Editorial Ethics in Publications “Although this may seem a paradox, all exact science is dominated by the idea of approximation. When a man tells you that he knows the exact truth about any- thing, you are safe in inferring that he is in inexact man.” Bertrand Russell (1872–1970) Uncertainty is the only certainty in science: nothing is clear- cut. That this is also the case about treatment for major human conditions, including osteoporosis, is an unfortunate reality. Investigations of novel and potentially better treat- ments are certainly warranted. Because of the high costs of success and failure, basic and clinical scientists are subject to predictable financially related pressures. In this arena, therefore, the reliability of what we do “know” is para- mount. Any errors that creep into our knowledge base can delay, distract, and mislead with potentially disastrous con- sequences. When these occur by honest mistake, it is most unfortunate, and correction of known errors is essential. If they arise and are allowed to stand uncorrected from con- scious action, it is indefensible. The importance of this issue relates to the critical marriage between peer review and ethics in the published word. These concepts are of major concern to the editorial team of a scientific journal. The structure of all scientific publication depends on the twin pillars of integrity of the authors and of reviewers. We de- pend on the honesty of the authors and rely on their asser- tions that they have had sufficient access to the data to be convinced of its reliability. We depend on honest and forth- right review unimpeded by any personal or financial bias for or against what is written. Failures or delays to report negative studies are as unac- ceptable as publishing patently false data. Concern about such publication biases has recently lead to the requirement by major medical journals for pharmaceutical companies to register all studies, including listing of primary and second- ary endpoints and preplanned analysis, before enrolling their first patient. The JBMR has agreed to a similar re- quirement as set out below. The JBMR editorial team strongly supports this improvement in transparency. Other critical ethical issues that deserve our careful con- cern as well as informed consent include primarily scientific internal and external validity and the threat from conflict of interest, and statistical issues and validity of endpoints and the fairness in the pursuit of human and animal studies. Other issues include honesty and transparency and appro- priate authorship. Each of these issues requires careful con- sideration by authors, reviewers, and editors. INFORMED CONSENT In human studies, investigators have daunting ethical re- sponsibilities as set out in the Declaration of Helsinki. Each participant must be adequately informed of the risks as well as potential benefits of participation. These points are usu- ally set out in the Patient Information and Informed Con- sent, overseen by any Research Ethics Committee or Insti- tutional Review Board (IRB). However, the availability and lucidity of these patient-directed materials do not over- ride the investigator’s primary responsibility to the partici- pant including the duty to ensure that the participants have a clear understanding of what is considered an acceptable risk (e.g., of fracture). In addition to making sure patients understand the risks, investigators must consider that there is the potential to achieve a meaningful outcome. This con- cept, which is part of the IRB review process, includes the requirement for valid comparators, adequate study size for reasonable power, and clear guidelines on dealing with po- tential conflict of interest. VALID COMPARATORS AND SURROGATE ENDPOINTS Valid comparators are a major concern, especially in os- teoporosis studies. Placebo controls give clear information about benefit of an intervention versus no treatment. How- ever, such studies are no longer justifiable in the high-risk group of men or women with osteoporosis and prior frac- tures, in whom the risk of subsequent fractures is substan- tial. The alternative of active comparator trials leads to un- attainable sample sizes for fracture endpoints. This situation has led to dependence on surrogate endpoints that may be reasonable within drug classes but not between types of agents. The early work with sodium fluoride rang this warning bell early in the osteoporosis field. Although fluoride increased BMD, obvious even on X-ray, placebo- controlled fracture studies showed no fracture reduction. Thus, surrogate endpoints have inherent limitations. Such studies, supported by animal data, may be appropriate for initial proof-of-principle human and dose-finding studies, but are inadequate for new types of agents, in which mecha- nism of action of effect may not be clear-cut. ANIMAL STUDIES Similar issues apply to animal studies (i.e., validity of the scientific question, optimal number of animals used (nei- ther too few nor too many), and scientific validity with meaningful outcomes versus surrogate endpoints). Appro- priate species must be used with humane care and treat- ment and in vitro testing, and alternate models should be fully explored to minimize any suffering and to minimize animal use. EXTRAPOLATION Some extrapolation of data from initial studies, from ini- tial time frames, and from initial study subjects is common JOURNAL OF BONE AND MINERAL RESEARCH Volume 21, Number 1, 2006 Published online on October 17, 2005; doi: 10.1359/JBMR.051010 © 2006 American Society for Bone and Mineral Research 1