Predictors of survival and functioning of arteriovenous fistula in patients on hemodialysis during a one-year follow-up

Using SPSS for Windows, version 22, data were examined, and the level of statistical significance was set at p≤0.05. Relative numbers, central tendency, and measures of variability were used to describe the sample of hemodialysis patients. Mann-Whitney and Chi-square tests were used to evaluate the differences in the investigated parameters. We performed logistic and linear univariable and multivariable regression analyses to examine predictors of arteriovenous fistula survival and functionality in hemodialysis patients followed up for one year.

Of a total of 120 patients included in the study, 101 (84.2%) survived and 19 (15.8%) died during the follow-up period. The average age of the subjects was 64.17±10.78 years (survivors 64.01±10.51 vs. deceased 65.00±12.39), 70 (58.3%) patients were men, while 50 (41.7%) of respondents were female. Patients on hemodialysis who survived the 1-year follow-up had statistically significantly more arteriovenous fistula as vascular access than Hickman catheter or arteriovenous graft (p = 0.001). Distal positioning of the arteriovenous fistula was statistically significantly more frequent than the proximal-located one (p = 0.038). Patients who completed the study during the follow-up period spent significantly more months on hemodialysis (74.50 ± 64.12 vs. 27.49 ± 27.14; p = 0.001), they had statistically substantially higher hemoglobin concentration (104.75 ± 11.85 vs. 96.68 ± 17.60; p = 0008), and had higher Kt/V index values (1.10 ± 0.23 vs. 0.97 ± 0.37; p=0.010) (Table 1).

Univariable logistic regression analysis showed that the predictors of one-year survival in hemodialysis patients are the positioning of arteriovenous fistula, type of vascular access, length of hemodialysis treatment, higher hemoglobin values, index Kt/V values, and creatinine concentration. In multivariable logistic regression analysis (Method Enter, Model Summary: χ²=71.467; Cox & Snell R²=0.243; Nagelkerke R²=0.417; Hosmer-Lemeshoe goodness-of-fit χ² 11.728, p=0.164), only the length of hemodialysis treatment was shown to be a significant predictor of survival, meaning that the shorter the time on hemodialysis, the higher the fatality rate (Table 2).

In hemodialysis patients who survived the one-year follow-up (N=101), the average length of functioning arteriovenous fistula was 63.04 ± 51.16 months. In the group of subjects with arteriovenous fistulas who survived there were 89 (91.8%) patients, while 8 (8.2%) patients with arteriovenous fistulas did not complete the follow-up period. In the univariable linear regression analysis, significant predictors of longer vascular access functioning were more months on hemodialysis, lower diastolic blood pressure, lower leukocytes and platelets, higher hemoglobin, creation of an arteriovenous fistula by a nephrologist and the start of hemodialysis treatment with central venous catheter. On the other hand, multivariable linear regression analysis (Method Enter, R Square Change 0.487, F Change 8.408, p=0.001) confirmed that only a greater number of months on hemodialysis and the creation of an arteriovenous fistula by a nephrologist were significant predictors of longer-functioning vascular access (Table 3).

Despite the visible progress of modern medicine, the mortality rate of hemodialysis patients remains alarmingly high, and after one year of treatment is between 10 and 25% [1, 10]. Such a wide range of mortality is probably due to the influence of risk factors that can be conditionally classified into short-term predictors that are varied and often contradictory [11]. The mortality rate of our subjects, in the projected period of one year of follow-up, was 15.8%, which is in accordance with the stated mortality.

Indicators of nutritional status, projected in body mass index, are powerful predictors of morbidity and mortality, which has long been considered the main factor in survival, is at least confusing and can be inaccurate, because it does not take into account body composition, so the effects of lean and fat mass mixed, which reduces its potential predictive value [4, 12]. Moreau-Gaudry et al. have determined through their research that a single measurement of serum creatinine (predialysis) is a strong determinant of patient survival, even after only one year of follow-up, which gives this parameter the power to predict survival, far more than other, perhaps overemphasized accepted parameters [13]. The results of our research confirmed that the group of patients, whose one-year follow-up ended fatally, had statistically significantly lower values of creatinine concentration, in correlation with the subjects who survived the controlled period. The value of serum creatinine, according to univariable statistical analysis, has a statistically significant negative relationship with the survival of dialysis patients. However, multivariable statistical analysis did not confirm its predictive significance for patient mortality in our study.

Hemoglobin has a negative correlation with the survival rate of hemodialysis patients, that is, a low hemoglobin concentration can increase the risk of death in hemodialysis patients [14]. And the results of our study confirmed these findings, hemoglobin was significantly higher among the subjects who survived. According to univariable statistical analysis, hemoglobin was negatively associated with the survival rate of hemodialysis patients in a one-year follow-up study. However, multivariable statistical processing did not confirm its predictive capacity.

The Kt/V index is useful as a reference for controlling and improving the amount and efficiency of dialysis and significantly affects the survival of hemodialysis patients [15, 16], which we found in our study, but without confirming its predictive significance as a mortality risk factor.

