Is my Study “Sex” or Is it “Gender”?

Investigators often ask whether their research fits better into studies of “sex” or studies of “gender”. In 2001, members of the Institute of Medicine addressed the differences between sex and gender (1), and these definitions have been widely accepted around the world, including by the WHO (2), the Council of Europe (3), and the Canadian Institute of Health Research (4), among others.

“Sex” is defined as any trait/phenotype that can be attributed to the presence or absence of sex steroids, to sexual or reproductive anatomy, or to sex chromosomes and the genes located on those chromosomes. Sex of an individual is typically defined as male or female. “Gender” is defined by the individual as to how they perceive themselves, often as woman or man, female or male, which in turn determines their behavior and how society then perceives them based on their personal perception. Gender is not only binary, and may not be static since it may change over time as the individual changes their perceptions. For purposes of this research definition, only the concept of gender identity is discussed

Human studies in which traits are different in men and women are defined as “gender differences”. In the 1990’s, the National Institutes of Health (NIH) began requiring that both men and women, and ethnic minorities, be included in grant-funded studies. However, in many studies the data were not analyzed separately for men and women, but as a single cohort, often because the studies were not statistically powered to find gender differences. It has only been in the past few years that studies in humans have made gender comparisons, and then only when the study design was present from inception.

An example of gender differences in hypertension is a study recently published in which individuals underwent measurement of central blood pressure (CBP) via radial artery applanation tonometry and other blood chemistries to evaluate cardiovascular disease (CVD) risk factors. This South Korean cohort included 2115 individuals, 50.4% were men, and the average age of women was slightly older than the men (59 years vs 57 years) (5). Of the women, 78.1% were postmenopausal, as determined by survey. Interestingly, the study found that the CVD risk vs CBP or systolic BP was a J-shaped curve for both men and women. At levels below optimal systolic BP (126 mm Hg), the CVD risk was similar for both men and women. However, above that optimal BP, and as BP increased further, the CVD risk increased faster in women than men. Postmenopausal women also had a greater CVD risk for the levels of BP than did premenopausal women, suggesting that the lack of estradiol and the implications there of (reduced nitric oxide, etc) could contribute to the CVD risk differences between pre- and postmenopausal women.

A study in humans can be determined to be “sex differences” if the trait/phenotype is produced by the presence or absence of sex steroids themselves or by a difference in sex chromosomes (XY as opposed to XX). One example is a recent review that discusses the clinical and morphological differences found with imaging in women compared to men as they are being evaluated for heart failure (6). The review points out that women have smaller heart volumes even when factored for body surface area or mass, and women have higher median left ventricular ejection fraction (6). In addition, women have a lower degree of left ventricular hypertrophy than men at all stages of ventricular dysfunction (6). These differences are likely due to genetic or chromosomal differences rather than differences in sex steroids, since these differences don’t change following menopause. Another example of sex differences is in women who have polycystic ovary syndrome that is characterized by elevated levels of androgens (7). When compared with women whose androgen levels are within normal limits, any physiological phenotype in PCOS women may be attributed to the presence of androgens.

Since animals do not choose gender identity, all studies in animals are considered to be studies of “sex”, and not “gender”, and animals are classified as either male or female sex based on sex steroids and anatomy. In 2014, the NIH began requiring that grants submitted had to include both sexes of animals unless the condition occurred in only a single sex, such as in models of preeclampsia in which only females were studied or in models of prostate cancer in which only males were studied (8). Therefore, the number of studies that have shown sex differences in a trait/phenotype in an animal model being studied has increased dramatically over the past several years.

In animal studies differences in mechanisms responsible for blood pressure control can be manipulated by changing the physiological levels of the sex steroids or removing them altogether. For example, in studies we performed many years ago in male and female spontaneously hypertensive rats (SHR), we determined that the higher blood pressure in males compared with females was due to the levels of androgens in males and the renin-angiotensin system (9). In the male SHR, gonadectomy lowered the blood pressure to the same levels as the blood pressure in the hypertensive females (see Figure 1), whereas gonadectomy in the females did not affect their blood pressure. Similarly, when ovariectomized females were given testosterone to the levels found in males, the blood pressure increased to similar levels as in the males. When the rats were given an angiotensin I converting enzyme (ACE) inhibitor, enalapril, the drug lowered the blood pressure to similar levels in all sex groups of SHR (Figure 1) (9). Interestingly, testosterone was no longer able to increase the blood pressure in the ovariectomized female SHR when ACE was blocked. As expected, we found these differences changed with aging in SHR as sex steroid levels changed and other mechanisms became involved and more important.

Figure 1. Contributions of sex steroids and the renin-angiotensin system to the differences in blood pressure in male and female SHR.

MAP is shown for 17- to 19-week-old SHR in the following groups: male, female, castrated male (cast), ovariectomized (ovx) female, and ovx female treated chronically with testosterone (ovx+T) starting at 11 weeks of age. Open bars indicate control rats; solid bars, enalapril-treated rats. ‡P<0.01 compared with intact males; *P<0.01 compared with females (cast or ovx); **P<0.01 compared with untreated rats of a similar gender. (Reproduced by permission of Hypertension).

With regard to the importance of sex chromosomal effects studied in animals, Arnold and colleagues developed the Four Core Chromosome Model in mice in which they moved the SRY gene, that is on the Y chromosome and determines male-ness, onto an autosome (10). In this way there are XY females and XY males, XX males and XX females, allowing an investigator to determine if a physiological trait is mediated by sex steroids or by sex chromosomes. More and more studies are showing that sex chromosomes impact physiological traits in animals.

In summary, research that includes humans should be termed “gender” or “gender differences” unless the reason for a particular trait is mediated by the presence or absence of sex steroids or sex chromosomes and thus is specific to a certain sex. Because animals are not able to gender identify and thus change their social environment, animal studies are considered “sex” or “sex differences”. For researchers who remain uncertain as to whether their studies fall into the category of “sex” or “gender”, the Canadian Institutes of Research have modules on sex and gender training they might find useful (11).