Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Translation of Research Evidence From Animals to Humans

Translation of Research Evidence From Animals to Humans To the Editor: Most medical therapies in use today were initially developed and tested in animals,1 yet animal experiments often fail to replicate when tested in rigorous human trials.2,3 We conducted a systematic review to determine how often highly cited animal studies translate into successful human research. Methods Methods The 7 leading scientific journals by citation impact factor (Journal Citation Reports, Thomson Scientific, Philadelphia, Pa, 2004) that regularly publish original animal studies were searched: Science, Nature, Cell, Nature Medicine, Nature Genetics, Nature Immunology, and Nature Biotechnology. Articles with more than 500 citations were retrieved under the assumption that such prominent findings would more likely be tested in subsequent human trials.4 A total of 2000 articles published between 1980 and 2000 were screened, reflecting advances in molecular biology and recombinant genetics. Articles were included if they investigated a preventive or therapeutic intervention in an in vivo animal model. When there were multiple animal studies of the same intervention, the most cited study was retained. Power calculations (α = 0.05, β = 0.05) estimated that 49 articles were needed to exclude a translation rate below 5%. Methods For each included study, a literature search identified human studies that translated the animal evidence. Successful translation was defined as replication in a randomized trial yielding results that were statistically positive according to primary outcome. Interventions and diseases analogous to those studied in the animal study were allowed. Methods MEDLINE, EMBASE, the Cochrane Central Register of Controlled Trials, the Cochrane Database of Systematic Reviews, the National Institutes of Health Clinical Trials Database, BIOSIS Previews, and the International Pharmaceutical Abstracts Database were searched from their inception through May 2006. Bibliographies of topic-specific review articles were manually searched for additional studies and experts were contacted if the search was negative. Methods The quality of the studies was assessed based on adapted standards for the conduct of animal research (Figure 1).5 Good quality was defined as a global methodology score of 50% or higher. Multivariable logistic regression was used to assess predictors of translation. The Pearson correlation test was used to determine if methodological quality of animal studies improved over time. Significance level was set at 2-sided P<.05. Analyses were conducted using SAS version 9.0 (SAS Institute Inc, Cary, NC). Figure 1. Methodological Quality of Animal Trials (n=76) View LargeDownload Results Results Seventy-six animal studies fulfilling inclusion criteria were identified (Figure 2; details of studies available in online eTable). No animal study was negative. The median citation count was 889 (range, 639-2233). The median publication year was 1992, yielding a median of 14 years for potential translation. Of the animal studies, 37 (49%) were rated as having good methodological quality. Most studies included dose-response gradients, clinically relevant outcomes, and long-term end points (Figure 1). Few studies included random allocation of animals, adjustment for multiple hypothesis testing, or blinded assessment of outcomes. Methodological quality did not improve during the study interval (r = −0.08, P = .47). Figure 2. Search Flow and Article Retrieval View LargeDownload Results eTable. Details of Animal Studies View LargeDownload Results Of the animal studies, 28 (37%; 95% confidence interval [CI], 26%-48%) were replicated in human randomized trials, 14 (18%) were contradicted by randomized trials, and 34 (45%) remain untested. Median time to replication was 7 years (range, 1-15 years). Global methodology score did not predict translation in unadjusted analyses (odds ratio [OR], 1.28 per 10% higher score; 95% CI, 0.97-1.69) or in analyses adjusted for citation rate and length of time available for human replication (OR, 1.27; 95% CI, 0.96-1.69). Animal studies incorporating dose-response gradients were more likely to translate to humans (OR, 3.3; 95% CI, 1.1-10.1). Other quality criteria, type of therapy, type of disease, species, journal, citation rate, length of follow-up, and year of publication did not predict subsequent translation. Eight replicated interventions were subsequently approved for use in patients. Comment Comment Only about a third of highly cited animal research translated at the level of human randomized trials. This rate of translation is lower than the recently estimated 44% replication rate for highly cited human studies.4 Limitations of this review include a focus on highly cited animal studies published in leading journals, which by their positive and highly visible nature may have been more likely to translate than less frequently cited research. In addition, this study had limited power to discern individual predictors of translation. Comment Nevertheless, we believe these findings have important implications. First, patients and physicians should remain cautious about extrapolating the findings of prominent animal research to the care of human disease. Second, major opportunities for improving study design and methodological quality are available for preclinical research. Finally, poor replication of even high-quality animal studies should be expected by those who conduct clinical research. Back to top Article Information Author Contributions: Dr Hackam had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Hackam, Redelmeier. Acquisition of data: Hackam. Analysis and interpretation of data: Hackam, Redelmeier. Drafting of the manuscript: Hackam, Redelmeier. Critical revision of the manuscript for important intellectual content: Hackam, Redelmeier. Statistical analysis: Hackam, Redelmeier. Obtained funding: Redelmeier. Administrative, technical, or material support: Redelmeier. Study supervision: Redelmeier. Literature retrieval: Hackam. Financial Disclosures: None reported. Funding/Support: Dr Hackam was supported by a Canadian Institutes of Health Research Fellowship Award, the Chisholm Memorial Fellowship, and the Clinician-Scientist Training Program of the University of Toronto. Dr Redelmeier was supported by the Canada Research Chair in Medical Decision Sciences, the Error Management Unit of Sunnybrook Health Sciences Centre, the National Institutes of Health Resuscitation Outcomes Consortium, and the Canadian Institutes of Health Research. Role of the Sponsors: The funding sources had no role in the design and conduct of the study; in the collection, management, analysis, and interpretation of the data; or in the preparation, review, or approval of the manuscript. Additional Information: The eTable is available. Acknowledgment: We are grateful to Mark Crowther, MD, MSc, Gideon Koren, MD, Philippe Poussier, MD, Joel Ray, MD, MSc, William Sibbald, MD, MBA, and Matthew Stanbrook, MD, PhD, for comments on a previous draft of the manuscript. No individual received compensation for their assistance. References 1. Guidance for Industry: Nonclinical Safety Evaluation of Drugs or Biologic Combinations. Rockville, Md: US Dept of Health and Human Services, Food and Drug Administration, Center for Drug Evaluation and Research; 2006 2. Kaste M. Use of animal models has not contributed to development of acute stroke therapies: pro. Stroke. 2005;36:2323-232416141431Google ScholarCrossref 3. Pound P, Ebrahim S, Sandercock P, Bracken MB, Roberts I. Where is the evidence that animal research benefits humans? BMJ. 2004;328:514-51714988196Google ScholarCrossref 4. Ioannidis JPA. Contradicted and initially stronger effects in highly cited clinical research. JAMA. 2005;294:218-22816014596Google ScholarCrossref 5. Stroke Therapy Academic Industry Roundtable. Recommendations for standards regarding preclinical neuroprotective and restorative drug development. Stroke. 1999;30:2752-275810583007Google ScholarCrossref http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png JAMA American Medical Association

