The translational roadblock, characterized by beneficial

Transcription

1 Special Report Methodological Quality of Experimental Stroke Studies Published in the Stroke Journal Time Trends and Effect of the Basic Science Checklist Jens Minnerup, MD*; Verena Zentsch, MD*; Antje Schmidt, MD; Marc Fisher, MD; Wolf-Rüdiger Schäbitz, MD The translational roadblock, characterized by beneficial effects of treatments in experimental studies and subsequent neutral or negative results in clinical trials, is a major problem in cerebrovascular medicine. 1 Growing evidence suggests that this may be due, at least in part, to low study quality of experiments resulting in a systematically reduced internal validity of animal studies. The lack of blinding and randomization in experimental studies, for instance, leads to an overestimation of treatment effects and may, therefore, contribute to different results of preclinical and clinical studies. 2 As a consequence, the Stroke Therapy Academic Industry Roundtable was founded to develop recommendations on quality characteristics of animal stroke studies. 3 Stroke is generally regarded as one of the most important journals in the area of cerebrovascular disease, publishing clinical and experimental studies with the latter focusing on basic mechanisms of cerebral ischemia and on the development of new stroke therapies. To insure the quality of experimental studies, authors need to declare in a checklist cornerstones of study conductance, such as randomization and blinding procedures, definition of inclusion and exclusion criteria, etc, when submitting an article to Stroke. However, whether implementation of the so-called Basic Science Checklist in the submission process in August 2011 improved the quality of preclinical studies is unknown so far. The aim of this study was to analyze time trends of quality and design of preclinical studies published in Stroke from January 2010 through December 2013 and whether implementation of the Basic Science Checklist in August 2011 had an effect on key quality measures. Methods Retrieving Publications and Selection of Studies We used the journal s Archive of All Online Issues 4 for the identification of eligible articles published from January 2010 to December Publications were categorized according to the journal s article categories as an original contribution (clinical or basic science), brief report, letter to the editor, and review article. We only included original basic science contributions in which the efficacy or side effects of a therapy were evaluated in an animal model of cerebral ischemia, intracranial bleeding, cerebral venous thrombosis, or intracranial aneurysms. Data Extraction We extracted data on the submission date, year, and month of print publication, the species used, and the stroke model used. We assessed whether the following items of the Basic Science Checklist were met 5 : (1) species, strains, and sources of animals defined, (2) statistical methods defined, (3) specific criteria for inclusions and exclusions specified, (4) randomization, allocation concealment, and blinding performed, and (5) postrandomization all excluded animals reported. The checklist as it seems during the submission process is provided in Figure I in the online-only Data Supplement. The study characteristics animals defined (strain, species, and source) and randomization, allocation concealment, and blinding are summarized as 1 item in the Basic Science Checklist but were also separately recorded for our analysis. In addition, we recorded whether the methods of randomization, blinding, and allocation concealment were stated as suggested by Macleod et al. 5 The item postrandomization excluded animals reported was met when the number of animals that were excluded after treatment allocation for different reasons, including mortality or other reasons, was reported for each treatment group. We further assessed the methodological quality of the included studies according to a previously published Stroke Therapy Academic Industry Roundtable recommendation derived quality scale. 6 8 The realization of the following aspects of each study was evaluated: (1) dose response relationship, (2) optimal time window of the treatment, (3) monitoring of physiological parameters (such as temperature, glucose level, or blood pressure), (4) assessment of at least 2 outcomes (infarct size and 1 functional outcome), (5) outcome assessment in the acute phase (1 7 days), (6) outcome assessment in the chronic phase (beyond 7 days), (7) animals with comorbidity (aged, diabetic, or hypertensive), (8) compliance with animal welfare regulations, (9) statement of potential conflict of interests, and (10) sample size calculation reported. Two authors (V.Z. and A.S.) independently extracted data. Disagreements were solved after discussion of the study details (J.M., V.Z., and A.S.). Received October 1, 2015; final revision received November 2, 2015; accepted November 4, From the Department of Neurology, University of Münster, Münster, Germany (J.M., V.Z., A.S.); Department of Neurology, Beth Israel Deaconess Medical Center, Harvard Medical School, Boston, MA (M.F.); and Department of Neurology, Bethel-Evangelisches Krankenhaus, Bielefeld, Germany (W.