Carotid artery intima-media thickness in Finnish families with familial combined hyperlipidemia

Transcription

1 Atherosclerosis 162 (2002) Carotid artery intima-media thickness in Finnish families with familial combined hyperlipidemia Kati Ylitalo a, Mikko Syvänne a, Riitta Salonen b,c, Ilpo Nuotio d,e, Marja-Riitta Taskinen a, *, Jukka T. Salonen b,c a Department of Medicine, Helsinki Uni ersity Central Hospital, Uni ersity of Helsinki, PO Box 340, Floor 11, Haartmaninkatu 4, 0029 Hus, Finland b The Research Institute of Public Health, Uni ersity of Kuopio, Kuopio, Finland c Inner Sa o Health Centre, Suonenjoki, Finland d Department of Human Genetics, The Gonda (Goldschmied) Neuroscience and Genetics Research Center, UCLA, Los Angeles, CA, USA e Department of Medicine, Turku Uni ersity Central Hospital, Uni ersity of Turku, Turku, Finland Received 30 May 2001; received in revised form 3 August 2001; accepted 21 August 2001 Abstract Background: Familial combined hyperlipidemia (FCHL) is the most common hereditary lipid disorder that predisposes the patients to premature coronary heart disease. Members of FCHL families are categorised as affected or unaffected according to serum lipid levels. This study is aimed to evaluate whether there is a difference in carotid artery wall thickness between asymptomatic FCHL family members who are affected and those who are unaffected according to the currently used lipid criteria. Methods and results: Carotid artery ultrasonography with intima-media thickness (IMT) measurements was performed for 148 members of 39 Finnish FCHL families. Study subjects who had no history of coronary heart disease or stroke were divided into two groups according to their serum total cholesterol and/or triglyceride levels. The average carotid IMT of the affected subjects ( mm) was not significantly different from that of their unaffected relatives ( mm), P=0.90. In multivariate analysis, age, gender, and pulse pressure, but no lipid variables, contributed significantly to the variation of carotid IMT. Conclusions: The IMT findings in FCHL family members indicate that the current lipid criteria alone are of limited value in predicting long-term risk of cardiovascular disease in asymptomatic members of FCHL families Elsevier Science Ireland Ltd. All rights reserved. Keywords: B-mode ultrasonography; Familial; Combined hyperlipidemia; Intima-media thickness; Atherosclerosis 1. Introduction * Corresponding author. Tel.: ; fax: address: marja-riitta.taskinen@hus.fi (M.-R. Taskinen). With a population prevalence of 1 2%, familial combined hyperlipidemia (FCHL) is the most common hereditary lipid disorder [1]. Up to 25% of survivors of premature myocardial infarction may have FCHL [1 3]. The affected FCHL family members exhibit hypercholesterolemia, hypertriglyceridemia, or both, along with elevated serum levels of apolipoprotein B (apo B). FCHL is a heterogeneous disorder that is most probably caused by several genetic defects and metabolic abnormalities. So far the members of FCHL families have been categorised as affected or unaffected according to different lipid or lipoprotein cut-off points [4]. However, the current lipid criteria, most often based on total cholesterol (TC) and triglycerides (TG), do not provide information on the degree of atherosclerosis in FCHL family members. B-mode ultrasonography is a method to assess early atherosclerotic changes of the arterial wall [5]. The association of carotid intima-media thickness (IMT) with the well-known risk factors for coronary heart disease (CHD), such as dyslipidemia, smoking and hypertension has been demonstrated in several studies [6 9]. Both cross-sectional and prospective studies have established the relationship between carotid artery IMT and cardiovascular disease [10 12] /02/$ - see front matter 2002 Elsevier Science Ireland Ltd. All rights reserved. PII: S (01)

