Materials and Methods
- Top of page
- Materials and Methods
Dogs evaluated between March 14, 2001 and June 12, 2007 at the Veterinary Teaching Hospital (VTH) of the Virginia Maryland Regional College of Veterinary Medicine with a history of diarrhea of at least 2 weeks duration were eligible for inclusion. On the basis of physical examination, CBC, and biochemical profile, no evidence of systemic diseases that would commonly cause diarrhea could be present. Other inclusion criteria were the availability of the radiologist (MML) to perform ultrasound examination and the owner's willingness to comply with all study requirements. The study was approved by the VTH Board. The ultrasound examination was usually completed within 24 hours before endoscopy or celiotomy. Additional diagnostic testing, localization of the origin of diarrhea to the small bowel, large bowel, or to both, formation of a diagnosis, and treatment were at the discretion of the managing clinician and were based on standard diagnostic criteria contained within the veterinary literature (Table 1).[12-28]
Table 1. Most common diagnoses
|Diagnosis||Number of Dogs (%)||Reference Number|
|IBD||35 (40)|| [12, 16-19] |
|Clostridium perfringens enterotoxicosis||12 (14)|| [20, 21] |
|GI neoplasia (excluding lymphoma and including 3 highly suspicious cases)||10 (12)|| [15, 22] |
|GI lymphoma (including 1 highly suspicious cases)||8 (9)|| [13, 23] |
|GI parasites||8 (9)|| [24, 25] |
|Lymphangiectasia||7 (8)|| [26, 27] |
|Idiopathic large bowel diarrhea||7 (8)||  |
|Campylobacteriosis||7 (8)|| [25, 28] |
|Food-responsive diarrhea||5 (5)||  |
A complete abdominal ultrasound examination1 was performed by a single board certified radiologist (MML) as previously described. The stomach was evaluated in both longitudinal and transverse planes, and followed to the duodenum. If necessary, right lateral intercostal windows were utilized for evaluation of the pylorus and proximal duodenum. The descending duodenum was followed caudally, evaluating the right limb of the pancreas concurrently. The jejunum, ileum, and colon were also evaluated.
The GI tract was evaluated for wall thickness, appearance of wall layers, luminal contents and diameter, and motility. Wall thickness was measured from the inner luminal interface to the outer serosal surface, and considered normal if within published reference ranges (stomach: 2–5 mm, duodenum: 3–6 mm depending on body weight, jejunum: 2–5 mm depending on body weight, ileum: 2–4 mm, and colon: 2–3 mm).[30, 31] Wall layers were considered normal if all layers were clearly visible, had normal relationship with each other, and were of normal echogenicity. Luminal diameter of the stomach and small bowel was assessed subjectively. If bowel dilatation was present, the affected segment was followed to identify an obstructing lesion. If gastric distension was noted, the pyloric area was carefully evaluated for obstruction. Motility of the stomach and bowel was measured and compared with published reference ranges (3–5 contractions/min).[4, 30] Mesenteric lymph nodes, if visible, were measured and assessed for shape and echogenicity. Lymph node size were considered normal if the thickness was less than 7–8 mm.[4, 32]
After discharge, the medical record was independently reviewed by 2 randomly selected (names drawn from a container) board certified internists, from a pool of 8, who did not manage that case. Cases were excluded if one or both reviewers did not agree with the original diagnosis. Each internist subjectively assessed the diagnostic utility of abdominal ultrasound examination, selecting a score from 1 to 5, from the following:
- Diagnosis was obtained via ultrasonography (including ultrasound-guided aspirate or biopsy). Additional biopsy via endoscopy or exploratory celiotomy was not necessary.
- Ultrasonography provided data that suggested endoscopy was not indicated and celiotomy should be performed. Ultrasonography suggested how to obtain a tissue biopsy (via endoscopy or celiotomy), making it very important for diagnosis.
- Ultrasonography provided important diagnostic information that helped assess other data, including endoscopic findings. Ultrasonography was important in arriving at a diagnosis.
- Ultrasonography provided descriptive information that did not affect assessment of other data obtained via endoscopy or celiotomy. The same diagnosis would have been reached without performing ultrasonography.
- Ultrasonography provided conflicting information that did not support, or may have hindered obtaining the final diagnosis.
In addition, each internist assessed whether the results of ultrasound examination otherwise contributed to case management (answering a yes/no question), independent of its contributions to the diagnosis of the cause of diarrhea. This yes/no question addressed findings involving other body systems that could be important to overall case management.
