Citations 0
July 6, 2005

Methodological Standards in Human vs Animal Clinical Trials

Author Affiliations

Letters Section Editor: Robert M. Golub, MD, Senior Editor.

JAMA. 2005;294(1):40-41. doi:10.1001/jama.294.1.40-a

To the Editor: Dr Pluta and colleagues found a significant benefit from the use of sodium nitrite in preventing delayed vasospasm following subarachnoid hemorrhage in a study of 14 monkeys.1 We are concerned about these conclusions given certain aspects of the study design.

Animal models allow well-controlled evaluation of new therapies prior to definitive studies in humans. However, these models do not eliminate the potential biases that are well-described in human studies. Human trials use random treatment allocation to distribute the variability among patients evenly between treatment groups.2 While the biological variation among subjects may be less in animal models than in human populations, genetic equivalence cannot be assumed, and there may be significant differences between treatment groups.3 While this study is strengthened by blinded outcome assessment, the 2 treatment groups may still have differed with respect to important baseline and prognostic covariates. Furthermore, the treating investigators were not blinded to treatment status during the intervention, which may have resulted in differential management of the groups. Human clinical trials that do not use randomization and blinding have been shown to overestimate the magnitude of treatment effects.4 Indeed, a human study of a therapy that did not utilize these rigorous methods would likely be viewed skeptically by clinicians.

First Page Preview View Large
First page PDF preview
First page PDF preview