Is Efficacy In Animal Models Enough To Show Possibility Of Direct Benefit To Children

• Loading metrics

Open Admission

Research in Translation

Research in Translation Inquiry in Translation manufactures discuss a item drug, handling, or public wellness intervention in the context of translation from early enquiry to clinical research, or clinical prove to do.

Run into all commodity types »

Tin can Beast Models of Disease Reliably Inform Human Studies?

• H. Bart van der Worp,
• David W. Howells,
• Emily S. Sena,
• Michelle J. Porritt,
• Sarah Rewell,
• Victoria O'Collins,
• Malcolm R. Macleod

x

• Published: March 30, 2010
• https://doi.org/10.1371/periodical.pmed.1000245

Figures

Commendation: van der Worp HB, Howells DW, Sena ES, Porritt MJ, Rewell Due south, O'Collins V, et al. (2010) Tin Beast Models of Disease Reliably Inform Human Studies? PLoS Med vii(3): e1000245. https://doi.org/10.1371/journal.pmed.1000245

Published: March 30, 2010

Copyright: © 2010 van der Worp et al. This is an open-access article distributed nether the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in whatever medium, provided the original author and source are credited.

Funding: This work was supported in function by the MRC Trials Methodology Hub and the National Health and Medical Research Council. The funders played no office in the determination to submit the article nor in its preparation.

Competing interests: Malcolm R. MacLeod is on the Editorial Lath of PLoS Medicine.

Abbreviations: ALS, amyotrophic lateral sclerosis; CAMARADES, Collaborative Approach to Meta-Analysis And Review of Fauna Data from Experimental Stroke; CONSORT, CONsolidated Standards Of Reporting Trials

Provenance: Commissioned; externally peer reviewed.

Linked Research Article

This Research in Translation discusses the following new study published in PLoS Biology:

Sena ES, van der Worp HB, Bathroom PMW, Howells DW, Macleod MR (2010) Publication bias in reports of animal stroke studies leads to major overstatement of efficacy. PLoS Biol 8(iii): e1000344. doi:x.1371/periodical. pbio.1000344

Publication bias confounds attempts to apply systematic reviews to appraise the efficacy of diverse interventions tested in experiments modeling acute ischemic stroke, leading to a 30% overstatement of efficacy of interventions tested in animals.

Summary Points

• The value of brute experiments for predicting the effectiveness of treatment strategies in clinical trials has remained controversial, mainly because of a recurrent failure of interventions apparently promising in creature models to interpret to the dispensary.
• Translational failure may be explained in part past methodological flaws in fauna studies, leading to systematic bias and thereby to inadequate information and wrong conclusions about efficacy.
• Failures besides result because of critical disparities, commonly disease specific, between the animal models and the clinical trials testing the handling strategy.
• Systematic review and meta-analysis of animal studies may aid in the selection of the most promising handling strategies for clinical trials.
• Publication bias may account for 1-third or more of the efficacy reported in systematic reviews of animal stroke studies, and probably also plays a substantial office in the experimental literature for other diseases.
• We provide recommendations for the reporting of aspects of study quality in publications of comparisons of handling strategies in animal models of disease.

Brute experiments have contributed much to our understanding of mechanisms of disease, just their value in predicting the effectiveness of handling strategies in clinical trials has remained controversial [1]–[3]. In fact, clinical trials are essential because animal studies practice not predict with sufficient certainty what volition happen in humans. In a review of animal studies published in seven leading scientific journals of high touch, almost one-third of the studies translated at the level of human randomised trials, and one-tenth of the interventions, were subsequently approved for employ in patients [1]. However, these were studies of high impact (median commendation count, 889), and less frequently cited brute research probably has a lower likelihood of translation to the dispensary. Depending on ane's perspective, this attrition rate of 90% may be viewed as either a failure or as a success, merely information technology serves to illustrate the magnitude of the difficulties in translation that beset even findings of high impact.

Recent examples of therapies that failed in big randomised clinical trials despite substantial reported benefit in a range of creature studies include enteral probiotics for the prevention of infectious complications of acute pancreatitis, NXY-059 for astute ischemic stroke, and a range of strategies to reduce lethal reperfusion injury in patients with astute myocardial infarction [4]–[vii]. In animal models of acute ischemic stroke, virtually 500 "neuroprotective" treatment strategies have been reported to improve result, only only aspirin and very early intravenous thrombolysis with alteplase (recombinant tissue-plasminogen activator) take proved effective in patients, despite numerous clinical trials of other treatment strategies [8],[9].

Causes of Failed Translation

The disparity betwixt the results of animal models and clinical trials may in part be explained by shortcomings of the clinical trials. For instance, these may have had insufficient statistical power to notice a truthful benefit of the treatment under written report. For practical or commercial purposes, the designs of some clinical trials have too failed to acknowledge the limitations of efficacy observed in animal studies, for example by allowing therapy at later fourth dimension points when the window of opportunity has passed [x],[eleven]. Secondly, the failure of apparently promising interventions to translate to the dispensary may as well be caused by inadequate animal data and overoptimistic conclusions about efficacy drawn from methodologically flawed brute studies. A tertiary possible explanation is the lack of external validity, or generalisability, of some animal models; in other words, that these do not sufficiently reverberate disease in humans. Finally, neutral or negative beast studies may be more likely to remain unpublished than neutral clinical trials, giving the impression that the first are more often positive than the 2d. This commodity aims to address the possible sources of bias that threaten the internal and external validity of animal studies, to provide solutions to amend the reliability of such studies, and thereby to improve their translation to the clinic.