Most dogs without cytologic or histologic evaluation received an open diagnosis and were withdrawn from the study, as diagnostic utility could not be assessed. Dogs with an open diagnosis that responded favorably to a hypoallergenic diet were reclassified as having food-responsive diarrhea and retained in the study. In addition, dogs in which gastrointestinal neoplasia was highly suspected, cytologic samples were highly suspicious of neoplasia or ultrasound identified additional masses within the abdomen or nodules were present on thoracic radiographs (both consistent with metastatic disease), were also included.
Owners were asked to maintain a daily record for 6 months of vomiting and diarrhea episodes, fecal consistency graded from 1 to 5, (1 liquid feces, 3 “cow-pie” consistency, and 5 formed), the presence of fecal blood or mucus, tenesmus, appetite, and treatments. Records were collected monthly and if incomplete, follow-up information was obtained via telephone communication with the owner or referring veterinarian. Retention within the study required follow-up information for a minimum of 5 months, or until the dog died or was euthanized. The 2 evaluators who scored the diagnostic utility of ultrasound assessed the follow-up information to ensure that it was consistent with the original diagnosis; if considered inconsistent by at least 1 reviewer, the dog was removed from the study.
Outcomes for each dog included (1) median diagnostic utility score, (2) categorical diagnostic utility score (category A - median diagnostic utility score ≤3 [ultrasound vital or beneficial] versus category B - median diagnostic utility score > 3 [ultrasound not helpful or of marginal value]), and (3) disease category (category I - diseases commonly have mass lesions that should be visible on ultrasound examination [ie, lymphoma, GI neoplasia, intussusception] versus category II - diseases not expected to have a mass lesion visible on ultrasound [ie, inflammatory bowel disease (IBD), Clostridium perfringens enterotoxicosis, food-responsive diarrhea]). Normal probability plots showed that age followed an approximately normal distribution. All other continuous variables were skewed. Agreement of the reviewers' diagnostic utility scores and the yes/no question regarding other contributions of ultrasonography were assessed using Kappa coefficient.
Association between median diagnostic utility score and age, duration of diarrhea (weeks), frequency of diarrhea/day, days/week with diarrhea, fecal grade, or weight loss as a percentage of body weight was assessed using scatter plots and the nonparametric Spearman correlation coefficient. Association between median diagnostic utility score and anatomic localization of diarrhea, appetite, and all diagnoses with a minimum of 5 cases was assessed using the Kruskal-Wallis test followed by Dunn's procedure for multiple comparisons. Association between median diagnostic utility score and weight loss (present or absent), disease category, vomiting, presence and type (abdominal or rectal) of a palpable mass was assessed using the Wilcoxon 2-sample test.
Association between utility and disease category and age was assessed using a Student's t-test. Differences in diarrhea duration or frequency, weight loss as a percentage of body weight, median reviewer score, or fecal grade between categories for each outcome were assessed using the Wilcoxon 2-sample test. Differences in vomiting, appetite, diarrhea localization, presence or absence and type of palpable mass between the 2 categories for each outcome were assessed using the Pearson chi-square or Fisher's exact test as appropriate. In addition, associations with categorical diagnostic utility score or disease category were assessed using multivariable logistic modeling.
Statistical significance was defined as P < .05. Analyses were performed using SAS.2
- Top of page
- Materials and Methods
One-hundred seven dogs entered the study. Eight were removed because of incomplete follow-up and an additional 8 because a diagnosis was not determined. Three others were removed because of incomplete follow-up and open diagnosis. A single dog was removed because 1 reviewer thought that follow-up information did not support the diagnosis. Eighty-seven dogs completed the study. Mean age was 6.7 years (±4.0 years), there were 44 females and 43 males, and 12 mixed breeds and 75 purebred dogs. The pure breeds included 4 Yorkshire Terriers, Cocker Spaniels, German Shepherds, and Labrador Retrievers, 3 Boxers, Beagles, Golden Retrievers, Rottweilers, Shetland Sheepdogs, and Standard Poodles, 6 breeds with 2 representatives, and 29 breeds with a single dog.
Median duration of diarrhea was 13 weeks (range; 2–312 weeks). Median number of diarrhea episodes/day was 3.5 (range; 0.5–12), and days/week with diarrhea was 7 (range; 0.25–7). Median fecal grade was 2 (range; 1–3.5). Twenty-eight (32%) dogs chronically vomited. Weight loss was reported for 50 of 81 (62%); its presence was not known in 6. Median weight loss estimate (known for 48 of 50), as a percentage of body weight before onset of diarrhea, was 12.6% (range; 3–60%). Appetite was normal in 40 (49%), decreased in 36 (44%), increased in 5 (6%), and unknown in 6 (7%). Six dogs (7%) had a palpable abdominal mass and 5 (6%) had a palpable rectal mass on initial physical examination. Diarrhea was localized to the small bowel in 20 (23%), the large bowel in 28 (32%), and to both in 39 (45%). The most common diagnoses are listed in Table 1. A single cause of diarrhea was present in 60 (69%) dogs, 2 diagnoses in 22 (25%), 3 diagnoses in 4 (5%), and 4 diagnoses present in 1 (1%).
