Calculated whole blood viscosity in non-diabetic subjects with asymptomatic carotid atherosclerosis: How insulin resistance may affect blood viscosity

1 Introduction

Asymptomatic carotid atherosclerosis (ACA) is an early marker of vascular damage and expresses the susceptibility to develop atherosclerotic disease, regardless of the presence of one or more risk factors. Subjects with evidence of ACA have a higher risk of short- and long-term cardiovascular events than controls grouped by age and sex. Therefore, the early stages of atherosclerosis are key for an early diagnosis and satisfactory prognostic results. All risk factors are related to the presence of endothelial dysfunction and, at the same time, the hemorheological alterations play a role in the pathogenesis and progression of the disease, as well as its prognosis [1–3].

Plasma viscosity, of which fibrinogen is one of the most important determinants as it influences the erythrocyte aggregation [4], is very important but, although several studies have confirmed its the role in the progression of cardiovascular diseases, the assessment has not always been considered in clinical practice [5, 6].

A hemorheological alteration appears to be a risk factor for the development of a cardiovascular disease due to its effect on atherogenesis, thrombogenesis and tissue ischemia. Hemorheological parameters may be correlated with the degree of carotid stenosis in both symptomatic and asymptomatic subjects, without implications with such risk factors [7–11]. Blood viscosity correlates with mean intimal thickening and reduced flow-mediated dilation [12] as well as such observation would tend to complicate the analysis of the cause-effect relationship between hemorheological alteration and atherosclerosis [13].

The determination of blood viscosity, which is related to cell mass (expressed by the haematocrit), plasma viscosity as well as erythrocyte aggregability and deformability, was carried out ex vivo using different types of viscometers such as rotational, capillary, and oscillatory ones. The same applies to the ex vivo assessment of both erythrocyte aggregability, determined directly with the use of aggregometers such as the Myrenne MA-1 and the Laser-assisted Optical Rotational Cell Analyzer (LORCA), and the erythrocyte deformability measured above all with filtration, with the use of micropipettes or with diffractometric techniques.

However, it is possible to calculate the blood viscosity from haematocrit and total plasma proteins or from haematocrit and plasma levels of fibrinogen. Some studies have measured blood viscosity in normal subjects with cardiovascular issues [14] and in subjects with insulin resistance [15].

Blood viscosity was calculated for the haematocrit and total plasma protein concentration and evaluated in relation to arterial pressure [16, 17], coronary and carotid atherosclerosis [13], coronary artery disease awaiting coronary artery bypass graft [18], microvascular angina [19] and myocardial infarction, to estimate the formation of a left ventricular thrombus within twelve months of an acute anterior infarction [20]. Starting from the same parameters, the calculated blood viscosity was examined in normal young men in relation to insulin sensitivity [21], in a group of subjects with prediabetes [22], in type 2 diabetics divided both by sex and diabetic retinopathy [23], in type 2 diabetics with metabolic syndrome [24], in subjects with hypertriglyceridemia undergoing plasmapheresis [25], in subjects with aortic sclerosis [26], in subjects with aortic valve sclerosis [27] and in scleroderma subjects with or without pulmonary hypertension [28].

As hypothesized by Merrill et al. [29], blood viscosity may be expressed as Yield Shear Stress (YSS) and calculated from the haematocrit values and fibrinogen plasma levels. The YSS illustrates, according to the Merrill formula, the maximum shear stress at which blood may cease to flow. This parameter up to now has been evaluated in type 1 diabetic subjects divided according to the glycometabolic control [30], in type 2 diabetic subjects with or without associated arterial hypertension [31] and in patients with multiple myeloma and in subjects with MGUS [32].

The indirect determination of blood viscosity, which as is known is a simple parameter with implicit complex implications, seems particularly useful when studies on the population are carried out, taking into account that, how it was estimated in the “de Simone’s study” [14], the haematocrit and total plasma proteins alone seem to predict more than 80% of the variability of the same blood viscosity, within a range of shear rates between 0.1 and 208 s-1.

Another consideration related to the calculated blood viscosity concerns the fact that all the formulas used so far do not consider neither the deformability nor the aggregability of red blood cells. The latter data is therefore an obstacle for the indirect assessment of blood viscosity in all clinical conditions characterized by primitive sclerocytemic hyperviscosity, which is generally observed in hereditary spherocytosis, beta-thalassemia, sickle cell anaemia [33] but also in hereditary elliptocytosis, chorea-acanthocytosis, ovalocytosis and hereditary stomatocytosis.

In a previous paper, [34] we examined the principal hemorheological determinants (whole blood viscosity, plasma viscosity, erythrocyte aggregation, haematocrit, plasma fibrinogen) in a cohort study, that also included the diabetic subjects, with ACA. The data obtained from this previous research showed that only whole blood viscosity, at high shear rate, distinguish the control group from these subjects, while the other hemorheological determinants show a particular trend when the whole group of these subjects were divided according to the number of cardiovascular risk factors and in relation to the medians of the surrogate markers reflecting the insulin resistance degree.

In this study, we examined the behaviour of the calculated blood viscosity using the Merrill formula in a group of non-diabetic subjects with asymptomatic carotid atherosclerosis (ACA) grouped by risk factors and especially in relation to the surrogate indices reflecting the insulin resistance degree.

2 Materials and methosd

3 Results

By splitting the whole group of ACA subjects according to the number of cardiovascular risk factors (Table 2) we observed that in the subgroup with 3-5 CRFs a significant increase in haematocrit, in plasma fibrinogen and in calculated whole blood viscosity (cWBV) is present. After having compared the ACA subject group in relation to the medians of the four surrogate indices of insulin resistance, (Table 3) we found for each group (carried out respectively for TG/HDL-C ratio, TyG, VAI and LAP) a constant increase in plasma fibrinogen that, however, reached a low degree of statistical significance only in the group considered for VAI. No variation instead has been observed for the haematocrit and cWBV.

