Individuals with high total cholesterol/hdl cholesterol ratios are insulin resistant

Transcription

1 Journal of Internal Medicine 1998; 243: Individuals with high total cholesterol/hdl cholesterol ratios are insulin resistant J. JEPPESEN, F. S. FACCHINI, & G. M. REAVEN From The Department of Medicine, Stanford University School of Medicine, Stanford, and Shaman Pharmaceuticals Inc., South San Francisco, CA, USA Abstract. Jeppesen J, Facchini FS, Reaven GM. (The Department of Medicine, Stanford University School of Medicine, Stanford, and Shaman Pharmaceuticals Inc., South San Francisco, CA, USA). Individuals with high total cholesterol/hdl cholesterol ratios are insulin resistant. Journal of Internal Medicine 1998; 243: Objectives. To define the pathophysiologic characteristics of patients at high risk for coronary heart disease due to an increased ratio of total cholesterol (TC) to high density lipoprotein-cholesterol (HDL-C). Design. Cross-sectional. Setting. Clinical Research Center. Subjects. One hundred-20 healthy, non-diabetic, normotensive, volunteers were screened for this study. From this pool, 40 individuals (20 females and 20 males) with the highest and the lowest TC/HDL-C ratios were selected for comparison. Main Outcome Measures. Values for body mass index (BMI), ratio of waist to hip girth (WHR), and blood pressure were obtained on all patients. In addition, measurements were made of fasting lipid and lipoprotein concentrations, plasma glucose and insulin responses to an oral glucose challenge, and insulin resistance as assessed by the insulin suppression test. Results. Age, BMI, and WHR were the same in the two groups. However, the group with a high TC/HDL- C ratio had higher (P < 0.05) systolic and diastolic blood pressures. In addition, patients with a high TC/HDL-C ratio had significantly higher (P < 0.001) very low density (VLDL) and low density lipoprotein (LDL)-cholesterol concentrations and lower HDLcholesterol concentrations, with significant (P < 0.001) correlations between the TC/HDL-C ratio and VLDL (r = 0.60), LDL (r = 0.54), and HDL (r = 0.73) cholesterol concentrations. Patients with a high TC/HDL-C ratio were also significantly (P < ) more insulin resistant, glucose intolerant with a greater plasma insulin response to oral glucose, and hypertriglyceridemic. Conclusions. The results indicate that an increase in LDL-cholesterol concentration is not necessarily the major contributor to a high ratio of TC/HDL-C. Furthermore, individuals with this epidemiologic designation are insulin resistant, and liable to all the other abnormalities associated with this metaboic defect. Keywords: cholesterol, cholesterol-hdl ratio, HDL cholesterol, insulin resistance, LDL cholesterol. Introduction The conclusion of a recent analysis of three large epidemiologic data sets [1], involving approximately 8000 individuals, was that the ratio of total/high density lipoprotein (HDL)-cholesterol concentration (TC/HDL- C) was a better predictor of subsequent coronary heart disease (CHD) than either total or low density lipoprotein (LDL)-cholesterol by themselves. Although this information is of substantial value in predicting who is at risk for CHD, it leaves unanswered the pathophysiological features that characterize these individuals, as well as implying that the ratio is a function of two related variables, cholesterol and HDL-cholesterol concentration. This is obviously not true. For example, increases in low density lipoprotein (LDL)-cholesterol are not associated with insulin resistance and/or increases in circulating insulin concentration [2]. In contrast, it appears that a low HDL-cholesterol concentration is independently associated with resistance to insulin-mediated glucose disposal and compensatory hyperinsulinaemia [3]. All three of these changes, insulin resistance [4, 5] hyperinsulinaemia [6], and a low HDL-cholesterol concentration [7] have been iden Blackwell Science Ltd 293

