Evaluation of feed protein quality by measuring plasma free amino acids in Atlantic salmon (Salmo salar L.) after dorsal aorta cannulation

Two successive experiments were conducted in order to assess plasma free amino acid (FAA) profiles as a method for evaluating protein quality of fish feeds for Atlantic salmon, Salmo salar (L.). In experiment 1, the importance of meal size and inter‐fish variation was assessed by using dorsal aorta cannulated fish and diets that contained different sources of fishmeal (menhaden versus herring) which in the case of herring, had been dried at either 70 or 100 °C. In experiment 2, an attempt was made to mimic a production situation by comparing the FAA profiles in salmon fed diets containing two commercially available fishmeals that had been produced in accordance with industrial standards (Norse‐LT94® and NorSeaMink®; Norsildmel AL, Fyllingsdalen, Norway). FAA profiles in plasma 6 h after feeding were compared with feed true protein digestibility as determined in mink. Cannulated fish, held in individual tanks, were hand‐fed twice daily to pellet rejection (satiety) and daily records of the actual rations consumed were maintained. A total of 24 different amino acids and other amino‐containing compounds were detected using high‐pressure liquid chromatography. Morning and evening meal size showed significant correlations. Meal size had a significant effect on blood levels of the majority of essential free amino acids (EAA) as well as the total sum of FAA (TFAA). In experiment 1, a marked inter‐individual effect was found, possibly because of incipient sexual maturation. FAA profiles were therefore corrected for meal size by linear regression while repeated sampling via the permanently implanted cannula allowed paired comparisons of the different test diets, minimizing inter‐individual variation. Significant differences in plasma FAA profile, EAA and TFAA were detected between fish fed all diets in both experiments. The preceding parameters for fish ingesting each feed were directly related to their respective mink protein digestibility in experiment 1, but not in experiment 2. Our results show that dietary protein quality can be differentiated by the aforementioned protocol, and by using fish with a low metabolic rate feed qualities could be ranked correctly as in experiment 1.