The diagnostic utility scores for each dog were moderately in agreement (P < .0001, Kappa = 0.55), differing in 23 (26%) cases. In these 23, 20 scores differed by only 1 category (15/20 were scored 3 or 4). In 57 (66%), the reviewers thought that the same diagnosis would have been reached without performing ultrasonography (score of 4). There were 15 (17%) with a median score of 3.5 in which 1 reviewer thought ultrasound examination provided important diagnostic information (score of 3), whereas the other felt it did not (score of 4). The median reviewers' scores are shown in Table 2.
Table 2. Median diagnostic utility scores
|Diagnostic Utility Score||Number of Cases (%)|
There were significant relationships between median diagnostic utility scores and the following: presence of weight loss (P = .0086), palpation of an abdominal or rectal mass (P = .0031), localization of diarrhea to mixed bowel versus large bowel (P = .0493), disease category (P < .0001), and diagnosis. Smaller utility scores occurred in cases of GI neoplasia compared with all the common diagnoses, except lymphoma.
The diagnostic utility of ultrasound was considered vital or beneficial (median score ≤3; utility category A) in 13 (15%). The following significant differences were found between these 13 dogs and the 74 diagnostic utility category B dogs, in which the ultrasound examination was not thought to be helpful (score of >3.5) or of marginal value (score of 3.5): palpation of an abdominal or rectal mass (46% versus 7%) (P = .0011), greater percentage of weight loss (median 18% [range; 0–60] versus median 4.2; range 0–59) (P = .0340), fewer dogs with decreased appetites (9% versus 39%) and more with normal appetites (82% versus 38%) (P = .0075), and a greater proportion of dogs with GI neoplasia (70%) and GI lymphoma (30%) and a lower proportion of dogs with IBD (0% versus 26%) (P = .0026). Multivariable logistic regression identified palpation of an abdominal or rectal mass as the only predictor of categorical utility score. Dogs with a mass were 30 times more likely to be placed in category A (odds ratio 30.5, 95% CI 5.5–169.6) (P < .0001).
There were 15 (17%) dogs in which at least 1 reviewer identified additional benefits of ultrasonography toward case management. The reviewers were moderately in agreement (P < .0001, Kappa 0.51) regarding the yes/no question, differing in only 9 (10%). Of these 15 dogs, 5 received median diagnostic utility scores of 4, 4 were graded as 3.5, whereas 6 others received scores of ≤3. The following (and the frequency) were identified in these 15 dogs that were not thought to be responsible for chronic diarrhea: abnormal hepatic echogenicity (3), adrenomegaly (3), thickened or irregular gall bladder wall (3), hepatic mass (2), dilated renal pelvis (2), splenic mass (2), urinary bladder mass (1), gastric mass (1), abnormal prostate (1), and abnormal pancreas (1). Eleven dogs had a single abnormality and 4 had 2.
Compared with disease category II, dogs with diseases likely to cause a mass detectable with ultrasound (disease category I) were older (mean 8.3 ± 3.5 versus 6.2 ± 4.1) (P = .0381), had diarrhea more days/week (first quartile; 7, range; 2.5–7 versus first quartile; 2.5, range 0.25–7) (P = .0217), experienced a greater percentage of weight loss (median 16%, range; 0–60% versus median 3%, range; 0–59%) (P = .0430), were more likely to have an abdominal or rectal mass palpated (38% versus 5%) (P = .0004), were more likely to have signs of mixed bowel diarrhea (67% versus 39%) than large bowel diarrhea (10% versus 39%) (P = .0251), and more likely to have a normal appetite (74% versus 36%) instead of a decreased appetite (21% versus 58%) (P = .0092). Multivariable logistic regression identified palpation of an abdominal or rectal mass as the only predictor of disease category. Dogs with a mass were 21 times more likely to be placed in disease category I (odds ratio 21.8, 95% CI 4.1–116.8) (P = .0003).
