Milk saturated fatty acids, odd- and branched-chain fatty acids, and isomers of C18:1, C18:2, and C18:3n-3 according to their duodenal flows in dairy cows: A meta-analysis approach

International audienceWe sought to establish predictive response models of milk fatty acid (FA) yields or concentrations from their respective duodenal flow, rumen digestive parameters, or diet characteristics in dairy cows, with a special focus on cis and trans isomers of C18:1, C18:2, odd- and branched FA, and mammary de novo synthesized FA. This meta-analysis was carried out using data from trials with nature of forage, percentage of concentrate, supplementation of diets with vegetable oils or seeds, and marine products' animal fats as experimental factors. The data set included 34 published papers representing 50 experiments with 142 treatments. Increasing duodenal C18 FA flow induced a quadratic increase in milk total C18 yield and a linear decrease in milk C4:0 to C14:0 concentration. Intra-experirnental predictive response models of individual milk cis C18:1 isomers (Delta 11 to 15 position) from their respective duodenal flows had coefficients of determination (R-2) ranging from 0.74 to 0.99, with root mean square error varying from 0.19 to 0.96 g/d, 0.02 to 0.10% of total FA, and 0.03 to 0.29% of C18 FA. Models predicting milk trans C18:1 isomer yields or concentrations had R-2 greater than 0.90 (except for trans-4 and trans-10 C18:1) with root mean square error varying from less than 0.1 to 5.2 g/d. Linear regressions for C18:2n-6, trans-10,cis-12 CLA, and trans-11,trans-13 CLA were calculated according to their respective duodenal flows. Quadratic models of milk C18:3n-3 yield or concentration from its duodenal flow had R-2 values above 0.97. Models of amounts desaturated from C18:0 into cis-9 C18:1 and trans-11 C18:1 into cis-9,trans-11 CLA indicated that the contribution of C18:0 and trans-11 C18:1 desaturation to respective cis-9 C18:1 and cis-9,trans-11 CLA yields in milk fat was 83.8% (+/- 0.75) and 86.8% (+/- 2.8). Furthermore, when cows were fed marine products, our results could indicate a lower mammary uptake of C18:0 and trans-11 C18:1 in proportion to their respective duodenal flow, with no associated change in mammary Delta(9)-desaturase activity. Yields or concentrations of C15:0, C17:0, iso-C15:0, iso-C17:0, anteiso-C15:0, and anteiso-C17:0 were dependent on their respective duodenal flow or concentration at duodenum, but synthesis of these FA from C3 units for linear-chain odd FA, and from C2 units for branched-chain FA was suggested, respectively. Several milk C18 FA concentrations were closely related to their duodenal concentrations with slopes of the linear models close to the bisector; this could reflect a priority for the use of these duodenal C18 FA by the mammary gland to favor their high concentration in plasma triglycerides arid nonesterified FA, which are preferentially taken up by the mammary gland