Gastrointestinal angiodysplasias diagnosed using video capsule endoscopy in 15 dogs

Abstract Background Angiodysplasia (AGD) is rarely diagnosed in dogs with gastrointestinal bleeding (GIB) and is reported in case reports in dogs. Objective Describe signalment, clinical and diagnostic features of dogs with gastrointestinal (GI) AGD diagnosed by video capsule endoscopy (VCE). Animals Dogs with overt or suspected GIB which underwent VCE. Methods Dogs for which a VCE was submitted for overt or suspected GIB from 2016 to 2021 were selected retrospectively. Medical records and full‐length VCE recordings where AGDs were initially detected, were reviewed by 2 trained internists. AGD was considered definitive if 2 readers detected it. Signalment, clinical signs, blood work, medications, concurrent diseases, findings of previous conventional endoscopy, and surgical exploration (if applicable) of dogs with AGD were recorded. Results Definitive AGD was diagnosed in 15 of 291 (5%) dogs (12 males, 3 females). Twelve (80%) had overt GIB, 11 (73%) had hematochezia, and 6 (40%) had microcytic and hypochromic anemia. AGD was missed by conventional endoscopy in 9/9 dogs and exploratory surgery in 3/3 dogs. Thirteen capsules were administered by mouth (1 incomplete study), and 2 via endoscopy directly into the duodenum. AGD was visualized in the stomach of 3 dogs, in the small intestine of 4, and in the colon of 13 dogs. Conclusion and Clinical Importance Although rare, AGD should be considered in dogs with suspected GIB after a negative conventional endoscopy or surgical exporation. Video capsuel endoscopy appears to be a sensitive test to identify AGD within the GI tract.

bleeding. AGD can affect any segment of the GI tract. In humans, AGD are most commonly found in the small bowel. 2,3 The exact etiology of the disease in humans is unknown. Clinical signs can vary from subclinical without any visible bleeding (occult GI bleeding) to visible hemorrhage (overt GI bleeding) in the form of hematemesis, hematochezia, or melena. 4,5 The GI hemorrhage can become chronic and intermittent resulting in iron deficiency anemia.
AGD in humans is recognized as an important cause of lower gastrointestinal bleeding (GIB) in elderly adults 6 but also occurs in younger patients. 7 The diagnosis of AGD in humans is often delayed because of the intermittent nature of the bleeding and the lack of sensitivity of the abdominal ultrasound, conventional endoscopy in cases of small bowel AGD, and surgical exploration to detect abnormal vessels. 8 The diagnosis is made based on results of endoscopic visualization of the gastrointestinal mucosa. Histopathological confirmation is not required. 1 Video capsule endoscopy (VCE) has many advantages over conventional endoscopy, as it is a less expensive, outpatient procedure with the further advantage of being able to evaluate the entire GI tract using a higher magnification. Therefore, VCE has become the diagnostic test of choice to detect AGD in humans. [9][10][11] Only 7 cases of GI vascular anomalies are reported in dogs, and all abnormalities were located in the rectum or colon. [12][13][14][15][16][17] This suggests that these disorders might be less common or underdiagnosed in dogs compared with people. [12][13][14][15][16][17] The median duration of clinical signs is 33 months (range: 0.25-60 months), [13][14][15][16][17][18] which illustrates that making a diagnosis of AGD is challenging. Three of 7 dogs underwent surgical exploration that did not identify the presence of AGD. 12,14,15 The AGD diagnosis was made with conventional endoscopy in all 7 cases. However, considering that in people AGD is most commonly diagnosed in the small intestine (SI), it is possible that AGD is missed in some dogs that only receive bidirectional endoscopy without further assessment of the complete small intestine.
The use of VCE has become more popular in dogs as a minimally invasive tool to detect GIB. [19][20][21][22][23] This procedure is well tolerated in dogs and yields a diagnosis in over 75% of dogs with overt GIB. 19 Because repeating VCE is less expensive and invasive than conventional endoscopy and surgical exploration, it is also a promising test for intermittent GIB.
The objective of this retrospective study was to report the signalment, clinical and diagnostic features of AGD diagnosed by VCE in a large cohort of dogs with suspected or overt GIB. reader 1(AD, nonblinded to history, signalment and initial diagnosis) and reader 2 (JSt, blinded to history, signalment and initial diagnosis) to confirm the diagnosis. Reader 1 and 2 had read more than 200 and 58 VCE studies respectively before this study. Esophageal, gastric, and small intestinal transit times were recorded. If the capsule reached recording capacity while still in the stomach, the study was considered "incomplete." If the capsule was delivered endoscopically directly into the small intestine or if it turned off while still in the SI, the study was considered "partial."