2 172 K. Ylitalo et al. / Atherosclerosis 162 (2002) We performed carotid ultrasonography with IMT measurements for 148 asymptomatic members of 39 well-characterised Finnish FCHL families. The study aimed to evaluate whether FCHL family members who are affected according to the current lipid criteria, but have no clinical symptoms of CHD, have greater carotid artery IMT than their unaffected relatives. 2. Subjects and methods 2.1. Study subjects The study subjects were originally recruited in Helsinki and Turku University Central Hospitals in Finland as a part of the EUFAM Study (European Multicenter Study on Familial Dyslipidemias). For both the EUFAM Study and the present substudy, the FCHL probands were required to be years of age, have verified CHD, and have TC and/or TG age and sex-specific Finnish 90th population percentiles. Members of the FCHL families were categorised as affected if their serum TC and/or TG exceeded the age and sex-specific 90th population percentiles [13]. Families that had at least one affected first degree relative in addition to the proband, were included in the study, and all accessible relatives were examined. None of the subjects was known to have renal or liver disease, hypothyroidism, or any other disease known to cause secondary hyperlipidemia. The EUFAM Study protocol has been presented in detail previously [14,15]. In principle, all non-diabetic members of our Finnish FCHL family cohort who were over 18 years of age, and lived at a reasonable distance from Helsinki were invited to the ultrasound study. Family members with a history of CHD or stroke were excluded; firstly, because individuals with clinical cardiovascular disease are known to have a high degree of atherosclerosis and thickened carotid artery walls [10,16,17]. Inclusion of the 39 probands in the group of affected family members (all probands are affected by definition) would have assigned in the affected group 39 subjects who by definition have CHD, and therefore most likely larger IMT than subjects with no clinical CHD. Secondly, in our cohort, FCHL patients with CHD are on lipid lowering medication with almost no exception, and it was considered unethical to pause their medication. Treatment might have affected their IMTs as well. Five relatives with both CHD and lipid lowering medication, and 11 relatives with lipid lowering medication were also excluded. A total of 148 members from 39 Finnish FCHL families ( four subjects/family) participated in the present study. The ethics committee of the Department of Medicine in Helsinki University Central approved the study protocol. Informed consent was obtained from all participants Ultrasound examinations Ultrasound scannings were performed with a Hewlett Packard Image Point M2410A ultrasound system (Hewlett Packard, Andover, USA) equipped with a 10 MHz linear array transducer. Scannings were videotaped with a Panasonic AG-MD830E PAL S- VHS VCR (Matsushita Electric Industrial Co., Ltd., Osaka, Japan). One physician (K.Y.) carried out all ultrasound examinations. Longitudinal images from three projections (anterolateral, lateral, and posterolateral) were displayed for each arterial segment. Images were focused and measurements carried out at a total of 28 sites: both the far wall (FW) and the near wall (NW) of six arterial segments: right and left distal 1 cm of common carotid artery (CCA), carotid bulb (CB), and proximal 1 cm of internal carotid artery (ICA). In CCA and CB all three projections were used, and in ICA a single angle with the best visualisation was used. IMT measurements were done by a single reader at Oy Jurilab Ltd ( IMT was measured from ultrasound images with a PC equipped with a video frame grabber interfaced to a S-VHS VCR. The Prosound software (Caltech, Pasadena, USA) [18] was used to measure the IMTs at a total of 28 sites corresponding to the 28 sites where the scanning was focused. All measurements were done in the diastole, when IMT is at its largest. Three variables were derived from each measurement: the minimum, the mean, and the maximum IMT. All outcome variables were first calculated for each subject. The average of all mean IMT measurements (mean-imt) over 28 sites (or fewer, if measurement could not be performed at all 28 sites) was chosen as the primary outcome variable. Likewise, the average of maximum IMTs (max-imt), the average of mean far-wall IMTs (FW-IMT), the average of mean near-wall IMTs (NW-IMT), and the average of maximum IMTs for each segment (CCA-IMT, CB-IMT, ICA-IMT), were calculated as secondary outcome variables. Plaques were not registered as their prevalence was expected to be low in subjects free of clinical cardiovascular disease Clinical measurements Venous blood samples for serum lipid and other biochemical determinations were collected in the morning after a 12 h fast within 2 weeks of the ultrasound examination. An oral glucose tolerance test (OGTT) with 75 g of glucose was performed at the initial study visits of the EUFAM project, 0 4 years before the ultrasound examination [14]. Fasting and area under the curve (AUC) values for glucose, insulin, and FFA presented in the current study are based on this OGTT. Lifelong exposure to smoking (cigarette years) was estimated as the product of the number of cigarettes