-R.S). Guest Editor for this article was Miguel A. Perez-Pinzon, PhD. *Drs Minnerup and Zentsch contributed equally. The online-only Data Supplement is available with this article at /-/DC1. Correspondence to Jens Minnerup, MD, Department of Neurology, University of Münster, Albert-Schweitzer-Campus 1, Münster, Germany. minnerup@uni-muenster.de (Stroke. 2016;47: DOI: /STROKEAHA ) 2015 American Heart Association, Inc. Stroke is available at DOI: /STROKEAHA

2 268 Stroke January 2016 Statistical Analysis To evaluate changes in study characteristics by year, we used the Mantel Haenszel test of trend for categorical and ordinal variables. We performed a logistic regression analysis for categorical variables and Student t test for continuous variables to determine the association of Basic Science Checklist items and whether the study was submitted before or after the implementation of the Basic Science Checklist in August Statistical significance was determined as P<0.05. Statistical analyses were carried out using the Statistical Package of Social Sciences version 22. Results A total of 1617 original contributions were published during the study period. Of these, 1357 (83.9%) were clinical science studies, 133 (8.2%) were basic studies, which evaluated the efficacy of a treatment, and 127 (7.9%) were basic studies that did not investigate the efficacy of a treatment but rather other aspects of cerebral ischemia, that is, its pathophysiology (Figure). The number of original contributions decreased from 348 in 2010 to 312 in The number of clinical science and basic science studies with or without a treatment investigation per year did not differ significantly during the study period (all P>0.05). Study characteristics of included basic science studies, which investigated a treatment, are given in Table 1. During the study period, the proportions of different stroke models and species used in the studies did not vary over time. With respect to Basic Science Checklist items, the proportion of articles that reported the combination of randomization, allocation concealment, and blinding increased (P=0.001). The proportion of each of the subitems randomization, allocation concealment, and blinding also increased over time (P=0.01, 0.001, and 0.004, respectively). None of the studies reported the methods of blinding or allocation concealment. The proportion of studies that met 2 or 3 checklist items decreased and those that met 4 or 5 items increased during the study period. Proportions of the other Basic Science Checklist characteristics and the mean sum of Basic Science Checklist items did not significantly change. Reporting of the Stroke Therapy Academic Industry Roundtable derived quality items investigation of a dose response relationship (P=0.02), Number of published original contributions Clinical science contribution Basic science contribution not investigating a treatment Basic science contribution investigating a treatment physiological monitoring (P=0.02), and compliance with animal welfare regulations (P=0.02) significantly increased. There was a trend toward an increased reporting of assessment of infarct size and functional outcome (P=0.08) and outcome assessment in the acute phase (P=0.05). The mean of the sum of all quality characteristics significantly increased from 7.04 in 2011 to 8.33 in 2013 (P=0.002). The proportion of studies with the lowest category of the sum of quality characteristics (0 5 items met) decreased during the study period. As required by Stroke, all studies stated a potential conflict of interest. Table 2 shows the results of the univariate analysis comparing whether Basic Science Checklist items of published articles submitted before and after implementation of the checklist in August 2011 were met. The proportion of studies that defined inclusion and exclusion criteria increased from 22.7% to 41.4% (P=0.02). More articles reported randomization, allocation concealment, and blinding after implementation of the checklist (6.7% versus 24.1%; P=0.007). Among subitems, allocation concealment (8.0% versus 25.9%; P=0.008), and blinding (70.7% versus 86.2%; P=0.04) significantly increased after checklist implementation. The majority of studies defined the animals used with no difference before and after checklist implementation (81.3% versus 74.1%; P=0.32). Particularly, all studies