Longitudinal association of carotid endothelial shear stress with renal function decline in aging adults with normal renal function: A population-based cohort study

The aim of this study was to investigate the associations between carotid wall shear stress (WSS) and renal function impairment (RFI) and albuminuria in aging adults. A total of 1,447 subjects aged 60 years and older with normal estimated glomerular filtration rate (eGFR ≥ 60 mL·min−1·1.72 m−2) and albumin/creatinine ratio (ACR < 30 mg·g−1) were enrolled between April 2007 and October 2009 in the Shandong area, China. Carotid WSS was assessed at baseline, and eGFR, which is based on serum creatinine and cystatin C, and ACR were assessed at baseline and at the annual follow-up visits. After an average of 62.9 months of follow-up, the reduction in eGFR and the increase in ACR were significantly higher in the Q1+2+3 group than the Q4 group, as classified by either the interquartile of the mean WSS or the interquartile of the peak WSS after adjustment for multi-variabilities, including the average blood pressures at every annual visit and baseline eGFR and ACR. For groups classified by mean WSS, the hazard ratios (95% confidence intervals) were 3.45 (1.36–8.75, p = 0.008) in the incident RFI and 3.24 3.22 (1.37–7.57, p = 0.009) in the incident albuminuria for the Q1+2+3 group compared with the Q4 group. Similar results were observed among groups classified by peak WSS. The Q1+2+3 group was associated with endothelial dysfunction and inflammation with respect to the Q4 group as classified by mean or peak WSS. The results indicate that carotid WSS plays an important role in RFI and albuminuria progression in aging adults. Lower WSS was associated with a higher risk of RFI and albuminuria compared with higher WSS.

Scientific RepoRts | (2019) 9:2051 | https://doi.org/10.1038/s41598-018-38470-x carotid artery (CCA) bifurcates, provide a common pathway through which precisely regulated systemic blood flows, and there is hemodynamic change 15,16 . The CCA is a well-established "observation window" for monitoring and measurement of systemic hemodynamic conditions in humans, thereby allowing precise regulation of blood flow 17 , and the carotid WSS may represent the overall hemodynamic condition of renal vessels 9,18,19 . Furthermore, chronic RFI is regarded as a part of a subclinical and generalized atherosclerotic-mediated vascular dysfunction 8,9 .
A previous study indicated that there was a close relationship between carotid WSS and renal function in aging adults 9 . Higher carotid WSS is related to higher eGFR and lower ACR. However, this study was a cross-sectional study and could not clearly illuminate the association between carotid WSS and renal function.
Based on this information, we hypothesized that baseline carotid WSS, serving as a marker of endothelial health, is associated with the development of future kidney disease and albuminuria. The main goal of this study was to conduct a community-based investigation of this association in an aging adult population.

Changes in eGFR and ACR over time.
In the Q 1+2+3 and Q 4 groups classified by MWSS, eGFR decreased and ACR increased at follow-up compared to the baseline. The reduction in eGFR and the increase in ACR were significantly higher in the Q 1+2+3 group than the Q 4 group after adjustment for age; sex; smoking; alcohol intake; histories of hypertension, diabetes, and dyslipidemia; medications of antihypertension, anti-dyslipidemia, glucose-lowering, and antiplatelet; and baseline blood pressures, blood lipids, fasting plasma glucose, and CCA-IMT (all adjustment p < 0.001, Fig. 1). However, the participants were grouped by PWSS, the reduction in eGFR and the increase in ACR were higher in the Q 1+2+3 group than the Q 4 group after adjustment of multi-variabilities (all adjustment p < 0.001, Fig. 1). study outcomes over time. Of the 1,434 participants, 54 (3.8%) developed RFI and 61 (4.3%) developed albuminuria over the follow-up period. The incidences of RFI and albuminuria were significantly higher in the Q 1+2+3 group than in the Q 4 group, which were grouped by MWSS after adjustment for multi-variabilities including baseline eGRF and ACR and the average blood pressures at every annual visit [for RFI: 49 (4.6%) vs. 5 (1.4%), p = 0.005 and for albuminuria: 55 (5.1%) vs. 6 (1.7%), p = 0.007]. The cumulative HRs (95% CIs) were 3.45 (1. 36-8.75, p = 0.008) in the incidence of RFI and 3.22 (1.37-7.57, p = 0.009) in the incidence of albuminuria for the Q 1+2+3 group compared with the Q 4 group after adjustment for multi-variabilities (Fig. 2). While the participants were grouped by PWSS, the incidences of RFI and albuminuria were 47 (4.4%) and 54 (5.0%) in the Q 1+2+3 group versus 7 (2.0%) and 7 (2.0%) in the Q 4 group (p = 0.048 and 0.017, respectively). The cumulative HRs (95% CIs) were 2.16 (0.99-4.71, p = 0.053) in the incidence of RFI and 2.49 (1.13-5.49, p = 0.024) in the incidence of albuminuria for the Q 1+2+3 group compared with the Q 4 group after adjustment for multi-variabilities (Fig. 2).
