Optimal Screening Program in Detecting Stenosis and Predicting Thrombosis in Hemodialysis Graft

A well-functioning vascular access (a system that allows the blood of the patient to leave the body, be brought to the artificial kidney or dialyzer to be cleaned and return the cleansed blood to the body) is essential for providing adequate life-sustaining treatment in patients with end stage renal disease on maintenance haemodialysis.

The preferred long-term vascular access is the arteriovenous fistula (AVF), which is created using the vessels of the patient by surgically connecting an artery with a superficial vein to increase the blood flow (Qa) in the venous system, which after few weeks will dilate (maturation) allowing the insertion of two needles, one to carry the blood to the dialyzer (arterial needle), and the other to return the cleansed blood to the body (venous needle) with the aid of a dialysis machine, equipped with a blood pump and pressure sensors at the level of the "arterial" and "venous" needle to monitor for needles dislodgment.

Unfortunately, the high prevalence of vascular disease of the haemodialysis patients makes difficult to create an adequate AVF in as many as 20 to 60% of the patients.

In these persons, a valid alternative is the arteriovenous graft: in graft method an artery is surgically connected to a vein with a short piece of synthetic soft tube which is implanted under the skin. Needles are inserted in the graft during the dialysis treatment.

Compared to an AV, however, graft is at higher risk of complications. The most frequent and feared complication is thrombosis (i.e. the formation of blood clot inside the graft).

Usually, thrombosis is the consequence of an underlying stenosis (i.e. a greater than 50% narrowing of the vessel or graft lumen by comparison with the lumen of a normal adjacent vessel or graft due to the thickening of the vessels and/or the graft wall) and its hemodynamic consequences of decreasing the access blood flow (Qa) and/or increasing pressure within the graft. Less frequently thrombosis occurs in the absence of significant stenosis and may be triggered by a drop in the patient's arterial blood pressure (hypotension).

All vascular access guidelines recommend regular non-invasive screening programs of grafts for timely identification of a stenosis associated with some type of functional or hemodynamic impairment, because its repair may prevent thrombosis and lengthen the useful life of the access.

Non-invasive screening methods include clinical monitoring (i.e. noting signs of access dysfunction during the haemodialysis session) and surveillance, which uses special equipment either to assess the hemodynamic consequences of stenosis by measuring Qa and static venous intra-access pressure ratio (VAPR) or to visualize the stenosis by means of duplex ultrasound (DU).

In graft, all guidelines have opted for surveillance. However, they also state that there is insufficient evidence to prefer one technique to another because of the lack of adequate comparative studies.

Indeed, the great majority of the studies have evaluated the ability of detecting stenosis and predicting incipient thrombosis (within 1 to 3 months) by one surveillance method, and very few have compared two-to-three techniques at the best (1).

The purpose of the investigators was to compare in the same graft population the diagnostic performance of all the currently available screening tools in detecting stenosis and predicting incipient thrombosis (to identify the best criteria for elective stenosis repair).

The primary aim of the study was to identify an optimal screening program to reduce the risk of thrombosis in graft by assessing the diagnostic accuracy (i.e. sensitivity, false positive rate, positive and negative predictive value, and area under the receiver operator curve) of all the available screening tools and the occurrence of acute hypotension and identify the best threshold/s for continuous variables in a) detecting stenosis, b) predicting incipient thrombosis (within a 4-month period following graft assessment for stenosis).

The following screening methods will be tested:

1. Monitoring: monitoring was considered positive when signs of graft dysfunction were noted during dialysis (difficult cannulation, aspiration of clots, inability to achieve the prescribed dialysis pump blood flow (Qb), excessive post-dialysis bleeding) or a >0.3 drop in single pool dialysis dose was documented;
2. Dynamic pressures: pressures were measured in the initial 5 minutes of dialysis, using 15 G needles and detected by the dialysis machine using the pressure sensors connected with the "arterial" (dynamic arterial pressure in mmHg: dAP) and "venous" needle (dynamic venous pressure in mmHg: dVP) at a standard Qb 250 ml/min. dAP was expressed as the ratio with Qb (dAP/Qb, in mmHg/ml/min);
3. Derived static intra-access venous pressure ratio (VAPR): was obtained in the initial 5 minutes of dialysis by the dVP, Qb, haematocrit and systemic systolic and diastolic blood pressure values, according to literature in mmHg/mmHg;
4. Access blood flow (Qa, in ml/min) was measured by the Ultrasound dilution method during dialysis (QaU) using the Transonic HD03 device, in the same dialysis session in which pressures were measured: each value is the mean of triplicate measurement;
5. Duplex Ultrasound (DU) was performed just before dialysis or in a non-dialysis day, using the Logiq 7 device (General Electric, Milwaukee). Grayscale and color imaging of the of the whole access circuit (feeding artery, graft and draining veins) was performed in the longitudinal and transverse plane to assess the presence of stenosis. Each stenosis was graded with regard to: (1) location (by defining "inflow stenosis" any stenosis located at the feeding artery, the arterial anastomosis or intragraft upstream the cannulation area, and "outflow stenosis" any stenosis located intragraft downstream the cannulation area, at the venous anastomosis, or at the draining veins); and (2) percent lumen reduction, determined by the ratio of the minimal intraluminal diameter at the stenosis to the nearby normal vessel or graft segment using an electronic calliper: a stenosis was considered significant (StD) when the lumen reduction was > 50% and/or downstream spectral doppler peak velocity of systolic blood flow (PSV) was > 400 cm/sec. Measurement of Qa is made in a straight portion of the brachial artery in the mid-third of the upper arm. The diameter of the blood flow was measured directly on the vessel thanks to b-flow color technology. Sampling volume was placed in the centre of the lumen and in the longitudinal plane. Typically, measurements were obtained over a sequence of 3 to 5 cardiac cycles (to allow for time-averaged mean velocities, TAV). TAV was calculated directly by the device from a doppler spectral waveform by the duplex scanner system. The Qa (in ml/min) is calculated by the device as the product of the artery diameter and the TAV. The mean value of at least 3 separate measurements was reported.

Subsequently all grafts underwent the biplanar Digital Subtraction Angiography (DSA), our gold standard for stenosis detection. The access was visualized in its entirety from the feeding artery to the right atrium seeking for the presence of stenosis. Each stenosis was graded with regard to location and degree as outlined for DU and was considered significant (StA) when the lumen reduction was > 50%. To ensure blinding the investigator performing DSA was unaware of the results of the other screening strategies.

Secondary aims of the study were to assess:

1. the relationship between Qa measurement by ultrasound dilution (QaU) and DU (QaD);
2. the intra-assay variability of QaU and QaD measurements;
3. The inter-assay variability of QaU, dAP/Qb, dVP, and VAPR measurements;
4. the concordance of 2 different radiologists in detecting presence of significant stenosis (>50%) at DSA.