Diabetes is a syndrome characterized by high glycemic levels that promote significant pathophysiological metabolic changes [1]. Atherosclerotic Cardiovascular Disease (ASCVD), the main cause of morbidity and mortality in people with Diabetes Mellitus (DM), may promote or be favored by free cholesterol deposited into the vessels' walls. The high levels of glycemia and insulin resistance are associated with negative impacts on High-density Lipoprotein (HDL), reverse cholesterol transport, and antioxidant activity [2-5].

Insulin Resistance (IR) is an impaired biological response to insulin by tissue targets. IR promotes a lipid metabolism modulation resulting in dyslipidemia, characterized by increasing Triglycerides (TG), decreased High-density Lipoprotein Cholesterol (HDL-C), and synthesis of a larger proportion of small and dense Low-density Lipoprotein (LDL) particles (sdLDL) [3, 6]. These three conditions are independent risk factors for Cardiovascular Disease (CVD), with the risk being increased by two to fourfold in diabetic patients. Regarding the risk stratification, women are more susceptible than men (i.e., three-fold in diabetic women and twice in diabetic men), as shown by the Framingham Heart Study, and the risk is 7.5 times higher in diabetics than in non-diabetic women [3, 7, 8].

The negative correlation between HDL-C levels and the incidence of CVD is already well established and evidenced in different epidemiological studies [9]. The estimation of cardiovascular risk through the use of surrogate markers, as the ratio of carried molecules by lipoproteins, has demonstrated an adequate way to predict risk since isolated markers sometimes do not effectively contemplate it [10]. High plasma TG levels have been reported to be associated with the increased risk of coronary heart disease because of direct TG-rich lipoprotein particles’ atherogenic effect, especially on Intermediate-density Lipoproteins (IDLs). When high levels of TG are associated with low HDL-C, coronary atheromas are more extensive than those initiated exclusively by isolated high concentrations of Low-density Lipoprotein Cholesterol (LDL-C) [11]. Consequently, the TG/HDL-C ratio, an independent predictor of cardiovascular disease, has been considered to estimate cardiovascular risk because it is a non-invasive accurate means for coronary disease detection, predicting its presence and extension [12, 13].

The investigation and a better understanding of lipoprotein metabolism kinetics in pathological conditions, such as coronary artery disease, can contribute to the management of diabetic patients with atherosclerotic and cardiovascular unfavorable conditions. This article aimed to evaluate if the prevalence of a TG/HDL-C ratio above 2.5 and the association of insulin and HOMA-IR with TG/HDL-C, apoB, non-HDL, and free cholesterol correlations can be flags of coronary obstructions among other biomarkers in women with established coronary atherosclerosis regardless of diabetes.

A cross-sectional study was carried out on 42 female patients, 41 to 75 years old, with cardiac catheterization confirmed coronary atherosclerotic disease, who had been scheduled for vascular angioplasty or coronary by-pass surgery at the Ana Nery Hospital, SESAB/UFBA, Salvador, Bahia, Brazil. The participants were selected by convenience sampling and were only included in the study if they signed a free, prior, and informed consent to participate. This study was approved by the Research Ethics Committee of the Ana Nery Hospital, letter no. 83/11, CEAA number 71148023.3.0000.0045 (November 28th, 2011).

Blood samples of each patient were collected before the interventional procedure. The samples were analyzed at the Clinical Biochemistry Laboratory, Faculty of Pharmacy, Federal University of Bahia (UFBA). The following biomarkers were measured: Total Cholesterol (TC), HDL-C, non-HDL-C (n-HDL-C), LDL-C, insulin, TG, Paraoxonase-1 (PON1), apolipoprotein A (apoA), apolipoprotein B (apoB), glycated hemoglobin (HbA1c), glycemia, and Free Cholesterol (FC). The HOMA-IR index (homeostasis model assessment – insulin resistance) has been calculated using the following formula: HOMA-IR = [glycemia (mmol/dL) * insulin]/22.5; HDL-C/ApoA, TG/HDL-C ratio, and index (log TG/HDL-C) have also been obtained.

The patients were stratified into two groups: diabetics and non-diabetics. The LDL-C was calculated with Friedewald’s formula: LDL-C = cholesterol total – HDL-C – (TG/5) [14]. Body Mass Index (BMI) was calculated as body weight/(height)², and BMI classification was considered following the World Health Organization (WHO) criteria [15]. The following cutoffs for lipid parameters were used: (a) isolated hypercholesterolemia: serum LDL-cholesterol levels ≥ 160 mg/dL; (b) isolated hypertriglyceridemia: serum triglyceride levels ≥ 150 mg/dL; (c) isolated HDL-C low levels: HDL-cholesterol levels < 50 mg/dL (female level); (d) high TG/HDL-C ratio: TG/HDL-C > 2.5; and (f) HOMA-IR ≥ 2.5 [16-19]. TG/HDL-C values greater than 2.5 in women were considered a risk factor for cardiovascular disease [17].