reported the species, and almost all studies (97.3% and 98.3%, respectively) reported the strain used. Also all but 1 study described their statistical methods. There was a trend toward an increased mean of the sum of checklist items (P=0.09). The proportion of studies, which met 2 or 3 items decreased from 78.7% to 51.7% (P=0.01) and the proportion of studies, which met 4 or 5 items increased from 10.7% to 31.0% (P=0.005) after checklist implementation. Discussion Systematic assessment of all animal experimental studies investigating a treatment published in Stroke between January 2010 and December 2013 revealed improvements in the reporting of key characteristics of scientific quality, such as Figure. Trends in published original contributions categorized as clinical science contribution, basic science contribution not investigating a treatment, and basic science contribution investigating a treatment Year of publication

3 Minnerup et al Quality of Experimental Studies Published in Stroke 269 Table 1. Characteristics of Animal Experimental Studies* Published in Stroke From 2010 to (n=28) 2011 (n=42) 2012 (n=24) 2013 (n=39) P Values for Trend Stroke model, n (%) Focal cerebral ischemia 20 (71.4) 27 (64.3) 18 (75.0) 27 (69.2) 0.89 Carotid artery ligation or bilateral stenosis 5 (17.9) 6 (14.3) 3 (12.5) 3 (7.7) 0.21 Intracerebral or subarachnoid hemorrhage 2 (7.1) 8 (19.0) 2 (8.3) 10 (25.6) 0.11 Aneurysm or arteriovenous malformation 0 (0.0) 1 (2.4) 2 (8.3) 1 (2.6) 0.42 Cerebral venous sinus thrombosis 1 (3.6) 0 (0.0) 0 (0.0) 0 (0.0) 0.17 Species, n (%) Mice 5 (17.9) 18 (42.9) 12 (50.0) 18 (46.2) 0.03 Rats 23 (82.1) 23 (54.8) 15 (62.5) 22 (56.4) 0.11 Rabbits 0 (0.0) 1 (2.4) 0 (0.0) 1 (2.6) 0.57 Basic Science Checklist, n (%) (1) Animals defined (strain, species, and source) 24 (85.7) 33 (78.6) 19 (79.2) 28 (71.8) 0.20 Species 28 (100.0) 42 (100.0) 24 (100.0) 39 (100.0) N/A Strain 28 (100.0) 40 (95.2) 24 (100.0) 38 (97.4) 0.86 Source of animals 24 (85.7) 34 (81.0) 19 (79.2) 28 (71.8) 0.16 (2) Statistical method defined 27 (96.4) 42 (100.0) 24 (100.0) 39 (100.0) 0.17 (3) Inclusion and exclusion criteria defined 9 (32.1) 8 (19.0) 8 (33.3) 16 (41.0) 0.17 (4) Randomization, allocation concealment, and 0 (0.0) 4 (9.5) 4 (16.7) 11 (28.2) blinding Randomization 14 (50.0) 27 (64.3) 17 (70.8) 31 (79.5) 0.01 Method of randomization stated 1 (3.6) 1 (2.4) 7 (29.2) 2 (5.1) 0.28 Allocation concealment 0 (0.0) 5 (11.9) 4 (16.7) 12 (30.8) Blinding 16 (57.1) 33 (78.6) 19 (79.2) 35 (89.7) (5) Postrandomization excluded animals reported 15 (53.6) 13 (31.0) 3 (12.5) 11 (28.2) 0.64 Sum of Basic Science Checklist items Mean (SD) 2.68 (0.77) 2.38 (0.94) 2.58 (1.02) 2.85 (1.20) or 1 item met, n (%) 2 (7.1) 6 (14.3) 3 (12.5) 7 (17.9) or 3 items met, n (%) 23 (82.1) 31 (73.8) 15 (62.5) 20 (51.3) or 5 items met, n (%) 3 (10.7) 5 (11.9) 6 (25.0) 12 (30.8) 0.02 STAIR-derived quality recommendations, n (%) Dose response relationship 2 (7.1) 11 (26.2) 11 (45.8) 12 (30.8) 0.02 Time window investigation 3 (10.7) 4 (9.5) 9 (37.5) 4 (10.3) 0.54 Physiological monitoring 15 (53.6) 31 (73.8) 20 (83.3) 31 (79.5) 0.02 Assessment of infarct size and functional outcome 15 (53.6) 22 (52.4) 17 (70.8) 27 (69.2) 0.08 Outcome assessment in the acute phase (1 7 d) 21 (75.0) 37 (88.1) 22 (91.7) 36 (92.3) 0.05 Outcome assessment in the chronic phase (>7 d) 10 (35.7) 15 (35.7) 10 (41.7) 15 (38.5) 0.73 Animals with comorbidity 1 (3.6) 2 (4.8) 1 (4.2) 4 (10.3) 0.25 Compliance with animal welfare regulations 25 (89.3) 40 (95.2) 24 (100.0) 39 (100.0) 0.02 Statement of potential conflict of interest 28 (100.0) 42 (100.0) 24 (100.0) 39 (100.0) N/A Sample size calculation 2 (7.1) 3 (7.1) 3 (12.5) 7 (17.9) 0.11 Sum of all quality characteristics Mean (SD) 7.04 (2.13) 7.31 (1.65) 8.46 (1.93) 8.33 (2.07) items met, n (%) 7 (25.0) 4 (9.5) 0 (0.0) 2 (5.1) items met, n (%) 20 (71.4) 37 (88.1) 20 (83.3) 33 (84.6) items met, n (%) 1 (3.6) 1 (2.4) 4 (16.7) 4 (10.3) 0.11 Percentages of stroke models and species sum up to >100% because some studies used >1 model or species. N/A indicates not applicable; and STAIR, Stroke Therapy Academic Industry Roundtable. *Only experimental stroke studies that investigated the efficacy or the side effects of a treatment. Includes middle cerebral artery occlusion, photothrombosis, and endothelin-1 injection. This criterion includes randomization, allocation concealment, and blinding but not the method of randomization. Sum of the items: (1) animals defined, (2) statistical method defined, (3) inclusion and exclusion criteria defined, (4) randomization, allocation concealment, and blinding, and (5) postrandomization excluded animals reported. Includes temperature, glucose level, or blood pressure. Includes Basic Science Checklist items and STAIR-derived quality recommendations.

