Introduction

Previous studies have shown that the relative risk of CHD incidence and mortality associated with type 2 diabetes compared with non-diabetes is stronger in women than in men [1, 2], and this difference is not explained by traditional risk factors such as blood pressure and dyslipidaemia [3]. Although early reports suggested that this difference was fully explained by the lower risk of CHD in women in the general population (i.e. without type 2 diabetes), rather than diabetes having an absolute greater detrimental effect in women [4], the Asia-Pacific Collaboration Study found a greater absolute effect, such that women with diabetes had a greater risk of CHD (7.7% experiencing a fatal CHD event) than men with diabetes (4.5%) [1]. Although not all studies observe greater absolute risk of CHD in female diabetic patients compared with male diabetic patients, many observe comparable absolute CHD risk or much smaller differences in absolute CHD or cardiovascular risk between male and female diabetic patients compared with sex differences among non-diabetic patients [3, 57]. The reason for the greater relative risk of CHD associated with diabetes in women compared with men is unclear and may be a consequence either of diabetes inducing a more adverse cardiovascular disease (CVD) risk profile in women combined with differences in treatment of CHD risk factors between men and women [8, 9], or be due to a need for women to undergo much larger metabolic perturbances to transit from non-diabetes to diabetes, i.e. in general women may have to ‘travel’ or ‘deteriorate’ more to get diabetes.

Previous studies have shown differences in lipid abnormalities to be more pronounced between diabetic and non-diabetic women than between diabetic and non-diabetic men [1012]. However, these differences in the lipid profile appear insufficient to explain the differences in clinical risk [3]. Recent attention has turned to emerging CV risk factors including endothelial dysfunction, inflammation and fibrinolysis, which have been associated with diabetes risk [13]. Women with type 2 diabetes may be subject to even more adverse changes in coagulation, inflammation and vascular function than men [1417]. Insulin resistance, a recognised contributor to hyperglycaemia and diabetes precedes the development of diabetes by years and has been associated with increased CVD risk including dyslipidaemia, hypertension, impaired fibrinolysis, inflammation and coagulation [18, 19]. Stronger associations have been reported between insulin resistance/the metabolic syndrome and inflammation and biomarkers of endothelial dysfunction in women [20]. However, few studies have explored sex differences in the emerging risk factor profile in type 2 diabetes, and the role adiposity measures and insulin resistance may play in sex differences in CVD risk factors in those with and without type 2 diabetes. The purpose of this study was to examine and compare the cardiovascular risk profile of diabetic and non-diabetic men and women, including conventional and novel risk markers, and to assess the role of adiposity and insulin resistance in contributing to the possible greater differences in atherogenic profile between diabetic and non-diabetic women and between diabetic and non-diabetic men.

Methods

Data from the British Regional Heart Study and the British Women’s Heart and Health Study were used. The British Regional Heart study is a prospective study of CVD involving 7,735 British men drawn from general practices in 24 British towns followed from 1978 to 1980 [21]. In 1998–2000, all surviving men, now aged 60–79 years, were invited for a 20-year follow-up examination; 4,252 men (77% of survivors) attended. In 1999–2001, a parallel study of 4,286 women (60% of those invited) of the same age and drawn from 22 of the same 24 towns was established, with the addition of one more study town (Bristol) [22]. Full details of the selection of participants and measurements have been reported [22]. The study population were predominantly (>95%) described as white by examining nurses. Similar protocols for data collection were used in both studies. In both studies, nurses administered questionnaires, made physical measurements and collected fasting venous blood samples, from which serum was stored at −70°C for subsequent analysis. Men and women completed detailed questionnaires on medical history, medication and lifestyle, including cigarette smoking, alcohol consumption and physical activity. Details of measurement and classification methods for smoking status, physical activity, social class, blood pressure and blood lipids in the two cohorts have been described [2125]. Serum insulin was measured using an ELISA assay, which does not cross-react with proinsulin. Insulin resistance was estimated according to the homeostasis model assessment (HOMA-IR) formula [26]. Fasting glucose measurements were available in 3,829 women (89%) and 4,032 men (95%). Prevalent diabetes was defined as either: (1) participant’s report of a previous doctor’s diagnosis of diabetes on questionnaire; (2) a previous diagnosis of diabetes in general practice records; or (3) fasting blood glucose ≥7.0 mmol/l. Pre-existing myocardial infarction (MI) was defined as a participant’s report of a previous doctor’s diagnosis of MI on questionnaire or a previous diagnosis of MI in general practice records. All participants provided written informed consent to the investigation and ethics approval was provided by all relevant local research ethics committees. All men and women with pre-existing MI were excluded from the analyses to avoid potential biases because many haemostatic and inflammatory markers are raised as a consequence of having an MI [27], and many of them will be on medication to control their risk factors.

