Skin microvascular reactivity and subendocardial viability ratio in relation to dyslipidemia and signs of insulin resistance in non‐diabetic hypertensive patients

The aim of this study was to evaluate the influence of dyslipidemia and insulin resistance for the development of microvascular dysfunction in non‐diabetic primary hypertension.

and insulin resistance, and impaired glucose uptake and increased peripheral resistance may in addition to endothelial dysfunction further deteriorate skin microvascular function 6,7 Skin microvascular dysfunction is present in individuals in the general population with cardiovascular (CV) risk factors and risk of developing type 2 diabetes 8,9 and in patients with type 2 diabetes and the metabolic syndrome. [10][11][12] Reduced endothelium-dependent vasodilatation of the skin is associated with coronary heart disease and future risk of CV events. 13 In addition, heat-induced maximum reactive hyperemia may represent total microvascular reactivity of the skin, where impaired vasodilatation after local heating is associated with increased CV risk and risk of acute coronary syndrome, further enhanced in diabetic patients. 13,14 Subendocardial viability ratio (SEVR) is a non-invasive estimate of myocardial oxygen supply and demand, calculated as the ratio of the area under the curve of the diastolic to systolic derived aortic pressure waveform. 15,16 SEVR is associated with coronary flow reserve, and a reduced SEVR is associated with increased CV risk and worse prognosis in patients with diabetes and with chronic kidney disease. [17][18][19][20] The metabolic syndrome and diabetes are associated with dyslipidemia and low levels of high-density lipoprotein cholesterol (HDL) levels, where impaired anti-oxidative capacity of HDL causes endothelial dysfunction and promotes inflammation and atherosclerosis. 21,22 The anti-atherogenic effects of HDL are attenuated in type 2 diabetes, reducing endothelium-dependent vasodilatation. 23 A disturbed glucose metabolism and impaired skin microvascular function is present in familial combined hyperlipidemia. 24,25 Hypercholesterolemia seems to impair the vasoprotective functions of HDL by alterations of HDL expression in endothelial cells, as shown in animal studies. 26 Thus, skin microvascular dysfunction is related to insulin resistance and dyslipidemia, where HDL seems to have an important role in regulating microvascular reactivity.
Several biomarkers of metabolic status have been shown to serve as proxies of insulin resistance. An increased ratio of triglyceride/HDL (TG/HDL) is associated with insulin resistance and the metabolic syndrome, and future risk of ischemic heart disease in the general population. [27][28][29] The triglyceride-glucose index (TyG), the product of fasting plasma glucose and triglycerides, is associated with insulin resistance and the metabolic syndrome, and future risk of developing diabetes. 30,31 Furthermore, TyG is associated with subclinical atherosclerosis and an increased risk of CV events. 32,33 Whether non-diabetic hypertensive patients have early changes in skin microvascular function due to early metabolic changes and dyslipidaemia with low HDL cholesterol has not been well studied.
Thus, the present study aimed to evaluate skin microvascular vasoreactivity and coronary microvascular function in relation to lipid profile with signs of dyslipidemia and early development of insulin resistance in non-diabetic hypertensive patients.

| MATERIAL S AND ME THODS
The Doxazosin-ramipril study investigated women and men above 18 years of age with untreated mild-to-moderate primary hypertension (office systolic blood pressure 141-180 mm Hg and/ or diastolic blood pressure 91-110 mm Hg). Exclusion criteria were ischemic heart disease, severe hypertension (180/110 mm Hg), chronic heart failure, arrhythmias, diabetes mellitus, and pregnancy. The primary aims of the study were to evaluate the effects of treatment with ramipril or doxazosin during 12 weeks on endothelial function and hemostasis, and the main results have been presented in detail elsewhere. 34,35 We, here, report crosssectional findings from vascular examinations and biochemistry regarding lipids and glucose control in untreated participants at baseline prior to randomization to active treatment. All vascular examinations were performed after overnight fasting and without intake of nicotine or caffeine, or any medications influencing endothelial function, in the supine position after 20 minutes of rest, at room temperature. 34 This study is registered at ClinicalTrials.gov (NCT02901977) and at EudraCT (# 2007-000631- 25), and was approved of by the appropriate Ethics committee. All subjects gave their oral and written consent to participate.