To investigate the effect of hypertension, dyslipidemia, and diabetes on the association between CCA WSS and the incidences of RFI and albuminuria, we stratified the participants in models ( Table 2). The cumulative HRs in the incidences of RFI and albuminuria were significantly higher in the Q 1+2+3 group compared with the Q 4 group as classified by either MWSS or PWSS (all p < 0.05), except for the cumulative HRs in the incidences of RFI between the Q 1+2+3 and Q 4 groups of the participants without dyslipidemia and diabetes as classified by PWSS (p = 0.083 and 0.053, respectively) after adjustment for multi-variabilities.

Changes in endothelial function and chronic inflammation over time.
In this study, we also assessed the differences in the changes in endothelial function and chronic inflammation over time between groups classified by MWSS and PWSS separately (Fig. 3). In the Q 1+2+3 and Q 4 groups, which were classified by MWSS and PWSS separately, serum NO levels decreased and serum ET-1, ICAM-1, VCAM-1, and hsCRP levels over the follow-up period significantly increased in the Q 1+2+3 group compared with the Q 4 group (all adjustment p < 0.001).
Changes in blood pressures, lipids, and fasting plasma glucose over time. Figure 4 shows the differences in the changes in blood pressure, and Fig. 5 depicts the differences in the changes in lipids and fasting plasma glucose over time. No significant differences between groups classified by MWSS and PWSS were observed (all p > 0.05).

Discussion
In this study, the major findings were that the lower mean and peak WSSs of CAA at baseline were significantly associated with (1) the subsequent reduction in eGFR and increase in ACR, (2) higher risks of the incidences of RFI and albuminuria, even after adjustment for confounders, and (3) the progression of vascular endothelial dysfunction and chronic low-grade inflammation.
In recent decades, the association between the local rheologic force of circulating blood and chronic kidney diseases has received increasing attention. Park 19 and Verbeke 20 reported that the mean and maximum wall shear rates of the brachial artery in healthy subjects were significantly higher than those in subjects with chronic kidney Scientific RepoRts | (2019) 9:2051 | https://doi.org/10.1038/s41598-018-38470-x disease and/or end-stage renal disease, even though their blood flow rates were similar. In the CCA, WSS was markedly lower in the prior-to-hemodialysis patients with end-stage renal failure than in the presumed healthy age-and sex-matched control subjects 21 . Our previous study showed that lower carotid artery WSS was closely correlated with lower eGFR and higher ACR 9 . The above-mentioned studies suggested that there is a close relationship between local WSS and chronic kidney diseases. However, it was difficult to illuminate the association between the local WSS and chronic kidney diseases in these cross-sectional studies.
