It is still a matter of debate as to which parameters should be used f
or noninvasive diagnosis of renovascular disease by renal Doppler sono
graphy (RDS). The accuracy of RDS in the detection of renal artery ste
nosis (RAS) was tested in 95 consecutive, moderate to severe hypertens
ive patients (I-II World Health Organization [WHO] stages). Reno-aorti
c ratio (RAR) for peak systolic velocity (PSV) was also calculated to
assist in the diagnosis of significant (>50%) RAS. Paired receiver-ope
rating characteristic (ROC) analysis was plotted for evaluating the re
lationship between sensitivity and specificity for each parameter. In
a subset of 57 kidneys, the influence of blood pressure and age on int
raparenchymal parameters was evaluated. Measurements of maximal peak s
ystolic velocity (PSV) at the site of stenosis, RAR for PSV, and minim
um acceleration index in the main renal artery showed high accuracy (a
reas under the ROC curve 0.97, 0.88, and 0.80, respectively). Among in
traparenchymal parameters, early systolic acceleration showed the best
area under the ROC curve (0.90), but provided a low positive predicti
ve value (29%) and was significantly influenced by blood pressure (mul
tiple r=0.56; p=0.001). Pulsatility and resistive indices were found t
o be less powerful as absolute values, and both significantly influenc
ed by blood pressure and age (multiple r = 0.60 and 0.50; p=0.001, p=0
.02, respectively). However, interindividual variance of intrarenal in
dices should be minimized by calculation of side difference, although
this procedure would become misleading or impossible in patients with
bilateral RAS or a single kidney, respectively. These results support
the use of extraparenchymal parameters for noninvasive detection of RA
S, and emphasize that intrarenal parameters cannot be considered as ab
solute values.