Abstract
We have produced compelling evidence that women are subject to a higher relative increase in their risk of coronary heart disease and stroke following a diagnosis of Type 2 diabetes. Thus, in terms of vascular risk, diabetes confers a female disadvantage. This excess risk could be due to three main factors. First, it is conceivable that this is merely a mathematical artifact caused by the relatively low background rate for cardiovascular diseases among women, compared with men. Second, it could be due to women receiving poorer care following their diagnosis of diabetes than men; for instance, due to physician bias. Third, certain underlying biological differences in women and men, most likely related to the distribution of body fat, could explain this female disadvantage.
Diabetes is a curse of modern life, with an estimated global prevalence of over 350 million [1]. Across the world it is on the rise, with a projected increase to 438 million by 2030 with most of the burden occurring in lower- and middle-income countries [1]. The strongest risk factor for new-onset diabetes is excess body weight, although increased age, increased blood pressure and family history of diabetes also have strong independent effects [2]. Those with diabetes have an increased risk of death from vascular conditions, certain cancers and respiratory and other infectious diseases [3,4], with cardiovascular disease (CVD) being the most common adverse outcome. Indeed, having diabetes approximately doubles a person's risk of contracting coronary heart disease (CHD) or stroke [3,5].
However, not all affected individuals have the same degree of excess risk. Our past meta-analyses [6–8] have established conclusively that women have a significantly and clinically important, higher excess risk of both CHD and stroke (fatal and combined fatal plus non fatal), consequent to diabetes than men. These meta-analyses have been conducted predominantly in Type 2 diabetes, which accounts for about 90% of all diabetes prevalence, and this report will largely concentrate on Type 2 diabetes. However, we have also reported an excess relative risk for coronary disease, stroke and all-cause mortality among women compared with men in Type 1 diabetes
The phenomenon of a sex difference in the effects of diabetes has long been recognized [10–13], although only recently quantified reliably. Since men have a higher risk of CVD than women, the glib statement, ‘diabetes removes the female advantage in cardiovascular disease’ – which has inspired the title of this report – has entered the medical lexicon. Women have a lower risk of CVD than men, but the difference, in relative terms is less when they, and their male comparator, have diabetes. Nevertheless, the ‘catching up’ process after diabetes is scientifically intriguing and worthy of serious research. If we can understand why the sex difference appears we will be able to suggest new policies that will not only restore the female advantage but may also lead to better consequences for men. This article will discuss potential explanations and how evidence for, or against, them may be obtained.
The artifactual explanation
One possible explanation is that this is a mathematical artifact: women, overall, have a lower risk of CVD than men and hence the same risk difference consequent to diabetes in each sex must translate to a higher relative effect in women. This is a mathematical truth
Moreover, it is inconceivable that there could be a meaningful universal quantification of average risk differences, owing to the considerable variation in background risks across populations. In contrast, relative risks are generally stable across populations, and thus make a far better metric for purposes of universal comparisons of risk. One exception to this stability is that relative risks tend to attenuate with age for all major cardiovascular risk factors.
Partially for all these reasons, in medical science, both in observational epidemiology and controlled clinical trials, the relative risk (or an approximation) is far more commonly reported than a risk difference. So even if there was a good reason to evaluate sex differences on an arithmetic scale it would be practically impossible to do so in a comprehensive way.
The accessibility explanation
Women could do worse when they develop diabetes because they do not have the same access to appropriate, cardio-protective, health care as do men who develop diabetes. There is a great wealth of evidence to show that women have, in the past, received poorer cardiovascular care than men – although campaigns such as ‘Go Red for Women’ [19] have prompted a great improvement, at least in industrialized nations. However, within the sphere of diabetes, much remains to be done to produce real equality. For example, the 2012–2013 report from the UK National Diabetes Audit [20], including about 2 million individuals with diabetes, found that women were 15% less likely than men to receive the care recommended by the national guidelines or to meet treatment targets. Standard care processes, incorporating annual checks for the effectiveness of diabetes treatment, management of cardiovascular risk factors and the emergence of early complications, were received by 58% of women, but 62% of men. Treatment targets for glycated hemoglobin, blood pressure and cholesterol were met by 34% of women, but 37% of men. Similarly, data from a US study of more than 147,000 outpatients with established heart disease, conducted 2010–2012, reported significant sex disparities in treatment and risk factor control to the detriment of women. For example, women were 37% less likely to be prescribed antiplatelet medication compared with men and, in individuals with diabetes, women were 25% less likely to achieve target cholesterol levels than men [21].