Internal Validity

Acceptable internal validity of an animal experiment implies that the differences observed between groups of animals allocated to different interventions may, apart from random error, be attributed to the treatment nether investigation [12]. The internal validity may be reduced by four types of bias through which systematic differences betwixt treatment groups are introduced (Table i). Just like whatsoever clinical trial, each formal animate being study testing the effectiveness of an intervention should be based on a well-designed study protocol addressing the design and conduct of the study, besides as the analysis and reporting of its results. Aspects of the blueprint, conduct, and analysis of an creature experiment that help to reduce bias and to improve the reliability and reproducibility of the results are discussed below. As the impact of study quality has been studied much more than extensively in clinical trials than in animate being studies, the backgrounds and recommendations regarding these issues are largely based on the clinical CONsolidated Standards of Reporting Trials (Espoused) statement, and to a smaller extent on published recommendations and guidelines for the deport and reporting of brute studies of acute ischemic stroke [13]–[17].

Randomisation

To prevent pick bias, treatment allocation should be based on randomisation (Box ane), a method that is almost ubiquitous in clinical treatment trials. In function, this prevents the investigator from having to choose which treatment a particular fauna will receive, a procedure which might result (consciously or subconsciously) in animals which are thought to exercise particularly well or particularly desperately being overrepresented in a particular treatment group. Foreknowledge of treatment group consignment may likewise lead to selective exclusion of animals based on prognostic factors [13]. These problems can arise with any method in which group allocation is known in accelerate or can be predicted. Such methods include both the use of predetermined rules (eastward.g., assignment in alternation or on the basis of the days of the week) or of open randomisation schedules. Picking animals "at random" from their cages also has the risk of conscious or hidden manipulation, and does not correspond true randomisation.

Box 1. Glossary

• Allotment concealment: Concealing the allocation sequence from those assigning animals to intervention groups, until the moment of consignment.
• Bias: Systematic distortion of the estimated intervention upshot abroad from the "truth," caused by inadequacies in the design, comport, or analysis of an experiment.
• Blinding (masking): Keeping the persons who perform the experiment, collect data, and assess outcome unaware of the treatment allocation.
• Eligibility criteria: Inclusion and exclusion criteria: the characteristics that define which animals are eligible to be enrolled in a study.
• External validity: The extent to which the results of an brute experiment provide a correct basis for generalisations to the homo status.
• Intention-to-treat analysis: Analysis of information of all animals included in the group to which they were assigned, regardless of whether they completed the intervention.
• Internal validity: The extent to which the design and conduct of the trial eliminate the possibility of bias.
• Power: The probability that a study volition notice a statistically pregnant effect of a specified size.
• Randomisation: Randomly allocating the intervention under study across the comparison groups, to ensure that group assignment cannot be predicted.
• Sample size: The number of animals in the written report

Definitions adapted from [13] and from Wikipedia (http://world wide web.wikipedia.org, accessed on nine November 2009).

Randomisation may appear redundant if the animals form a homogeneous grouping from a genetic and ecology perspective, as often is the case with rats and other rodents. Nonetheless, it is not just the creature itself but mainly the induction of the disease that may requite rise to variation. For instance, in that location is a large variation in infarct size in most rat models of ischaemic stroke non only because of interindividual differences in collateral circulation—fifty-fifty in inbred strains—only also considering in some animals the artery is occluded ameliorate than in others and because the models are inherently vulnerable to complications that may affect outcome, such as periprocedural hypotension or hypoxemia. It is because of this variation that randomisation, ideally occurring later on the injury or affliction has been induced, is essential.

In clinical trials, automated randomisation techniques such equally random number generation are most commonly used, just manual methods (such every bit tossing a coin or throwing die) are also acceptable as long equally these cannot be manipulated. By preference, such transmission techniques should be performed past an independent person.

Blinding

In studies that are blinded throughout their course, the investigators and other persons involved volition not be influenced past knowledge of the treatment assignment, thereby preventing functioning, detection, and attrition bias. Noesis of handling assignment may subconsciously or otherwise affect the supply of boosted intendance, issue assessment, and decisions to withdraw animals from the experiment.

In contrast to allocation darkening (Box i), blinding may not ever be possible in all stages of an experiment, for instance when the treatment under investigation concerns a surgical procedure. All the same, blinding of outcome assessment is almost always possible.

In clinical trials, the virtually common course of blinding is double blinding, in which the patients, the investigators, and the caregivers are unaware of the intervention assignment. Considering the patient does not know which handling is beingness administered, the placebo upshot will be similar beyond the comparison groups. As animals are not susceptible to the placebo effect, double blinding is not an issue in animal studies. Even so the influence that unblinded animal handling can have on performance in neurobehavioural tasks [18], the fact that in some articles of animal studies "double blinding" is reported raises questions about the authors' knowledge of blinding likewise as near the review and editorial processes of the journals in which the studies were published [xix],[20].

