ARTICLE IN PRESS The Veterinary Journal ■■ (2015) ■■–■■
Contents lists available at ScienceDirect
The Veterinary Journal j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / l o c a t e / t v j l
Development and validation of an endoscopic activity score for canine inflammatory bowel disease J.E. Slovak a,*, C. Wang b, Y. Sun b, C. Otoni c, J. Morrison d, K. Deitz d, D. LeVine d, A.E. Jergens d a
Department of Veterinary Clinical Science, College of Veterinary Medicine Washington State University, PO Box 646610, Pullman, WA 99164, USA Department of Veterinary Diagnostic and Production Animal Medicine, College of Veterinary Medicine Iowa State University, 1600 S. 16th St, Ames, IA 50011, USA c VCA Arboretum View Animal Hospital, 2551 Warrenville Rd, Downers Grove, IL 60515, USA d Department of Veterinary Clinical Science, College of Veterinary Medicine Iowa State University, 1600 S. 16th St, Ames, IA 50011, USA b
A R T I C L E
I N F O
Article history: Accepted 27 December 2014 Keywords: Gastrointestinal endoscopy IBD Inter-observer agreement CIBDAI Mucosal appearance
A B S T R A C T
The aim of this study was to develop and prospectively validate a simple endoscopic score of disease activity for dogs with inflammatory bowel disease (IBD). Archived endoscopic still images and video recordings of gastric, duodenal, and colonic endoscopic examinations were displayed to novice and experienced endoscopists for assessment of inflammatory activity using established descriptions. The mucosal appearances evaluated were normal tissue, erosions, friability, increased granularity, lymphangiectasia (duodenum), and mass (colon). Fleiss and Cohen’s Kappa statistics were used to estimate the inter-observer agreement of the index. For duodenal assessment, there were statistically significant (P < 0.05) differences in inter-observer agreement, with experienced endoscopists performing better than novice endoscopists in the accurate identification of mucosal appearance of the duodenum. In contrast, there was no significant difference between novice and experienced endoscopists in their interpretation of gastric (P = 0.10) and colonic (P = 1.0) mucosal appearances. Validation studies using endoscopic video clips to assess the same endoscopic parameters by quantitative (lesion number and severity) and qualitative (presence of mucosal lesions) methods showed moderate-to-substantial agreement between experienced endoscopists. Based on the observations that the quantitative and qualitative scores of mucosal appearances are virtually identical, and that qualitative assessment was performed more quickly and objectively than quantitative assessment, we propose a simple endoscopic activity score based on qualitative criteria alone in dogs with inflammatory bowel disease. © 2014 Elsevier Ltd. All rights reserved.
Introduction The pathogenesis of canine inflammatory bowel disease (IBD) probably involves interplay between the mucosal immune system and the intestinal microbiota, similar to human IBD (i.e. Crohn’s disease, CD, and ulcerative colitis, UC; Allenspach, 2011; Sartor, 2006; Xavier and Podolsky, 2007). Diagnostic tests including dietary trials, routine hematology, parasitic and bacteriologic fecal analyses, radiographic imaging, and histopathologic examination of intestinal biopsy specimens serve to eliminate other causes of chronic enteropathy. Gastrointestinal (GI) endoscopy is a useful and relatively non-invasive technique to diagnose IBD. It allows direct visualization of the mucosa for the acquisition of targeted biopsies to evaluate
* Corresponding author. Tel.: +1 507 3178960. E-mail address: [email protected] (J.E. Slovak).
the severity and extent of intestinal inflammation (Roth et al., 1990; Zoran, 2001). Determination of an inflammatory state is critical for defining disease activity and for tailoring IBD therapy. Endoscopic measures of mucosal inflammation in human IBD have been in use for over 40 years; however, no standardized model has been established (Truelove and Witts, 1955; Powell-Tuck et al., 1978; Seo et al., 1992; Lichtiger et al., 1994; Walmsley et al., 1998). Most indices for CD and UC are based on observations of mucosal erythema, increased granularity, vascular pattern, spontaneous bleeding, and mucosal damage (mucus, fibrin, exudates, erosions and ulcer), using simple scoring systems to define inflammatory activity. To date, no similar validated endoscopic score exists in veterinary medicine, and there is only limited trial data (Allenspach et al., 2007; Garcia-Sancho et al., 2007) evaluating the duodenal appearance of canine IBD, despite the suitability of the dog as a spontaneous animal model for intestinal inflammation (Jergens and Simpson, 2012). We recently reported that the inter-observer agreement for duodenal
http://dx.doi.org/10.1016/j.tvjl.2014.12.030 1090-0233/© 2014 Elsevier Ltd. All rights reserved.
Please cite this article in press as: J.E. Slovak, et al., Development and validation of an endoscopic activity score for canine inflammatory bowel disease, The Veterinary Journal (2015), doi: 10.1016/j.tvjl.2014.12.030
ARTICLE IN PRESS J.E. Slovak et al./The Veterinary Journal ■■ (2015) ■■–■■
2
Fig. 1. Representative still images used in the development phase of the endoscopic study. (A) normal stomach; (B) gastric erosions; (C) gastric friability; (D) gastric granularity; (E) normal duodenum; (F) duodenal erosions; (G) duodenal friability; (H) duodenal granularity; (I) duodenal lymphatic dilatation; (J) normal colon; (K) colonic erosions; (L) colonic friability; (M) colonic granularity; (N) colonic mass.
endoscopic assessment of canine IBD differed among trainee and experienced endoscopists on the basis of operator experience (Slovak et al., 2014). The aim of the present study was to develop and validate an endoscopic disease activity score for the mucosal appearance of the stomach, duodenum and colon of dogs with IBD. Materials and methods Study design The study was comprised of four parts. In the first part (development phase), the most relevant duodenal endoscopic variables (still images) in dogs with IBD were selected and scored for inter-observer agreement between novice and experienced endoscopists. For the second part of the development phase, an activity score based on endoscopic variables (still images) affecting the gastric and colonic mucosa was designed and assessed for inter-observer agreement among novice and experienced endoscopists. In the third part (the validation phase), representative video clips of approximately 5 min duration of gastroenteroscopic and/or colonoscopic examinations were assessed simultaneously but scored independently by two experienced endoscopists using the endoscopic criteria derived from the development phase. In the fourth phase, a simplified score was proposed based on the validation test results of endoscopic activity observed in the stomach, duodenum, and colon of dogs with IBD. Endoscopic examinations The details of the initial duodenal examination tests are described elsewhere (Slovak et al., 2014). In brief, the duodenal endoscopic mucosal appearance obtained from 25 dogs diagnosed with IBD at a single study center was assessed for inter-observer agreement. For the development phase of the study, 27 dogs diagnosed with IBD from 2002 to 2012 underwent gastroscopic examination, and 23 dogs with IBD from 2002 to 2011 undergoing colonoscopy were evaluated at the same study center. Endoscopic mucosal lesions were assessed for inter-observer agreement. Archived endoscopic still images from individual GI procedures were retrieved from a computerized database and reviewed. A total of 30 gastric, 35 duodenal and 30 colonic still images were selected for operator evaluation. A canine IBD activity index (CIBDAI) score as described by Jergens et al. (2003) was assigned retrospectively during the development phase of the study. For each dog in the validation phase, a CIBDAI score was prospectively assigned and a video recording of the entire gastroduodenoscopic procedure (n = 23) and colonoscopic procedure (using 10 of the same dogs) was performed. Approximately 5 min of representative endoscopic video recording, including the insertion phase of obtaining mucosal biopsies of each organ(s), was evaluated. Two gastroenterologists (JES, AEJ), experienced in the examination of small and large intestinal mucosa in dogs with chronic enteropathies, then reviewed these clips and independently scored the presence/absence and severity of endoscopic mucosal abnormalities using criteria derived from the development phase of the study. Endoscopic data collection and interpretation Endoscopic procedures were performed using a commercial video endoscope (Olympus GIF-160, Olympus Optical) with still images and video recordings of GI mucosa captured by the endoscopist. The file size of the downloaded images was approximately 100 kb, with a pixel array of 640 × 480 and 24-bit color. These still images were then arranged in a presentation for testing purposes (Microsoft Office
Table 1 List of mucosal appearances evaluated endoscopically (development phase). Appearance Normal mucosa Friability Granularity Erosion Hyperemia Lymphatic dilatation Mass
Definition No macroscopic lesions to mucosal surface Bleeding on contact with endoscope or biopsy forceps Alteration in the texture of the mucosa Superficial linear mucosal defect(s) with hemorrhage Gradations of mucosal redness (pale to red) Multifocal to diffuse white foci within the mucosa Abnormal growth of tissue projecting into lumen
PowerPoint, Mac 2011 14.3.9). The images were assessed by three experienced and five novice endoscopists for inflammatory activity. Experienced endoscopists were defined as individuals with advanced clinical training (rotating internship and residency trained in small animal internal medicine) and active and consistent operator participation in a minimum of 50 GI endoscopy procedures (primary clinician/ case responsibility) over the preceding 24 months. These operators were experienced and familiar with mucosal lesions as identified using GI endoscopy. Neither JES nor AEJ were included in this experienced operator group. Novice endoscopists had minimal endoscopic training, lacked consistent endoscopic operator experience, and had participated in less than five procedures over the same 24-month period. Images were randomized using a web-based randomization program1 and assessed independently by each endoscopist for mucosal appearance as originally determined by JES and AEJ. Neither the clinical data nor the date on which the image was taken was made known to the endoscopists. The endoscopic parameters evaluated for the stomach included: granularity (n = 6), friability (n = 6), erosions (n = 8), hyperemia (n = 3), normal pre-biopsy mucosa (n = 3), and normal post-biopsy mucosa (n = 4). Parameters for the small intestine included: granularity (n = 6), friability (n = 6), erosions (n = 7), lymphatic dilation (n = 5), hyperemia (n = 5), normal pre-biopsy mucosa (n = 4), and normal post-biopsy mucosa (n = 2). For the colon, the following parameters were evaluated: granularity (n = 7), friability (n = 6), erosions (n = 7), mass (n = 2), normal pre-biopsy mucosa (n = 5), and normal post-biopsy mucosa (n = 3; Fig. 1). Written definitions of each endoscopic parameter were made available to all endoscopists prior to still image testing (Table 1). If an individual image was interpreted as having more than one mucosal abnormality, the endoscopist was asked to identify the predominant lesion. For the validation phase, the results of the archived image assessment were validated using video clips on a test sample of dogs with IBD. These dogs were different from those used in the development phase. Five-minute video streams most representative of an endoscopic procedure were evaluated by a pair of experienced endoscopists (JES and AEJ) using both quantitative (i.e. 0–2 scoring based on the presence/extent of abnormal mucosal appearance) and qualitative (i.e. scoring based only on the presence of abnormal mucosal appearances) indices of endoscopic activity. For all dogs with IBD, the insertion phase (with mucosal biopsy) of the endoscopic procedure was used to produce video clips for scoring. This concept was important, since endoscopic disease activity assessment during insertion vs. withdrawal could likely affect interpretation of friability but not erosions, enhanced granularity, lymphatic dilatation or mass lesions (Samuel et al., 2013).
