Corresponding author: Michelle Im Hof, Department of Clinical Veterinary Medicine, Division of Small Animal Internal Medicine, Vetsuisse Faculty of the University of Bern, Länggassstrasse 128, 3001 Bern, Switzerland; e-mail:firstname.lastname@example.org.
Inflammatory bowel disease (IBD) and food-responsive diarrhea (FRD) are chronic enteropathies of dogs (CCE) that currently can only be differentiated by their response to treatment after exclusion of other diseases. In humans, increased urinary concentrations of leukotriene E4 (LTE4) have been associated with active IBD.
To evaluate urinary LTE4 concentrations in dogs with IBD, FRD, and healthy controls, and to assess correlation of urinary LTE4 concentrations with the canine IBD activity index (CIBDAI) scores.
Eighteen dogs with IBD, 19 dogs with FRD, and 23 healthy control dogs.
In this prospective study, urine was collected and CIBDAI scores were calculated in client-owned dogs with IBD and those with FRD. Quantification of LTE4 in urine was performed by liquid chromatography-tandem mass spectrometry and corrected to creatinine.
Urinary LTE4 concentrations were highest in dogs with IBD (median 85.2 pg/mg creatinine [10th–90th percentiles 10.9–372.6]) followed by those with FRD (median 31.2 pg/mg creatinine [10th–90th percentiles 6.2–114.5]) and control dogs (median 21.1 pg/mg creatinine [10th–90th percentiles 9.1–86.5]). Urinary LTE4 concentrations were higher in dogs with IBD than in control dogs (P = .011), but no significant difference between IBD and FRD was found. No correlation was found between urinary LTE4 concentrations and CIBDAI.
Conclusions and Clinical Importance
The higher urinary LTE4 concentrations in dogs with IBD suggest that cysteinyl leukotriene pathway activation might be a component of the inflammatory process in canine IBD. Furthermore, urinary LTE4 concentrations are of potential use as a marker of inflammation in dogs with CCE.
Inflammatory bowel disease (IBD) refers to a group of chronic gastrointestinal disorders of unknown origin characterized by mucosal inflammation.[1-3] Diagnosis of IBD is currently defined by a combination of chronic persistent or recurrent gastrointestinal signs; histopathologic evidence of mucosal inflammation; inability to document other causes of gastrointestinal inflammation; inadequate response to dietary, antibiotic, and anthelmintic therapies alone; and clinical response to anti-inflammatory or immunosuppressive agents. Food-responsive diarrhea (FRD) refers to inflammatory enteropathies for which clinical signs can be controlled by feeding a specific diet without the need for immunosuppressive treatment. Currently, the only way to differentiate dogs with FRD from those with IBD is the treatment success with feeding an elimination diet for up to 8 weeks.[5, 6]
Disease severity and response to treatment in dogs with IBD and FRD are currently assessed with clinical scoring systems such as the canine IBD activity index (CIBDAI) and the canine chronic enteropathy activity index.[5, 7] The former is a simple numeric index based on the presence and frequency of cardinal gastrointestinal signs, such as appetite, vomiting, feces consistency, frequency of defecation, weight loss, and attitude/activity. A score of 0–3 indicates clinically insignificant disease, a score of 4–5 indicates mild IBD, a score of 6–8 indicates moderate IBD, and a score of ≥9 indicates severe IBD.[7, 8] Unfortunately, the usefulness of the CIBDAI scoring system is limited because it is based on partially subjective assessments of clinical signs and does not correlate well with histologic severity.[5, 6, 9]
In an effort to provide a more objective method for assessing disease severity in dogs with IBD and to help differentiate them from dogs with other causes of canine chronic enteropathy (CCE), novel biomarkers including insulin-like growth factor (IGF-1), C-reactive protein, and perinuclear antineutrophilic cytoplasmic antibody have been recently investigated.[5, 7, 9-11] Although not all of these assays are commercially available, these results suggest that biomarkers might be helpful in assessing disease severity in dogs with IBD or differentiating between IBD and FRD.
Cysteinyl leukotrienes, such as leukotriene C4, leukotriene D4, and leukotriene E4 (LTE4), are proinflammatory 5-lipoxygenase-derived products that contribute to the inflammatory response by increasing microvascular permeability, enhancing chemotaxis, and increasing secretion of colonic mucosa.[12-15] LTE4 is the major urinary metabolite of this enzymatic pathway and provides the best measure of in vivo systemic cysteinyl leukotriene production in humans. Leukotrienes derived from the 5-lipoxygenase pathway play a major role in the inflammatory response in humans with IBD. Colonic biopsies from patients with active IBD show increased cellular expression of 5-lipoxygenase, 5-lipoxygenase-activating protein, and leukotriene A4 hydrolase, indicating increased tissue synthesis of leukotriene B4. Furthermore, urinary excretion of LTE4 is increased in patients with Crohn's disease and those with ulcerative colitis compared with controls. In addition, LTE4 concentrations are higher in patients with active disease than patients in remission.
