Study Population

Consecutive patients with FH meeting the following criteria were recruited from our outpatient clinic: 1) genetic diagnosis, 2) fasting LDL-C ≥ 4.9 mmol/L (190 mg/dL), or 3) fasting LDL-C ≥ 4.1 mmol/L (160 mg/dL) with at least 1 biological parent with FH and a family history of premature atherosclerotic disease defined as onset of clinical atherosclerotic disease before age 55 years in men or age 65 years in women in first- or second-degree relatives. Subjects were excluded if they had signs or symptoms of diabetes or cardiovascular disease, a smoking history, or received lipid-lowering medication. Healthy control subjects recruited within the same age range as the patients from the family and friends of hospital staff served as controls. Height and weight were measured in each patient by the same observer using a wall-mounted stadiometer and electronic scale. Body mass index was calculated as kilograms/meters squared. Systolic and diastolic blood pressures were measured in the right arm with a Dinamap automated vital signs monitor (Criticon Inc., Tampa, FL). The study was approved by the human research ethics committee at Nihon University School of Medicine, and all subjects or their parents gave informed consent.

Ultrasound Examination

The cervical carotid arteries were examined using an SSA-660A ultrasound scanner (Toshiba, Tokyo, Japan) equipped with a 7-11 MHz extended-frequency linear probe focused at the level of the far wall in the common carotid artery (CCA). Scan settings (maximum dynamic range 65 dB, gain 80 dB, post-process 2 or 4, edge enhance 1) were preset at machine start-up and not altered during the examinations. The axial resolution of the M-mode system was 0.01 mm. Longitudinal images of CCA were obtained by routine B-mode ultrasound imaging of subjects in the supine position with the head turned slightly to the left under a condition of constant two-dimensional Doppler gain. The IMT of the CCA far wall was measured with the electronic calipers of the machine. The mean IMT was calculated for each subject as the average of three consecutive maximal far wall thickness measurements obtained in the CCA 1 to 2 cm proximal to the bulb ( Figure 1 A ). In our laboratory, the coefficients of variance for interobserver and intraobserver variability for the mean IMT were 7.3% and 5.0%, respectively.

(A) Conventional B-mode image of CCA. Rectangle indicates the magnified portion around the far wall of CCA. (B) Maximal magnified B-mode image (×8.6). Rectangle indicates the ROI in the far wall of the IMC. (C) Histogram of ROI. This feature was automatically extracted by Image J 1.38x software (National Institutes of Health, Bethesda, MD). (D) Angle of measurement in gray-level co-occurrence matrix. Dashed lines represent the direction of resolution in gray-level co-occurrence matrix, and 90 degrees indicates the vertical direction of resolution in the ROI in this study. CCA, Common carotid artery; ROI, region of interest.

A single frame of CCA from a cine-loop was selected in the cardiac diastole when the arterial lumen diameter was at a minimum. All digital images (unmagnified and magnified) were stored on magneto-optical disks as Tagged Image File Format files (resolution of 576 pixels [height] × 768 pixels [width]) and subsequently transferred to a PC and normalized for a gray scale from 0 (black) to 255 (white). Blood was used as a reference for black, and the adventitia was the reference for white. All textural characteristics of carotid IMC were evaluated off-line by Image J 1.38x software (National Institutes of Health, Bethesda, MD), including a set of first-order statistics (mathematic descriptors of the shape of the frequency distribution of gray-level histogram) and second-order statistics (mathematic descriptors of the spatial distribution of gray level within the region of interest [ROI]) parameters. The ROI was placed manually over a 0.5 cm length of the far wall IMC, which was evaluated for mean IMT ( Figure 1 B). Basically, as a set of first-order statistics, gray-scale median, skewness, and kurtosis represent the gray-scale median, the third-order moment about the mean, and the fourth-order moment about the mean, respectively. Gray-scale median represents the median of the frequency distribution of gray tones of the pixels included in the ROI in a scale of 256 gray tones, that is, dark (hypoechoic) regions were associated with a gray-scale median that tends to approach 0, whereas bright (hyperechoic) regions were associated with a gray-scale median that tends to approach 255 ( Figure 1 C). For a set of second-order statistics, which are called the gray-level co-occurrence matrix demonstrated by Haralick et al. , we used an interpixel distance (d) of 1 (i.e., all pixels) and an angle of 90 degrees ( Figure 1 D). The following features were calculated: entropy, angular second moment, and contrast, which represent the disorder or complexity of the texture, uniformity of texture, and change in density, respectively. Entropy is larger for surfaces that do not have a uniform texture. Angular second moment is a measure of local homogeneity and the opposite of entropy. Contrast is a measure of the amount of local variation in the images. The mathematic definitions of these texture parameters have been described.