To achieve a minimum test power of 80%, capable of identifying maximum absolute differences of 7.14 units between the medians, using a paraoxonase reference standard deviation of 10% (for high values compatible with CAD), 42 females were selected to participate in the study. The critical level chosen for significance was 0.05 (5%) for a 95% confidence interval (GraphPad StatMate 2.0).

The data descriptive analysis was performed by using centrality and dispersion estimates to obtain summary measures. After that, a normality test D’Agostino-Pearson was performed, followed by Pearson’s linear correlation to determine the association between insulin and the other biomarkers determined. The outlier detection was done by using Grubb’s test. To compare biochemistry parameters between diabetics and non-diabetic patients, the two-sample t-test was used, and a linear regression was performed to obtain possible associations among insulin and glycemia, TG, and HOMA-IR; and TG and glycemia and HOMA-IR. Also, a linear correlation was made of TG/HDL-C ratio and log TG/HDL-C index with insulin and HOMA-IR. Data analysis was considered significant when the obtained differences showed critical levels (p) less than 5% (p < 0.05) for a 95% confidence interval. For the statistical tests, the Stata 14.2 software (StataCorp. 2015. Stata Statistical Software: Release 14. College Station, TX: StataCorp LP), GraphPad InStat v.3.05, and GraphPad Prism 5.01 software (GraphPad Software, Inc., CA, USA) were used.

42 coronary atherosclerotic female patients with a mean age of 59.64 (± 8.05) were conveniently included in the study. 30 (71.43%) of the patients were diabetic (diabetes mellitus type 2), and 12 (28.57%) were non-diabetics. None of the patients showed low weight or class III obesity (BMI ≥ 40.0 kg/m2). Table 1 shows the patient's demographics and clinical and behavioral characteristics.

Table 2 shows the glycemia, insulin levels, and Insulin Resistance index (HOMA-IR) of the participating patients.

The general prevalence of isolated hypercholesterolemia, hypertriglyceridemia, and low HDL-C levels is shown in Table 3.

Insulin was found to be positively correlated with triglycerides, apoB, non-HDL cholesterol, and free cholesterol (non-esterified cholesterol) in diabetics (Table 4). Insulin was not significantly correlated with the other biomarkers tested in the diabetics and with all other biomarkers in the non-diabetics.

In a linear regression model with insulin as a dependent variable, the levels of glycemia, triglycerides, and HOMA-IR in diabetics were found to be associated (p < 0.001, R² = 0.9868); however, in non-diabetics, the association was only found between insulin and HOMA-IR (p = 0.002, R² = 0.9031). On the other hand, using triglycerides as a dependent variable, its association was observed only with HOMA-IR in both groups, i.e., diabetics (p = 0.006, R² = 0.2504) and non-diabetics (p = 0.014, R² = 0.4697). The model was selected based on the best coefficient of determination (R²).

ApoB: apolipoprotein B; n-HDL-C: Cholesterol non-HDL. r: Pearson correlation test, significant at p < 0.05 for a 95% confidence interval.

In this study, the relatively young age of patients with coronary diseases stood out, showing the large occurrence of premature atherosclerosis in young women. Premature atherosclerosis is defined as an ischemic cardiovascular event that occurs before 65 years of age in women and 55 years of age in men [18]. The women’s natural cardiovascular protection is conferred mostly by estrogen action. However, during the climacteric, a biological phase that starts around 40 years, the ovarian function decreases, reducing estrogen synthesis and influencing cardiovascular protection [19-21]. We noticed a significant percentage of women with previous episodes of acute myocardial infarction among diabetics, which not only corroborates this condition of premature atherosclerosis, but also shows evidence of a more advanced stage.

Other comorbidities, such as being overweight and obesity, also impact atherosclerosis prematurity. Approximately 70% of patients have a BMI over 25 kg/m², evidencing overweight and obesity as factors that increase the risk of cardiovascular disease. Obesity and diabetes mellitus, along with metabolic syndrome, are interrelated conditions, with many pathophysiological mechanisms that, in succession, lead to cardiovascular complications [22].

