Low Insulin Sensitivity Measured by Both Quantitative Insulin ...

0013-7227/02/$15.00/0 Printed in U.S.A.

The Journal of Clinical Endocrinology & Metabolism 87(11):5092–5097 Copyright © 2002 by The Endocrine Society doi: 10.1210/jc.2002-020703

Low Insulin Sensitivity Measured by Both Quantitative Insulin Sensitivity Check Index and Homeostasis Model Assessment Method as a Risk Factor of Increased IntimaMedia Thickness of the Carotid Artery ¨ IVA ¨ NSALO, OUTI PELKONEN, ILKKA SURAMO, ULLA RAJALA, MAURI LAAKSO, MARKKU PA ¨ NEN-KIUKAANNIEMI SIRKKA KEINA

AND

Departments of Public Health Science and General Practice (U.R., M.L., S.K.-K.) and Diagnostic Radiology (M.P., O.P., I.S.), University of Oulu; and Unit of General Practice (M.L., S.K.-K.), Oulu University Hospital, 90220 Oulu, Finland The present study evaluated the association of ultrasonographic manifestations of carotid atherosclerosis with glucose status, various components of the insulin resistance syndrome, and insulin sensitivity measured by a novel quantitative insulin sensitivity check index (QUICKI ⴝ 1/[log(I0) ⴙ log (G0)]). Carotid ultrasonographic measurements were performed on 54 diabetic subjects, 97 subjects with impaired glucose tolerance and 57 normoglycemic subjects. QUICKI and insulin resistance measured by a HOMA (homeostasis model assessment) method had a high negative correlation (r ⴝ ⴚ0.995, P < 0.001). QUICKI was lower in diabetic subjects (0.319 ⴞ 0.022) than in subjects with impaired glucose tolerance (0.334 ⴞ 0.027) or normoglycemia (0.335 ⴞ 0.022, P ⴝ 0.002). There was an increasing trend in the mean and maximal intima-media thickness (IMT) of the common carotid artery (CCA) with worsening of glucose status. The

H

IGH-RESOLUTION B-MODE ULTRASOUND imaging of the carotid artery has been shown to reflect histopathologically verified atherosclerosis (1, 2) and is, therefore, widely applied in studies of the occurrence and risk factors of early atherosclerosis (3, 4). Previous studies have shown that subjects with type 2 diabetes (5–10) and even those with impaired glucose tolerance (IGT) (7, 9, 10) may have greater carotid intima-media thickness (IMT) than nondiabetic subjects. Nevertheless, one recent Finnish study (11) failed to demonstrate any significant differences in the maximal intima-media thickness of the common carotid artery (CCA) between elderly subjects with diabetes, IGT, and normal glucose tolerance. Some prospective studies have shown that high fasting or postchallenge insulin levels have been predictive of coronary heart disease in men (12–16). The prospective epidemiological evidence of hyperinsulinemia increasing the incidence of coronary heart disease in women is inconsistent (12, 17). The inAbbreviations: BIF, Bifurcation; BMI, body mass index; CCA, common carotid artery; CI, confidence interval; CPIMT, combined plaqueintima-media thickness; DM, diabetes mellitus; HDL, high-density lipoprotein; HOMA, homeostasis model assessment; HOMA␤, HOMA method to estimate ␤-cell function; HOMA-IR, HOMA method to estimate insulin resistance; ICA, internal carotid artery; IGT, impaired glucose tolerance; IMT, intima-media thickness; NGT, normoglycemia; OGTT, oral glucose tolerance test; QUICKI, quantitative insulin sensitivity check index.

maximal IMT of the CCA correlated inversely with QUICKI (r ⴝ ⴚ0.158, P ⴝ 0.027). The prevalence of severe CCA atherosclerosis (maximal IMT of the CCA > 1.2 mm) was 41% in men and 16% in women (P < 0.001). It was also associated with a long (>26 yr) smoking history. The prevalence of severe CCA atherosclerosis was 11% in the highest QUICKI tertile, 36% in the middle tertile, and 33% in the lowest tertile (P ⴝ 0.002). Systolic blood pressure was higher and high-density lipoprotein cholesterol lower in subjects with severe CCA atherosclerosis, compared with those without it. In multiple regression analysis, the adjusted odds ratio for severe CCA atherosclerosis was 5.7 (95% confidence interval, 2.2–15.1) in subjects in the two lowest tertiles of QUICKI, compared with those in the highest tertile. (J Clin Endocrinol Metab 87: 5092–5097, 2002)

sulin resistance syndrome, a cluster of cardiovascular risk factors, has been suggested to be the cause of cardiovascular disease (18, 19). Some studies have shown hyperinsulinemia to be associated with an increased carotid IMT (6, 8, 9, 20). Nevertheless, up until the present, studies on this topic have been scarce, and a population-based study carried out in northern Finland did not find any positive association between insulin parameters and carotid IMT (21). The gold standard for measuring insulin resistance is the hyperinsulinemic euglycemic glucose clamp technique, but it is not easy to apply in epidemiological studies. Therefore, methods for estimating insulin sensitivity from fasting insulin and glucose values have been developed (22–25). The older of these methods, homeostasis model assessment (HOMA), allows also for an estimation of ␤-cell function (22). Recently, a quantitative insulin sensitivity check index (QUICKI) with a high correlation (r ⫽ 0.78) with insulin sensitivity from the glucose clamp technique has been developed (24, 25). QUICKI can be determined from fasting insulin and glucose values according to the equation: QUICKI ⫽ 1/[log(I0) ⫹ log (G0)], in which I0 is fasting insulin and G0 is fasting glucose. The present study evaluated the association of ultrasonographic manifestations of carotid atherosclerosis with glucose status, various components of the insulin resistance syndrome, and insulin sensitivity measured by QUICKI. The

