Prevalence of gastric and duodenal ulceration in 691 nonsurviving foals (1995–2006)


  • Portions of this work were presented in abstract form at the Dorothy Russell Havemeyer Foundation Equine Neonatal Septicemia Workshop V, Salem, MA, 2008, and the ACVIM Veterinary Medical Forum, 2010.


Reason for performing study: Gastric ulcer disease is reported to be a significant cause of morbidity in foals, but the prevalence of ulcers in this population has not recently been evaluated.

Objectives: To determine the prevalence of gastric and duodenal ulceration in nonsurviving foals, and the association of ulceration with the body system of primary diagnosis. Secondary objectives were to evaluate a potential association between age and ulcer prevalence and to evaluate the use of antacid medication in the neonatal hospital population during the study years.

Methods: Necropsy records were searched for all equine accessions from 1995 to 2006. Foals aged from one day to 6 months were included. Year, age, breed, sex, diagnosis and the presence of glandular, nonglandular and/or duodenal ulceration were recorded. Diagnoses were divided into groups based on the body system of primary diagnosis, with multiple diagnoses possible. A computerised database was searched for antacid treatment of all neonatal admissions.

Results: The overall prevalence of ulcers was 22%, with nonglandular ulcers predominating. Ulceration was significantly associated with gastrointestinal disease. There was no significant change in ulcer prevalence over time, although there was a significant decrease in the use of antacid medications in the later study years. Neonatal foals (<1 month) had a lower ulcer prevalence than older foals.

Conclusions: The prevalence of ulcers in foals, although low, has remained stable over time. Gastric or duodenal ulcers are associated with a primary diagnosis of gastrointestinal disease and are less prevalent in neonates.

Potential relevance: The prevalence of ulcers in nonsurviving foals is low. Foals with gastrointestinal disease are more likely to have ulcers than foals with other primary diagnoses, and older foals are more likely to have ulceration than neonates. The prevalence of ulceration did not appear to be related to hospital-wide antacid medication use in neonates.


Gastroduodenal ulcer disease (GDUD) is a significant cause of morbidity and mortality in late suckling and early weanling foals (Becht and Byars 1986), with a prevalence of 25% at necropsy (Wilson 1985) and 51% with gastroscopy (Murray 1989; Murray et al. 1990a, b). In mature horses lesions are located primarily in the nonglandular and/or glandular mucosa, with very infrequent duodenal involvement. In weanling age foals, lesions are frequently located in the pylorus and/or duodenum and may create a gastric outflow obstruction. In those cases, ulceration of the gastric squamous epithelium and, in many cases, the oesophagus occurs secondary to reflux acid exposure (Murray 1991; Andrews et al. 1999). Clinical signs of ulceration in foals, including ill thrift, inappetence, ptyalism, bruxism, colic and reflux, are related to the portion of the proximal gastrointestinal tract affected and result primarily from acid exposure (Becht and Byars 1986; Nappert et al. 1989; Murray 1999). Foals with documented outflow obstruction can have a favourable outcome following surgical correction (Coleman et al. 2009; Zedler et al. 2009). The disease may develop in otherwise healthy foals but has been associated with systemic diseases or stresses, including diarrhoea, musculoskeletal disease, weaning or shipment (Rebhun et al. 1982; Gross and Mayhew 1983; Becht and Byars 1986; Murray 1989).

Gastric ulceration is reportedly common in sick neonatal foals and can be associated with gastric rupture (Nappert et al. 1989). Although the ontogeny of gastric acid secretion in neonatal foals is not completely known, term neonatal foals are capable of producing gastric acid and maintain an acidic gastric pH between feedings by age 2 days (Sanchez et al. 1998). However, the intragastric pH in critically ill neonates is more variable and often alkaline (Sanchez et al. 2001). Therefore, the pathogenesis of gastric ulcers in this population remains unclear, but may be related to alterations in mucosal perfusion or other mechanisms of mucosal defence (Ryan and Sanchez 2005; Sanchez et al. 2008a). Although gastric ulcer prophylaxis is common in many referral hospitals, the prevalence of gastric ulcers in nonsurviving neonates was unaffected by ulcer prophylaxis and in one post mortem study no foals had died as a result of gastric ulceration in one post mortem study (Barr et al. 2000).

The objectives of the study reported here were to determine the prevalence of gastric and duodenal ulceration in nonsurviving neonatal and weanling-aged foals and to determine the association, if any, of ulceration with the body system of primary diagnosis, age or sex. A secondary objective was to determine any changes over time in either ulcer prevalence or the use of antacid medications over time in the overall hospital population. Our hypotheses were that gastric and/or duodenal ulceration would be associated with the primary diagnoses of gastrointestinal and musculoskeletal disease and that the prevalence of ulceration would not change over time.


