Does separating the resistive index into pre- and postglomerular resistance and vascular compliance improve the diagnostic accuracy of renal transplant doppler ultrasound?

Boas FE, Desser TS, and Kamaya A. (2011) AJR 196(5): A31-35. Presented at the ARRS meeting, Chicago, 2011-05-03.

Purpose: Doppler ultrasound resistive index (RI) is a widely used parameter in the assessment of renal transplant function, but current work suggests that its sensitivity and specificity for acute rejection is low. Although RI (which is a measure of pulsatile flow) does increase with post-glomerular vascular resistance, it also increases with increased vascular compliance (decreased atherosclerosis), decreased pre-glomerular vascular resistance, increased pulse pressure, and increased heart rate. We postulated that separating out these potentially confounding factors and directly calculating vascular resistance and compliance, would allow for more accurate assessment of kidney transplants.

Materials and methods: 47 kidney transplant ultrasounds were analyzed. RI was measured at the segmental renal arteries. We modeled the transplanted kidney vasculature as a tube with a pre-glomerular and post-glomerular vascular resistance, and a vascular compliance (three-element Windkessel model). This model accurately describes in vivo blood flow in rat kidneys. The arterial blood pressure was modeled using a range of waveforms from the MGH/MF Waveform Database, adjusted to match the patient's measured heart rate, systolic blood pressure, and diastolic blood pressure. The parameters of the model were adjusted to fit the experimental mid renal artery velocity waveforms. The final diagnosis was determined from pathology (when available) and the nephrology clinic visit notes.

Results: Normal transplant kidneys have an average RI of 0.71 ± 0.11, and kidneys in acute rejection have an RI of 0.77 ± 0.11. Using a cutoff of 0.8 results in a sensitivity of 38% and specificity of 63% for acute rejection. A resistive index > 1 was only seen in renal vein thrombosis (2 cases). Waveforms for acute rejection, chronic rejection, hydronephrosis, and delayed graft function all had a large amount of overlap with normal waveforms (Figure A). Using the three-element Windkessel model, we showed that these 4 diagnoses are associated with vascular resistances and compliances within 1 standard deviation of normal, and none of the differences were statistically significant (Figure B). On the other hand, renal vein thrombosis is associated with a non-significant increase in post-glomerular resistance, and a significant decreases in pre-glomerular resistance (p=0.01) and vascular compliance (p=0.0003).


Figure A: Average renal artery waveforms in normal kidney transplants, acute rejection, and renal vein thrombosis.


Figure B: Pre- and post-glomerular resistance and vascular compliance for each diagnosis.


Conclusion: Doppler ultrasound of kidney transplants has limited value in diagnosing acute rejection. We examined resistive index, pre-glomerular resistance, post-glomerular resistance, vascular compliance, and the shape of the mid renal artery velocity waveform. None of these variables could reliably predict acute rejection, chronic rejection, hydronephrosis, or delayed graft function. However, resistive index > 1 was associated with renal vein thrombosis in 2 out of 2 cases.


NormalDecreased vascular compliance (C)



Increased pre-glomerular resistance (R1)Increased post-glomerular resistance (R2)

Figure C: The shape of the renal artery flow waveform (flow velocity vs time) depends on the pre- and post-glomerular resistance and the vascular compliance (three-element Windkessel model).


Figure D: Intra-renal resistive index for each diagnosis.


View the:

Ed Boas' home page | Stanford radiology | CT metal artifact reduction