| Blood pressures measurements
Brachial blood pressure readings were obtained in the supine position by an oscillometric device (OMRON 705IT, OMRON Healthcare Co, Ltd. Kyoto Japan) on the dominant arm with an appropriately sized cuff, as a mean of three readings 1 min apart. LDF and local heating of forearm skin to +44° C for 6 min was used to evaluate the maximum cutaneous reactive hyperemia as a measure of total skin microvascular reactivity. 34,36 Assessment of skin microvascular function by LDF and transdermal iontophoretic drug administration, and by LDF and local heating, are validated and reproducible methods. 14,37,38 Microvascular function was also expressed as maximal cutaneous vascular conductance (CVC), calculated as cutaneous blood flow (in PU) divided by brachial mean arterial pressure (diastolic +1/3 (systolic -diastolic) blood pressure, in mm Hg) to account for blood pressure differences between patients. 38 2.3 | Subendocardial viability ratio as a marker of coronary microvascular function SEVR is a non-invasive estimate of myocardial oxygen supply and demand calculated as the ratio of (aortic diastolic pressure x time integral) to the (aortic systolic blood pressure x time integral), which is taken to represent the subendocardial perfusion capacity relative to myocardial contraction, that is, myocardial perfusion relative to cardiac workload. [15][16][17] SEVR was derived from a general transfer function using pulse wave analysis (SphygmoCor, AtCor Pty Ltd, West Ryde, NSW, Australia) with applanation tonometry (Millar Instruments, Houston, TX, USA), as described elsewhere. 17,39 Measurements were made prior to, and separate from, the evaluation of endotheliumdependent vasodilation of the resistance arteries with beta 2-adrenoceptor agonist stimulation (see below) to exclude confounding influence. SEVR has been validated with invasive coronary artery measurements in hypertensive patients without coronary heart disease and is associated with coronary flow reserve. 17 However, SEVR may overestimate the diastolic pressure x time integral as it is derived from blood pressure measurements and not from left ventricular end diastolic pressure. 40 SEVR is dependent of heart rate and the slope of diastolic decay, as well as to age and sex. 17,18 This potential confounding was addressed by additional multivariable analyses.

| Endothelial function in resistance arteries and large arteries
To evaluate endothelium-dependent vasodilation of the resistance arteries we used applanation tonometry and pulse wave analysis with additional beta 2-adrenoceptor agonist stimulation (terbutaline Germany). Relative changes in artery diameter were calculated from rest to 4 min following GTN administration. 34 To calculate the endothelium-dependent in relation to endothelium independent vasodilatation, the endothelial functional index was calculated as the FMD/GTN ratio, as described elsewhere. 34  the TyG-index, calculated as ln[fasting plasma glucose (mg/dl) x triglycerides (mg/dl)/2], and the TG/HDL ratio. 30

| Microvascular reactivity in relation to lipid profile
ACh-mediated peak flux was related to HDL, but not to LDL, or the LDL/HDL ratio (Table 3; Figure 1A,B). Heat-induced maximum peak flux was related to HDL and to the LDL/HDL ratio, but not to LDL ( Table 3; Figure 1C,D). SNP-mediated peak flux and the ACh peak flux/SNP peak flux ratio did not relate to HDL, LDL, or the LDL/ HDL ratio ( Table 3). The results for the relative changes in peak flux after ACh, SNP and heat stimulation were similar (data not shown).
CVC values for ACh-mediated peak flux and heat-induced peak flux showed similar results (Table 3). Multivariable analysis confirmed that microvascular reactivity was independently related to HDL, for endothelium-dependent peak flux (p = .011) and heat-induced maximum peak flux (p = .017), respectively. In contrast to skin microcirculation, indices of coronary artery microvascular function assessed by SEVR were inversely related to LDL, but was unrelated to HDL, and the LDL/HDL ratio (Table 3; Figure 2A,B). However, this relation to LDL was not retained in a multivariable analysis (p = .24).

| Macrovascular endothelial function in relation to lipid profile
Endothelium-dependent vasodilatation in conduit arteries (as assessed by FMD) tended to relate to HDL, but not to LDL, or the LDL/ HDL ratio (Table 3). Endothelium independent vasodilation (as assessed by GTN) did not relate to HDL, LDL, or the LDL/HDL ratio (Table 3). However, the endothelial functional index (i.e., FMD/GTN) was related to HDL but not to LDL, or the LDL/HDL ratio (Table 3; Figure 3A,B). Endothelial function in resistance arteries (as evaluated by the RI change) related inversely to HDL, and related to the LDL/HDL ratio, but not to LDL, that is, a smaller reduction in RI change after beta 2-agonist stimulation was related to dyslipidemia with low HDL (Table 3; Figure 3C,D). Multivariable analysis confirmed an independent relation between RI and HDL (p = .017) but not to Endothelial functional index and HDL (p = .15).