Sample Size Adding

Selection of target sample size is a disquisitional factor in the design of any comparing study. The study should be big plenty to accept a high probability of detecting a treatment upshot of a given size if such an outcome truly exists, but likewise pay attending to legal requirements and ethical and practical considerations to keep the number of animals every bit pocket-size as possible. The required sample size should be determined before the start of the report with a formal sample size calculation, of which the primal elements of statistical significance (α), effect size (δ), power (ane–β), and standard deviation of the measurements have been explained in numerous articles [thirteen],[21]. Unfortunately, the assumptions on variation of the measurements are often based on incomplete data, and small-scale errors tin lead to a study that is either under- or overpowered. From an ethical indicate of view, underpowered studies are undesirable, every bit they might lead to the false determination that the intervention is without efficacy, and all included animals will take been used to no benefit. Overpowered studies would also be unethical, but these are much less prevalent.

Monitoring of Physiological Parameters

Depending on the affliction under investigation, a range of physiological variables may affect outcome, and inadequate control of these factors may lead to erroneous conclusions. Whether or not physiological parameters should be assessed, and for how long, therefore depends on the model and on the tested condition.

Eligibility Criteria and Drop-Outs

Because of their complication, many creature models are inherently vulnerable to complications—such as inadvertent claret loss during surgery to induce cognitive or myocardial ischemia—that are not related to the handling nether report but that may have a large outcome on outcome. Given the explanatory character of preclinical studies, information technology is justifiable to exclude animals with such complications from the analyses of treatment effects, provided that the eligibility criteria are predefined and not determined on a post-hoc basis, and that the person responsible for the exclusion of animals is unaware of the treatment assignment.

In clinical trials, inclusion and exclusion criteria are normally practical before enrolment in the study, but for the reason above, in animal studies it is justifiable also to apply these criteria during the course of the study. Nonetheless, these should exist limited to complications that are demonstrably not related to the intervention nether study, every bit this may otherwise lead to attrition bias. For example, if a potential novel treatment for colorectal cancer increases instead of reduces neoplasm progression, thereby weakening the animals and increasing their susceptibility to infections, exclusion of animals dying prematurely because of respiratory tract infections may lead to selective exclusion of animals with the largest tumours and mask the detrimental result of the novel intervention.

Statistical Analysis

The statistical analysis of the results of fauna experiments has been given elaborate attending in review articles and books [22]. Nevertheless, even when data appear elementary and their assay straightforward, inadequate techniques are oftentimes used. Common examples include the employ of a t-test for nonparametric information, calculating means and standard deviations for ordinal data, and treating multiple observations from i animal as contained.

In clinical trials, an intention-to-treat analysis is mostly favoured considering it avoids bias associated with nonrandom loss of participants [13]. Every bit explained to a higher place, the explanatory character of most studies justifies the use of an analysis restricted to data from animals that have fulfilled all eligibility criteria, provided that all animals excluded from the analysis are deemed for and that those exclusions accept been fabricated without noesis of handling grouping allocation.

Control of Study Conduct

The careers of investigators at academic institutions and in industry depend in part on the number and impact of their publications, and these investigators may be all too aware of the fact that the prospect of their work being published increases when positive results are obtained. This underscores non only the importance of randomisation, allocation concealment, and blinding, but as well the need for acceptable monitoring and auditing of laboratory experiments past tertiary parties. Indeed, adopting a multicentre approach to animal studies has been proposed, as a way of securing transparent quality control [23].

Bias in Animal Studies

The presence of bias in creature studies has been tested near extensively in studies of acute ischemic stroke, probably because in this field the gap between the laboratory and the dispensary is both very large and well recognised [8]. In systematic reviews of different interventions tested in beast models of acute ischemic stroke, other emergencies, Parkinson's disease, multiple sclerosis, or amyotrophic lateral sclerosis, generally about a third or less of the studies reported random allocation to the treatment group, and even fewer studies reported darkening of treatment allocation or blinded event assessment [ii],[16],[xix],[24],[25]. Even when reported, the methods used for randomisation and blinding were rarely given. A priori sample size calculations were reported in 0%–three% of the studies (Table 2).

Complications of the disease and/or handling under report were reported in 19% of the studies of hypothermia for acute ischemic stroke. All but one of these complications concerned premature death, and about 90% of these animals were excluded from the analyses [20]. In another review of several treatment strategies for acute ischemic stroke, only one of 45 studies mentioned predefined inclusion and exclusion criteria, and in only 12 manufactures (27%) exclusion of animals from assay was mentioned and substantiated. It is difficult to believe that in every other study every single experiment went as smoothly as the investigators had planned [19].

2 factors limit the interpretation of the to a higher place-mentioned data. First, the assessment of possible confounders in systematic reviews was based on what was reported in the articles, and may have been incomplete because the authors considered these aspects of written report design not sufficiently relevant to be mentioned. In improver, definitions of randomisation, allocation concealment, and blinding might vary across studies, and, for example, randomly picking animals from their cages may take been chosen "randomisation." Indeed, a survey of a sample of authors of publications included in such reviews suggested that this was sometimes the case [26].

Quality Checklists

At least iv dissimilar simply largely overlapping study-quality checklists have been proposed for utilise in animate being studies of focal cerebral ischemia. These checklists have included items relating first to the range of circumstances under which efficacy has been shown and second to the characteristics that might human activity as a source of bias in private experiments [sixteen].

