Materials and Methods

The experimental protocol was approved by the ethics committee for animal experiments at the University of Helsinki, Finland. Medical records from January 1991 to December 1993 at the University Veterinary Teaching Hospital were reviewed, and owners of dogs with a diagnosis of EPI were contacted. The diagnosis of EPI was made on the basis of typical clinical signs, low serum trypsinlike immunoreactivity,¹⁴ and negative results for fecal protease in a soybean stimulation test.¹⁵

Four feeding periods were conducted during a period of 14 weeks in each dog's home. During the 2-week control period, all dogs were fed the diet they customarily received at home. Thereafter, 3 special diets were fed for 3 weeks each. Dogs were randomly allocated to 1 of 6 possible sequences of the diets, with 3 or 4 dogs in each group. Dogs' owners were unaware of the order in which the special diets were fed. All food cans were similar in appearance and were labeled only with a code number. Dogs were allowed to become accustomed to each new diet for 1 week before the commencement of each recording period, so that data were recorded for the last 2 weeks of each 3-week period. Except for the different diets, dogs lived in their home environments with no changes.

The original diets (ie, fed before the study) differed from dog to dog and usually consisted of commercial dry kibble plus homemade food (eg, beef, chicken, rice, and potato). The special diets were a high-fat diet,^a a low-fat high-fiber diet,^b and a low-residue (ie, highly digestible, moderate-fat) diet.^c The 3 special diets were fed in canned form. Nutrient and energy contents of the diets as reported by the manufacturer were summarized (Table 1). Bacterial culture was performed on samples of each special diet. Owners fed all diets twice a day in a quantity necessary to maintain the dogs' body weight. Dogs were weighed by owners every other day during the first week and then once per week for the next 2 weeks of each feeding period. Total daily consumption of the special diets was recorded by the owners. Mean daily consumption of each diet was calculated from values for the last week of each feeding period. Enzyme supplementation^d was maintained throughout the study at a dosage of 3 g/25 kg of body weight (3 g/55 lb) administered orally per meal.

Dogs were assessed by owners daily for 21 days and scored for the following clinical signs: appetite, defecation frequency, consistency of feces, flatulence, borborygmus, and coprophagia. Data were recorded such that the lowest score of 0 represented a clinical sign least typical of EPI, and the highest score of 3 represented a sign most typical of EPI (Appendix). An EPI index was calculated by adding the daily scores for each sign.

During the last week of each diet, owners collected dogs' feces over a 24-hour period. Feces were stored at −20°C while awaiting processing, and all samples were analyzed at the end of the study. For analyses, feces were thawed, weighed, and homogenized. Fecal fat excretion was determined by a modification of the van de Kamer method.¹⁶ Fecal enzyme activities of protease, amylase, and lipase were determined via radial gel diffusion in a described method.¹⁷

Statistical analysis—The Friedman test for multiple treatments with same subjects and multiple comparison tests were used to evaluate differences among diets in feeding-period mean EPI indices, clinical signs, consumption, fecal variables, and body weight. Pairwise comparisons between mean ranks for all 4 diets were performed. Overall error rate was controlled by setting the error rate for each test to 0.05 divided by the number of comparisons. One dog (dog No. 8) did not complete the high-fiber diet feeding period and was excluded from those analyses. For all dogs, the Kruskal-Wallis 1-way ANOVA was applied, incorporating the daily EPI index values as input, to determine whether there were differences in EPI indices among diets in any given dog. The Spearman correlation coefficient was used to test associations among variables. For all analyses, values of P < 0.05 were considered significant.

Results

Twenty-one dogs with EPI were included in the study. Breeds represented were the German Shepherd Dog (11 dogs), Rough-coated Collie (9), and mixed-breed dogs (1). Dogs' ages ranged from 2 to 6 years. Dogs had been given enzyme replacement treatment for at least 6 months prior to the study. All dogs were in good clinical condition despite the illness, and none were receiving medication other than enzyme supplementation. Comparison of mean EPI indices for each feeding period in all 20 dogs that completed the study revealed no significant differences in scores among the 4 feeding periods (P = 0.195). In general, the mean EPI index was lowest (best) while dogs were consuming the original diets. Among the special diets, lowest (best) mean values for EPI index were associated with consumption of the high-fiber diet and median values for EPI index for that diet were better than those associated with the original diet (Figure 1; Table 2).

