Research Institute of Fish Culture and Hydrobiology, University of South Bohemia, Vodnany, Czech Republic.
Br J Nutr. 2009 Dec;102(12):1709-22. doi: 10.1017/S0007114509991279.
Meta-analysis was applied to quantify the effect of replacement of fish meal by soyabean products in diets on fish growth. Measurement of growth in different units among studies required the use of a standardised effect size (Hedges' d). From a total of ninety-nine studies concentrating on fish meal replacement by defatted soyabean meal, 53 % were eliminated due to, among others, absence of a fish meal control diet (n 18), or no statistical differences or measurement of dispersion (n 6) indicated. Replacement of 4 to 40 % fish meal by soyabean meal (inclusion levels of 71-366 g/kg) resulted in a mean effect size of - 0.1142 (95 % CI - 0.4665, 0.2382) obtained in forty-eight comparisons evaluated with seventeen different fish species. However, at higher fish meal replacement levels the 95 % CI calculated for combined effect sizes did not overlap with zero. With soya protein concentrate replacing 25 to 100 % of fish meal in diets for seven fish species, methionine supplementation (mean - 2.4373 (95 % CI - 3.9004, - 0.9742); n 10) did not have a substantial influence on the magnitude of cumulative effect sizes relative to no supplementation (mean - 2.7306 (95 % CI - 3.7991, - 1.6620); n 16). Information on other soyabean products (full-fat soyabeans, soya flour) used as protein sources in fish diets was found as too limited for analysis and definite conclusions. The present study contributes by putting a numerical value to the magnitude of growth differences in fish when replacing dietary fish meal by soyabean products.
采用荟萃分析量化了用大豆制品替代鱼粉对鱼类生长的影响。由于研究中使用的不同单位的生长测量需要使用标准化效应大小（Hedges' d），因此需要从总共 99 项集中研究脱脂豆粕替代鱼粉的研究中淘汰 53%的研究。淘汰的原因包括缺乏鱼粉对照饲料（n=18），或没有统计学差异或没有进行分散度测量（n=6）等。用大豆粉替代 4%至 40%的鱼粉（添加水平为 71-366g/kg），在 48 次比较中得到了-0.1142 的平均效应大小（95%置信区间为-0.4665，0.2382），这是用 17 种不同鱼类评估的结果。然而，在更高的鱼粉替代水平下，计算的合并效应大小的 95%置信区间与零没有重叠。用大豆浓缩蛋白替代七种鱼类饲料中 25%至 100%的鱼粉时，蛋氨酸补充（平均-2.4373（95%置信区间-3.9004，-0.9742）；n=10）对累积效应大小的大小没有实质性影响与不补充相比（平均-2.7306（95%置信区间-3.7991，-1.6620）；n=16）。关于其他大豆制品（全脂大豆、豆粉）用作鱼类饲料中蛋白质来源的信息，由于分析和确定结论的信息有限，因此无法获得。本研究通过对用大豆制品替代饮食中的鱼粉时鱼类生长差异的幅度进行数值化，做出了贡献。
