纽约（路透卫生）12月14日发表的一份报告显示，当一个临床试验研究一个药物的时候，结论常常完全不同于最初在动物身上实验的发现，以上结论是在系统回顾了英国医学杂志（BMJ）网络发表的两类这样的研究后作出的。

　　来自伦敦热带医学与卫生学校的Roberts博士和他的同事，发现6种干预在临床试验中与实验是不一致的（有益或有害），然后研究者们检索medline和其他的数据库调查动物实验中的干预是否与临床一致。这6种干预包括：糖皮质激素治疗脑外伤，抗纤药治疗出血，溶栓要治疗缺血性中风，替拉扎特治疗缺血性中风，出生前使用糖皮质激素预防新生儿呼吸窘迫综合征，膦酸盐治疗骨质疏松症。其中三项研究临床结果与动物实验不一致。糖皮质激素在临床试验中不能改善脑外伤，但是它在动物实验中有益。抗纤药在临床试验中减少出血，但在动物实验中是非结论性。值得指出的是，替拉扎特在动物实验中对于中风有益，而临床试验却证实有害。其他三个干预手段则临床及动物研究中或多或少的一致。溶栓治疗在临床及动物研究中都证实可以提高中风的结局，在动物实验中，磷酸盐可以增加骨骼矿物质的密度，出生前使用糖皮质激素可以改善呼吸窘迫综合征，但是仅仅在临床试验中发现生存方面获益。
   
　　在路透卫生的综述中，共同研究者Ian Roberts博士说，“不完善的方法、发表偏倚、简单的动物模型不能反映人类疾病”是实验与临床试验不一致的原因。他相信“更多的动物研究的系统回顾是需要的”以保证实验与临床试验相关。

英文原文：
　　　　　　　　Animal Studies and Clinical Trial Findings Often at Odds

　　NEW YORK (Reuters Health) Dec 14 - When a clinical trial investigates a particular drug, the conclusion is often completely different from the findings first generated in animal studies, according to a systematic review of data from both types of studies published online by the British Medical Journal today.

　　For example, in clinical trials, tirilazad therapy for ischemic stroke increases the risk of death and dependency, whereas in animal studies, it reduces infarct volume and seems to improve neurobehavioral scores.

　　In an interview with Reuters Health, study co-author Dr. Ian Roberts said "poor methodology, publication bias, and animal models that simply do not reflect human disease" are to blame for the discordant findings. He believes that "more systematic reviews of animal studies are needed" to ensure that the findings will be relevant to the design of clinical trials.

　　Dr. Roberts, from the London School of Hygiene and Tropical Medicine, and colleagues identified six interventions that have shown an unambiguous treatment effect (either beneficial or harmful) in clinical trials. The investigators then conducted a search of Medline and other databases for studies investigating these interventions in animal models.

　　The six interventions studied included: corticosteroids for traumatic head injury, antifibrinolytics for hemorrhage, thrombolysis for ischemic stroke, tirilazad for ischemic stroke, antenatal corticosteroids to prevent neonatal respiratory disease syndrome, and bisphosphonates to treat osteoporosis.

　　For three of the interventions, the results of the clinical trials did not match those of the animal studies.

　　Corticosteroids did not improve head injury in clinical trials, but in animal studies, they had a beneficial effect. Antifibrinolytics reduced bleeding in clinical trials, but animal study findings were inconclusive. And, as noted, tirilazad for stroke proved harmful in clinical trials, but beneficial in animal studies.

　　For the other three interventions, the findings from both types of studies were more or less in agreement:

　　Thrombolysis for ischemic stroke improved outcomes in both clinical trials and animal studies. Bisphosphonates consistently increased bone mineral density. Antenatal steroids improved respiratory distress in both types of studies, but a survival benefit was only seen in clinical trials.

　　To improve the quality of animal studies, Dr. Roberts noted that there is a group called SABRE (www.sabre.org.uk), which is campaigning for more systematic reviews of animal experiments. "The goal is to make animal experimentation a little bit more serious as a scientific area. If the research is not valid, the results are not precise, and strong publication bias still remains, then animal studies are never going to be that useful."

　　　　　　　　　　　　　　　　　　　　　　　　　　摘自《生物谷》