A total of 873 participants (573 men and 300 women) aged 35 - 69 years who enrolled in the baseline survey of a prospective cohort study in Tokushima Prefecture, Japan from November 2009 to January 2012 and who underwent ba-PWV measurement at the baseline survey were included in this cross-sectional study. The subjects were mostly office workers. This study was conducted as part of the prospective cohort study, that is named as the Japan Multi-Institutional Collaborative Cohort (J-MICC) Study, as described previously [10]. In brief, the aim of the J-MICC Study is to examine the associations of lifestyle and genetic factors and their interactions with lifestyle-related diseases. All participants in the J-MICC Study provided written informed consent. The ethics committees of Nagoya University School of Medicine (the affiliation of the former principal investigator [Nobuyuki Hamajima]), Aichi Cancer Center (the affiliation of the current principal investigator [Hideo Tanaka]), and the University of Tokushima Graduate School approved the protocol of this study.

Individual lifestyle characteristics over the past year including medical history and treatment for diseases, smoking habit, alcohol drinking status, and leisure-time exercise were obtained through a structured self-administered questionnaire; trained staffs reviewed the accuracy of the mention contents to the questionnaire at the survey.

Leisure-time exercise was estimated based on the International Physical Activity Questionnaire (IPAQ) [11]. Exercise was divided into three levels: light exercise such as walking or hiking, moderate exercise such as light jogging or swimming, and vigorous exercise such as marathon running or competitive sports. The degrees of leisure-time exercise for the 3 levels were expressed as MET-hours per week (MET level x hours of activity x events per week) and summed [12]. In this estimation, 3.4 METs was assigned for light exercise, 7.0 METs for moderate exercise, and 10.0 METs for vigorous exercise.

Additionally, we obtained information regarding a firstdegree family history of diabetes (i.e., positive, negative, or unknown). When analyzed, a response of “positive” was regarded as having a family history of diabetes, whereas “negative” and “unknown” were regarded as not having it.

Body height was obtained from the questionnaire and body weight was measured to the nearest 0.1 kg at the survey. Body mass index was calculated as weight (in kg) divided by height (in m) squared. Venous blood was aspirated from each participant and serum was separated within 3 hours. Serum lipid levels were measured at an external laboratory (BML Inc., Tokyo, Japan). Total cholesterol and triglycerides were determined by an enzyme assay, and high-density lipoprotein (HDL) cholesterol was determined by a direct method.

ba-PWV was measured using a waveform analyzer (model BP-203RPE III; Colin, Co. Ltd., Komaki, Japan) as described previously [13]. Briefly, the subject was examined while resting in the supine position in an air-conditioned room. Extremity blood pressure was measured using an oscillometric method, and the ankle-brachial pressure index (ABI) was automatically calculated. Heart rate was recorded simultaneously. ba-PWV was calculated by time-phase analysis between the right brachial artery pressure and volume waveforms at both ankles. To reduce inter-observer variation, all ba-PWV measurements were performed by a single researcher throughout the study. Individual ba-PWV and ABI data are expressed as the means of the bilateral ba-PWV and ABI, respectively.

Among the 873 participants (573 men and 300 women) initially included in this cross-sectional study, we excluded 68 with a history of ischemic heart disease (n = 17), stroke (n = 9) and diabetes (n = 52). We excluded another 5 subjects who had a low right or left ABI (ABI ≤ 0.9), which suggested peripheral arterial occlusive disease and their ba-PWV values might be unreliable. After excluding an additional 13 subjects for whom serum lipid level data were missing, 787 non-diabetic subjects (502 men and 285 women) were included for analysis.

Continuous variables are expressed as mean ± standard deviation, whereas those with skewed distribution are expressed as median (25 percentile, 75 percentile). Categorical variables are expressed as proportion (%). Student’s t-test, Wilcoxon rank sum test, or Fisher’s exact test was used to compare the baseline characteristics between sexes and between subjects with and without a family history of diabetes. Prevalent hypertension was defined as systolic blood pressure ≥140 mmHg or diastolic blood pressure ≥90 mmHg, or antihypertensive agent use. Hypercholesterolemia was defined as serum total cholesterol ≥220 mg/dl or receiving treatment for hypercholesterolemia, and low HDL cholesterol was as serum HDL cholesterol <40 mg/dl. Elevated triglycerides was defined as serum triglycerides ≥150 mg/dl.

We used analysis of covariance to evaluate the associations between a family history of diabetes and baPWV values, in men and women separately, after adjusting for the probable covariates. Adjusted covariates were as follows: 1) age (continuous) and systolic blood pressure (<120, 120 to <140, 140 to <160, or ≥160 mmHg with no medical treatment, or anti-hypertensive agent use) (model 1), 2) age, systolic blood pressure, body mass index (kg/m², quartiles), smoking habit (current, past, and never), alcohol drinking status (current, past, and never), leisure-time exercise (MET-hours/week, quartiles), hypercholesterolemia (no/yes), Low HDL cholesterol (no/yes), elevated triglycerides (no/yes), and heart rate (continuous) (model 2). Additionally, we evaluated the association between a family history of diabetes and high ba-PWV by logistic regression in all non-diabetic subjects adjusting for sex with the same covariates in the analysis of covariance. High ba-PWV was defined as a value exceeding the sex-specific median value (1397 and 1213 cm/s in men and women, respectively). In the logistic regression analyses, categorical variables were converted into dummy variables, and these dummy variables except for reference categories were included in the model.

All calculations and statistical tests were performed using SAS, version 8.2 (SAS Institute Inc., Cary, NC, USA). All statistical tests were based on 2-sided probabilities, and P values <0.05 were considered statistically significant.

Age was significantly and positively correlated with baPWV in both sexes (r = 0.458, P < 0.001 in men, and r = 0.553, P < 0.001 in women). Systolic blood pressure was also strongly and positively correlated with ba-PWV in both sexes (r = 0.666, P < 0.001 in men, and r = 0.710, P < 0.001 in women).

Table 1 shows the baseline characteristics of the nondiabetic subjects by sex. Men showed significantly higher years of ages, body mass index, leisure-time exercise level, systolic and diastolic blood pressure, serum triglyceride level, and significantly lower serum HDL cholesterol level than women. Men also exhibited significantly higher rates of current smoking and drinking, and higher prevalence of hypertension, low HDL cholesterol, and elevated triglycerides than women. The ba-PWV values were considerably higher in men than in women. The prevalence rate of a family history of diabetes was not different between sexes.

Table 2 presents the baseline characteristics of the subjects with and without a family history of diabetes in non-diabetic subjects by sex. In men, although serum triglyceride level was significantly higher in subjects with a family history of diabetes than those without that trait, the other characteristics including ba-PWV values were not different between subjects with and without a family history of diabetes. In women, subjects with a family history of diabetes showed significantly lower prevalence of hypertension and low HDL cholesterol, and significantly lower systolic and diastolic blood pressure, and ba-PWV values.