Subsequently, we have correlated each surrogate index of insulin resistance with the cWBV. The analysis of this correlation, carried out with the Spearman rank, showed that no correlation is present in the entire group of ACA subjects and in the subgroup with 1-2 CRFs while in the subgroup with 3-5 CRFs (Table 4), even if at worst or at low degree of statistical significance, there is a positive relationship between cWBV and TG/HDL-C ratio and between cWBW and TyG.

It may be argued that the data highlights that in the group of non-diabetic ACA subjects the cWBV as well as the haematocrit and the plasma fibrinogen clearly distinguish ACA subjects with 3-5 CRFs from those with 1-2 CRFs. The presented data for the cWBV does not conflict with our previous studies where the measured whole blood viscosity (even if at high shear rate only) emphasised the same division. Considering that the evaluation of the cWBV was carried by using the Merrill formula, it may be argued that in ACA subjects with more CRFS the behaviour of haematocrit and fibrinogen has a direct consequence on cWBV. Keeping in mind that in this group of ACA subjects there is a prevalence of a cardiovascular and/or cardiometabolic clustering (as it may be inferred by the high percentage of CRFs discovered in each subject of this group) an increase of cWBV is expected. As it is known, the alteration of the hemorheological profile has been described in subjects with impaired plasma lipid pattern, in subjects with arterial hypertension, in subjects with overweight and obesity, in subjects with a family history of cardiovascular disease, in smokers and ex-smokers, in metabolic syndrome and in subjects with prediabetes; this last metabolic condition seems even associated with asymptomatic carotid atherosclerosis [45]. There is much literature on this and on the effects on hemorheological profiles, that the CRFs present in each subject of this ACA group, associated or not to a condition of insulin resistance, influence certainly haematocrit behaviour [46–64].

Having divided this ACA subject group by the medians of the different surrogate indices that reflect the insulin resistance degree, no difference has been observed, relatively to each subgroup examined respectively for TG/HDL-C, TyG, VAI and LAP, about haematocrit and cWBV while in each division there is a constant and light increase of the plasma fibrinogen that results at worst of statistical significance when measured in relation to the VAI index. Plasma fibrinogen plays its role in atherothrombosis through different mechanisms, such as: promotion of atherosclerosis by proliferation of endothelial and smooth vascular muscle cells, component of platelet aggregation, modulator of the amount of deposited fibrin and size of blood clot, increasing plasma viscosity. Several authors claimed that plasma fibrinogen is also a marker of clinical and asymptomatic carotid atherosclerosis [65–69]. As it is known, plasma fibrinogen is also the main component of the coagulation system with inevitable implications in the pathophysiology of the atherosclerotic carotid disease. Interesting is also the data regarding the plasma fibrinogen behaviour in clinical disorders associated with insulin resistance [70–76]. Besides influencing plasma viscosity and erythrocyte aggregation, plasma fibrinogen (in the clinical disorders associated with insulin resistance) is quickly synthetized by the liver under stimulation of IL-6, TNF-alpha and free fatty acids. It must be also underscored that insulin resistance results also in an increase of the plasminogen activator inhibitor –1 and thrombin-activator fibrinolysis inhibitor that reduces fibrinolytic activity [77].

In the whole group, but also in the subgroup of ACA subjects with 1-2 CRFs, no correlation was observed between cWBV and each surrogate index of insulin resistance degree while in the subgroup with 3-5 CRFs has been found a positive relationship, at worst and/or at low degree of statistical significance, between cWBV and TH/HDL-C and between cWBV and TyG. Moreover, many authors have assessed that between insulin sensitivity and blood viscosity there is a strong negative correlation while positive was between insulin resistance and blood viscosity [78–82]. To date, this correlation was observed not only for the whole blood viscosity but also for some determinants of the hemorheological profile such as erythrocyte aggregation [83] and plasma viscosity [84, 85]. For this research we only calculated whole blood viscosity in non-diabetic subjects with ACA, clearly demonstrates the presence of a correlation between cWBV and the two surrogate indices of insulin resistance which is superimposable to what previously had been observed [34] after having examined the correlation between the same indices with the measured whole blood viscosity. This last assertion seems to confirm the usefulness to resort to the calculated or estimated whole blood viscosity, according to the “de Simone and/or to the Merrill formula”, when the research takes into consideration a large group of subjects or concerns an observational study. Moreover, keeping in mind that the blood viscosity may be considered a marker or predictor of the cardiovascular damage, we argue that in each clinical disorder, though asymptomatic, in which results are evident, even if with alone instrumental investigation such as coronary angiography, ultrasound of large arteries, cerebral angiography, it is necessary to determine directly (measured) or indirectly (calculated or estimated) the whole blood viscosity. It must be underlined that the vascular geometry of the carotid area modifies the hemodynamic profile, and this variation affects blood viscosity. Moreover, many authors have analysed the role of the modified shear stress in the pathogenesis of atherosclerosis [86–91] while Sloop [92] elaborated the hemorheological-hemodynamic theory of atherogenesis, which is useful to explain the first stages of atherothrombosis.

In conclusion, the cWBV in non-diabetic subjects with ACA operates a clear distinction regarding the number of CRFs, which result unaltered with reference to the degree of insulin resistance and only in the subgroup of ACA subjects with 3-5 CRFs it results correlated at a low degree of significance, with TH/HDL-C and TyG.