This is a single-center retrospective study that includes patients who attended a specific appointment for vascular access planning consultation at Oporto University Hospital Center between January 2019 and December 2021. Inclusion criteria were (a) age over 18 years and (b) the purpose of the consultation is vascular mapping for the creation of first vascular access. Patients may or may not have had previous hemodialysis sessions through a CVC: among the evaluated patients, there were 114 incident patients (CKD stage 4 or 5) and 53 prevalent patients (under hemodialysis through CVC).

In the vascular access consultation, a nephrologist evaluated the patient and performed the physical exam and vascular mapping through CDU. After patient evaluation, the nephrologist proposed to the vascular surgeon team (five surgeons and respective residents) a specific type and location of vascular access. The proposed access could be an AVF, arteriovenous graft (AVG) or CVC (we do not advocate vascular access planning in patients on peritoneal dialysis, waiting for a living donor kidney transplant in short term or when we do not expect to start dialysis within the next 6 months).

The following data were collected: demographic data, weight, height, medical history (diabetes mellitus – DM, peripheral arterial disease – PAD, ischemic coronary disease, previous stroke), and characterization of chronic kidney disease (etiology, proteinuria, serum creatinine and glomerular filtration rate on the date of the first consultation and the date of construction of the vascular access).

During the follow-up, the following data were recorded: date and type of access proposed by the nephrological team, date and type of vascular access constructed by the vascular surgery team, date of the beginning of hemodialysis, and access used in the first hemodialysis.

Patients were followed by the nephrologist in vascular access consultation until one month after the first hemodialysis through a functioning vascular access. During follow-up, vascular access complications, number and type of vascular interventions, and the need for subsequent vascular access creation were evaluated.

The investigation followed the principles outlined in the Declaration of Helsinki and was approved by the Institutional Review Board of the Oporto University Hospital Center.

An ultrasound General Electric with high frequency (7–15MHz) linear probe was used for measurements.

Physical and DU assessment always began by non-dominant arm blood vessels; patients were in a supine position without angling the elbow joint to avoid vessels compression. Both the superficial and deep venous systems were examined from the wrist to the axillar vein and axillar artery, if technically possible.

When the venous anatomy was acceptable for an AVF, arterial examination was performed by palpating arterial pulses. If the radial artery was not suitable, the ulnar and brachial arteries were examined as alternative sources. Evaluation of the dominant arm was performed only when the non-dominant arm evaluation was unsatisfactory.

In our study, only vessels that met the minimum criteria were chosen – a venous luminal diameter of ≥2.5mm for distal AVF and ≥3.0mm for proximal AVF (using a tourniquet) and continuity with proximal veins in the arm; and arterial luminal diameter of ≥2.0mm. Vein compressibility and distensibility, distance from the skin surface, continuity with the deep venous system in the upper arm, and confirmation of the absence of ipsilateral central venous stenosis or occlusion were performed.

Vein distensibility was evaluated in a subjective way before and after a tourniquet placement, and we did not measure vein distensibility by DU. The presence of venous congestion and a reduced or absent respiratory variability in the axillar vein implied central stenosis exclusion by angiography before AV access creation.

The distensibility of the arterial wall was assessed by evaluation of the Doppler wave form in the radial artery during reactive hyperemia induced by reopening the first after it had been clenched for 2min (change from the high-resistance triphasic to low-resistance biphasic wave-form) and resistance index (RI) was measured and considered the target <0.7.

Arterial inflow was also evaluated to exclude arterial stenosis. In this field, we did not propose AVF creation in three specific conditions. First, AVF was not created in the presence of calcifications of the feeding artery wall accompanied by a negative reactive hyperemia test in artery. Second, anastomosis was not created distal to a stenosis above 50% in radial artery. Third, we did not advocate AVF creation in the presence of cubital artery with stenosis, a diameter <1.5mm or absent associated with radial artery arteriopathy due to the increased risk of primary failure with distal AVF and of ischemia with proximal AVF. When high bifurcation of the brachial artery is found, resulting in deeper and larger ulnar artery and smaller and more superficial radial artery, the larger and deeper ulnar artery is recommended for anastomosis.