4 270 Stroke January 2016 Table 2. Comparison of Quality Characteristics of Animal Experimental Studies* Submitted to Stroke Before and After the Implementation of the Basic Science Checklist in August 2011, Univariate Analysis Submission Before Checklist Implementation (n=75) Submission After Checklist Implementation (n=58) OR 95% CI P Values Basic Science Checklist, n (%) (1) Animals defined (strain, species, and 61 (81.3) 43 (74.1) source) Species 75 (100.0) 58 (100.0) N/A N/A NA Strain 73 (97.3) 57 (98.3) Source of animals 62 (82.7) 43 (74.1) (2) Statistical method defined 74 (98.7) 58 (100.0) (3) Inclusion and exclusion criteria defined 17 (22.7) 24 (41.4) (4) Randomization, allocation concealment, 5 (6.7) 14 (24.1) and blinding Randomization 46 (61.3) 43 (74.1) Method of randomization stated 5 (6.7) 6 (10.3) Allocation concealment 6 (8.0) 15 (25.9) Blinding 53 (70.7) 50 (86.2) (5) Postrandomization excluded animals 29 (38.7) 23 (39.7) reported Sum of Basic Science Checklist items, mean (SD) Mean (SD) 2.48 (0.86) 2.79 (1.17) N/A N/A or 1 item met, n (%) 8 (10.7) 10 (17.2) or 3 items met, n (%) 59 (78.7) 30 (51.7) or 5 items met, n (%) 8 (10.7) 18 (31.0) Percentages of stroke models and species sum up to >100% because some studies used >1 model or species. CI indicates confidence interval; N/A, not applicable; and OR, odds ratio. *Only experimental stroke studies that investigated the efficacy or side effects of a treatment were included. This criterion includes randomization, allocation concealment, and blinding but not the method of randomization. Sum of the items: (1) animals defined, (2) statistical method defined, (3) inclusion and exclusion criteria defined, (4) randomization, allocation concealment, and blinding, and (5) postrandomization excluded animals reported. definition of inclusion and exclusion criteria, randomization, allocation concealment, blinding, and physiological monitoring. The overall quality measured by a sum score of quality items also improved over time. When comparing whether items of the Basic Science Checklist were met in published studies submitted before and after checklist implementation, we found an increased reporting of inclusion and exclusion criteria definition, allocation concealment, blinding, and an increase of studies with the highest category of the checklist item sum score. However, relevant components of quality did not improve over time or improved but still had a low prevalence, including inclusion and exclusion criteria defined, method of randomization stated, allocation concealment, and postrandomization excluded animals reported. Our results should be interpreted in the context of previous systematic reviews and meta-analyses of preclinical studies. Systematic analyses of experimental stroke studies investigating neuroprotective treatments found that randomization of animals was reported in 42% to 50% compared with 61% before and 74% after checklist implementation in our study. 7,9 Studies on Alzheimer disease, multiple sclerosis, and cancer reported randomization in only 22%, 16%, and 28%, respectively Although these comparisons suggest that the stroke field does better than other research areas, our study reveals that only few studies describe whether a proper procedure for randomization was followed. Reporting the method of randomization is important as picking animals at random from a cage are considered unlikely to provide adequate randomization. 5 Blinded assessment of outcome and allocation concealment are further key quality measures. Systematic evaluations found blinding to be reported in 40% to 58% of experimental stroke studies, 7,9 in 22% of studies on Alzheimer disease, 12 and in 16% of studies on multiple sclerosis, 11 compared with 71% before and 86% after checklist implementation in the present analysis. A review of 100 articles on cancer showed that only 2% reported that observers were blinded to treatment. 10 However, 1 might consider that these studies were published earlier and quality might have improved within the last years. Before implementation of the checklist, reporting of allocation concealment in our study was within the range of previously published studies and increased by 3-fold after checklist implementation. 13 However, allocation concealment is still reported in only 26% of the analyzed studies. Outliers in animal studies are not unusual, and their exclusion can influence the study results. For example, a treatment may seem effective if it kills the most severely affected animals and their neurological outcome and infarct volume are not considered for analyses. 14 Therefore, the definition of

