Association between diet composition and in vitro selenium bioaccessibility in pet foods

Introduction: The extremely variable levels of Se in raw materials can make it difficult to formulate pet foods that meet the current recommended allowance of 17.9 µg Se/MJ ME and stay below the legal upper limit of 33.9 µg Se/MJ ME (FEDIAF, 2012). Since many factors influence Se availability in other species and information on these factors in pet food is scarce, a dietary recommendation based on total dietary Se may not be accurate enough. Therefore, the aim of this study is to identify dietary factors associated with in vitro bioaccessibility (BAiv) of Se in dog and cat foods. Materials & Methods: Sixty-two pet diets (cat (n=10) and dog (n=52)) were selected, of which fifty-four were commercially available and eight were unprocessed diets. The unprocessed diets (kibble (n=4); can (n=4)) came from the same batch as the corresponding processed diet. Diets were analysed for DM, ash, CP, Cfat, S, AA profile, TDF, GE, and total Se. An in vitro digestion trial (adapted from Hervera et al., 2007) was performed. Residues (undigested fraction) were analysed for DM, ash, and CP and filtrates (digested fraction) for total Se. Data was analysed using the GLM and REG procedures of SAS, version 9.3. Results & Discussion: Total dietary selenium concentration was on average 0.98 mg/kg DM (range: 0.18-11.96). In the complete dataset only a significant correlation was found between Se BAiv and processing type (p=0.0002). Canned and steamed meat diets had a significantly lower Se BAiv (58% (29-91) and 47% (39-54), resp.) than pellets and raw meat diets (79% (62-101) and 91% (79-98), resp.). These findings indicate an effect of processing on Se BAiv. Among the extruded diets (n=19) a positive correlation was found between Se BAiv (72% (37-97)) and protein digestibility, which is in accordance with results in milk for human consumption (Shen et al., 1996). The opposite was found for canned diets (n=16), which also might express an effect of processing, or differences in Se speciation between the diets, on which processing may have a different effect. It might also be due to an effect of Maillard reactions, which yield volatile seleniferous compounds (Tsai et al., 1998), or AA cross-linking, although there was no difference in total Se and protein digestibility, respectively, between the unprocessed and processed diets. This study showed a difference in Se BAiv between processing types. In vivo work should validate these in vitro findings and show if recommendations on Se inclusion levels in complete and balanced diets need to take type of dietary processing into account