All equine necropsy reports from 1995 to 2006 from the University of Florida Veterinary Medical Center were reviewed. Foals aged <6 months were included. Aborted fetuses were excluded regardless of gestational age. Neonates were defined as foals aged <1 month and weanlings as those aged 1–6 months. Year, age, breed, sex, diagnosis, and the presence of glandular, nonglandular and/or duodenal ulceration were recorded. Diagnoses were divided into the following groups based on body system of primary post mortem diagnosis: sepsis, prematurity, gastrointestinal, musculoskeletal and other (including respiratory, neurological, urogenital and cardiac). The disease categories used for analysis were those that occurred most commonly in the study population. The foal was classified as having multiple diagnoses if a diagnosis was given in more than one body system. Prematurity and sepsis were valid diagnoses for neonatal foals only. Prematurity was defined as known gestational age <320 days, physical characteristics consistent with prematurity such as small body stature, short silky hair coat, domed forehead and curled ears, or as stated by the submitting veterinarian. Sepsis was defined as a positive ante mortem blood culture reported by submitting veterinarian, or the presence of pathological features of sepsis. The location of ulceration (glandular, nonglandular or duodenal) was recorded. Unless gastric and/or duodenal ulceration was the only lesion identified, it was not included as a gastrointestinal lesion.

A computerised database of all neonatal foal admissions (not limited to nonsurvivors) was searched for use of antacid therapy in foals <30 days during the study years. Antacids were omeprazole, ranitidine or cimetidine, with no distinction made between types of antacid medication. Foals were grouped by year of presentation (1995–2000 or 2001–2006) to determine temporal changes, if present in the prevalence of ulceration and disease category in the study population, or hospital-wide use of antacid medications.

The data were analysed for an association between ulceration (regardless of location within the proximal gastrointestinal tract) and sex or disease category, and for an association between age (neonates [age <1 month] vs. weanlings [age ≥1 month] and ulcer location (nonglandular, glandular or duodenal) by Chi-square analysis. When significant associations were present, post hoc Chi-square analysis was performed using Bonferonni correction. Hospital-wide antacid use and overall prevalence of ulceration (regardless of ulcer location or foal's age) was compared between the years 1995–2000 and 2001–2006 by Chi-square analysis. All analyses were performed using SigmaPlot for Windows version 11.0a. Significance was set at P<0.05.


There were 691 foals included in the study, 354 neonates and 337 weanlings. There were 319 fillies and 367 colts. Sex was not recorded for 5 foals. There was no significant difference in the prevalence of ulceration between colts and fillies (P = 0.805). There were 429 Thoroughbreds, 79 Quarter Horses, 32 Arabians, 28 Paint horses, 20 Paso Finos, 17 Appaloosas, 15 Standardbreds, 13 Warmbloods and fewer than 10 of each of the following breeds: miniature horse, pony, Tennessee Walking Horse, Morgan, Percheron, Shire, Clydesdale, Andalusian and Friesian. The breed distribution approximately mirrored that of the hospital population during those years.

Data for ulcer location and foal age are presented in Table 1. There was evidence of ulceration in 155 foals (22%), regardless of site. Ulcers were most common within the nonglandular mucosa (70), followed by glandular mucosa (25), nonglandular and glandular mucosae (25), and nonglandular and duodenal mucosae (20). Neonatal foals were less likely to have ulceration at any site or nonglandular ulceration, relative to weanlings (P<0.001). The prevalence of glandular (P = 0.066) or duodenal ulceration (P = 0.483) did not differ significantly between the age categories. Only one neonatal foal in the study had a perforated gastric ulcer (along the margo plicatus).

Table 1. Ulcer distribution by location and age in nonsurviving foals between 1995 and 2006
 <1 week1–2 weeks2–4 weeks1 month2 months3 months4 months5 months6 monthsAll ages
  1. Ulceration was grouped by location at necropsy (S = squamous; G = glandular; D = duodenal; SG = squamous and glandular; SD = squamous and duodenal; SGD = squamous, glandular and duodenal). Ulceration was most common within the nonglandular mucosa and least common in duodenal mucosa.

Total ulcers (%)31 (12)12 (26)11 (20)25 (36)22 (22)25 (41)13 (33)12 (32)4 (14)155 (22)
Total foals25247556910161393829691

There were 230 (65%) neonatal foals with a diagnosis of sepsis and 47 (13%) with prematurity. Most neonatal foals (142; 62%) had multiple diagnoses. Sixty-four (18%) neonatal foals were diagnosed with gastrointestinal disease and 16 had evidence of ulceration. The most common (122; 36%) primary diagnosis for weanlings was in a body system classified as other. Gastrointestinal disease was the primary diagnosis for 103 weanlings, of which 51 (50%) had evidence of ulceration. For all foals, regardless of age, gastrointestinal disease was significantly associated with the presence of ulceration (P<0.001; Fig 1). There were no significant associations between the presence of ulceration and any other classification of disease.

