The water-soluble, non-starch polysaccharides (NSP) occurring in rye, wheat and barley are believed responsible for the reduction of growth performance and the digestibility of lipids, protein and starch in broiler chickens fed these feedstuffs (Choct and Annison 1992a, Fengler and Marquardt 1988, White et al. 1981). On the basis of a literature review, Annison (1993) proposed that dietary soluble NSP inhibit nutrient absorption in broiler chickens not only by raising the viscosity of the digesta but also by enhancing bacterial fermentation. It is not known how the effect of soluble NSP is mediated by enhanced bacterial fermentation, but the addition of antibiotics to the diet of broiler chickens can diminish the inhibitory effect of NSP on nutrient absorption (Misir and Marquardt 1978).

Viscous plant NSP generally are readily fermentable (Roberfroid 1993), and thus it is difficult to assess separately the antinutritive effects of viscosity and fermentability. The effect of viscosity per se can be studied by the use of non-fermentable carboxymethylcellulose (CMC) preparations with different degrees of viscosity. Thus, in the present study water-soluble CMC with various polymer lengths were fed to broiler chickens to investigate their effect on growth performance and on digestibility of macronutrients.

MATERIALS AND METHODS

Table 1. Composition of the experimental diets [View Table]

1

For the determination of the in vitro fermentability of the CMC preparations, a 32 g/100 g fecal slurry was made by using fresh human feces and sodium phosphate buffer (0.1 mol/L, pH 6.5). The slurry was kept under a continuous flow of nitrogen gas. Five milliliters of the slurry was mixed with 5 mL of a sodium phosphate buffer (0.1 mol/L) without addition or containing 2 g/100 mL citrus pectin, LCMC or HCMC. The citrus pectin served as a positive control and the buffer without addition as a negative control. For each condition, five plastic tubes were used to determine the fermentation products after 0, 1, 3, 6 and 24 h. The incubation tubes were placed in a water bath at 37°C. At each time interval, one tube was removed and centrifuged for 10 min at 4260 × g and 4°C, and the supernatant was collected and stored at 20°C. The amounts of acetic, propionic, butyric and valeric acids in the supernatant were determined by the method of Tangerman et al. (1983).

The viscosity of nondiluted supernatants prepared from the contents of duodenum plus jejunum and ileum was measured using a cone-plate viscometer (Brookfield digital DV-II+, Brookfield Engineering Labs, Stoughton, U.K.) maintained at 40°C. The values were recorded at a shear rate of 45 s¹. Viscosity was expressed as milli-Pascal seconds (mPa·s). The pH of intestinal contents was determined immediately after collection (PHM 82 standard pH meter, Radiometer Nederland, Zoetermeer, The Netherlands). Plasma triglyceride and cholesterol concentrations were determined with the use of commercial test kits (CHOD-PAP and GPO kits, Boehringer-Mannheim GmbH, Mannheim, Germany). Chromium in 1-g samples of food and freeze-dried chyme and excreta was determined by atomic absorption spectrophotometry (SpectrAA-10, Varian Nederland BV, Houten, The Netherlands) after the samples had been ashed at 550°C and oxidized by the addition of 6 mL of potassium bromate (30 g/L) and 3 mL of a solution containing magnesium sulfate (0.67 g/L) and orthophosphoric acid (824.5 g/L). Total lipid contents of food and excreta were determined by extraction of the samples with hexane in a Sohxlet tube after they had been boiled in hydrochloric acid (3 mol/L) for 30 min. The determination of nitrogen in food, ileal digesta and excreta was performed with the Kjeldahl method. Fecal nitrogen in the excreta was calculated as total nitrogen minus nitrogen in uric acid. Uric acid was analyzed by the method of Terpstra and De Hart (1974). Starch in food and ileal digesta was analyzed according to the NEN 3574-procedure (NNI 1974). The concentration of ATP in chyme was estimated by the luciferin-luciferase (EC 1.13.12.7) method as described by Bach-Knudsen et al. (1991), with the use of ATP monitoring kits (Microbial Biomass Testkit Lumac 93221-1, ATP Standard Lumac 9263-8 Perstorp Analytical, Oud Beijerland, The Netherlands) and a luminometer (Lumac Biocounter M1500, Perstorp Analytical, Oud Beijerland, The Netherlands).