As all patients used simvastatin, the prevalence of isolated high LDL-C levels was low. However, triglyceride levels, which are not affected by that statin, were higher in both groups of patients in the study, being more prevalent in diabetic patients. The dyslipidemia profile was found appropriate for diabetic patients, and interestingly, these alterations were observed just before the onset of diabetes (pre-diabetes patients), being associated with insulin resistance [23].

The higher the insulin concentration, the greater the concentrations of lipoprotein biomarkers conferring cardiovascular risk. Deficiency of adequate cellular insulin signaling, due to absence or resistance, initiates a sequence of enzymatic abnormalities, involving the activity of the enzymes lipoprotein lipase and Cholesteryl Ester Transfer Protein (CETP) in diabetic patients. In the case of the lipoprotein lipase, the enlarged activity promotes an increase in the triglycerides levels [24, 25], and in the case of CETP, the biosynthesis of large and triglycerides-rich VLDL begins, resulting in a more atherogenic small and dense apoB richer LDL [26, 27]. Responsible for the exchange of esterified cholesterol and triglycerides from HDL to LDL in diabetic patients, CETP interferes with the metabolic pathways of free and esterified cholesterol [28].

Insulin Resistance score (HOMA-IR) is considered an independent risk factor of cardiovascular disease in diabetes patients [29]. Considering the cutoff of 2.5 [17], we observed diabetic patients to have a higher risk of developing another cardiovascular episode. Besides, Pivatto et al. [30] identified HOMA-IR as the main predictive factor in metabolic syndrome development. So, a more careful follow-up should be conducted for these patients.

Lipid ratios are better predictors of coronary artery disease compared to decreasing lipids evaluation, reflecting the interaction between atherogenic and protector lipid fractions [17, 31]. Particularly, the TG/HDL-C ratio is related to insulin resistance, atherogenesis, evidence of more dense and atherogenic LDL particles (type B lipid profile), arterial stiffness, and acute cardiovascular event in diabetes patients, being an important cardiovascular risk predictor [12, 28, 31, 32]. In our study, in both groups, we found an elevated prevalence of a high TG/HDL-C ratio with a median higher than 2.5, indicating a lipid profile more atherogenic inclusive in the non-diabetic group. The TG/HDL-C ratio evidenced a higher probability of other cardiovascular episodes.

TG/HDL-C ratio is also an indicator of Insulin Resistance (IR), the main risk factor for the development of diabetes mellitus [12, 24, 33]. So, the ratio may be used as a predictor of incident diabetes mellitus [34]. As we found an elevated TG/HDL-C ratio above the cut-off among non-diabetic patients, the risk of developing diabetes mellitus was higher.

The positive linear correlation between insulin and free cholesterol indicates possible metabolic errors in diabetic patients' lipid enzyme activity, which may increase cardiovascular metabolic risk [26]. In the normal cell signaling response to insulin, degradation of apoB is stimulated [27], although in cases of diabetes, with unsatisfactory insulin response, apoB degradation does not occur as expected. As apoB is constitutive of lipoproteins and inclusive of atherogenic fractions (non-HDL lipoproteins), the high levels are associated with the elevation of cardiovascular risk [30, 35, 36], which explains the obtained correlation between apoB and insulin. The slow apoB degradation is a result of the lipoprotein lipase’s decreased activity, which also stimulates VLDL synthesis, promoting an increase in atherogenic lipid fractions and cardiovascular risk [27].

In the diabetic panorama, insulin resistance also involves, as a consequence of the reduction in the ABCA1 activity, decreased HDL protection against atherosclerosis and modified Lecithin-Cholesterol Acyltransferase (LCAT) activity, which esterifies free cholesterol into HDL particles [28, 37, 38]. The alteration of all enzymatic and transfer protein activities observed in diabetic patients results in lipid metabolism and lipoprotein remodeling modifications originating in highly atherogenic lipoprotein particles [27].

Our study has evidenced an elevated prevalence of female patients with a high TG/HDL-C ratio and the association between insulin and HOMA-IR regardless of whether they are diabetic or not. The findings of a high prevalence of TG/HDL-C ratio higher than 2.5 and the association of insulin and HOMA-IR with apoB, non-HDL-C, and free cholesterol, should be better evaluated to be used as flags of female precocious coronary atherosclerosis independent of being diabetic. With the use of simvastatin, the prevalence of high levels of isolated LDL-C has been found to be low, with HDL functionality being improved. The use of statins has been found to limit lipoprotein modifications and cardiovascular risk scenarios. Future appropriately designed studies are needed to support the use of surrogate biomarkers in specific at-risk populations in order to achieve a positive clinical outcome.