Box plots of EPI indices of 21 dogs with EPI during 2- week-long feeding trials for each of 4 diets with different nutrient content. Box indicates interquartile range, line in box indicates median, whiskers indicate range, and circles indicate outliers. ^*One dog did not complete the feeding period for the high-fiber diet. See Appendix for scoring system used to calculate EPI indices.

Citation: Journal of the American Veterinary Medical Association 228, 2; 10.2460/javma.228.2.225

• Download Figure
• Download figure as PowerPoint slide

View Full Size

Box plots of EPI indices of 21 dogs with EPI during 2- week-long feeding trials for each of 4 diets with different nutrient content. Box indicates interquartile range, line in box indicates median, whiskers indicate range, and circles indicate outliers. ^*One dog did not complete the feeding period for the high-fiber diet. See Appendix for scoring system used to calculate EPI indices.

Citation: Journal of the American Veterinary Medical Association 228, 2; 10.2460/javma.228.2.225

• Download Figure
• Download figure as PowerPoint slide

Box plots of EPI indices of 21 dogs with EPI during 2- week-long feeding trials for each of 4 diets with different nutrient content. Box indicates interquartile range, line in box indicates median, whiskers indicate range, and circles indicate outliers. ^*One dog did not complete the feeding period for the high-fiber diet. See Appendix for scoring system used to calculate EPI indices.

Citation: Journal of the American Veterinary Medical Association 228, 2; 10.2460/javma.228.2.225

• Download Figure
• Download figure as PowerPoint slide

Interindividual variation in EPI index scores associated with the different diets was high. The effect of the different diets on the EPI index of each dog was assessed by comparing daily indices during the feeding periods. In 20 of the 21 dogs, significant differences among feeding periods were observed. The original diet resulted in the lowest or best EPI index score in 8 of 20 (40%) dogs (dog No. 8 excluded), the high-fat diet was best in 5 of 20 (25%) dogs, the high-fiber diet was best in 4 of 20 (20%) dogs, and the low-residue diet was best in 2 of 20 (10%) dogs. In 1 dog, the lowest EPI index scores during feeding of the original and high-fat diets were equal. In some dogs, EPI index scores for 2 diets differed only slightly. Results indicated that the effect of the study diets on dogs' well-being was unpredictable.

The diets appeared to influence certain clinical signs associated with EPI (Table 2). During the highfiber and low-residue diet feeding periods, frequency of defecation was significantly higher than during feeding of the original diet. Fecal consistency was significantly firmer with the high-fiber diet than with the other 2 special diets. Flatulence and borborygmus were decreased with the high-fiber diet, compared with the other diets. Coprophagia was seldom observed in any feeding period.

Dogs' mean daily consumption of the less energydense high-fiber diet was significantly higher than consumption of the other special diets. Mean consumption of fat varied significantly among the special diets and was lowest with the high-fiber diet and highest with the high-fat diet. Mean fluctuation in body weight of the dogs was ± 800 g during all feeding periods. Pathogenic bacteria were not detected via bacterial culture of any samples from the special diets.

Volume of feces and fecal fat excretion were affected by diet. Fecal volume was significantly increased during the high-fiber diet feeding period, compared with volume observed with the original diets. Fecal fat excretion was significantly higher during the high-fat diet feeding period than with the other diets (Figure 2); mean daily fecal fat excretion was 0.7 g/kg of body weight, a value that is consistent with mild steatorrhea, whereas during feeding with the original high-fiber and low-residue diets, results of tests for fecal fat were within reference range.⁶

Box plots of fecal fat content in the same dogs as in Figure 1. BW = Body weight. See Figure 1 for remainder of key.

Citation: Journal of the American Veterinary Medical Association 228, 2; 10.2460/javma.228.2.225

• Download Figure
• Download figure as PowerPoint slide

View Full Size

Box plots of fecal fat content in the same dogs as in Figure 1. BW = Body weight. See Figure 1 for remainder of key.

Citation: Journal of the American Veterinary Medical Association 228, 2; 10.2460/javma.228.2.225

• Download Figure
• Download figure as PowerPoint slide

Box plots of fecal fat content in the same dogs as in Figure 1. BW = Body weight. See Figure 1 for remainder of key.