It is debatable whether this sex inequality is due to women underestimating, or ignoring, their risk, relative to men, or whether it is due to structural biases in cardiovascular health provision. It seems unlikely that women will, in general, be less aware of the adverse consequences of unhealthy risk behaviors, relative to men [22]. Nevertheless, recent data from a very large study in the USA and its territories [23] found adherence to antidiabetic medication to be slightly lower among women. Possibly women have a different concept of personal risk to men, at least in terms of its importance for them relative to other members of their immediate family. Doctors have undoubtedly been liable to treat CVD as predominantly a ‘man's disease’ in the past, and perhaps some doctors still think this way, even if not consciously so. But any explanation of sex differences, in diabetes, relating to access has to consider why no sex difference is seen in hypertension [15], which is arguably the strongest risk factor for CVD.
The biological explanation
Women and men are of the same species and largely subject to the same environmental exposures but are, intrinsically, biologically different. Hence, it is reasonable to propose that some of the ‘catching-up’ in vascular risk by women after they develop diabetes is due to inherent physiological, hormonal or genetic differences.
There is evidence to suggest that women transition to diabetes in a worse state than men, in regard to other CVD risk factors, especially in relation to excess body weight [24]. In this case, it may be that it is not diabetes itself that is driving the excess relative risk in women with diabetes. For example, a study of around 95,000 Scots [25] found that, when first diagnosed as having diabetes, women had a BMI of almost 2 kg/m2 higher than men. This difference might be explained by differential patterns of preferential adiposity storage in women and men. Women generally have greater subcutaneous fat storage reserves than men, and therefore need to gain more weight before less hazardous subcutaneous storage becomes exhausted and excess adipose tissue is placed into visceral and ectopic tissues, which are linked to insulin resistance and diabetes. Thus, women have greater cardiometabolic reserves than men and have to put on more weight before they reach the point of having developed diabetes. This more efficient fat storage by women, before the menopause, is likely to be nature's response to the needs of child bearing. However, the hypothesis that diabetes is merely a marker of a more deleterious vascular state in women is not supported by epidemiological analyses that adjust diabetes for other classical risk factors and still find a female disadvantage in relative risks for CVD. On the other hand, such analyses generally do not adjust for such factors at the time when diabetes is diagnosed.
Many researchers have shown that adverse changes in other CVD risk factors are greater for women than men, after they transition to diabetes. For example, a cross-sectional study of 7529 British women and men [26] concluded that diabetes had a greater adverse influence on adiposity, insulin resistance and downstream blood pressure, lipids, endothelial dysfunction and systemic inflammation in women compared with men. Whether this is due to some single, driving, biological change is debatable, but obesity would seem to be the most likely candidate because obesity is a mediator of several CVD risk factors. Sex differences in the distribution of fat in the bodies of women and men, naturally caused by the hormones, may also play a role: in women fat tends to accumulate more in the buttocks, hips and thighs and, linked to this, women have less visceral fat than men. Perhaps diabetes alters the prevailing sex imbalances in fat distribution, causing women to move towards the CVD risk profile of men.
Although these potential biological explanations, mostly related to body fat, have been expressed in terms of characterizations at the time of, or subsequent to, the development of diabetes, this is actually a gross oversimplification. The very definition of ‘diabetes’ in clinical practice is arbitrary. Although it is well understood that diabetes is a state where the pancreas produces insufficient insulin or the body cannot respond sufficiently to the insulin that is produced, clinical diagnosis depends on blood glucose (or some other measure of glycemia, such as glycated hemoglobin) being at or beyond a certain threshold. In fact, cardiovascular risk increases monotonically with increasing glucose above a level well below any accepted threshold (hence the invention of the clinical state of ‘prediabetes’, which only shifts the problem downward) [5,27]. So biological explanations must, unlike accessibility explanations, account for longitudinal changes in glycemia and (for instance) adiposity.
This process of ‘developing diabetes’ leads to another possible explanation of the excess vascular risk conferred by diabetes in women compared with men. Possibly this differential is driven by women having a greater cumulative exposure to hyperglycemia in the prediabetic state compared with men. As outlined above, women, on average, have to become more overweight than men to develop overt diabetes. Consequently it is likely to take longer for women to develop diabetes than men and thus they live in a ‘prediabetic’ state for a longer period of time. Men, on average, have prediabetes for 8 years, and women for 10 years, before they progress to overt diabetes [28]. We have hypothesized, that during this time a woman's metabolic profile – that incorporates lipids, blood pressure, inflammatory markers and other vascular risk factors – continues to deteriorate relative to men, and that by the time they are clinically diagnosed with diabetes, considerable vascular damage resulting from living in a state of prolonged – but silent – suboptimal glycemia has already occurred.