3 K. Ylitalo et al. / Atherosclerosis 162 (2002) smoked daily and the number of years smoked. Subjects were also categorised as smokers (current and ex-smokers) and non-smokers. Waist and hip circumferences, weight and height were measured, and body mass index (BMI) recorded as weight (kg)/height (m) 2. Blood pressure was measured with a mercury sphygmomanometer after the ultrasound examination with the subject still in a supine position. Subjects were coded as hypertensive, if the systolic blood pressure (SBP) was 140 mmhg, the diastolic blood pressure (DBP) was 90 mmhg, or the subject used antihypertensive medication. Pulse pressure was calculated as the difference between the SBP and the DBP Measurement of lipids and other metabolic parameters Measurement of lipids and lipoproteins has been previously described in detail [14]. Briefly, serum TC, TG and HDL-cholesterol (after precipitation procedures) concentrations were determined enzymatically, and serum total apo B, and apolipoprotein A-I (apo A-I) concentrations by an immunoturbidimetric method. Commercial standards used for apolipoprotein assays had been calibrated against the reference material from the Centers for Disease Control and Prevention. LDL was separated by sequential flotation as described [19]. Blood glucose concentrations were determined by the glucose dehydrogenase method, serum free insulin concentrations by radioimmunoassay, and concentrations of FFA using the microfluorometric method of Miles [20] Statistical analyses Statistical comparisons of clinical parameters were performed with the version 9.0 of the SPSS for Windows software (SPSS Inc., Chicago, USA). Data are presented as mean S.D. for continuous variables, and as frequencies or percentages for categorical variables. Variables with non-normal distribution were log 10 transformed. AUC values were calculated using the trapezoid rule. Continuous variables were compared between the two groups by two-way analysis of variance (ANOVA). Family number (that indicates belonging to a certain family) was used as a random factor in two-way ANOVA to at least partly correct for the nonindependence of the study subjects. The frequency distribution of the categorical variables in the two groups was compared by the 2 -test or Fisher s exact test. P-values are presented assuming all study subjects are independent. Correlations were calculated using multivariate analysis, in which family number was always included as an independent variable. To examine age or gender adjusted partial correlations, age or gender was entered in multivariate analyses as an independent variable. Multivariate analysis was performed with the log 10 -transformed mean-imt as the dependent variable. All variables with an at least moderate (P 0.20) correlation (adjusted only for family number) with mean-imt, were chosen for the multivariate analysis. Finally, the following continuous variables were selected for the multivariate model: age, LDL-cholesterol (LDL-C), TG (log), apo B, glucose-auc (log), FFA-AUC (log), BMI (log), waist-hip ratio, pulse pressure (log), and cigarette years. LDL-C, pulse pressure and glucose-auc were selected as representatives of the pertinent parameters to avoid problems of multicollinearity. In addition to family number, gender was always included in the model. As a next step, variables were removed from the model until the best fitting model with the maximum adjusted multiple R 2 was achieved. The present study had an 80% probability to detect a 0.07 mm difference in the mean-imt (with a S.D. of 0.15 mm) between the two groups at a significance level of Results 3.1. Subject characteristics The mean age of study subjects was similar in the two groups (Table 1). Gender distribution was equal among the affected and the unaffected. By definition, the lipid values were higher in the affected family members than in the unaffected relatives (Table 2). In addition to TC and TG values of the sonography visit, the maximum values of TC and TG (TC-max and TG-max) measured at any of the EUFAM Study visits