The aim of this study is to: (1) investigate the similarity and differentia of risk factors of the two diseases, and (2) to explore serum-based diagnostic biomarkers to predict AAA in patients with PAD.

Six hundred and eighty-four patients who were diagnosed with AAA and/or PAD at the department of vascular surgery, Xiangya Hospital, Central South University from January 2010 to May 2019 were enrolled in this study. Patients can be divided into different groups based on diagnosis: including the AAA only group (n = 320), in which patients were diagnosed with AAA, but not with PAD; the PAD only group (n = 320); in which patients were diagnosed with PAD but not with AAA; and the PAD + AAA group (n = 44), in which patients were diagnosed with both AAA and PAD (Figure 1). All procedures were approved by the Institute Review Board of Xiangya Hospital, Central South University (IRB No. 201905267). Patients were informed, and consent forms were obtained for research purposes.

Including criteria for AAA: (1) Age >40 years old; (2) AAA was confirmed by CTA or MRA.

The exclusion criteria are as follows: (1) infectious AAA; (2) secondary AAA with certain causes (such as injury); (3) severe mental illness leading to incapability of cooperation; (4) severe infectious or autoimmune diseases; (5) pregnancy or malignant tumor; and (6) thrombotic diseases such as venous thromboembolism.

The inclusion criteria for PAD were: (1) Age > 40 years; (2) High-risk factors such as smoking, diabetes, hypertension, hyperlipidemia, and others; (3) Symptoms of the ischemic lower extremity; (4) The pulse of the distal arteries of the ischemic limb is weakened or disappeared; (5) Ankle-brachial index (ABI) of ≤ 0.9; (6) Artery stenosis or occlusion were confirmed by CTA or ultrasound. The clinical diagnosis of PAD can be made if four of the above diagnostic criteria were met.

The exclusion criteria of PAD are as follows: (1) severe infectious or connective tissue diseases; (2) pregnancy or malignant tumor; (3) severe mental illness leading to the incapability of cooperation; and (4) thrombotic diseases such as venous or arterial thromboembolism.

Demographic characteristics, history, and personal history were recorded. Peripheral blood was collected at the time of administration. Specifically, levels of D-dimer were obtained at the time of admission by enzyme-linked immunosorbent assay. Laboratory tests were performed and data were collected.

Kolmogorov-Smirnov test was used to determine whether quantitative data were normally distributed. Quantitative data are shown as a mean ± standard deviation (SD) for normally distributed variables or median (interquartile range) for non-normally distributed variables. Qualitative data are presented as relative frequencies. Chi-square test or Fisher exact test was performed among qualitative data, and Student's t-test or Mann Whitney U-test was conducted among quantitative data. Multivariate logistic regression analysis was utilized to determine the differential distribution of biomarkers. Receiver operating characteristics (ROC) curves and the area under the curve (AUC) were performed to further assess the diagnostic efficacy of biomarkers. All statistical analyses were performed using the Statistical Product and Service Solutions (SPSS) 22.0 statistical software package (IBM Analytics, Armonk, NY, USA).

Demographic and medical history were recorded. To screen the potential differentially distributed parameters in patients with AAA and patients with PAD, statistical analysis was performed in AAA only group (n = 320) and PAD only group (n = 320). As shown in Table 1, the incidence of diabetes mellitus decreased significantly in AAA only group compared to PAD only group (11.25 vs. 31.56%, P < 0.01). No obvious difference in age, gender, hypertension, dyslipidemia, coronary heart disease, cerebral infarction, smoking, or drinking was founded between AAA only group and PAD only group (P > 0.05).

To screen the potential differentially distributed serum markers among patients with PAD and patients with AAA, peripheral blood samples were collected at the time of admission, and laboratory tests were performed (liver function, renal function, lipids, coagulation, and platelet count). As shown in Table 2, levels of total bilirubin, direct bilirubin, and D-dimer were increased, while levels of fibrinogen and platelet were decreased significantly in AAA only group when compared with those in PAD only group (P < 0.001). No obvious difference in TBA, BUN, SCR, uric acid, fasting glucose, low-density lipoprotein cholesterol, high-density lipoprotein cholesterol, triglyceride, total cholesterol, PT, or hemoglobin were found between AAA only group and PAD only group (P > 0.05). This suggests that patients with AAA may experience relatively high levels of bilirubin and D-dimer, while patients with PAD may show relatively high levels of fibrinogen and platelet.

