1
Harnessing the Power of Sex Differences: What a Difference Ten
Years Did Not Make
Rebecca K. Rechlin
#1,
5
, Tallinn F.L. Splinter
#2,
5
, Travis E. Hodges
1,
5
, Arianne Y. Albert
3,5
,
Liisa A.M. Galea*
4,
5
1
Department of Psychology,
2
Department of Biology,
3
Women's Health Research Institute of British
Columbia,
4
Djavad Mowafaghian Centre for Brain Health,
5
Women’s Health Research Cluster,
University of British Columbia, Vancouver, BC, Canada
# both authors contributed equally
*Address all correspondence to:
L. A. M. Galea, PhD
Djavad Mowafaghian Centre for Brain Health
2215 Wesbrook Mall
Vancouver, British Columbia
V6T 1Z3, Canada
E-mail: liisa.galea@ubc.ca
.CC-BY-NC-ND 4.0 International licensemade available under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
The copyright holder for this preprintthis version posted July 1, 2021. ; https://doi.org/10.1101/2021.06.30.450396doi: bioRxiv preprint
2
Highlights
68% of Neuroscience and Psychiatry papers reported the use of both sexes in 2019
Only 19% of studies in 2019 used sex consistently throughout the study analyses
Of the studies that used males and females, 59% did not include sex in the analyses
Only 5% of studies in 2019 used sex as a discovery variable in their analyses
Male only papers were 8.4 times more prevalent than female-only papers
.CC-BY-NC-ND 4.0 International licensemade available under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
The copyright holder for this preprintthis version posted July 1, 2021. ; https://doi.org/10.1101/2021.06.30.450396doi: bioRxiv preprint
3
Abstract
Sex differences exist in a variety of neurological and psychiatric diseases but most past research
has been conducted in males. Multiple mandates have been initiated across funding agencies and
scientific publishers for research to include males and females in research. What has been lacking
in the literature is a detailed assessment of how sex is incorporated into the design (e.g. balanced
design) and into the analyses (e.g. covariate). Here we investigated papers in 2009 and 2019 in six
journals in Neuroscience and Psychiatry. We found a 30% increase in the percentage of papers that
included both sexes from 2009 to 2019 to 68% in 2019. Despite this increase, only 19% of all
studies used an optimal design for discovery of possible sex differences and only 5% analyzed sex
as a discovery variable in 2019. The percentage of single sex papers remains unchanged across
the ten years (3% for female-only, 27% for male-only). Neuroscience had fewer papers (20%) that
analyzed by sex compared to Psychiatry (61%). Overall, in 2019, only 5% (up from 2% in 2009) of
studies used an optimal design and analysis for discovery of possible sex differences. Therefore,
little progress has been made across the last ten years in harnessing the power that sex differences
can afford in research for discovery and therapeutic potential for neurological and psychiatric
disease to improve the health of men, women and gender diverse individuals.
.CC-BY-NC-ND 4.0 International licensemade available under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
The copyright holder for this preprintthis version posted July 1, 2021. ; https://doi.org/10.1101/2021.06.30.450396doi: bioRxiv preprint
4
Introduction
It has become clear that the consideration of sex in published reports is essential to our
understanding of disease and the biological mechanisms that contribute to the etiology,
manifestation and treatment of disease (Mauvais-Jarvis et al., 2020). The study and disaggregation
of sex differences are critical to our understanding of precision medicine in finding effective
treatments for disease. Sex differences have been observed in the prevalence and manifestation of
a number of neurological and psychiatric diseases (Irvine et al., 2012, Eid et al., 2019). Females are
more likely to be diagnosed with multiple sclerosis, major depressive disorder, and have a greater
lifetime risk of Alzheimer’s Disease compared to males, whereas males are more likely to be
diagnosed with autism spectrum disorder, attention and hyperactivity disorder, and Parkinson’s
Disease (Gillies et al. 2014; Mauvais-Jarvis et al., 2020; Gutiérrez-Lobos et al., 2002; Liu et al.,
2019; Anstey et al., 2021; Nebel et al., 2018). Even in diseases that do not show strong sex
differences in prevalence, age of disease onset or manifestation can be different between the sexes
(Häfner et al., 1992; Liu and Mager, 2016). Sex differences in incidence can shift with age as
schizophrenia is more likely to emerge in males earlier in life than in females, but later in life in
females compared to males (Häfner et al., 1992) and the incidence of stroke increases dramatically
postmenopause in females (Reeves et al., 2008). Perhaps more concerning, there are notable
differences in time to diagnosis (Westergaard et al., 2019), disease progression (Irvine et al., 2012;
Golden and Voskuhl, 2017), vaccine response (Fischinger et al., 2019) and treatment efficacy/drug
response (Zucker and Predergast, 2020; Soldin & Mattison, 2009; Sohrabji et al., 2017). Harnessing
the knowledge that males and females can differ on a number of disease-related outcomes will be
fruitful in not only understanding disease but also in determining whether sex-specific risk factors for
disease may warrant further attention. For example, the manifestation of cardiovascular disease can
be different between the sexes (Liu and Mager, 2016), prompting calls for changes to the diagnostic
guidelines for cardiovascular disease (Trutter et al., 2020). Thus, to make headway for precision
medicine and most effective treatment and diagnoses, sex must be taken into consideration.
Many of the health disparities in treatment and diagnosis have been attributed to the lack of
research in females and inclusion of women in clinical trials (Feldman et al., 2019; Lee, 2018;
Yakerson, 2019). In an effort to increase the enrolment of women in clinical research, the United
States Congress passed The Revitalization Act of 1993. This Act, stated that women and minorities
must be included as subjects in clinical research funded by the National Institutes of Health (NIH).