5 Minnerup et al Quality of Experimental Studies Published in Stroke 271 exclusion criteria and reporting of postrandomization excluded animals are part of the Basic Science Checklist. Reporting a definition of inclusion and exclusion criteria almost doubled from 23% to 41% after checklist implementation, whereas the description of postrandomization excluded animals remained stable (39% before and 40% after checklist implementation). Definition of exclusion and inclusion criteria and reporting of excluded animals have been hardly investigated in systematic studies so far. An analysis of studies published in the Journal of Cerebral Blood Flow and Metabolism in 2008 found that 19% reported inclusion and exclusion criteria and 8% mortality. 11 However, reporting excluded animals might be overstated in our study because this item was fulfilled once mortality was reported, although exclusion for other reasons might have occurred that was not reported. Indeed, the relatively high number of excluded animals in a recent preclinical randomized controlled multicenter trial suggests that exclusion of animals for different reasons is more frequent as reported in monocenter studies. 15 The need for improving the quality of animal studies is obvious as inadequate study designs and reporting were shown to correlate with overstated findings. 2,16 Strategies for achieving this goal were adopted from the clinical trial community, which developed and implemented the Consolidated Standards or Reporting Trials (CONSORT) guidelines. Analogous to the CONSORT statement, the Animal Research Reporting In vivo Experiments guidelines that were published in 2010 provide recommendations for the design and reporting of animal experimental studies. 17 The question is how to enforce following such recommendations. In a landmark article on how to improve the methodological reporting of animal studies, Landis and et al 14 proposed checklist implementation during submission. Our results support this strategy. However, a definitive causal relationship between the implementation of the Basic Science Checklist and the observed improvements in quality cannot be made. Actually, quality characteristics not included in the checklist were also increasingly met during the study period, including physiological monitoring, compliance with animal welfare regulations, and dose response relationship investigation, thus suggesting that a general trend toward an improved quality exists, which is independent of checklist implementation. Our study has strengths and limitations. A limitation of our study might be that there is no final proof for Stroke Therapy Academic Industry Roundtable derived quality criteria to improve translation because no treatment, either following or not following these recommendations, has been shown to predict a positive phase III clinical trial. However, the importance of quality standards in preclinical studies was unequivocally demonstrated. 2 Another limitation is that some checklist items can hardly be judged unambiguously. Two examples are discussed above, that is, the reporting of excluded animals and randomization. In addition, although blinding and allocation concealment were frequently stated, their methods were not reported in any study. We further cannot exclude the possibility that some studies may have met key methodological parameters but did not report them. However, previous analyses showed that deficiencies in reporting quality standards itself are associated with overstated findings. Strengths of our study include the comprehensive analysis of experimental studies published in Stroke, a leading journal in the field, during 4 years that sufficiently cover the period before and after implementation of the Basic Science Checklist. Data were extracted independently by 2 assessors, and disagreements were solved after discussion with a third party. Conclusions In this first systematic analysis of the methodological quality of experimental studies published in Stroke, we found that relevant key measures, including randomization, blinding, and allocation