An aged (estimated 28-year-old) Arabian mare was referred for evaluation of colic. On the day before presentation, the mare was lethargic with a decreased appetite, prompting administration of flunixin meglumine1 (0.5 mg/kg, PO) by the owner. On the morning of admission, the mare was found recumbent in the stall with evidence of rolling during the night. Examination by the referring veterinarian revealed a normal rectal temperature (101.0°F), tachycardia (84 beats/min), tachypnea (50 breaths/min), and bright pink oral membranes. Rectal examination findings were normal, but 8 L of gastric reflux were recovered on passage of a nasogastric tube. Flunixin meglumine2 (1.1 mg/kg, IV) was administered before referral. Additional history included that the mare was in poor body condition when acquired 2 years previously. The mare's condition improved over the following 6 months. No prior colic episodes were reported.
On presentation, the mare was lethargic but responsive. Rectal temperature was normal (99.8°F), but heart (100 beats/min) and respiratory (50 breaths/min) rates were increased. The mare stood quietly in the stocks and did not show obvious signs of abdominal pain. Oral membranes were bright pink and borborygmi were decreased in all quadrants. Extremities were cool and peripheral pulses, including digital pulses, were difficult to palpate. A few small, dry fecal balls were passed shortly after admission, but rectal examination did not reveal any abnormalities. A nasogastric tube was passed and 10 L of gastric reflux were obtained. Weight was 393 kg and body condition score was 6/9.
Transabdominal ultrasonographic examination revealed the stomach to extend to the 16th intercostal (IC) space, but no other abnormalities were found. Abdominocentesis yielded a straw-colored peritoneal fluid; only a mild increase in TS was quantified by refractometry (2.8 g/dL; reference interval [RI] <2.5 g/dL). A rapid blood chemistry analysis revealed hyperglycemia (133 mg/dL; RI 85–115 mg/dL), mild hypernatremia (142 mmol/L; RI 135–140 mmol/L), decreased ionized Ca++ concentration (4.7 mg/dL; RI 5.6–6.1 mg/dL), increased lactate concentration (3.3 mmol/L; RI <0.8 mmol/L), and azotemia (BUN 31 mg/dL; RI 15–25 mg/dL; Cr 2.7 mg/dL, RI 0.6–1.6 mg/dL). A CBC revealed increases in PCV (57%; RI 30–45%) and TS (8.3 g/dL; RI 5.9–7.5 g/dL) and a mature neutrophilic leukocytosis (segmented neutrophils, 12,230/μL; RI 1,940–7,400/μL) with a total leukocyte count of 13,250/μL (RI 5,100–13,120/μL). Estimated fibrinogen concentration (heat precipitation method) was mildly increased (0.5 g/dL; RI 0.2–0.4 g/dL).
Initial treatment consisted of IV fluid therapy (a 2 L bolus of 7.2% NaCl solution followed by LRS supplemented with calcium borogluconate [50 mL/L of a 23% solution] and dextrose [20 g/L] at 5 mL/kg/h), oxytetracycline HCl3 (6.6 mg/kg, IV, q24h), flunixin meglumine2 (0.5 mg/kg, IV, q12h), and lidocaine HCl4 (1.3 mg/kg, IV bolus, followed by a continuous rate infusion [CRI] of 0.05 mg/kg/min, IV). The mare's feet were placed into ice boots and gastric decompression was performed every 2 hours via an indwelling nasogastric tube. A sample of gastric reflux obtained at admission and 5 serial fecal samples yielded negative culture results for Salmonella spp. PCR testing of blood and feces for Neorickettsia risticii and ELISA testing5 of fecal fluid for Clostridium difficile toxins A and B and Clostridium perfringens enterotoxin all yielded negative results.
During the initial night of hospitalization the mare remained quiet and no further colic signs were observed. One small pile of feces was passed and the volume of reflux diminished (only 6 additional liters were recovered over the next 12 hours), membrane color improved, and the heart rate decreased to 52 beats/min. By the following morning (day 2), the mare appeared adequately hydrated on clinical examination, and PCV and TS had decreased to 43% and 6.2 g/dL, respectively. No abnormal findings were detected on transabdominal ultrasonography. Consequently, IV fluids and the lidocaine CRI were discontinued, the nasogastric tube was removed, and the mare was offered a small amount of pelleted feed in the afternoon. Appetite was present but diminished. The mare remained comfortable throughout the night although heart rate remained elevated (50–68 beats/min) and only 2 pounds of feed were consumed. However, by midday on day 3, the mare started to paw intermittently and rolled a couple of times and 5 L of reflux were obtained by nasogastric intubation. Supportive treatment with IV fluids and a lidocaine CRI was reinstituted. From this colic episode onward, repeated cycles of clinical improvement followed by intermittent mild to moderate abdominal pain were observed.