5092

Rajala et al. • Low QUICKI and Increased IMT of the CCA

HOMA method was used in estimating insulin resistance (HOMA-IR) and ␤-cell function (HOMA␤). The correlation between insulin sensitivity measured by QUICKI and insulin resistance measured by HOMA was evaluated. Materials and Methods From 1990 –1992, a population-based study (26) was carried out in northern Finland to assess the prevalence of diabetes mellitus (DM) and IGT. All the 1008 subjects born in 1935 and living in Oulu, a city of 100,000 inhabitants, on October 1, 1990, were invited to participate in the study. Altogether 768 subjects attended the oral glucose tolerance tests (OGTTs), and the participation rate was 78%. In 1994 and 1996 –1998, two follow-up studies were carried out (27, 28). The subjects with IGT at the baseline were invited to participate in the first of the follow-up studies, and all participants of the 1990 –1992 study were invited to attend the second follow-up study. OGTTs were carried out at these follow-up visits. Questionnaires, interviews, clinical examinations, and laboratory tests were used to collect data during the follow-up study in 1996 –1998. Self-reported use of antihypertensive medication was recorded in the questionnaire, and 44% (91 of 208) of the present study subjects used some antihypertensive medication. ␤-blockers were the most common drugs. The use of diuretics, calcium channel blockers, and angiotensinconverting enzyme inhibitors was less frequent than that of ␤-blockers. Four measurements of blood pressure were made by the physician, using both upper limbs and both sitting and recumbent positions. The mean value of these four measurements was used in the analyses. Hypertension was defined as either a systolic blood pressure of 160 mm Hg or more or a diastolic blood pressure of 95 mm Hg or more or being on antihypertensive drug treatment regardless of the blood pressure values (29). Height and weight in light clothing were measured in the clinical examination, and the body mass index (BMI) (kilograms per square meter) was calculated. The waist to hip ratio (WHR) was a measure of minimal waist girth/maximal hip girth to the nearest centimeter. A standardized 75-g OGTT was performed according to the instructions of the World Health Organization Study Group (30). After a fast of 10 –12 h, a venous blood sample was drawn at 0800 – 1000 h to obtain a fasting glucose value and values for fasting immunoreactive insulin, total cholesterol, high-density lipoprotein (HDL) cholesterol and triglycerides. After that, a 75-g glucose load was given to the participants. At 2 h, a venous sample was collected for the 2-h glucose value. The concentration of blood glucose was determined with the hexokinase-glucose-6-phosphate dehydrogenase method (Merck Diagnostica, Darmstadt, Germany). The coefficients of variation for withinrun studies were between 1.7% and 2.8%, depending on the glucose concentration. For day-to-day studies, the coefficient of variation was 2.8% at the upper end of the reference range. To determine insulin resistance (HOMA-IR) and ␤-cell function (HOMA␤), fasting plasma glucose concentration is needed. To convert the whole blood glucose concentrations to plasma glucose concentration, we used the following equation: plasma glucose (mmol/liter) ⫽ 0.558 ⫹ 1.119* whole blood glucose (mmol/liter). The equation is based on 294 paired samples of whole blood and plasma glucose concentrations drawn in the Diabetes and Genetic Epidemiology Unit at the National Public Health Institute in Finland from a standard 75-g OGTT in 74 subjects (Qiao, Q., J. Lindstro¨ m, T. Valle, and J. Tuomilehto, manuscript submitted for publication). Serum immunoreactive insulin concentration was measured by RIA using the Phadeseph Insulin RIA100 kit (Pharmacia Diagnostics AB, Uppsala, Sweden), which also detects proinsulin and proinsulin conversion products with considerable sensitivity. The cross-reactivity of proinsulin in this assay is about 41%. Insulin levels were not analyzed from the samples of the diabetic patients with insulin treatment. The urinary albumin and creatinine concentrations were measured from an overnight spot urine sample. The highest decile of the urinary albumin to creatinine ratio (ⱖ2.5 mg/mmol in men and ⱖ3.0 mg/mmol in women) was used as a measure of microalbuminuria. Subjects with an albumin to creatinine ratio of more than 25 mg/mmol were considered macroalbuminuric and excluded from the analyses. All participants of the 1990 –1992 study, with the exception of diabetic

