In this post-hoc analysis two surrogate indexes of IR, containing metabolic and anthropometric parameters, namely VAI and TG/HDL-C index, were significantly associated with a first MI in females, but not in males, independently of traditional CV risk factors and glycaemic state, the latter characterized by OGTT.
In the present population, there were sex differences in anthropometric and metabolic characteristics. Females displayed a better cardio-metabolic risk profile than males since they were less often smokers, had lower BMI, WC and triglycerides levels, and were less insulin resistant than males. In general, female sex is characterized by the storage of adipose tissue in subcutaneous sites as compared with preferential visceral deposition in males (24). Previous studies have shown that in a population with normal blood glucose levels females are more insulin sensitive than males (25), although this sex advantage disappears in females with diabetes (26). Additionally, the VIRGO study found that young females with MI only had a slightly more favourable lipid panel compared with males, suggesting that the sex difference in outcomes after MI cannot be explained by dyslipidaemia only (27). Several clamp studies, investigating how sex can affect insulin sensitivity in individuals with NGT, showed higher insulin-stimulated glucose disposal in females than in males (28–36). In our comparison, females displayed lower fasting plasma glucose and insulin levels, and consequently, lower HOMA-IR than their male counterparts. This is in accordance with previous work showing a higher prevalence of IFG in males than in females (8). However, HOMA-IR might not be the most accurate way to express IR in females, considering that it mainly represents hepatic IR, whereas females would be more exposed to peripheral IR (37). Using indexes that include anthropometric characteristics and lipids could help clarify the mechanisms underlying the sex differences across glycaemic states.
Another important finding is that, in both sexes, patients with a first MI had consistently higher values of all IR indexes compared with controls, not only in the general study population but also in the subgroup with NGT. This is remarkable, as we classified NGT quite strictly, excluding not only patients with IGT and T2DM but also those with IFG. This, together with the fact that patients, all surviving a first MI, were relatively healthy, further supports the hypothesis that a certain degree of metabolic derangement exists in CAD, even without glycaemic perturbations (38).
Thus, even if VAI and TG/HDL-C index were lower in females than in male, they might capture an early stage of metabolic disturbance that is not pictured by the glycaemic state but is still clinically important and they could be better predictors of MI. This is in line with a recent study highlighting that VAI is associated with CV events in normal weight and over-weight subjects, but not in those who had obesity (39), suggesting that indexes derived from multiple parameters, both anthropometric and laboratory, are more able to identify CV risk in healthier populations.
Additionally, a Chinese study shows that VAI is significantly associated with intracranial atherosclerotic stenosis in middle-aged and elderly females (40). Accordingly, in a recent clamp-based study in patients without diabetes, there was greater deterioration of insulin sensitivity and greater fat accumulation in females than in males (11). Therefore, we hypothesized that VAI and TG/HDL-C index could have a better performance than OGTT-derived glycaemic status and HOMA-IR in identifying clinically meaningful IR, with the advantage that they are more practical than OGTT or clamps. Indeed, both VAI and TG/HDL-C index are based on anthropometrics and lipid profiles, (and do not at all include insulin levels), which are routinely assessed, and they are quite well validated.
Assessing VAI and TG/HDL index in females could contribute to the early identification of metabolic deterioration and to instituting preventive measures, such as lifestyle modifications, that could prevent them from developing T2DM and hence losing advantages in terms of CV risk. Indeed, with CV risk stratification being widely based on the limited number of traditional risk factors derived from the Framingham study, our findings provide a rationale for further exploring the possibility of finding risk factors that are more specific for CV risk in females.
The INTERHEART study tried to widen the view on CV risk factors but confirmed that the two most important ones were smoking and increased lipids, followed by hypertension and diabetes, and the importance of all CV risk factors was similar in both sexes, regardless of age and geographical region (41). Although females remain less represented in CV clinical trials than men, recent years have brought a better understanding of the sex differences in the biological processes accounting for CV risk factors, allowing for an expansion of the number of factors that might play a key role in CV risk. The main challenge in the following years will be to more equitably include both sexes in trials, aiming at adopting efficient preventive measures in females and males, separately.
Strengths and limitations
The present study has several strengths. Firstly, the study cohort was recruited from 17 Swedish hospitals, covering a nationwide geographical area and various educational and socioeconomic states. The whole population is well characterised and relatively young and healthy, as it includes only patients with a first MI and a well-matched control population. Some limitations should be underlined. As a post-hoc investigation it can only be hypothesis-generating and the present findings need further confirmation. The proportion of females in the complete population (19%) was relatively low and, even if this is typical for a MI population with an upper age limit of 75 years, it restricts the power of the analyses. Additionally, IR was not evaluated with the gold standard method, the Euglycaemic Hyperinsulinemic Clamp, but with surrogate indexes, however validated and more clinically practical.