Haemostatic and inflammatory variables

Blood was anticoagulated with K2-EDTA (1.5 mg/ml) for measurement of white blood cell (WBC) count and plasma viscosity at 37°C in a semi-automated capillary viscometer (Coulter Electronics, High Wycombe, UK). Blood viscosity was calculated from plasma viscosity and microhaematocrit [19]. Blood was also anticoagulated with 0.109 mol/l trisodium citrate (9:1, vol./vol.) for measurement of clottable fibrinogen (Clauss method); as well as coagulation factors VII, VIII and IX in an MDA automated coagulometer. Plasma levels of tissue plasminogen activator antigen (t-PA) and D-dimer were measured with ELISA (Biopool AB, Umea, Sweden), as was von Willebrand factor (VWF) antigen (DAKO, High Wycombe, UK). C-reactive protein (CRP) was assayed by ultra-sensitive nephelometry (Dade Behring, Milton Keynes, UK). Interleukin-6 (IL-6) was assayed using a high-sensitivity ELISA (R & D Systems, Oxford, UK).

Anthropometric measurements

Measurements included height, weight, waist and hip circumference. Subjects were measured standing in light clothing without shoes. Height was measured with a Holtain stadiometer to the last complete 0.1 cm and weight with a Soehnle digital electronic scale to the last complete 0.1 kg. BMI (weight/height2 in kg/m2) was calculated for each participant. Waist circumferences (WC) were measured in duplicate with an insertion tape (CMS Ltd, London, UK); hip circumference was measured at the point of maximum circumference over the buttocks. The waist measurement was taken from the midpoint between the iliac crest and the lower ribs measured at the sides.

Statistical analysis

The distribution of CRP, IL-6, WBC count, blood viscosity and fibrin D-dimer were highly skewed and log transformation was used. Analysis of covariance was used to obtain adjusted mean levels according to sex and diabetes status. Adjustments were made in a stepwise manner, adjusting first for WC and then in addition for insulin resistance (HOMA-IR) to assess whether adiposity and insulin resistance explained the sex differences. Although duration of diabetes was not available in women, to take into account possible differences in duration of diabetes between men and women, we also adjusted in addition for glycated haemoglobin (HbA1c) as HbA1c generally rises with duration. Comparisons between diabetic and non-diabetic patients were carried out separately in men and women. Test for sex × diabetes interactions in their association with cardiovascular risk factors was assessed by fitting a sex × diabetes interaction term to the multivariable models. STATA was used for statistical analyses.

Results

Table 1 shows the baseline characteristics in all men and women without pre-existing MI. Overall, men tended to have significantly higher levels of adverse CVD risk factors than women; in particular higher mean levels of systolic and diastolic blood pressure, t-PA, fibrinogen, blood viscosity and fibrin D-dimer, and lower levels of HDL-cholesterol. However, women had higher levels of BMI, blood cholesterol and LDL-cholesterol, VWF, FVIII and WBC than men. There was no strong evidence of difference in levels of triacylglycerol, glucose, HOMA-IR, CRP or IL-6 between older men and women.

Table 1 Characteristics and level of atherosclerotic risk factors by sex in subjects with no pre-existing MI

Table 2 shows the age-adjusted mean levels of CVD risk factors by diabetes status in men and women separately. Non-diabetic men had significantly higher levels of systolic and diastolic blood pressure, HOMA-IR, IL-6, t-PA, fibrinogen, fibrin D-dimer, blood viscosity and lower levels of HDL-cholesterol than their female counterparts, but the magnitude of difference in these risk factors between males and females with diabetes was less marked, particularly for HOMA-IR (mean sex difference diabetic vs non-diabetic patients: −0.05 vs 0.35). The higher levels of WBC and FVIII seen overall in women were more marked in diabetic than in non-diabetic patients.

Table 2 Age-adjusted mean levels of biological and haemostatic and inflammatory markers in non-diabetic and diabetic men and women with no MI

Men and women with diabetes had higher levels of WC, BMI, systolic blood pressure, triacylglycerol, fasting blood glucose, HbA1c, HOMA-IR and haemostatic and inflammatory markers (except for fibrin D-dimer) and lower levels of HDL-cholesterol, cholesterol, LDL-cholesterol and lung function (forced expiratory volume in 1 s; FEV1) compared with their non-diabetic counterparts. Diabetes was associated with higher diastolic blood pressure in women but not in men. The differences in age-adjusted mean WC, BMI, blood pressure, HDL-cholesterol, triacylglycerol, HOMA, CRP, WBC, t-PA, blood viscosity and factor VIII between diabetic and non-diabetic women was greater than between diabetic and non-diabetic men. There was strong statistical evidence for sex heterogeneity in the association of diabetes with WC, BMI, diastolic blood pressure, HDL-cholesterol, HOMA-IR, t-PA and FVIII (test for diabetes × sex interaction all p values <0.05) and some evidence for blood viscosity (p = 0.06). There was little statistical evidence for heterogeneity between men and women for diabetes risk factor associations for CRP, systolic blood pressure and triacylglycerols despite the point estimates for the associations with these risk factors being greater in females as with other risk factors. Exclusion of men and women on statins made little difference to the findings seen in Table 2.