J Clin Endocrinol Metab, November 2002, 87(11):5092–5097 5093

subjects on antidiabetic medication, were invited to participate in the OGTTs during the follow-up study in 1996 –1998, and 244 men (81% of the original study participants alive and living in Oulu during the follow-up study) and 347 (83%) women attended it. Clinical diabetes was defined as either diabetes diagnosed by a physician before the baseline or as two elevated blood glucose values (either fasting blood glucose values of ⱖ6.1 mmol/liter or 2-h OGTT values of ⱖ11.1 mmol/ liter) during this study in 1992–1998. Sixty subjects had clinical diabetes and 31 subjects one elevated blood glucose value (either a fasting value of ⱖ6.1 or a 2-h OGTT value of ⱖ11.1 mmol/liter once). On the basis of the 1996 –1998 OGTT results, 102 subjects were classified as having IGT and 404 as having normoglycemia (NGT). Here, we report the associations of carotid intima-media thickness measured by ultrasound with variables measured during the follow-up study in 1996 –1998, emphasizing glucose status, various components of the insulin resistance syndrome and insulin sensitivity measured by QUICKI. The HOMA method was used in estimating insulin resistance and ␤-cell function. The correlation between insulin sensitivity measured by QUICKI and insulin resistance measured by HOMA was evaluated. All subjects with clinical diabetes (except one with cancer) and those with IGT were invited to participate in the present study. In addition, for each diabetic subject, a control subject with normoglycemic test results in OGTTs in 1992 and 1996 –1998 was invited to participate. The diabetic and normoglycemic groups were matched for gender, current smoking status, and BMI class.

Carotid ultrasonography Carotid ultrasound examinations were carried out between February 15 and September 29, 2000, at the Department of Diagnostic Radiology, Oulu University Hospital. The carotid ultrasound examination was carried out using a color Doppler ultrasound system (Toshiba PowerVision 7000 or 8000) with a scanning frequency of 9 MHz in B-mode, following the same protocol throughout, by a single trained radiologist (M.P.) blinded to the glucose status of the participant. Each carotid system was imaged in anterior oblique and lateral planes, transversally and longitudinally, with the subject supine and the head turned away from the examiner at an angle of 45 degrees. The examiner consistently aimed at the clearest possible image of the near and far walls of the carotid arteries. The scan head was kept perpendicular to the arterial walls and the transducer usually angled laterally to give the optimal visualization. Doppler scanning was used to identify the vessels and evaluate flow disturbances. Each scan of the common carotid artery began just above the clavicle and was moved cephalically through the bifurcation (BIF) and along both the internal and external branches as far distally as possible. The whole scanning procedure was recorded on a SuperVHS videocassette recorder (Panasonic, Osaka, Japan). All measurements were made about 6 months later from the video image on the monitor of the ultrasound device, using its electronic calipers. The dimensions measured were the IMT and the size and number of atheromatous plaques. The IMT, defined as the distance between the medial-adventitial interface and the luminal-intimal interface, was measured on each side from the CCA, BIF, and internal carotid artery (ICA) at the point at which each seemed thickest, avoiding sites with atheromatous plaque. In the case of atheromatous plaque, the combined plaque-IMT (CPIMT) was measured. If there was no plaque at the site of measurement, the IMT value alone was used for CPIMT. The IMT was measured with the instrument’s electronic calipers to the nearest 0.1 mm. The following measurements were used in the analyses: mean IMT CCA, maximal IMT CCA, mean IMT CCA/BIF/ICA, maximal IMT CCA/BIF/ICA, maximal CPIMT CCA/BIF/ICA, and the number of atheromatous plaques. An arbitrary cutoff point (ⱖ1.2 mm) of maximal IMT CCA was chosen in such a way that the 27% of the population with the highest values was selected and defined as having severe CCA atherosclerosis. Fifty-four (92%) of the subjects with clinical diabetes, 97 (95%) of those with IGT, and 57 (97%) of the normoglycemic subjects underwent the ultrasound measurements.

Statistical methods The differences between the group means of the continuous variables were tested for significance by ANOVA and Student’s t, Mann-Whitney U, and Kruskall-Wallis tests when appropriate. The dichotomized vari-

5094

J Clin Endocrinol Metab, November 2002, 87(11):5092–5097


ables were compared with the ␹2 test. Pearson’s correlation coefficients were calculated among the variables of interest. Multiple logistic regression analyses with severe CCA atherosclerosis as a dependent variable were made. In multiple logistic regression analyses, the log likelihood statistic was used in model building. The possible interaction of QUICKI with glucose status was tested by an interaction term. Variables with skewed distribution (insulin, triglycerides, HOMA-IR, HOMA␤, and albumin to creatinine ratio) were analyzed after logarithmic transformation. P values less or equal to 0.05 were considered statistically significant. All the statistical analyses were made with the SPSS for Windows software (version 10.0; SPSS, Inc., Chicago, IL).

Results

The glucose status characteristics of the study population are shown in Table 1. The proportion of women was higher among the subjects with IGT than among the subjects with DM or NGT. Systolic and diastolic blood pressure and the prevalence of hypertension increased with deterioration of glucose status. BMI and the prevalence of central obesity were highest in subjects with DM, but there was no increasing trend with the degree of glucose status. Fasting immunoreactive insulin values did not differ statistically significantly among the groups, but diabetic subjects had lower QUICKI than the other groups. HOMA-IR was higher in subjects with DM than in subjects with NGT (P ⫽ 0.002) and in subjects with IGT (P ⫽ 0.004). But there was no statistically significant difference in HOMA-IR (P ⫽ 0.729) of subjects with NGT and IGT. HOMA␤ was lower in subjects with DM, compared with subjects with IGT (P ⫽ 0.002) and NGT (P ⫽ 0.012). But there was no statistically significant difference in HOMA␤ (P ⫽ 0.692) of subjects with NGT and IGT. The prevalence of microalbuminuria tended to be higher in subjects with DM than in the other groups. QUICKI and HOMA-IR had a high negative correlation (r ⫽ ⫺0.995, P ⬍ 0.001). When QUICKI and HOMA-IR were divided into tertiles, almost the same subjects were included

