejunoileostomy 24 Ref Ref End-to-end jejunojejunostomy 21 0.30 0.09-1.02 0.06§ 0.13 0.00-4.78 0.26 Side-to-side jejunocaecostomy 6 0.60 0.10-3.63 0.58 0.08 0.00-3.36 0.19 Post-operative reflux 107 0 NO 66 Ref Ref YES 41 20.71 7.39-58.09 ˂0.01*,§ 129.83 9.03-1866.70 ˂0.01* Surgeon, nr cases 145 0 Surgeon 1, 16 0.38 0.08-1.69 0.20 Surgeon 2, 19 0.84 0.20-3.63 0.84 Surgeon 3 , 6 0.83 0.16-9.54 0.83 Surgeon 4¥, 6 - - - Surgeon 5, 8 1.04 0.17-6.40 0.97 Surgeon 6, 17 0.56 0.13-2.41 0.43 Surgeon 7, 32 0.63 0.17-2.33 0.48 Surgeon 8, 13 0.73 0.15-3.47 0.69 Surgeon 9, 15 0.23 0.05-1.12 0.07 Surgeon 10, 14 Ref

*: P value ≤ 0.05 is considered significant; §: P value ≤ 0.2 selected for the multivariable binary logistic regression model; ¥: Surgeon 4 has been removed from the analysis because all horses were euthanized during anaesthesia for reasons unrelated to the surgeon; Ref: referent

S4: Univariable and multivariable associations between potential explanatory variables and occurrence of colic after hospital discharge. Horses available for analysis: 70. The odds ratios (OR) represent the effect of the specific variable in relation to the occurrence of colic after hospital discharge with an OR between 0 and 1 as protective factor and an OR ˃ 1 determining a risk-factor for occurrence of colic after hospital discharge.

Univariable analysis Multivariable analysis Variable Number Missing OR 95% CI P-value OR 95% CI P-value available data (%) Continuous variables Age (years) 62 11 1.00 1.00-1.01 0.49 Weight (kg) 62 11 1.00 1.00-1.01 0.41 Duration colic signs (min) 61 13 1.05 0.95-1.16 0.37 Heart rate (beats/min) 62 11 1.02 0.98-1.06 0.39 Respiratory rate (respirations/min) 44 37 Packed cell volume (%) 62 11 0.96 0.88-1.05 0.41 Base excess (mEq/L) 62 11 1.08 0.94-1.24 0.31 Blood lactate (mmol/L) 11 84 Peritoneal fluid lactate (mmol/L) 18 74 Length involved intestine (cm) 58 17 1.00 1.00-1.00 0.68 Duration anaesthesia (min) 62 11 1.01 0.99-1.02 0.40 Duration hospitalization (days) 62 11 1.05 0.98-1.11 0.15§ 1.08 1.01-1.15 0.03* Categorical variables Breed 62 11 Warmblood 52 Ref Others 10 1.06 0.24-4.64 0.94 Sex 62 11 Mare 12 0.18 0.021-1.58 0.12§ Stallion 14 1.11 0.31-4.05 0.87 Gelding 36 Ref Ileum involved 62 11 NO 15 Ref YES 47 0.77 0.22-2.67 0.67 Resection performed 62 11 NO 40 Ref YES 22 1.71 0.56-5.28 0.35 Post-operative reflux 62 11 NO 52 Ref YES 10 0.56 0.11-2.95 0.50 Windsucking/crib-biting 54 23 NO 29 Ref YES 25 1.49 0.43-5.22 0.53 *: P value ≤ 0.05 is considered significant; §: P value ≤ 0.2 selected for the multivariable binary logistic regression model; Ref: referent

S5: Univariable and multivariable associations between potential explanatory variables and cumulative probability of survival of horses discharged from the hospital (univariable and multivariable Cox regression analysis). Horses available for analysis: 70. The hazard ratios (HR) represent the effects of the specific variables in relation to the cumulative probability of survival after hospital discharge with a HR between 0 and 1 as protective factor and a HR ˃ 1 determining a risk-factor for the cumulative probability of survival after hospital discharge. Censored horses: non-colic related death and horses still alive at the end of their specific follow-up period.

Univariable analysis Multivariable analysis Variable Number Missing Number Hazard 95% CI P-value Hazard 95% CI P-value available data (%) censored (%) Ratio Ratio Continuous variables Age (months) 63 10% 54 (85.7%) 1.00 0.99-1.01 0.68 Weight (kg) 63 10% 54 (85.7%) 1.00 0.99-1.00 0.46 Duration colic signs (min) 62 11% 53 (85.5%) 0.90 0.72-1.12 0.35 Heart rate (BPM) 63 10% 54 (85.7%) 1.03 0.99-1.08 0.18§ Respiratory rate 42 40% 34 (81.0%) Packed cell volume (%) 63 10% 54 (85.7%) 0.91 0.81-1.03 0.13§ 0.85 0.73-0.99 0.04* Base excess (mEq/L) 63 10% 54 (85.7%) 1.06 0.90-1.25 0.49 Blood lactate (mmol/L) 7 90% Peritoneal fluid lactate (mmol/L) 19 73% Duration anaesthesia (min) 63 10% 54 (85.7%) 1.00 0.99-1.01 0.66 Length involved intestine (cm) 59 16% 50 (84.9%) 1.00 0.10-1.00 0.85 Length resection (cm) 23 0% 18 (78.3%) 1.00 1.00-1.00 0.77 Categorical variables Breed 63 10% 54 (85.7%) Warmblood 53 Ref Other breeds 10 1.34 0.27-6.56 0.72 Sex 63 10% 54 (85.7%) Mare 13 0.50 0.06-4.12 0.52 Stallion 14 0.43 0.05-3.53 0.43 Gelding 36 Ref Ileum involved 63 10% 54 (85.7%) NO 15 Ref YES 48 0.61 0.12-3.10 0.55 Resection 63 10% 54 (85.7%) NO 40 Ref Ref YES 23 2.65 0.70-10.06 0.15§ 5.28 1.13-24.67 0.03* Type anastomosis 23 0% 18 (78.3%) End-to-end jejunoileostomy 9 Ref End-to-end jejunojejunostomy 12 0.69 0.10-5.02 0.72 Side-to-side jejunocaecostomy 2 1.47 0.12-17.39 0.76 *: P value ≤ 0.05 is considered significant; §: P value ≤ 0.2 selected for the multivariable Cox- regression model, Ref: referent