are shown. The affected subjects were more obese than the unaffected (P=0.039), and consequently also had higher glucose-auc and insulin-auc values during the OGTT (P=0.026 and P=0.024, respectively). Thirty five subjects (23 affected, 12 unaffected) were diagnosed to have hypertension. Six affected and two unaffected subjects received medication for hypertension. They were allowed to take their medication in the morning before the blood pressure measurement. On the average 24.4 of 28 measurements/subject (24.3 measurements in the affected and 24.5 measurements in unaffected subjects, P=0.22) were available for calculation of the mean-imt. ICA could not be visualised in two affected subjects. Intra-sonographer variability in the scannings was estimated by scanning 16 subjects twice on two different occasions within 3 weeks. The Spearman s correlation coefficient between the mean- IMTs of the paired scannings was (P), the coefficient of variation 2.4%, and the absolute difference (mean S.D.) mm. For the reader, Spearman s correlation coefficient between the mean-

4 174 K. Ylitalo et al. / Atherosclerosis 162 (2002) IMTs of two paired readings was (P), the coefficient of variation 0.29%, and the absolute difference mm Intima-media thickness The average mean-imt of all subjects was mm. The average mean-imt of the affected family members ( mm) did not differ from that of the unaffected subjects ( mm) (P=0.90). There was no significant difference between the affected and unaffected subjects with regard to other IMT outcome variables either (Table 3). Since age had a significant effect on IMT, the mean- IMT was examined also in three age groups, each spanning approximately 15 years: 19 35, 36 50, and over 50 years. The number of affected and unaffected subjects was comparable in each age group. No significant differences in mean-imt could be observed between the affected and unaffected in any age group (Fig. 1). There was no significant difference in mean- IMT between males ( mm) and females ( mm, P=0.26), or smokers (current and ex-smokers) ( mm) and non-smokers ( mm, P=0.19). Hypertensive subjects had higher mean-imt ( mm) than the normotensive subjects ( mm) (P) Correlation analyses The strongest correlation of mean-imt was found with age (r=0.81, P). Of the lipid and lipoprotein variables, mean-imt correlated statistically significantly with TC, LDL-cholesterol, TG and apo B (Table 4). Significant correlations were also observed between mean-imt and fasting glucose, and glucose- AUC, but not with insulin or FFA variables. Mean- IMT correlated significantly with the following clinical variables: BMI, waist-hip ratio, SBP, DBP, pulse pressure and cigarette years. When adjusted for age, most of the above mentioned correlations tended to weaken (Table 4). Adjustment for gender did not have a significant influence on any of the correlations presented (data not shown) Multi ariate analysis In the final model that explained 75.4% of the variation in mean-imt (adjusted multiple R 2 =0.754), age had by far the highest standardised coefficient (0.747, P), followed by log-transformed pulse pressure (0.212, P) and gender ( 0.193, P=0.002, indicating that men had higher average mean-imt than women). Apo B, waist-hip ratio and cigarette years did not reach statistical significance in the model (Table 5). 4. Discussion To the best of our knowledge, this is the first study that examined the carotid artery intima-media thickness in an FCHL cohort. No significant differences in carotid IMT between the affected and the unaffected FCHL family members could be detected when subjects were categorised according to the 90th population percentiles for TC and TG. Exclusion of family members Table 1 Clinical characteristics of the study subjects Variable Affected Unaffected P-value n Mean S.D. n Mean S.D. Age (years) years years years Males (%, M/F) 40 (31/46) (26/45) 0.74 Cigarette years Smokers (%) Drinking weekly (%) HRT (n) BMI (kg/m 2 ) Waist-hip ratio SBP (mmhg) DBP (mmhg) Pulse pressure (mmhg) Hypertension (n) Smokers include current and ex-smokers. HRT denotes hormone-replacement therapy. BMI is body mass index; SBP, systolic blood pressure and DBP, diastolic blood pressure. Pulse pressure was calculated as SBP DBP. For statistical comparisons, two-way ANOVA (with family number as a random factor) was used for continuous and 2 or Fisher s exact test for categorical variables.

