Ankle‐brachial index, arterial stiffness, and biomarkers in the prediction of mortality and outcomes in patients with end‐stage kidney disease

Abstract Background Although ankle‐brachial index (ABI) and brachial‐ankle pulse wave velocity (baPWV) are significant predictors of major adverse cardiovascular event (MACE), their prognostic value in association with biomarkers has not been fully evaluated in patients with end‐stage kidney disease (ESKD). Hypothesis We hypothesized that ABI/baPWV would provide better prognostic value independent of biomarkers in ESKD patients. Methods This study included 104 ESKD patients treated with maintenance hemodialysis who underwent ABI and baPWV examinations and laboratory tests, including brain‐natriuretic peptide, high‐sensitive cardiac troponin T (hs‐cTnT), and high‐sensitive C‐reactive protein (hs‐CRP). MACE was defined as a composite event of all‐cause death, acute coronary syndrome, and stroke. Results During a mean follow‐up of 3.6 ± 1.7 years, a total of 51 MACE were observed. The independent factors associated with MACE were age >75 years (adjusted hazard ratio [HR], 2.15; P < .05), abnormal ABI (adjusted HR, 2.01; P < .05), left ventricular ejection fraction (LVEF) <50% (adjusted HR, 3.33; P < .001), the upper tertile of hs‐cTnT (adjusted HR, 2.77; P < .05), and hs‐CRP (HR, 1.96; P < .05). However, baPWV did not remain as an independent predictor of MACE in the entire cohort and also in patients without abnormal ABI. The combination of predictors improves the predictive value of MACE, providing increased HR with 4.00 for abnormal ABI + hs‐CRP, 4.42 for abnormal ABI + hs‐cTnT, and 7.04 for abnormal ABI + LVEF <50% (all P < .001). Conclusion Abnormal ABI is a robust predictor of MACE independent of biomarkers and their combination provides better risk stratification compared with a single predictor in ESKD patients.


| INTRODUCTION
End-stage kidney disease (ESKD) is characterized by multiple risk factors which pose increased risks of mortality and major adverse cardiovascular events (MACE). 1,2 The mechanisms underlying the association between ESKD and MACE include highly prevalent comorbidities such as coronary artery disease, cerebrovascular disease, and peripheral artery disease (PAD). [2][3][4] Accumulating evidence suggests that arterial stiffness and anklebrachial index (ABI) are well correlated with atherosclerotic burden and are strong predictors of MACE. [5][6][7][8] Among the various indices of arterial stiffness, brachial-ankle (ba) pulse wave velocity (PWV) is an established tool for the noninvasive assessment of arterial stiffness in clinical practice. 9,10 However, measurements of baPWV can be influenced by the presence of PAD and hemodynamic condition after hemodialysis (HD) session. [11][12][13] Therefore, the prognostic value of baPWV in ESKD patients is still discussed. 14 With the widespread clinical application of standardized assays, several biomarkers and multiple biomarker approach enable the accurate prediction of the long-term prognosis of ESKD patients treated with HD. [15][16][17] Previous studies have demonstrated that cardiac troponin (cTn), brain-natriuretic peptide (BNP), and C-reactive protein (CRP) are the robust predictors of MACE independent of traditional risk factors and two-dimensional thoracic Doppler echocardiography (TTDE) findings. [18][19][20][21] However, the prognostic significance of ABI/PWV in association with biomarkers have not been fully investigated in ESKD patients. In addition, the heterogeneous risks of ESKD patients might require risk assessment approach from multiple aspects. The aim of this study was to investigate the prognostic value of ABI/PWV in association with biomarkers and to examine whether biomarkers would provide additional predictive value on ABI/PWV in ESKD patients treated with HD.

| Study population
During the period between April 2012 and February 2013, a total of 104 consecutive ESKD patients treated with HD for more than 1 year at the Ishikiri-seiki Hospital were enrolled in this study. The exclusion criteria were as follows: (a) patients who were hospitalized within the past 1 month due to acute coronary syndrome (ACS) or heart failure; (b) patients with systemic infection or chronic inflammatory diseases; and (c) patients with symptomatic severe valvular heart diseases requiring surgical intervention. The present study was approved by the Ethics Committee of Ishikiri-seiki Hospital and was conducted in accordance with the institutional regulations and the Declaration of Helsinki. All the study patients gave written informed consent for the enrollment in this study. All the patients were required to provide baseline plasma samples   before HD session on the day underwent HD session, including   hemoglobin, hemoglobin A1c, cholesterol profiles, BNP, high-sensitive cTnT (hs-cTnT), and high-sensitive CRP (hs-CRP). The samples were centrifuged within 60 minutes and stored at −80 C for further analysis. hs-cTnT was measured by an electro chemiluminescence immunoassay (ECLIA) with the EClusys hs-cTnT Roche diagnostic assay. In the hs-cTnT assay, the 99th percentile of the upper reference limit and the lower limit of detection were 0.014 and 0.003 ng/mL, respectively. BNP was measured with a specific immunoradiometric assay for human BNP (ARCHITECT BNP-JP; ABBOTT JAPAN Co, Ltd, Tokyo, Japan). The inter-and total coefficient variation for BNP was 1.1% to 5.1% and 2.3% to 5.3%, respectively. Hs-CRP was measured using N-latex CRP II (Dade Behring Inc., Marburg, Germany) with a coefficient variation of 3.1% at 0.5 mg/L. over the surface of the body. ABI was calculated as the ratio of the systolic blood pressure measured at the ankle to that at the brachial artery, with no blood access. The lower value of ankle blood pressure was used for the assessment of ABI. Abnormal ABI was defined as <0.90 or >1.40. 6 TTDE examination was also performed after the HD session at the study enrollment using a commercially available system (the Vivid 7; General Electric, Milwaukee, Wisconsin) by trained cardiac echo sonographers. Left ventricular ejection fraction (LVEF) was calculated by using Simpson method from apical 4-and 2-chamber views, as previously reported. 21 In the present study, LVEF <50% was defined as decreased LVEF.