Cess of overall methodological quality of individual studies with these checklists is limited by controversy nearly the limerick of the checklists and, more than importantly, because the weight of each of the individual components has remained uncertain. For case, in the virtually frequently used CAMARADES checklist, "acceptable resource allotment concealment" may have a much larger affect on effect size than "compliance with regulatory requirements" [16].

Does Methodological Quality Matter?

Several systematic reviews and meta-analyses have provided empirical evidence that inadequate methodological approaches in controlled clinical trials are associated with bias. Clinical trials in which authors did not written report randomisation, fairly conceal treatment allocation, or use double blinding yielded larger estimates of treatment furnishings than trials in which these study quality issues were reported [12],[27]–[32].

The touch on of methodological quality on the result size in animal studies has been examined less extensively. In animal studies testing interventions in emergency medicine, the odds of a positive upshot were more than three times as large if the publication did not written report randomisation or blinding as compared with publications that did report these methods [33]. In systematic reviews of FK-506 or hypothermia for acute ischemic stroke, an inverse relation was institute between effect size and report quality, equally assessed by a ten-item report-quality checklist [xx],[34]. The aforementioned review on hypothermia found large overstatements of the reduction in infarct volume in animal stroke studies without randomisation or blinded outcome assessment when they were compared with randomised or blinded studies, but a meta-analysis of 13 meta-analyses in experimental stroke describing outcomes in a total of 15,635 animals found no statistically pregnant outcome of these quality items on effect size. In this meta-meta-assay, only resource allotment darkening was associated with a larger effect size [35].

A limitation of the meta-analyses assessing the effect of study quality aspects on issue size is the fact that no consideration has been given to possible interactions betwixt quality items, and that only univariate analyses were performed. However, private quality aspects that may affect the results of meta-analyses of animal studies are unlikely to operate independently. For example, nonrandomised studies may be more likely than randomised studies to condone other quality problems, such as allocation concealment or blinding, or to utilize shorter delays for the initiation of treatment, all of which may bear on study results. The relative importance of the various possible sources of bias is therefore not notwithstanding known and is the field of study of ongoing research.

External Validity

Even if the design and conduct of an animal study are sound and eliminate the possibility of bias, the translation of its results to the dispensary may fail considering of disparities between the model and the clinical trials testing the treatment strategy. Common causes of such reduced external validity are listed in Box ii and are not limited to differences between animals and humans in the pathophysiology of affliction, but besides include differences in comorbidities, the use of co-medication, timing of the administration and dosing of the study treatment, and the selection of effect measures. Whereas the issues for internal validity probably employ to the majority of animal models regardless of the disease under written report, the external validity of a model will largely exist determined past disease-specific factors.

Box 2. Common Causes of Reduced External Validity of Animal Studies

• The induction of the disease under study in animals that are immature and otherwise good for you, whereas in patients the illness mainly occurs in elderly people with co-morbidities.
• Cess of the upshot of a treatment in a homogeneous grouping of animals versus a heterogeneous group of patients.
• The use of either male or female animals only, whereas the disease occurs in male and female patients alike.
• The use of models for inducing a disease or injury with bereft similarity to the human condition.
• Delays to kickoff of treatment that are unrealistic in the clinic; the use of doses that are toxic or not tolerated by patients.
• Differences in event measures and the timing of effect assessment between animal studies and clinical trials.

Stroke Models

As mentioned above, the translation of efficacy from animal studies to human affliction has perhaps been least successful for neurological diseases in full general and for ischaemic stroke in particular. Equally at that place is also no other beast model of disease that has been more than rigorously subjected to systematic review and meta-assay, stroke serves equally a skillful example of where difficulties in translation might arise.

The incidence of stroke increases with historic period, and stroke patients commonly have other health issues that might increase their stroke risk, complicate their clinical course, and affect functional outcome. Of patients with acute stroke, upward to 75% and 68% take hypertension and hyperglycaemia, respectively [ix],[36]. While it is important to know whether candidate stroke drugs retain efficacy in the face of these comorbidities, only about ten% of focal ischaemia studies have used animals with hypertension, and fewer than 1% have used animals with induced diabetes. In add-on, animals used in stroke models were well-nigh invariably young, and female person animals were highly underrepresented. Over 95% of the studies were performed in rats and mice, and animals that are maybe biologically closer to humans are inappreciably ever used [xvi],[19]. Moreover, well-nigh brute studies have failed to acknowledge the inevitable delay betwixt the onset of symptoms and the possibility to kickoff treatment in patients. In a systematic review of animal studies of five dissimilar neuroprotective agents that had as well been tested in 21 clinical trials including a total of more than 12,000 patients with acute ischaemic stroke, the median time between the onset of ischaemia and start of treatment in the animal studies was simply 10 minutes, which is infeasible in the clinic [19]. In the large majority of clinical trials, functional result is the primary measure out of efficacy, whereas brute studies usually rely on infarct volume. Several studies have suggested that in patients the relation between infarct volume and functional issue is moderate at best [37],[38]. Finally, the usual time of outcome assessment of 1–three days in beast models contrasts sharply with that of 3 months in patients [xix]. For these reasons, it is not surprising that, except for thrombolysis, all treatment strategies proven effective in the laboratory take failed in the dispensary.