in the best, middle, and worst tertiles. The only exception was one subject in the middle QUICKI tertile, who was included in the worst HOMA-IR tertile. Mean and maximal IMT of the CCA, mean and maximal IMT of the CCA/BIF/ICA, and maximal CPIMT according to glucose status are presented in Table 2. There was an increasing trend in the mean and maximal IMT of the CCA with worsening of glucose status, although the differences were not statistically significant. The correlations between the maximal IMT of the CCA and the background variables are presented in Table 3. Carotid IMT showed a statistically significant correlation with HOMA-IR, WHR, and systolic and diastolic blood pressure and inversely with HDL cholesterol and QUICKI. The correlation with triglycerides was of borderline significance. Fasting insulin tended to correlate with maximal IMT of the CCA. The mean of maximal IMT CCA was 1.16 mm ⫾ 0.42 in men and 0.95 mm ⫾ 0.32 in women (P ⬍ 0.001). The subjects with a long (ⱖ26 yr) smoking history had greater maximal IMT CCA than the subjects whose smoking history was 25 yr or shorter or who had never smoked (1.24 mm ⫾ 0.50 vs. 1.00 mm ⫾ 0.28 and 1.01 mm ⫾ 0.37, respectively, P ⫽ 0.013). The subjects in the highest QUICKI tertile tended to have lower maximal IMT CCA, compared with the subjects in the middle and lowest QUICKI tertiles (0.96 mm ⫾ 0.29 vs. 1.09 mm ⫾ 0.36 and 1.09 mm ⫾ 0.48, respectively; P ⫽ 0.074). The prevalence of severe CCA atherosclerosis was higher in men (41%) than in women (16%) (P ⬍ 0.001). Its prevalence was 50% in subjects with a long (ⱖ26 yr) smoking history, 22% in subjects who had smoked for 1–25 yr, and 24% in subjects who had never smoked (P ⫽ 0.004). It was 11% in the highest QUICKI tertile, 36% in the middle tertile, and 33% in the lowest tertile (P ⫽ 0.002). Instead, severe CCA atherosclerosis was not associated with glu-

TABLE 1. Characteristics of the study population in relation to the glucose status

n Women (%) BMI (kg/m2) Systolic BP (mm Hg) Diastolic BP (mm Hg) Fasting BG (mmol/liter) 2-h BG (mmol/liter) Fasting insulin (mU/liter) Insulin sensitivity (QUICKI) HOMA-IR HOMA-␤ Total cholesterol (mmol/liter) LDL cholesterol (mmol/liter) HDL cholesterol (mmol/liter) Triglycerides (mmol/liter) Microalbuminuria (%) WHR ⬎0.90 in men, ⬎0.85 in women (%) Hypertension (%) Use of cholesterol-lowering drugs (%) Lifetime smoking (%) No 1–25 yr ⱖ26 yr

DM subjects

IGT subjects

NGT subjects

P

54 46 30.4 ⫾ 4.4 162 ⫾ 20 89 ⫾ 9 6.7 ⫾ 1.8 11.3 ⫾ 3.1 13 (6 –29) 0.319 ⫾ 0.022 4.3 (1.9 –12.4) 67 (17–214) 5.73 ⫾ 0.88 3.61 ⫾ 0.68 1.33 ⫾ 0.40 1.60 (0.58 –5.75) 19 74 59 9

97 63 28.3 ⫾ 4.3 158 ⫾ 21 88 ⫾ 9 5.0 ⫾ 0.5 8.7 ⫾ 0.7 13 (3–56) 0.334 ⫾ 0.027 3.3 (0.9 –16.9) 91 (20 –389) 5.91 ⫾ 1.02 3.80 ⫾ 0.91 1.40 ⫾ 0.38 1.38 (0.49 –5.95) 9 57 51 15

57 46 29.6 ⫾ 4.7 144 ⫾ 16 84 ⫾ 8 4.9 ⫾ 0.5 5.7 ⫾ 1.2 11 (5–36) 0.335 ⫾ 0.022 3.0 (1.2–7.9) 82 (43–222) 5.71 ⫾ 0.89 3.65 ⫾ 0.77 1.41 ⫾ 0.40 1.33 (0.56 – 4.70) 6 60 18 11

0.050a 0.016a ⬍0.001a 0.020a ⬍0.001a ⬍0.001a 0.389 0.002a 0.005a 0.008a 0.322 0.339 0.432 0.089 0.063 0.146 ⬍0.001a 0.543

64 19 17

46 35 19

52 31 17

0.155

Data represent n, mean ⫾ SD, or percentage, except for fasting insulin, triglycerides, HOMA-IR, and HOMA-␤, which represent median (range). BG, Blood glucose; BP, blood pressure. a P value ⱕ 0.05.



TABLE 2. Mean and maximal IMT (mm) and maximal CPIMT of the carotid arteries in relation to the glucose status

Mean IMT CCA Maximal IMT CCA Mean IMT CCA/BIF/ICA Maximal IMT CCA/BIF/ICA Maximal CPIMT CCA/BIF/ICA Data are mean ⫾

SD.