However, implementation of the requirement of women and minorities has not translated into
analysis by sex or race/ethnicity (Geller et al., 2018). The importance of sex consideration in
research led the NIH to further mandate the inclusion of sex as a biological variable (SABV) in
biomedical research in 2016 (Clayton and Collins, 2014). However, this mandate, much like the one
for clinical trials in 1993, did not include specifications as to the analysis of the data by sex (Mazure
and Jones, 2015) nor did it specify sample size requirements (Miller et al.,2017; McCullough et al.,
2014; Tannenbaum et al., 2016). Other countries have followed suit with some notable differences
as the Canadian Institutes of Health Research (CIHR) mandated Sex and Gender-Based Analysis
(SGBA) policy in 2019, and Horizon Europe has indicated the need for inclusive intersectionality
analyses of gender and sex as of December 2020. Despite the act of prescriptive guidelines from
NIH there are a number of reviews with suggestions on the appropriate incorporation of SABV and
SGBA in the literature (Miller et al., 2017; McCarthy, 2015; McCarthy et al., 2012, Shansky and
Woolley, 2016). However, despite the mandates and recommendations there have been
implementation issues of the mandate as reviewers and authors of papers may be applying SABV
and SGBA inconsistently perhaps given the lack of official guidelines (Galea et al., 2020; Woitowich
and Woodruff, 2019).
.CC-BY-NC-ND 4.0 International licensemade available under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
The copyright holder for this preprintthis version posted July 1, 2021. ; https://doi.org/10.1101/2021.06.30.450396doi: bioRxiv preprint
5
It is important to note that the biomedical and clinical research community is beginning to make
corrections for a long standing bias of using males predominately in research. With the publication
in 2011 by Beery and Zucker it became clear that, although there was considerable variation by
research field, the majority of studies were not using both sexes (Beery and Zucker, 2011). Studies
in human populations were more likely to use both males and females across the ten disciplines
examined compared to studies using animals (Beery and Zucker, 2011). Recently, a ten year follow
up was done comparing the these data from 2009 across the ten various biological disciplines with
publications from 2019, demonstrating there has been an increase in the inclusion of both sexes in
research from 20% in 2009 to 49% in 2019 in articles, with Neuroscience having one of the largest
increases in sex inclusion (Woitowich et al. 2020). Even though a greater proportion of
neuroscience studies are including both sexes in recent years, there are issues in how these sexes
have been included, as approximately one third of sex-inclusive studies surveyed by Woitowich and
colleagues in 2020 did not specify sex in the sample size descriptions. Perhaps less known is that
the majority of studies that used both males and females failed to analyze the data by sex with
Neuroscience at approximately 80% in 2009 (Beery and Zucker, 2011) and 44% of Neuroscience
papers in 2017 (Mamlouk et al., 2020) failing to analyze by sex. Furthermore, it was disappointing to
see that there has been lack of an increase in analyses by sex as only one discipline
(Pharmacology) improved in analyses of sex in their papers over the ten years, as overall there was
in fact a decrease by 8% in the papers that indicated they used sex in their analyses (Woitowich et
al., 2020). Furthermore, sex bias favouring males is still prevalent in neuroscience research (Will et
al., 2017; Woitowich et al., 2020). In fact a study from Will and colleagues (2017) indicated that the
use of solely males in studies increased from 2010 to 2014, while the number of female studies
remained at a constant low value, approximately 5% of Neuroscience papers. Thus, across the 10
years, studies indicate that although the sex omission rate is decreasing across disciplines, the use
of sex in the analyses and the large differential in single-sex studies favoring males have not
appreciably changed over the years (Woitowich et al., 2020).
What has been lacking in the literature is a detailed assessment of not only how sex is reported in
papers (sample sizes disclosed, whether the study design is balanced, sex used consistently
throughout the studies within the papers) but also how males and females are included in any
analyses. Often in clinical studies, sex is used as a covariate which removes the statistical linear
association between two variables and does not inform on the effect of sex. Therefore, in the
present study, we examined not only whether a statistical analysis was done in the studies but what
type of analysis was done to determine whether sex was controlled for, via a covariate analyses, or
explicitly examined as a discovery variable. We were also interested in how many papers used an
experimental design that was optimal for discovery of potential sex differences (balanced design,
sex used through throughout the experiments).
Given that the original studies (Beery and Zucker, 2011; Woitowich et al., 2020) were a sampling of
the first 10-20 papers in the year, we wanted to do an exhaustive analyses of two disciplines by
looking in depth at the various ways that sex was used in analyses. Given the prominent sex
differences in neurological and psychiatry disorders, we chose to do a detailed examination of
journals that targeted Neuroscience and Psychiatry. As the mandates for inclusion of males and
females in biomedical research were in place in 2016, we examined two years over the ten-year
period of 2009 to 2019 as was done by Woitowich and colleagues and as these were dates before
and after the mandates. Therefore, we assessed sex bias/omission, design of experiment (balanced
ratios, consistent use of sex) and determined how papers were analyzing by sex. We hypothesized
that there would be an increase in the number of papers that included both sexes from 2009 to 2019
in Neuroscience and Psychiatry papers, but also that there would be an increase in experimental
design that was not optimized to examine sex as a biological variable. We also expected that the
.CC-BY-NC-ND 4.0 International licensemade available under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
The copyright holder for this preprintthis version posted July 1, 2021. ; https://doi.org/10.1101/2021.06.30.450396doi: bioRxiv preprint
