Dear editor,
Recently in this Journal, Sergey Moiseev et al. reported that the mortality rate in intensive care unit patients with severe COVID-19 was higher in males older than 50 years old in contrast to females of similar age based on a nationwide cohort study.1 The sex differences between mortality rates cannot be explained by comorbidities, given the increased occurrence of chronic illnesses in females that may worsen survival in COVID-19 patients. The results initially supported that sex steroid hormones underlie sex-related differences in COVID-19 mortality. However, the alternations of sex hormones were accompanied by changes of many risk factors for severity and mortality of COVID-19, such as obesity, diabetes, and other comorbidities. It was unconvincing to judge the causality between sex hormones and COVID-19 outcomes based on epidemiological data due to above-confounding factors. Mendelian randomization (MR) analysis is an analytic method to estimate the causal effect, which overcome the limitations of measurement errors and confounding frequently encountered in observational studies. In this study, we performed a sex-stratified two-sample MR analysis to explore the causal relationship of serum sex hormone levels including bioavailable testosterone (BAT), total testosterone (TT), and sex hormone binding globulin (SHBG) on COVID-19 outcomes.
We obtained summary statistics for sex hormones including BAT, TT, and SHBG from previous genome-wide association analyses study based on genetic data in UK Biobank.2 The datasets of sex hormones could be downloaded from https://www.ebi.ac.uk/gwas/. The GWAS summary statistics for COVID-19 susceptibility (C2: covid vs population), hospitalization (B2: hospitalized covid vs population) and severe disease (A2: critically ill covid vs population) outcomes were obtained from the COVID-19 Host Genetics Initiative (https://www.covid19hg.org/, Release 5, European ancestry cohorts but excluded UKBB). The selection of instrumental variables was followed by previous procedure (1) selected SNPs which effect on sex hormones was significant (P<5 × 10–8); (2) matched these SNPs with the outcome dataset by rsid; (3) obtained independent SNPs through clumping procedure based on linkage disequilibrium with r 2 < 0.001 or the physical distance more than 10 000 kb via PLINK; (4) removed SNPs that were significantly associated with the outcome.3 We calculated the proportion of variance explained for all remaining instruments to evaluate the strength of instrument variables. To investigate the causal effect, we performed the MR analysis via inverse variance weighted (IVW), weighted median, MR-Egger, MR-RAPS, MR-Lasso, and MRPRESSO method within males and females separately. All computations were performed using the Mendelian Randomization package.
In the study, we obtained 208, 141, and 216 instruments separately of serum TT, BAT, and SHBG levels when studied causal effect of sex hormones and COVID-19 susceptibility in females, which could explain 7.42%, 5.44% and 10.95% of the variance of TT, BAT and SHBG respectively. The number of instruments and the PVE for instruments used in analysis of causal effect of sex hormone on COVID-19 hospitalization and severe disease was approximately same as analysis in COVID-19 susceptibility. The results showed that each one standard deviation (SD) increase in serum BAT levels was associated with a higher risk of COVID-19 hospitalization (OR:1.214, 95% CI:1.011–1.458, p = 0.038) and COVID-19 severe disease (OR:1.336, 95% CI:1.024–1.744, p = 0.033) based on IVW method, and the results were verified by MR-Lasso, MRPRESSO method (Fig. 1 ). For a SD increase in serum BAT level, we observed no statistically significant effect upon odds of COVID-19 susceptibility (p = 0.800) (Fig. 1). Similarly, we observed no significant difference in risk of COVID-19 outcomes associated with an SD increase in serum TT or SHBG levels based on all MR analysis methods (Fig. 1). The results suggested a causal effect of increased serum BAT levels on the higher risk of COVID-19 hospitalization and severe disease in females. No statistically significant difference in risk of COVID-19 outcomes associated with a SD increase in serum BAT, TT or SHBG levels was found in males (Fig. 2 ). The results demonstrated no causal relationship between sex hormones and COVID-19 outcomes in males.
Fig. 1.
The Mendelian randomization analysis of casual effect between sex hormones and COVID-19 outcomes in females. IVW: inverse variance weighted; BAT: bioavailable testosterone; TT: total testosterone; SHBG: sex hormone binding globulin.
Fig. 2.