These findings in human patients suggest that urinary LTE4 concentrations might correlate with disease severity in dogs with IBD and might help to differentiate dogs with IBD from those with FRD and healthy dogs. The aims of this study were to evaluate urinary LTE4 concentrations in dogs with IBD, dogs with FRD, and healthy controls and to assess correlation of urinary LTE4 concentrations with CIBDAI scores in these dogs.
Materials and Methods
Healthy Control Dogs
The control group consisted of 23 privately owned dogs (blood donors and staff-owned) that were determined to be healthy based on history, physical examination, and a lack of abnormalities on complete blood count, serum biochemistry profile, and urinalysis. For each dog, a urine sample was collected by cystocentesis and immediately stored at −20°C until LTE4 quantification. Additionally, 3 urine samples each were collected from 3 dogs 4 hours apart during daytime.
Dogs with Chronic Enteropathies
Thirty-seven dogs with signs of chronic gastrointestinal disease that were presented to the Small Animal Teaching Hospital of the University of Bern between November 2003 and June 2009 were evaluated. The initial diagnostic work up for all of these patients included a complete blood count, serum biochemistry profile, urinalysis, fecal parasitology, fecal culture, abdominal ultrasound, serum trypsin-like immunoreactivity, cobalamin and folate concentrations, and gastrointestinal endoscopy. None of the dogs had been treated with antibiotics, corticosteroids, or antacids in the 2 weeks before entering the study. Selection criteria included a history of chronic diarrhea of at least 6 weeks duration with or without vomiting, exclusion of identifiable underlying disorders, and histopathologic evidence of intestinal inflammatory cellular infiltrate. Histopathologic evaluation was performed by 3 board certified pathologists from 2003 to 2009. At the time of initial presentation, disease severity was evaluated by calculation of CIBDAI scores, and urine was collected by cystocentesis and immediately stored at −20°C until LTE4 quantification.
Following the initial evaluation, all dogs were fed an elimination diet1 exclusively. Dogs in which clinical signs improved or resolved within 14 days were classified as having FRD. Dogs that did not improve within the first 14 days of dietary treatment were classified as having IBD and were treated with prednisolone (1 mg/kg q12h) for 14 days followed by a tapering dosage over 10 weeks. All experimental procedures for both healthy dogs and those with chronic enteropathies were approved by the Cantonal Animal Care and Use Committee for Animal Experimentation, Bern, Switzerland.
Measurement of Urinary LTE4 Concentrations
Quantification of LTE4 was performed by liquid chromatography-tandem mass spectrometry as previously described. The lower limit of quantification was 10 pg/mL. Interassay and intraassay precision was <6%. Concentrations of LTE4 were corrected to urinary concentrations of creatinine, as measured with a Jaffe kinetic-based assay2 and were expressed as pg LTE4/mg creatinine.
Data were analyzed by means of nonparametric methods by use of a commercial statistical software package.3 All quantitative data were presented as median with 10th and 90th percentiles. Comparison of urinary LTE4 concentrations among dogs with IBD, those with FRD, and control dogs was performed with a Kruskal-Wallis one-way ANOVA with Dunn's test at which the latter was considered significant if z-value >2.394. A receiver-operating characteristic (ROC) curve was used to select cutpoint values for sensitivity and specificity reporting. In dogs with IBD and those with FRD, correlation between urinary LTE4 concentrations and CIBDAI was analyzed using a Spearman rank test. A P value <.05 was considered significant.
The control group (n = 23) included 12 females and 11 males of which 10 and 5 were neutered, respectively. The dogs were 1.25–10.5 years old (median 5.75 years). Breeds included mixed breeds (n = 7), Belgian Malinois (n = 3), Labrador Retriever (n = 2), and 1 each Doberman Pinscher, Standard Poodle, Flat Coated Retriever, Dutch Shepherd, German Shepherd, Irish Terrier, Beagle, Coton de Tulear, Jack Russell Terrier, West Highland White Terrier, and Dachshund. Body weights ranged from 5.5 to 38 kg (median 21 kg). Dogs classified as having IBD (n = 18) included 9 females and 9 males of which 6 and 2 were neutered, respectively. The dogs were 1.0–12.7 years old (median 6.5 years). Breeds included mixed breeds (n = 4), Rottweiler (n = 3), Pug (n = 2), and 1 each Beauceron, Dalmatian, German Shepherd, Coton de Tulear, Papillon, Cavalier King Charles Spaniel, Yorkshire Terrier, Australian Shepherd, and Dachshund. Body weights ranged from 2.9 to 57 kg (median 18.35 kg). Dogs classified as having FRD (n = 19) included 8 females and 11 males of which 4 and 5 were neutered, respectively. The dogs were 8 months–11.2 years old (median 3 years). Breeds included mixed breeds (n = 5), Labrador Retriever (n = 2), and 1 each Newfoundland, Alaskan Malamute, English Setter, Golden Retriever, Berger Blanc Suisse, Bernese Mountain Dog, German Shepherd, French Bulldog, Cairn Terrier, West Highland White Terrier, Shih-Tzu, and Dachshund. Body weights ranged from 6.4 to 49 kg (median 27.0 kg).