The diabetes × sex interactions seen for diastolic blood pressure, HOMA, t-PA, WBC and FVIII changed little with adjustment for differences in WC, but were attenuated after further adjustment for HOMA-IR (Table 3). The diabetes × sex interaction for the association with WBC remained even after these adjustments. Adjustment for HOMA-IR reversed the positive association between diabetes and t-PA, which became inverse after adjustment for HOMA-IR in both sexes.

Table 3 Adjusted mean differences in biological and haemostatic and inflammatory markers between diabetic and non-diabetic participants (Diab–Non-diab) with no MI and in mean levels between men and women according to diabetes status

Glycated haemoglobin levels (HbA1c) were similar in diabetic men and diabetic women and adjustment for HbA1c had little effect on the above findings.

Discussion

Men have much higher absolute risk of CHD than women, although the gap narrows after the menopause. The majority of studies have shown that the relative risk of CHD associated with diabetes is far greater in women than in men [1, 2], resulting in comparable or smaller sex differences in absolute CHD risk in diabetic patients [3, 57]. The reasons why diabetes in women increases the relative risk of CHD more than in men compared with their non-diabetic counterpart is not clear, but a possible explanation may be that diabetes has a greater adverse effect on CVD risk factors in women than in men.

In this large study of older men and women sampled across the UK, non-diabetic women tended to have a more favourable risk profile than non-diabetic men and in particular were less insulin resistant than men but this difference was markedly diminished in the diabetic state. Contrary to findings with most CVD risk factors, women without diabetes had significantly higher levels of VWF and WBC than men without diabetes; this sex difference was more marked in women and men with diabetes. Although absolute levels were not higher, diabetic women showed greater relative differences in abdominal adiposity (WC), insulin resistance, lipids (low HDL-cholesterol), diastolic blood pressure, inflammation (WBC), endothelial dysfunction (t-PA) and coagulation (FVIII) than diabetic men when compared with their non-diabetic counterparts. The sex differences in the association between diabetes and these risk markers were to some extent explained by adiposity and more strongly by insulin resistance (HOMA-IR).

Our study supports and extends previous studies by examining the cardiovascular risk profile of a wide range of novel risk markers including markers of inflammation and endothelial dysfunction in non-diabetic and diabetic men and women and assessing the role of insulin resistance (HOMA-IR) in explaining these sex differences. Our findings are in keeping with the notion that as women go from non-diabetes to diabetes, they have to ‘travel’ further, i.e. they need to put on more weight, and deteriorate their insulin sensitivity and related risk factors to a greater extent than do men. They do so in part because the average middle-aged women (in differing parts of the world) are at lower risk of diabetes (and thus more insulin sensitive) than the average man as shown by recent studies [28].

Women with diabetes, compared with their non-diabetic counterparts, showed a greater difference in central adiposity than was seen when comparing diabetic and non-diabetic men, consistent with the finding that WC has been shown to be a stronger predictor of diabetes in women than in men [25]. The relatively greater difference between men and women in adiposity partially explains the greater difference in insulin resistance seen between diabetic women and non-diabetic women compared with men. The stronger relative association between diabetes and HDL-cholesterol and diastolic blood pressure in women than in men observed in this study has been noted in other studies [3, 9], and we have shown this to be associated with the relative greater adiposity and insulin resistance (HOMA-IR) in women. Unlike previous studies we did not find any significant interaction between diabetes status and LDL-cholesterol [11]. This difference in findings may relate to age as in this study non-diabetic women had higher (not lower as commonly seen in younger populations) LDL-cholesterol and cholesterol levels than their male counterparts [11]. Moreover, both male and female diabetic patients showed lower not higher levels of LDL-cholesterol and total cholesterol than non-diabetic patients, which was not explained by use of statins. The lower level of LDL-cholesterol in diabetic patients compared with non-diabetic patients has been observed in the Strong Heart Study [9] and is largely linked to alteration in LDL particle composition in diabetes with appearance of smaller denser particles that contain less cholesterol.