1
See: http://www.randomizer.org (accessed 24 December 2014).
Please cite this article in press as: J.E. Slovak, et al., Development and validation of an endoscopic activity score for canine inflammatory bowel disease, The Veterinary Journal (2015), doi: 10.1016/j.tvjl.2014.12.030
ARTICLE IN PRESS J.E. Slovak et al./The Veterinary Journal ■■ (2015) ■■–■■
Table 2 Quantitative assessment of mucosal appearance for endoscopic activity (validation phase). Appearance Friability
Granularity
Erosions
Lymphatic dilatationa
Score
Description
0 1 2 0 1 2 0 1 2 0 1 2
Absent Mild bleeding to touch Marked bleeding to touch Normal mucosal texture Mucosal texture increased Marked mucosal texture Absent Only few erosions Diffuse erosions Absent Focal–multifocal white foci Diffuse white foci
Maximum gastroscopy score, 6; maximum enteroscopy score, 8; maximum colonoscopy score, 6. a Defined only during enteroscopy.
Table 3 Qualitative assessment of mucosal appearance for endoscopic activity (validation phase). Appearance Friability Granularity Erosions Lymphatic dilatationa
Scoreb 0 1 0 1 0 1 0 1
Description Absent Present Absent Present Absent Present Absent Present
a
Defined only during enteroscopy. Maximum gastroscopy score, 3; maximum enteroscopy score, 4; maximum colonoscopy score, 3. b
Analysis of video clips in the validation phase was performed using both quantitative and qualitative grading criteria. For quantitative scoring, a 0–2 point system (0, absent; 1, mild-moderate; 2, moderate-severe) was used to score endoscopic activity in each organ. In this regard, the stomach was scored for granularity, friability, and erosions with a maximum lesion score of six points; the duodenum was scored for granularity, friability, erosions, and lymphatic dilation with a maximum lesion score of eight points; and the colon was scored for granularity, friability, and erosions with a maximum lesion score of six points (Table 2). Additionally, qualitative scoring, based only on the presence of abnormal mucosal appearances (and not their extent or severity), was performed on each video clip (Table 3). Data analysis Data pertaining to the development phase was collected from each operator using pre-designed (Microsoft Office Excel, Mac 2011 14.3.9) spread sheets for statistical analysis. Fleiss Kappa coefficients were calculated to assess inter-observer agreement among multiple raters within experienced and novice groups and then tested against the null value 0 using a freeware program (the irr package in R, R Core Team).2 Fleiss Kappa interpretation is similar to that of Cohen’s Kappa but allows for comparison between more than two evaluators. A mixed effects logistic regression model was then used to analyze agreement on descriptive appearance among operators vs. the ‘standard’ established by JES/AEJ for the stomach, duodenum, and colon using commercially available software (Glimmix procedure, SAS Institute, version SAS 9.4). The response variable was agreement with the ‘standard’ established by JES/AEJ, i.e. 1 for agreed score and 0 for a disagreed score. The fixed effect was experience level, and the random effect was the sample. For the validation phase, the video clips were scored by two experienced endoscopists (JES/AEJ) and results were recorded using the 0–2 point assessment system. The inter-observer agreement between both endoscopists was assessed using Cohen’s Kappa coefficient. Additionally, the agreement between endoscopists was compared using a lesion present (1) or absent score (0) qualification. The difference in the distribution of the endoscopy scores between experienced operators was assessed using the test of symmetry. A P value < 0.05 was considered significant for all statistical evaluations.
2
See: http://www.R-project.org (accessed 24 December 2014).
3
Results Development phase involving mucosal appearances of the stomach, duodenum, and colon The baseline characteristics for dogs enrolled in the development and validation phases are presented in Table 4. Thirty endoscopic images of normal and inflamed gastric mucosa captured from 27 dogs were reviewed. The dogs ranged in age from 1 to 13 years (mean, 8 years). There were 14 spayed females, eight neutered males, four intact males, and one intact female included and the following breeds were represented: two Weimeraners, two American Eskimos, two mixed breeds, two Vizslas, two Labrador retrievers, two West Highland white terriers, and one each of Pomeranian, Bernese Mountain dog, Wheaten terrier, Sheltie, Chow Chow, Shar Pei, Pembroke Welsh corgi, English bulldog, Cavalier King Charles spaniel, Basset hound, Samoyed, Boston terrier, Miniature schnauzer, Gordon setter, and Akita. All dogs exhibited chronic GI signs and fulfilled rigorous diagnostic criteria consistent with a diagnosis of idiopathic IBD (Jergens et al., 1992, 2003). Evaluation of Fleiss Kappa statistics testing for agreement within experienced and novice operator groups for gastric mucosal evaluation was K = 0.04 and K = 0.15, respectively, indicating very little agreement. There was no significant difference (P = 0.10) in inter-observer agreement between experienced and novice operators for gastric mucosal interpretation. Thirty-five endoscopic images of normal and inflamed duodenal mucosa captured from 25 dogs were viewed and assessed by endoscopists. The dogs ranged in age from 1 to 11 years (mean, 6.8 years). This group of dogs with IBD consisted of 11 spayed female and 14 neutered male dogs of the following breeds: four West Highland White terriers, three Golden retrievers, two mixed breed dogs, two Boxers, two Labrador retrievers, two Shih Tzus, two Yorkshire terriers and one each of Wheaten terrier, German shepherd dog, Vizsla, English bulldog, Cocker spaniel, Gordon setter, Beagle, and Miniature poodle. Fleiss Kappa statistics performed for duodenal mucosal appearance agreement among experienced operators was K < 0.01 and for novice endoscopists was K < 0.02. Inter-observer agreement in duodenal assessment between novice and experienced endoscopists were significantly different (P < 0.05). Although agreement within the groups was poor, when compared with the reference standard, the experienced endoscopists performed better than novice operators. Thirty endoscopic images of normal and inflamed colonic mucosa obtained from 23 dogs were scored for endoscopic appearance by both operator groups. All dogs undergoing colonoscopy were diagnosed with IBD, as described previously. The dogs with IBD in this group ranged in age from 1 to 11 years (mean, 6.8 years). There were 11 spayed females, 11 neutered males, and one intact male representing the following breeds: four Boxers, three Shih Tzus, two German shorthaired pointers, two Labrador retrievers, and one each of West Highland white terrier, Miniature Dachshund, Brittany Spaniel, Basset hound, Mixed breed, Weimeraner, German wirehaired pointer, Australian shepherd, Rottweiler, Siberian Husky, Beagle, Pembroke Welsh corgi, and Bichon Frise. Fleiss Kappa statistics testing performed for assessment of inter-observer agreement among experienced and trainee endoscopists for colonic evaluation was K = 0.22 and K = 0.15, respectively. There was no significant difference between experienced and novice endoscopists in their interpretation of colonic mucosal appearance. Validation phase and design of an endoscopic activity score for canine IBD Twenty-three dogs with a clinical diagnosis of IBD were prospectively enrolled in a separate study and their archived images were utilized in the validation phase of the study. They ranged in
Please cite this article in press as: J.E. Slovak, et al., Development and validation of an endoscopic activity score for canine inflammatory bowel disease, The Veterinary Journal (2015), doi: 10.1016/j.tvjl.2014.12.030
ARTICLE IN PRESS J.E. Slovak et al./The Veterinary Journal ■■ (2015) ■■–■■
4
Table 4 Clinical characteristics of dogs with inflammatory bowel disease (IBD) enrolled in the study phases. Parameter n Age at diagnosis (median/range, years) Sex (female/male) Disease duration (median/range, months) CIBDAI at diagnosis (median/range) Histopathologic inflammationa Normal-mild Moderate-severe
Development phase (stomach)
Development phase (duodenum)
Development phase (colon)
Validation phase (all GI organs)
27
25
23
8/1–13 15/12 7/0.25–15 5/1–10
6.8/1–11 11/14 18/1–36 5.9/3–10
6.8/1–11 11/12 12/0.5–24 6/2–10
S/D – 23 C – 10 6.7/1–14 14/9 12/0.5–24 8.5b/3–14
67% 33%
24% 76%
43% 57%
36% 64%
GI, gastrointestinal; S, stomach; D, duodenum; C, colon; CIBDAI, canine IBD activity index. a Overall (composite) histology score across one or more alimentary organs. CIBDAI score was unavailable in four dogs undergoing gastroscopy; disease duration (>3 weeks) unavailable in six dogs undergoing gastroscopy; CIBDAI score and disease duration (>3 weeks) unavailable in one dog undergoing enteroscopy; CIBDAI score unavailable in two dogs undergoing colonoscopy; disease duration (>3 weeks) unavailable in three dogs undergoing colonoscopy. b CIBDAI score prospectively scored for validation phase.