The Mendelian randomization analysis of casual effect between sex hormones and COVID-19 outcomes in males. IVW: inverse variance weighted; BAT: bioavailable testosterone; TT: total testosterone; SHBG: sex hormone binding globulin.
In summary, we performed a sex-stratified two-sample Mendelian randomization analysis and demonstrated a causal relationship of increased BAT levels and higher risk of COVID-19 hospitalization and severe disease in females, not in males. A null causal relationship was observed for TT or SHBG levels with COVID-19 outcomes in females and males. The relationships between gender difference, sex hormone difference and COVID-19 outcomes have been reported based on several observational studies1 , 4 and various hypotheses have been postulated to explain the relationships, including a sex-dependent difference in immune responses, sex-related expression difference of angiotensin-converting enzyme 2 and transmembrane protease serine 2.[5], [6], [7] Our study demonstrated a direct causal effect of BAT levels and COVID-19 hospitalization and severe disease in females based on MR analysis. The MR analysis exploited the natural random allocation of genetic variants and limited the potential confounding and reverse causal effect, thus providing powerful evidence for the causal effect of testosterone on COVID-19 outcomes.
In addition, the BAT referred to testosterone loosely bound to albumin and free form testosterone, which participated in the biological process in vivo and therefore may become more relevant to proposed causal relationship. This is consistent with our results that only BAT levels have a causal effect on COVID-19 outcomes, but not serum TT and SHBG levels. Another innovative point in our present study was that the causal relationship of BAT and COVID-19 outcomes was only observed in females, but not in males. We inferred that the causal relationship between BAT and COVID-19 outcomes may be non-linear, and both high BAT and low BAT levels were risk factor for the poor prognosis of COVID-19. Further studies based on male population excluded patients with testosterone deficiency are needed to verify the relationship. In the study, we were unable to obtain gender-specific GWAS data of COVID-19 outcomes from update datasets, further study should be performed. Then, the findings that BAT increased the risk of COVID-19 hospitalization and severe disease in females helped better understand the role of sex hormones in COVID-19 occurrence and progression, and provided evidence for hormone therapy in the treatment of COVID-19 in females.
Funding
This work was supported in part by grants from the National Natural Science Foundation of China (81770860) and Key Research and Development Plan of Shandong Province (2017CXGC1214).
Declaration of Competing Interest
The authors declare that there is no conflict of interest.
References
- 1.Moiseev S., Brovko M., Tao E., Bulanov N., Akulkina L., Fomin V. Sex differences in mortality in the intensive care unit patients with severe COVID-19. J Infect. 2021;82(2):282–327. doi: 10.1016/j.jinf.2020.09.031. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Ruth K.S., Day F.R., Tyrrell J., et al. Using human genetics to understand the disease impacts of testosterone in men and women. Nat Med. 2020;26(2):252–258. doi: 10.1038/s41591-020-0751-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Wang Y., Guo P., Liu L., Zhang Y., Zeng P., Yuan Z. Mendelian randomization highlights the causal role of normal thyroid function on blood lipid profiles. Endocrinology. 2021;162(5) doi: 10.1210/endocr/bqab037. bqab037. [DOI] [PubMed] [Google Scholar]
- 4.de Lusignan S., Joy M., Oke J., et al. Disparities in the excess risk of mortality in the first wave of COVID-19: cross sectional study of the English sentinel network. J Infect. 2020;81(5):785–792. doi: 10.1016/j.jinf.2020.08.037. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Giefing-Kröll C., Berger P., Lepperdinger G., Grubeck-Loebenstein B. How sex and age affect immune responses, susceptibility to infections, and response to vaccination. Aging Cell. 2015;14(3):309–321. doi: 10.1111/acel.12326. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Asselta R., Paraboschi E.M., Mantovani A., Duga S. ACE2 and TMPRSS2 variants and expression as candidates to sex and country differences in COVID-19 severity in Italy. Aging. 2020;12(11):10087–10098. doi: 10.18632/aging.103415. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Mjaess G., Karam A., Aoun F., Albisinni S., Roumeguère T. COVID-19 and the male susceptibility: the role of ACE2, TMPRSS2 and the androgen receptor. Prog Urol. 2020;30(10):484–487. doi: 10.1016/j.purol.2020.05.007. [DOI] [PMC free article] [PubMed] [Google Scholar]