| Patient characteristics and outcomes
Baseline patient characteristics are presented in Table 1

| Prognostic value of biomarkers and noninvasive examinations
In the entire group, univariate Cox hazard analysis showed that age  (Table 2).
Similarly, univariate and multivariable Cox hazard analysis were performed to investigate the prognostic value of baPWV in the normal ABI group (Table 3). In univariate analysis, statistically significant predictors of MACE were age >75 years (HR, 3.33; P = .002), LVEF   In the present study, we found that abnormal ABI was a strong predictor of MACE independent of biomarkers, whereas baPWV was not an independent predictor for the entire group and also for the patients with normal ABI. baPWV has been reported to be correlated with the direct measure of aortic PWV and carotid-femoral PWV, which has been shown to be a strong predictor of mortality in ESKD Note: The outcome measures of baPWV and biomarkers were divided into tertiles of the normal ABI group (ABI between ≥0.9 and ≤1.4; n = 73). Abbreviations: ABI, ankle-brachial index; ba-PWV, brachial-ankle-pulse wave velocity; BNP, brain-natriuretic protein; CI, confidence interval; HR, hazard ratio; hs-TnT, high-sensitive troponin T; LVEF, left ventricular ejection fraction; MACE, major adverse cardiovascular event; PWV, pulse wave velocity.

| DISCUSSION
patients. 10,23 However, because of technical issues such as measurement protocols and reproducibility, 22 the use of baPWV for the accurate risk prediction of ESKD patients is still discussed. 14 Ho-Ming, et al showed that the ba-PWV measurements were influenced by blood volume changes after HD sessions, whereas ABI was not. 11 These observations support the use of ABI for the risk stratification of ESKD patients treated with HD.
Furthermore, we found that the combination of upper tertile of hs-cTnT or hs-CRP in addition to abnormal ABI showed higher risk compared to abnormal ABI alone. A possible explanation is that elevations of these biomarkers represent the target organ damage and systemic inflammation, which is difficult to be evaluated by ABI alone.
Abnormal ABI has been shown to be well correlated with systemic disease burden of atherosclerosis, which is a strong predictor of MACE. 6 is a potential mechanism for increased risk of MACE.
The findings of this study suggest that the use of ABI rather than baPWV may provide prognostic information independent of other robust predictors such as decreased LVEF and biomarkers in this population. Considering its noninvasive nature and its common use in the clinical practice of ESKD patients, ABI and biomarker approach could be of importance to the management and the risk prediction of ESKD patients. Future large population studies are needed to determine if the combination of these independent predictors provides better risk prediction of MACE in ESKD patients treated with HD.

| STUDY LIMITATIONS
Our study has several limitations. First, this study included relatively small number of patients with and without normal ABI. Our findings regarding the predictors of MACE should be interpreted with caution, although baPWV did not remain as an independent predictor of MACE even in the normal ABI group. Second, we employed baPWV as a marker of arterial stiffness because of its common use in clinical practice. Different markers such as aortic stiffness and arterial stiffness gradient to assess arterial stiffness may contribute to the prediction of MACE in association with biomarkers and other noninvasive tests. 7 Finally, we evaluated the markers at a single time point of the study enrollment. Arterial stiffness is associated with organ damage which cannot be entirely explained by the established cardiovascular risk factors. 32 Several studies have suggested the potential benefit to assess changes in arterial stiffness as the treatment response. Future studies are needed to investigate the impact of targeting changes in arterial stiffness after optimized therapy in predicting MACE in ESKD patients.

| CONCLUSIONS
This study demonstrated that abnormal ABI was a robust predictor of MACE independent of biomarkers and decreased LVEF, whereas ba-PWV was not. The combination of the presence or absence of these predictors could provide better risk stratification compared with abnormal ABI alone in ESKD patients treated with HD.