| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
ANIMAL NUTRITION |
* Sao Paulo State University, Faculty of Agrarian and Veterinary Sciences, Jaboticabal, SP 14884-900, Brazil; and
University of Sao Paulo, Faculty of Veterinary Medicine and Animal Sciences, Sao Paulo, SP 05508-270, Brazil
 |
Abstract |
|---|
 Top Abstract INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Key Words: digestion • feline • ingredient • meal response test • starch
|
INTRODUCTION |
|---|
|
TopAbstractINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
; Morris et al., 1977; Wilde and Jansen, 1989; Kienzle, 1994b).
Although both cats and dogs belong to the order Carnivora, significant nutritional and metabolic differences have been shown between the 2 species (Kienzle, 1993; Morris, 2001). Cats are considered to exhibit decreased apparent nutrient digestibility than their canine counterparts (Kendall et al., 1982). Some investigations showed that cats are unable to metabolize sugars in large quantities (Kienzle, 1993, 1994b; Washizu et al., 1999). However, few studies have looked at the digestibility of cat foods containing different starch sources.
It is accepted that carbohydrates, primarily starches, are the principal nutrients that determine and modify the postprandial glucose and insulin curves in dogs and humans (Nguyen et al., 1994; Wolever and Bolognesi, 1996). For cats, however, the few studies carried out indicate comparatively small serum glucose and insulin variations after the consumption of different starch and sugar types (Kienzle, 1994a; Bouchard and Sunvold, 2000; Appleton et al., 2004). Therefore, the objective of the present research was to investigate the effects of cassava flour, brewers rice, corn, sorghum, peas, and lentils on the digestibility and postprandial glucose and insulin responses of healthy cats.
|
MATERIALS AND METHODS |
|---|
|
TopAbstractINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Animals
Thirty-six mixed-breed, neutered, male and female cats, aged 4.9 ± 0.7 yr and weighing 4.34 ± 0.76 kg [i.e., not obese (body condition score between 4 and 6; Laflamme, 1997)], were used in the digestibility and meal response tests. The cats were kept in the Laboratory of Nutrition and Nutritional Diseases at Sao Paulo State University (Jaboticabal, Brazil). During the digestibility and postprandial response experiments, the cats were individually housed in 0.9 x 0.8 x 0.9 m stainless steel metabolic cages. Water was available ad libitum throughout the duration of the experiment.
Diets
In total, 6 diets were tested, and the ingredient composition of each is reported in Table 1. Each diet incorporated 1 of 6 carbohydrates as its exclusive source of starch: corn, brewers rice, sorghum, peas, lentils, or cassava flour. Due to the chemical composition of the carbohydrate sources being evaluated, additional ingredients were incorporated to obtain diets that contained percentages (DM basis) of starch, fat, calcium, and phosphorus as balanced as possible. Isolated soybean protein, poultry by-product meal, and poultry fat were used to equalize the formulations. All diets contained the same amount of added salt, vitamins, and trace minerals. The amount of total dietary fiber (TDF) varied according to the concentration of these nutrients found in the carbohydrate sources (Table 2). Diets were formulated in accordance with the AAFCO (2003) nutrient guide for cats and balanced to meet maintenance requirements before being extruded and kibbled under identical processing conditions in a single-screen extruder at the College of Agricultural and Veterinarian Sciences, Sao Paulo State University (Mab 400S, Extrucenter, Monte Alto, Brazil). The food manufacturing quality was controlled every 20 min by adjusting the density (g/L) of each food preparation to achieve the same kibble parameters (i.e., size, expansion) and, indirectly, the same gelatinization (Table 2).
|
|
The 36 cats were divided into 3 groups of 12 animals each, allowing 2 cats to be fed the same diet within each group. The experiment was conducted in 3 periods with 1 group evaluated during each period. The digestibility assay was carried out through quantitative collection of feces, according to AAFCO (2003) guidelines. A 10-d test-diet adaptation phase preceded a 10-d collection of feces in each experimental period. The quantity of diet provided was calculated using standard equations that determine proper energy requirements for cat maintenance (ME, kcal = 70 x kg of BW), in accordance with NRC (1986) data. Each day, food was weighed and divided into 2 equal portions, placed in stainless steel bowls, and left out at 0900 and 1700 h. Bowls were removed before the next meal, and any remaining food was weighed and recorded. On the first day of fecal collection, all feces were removed from the cages and discarded before 0800 h. Fecal output was collected from this point on for the next 10 d at each mealtime. Samples were frozen (–15°C) as they were collected and pooled by cat.