Other Acute Affliction Models

Differences betwixt animal models and clinical trials like to those mentioned above have been proposed every bit causes of the recurrent failure of a range of strategies to reduce lethal reperfusion injury in patients with acute myocardial infarction [vi],[7]. The failure to acknowledge the presence of often severe comorbidities in patients, and brusk and clinically unattainable onset-to-handling delays, accept besides limited the external validity of fauna models of traumatic brain injury [ii].

Chronic Affliction Models

The external validity of models of chronic and progressive diseases may also be challenged by other factors. For the handling of Parkinson's disease, researchers take mainly relied on injury-induced models that mimic nigrostriatal dopamine deficiency merely do non recapitulate the dull, progressive, and degenerative nature of the affliction in humans. Whereas in clinical trials interventions were administered over a prolonged period of time in the context of this slowly progressive disease, putative neuroprotective agents were administered earlier or at the same fourth dimension as an acute Parkinson'south disease-like lesion was induced in the typical underlying animal studies [39].

Based on the identification of unmarried point-mutations in the factor encoding superoxide dismutase ane (SOD1) in near 3% of the patients with amyotrophic lateral sclerosis (ALS), mice carrying 23 copies of the human SOD1G93A transgene are considered the standard model for therapeutic studies of ALS. Apart from the fact that this model may exist valid only for patients with SOD1 mutations, the mice may endure from a phenotype that is and then aggressive and so overdriven by its 23 copies of the transgene that no pharmacological intervention outside of the directly inhibition of SOD1 will e'er impact ALS-related survival. In improver, it has been suggested that these mice may be more susceptible to infections and other non-ALS related illnesses and that it is this illness rather than the ALS that is alleviated by the experimental treatment. Consistent with this hypothesis, several of the compounds reported equally efficacious in SOD1G93A mice are wide-spectrum antibiotics and full general anti-inflammatory agents [xl].

Publication Bias

Decisions to assess the consequence of novel handling strategies in clinical trials are, ideally, based on an understanding of all publicly reported information from preclinical studies. Systematic review and meta-analysis are techniques adult for the analysis of data from clinical trials and may be helpful in the selection of the most promising strategies [xvi]. However, if studies are published selectively on the basis of their results, even a meta-analysis based on a rigorous systematic review volition exist misleading.

The presence of bias in the reporting of clinical trials has been studied extensively. There is potent empirical testify that clinical studies reporting positive or significant results are more likely to be published, and that outcomes that are statistically significant have higher odds of being reported in total rather than as an abstract. Such publication bias will lead to overestimation of treatment effects and can make the readily bachelor show unreliable for decision making [41].

Unfortunately, the presence of publication bias in animal studies has received much less attention. In a recent systematic review of studies testing the efficacy of interventions in animal models of man disease, only six reported testing for the presence of publication bias, and such bias was constitute in 4 [34],[42]–[46]. No study gave quantitative estimates of the bear on on upshot size of publication bias [47].

In a subsequent meta-analysis of 525 publications [47] included in systematic reviews of 16 interventions tested in fauna studies of acute ischaemic stroke, Egger regression and Trim and Fill assay suggested that publication bias was widely prevalent. The analyses suggested that publication bias might account for around one-third of the efficacy reported in systematic reviews of beast stroke studies. Considering this meta-analysis included all reported experiments testing an effect of an intervention on infarct size, and non simply the experiment with the largest effect size from each publication, at least some experiments testing ineffective doses (due east.g., at the lower finish of a dose-response bend) were included. For this reason, this meta-assay is more probable to underestimate than to overestimate the outcome of publication bias. It is therefore probably more revealing that of the 525 publications, only x (ii%) did not report at least i meaning effect on either infarct book or neurobehavioural score [47]. Although unproven, it appears unlikely that the animal stroke literature is uniquely susceptible to publication bias.

Nonpublication of the results of brute studies is unethical not only because it deprives researchers of the authentic data they demand to gauge the potential of novel therapies in clinical trials, but also because the included animals are wasted because they practice non contribute to accumulating noesis. In add-on, inquiry syntheses that overstate biological furnishings may lead to further unnecessary animal experiments testing poorly founded hypotheses.

Applied Improvement Strategies

Although there is no direct evidence of a causal relationship, information technology is likely that the recurrent failure of apparently promising interventions to improve outcome in clinical trials has in function been caused past inadequate internal and external validity of preclinical studies and publication bias favouring positive studies. On the basis of aplenty empirical evidence from clinical trials and some evidence from preclinical studies, we suggest that the testing of treatment strategies in fauna models of illness and its reporting should adopt standards similar to those in the clinic to ensure that conclusion making is based on high-quality and unbiased data. Aspects of study quality that should be reported in any manuscript are listed in Box 3.

