APS Advises NIH on Sex as a Biological Variable

The following comments were submitted in response to questions posed by the NIH in NOT-OD-14-128, a request for information on consideration of sex as a biological variable in biomedical research.

Whether consideration of sex as a biological variable is an issue affecting the reproducibility, rigor, and/or generalizability of research findings.

  • Accounting for sex differences in preclinical research is a critical first step towards the development of personalized medicine.
  • Because sex as a biological variable is an issue of particular concern with respect to generalizability of research findings, it should be considered wherever appropriate and feasible. An important first step would be to require researchers to report the sex of all animal subjects and of the organisms from which all biological materials used in experiments were derived.
  • Reproducibility of findings in animal and cell studies is a related issue that depends in part on genetic and epigenetic differences between animals and cell lines. However, sex is just one of several relevant factors that affect reproducibility.

Areas of science (e.g., cancer, neuroscience) or phases of research (e.g., basic, translational) conducted with animals that have the greatest opportunity or need for considering sex as a biological variable.

  • Translational and clinical research with immediate implications for human medicine should include consideration of sex differences unless there is a sound reason not to do so. Analysis of sex differences is particularly important in proof of concept studies such as translational studies in animal models.
  • Sex is a variable with broad implications for pharmacology and the development of therapeutics because of differences in drug absorption, distribution, metabolism and excretion. Some animal studies investigating effects of sex may need to take into account the estrous state of the animal.
  • Research into conditions and diseases that affect only one sex should be exempted from requirements to consider sex differences. Examples include pregnancy, menopause, cervical cancer, and prostate cancer.
  • Basic research exploring fundamental biological principles without a direct translational or clinical relevance may not always need to consider sex as a biological variable, and policies should be put in place to allow exemptions when determined to be appropriate.

Areas of science or phases of research conducted with cells and/or tissues that have the greatest opportunity or need for considering sex as a biological variable.

  • Some research performed at the molecular and cellular level may be less likely to be affected by sex differences unless there is a chromosomal effect. However, when research on cellular and molecular mechanisms has a strong and direct relevance to more complex systems at the organ and whole organism level, the inclusion of material derived from both sexes may facilitate consideration of sex as a variable in later studies.

Main impediments (e.g. scientific, technical, and other) to considering sex as a biological variable in research.

  • For certain patented animal strains, producers currently only sell male animals to prevent others from developing their own animal colonies. Guidance should be provided on how to deal with this type of situation.
  • In naturally-occurring diseases in nonhuman primates, sex should certainly be reported but it may not be controllable. When such diseases occur sporadically, animals of only one sex may be available at any given time. In addition, since every effort is made to minimize the numbers of non-human primates in research, there may be too few subjects to allow statistical analysis of sex differences.
  • Depending on the species, it may be inappropriate to house animals of different sexes in the same area. If adequate facilities are not available to house the sexes separately, requiring the inclusion of both sexes could significantly affect the reliability of results, especially when experiments depend on the animals being in a calm, steady and conscious state. This is particularly true of some larger animal species. If facilities must be expanded to accommodate both sexes, this will add significant cost and time delays.
  • Since many sex-based differences are due to differences in sex steroid levels, for some studies it may be necessary to study female animals at either one specific phase of the estrous cycle or at various phases of the cycle. When females in various stages of the estrous cycle must be studied, the number of subjects will have to be increased substantially. In some cases this could be addressed by spaying the animals to eliminate hormonal fluctuations.
  • Additional resources will be needed to increase the number of animals or cell lines in an existing study because of the need to cover the costs of acquiring and maintaining the additional animals and materials, along with the cost of conducting additional procedures, assays etc.
  • Researchers will need time to develop new experimental designs that accommodate analysis of sex differences, and they will also need additional time to complete and analyze the additional studies.

Ways in which NIH can facilitate the consideration of sex as a biological variable in NIH-supported research.

  • NIH and members of the scientific community, including professional societies, should undertake a broad campaign of education outlining why consideration of sex as a biological variable is important and how to begin to incorporate it. This campaign should be targeted to a wide audience including students, faculty, and reviewers.
  • As a first step the NIH should begin asking grant applicants to provide information about the sex of all animal subjects and of the organisms from which all biological materials used in experiments were derived. Determination of whether additional experiments will be required to address sex differences should be considered subsequently.
  • NIH should provide guidelines for when using both sexes is required, and when it is not.
  • NIH should also develop guidelines for predictive information that can be used to design drug studies involving both sexes.
  • Some investigators may need training how to implement experimental protocols using both sexes. Related education should include how to do the best power analysis to make sure the investigator has the correct study design and how to craft a justification statement when exemptions are sought. These areas of education would best be targeted to faculty writing the grants.
  • In addition, NIH reviewers and grant management staff will need training in experimental design that includes both sexes. They will also need clear guidelines for evaluating justification statements when exemptions are sought and for determining when exemptions are appropriate. Policies should also be developed for reviewing progress reports of funded grants to determine whether both sexes are being used as proposed in the original application.
  • NIH should consider a request for applications to develop training programs, similar to the BEST program for broadening graduate training.
  • In some cases it may not be clear at the outset whether sex differences may play a role in the outcome of research. The APS recommends that NIH provide funds for pilot projects to determine whether sex differences are likely to exert a significant effect in the research being undertaken.
  • Modifying experimental design to include consideration of sex as a biological variable will add cost to research budgets. Providing additional resources to investigators who are increasing the scope of their proposed research is essential. NIH should also consider additional incentive such as RFAs or supplements to understand the implications of previously published findings in animals or specimens of the other sex.
  • Given the practical and financial implications of incorporating of sex as a biological variable, this policy should be phased in, and priority should be given to areas such as translational research.
  • NIH should allow investigators adequate time to modify their experimental designs to optimize the value of previous work.

Any additional comments you would like to offer to NIH about the development of policies for considering sex as a biological variable in research involving animals, tissues, or cells.

In addition to sex, other biological variables may be important to consider, including:

  • age of the animals being studied
  • genomic variation
  • microbiome variation between colonies of animals
  • differences related to the time of day that animals are handled and disruption of circadian rhythms
  • differences based on the type and brand of feed animals receive

Because resources are not available to explore every possible variable, it will be necessary to consider where the best value lies for each particular research study. For each of the variables listed above, and for sex, detailed factual reporting is a critical first step toward understanding the role that each variable plays in determining experimental outcomes.