age from 1 to 14 years (mean, 6.7 years) and included 14 spayed females, eight neutered males, and one intact male dog. Their breeds included: four West Highland white Terriers, three Mixed breeds, two Shih Tzu’s, two German shepherd dogs, two Labrador retrievers, two Boxers, and one each of German shorthaired pointer, Gordon Setter, Boston terrier, Scottish terrier, Vizsla, and Collie. All 23 dogs had gastroduodenoscopic evaluations performed and 10 of these same dogs also underwent colonoscopy. For quantitative scoring (0, absent; 1, mild-to-moderate abnormal appearance; 2, moderate-to-severe abnormal appearance) of video clips between the two experienced endoscopists, there was almost perfect agreement (K = 0.87) between operators when total scores of the stomach were compared. Moreover, there was moderate agreement (K = 0.6) in total scores for duodenal mucosal assessment between experienced endoscopists and substantial agreement (K = 0.74) between operators for colonic mucosal assessment. When assessing individual endoscopic parameters for each organ using this quantitative scheme, there was substantial agreement in identification of gastric erosion (K = 0.84), granularity (K = 1) and friability (K = 0.8). Assessment of the small intestine showed substantial agreement in identification of mucosal erosion (K = 0.84) and moderate agreement for granularity (K = 0.66), friability (K = 0.67), and lymphatic dilation (K = 0.62). The assessment of colonic mucosal appearance showed substantial agreement for erosion (K = 1), granularity (K = 1), and friability (K = 1). Similarly, qualitative mucosal assessment (i.e. scoring based simply on the presence of an abnormal appearance) showed almost perfect agreement between experienced endoscopists when comparing cumulative scores of the stomach (K = 0.87). There was moderate agreement in the cumulative scores for duodenal mucosal assessment between experienced endoscopists (K = 0.58) and there was substantial agreement in cumulative scores between experienced endoscopists when performing colonic mucosal assessment (K = 0.8). In the validation phase of this study, the quantitative and qualitative scores of mucosal appearances were virtually identical (Table 5) and the presence of abnormal mucosal appearances/ organ was more quickly and objectively assessed by experienced operators.
tectable histopathologic lesions, especially inflammatory and neoplastic disorders involving the GI mucosa (Roth et al., 1990). In this large clinical trial, flexible GI endoscopy was used to examine the stomach, duodenum, and colon of 58 dogs with IBD and to acquire mucosal biopsies for histopathologic characterization. Most dogs in this earlier report had abnormal endoscopic appearances, interpreted as erythema, friability, enhanced granularity, and ulceration or erosion associated with predominantly lymphocyticplasmacytic mucosal inflammation. More recent trials have evaluated the association between endoscopic scores of the duodenum and colon and long-term outcome in dogs with chronic enteropathies (Allenspach et al., 2007) or have utilized endoscopic evaluation to assess mucosal healing pre- vs. post-treatment in dogs with lymphocytic-plasmacytic gastroenteritis (Garcia-Sancho et al., 2007). Given the ubiquitous nature of GI endoscopy in clinical practice, it is surprising that no studies have investigated the inter-observer variability in the endoscopic assessment of mucosal lesions using a validated scoring system. The current study investigated the inter-observer agreement in the endoscopic assessment of mucosal appearance of the GI tract to design and validate a simple and easy-to-use endoscopic scoring system for dogs with IBD, using a combination of retrospectively archived still images and prospectively acquired video clips. In the validation phase, we found moderate-to-substantial inter-observer agreement and reproducibility among experienced endoscopists for increased granularity, friability, erosions, and lymphatic dilatation (duodenal mucosa) involving the stomach, duodenum, and colon. A composite score that incorporates only the presence/absence of these four mucosal variables, and not their magnitude or the extent of their severity, is proposed for use in the endoscopic assessment of canine IBD. The first step in development of this endoscopic score was to identify which mucosal parameters in the stomach, duodenum, and colon would be evaluated by novice and experienced endoscopists
Discussion
Stomach Duodenum Colon
Endoscopic evaluation plays an important role in the diagnosis and management of dogs with chronic enteropathy. Abnormal endoscopic observations in dogs with histories of regurgitation, vomiting, and/or diarrhea have been previously associated with de-
Table 5 Kappa statistic (K) valuesa for validation tests for inter-observer agreement (IOA) used in the endoscopic activity score. Organ
Quantitative score
Qualitative score
0.87 0.60 0.74
0.87 0.58 0.80
a K < 0 represents no agreement, K = 0–0.20 represents slight agreement, K = 0.21– 0.40 represents fair agreement, K = 0.41–0.60 represents moderate agreement, K = 0.61–0.80 represents substantial agreement and K = 0.81–1 represents almost perfect agreement.
Please cite this article in press as: J.E. Slovak, et al., Development and validation of an endoscopic activity score for canine inflammatory bowel disease, The Veterinary Journal (2015), doi: 10.1016/j.tvjl.2014.12.030
ARTICLE IN PRESS J.E. Slovak et al./The Veterinary Journal ■■ (2015) ■■–■■
for inter-observer agreement. One approach was to identify the most common mucosal lesions (i.e. hyperemia, friability, enhanced granularity, erosions, lymphatic dilatation, and/or mass) in dogs with IBD based on earlier studies and the endoscopic experiences of AEJ and others over many years (Roth et al., 1990; Jergens et al., 1992, 2003, 2010; Allenspach et al., 2007; Slovak et al., 2014). These same general endoscopic parameters have been shown to be associated with clinical GI signs (Jergens et al., 2003; Allenspach et al., 2007; Garcia-Sancho et al., 2007; Heilmann et al., 2014), evidence of mucosal healing (Allenspach et al., 2007; Garcia-Sancho et al., 2007), histopathologic interpretation (Roth et al., 1990; Jergens et al., 1992; Allenspach et al., 2007; Larson et al., 2012), and prognostic outcome (Allenspach et al., 2007) in separate clinical trials. We observed, as described in the human literature, that observer variation for graded characteristics (i.e. mucosal hyperemia – is it pale, pink, or red?) was quite high, while that for discontinuous variables (i.e. presence or absence of erosions) was generally low (Baron et al., 1964). A discrepancy between mucosal hyperemia and the presence of histopathologic abnormalities in dogs with chronic enteropathy has been previously reported (Roth et al., 1990). Similar to Roth et al. (1990), we have also shown that operator experience plays an important role in duodenal endoscopic assessment, with novice endoscopists less likely to identify mucosal lesions or misinterpret normal vs. abnormal mucosa (Slovak et al., 2014). Still images were used exclusively in the development phase of this study since: (1) they are widely used in almost all clinical practices and are often retained as a component of the dog’s medical record, and (2) they are frequently used to communicate results of the diagnostic procedure to other veterinarians involved in case management and to clients (Washabau et al., 2010). Additionally, a manageable number of still images (vs. video clips) could be more easily evaluated by participating clinicians, which assured good compliance in the trial. Accurate assessment of endoscopic activity from archived still images of humans with UC has been previously reported (Osada et al., 2010). In the second part of the study, these same endoscopic parameters were tested in the stomach and colon, regardless of the level of reproducibility found in the previous inter-observer agreement study (Slovak et al., 2014). As noted in the results, there were no significant differences in inter-observer agreement between novice and experienced operators regarding either gastric or colonic mucosal interpretation. This observation that the endoscopic parameters chosen for evaluation showed no significant difference in reproducibility between operator groups was both interesting and reassuring. In this regard, these data suggested that: (1) different endoscopists could reliably recognize specific endoscopic appearances in the stomach and colon that they felt were important indicators of mucosal inflammation and (2) those abnormal endoscopic appearances previously identified by operators using still images could be reliably tested in prospective clinical trials using video clips. Our results demonstrated consistently good inter-observer agreement among the experienced endoscopists using defined mucosal appearances to define endoscopic activity in two ways – quantitatively and qualitatively. For the assessment of overall endoscopic severity, our study showed that the strength of agreement for quantitative vs. qualitative scoring was virtually identical (Table 5). When assessing individual endoscopic parameters quantitatively by organ, we observed the lowest inter-observer agreement for duodenal scoring for each descriptor, although at least moderate agreement was documented between experienced operators. A similar observation was made when using the qualitative duodenal scoring criteria, where the inter-observer agreement was reduced approximately 15% in comparison with gastric or colonic mucosal assessment. The reasons for the lower duodenal inter-observer agreement are unclear, but could possibly be related to differences in the experience level of the two operators in the evaluation of small in-
5
