Application of the Navigation Guide systematic review methodology to the evidence for developmental and reproductive toxicity of triclosan.
We strongly support EHP ’s proposal encouraging authors who perform animal studies to adhere to the ARRIVE guidelines (http://www.nc3rs.org.uk/ARRIVE/) (Tilson and Schroeder 2013), a 20‐item checklist for the reporting of key elements necessary to describe a study compre hensively and transparently (Kilkenny et al. 2010). In the clinical sphere, systematic review methods have been at the forefront of evidence‐based research for the past 20 years and have demon‐ strated that poor reporting, particularly in domains associated with increased risk of bias, can have a signifi cant impact on the results of a study (Juni et al. 2001). By supporting the ARRIVE guidelines, EHP recognizes the importance of reporting—and ultimately implementing—methodo logi cal approaches that can influence study quality. The ARRIVE guide lines have already been adopted by many journals, including several high‐impact publica tions from the Nature, BioMed Central, and PLoS publishing groups. EHP will join these other prestigious journals in leading the way in recognizing the importance of reporting methodological elements in the toxicological sciences and ultimately strengthening the scientific approach. Although we endorse the use of an approach such as the ARRIVE guidelines, the ARRIVE guidelines do not currently include all important reporting elements in a clearly defined manner. In addition, because the ARRIVE guidelines were developed in the clinical sphere, some elements that are pertinent to toxicological research are not included. We encourage EHP to simulta‐ neously address those issues, including the following: • Authors should adequately report sample sizes per group (Landis et al. 2012) and include explicit details of any losses to follow‐up. We urge EHP to encourage authors to perform and report details of an a priori sample size calculation. • Authors should describe any animals with “peculiarities”; this criterion is described in the “Gold Standard Publication Checklist” by Hooijmans et al. (2010) but is not neces sarily covered by the “adverse events” criteria in the ARRIVE guidelines. • Authors should report where and when the study was performed to aid in assessing whether the cohort of animals is unique from other published studies. • Whenever possible, authors should include in the study doses that are environmentally rele vant (to humans) and a measured concentration in the animal for comparison/integration with human biomarkers. Authors should also report all of their funding sources and include a statement regarding potential conflicts of interest, including when none exist. “Conflict of inter‐ est” risk‐of‐bias domain has been proposed— but not yet adopted—by the Cochrane Collaboration and GRADE (Grading of Recommendations Assessment, Development and Evaluation) as an important risk of bias. This is based on empirical data from studies of the health effects of tobacco (Barnes and Bero 1997, 1998), the safety and efficacy of pharma ceuticals (Bero et al. 2007; Lexchin et al. 2003; Lundh et al. 2012), and medical procedures (Popelut et al. 2010; Shah et al. 2005) that have shown that source of funding influences the study outcome. A criticism of the ARRIVE guidelines is that they are not topic specific. However, we believe that risk‐of‐bias domains used in human experi mental studies that have an empirical basis—including sequence generation, alloca tion concealment, blind‐ ing, incomplete outc ome data, and selective reporting—are directly relevant to toxicologi‐ cal studies. In the clinical sphere, these five criteria address nearly all issues that bear on the quality of human experi mental evidence (Balshem et al. 2011). Further, these elements have been shown in the preclinical animal literature to influence study outcomes (Landis et al. 2012; Vesterinen et al. 2010). Therefore, in the absence of evidence to the contrary, the field of toxicology should proceed based on such comparability. At the same time, it would be of enormous bene fit to the field of toxicology research for EHP to adopt the ARRIVE guidelines and to simultaneously encourage empirically based research to assess which criteria are most criti cal to the results of various types of toxico logi cal studies that provide the evidence for decision making in environmental health. Finally, for the EHP editors to highlight their commitment to improving the quality of the research published in EHP, they could commis sion or encourage research that assesses the current standard or quality of their publica tions and then repeat the exer cise in a few years to see if the guidelines have resulted in improve ments to both reporting and quality of toxico logi cal studies [for example, see Dirnagl and Lauritzen (2011)]. The views expressed in this letter are those of the authors and do not reflect the views of the Oak Ridge Institute for Science and Education. The authors declare they have no actual or potential competing financial interests. Hanna M. Vesterinen,1 Paula I. Johnson,1 Erica Koustas,2 Juleen Lam,3 Patrice Sutton,1 and Tracey J. Woodruff 1 1Program on Reproductive Health and the Environment, University of California, San Francisco, Oakland, California, USA; 2Oak Ridge Institute for Science and Education, Washington, DC, USA; 3Department of Health Policy and Management, Johns Hopkins Bloomberg School of Public Health, Baltimore, Maryland, USA E-mail: VesterinenH@obgyn.ucsf.edu