Fecal samples were scored according to the following system: 1 = watery – liquid that can be poured; 2 = soft, unformed – stool assumes shape of container; 3 = soft, formed, moist – softer stool that retains shape; 4 = hard, formed, dry stool – remains firm and soft; 5 = hard, dry pellets – small, hard mass. Fecal pH was determined by mixing 10 mL of distilled water with 5 g of feces and measuring the result with a pH meter (model Q-400-Bd, Quimis, Diadema, Sao Paulo, Brazil).
Postprandial Response Tests
After the digestibility trials, cat postprandial glucose and insulin responses were measured using a sampling protocol based on Appleton et al. (2004), minus the samples at 14, 16, and 18 h after the meal. As previously stated, the 36 cats were divided into 3 groups of 12 cats each, with 2 cats receiving the same diet within each group. One group was evaluated per period. The cats were allowed to adapt to their diets for 7 d before samples were taken, and during this period, the animals were fed once per day and conditioned to ingest their food within 15 min. Subsequently, the cats were deprived of food for 24 h before the initiation of the glucose-insulin response. Two days prior, each cat had been aseptically catheterized under light sedation with levomepromazine (Neozine, Aventis, Morumbi, Sao Paulo, Brazil), with a central venous catheter (Intracath, 30.5 cm, 1.1 mm, Becton Dickinson Vascular Access, Sandy, UT) inserted into a jugular vein. Catheters were flushed twice daily with dilute, heparinized saline (20 IU/mL) to maintain patency. Blood samples were taken prefeeding (baseline sample, time 0) and 1, 2, 4, 6, 8, 10, and 12 h postfeeding (the times were counted from the end of the meal). Cats were fed their total daily energy requirement (NRC, 1986) and allowed a maximum of 15 min to consume their diets; only those that finished their allotted diet in 15 min were tested. Blood was collected at the same time, beginning at 0700 h. Each 1.5-mL blood sample was taken using a syringe and transferred to a Na-heparin tube, centrifuged (2,000 x g for 5 min), and the plasma was divided equally into 2 Eppendorf tubes. Plasma samples for glucose measurement were kept under refrigeration (4°C) for a maximum of 2 h before analysis; plasma samples for analysis of insulin were frozen (–70°C) for a maximum of 2 mo before they were analyzed.
Laboratory Analyses
At the end of the collection period, feces were thawed, homogenized, and pooled by cat. Before performing laboratory tests, feces were dried in a forced-air oven at 55°C for 72 h (Fanem, Sao Paulo, Brazil) and ground in a cutting mill with a 1-mm sieve. Food samples were ground in a similar fashion. Diet and feces were analyzed for DM, OM, ash, CP (Kjeldahl method), acid-hydrolyzed fat, phosphorus, and calcium using AOAC (1995) methods.
Determination of TDF was carried out according to Prosky et al. (1992). Gross energy content of diets and fecal matter was determined using a bomb calorimeter (model 1261, Parr Instrument Company, Moline, IL). The total amount of starch was analyzed according to the guidelines set out by Miller (1959) and Hendrix (1993), whereas amylose and amylopectin contents were determined following the methodology of Knutson (1986). All analyses were carried out in duplicate, with a CV below 5%.
Plasma glucose concentrations were determined by a glucose oxidase test (GOD-ANA, Labtest Diagnóstica S.A., Lagoa Santa, Minas Gerais, Brazil) using a semi-automated glucose analyzer (Labquest model BIO-2000, Labtest Diagnóstica S.A.). Plasma insulin was measured by RIA using a commercially available kit (human insulin as a standard; Diagnostic Products Corporation, Los Angeles, CA; I125 as tracer) that was validated for cats (Nelson et al., 1990). The intraas-say CV for insulin was 7.6%, and the SE was 0.12 µIU/ mL.
Calculations
Nutrient apparent digestibility values were determined for each experimental diet. Changes in plasma glucose and insulin concentrations were calculated for each postprandial period. Responses were compared for the average and maximum increase, the average and maximum incremental increase (the difference between the absolute glucose or insulin concentration of the sample and the baseline concentration), and the time to peak increase. The integrated area under the postprandial glucose and insulin response curves and the integrated area under the incremental postprandial glucose and insulin curves were calculated by the trapezoidal method. The software ORIGIN (Microcal Software Inc. Northampton, MA) was used for area under curve computing.