In this study, we found that eGFR significantly decreased and ACR significantly increased over time with lower WSS compared with higher WSS, as classified by either MWSS or PWSS. After an average of 62.9 months of follow-up, the incidences of RFI and albuminuria were more than three-fold in the Q 1+2+3 MWSS group than the Q 4 MWSS group and more than two-fold in the Q 1+2+3 PWSS group compared with the Q 4 PWSS group, even after adjustment for confounders, including baseline CCA-IMT, eGFR, and ACR and the average blood pressure at each annual follow-up visit. Although there were no significant differences in the incidences of RFI in participants without dyslipidemia and diabetes between the Q 1+2+3 and Q 4 groups, the risks of RFI and albuminuria were higher in the Q 1+2+3 group than the Q 4 group as classified by MWSS and PWSS in participants with/without hypertension, with dyslipidemia, and with diabetes. Moreover, consistent with our previous study 9 , eGFR was significantly lower and ACR was higher in the Q 1+2+3 groups than in the Q 4 groups as classified by MWSS and PWSS, respectively. The results indicated that carotid WSS might be important for the development and progression of RFI and albuminuria in older populations. Chronic kidney disease and/or end-stage renal disease negatively affects vascular function and thus causes divergent vascular pathologies, including atherosclerosis, endothelial dysfunction, vascular calcification, and hemodynamic disturbance [19][20][21] . However, whether hemodynamic disturbances such as disturbances in carotid WSS affect the decline of renal function in subjects with normal renal function remain unclear. In this study, all participants who were eligible and enrolled at baseline had eGFR ≥ 60 mL·min −1 ·1.72 m −2 and ACR < 30 mg·g −1 . These results suggested that carotid artery rheologic forces might play a crucial role in the development of RFI and albuminuria in aging adults with normal renal function. Among the many reasons for the association between the disturbances in carotid WSS and RFI, the primary reason might be that lower WSS upregulates endothelial dysfunction and inflammation that represents a marker of vascular damage 9,[11][12][13]22 .
Hemodynamic forces are essential for vascular endothelial functions. Associations have been well established between lower and/or oscillatory shear stress and endothelial dysfunction 23 . In this study, changes in endothelial function and chronic inflammation over time were assessed using serum biomarkers. The results showed that serum NO levels decreased, and serum ET-1, ICAM-1, VCAM-1, and hsCRP levels significantly increased in the Q 1+2+3 group compared with the Q 4 group that were classified either by MWSS or by PWSS. ACR is not only one of the renal functional markers but is also regarded as one of the most important markers of generalized endothelial dysfunction 24,25 . Endothelial dysfunction leads to impairment of the endothelial barrier, which results in increased transvascular albumin leakage 24, 26 . Endothelial dysfunction has been regarded as an early event in atherosclerosis and is associated with kidney disease [27][28][29] . Endothelial dysfunction results in the production of reactive oxygen species; increased endothelial permeability; increased intracellular Ca 2+ concentrations; and expression of cytokines, chemokines, and adhesion molecules, which facilitate leukocyte transmigration into the extracellular matrix underneath, and finally vascular calcification and RFI 27,30-32 .
We did not find the significant differences in the changes in blood pressures and plasma levels of lipids and glucose between groups classified by MWSS and PWSS in the duration of follow-up in this study. This may be attributable to the fact that this study was an observational study, and we were unable to control for medication schedules such as treatment for hypertension, dyslipidemia, or diabetes. However, our results further demonstrated that endothelial shear stress could play an important role in the progression of renal function decline.
A major strength of this study is that it is a longitudinal prospective cohort study. It enabled us to explore the potential longitudinal associations between WSS and RFI progression in subjects with normal renal function that are difficult to explain by cross-sectional relationships. Another is that we have combined serum creatinine and cystatin C to estimate GFR. The precision of these equations has been validated within the Chinese population and can be used to reduce the bias that results from GFR estimated by creatinine and cystatin C alone [33][34][35] .
