Elevation of pressure measured in the drip chamber of the venous cannula downstream of the dialysis pump and membrane unit can occur with venous outflow obstruction. Since venous side stenoses are a common cause of access failure, serial measurement of venous line pressures is an accepted surveillance technique. Accuracy of the pressure measurement is highly dependent on proper matching of the dynamic response of the catheter—transducer system, transducer zeroing relative to height differences in the system, and elimination of air bubbles or blood clots. Both static (measured under no dialyzer flow condition) and dynamic (with dialyzer flow) venous line pressures have been utilized. Measurement of line pressure with the dialysis flow rate held constant at 200–250 mL/min for several minutes is the preferred technique. Testing is performed prior to proceeding with hemodialysis at a flow rate of 300–400 mL/min with a line pressure <150 mmHg considered normal.

The diagnostic accuracy of venous line pressures in detection of >50% DR stenoses is sufficient for surveillance testing. A dynamic venous pressures >150 mmHg during three consecutive dialysis sessions had an 86% sensitivity and 93% specificity for detection of >50% venous outflow stenosis by confirmed fistulography [10]. Mean dynamic venous line pressure was significantly higher in patients with >50% DR angiographic-confirmed stenosis (126 ± 35 mmHg) than in accesses without stenosis (95 ± 22 mmHg) [19]. When used alone, venous line pressure measurements have not been predictive of access thrombosis [5, 18, 19]. Testing is most sensitive to the development of venous anastomotic or outflow stenosis since lesions proximal to venous cannula will not be detected. Serial venous line pressure measurements can recognize the “failing” access, and a threshold value in the 150–200 mmHg is an appropriate indication to proceed with additional diagnostic testing such as duplex ultrasound.

Serial volume flow measurement of the patent dialysis access holds promise as the most accurate method for access surveillance. The relationship between low volume flow rate and thrombosis risk has been clearly demonstrated. By comparison, the relationship between presence of stenosis and conduit flow rates remains to be fully defined. A duplex-detected stenosis with a <2-mm residual lumen diameter, peak systolic velocity >400 cm/s, and peak systolic velocity ratio >2 across the stenosis indicates >50% DR but is not always associated with low (<500–800 mL/min) access volume flow. Duplex-derived volume flow calculations and Doppler ultrasonic techniques have been used to estimate time-averaged blood flow rates. Since the diameter of the vessel and time-averaged velocity can be measured by ultrasound methods, volume flow can be calculated using the formula:
where v is the time- and spatially averaged velocity over the lumen cross section and d and A are the lumen diameter and cross-sectional area at the site of velocity measurement. Assumptions made in this calculation are that blood flow is minimally disturbed and axial symmetric at the recording site, and the lumen cross section is circular. Testing should not be performed during or immediately after dialysis, since reduction in blood pressure caused by hypovolemia is a potential source of error by providing low volume flow estimates. Duplex ultrasound volume flow should be measured on a non-dialysis day in a normotensive patient. The software package for this calculation is available on essentially all contemporary duplex ultrasound units. For upper extremity autogenous accesses with variable vein diameters and branching patterns, volume flow measurement from the inflow artery is an appropriate option. Sites of flow measurement should be carefully selected—no stenosis or lumen narrowing should be imaged, and pulsed-Doppler velocity spectra should demonstrate only mild-to-moderate spectral broadening (Fig. 33.1). Vessel diameter (anterior-posterior dimension) is measured with transducer scanlines perpendicular to the long axis of the conduit in a longitudinal/sagittal view. The pulsed-Doppler sample volume is sized to encompass the entire lumen, and the velocity spectra recorded at a Doppler angle of 60°. The time-averaged velocity is calculated over 2–3 pulse cycles. Measurements are obtained in the inflow artery and at two to three locations along the access conduit, typically several (3–4) vessel diameters downstream of the arterial anastomosis, at mid-graft, and proximal to any venous anastomosis (Fig. 33.2). Experimental validation of duplex-derived volume flow measurements have been performed in baboons with an average error of 13% and good correlation (r = 0.9) with timed blood collection at flow rates in the range of 300 mL/min [20]. Similar validation data does not exist for the high volume and disturbed flow conditions present in arteriovenous access conduits, but it is estimated the reproducibility error is approximately 20–25%.

Access flow can also be measured during hemodialysis sessions using the transit-time, ultrasonic dilution method. Separate ultrasound transducers are placed on the arterial and venous dialysis tubing, and the lines are reversed so that the arterial line is downstream of the venous line within the access conduit. The dialysis circuit flow is fixed at 200–300 mL/min and ultrafiltration turned off. Rapid injection of 5–10 mL of normal saline at body temperature into the venous line dilutes the red cell mass in blood flowing through the access and results in alteration of the Doppler-derived velocity waveform recorded by the arterial line transducer. The measured areas under the perturbed velocity versus time curve at the venous (Sv) and arterial (Sa) lines and the known dialyzer flow rate (Qb) allows calculation of the access flow rate by the relation:

Accuracy of the access flow calculation appears independent of dialyzer flow rates between 177 and 350 mL/min but requires careful positioning of the arterial needle within the centerstream of the access flow [21]. Two or three access flow measurements should be made for reproducibility as the error between consecutive measurements averages 5% [20]. Clinical comparison of ultrasound dilution and Doppler-derived flow rate measurements has yielded acceptable agreement (correlation coefficients 0.79–0.83) over a wide range of access flow [17, 22, 23]. The ultrasound dilution technique does not have a tendency to over- or underestimate flow rates except in accesses with conduit stenosis where dilution measurements were lower than obtained by duplex ultrasound [22].