2 294 J. JEPPESEN et al. tified as risk factors for CHD per se as well as belonging to a cluster of other abnormalities associated with increased risk of CHD [8]. Thus, although the epidemiologic significance of a high TC/HDL-C ratio is relatively clear, there is not a great deal of information concerning the metabolic implications of this designation. The study to be described was initiated to address this latter issue. Methods The study population consisted of 120 healthy volunteers, defined as healthy on the basis of an unremarkable medical history, physical examination, a body mass index < 30 kg m 2 for females and < 32 kg m 2 for males, blood pressure < 160/90 mm Hg, normal values for routine urinalysis, haemogram, and blood chemistry, a normal electrocardiogram, and a non-diabetic 75-g oral glucose tolerance test [9]. None of the patients were taking any medication known to affect carbohydrate or lipid metabolism. The patients abstained from alcohol and smoking and performed no heavy exercise for 24 h before being tested. The study was approved by Stanford University Human Subjects Committee, and each subject gave written, informed consent before being admitted to the General Clinical Research Center of Stanford Medical Center. In order to obtain two groups with widely divergent ratios of TC/HDL-C cholesterol, the 40 individuals (20 males and 20 females) with the highest and lowest ratios were identified. The baseline characteristic of the two groups thus formed are shown in Table 1, and it can be seen that they were also similar in terms of age, body mass index (BMI), and ratio of waist to hip girth (WHR). However, both systolic and diastolic blood pressures were somewhat higher in those with the Table 1 Baseline characteristics of the two study groups (mean ± SE) Low High Variable TC/HDL-C* TC/HDL-C* (2.9 ± 0.1) (4.6 ± 0.1) p Gender (female/male) 20/20 20/20 NS Age ( years) 46 ± 2 47 ± 2 NS Body mass index (kg/m 2 ) 24.3 ± ± 0.3 NS Waist/hip ratio (WHR) 0.87 ± ± 0.01 NS Systolic BP (mm Hg) 115 ± ± 3 < 0.05 Diastolic BP (mm Hg) 72 ± 1 76 ± 1 < 0.05 * TC = total cholesterol. HDL-C = high density lipoproteincholesterol. highest TC/HDL-C ratio. By selection, there was no overlap between the ratios of the two groups. The following measurements were made on all patients. All studies were initiated after a 14 h fast, and performed in random order. 1 Plasma glucose [10] and insulin [11] concentrations were determined before and 30, 60, 120, and 180 min after a 75 g oral glucose load. The area under the resultant concentrations curves was calculated by the trapezoidal method and defined as the glucose and insulin responses to oral glucose. 2 Resistance to insulin-mediated glucose disposal was determined by a modification of the insulin suppression test [12]. This technique has been described in detail elsewhere [13]. Briefly, each subject received a continuous intravenous infusion of somatostatin (5 mg min 1 ), insulin (25 mu/m 2 /min), and glucose (240 mg m 2 min 1 ) via an indwelling teflon catheter in a superficial antecubital vein. Venous blood samples were obtained from a similar catheter inserted in a contralateral antecubital vein kept open with a 0.9% NaCl infusion containing 20 meq L 1 KCL. The continuous infusion was given for 180 min, and blood was obtained before and 30, 60, 90, 120, 150, 160, 170, and 180 min after starting the infusion for measurement of plasma glucose and insulin. The mean value of the four measurements made during the last 30 min of the test were used to calculate the steady-state plasma insulin (SSPI) and steady-state plasma glucose (SSPG) concentration. Given the similarity of the SSPI levels in all individuals, the SSPG concentration provides a measure of insulin-mediated glucose disposal, i.e. the higher the SSPG, the more insulin resistant the research subject. 3 Before each of the tests described above, blood was obtained in EDTA tubes for measurement of fasting plasma triglyceride (TG), cholesterol, and lipoprotein TG and cholesterol concentrations [14, 15]. These samples were subjected to sequential density ultracentrifugation [16] to separate very low-density lipoprotein (VLDL), intermediate low-density lipoprotein (IDL), low-density lipoprotein (LDL), and high-density lipoprotein (HDL) fractions at densities below 1.006, and at and mg dl 1, respectively, and the TG and cholesterol contents of the various fractions determined. The results are expressed as the mean ± SEM and were analysed by STATVIEW Ô. Non-normally distributed variables: fasting plasma TG, VLDL-TG, VLDL-cholesterol, HDL-cholesterol, SSPG, and insulin response to oral glucose, were normalized