This study has several limitations. First, the participants were predominantly recruited from the Han ethnic population in the Shandong area, China. There were essentially ethnic, racial, and geographical limitations. Genetic background has been reported to be strongly associated with eGFR 36 . Thus, multi-ethnic and multi-national studies are needed. Second, we did not measure the renal arteries' WSS. The renal arteries' WSS tends to be more reasonable, which could be a precise reflection of the renal rheological forces in comparison to the carotid WSS. Third, RFI is a chronic disease with multi-risk factors 37,38 . Thus, the histories and medications of hypertension, diabetes, and dyslipidemia and the antiplatelet medication were accounted for in the analysis models, as these diseases and medications might influence the carotid WSS and renal function and introduce a  39 . In addition, we did not investigate the exercise and dietary habits of patients, which may affect the carotid WSS and renal function 40 .
In conclusion, our findings suggested that CCA WSS plays an important role in RFI progression in aging adults with normal renal function. Lower WSS was associated with a higher risk of RFI compared with higher WSS. This association might be via vascular endothelial dysfunction and low-grade inflammation pathways. patients and Methods study population and design. This study was based on data from an observational and population-based cohort study in the Shandong area, China 41 . The major objectives of the original cohort were to investigate determinants of various chronic diseases in a general population with ages 15 years and older (Registration number: ChiCTR-EOC-17013598), and the cohort was expanded to 21,000 participants between April 2007 and October 2009. The exclusion criteria of this study were as follows: age < 60 years, eGFR < 60 mL·min −1 ·1.72 m −2 and ACR ≥ 30 mg·g −1 , secondary hypertension, heart failure, active malignancy, severe liver disease, including chronic hepatitis and cirrhosis, and drug abuse. Finally, 1,447 older subjects were eligible and enrolled in this study. This study was approved by the Research Ethics Committee of the Institute of Basic Medicine, Shandong Academy of Medical Sciences and compliance with the "Declaration of Helsinki". Written informed consents were obtained from all participants.

Ultrasonography of common carotid artery and Wss evaluation. Ultrasound examinations
were performed by experienced ultrasonographers who were blinded to the participants' clinical details and during morning hours in a quiet room with a temperature between 22 °C and 25 °C. Before examination, the participants were not allowed to use/consume vasoactive medications (including nitrates, calcium antagonists, angiotensin-converting enzyme inhibitors, and angiotensin antagonists), smoke, alcohol, tea, and caffeine for 24 h and underwent fasting for 12 h. Duplex ultrasonography of the left and right common carotid artery was examined using a high-resolution ultrasound with a 7.5-MHz linear array transducer and electrocardiogram triggering (Vivid i, GE Medical Systems Ultrasound Israel Ltd., Tirat Carmel, Israel). Intima-media thickness of CCA (CCA-IMT) was measured by recording ultrasonographic images on the far wall of the CCA. CCA-IMT was defined as the distance from the interface of the CCA lumen-intima (first echogenic line) to the collagen-containing upper layer of the CCA adventitia (second echogenic line). Peak systolic velocity (V PS ), end diastolic velocity (V ED ), and mean velocity (V M ) were measured 1-2 cm below the bifurcation for three cardiac cycles, and their computed means were used for further analysis. The internal diameter (ID) was defined as the distance from the leading edge of the echo produced by the intima-lumen interface of the near wall to the leading edge of the echo produced by the lumen-intima interface of the far wall 9,18,42 . The IDs at the R (ID R ) and peak T (ID T ) waves on the electrocardiogram were acquired using B-mode tracings with automatic border detection function. The carotid artery wall was assumed to be rigid with blood as a Newtonian fluid 42 , and WSS was calculated using the following formula 9,43,44 : mean WSS = (8 × η × V M /ID R) and peak WSS = (8 × η × V PS /ID T ), where WSS is the carotid artery wall shear stress (Pa), η is the blood viscosity and equal to 0.0035 Pa·s; V is the velocity (m/s), and ID is the lumen diameter (m). WSS was calculated separately for mean and peak systolic velocity.   Table 2. Probability hazards of RFI and albuminuria over time stratified by hypertension, dyslipidemia, and diabetes. RFI indicates renal function impairment; MWSS, mean wall shear stress; and PWSS, peak wall shear stress. The models were adjusted for age; sex; smoking; alcohol intake; medications for antihypertension, antidyslipidemia, glucose-lowering, and antiplatelet; baseline blood pressures, blood lipids, fasting plasma glucose, common carotid artery intima-media thickness, estimated glomerular filtration rate, and albumin/creatinine ratio; and the average systolic and diastolic blood pressure levels at every annual visit.  Urinary albumin excretion assessment. Urinary albumin excretion was assessed annually using urinary ACR that was calculated using the urinary albumin from the morning first void sterile spot and serum creatinine as previously described 9 . The threshold of albuminuria was 30 mg·g −1 of ACR in categorical analyses.

