Study participants characteristics

General characteristics of the 13,144 participants by gender were shown in Table 1. The mean age was 52.2 ± 13.1 years old. 7181 subjects of them (54.6%) were man. The average BMI of all participants was 25.2 ± 3.81 kg/m2, and the average WC was 88.5 ± 11.7 cm respectively. The average ABI was 1.08 ± 0.09, and the average IABPD was 3.55 ± 3.79 mmHg respectively. Furthermore, the prevalence of ABI ≤ 0.9 and IABPD ≥ 15 mmHg in this study population was 1.9% and 1.4% respectively.

Table 1 Clinical characteristics of study participants according to genderBMI and WC values according to different ABI and IABPD categories

The BMI and WC according to different ABI and IABPD categories were calculated and compared. WC was significantly higher in subjects with ABI ≤ 0.9 than that in subjects with ABI > 0.9 (P < 0.001, Table 2). However, BMI was not significantly different in subjects with ABI ≤ 0.9 and with ABI > 0.9 (P = 0.844, Table 2). At the same time, the WC and BMI were significantly higher in subjects with IABPD ≥ 15 mmHg than those in subjects with IABPD < 15 mmHg respectively (both P < 0.001, Table 2).

Table 2 BMI and WC values according to different ABI and IABPD categoriesIndependence of BMI and WC associated with different ABI and IABPD categories

In order to evaluate the independence of BMI and WC associated with different ABI and IABPD categories, multiple logistic regression analysis was used to calculate the OR and 95% CI of BMI and WC associated with ABI ≤ 0.9 and IABPD ≥ 15 mmHg respectively with adjustment for other potential confounders including age, men, smoking, drinking, hypertension, diabetes mellitus, lipid disorders, and chronic kidney disease. These indicators of obesity entered regression equation as continuous variables respectively. We found that WC was independently associated with ABI ≤ 0.9 (OR 1.014, 95% CI 1.002–1.026, P = 0.017, Table 3). Nevertheless, BMI was not independently associated with ABI ≤ 0.9 using this multiple logistic regression analysis. At the same time, the data showed that BMI and WC were independently associated with IABPD ≥ 15 mmHg respectively (OR 1.139, 95% CI 1.100–1.181, P < 0.001, and OR 1.058, 95% CI 1.044–1.072, P < 0.001, Table 3).

Table 3 Independence of BMI and WC associated with ABI ≤ 0.9 and IABPD ≥ 15 mmHgPrevalence of ABI ≤ 0.9 and IABPD ≥ 15 mmHg with different categories of BMI

As we mentioned in the above section, though we cannot discover a linear relationship between BMI and ABI statistically, we still try to explore the prevalence of ABI ≤ 0.9 in study subjects when they were categorized as four groups according to BMI (< 20, 20 to < 25, 25 to < 30, and ≥ 30). As a result, we found that prevalence of ABI ≤ 0.9 was displayed with a U-shaped pattern according to different BMI categories (Fig. 2). Prevalence of ABI ≤ 0.9 in subjects with BMI < 20 and BMI ≥ 30 was significantly higher compared with that in subjects with BMI 20 to < 25 respectively (both P < 0.001).

Fig. 2

Prevalence of ABI ≤ 0.9 and IABPD ≥ 15 mmHg in different categories of BMI. ABI: ankle-brachial index; BMI: body mass index; IABPD: interarm systolic blood pressure difference. N = 925 for BMI < 20, 5643 for BMI ≥ 20 to < 25, 5203 for BMI ≥ 25 to < 30, and 1373 BMI ≥ 30. Prevalence of ABI ≤ 0.9 in subjects with BMI < 20 and BMI ≥ 30 was significantly higher compared with that in subjects with BMI ≥ 20 to < 25 respectively (both P < 0.001). Prevalence of IABPD ≥ 15 mmHg was significantly increased with incremental BMI (P for trend < 0.001)

At the same time, we also tried to observe the prevalence of IABPD ≥ 15 mmHg when study subjects were categorized as four groups according to BMI. A different trend was discovered that prevalence of IABPD ≥ 15 mmHg was significantly increased with incremental BMI (P for trend < 0.001, Fig. 2).

Relationship between BMI and abnormalities of peripheral arteries

The above data showed that, unlike WC, relationship between BMI and abnormalities of peripheral arteries appeared to be different. Thus, we further carefully evaluated whether various BMI categories (< 20, 20 to < 25, 25 to < 30, and ≥ 30) were associated with ABI ≤ 0.9 and IABPD ≥ 15 mmHg using multiple logistic regression analysis. The data displayed that, compared with BMI 20 to < 25, the risk of ABI ≤ 0.9 was significantly increased when BMI < 20 or ≥ 30 respectively (OR 2.595, 95% CI 1.745–3.858, P < 0.001, and OR 1.618, 95% CI 1.087–2.410, P = 0.018, Table 4). However, the risk of IABPD ≥ 15 mmHg tended to be increased when participants had bigger BMI. Compared with BMI 20 to < 25, the risk of IABPD ≥ 15 mmHg was significantly increased when BMI ≥ 30 (OR 3.218, 95% CI 2.133–4.855, P < 0.001, Table 4).

Table 4 Various categories of BMI associated with ABI ≤ 0.9 and IABPD ≥ 15 mmHg

Furthermore, we also explored the nonlinear relationship between BMI and the risk of ABI ≤ 0.9 using a restricted cubic spline model by multivariable adjustment. Restricted cubic spline analysis (Fig. 3) indicated a significant U-shaped relationship between BMI and the risk of ABI ≤ 0.9 (P for non-linearity < 0.001).

Fig. 3

Nonlinear relationship between BMI and the risk of ABI ≤ 0.9. Y-axis stands for the odds ratio (OR) and 95% CI of body mass index (BMI) (independent variable) associated with ankle-brachial index (ABI) ≤ 0.9 using restricted cubic spline analysis, with adjustment for other potential confounders including age, men, smoking, drinking, hypertension, diabetes mellitus, lipid disorders, and chronic kidney disease. A significant U-shaped relationship between BMI and the risk of ABI ≤ 0.9 was exhibited (P for non-linearity < 0.001)