3 CHOLESTEROL LEVELS & INSULIN RESISTANCE 295 by logarithmic transformation. For sake of simplicity only non-normalized data are presented. Student s t- test were used to compare mean values in the two groups, and Spearman s rank correlation coefficients were calculated to identify relationships between variables of interest. Finally, multivariate analysis was performed to evaluate the contributions of selected variables to the ratio of TC/HDL-C. Results Plasma and lipoprotein cholesterol concentrations of the two experimental groups are listed in Table 2. Patients with the high ratio of TC/HDL-C had significantly higher concentrations of plasma (P < 0.001), VLDL (P < 0.001), LDL (P < 0.001), and lower (P < 0.001) HDL-cholesterol concentrations. Although all of these differences were statistically significant, the relative magnitude of differences in VLDL-cholesterol was approximately four-times greater (135%) than that seen in the other four variables (20 33%). It is also of note that the absolute increase in mean VLDL and LDL cholesterol concentrations were reasonably comparable, i.e vs mmol L 1, respectively. Since significant differences existed between the two experimental groups in terms of cholesterol and all lipoprotein cholesterol classes, an effort was made to see which of these variables were most closely correlated with the TC/HDL-C ratio in all 80 patients. Table 2 Plasma and lipoprotein cholesterol (Chol) concentrations in the two experimental groups (mean ± SEM) Variable Low High mmol L 1 TC/HDL-C* TC/HDL-C* p Plasma Chol 4.37 ± ± 0.10 < VLDL- Chol 0.34 ± ± 0.08 < LDL- Chol 2.30 ± ± 0.10 < HDL- Chol 1.53 ± ± 0.03 < * TC = total cholesterol. HDL-C = high density lipoproteincholesterol. Table 3 Spearman rank correlation coefficients (R) between total/hdl-cholesterol ratio and plasma and lipoproteins cholesterol concentrations in all patients Variable R p Plasma cholesterol 0.53 < VLDL-cholesterol 0.58 < LDL-cholesterol 0.53 < HDL-cholesterol 0.75 < The results of these calculations are shown in Table 3, and these relationships are depicted in Fig. 1. It is apparent that all of the cholesterol measurements were correlated with the TC/HDL-C ratio. To explore further differences between the characteristics of the two groups, Table 4 lists the changes in TG concentrations. It can be seen from these data that plasma, VLDL, and LDL-TG concentrations were higher (P < 0.001) in those with the high TC/HDL-C ratio. It is also apparent that the increase in VLDL-TG was responsible for essentially all of the increase in plasma TG concentration. The results of the measurements of the total glucose and insulin responses to the oral glucose challenge, and the steady-state plasma glucose r = 0.58 P < Ð VLDL cholesterol (mmol L Ð1 2.5 ) r = 0.53 P < LDL cholesterol (mmol L Ð ) r = 0.75 P < HDL cholesterol (mmol L Ð1 ) Fig. 1 The relationship between the ratio of TC/HDL-C and VLDL, LDL, and HDL cholesterol in the entire population.

4 296 J. JEPPESEN et al. Table 4 Plasma and lipoprotein triglyceride (TG) concentration in the two experimental groups (mean ± SEM) Variable Low High mmol L 1 TC/HDL-C* TC/HDL-C* p Plasma TG 0.82 ± ± 0.12 < VLDL-TG 0.45 ± ± 0.10 < LDL-TG 0.18 ± ± 0.01 < HDL-TG 0.10 ± ± 0.01 NS * TG = total cholesterol. HDL-C = high density lipoprotein cholesterol. Table 5 Total integrated glucose and insulin responses to oral glucose and insulin resistance (SSPG) in the two experimental groups (mean ± SEM) Variable Low High TC/HDL-C* TC/HDL-C* p Glucose response to oral glucose (mmol L 1 h 1 ) 16.6 ± ± 0.6 < 0.05 Insulin response to oral glucose (pmol L 1 h 1 ) 660 ± ± 96 < SSPG (mmol L 1 ) 5.2 ± ± 0.6 < * TC = total cholesterol. HDL-C = high-density lipoprotein cholesterol. concentration at the end of the infusion study are given in Table 5. The glucose response was modestly (approximately 10%), but significantly (P < 0.05) higher in those with the higher ratio of TC/HDL-C. The increase in the total integrated insulin response and the SSPG were increased to a proportionately greater degree in this group (approximately 50%), and the changes were statistically significant for both variables (P < ). Finally, in an attempt to evaluate the relative contribution of various factors to the TC/HDL-C ratio, multivariate analysis was performed. For this purpose we considered the impact of concentrations of LDL, VLDL, and HDL cholesterol, as well of that of the estimate of insulin resistance (SSPG) in the entire population. The results of this analysis is shown in Table 6. Model 1 in Table 6 presents the effect of differences in age, gender, body mass index, and waist/hip ratio on the variability of the TC/HDL-C ratio, and indicates that R2 = If LDL-cholesterol is added to the model (Model 2) the R2 value increases to 0.48, and the value for the standard B coefficient indicates that this addition significantly (P < 0.001) improved the model. The results in