DM subjects (n ⫽ 54)

IGT subjects (n ⫽ 97)

NGT subjects (n ⫽ 57)

P

0.99 ⫾ 0.31 1.12 ⫾ 0.47 1.03 ⫾ 0.26 1.44 ⫾ 0.48 2.03 ⫾ 0.73

0.92 ⫾ 0.25 1.04 ⫾ 0.39 0.98 ⫾ 0.26 1.41 ⫾ 0.59 2.05 ⫾ 0.93

0.90 ⫾ 0.22 0.99 ⫾ 0.27 0.97 ⫾ 0.22 1.39 ⫾ 0.46 1.88 ⫾ 0.88

0.137 0.173 0.340 0.473 0.493

BIF, Bifurcation.

TABLE 3. Bivariate correlations of the maximal IMT of the CCA and background variables Background variable

r

P

BMI WHR Systolic BP Diastolic BP Fasting insulin Insulin sensitivity (QUICKI) HOMA-IR HOMA-␤ LDL cholesterol HDL cholesterol Triglycerides Albumin/creatinine ratio Fasting glucose 2-h glucose

0.005 0.264 0.248 0.148 0.095 ⫺0.158 0.152 ⫺0.004 0.071 ⫺0.211 0.128 0.013 0.090 0.064

0.945 ⬍0.001a ⬍0.001a 0.036a 0.183 0.027a 0.033a 0.958 0.317 0.002a 0.066 0.848 0.194 0.391

a

P ⱕ 0.05.

cose status or hypertension. The mean QUICKI and HDL cholesterol were lower, but HOMA-IR, fasting insulin, and triglycerides were higher in subjects with severe CCA atherosclerosis, compared with those with maximal IMT CCA of less than 1.2 mm (Table 4). Systolic blood pressure was higher and diastolic blood pressure tended to be higher in the former than the latter group. Instead, there was no statistically significant difference among the groups in total cholesterol, low-density lipoprotein cholesterol, fasting glucose, or 2-h glucose levels. The interaction term of QUICKI with glucose status was not statistically significant (P ⫽ 0.373). Therefore, the multiple regression analyses were made in the whole study population. The results of multiple regression analyses are shown in Table 5. Two models are presented. Low insulin sensitivity (the two lowest tertiles of QUICKI, compared with the highest tertile), male gender, long-lasting smoking, and high systolic blood pressure were independently associated with severe CCA atherosclerosis. The adjusted odds ratio of the subjects in the two lowest tertiles of QUICKI, compared with those in the highest tertile, was 5.7 [95% confidence interval (CI) 2.2–15.1]. The odds ratio for male gender was 3.9 (95% CI, 1.8 – 8.9); long-lasting smoking, 2.9 (95% CI, 1.0 –7.9); and an increase of 10 mm Hg in systolic blood pressure, 1.3 (95% CI, 1.1–1.6). Twenty-three percent of the study subjects had no plaques, 37% had one or two plaques, and 40% three or more plaques. The proportion of subjects without plaques was higher among those with NGT than among those with IGT or DM (37% vs. 18% and 19%, respectively, P ⫽ 0.015). The corresponding prevalences in the highest, middle, and lowest QUICKI tertiles were 25%, 20%, and 23%, respectively (P ⫽ 0.781). The total number of plaques tended to be higher

among subjects in the lowest two QUICKI tertiles, compared with those in the highest tertile (median 2, range 0 – 8 vs. median 1, range 0 –11, P ⫽ 0.208). Discussion

Low insulin sensitivity measured by QUICKI, male gender, long-lasting smoking, and high systolic blood pressure were the major determinants of severe carotid atherosclerosis in the present study population. Subjects in the two lowest QUICKI tertiles had a 5-fold risk for severe carotid atherosclerosis, compared with those in the highest QUICKI tertile. Subjects with DM had lower QUICKI than subjects with IGT or NGT. There was an increasing trend in the maximal IMT of the common carotid artery along with the deterioration of glucose tolerance from NGT to IGT and DM, although the differences were not statistically significant. The proportion of subjects without plaques was higher among those with NGT than among subjects with IGT or DM. Laakso et al. (31) were the first to demonstrate that insulin resistance, measured by the euglycemic hyperinsulinemic clamp technique, is associated with carotid atherosclerosis. They compared 30 normotensive, nonobese subjects with asymptomatic plaques detected by ultrasound and 13 control subjects without plaques and observed that the former group was more insulin resistant than the latter. In population studies, direct measures of insulin resistance are difficult to obtain, and insulin concentrations are often used as a surrogate. Some previous studies (6, 8, 9, 20) have shown that either fasting or postchallenge hyperinsulinemia has been associated with an increased carotid IMT. One populationbased study (21) carried out in the same geographical area as the present study did not find any positive association between insulin parameters and carotid IMT. The contradictory results of the previous studies may be explained by the fact that fasting and postchallenge hyperinsulinemia are not satisfactory measures of insulin resistance. To our knowledge, there are no previous reports of the association between carotid atherosclerosis and low insulin sensitivity measured by QUICKI. The correlation coefficient of QUICKI with insulin sensitivity from the glucose clamp technique has been as high as 0.78 being higher for obese than nonobese subjects (24). The results of the present study also showed that insulin sensitivity measured by QUICKI and insulin resistance measured by HOMA had a high negative correlation. Both of these methods are based on simple equations derived from fasting insulin and glucose values and are useful methods in epidemiological studies. Insulin sensitivity was evaluated by the frequently sampled iv glucose tolerance test with minimal model analysis in a large epidemiological study, Insulin Resistance Athero-