Statistical Analyses
Data were analyzed in a completely randomized block design using the GLM procedures (SAS Inst. Inc., Cary, NC). The individual cat was considered the experimental unit. Model sums of squares were separated into diet, period (blocks), and animal effects. Where significant (P < 0.05), differences were detected in the ANOVA F test for nutrient intakes, feces characteristics, or apparent total tract nutrient digestibilities; multiple comparisons of means were made using Tukey’s test (P < 0.05). Repeated-measures ANOVA with 2 among-animals factors (diet and period) and 1 within-animals factor (time of sampling) was the statistical method chosen to evaluate the effects of diet and time on postprandial plasma changes. Pairwise means comparisons were also made through Tukey’s test (P < 0.05) when the ANOVA F-test results were statistically significant. All data complied with the assumptions of ANOVA models.
|
RESULTS |
|---|
|
TopAbstractINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
The chemical compositions of the starch sources and experimental diets are given in Table 2
. The DM and OM ingredient composition were similar for all sources. Crude protein concentration was greatest in peas and lentils and lowest in cassava flour. Fat concentration was greatest in corn and sorghum. Regarding carbohydrate composition, cassava flour and brewers rice had the greatest amounts of starch; lentils and peas had the greatest concentrations of fiber.
All diets contained comparable concentrations of DM and OM. Mean starch concentration was 35.1%, and the sorghum-based diet contained 5% more starch than the pea diet. The concentrations of CP varied due to its usage in completing the formulations, and the fiber varied according to its concentration in the tested ingredients. Regarding food processing, kibble densities were comparable among the experimental diets, especially after drying (Table 2), indicating similar cooking.
Digestibility
Daily nutrient intake, apparent total tract digestibility values, and cat fecal characteristics are presented in Table 3. Differences in TDF and fat ingestion among diets were verified (P < 0.05) and can be explained by differences in food composition. Variations in fat ingestion were low and likely did not interfere with digestibility determinations. Ingestion of the other nutrients was similar among diets.
|
Regarding fecal evaluation, no differences were observed in fecal scores, pH, feces excretion (DM basis), or fecal output. Fecal DM was greater for cats fed the brewers rice diet than the cassava flour, sorghum, lentil, and pea diets (P < 0.05). Wet fecal production was lower for cats fed the brewers rice diet than the lentil, cassava, sorghum, and pea diets (P < 0.05).
Postprandial Responses of Glucose and Insulin
There were no differences in ingestion of starch, DM, OM, CP, and ME (Table 4). Fat ingestion was lower for cats fed the lentil diet compared with the brewers rice diet (P < 0.05). Cats fed the sorghum, lentil, and pea diets showed greater TDF ingestion than those fed the brewers rice and cassava flour diets (P < 0.05).
|
, the mean glucose concentration, the maximum glucose concentration, and the area under the glucose curve were greater for the corn-based than for the cassava flour, sorghum, lentil, and pea diets (P < 0.05). For the mean incremental glucose concentration, the maximum incremental glucose concentration, and the area under the incremental glucose curve, the corn-based diet was greater than the lentil diet only (P < 0.05). When compared with baseline values, only the corn diet stimulated a significant increase in the glucose response, occurring at 4 and 10 h post-meal (P < 0.05), as shown in Figure 1.
|
|
. |
DISCUSSION |
|---|
|
TopAbstractINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
and Svihus et al. (2005), these differences occur because the proportions of hull, pericarp, aleurone, and germ in the whole grains differ, making their individual chemical compositions unique. Brewers rice is hulled and polished; cassava flour is the residue obtained after hulling, grinding, and removing the soluble starch of the cassava root: this explains the reduced TDF concentrations of these ingredients. The proportion of amylose to amylopectin is also unique to each starch source. Usually, legume starches have greater amylose concentrations than cereal starches (Biliaderis, 1991).
Measurements of digestibility indicate that, despite the low carbohydrate content of the carnivorous diets characteristic of the Felidae, cats can efficiently digest starch-based diets that are adequately ground and cooked. The digestibility values of these diets compare with the starch digestibilities of other species such as dogs and rats (Morris et al., 1977; Holm et al., 1988; Murray et al., 1999).
Most commercial dry, extruded cat foods are composed of rice or corn. Previous studies in cats using cooked, finely ground corn indicated starch digestibilities greater than 90% (Pencovic and Morris, 1975; Morris et al., 1977; Wilde and Jansen, 1989). Studies using dogs also demonstrated that diets based on corn or rice have high digestibilities (Walker et al., 1994; Murray et al., 1999) and that brewers rice-based diets have a similar or greater nutrient digestibility than corn diets (Belay et al., 1997; Twomey et al., 2002). In the current study, brewers rice- and corn-based diets showed similar starch digestibilities in cats.