In addition, we investigated the changes in textural characteristics of the far wall IMC on visual inspection of maximal magnified B-mode images (×8.6) over a 0.5-cm length within a ROI. Typically, longitudinal images of the carotid arterial wall on the B-mode ultrasound demonstrate two parallel echogenic lines separated by a relatively hypoechoic central region, referred to as the “IMC.” The more proximal line has been termed the “lumen-intima interface,” and the distal line has been termed the “media-adventitia interface.” Changes in textural characteristics of the IMC were scored using visual scoring, which is a modification of the classification proposed by Belcaro et al. Three score patterns for visual scoring were defined as follows: 1, normal wall (intima media and adventitia clearly separated); 2, proximal interface disruption (appearance of hypoechoic components in the proximal echogenic layer); 3, intima-media granulation (mosaic pattern with mixed hypoechoic and echogenic components in IMC) ( Figure 2 ). Evaluation was performed by consensus between two independent observers who were blinded to the subject status by a randomized code covering all of the data. A third observer resolved any disagreement in evaluations between the two observers. In our laboratory, interobserver and intraobserver agreement studies for visual scoring using repeated recordings demonstrated Kappa coefficients of 0.78 and 0.89, respectively.

Representative typical visual scoring by magnified B-mode ultrasound images of the IMC. Rectangle represents the ROI in the far wall. Values of mean IMT, angular second moment, and entropy are indicated in each case. In this study, medial infiltration was defined as the findings over visual scoring 2. (A) Control. (B) Visual scoring 1: normal wall (intima media and adventitia clearly separated). (C) Visual scoring 2: proximal interface disrupted by hypoechoic components in the proximal echogenic layer. (D) Visual scoring 3: intima-media granulation with mixed hypoechoic and echogenic components. IMT, Intima-media thickness.

The patients were categorized into three groups according to the mean IMT based on differences from our age-matched normal control values for IMT (group I, with normal IMT (SD < +1.5): IMT < 0.55 mm; group II, with moderately increased IMT (+1.5 ≤ SD ≤ +2.5): 0.55 mm ≤ IMT ≤ 0.60 mm; group III, with markedly increased IMT (+2.5 < SD): 0.60 mm < IMT). Lipid profile, anthropometric variables, textural parameters, and visual scoring were assessed. These variables were compared among the groups.

Laboratory Methods

Venous blood samples were obtained in the morning after an overnight fast. Serum total cholesterol, triglyceride, and high-density lipoprotein cholesterol were determined using conventional enzymatic methods. The concentration of LDL-C was calculated using the Friedewald formula. Functional LDL receptor activity was measured in peripheral lymphocytes via flow cytometry using fluorescent-labeled LDL.

Data Analysis

Data are presented as the mean ± SD. To compare differences in variables among the four groups, one-way analysis of variance was performed. When values of P < .05 were obtained on analysis of variance, multiple comparisons between the two groups were performed by post hoc test (Scheffé’s F). For categoric variables, chi-square test was performed. Stepwise multiple linear regression analysis was used to identify significant determinants of IMT. Two-sided values of P < .05 were considered significant. All statistical analyses were performed using SPSS version 10.0 (SPSS Inc., Chicago, IL).