RESULTS

Table 2. Viscosity of supernatant, pH and ATP concentrations of digesta in broiler chickens fed diets with low (LCMC) or high (HCMC) viscous carboxymethylcellulose1 [View Table]

The pH of the intestinal contents was not affected by the type of CMC. Consumption of HCMC vs. LCMC produced a rise in ATP levels, but the difference only approached significance for the chyme of duodenum plus jejunum (P = 0.062).

The feeding of HCMC instead of LCMC raised water intake and reduced weight gain (Table 3). The feed to gain ratio was significantly elevated after consumption of the HCMC diet compared with the LCMC diet.

Table 3. Weight gain, food intake, water intake and feed conversion ratio of broiler chickens fed diets with low (LCMC) or high (HCMC) viscous carboxymetylcellulose from 21 to 35 d of age1 [View Table]

The full and empty weights of small intestine and colon, expressed relative to body weight, were significantly raised by HCMC vs. LCMC in the diet (Table 4). Moreover, HCMC consumption, when compared with LCMC, produced an increase in length of the small intestine, caeca and colon. The full weight of the caeca was not significantly influenced by the type of CMC in the diet, but the empty weight was significantly elevated. Pancreas and liver weight were not different for broilers fed the diets with either HCMC or LCMC.

Table 4. Weights of the intestinal segments, pancreas and liver in broiler chickens fed diets with low (LCMC) or high (HCMC) viscous caroxymethylcellulose1 [View Table]

Consumption of HCMC instead of LCMC resulted in lower apparent ileal digestibility of starch and lower apparent fecal digestibility of nitrogen and lipids (Table 5). Ingestion of HCMC rather than LCMC tended (P = 0.077) to result in lower group mean plasma concentrations of triglycerides (0.44 ± 0.10 compared with 0.58 ± 0.11 mmol/L). Plasma cholesterol concentrations for the HCMC and LCMC groups were 2.45 ± 0.42 and 2.93 ± 0.18 mmol/L, respectively (P = 0.183).

Table 5. Digestibility of nitrogen, starch and lipids in broiler chickens fed diets with low (LCMC) or high (HCMC) viscous carboxymethylcellulose1 [View Table]

DISCUSSION

When compared with LCMC, HCMC had an antinutritive effect in broiler chickens. Consumption of HCMC depressed growth performance, raised water intake and inhibited macronutrient digestion. Similar results were obtained earlier in broiler chickens after consumption of isolated viscous NSP of plant origin, such as water-soluble wheat pentosans (Choct and Annison 1992a), water-soluble rye pentosans (Antoniou and Marquardt 1981, Fengler and Marquardt 1988) and barley -glucans (White et al. 1981). Our data indicate that a raised viscosity per se of intestinal contents may be associated with the antinutritive effects.

Consumption of HCMC compared with LCMC depressed the apparent digestibilities of starch, nitrogen and lipids. The mechanism by which an increase in digesta viscosity may affect macronutrient digestion is not clear. Possible processes involved are a reduction of the diffusion rate of digestive enzymes and bile acids (Ebihara and Schneeman 1989, Isaksson et al. 1982) and less mixing of chyme (Edwards et al. 1988). The latter may result in less contact of digesta with the absorptive surface. In the contents of duodenum plus jejunum from chickens fed HCMC, the ATP concentrations were elevated, pointing to increased microbial populations. Because HCMC itself is not fermented, the rise in microbial activity could be due to accumulation of nondigested material. In addition, an increase in viscosity of digesta may raise retention time (van der Klis et al. 1993), which would also stimulate fermentation. The primary effect of HCMC may be the elevation of the viscosity of intestinal digesta, causing depressed macronutrient digestion and enhanced bacterial fermentation. The material sustaining fermentation consists at least of carbohydrates and/or protein rather than lipids, because the latter compounds will not serve as bacterial substrates.