Translation of Research Evidence From Animals to Humans

Hackam, Daniel G.; Redelmeier, Donald A.
JAMA , Volume 296 (14) – Oct 11, 2006
Download PDF
4 pages

Loading next page...
 
/lp/american-medical-association/translation-of-research-evidence-from-animals-to-humans-vCP0ECowra

References (5)

Publisher
American Medical Association
Copyright
Copyright © 2006 American Medical Association. All Rights Reserved.
ISSN
0098-7484
eISSN
1538-3598
DOI
10.1001/jama.296.14.1731
Publisher site
See Article on Publisher Site

Abstract

To the Editor: Most medical therapies in use today were initially developed and tested in animals,1 yet animal experiments often fail to replicate when tested in rigorous human trials.2,3 We conducted a systematic review to determine how often highly cited animal studies translate into successful human research. Methods Methods The 7 leading scientific journals by citation impact factor (Journal Citation Reports, Thomson Scientific, Philadelphia, Pa, 2004) that regularly publish original animal studies were searched: Science, Nature, Cell, Nature Medicine, Nature Genetics, Nature Immunology, and Nature Biotechnology. Articles with more than 500 citations were retrieved under the assumption that such prominent findings would more likely be tested in subsequent human trials.4 A total of 2000 articles published between 1980 and 2000 were screened, reflecting advances in molecular biology and recombinant genetics. Articles were included if they investigated a preventive or therapeutic intervention in an in vivo animal model. When there were multiple animal studies of the same intervention, the most cited study was retained. Power calculations (α = 0.05, β = 0.05) estimated that 49 articles were needed to exclude a translation rate below 5%. Methods For each included study, a literature search identified human studies that translated the animal evidence. Successful translation was defined as replication in a randomized trial yielding results that were statistically positive according to primary outcome. Interventions and diseases analogous to those studied in the animal study were allowed. Methods MEDLINE, EMBASE, the Cochrane Central Register of Controlled Trials, the Cochrane Database of Systematic Reviews, the National Institutes of Health Clinical Trials Database, BIOSIS Previews, and the International Pharmaceutical Abstracts Database were searched from their inception through May 2006. Bibliographies of topic-specific review articles were manually searched for additional studies and experts were contacted if the search was negative. Methods The quality of the studies was assessed based on adapted standards for the conduct of animal research (Figure 1).5 Good quality was defined as a global methodology score of 50% or higher. Multivariable logistic regression was used to assess predictors of translation. The Pearson correlation test was used to determine if methodological quality of animal studies improved over time. Significance level was set at 2-sided P<.05. Analyses were conducted using SAS version 9.0 (SAS Institute Inc, Cary, NC). Figure 1. Methodological Quality of Animal Trials (n=76) View LargeDownload Results Results Seventy-six animal studies fulfilling inclusion criteria were identified (Figure 2; details of studies available in online eTable). No animal study was negative. The median citation count was 889 (range, 639-2233). The median publication year was 1992, yielding a median of 14 years for potential translation. Of the animal studies, 37 (49%) were rated as having good methodological quality. Most studies included dose-response gradients, clinically relevant outcomes, and long-term end points (Figure 1). Few studies included random allocation of animals, adjustment for multiple hypothesis testing, or blinded assessment of outcomes. Methodological quality did not improve during the study interval (r = −0.08, P = .47). Figure 2. Search Flow and Article Retrieval View LargeDownload Results eTable. Details of Animal Studies View LargeDownload Results Of the animal studies, 28 (37%; 95% confidence interval [CI], 26%-48%) were replicated in human randomized trials, 14 (18%) were contradicted by randomized trials, and 34 (45%) remain untested. Median time to replication was 7 years (range, 1-15 years). Global