concealment, improved during the study period. Quality standards were more frequently met in studies submitted after implementation of the Basic Science Checklist, suggesting that the checklist contributed to the improved quality. However, we also identified deficiencies, such as low rates of reporting the methods of randomization, inclusion and exclusion criteria, excluded animals, and allocation concealment. For further improvements, we, therefore, propose mandatory reporting of key methodological parameters in the published article and not only during submission. Such reporting can be implemented at the end of the article like the disclosure statement or in a separate formal panel. Besides reporting quality characteristics, adherence to checklist items should be mandatory for article acceptance. Moreover, information on quality characteristics should be more detailed and unambiguous. Sources of Funding J. Minnerup was supported by the Else Kröner-Fresenius-Stiftung (2014_EKES.16), and A. Schmidt was supported by the Medical faculty of the University of Muenster. None. Disclosures References 1. O Collins VE, Macleod MR, Donnan GA, Horky LL, van der Worp BH, Howells DW. 1,026 experimental treatments in acute stroke. Ann Neurol. 2006;59: doi: /ana Crossley NA, Sena E, Goehler J, Horn J, van der Worp B, Bath PM, et al. Empirical evidence of bias in the design of experimental stroke studies: a metaepidemiologic approach. Stroke. 2008;39: doi: / STROKEAHA Stroke Therapy Academic Industry Roundtable (STAIR). Recommendations for standards regarding preclinical neuroprotective and restorative drug development. Stroke. 1999;30: Stroke. Archive of All Online Issues. Accessed August 2, Macleod MR, Fisher M, O Collins V, Sena ES, Dirnagl U, Bath PM, et al. Good laboratory practice: preventing introduction of bias at the bench. Stroke. 2009;40:e50 e52. doi: /STROKEAHA Minnerup J, Heidrich J, Wellmann J, Rogalewski A, Schneider A, Schäbitz WR. Meta-analysis of the efficacy of granulocyte-colony stimulating factor in animal models of focal cerebral ischemia. Stroke. 2008;39: doi: /STROKEAHA Minnerup J, Wersching H, Diederich K, Schilling M, Ringelstein EB, Wellmann J, et al. Methodological quality of preclinical stroke studies is not required for publication in high-impact journals. J Cereb Blood Flow Metab. 2010;30: doi: /jcbfm Minnerup J, Heidrich J, Rogalewski A, Schäbitz WR, Wellmann J. The efficacy of erythropoietin and its analogues in animal stroke models: a meta-analysis. Stroke. 2009;40: doi: / STROKEAHA

6 272 Stroke January Philip M, Benatar M, Fisher M, Savitz SI. Methodological quality of animal studies of neuroprotective agents currently in phase II/III acute ischemic stroke trials. Stroke. 2009;40: doi: / STROKEAHA Hess KR. Statistical design considerations in animal studies published recently in cancer research. Cancer Res. 2011;71:625. doi: / CAN Vesterinen HM, Sena ES, ffrench-constant C, Williams A, Chandran S, Macleod MR. Improving the translational hit of experimental treatments in multiple sclerosis. Mult Scler. 2010;16: doi: / Sena ES, Currie GL, McCann SK, Macleod MR, Howells DW. Systematic reviews and meta-analysis of preclinical studies: why perform them and how to appraise them critically. J Cereb Blood Flow Metab. 2014;34: doi: /jcbfm Frantzias J, Sena ES, Macleod MR, Al-Shahi Salman R. Treatment of intracerebral hemorrhage in animal models: meta-analysis. Ann Neurol. 2011;69: doi: /ana Landis SC, Amara SG, Asadullah K, Austin CP, Blumenstein R, Bradley EW, et al. A call for transparent reporting to optimize the predictive value of preclinical research. Nature. 2012;490: doi: /nature Llovera G, Hofmann K, Roth S, Salas-Pérdomo A, Ferrer-Ferrer M, Perego C, et al. Results of a preclinical randomized controlled multicenter trial (prct): anti-cd49d treatment for acute brain ischemia. Sci Transl Med. 2015;7:299ra121. doi: /scitranslmed.aaa Macleod MR, van der Worp HB, Sena ES, Howells DW, Dirnagl U, Donnan GA. Evidence for the efficacy of NXY-059 in experimental focal cerebral ischaemia is confounded by study quality. Stroke. 2008;39: doi: /STROKEAHA Kilkenny C, Browne WJ, Cuthill IC, Emerson M, Altman DG. Improving bioscience research reporting: the ARRIVE guidelines for reporting animal research. PLoS Biol. 2010;8:e doi: /journal. pbio Key Words: animals cerebral infarction checklist models, animal stroke