Introduction
It is well known that dyslipidemia, with characteristic features such as high plasma triglyceride (TG) concentration, low high-density lipoprotein (HDL) cholesterol concentration and increased concentration of small dense low-density lipoprotein (LDL) cholesterol particles, is associated with insulin resistance, and is one of the major risk factors for cardiovascular disease in patients with diabetes mellitus [
1]. The pathogenesis of diabetes dyslipidemia is not precisely known. However, studies suggest that insulin resistance plays a pivotal role in the development of this condition [
2–
5].
Previous studies have proved that both men and women with diabetes had an increased prevalence of hypertrigyceridemia and low HDL cholesterol (HDL-C) levels, but no significant differences of their total cholesterol (TC) from non-diabetics were found [
6,
7]. LDL cholesterol (LDL-C) levels are higher in women with diabetes than in non-diabetic women, as shown in the UK Prospective Diabetes Study (UKPDS) [
6], however, in the Framingham Heart study, LDL-C levels do not differ from diabetic to non-diabetic counterparts [
7]. Studies have also revealed that the individuals without diabetes but with higher levels of TG, TC, LDL-C and lower levels of HDL-C are more likely to develop diabetes in the future. This trend is more evident in women than in men [
8,
9].
Association of dyslipidemia with high levels of fasting plasma glucose (FPG) in type 2 diabetes mellitus is ascribed to free fatty acid release secondary to insulin resistance [
1–
5]. Among non-diabetes, however, whether this association still exists, and whether dyslipidemia is an independent risk factor for high FPG levels are not clear. Few studies have examined the relation of plasma lipid to FPG in non-diabetics. Therefore, this study was intended to evaluate the association between plasma lipids and FPG in non-diabetic urban Chinese.
Methods
Study population
The study samples were selected from a health survey in Yinze District in Taiyuan, capital city of Shanxi Province in China, from April to October 2008. Three communities were selected, and the residents were invited. In total, 4798 subjects attended the survey. The exclusion criteria of the study were: (1) individuals with missing or poorly recorded information, (2) subjects with less than 35 or more than 75 years of age. Finally, complete data were obtained from 4073 subjects aged 35–75 years (1245 men, and 2828 women).
In this cross-section study, we focused on the relationship between plasma lipids and FPG in non-diabetics. Thus, 3460 non-diabetic subjects (1027 men and 2433 women) were finally recruited in this study after removing 613 patients with diabetes mellitus. Diabetes mellitus was defined as having a FPG greater than or equal to 7.0 mmol/L or being on treatment for diabetes. Impaired fasting glucose (IFG) was defined as having a FPG≥6.1 mmol/L and<7.0 mmol/L [
10]. Men and women were classified into tertiles of increasing HDL-C, LDL-C, TC, or TG, respectively.
Physical examinations
A standard questionnaire assessing demographic information, clinical history, and lifestyle was completed by trained research staff. Participants underwent routine physical examinations including the following measurements: height, weight, waist circumference, hip circumference, and resting blood pressure.
Height was measured in meters (without shoes), and weight in kilograms (without heavy clothing). Body mass index (BMI) was calculated as weight in kilograms divided by the square of height in meters. We measured waist circumference on standing subjects with a soft tape at the midpoint between the bottom of rip cage and the top of lateral border of iliac crest during minimal respiration. The measurement of blood pressure was based on the criteria of JNC7 [
11], two measurements were obtained in a sitting position after a 30-min rest, the measurements were taken in 5-min intervals, and mean values were calculated.
Laboratory analyses
Overnight fasting venous blood specimens were collected. FPG was measured by using a modified hexokinase enzymatic method. TC, TG, HDL-C, and LDL-C were measured using reagents (Kyowa Medex Co., Ltd., Tokyo, Japan) and were performed in Unicel Dxc 800 (Beckman Coulter, Fullerton, CA, USA). All laboratory assays were completed without knowledge of participants’ medical histories. Lipids and FPG concentrations were reported as mmol/L.