5 K. Ylitalo et al. / Atherosclerosis 162 (2002) Table 2 Biochemical characteristics of the study subjects Variable Affected Unaffected P-value n Mean S.D. n Mean S.D. TC (mmol/l) TC-max (mmol/l) LDL-C (mmol/l) HDL-C (mmol/l) TG (mmol/l) TG-max (mmol/l) Apo B (mg/dl) Apo A-I (mg/dl) Glucose (mmol/l) Glucose AUC (mmol/l per h) Insulin (mu/l) Insulin AUC (mu/l per h) FFA AUC ( mol /l per h) TC indicates total cholesterol; LDL-C, LDL-cholesterol; HDL-C, HDL-cholesterol and TG, triglycerides. TC-max and TG-max are the highest TC and TG values measured for the subject in the EUFAM study. FFA denotes free fatty acids. AUC-values were calculated from 0, 30, 60 and 120 min values of the corresponding variables during the oral glucose tolerance test. Two-way ANOVA (with family number as a random factor) was used for statistical comparisons. with CHD (and probably with the highest lipid values) may have attenuated the difference in IMT between the affected and unaffected relatives. However, subjects who already have the diagnosis of CHD, are known to represent a population with a high degree of atherosclerosis. We aimed to examine specifically those family members who are symptomless at the time, but have been assigned a high risk of CHD because of their dyslipidemia and/or family history of early CHD. The present study did not involve an unrelated control group. The lack of a difference in carotid IMT between the two groups could also reflect an increased CHD risk in all FCHL relatives, as compared with normal population. However, this does not seem likely considering the relatively thin arterial walls in the present cohort. Another explanation for the present result may be, as suggested by Sharrett et al. [21], that different factors may be involved in different stages of atherogenesis. Clinical events in FCHL may be more related to atherothrombotic factors than to classical risk factors that are associated with wall thickness. In the present study cohort the lipid values tended to be lower at the sonography visit than at the earlier study visits of the EUFAM Study. This may be due to a change of diet and other lifestyle factors after the subjects became aware of their dyslipidemia. This phenomenon has been recently described in a Dutch FCHL cohort: only 74% of subjects with FCHL diagnosis remained affected over a 5-year period [22]. Therefore we also present the highest TC and TG values measured for each subject during the EUFAM Study. FCHL is characterised by elevated serum levels of apo B [4]. However, the cut-off limit depends on the population and the assay of apo B in particular, or its standardisation. In a Dutch FCHL cohort, all subjects with TC, TG or LDL-C exceeding the cut-off limits had serum apo B concentration 125 mg/dl [23]. Total plasma apo B has proved to be a relevant estimate of CHD risk in men [24], and possibly to provide information that would not be obtained from the conventional lipid measurements [24,25]. Since the currently used lipid criteria could not distinguish asymptomatic FCHL family members with a high degree of atherosclerosis from those with a low degree of atherosclerosis, we performed further analysis by dividing the present cohort into two groups: those with maximum serum apo B equal to or greater than 110 mg/dl (n=69) and those with serum apo B less than 110 mg/dl (n=79). The apo B cut-off was derived from the median of the Quebec Heart Study control population (consisting of men only) free of CHD [26]. Subjects in the high apo B category had a significantly greater mean- IMT than subjects in the low apo B category ( vs mm, P). The difference remained significant when adjusted for age. This encourages reevaluation of the current FCHL criteria. In the present study, all lipid variables mentioned in Table 2 (except for HDL-C and apo A-I) correlated significantly with mean-imt. Adjustment for age markedly attenuated these correlations confirming the association of lipid variables with age. Age was the most important predictor of variation of the mean-imt in multivariate analysis. Strong evidence exists for the role of age as a major determinant of carotid IMT [7,8,27]. However, it is more likely that age as such does not cause thickening of the arterial wall, but the relationship is more probably explained by lifelong exposure to various CHD risk factors, such as hyperlipidemia, hypertension and smoking.