Of all the types of vascular access, arteriovenous fistula is, according to general recommendations, the vascular approach of choice for hemodialysis because it has the lowest complication rate and is superior in terms of survival. Patients who start hemodialysis with a catheter have a 70 to 77% higher mortality rate than those who start hemodialysis with an arteriovenous fistula [17, 18], so it is no surprise that vascular access is one of the factors that determine the outcome of treatment [19]. This suggestion is true only when patients with mature arteriovenous fistulae are included for outcome determination [8]. The group of our subjects who survived the one-year control period had, as vascular access, a statistically significantly higher number of arteriovenous fistulas and they were positively associated with survival. However, arteriovenous fistulas did not confirm their predictive capacity by multivariable statistical method.

Guidelines for vascular access, whenever feasible, recommend creating an arteriovenous fistula in the distal part of the upper extremity as the first choice [20]. Our previous research is in accordance with the recommendations on the distal position, as the primary location of arteriovenous anastomosis, where we confirmed the predictive significance of the distal fistula on the survival of patients on hemodialysis [21]. Starting from the fact that every part of the vascular system represents an extremely important potential for patients who need an arteriovenous fistula, our position was that, wherever possible, an initial attempt should be made to create an anastomosis in the distal part of the forearm. Therefore, the results of our research confirmed, by univariable statistical analysis, the importance of the distal position of the arteriovenous fistula on the survival of our subjects, but without confirmed predictive influence.

Platelets play a role in the development of vascular access stenosis, and the key event in the development of stenosis is intimal hyperplasia. It is apostrophized that higher platelet aggregability is found in patients with vascular access dysfunction and that for the risk of thrombosis, platelet activity is more important than platelet count [22]. Through our research, we confirmed the negative association of the number of platelets with the function of the arteriovenous fistula through univariable statistical analysis, but the predictive significance of platelets was not confirmed.

Leukocytes are an important link in a complex multi-step process that, by activating cytokines and chemokines, inflammatory damages endothelial cells leading to intimal hyperplasia. It is a very complex mechanism that creates the conditions for stenosis and thrombosis of the arteriovenous fistula [23]. In our research, the significant influence of leukocytes on the functioning of the arteriovenous fistula was confirmed a negative association between leukocytes and fistula function was determined, without predictive significance.

Some studies have found a reduced rate of functional fistulas in patients with elevated blood pressure, most likely due to impaired endothelial cell functions and the consequent development and progression of atherosclerosis [24, 25]. In our subjects who survived the one-year follow-up period, a negative association was found between functionality and the diastolic component of blood pressure, but its value was not confirmed by multivariable statistical analysis.

KDOKI guidelines recommend the use of a central venous catheter for periods not exceeding seven days, due to possible stenosis and occlusion of the central vein resulting in significant long-term effects, such as venous hypertension, which affects the reduction of flow through the arteriovenous fistula and its inadequate function [26]. In our study, 55.8% of subjects started hemodialysis with central venous catheters. The policy of our Clinic is, whenever there is a possibility to create an arteriovenous fistula in time, before the start of hemodialysis treatment, that is, for patients to start hemodialysis with matured fistulas and to avoid the placement of central venous catheters. We believe that this attitude had a decisive influence on the functioning of the arteriovenous fistula, which was confirmed by univariable statistical analysis, but without predictive significance.

It is currently believed that countries where nephrologists have a significant role in professional coordination, including participation in the organization of treatment and creation of vascular access, achieve good results [27]. The results of our previous research showed that arteriovenous fistulas created by vascular surgeons had a 38% higher chance of failure compared to fistulas created by nephrologists. One of the conclusions was the use of radiological diagnostics and a detailed preoperative clinical examination of blood vessels by a nephrologist [28]. The results of our current study confirmed the negative association of arteriovenous fistula functioning concerning the information about who created the fistula but without predictive significance.

The risk of mortality increased concerning the length of dialysis treatment [4]. The uremic syndrome causes a complex mixture of organ dysfunction [29] hurts multiple organ systems and metabolic pathways and directly affects the increase in the mortality rate of dialysis patients [30]. The relationship between uremia, altered immune response, and impaired vascular biology is well recognized, which is why it can be argued that uremia can affect the outcome of arteriovenous fistula [31]. Survival of dialysis patients decreases with age and length of treatment, due to a large number of uremic complications and more frequent comorbidities [32]. Our result of the multivariable regression shows that the shorter the time on hemodialysis, the higher the mortality, that is, those who died during the 1-year follow-up had lower length on hemodialysis treatment, compared with those who survived, perhaps due to shorter follow-up time, etiological reason for end-stage renal failure, or higher number of comorbidities.

The study had several limitations. First, the study was organized as an observational one, so the results may indicate an association and not a cause, which suggests that the results must be interpreted with caution to determine the possible association between the mortality rate and the functionality of the arteriovenous fistula, and how the outcome of the treatment and the functioning of the fistula can be improved after adjustment with independent risk factors. Likewise, we did not analyze other important factors that may affect the functionality of the arteriovenous fistula, nor did we define the cause of mortality in our subjects, which could have an impact on the results of our research, such as the presence of various comorbidities and the etiology of renal failure.