Analytical variable specification
Smoking/drinking status, physical exercise and education status were defined using information obtained from the questionnaire. The categories of smoking/drinking were never, former, and current smoking/drinking. An ever smoker was defined as one who had smoked at least 1 cigarette daily for 1 year or more. And an ever drinker was defined as one who had averagely drank at least 10 g alcohol, 100 g wine or 1 bottle of beer per week for 1 year or more. Current smokers/drinkers were ever smokers/drinkers who were still smoking/drinking at the time of the interview, and former smokers/drinkers were defined as those who stopped smoking/drinking for at least 1 year. Physical exercise was defined as seldom (<1 time per week), occasional (1–4 times per week), and always (≥5 times per week) with more than half an hour each time. Categories of education status included≤Primary school,<Bachelor degree, and≥Bachelor degree.
Statistical analyses
Data were entered by using Microsoft Office Excel (Microsoft Corp., Redmond, Washington, USA). Categorical variables such as education, smoking, drinking status, and physical activity were given as frequency (%). Continuous variables such as FPG, blood pressure, and plasma lipids were given as means±SD. The two side t tests and Pearson’s Chi-square tests were used. Univariate variance analyses were used for comparison of FPG among different categories of lipid levels in men or in women. Multivariable linear regression analysis was performed to test for independent predictors of HDL-C, LDL-C, TC, and TG in non-diabetic women. The model included age, BMI, waist/hip ratio, education, smoking, alcohol drinking, and physical exercise. Logistic regression was used to estimate the adjusted odds ratios (ORs) and 95% confidence interval (CI) of impaired fasting glucose (IFG). We adjusted potential confounders including age, BMI, waist/hip ratio, education, smoking, alcohol drinking, and physical exercise. Statistical analyses were performed using SPSS (version 15.0, SPSS Inc., Chicago, IL) software. All reported P values are two tailed.
Results
Gender-specific general characteristics of the 3460 non-diabetic subjects (1027 men and 2433 women) categorized by HDL-C level were shown in Table 1. We found a clear decreasing trend in levels of SBP, DBP, BMI, waist hip ratio, FPG, uric acid, LDL-C, TC, and TG from women with low HDL-C to those with higher HDL-C. The same trend was found in men except SBP, DBP, and FPG. A general decrease in the prevalence of IFG with increased concentrations of HDL-C was found in women (12.3%, 8.5%, and 5.5% for successive tertiles).
Adjusted mean FPG concentrations according to tertiles of increasing plasma lipids are listed in Table 2. In women, a stepwise increase in HDL-C was associated with decreasing FPG levels (lowest tertiles, FPG: 5.376±0.018; middle tertiles, 5.324±0.018; highest tertiles, 5.276±0.018 mmol/L; P = 0.001). Reversely, an increasing trend in FPG levels from lowest tertiles to highest tertiles of LDL-C, TC, and TG was found. For example, adjusted means of FPG from lowest to highest tertiles of TG were 5.266±0.018, 5.304±0.018, and 5.405±0.018 mmol/L, respectively (other data not shown here). In men, no significant relationship between plasma lipids and FPG levels was detected (Table 2).
Linear regression analysis was used to identify whether the relationship of FPG with HDL-C, LDL-C, TC, and TG in women non-diabetic subjects was independent or not (Table 3). After multivariable adjustment HDL-C (β = -0.074, P = 0.006), LDL-C (β = 0.119, P = 0.029), and TG (β = 0.070, P = 0.011) were significant and independent associated with FPG. Furthermore, age, BMI, and waist/hip ratio (WHR) were also associated with FPG.
We further evaluated the relative risk of IFG in relation to varying concentration of plasma lipids in men and women, respectively (Table 4). After adjusting for confounding by age, BMI, WHR, smoking, drinking, education, and physical exercise, we found that women in the first tertile with lower HDL-C level had a 1.75-fold increase in risk of IFG compared with women in the third tertile with higher HDL-C level (OR: 1.75; 95% CI: 1.20–2.56). Women in the third tertile with higher TG level had higher risk for IFG as compared with women in the first tertile with low TG level (OR: 1.61; 95% CI: 1.09–2.36).
