Table 3.

Explanatory Accompaniment to the Checklist for Methodological and Reporting Features of Studies Involving Experimental Interventions in Animals

Study designThe protocol should be written before commencing the study. If the study is exploratory (eg, the first test of a new therapy or manipulation in vivo), or if it is primarily mechanistic (ie, its primary purpose is to elucidate mechanisms rather than establish efficacy), definition of primary and secondary end points is not necessary. However, if the study is confirmatory (ie, its purpose is to validate the effect of a therapy or intervention), primary and secondary end points should be defined and reported. Furthermore, if the study is confirmatory and if multiple groups are compared with respect to the primary end point, corrections for a type I error should be used.
Inclusion and exclusion criteriaA priori defined criteria to include and exclude animals need to be established and described. These may be various parameters like animals’ species, sex, strain, weight, developmental state, source nomenclature, and genotype, drug or test naive, and so on. Any criteria used for inclusion/exclusion after the protocol is started (eg, loss of cardiac function after infarction) should also be reported.
RandomizationRandomization is considered the gold standard for studies intended to test hypotheses and should be used for assignment of animals to control or experimental groups. Certain studies, such as those with only a single arm (intervention only) or those with exploratory or mechanistic aims (vide supra), may not need to use randomization. Authors are asked to provide clear explanations of such exceptions. The type and methods used to randomize animals to various study arms should be reported.
a. Type of randomization: for example, simple or stratified. If stratified, please describe which variables/factors were used for randomization and why. It is also recommended that investigators report measures of successful randomization, by providing baseline characteristics of animals in various comparison groups.
b. Process of randomization: please provide details of the process used for randomization. These may include (but are not limited to) use of electronically generated lists, sealed envelopes, coin toss, dice roll, etc.
c. Other procedural details pertaining to randomization include describing the independence of team members that generate randomization sequences, perform randomization and group allocation, and conduct experimentation.
Allocation concealment is an important aspect of randomization. If the study team member(s) performing randomization have knowledge of treatment assignment groups before selection of animals, then the allocation is not concealed, and a selection bias can be introduced. Allocation concealment can be achieved if deidentified codes are generated and provided to team members performing the randomization. This usually entails an independent team member producing randomization sequences and lists. The journal encourages use and reporting of such procedures.
BlindingBlinding is essential to avoid bias and must be used whenever possible. Blinding or masking refers to
a. Inability of the investigator(s) that administers the study intervention to discern treatment from the placebo/vehicle. This can be achieved by providing the investigator(s) with animals having codes for group allocation, as well as study drug/product and placebo using a concealment procedure. The journal strongly recommends using and reporting the blinding procedures used for administration of treatment/placebo, wherever possible. For experimentation that in itself is surgical in nature, masking of the investigator(s) performing the study intervention may not be possible. We encourage appropriate reporting of these aspects in the article.
b. Besides the masking of intervention, it is also important that all outcomes be assessed in a blinded manner. Although blinding the investigator(s) who administers the treatment may not be possible in all instances, blinded assessment of functional (eg, echocardiographic), histological (eg, infarct size), and other outcomes is almost universally possible. This is achieved by using an effective tracking mechanism for the experimental animals and via independence of the study team members performing outcome ascertainment from those administering the study intervention. Authors are expected to provide an account of the experimental phases that were blinded to various investigator(s) and the steps undertaken to maintain blinding while transitioning from one experimental phase to the other.
Sample size and power calculationsIn confirmatory studies involving formal hypothesis testing (vide supra), authors should provide an explanation of power and sample size calculations. The rationale for using the chosen sample size should be provided. The rationale for postulating a certain effect size of the intervention on an a priori-determined primary outcome(s) should be presented, along with the anticipated variability. The sources that form the bases of power and sample size calculations should be cited. We do recognize that in exploratory or mechanistic experiments (vide supra), a formal power and sample size calculation may not be possible or meaningful. These exploratory or mechanistic analyses should be presented as such, and an explanation of why power calculations were not formally performed should be provided.