Citation: Journal of the American Veterinary Medical Association 228, 2; 10.2460/javma.228.2.225

• Download Figure
• Download figure as PowerPoint slide

A significant decrease in fecal protease and amylase activities was observed during the high-fiber and low-residue diet feeding periods, compared with the other diets. Values for fecal lipase activity associated with the high-fiber diet were significantly lower than with the original diet (Table 2). There was a significant negative correlation (r_s = −0.36; P < 0.001) between fecal lipase activity and fecal fat excretion. No relationship between fecal protease activities and EPI index was detected.

During the study, 9 dogs had sporadic gastroin-testinal disturbances and a short course of orally administered tylosin^e (dose, 20 mg/kg [9 mg/lb]) was given every 12 hours for 1 to 3 days. These problems were most apparent with the low-residue diet, affecting 7 dogs. In 1 dog, the high-fat diet was discontinued 2 days earlier than planned because of excessive flatulence, and 1 dog refused to eat the high-fiber diet.

Discussion

Clinical feeding studies have yielded contrasting results on the benefits of special diets as treatment for dogs with EPI. We previously investigated^11,12 the use of 2 special diets, and our findings indicated that a highly digestible low-fiber diet alleviated certain clinical signs of EPI, but that the low-fat diet resulted in no significant changes in clinical signs.Advantages of feeding special diets were minimal in our recent study¹⁰ of long-term treatment of EPI in dogs. The chief aim of the present study was to elucidate how owners of dogs with EPI should feed affected pets. Therefore, dogs were fed 4 diets of substantially different composition, and an EPI index score was calculated. Because the composition of dogs' original diets was not analyzed, it is possible that the original diets in some dogs did not differ substantially in content from certain special diets. The EPI index has been validated as a reliable means of comparing effects of various feeding regimens,^11,12 with values increasing when enzyme supplementation is decreased in dogs with EPI.¹² Use of the index does, however, pose certain practical difficulties. When owners act as observers, interobserver variation is high because evaluation of some signs is subjective. Moreover, certain factors, including stress and the stealing of food, cannot be controlled in studies such as this. In our study, the effects of single erroneous evaluations were minimized by calculation of mean scores over a protracted period. The time required for adaptation when diets are changed is not known; it is possible that the adaptive intervals in our trial should have been longer. The feeding of the special diets was carried out with owners unaware of diets' compositions with the intention of diminishing subjective effects during owner scoring.

In individual dogs, values of the daily EPI indices varied significantly among the different diets. This result indicates that diet plays an important role in the treatment of dogs with EPI. However, results also indicated that there was marked interindividual variation among dogs. This observation is problematic with regard to using the information to formulate instructions for feeding affected dogs and was the primary reason why no significant differences were detected in the mean EPI index values among the 4 diets. The original diet fed at home resulted in the lowest EPI index score for 8 of 20 (40%) dogs and for 1 dog in which the original and high-fat diets yielded equal EPI scores.

An unexpected finding was that 5 of 20 dogs fed the high-fat diet and 4 fed the high-fiber diet had low (good) EPI index scores because those diets have been proposed to adversely affect dogs with EPI.¹⁸ For purposes of comparing the influence of different diets by use of the EPI index, it should be remembered that the index incorporates 6 clinical signs. One diet might increase the severity of 1 clinical sign while simultaneously decreasing the severity of another sign, thus having a net effect of zero on the EPI index. This was observed particularly in dogs fed the high-fiber diet, which increased the severity of certain clinical signs and decreased the severity of others.

Some investigators^2,19 have concluded that a highfiber diet should be avoided in dogs with EPI because fiber may further diminish the absorption of other nutrients. Moreover, certain fibers inhibit the activity of pancreatic enzymes, especially lipase.¹⁹ The highfiber diet is restricted in fat and energy content and thus is recommended for use when weight reduction is desired. Although there were no marked changes in body weights of dogs during the 3-week feeding periods, some dogs had mild weight loss, likely a result of the difficulty in consuming sufficient food to meet daily energy requirements. Increased food consumption and poor digestion caused significant increases in fecal mass. The frequency of defecation was also increased, likely as a result of the influence of fiber on intestinal transit time. In addition to the negative effects of fiber, positive effects were also noticed; fecal consistency was better (ie, firmer) than that observed with the other diets, and flatulence and borborygmus were reduced.