Forearm AVF location was preferred compared to upper-arm location. To make a final decision clinical data, risk factors associated with maturation failure, and CDU findings were considered. The aim was the creation of vascular access with more probability of success, reducing the time of dialysis CVC and the risk of maturation failure.

When patients did not have arteries and veins suitable for AVF, we immediately attempted an AVG. The criteria for good arterial inflow and venous outflow for AVG were brachial artery lumen diameter ≥3.0mm and axillar vein lumen diameter ≥4.0mm, respectively.

A postoperative surveillance scan was planned at 4–6 weeks after access creation to assess access maturation: patency, Qa, artery and vein diameter, presence of stenosis formation and measured flow in the fistula (ml/min), the vein diameter, and any abnormalities.

The AV access creation involved a multidisciplinary strategy.

The type and location of AV access proposal were made by the nephrologist, but surgeons have access to the CDU in the operating room, they can repeat the evaluation and they had the final word in relation to the choice of AV access.

Five types of possible vascular access were recorded: radial-cephalic AVF (fed by radial artery), proximal AVF (brachio-cephalic or brachio-basilic, fed by brachial artery), AVG, and CVC.

The primary outcomes were the evaluation of primary failure (PF) and failure of vascular access maturation at eight weeks (MF). Secondary outcomes were the evaluation of primary patency (PP) and primary assisted patency (PAP) at 6, 12, and 24 months.

PF was defined as access failure due to early technical failures (intra-operative thrombosis or other complications or abandonment of the newly created AVF).

MF is defined as insufficient access flow to maintain dialysis or the inability to cannulate an AVF, if required, at eight weeks after surgery.

PP is the interval from the time of access placement until any intervention designed to maintain or reestablish patency, access thrombosis, or the time of measurement of patency. PAP is the interval from the time of access placement until access thrombosis or the time of measurement of patency, including surgical or endovascular interventions designed to maintain the functionality of a patent access.15

Categorical variables were expressed as absolute counts and percentages. Continuous variables were presented as mean±standard deviation if normally distributed.

Survival evaluations were used to estimate PP and PAP with Kaplan–Meier curves. The association between the variables and each outcome was studied using Chi-square test and Fisher exact test, according to the sample size, in nominal variables. To continues variables, we used univariate logistic regression.

The log rank test was used to assess the differences between the groups in the Kaplan–Meir analysis. The inter-rater reliability between the nephrologist's proposal and the surgical team's decision was evaluated using the Cohen's kappa coefficient.

In all these tests, we obtained the odds ratio with a confidence interval of 95%; the significance level was 0.05. All statistical analyses were performed using SPSS-IBM.

In this study, 167 patients (62.4±15.8 years) were evaluated in vascular access consultation with the aim of creating a vascular access for hemodialysis (Table 1). Gender distribution was: 100 men (59.9% of patients) and 67 women (40.1%). The comorbidities, in order of frequency, were: DM (40.7%), ischemic coronary disease (24%), PAD (14.4%), and previous stroke (10.8%).

The main cause of chronic kidney disease was diabetic nephropathy (28.1%) followed by glomerulonephritis (18.0%) and polycystic kidney disease (7.2%); 14.4% of patients have a multifactorial etiology and 21.0% of patients had no determined etiology.

The time between the first evaluation and the decision to create a vascular access was 2.6±5.8 months. The time between the decision to the creation and the surgical construction of vascular access was 27.3±28.6 days.

The vascular accesses proposed by nephrologist were radial-cephalic AVF in 70 patients (41.9%), brachio-cephalic AVF in 50 patients (29.9%), brachio-basilic AVF in 34 patients (20.4%), AVG in 8 patients (4.8%) and CVC in 2 patients (1.2%) (Table 2). In three patients (1.8%) a construction of vascular access was not proposed because a short-term living donor kidney transplant is predictable or because they chose peritoneal dialysis.

A vascular access was constructed in 141 patients: radial-cephalic AVF in 57 patients (40.4%), brachio-cephalic AVF in 54 patients (38.3%), brachio-basilic AVF in 27 patients (19.1%), and AVF graft in 3 patients (2.1%).