methodology score did not predict translation in unadjusted analyses (odds ratio [OR], 1.28 per 10% higher score; 95% CI, 0.97-1.69) or in analyses adjusted for citation rate and length of time available for human replication (OR, 1.27; 95% CI, 0.96-1.69). Animal studies incorporating dose-response gradients were more likely to translate to humans (OR, 3.3; 95% CI, 1.1-10.1). Other quality criteria, type of therapy, type of disease, species, journal, citation rate, length of follow-up, and year of publication did not predict subsequent translation. Eight replicated interventions were subsequently approved for use in patients. Comment Comment Only about a third of highly cited animal research translated at the level of human randomized trials. This rate of translation is lower than the recently estimated 44% replication rate for highly cited human studies.4 Limitations of this review include a focus on highly cited animal studies published in leading journals, which by their positive and highly visible nature may have been more likely to translate than less frequently cited research. In addition, this study had limited power to discern individual predictors of translation. Comment Nevertheless, we believe these findings have important implications. First, patients and physicians should remain cautious about extrapolating the findings of prominent animal research to the care of human disease. Second, major opportunities for improving study design and methodological quality are available for preclinical research. Finally, poor replication of even high-quality animal studies should be expected by those who conduct clinical research. Back to top Article Information Author Contributions: Dr Hackam had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Hackam, Redelmeier. Acquisition of data: Hackam. Analysis and interpretation of data: Hackam, Redelmeier. Drafting of the manuscript: Hackam, Redelmeier. Critical revision of the manuscript for important intellectual content: Hackam, Redelmeier. Statistical analysis: Hackam, Redelmeier. Obtained funding: Redelmeier. Administrative, technical, or material support: Redelmeier. Study supervision: Redelmeier. Literature retrieval: Hackam. Financial Disclosures: None reported. Funding/Support: Dr Hackam was supported by a Canadian Institutes of Health Research Fellowship Award, the Chisholm Memorial Fellowship, and the Clinician-Scientist Training Program of the University of Toronto. Dr Redelmeier was supported by the Canada Research Chair in Medical Decision Sciences, the Error Management Unit of Sunnybrook Health Sciences Centre, the National Institutes of Health Resuscitation Outcomes Consortium, and the Canadian Institutes of Health Research. Role of the Sponsors: The funding sources had no role in the design and conduct of the study; in the collection, management, analysis, and interpretation of the data; or in the preparation, review, or approval of the manuscript. Additional Information: The eTable is available. Acknowledgment: We are grateful to Mark Crowther, MD, MSc, Gideon Koren, MD, Philippe Poussier, MD, Joel Ray, MD, MSc, William Sibbald, MD, MBA, and Matthew Stanbrook, MD, PhD, for comments on a previous draft of the manuscript. No individual received compensation for their assistance. References 1. Guidance for Industry: Nonclinical Safety Evaluation of Drugs or Biologic Combinations. Rockville, Md: US Dept of Health and Human Services, Food and Drug Administration, Center for Drug Evaluation and Research; 2006 2. Kaste M. Use of animal models has not contributed to development of acute stroke therapies: pro. Stroke. 2005;36:2323-232416141431Google ScholarCrossref 3. Pound P, Ebrahim S, Sandercock P, Bracken MB, Roberts I. Where is the evidence that animal research benefits humans? BMJ. 2004;328:514-51714988196Google ScholarCrossref 4. Ioannidis JPA. Contradicted and initially stronger effects in highly cited clinical research. JAMA. 2005;294:218-22816014596Google ScholarCrossref 5. Stroke Therapy Academic Industry Roundtable. Recommendations for standards regarding preclinical neuroprotective and restorative drug development. Stroke. 1999;30:2752-275810583007Google ScholarCrossref

Journal

JAMA – American Medical Association

Published: Oct 11, 2006

There are no references for this article.