Considering Sex as a Biological Variable in Research – Your Input is Requested


We as a scientific community have made major progress toward balancing the number of men and women who volunteer as participants in biomedical research studies; in fact, women now account for roughly half of the participants in NIH-funded clinical trials. However, we haven’t seen a similar pattern in the pre-clinical research involving animals and cells. As described in a May Nature article by NIH Director Francis Collins and NIH Associate Director for Research on Women’s Health Janine Clayton, research studies using animal models frequently include only males, and studies using cell models often do not consider the sex of the individual  from which the cells were obtained. And, even if a study design includes male and female specimens or animals, the resulting data may not be analyzed or reported in a way that helps us understand how sex, as a biological research variable, influences the findings.

Sex is a critical variable when trying to understand the biological and behavioral systems that fundamentally shape human health. As such, it also is a critical element to be considered when designing rigorous preclinical research experiments that inform human studies. Thus, as announced in May, NIH intends to develop and implement policies requiring NIH applicants to consider sex as a variable in biomedical research involving animals and cells. As I described in a blog post earlier this year, NIH will be taking a phased approach to this initiative.

We are starting by gathering input from you. So today we announced a formal request for information (RFI) to get input from the research community, and others. As described in the RFI, we want to hear your thoughts on several topics – for example, whether consideration of sex as a biological variable is an issue affecting the reproducibility of research findings; what areas of science or phases of research present the greatest needs or opportunities for understanding data disaggregated by sex; and the major challenges to considering sex as a biological variable in research. We also want to hear your ideas about how NIH can facilitate this aspect of your research, and any other feedback related to NIH’s development of policies for considering sex as a biological variable in research involving animals, tissues, or cells.

Your input will help us shape policies that enhance the rigor of biomedical research, and in turn strengthen the foundation of clinical research.

As always, you are welcome to weigh in on the comment section of the blog, but please refer to the RFI for information on how to submit formal feedback to NIH. We look forward to hearing from you! I also hope you will stay tuned this fall for more information from NIH on how we are working to strengthen science by studying sex.


  1. I am intrigued by this concept and encourage everyone to respond to the RFI. I believe that interesting new data will be generated.

    However, it also feels a bit uncharacteristic of the NIH. What I mean is, generally, the NIH is interested in investigator-driven projects, while the implementation of policy on sex-driven differences (at least the policies I have read about) sounds like NIH-directed science.

    Indeed, since I do not have past training on distinguishing, describing, or characterizing sex-based differences, maybe I am not the target audience. Is it possible to identify investigators who already have this in their toolkit and provide supplemental $. Or announce specific RFAs to address specific questions.

  2. Rhesus monkeys are now used extensively for AIDS vaccine research, for study of AIDS pathogenesis, and for study of viral reservoirs toward a “cure”. Rhesus monkeys are bred in captivity for these and other research purposes and are available in limited supply. Breeding is often accomplished through the use of “harem breeding groups”. A harem breeding group will contain one male, 10-12 females, and the male is expected to keep the females pregnant essentially all of the time. No matter what the breeding program, each female represents one potential new birth per year, but that is not the case for males. Thus, females are much more important for being able to continue to supply maximal numbers of rhesus monkeys for these important research programs. Consequently, male rhesus monkeys are typically more readily available for research protocols. Forced use of increased numbers of females will eventually have severe negative consequences on the availability of rhesus monkeys for these important research programs.

  3. In theory, it is acceptable that preclinical NHP studies designed to develop product that will eventually moved to clinical trials take the possibility of sex-related differences more into account. Having said that, the idea of routinely doubling the size (and budget) of any NHP experiment we conduct just in case there could be a sex difference does not make sense if the work is at the stage in which the potential for clinical development is still unclear.

  4. I applaud this move to ensure that pre-clinical data comes from both sexes of the test animals used. There are clear differences in humans, I would expect this to be presaged by differences in animals. This is not “directive”; it is merely ensuring good and believable scientific experiments. I expect we will see changes biochemically, metabolically, neurologically as well as morphologically.

  5. I agree that for studies on juvenile or adult mice, especially those for pre-clinical testing of drugs or treatments that might be done on humans, that data for males and females should be analyzed separately. However, for those of us who study embryonic events and tissues in mice, this will be a huge additional burden on resources, time, and mouse costs. It would double the number of litters to be collected, genotyped, and double the number of embryos needed to analyze to get enough of each sex. Unless we were studying sexually dimorphic tissues, I don’t see how this information is scientifically important enough to justify doubling our cost and work, when the animals are nowhere near sexual maturity. Mouse studies are already becoming unsustainable with rising per diems and shrinking R01’s. I simply cannot afford it. If NIH is going to demand this, they need to give me a larger R01, not a smaller one each year.