Figure 1.

Association of body system of primary diagnosis with the presence of ulcers at necropsy in nonsurviving foals. Foals (n = 691) were grouped by body system of primary diagnosis at necropsy (GI = gastrointestinal; MS = musculoskeletal; Mult = more than one body system; Other = including respiratory, neurological, urogenital and cardiac). Ulcers were in nonglandular, glandular and duodenal mucosae. Asterisk (*) indicates a significant (P<0.05) association between gastrointestinal disease and ulceration regardless of site.

Medical record data were available for 1103 neonatal foals. There were significantly (P<0.001) fewer neonates treated with antacid medications from 2001–2006 (228/620) relative to 1995–2000 (325/483; Fig 2). There was no significant association between year (1995–2000 vs. 2001–2006) and ulceration when evaluating neonates (0.090) or all foals (P = 0.169). There was no significant association between time period and overall disease category in either all foals (P = 0.480) or neonates (P = 0.239) or the prevalence of sepsis in neonates (P = 0.469); the prevalence of prematurity was higher in neonates in the later years (2001–2006) relative to the early years (1995–2000; P = 0.037).

Figure 2.

Prevalence of ulceration in nonsurviving neonatal foals (n = 354) and antacid treatment by 5 year period (1995–2000 and 2001–2006) in neonatal foals (n = 1103). No distinction was made between types of antacid treatment (cimetidine, ranitidine or omeprazole). Asterisk (*) indicates a significant (P<0.05) decrease in antacid treatment in the 5 year period 2001–2006 compared to 1995–2000.


Nonsurviving foals in this study had a 22% prevalence of gastric and/or duodenal ulceration diagnosed at post mortem examination. This prevalence is similar to that reported in a post mortem study (25.2%) from 1980 to 1984 (Wilson 1985). In the 3 decades since post mortem prevalence was estimated, there have been advances in the understanding of gastric acidity in neonatal foals (Baker and Gerring 1993; Sanchez et al. 1998, 2001), the efficacy of antacid medications (Sanchez et al. 1998, 2001, 2004; Ryan et al. 2005; Javsicas and Sanchez 2008), and the treatment of critically ill foals (Furr et al. 1997; Corley et al. 2005, 2008b). Despite improvements in diagnostic capabilities and therapeutic interventions, the prevalence of ulceration in nonsurviving foals has remained stable over time. Further experimental studies and prospective clinical trials are needed to determine the pathogenesis and true clinical relevance of gastric ulceration in foals.

The prevalence of ulcers identified in these post mortem studies is lower than the 51% prevalence of lesions identified gastroscopically in clinically normal foals in England and Ireland (Murray et al. 1990a, b). Desquamation of the nonglandular epithelium was considered a lesion via gastroscopy in some reports (Murray 1989; Murray and Mahaffey 1993) but not in the current study. This is probably a normal phenomenon in foals (Merritt 2003) and may partially account for the higher prevalence of lesions at gastroscopy, despite the fact that those observations were performed in clinically normal foals instead of nonsurvivors. It is also possible that mild lesions identified at gastroscopy, such as hyperkeratosis or punctate erosions within the squamous mucosa, are not identified as lesions upon post mortem examination. Thus, post mortem evaluation may underestimate the true prevalence of ulcers by selecting for animals with more severe disease.

The dichotomy in the prevalence of ulcers dependent on method of diagnosis is also apparent in mature horses, where 10.3% of horses are diagnosed with ulcers at post mortem (Sandin et al. 2000), and from 11% of horses in a university riding programme (Chameroy et al. 2006) to 100% of Thoroughbreds actively racing are diagnosed using gastroscopy (Murray et al 1996). The gastroscopy data are often skewed in that many include horses in athletic training, regardless of discipline, or with clinical signs of gastrointestinal disease, whereas most post mortem studies include horses with potentially more variable backgrounds.