Further diagnostic evaluation on day 3 included upper intestinal endoscopy that revealed a normal esophagus and one 2 × 3 cm squamous gastric ulcer just above the margo plicatus. In addition, several raised nodular lesions (estimated 5–7 mm diameter) were visualized in the proximal duodenum, some of which had central cavitation (Fig 1). Four duodenal biopsy specimens consisted of superficial portions of small intestinal villi lined by tall columnar epithelial cells. No appreciable inflammatory cell infiltrates were found in the superficial subepithelial stroma. These findings were considered nondiagnostic. No abnormal findings were detected on rectal palpation and a rectal mucosal biopsy was characterized by a moderate lymphoplasmacytic inflammatory cell infiltrate within the mucosal lamina propria and submucosa, supportive of inflammatory bowel disease (IBD).
A total of 9 L of gastric reflux was obtained over the subsequent 4 hours, after which no net reflux was recovered for the remainder of the night. The tube was removed in the morning, and the mare was again offered feed by alternating grazing grass and small amounts of hay chaff. Appetite during this period was the best since admission. During this same time period, 2 small piles of feces were passed. Omeprazole paste6 (4.4 mg/kg, PO, q24h) was added to the treatment regime. Hand walking was used to control mild colic signs and no analgesic medication was administered.
When gastrointestinal discomfort, primarily manifested by a decrease in appetite and a persistently elevated heart rate (56–60 beats/min), returned on day 4, repeat physical, rectal, and transabdominal examination findings were unremarkable, although feces had become soft. Repeat abdominocentesis yielded serosanguinous fluid high in protein concentration (TS 6.0 g/dL) with an increased total nucleated cell count (7,210/μL; RI <5,000/μL). Nondegenerate neutrophils predominated with lesser numbers of macrophages, small lymphocytes, and mesothelial cells. Phagocytized nucleated cells, and less commonly erythrocytes, were seen within macrophages, but no atypical cells were seen. Bacterial culture of the peritoneal fluid yielded no growth. A CBC revealed a normal leukocyte count (6,320/μL), but fibrinogen concentration had increased to 0.7 g/dL and a serum chemistry profile revealed mild hypoalbuminemia (3.0 g/dL; RI 3.6–4.8 g/dL), increases in total bilirubin (4.0 mg/dL; RI 0.1–2.1 mg/dL) and triglyceride (472 mg/dL; RI 9–71 mg/dL) concentrations, and increased activities of alkaline phosphatase (AP, 339 U/L; RI 95–311 U/L) and gamma-glutamyl transferase (GGT, 50 U/L; RI 6–17 U/L).
Because appetite remained poor during the night of day 4 and only small amounts of feces had been passed since presentation, a nasogastric tube was passed again on day 5 and, after no reflux was recovered, enteral fluid therapy (6 L of alternating isotonic NaCl or KCl solutions administered q4h via the indwelling nasogastric tube) was started. Enteral fluid therapy was well tolerated by the mare, but heart rate remained elevated (60–68 beats/min). Little clinical improvement was evident by the morning of day 6 and the mare was again mildly colicky. Abdominal ultrasonography had unremarkable findings, but the stomach extended to the 15th IC space. Abdominal and thoracic radiographs did not reveal significant accumulation of radiopaque material (sand or an enterolith) or other abnormalities.