Box 3. Aspects of Study Quality to Exist Reported in the Manuscript

• Sample size adding: How the sample size was determined, and which assumptions were made.
• Eligibility criteria: Inclusion and exclusion criteria for enrolment.
• Treatment allocation: The method by which animals were allocated to experimental groups. If this allotment was past randomisation, the method of randomisation.
• Allocation concealment: The method to implement the allocation sequence, and if this sequence was concealed until consignment.
• Blinding: Whether the investigators and other persons involved were blinded to the treatment allocation, and at which points in time during the study.
• Flow of animals: Flow of animals through each stage of the study, with a specific attention to animals excluded from the analyses. Reasons for exclusion from the analyses.
• Control of physiological variables: Whether and which physiological parameters were monitored and controlled.
• Control of study deport: Whether a 3rd party controlled which parts of the conduct of the study.
• Statistical methods: Which statistical methods were used for which analysis.

Recommendations based on [13],[17].

Non only should the disease or injury itself reflect the condition in humans equally much as possible, merely age, sex, and comorbidities should also exist modelled where possible. The investigators should justify their selection of the model and outcome measures. In turn, human clinical trials should be designed to replicate, as far equally is possible, the circumstances under which efficacy has been observed in animals. For an adequate interpretation of the potential and limitations of a novel treatment strategy, a systematic review and meta-analysis of all available evidence from preclinical studies should be performed before clinical trials are started. Evidence of benefit from a single laboratory or obtained in a single model or species is probably not sufficient.

Finally, the recognition of substantial publication bias in the clinical literature has led to the introduction of clinical trial registration systems to ensure that those summarising enquiry findings are at least aware of all relevant clinical trials that have been performed [48]. Given that a framework regulating animal experimentation already exists in many countries, nosotros suggest that this might be exploited to allow the maintenance of a central annals of experiments performed, and registration referenced in publications.

Five Key Papers in the Field

Hackam 2006 [1]: Shows that about a third of highly cited animal enquiry translates at the level of homo randomised trials.

Sena 2007 [16]: Proposes minimum standards for the range and quality of pre-clinical creature information earlier these are taken to clinical trials.

Dirksen 2007 [half-dozen]: Provides an overview of the various strategies that inhibit reperfusion injury after myocardial infarction and discusses potential mechanisms that may have contributed to the discrepancy between promising pre-clinical information and the disappointing results in randomised clinical trials.

Scott 2008 [40]: Elaborate report suggesting that the majority of published effects of treatments for amyotrophic lateral sclerosis are almost likely measurements of noise in the distribution of survival means as opposed to bodily drug result.

Sena 2010 [47]: The commencement study to estimate the affect of publication bias on the efficacy reported in systematic reviews of animal studies.

Author Contributions

ICMJE criteria for authorship read and met: HBvdW DWH ESS MJP SR VO MRM. Wrote the showtime draft of the paper: HBvdW. Contributed to the writing of the newspaper: DWH ESS SR VO MRM.