Although no previous research detailing the effects of cassava flour-based diets on cats was encountered, in dogs, it was shown that the apparent digestibilities of DM, CP, and fat for brewers rice- and cassava flour-based diets do not differ (Kamalu, 1991). These findings were supported in the current study, which characterized these ingredients as being highly digestible by cats.
Lower digestibility values for the lentil- and pea-based diets compared with those for the other starch sources were also described by other authors working with dogs, who attributed the differences between cereals and legumes to the starch and TDF compositions of the ingredients (Bednar et al., 2001). The elevated amounts of resistant and slowly digestible starches found in lentils (Cummings and Englyst, 1995) results in the slower digestion and absorption rates of this legume in humans (Lin et al., 1992).
Among the cereals, including rice and corn, sorghum is considered to have decreased starch digestibility (Svihus et al., 2005), a finding confirmed in our study. According to Rooney and Pflugfelder (1986), the protein matrix of the hard outer endosperm closely surrounds sorghum starch; this complex interaction between protein and starch restricts digestibility. Additionally, antinutritional substances such as tannin can contribute to enzymatic inhibition, which may delay sorghum starch digestion.
Studies on starch usually separate ileal from total tract digestibility, because the fraction of starch that escapes enzymatic digestion in the small intestine can be fermented by the microbiota in the large intestine, leading to overestimates of digestibility. However, Kienzle (1993) showed that amylase activity in the gastrointestinal tract of cats is mainly of bodily origin, which led to the conclusion that there is little interference with total apparent starch digestibility.
Another concern is the diet adaptation period. Kienzle (1993) observed that cats fed starch-rich diets early in life show greater amylase activity than cats first fed the same diets as adults. In the current study, all cats were given commercial dry-extruded foods rich in starch since birth, likely contributing to the elevated digestibilities encountered. Finally, another factor to be considered is the food processing itself (Riaz, 2003), in which the milling and extrusion of experimental diets may affect the utilization of the starches.
Measurements of postprandial responses indicate that differences among diets were not as pronounced for cats as those normally observed for dogs and humans. Existing studies are still inconclusive regarding the effect of starch on cat postprandial glucose and insulin responses; large individual variations of these characteristics have been encountered (Appleton et al., 2004). Data obtained in this study, like in other studies with this species (Kienzle, 1994a; Bouchard and Sunvold, 2000), suggest that starch has only a minor effect on the glucose and insulin responses of cats. This may be explained by the metabolic peculiarities of felines, among them the preferential usage of AA as a source of energy and lower enzymatic capabilities to digest starch and to metabolize dietary glucose (Kienzle, 1993, 1994a; Washizu et al., 1999; Morris, 2001), which can slow and prolong starch digestion and absorption. Reinforcing this observation is the time required for glucose elimination after an intravenous or oral glucose tolerance test, which is prolonged in cats compared with dogs and humans (Kienzle, 1994a; Bouchard and Sunvold, 2000; Appleton et al., 2004).
Increased concentration of glucose and insulin, as observed in the digestibility experiment, depends on an integrated evaluation of the diet, including consideration of factors intrinsic to starch, such as the digestion rate, the amylase:amylopectin ratio, and the amount of resistant starch and also extrinsic factors such as processing method, diet composition, and the amount ingested (Brand et al., 1985; Heaton et al., 1988; Wolever and Bolognesi, 1996; Nguyen et al., 1998). In the current study, diets were formulated to have similar chemical compositions. Variations in DM, OM, starch, and CP ingestion were small and not significant. There were differences between the fat and TDF ingestion for different diets. For humans, Wolever and Bolognesi (1996) have suggested that in practical diets, the apparent effects of protein and fat on the glucose response would be negligible. Interestingly, the quantity of starch ingested corresponds to between 46 and 64% of the glucose response variation, being at times even more important than the type of starch consumed. In the present study, ingestion of starch during the meal-response testing ranged from 7.1 to 10.2 g/kg of BW0.67 for all diets. This is not a significant difference; however, it could reduce the glucose and insulin responses of cats fed the lentil and pea diets.