6 176 K. Ylitalo et al. / Atherosclerosis 162 (2002) Table 3 Intima-media thickness of the study subjects Variable Affected (n=) Unaffected (n=71) P-value Mean S.D. Range Mean S.D. Range Mean-IMT (mm) FW-IMT (mm) NW-IMT (mm) Max-IMT (mm) CCA-IMT (mm) CB-IMT (mm) ICA-IMT (mm) Mean-IMT is the average of all mean IMT measurements over the 28 sites. FW-IMT is the average of mean far-wall IMT measurements and NW-IMT the average of mean near-wall IMT measurements. Max-IMT is the average of maximum IMTs over the 28 sites. CCA-IMT, CB-IMT and ICA-IMT represent the average of maximum IMTs of the respective segments. P-values were calculated by two-way ANOVA with the family number as a random factor. No formal correction for multiple testing was performed. SBP has been consistently related to carotid atherosclerosis in several studies, whereas DBP has not, or the relationship has been negative [6,8,9,28]. In the current study, all blood pressure variables correlated with mean-imt, but also with age (data not shown). Pulse pressure was chosen to represent the BP variables in the multivariate analysis, because there is evidence that pulse pressure may be an even better predictor of CHD than SBP [8,29]. In the multivariate analysis pulse pressure was, after age, the most important predictor of carotid IMT variation in FCHL family members. The antihypertensive medication used by eight study subjects may have caused some bias in the results, and made the observed relationship between BP and carotid IMT weaker. In our FCHL cohort, adjustment for age weakened the correlations between mean-imt and fasting glucose and glucose-auc. This is in line with reports showing only a weak, or no correlation between fasting plasma glucose and carotid IMT in non-diabetic populations [8,30].It was of interest that in multivariate analysis no lipid parameter contributed significantly to the variation of carotid IMT, though several significant correlations were observed. The minor influence of lipids on IMT in the FCHL cohort raises the question whether the carotid MT is an indicator of atherosclerosis, a disorder of arterial intima. The intimal and medial layers of the arterial wall cannot be distinguished from each other by ultrasound. Therefore it is possible that the strong association of carotid IMT with pulse pressure observed in this study is not solely a result of increased atherosclerosis, but also of medial hypertrophy. However, this does not abolish the importance of carotid artery IMT as an established indicator for increased CHD risk [11,12]. Unfortunately, the lack of standardisation and normal values of IMT measurements prevent its use in defining individual CHD risk. We have for the first time assessed the degree of early atherosclerosis in an asymptomatic FCHL cohort consisting of 148 family members from 39 Finnish FCHL families. The IMT findings in FCHL family members indicate that the current lipid criteria alone are of limited value in predicting long-term risk of cardiovascular disease in asymptomatic members of FCHL families. In multivariate analysis, age, gender and pulse pressure contributed significantly to the variation of the mean-imt. Thus the results underline firstly, the importance of investigating all risk factors and not only hyperlipidemia, in FCHL family members. Secondly, the results encourage identification of better diagnostic and prognostic criteria for FCHL family members. Fig. 1. Carotid artery mean intima-media thickness of the affected (black bars) and unaffected (gray bars) FCHL family members in three age groups. The bars represent mean S.D. The small boxes denote the mean of each group. The mean S.D. of the mean-imt for each group is shown below the bar. Number of subjects in each group is shown below the x-axis. There were no significant differences in mean-imt between the affected and unaffected in any age group.