References

1. 1. Hackam DG, Redelmeier DA (2006) Translation of inquiry evidence from animals to humans. JAMA 296: 1731–1732.
   • View Article
   • Google Scholar
2. 2. Perel P, Roberts I, Sena E, Wheble P, Briscoe C, et al. (2007) Comparison of treatment effects between brute experiments and clinical trials: systematic review. B M J 334: 197.
   • View Article
   • Google Scholar
3. three. Hackam DG (2007) Translating animal research into clinical do good. B M J 334: 163–164.
   • View Article
   • Google Scholar
4. 4. Besselink MG, van Santvoort HC, Buskens E, Boermeester MA, van Goor H, et al. (2008) Probiotic prophylaxis in predicted severe acute pancreatitis: a randomised, double-bullheaded, placebo-controlled trial. Lancet 371: 651–659.
   • View Article
   • Google Scholar
5. 5. Shuaib A, Lees KR, Lyden P, Grotta J, Davalos A, et al. (2007) NXY-059 for the treatment of astute ischemic stroke. N Engl J Med 357: 562–571.
   • View Commodity
   • Google Scholar
6. 6. Dirksen MT, Laarman GJ, Simoons ML, Duncker DJ (2007) Reperfusion injury in humans: a review of clinical trials on reperfusion injury inhibitory strategies. Cardiovasc Res 74: 343–355.
   • View Commodity
   • Google Scholar
7. vii. Yellon DM, Hausenloy DJ (2007) Myocardial reperfusion injury. N Engl J Med 357: 1121–1135.
   • View Article
   • Google Scholar
8. eight. O'Collins VE, Macleod MR, Donnan GA, Horky LL, van der Worp BH, et al. (2006) one,026 experimental treatments in acute stroke. Ann Neurol 59: 467–477.
   • View Commodity
   • Google Scholar
9. 9. Van der Worp HB, Van Gijn J (2007) Clinical practice. Acute ischemic stroke. Northward Engl J Med 357: 572–579.
   • View Article
   • Google Scholar
10. 10. Grotta J (2001) Neuroprotection is unlikely to be effective in humans using current trial designs. Stroke 33: 306–307.
    • View Article
    • Google Scholar
11. 11. Gladstone DJ, Black SE, Hakim AM, Heart and Stroke Foundation of Ontario Middle of Excellence in Stroke Recovery (2002) Toward wisdom from failure. Lessons from neuroprotective stroke trials and new therapeutic directions. Stroke 33: 2123–2136.
    • View Article
    • Google Scholar
12. 12. Juni P, Altman DG, Egger M (2001) Systematic reviews in health care: Assessing the quality of controlled clinical trials. B Chiliad J 323: 42–46.
    • View Article
    • Google Scholar
13. thirteen. Altman DG, Schulz KF, Moher D, Egger M, Davidoff F, et al. (2001) The revised Espoused argument for reporting randomized trials: explanation and elaboration. Ann Intern Med 134: 663–694.
    • View Article
    • Google Scholar
14. 14. Stroke Therapy Academic Industry Roundtable (STAIR) (1999) Recommendations for standards regarding preclinical neuroprotective and restorative drug development. Stroke 30: 2752–2758.
    • View Commodity
    • Google Scholar
15. 15. Dirnagl U (2006) Bench to bedside: the quest for quality in experimental stroke research. J Cereb Blood Flow Metab 26: 1465–1478.
    • View Article
    • Google Scholar
16. 16. Sena Due east, Van der Worp HB, Howells D, Macleod Yard (2007) How can nosotros amend the pre-clinical evolution of drugs for stroke? Trends Neurosci 30: 433–439.
    • View Commodity
    • Google Scholar
17. 17. Macleod MR, Fisher M, O'Collins Five, Sena ES, Dirnagl U, et al. (2009) Practiced laboratory practice: preventing introduction of bias at the bench. Stroke forty: e50–e52.
    • View Article
    • Google Scholar
18. 18. Rosenthal R (1966) Experimenter furnishings in behavioral enquiry. New York: Appleton-Century-Crofts.
19. xix. Van der Worp HB, de Haan P, Morrema E, Kalkman CJ (2005) Methodological quality of fauna studies on neuroprotection in focal cerebral ischaemia. J Neurol 252: 1108–1114.
    • View Article
    • Google Scholar
20. 20. Van der Worp HB, Sena ES, Donnan GA, Howells DW, Macleod MR (2007) Hypothermia in animal models of acute ischaemic stroke: a systematic review and meta-analysis. Brain 130: 3063–3074.
    • View Article
    • Google Scholar
21. 21. Campbell MJ, Julious SA, Altman DG (1995) Estimating sample sizes for binary, ordered categorical, and continuous outcomes in two grouping comparisons. B M J 311: 1145–1148.
    • View Commodity
    • Google Scholar
22. 22. Festing MF, Altman DG (2002) Guidelines for the pattern and statistical analysis of experiments using laboratory animals. ILAR J 43: 244–258.
    • View Article
    • Google Scholar
23. 23. Bath PM, Macleod MR, Green AR (2009) Emulating multicentre clinical stroke trials: a new paradigm for studying novel interventions in experimental models of stroke. Int J Stroke 4: 471–479.
    • View Article
    • Google Scholar
24. 24. Lee DS, Nguyen QT, Lapointe N, Austin PC, Ohlsson A, et al. (2003) Meta-analysis of the furnishings of endothelin receptor blockade on survival in experimental heart failure. J Carte du jour Fail 9: 368–374.
    • View Article
    • Google Scholar
25. 25. Macleod MR, Van der Worp HB, Sena ES, Howells DW, Dirnagl U, et al. (2008) Evidence for the efficacy of NXY-059 in experimental focal cerebral ischaemia is confounded by report quality. Stroke 39: 2824–2829.
    • View Article
    • Google Scholar
26. 26. Samaranayake S (2009) Report Quality in Experimental Stroke. Camarades Monograph Number 2, http://www.camarades.info/index_files/CM2.pdf. (accessed 22/12/09).
27. 27. Miettinen OS (1983) The demand for randomisation in the study of intended furnishings. Stat Med 2: 267–271.
    • View Commodity
    • Google Scholar
28. 28. Schulz KF, Chalmers I, Hayes RJ, Altman DG (1995) Emperical evidence of bias. Dimensions of methodological quality associated with estimates of treatment furnishings in controlled trials. JAMA 273: 408–412.
    • View Article
    • Google Scholar
29. 29. Noseworthy JH, ebers GC, Vandervoort MK, Farquhar RE, Yetisir Eastward, et al. (1994) The impact of blinding on the results of a randomized, placebo-controlled multiple sclerosis clinical trial. Neurology 44: 16–xx.