testinal mucosa i.e. extensive (30 years) vs. moderate (3 years) clinical experience in performance of advanced GI endoscopic procedures. Interestingly, even among expert human endoscopists, the extent of complete or partial agreement for endoscopic activity of UC (based on score 0–3) can range from 21% to 46% of patients (Walsh et al., 2009). The endoscopic activity score for canine IBD was broadly constructed from the observations of the six established activity indices used in human IBD, including the Baron (Baron et al., 1964), Blackstone (Osada et al., 2010), Schroeder (Schroeder et al., 1987; Osada et al., 2010), and Matts (Matts, 1961), scores for UC, which are all based on minor modifications of the original Baron score, the Crohn’s disease endoscopic index of severity (CDEIS; Mary and Modigliani, 1989) and simple endoscopic score for Crohn’s disease (SES-CD; Daperno et al., 2004). Similar to these human indices, the proposed canine score uses the mucosal appearance of erosions, friability, and enhanced granularity to define endoscopic activity with observations of lymphatic dilatation and mass lesions limited to duodenal and colonic mucosal assessment, respectively. The infrequent occurrence of colonic mass lesions (vs. other colonic mucosal indices) in dogs with IBD provided the rationale for excluding this parameter from further evaluation in the validation phase of this study. Moreover, the proposed index was associated with relatively high inter-observer agreement and reproducibility for the experienced endoscopists in the validation phase. It is noteworthy that we provide new information on the inter-observer agreement for gastric and colonic mucosal assessment in dogs with IBD, which has not been previously reported. Taken together, these findings suggest that an endoscopic activity score for canine IBD, developed and validated using still images and video clips, would be of practical value to all clinicians using current endoscopic technology. There are some potential limitations in our study. First, we utilized only a single referral center for this clinical trial. Whether our results for inter-observer agreement across different study centers, endoscopic interpretation by non-specialist clinicians, or endoscopic evaluation of dogs with different enteropathies are generalizable to other veterinary practices was not assessed. Additionally, our paper discusses inter-observer agreement, comparisons within the group vs. intra-observer agreement, and comparisons of results obtained from individual clinicians. Future studies should include both inter- and intra-observer comparisons. Most dogs enrolled in our study had active moderate-to-severe clinical disease activity based on CIBDAI scores. We did not perform subset analyses to determine whether these same endoscopic variables would have comparable inter-observer agreement in dogs presenting with quiescent disease or mild clinical severity. Further, we did not report on the abnormal appearance of the ileal mucosa in dogs with IBD. Endoscopic examination of the ileum was not performed in all dogs undergoing colonoscopy due to technical difficulties associated with general anesthesia and/or the inability to traverse the ileocolic sphincter (i.e. incomplete bowel cleansing). The relative importance of these potential problems is likely to be clarified by performing large multi-center studies with broad application of the endoscopic score in research and in clinical decision-making. Conclusions The proposed qualitative endoscopic activity score for canine IBD meets the following requirements as described for human endoscopic indices: (1) it is well correlated with the endoscopist’s overall assessment of abnormal mucosal appearance; (2) the quality of the assessment is consistent when evaluated in a set of patients different than that used for its development; (3) it incorporates mucosal parameters intuitively considered as important by experienced endoscopists; (4) it incorporates mucosal parameters previously shown to be reproducibly collected, and is itself reproducible; and
Please cite this article in press as: J.E. Slovak, et al., Development and validation of an endoscopic activity score for canine inflammatory bowel disease, The Veterinary Journal (2015), doi: 10.1016/j.tvjl.2014.12.030
ARTICLE IN PRESS J.E. Slovak et al./The Veterinary Journal ■■ (2015) ■■–■■
6
(5) it is simple to calculate. Importantly, this validated endoscopic index should be correlated with clinical indices and laboratory biomarkers of inflammation in future studies (Mary and Modigliani, 1989). Conflict of interest statement None of the authors of this paper has a financial or personal relationship with other people or organizations that could inappropriately influence or bias the content of this paper. Acknowledgements The authors thank Amy Hodnefield and Carrie Schwake for technical assistance. Portions of this manuscript were presented as research abstracts at the 2013 ACVIM Forum, Seattle, WA, and the 2013 ECVIM Congress, Liverpool, England. References Allenspach, K., 2011. Clinical immunology and immunopathology of the canine and feline intestine. Veterinary Clinics of North America. Small Animal Practice 41, 345–360. Allenspach, K., Wieland, B., Grone, A., Gaschen, F., 2007. Chronic enteropathies in dogs: Evaluation of risk factors for negative outcome. Journal of Veterinary Internal Medicine 21, 700–708. Baron, J.H., Connell, A.M., Lennard-Jones, J.E., 1964. Variation between observers in describing mucosal appearances in proctocolitis. British Medical Journal 1, 89–92. Daperno, M., D’Haens, G., Van Assche, G., Baert, F., Bulois, P., Maunoury, V., Sostegni, R., Rocca, R., Pera, A., Gevers, A., et al., 2004. Development and validation of a new, simplified endoscopic activity score for Crohn’s disease: The SES-CD. GI Endoscopy 60, 505–512. Garcia-Sancho, M., Rodriguez-Franco, F., Sianz, A., Mancho, C., Rodriquez, A., 2007. Evaluation of clinical, macroscopic, and histopathologic response to treatment in non-hypoproteinemic dogs with lymphocytic-plasmacytic enteritis. Journal of Veterinary Internal Medicine 21, 11–17. Heilmann, R.M., Grellet, A., Allenspach, K., Lecoindre, P., Day, M.J., Preistnall, S.L., Toresson, L., Procoli, F., Grutzner, N., Suchodolski, J.S., et al., 2014. Association between fecal S100A12 concentration and histologic, endoscopic, and clinical disease severity in dogs with idiopathic inflammatory bowel disease. Veterinary Immunology Immunopathology 158, 156–166. Jergens, A.E., Simpson, K.W., 2012. Inflammatory bowel disease in veterinary medicine. Frontier Bioscience 4, 1404–1419. Jergens, A.E., Moore, F.M., Haynes, J.S., Miles, K.G., 1992. Idiopathic inflammatory bowel disease in dogs and cats: 84 cases (1987–1990). Journal of the American Veterinary Medical Association 201, 1603–1608. Jergens, A.E., Schreiner, A.C., Frank, D.E., Niyo, Y., Ahrens, F.E., Eckersall, P.D., Benson, T.J., Evans, R., Pressel, M., Ackermann, M.L., et al., 2003. A scoring index for disease activity in canine inflammatory bowel disease. Journal of Veterinary Internal Medicine 17, 291–297. Jergens, A.E., Crandell, J., Morrison, J.A., Deitz, K., Pressel, Ackermann, M., Suchodolski, J.S., Steiner, J.M., Evans, R., 2010. Comparison of oral prednisone and prednisone
combined with metronidazole for induction therapy of canine inflammatory bowel disease: A randomized-controlled trial. Journal of Veterinary Internal Medicine 24, 269–277. Larson, R.N., Ginn, J.A., Bell, C.M., Davis, M.J., Foy, D.S., 2012. Duodenal endoscopic findings and histopathologic confirmation of intestinal lymphangiectasia in dogs. Journal of Veterinary Internal Medcine 26, 1087–1092. Lichtiger, S., Present, D.H., Kornbluth, A., Glernt, I., Bauer, J., Galler, G., Michelassi, F., Hanauer, S., 1994. Cyclosporine in severe ulcerative colitis refractory to steroid therapy. New England Journal of Medicine 330, 1841–1845. Mary, J.Y., Modigliani, R., 1989. Development and validation of an endoscopic index of the severity for Crohn’s disease: A prospective multicenter study. Gut 30, 983–989. Matts, S.G., 1961. The value of rectal biopsy in the diagnosis of ulcerative colitis. Quarterly Journal of Medicine 30, 393–407. Osada, T., Ohkusa, T., Yokoyama, T., Shibuya, T., Sakamota, N., Beppu, K., Nagahara, A., Otaka, M., Ogihara, T., Watanabe, S., 2010. Comparison of several activity indices for the evaluation of endoscopic activity in UC: Inter-and intra-observer consistency. Inflammatory Bowel Disease 16, 192–198. Powell-Tuck, J., Bown, R.L., Lennard-Jones, J.E., 1978. A comparison of oral prednisolone given as single or multiple daily doses for active proctocolitis. Scandinavian Journal of Gastroenterology 13, 833–837. Roth, L., Leib, M.S., Davenport, D.J., Monroe, W.E., 1990. Comparisons between endoscopic and histologic evaluation of the gastrointestinal tract in dogs and cats: 75 cases (1984–1987). Journal of the American Veterinary Medical Association 4, 635–638. Samuel, S., Bruining, D.H., Loftus, E.V., Jr., Thia, K.T., Schroeder, K.W., Tremaine, W.J., Faubion, W.A., Kane, S.V., Pardi, D.S., De Groen, P.C., et al., 2013. Validation of the ulcerative colitis colonoscopic index of severity and its correlation with disease activity measures. Clinical Gastroenterology and Hepatology 11, 49–54. Sartor, R.B., 2006. Mechanisms of disease: Pathogenesis of Crohn’s disease and ulcerative colitis. National Clinical Practical Gastroenterology and Hepatology 3, 390–407. Schroeder, K.W., Temaine, W.J., Ilstrup, D.M., 1987. Coated oral 5-aminosalicylic acid therapy for mildly to moderately active ulcerative colitis, a randomized study. New England Journal of Medicine 317, 1625–1629. Seo, M., Okada, M., Yao, T., Ueki, M., Arima, S., Okumura, M., 1992. An index of disease activity in patients with ulcerative colitis. American Journal of Gastroenterology 87, 971–976. Slovak, J.E., Wang, C., Morrison, J.A., Deitz, K.L., LeVine, D.N., Otoni, C., King, R.R., Gerber, L.E., Hanson, K.R., Lundberg, A.P., et al., 2014. Endoscopic assessment of the duodenum in dogs with inflammatory bowel disease. Journal of Veterinary Internal Medicine 28, 1442–1446. Truelove, S.C., Witts, L.J., 1955. Coritsone in ulcerative colitis; final report on therapeutic trial. British Medical Journal 2, 1041–1048. Walmsley, R.S., Ayres, R.C., Pounder, R.E., Allan, R.N., 1998. A simple clinical colitis activity index. Gut 43, 29–32. Walsh, A.J., Brain, O., Keshav, S., 2009. How variable is the Mayo score between observers and might this affect trial recruitment or outcome? Gastroenterology 136 (Suppl. 1), A677 (Abstract). WSAVA International Gastrointestinal Group: Washabau, R.J., Day, M.J., Willard, M.D., Hall, E.J., Jergens, A.E., Mansell, J., Minami, T., Bilzer, T.W., 2010. Endoscopic, biopsy, and histopathologic guidelines for the evaluation of gastrointestinal inflammation in companion animals. Journal of Veterinary Internal Medicine 24, 10–26. Xavier, R.J., Podolsky, D.K., 2007. Unravelling the pathogenesis of inflammatory bowel disease. Nature 448, 427–434. Zoran, D.L., 2001. Gastroduodenoscopy in the dog and cat. Veterinary Clinics of North America. Small Animal Practice 31, 631–656.
Please cite this article in press as: J.E. Slovak, et al., Development and validation of an endoscopic activity score for canine inflammatory bowel disease, The Veterinary Journal (2015), doi: 10.1016/j.tvjl.2014.12.030
Contents lists available at ScienceDirect
The Veterinary Journal j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / l o c a t e / t v j l
Development and validation of an endoscopic activity score for canine inflammatory bowel disease J.E. Slovak a,*, C. Wang b, Y. Sun b, C. Otoni c, J. Morrison d, K. Deitz d, D. LeVine d, A.E. Jergens d a
Department of Veterinary Clinical Science, College of Veterinary Medicine Washington State University, PO Box 646610, Pullman, WA 99164, USA Department of Veterinary Diagnostic and Production Animal Medicine, College of Veterinary Medicine Iowa State University, 1600 S. 16th St, Ames, IA 50011, USA c VCA Arboretum View Animal Hospital, 2551 Warrenville Rd, Downers Grove, IL 60515, USA d Department of Veterinary Clinical Science, College of Veterinary Medicine Iowa State University, 1600 S. 16th St, Ames, IA 50011, USA b
A R T I C L E
I N F O
Article history: Accepted 27 December 2014 Keywords: Gastrointestinal endoscopy IBD Inter-observer agreement CIBDAI Mucosal appearance
A B S T R A C T
The aim of this study was to develop and prospectively validate a simple endoscopic score of disease activity for dogs with inflammatory bowel disease (IBD). Archived endoscopic still images and video recordings of gastric, duodenal, and colonic endoscopic examinations were displayed to novice and experienced endoscopists for assessment of inflammatory activity using established descriptions. The mucosal appearances evaluated were normal tissue, erosions, friability, increased granularity, lymphangiectasia (duodenum), and mass (colon). Fleiss and Cohen’s Kappa statistics were used to estimate the inter-observer agreement of the index. For duodenal assessment, there were statistically significant (P < 0.05) differences in inter-observer agreement, with experienced endoscopists performing better than novice endoscopists in the accurate identification of mucosal appearance of the duodenum. In contrast, there was no significant difference between novice and experienced endoscopists in their interpretation of gastric (P = 0.10) and colonic (P = 1.0) mucosal appearances. Validation studies using endoscopic video clips to assess the same endoscopic parameters by quantitative (lesion number and severity) and qualitative (presence of mucosal lesions) methods showed moderate-to-substantial agreement between experienced endoscopists. Based on the observations that the quantitative and qualitative scores of mucosal appearances are virtually identical, and that qualitative assessment was performed more quickly and objectively than quantitative assessment, we propose a simple endoscopic activity score based on qualitative criteria alone in dogs with inflammatory bowel disease. © 2014 Elsevier Ltd. All rights reserved.
Introduction The pathogenesis of canine inflammatory bowel disease (IBD) probably involves interplay between the mucosal immune system and the intestinal microbiota, similar to human IBD (i.e. Crohn’s disease, CD, and ulcerative colitis, UC; Allenspach, 2011; Sartor, 2006; Xavier and Podolsky, 2007). Diagnostic tests including dietary trials, routine hematology, parasitic and bacteriologic fecal analyses, radiographic imaging, and histopathologic examination of intestinal biopsy specimens serve to eliminate other causes of chronic enteropathy. Gastrointestinal (GI) endoscopy is a useful and relatively non-invasive technique to diagnose IBD. It allows direct visualization of the mucosa for the acquisition of targeted biopsies to evaluate
* Corresponding author. Tel.: +1 507 3178960. E-mail address: [email protected] (J.E. Slovak).
the severity and extent of intestinal inflammation (Roth et al., 1990; Zoran, 2001). Determination of an inflammatory state is critical for defining disease activity and for tailoring IBD therapy. Endoscopic measures of mucosal inflammation in human IBD have been in use for over 40 years; however, no standardized model has been established (Truelove and Witts, 1955; Powell-Tuck et al., 1978; Seo et al., 1992; Lichtiger et al., 1994; Walmsley et al., 1998). Most indices for CD and UC are based on observations of mucosal erythema, increased granularity, vascular pattern, spontaneous bleeding, and mucosal damage (mucus, fibrin, exudates, erosions and ulcer), using simple scoring systems to define inflammatory activity. To date, no similar validated endoscopic score exists in veterinary medicine, and there is only limited trial data (Allenspach et al., 2007; Garcia-Sancho et al., 2007) evaluating the duodenal appearance of canine IBD, despite the suitability of the dog as a spontaneous animal model for intestinal inflammation (Jergens and Simpson, 2012). We recently reported that the inter-observer agreement for duodenal
http://dx.doi.org/10.1016/j.tvjl.2014.12.030 1090-0233/© 2014 Elsevier Ltd. All rights reserved.
Please cite this article in press as: J.E. Slovak, et al., Development and validation of an endoscopic activity score for canine inflammatory bowel disease, The Veterinary Journal (2015), doi: 10.1016/j.tvjl.2014.12.030
ARTICLE IN PRESS J.E. Slovak et al./The Veterinary Journal ■■ (2015) ■■–■■
2
Fig. 1. Representative still images used in the development phase of the endoscopic study. (A) normal stomach; (B) gastric erosions; (C) gastric friability; (D) gastric granularity; (E) normal duodenum; (F) duodenal erosions; (G) duodenal friability; (H) duodenal granularity; (I) duodenal lymphatic dilatation; (J) normal colon; (K) colonic erosions; (L) colonic friability; (M) colonic granularity; (N) colonic mass.
endoscopic assessment of canine IBD differed among trainee and experienced endoscopists on the basis of operator experience (Slovak et al., 2014). The aim of the present study was to develop and validate an endoscopic disease activity score for the mucosal appearance of the stomach, duodenum and colon of dogs with IBD. Materials and methods Study design The study was comprised of four parts. In the first part (development phase), the most relevant duodenal endoscopic variables (still images) in dogs with IBD were selected and scored for inter-observer agreement between novice and experienced endoscopists. For the second part of the development phase, an activity score based on endoscopic variables (still images) affecting the gastric and colonic mucosa was designed and assessed for inter-observer agreement among novice and experienced endoscopists. In the third part (the validation phase), representative video clips of approximately 5 min duration of gastroenteroscopic and/or colonoscopic examinations were assessed simultaneously but scored independently by two experienced endoscopists using the endoscopic criteria derived from the development phase. In the fourth phase, a simplified score was proposed based on the validation test results of endoscopic activity observed in the stomach, duodenum, and colon of dogs with IBD. Endoscopic examinations The details of the initial duodenal examination tests are described elsewhere (Slovak et al., 2014). In brief, the duodenal endoscopic mucosal appearance obtained from 25 dogs diagnosed with IBD at a single study center was assessed for inter-observer agreement. For the development phase of the study, 27 dogs diagnosed with IBD from 2002 to 2012 underwent gastroscopic examination, and 23 dogs with IBD from 2002 to 2011 undergoing colonoscopy were evaluated at the same study center. Endoscopic mucosal lesions were assessed for inter-observer agreement. Archived endoscopic still images from individual GI procedures were retrieved from a computerized database and reviewed. A total of 30 gastric, 35 duodenal and 30 colonic still images were selected for operator evaluation. A canine IBD activity index (CIBDAI) score as described by Jergens et al. (2003) was assigned retrospectively during the development phase of the study. For each dog in the validation phase, a CIBDAI score was prospectively assigned and a video recording of the entire gastroduodenoscopic procedure (n = 23) and colonoscopic procedure (using 10 of the same dogs) was performed. Approximately 5 min of representative endoscopic video recording, including the insertion phase of obtaining mucosal biopsies of each organ(s), was evaluated. Two gastroenterologists (JES, AEJ), experienced in the examination of small and large intestinal mucosa in dogs with chronic enteropathies, then reviewed these clips and independently scored the presence/absence and severity of endoscopic mucosal abnormalities using criteria derived from the development phase of the study. Endoscopic data collection and interpretation Endoscopic procedures were performed using a commercial video endoscope (Olympus GIF-160, Olympus Optical) with still images and video recordings of GI mucosa captured by the endoscopist. The file size of the downloaded images was approximately 100 kb, with a pixel array of 640 × 480 and 24-bit color. These still images were then arranged in a presentation for testing purposes (Microsoft Office
Table 1 List of mucosal appearances evaluated endoscopically (development phase). Appearance Normal mucosa Friability Granularity Erosion Hyperemia Lymphatic dilatation Mass
Definition No macroscopic lesions to mucosal surface Bleeding on contact with endoscope or biopsy forceps Alteration in the texture of the mucosa Superficial linear mucosal defect(s) with hemorrhage Gradations of mucosal redness (pale to red) Multifocal to diffuse white foci within the mucosa Abnormal growth of tissue projecting into lumen
PowerPoint, Mac 2011 14.3.9). The images were assessed by three experienced and five novice endoscopists for inflammatory activity. Experienced endoscopists were defined as individuals with advanced clinical training (rotating internship and residency trained in small animal internal medicine) and active and consistent operator participation in a minimum of 50 GI endoscopy procedures (primary clinician/ case responsibility) over the preceding 24 months. These operators were experienced and familiar with mucosal lesions as identified using GI endoscopy. Neither JES nor AEJ were included in this experienced operator group. Novice endoscopists had minimal endoscopic training, lacked consistent endoscopic operator experience, and had participated in less than five procedures over the same 24-month period. Images were randomized using a web-based randomization program1 and assessed independently by each endoscopist for mucosal appearance as originally determined by JES and AEJ. Neither the clinical data nor the date on which the image was taken was made known to the endoscopists. The endoscopic parameters evaluated for the stomach included: granularity (n = 6), friability (n = 6), erosions (n = 8), hyperemia (n = 3), normal pre-biopsy mucosa (n = 3), and normal post-biopsy mucosa (n = 4). Parameters for the small intestine included: granularity (n = 6), friability (n = 6), erosions (n = 7), lymphatic dilation (n = 5), hyperemia (n = 5), normal pre-biopsy mucosa (n = 4), and normal post-biopsy mucosa (n = 2). For the colon, the following parameters were evaluated: granularity (n = 7), friability (n = 6), erosions (n = 7), mass (n = 2), normal pre-biopsy mucosa (n = 5), and normal post-biopsy mucosa (n = 3; Fig. 1). Written definitions of each endoscopic parameter were made available to all endoscopists prior to still image testing (Table 1). If an individual image was interpreted as having more than one mucosal abnormality, the endoscopist was asked to identify the predominant lesion. For the validation phase, the results of the archived image assessment were validated using video clips on a test sample of dogs with IBD. These dogs were different from those used in the development phase. Five-minute video streams most representative of an endoscopic procedure were evaluated by a pair of experienced endoscopists (JES and AEJ) using both quantitative (i.e. 0–2 scoring based on the presence/extent of abnormal mucosal appearance) and qualitative (i.e. scoring based only on the presence of abnormal mucosal appearances) indices of endoscopic activity. For all dogs with IBD, the insertion phase (with mucosal biopsy) of the endoscopic procedure was used to produce video clips for scoring. This concept was important, since endoscopic disease activity assessment during insertion vs. withdrawal could likely affect interpretation of friability but not erosions, enhanced granularity, lymphatic dilatation or mass lesions (Samuel et al., 2013).
1
See: http://www.randomizer.org (accessed 24 December 2014).
Please cite this article in press as: J.E. Slovak, et al., Development and validation of an endoscopic activity score for canine inflammatory bowel disease, The Veterinary Journal (2015), doi: 10.1016/j.tvjl.2014.12.030
ARTICLE IN PRESS J.E. Slovak et al./The Veterinary Journal ■■ (2015) ■■–■■
Table 2 Quantitative assessment of mucosal appearance for endoscopic activity (validation phase). Appearance Friability
Granularity
Erosions
Lymphatic dilatationa
Score
Description
0 1 2 0 1 2 0 1 2 0 1 2
Absent Mild bleeding to touch Marked bleeding to touch Normal mucosal texture Mucosal texture increased Marked mucosal texture Absent Only few erosions Diffuse erosions Absent Focal–multifocal white foci Diffuse white foci
Maximum gastroscopy score, 6; maximum enteroscopy score, 8; maximum colonoscopy score, 6. a Defined only during enteroscopy.
Table 3 Qualitative assessment of mucosal appearance for endoscopic activity (validation phase). Appearance Friability Granularity Erosions Lymphatic dilatationa
Scoreb 0 1 0 1 0 1 0 1
Description Absent Present Absent Present Absent Present Absent Present
a
Defined only during enteroscopy. Maximum gastroscopy score, 3; maximum enteroscopy score, 4; maximum colonoscopy score, 3. b
Analysis of video clips in the validation phase was performed using both quantitative and qualitative grading criteria. For quantitative scoring, a 0–2 point system (0, absent; 1, mild-moderate; 2, moderate-severe) was used to score endoscopic activity in each organ. In this regard, the stomach was scored for granularity, friability, and erosions with a maximum lesion score of six points; the duodenum was scored for granularity, friability, erosions, and lymphatic dilation with a maximum lesion score of eight points; and the colon was scored for granularity, friability, and erosions with a maximum lesion score of six points (Table 2). Additionally, qualitative scoring, based only on the presence of abnormal mucosal appearances (and not their extent or severity), was performed on each video clip (Table 3). Data analysis Data pertaining to the development phase was collected from each operator using pre-designed (Microsoft Office Excel, Mac 2011 14.3.9) spread sheets for statistical analysis. Fleiss Kappa coefficients were calculated to assess inter-observer agreement among multiple raters within experienced and novice groups and then tested against the null value 0 using a freeware program (the irr package in R, R Core Team).2 Fleiss Kappa interpretation is similar to that of Cohen’s Kappa but allows for comparison between more than two evaluators. A mixed effects logistic regression model was then used to analyze agreement on descriptive appearance among operators vs. the ‘standard’ established by JES/AEJ for the stomach, duodenum, and colon using commercially available software (Glimmix procedure, SAS Institute, version SAS 9.4). The response variable was agreement with the ‘standard’ established by JES/AEJ, i.e. 1 for agreed score and 0 for a disagreed score. The fixed effect was experience level, and the random effect was the sample. For the validation phase, the video clips were scored by two experienced endoscopists (JES/AEJ) and results were recorded using the 0–2 point assessment system. The inter-observer agreement between both endoscopists was assessed using Cohen’s Kappa coefficient. Additionally, the agreement between endoscopists was compared using a lesion present (1) or absent score (0) qualification. The difference in the distribution of the endoscopy scores between experienced operators was assessed using the test of symmetry. A P value < 0.05 was considered significant for all statistical evaluations.
2
See: http://www.R-project.org (accessed 24 December 2014).
3
Results Development phase involving mucosal appearances of the stomach, duodenum, and colon The baseline characteristics for dogs enrolled in the development and validation phases are presented in Table 4. Thirty endoscopic images of normal and inflamed gastric mucosa captured from 27 dogs were reviewed. The dogs ranged in age from 1 to 13 years (mean, 8 years). There were 14 spayed females, eight neutered males, four intact males, and one intact female included and the following breeds were represented: two Weimeraners, two American Eskimos, two mixed breeds, two Vizslas, two Labrador retrievers, two West Highland white terriers, and one each of Pomeranian, Bernese Mountain dog, Wheaten terrier, Sheltie, Chow Chow, Shar Pei, Pembroke Welsh corgi, English bulldog, Cavalier King Charles spaniel, Basset hound, Samoyed, Boston terrier, Miniature schnauzer, Gordon setter, and Akita. All dogs exhibited chronic GI signs and fulfilled rigorous diagnostic criteria consistent with a diagnosis of idiopathic IBD (Jergens et al., 1992, 2003). Evaluation of Fleiss Kappa statistics testing for agreement within experienced and novice operator groups for gastric mucosal evaluation was K = 0.04 and K = 0.15, respectively, indicating very little agreement. There was no significant difference (P = 0.10) in inter-observer agreement between experienced and novice operators for gastric mucosal interpretation. Thirty-five endoscopic images of normal and inflamed duodenal mucosa captured from 25 dogs were viewed and assessed by endoscopists. The dogs ranged in age from 1 to 11 years (mean, 6.8 years). This group of dogs with IBD consisted of 11 spayed female and 14 neutered male dogs of the following breeds: four West Highland White terriers, three Golden retrievers, two mixed breed dogs, two Boxers, two Labrador retrievers, two Shih Tzus, two Yorkshire terriers and one each of Wheaten terrier, German shepherd dog, Vizsla, English bulldog, Cocker spaniel, Gordon setter, Beagle, and Miniature poodle. Fleiss Kappa statistics performed for duodenal mucosal appearance agreement among experienced operators was K < 0.01 and for novice endoscopists was K < 0.02. Inter-observer agreement in duodenal assessment between novice and experienced endoscopists were significantly different (P < 0.05). Although agreement within the groups was poor, when compared with the reference standard, the experienced endoscopists performed better than novice operators. Thirty endoscopic images of normal and inflamed colonic mucosa obtained from 23 dogs were scored for endoscopic appearance by both operator groups. All dogs undergoing colonoscopy were diagnosed with IBD, as described previously. The dogs with IBD in this group ranged in age from 1 to 11 years (mean, 6.8 years). There were 11 spayed females, 11 neutered males, and one intact male representing the following breeds: four Boxers, three Shih Tzus, two German shorthaired pointers, two Labrador retrievers, and one each of West Highland white terrier, Miniature Dachshund, Brittany Spaniel, Basset hound, Mixed breed, Weimeraner, German wirehaired pointer, Australian shepherd, Rottweiler, Siberian Husky, Beagle, Pembroke Welsh corgi, and Bichon Frise. Fleiss Kappa statistics testing performed for assessment of inter-observer agreement among experienced and trainee endoscopists for colonic evaluation was K = 0.22 and K = 0.15, respectively. There was no significant difference between experienced and novice endoscopists in their interpretation of colonic mucosal appearance. Validation phase and design of an endoscopic activity score for canine IBD Twenty-three dogs with a clinical diagnosis of IBD were prospectively enrolled in a separate study and their archived images were utilized in the validation phase of the study. They ranged in
Please cite this article in press as: J.E. Slovak, et al., Development and validation of an endoscopic activity score for canine inflammatory bowel disease, The Veterinary Journal (2015), doi: 10.1016/j.tvjl.2014.12.030
ARTICLE IN PRESS J.E. Slovak et al./The Veterinary Journal ■■ (2015) ■■–■■
4
Table 4 Clinical characteristics of dogs with inflammatory bowel disease (IBD) enrolled in the study phases. Parameter n Age at diagnosis (median/range, years) Sex (female/male) Disease duration (median/range, months) CIBDAI at diagnosis (median/range) Histopathologic inflammationa Normal-mild Moderate-severe
Development phase (stomach)
Development phase (duodenum)
Development phase (colon)
Validation phase (all GI organs)
27
25
23
8/1–13 15/12 7/0.25–15 5/1–10
6.8/1–11 11/14 18/1–36 5.9/3–10
6.8/1–11 11/12 12/0.5–24 6/2–10
S/D – 23 C – 10 6.7/1–14 14/9 12/0.5–24 8.5b/3–14
67% 33%
24% 76%
43% 57%
36% 64%
GI, gastrointestinal; S, stomach; D, duodenum; C, colon; CIBDAI, canine IBD activity index. a Overall (composite) histology score across one or more alimentary organs. CIBDAI score was unavailable in four dogs undergoing gastroscopy; disease duration (>3 weeks) unavailable in six dogs undergoing gastroscopy; CIBDAI score and disease duration (>3 weeks) unavailable in one dog undergoing enteroscopy; CIBDAI score unavailable in two dogs undergoing colonoscopy; disease duration (>3 weeks) unavailable in three dogs undergoing colonoscopy. b CIBDAI score prospectively scored for validation phase.