The absolute value of circulating parameters in Table 2 cannot fully reflect the health status of patients since whether those markers are in the normal range is more clinically relevant. Thus, patients in the AAA only group and PAD only group are sub-grouped into a normal or abnormal group (Table 3). Statistical analysis was carried out to validate the association of variables with AAA and PAD. As shown in Table 3, patients with AAA-only manifested a higher possibility of having an abnormal status of D-dimer, fibrinogen, and platelet count, and a lower possibility of diabetes mellitus when compared with patients with PAD-only (P < 0.001). No obvious difference in the abnormal status of total bilirubin and direct bilirubin was identified between AAA only group and PAD only group (P > 0.05), which is reasonable due to the median (Q1/Q3) value of total bilirubin and direct bilirubin in Table 2 are in the normal range (normal range for total bilirubin and is ≤ 17.1 and ≤ 6.8, respectively).

Together, the above results suggested that diabetes mellitus, D-dimer, fibrinogen, and platelet count were differentially distributed in patients with AAA when compared to patients with PAD.

The AAA and PAD share common risk factors including atherosclerosis, age, gender (male), smoking and hypertension, and others. This makes it clinically realistic that a patient with PAD is concomitant with AAA, which is easily overlooked by clinicians, although management strategies are altered totally. To find out variants to distinguish between patients with AAA and patients with PAD, multivariable logistics regression analysis was performed in the AAA group (n = 364, Figure 1B) and PAD group (n = 364, Figure 1B). Results in Table 4 indicated that total bilirubin, direct bilirubin, D-dimer, fibrinogen, and platelet count could be used to distinguish between patients with AAA and patients with PAD (P < 0.05).

To further evaluate the diagnostic efficacy of the mentioned markers, the receiver operating characteristic (ROC) curve of the five risk factors mentioned above was performed in the AAA group (n = 364, Figure 1B) and PAD group (n = 364, Figures 1B, 2). Figure 2F summarizes detailed information on ROC curves in Figure 2. As indicated in Figure 2F, D-dimer may be the most optimal biomarker for the distinction between patients with AAA and patients with PAD with the area under the curve (AUC) of 0.7034.

As mentioned above, the main objective of this study was to find out ideal biomarkers for the prediction of AAA in patients with PAD. Total bilirubin, direct bilirubin, D-dimer, fibrinogen, and platelet may potentially be predictive biomarkers of AAA, though OR value or AUC was not that promising between patients with AAA (n = 364) and patients with PAD (n = 364). To further validate the efficacy of these biomarkers in a more clinically realistic condition, statistical analysis was performed again between the AAA + PAD group (n = 44) and the PAD group (n = 364).

As shown in Table 5, multivariate logistic regression analysis showed that the serum levels of D-dimer increased in AAA + PAD group compared with the PAD group (OR: 2.630, 95% CI:1.639–4.221; P < 0.001). No significant difference in fibrinogen and platelet count was found between the AAA + PAD group and the PAD group (P > 0.05).

Furthermore, the ROC curve of the D-dimer was carried out (Figure 3), and AUC was 0.867. The Youden index suggested that the best cutoff value of D-dimer was 0.675 mg/L (Table 6 with partial data provided).

To better visualize the predictive efficacy of D-dimer for AAA, the 364 patients with PAD were divided into two groups based on D-dimer: D-dimer > 0.65 mg/L, and D-dimer ≤ 0.65 mg/L. As shown in Figure 4, 41.46% of patients with PAD were concomitant with AAA when D-dimer was > 0.65 mg/L, while only 3.55% of patients with PAD were concomitant with AAA when D-dimer was ≤ 0.65 mg/L.