Low volume flow rates (<300–500 mL/min) have been shown to be predictive of PTFE bridge graft failure (Table 33.1) [5, 17–19, 24–29]. The thrombotic risk increases whether measured by Doppler-derived or ultrasound dilution techniques . The association between low volume flow and AVF patency is not as strong as for prosthetic bridge grafts likely due to lower thrombotic potential of the autologous venous conduit. The combination of access stenosis with low volume flow is also predictive of access failure [5, 19]. Flow rates were significantly less in conduits with stenosis than in functioning conduits with a maintained stenosis-free patency. In fact, the volume flow rate values in stenotic conduits were similar to those measured in accesses with subsequent thrombotic events [5]. These observations confirm access conduit stenosis associated with low or interval reduction in volume flow increases thrombotic risk. It should be noted that in some access conduits with low volume flow rates, a stenosis may not be detected. In these instances, a careful search for inflow or central vein occlusive disease should be conducted. Mechanisms of thrombosis in low-flow, non-stenotic access includes interval development of heart failure, hypovolemia, hypotension, hypercoagulability, or extrinsic pressure on the graft (e.g., particularly during access decannulation).

Controversy exists regarding the threshold access volume flow that should prompt intervention or additional imaging studies such as contrast fistulography. Based on current retrospective and uncontrolled prospective data, PTFE bridge grafts with access flow rates below 700–800 mL/min and reduced peak systolic velocities in the inflow brachial or radial artery <100 cm/s should be imaged with duplex scanning. Strauch et al. [7] originally described a lower volume flow cutoff (350 cm/s) to identify grafts that will likely fail without intervention. A recent comprehensive review documented the accuracy of various “low” volume flow thresholds for predicting access failure or need for revision [29]. In general, a very low-flow threshold (300–400 mL/min range) possesses the highest specificity for predicting subsequent access failure but risks losing patency in other access not meeting these criteria (i.e., high false-negative rate). Alternatively, higher values of a low-flow threshold (700–800 mL/min range) may minimize subsequent loss of patent accesses (higher sensitivity) but could result in increased numbers of nontherapeutic fistulograms (i.e., lower specificity due to more non-stenotic accesses subjected to fistulograms). Back et al. [27] found a threshold volume flow rate of 800 mL/min was a better discriminant of failing and functional AVFs and bridge grafts (with accuracy of 77%) than a flow rate greater than or less than 500 mL/min (accuracy of 67%). Regardless of the threshold levels chosen, grafts demonstrated to have flow-reducing stenoses should be more closely monitored for concomitant reduction in volume flow or observed access dysfunction.

Duplex scanning of a dialysis access provides both imaging and hemodynamic assessment of the arterial inflow, conduit, and venous outflow with precision comparable to contrast fistulography. The subcutaneous location of the access conduits allows the use of high frequency (7.5–10 MHz) linear array transducers to obtain high-resolution vessel imaging. Both transverse and longitudinal/sagittal B-mode, color Doppler, and power Doppler imaging can be used to image the access, anastomotic regions, and venous outflow including the central veins to the extent possible. The technique of duplex mapping is identical to peripheral artery scanning with color/power Doppler imaging for stenosis and pulsed-wave Doppler velocity spectra recording for stenosis classification in DR categories of <50 and >50% based on peak systolic velocity (PSV) and end-diastolic velocity (EDV). The velocity spectra of dialysis access hemodynamics are one of high-flow and low resistance with “normal” dialysis conduit PSV > 150 cm/s (Fig. 33.1). The resistive index should be 0.7 or less. Spectral broadening is typically present especially within the arterial anastomotic regions due to PSV > 200 cm/s and vessel tortuosity producing highly disturbed flow conditions. Proper selection of the color bar and velocity scale helps to minimize aliasing artifacts associated with color Doppler imaging. Access graft stenosis is identified by color flow imaging of a narrowed lumen and at least doubling of peak systolic velocities compared with adjacent graft segments (Fig. 33.3). In AVF stenoses, a perivascular color artifact may be present and represents a tissue bruit caused by turbulent flow and vessel wall/tissue vibration. The duplex criteria for classification criteria dialysis access stenosis severity is based on PSV in the stenosis and reduction of conduit flow velocity in a non-stenotic segment (Table 33.2). In general, a >50% DR, flow-reducing stenosis is associated with PSV > 400 cm/s and a conduit and inflow artery PSV < 150 cm/s. When the access is thrombosed, the inflow artery will demonstrate a triphasic, high-resistance velocity spectra, and no Doppler signal will be recorded within the occluded access conduit. In general, a hemodynamically significant (>50% DR) stenosis that impairs volume flow and should be repaired has the duplex features of a PSV > 400 cm/s, EDV > 250 cm/s, a local PSV ratio >2, and a residual color or power Doppler flow lumen than 2–3 mm in diameter.