  6. This directive is exceptionally ill-conceived, given the funding climate. Some practical consideration of costs is warranted. If I am studying lung cancer and using cohorts of 6 females per cage, what happens when I have to add 6 males? They must be housed separately in 6 cages because of aggression, so now I have 7 cages per sample point instead of one. The maintenance charges increase 7x, not just 2x, and the size of animal facilities must increase accordingly. This means higher overhead, less time and money for mechanistic studies, and fewer personnel supported by a grant. If there is really some reason to suspect sex differences in lung cancer etiology, why not issue an RFA instead of forcing all lung cancer researchers to address it superficially, and slowing progress in all other cancer research? And what of cell biology? Are we to test whether every protein-protein interaction, every type of channel regulation, and every transcription factor binding event occurs in cell lines from both sexes? Did HEK293 come from a male or female embryo? Is this really the best use of our time and resources?

  7. I firmly urge NIH to put the brakes upon the proposed policy for the sake of maintaining public sanity and preserving competitiveness of American research. The implementation of the proposal would slow down the pace of ongoing innovative research and would stifle potential discoveries. It addresses trivial issues, would not result in any actual improvement, would increase waste and the paperwork burden for all of us, and would further demoralize the researchers. Let me elaborate.

    There are fields of science that are sex-specific already (let’s say, studies of pregnancy, lactation, ovary cancer, or any study where sex differences is the subject matter). This is a very obvious case, where the implementation of policy would result only in adding half-a-page to all grant, IRB and IACUC applications and progress reports.

    There are fields where sex differences are irrelevant, for instance, the development of techniques to measure brain tissue elasticity, conductivity, T2, diffusion tensor, among many others. Such fields include most of biophysics, bioengineering and molecular biology. The only palpable result of the policy implementation here would be, again, an increase in paperwork.

    Finally, there are fields where sex may be a confounding factor (such as studies of, say, metabolism, where inadvertent inclusion of female animals in certain phases of their cycle may confound the metabolic response patterns). In this case you’d want to use as homogeneous pool of male animals as possible. All researchers concerned already know this and do this. If they don’t, this can easily be weeded in peer review. It is just a matter of good science versus junk science. No formal policy needed.

    Here, implementation of the policy may result in two courses of action: either heterogenizing the pool of animals (or cells), or duplicating the study using two separate pools of animals (cells) of different sexes. The first option would compromise the validity of results and obscure the mechanistic implications due to an inhomogeneous pool of animals. (By the way, this is exactly what plagues much of clinical research now.) If sex differences are not the purpose of the study, the second option would blur the focus, increase waste, and force a lot of already busy people to constantly write justifications and apply for waivers. If there were any value and resources to address sex differences in this scenario, good professionals would already do this without being nudged by policymakers. If they are not good professionals, maybe their data are not trustable anyway.

    In summary, although one can imagine a couple of isolated cases where the proposed policy could lead to a positive outcome, the overall impact of its implementation would be overly negative. It would jeopardize scientific progress and threaten the very competitiveness of American laboratories in the world.

  8. Is there any scientific evidence that the cost and increased number of animals in each experiment is going to exceed the benefits? My IACUC says I am supposed to reduce the number of animals as much as possible, have I done that by adding another parameter (out of hundreds) that we do not know whether it affects the outcome or not.

    As we all know science is a zero sum game. If scientists are forced to double (or more) the size of their experiments then they will only be able to conduct half as many experiments. Where is the scientific evidence that this is a worth the cost. So it will take twice as long to get to that cure for lung cancer because at every step we had to prove that there were no male-female difference in the mice. Will the families of those dead lung cancer patients thank us or think we are insane?

  9. As a woman I theoretically support the idea of testing in both sexes. But as a researcher I think we should carefully select when we need testing in both sexes and if the numbers always need to be equal for scientific integrity.
    This proposal would not only drive up costs markedly by increasing the requisite animal numbers but females can be much more expensive because they are the breeding stock for any colony. A breeding stock of many females and a few males can be successful but the opposite is not true.
    For animal welfare, we also always need to look at using as few animals as possible to support or refute any hypothesis.

Before submitting your comment, please review our blog comment policies.

Leave a Reply

Your email address will not be published. Required fields are marked *