Vascular endothelial function and chronic inflammation assessment.
In this study, the vascular endothelial function and chronic inflammation were assessed using the serum levels of NO, endothelin (ET)-1, intercellular adhesion molecule (ICAM)-1, vascular cell adhesion molecule (VCAM)-1, and hypersensitive C-reactive protein (hsCRP). NO levels were assessed using the quantification of nitrite by Griess assay 45 . The reagents were purchased from Sigma (St. Louis, MO, USA). ET-1, ICAM-1, VCAM-1, and hsCRP levels were assessed using enzyme-linked immunosorbent assay kits (Bender MedSystems, Vienna, Austria) following the manufacturer's instructions. Each sample was tested in duplicate, and the mean value was used for further analysis.  Follow-up. With the help of family physicians and nurses, participants were visited annually after the baseline survey. Demographic and clinical characteristics of participants including medications (antihypertensive, anti-dyslipidemia, glucose-lowering, and antiplatelet) were acquired at baseline and every annual follow-up visit. For the participant with hypertension, dyslipidemia, or diabetes, the corresponding therapy had been advised at every follow-up visit when the participant was willing to receive treatment. However, there was no unified therapeutic regimen in our study.
outcomes. The progression of renal dysfunction and albuminuria was assessed using the changes in eGFRs and ACR, respectively, over time. Putative RFI was identified as eGFR < 60 mL·min −1 ·1.72 m −2 7 . Putative albuminuria was identified as ACR ≥ 30 mg·g −1 6,7 .
statistical analysis. SPSS for Windows software package, version 24.0 (SPSS Inc., Chicago, IL, USA), was used to perform the statistical analyses. The continuous data were expressed as means and standard deviations or median with interquartile range (IQR, range: 25th to 75th percentile) depending on the normality of the data, which was determined by the Kolmogorov-Smirnov test. Categorical data were presented as numbers with percentages. Participants were classified into two groups (Q 4 group vs. Q 1+2+3 group) by the interquartile of mean WSS and peak WSS separately. The comparisons of continuous data between the groups were performed using the Student's t-test or Wilcoxon W test depending on the normality of the data. The chi-square test was used to assess the differences in categorical data. A linear mixed model was used to compare the changes in eGFR, ACR, and endothelial function over time among the groups. A logistic regression model was used to compare the odds of the incidences of RFI and albuminuria among the groups. Cumulative incidences of RFI and albuminuria were assessed using the Kaplan-Meier method and comparisons were assessed using log-rank test among groups. The hazard ratios (HRs) with 95% confidence interval (CI) in the incidences of RFI and albuminuria were evaluated using Cox proportional hazards models after adjustment for age; sex (male vs. female); smoking (no vs. yes); alcohol intake (no vs. yes); histories of hypertension, diabetes, and dyslipidemia (no vs. yes); medications for antihypertension, anti-dyslipidemia, glucose-lowering, and antiplatelet (no vs. yes); baseline blood pressures, blood lipids, fasting plasma glucose, CCA-IMT, eGFR, and ACR; and the average blood pressures at every annual visit. A two-sided p < 0.05 was considered statistically significant.