Framingham Risk Score and clinical outcomes of forearm autologous arteriovenous fistulas in end-stage renal disease patients: a retrospective cohort study.

Background The failure of autologous arteriovenous fistulas (AVF) and cardiovascular disease (CVD) share several common risk factors. The Framingham risk score (FRS) can predict the occurrence of CVD. Accordingly, we conducted a retrospective cohort study to evaluate if the FRS can predict the AVF failure in end-stage renal disease (ESRD) patients. Methods In this study, 188 ESRD patients (with the mean age 57.2 ± 13.8 years, 64.9% of them were males) were enrolled with a median follow-up time of 41 months. AVF failure was defined as the presence of stenosis or thrombosis, and loss of the ability to receive prescribed hemodialysis. Primary patency of AVFs, and the association between FRS and AVF failure or death were analyzed by the Kaplan-Meier method or Cox proportional hazards model, accordingly. The primary patency of AVFs was 84.6%, 79.3%, 76.6%, 76.1% and 76.1% at 12, 24, 36, 48 and 60 months, respectively. However, there was no significant difference of the primary patencies among the 1st Quartile, 2nd Quartile, 3rd Quartile and 4th Quartile of FRS (log-rank test, p =0.15). Moreover, FRS was not associated with all-cause or CVD mortality after adjusting for other confounders in the Cox proportional hazards models. This study did not find FRS associated with AVF failure or mortality among ESRD patients.

, 76% in Korea [ 9], 67% in Europe and only 64% in the United States [ 10]. Therefore, AVF failure is also a major issue of vascular access related to the HD. A recent study demonstrated that the primary and secondary patency of AVFs at one year were only 64% and 79% [ 11], which further confirms the importance of intensive examinations for the early detection of clarifying risk factors to prevent the occurrence of AVF failure. However, to date most of the studies about the risk factors for AVF failure were still controversial [ 12,13].
The Framingham risk score (FRS) has been well validated for cardiovascular risk stratification, and was revised in 2008 and became the Framingham general CVD risk score [ 14]. Now, FRS is the most commonly used algorithm in clinical practice to predict cardiovascular disease (CVD) risk.
Recent studies demonstrated that the failure of AVF and CVD share several common risk factors, such as elderly, hypertension, diabetes and arterial stiffness [15][16][17]. However, the relationship between FRS and the AVF failure is still uncertain. Accordingly, we conducted a retrospective cohort study to evaluate if the FRS can correctly predict the AVF failure in ESRD patients.

Method Study population
The study was a single-center retrospective cohort study, ESRD patients were enrolled between January 2013 and June 2016. The deadline of follow-up time was November 10 th , 2018. AVFs were created at the cephalic vein to radial artery, of the left or right lower half of the forearm in CKD stage 4-5 patients evaluated by the estimated filtration rate (eGFR). eGFR was calculated using the CKD-EPI equation [ 18]. Primary patency data were available during the follow-up. The exclusion criteria were as follows: 1) age < 18 years; 2) presence of malignant disease, hypotension, active bleeding, or severe liver disease; 3) no known AVF patency at the end of the follow-up period. After excluding relevant indicators, 188 CKD stage 4-5 patients were eligible for inclusion in the study, with a median follow-up time of 41 months.
The study was conducted in accordance with the Declaration of Helsinki, and also under the approval of the Ethics Committee of Qianfoshan Hospital affiliated to Shandong University. The informed consent from patients in the study was waived, however, patients were informed about the registration of all individuals with treated ESRD by the nephrology clinic as well as their right to not participate in the study.

Clinical and laboratory measurements
Demographic characteristics including age, sex, smoking status and use of prescription drugs, primary renal diseases and comorbidities (hypertension, diabetes and cardiovascular disease) were recorded. Laboratory findings, including hemoglobin, calcium, phosphorus, magnesium, parathyroid hormone (PTH), urea nitrogen, serum creatinine, fasting blood glucose, serum uric acid, serum total cholesterol (TC), and triglycerides (TG), high-density lipoprotein (HDL) cholesterol, and low-density lipoprotein (LDL) cholesterol were measured by the automatic biochemistry analyzer in the central laboratory of Qianfoshan Hospital. Cardiovascular disease (CVD) included ischemic heart disease, heart failure, stroke and peripheral vascular disease was recorded.
Framingham risk score (FRS) FRS were calculated for the total study population based on the following characteristics: age, smoking, diabetes mellitus, systolic blood pressure (mmHg), HDL cholesterol concentration (mg/dL), and TC concentration (mg/dL) [ 14].

Failed AVF
Failed AVF was defined as the presence of stenosis or thrombosis, and loss of the ability to receive prescribed hemodialysis [ 13]. Screening was conducted by ultrasonography to assess the degree and location of stenosis or thrombosis and to specify the outflow veins.

Outcomes
Clinical follow-up was undertaken in the dialysis unit where details of complications with AVF use were recorded. Patients were also contacted by telephone to incorporate a structured interview that included questions on frequency and success at HD, complications (such as stenosis and thrombosis), and any subsequent treatment [ 19]. End points including AVFs failure, all-cause and CVD death.
Sociodemographic characteristics, such as health history (eg, hypertension and diabetes) and lifestyle behavior (eg, smoking and habitual drinking) were obtained by means of questionnaire. The body mass index (BMI) was calculated as weight (in kilograms) divided by height squared (in square meters). Diabetes was defined as fasting blood glucose ≥ 7.0 mmol/L or by the use of hypoglycemic agents or by self-reported history of diabetes. Blood pressure was measured using a sphygmomanometer, and three measurements were taken at 5min intervals. The mean of the three readings was calculated, unless the difference between the readings was greater than 10 mmHg, in which case the mean of the two closest measurements was used. Hypertension was defined as systolic blood pressure of more than 140 mm Hg or diastolic blood pressure of more than 90 mm Hg, or both, or patients already being priscribed by antihypertensive medicaments. The pulse of waveforms of the right carotid and femoral arteries (cfPWV) was assessed using the SphygmoCor device (AtCor Medical LtD., Sydney, Australia) as previously described [15].
All of the study investigators and staff members completed a training program to learn the methods and procedures of the study.