Urinary LTE4 Concentrations
Urinary LTE4 concentrations were highest in dogs with IBD (median 85.2 pg/mg creatinine [10th–90th percentiles 10.9–372.6 pg/mg creatinine]) followed by those with FRD (median 31.2 pg/mg creatinine [10th–90th percentiles 6.2–114.5 pg/mg creatinine]) and control dogs (median 21.1 pg/mg creatinine [10th–90th percentiles 9.1–86.5 pg/mg creatinine]). Median urinary LTE4 concentrations among the 3 groups were different (P = .011). Urinary LTE4 concentrations were significantly higher in dogs with IBD than in control dogs (z-value = 2.991) but were not significantly different between dogs with IBD and those with FRD (z-value = 1.88) or between dogs with FRD and controls (z-value = 1.041) (Fig 1). Within the 3 groups, LTE4 was not normally distributed as determined by the Kolmogorov–Smirnov and the Shapiro–Wilk tests. The ROC curve analysis suggested an optimal urinary LTE4 cutpoint value of 90.59 pg/mg creatinine for the diagnosis of IBD. The sensitivity and specificity for predicting IBD at this cutpoint were 50 and 96%, respectively (Fig 2A). Another potential cutpoint value of 51.85 LTE4 (pg/mg creatinine) revealed a sensitivity and specificity of 72 and 78%, respectively (Fig 2A). Results of ROC curves analyses of urinary LTE4 (pg/mg creatinine) for IBD, FRD, and CCE (=IBD + FRD) are shown in Figure 2A–C. A cutpoint value of 98.2 LTE4 (pg/mg creatinine) revealed a sensitivity of 28% and specificity of 96% for predicting FRD. A cutpoint value of 90.59 LTE4 (pg/mg creatinine) revealed a sensitivity of 39% and a specificity of 96% for predicting CCE. Urinary LTE4 concentrations of 3 samples collected from 3 healthy dogs 4 hours apart were 5.8, 16.1, 15.8; 22.9, 15.9, 24.1; and 35.8, 32.8, 44.8 pg/mg creatinine, respectively.
Correlation between LTE4 and CIBDAI
In dogs with IBD and FRD, no correlation was found between urinary LTE4 concentrations and clinical severity as assessed with the CIBDAI (P = .11).
This report describes measurement of urinary LTE4 concentrations in healthy dogs and in those with chronic enteropathies. Our findings are similar to those previously described in human patients and suggest that measurement of urinary LTE4 might be useful for the noninvasive assessment of cysteinyl leukotriene pathway activation in animals, as it is in people. In addition, our findings suggest that measurement of urinary LTE4 could be considered in future investigations aimed at the development of new tools for the evaluation of other inflammatory diseases in dogs.
In this study, urinary LTE4 concentrations were significantly higher in dogs with IBD than in healthy dogs, suggesting that urinary LTE4 might indeed be a candidate biomarker for the differentiation of dogs with IBD from those with noninflammatory enteropathies. We did not find significant differences between dogs with IBD and those with FRD or between dogs with FRD and healthy controls. However, because of the large variability within each group and the relatively small sample size, this investigation might have been underpowered. A sample size calculation, based on log-transformed values of LTE4 of the current study, under the assumption of a power of 0.8, indicates that 2 groups of 132 dogs each would be necessary to detect a potential difference in mean LTE4 values between healthy dogs and dogs with FRD. Two groups with 44 dogs each would be necessary to detect a potential difference between dogs with FRD and IBD. Another potential reason for the lack of significant difference in urinary LTE4 concentration between dogs with IBD and those with FRD is that some dogs might have been misclassified because of the short duration of the dietary trial. Previous recommendations for the duration of elimination diet trials have ranged from 2 to 6–8 weeks.[5, 22] In this investigation, a duration of only 2 weeks was chosen because longer durations were thought to hinder client compliance and enrollment in the study. Therefore, it is possible that some dogs with FRD were erroneously assigned to the IBD group because they failed to have a response to short-term dietary management, but would have had a response given more time on the diet.