Regarding food processing, the ingredients were ground in the same milling machine with the same sieve, and the degree of cooking was controlled by means of kibble density, which was similar among diets. However, starch cooking is an extrinsic factor that can influence starch digestibility and the glucose response. For example, if the gelatinization index (not measured here) was different among the diets, it could have affected the experimental results.
Dietary fiber content is an additional factor that can alter the postprandial glucose and insulin responses (Wolever, 1990; Graham et al., 1994). The sorghum-, lentil-, and pea-based diets resulted in the greatest ingestions of TDF, which may have played a role in delaying and prolonging the glucose absorption period as well as lessening the variation in glucose and insulin concentration. Some fiber types (e.g., soluble fiber) slow gastrointestinal transit and gastric emptying and decrease starch hydrolysis and, consequently, the rate of glucose absorption. However, some studies have demonstrated that for diets with usual levels of fiber, variations in the intake of this ingredient did not significantly affect postprandial responses (O’Dea et al., 1980; Nguyen et al., 1998). Therefore, it is conceivable that factors other than fiber are responsible for the slower, prolonged glucose and insulin responses observed for diets incorporating lentils and peas.
Other studies have shown that the amylose content of starch is directly related to starch digestibility (Biliaderis, 1991). In contrast, the glucose response is inversely related to the amylose content of starch (Goddard et al., 1984). Normally, legume starches have high concentrations of amylose and, consequently, are more likely to form resistant starch. Therefore, legume starches may provoke lesser glucose and insulin responses than cereal starches in animals and humans (Lee et al., 1985). Moreover, legume starches have only small concentrations of free sugars and rapidly digestible starch. The slow and incomplete digestion of legume starches is probably related to properties of the starch granule (e.g., amylose:amylopectin ratio) and its physical association with the plant cell wall (fiber), which contribute to reducing total starch gelatinization compared with that of the cereal starches (Englyst and Hudson, 1996; Englyst et al., 1996). The current study confirmed that cats fed lentil and pea diets (2 leguminous plants) showed decreased blood glucose and increased insulin compared with the other diets, with the exception of the sorghum diet.
Rice is classified as having a high glycemic index (Jenkins et al., 1981; Goddard et al., 1984), leading to large and rapid alterations in blood glucose and insulin concentrations in humans. This characteristic is due to its small quantity of amylose (Belay et al., 1997), associated with a low concentration of TDF. Studies using cats that compared brewers rice with corn (Bouchard and Sunvold, 2000) or brewers rice with a corn-sorghum blend (Appleton et al., 2004) also observed that brewers rice was capable of stimulating greater blood glucose and insulin concentrations. In our study, however, the brewers rice- and corn-based diets presented similar results for both the glucose and insulin measurements (Table 5), and only corn provoked a rise in glycemia in relation to the baseline values (Figure 1). For insulin, both corn and brewers rice caused an increase in blood concentration in relation to the baseline values.
Bouchard and Sunvold (2000) did not encounter differences in the glucose and insulin responses in a comparison of corn with sorghum. In our study, we observed a lesser glucose response for sorghum compared with corn (mean concentration, maximum concentration, and area under curve) and a similar insulin response for both. Studies using dogs also revealed a small postprandial glucose response for sorghum diets (Bouchard and Sunvold, 1999). The decreased response to sorghum may be explained by the characteristics of the starch granule of this grain (Rooney and Pflugfelder, 1986; Svihus et al., 2005) and possibly by the increased ingestion of TDF allowed by this diet; however, studies of sorghum fiber and glycemia in cats were not found in the literature.
The current study showed that all sources of starch showed satisfactory digestibilities, indicating that they can be used in the formulation of cat food. The brewers rice and cassava flour diets had the greatest digestibilities, whereas the lentil and sorghum diets had the least. Changes in blood glucose concentrations were greater for the corn diet compared with the cassava flour, sorghum, pea, and lentil diets. However, only corn increased the blood glucose concentration over the baseline values. However, differences in blood insulin response were not so evident among the diets, occurring only between the corn and pea diets. Nevertheless, in addition to corn, brewers rice, pea, and sorghum also raised the postprandial insulin concentration above the baseline value for at least 1 point on the response curve. Therefore, more studies are necessary to evaluate the effectiveness of using different starch sources for glucose or insulin modulation in the cat.
|
Footnotes |
|---|
2 Corresponding author: aulus.carciofi{at}gmail.com
Received for publication June 14, 2007. Accepted for publication May 2, 2008.
|
LITERATURE CITED |
|---|
|
TopAbstractINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Values greater than baseline for glucose concentrations (P < 0.05).