Data reportingControl groups may either be placebo concurrent, no-treatment concurrent, active-treatment concurrent, or dose/timing-comparison concurrent. This needs to be clearly stated. It is required that investigators use a concurrent control for each intervention group. Using singular control data across several experiments conducted during a course of time is similar to using historical controls in clinical research and is a potential source of error.
The number of animals included and analyzed in intervention and control groups should be clearly reported. The authors may consider using a flow diagram (as supplemental material) showing the number of animals available for randomization, number of animals randomized to various treatment/control groups, number of animals in which baseline data were collected, number of animals included in the intervention, number of animals in which follow-up/outcomes were assessed, and number of animals that were lost at any stage during the experimental protocol.
Animal characteristics, such as age, developmental stage, sex, species, genetic strain, and comorbidities, should be reported, along with relevant baseline data (eg, baseline left ventricular function).
Number of deaths or other events leading to exclusion should be reported in all experimental groups. If none, please state so.
When reporting results, investigators are encouraged to provide summary data as tables (ie, authors should not limit data to figures). At times, figures in themselves do not allow for precise reporting of measures of central tendency and spread for different time points. If data are presented as figures, tabulated results can be provided in supplementary materials. When sample sizes are small, dot plots are preferable to bar graphs because they provide more information on variability.
Statistical methodsAuthors are required to provide details on the statistical methods used for each analysis and hypothesis tested. Descriptive analyses should use measures of central tendency and spread suitable to the distribution of the data. Measures of spread (eg, SD or SE) should be suitably used, reported, and interpreted. Use of 95% confidence intervals is important to assess the size of the effect of interventions. Nonparametric tests are more appropriate for data that are not distributed normally. Commonly used tests, such as the Student t test, might not be appropriate when the sample size in each group is small. The levels of significance used for hypothesis testing should be reported, and if any adjustments were done for multiple or repeated testing. Corrections for multiple hypotheses testing and multiple comparisons to avoid type I errors are essential, unless the study is exploratory or mechanistic in nature (vide supra). When the primary outcome of a study is negative, the probability of a type II error should be reported because sample sizes are often insufficient to rule out the null hypothesis with a high level of confidence.
Experimental details, ethics, and funding statementsIt is important to provide as many experimental details as necessary for interpretation, comparison, and, importantly, full replication of the study. These details include (but are not limited to) formulation and dosage of therapeutic agent, site and route of administration, use of anesthesia and analgesia, temperature control during experimentation, and postprocedural monitoring. Temperature is particularly important in open-chest models. Further details on primary and secondary outcomes and methods used for analysis of outcomes should also be provided or clearly referenced.
Authors should give due consideration to the use of animals with both sexes. Assuming only males or females are used, reasons need to be provided.
The journal requires that all studies reporting animal experimentation abide by the respective institutional animal welfare and ethics regulations and reporting of approvals. Funding and conflict of interest statements should also be provided. The funding of the study by a drug development or technology development industry should be clearly stated. Any financial interest of the authors in the product being tested should be clearly indicated.
  • The above table is a modified version of a similar table promulgated by Stroke in 2011 and revised and expanded in 2016. The purpose of the table is to provide important metrics for the conduct and reporting of preclinical studies that involve assessment of experimental interventions in vivo. The recommendations provided herewith are intended to (i) clearly communicate the desired standards of rigor to authors and investigators and (ii) facilitate the journal’s overall mission of promoting robust, rigorous, and reproducible scientific inquiry and experimentation. We recognize the complexity and diversity of many studies, and the existence of different levels of emphasis on in vivo data in different studies. The recommendations outlined herein are meant to be used by reviewers and editors as guidelines, not as rigid criteria for accepting or rejecting articles; thus, decisions on acceptance or rejection will continue to be made on a case-by-case basis depending on the individual content and message of each submitted article. In addition, we recognize that certain standards outlined above may not apply to exploratory, hypothesis-generating, or mechanistic animal studies. We encourage and welcome such studies; however, in these cases, we encourage the authors to address the areas outlined here and describe why adherence to particular standards was not practical or meaningful. We promote the use of online supplements for providing detailed experimental protocols and communicating any secondary results not reported in the main article.