Because of the persistence of tachycardia and intermittent colic signs over several days and a lack of a diagnosis, an exploratory laparotomy was performed on day 6. The duodenum and proximal jejunum were minimally distended, but appeared mildly inflamed as characterized by slight serosal reddening and edema. Retroflexion of the pelvic flexure was also found. As the colon was largely empty, the displacement was not considered a significant finding. Several biopsies of the small (proximal and midjejunum) and large intestine (left dorsal and left ventral colon) were taken. Proximal jejunal sections showed mild to moderate lymphoplasmacytic and mild neutrophilic mucosal infiltrates, and colon sections also contained inflammatory cell infiltrates, predominantly lymphoplasmacytic, in both the mucosa and submucosa. These findings were also supportive of IBD. In addition, patchy granulating fibrosis was found on the serosal surfaces of both small intestine and colon, supportive of peritonitis detected by abdominocentesis. Before surgery, antimicrobial therapy was changed to potassium penicillin G7 (22,000 IU/kg, IV, q6h) and gentamicin8 (6.6 mg/kg, IV, q24h) and postoperatively dextrose (20 g/L) was added to IV LRS solution in an attempt to correct hypertriglyceridemia. Flunixin meglumine (1.1 mg/kg, IV, q12h) and a lidocaine HCl CRI were also reinstituted for 36 hours postoperatively. Further, dexamethasone9 (0.1 mg/kg, IV, q24h) was administered for 3 days followed by 0.05 mg/kg, IV, q24h for treatment of suspected IBD. Postoperatively, the mare remained lethargic, but slowly developed a fair appetite for hay chaff. Daily abdominal ultrasonographic examination revealed nondistended small intestine with normal wall thickness (<4 mm) in both inguinal areas (stomach size was not determined because of the presence of an abdominal bandage). Small intestinal motility was initially sluggish but improved during subsequent evaluations. Clinicopathologic data on day 7 revealed a mature neutrophilia (10,390/μL) and a normal fibrinogen concentration (0.4 g/dL), whereas a serum chemistry profile performed on day 9 revealed hypoalbuminemia (2.7 g/dL) and an increased triglyceride concentration (489 mg/dL). Four days postoperatively a right forelimb lameness was detected (Grade III/V) and lateral front foot radiographs revealed evidence of chronic bilateral laminitis with moderate rotation. Hoof angle was raised and sole support was provided by application of putty elastomer and plastic shoes and phenylbutazone10 (2.2 mg/kg, PO, q12h) was started. Endogenous cortisol concentration measured 22 hours after IV administration of 20 mg of dexamethasone (as therapy for IBD) was 99 nmol/L providing support for pituitary pars intermedia dysfunction (PPID) because a value of <30 nmol/L would have been expected after administration of this dose of dexamethasone. Insulin concentration (371 pmol/L; RI <300 pmol/L) was also elevated supporting insulin resistance, although treatment with dexamethasone could have also decreased insulin sensitivity in the mare. Laminitis pain improved, heart rate returned to normal, and appetite for hay chaff, grass during walks, and pellets progressively improved from days 9 to 12. Small piles of feces were regularly produced. However, on day 12 (6 days postoperatively), more severe colic signs developed. Distended loops of small intestine were palpated on rectal examination and detected ultrasonographically in both inguinal areas. When repeated sedation with detomidine11 (6–10 μg/kg, IV) failed to control colic signs, the owners elected euthanasia.
Postmortem examination revealed 2–3 L of serosanguinous peritoneal fluid and the stomach, duodenum, and proximal 10 feet of jejunum were moderately distended. Serosal surfaces were irregularly discolored red, but no devitalized bowel, adhesions, or other obvious cause for distension was apparent. The mucosal surface of the proximal duodenum had patchy areas of hemorrhage as well as numerous raised 5–7 mm diameter nodular lesions, along with a solitary larger (~2 cm diameter) lesion (Fig 2). The remainder of the intestinal tract was considered grossly normal. Histopathological examination of multiple duodenal sections revealed multifocal nodules that consisted of markedly hyperplastic Brunner's glands with occasional cystic dilation interspersed with nests of glands resembling exocrine pancreatic tissue. Hyperplastic Brunner's glands expanded the submucosa and extended through the muscularis mucosa into the mucosal lamina propria. These glands were well-differentiated and consisted of polygonal to columnar epithelial cells arranged in tubules and acini with abundant amphophilic vacuolated cytoplasm and small round basally located nuclei (Fig 3). Occasionally, these cells lined large cystic cavities. The interspersed nests of exocrine pancreatic tissue consisted of columnar epithelial cells with eosinophilic, granular, cytoplasm arranged in acini. The solitary larger lesion was similar in microscopic appearance but was termed a Brunner's gland adenoma based on the larger size of this lesion.
Brunner's glands, or duodenal glands, are present in the proximal small intestinal submucosa of all mammals. Alkaline fluid secreted by these glands plays an important role in digestive function by buffering gastric acid. The structure of Brunner's glands in horses is similar to that of other mammals, being arranged in coiled tubules within the submucosa. Ducts of these glands course through the muscularis mucosa before emptying into intestinal crypts. Two main cell types are found in the acinus: serous and mucous cells, with the latter predominating.[1, 2] Although mucous cells are present in all duodenal glands, serous cells are mainly found in the most proximal duodenal glands extending ~10 cm beyond the pylorus.
