14.2 Duplex Evaluation of AV Fistulas & Grafts

Understand the circuit and the indication

An arteriovenous fistula (AVF) directly connects an artery and native vein. An arteriovenous graft (AVG) uses a conduit between arterial inflow and venous outflow. Both create a low-resistance circuit for dialysis, but their anatomy and common failure sites differ. Review the operation and identify inflow and outflow rather than assuming the arm's visible direction reveals the circuit.

A radiocephalic fistula connects radial artery and cephalic vein, a brachiocephalic fistula uses brachial inflow and cephalic outflow, and a brachiobasilic access uses the basilic vein with a surgical strategy to make it cannulatable. Grafts have arterial and venous limbs whose orientation depends on the operation. Use the operative record and the actual flow direction to label them correctly.

Duplex is useful for maturation problems, suspected stenosis or thrombosis, swelling, difficult cannulation, abnormal dialysis performance, and other clinical questions. KDOQI distinguishes clinical monitoring from instrument-based surveillance. Findings from surveillance are supplementary; routine preemptive treatment of an asymptomatic narrowing is not justified solely because an isolated flow value crosses a threshold.

Examine the whole pathway

Use a linear probe for superficial segments and a lower-frequency approach when needed for deeper outflow. Scan in multiple planes and optimize for the actual flow. Avoid excessive pressure on vulnerable skin or an aneurysmal segment. Document inaccessible central veins rather than declaring central patency from a single peripheral waveform.

A patent low-resistance shunt usually increases forward diastolic flow in the feeding artery and arterializes the venous waveform. High velocities and broadening near an anastomosis can be expected from geometry. They are not independently proof of stenosis. Conversely, loss of the expected low-resistance pattern can suggest reduced shunt flow or thrombosis, but the full circuit and clinical findings determine the interpretation.

Measuring volume flow

Volume flow (Q) is cross-sectional area multiplied by mean velocity. With diameter in centimeters and mean velocity in cm/s:

$$Q = \pi\left(\frac{d}{2}\right)^2 \times V_{\text{mean}} \times 60$$

The result is mL/min because one cubic centimeter is one milliliter. The velocity must be a cross-sectional/time mean appropriate to the system's flow method, not simply PSV or the time-averaged maximum envelope. Review the machine's algorithm and protocol; a mislabeled mean can create a plausible but incorrect flow value.

Choose the protocol-specified straight, non-tapering measurement site away from marked turbulence. Measure diameter accurately in an appropriate perpendicular plane and use representative angle-corrected flow data. The sample volume should capture the lumen as required by the validated method. Average representative cycles and repeat measurements when rhythm or flow is variable. Site choice may differ for a native fistula and a prosthetic graft.

Worked flow calculation

For diameter 0.60 cm and appropriate mean velocity 50 cm/s, area = π × 0.30² = 0.2827 cm². Flow = 0.2827 × 50 × 60 = approximately 848 mL/min. Using the 6 mm diameter as “6 cm” would create a hundredfold area error. Substituting a PSV of 150 for the actual mean of 50 would overestimate flow threefold.

Diameter error is squared. Measuring 10% too large multiplies calculated area/flow by 1.1² = 1.21, a 21% overestimate. Measuring 10% too small multiplies it by 0.9² = 0.81, a 19% underestimate. Do not state that every 10% error has the same signed percentage effect.

Interpret changes with clinical monitoring

Historical screening values such as flow below 400–500 mL/min in a fistula or below 600 mL/min in a graft can prompt concern in some protocols. They are not universal intervention triggers. A substantial decline from the patient's baseline, difficult cannulation, prolonged bleeding, abnormal thrill/bruit, poor dialysis delivery, or elevated venous pressures strengthens concern for dysfunction.

For example, an AVG flow decreasing from 1150 to 480 mL/min is a significant change worth verifying and communicating. Check technique, blood pressure, rhythm, and the full circuit. It does not independently prove that thrombosis is imminent or mandate angioplasty in an otherwise asymptomatic patient. KDOQI does not recommend preemptive angioplasty of AVF/AVG stenosis without clinical indicators merely to improve patency.

Very high flow can contribute to cardiac strain or hand ischemia in susceptible patients, but no single flow cutoff proves high-output heart failure. Symptoms, cardiac output, access flow fraction, underlying heart disease, and the clinical assessment matter. Report the measurement and relevant observations; the treating team decides the response.

Finish with an actionable report

Describe the access configuration, indication, examined extent, patency, lesions, measurements, flow method/site, baseline comparison, and limitations. Identify concerning skin changes, absent flow, or symptoms requiring urgent communication. Duplex supports a clinical decision by describing the entire circuit accurately, not by declaring a universal “normal” range detached from the patient's access.

KDOQI 2019 vascular access guideline — checked October 18, 2026; see monitoring/surveillance and statements 13.6–13.9.