- Top of page
- Materials and Methods
Abdominal ultrasonography did not substantially contribute to determination of the diagnosis in the majority of dogs (59/87, 68%) with chronic diarrhea. In most, the diagnosis required endoscopy, or less commonly, celiotomy. However, abdominal ultrasonography was vital to the diagnostic evaluation of 9/87, (10%) dogs, and beneficial in 4/87 (5%) more. In 15/87 (17%) others, the ultrasound examination was of questionable benefit.
In addition, of the 15 dogs in which the ultrasound otherwise contributed to case management, 9 received a diagnostic utility score of 3.5 or 4 (ultrasonography not beneficial). By including these 9 dogs with the 13 other dogs, in which the diagnostic utility of ultrasound was vital or beneficial, ultrasonography was beneficial in the diagnosis and management of 22 (25%) dogs.
Of the factors directly associated with better diagnostic utility, the presence of weight loss or palpation of an abdominal or rectal mass should support the decision to perform abdominal ultrasound in dogs with chronic diarrhea. However, some of the dogs in this study with utility scores ≤3, did not have either weight loss or a palpable mass. In dogs without weight loss or a palpable mass diagnostic utility could be increased by considering factors indirectly associated with utility, ie, those related to classification of dogs as utility category A or disease category I. Utility category A dogs had a greater percentage of weight loss and fewer dogs with a decreased appetite than those in utility category B. Although difficult to determine disease category, based on history and physical examination, the presence of the following increased the likelihood of a dog being classified as disease category I: increased age, more days/week with diarrhea, greater percentage of weight loss, greater percentage of dogs with normal versus decreased appetite, and a greater proportion of mixed bowel diarrhea than large bowel. The authors cannot explain why more dogs in both utility category A and disease category I would have a normal appetite, because GI neoplasia was a common diagnosis in both groups, and affected dogs would be expected to have decreased appetite.
Another potential benefit of abdominal ultrasonography was identified, but not further evaluated. There were 11 dogs that were initially entered into the study, but withdrawn because a clinical diagnosis was not reached upon discharge. In these dogs, additional diagnostic tests, such as gastroduodenoscopy, celiotomy, or fine-needle aspiration, were not done because a diet trial, administration of an anthelmintic, or both had not been performed, or the owner elected not to pursue further diagnostic testing. In many of these cases, the reviewers commented that the normal ultrasound examination could be considered beneficial to case management, because it supported the clinicians' decisions not to perform additional diagnostic tests and to institute therapeutic trials. This was an unforeseen beneficial role for ultrasonography that was not considered when the diagnostic utility scoring system was designed and implemented and was also identified in a previous study assessing utility in vomiting dogs.
Accuracy of each diagnosis was vital to the validity of the results because utility scores were based on the contribution of ultrasonography to arriving at each diagnosis. Following the clinical signs of each dog for a minimum of 5 months, or until death or euthanasia validated the diagnosis.
The authors cannot compare results of this study to other studies in dogs with chronic diarrhea. Ultrasonographic findings in dogs and/or cats have been reported for IBD,[1-5] GI lymphoma,[1, 4, 6-8] intestinal neoplasia,[4, 7-9] intussusception,[7, 10] and GI foreign bodies.[7, 11] However, diagnostic utility of ultrasonography, when other diagnostic tests are utilized in dogs with chronic diarrhea, was not assessed in these reports. Results of the present study were remarkably similar to those in a previous report evaluating diagnostic utility in dogs with chronic vomiting, in which abdominal ultrasonography did not substantially contribute to the diagnosis in 69% of dogs (versus 68% in the present study). In addition, in that vomiting study, ultrasound contributed to case management, independent of its value for the diagnosis of the primary problem in 12% of dogs (versus 17% in the present study). Combining dogs with utility scores ≤3 with those in which ultrasound otherwise contributed to case management resulted in ultrasound being beneficial in diagnosis or case management in 27% of dogs in the vomiting study and 25% in the present study. In the vomiting study, the following factors were associated with increased utility: increasing age, greater number of vomiting episodes/week, presence of weight loss, a greater percentage of lost body weight, and final diagnosis of GI lymphoma or gastric adenocarcinoma.
Although assignment of a diagnostic utility score and assessment (yes/no question) of the value of ultrasonography in case management, independent of its contributions to the diagnosis of the cause of diarrhea, were subjective evaluations, there was a moderate level of agreement between reviewers, as found in the vomiting study. Reviewers assessed each medical record independently, and agreed on the diagnostic utility score in 74% of cases and on the value of ultrasonography, independent of the diarrhea problem, in 90% of dogs. This was despite a wide variety of individual opinions regarding case management, as 8 board certified internists reviewed medical records. Objectivity was maintained because only clinicians who were not directly involved with an individual case were eligible for medical record review of that case.