7 K. Ylitalo et al. / Atherosclerosis 162 (2002) Table 4 Results of the correlation analyses Variable n Mean-IMT* fam P-value Mean-IMT fam+age P-value Age (years) TC (mmol/l) LDL-C (mmol/l) HDL-C (mmol/l) TG (mmol/l) Apo B (mg/dl) Apo A-I (mg/dl) Glucose (mmol/l) Glucose AUC (mmol/l per h) Insulin (mu/l) Insulin AUC (mu/l per h) FFA AUC ( mol /l perh) BMI (kg/m 2 ) Waist-hip ratio SBP (mmhg) DBP (mmhg) Pulse pressure (mmhg) Cigarette years Abbreviations as in Tables 1 3. Partial correlations of the mean-imt with the selected variables were calculated by multivariate analyses with * the family number as an independent variable and family number and age as independent variables. Table 5 Multivariate model of mean intima-media thickness in FCHL family members Effect Coefficient Std error Std coefficient t Tolerance P-value Constant Family number Age Gender Pulse pressure (log) Cigarette years Waist-hip ratio Apo B Std denotes standard(ised). Adjusted multiple R 2 of the model is 0.754, ANOVA F-ratio is 62.2 (P). Acknowledgements The image analyses were carried out at Oy Jurilab Ltd ( Arja Malkki s excellent work in reading the scannings and expert help in setting up the equipment is gratefully acknowledged. We thank Kimmo Ronkainen for expert data management. As always, the skilful assistance of the personnel at the Research Laboratory of Helsinki University Central Hospital is greatly appreciated. We thank the FCHL family members for participating in this study. This work was supported by grants from the Research Foundation of Orion Corporation, The Finnish Medical Foundation, the Helsinki University Central Hospital Research Funds, the Clinical Research Funds of Helsinki University Central Hospital, and the Finnish Heart Foundation. The EUFAM study is supported by the European Commission, contract number BMH4- CT References [1] Goldstein JL, Schrott HG, Hazzard WR, Bierman EL, Motulsky AG. Hyperlipidemia in coronary heart disease. II. Genetic analysis of lipid levels in 176 families and delineation of a new inherited disorder, combined hyperlipidemia. J Clin Invest 1973;52: [2] Genest JJ Jr, Martin-Munley SS, McNamara JR, et al. Familial lipoprotein disorders in patients with premature coronary artery disease. Circulation 1992;85: [3] Nikkilä EA, Aro A. Family study of serum lipids and lipoproteins in coronary heart disease. Lancet 1973;1: [4] Grundy SM, Chait A, Brunzell JD. Familial combined hyperlipidemia workshop. Arteriosclerosis 1987;7: [5] Salonen JT, Salonen R. Ultrasound B-mode imaging in observational studies of atherosclerotic progression. Circulation 1993;87(Suppl II):II [6] O Leary DH, Polak JF, Kronmal RA, et al. Stroke 1992;23: [7] Crouse JR, Toole JF, McKinney WM, et al. Risk factors for extracranial carotid atherosclerosis. Stroke 1987;18:

8 178 K. Ylitalo et al. / Atherosclerosis 162 (2002) [8] Zanchetti A, Bond MG, Hennig M, et al. Risk factors associated with alterations in carotid intima-media thickness in hypertension: baseline data from the European Lacidipine Study on Atherosclerosis. J Hypertens 1998;16: [9] Salonen R, Salonen JT. Carotid atherosclerosis in relation to systolic and diastolic blood pressure: Kuopio Ischaemic Heart Disease Risk Factor Study. Ann Med 1991;23:23 7. [10] Burke GL, Evans GW, Riley WA, et al. Arterial wall thickness is associated with prevalent cardiovascular disease in middleaged adults. The Atherosclerosis Risk in Communities (ARIC) Study. Stroke 1995;26: [11] Chambless LE, Heiss G, Folsom AR, Rosamond W, Szklo M, Sharrett AR, Clegg LX. Association of coronary heart disease incidence with carotid arterial wall thickness and major risk factors: the Atherosclerosis Risk in Communities (ARIC) Study, Am J Epidemiol 1997;146: [12] O Leary DH, Polak JF, Kronmal RA, Manolio TA, Burke GL, Wolfson SK. Carotid-artery intima and media thickness