age from 1 to 14 years (mean, 6.7 years) and included 14 spayed females, eight neutered males, and one intact male dog. Their breeds included: four West Highland white Terriers, three Mixed breeds, two Shih Tzu’s, two German shepherd dogs, two Labrador retrievers, two Boxers, and one each of German shorthaired pointer, Gordon Setter, Boston terrier, Scottish terrier, Vizsla, and Collie. All 23 dogs had gastroduodenoscopic evaluations performed and 10 of these same dogs also underwent colonoscopy. For quantitative scoring (0, absent; 1, mild-to-moderate abnormal appearance; 2, moderate-to-severe abnormal appearance) of video clips between the two experienced endoscopists, there was almost perfect agreement (K = 0.87) between operators when total scores of the stomach were compared. Moreover, there was moderate agreement (K = 0.6) in total scores for duodenal mucosal assessment between experienced endoscopists and substantial agreement (K = 0.74) between operators for colonic mucosal assessment. When assessing individual endoscopic parameters for each organ using this quantitative scheme, there was substantial agreement in identification of gastric erosion (K = 0.84), granularity (K = 1) and friability (K = 0.8). Assessment of the small intestine showed substantial agreement in identification of mucosal erosion (K = 0.84) and moderate agreement for granularity (K = 0.66), friability (K = 0.67), and lymphatic dilation (K = 0.62). The assessment of colonic mucosal appearance showed substantial agreement for erosion (K = 1), granularity (K = 1), and friability (K = 1). Similarly, qualitative mucosal assessment (i.e. scoring based simply on the presence of an abnormal appearance) showed almost perfect agreement between experienced endoscopists when comparing cumulative scores of the stomach (K = 0.87). There was moderate agreement in the cumulative scores for duodenal mucosal assessment between experienced endoscopists (K = 0.58) and there was substantial agreement in cumulative scores between experienced endoscopists when performing colonic mucosal assessment (K = 0.8). In the validation phase of this study, the quantitative and qualitative scores of mucosal appearances were virtually identical (Table 5) and the presence of abnormal mucosal appearances/ organ was more quickly and objectively assessed by experienced operators.
tectable histopathologic lesions, especially inflammatory and neoplastic disorders involving the GI mucosa (Roth et al., 1990). In this large clinical trial, flexible GI endoscopy was used to examine the stomach, duodenum, and colon of 58 dogs with IBD and to acquire mucosal biopsies for histopathologic characterization. Most dogs in this earlier report had abnormal endoscopic appearances, interpreted as erythema, friability, enhanced granularity, and ulceration or erosion associated with predominantly lymphocyticplasmacytic mucosal inflammation. More recent trials have evaluated the association between endoscopic scores of the duodenum and colon and long-term outcome in dogs with chronic enteropathies (Allenspach et al., 2007) or have utilized endoscopic evaluation to assess mucosal healing pre- vs. post-treatment in dogs with lymphocytic-plasmacytic gastroenteritis (Garcia-Sancho et al., 2007). Given the ubiquitous nature of GI endoscopy in clinical practice, it is surprising that no studies have investigated the inter-observer variability in the endoscopic assessment of mucosal lesions using a validated scoring system. The current study investigated the inter-observer agreement in the endoscopic assessment of mucosal appearance of the GI tract to design and validate a simple and easy-to-use endoscopic scoring system for dogs with IBD, using a combination of retrospectively archived still images and prospectively acquired video clips. In the validation phase, we found moderate-to-substantial inter-observer agreement and reproducibility among experienced endoscopists for increased granularity, friability, erosions, and lymphatic dilatation (duodenal mucosa) involving the stomach, duodenum, and colon. A composite score that incorporates only the presence/absence of these four mucosal variables, and not their magnitude or the extent of their severity, is proposed for use in the endoscopic assessment of canine IBD. The first step in development of this endoscopic score was to identify which mucosal parameters in the stomach, duodenum, and colon would be evaluated by novice and experienced endoscopists
Discussion
Stomach Duodenum Colon
Endoscopic evaluation plays an important role in the diagnosis and management of dogs with chronic enteropathy. Abnormal endoscopic observations in dogs with histories of regurgitation, vomiting, and/or diarrhea have been previously associated with de-
Table 5 Kappa statistic (K) valuesa for validation tests for inter-observer agreement (IOA) used in the endoscopic activity score. Organ
Quantitative score
Qualitative score
0.87 0.60 0.74
0.87 0.58 0.80
a K < 0 represents no agreement, K = 0–0.20 represents slight agreement, K = 0.21– 0.40 represents fair agreement, K = 0.41–0.60 represents moderate agreement, K = 0.61–0.80 represents substantial agreement and K = 0.81–1 represents almost perfect agreement.
Please cite this article in press as: J.E. Slovak, et al., Development and validation of an endoscopic activity score for canine inflammatory bowel disease, The Veterinary Journal (2015), doi: 10.1016/j.tvjl.2014.12.030
ARTICLE IN PRESS J.E. Slovak et al./The Veterinary Journal ■■ (2015) ■■–■■
for inter-observer agreement. One approach was to identify the most common mucosal lesions (i.e. hyperemia, friability, enhanced granularity, erosions, lymphatic dilatation, and/or mass) in dogs with IBD based on earlier studies and the endoscopic experiences of AEJ and others over many years (Roth et al., 1990; Jergens et al., 1992, 2003, 2010; Allenspach et al., 2007; Slovak et al., 2014). These same general endoscopic parameters have been shown to be associated with clinical GI signs (Jergens et al., 2003; Allenspach et al., 2007; Garcia-Sancho et al., 2007; Heilmann et al., 2014), evidence of mucosal healing (Allenspach et al., 2007; Garcia-Sancho et al., 2007), histopathologic interpretation (Roth et al., 1990; Jergens et al., 1992; Allenspach et al., 2007; Larson et al., 2012), and prognostic outcome (Allenspach et al., 2007) in separate clinical trials. We observed, as described in the human literature, that observer variation for graded characteristics (i.e. mucosal hyperemia – is it pale, pink, or red?) was quite high, while that for discontinuous variables (i.e. presence or absence of erosions) was generally low (Baron et al., 1964). A discrepancy between mucosal hyperemia and the presence of histopathologic abnormalities in dogs with chronic enteropathy has been previously reported (Roth et al., 1990). Similar to Roth et al. (1990), we have also shown that operator experience plays an important role in duodenal endoscopic assessment, with novice endoscopists less likely to identify mucosal lesions or misinterpret normal vs. abnormal mucosa (Slovak et al., 2014). Still images were used exclusively in the development phase of this study since: (1) they are widely used in almost all clinical practices and are often retained as a component of the dog’s medical record, and (2) they are frequently used to communicate results of the diagnostic procedure to other veterinarians involved in case management and to clients (Washabau et al., 2010). Additionally, a manageable number of still images (vs. video clips) could be more easily evaluated by participating clinicians, which assured good compliance in the trial. Accurate assessment of endoscopic activity from archived still images of humans with UC has been previously reported (Osada et al., 2010). In the second part of the study, these same endoscopic parameters were tested in the stomach and colon, regardless of the level of reproducibility found in the previous inter-observer agreement study (Slovak et al., 2014). As noted in the results, there were no significant differences in inter-observer agreement between novice and experienced operators regarding either gastric or colonic mucosal interpretation. This observation that the endoscopic parameters chosen for evaluation showed no significant difference in reproducibility between operator groups was both interesting and reassuring. In this regard, these data suggested that: (1) different endoscopists could reliably recognize specific endoscopic appearances in the stomach and colon that they felt were important indicators of mucosal inflammation and (2) those abnormal endoscopic appearances previously identified by operators using still images could be reliably tested in prospective clinical trials using video clips. Our results demonstrated consistently good inter-observer agreement among the experienced endoscopists using defined mucosal appearances to define endoscopic activity in two ways – quantitatively and qualitatively. For the assessment of overall endoscopic severity, our study showed that the strength of agreement for quantitative vs. qualitative scoring was virtually identical (Table 5). When assessing individual endoscopic parameters quantitatively by organ, we observed the lowest inter-observer agreement for duodenal scoring for each descriptor, although at least moderate agreement was documented between experienced operators. A similar observation was made when using the qualitative duodenal scoring criteria, where the inter-observer agreement was reduced approximately 15% in comparison with gastric or colonic mucosal assessment. The reasons for the lower duodenal inter-observer agreement are unclear, but could possibly be related to differences in the experience level of the two operators in the evaluation of small in-
5
