Effect of dietary fat to starch content on fecal microbiota composition and activity in dogs1.

Dietary fat is known to modulate the hindgut microbiota in rodents; however, there is no clear evidence on the impact of high-fat diets on canine gut microbiota. The purpose of this study was to investigate the effect of feeding of diets differing in the amount of ME provided by fat and starch on the composition and activity of canine fecal microbiota. Twelve adult (3 to 7 yr of age) spayed Beagle dogs received a low-fat-high-starch diet (LF-HS; approximately 23%, 42%, and 25% ME provided by fat, starch, and CP, respectively) and a high-fat-low-starch diet (HF-LS; approximately 43%, 22%, and 25% ME provided by fat, starch, and CP, respectively) following a 2-period crossover arrangement. The higher amount of fat in the HF-LS diet was provided by lard, whereas the higher amount of starch in the LF-HS diet was provided primarily by maize and broken rice. Each period lasted 7 wk and included 4 wk for diet adaptation. Dogs were fed to meet their daily energy requirements (set at 480 kJ ME/kg BW0.75). Fecal samples were collected on weeks 5 and 6 of each period for the analysis of bacterial richness, diversity, and composition [by Ion-Torrent next-generation sequencing], bile acids, ammonia, and VFA. Additional fecal samples were collected from four dogs per diet and period to use as inocula for in vitro fermentation using xylan and pectin as substrates. Gas production was measured at 2, 4, 6, 9, 12, and 24 h of incubation. On week 7, blood samples were collected at 0- and 180-min postfeeding for the analysis of bacterial lipopolysaccharide (LPS). Feeding the HF-LS diet led to a greater (P < 0.05) fecal bile acid concentration compared with the LF-HS diet. Bacterial richness and diversity did not differ between diets (P > 0.10). However, dogs showed a lower relative abundance of Prevotella (P < 0.01), Solobacterium (P < 0.05), and Coprobacillus (P ˂ 0.05) when fed of the HF-LS diet. Fecal ammonia and VFA contents were not affected by diet (P > 0.10). Relative to the LF-HS diet, in vitro fermentation of xylan using feces of dogs fed the HF-LS diet produced less gas at 6 h (P < 0.01) and 9 h (P < 0.05). Blood LPS did not increase at 180-min postfeeding with either diet (P < 0.10). These findings indicate that feeding a HF-LS diet to dogs does not affect bacterial diversity or fermentative end products in feces, but may have a negative impact on Prevotella and xylan fermentation.

Prediction of crude protein digestibility of animal by-product meals for dogs by the protein solubility in pepsin method.

Animal by-product meals have large variability in crude protein (CP) content and digestibility. In vivo digestibility procedures are precise but laborious, and in vitro methods could be an alternative to evaluate and classify these ingredients. The present study reports prediction equations to estimate the CP digestibility of meat and bone meal (MBM) and poultry by-product meal (PM) using the protein solubility in pepsin method (PSP). Total tract CP digestibility of eight MBM and eight PM samples was determined in dogs by the substitution method. A basal diet was formulated for dog maintenance, and sixteen diets were produced by mixing 70 % of the basal diet and 30 % of each tested meal. Six dogs per diet were used to determine ingredient digestibility. In addition, PSP of the MBM and PM samples was determined using three pepsin concentrations: 0·02, 0·002 and 0·0002 %. The CP content of MBM and PM ranged from 39 to 46 % and 57 to 69 %, respectively, and their mean CP digestibility by dogs was 76 (2·4) and 85 (2·6) %, respectively. The pepsin concentration with higher Pearson correlation coefficients with the in vivo results were 0·0002 % for MBM (r 0·380; P = 0·008) and 0·02 % for PM (r 0·482; P = 0·005). The relationship between the in vivo and in vitro results was better explained by the following equations: CP digestibility of MBM = 61·7 + 0·2644 × PSP at 0·0002 % (P = 0·008; R (2) 0·126); and CP digestibility of PM = 54·1 + 0·3833 × PSP at 0·02 % (P = 0·005; R (2) 0·216). Although significant, the coefficients of determination were low, indicating that the models were weak and need to be used with caution.