In this study, we show that although the absolute number of COVID-19 cases were higher for females, females have a lower COVID-19 incidence rate when high-risk populations, namely health care workers and long-term care residences, were disaggregated. These two groups, which are predominantly females, were at a much higher risk of contracting SARS-CoV-2 likely due to higher exposure rates in hospitals and outbreaks in long-term care facilities. However, we also found that women at the age of 80 or older had a higher chance of developing a diagnosable case of COVID-19. Our finding that overall, females have a lower incidence of COVID-19 is consistent with early reports from several countries (2, 4–9) and likely explains why later data show no clear difference in COVID-19 incidence between sexes. Since many countries/regions applied state or regional lockdowns, many new cases are essential workers, such as health care professionals, who are predominantly females. It is very likely that if men and women had a similar chance of SARS-CoV-2 exposure, there would be fewer cases for female patients younger than 80 years old.
After puberty and before menopause, women have much higher circulating estrogens, especially estradiol, than males and prepubertal and postmenopausal females. In this study, we found that females had a lower COVID-19 incidence rate than males unless they were 80 or older. Interestingly, when comparing the age groups of 20–49 and 60 or older, we found only females of the reproductive age group had a lower incidence rate than males. This result suggests that estrogens may exert protective effects in reducing COVID-19 infection. However, females from 60 s to 70 s still have lower incidence rate than males. Since there are no major differences in estrogen levels between sexes after menopause, it is highly possible that additional factors are also important in determining susceptibility to COVID-19.
One of the possible mechanisms by which estrogens reduce COVID-19 incidence is to inhibit SARS-CoV-2 entry into host cells. Angiotensin-converting enzyme II (ACE2) and transmembrane protease serine 2 (TMPRSS2) are located on the cell membrane and mediate the entry of SARS-CoV-2 into cells (20, 21). Using normal human bronchial epithelial cells from a single female donor, it was reported that treatment with estradiol significantly reduced the mRNA levels of ACE2, but not TMPRSS2 (22), suggesting that estrogens may inhibit the entry of SARS-CoV2 into cells. However, it has also been reported that ACE2 may have protective effects against SARS-induced lung damage (23). A recent analysis has shown higher ACE2 levels in Asian women and younger age groups but lower levels in men and older age groups (24). In an unpublished study, Wang et al (2020) used computational approaches to analyze RNA-seq datasets that involved drug treatments and identified estrogenic and androgenic compounds as transcriptional modulators of TMPRSS2. Specifically, they found that estradiol and a phytoestrogen, genistein, reduced TMPRSS2 mRNA levels. In contrast, testosterone and synthetic androgens induced the expression of TMPRSS2 (25). However, a recent study found the expression of TMPRSS2 mRNA in lung tissues was not significantly different between males and females (26). Thus, whether or not estrogens regulate ACE2 and TMPRSS2 to reduce SARS-CoV-2 infection requires further investigation.
A recent study comprehensively compared differences in immune responses between male and female COVID-19 patients and found that males had higher circulating innate inflammatory cytokines, interleukin (IL)-8 and IL-18, and stronger induction of non-classical monocytes while females had a more robust T cell activation and higher interferon (INF)α2 (27). These differences in immune response may explain the different COVID-19 incidence and clinical outcomes between sexes. Although the mechanisms underlying the sex-differences in immune responses are not well understood, sex hormone receptors, notably the estrogen and androgen receptors (ER and AR respectively), are widely expressed in immune cells and play a role in regulating immunity (28). Estrogens have been reported to have immunostimulatory or immunosuppressive effects depending on concentration and cell types, while testosterone is immunosuppressive (29). Females may elicit stronger innate immune responses to viral infections due to greater production of type I interferons (IFNs), which induce anti-viral effects and type 1 immunity (29). Type I IFNs are released by plasmacytoid dendritic cells through stimulation of toll-like receptors (TLR), specifically in coronaviruses by recognition of single-stranded RNA by TLR7 (30). IFN-α release was significantly higher in female than male lymphocytes upon TLR7 stimulation (31). Estrogens may contribute to this phenomenon, as estradiol promotes TLR-activated IFNα release through ERα (30).
Similar to many previous studies (14, 32, 33), we also found that female patients have lower hospitalization, ICU admission, and case fatality rates than males. However, to our surprise, the sex-based differences were greater in the older age groups than in the reproductive age group. Although this could be due to the low rate of hospitalization and death in younger patients, the fact that this sexual difference is greater in postmenopausal women suggests that other factors, rather than estrogens, are more important in reducing the severity of COVID-19. Several genes on the X-chromosome are known to regulate immune response and it has been suggested that females have a lower viral load and inflammation (34). In addition, sex differences in chronic diseases are also commonly observed. Some of these diseases, such as diabetes and cardiovascular disease, are risk factors for poor COVID-19 outcomes (3, 33, 35, 36). According to the Public Health Agency of Canada (PHAC), the prevalence of diabetes and heart diseases are higher in males than in females, especially in the age groups of 45 and older (37, 38). These conditions may explain the higher ICU and hospitalization rates in males, particularly in the older age populations.
One of the interesting findings from this study is the differences in symptom presentation between males and females. Analyses of sex disaggregated data of symptoms show a higher rate of fever in males, while several other symptoms, such as sore throat, runny nose, nausea, diarrhea, shortness of breath, and headache are more frequently reported by females. A higher rate of nausea and diarrhea may indicate greater gastrointestinal involvement. The intestines may be an active site of viral activity as enterocytes do express ACE2, and stool samples of COVID-19 patients have tested positive for SARS-CoV2 (39). Meanwhile, other symptoms more likely in females, such as sore throat, runny nose, and headache, are associated with upper respiratory infection and myositis. The difference in immune function may explain these results; a greater antiviral response in various tissues manifests more symptoms (40–42). Notably, the sex differences in symptom manifestation are only observed in the age group of 20–49, but not 60 or older, suggesting a possible involvement of estrogens. Interestingly, all these female-leaning symptoms, except shortness of breath, are more frequently observed in patients who recovered from COVID-19. Therefore, it is possible that estrogens may also play a role in reducing COVID-19 fatality.
Our observation that there is a correlation between age and hospitalization, ICU admission, recovery time, and fatality is in agreement with previous studies in other countries. Many studies have demonstrated that age is a risk factor for poor COVID-19 outcomes. Analysis of COVID-19 cases in China, the U.K., and Italy consistently demonstrate that increasing age is associated with poorer outcomes (3, 43–45). For patients younger than 20 years old, we observed a very low number of cases, hospitalization, and ICU admission, and no death. These young patients also had faster recovery and fewer symptoms than older age groups. These findings are also consistent with the current consensus that children and teens are less likely to be infected by SARS-CoV-2 and are more likely to have a mild case if they get infected. For example, a study using data from China, Italy, Japan, Singapore, Canada, and South Korea, found higher symptom rates in older age groups, estimating the rate of clinical presentation of SARS-CoV-2 to range from 20% in children and teenagers to 70% in seniors (46).