FC 101: The Impact of Vascular Access Types on Survival in Hemodialysis: Time for a Shift in Paradigm? A Prospective Cohort Study

Abstract

BACKGROUND AND AIMS Despite arteriovenous native fistula (AVF) is the vascular access (VA) of choice due to a better outcome than central venous catheter (CVC) [1], catheter use is still common among haemodialysis (HD) patients. Our study aims to describe the survival and epidemiological features of a cohort of dialysis patients with a focus on the role of VA type. METHOD Our cohort comprises a prospective follow-up conduced from 2001 to 2020, which led to recruit 754 patients in HD. We identified a subgroup of patients survived less than 7 months (n 124) and a subgroup of patients survived more or equal than 7 months and up to 60 months (n = 472). In addition to this, every subgroup was subdivided into two different age groups, including patients of age ≤74 years old and those ≥75 years old. Other analysed characteristics were demographic features and baseline clinical data, like primary cause of ESKD, co-morbidities at the beginning of the HD treatment and type of VA both at start of HD and at the end of the follow-up. Starting HD without planning, namely being ‘late referral’, was also considered. Descriptive analysis was performed on baseline data. Moreover, we analysed predictive variables of death for any cause with Univariate and Multivariate Logistic Regression. Predictors of survival were analysed through univariate and multivariate Cox regression. RESULTS At the multivariate logistic regression, the use of tunnelled CVC at the start of HD resulted protective against death from any cause (aOR, 0.43; P = 0.017) in the whole cohort (n = 754). In the subgroup analysis of patients (n = 124) with a follow-up < 7 months, malignancy (aOR, 4.57; P = 0.002), severe cardiomyopathy (aOR, 4.35; P = 0.001) and vascular disease (aOR, 3.22; P = 0.008) were estimated as significantly predictors of death. At the multivariate Cox analysis, being ‘late referral’ was associated with decreased survival within 6 months (aHR, 3.79; P = 0.001). In the subgroup of elderly (≥75 years old; n = 201/472) survived within 7–60 months, multivariate logistic regression showed that the use of tunnelled CVC at the start of HD (Figure 1) resulted even more protective against death (aOR 0.25; P = 0.021). Moreover, vascular disease resulted as the main risk factor for death (aOR 5.11; P = 0.000), and it was confirmed as the only independent risk factor at the COX analysis (aHR 1.58; P = 0.017; see Figure 2). CONCLUSION Vascular disease is the main risk factor for death in haemodialysis patients. Furthermore, cardiovascular disease influences patient outcome more than the VA type, confirming the assumption that the choice of VA might be influenced by several issues as age and comorbidities [2, 3]. In the short-term survival subgroup, neither VA nor age emerged to be predictor of mortality, suggesting that health status before starting maintenance HD might explain an important part of the risk often attributed to catheter use.[4] Moreover, starting RRT without proper planning (‘late referral’) was extremely relevant in determining short-term survival. Tunnelled CVC might be the option of choice as VA, specifically in the elderly with vascular disease. In conclusion, our long-term experience suggests that the choice of VA should not necessarily follow the ‘fistula first’ approach, but it should be based on dialytic life expectancy and comorbidities.