Another, more speculative, mediator of the excess vascular risk among women with diabetes may be due to their exhibiting a greater degree of ‘diabetic cardiomyopathy’ compared with men with diabetes. Diabetic cardiomyopathy has been reported to be present among individuals with diabetes (both Type 1 and Type 2) and is characterized by metabolic, structural and functional abnormalities that are unrelated to myocardial ischemia, which can ultimately lead to left ventricular hypertrophy and other cardiac disturbances including heart failure. In a recent study of 100 adults (44% women) [29] with diabetes and no prior history of structural CHD, evidence of echocardiographic diabetic cardiomyopathy was diagnosed in 48% of the group, approximately two-thirds of whom were women. These findings are consistent with prospective data from the Strong Heart Study and the Framingham Heart Study, both of which have reported an independent relationship between diabetes and left-ventricular mass, with some evidence of a greater effect in women than in men with diabetes [30,31]. These findings are intriguing, not least because they may help to explain reports that women with diabetes have twice the excess risk of heart failure compared with similarly affected men, lending further support to the existence of important and clinically meaningful sex differences in diabetes-related risk of vascular disease [32–34]. Importantly, the CALIBER study [35] found higher hazard ratios for heart failure as the first presentation of CVD in women than men, in both younger and older age groups, suggesting that the sex inequality in heart failure is not entirely explained by the sex inequality in CHD and stroke discussed earlier.
Other possible explanations for the excess cardiovascular risk from diabetes would include confounding effects from risk factors not considered above. Female-specific factors, such as complications in pregnancy, age at menopause, contraceptive practices and use of hormone replacement therapy, may have also have a role. To our knowledge, there is no compelling evidence for any of these theories.
What evidence is needed?
The female disadvantage in diabetes may not be inevitable. Even if it is, health policies might be appropriately skewed to equalize the outcomes. The key first step is to provide a reliable evidence base encompassing all possible explanations for the current disadvantage that women face. A good start to this would be to have a general policy that sex-specific results are always presented whenever studies are carried out in diabetes. Often this should include both risks and relative risks. Our perusal of the literature when undertaking our systematic reviews has suggested that this is often not the case in epidemiological research, although anecdotal evidence is that many, but not all, clinical trials have subgroup analyses by sex. Women-specific studies are, clearly, of immense value in general terms for studying women's health in diabetes, but cannot reliably contribute to studying the female disadvantage in diabetes; studies of only men have the same problem. Although between-study comparisons of the sexes are possible, these will suffer from differences in study design, setting and conduct.
In general it should be simple to use existing contemporary data sources to quantify sex-specific effect sizes, comparing outcomes between those with and without diabetes. Similarly accessibility issues can be explored using basic survey data, understanding that this is likely to vary by location and social customs. A greater problem lies in exploring the biological explanation for sex differences in diabetes, which requires detailed longitudinal clinical data, beginning at a time well before diabetes is diagnosed. These are unlikely to be available, with the numbers that are required for reliable estimation of sex differences, from any research databases. Linked routine clinical ‘big data’ may provide good evidence, although glycemia may not often be measured before a problem is indicated. Perhaps another kind of ‘big data’, self-managed health checkups involving mobile technology with data capture and transmission, may soon be common and the anonymous data made available for this research. Addressing these information gaps is crucial to understanding why women with diabetes are relatively disadvantaged.
Future perspective
We have published compelling evidence that women are at a relative disadvantage when they have diabetes, at least in terms of cardiovascular disease – the most common sequela. Although this has long been suspected, to our knowledge, no comprehensive research program has previously been mounted to address this issue. Our long-term plan is to take on the first stage of this challenge: using existing datasets to create an evidence base from which appropriate sex-specific health policies can be formulated. We already have a large database of etiological associations between diabetes and CVD, by sex, which now requires updating, for example, by adding studies and by stratifying results by subjects’ age. Our future plan is to use the large databases, especially those that are available in the UK, to quantify sex-specific differences in accessibility and biological differences associated with diabetes, including genetic studies. Within the next few years we expect to provide the evidence needed to inform policies that will reduce, or even remove, the female disadvantage in diabetes.
Executive summary
We have estimated that the excess risk for a coronary event conferred by Type 2 diabetes is 44% higher in women than men; this excess is much greater for Type 1 diabetes.
We have, likewise, estimated that the excess risk for stroke conferred by Type 2 diabetes is 27% higher in women than men; this excess is similar for Type 1 diabetes.
It is possible, though unlikely, that this apparent female disadvantage is merely an arithmetic nicety, caused by the higher risks of cardiovascular risks for men compared with women in healthy populations.
A more likely explanation is that women receive poorer quality cardio-preventive care following diabetes than do men, although this issue is less likely to be a problem in modern societies where women have achieved a considerable degree of emancipation.
Women may also take less care of themselves after a diagnosis of diabetes than do men, but clear evidence is lacking.
A key factor in this female disadvantage may well be the greater subcutaneous fat storage reserves in women, meaning that they need to gain more weight before the state of diabetes is reached.
To help guide policies to eradicate, or reduce, the problem, we propose to explore several of the scientific hypotheses for the female disadvantage in diabetes using Big Data – linked routine clinical databases.
Footnotes
The authors have no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.
No writing assistance was utilized in the production of this manuscript.