5096

J Clin Endocrinol Metab, November 2002, 87(11):5092–5097


TABLE 4. Characteristics of the subjects with severe carotid atherosclerosis (maximal IMT CCA ⱖ1.2 mm) compared with subjects with maximal IMT CCA of ⬍1.2 mm

Systolic BP (mm Hg) Diastolic BP (mm Hg) Total cholesterol (mmol/liter) LDL cholesterol (mmol/liter) HDL cholesterol (mmol/liter) Triglycerides (mmol/liter) Fasting BG (mmol/liter) 2-h BG (mmol/liter) Fasting insulin (mU/liter) Insulin sensitivity (QUICKI) HOMA-IR HOMA-␤ Data represent n, mean ⫾ glucose; BP, blood pressure. a P ⱕ 0.05.

SD,

Subjects with IMT of ⱖ1.2 mm (n ⫽ 57)

Subjects with IMT of ⬍1.2 mm (n ⫽ 151)

P

161 ⫾ 22 89 ⫾ 9 5.81 ⫾ 1.06 3.78 ⫾ 0.95 1.25 ⫾ 0.38 1.60 (0.76 – 4.47) 5.54 ⫾ 1.27 8.33 ⫾ 2.43 14.0 (5–53) 0.323 ⫾ 0.020 3.9 (1.4 –16.9) 90 (17–290)

153 ⫾ 20 86 ⫾ 9 5.80 ⫾ 0.91 3.69 ⫾ 0.76 1.44 ⫾ 0.38 1.38 (0.49 –5.95) 5.37 ⫾ 1.27 8.15 ⫾ 2.43 11.0 (3–56) 0.333 ⫾ 0.027 3.2 (0.9 –15.9) 80 (20 –389)

0.030a 0.055 0.976 0.454 0.001a 0.025a 0.379 0.663 0.036a 0.004a 0.012a 0.326

or percentage except for fasting insulin, triglycerides, and HOMA-␤, which are median (range). BG, Blood

TABLE 5. Multiple regression analysis with the severe CCA atherosclerosis (maximal IMT CCA ⱖ1.2 mm) as dependent variable Model 1

Male gender Lifetime smoking ⱖ26 years QUICKI (two lowest vs. the highest tertile) Systolic BP (for an increase of 10 mm Hg)

Model 2

OR

95% CI

OR

95% CI

3.5 2.6

1.6–7.6 1.0– 6.6

3.9 2.9

1.8– 8.9 1.0–7.9

4.6

1.8–11.7

5.7

2.2–15.1

1.3

1.1–1.5

1.3

1.1–1.6

Model 2 was adjusted by glucose status and LDL cholesterol. BP, Blood pressure.

sclerosis Study (IRAS), carried out in the United States (32). The frequently sampled iv glucose tolerance test is a heavy method to carry out because 12 blood samples for measurement of glucose and insulin are needed for estimating insulin sensitivity on a computer program. The IRAS study population consisted of equal numbers of participants in different glucose tolerance categories (DM, IGT, NGT). Low insulin sensitivity was independently associated with carotid atherosclerosis. The association was reduced, but not completely eliminated, by adjusting for traditional cardiovascular risk factors. According to the suggestion of the authors (32), the association of low insulin sensitivity with carotid atherosclerosis is in part mediated by the relationship of low insulin sensitivity with confounding factors, including changes in lipids, hypertension, glucose tolerance, and adiposity. In the present study, we focused on subjects with DM and IGT. The study subjects were selected from a populationbased cohort by inviting all subjects with clinical diabetes and IGT and, in addition, by selecting a normoglycemic control subject for each diabetic subject. Seventy-three percent of the study subjects had clinical diabetes or IGT, and 41% of the study subjects were obese (BMI ⱖ 30 kg/m2). We suggest that the results of this study can be generalized to apply to elderly Finnish subjects with DM and IGT, but because of the small number of normoglycemic subjects,

some caution should be observed when generalizing the results to apply to them. The mean and maximal IMT of the common carotid artery tended to be higher in diabetic than in nondiabetic subjects. Furthermore, the proportion of subjects with plaques was higher among diabetic than normoglycemic subjects. Many previous studies have shown that diabetic subjects have increased carotid IMT, compared with nondiabetic subjects (5–10). In the present study, the results concerning IMT did not quite reach statistical significance, which may be explained by the relatively small number of subjects and the low statistical power. In addition, the diabetic subjects of this study were mostly newly diagnosed: only 12 of the 54 diabetic subjects had been diagnosed before 1992. In the present study, although diabetic subjects did not have increased carotid IMT, they had lower insulin sensitivity than subjects with IGT or NGT. In addition to low insulin sensitivity, hyperglycemia per se can induce a large number of alterations at the cellular level of the vascular tissue that potentially accelerate the atherosclerotic process (33). In the present study, the mean and maximal IMT of the common carotid artery in subjects with IGT was intermediate, compared with subjects with DM and those with NGT, resembling more closely the values of normoglycemic than diabetic subjects. Some of the previous studies have demonstrated an increased carotid IMT in subjects with IGT, compared with normoglycemic subjects (7, 9), but some others have failed to detect any difference between the groups (8, 11). The present study subjects were classified as having impaired glucose tolerance based on the results of one OGTT test. The group of subjects with IGT is heterogeneous, however, and on repeat testing, some subjects revert to NGT (34). Kawamori (9) has shown that IGT subjects with fasting and postchallenge hyperinsulinemia had increased carotid IMT, compared with IGT subjects with low insulin values. The former group had higher BMI, diastolic blood pressure, and triglyceride levels than the latter group (i.e. the former group was more insulin resistant than the latter group). One population-based study (21) carried out in the same geographical area as the present study did not find any positive association between insulin parameters and carotid