as a risk factor for myocardial infarction and stroke in older adults. New Engl J Med 1999;340: [13] Vartiainen E, Puska P, Jousilahti P, Korhonen HJ, Tuomilehto J, Nissinen A. Twenty-year trends in coronary risk factors in North Karelia and in other areas of Finland. Int J Epidemiol 1994;23: [14] Ylitalo K, Porkka KVK, Meri S, et al. Serum complement and familial combined hyperlipidemia. Atherosclerosis 1997;129: [15] Porkka KVK, Nuotio I, Pajukanta P, et al. Phenotype expression in familial combined hyperlipidemia. Atherosclerosis 1997;133: [16] Wofford JL, Kahl FR, Howard GR, McKinney WM, Toole JF, Crouse JR III. Relation of extent of extracranial carotid artery atherosclerosis as measured by B-mode ultrasound to the extent of coronary atherosclerosis. Arterioscler Thromb 1991;11: [17] Craven TE, Ryu JE, Espeland MA. Evaluation of the associations between carotid artery atherosclerosis and coronary artery stenosis. A case control study. Circulation 1990;82: [18] Selzer RH, Hodis HN, Kwong-Fu H, Mack WJ, Lee PL, Liu C, Liu C. Evaluation of computerized edge tracking for quantifying intima-media thickness of the common carotid artery from B- mode ultrasound images. Atherosclerosis 1994;111:1 11. [19] Taskinen M-R, Kuusi T, Helve E, Nikkilä E, Yki-Järvinen H. Insulin therapy induces antiatherogenic changes of serum lipoproteins in noninsulin-dependent diabetes. Arteriosclerosis 1988;8:168. [20] Miles J, Glasscock R, Aikens J, Gerich J, Haymond M. A microfluorometric method for the determination of free fatty acids in plasma. J Lipid Res 1983;24:96 9. [21] Sharrett AR, Patsch W, Sorlie PD, Heiss G, Bond MG, Davis CE. Associations of lipoprotein cholesterols, apolipoproteins A-I and B, and triglycerides with carotid atherosclerosis and coronary heart disease. The Atherosclerosis Risk in Communities (ARIC) Study. Arterioscler Thromb 1994;14: [22] Veerkamp M, de Graaf J, Bredie S, Hendriks J, Stalenhoef A. Diagnosis of familial combined hyperlipidaemia (FCH) based on lipid phenotype expression in 32 families: results of a 5-year follow-up study. Eur J Clin Invest 2001;31:20. [23] Demacker PN, Veerkamp MJ, Bredie SJ, Marcovina SM, de Graaf J, Stalenhoef AF. Comparison of the measurement of lipids and lipoproteins versus assay of apolipoprotein B for estimation of coronary heart disease risk: a study in familial combined hyperlipidemia. Atherosclerosis 2000;153: [24] Lamarche B, Moorjani S, Lupien PJ, Cantin B, Bernard PM, Dagenais GR, Despres J-P. Apolipoprotein A-I and B levels and the risk of ischemic heart disease during a five-year follow-up of men in the Quebec cardiovascular study. Circulation 1996;94: [25] Sniderman AD, Cianflone K. Measurement of apoproteins: time to improve the diagnosis and treatment of the atherogenic dyslipoproteinemias. Clin Chem 1996;42: [26] Lamarche B, Tchernof A, Mauriege P, Cantin B, Dagenais GR, Lupien PJ, Despres J-P. Fasting insulin and apolipoprotein B levels and low-density lipoprotein particle size as risk factors for ischemic heart disease. J Am Med Assoc 1998;279: [27] Salonen R, Salonen JT. Determinants of carotid intima-media thickness: a population-based ultrasonography study in Eastern Finnish men. J Int Med 1991;229: [28] Arnett DK, Tyroler HA, Burke G, Hutchinson R, Howard G, Heiss G. Hypertension and subclinical carotid artery atherosclerosis in blacks and whites. The Atherosclerosis Risk in Communities Study. Arch Intern Med 1996;156: [29] Franklin SS, Sutton-Tyrrell K, Belle SH, Weber MA, Kuller LH. The importance of pulsatile components of hypertension in predicting carotid stenosis in older adults. J Hypertens 1997;15: [30] Temelkova-Kurktschiev T, Koehler C, Schaper F, et al. Relationship between fasting plasma glucose, atherosclerosis risk factors and carotid intima media thickness in non-diabetic individuals. Diabetologia 1998;41: