Skip to main content
SearchLogin or Signup

Review 2: "Estrogen and COVID-19 symptoms: associations in women from the COVID Symptom Study"

This is a reliable study that shows the protective role of estrogens against COVID-19 severe complications among 1.6 million UK women. Novel findings show potential increased risk amongst postmenopausal women, and a potentially protective role of COCP in pre-menopasual women.

Published onAug 21, 2020
Review 2: "Estrogen and COVID-19 symptoms: associations in women from the COVID Symptom Study"
1 of 2
key-enterThis Pub is a Review of
Estrogen and COVID-19 symptoms: associations in women from the COVID Symptom Study

Background: Men and older women have been shown to be at higher risk of adverse COVID-19 outcomes. Animal model studies of SARS-CoV and MERS suggest that the age and sex difference in COVID-19 symptom severity may be due to a protective effect of the female sex hormone estrogen. Females have shown an ability to mount a stronger immune response to a variety of viral infections because of more robust humoral and cellular immune responses. Objectives: We sought to determine whether COVID-19 positivity increases in women entering menopause. We also aimed to identify whether premenopausal women taking exogenous hormones in the form of the combined oral contraceptive pill (COCP) and post-menopausal women taking hormone replacement therapy (HRT) have lower predicted rates of COVID-19, using our published symptom-based model. Design: The COVID Symptom Study developed by Kings College London and Zoe Global Limited was launched in the UK on 24th March 2020. It captured self-reported information related to COVID-19 symptoms. Data used for this study included records collected between 7th May - 15th June 2020. Main outcome measures: We investigated links between COVID-19 rates and 1) menopausal status, 2) COCP use and 3) HRT use, using symptom-based predicted COVID-19, tested COVID-19, and disease severity based on requirement for hospital attendance or respiratory support. Participants: Female users of the COVID Symptom Tracker Application in the UK, including 152,637 women for menopause status, 295,689 for COCP use, and 151,193 for HRT use. Analyses were adjusted for age, smoking and BMI. Results: Post-menopausal women aged 40-60 years had a higher rate of predicted COVID (P=0.003) and a corresponding range of symptoms, with consistent, but not significant trends observed for tested COVID-19 and disease severity. Women aged 18-45 years taking COCP had a significantly lower predicted COVID-19 (P=8.03E-05), with a reduction in hospital attendance (P=0.023). Post-menopausal women using HRT or hormonal therapies did not exhibit consistent associations, including increased rates of predicted COVID-19 (P=2.22E-05) for HRT users alone. Conclusions: Our findings support a protective effect of estrogen on COVID-19, based on positive association between predicted COVID-19 and menopausal status, and a negative association with COCP use. HRT use was positively associated with COVID-19 symptoms; however, the results should be considered with caution due to lack of data on HRT type, route of administration, duration of treatment, and potential comorbidities. Trial registration: The App Ethics has been approved by KCL ethics Committee REMAS ID 18210, review reference LRS-19/20-18210

RR:C19 Evidence Scale rating by reviewer:

  • Reliable. The main study claims are generally justified by its methods and data. The results and conclusions are likely to be similar to the hypothetical ideal study. There are some minor caveats or limitations, but they would/do not change the major claims of the study. The study provides sufficient strength of evidence on its own that its main claims should be considered actionable, with some room for future revision.



Study Summary and Strengths: This study showcases a large, rich data set of women’s symptom reports, self-reported COVID-19 test results, symptom-based COVID-19 predicted status, age, attributes of sex hormone status, and a subset of detailed epigenetic profiles. The authors choose to focus on the cross-section of age and sex hormone status to attempt to understand underlying factors contributing to sex-differences in COVID-19 infection rate, symptoms, and severity. This choice of focus is very important, as women – particularly women during and after the perimenopausal transition, are broadly understudied, and the effects of HRT are even more understudied. To my knowledge, this is the first paper that investigates the interaction of sex, age, and endogenous vs. endogenous sex hormones in relation to COVID-19. The study finds that post-menopausal individuals have a higher rate of predicted (but not of test-confirmed) COVID-19, and that women using COCP have lower predicted COVID-19 rate.  An additional strength is that the study does not overinterpret its results and acknowledges the heterogeneity of HRT and need for its additional investigation.

Room for Improvement: The presentation of the results in table format without the use of any visual/graphic display of data is a large weakness that could limit the reader’s ability to interpret the findings.

Additionally, the piece is framed by calling out the sex-discrepancy between Covid-19 complications between men and women. However, the analyses presented are entirely within women, begging the question: “Are all the groups shown here still at less risk than men?”. Overall, as each comparison is made with respect to a different “control” it is difficult to assess the overall relative state of each group. Some suggestions and questions are detailed under major comment 2.

Finally,  the authors state that allowing users to input detail on their HRT would be too onerous for an in-app interaction, I would strongly suggest that if the study is to continue through the 2nd/3rd waves that participants be given the opportunity to share HRT type. This study is generating data that will prove useful beyond the bounds of COVID-19, and such information could prove important in the long term better understanding of menopause that the authors argue deserves attention.

Recommendation: Request revision of tables and inclusion of graphic display of data (see ideas described below). As COVID-19 related studies are emerging quickly, and combed daily by interested parties, ease of information-intake is of great importance. Scatters would more rapidly illustrate the findings to a wider audience. For hopeful ease of integration, the graphic additions I recommend are hopefully not onerous to create given the existing organization of the data. Revise conclusions to mention “estrogen-associated hormonal status” instead of making a conclusion about estrogen directly. Regarding choice of statistical model: I would recommend relying on other reviewers’ discretion on appropriateness of the association analyses; seems logical but not my area of expertise.

Specific Comments:


1.     Demographic results are not reported that I can see. If this data is not available this should be stated.

2.     A 3D scatter of age, hormone status and predicted COVID would much more strongly support the results listed here.

3.     What is the control group for each comparison? It appears that each comparison has a different control group which makes precludes comparing across groups.

-        Menopause vs. Premenopause: Why is the comparison group of “premenopausal women” 45 years old? Most 45 year old women are experiencing perimenopausal changes, including increased frequency of anovulation, lower estrogen and progesterone, shorter cycles and elevated LH. Why not compare to young regularly cycling premenopausal women (age <35)? If aiming for as close to an age match as possible, it would be more accurate to indicate that these individuals are likely perimenopausal.

-        COCP vs. Regular Cycling: It seems odd to include postmenopausal women in the COCP group, as these drugs interact differently with menopausal physiology than they do with younger physiology. Was a combination of postmenopausal + COCP and HRT evaluated? Or has the COCP group been evaluated on premenopausal women separately?

-        General: Is a male control group possible in order to illustrate relative “advantages” by female age and hormonal status?

4.     The authors to state that they investigated if higher levels of estrogen are linked to a reduction in rate and severity of covid-19 among women. However, because measurements of estrogen are not included (only guesses/proxies), this seems inappropriate. In general, the authors are probably correct, but I would suggest “some physiological states associated with higher estrogen levels, e.g., premenopause and  COCP use, are linked to a reduction in rate and severity of  predicted covid-19”.

-        In both tables, could results be organized by OR  rather than alphabetization, and could predicted and tested covid19 be moved to the top, as they are what readers will care most about?


5.     Because estrogen is not measured in this study (inferred from HRT and COCP for those groups only), it may be inappropriate to include it in the title. Something like “Female Age and Reproductive Status… “ might be more appropriate.

6.     On page 4 the questions relating to categorization of women (menstruating regularly, menopausal, menopausal + HRT, COCP, menopausal +COCP) were briefly described. Pregnancy was not discussed. Were any metrics included by which women who were pregnant or became pregnant between Spring and Summer 2020 employed? This would be a minority of women, but if not assessed, a statement to this effect would be useful.

7.     A brief summary of the predicted COVID-19 algorithm would be appropriate, and a defense of the choice to classify positive predicted COVID-19 as >50% (which seems like it would generate a very high false positive rate).

8.     I would suggest not calling estrogen a female sex hormone, as it plays an important role in both sexes (which the authors acknowledge), I would suggest merely “sex hormone”.

9.     Because tested COVID-19 has an OR of 1, does this imply that menopausal women have “lost” the COVID-19 advantage of women generally? This seems like it would be worth calling out specifically


No comments here