Lipid digestibility was more depressed by HCMC than was the digestibility of protein or starch. A similar pattern was noted in birds fed diets containing isolated water-soluble pentosans (Choct and Annison 1992a and 1992b, Fengler and Marguardt 1988). In young birds, the low bile acid concentration in the digesta may limit lipid absorption (Iñarrea et al. 1989, Polin et al. 1980). Various authors have suggested that the low lipid digestibility in broiler chickens fed diets with a high content of gelling NSP may be due to bacterial overgrowth in the small intestine and subsequent excessive deconjugation of bile acids, which reduces their efficacy in solubilizing lipids (Huhtanen and Pensack 1965, Salih et al. 1991). A marked increase in microbial cholytauryl hydrolase activity was observed in chicks fed rye (Feighner and Dashkevicz 1988). This observation could imply that a high viscosity of intestinal digesta may not be the primary cause for the lowered lipid digestion seen after HCMC consumption. However, it cannot be excluded that a viscosity-induced lowering of diffusion of bile acids through the digesta had a major effect on lipid digestion in birds fed HCMC. Addition of sodium taurocholate to a rye-based diet improved apparent lipid digestibility in broiler chickens (Campbell et al. 1983). In HCMC-fed birds there might be a lowered bile acid efficacy in the digesta, which could then explain not only the low lipid digestibility but also the low plasma triglyceride and cholesterol concentrations as secondary features. Increased viscosity of intestinal contents reduced cholesterol absorption and plasma cholesterol concentration in hamsters fed hydroxypropyl methylcellulose with varying viscosity (Carr et al. 1996, Gallaher et al. 1993).

Volatile fatty acids and polyamins produced by gut bacteria have stimulatory effects on the proliferation rate and secretory activity of intestinal mucosa (Furuse et al. 1991, Osborne and Seidel 1989, Sakata 1987). In HCMC-fed birds, with an increase in microbial fermentation there would be more loss of endogenous nitrogen, leading to a reduction of apparent nitrogen digestibility, as was seen. Angkanaporn et al. (1994) demonstrated that water-soluble pentosans in the diet significantly raised the endogenous losses in broiler birds. Larsen et al. (1993) found an increase in endogenous nitrogen loss when highly viscous CMC was added to the diets of rats, whereas ileal true nitrogen digestibility was not affected. Thus, viscous CMC did not influence nitrogen digestibility in those rats, at least not at the level of the terminal ileum. The increase in bacterial fermentation in the broilers fed HCMC may have been relatively small. There was a rise in ATP concentration only in the duodenum-jejunum, with no lowering of the pH of the digesta. Although part of the observed hypertrophic effect of HCMC on the small intestine, caeca and colon could have been the result of the increase in microbial fermentation, HCMC itself could also have caused the effect. Recently, Pell et al. (1995) observed a trophic effect on the small intestinal length and cecal weight in germ-free mice fed a diet with guar gum. Thus, the trophic effect of HCMC on the length of the small intestine may have been unrelated to the enhanced microbial fermentation. Extra losses of endogenous nitrogen could also be of pancreatic origin, because diets containing viscous fibers have been shown to raise the secretion of pancreatic juice in rats (Ikegami et al. 1991).

In conclusion, the feeding of HCMC instead of LCMC to broiler chickens lowered the apparent digestibility of starch, protein and lipids. The HCMC raised the viscosity of digesta, which may interfere with starch and lipid digestion. The accumulation of undigested carbohydrates would stimulate microbial fermentation. The observed reduction in apparent protein digestibility could have been the result of greater losses of endogenous nitrogen.

ACKNOWLEDGMENT