| MATERIALS AND METHODS
A lesion of AGD was defined by the presence of a clearly demarcated, bright-red, flat lesion, consisting of tortuous and clustered capillary dilatations within the mucosal layer. 24 The number of frames where AGD lesions were visible was recorded as follows: 1 frame, 2 to 5 frames, 6 to 10 frames or more than 10 frames. If both investigators (reader 1 and 2) observed a lesion of AGD in ≥2 frames, it was considered definitive. If only 1 reader observed a lesion of AGD on ≥2 frames, it was considered suspicious. For each dog, the presence of a AGD lesion was recorded along with the location (stomach, SI or colon) as well as the probability of bleeding based on a previously reported score. 25 (Table 1). Nine dogs were fed a hypoallergenic diet for a suspected (4) or confirmed (5) inflammatory bowel disease by histopathology.

| Concurrent diseases of the 15 dogs with definitive GI AGD lesions
Based on the diagnostic algorithm for immune-mediated hemolytic anemia (IMHA), 26 testing was supportive of IMHA in 1 dog based on the presence of spherocytes and autoagglutination on the blood smear 26 and suspicious for IMHA in 9 dogs. Direct agglutination tests were performed on 3 dogs and were negative.

| Ongoing treatment of the 15 dogs with definitive GI AGD lesions
Eight dogs had been treated with prednisone (median: 2 mg/kg/day; range: 1-4 mg/kg/day) for a median of 3.25 months (range: 2 weeks to 5 months) for suspected IMHA. In 2 dogs, the prednisone was discontinued 2 weeks before the capsule was administered. At the time of capsule administration, 5 dogs were on immunosuppressive dose of prednisone. Two of the dogs had been on budesonide, 1 for 1 month and 1 for 6 months (0.14 mg/kg q24h and 0.1 mg/kg q24h respectively). Fifteen dogs were receiving proton pump inhibitors, and 15 were receiving sucralfate. One dog was administered maropitant, and Vitamin B12. One was treated with barium, and another 1 with Yunnan Baiyao. One dog was administered pimobendan, and 1 was administered iron supplementation. Eight dogs received a median of 3 transfusions (range 1-8) before the VCE over a median of 30 days (range 1-55).

| Clinicopathological data in dogs with definitive GI AGD lesions
All dogs had a complete blood count and 14 had a biochemistry profile performed within 1 week before the VCE. All 15 dogs were T A B L E 1 Number of dogs with suspected or overt GIB, hematochezia, melena, hematemesis, and a combination of hematochezia and melena, among the 15 dogs with definitive angiodysplasia and among the 276 dogs without definitive angiodysplasia respectively. Hematemesis 0 (0) 44 (16) .1385 Hematochezia and melena 8 (53) 32 (12) .0002 Note: Angiodysplasia was diagnosed by VCE by 2 readers independently. If only 1 reader diagnosed AGD, it was considered a suspicious lesion, and the dog was classified as not having AGD. Abbreviations: AGD, angiodysplasia; GI, gastrointestinal; GIB, gastrointestinal bleeding; NA, not applicable; VCE, video capsule endoscopy.
anemic. The median hematocrit was 15% (range: 9-37%).   Reader 1 and 2 detected AGD in more than 10 frames in 14 dogs, and in 5 to 10 frames in 1 dog. Angiodysplasias were observed in the stomach in 3 dogs (Figure 1). One of these studies was incomplete.
One out of 3 had colonic AGD as well ( Table 2). AGD was visualized in the first third of the small intestine in 1 dog (Figure 2), in the distal third of the small intestine in 3 dogs (Figure 3), and in the colon in 13 dogs (Figures 4 and 5). Eleven of 12 dogs with complete VCE studies had colonic AGD. The 2 dogs with capsules administered endoscopically had colonic AGD (Table 2).   Figure 1).  This study suggests that VCE might be more sensitive than conventional endoscopy and surgery to detect GI AGD in dogs.