IMT. Technical differences in performing the ultrasound examinations cannot explain the contradictory results because the same trained radiologist (M.P.) made the examinations in both studies. Laboratory measurements of insulin were different in these studies, and QUICKI was not yet available for measuring insulin sensitivity in 1998 when Rantala et al. (21) published their results. The mean age of the study subjects was about 10 yr higher in the present study, compared with the other study population. These differences between the studies may explain the contradictory results. Male gender, smoking, and high systolic blood pressure were associated with an increased IMT of the common carotid artery in the present study. The association of male gender (4, 6, 10, 35), smoking (4, 5, 9, 10, 21, 35), and high systolic blood pressure (4, 5, 21, 35) or hypertension (4, 21, 35) with carotid IMT has also been a well-established finding in previous studies. In conclusion, low insulin sensitivity measured by QUICKI, male gender, long-lasting smoking, and high systolic blood pressure were the major determinants of severe carotid atherosclerosis in the present study population. The results of this study showed that subjects in the two lowest QUICKI tertiles had a 5-fold risk for severe carotid atherosclerosis, compared with those in the highest QUICKI tertile. QUICKI and insulin resistance measured by HOMA method had a high negative correlation, and the results were virtually the same by using either QUICKI or HOMA in evaluating insulin sensitivity. Acknowledgments Received May 6, 2002. Accepted August 13, 2002. Address all correspondence and requests for reprints to: Ulla Rajala, M.D., Ph.D., Department of Public Health Science and General Practice, University of Oulu, Aapistie 1, 90220 Oulu, Finland. E-mail: urajala@ cc.oulu.fi.

References 1. Pignoli P, Tremoli E, Poli A, Oreste P, Paoletti R 1986 Intimal plus medial thickness of the arterial wall: a direct measurement with ultrasound imaging. Circulation 74:1399 –1406 2. Wong M, Edelstein J, Wollman J, Bond G 1993 Ultrasonic-pathological comparison of the human arterial wall: verification of intima-media thickness. Arterioscl Thromb 13:482– 486 3. Heiss C, Sharrett AR, Barnes R, Chambless LE, Szklo M, Alzola C, ARIC Investigators 1991 Carotid atherosclerosis measured by B-mode ultrasound in populations: associations with cardiovascular risk factors in the ARIC study. Am J Epidemiol 134:250 –256 4. O’Leary DR, Polak JF, Kronmal RA, Kittner SJ, Bond MG, Wolfson SK, Bommer W, Price TR, Gardin JM, Savage PJ 1992 Distribution and correlates of sonographically detected carotid artery disease in the cardiovascular health study. Stroke 23:1752–1760 5. Salonen R, Salonen JT 1991 Determinants of carotid intima-media thickness: a population-based ultrasonography study in eastern Finnish men. J Intern Med 229:225–231 6. Folsom AR, Eckfeldt JH, Weizman S, Ma J, Chambless LE, Barnes RW, Cram KB, Hutchinson RG, for the Atherosclerosis Risk in Communities (ARIC) Study Investigators 1994 Relation of carotid artery wall thickness to diabetes mellitus, fasting glucose and insulin, body size, and physical activity. Stroke 25:66 –73 7. Yamasaki Y, Kawamori R, Matsushima H, Nishizawa H, Kodama M, Kubota M, Kajimoto Y, Kadama T 1995 Asymptomatic hyperglycemia is associated with increased intimal plus medial thickness of the carotid artery. Diabetologia 38:585–591 8. Niskanen L, Rauramaa R, Miettinen H, Haffner SM, Mercuri M, Uusitupa M 1996 Carotid artery intima-media thickness in elderly patients with NIDDM and in nondiabetic subjects. Stroke 27:1986 –1992 9. Kawamori R 1998 Asymptomatic hyperglycaemia and early atherosclerotic changes. Diabetes Res Clin Pract 40(Suppl 40):S35–S42