The decision to perform diagnostic tests is complex and based on many factors, including personal preference, institutional bias, severity and chronicity of clinical signs, client's wishes, and test invasiveness, availability, expense, and diagnostic utility. This study provides the only evidence available in the veterinary literature regarding diagnostic utility of abdominal ultrasound examination in dogs with chronic diarrhea. On the basis of study's findings, clinicians should expect an abdominal ultrasound examination to be vital or beneficial to obtaining a diagnosis or to identify other factors important in case management in approximately 25% of dogs with chronic diarrhea. Clinical factors were identified that veterinarians can utilize to help determine which dogs with chronic diarrhea should be expected to have a greater diagnostic utility of ultrasound. The authors recommend performing abdominal ultrasound in dogs with chronic diarrhea that have lost weight, especially a large percentage of weight, have a palpable abdominal or rectal mass, have mixed bowel diarrhea, are older, have diarrhea on most days of the week, or are otherwise suspected of having GI neoplasia. In addition, if the decision to offer therapeutic trials versus additional invasive tests remains unclear, or the owner does not wish to pursue further diagnostic testing, a normal ultrasound examination supports the decision to pursue the more conservative path. Conversely, the authors do not recommend routinely performing ultrasound examination in dogs with chronic diarrhea that have not lost weight, do not have a palpable abdominal or rectal mass, or are likely to have IBD. Diagnosis in these cases usually can be made using endoscopy. In only 2 (7%) of 31 dogs without weight loss (one of which had a palpable abdominal mass), only 1 of 28 (4%) without weight loss or a palpable abdominal or rectal mass (cecal inversion), and none of the 35 dogs with IBD was ultrasound examination useful for diagnosis. In these dogs, money saved by not performing ultrasonography could be utilized for endoscopy, which should provide a definitive diagnosis. There were 5 dogs in which ultrasound identified an abdominal mass that was not palpated on physical examination. The weight loss status was known for only 3 dogs. Two had lost 18 or 19% of body weight and the third with cecal inversion, listed above, that had not lost weight. Utilizing the criteria identified in this study, ultrasound should have been performed in 2 dogs because of the large amount of weight loss. Not performing ultrasound in the 3rd dog with cecal inversion would have resulted in performance of an unnecessary colonoscopy, rather than celiotomy.
There are several limitations of this study. Most of the dogs were referred for chronic diarrhea. The diagnostic utility of ultrasound could be different in the population of dogs managed at primary care clinics. The duration of clinical signs would be expected to be shorter in these cases and disease processes less advanced, potentially making lesions more difficult to detect. Secondly, ultrasonography is highly operator and equipment dependent. This study utilized a single, highly experienced, and skilled ultrasonographer, using high quality equipment. It is possible that the utility of ultrasonography would be lower if less experienced clinicians or those using lesser quality equipment perform examinations. Thirdly, the authors' institutional bias during the past 25 years includes histology for the diagnosis of dogs with chronic diarrhea, most often obtained via gastroduodenoscopy. This approach is well supported within the veterinary literature[5, 12, 16, 33-37] and was the basis for the design of the utility scoring system. It is possible that a different scoring system, not based so heavily on endoscopic biopsy, could have produced different results. Fourthly, we did not define a standard diagnostic plan, criteria for establishing a diagnosis, or specific treatment protocols for each diagnosis; these decisions were left to managing clinicians. However, standard diagnostic criteria from the veterinary literature were utilized to establish a diagnosis.[12-28] Although this added variability to case management and treatment response, this may have actually strengthened results because it more accurately reflects global case management. However, despite the lack of specific diagnostic and therapeutic criteria, both reviewers had to agree with the clinical diagnosis for each dog and that the response to treatment was consistent with this diagnosis. Fifthly, the number of dogs in which ultrasound examination was useful for diagnosis was small. These numbers precluded performing multivariable regression analyses directly relating factors to utility. Larger numbers of dogs may have identified more and stronger characteristics associated with better diagnostic utility in dogs with chronic diarrhea. Finally, it is possible that multivariable evaluation of routine laboratory results might have identified other factors that could be associated with increased diagnostic utility of ultrasound.
In summary, ultrasonography was vital or beneficial to obtaining a diagnosis in dogs with chronic diarrhea in 13/87 (15%) of cases, not helpful in 59/87 (68%), and of marginal value in 15/87 (17%). Including other abnormalities identified with ultrasonography independent of the diagnosis of diarrhea, abdominal ultrasound was helpful in 22/87 (25%), not helpful in 54/87 (62%), and of marginal value in 11/87 (13%) of dogs with chronic diarrhea.