Abstract
Lower COVID-19 mortality rates in women than in men
At the time of writing in the first week of May the COVID-19 infection curves were beginning to flatten and decline. And with that, as the registries of infection accumulated and trends became evident, patterns of susceptibility were beginning emerge. Geographically, people in inner cities and deprived urban areas have suffered higher mortality rates that those in rural locations according to the UK’s Office for National Statistics. Patterns of COVID-19 deaths were in April reported by the ONS to be 86 per 100,000 in London, against national average of 36 per 1000,000. The same data also showed a far higher mortality rate among black Afro-Caribbeans than that of white people. Of course, the explanations will be complex, but a simple reading of the data suggested that poverty and population density were contributing factors to mortality.
As early as March a report in The Lancet suggested another trend in susceptibility: that more men than women were dying, ‘potentially due to sex-based immunological or gendered differences’, even though initial infection rates appeared comparable between the two sexes. 1 Almost two months later an analysis of 5700 patients hospitalised with COVID-19 in the New York area found that only one-third of them requiring intensive care were female. 2 The study also raised the impact of other risk factors – notably co-morbidities (hypertension, obesity, and diabetes) and age – but results clearly showed that mortality rates were higher for male than female patients at every 10-year age interval above 20 years.
Similarly, in late April a preliminary study following a case series of patients hospitalised in Beijing found that the male cases ‘tended to be more serious than women’s’, with a male mortality rate 2.4 times higher than that of female (70% mortality rate in men, 30% in women). 3 Again, the investigators found that, while men and women had the same susceptibility to infection, men were more prone to dying. These figures are supported by the WHO, which in ‘Epi week 17’ (20–26 April), reported that 61% of all COVID-19 deaths in Europe were in men.
The distinction was so marked in assessing the burden of the disease that scientists were quickly searching first for a biological explanation and next for a gender-based treatment. Among the proposed explanations were that the protein by which coronaviruses are sensed is encoded on the X chromosome, and is thus expressed at twice the dose in females than in males, thereby amplifying the immune response to coronavirus in females. Or it may simply be that men have higher rates of comorbidity than women associated with their lifestyles.
However, based on a hypothesis that this female immune response may also be in part regulated by sex hormones, two groups in the USA are now separately testing estrogen and progesterone treatments for men in two short COVID-19 clinical trials. 4 Explaining the progesterone trial, one investigator told the New York Times that hospitalised men ‘are clearly doing worse than women’ and that pregnant women, who usually have high levels of estrogen and progesterone, tend to have mild courses of the disease’. ‘So something about being a woman is protective,’ she said ‘and that makes us think about hormones.’ Certainly, the rationale seems speculative at best, especially as many of the female patients who are outliving their male counterparts are postmenopausal anyway, and deficient in estrogen and progesterone. Nevertheless, with such a high mortality rate for hospitalised men, even a small and not clearly understood benefit would have a major impact.
References
Adverse pregnancy outcomes associated with menopausal symptoms and cardiovascular risk
A study of almost 2000 mid-life women suggests that female reproductive history – specifically adverse pregnancy outcomes – in earlier life is associated with a greater likelihood of severe vasomotor symptoms and a higher cardiovascular risk profile in later life than in those without such reproductive exposures.1
As background to the study the authors – from the CARDIA (Coronary Artery Risk Development in Young Adults) study in the USA – explain that adverse pregnancy outcomes, including preterm birth and small‐for‐gestational‐age, are known to be associated with excess cardiovascular disease risk later in life. For example, a report from the SWAN study in 2017 suggested that a history of preterm birth is associated with higher blood pressure in later midlife. 2 At the other end of the reproductive lifespan, vasomotor symptoms have also been associated with higher levels of cardiovascular risk factors such as raised blood pressure. Given these ‘potentially shared etiologies’, the authors suggest, there may be a continuing link between adverse pregnancy outcomes, vasomotor symptoms and cardiovascular risk. This study, therefore, aimed to test that possibility – whether women with a history of preterm birth and small-for-gestational age deliveries did report a greater degree of vasomotor symptoms than those with normal deliveries, and whether women with a history of both adverse pregnancy outcomes and vasomotor symptoms had a higher risk of cardiovascular disease.
