The COVID-19 pandemic, caused by SARS-CoV-2, has shown anecdotal evidence of sex disparities in morbidity and mortality. While some regional analyses hinted at a male bias towards severe disease, a comprehensive global analysis was lacking. This study aimed to address this gap by conducting a large-scale meta-analysis to determine if the observed sex bias is a global phenomenon and to understand whether this bias is due to higher infection rates in males or reflects a true difference in disease severity. This is crucial because sex remains an underappreciated factor in infectious disease outcomes. The research question focuses on whether sex is a significant risk factor for severe COVID-19, measured by ITU admission and mortality, independent of infection rates. The study's purpose is to provide a statistically robust, globally representative assessment of the impact of sex on COVID-19 outcomes. The importance lies in informing public health responses and clinical management strategies, potentially leading to improved patient care and resource allocation. Understanding the biological mechanisms underlying sex differences in COVID-19 severity could lead to the development of targeted interventions.

Previous coronavirus outbreaks, such as the SARS-CoV-1 and MERS outbreaks, also demonstrated a male bias towards severe disease and increased mortality. These observations, along with the emerging evidence in COVID-19, highlight the need for a more thorough understanding of the role of sex in influencing the immune response to these viruses and the resulting disease severity. The literature review likely encompassed studies reporting sex-disaggregated data on COVID-19 infections, ITU admissions, and mortality, as well as studies investigating sex differences in immune responses to other viral infections.

The study collected data from 90 reports encompassing 46 countries and 44 US states, totaling 3,111,714 COVID-19 cases. Data included total infections, deaths, and intensive therapy unit (ITU) admissions disaggregated by sex. The data collection period was from January 1st, 2020, to June 1st, 2020. Reports were included if they contained sex-specific data on case numbers, ITU admissions, or mortality and met certain sample size requirements. Data were extracted and entered into a structured online tool. For sources reporting percentages instead of counts, counts were calculated by rounding the nearest integer. Reports with a small proportion of cases with unknown sex were included, minimizing potential bias. Meta-analyses were performed to estimate the overall proportion of male cases and the odds ratios (ORs) for ITU admission and death associated with male sex. A random-effects model was used, weighting studies by their size and precision. Sensitivity analyses were conducted to assess the robustness of the results, including the use of generalized linear mixed models. Forest plots were used to visualize the results. The data analysis involved using statistical methods to combine data from multiple studies and assess the overall effect of sex on COVID-19 outcomes. Inverse variance methods and Mantel-Haenszel/DerSimonian-Laird methods were used to calculate effect sizes. The significance of the findings was determined using two-sided p-values.

The meta-analysis revealed no significant difference in the proportion of male and female COVID-19 cases (proportion of male cases = 0.5, 95% CI = 0.48–0.51, p = 0.56). However, male sex was significantly associated with increased odds of ITU admission (OR = 2.84; 95% CI = 2.06, 3.92; p = 1.86 × 10−10) and increased odds of death (OR = 1.39; 95% CI = 1.31, 1.47; p = 5.00 × 10−30). Sensitivity analyses suggested that the estimated OR for mortality in men might be an underestimate. The findings indicated a consistent global pattern; the sex bias was observed across various regions. The study's findings are statistically significant, with p-values indicating strong evidence against the null hypothesis (no association between sex and disease severity). The magnitude of the effect is substantial, with males demonstrating substantially higher odds of ITU admission and death than females. The consistency of these findings across numerous studies and diverse geographical locations reinforces their robustness.

The findings confirm a significant global sex bias in COVID-19 outcomes, with males experiencing a higher risk of severe disease and death despite similar infection rates. This suggests that biological factors, rather than differences in infection rates or access to healthcare, are primarily driving this disparity. The observed sex differences likely stem from fundamental differences in the innate and adaptive immune systems between males and females. Females generally exhibit a stronger humoral immune response (higher antibody production) and increased type I interferon production, an important antiviral cytokine. These immune responses may contribute to a more effective early response to the virus and better protection against severe disease. Conversely, testosterone, a male sex hormone, may suppress immune responses. These findings have important implications for developing targeted interventions and strategies for mitigating the impact of COVID-19.

This large-scale meta-analysis provides robust evidence for a global sex bias in COVID-19 severity, with men at significantly higher risk of ITU admission and death. The observed differences likely reflect underlying biological mechanisms related to sex-based differences in immune responses. This highlights the urgent need for sex-specific approaches to clinical management, public health strategies, and vaccine development. Future research should focus on further elucidating the biological mechanisms underlying these sex differences and exploring the potential for sex-specific therapeutic interventions.

The study relies on aggregated data from various sources, which might introduce heterogeneity in data collection methods and reporting practices. Potential biases related to testing and reporting practices might exist. The lack of detailed individual-level data limited the ability to fully control for potential confounding factors, such as comorbidities and lifestyle differences.