The prevalence of gastric and/or duodenal ulcers in foals aged <30 days was 15% in this study, which was significantly lower than noted in older foals. In addition to the relatively low incidence of ulceration, there was only one neonatal foal with a ruptured gastric ulcer. The prevalence of ulceration was unchanged between the 6 year periods of this study, despite a significant decrease in the incidence of antacid medication usage in the hospital population overall. The prevalence of disease categories other than prematurity was also not significantly different between the time periods. The increased incidence of prematurity in the later study years, albeit interesting, does not appear related to other study findings. Despite the fact that a correlation between antacid therapy and ulceration was not directly evaluated in this population, both the prevalence and the lack of association with antacid medication usage are consistent with those of a previous report (Barr et al. 2000). Therefore, it is possible that gastric and/or duodenal ulcers develop in foals for reasons other than gastric acidity alone. It has been previously reported that premature neonatal foals often maintain an alkaline intragastric pH and that sick neonatal foals have inconsistent gastric pH profiles compared with clinically normal, term neonatal foals (Sanchez et al. 2001). In that report, foals with an alkaline gastric pH were more likely to be nonsurvivors than those with an acidic pH profile. These data cumulatively question the need for therapeutic prophylaxis with antacid medications in this population of foals. Further studies are needed to determine the prevalence of gastric ulceration in hospitalised neonatal foals.

Compared to neonatal foals, weanlings had a higher prevalence of gastric ulceration (30%). Foals in this age range often develop gastric and/or duodenal outflow obstruction as a result of ulceration within the pylorus or duodenum (Andrews and Nadeau 1999). It remains unclear why foals in this age group develop outflow obstruction whereas mature horses typically do not. Multiple ulcerogenic factors have been identified in the mature horse, including gastric pH, changes in intra-abdominal pressure during exercise, diet and meal feeding (Murray 1992; Lorenzo-Figueras and Merritt 2002). Associations have been made between previous diseases and/or stressors and the development of gastric and/or duodenal ulceration in weanlings (Rebhun et al. 1982; Gross and Mayhew 1983; Becht and Byars 1986; Murray 1989).

The results of this study suggest that gastric and/or duodenal ulceration is more frequently associated with gastrointestinal disease than other categories of primary disease in foals. It is impossible in a post mortem study to determine whether the ulcers were the cause or the effect of gastrointestinal disease. Gastrointestinal disease is often associated with large fluctuations in fluid and electrolyte balance, abnormal intestinal motility, intestinal inflammation and visceral pain. These factors, in addition to gastric acidity, may be responsible for ulcer development in this population. Administration of supra-label doses of phenylbutazone (Traub et al. 1983; Geor et al. 1989) or appropriate doses of flunixin meglumine for 30 days (Traub-Dargatz et al. 1988) induced gastric ulcers in healthy weanling-aged foals. Flunixin meglumine is often administered to foals with gastrointestinal disease for both analgesic and anti-inflammatory purposes. The duration of nonsteroidal anti-inflammatory drug (NSAID) therapy required to induce gastric ulceration in foals remains unknown, but it is probably more than the few doses used clinically to control pain and inflammation in foals with colic or diarrhoea. An unexpected result was the lack of association of musculoskeletal disease with ulceration. It was hypothesised that foals with musculoskeletal disease would have been treated with NSAIDs for pain management, and that therefore foals with lameness would develop ulcers from chronic NSAID administration. Treatment with ranitidine reduced the severity of gastric ulceration in a phenylbutazone toxicity model (Geor et al. 1989). Data regarding neither actual treatment with NSAIDs or antacids nor the hospital-wide use of antacid or NSAID therapy in weanlings were available in this study. It is possible that foals with musculoskeletal disease were more likely than those with gastrointestinal disease to be treated with NSAIDs and prophylactic antacids, thus preventing a positive association between musculoskeletal disease and ulceration.

In conclusion, the prevalence of gastric and/or duodenal ulceration in nonsurviving foals remains consistent with previous reports. The relatively low prevalence of ulceration in neonates, which remained unchanged despite an overall decrease in antacid therapy, suggests that antiulcer prophylaxis may be unnecessary in this population. However, foals with gastrointestinal disease, regardless of age, may be predisposed to the development of gastric and/or duodenal ulceration. Because only one of 691 foals had perforating gastric rupture and the prevalence of ulceration appears unrelated to frequency of antiulcer therapy in the neonatal population, such prophylaxis may be unnecessary.

Conflicts of interest

No conflicts of interest have been declared.

Sources of funding



The authors appreciate the contributions of the veterinary pathology faculty and residents for access to the necropsy reports. They also acknowledge Drs Heidi Banse and Amy Steiler for their contributions to data entry. Portions of this work were presented in abstract format the Dorothy Russell Havemeyer Foundation Equine Neonatal Septicemia Workshop V, Salem, Massachusetts, 2008, and the ACVIM Veterinary Medical Forum, 2010.

Manufacturer's address

a Systat Software, Inc., San Jose, California, USA.

Author contributions The authors contributed equally to this work.