F I G U R E 3
Five percent of dogs which had a VCE for suspected or overt GIB had AGD. This number is comparable to humans, where 2% to 8% of GIB are secondary to AGD. 28 In humans, the frequency of AGD in the colon being a cause of lower intestinal bleeding is 40% among adults, 28 but the frequency is unknown for children and for dogs.
The most common clinical signs among the 15 dogs with definitive AGD were hematochezia, lethargy, melena and diarrhea. This is consistent with reports of disease in dogs. 12 Hematochezia and hematochezia combined with melena were significantly more common in the group of dogs with GIB and AGD compared to the group without AGD. The results were not adjusted for comparison to avoid underpowering the tests and increasing type II error rates. 29 These findings suggest that AGD should be considered in adult dogs with GIB and concurrent hematochezia or hematochezia and melena.
In this study, diagnosis occurred a median of 10 weeks after the start of the clinical signs or unexplained anemia. Similarly, in humans a delay in diagnosis of 2 months to 3 years has been reported. 8,30 The delay in diagnosing AGD in dogs is likely because of the intermittent characteristic of the GIB, and the possible lack of sensitivity of the conventional endoscopy and surgical exploration. Additionally, bleeding lesions may be identified on traditional endoscopy but misinterpreted as erosive or ulcerative lesions instead of abnormal vessels.
AGD cannot be identified using an AUS; however, it can, in some patients, be observed by conventional endoscopy. In the veterinary literature, AGD was diagnosed through colonoscopy in 7 dogs. 14,15 However, AGD should not be ruled out based on the absence of abnormal vessels on conventional endoscopy because intermittent bleeding, localization of AGD lesions within the mid SI, lower magnification (versus VCE), and overinflation of the bowel with air can cause lesions to blanche and be invisible. Furthermore, the AGD can easily be confused with iatrogenic minor mucosa trauma from the endoscope. The appearance of the lesions can also be influenced by the patient's blood pressure, degree of anemia, and effective blood volume, as well as the routine use of opioids 31 that are vasoconstrictive. 32 In humans, AGD is rarely detected during surgical exploration because the lesions are not always visible on the serosal surface of the GI. 30 Similarly, in 2 veterinary case reports and in 3 dogs in our study, surgical exploration failed to detect AGD. 14,15 Moreover, in humans, AGD can be missed during histopathologic observation of GI biopsies. Most lesions are not diagnosed histologically because of their small size (shrinkage with processing of the specimen, or error in targeting the biopsy). 33 Similarly, in our study, surgical biopsies of the colon did not reveal AGD in dogs with confirmed AGD on endoscopy.
In humans, VCE is the preferred method of small-bowel evaluation after inconclusive conventional endoscopy because of higher or at least equivalent diagnostic performance as compared with other, more invasive techniques, such as push enteroscopy, mesenteric angiography, computed tomography and intraoperative enteroscopy. 10,34,35 VCE is a noninvasive procedure with a high rate of complete small bowel exploration (>90%) and high diagnostic performance (>60%) used to diagnose GIB lesions in humans. 10,11 Compared to conventional endoscopy, VCE facilitates visualization because of the higher magnification, the evaluation of the entire GI tract, and the absence of insufflation, sedation, or general anesthesia that can mask AGD. Besides, because of its lower cost and lack of invasiveness, it can be repeated easily. In 3 dogs in our study, the capsules identified AGD in the distal small intestine that could not previously be reached by a conventional endoscope.
VCE has several limitations. Lesion localization is estimated according to the transit time and the use of the pylorus and cecum as landmarks, but these lack precision 36 and can be problematic if surgical resection is indicated. Nonetheless, VCE can help in deciding the best insertion route for endoscopy if endoscopic cauterization is considered (upper or lower).
In humans, 1 study compared capsule endoscopy and intraoperative enteroscopy in 47 patients with obscure GIB. 37 The sensitivity and specificity of VCE were 95% and 75%, respectively. The positive predictive value of VCE was 95%, and the negative predictive value was 86%. 37 The sensitivity and specificity of VCE to detect AGD is still unknown in people and in dogs.
The ability of VCE to detect a true AGD is affected by the diffi-  The presence of inflammatory bowel disease in 5 dogs in our study and 1 dog in a previous report 16 suggests a potential association between AGD and IBD in dogs.
Aortic stenosis, 55 14,16 Similarly, AGD is more common in human patients aged >60 years. 6 The etiology of AGD is unknown, but 1 potential age-related explanation is a partial obstruction of submucosal veins leading to a weakening of the precapillary sphincters, creating congestion. 60 In humans, it has been suggested that chronic obstruction of the submucosal veins leads to capillary dilatation and eventually to the failure of a precapillary sphincter. 1,30,61 In our study, the majority of dogs diagnosed with definitive AGD were males (12/15, 80%). Similarly, 6/7 cases reported in the veterinary literature were male. [12][13][14][15][16][17] A statistical sex predilection could not be examined because of the small number of dogs in our study, nor has 1 been documented in humans. 62 In our study, 3 of 15 dogs with definitive AGD were Miniature Schnauzers born ≥6 years prior and living in North America. The dogs were not related but a breed predisposition is possible. No other Schnauzers were previously reported. [12][13][14]17 Further studies are needed to determine if some breeds are predisposed to AGD.
This study has several limitations. First, as per its retrospective nature, the management of the cases was not standardized, and some data may have been missed in the medical records. Second, it is possible that the interpretation of VCE findings by reader 1 was biased because this reader was not blinded to signalment, clinical history and initial VCE findings. Therefore, a lesion was only considered as definitive AGD if both readers were in agreement as reader 2 was blinded to the dog's information and also to the selection criterion for the VCE studies.
Third, the selection of VCE studies that were reviewed by the 2 readers (AD, JSt) was based on the initial diagnosis made on the first reading. Therefore, it is possible that AGD lesions were missed in the 276 videos that were not reviewed. However, since the initial readings were performed by experienced board-certified internists, the risk of missed lesions is considered low.
It was not possible to review the medical records of the 273 dogs who underwent VCE for suspected or overt GIB during the same period of time and for which no AGD was detected initially. For this reason, no solid conclusion can be offered regarding a possible correlation of breed, IMHA, or IBD and GI AGD.