Obesity as a risk factor for COVID‐19 mortality in women and men in the UK biobank: Comparisons with influenza/pneumonia and coronary heart disease

Abstract Obesity is associated with severe COVID‐19 outcomes, yet, it is unclear whether the risk of COVID‐19 mortality associated with obesity is similar between the sexes. We used data from the UK Biobank to assess the risk of COVID‐19 mortality associated with various anthropometric measures in women and men. To put these results in context, we also compared these estimates with those for mortality from influenza/pneumonia and coronary heart disease (CHD). The analyses included 502 493 individuals (54% women), of whom 410 (36% women) died from COVID‐19, 549 (36% women) died from influenza/pneumonia and 3355 (19% women) died from CHD. A higher body mass index (BMI), waist circumference, waist‐to‐hip ratio and waist‐to‐height ratio were each associated with a greater risk of death from COVID‐19, influenza/pneumonia and CHD in both sexes, with the exception of the association between higher BMI and the risk of influenza/pneumonia death in men. A higher BMI was associated with a stronger risk of COVID‐19 mortality in women than men; the women‐to‐men ratio of hazard ratios was 1.20 (95% confidence interval 1.00; 1.43). This study demonstrates the role of obesity in COVID‐19 mortality and shows that the relative effects of a higher BMI on COVID‐19 mortality may be stronger in women than men.


| INTRODUCTION
Although women and men are approximately as likely to be infected by coronavirus disease 2019 (COVID-19), men have higher death rates from COVID-19 in almost all countries where data are available. 1 Obesity has been identified as one of the key factors associated with severe COVID-19 outcomes, 2,3 some of which may be explained by the adverse effects of obesity on diabetes. However, it is unclear whether the risk of COVID-19 mortality associated with obesity is similar between the sexes. We thus assessed the risk of COVID-19 mortality associated with various anthropometric measures in women and men in the UK Biobank. For comparability, we also evaluated the association of these measures with mortality from influenza or pneumonia, the leading causes of death from respiratory disease in usual circumstances, and coronary heart disease (CHD), a condition which has a well-established association with obesity.
T A B L E 1 Adjusted hazard ratios (HRs) and women-to-men ratios of HRs with 95% confidence intervals for death from COVID-19, influenza/pneumonia or coronary heart disease associated with body mass index, waist circumference, waist-to-hip ratio and waist-to-height ratio Abbreviation: BMI, body mass index.
Analyses are adjusted for age, Townsend index, smoking status, and ethnicity. HRs for continuous variables are per 1-SD higher value, taking the overall SD from the sex-combined baseline data. HRs for BMI categories are presented on a floating absolute scale, with the healthy weight group as the reference group.

| METHODS
The UK Biobank is a prospective, population-based cohort study of women and men aged 40 to 69 years at baseline between 2006 and 2010. 4 Follow-up for cause-specific mortality was conducted to June 30, 2020 through linkage with the National Death Register.
Cox regression was used to estimate the sex-specific hazard ratios (HRs) and 95% confidence intervals (CIs) for mortality from COVID-19, influenza/pneumonia and CHD for overweight and obesity (defined as a body mass index [BMI] ≥25 to <30 kg/m 2 and ≥30 kg/m 2 , respectively) and an overall 1-standard deviation (SD) increase in BMI (SD 4.8), waist circumference (SD 13.5), waistto-hip ratio (SD 0.09) and waist-to-height ratio (SD 0.08). Influenza and pneumonia were taken together to avoid unreliable estimation due to small numbers. CIs were estimated using floating absolute risks. 5 Adjustments were made for age, smoking status (never/ex/current), socio-economic status (determined using the Townsend index of area deprivation) and ethnicity (white or not). Interactions between each variable and sex were added to the model, so as to obtain the women-to-men ratio of HRs (RHRs) for each risk factor. 6 Penalized smoothing splines were used to examine the shape of associations between BMI and the study endpoints. In secondary analyses, we additionally adjusted for diabetes. We also stratified our analyses on
F I G U R E 1 Penalized spline plots with adjusted hazard ratios and 95% confidence intervals for death from COVID-19, influenza/pneumonia or coronary heart disease (CHD) associated with body mass index. Analyses are adjusted for age, Townsend index, smoking status and ethnicity A higher BMI, waist circumference, waist-to-hip ratio and waistto-height ratio were each associated with a greater risk of death from COVID-19, influenza/pneumonia and CHD in both sexes ( Table 1).
The only exception was the association between higher BMI and the risk of influenza/pneumonia death in men, which may be attributable to the apparent underlying U-shaped relationship (Figure 1) (Table S1).
Compared with their white counterparts, the association between obesity measures and COVID-19 mortality was stronger among nonwhite women, but broadly similar in non-white men (

| DISCUSSION
This population-based study of over 500 000 women and men in the UK Biobank shows that higher levels of adiposity measures were associated with a higher risk of death from COVID-19, of an order of magnitude similar to that seen for other respiratory diseases and CHD. The findings for COVID-19 are in agreement with previous studies that highlighted the role of obesity in COVID-19 severity and mortality, although these studies did not investigate the sex-specific associations. 2,3 They also add to the findings of previous studies in the UK Biobank that showed that both BMI and waist circumference were associated with an increased chance of testing positive for COVID-19, especially among non-white people. 7 We show that the associations between higher BMI, and obesity compared to healthy weight, with death from COVID-19 and death from influenza or pneumonia, were each approximately 20% greater in women than men.
Our analyses also suggest that the risk of COVID-19 mortality associated with obesity may be particularly strong among women from non-white backgrounds, although the number of non-white individuals was relatively small. In agreement with our previous meta-analysis, 8 there was no sex difference in the effects of adiposity on death from CHD. By and large, results were similar for the other three measures of adiposity considered. We are unable to explain the exception to this general agreement in that we found a higher effect on COVID-19 mortality of waist-to-hip ratio in men than women; this requires further study. Otherwise, given that the association between BMI and COVID-19 mortality was stronger, not weaker, in women than men, the greater risk of death from COVID-19 in men is unlikely to be attributable to a greater relative impact of obesity on death from COVID-19 in men. Although our analyses relied on single measurements of obesity markers at study baseline, we anticipate that the impact of any changes in obesity levels during follow-up and the risk of COVID-19 would be similar for women and men.
These results indicate that, if causal, obesity prevention strategies aimed at reducing the burden of several chronic diseases should also lead to better outcomes among both women and men affected by COVID-19. Furthermore, the striking similarity in most of the observed effects of adiposity on COVID-19 and the other two respiratory diseases considered, both in sex-specific and sexcomparative terms, suggests that obesity is likely to be a key driver of mortality in any future viral epidemic, particularly amongst women.