To the authors' knowledge, Brunner's gland hyperplasia and adenoma associated with recurrent gastric distension and colic have not been previously reported in horses. In humans, Brunner's gland hyperplasia is found in a low percentage of gastrointestinal endoscopic evaluations and is often considered an incidental finding. Whereas Brunner's gland hyperplasia is characterized by nodules less than 1 cm in diameter, larger lesions are described as Brunner's gland adenoma. These adenomas are rare with an incidence in humans of less than 0.01%. However, there are reports of Brunner's gland hyperplasia and adenoma causing abdominal pain because of intermittent gastric outflow obstruction with or without ulceration.[6-8] Brunner's gland hyperplasia has also been implicated as a contributing factor to pancreatitis. In all instances, attributing gastric distension or outflow obstruction to Brunner's gland hyperplasia is a diagnosis of exclusion.
In the mare in this report, passage of a nasogastric tube yielded gastric reflux that, once removed, ameliorated colic signs on several occasions. This favorable clinical response after gastric decompression supported gastric outflow obstruction and distension as a cause for recurrent colic signs. Although bowel biopsies were supportive of IBD, a disease that could have contributed to the mare's colic signs, recurrent gastric distension and reflux is not a typical finding with IBD. The most striking finding during evaluation of this mare was the unusual cavitary, nodular lesions visualized during duodenoscopy. Further, until the terminal colic bout, small intestinal distension was not found on abdominal ultrasonography and little further reflux was recovered following initial gastric decompression. Of interest, repeated abdominal ultrasonographic examinations failed to reveal significant abnormalities (thickening or distension) of the duodenum along the right side of the abdomen or the jejunum in both inguinal areas. Next, serial serum chemistry analyses demonstrated mild increases in AP and GT activity on day 4. Although not well documented with small intestinal disease, many clinicians would interpret these results as supportive of partial outflow obstruction of pancreatic and bile fluid. In addition, other causes of gastric outflow obstruction in horses, including pyloric masses or pyloric hypertrophy and partial outflow obstruction (more common in foals), were not found. Thus, by exclusion of these and other disorders, including pancreatitis and neoplastic diseases such as squamous cell carcinoma, lymphosarcoma, or adenocarcinoma, we speculate that Brunner's gland hyperplasia and adenoma may have contributed to intermittent gastric distension and colic signs in this mare. It warrants mention that the mucosal nodules and adenoma in the duodenum of this mare were not detected by palpation of the small intestine either during surging or at postmortem examination, before the bowel was opened.
The pathogenesis of Brunner's gland hyperplasia and adenoma formation is poorly understood but is postulated to be a compensatory response to gastric acid hypersecretion.[3, 4] Brunner's gland hyperplasia has also been found with chronic pancreatitis, chronic renal failure, and Helicobacter pylori infection.[12-14] The mare in this report was in poor body condition when acquired by the owners and could have had chronic gastric ulcer disease at that time; however, body condition improved with proper nutrition and gastroscopy during hospitalization revealed only mild squamous ulcer disease and the pylorus appeared normal. The only other concurrent disorders included PPID, insulin resistance and chronic laminitis but any role of these disorders in development of Brunner's gland hyperplasia is unknown. Of interest, the mare was reported to consistently pass manure piles that were smaller than those of its herd mate, and this observation was also made during hospitalization. Whether this observation was coincidental, reflected an underlying GI motility disturbance, or was related to Brunner's gland hyperplasia also remains unclear. The interspersed nests of exocrine pancreatic tissue found in this case are another uncommon finding. Although heterotopic pancreatic tissue has been previously reported in humans[8, 14] the significance and relationship with Brunner's gland hyperplasia is not known.
Based on episodic gastric distension for which other causes were excluded, we speculate that Brunner's gland hyperplastic lesions, and solitary adenoma formation, may have caused proximal small intestinal dysfunction leading to a functional gastric outflow obstruction and recurrent signs of colic in this mare. In contrast, an unidentified gastrointestinal problem leading to recurrent motility disturbances may also have led to Brunner's gland hyperplasia. Further investigation of the prevalence of Brunner's gland hyperplasia in normal horses of all ages and those with gastrointestinal disease is warranted.