Discussion
The data presented here showed that, in non-diabetic women, increased levels of fasting glucose and high prevalence of IFG were significantly and independently associated with low HDL-C levels and increased levels of LDL-C, TC and TG (features of dyslipidemia). We also noted that the risk of IFG was negatively associated with HDL-C and positively associated with TG in non-diabetic women. Nevertheless, no significant association was found between dyslipidemia and FPG in non-diabetic men.
Blood lipid and lipoprotein levels are known to be important predictors for development of type 2 diabetes mellitus [
8,
9]. The main cause of diabetic dyslipidemia is well acknowledged to be secondary to insulin resistance [
1–
5]. Demonstration that dyslipidemia was associated with high levels of FPG in non-diabetic women in this study was noteworthy. We attempt to explain the association by the following three reasons. First, in the prediabetic phase, insulin resistance is already in place [
12,
13]. Second, dyslipidemia may inflect insulin secretion or may result in insulin resistance. The association of hyperinsulinemia with low HDL-C levels has already been evaluated [
14,
15]. It was also found that reconstituted HDL (rHDL) induced elevation in plasma insulin [
16]. Third, dyslipidemia and dysglycemia may have some common mediators. For example, tumor necrosis factor (TNF) which is implicated in obesity-related diabetes mellitus can lower HDL-C levels [
17,
18]. The evidence in our study strongly showed that dyslipidemia may be a very early sign of a failing pancreas before the dysglycemia reaches the definition of diabetes.
No significant association among men was found. It may be due to different lifestyle factors such as alcohol intake, cigarette smoking, and physical exercise. More men than women have alcohol and smoking habits. Drinking and smoking can significantly affect the lipids levels and make the association of dyslipidemia with dysglycemia more complicated [
19–
22].
A number of limitations of the present study should be noted. First, the oral glucose tolerance test (OGTT) and fasting plasma insulin were not measured due to the size of the population (4798 subjects enrolled). We could not exclude latent diabetic subjects who may have a normal FPG level but have 2-h postprandial glucose>11.1 mmol/L, and also could not diagnose subjects as impaired glucose tolerance (IGT) [
10]. Insulin resistance could not be evaluated in these non-diabetic subjects. Because OGGT was not performed, we considered the non-diabetic as non-impaired glucose tolerance subjects. There are few studies regarding the association between postprandial glucose and lipid profiles in non-diabetic subjects. Only two studies demonstrated that dietary strategies and β-glucans can improve both serum lipids and postprandial glucose [
23,
24]. Thereby this limitation cannot deny the association between dyslipidemia and high level of FPG we found in Chinese non-diabetic women. Second, the study is limited by the cross-sectional design, which cannot establish causality and temporal relationship between dyslipidemia and evaluated FPG in non-diabetes. However, the temporal association of dyslipidemia and type 2 diabetes has been well established in previous studies [
1–
5]. It is well accepted that hyperlipidemia is not only the disorder of glucose metabolism but also an early event that contributes to the development of type 2 diabetes. Third, our study has a higher proportion of women than men. It may influence the conclusion drawn from men. This may be due to the fact that women were more health conscious and assertive in attending the health survey than men. In addition, male residents in China often need to work hard to raise the family, so relatively few male have free time to attend the health survey. Fourth, plasma lipids and FPG levels were measured only once, which might have led to random errors.
In summary, three conclusions were drawn from Chinese non-diabetic women: (1) dyslipidemia is associated with high levels of FPG; (2) TG, HDL-C, and LDL-C are predictors of IFG independent of BMI and waist/hip ratio; (3) low HDL-C and high TG are risk factors of IFG. No significant association was detected in men. Our findings implied that the linkage between dyslipidemia and dysglycemia may be not just the insulin resistance. Dysglycemia may be secondary to dyslipidemia or dyslipidemia and dysglycemia may have some common pathogenesis, such as inflammation [
12,
13]. These hypotheses need to be studied in prospective settings to evaluate association of dyslipidemia to FPG among subjects without insulin resistance.
Higher Education Press and Springer-Verlag Berlin Heidelberg
PDF
(108KB)