| Microvascular reactivity and endothelial function in relation to markers of metabolic status
Peak flux to ACh was inversely related to the TG/HDL ratio and tended to inversely relate to TyG (Table 4; Figure 4A,B). SNPinduced peak flux, the ACh peak flux/SNP peak flux ratio, and heat-induced maximum peak flux were unrelated to TyG and TG/ HDL ratio (Table 4; Figure 4C,D). Corresponding CVC values for skin microvascular function showed similar results (Table 4). In contrast, coronary microvascular function assessed by SEVR did not relate to TyG or the TG/HDL ratio (Table 4) and large artery function (FMD, GTN, endothelial function index and RI change) was unrelated to measurements of insulin resistance (data not shown).

| DISCUSS ION
This appears to be the first study in non-diabetic hypertensive patients to explore endothelial function in different vascular beds in relation to signs of dyslipidemia and insulin resistance. Our main findings are, first, that impaired skin microvascular function related with signs of dyslipidemia with lower HDL cholesterol levels, and with insulin resistance. Second, heat-induced maximum reactive hyperemia, appears to be a robust method to evaluate total skin microvascular function in non-diabetic primary hypertension. Third, skin microvascular function and SEVR, as a marker of coronary microvascular function, showed different associations to lipid profile.
Our results showed that reduced ACh-mediated peak flux was associated with low HDL, and RI was inversely related to HDL.
These findings confirm the interrelation between HDL and the endothelium with a vasoprotective role of HDL. 26 Of note, the function of the HDL is important for the NO pathway, oxidative capacity, vascular inflammation, and for endothelial function. 21 F I G U R E 3 Relations between (A) endothelial functional index (EFI) and HDL, (B) EFI and LDL/HDL ratio, (C) relative change in reflection index (RI) before and after beta 2-adrenoceptor agonist stimulation and HDL, and (D) RI and LDL/HDL ratio. EFI, calculated as the FMD to GTN ratio, is an index of endothelium-dependent vasodilation. A lower EFI represents impaired endothelial function. RI change (%) is the relative change of height of the reflecting radial pulse waveform before and after beta 2-adrenoceptor agonist stimulation.
A smaller relative change in RI (%) represents impaired endothelial function HDL. In hypertension, skin blood flow is impaired due to a reduced endothelium-dependent vasodilatation, and to a diminished neurogenic response after local heating. 3 Heat-induced maximum reactive hyperemia is induced by an early direct axon mediated reflex, and by a delayed response mediated mainly through NO. 38 Thus, evaluation of global microvascular function by heat-induced maximum reactive hyperemia provides information about total skin microvascular reactivity. Furthermore, our data support the contention that heat-induced maximum reactive hyperemia is reduced in primary hypertension and relates to an early stage of dyslipidemia with low HDL.
We used the TG/HDL ratio and TyG as markers of insulin resistance. Our results showed that an impaired response to AChmediated vasodilatation related inversely to the TG/HDL ratio. Evaluation of skin microvascular function is subject to confounding by external influence, including changes in heart rate, blood pressure, skin temperature, and vascular sympathetic tone. However, the experimental conditions were strictly standardized with vascular examinations performed on one single occasion. 14,34 In addition, we used CVC as internal validation for skin microvascular measurements to account for differences in baseline mean arterial pressure among patients, with similar results. We did not evaluate myocardial microvascular function invasively. However, indirect assessment of microvascular function by SEVR has been validated to invasive measurements of myocardial microcirculation and to cardiovascular outcome, as discussed above. We used the TG/HDL ratio and

| PER S PEC TIVE S
Our results in patients with uncomplicated non-diabetic hypertension suggest early signs of microvascular dysfunction associated to dyslipidemia with low HDL and insulin resistance. This is in consort with observations in type 2 diabetes and in the metabolic syndrome.

ACK N OWLED G M ENT
We thank Ms. E. Andersson, J. Rasck and E. Wallén Nielsen for expert technical assistance.

CO N FLI C T S O F I NTE R E S T
The authors declare that they have no conflicts of interest.

AUTH O R CO NTR I B UTI O N
AJ designed the study, interpreted data, and wrote manuscript. MK interpreted data and gave expert comments on the manuscript.
TK conceptualized the study, interpreted data, and edited the manuscript.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author upon reasonable request.