Herman Rod A, Ekmay Ricardo
Dow AgroSciences LLC, 9330 Zionsville Road, Indianapolis, IN 46268, United States.
Regul Toxicol Pharmacol. 2014 Feb;68(1):171-4. doi: 10.1016/j.yrtph.2013.07.003. Epub 2013 Jul 11.
The use of whole-food (grain meal contained in feed) animal-feeding studies to support the safety assessment of genetically modified crops has been contentious. This may be, in part, a consequence of poorly agreed upon study objectives. Whole-food animal-feeding studies have been postulated to be useful in detecting both expected and unexpected effects on the composition of genetically modified crops. While the justification of animal feeding studies to detect unexpected effects may be inadequately supported, there may be better justification to conduct such studies in specific cases to investigate the consequences of expected compositional effects including expression of transgenic proteins. Such studies may be justified when (1) safety cannot reasonably be predicted from other evidence, (2) reasonable hypothesis for adverse effects are postulated, (3) the compositional component in question cannot be isolated or enriched in an active form for inclusion in animal feeding studies, and (4) reasonable multiples of exposure can be accomplished relative to human diets. The study design for whole-food animal-feeding studies should be hypotheses-driven, and the types of data collected should be consistent with adverse effects that are known to occur from dietary components of biological origin.
利用全食物（饲料中所含谷物粉）动物喂养研究来支持转基因作物的安全性评估一直存在争议。这在一定程度上可能是由于研究目标未达成充分共识所致。全食物动物喂养研究被假定有助于检测转基因作物成分方面的预期和非预期效应。虽然通过动物喂养研究检测非预期效应的理由可能并不充分，但在特定情况下开展此类研究以调查预期成分效应（包括转基因蛋白表达）的后果可能有更好的理由。当出现以下情况时，此类研究可能是合理的：（1）无法根据其他证据合理预测安全性；（2）提出了关于不良影响的合理假设；（3）所涉成分无法以活性形式分离或富集以纳入动物喂养研究；（4）相对于人类饮食能够实现合理的暴露倍数。全食物动物喂养研究的设计应以假设为导向，收集的数据类型应与已知由生物来源的饮食成分产生的不良影响相一致。