The study was a follow-up of women initially recruited in 1985-86 from four metropolitan areas of the USA. The 1866 women eligible for this study had each reported menopause or hysterectomy, vasomotor symptoms since the age of 40, absent cycles, or use of medication (prescription or over‐the‐counter) to manage their menopausal symptoms. They were then assigned to one of four groups: no reproductive exposures (ie, no adverse pregnancy outcomes and no vasomotor symptoms); either adverse pregnancy outcomes or vasomotor symptoms only; or both.
Results showed that, while there was no difference in the proportion of women who experienced vasomotor symptoms between those with or without a history of adverse pregnancy outcome (78% versus 76%), women with such history were more likely to have severe symptoms (36% versus 30%, P<0.02). Moreover, assessing cardiovascular risk using a 10-year risk calculator (unadjusted and then adjusted for BMI, education, hysterectomy, and use of hormone therapy) found women with a history of both adverse pregnancy outcomes and vasomotor exposures had higher risk scores than those with neither exposure or only the vasomotor exposure.
In view of their results the investigators describe the combined effect of pregnancy outcomes and menopausal symptoms as ‘important’, and that ‘obtaining a life‐course reproductive history could help clinicians assess risk in women’.
A similar conclusion was drawn in a recent US study based on data from the UK Biobank. In this case investigators from Harvard Medical School found that a natural and surgical premature menopause (before age 40 years) were also associated with a small but statistically significant increased risk for a composite of cardiovascular diseases in later life. 3 Again, an event from an individual’s reproductive history – in this case derived from a cohort of more than 500,000 – associated with postmenopausal cardiovascular risk.
References
Time for a ‘rethink’ on studies of HRT and breast cancer risk
The meta-analysis of hormone therapy (HRT) and breast cancer risk from the Collaborative Group on Hormonal Factors in Breast Cancer published last year in The Lancet prompted a ‘new information’ alert from the UK’s MHRA (Medicine’s and Healthcare products Regulatory agency). 1 The study, pooling patient data on more than 108,000 women, was said by the MHRA to confirm that women who use HRT for longer than one year have a higher risk of breast cancer than never users but also – as ‘new’ – that some of this increased risk remains for more than 10 years. The MHRA thus reaffirmed its advice that ‘HRT should only be initiated for relief of menopausal symptoms that adversely affect quality of life’.
The study, quantifying and dramatising risk as its forerunner the Million Women Study had done, found that five years of HRT would result in an increased incidence of breast cancer of one in every 50 women taking continuous combined treatment, one in every 70 women taking sequential, and one in every 200 women taking estrogen alone. And this incidence was apparently doubled if they took it for 10 years.
That widely reported study is now the subject of a new review which, after raising several methodological objections, actually concludes that it ‘really does not add any clarity to our understanding’ of the relationship between HRT and breast cancer risk. 2 Thus, the review authors, who have form in examining such meta-analyses, ask: ‘We are now left in confusion regarding HRT and breast cancer risk. Does HRT cause an increase, do all forms of HRT have this effect, and does the magnitude of any risk have true clinical implications?’
Their objections lie in four methodological domains: the study population; previous randomised trials; type of HRT; and biological plausibility. One requirement of the first is that the population should be ‘appropriate and representative’, yet some large studies are excluded and the biggest contributor by far is the Million Women Study itself, ‘carrying its flaws into the current study’. There is also a discrepancy with results from the Women’s Health Initiative, particularly with estrogen alone, which found a significant decrease in breast cancer risk after seven years compared to placebo, but in this latest meta-analysis the risk remains (RR 1.34). Among the biologically implausible findings is the study’s conclusion that ‘HRT of any formulation, in any dose, prescribed by any route will result in an increased incidence (cause) of breast cancer from between a few months and several decades of exposure’. Stevenson and Farmer urge us – as before – to ‘think again’ and assume caution in interpreting these studies. ‘Overall,’ they write, ‘the benefits of HRT far outweigh any risks, patients should be reassured, and its use should not be restricted.’