Another consideration in the potential application of urinary LTE4 as a biomarker for IBD would be the importance of ruling out the presence of concomitant inflammatory diseases in other organ systems. In human patients, increased urinary LTE4 concentrations have been associated with bronchial obstructive disease, coronary artery disease, cardiac ischemia, and type 1 diabetes as well as with IBD.[18-20, 23-26] Therefore, a thorough diagnostic evaluation (as was performed in the dogs enrolled in this study) would be important to rule out other inflammatory diseases if the use of urinary LTE4 to assess disease type and severity in dogs with suspected IBD were to be considered.
According to ROC analysis, there was not a clear cutpoint value with high sensitivity and specificity to distinguish between the 3 study groups. When cutpoint values greater than 90.59 pg/mg creatinine were used as abnormal, urinary LTE4 was highly specific (96%), although not particularly sensitive (50%) in discriminating dogs with IBD from healthy dogs. Considering a 2nd cutpoint value of 51.85 LTE4 pg/mg creatinine, sensitivity would increase (72%) whereas specificity would decrease (78%). This finding suggests the former cutpoint value (90.59 LTE4 pg/mg creatinine) to be useful as a confirmation test for IBD rather than a screening test. However, it is important to note that the ROC analysis has been evaluated only in the 3 selected groups enrolled in this study.
In this study, CIBDAI scores from the IBD and FRD dogs did not correlate with urinary LTE4 concentrations. This may be because of the fact that the CIBDAI scoring system is a partially subjective assessment that reflects the clinical severity of the disease, whereas urinary LTE4 concentration is an objective measurement that reflects the degree of inflammation. This discrepancy between clinical scores and pathologic severity has also been reported by others, revealing weak association between clinical findings and histopathologic lesions in dogs with IBD.[5, 6, 9, 27] The reason for this discrepancy remains unknown.
Previous investigations in people with Crohn's disease or ulcerative colitis have found significantly higher concentrations of urinary LTE4 in patients with active disease than in those in clinical remission. Although urinary LTE4 concentrations were not assessed after treatment in the dogs described here, the differences detected between IBD dogs and healthy controls support the need for future studies to assess posttreatment urinary LTE4 concentrations in dogs with IBD, and to correlate these concentrations with degree of clinical response to help determine whether urinary LTE4 concentrations would be useful as a monitoring tool in dogs being treated for IBD.
The finding of increased urinary LTE4 concentrations in dogs with IBD suggests that cysteinyl leukotriene pathway activation may contribute to the inflammation associated with canine IBD. This suggestion is supported by another study, in which COX-2 in the colon and duodenum as well as 5-LO in the colon was significantly higher in IBD and FRD compared with healthy control dogs.4 Additional studies should be aimed at further evaluating pathway activation on a cellular basis in dogs with IBD such as immunohistochemical and molecular analysis of the enzymes and metabolites associated with the 5-lipoxygenase pathway in intestinal biopsies. The potential role of cysteinyl leukotriene pathway activation in canine IBD is important in part because it may be a potential target for pharmacologic intervention through the use of 5-lipoxygenase inhibitors or leukotriene receptor antagonists. Although the use of these compounds has not yet been evaluated in dogs, the 5-lipoxygenase inhibitor zileuton has been shown to prevent the development of inflammatory events in a model of dextran sodium sulfate-induced colitis in rats.
There are several limitations of this study that should be noted. First, the assay used to measure urinary LTE4 is based on a liquid chromatography-tandem mass spectrometry that has high selectivity and specificity for LTE4 in humans. Additional steps should be taken to further validate the use of this assay in dogs and to assess its sensitivity and specificity for detecting canine LTE4 in urine. In addition, although LTE4 is stable in urine and is considered the best marker of cysteinyl leukotriene production in people, instances of sudden peaks in urinary LTE4 concentrations secondary to bronchial provocation with allergens in atopic asthmatic subjects have been documented. Because these peaks are generally of short duration, collection of urine at short intervals is needed to detect the transient increase in LTE4. Although urinary LTE concentrations seemed to vary only little over time in the 3 healthy dogs in this study, it is not known whether transient increases cause hour-to-hour or day-to-day variations in urinary LTE4 concentrations in dogs. Therefore, collection of multiple urine samples at short intervals might be of importance for accurate estimation of cysteinyl leukotriene production in dogs.
In summary, we conclude that LTE4 concentrations can be measured in canine urine and are significantly higher in dogs with IBD than in healthy dogs. Future studies should focus on evaluating urinary LTE4 concentrations in a larger number of dogs with IBD and in dogs with other chronic enteropathies, both before and after treatment.
Purina Veterinary Diets LA salmon and rice, Société des Produits Nestlé SA, Vevey, Switzerland or Biomill fish and rice, Biomill SA, Granges-près-Marnand, Switzerland
Modular, Roche Diagnostic, France
NCSS Version 07.1.20, 2007, Kaysville, UT
Dumusc S, et al. COX-2 and 5-LO in canine chronic enteropathies. 19th ECVIM-CA Congress, Porto, Portugal, 2009 (abstract)