Statistical analyses
Data are presented as proportions for categorical variables and mean ± SD or median [interquartile range (IQR)] for continuous variables. T test or Wilcoxon rank-sum test of variables was used to study the significance of differences in continuous variables between two groups accordingly.
Chi-square test was used to study the difference in the distribution of categorical variables.
Outcomes of AVF were evaluated using the Kaplan Meier method, and the log-rank test was used to test the differences between different groups. The cumulative primary patency rates were calculated for 12, 24, 36, 48, and 60 months. Cox proportional hazards models were used to explore the risk factors for AVF failure. Baseline variables including age, sex, smoking status, comorbidities (including hypertension, diabetes and CVD), statins, antiplatelet drugs, calcium, phosphorus, PTH, triglycerides, total cholesterol, HDL, LDL and FRS that were considered clinically relevant were entered into the multivariable regression models. The FRS was analyzed in quartiles. The 25, 50 and 75 percentile of FRS was 12.3, 17.0 and 22.0, respectively. Furthermore, we used the Cox proportional hazards analyses to explore the association between the FRS and mortality. Crude and adjusted odds ratios (ORs) with 95% confidence intervals (CIs) were reported.
All analyses were performed with SPSS statistical package, version 16.0 (SPSS, Inc., Chicago, IL). All p-values are two tailed. A p value < 0.05 was considered statistically significant.

Baseline characteristics
The baseline characteristics of the participants stratified according to the gender were shown in Table 1. The mean age of participants was 57.2 ± 13.8 years (range from 20 to 86 years), and 64.9% of them were males. Diabetic kidney disease (37.8%) was the major cause of ESRD, and 37.2% of participants had CVD. Moreover, the FRS was not significantly different between males and females, which was 17.1 ± 6.2 and 16.1 ± 5.6 (p = 0.29), respectively.

Clinical outcomes of AVFs
During the follow-up period, 45 (23.9%) patients experienced AVF failure, and the percentage of AVFs failure in females was higher than in males, which was 31.8% and 19.7%, respectively. In total participants, 11 patients were lost to follow-up, 63 patients died and 114 patients remained alive,

Factors associated with AVF failure
We confirmed the proportional hazard assumption and conducted Cox proportional hazards regression analyses for risk factors associated with AVF failure. After adjusting for age and sex, age (OR 1.02, 95% CI, 1.0-1.04) was independent predictors of AVF failure, however, sex, DKD, smoking, hypertension, CVD, statins, antiplatelet drugs, calcium, phosphorus, PTH, TC, TG, HDL, LDL and FRS were not associated with AVF failure (Table 2).
Relationship between FRS and the mortality.

Discussion
Successful establishment of an optimally functioning AVF is a highly desirable result in ESRD patients receiving HD. In fact, there are large international differences exist in successful use of AVF.
Successful AVF use was 87% in Japan, and only 64% in the United States [ 10]. In our dialysis unit, AVF placement in distal upper-extremity sites was the first choice for AVF creation when feasible [ 19].
This cohort study demonstrated that the primary patency of AVFs was 91.5% and 85.1% at 12 and 24 months, respectively. Moreover, the failure rate was 21.3% at 60 months, which further confirms the importance of early detection of clarifying risk factors to prevent AVFs failure.
In the Cox proportional analysis, our study demonstrated that age was positively associated with AVF failure, which probably due to higher vascular stiffness caused by numerous comorbidities, such as hypertension and diabetes. Some of these disease, such as severe arteriopathy are classic predictors of AVF failure [ 16,17]. Cui et al [ 20] also found that the primary failure rate was higher for the fistulas compared with the grafts in the elderly patients, which meaned that AVF as first option for HD not necessary in elderly patients. However, a recent multi-center cohort study demonstrated that only post surgery of AVF blood flow and diameter moderately predicted unassisted and overall AVF patency, but, the other factors such as age and sex did not further predict AVF dysfunction [ 12].
There have been studies investigating the molecular mechanisms responsible for AVF failure, such as inflammation and uremia [ 21,22], which were also the risk factors for CVD. Recent studies demonstrated that the failure of AVF and CVD share several common risk factors, such as elderly, hypertension, diabetes, hyperphosphatemia and arterial stiffness [15][16][17]23]. FRS is the most commonly used algorithm in clinical practice to predict cardiovascular disease (CVD) risk. Until now, the relationship between FRS and the AVF failure is still uncertain. In our study, participants in the lowest FRS group had better patency rates than in the higher Quartiles of FRS at 12 months.
However, there was no significant different in primary patencies among these four groups at 60 months tested by the log-rank test. Notably, most of current studies lacking a core outcome of mortality among ESRD patients receiving AVF creation [ 13]. Moreover, this study did not find FRS associated with mortality among these ESRD patients.
This study also has limitations that deserve attention. First, this is a single-center study, and our prediction models are not calibrated to any other centers, therefore, selection bias in the study limited the extension of the results from this study to other populations. Second, there was a lack of ultrasound findings, such as the presence of vascular calcification and post operative venous dilatation, which might provide important prognostic information. Third, the study sample was relatively small, the negative results could be caused by the limited power.

Conclusion
In summary, this study did not find FRS associated with AVF failure or mortality among ESRD Increased expression of TGF-beta1 and IGF-I in inflammatory stenotic lesions of hemodialysis fistulas.