testinal mucosa i.e. extensive (30 years) vs. moderate (3 years) clinical experience in performance of advanced GI endoscopic procedures. Interestingly, even among expert human endoscopists, the extent of complete or partial agreement for endoscopic activity of UC (based on score 0–3) can range from 21% to 46% of patients (Walsh et al., 2009). The endoscopic activity score for canine IBD was broadly constructed from the observations of the six established activity indices used in human IBD, including the Baron (Baron et al., 1964), Blackstone (Osada et al., 2010), Schroeder (Schroeder et al., 1987; Osada et al., 2010), and Matts (Matts, 1961), scores for UC, which are all based on minor modifications of the original Baron score, the Crohn’s disease endoscopic index of severity (CDEIS; Mary and Modigliani, 1989) and simple endoscopic score for Crohn’s disease (SES-CD; Daperno et al., 2004). Similar to these human indices, the proposed canine score uses the mucosal appearance of erosions, friability, and enhanced granularity to define endoscopic activity with observations of lymphatic dilatation and mass lesions limited to duodenal and colonic mucosal assessment, respectively. The infrequent occurrence of colonic mass lesions (vs. other colonic mucosal indices) in dogs with IBD provided the rationale for excluding this parameter from further evaluation in the validation phase of this study. Moreover, the proposed index was associated with relatively high inter-observer agreement and reproducibility for the experienced endoscopists in the validation phase. It is noteworthy that we provide new information on the inter-observer agreement for gastric and colonic mucosal assessment in dogs with IBD, which has not been previously reported. Taken together, these findings suggest that an endoscopic activity score for canine IBD, developed and validated using still images and video clips, would be of practical value to all clinicians using current endoscopic technology. There are some potential limitations in our study. First, we utilized only a single referral center for this clinical trial. Whether our results for inter-observer agreement across different study centers, endoscopic interpretation by non-specialist clinicians, or endoscopic evaluation of dogs with different enteropathies are generalizable to other veterinary practices was not assessed. Additionally, our paper discusses inter-observer agreement, comparisons within the group vs. intra-observer agreement, and comparisons of results obtained from individual clinicians. Future studies should include both inter- and intra-observer comparisons. Most dogs enrolled in our study had active moderate-to-severe clinical disease activity based on CIBDAI scores. We did not perform subset analyses to determine whether these same endoscopic variables would have comparable inter-observer agreement in dogs presenting with quiescent disease or mild clinical severity. Further, we did not report on the abnormal appearance of the ileal mucosa in dogs with IBD. Endoscopic examination of the ileum was not performed in all dogs undergoing colonoscopy due to technical difficulties associated with general anesthesia and/or the inability to traverse the ileocolic sphincter (i.e. incomplete bowel cleansing). The relative importance of these potential problems is likely to be clarified by performing large multi-center studies with broad application of the endoscopic score in research and in clinical decision-making. Conclusions The proposed qualitative endoscopic activity score for canine IBD meets the following requirements as described for human endoscopic indices: (1) it is well correlated with the endoscopist’s overall assessment of abnormal mucosal appearance; (2) the quality of the assessment is consistent when evaluated in a set of patients different than that used for its development; (3) it incorporates mucosal parameters intuitively considered as important by experienced endoscopists; (4) it incorporates mucosal parameters previously shown to be reproducibly collected, and is itself reproducible; and
Please cite this article in press as: J.E. Slovak, et al., Development and validation of an endoscopic activity score for canine inflammatory bowel disease, The Veterinary Journal (2015), doi: 10.1016/j.tvjl.2014.12.030
ARTICLE IN PRESS J.E. Slovak et al./The Veterinary Journal ■■ (2015) ■■–■■
6
(5) it is simple to calculate. Importantly, this validated endoscopic index should be correlated with clinical indices and laboratory biomarkers of inflammation in future studies (Mary and Modigliani, 1989). Conflict of interest statement None of the authors of this paper has a financial or personal relationship with other people or organizations that could inappropriately influence or bias the content of this paper. Acknowledgements The authors thank Amy Hodnefield and Carrie Schwake for technical assistance. Portions of this manuscript were presented as research abstracts at the 2013 ACVIM Forum, Seattle, WA, and the 2013 ECVIM Congress, Liverpool, England. References Allenspach, K., 2011. Clinical immunology and immunopathology of the canine and feline intestine. Veterinary Clinics of North America. Small Animal Practice 41, 345–360. Allenspach, K., Wieland, B., Grone, A., Gaschen, F., 2007. Chronic enteropathies in dogs: Evaluation of risk factors for negative outcome. Journal of Veterinary Internal Medicine 21, 700–708. Baron, J.H., Connell, A.M., Lennard-Jones, J.E., 1964. Variation between observers in describing mucosal appearances in proctocolitis. British Medical Journal 1, 89–92. Daperno, M., D’Haens, G., Van Assche, G., Baert, F., Bulois, P., Maunoury, V., Sostegni, R., Rocca, R., Pera, A., Gevers, A., et al., 2004. Development and validation of a new, simplified endoscopic activity score for Crohn’s disease: The SES-CD. GI Endoscopy 60, 505–512. Garcia-Sancho, M., Rodriguez-Franco, F., Sianz, A., Mancho, C., Rodriquez, A., 2007. Evaluation of clinical, macroscopic, and histopathologic response to treatment in non-hypoproteinemic dogs with lymphocytic-plasmacytic enteritis. Journal of Veterinary Internal Medicine 21, 11–17. Heilmann, R.M., Grellet, A., Allenspach, K., Lecoindre, P., Day, M.J., Preistnall, S.L., Toresson, L., Procoli, F., Grutzner, N., Suchodolski, J.S., et al., 2014. Association between fecal S100A12 concentration and histologic, endoscopic, and clinical disease severity in dogs with idiopathic inflammatory bowel disease. Veterinary Immunology Immunopathology 158, 156–166. Jergens, A.E., Simpson, K.W., 2012. Inflammatory bowel disease in veterinary medicine. Frontier Bioscience 4, 1404–1419. Jergens, A.E., Moore, F.M., Haynes, J.S., Miles, K.G., 1992. Idiopathic inflammatory bowel disease in dogs and cats: 84 cases (1987–1990). Journal of the American Veterinary Medical Association 201, 1603–1608. Jergens, A.E., Schreiner, A.C., Frank, D.E., Niyo, Y., Ahrens, F.E., Eckersall, P.D., Benson, T.J., Evans, R., Pressel, M., Ackermann, M.L., et al., 2003. A scoring index for disease activity in canine inflammatory bowel disease. Journal of Veterinary Internal Medicine 17, 291–297. Jergens, A.E., Crandell, J., Morrison, J.A., Deitz, K., Pressel, Ackermann, M., Suchodolski, J.S., Steiner, J.M., Evans, R., 2010. Comparison of oral prednisone and prednisone
combined with metronidazole for induction therapy of canine inflammatory bowel disease: A randomized-controlled trial. Journal of Veterinary Internal Medicine 24, 269–277. Larson, R.N., Ginn, J.A., Bell, C.M., Davis, M.J., Foy, D.S., 2012. Duodenal endoscopic findings and histopathologic confirmation of intestinal lymphangiectasia in dogs. Journal of Veterinary Internal Medcine 26, 1087–1092. Lichtiger, S., Present, D.H., Kornbluth, A., Glernt, I., Bauer, J., Galler, G., Michelassi, F., Hanauer, S., 1994. Cyclosporine in severe ulcerative colitis refractory to steroid therapy. New England Journal of Medicine 330, 1841–1845. Mary, J.Y., Modigliani, R., 1989. Development and validation of an endoscopic index of the severity for Crohn’s disease: A prospective multicenter study. Gut 30, 983–989. Matts, S.G., 1961. The value of rectal biopsy in the diagnosis of ulcerative colitis. Quarterly Journal of Medicine 30, 393–407. Osada, T., Ohkusa, T., Yokoyama, T., Shibuya, T., Sakamota, N., Beppu, K., Nagahara, A., Otaka, M., Ogihara, T., Watanabe, S., 2010. Comparison of several activity indices for the evaluation of endoscopic activity in UC: Inter-and intra-observer consistency. Inflammatory Bowel Disease 16, 192–198. Powell-Tuck, J., Bown, R.L., Lennard-Jones, J.E., 1978. A comparison of oral prednisolone given as single or multiple daily doses for active proctocolitis. Scandinavian Journal of Gastroenterology 13, 833–837. Roth, L., Leib, M.S., Davenport, D.J., Monroe, W.E., 1990. Comparisons between endoscopic and histologic evaluation of the gastrointestinal tract in dogs and cats: 75 cases (1984–1987). Journal of the American Veterinary Medical Association 4, 635–638. Samuel, S., Bruining, D.H., Loftus, E.V., Jr., Thia, K.T., Schroeder, K.W., Tremaine, W.J., Faubion, W.A., Kane, S.V., Pardi, D.S., De Groen, P.C., et al., 2013. Validation of the ulcerative colitis colonoscopic index of severity and its correlation with disease activity measures. Clinical Gastroenterology and Hepatology 11, 49–54. Sartor, R.B., 2006. Mechanisms of disease: Pathogenesis of Crohn’s disease and ulcerative colitis. National Clinical Practical Gastroenterology and Hepatology 3, 390–407. Schroeder, K.W., Temaine, W.J., Ilstrup, D.M., 1987. Coated oral 5-aminosalicylic acid therapy for mildly to moderately active ulcerative colitis, a randomized study. New England Journal of Medicine 317, 1625–1629. Seo, M., Okada, M., Yao, T., Ueki, M., Arima, S., Okumura, M., 1992. An index of disease activity in patients with ulcerative colitis. American Journal of Gastroenterology 87, 971–976. Slovak, J.E., Wang, C., Morrison, J.A., Deitz, K.L., LeVine, D.N., Otoni, C., King, R.R., Gerber, L.E., Hanson, K.R., Lundberg, A.P., et al., 2014. Endoscopic assessment of the duodenum in dogs with inflammatory bowel disease. Journal of Veterinary Internal Medicine 28, 1442–1446. Truelove, S.C., Witts, L.J., 1955. Coritsone in ulcerative colitis; final report on therapeutic trial. British Medical Journal 2, 1041–1048. Walmsley, R.S., Ayres, R.C., Pounder, R.E., Allan, R.N., 1998. A simple clinical colitis activity index. Gut 43, 29–32. Walsh, A.J., Brain, O., Keshav, S., 2009. How variable is the Mayo score between observers and might this affect trial recruitment or outcome? Gastroenterology 136 (Suppl. 1), A677 (Abstract). WSAVA International Gastrointestinal Group: Washabau, R.J., Day, M.J., Willard, M.D., Hall, E.J., Jergens, A.E., Mansell, J., Minami, T., Bilzer, T.W., 2010. Endoscopic, biopsy, and histopathologic guidelines for the evaluation of gastrointestinal inflammation in companion animals. Journal of Veterinary Internal Medicine 24, 10–26. Xavier, R.J., Podolsky, D.K., 2007. Unravelling the pathogenesis of inflammatory bowel disease. Nature 448, 427–434. Zoran, D.L., 2001. Gastroduodenoscopy in the dog and cat. Veterinary Clinics of North America. Small Animal Practice 31, 631–656.
Please cite this article in press as: J.E. Slovak, et al., Development and validation of an endoscopic activity score for canine inflammatory bowel disease, The Veterinary Journal (2015), doi: 10.1016/j.tvjl.2014.12.030