The created access corresponds to the proposed access in 124 patients (87.9%). Twenty-six patients did not create vascular access for some reasons: death, refusal, or modification of option.

Among the 141 vascular access created, 22 (15.6%) primary failures were registered (Table 3). Sex was not a statistically significant determinant for the occurrence of primary failure (OR=1.775 (0.710–4.435), p=0.216).

The mean age of patients with PF was 58.3±15.7 years; age was not significantly associated with the occurrence of primary failure (OR=0.983 (0.956–1.010), p=0.209).

All PF occurred in AVF, with the following locations: 15 radial-cephalic AVF (PF 26.3%), three brachio-cephalic AVF (PF 5.6%), and four brachio-basilic AVF (PF 14.8%).

Radial-cephalic AVFs were associated with a higher risk of PF in this study (OR=3.929 (1.485–10.392), p=0.004). Brachio-cephalic AVFs have a lower risk of PF in this study (OR=0.211 (0.059–0.750), p=0.010). Brachio-basilic AVFs and AVG were not statistically significantly correlated with the occurrence of PF.

In this study, DM was associated with a higher risk of PF (OR=3.867 (1.235–12.113), p=0.014), while previous stroke, PAD, and ischemic coronary disease did not significantly influence this event. However, in evaluation by the etiology of chronic kidney disease, diabetic nephropathy appeared to be a protective factor against PF (OR=0.198 (0.044–0.887), p=0.021).

The glomerular filtration rate at first evaluation and at the time of the creation of vascular access did not significantly influence the occurrence of PF.

In this study, the incidence of maturation failure at eight weeks was 4.8% for AVF (Table 3); all patients required surgical or endovascular intervention to allow access maturation.

The mean age was 61.5±26.7 years; three patients had DM (37.5%), and one patient had PAD (12.5%). The cause identified for MF was inflow stenosis in 75% of cases, outflow stenosis in 12.5%, and central stenosis in 12.5%.

Four maturation failures were identified in radial-cephalic AVF and four in brachio-cephalic AVFs.

In our population, the PP at 6, 12 and 24 months was 76.3%, 70.4% and 49.2%. The assessment of PP according to location is described in Table 4. In the assessment by location, no statistically significant differences between the groups were found (p=0.128) (Fig. 1).

Fig. 1.

Kaplan–Meier curves of primary patency and primary patency by location.

(0,14MB).

PP excluding PF was 89.7% at six months, 82.5% at 12 months, and 58.6% at 18 months.

Primary assisted patency was 84.8% at 6 and 12 months and 81.3% at 24 months (Table 4). PAP was different according to the location of vascular access (p=0.009) (Fig. 2).

Fig. 2.

Kaplan–Meier curves of primary assisted patency and primary assisted patency by location.

(0,19MB).

In 12 patients (8.5%) there was a discrepancy between the nephrologist's proposal and the vascular surgeon's final decision in the operating room (Table 5). High reliability was documented between the nephrologist and surgical team assessments (k=0.874; p<0.01; concordance=91.5%).

In this group, most patients were male (n=9; 75%) and seven were incident patients (58.3%). The rate of primary failure and maturation failure in this group was 16.6%.

In five patients, the nephrologist proposed the construction of a radial-cephalic fistula, and the surgical team opted for constructing a brachio-cephalic fistula; in this group, no further complications were identified in the development of vascular access.

In three patients with proposed brachio-cephalic fistula, the surgical team opted to construct a radial-cephalic fistula; primary failure was documented in one of them.

In one patient, there was disagreement regarding the laterality of vascular access: a brachio-cephalic fistula was proposed in the right upper limb and was created in the left upper limb; the fistula had maturation failure and need for endovascular intervention due to recurrent stenosis of the cephalic vein.

In two patients, the construction of a vascular graft was proposed by nephrologist: in one patient, the surgical team initially constructed a brachio-basilic fistula, which had a primary failure, and subsequently, a vascular graft was constructed; in another patient, a radial-cephalic fistula was successfully constructed in the contralateral limb.