Because enzyme supplementation alone cannot restore normal fat absorption in dogs with EPI, a lowfat diet has been considered necessary to achieve a favorable response. However, results of 1 study²⁰ indicate that a diet with a high proportion of fat corrects steatorrhea in dogs with EPI. Orally administered lipase is destroyed by the acidic conditions in the stomach, and up to 90% of lipase activity is lost before lipase reaches the duodenum.⁷ In dogs with small intestinal disease, fat absorption may also be negatively influenced by bacterial deconjugation of bile salts, an action that produces metabolites that in turn may result in diarrhea.^5,9,21 Simpson et al⁹ reported that during the initial period of enzyme treatment, a low-fat diet is important for achievement of a satisfactory treatment response. Results of an earlier study¹² with low-fat diets in affected dogs, however, did not reveal an association between a decrease in fat intake and alleviation of clinical signs. In the present trial, we studied the response of dogs with EPI to excessive dietary fat. The EPI index scores of dogs fed the high-fat diet were not significantly different from those of dogs fed the original diets, although the content of the original diet was unknown. However, we only measured the quantitative fecal fat excretion over a 24-hour period because the methodology for conducting the measurements is time consuming and unpleasant. Single-day collection of feces has been reported^16,22 to provide reliable results, although a 72-hour collection period is preferable. The effect of the dietary fat content was obvious during the high-fat feeding period in that fecal fat output was significantly increased, compared with output during the other feeding periods. During the feeding periods for the original, low-residue, and high-fiber diets, mean fecal fat excretion was within reference range, indicating that fat absorption returns to normal with enzyme supplementation in most dogs. The fecal consistency in dogs fed a high-fat diet was much poorer than when fed their original diet.

Findings from an earlier study¹¹ indicated that clinical signs of EPI may be alleviated when a dog's original diet is changed to a low-residue diet. However, in the present study, the low-residue diet was not superior to the other diets but, rather, yielded results that were contradictory in certain respects. That diet was only the best diet (as per the calculated EPI index) for 2 dogs. An unexpected finding associated with feeding the low-residue diet was that fecal mass was greater with that diet than with the original and high-fat diets. In another study,² consistent reductions in fecal mass were detected in dogs fed a lowfiber diet. In the present study, the low-residue diet induced unexpected gastrointestinal tract disturbances in some dogs and such disturbances were likely responsible for the high EPI scores. The etiology of the gastrointestinal tract disturbances associated with the low-residue diet remains undetermined. An explanation for the divergent results in the previous¹¹ and present studies with low-residue diets may lie in the different form of the feeds used. In the first study,¹¹ a dry diet was used, which was in contrast to the canned form used in the present study. Nevertheless, feeding dry versus canned food has not been reported to significantly change fecal moisture.²³ The possibility of bacterial contamination of the diets in this study was ruled out by bacterial culture of each diet.

No protease activity can be detected in the feces of untreated dogs with EPI. Detected amounts of amylase and lipase are minimal, and these enzymes are probably not of pancreatic origin.^2,15 In fecal samples in our study, amylase and protease activities were variable, but in many samples, no lipase activity was detected. This is probably a result of destruction of lipase during passage through the intestinal tract. Storage of fecal samples before analyses may have affected results because lipase may be inactivated during storage.¹⁵ The detection of high fecal protease activity in feces of dogs with EPI during enzyme supplementation may be considered an indication that the treatment is adequate and that consumption or destruction of enzymes during intestinal passage is not excessive. However, in the present study, there was no correlation between fecal protease activity and EPI indices. Conversely, a significant negative correlation between fecal lipase activity and fecal fat excretion was observed, a finding that might indicate that orally administered enzyme supplementation should be increased if a high-fat diet is being consumed. A dietary influence on fecal enzyme excretion was also observed in that the lowest values for fecal activities of protease, amylase, and lipase were detected in dogs fed the high-fiber diet. Dietary fiber is known to bind to pancreatic enzymes, and certain types of fiber may also have decreased duodenal pH, thus enhancing enzyme inactivation.^24–26

Our findings indicate that by changing the composition of the diet, the severity of some clinical signs of EPI can be decreased. However, responses to diets varied among individual dogs. Large interindividual variation may explain why none of the 4 diets was superior to the others with regard to mean EPI index scores. The most suitable diet for a given dog with EPI is likely best discerned by trial and error; thus, an individually planned feeding regimen is recommended for affected dogs.