In many countries, basic animal research receives more funding than clinical research on the assumption that it would benefit humans in the long-run [1]. Animal experiments are pivotal to decide what treatment should be taken forward in clinical trials, but the reported treatment effect need to be valid and precise. Biased or imprecise results from animal experiments may result in exposing patients to unnecessary risks and wasting scarce research funds. There is a need to report accurate estimates of treatment effects, not only as the previous step to translate treatments to the clinical setting, but also to be used as a landmark for future animal studies. Repeatable, reliable, and valid treatment effect estimates are the starting point for many researchers who want to investigate any intervention.

Targeting myocardial injury that paradoxically occurs with the acute reperfusion of ischaemic myocardium remains one of the top ten unmet clinical needs in cardiology [2]. Although myocardial reperfusion is needed to salvage viable myocardium in patients presenting with acute myocardial infarction (AMI), it comes at a price in what is known as reperfusion-induced injury (IRI). Therapies aimed to protect the heart against IRI, known as cardioprotective therapies [3], include local ischaemic preconditioning (IPC), whereby brief cycles of coronary occlusion and reperfusion elicit protection from a prolonged IRI. This procedure has emerged as the paradigm of cardioprotection [3] in the animal setting being highly reproducible between laboratories and across species [4]. Both local IPC and remote ischaemic preconditioning (RIPC) have become pivotal in the assessment of novel cardioprotective interventions. First, they are used as positive controls when testing for other cardioprotective therapies. Second, they have been used to unveil the molecular architecture behind the protection against IRI, as it is believed that most cardioprotective interventions share common signalling pathways.

Whilst it is well known that both IPC and RIPC protect the heart against IRI, many unanswered questions remain open to debate. For instance, no consensus has been reached on how many cycles should be applied for optimal protection and what their duration and timing should be [5], or to which extent the interval between preconditioning stimulus and the index ischaemia affects the outcome. There is also little evidence on whether local IPC or RIPC provide similar treatment size, and most of this evidence comes from meta-analyses using aggregated data in exceedingly heterogeneous experimental conditions [4, 6]. In this small study, we combined all the available animal data collected over the last 3 years regarding the use of IPC in the Langendorff perfused isolated mouse heart model and RIPC in the in vivo mouse model from our Institution. The aim is to use this pool of more than 90 animals to provide accurate estimate of the effect size of both IPC and RIPC.

All work was conducted in accordance with the Guidelines on the Operation of the Animals (Scientific Procedures) Act 1986, published by The Stationery Office (London, UK), conforming with National Institute of Health Guidelines for the Care and Use of Laboratory Animals.

Our present data indicate that local IPC reduces myocardial infarct size by a mean difference of 24.1% (95% CI 19.5 to 28.6%) and RIPC diminishes myocardial damage by a mean difference of 20.8% (95% CI 14.7 to 26.9%). Notably, the direction of the results was consistent across sets of experiments, and the overall estimate remained consistent in a sensitivity analysis using standardized mean differences. In this study, we pooled the results from eight separate experiments testing for two interventions carried out in our Institute over the last 3 years in order to obtain precise estimates of efficacy.

Efforts to avoid bias and random error are as important when reviewing the results of experimental models as when reviewing the results of human studies, given that this evidence would be used to determine which interventions are taken forward in clinical trials. This would apply for all cardioprotective interventions which can potentially be translated to the clinical setting and would imply that further efforts are needed to accurately estimate treatment effects discounting both the effect of chance and biological variability, before concluding that these results can be translated to the clinical setting thought a randomized trial. Notably, previous meta-analysis in different fields of preclinical research shows that publication bias and methodological flaws are often responsible for systematic effect size overestimation and thereby incorrect conclusion about efficacy [1]. Thus, it is needed to estimate treatment effect using standardized protocols in relatively homogenous experimental conditions. Using data already collected is in accordance with the principles of the 3Rs (reduction and replacement of animals and refinement of procedures), which represent the cornerstone of sensible and streamlined animal research [12].

Since the clinical feasibility of local and remote ischaemic preconditioning in IRI is mainly limited to elective intervention, such as interventional or surgical coronary revascularization, we believe that our results have implication in other equally important aspects. We expect these findings to be used as the standard for other cardioprotective interventions to determine by how much myocardial IS can be reduced following an IRI insult. Our observations may contribute to more successful development of new therapies aimed to reduce myocardial IRI, as cardioprotective therapies providing a limited treatment effect in comparison with expected IS reduction provided by the gold-standard (IPC or RIPC) should be re-assessed before moving the research towards the clinical setting. Moreover, taking into account that our estimate effect size is the result of pooling more than 90 animals in a single dataset, we believe that these results are of potential use when establishing positive controls within a set of experiments. IPC has become the paradigm to study molecular signalling in cardioprotection [8, 10, 13] and researchers can use our values to compare whether their IPC models sufficiently reduce myocardial infarct size or whether they need to change the protocol to obtain a more suitable model. Small rodent animals of AMI have been extensively used over the last decades to elucidate the pathologic responses to IRI, and their use has provided substantial insights of the molecular signalling underlying most of the cardioprotective interventions, thus unveiling potential drug targets.

The magnitude of infarct size reduction using either IPC or RIPC estimated at individual animal-level in our study is similar to those reported in previous meta-analyses using aggregated data in highly heterogeneous backgrounds [4, 6, 14]. Most of these studies lack the consistency of using a single standardized protocol, as most of them have pooled data from both local and remote conditioning and have pooled together results from different protocols (i.e. number and duration of cycles), and some of these results have been reported for different organs (i.e. heart, kidney). Notably, our results define with precision the magnitude of the effect of the protection afforded by IPC in an adequately-sized subset of experiments in two well-defined and recognized models using standard protocols consistently in the same organ.

Given that all the experiments were performed in the same laboratory, we observed less heterogeneity in our statistical analysis in comparison with previous meta-analysis of pre-clinical studies. This is a double-edged sword, as in one hand we obtained more precise estimates, but on the other hand, this should be understood as a limitation to generalize our results. Moreover, caution should be taken when extrapolating our findings to other settings, given that we used young animals with no comorbidities. The failure to translate cardioprotective therapies into the clinical setting has been attributed to a disconnection between animal models and the clinical setting [15]. It is important to appreciate that our study was not designed to address this disconnection. However, it would be important to undertake similar studies to that of the above, using comorbid animals to better understand their influence on myocardial infarction.

In conclusion, using a relatively large animal data set, we found that local IPC reduces myocardial infarct size by 24.1% and RIPC by 20.8% in the ex vivo and in vivo mouse models of IRI, respectively. These values may become reference values for other researchers to either establish their positive controls or to determine by how much myocardial infarct size can be reduced by novel cardioprotective interventions following an IRI insult.