Table 6 also show that the addition of either VLDL (Model 3) or HDL (Model 4) cholesterol to the model increased the R2 value, with both measures of lipoprotein cholesterol making independent contributions to the TC/HDL-C ratio that are as strong as that of LDL cholesterol. Finally, when the separate contributions of SSPG and LDL cholesterol are evaluated (Model 5), it can be seen that the standard B coefficients of these two variables are comparable, and that the R2 value due to LDL-cholesterol alone increases from 0.48 to Discussion There seems to be considerable evidence that the ratio of TC/HDL-cholesterol is a powerful predictor of risk of CHD [1]. In the present instance, a ratio of 4.6 in the high TC/HDL-C group represents an approximate two-fold increase in risk for CHD as compared to the low TC/HDL-C patients. Although this information is clearly important to patient and physician, its utility is limited to the degree that the pathophysiological significance of the ratio can be understood and acted upon. In this context, the current results are of considerable interest. For example, it can be seen from Table 2 that patients in the high TC/HDL-C had an increment in VLDL-cholesterol that was Table 6 Multivariate relationship between the ratio of total cholesterol to HDL-cholesterol and LDL cholesterol, VLDL cholesterol, HDL cholesterol and steady-state plasma glucose (SSPG) concentration in the total population Variable Model 1 Model 2 Model 3 Model 4 Model 5 R 2 = 0.31 R 2 = 0.48 R 2 = 0.61 R 2 = 0.80 R 2 = 0.59 LDL cholesterol 0.46* 0.45* 0.43* 0.43* VLDL cholesterol 0.40* HDL cholesterol 0.64* SSPG 0.36* Model 1 presents the R 2 value due to differences in gender, age, body mass index, and waist to hip ratio, independent of differences in lipoprotein cholesterol concentration or SSPG. In the remaining four models adjustments have been made for these variables. The numbers in the Table are the standard B coefficients. * = P <

5 CHOLESTEROL LEVELS & INSULIN RESISTANCE 297 approximately as great as that in LDL-cholesterol. Furthermore, these individuals also had substantially lower HDL-cholesterol concentration. Indeed, changes in all these variables were significantly related to the TC/HDL-C ratio as seen in Table 3. The results shown in Table 6 provide a quantitative estimate of the importance that differences in the cholesterol concentration of the various lipoprotein classes play in determining the TC/HDL-C ratio. These data clearly indicate that LDL, VLDL, and HDL make comparable contributions to the TC/HDL-C ratio in the 120 normal volunteers studied. This observation is of more than casual interest, since the metabolic defects leading to an increase in LDL-cholesterol are certainly not those that would be associated with an increase in VLDL-cholesterol and a decrease in HDL-cholesterol. In addition to the ambiguity of not knowing from the ratio of TC/HDL-C the pathophysiological mechanism responsible for this phenomenon, it is also clear from the results presented that there are multiple changes seen in patients with a high TC/HDL-C ratio that may significantly contribute to their increased risk for CHD. More specifically, patients with a high TC/HDL-C ratio also had higher blood pressures (Table 1), increased TG concentrations (Table 4), and were significantly more insulin resistant, glucose intolerant, and hyperinsulinemic (Table 5). All of these latter changes have been identified as risk factors for CHD [4 8, 17, 18]. Thus, the use of a ratio that emphasizes measurement of cholesterol concentrations may actually identify patients at increased risk of CHD for reasons unrelated to cholesterol metabolism. Once the characteristics of patients identified as being at increased risk for CHD on the basis of their ratio of TC/HDL-C have been identified, it becomes clear that they share the cluster of abnormalities that were initially demonstrated to be secondary to resistance to insulin-mediated glucose disposal [2 5, 8, 20]. Indeed, as shown in Table 6, the SSPG concentration (the estimate of insulin resistance used in this study) was as significant a predictor of the TC/HDL-C ratio as any measure of lipoprotein cholesterol concentration. This is not meant to imply that insulin resistance, per se, is a predictor of CHD, but rather that the cluster of abnormalities associated with insulin resistance may be very important in this context. Initially [8], the abnormalities associated with insulin resistance were said to include some degree of