10. Bonora E, Kiechl S, Oberhollenzer F, Egger G, Bonadonna RC, Muggeo M, Willeit J 2000 Impaired glucose tolerance, type II diabetes mellitus and carotid atherosclerosis: prospective results from the Bruneck study. Diabetologia 43: 156 –164 11. Tuomilehto J, Qiao Q, Salonen R, Nissinen A, Salonen JT 1998 Ultrasonographic manifestations of carotid atherosclerosis and glucose intolerance in elderly eastern Finnish men. Diabetes Care 21:1349 –1352 12. Welborn TA, Wearne K 1979 Coronary heart disease incidence and cardiovascular mortality in Busselton with reference to glucose and insulin concentrations. Diabetes Care 2:154 –160 13. Ducimetiere P, Eschwege E, Papoz L, Richard JL, Claude JR, Rosselin G 1981 Relationship of plasma insulin levels to the incidence of myocardial infarction and coronary heart disease mortality in a middle-aged population. Diabetologia 19:205–210 14. Pyo¨ra¨la¨ K, Savolainen E, Kaukola S, Haapakoski J 1985 Plasma insulin as coronary heart disease factor: relationship to other risk factors and predictive value during 91⁄2-year follow-up of the Helsinki Policemen Study population. Acta Med Scand 701(Suppl):38 –52 15. Fontbonne A, Charles MA, Thibult N, Richard JL, Claude JR, Warnet JM, Rosselin GE, Eschwe`ge E 1991 Hyperinsulinaemia as a predictor of coronary heart disease mortality in a healthy population: the Paris Prospective Study, 15-year follow-up. Diabetologia 34:356 –361 16. Pyo¨ra¨la¨ M, Miettinen H, Laakso M, Pyo¨ra¨la¨ K 1998 Hyperinsulinemia predicts coronary heart disease risk in healthy middle-aged men. The 22-year follow-up results of the Helsinki Policemen Study. Circulation 98:398 – 404 17. Folsom AR, Szklo M, Stevens J, Liao F, Smith R, Eckfeldt JH 1997 A prospective study of coronary heart disease in relation to fasting insulin, glucose, and diabetes. The Atherosclerosis Risk in Communities (ARIC) Study. Diabetes Care 20:935–942 18. Reaven GM 1988 Role of insulin resistance in human disease. Diabetes 37: 1595–1607 19. Yudkin JS 1996 Hyperinsulinemia, insulin resistance, microalbuminuria and the risk of coronary heart disease. Ann Intern Med 28:433– 438 20. Fujii K, Abe I, Ohya Y, Onaka U, Tominaga M, Ohmori S, Fujishima S, Kobayashi K, Fujishima M 1997 Association between hyperinsulinemia and intima-media thickness of the carotid artery in normotensive men. J Hypertens 15:167–172 21. Rantala AO, Païva¨nsalo M, Kauma H, Lilja M, Savolainen MJ, Reunanen A, Kesa¨niemi YA 1998 Hyperinsulinemia and carotid atherosclerosis in hypertensive and control subjects. Diabetes Care 21:1188 –1193 22. Matthews DR, Hosker JP, Rudenski AS, Naylor BA, Treacher DF, Turner RC 1985 Homeostasis model assessment: insulin resistance and ␤-cell function from fasting plasma glucose and insulin concentration in man. Diabetologia 28:412– 419 23. Van Haeften T 1998 Glucose tolerance, insulin sensitivity, and the homeostasis model assessment method. Diabetes Care 21:673– 674 24. Katz A, Nambi SS, Mather K, Baron AD, Follman DA, Sullivan G, Quon MJ 2000 Quantitative insulin sensitivity check index: a simple, accurate method for assessing insulin sensitivity in humans. J Clin Endocrinol Metab 85:2402– 2410 25. Hrebicek J, Janout V, Malincikova J, Horakova D, Cizek L 2002 Detection of insulin resistance by simple quantitative insulin check index OUICKI for epidemiological assessment and prevention. J Clin Endocrinol Metab 87: 144 –147 26. Rajala U, Keina¨nen-Kiukaanniemi S, Uusima¨ki A, Reijula K, Kivela¨ S-L 1995 Prevalence of diabetes mellitus and impaired glucose tolerance in a middle-aged Finnish population. Scand J Prim Health Care 13:222–228 27. Qiao Q, Keina¨nen-Kiukaanniemi S, Rajala U, Uusima¨ki A, Kivela¨ S-L 1996 Risk for diabetes and persistent impaired glucose tolerance among middleaged Finns. Diabetes Res Clin Pract 33:191–198 28. Rajala U, Qiao Q, Laakso M, Keina¨nen-Kiukaanniemi S 2000 Antihypertensive drugs as predictors of type 2 diabetes among subjects with impaired glucose tolerance. Diabetes Res Clin Pract 50:231–239 29. World Health Organization 1978 Arterial Hypertension, Technical Report, Series no. 638. Geneva: World Health Organization 30. World Health Organization 1985 Diabetes mellitus: report of a WHO Study Group, Technical Report Series no. 727. Geneva: World Health Organization 31. Laakso M, Sarlund H, Salonen R, Suhonen M, Pyo¨ra¨la¨ K, Salonen JT, Karhapaä¨ P 1991 Asymptomatic atherosclerosis and insulin resistance. Arterioscler Thromb 11:1068 –1076 32. Howard G, O’Leary DH, Zaccaro D, Haffner S, Rewers M, Hamman R, Selby JV, Saad MF, Savage P, Bergman R 1996 Insulin sensitivity and atherosclerosis. Circulation 93:1809 –1817 33. Aronson D, Rayfield EJ 2002 How hyperglycemia promotes atherosclerosis: molecular mechanisms. Cardiovasc Diabetol 1:1 34. Riccardi G, Vaccaro O, Rivellese A, Pignalosa S, Tutino L, Mancini, M 1985 Reproducibility of the new diagnostic criteria for impaired glucose tolerance. Am J Epidemiol 121:422– 429 35. Païva¨nsalo M, Rantala A, Kauma H, Lilja M, Reunanen A, Savolainen M, Kesa¨niemi A, Suramo I 1996 Prevalence of carotid atherosclerosis in middleaged hypertensive and control subjects. A cross-sectional systemic study with duplex ultrasound. J Hypertens 14:1433–1439