Insulin resistance has been associated with abnormalities of coagulation, inflammation, endothelial dysfunction and fibrinolysis [18, 19]. The greater relative risk of CHD associated with diabetes in women may be explained by diabetes producing a more adverse change in the coagulation, inflammation and fibrinolytic system than men [1417]; or, as explained above, changes in such pathways are simply greater due to a greater deterioration in insulin resistance in women. We examined several markers of inflammation that have been associated with diabetes and CHD including CRP, IL-6 and WBC [13, 23], and all showed differing patterns of associations. In this older population there were no sex differences in levels of CRP in those with or those without diabetes. Non-diabetic men showed significantly higher levels of IL-6 than non-diabetic women but this sex difference was abolished in diabetic patients. However, both non-diabetic and diabetic women showed significantly higher levels of WBC than their male counterparts and the difference was more marked in diabetic patients. A significant sex × diabetes interaction was seen for WBC, which was robust to adjustment for WC and HOMA-IR, but no sex difference was found for the association with CRP or IL-6. Sex × diabetes interactions with CRP have been observed in some but not all studies [14, 29]. Our finding of a marked sex × diabetes interaction in the association with WBC has some consistency with the finding that WBC has been shown to be more strongly correlated with insulin sensitivity in older women than in men [30].

Circulating biomarkers of endothelial dysfunction such as VWF and t-PA have been shown to predict CHD [23]. Women had significantly higher levels of VWF (one marker of endothelial dysfunction) and FVIII than men in both non-diabetic and diabetic participants; circulating factor VIII levels partly reflect levels of its endothelial-derived carrier protein, VWF (as shown by the strong correlation (0.68) between factor VIII and VWF in the present study). However, men had significantly higher levels of t-PA antigen than women in both groups but the sex difference diminished in diabetic patients. Diabetes was associated with a more adverse effect on t-PA in women than in men, which is consistent with the stronger influence of diabetes on plasminogen activator inhibitor 1 (PAI-1) in women observed in previous studies [18]. Elevated t-PA antigen (which largely reflects circulating inactive tPA–PAI-1 complexes) is considered an integral feature of the insulin resistance syndrome [19]; the greater difference in t-PA between diabetic and non-diabetic women was entirely explained by insulin resistance and WC. Thus, the greater increase in endothelial dysfunction may also contribute to the loss of cardiovascular protection in diabetic women.

Although the majority of studies have shown the relative risk of CHD associated with diabetes is far greater in women [1, 2] than in men, it is important to note that the absolute risk of CHD has not always been shown to be greater in diabetic women than in diabetic men [3, 57, 31]. Nevertheless many of these cohort studies have shown comparable absolute risk of coronary or cardiovascular events in diabetic women compared with diabetic men or diminished sex differences in absolute risk in the diabetic state [3, 57]. This is consistent with the pattern seen for many of the CVD risk factors observed in this study, where the more favourable pattern of risk factors seen in women is diminished in the diabetic state suggesting that the greater excess in CVD risk factors in diabetic women may contribute to the greater increase in relative risk of CHD in diabetic women.

The strengths and limitations of the present study require careful consideration. The study population is not strictly a random population sample, being influenced by survival and response, both of which will tend to lead to under-representation of selected groups of individuals such as smokers or obese participants. Although this may affect the average levels of the biological markers in the population, there is no reason to believe that under-representation per se should bias the relationships between diabetes and the biological markers studied. However, the average HbA1c (which generally increases with duration) was near identical in diabetic women compared with diabetic men suggesting similar duration in male and female diabetic patients. We cannot extend our findings to other ethnic groups or to younger participants. The cross-sectional nature of our study cannot provide evidence as to whether development of type 2 diabetes mellitus in women exerts or requires more adverse changes in coagulation, inflammation and vascular function than in men, although this hypothesis would fit with the current data. Nor can we provide direct evidence as to whether these greater differences seen particularly for insulin resistance, HDL-cholesterol and markers of activated coagulation, inflammation and endothelium between diabetic women and non-diabetic women compared with non-diabetic men and diabetic men contribute to the greater relative risk of CHD associated with diabetes in women. Both these aspects require future study.

In conclusion, in this large study of older adults aged 60–79 years, women tended to have more favourable cardiovascular risk factors than men, but many of these advantages were diminished or abolished in the diabetic state. Diabetes was associated with increased CVD risk factors in both men and women, but diabetic women showed greater relative excess in many CVD risk factors than diabetic men; these excesses were to some or a large degree explained by the greater increase in adiposity and insulin resistance associated with diabetes in women than in men. Our results therefore are consistent with the hypothesis that women have to undergo a greater metabolic deterioration to develop diabetes than do men and as such, many insulin resistance-related risk factors, as observed herein, must change to a greater extent. The greater relative excess in many CVD risk factors in diabetic women may help to explain the increased relative risk of CHD in women with diabetes compared with men.