glucose intolerance, hyperinsulinaemia, and dyslipidaemia characterized by increases in VLDL-cholesterol and TG and decreases in HDL-cholesterol, and a tendency to increased blood pressure. More recently, evidence has been summarized [19] showing that such individuals also have smaller and denser LDLparticles, enhanced postprandial lipaemia, and dysfibrinolysis, further accentuating their risk of CHD. It should be noted at this point that the marked differences in the TC/HDL-C ratio of the two experimental groups, and all of the other differences noted between them, were seen despite the fact that the two groups were essentially identical in terms of age, BMI, and WHR. Conclusion In conclusion, the results presented indicate that individuals at increased risk for CHD on the basis of their ratio of TC/HDL-C suffer from several additional metabolic defects that increase risk of CHD. As such, the data demonstrated suggests that it is inappropriate to conclude that the major reason for the epidemiologic relationship between an increased TC/HDL-C ratio and CHD is an increase in LDL-cholesterol concentration. Acknowledgements Supported by NIH Research Grants (HL-08506, RR ). References 1 Kinosian B, Glick H, Garland G. Cholesterol and coronary heart disease: predicting risks by levels and ratios. Am Intern Med 1994; 121: Sheu WH-H, Shieh S-M, Fuh MM-T, Shen D D-C, Jeng C-Y, Chen Y-DI, Reaven GM. Insulin resistance, glucose intolerance, and hyperinsulinemia. Hypertriglyceridemia versus hypercholesterolemia. Arterio and Thromb 1993; 13: Laws A, Reaven GM. Evidence for an independent relationship between insulin resistance and fasting plasma HDL-cholesterol, triglyceride and insulin concentrations. J Int Med 1992; 231: Shinozaki K, Suzuki M, Ikebuchi M, Hara Y, Harano Y. Demonstration of insulin resistance in coronary heart disease documented with angiography. Diabetes Care 1996; 19: Howard G, O Leary DH, Zaccaro D, Haffner S, Rewers M, Hamman R, et al., for the IRAS Investigators. Insulin sensitivity and atherosclerosis. Circulation 1996; 93: Pyörälä K. Relationship of glucose tolerance and plasma insulin to the incidence of coronary heart disease: results from two population studies in Finland. Diabetes Care 1979; 2:

6 298 J. JEPPESEN et al Bainton D, Miller NE, Bolton CH, Yarnell JWG, Sweetnam PM, Lewis B, Elwood PC. Plasma triglyceride and high density lipoprotein cholesterol as predictors of ischaemic heart disease in British men: the Caerphilly and Speedwell Collaborative Heart Disease Studies. Br Heart J 1992; 68: Reaven GM. Role of insulin resistance in human disease. Diabetes 1988; 37: National Diabetes Data Group. Classification and diagnosis of diabetes mellitus and other categories of glucose intolerance. Diabetes 1979; 28: Kadish AH, Little RL, Sternberg JC. A new and rapid method for the determination of glucose by measurement of rate of oxygen consumption. Clin Chem 1968; 14: Hales CN, Randle PJ. Immunoassay of insulin and insulinantibody precipitate. Biochem J 1963; 88: Greenfield MD, Doberne L, Kraemer FB, Tobey TA, Reaven GM. Assessment of insulin resistance with the insulin suppression test and the euglycemic clamp. Diabetes 1981; 30: Shen DC, Shieh SM, Fuh MT, Wu DA, Chen YD, Reaven GM. Resistance to insulin-stimulated-glucose uptake in patients with hypertension. J Clin Endocrinol Metab 1988; 66: Wahlefield AW. Triglycerides: determination after enzymatic hydrolysis. In: Bergmeyer HV, ed Methods of Enzymatic Analysis. New York: Academic Press 1974; Allain CC, Poon LS, Chan CG, Rischmond W, Fu PC. Enzymatic determination of total cholesterol. Clin Chem 1974; 20: Havel RJ, Eder HA, Bragdon JH. The distribution of ultracentrifugally separated lipoproteins in human serum. J Clin Invest 1955; 34: Collins R, Peto R, MacMahon S, Herbert P, Fiebach NH, Eberlein J, et al. Blood pressure, stroke, and coronary heart disease. Pt. 2: Short-term reductions in blood pressure: Overview of randomized drug trials in their epidemiological context. Lancet 1990; 335: Austin MA. Plasma triglyceride and coronary heart disease. Arterioscler Thromb 1991; 11: Vaccaro O, Ruth KJ, Stamler J. Relationship of postload plasma glucose to mortality with 19-yr follow-up. Diabetes Care 1992; 13: Reaven GM. Pathophysiology of insulin resistance in human disease. Physiol Rev 1995; 75: Received 22 April 1997; accepted 6 November Correspondence: Dr Gerald M. Reaven MD, Shaman Pharmaceuticals Inc., 213 East Grand Avenue, South San Francisco, CA , USA (fax: ; greaven@shaman.com).