    • View Commodity
    • Google Scholar
30. 30. Schulz KF, Grimes DA (2002) Blinding in randomised trials: hiding who got what. Lancet 359: 696–700.
    • View Article
    • Google Scholar
31. 31. Schulz KF, Grimes DA (2002) Allocation concealment in randomised trials: defending against deciphering. Lancet 359: 614–618.
    • View Article
    • Google Scholar
32. 32. Schulz KF, Grimes DA (2005) Sample size calculations in randomised trials: mandatory and mystical. Lancet 365: 1348–1353.
    • View Article
    • Google Scholar
33. 33. Bebarta Five, Luyten D, Heard Thou (2003) Emergency medicine animal research: Does employ of randomization and blinding affect the results? Acad Emerg Med ten: 684–687.
    • View Commodity
    • Google Scholar
34. 34. Macleod MR, O'Collins T, Howells DW, Donnan GA (2004) Pooling of animal experimental information reveals influence of report design and publication bias. Stroke 35: 1203–1208.
    • View Commodity
    • Google Scholar
35. 35. Crossley NA, Sena E, Goehler J, Horn J, van der WB, et al. (2008) Empirical evidence of bias in the blueprint of experimental stroke studies: a metaepidemiologic approach. Stroke 39: 929–934.
    • View Commodity
    • Google Scholar
36. 36. Van der Worp HB, Raaijmakers TW, Kappelle LJ (2008) Early complications of ischemic stroke. Curr Treat Options Neurol 10: 440–449.
    • View Commodity
    • Google Scholar
37. 37. Saver JL, Johnston KC, Homer D, Wityk R, Koroshetz W, et al. (1999) Infarct volume as a surrogate or auxiliary result measure in ischemic stroke clinical trials. Stroke 30: 293–298.
    • View Commodity
    • Google Scholar
38. 38. The National Institute of Neurological Disorders and Stroke (NINDS) rt-PA Stroke Study Group (2000) Upshot of intravenous recombinant tissue plasminogen activator on ischemic stroke lesion size measured by computed tomography. Stroke 31: 2912–2919.
    • View Article
    • Google Scholar
39. 39. Kimmelman J, London AJ, Ravina B, Ramsay T, Bernstein Chiliad, et al. (2009) Launching invasive, start-in-human being trials against Parkinson'southward disease: ethical considerations. Mov Disord 24: 1893–1901.
    • View Article
    • Google Scholar
40. 40. Scott S, Kranz JE, Cole J, Lincecum JM, Thompson K, et al. (2008) Pattern, ability, and interpretation of studies in the standard murine model of ALS. Amyotroph Lateral Scler 9: 4–15.
    • View Article
    • Google Scholar
41. 41. Dwan One thousand, Altman DG, Arnaiz JA, Bloom J, Chan AW, et al. (2008) Systematic review of the empirical testify of study publication bias and upshot reporting bias. PLoS I iii: e3081.
    • View Article
    • Google Scholar
42. 42. Dirx MJ, Zeegers MP, Dagnelie PC, van den Bogaard T, van den Brandt PA (2003) Energy brake and the chance of spontaneous mammary tumors in mice: a meta-assay. Int J Cancer 106: 766–770.
    • View Article
    • Google Scholar
43. 43. Macleod MR, O'Collins T, Horky LL, Howells DW, Donnan GA (2005) Systematic review and metaanalysis of the efficacy of FK506 in experimental stroke. J Cereb Claret Menstruation Metab 25: 713–721.
    • View Article
    • Google Scholar
44. 44. Juutilainen J, Kumlin T, Naarala J (2006) Do extremely depression frequency magnetic fields enhance the effects of ecology carcinogens? A meta-analysis of experimental studies. Int J Radiat Biol 82: i–12.
    • View Article
    • Google Scholar
45. 45. Benatar M (2007) Lost in translation: treatment trials in the SOD1 mouse and in human being ALS. Neurobiol Dis 26: 1–13.
    • View Article
    • Google Scholar
46. 46. Neitzke U, Harder T, Schellong Yard, Melchior K, Ziska T, et al. (2008) Intrauterine growth restriction in a rodent model and developmental programming of the metabolic syndrome: a disquisitional appraisement of the experimental evidence. Placenta 29: 246–254.
    • View Commodity
    • Google Scholar
47. 47. Sena ES, Van der Worp HB, Bathroom PMW, Howells DW, Macleod MR (2010) Publication bias in reports of animal stroke studies leads to major overstatement of efficacy. PloS Biology. PLoS Biol viii(3): e1000344.
    • View Article
    • Google Scholar
48. 48. De Angelis C, Drazen JM, Frizelle FA, Haug C, Hoey J, et al. (2004) Clinical trial registration: a argument from the International Committee of Medical Journal Editors. N Engl J Med 351: 1250–1251.
    • View Commodity
    • Google Scholar
49. 49. Willmot M, Gibson C, Grayness Fifty, Potato Due south, Bathroom P (2005) Nitric oxide synthase inhibitors in experimental ischemic stroke and their furnishings on infarct size and cerebral blood flow: a systematic review. Complimentary Radic Biol Med 39: 412–425.
    • View Article
    • Google Scholar
50. 50. Willmot 1000, Greyness L, Gibson C, White potato Due south, Bathroom PM (2005) A systematic review of nitric oxide donors and L-arginine in experimental stroke; furnishings on infarct size and cerebral claret menses. Nitric Oxide 12: 141–149.
    • View Article
    • Google Scholar
51. 51. Gibson CL, Gray LJ, Tater SP, Bath PM (2006) Estrogens and experimental ischemic stroke: a systematic review. J Cereb Blood Menses Metab 26: 1103–1113.
    • View Article
    • Google Scholar
52. 52. Gibson CL, Grey LJ, Bathroom PM, Potato SP (2008) Progesterone for the handling of experimental brain injury; a systematic review. Brain 131: 318–328.
    • View Article
    • Google Scholar
53. 53. Banwell V, Sena ES, Macleod MR (2009) Systematic review and stratified meta-analysis of the efficacy of interleukin-1 receptor adversary in brute models of stroke. J Stroke Cerebrovasc Dis xviii: 269–276.
    • View Article
    • Google Scholar

Source: https://journals.plos.org/plosmedicine/article?id=10.1371%2Fjournal.pmed.1000245

Posted by: oliveirahileboseek.blogspot.com

Share this post