Although the technique of renal arterial US scanning has been well established for years, a lot of difficulties in reliably identifying main and accessory renal arteries remain [8–10, 25]. Most of these dificulties are related to the patient obesity, the presence of bowel gas, excessive respiratory movement, and the depth and tortuosity of the renal arteries [8, 16]. The time expended in the examination can be too long as almost 60 minutes , and failure of technique varies from 9 to 25%. In our study, non-enhanced Doppler US showed a feasibility rate of 100%, similar to some single centers, but higher than the majority of studies using this technique (58–90%) . Indeed, two recently published studies reported feasibility not exceeding 11% and 12% [22, 23]. One of the reasons of our high rate of feasibility probably is related to the quality of the machine, which allowed a scan imaging with an excellent definition. To our knowledge, the present study is the first randomized study in a selected group of hypertensive patients in which renal arterial color Doppler flow US with and without a continuous infusion of PESDA was compared against the reference standard of angiography. The infusion of PESDA did not alter the feasibility that remains 100%. In a multicentric study  using Levovist as the US contrast, the infusion increased by 20% the number of patients in whom renal arteries could be evaluated, including difficult cases such as those involving patients who are obese and patients with impaired renal function. However, some centers participating of the study also presented a feasibility of 100% and the Levovist infusion did not interfere in the results. In our study only one obese patient had a better visualization of renal artery after contrast infusion, and in all patients with renal failure, the non-enhanced US was able to localize renal arteries.
The most important conclusion from this study is that both sensitivity and specificity of Doppler US of renal arteries are strongly dependent on the criterion used, and the infusion of PESDA contrast seems not to improve it significantly, although we observe a slight increase in specificity. Thus, the best sensitivity was obtained when based on resistance index <0.8 (82.9%) but at expense of a low specificity (56.8%). On the other hand, the best specificity was obtained with renal aortic ratio >3 (89.2%), but the sensitivity was low (56.2%). In addition, the sensitivity and specificity for a peak systolic velocity of 1.5 m/sec showed intermediate values, respectively, 61.9% and 64.9%. An analysis of previously published studies [8–10, 12, 15, 16, 20] based on non enhanced Doppler evaluation of the renal artery clearly shows that the diagnostic criteria and respective threshold values fluctuate from one report to the other. Miralles et al  reported a sensitivity of 87.3% and a specificity of 91.5% for a higher peak systolic velocity (>1.98 m/sec) and a higher renal aortic ratio (>3.3), while Olin et al  reported a sensitivity of 98% and specificity of 98% for a quite similar criteria. Helenon et al  quoted a sensitivity of 89% and a specificity of 99% with use of a peak systolic velocity cutoff point similar to our study (1.5 m/sec) but taking into account the presence of poststenotic turbulence and not renal aortic ratio. Moreover, in the multicentric study cited above comparing non enhanced and enhanced Doppler US , renal aortic ratio was more accurate than peak systolic velocity in the diagnosis of a renal arterial stenosis greater than 50%, but it was difficult to determine a precise cutoff point. In the same study, the authors demonstrated, in terms of patients, a sensitivity of 80.0% and a specificity of 80.8%, but according to renal arteries the sensitivity was lower (66.7%) and the specificity was higher (90.4%). These latter results were quite similar to our results based on RAR criteria, also evaluated according to renal arteries. These facts, determination of accuracy in terms of renal arteries and not in terms of patients, can explain in part the differences encountered between our conclusions and those from the studies mentioned above.
The continuous infusion of PESDA contrast increased moderately the specificity for renal aortic ratio criteria from 89.2% to 97.3% but at the expense of a significant decrease of sensitivity from 56.2% to 33.3%. For the another criteria, peak systolic velocity and resistance index the infusion of PESDA decreased mildly or did not affect the sensitivity and specificity. MISSOURIS et al  have reported an increase of sensitivity from 85% to 94% and of specificity from 79% to 88% after injection of microbubble Levovist® in hypertensives with renal artery stenosis. In a more recent multicentric study the contrast Levovist did not affect either sensitivity or specificity: sensitivity was 80.0%–83.7%, whereas specificity moderately increased from 80.8% to 83.6% or 86.2%, depending on the subgroups of comparable patients. In two single-center studies in which the value of Doppler US examination after intravenous injection of contrast agents for the diagnosis of renal arterial was also evaluated it was demonstrated an improvement in sensitivity, which increased from 83% to 95% in one study and from 75% to 100% in the other [22, 23]. However, both of these studies were based on a limited number of patients with a very low feasibility rate of 11% and 12% at baseline examination, respectively. In addition, Melany et al  reported that contrast Levovist injection did not improve specificity, as we also demonstrated with PESDA infusion.
There is no consensus whether contrast agent injection potentially reduces examination duration. In our study, we found a significant reduction of mean examination time after contrast infusion (17%). In other study, it was reported that the use of Levovist dramatically reduced the mean examination time from 24.5 minutes to 13.5 minutes . This advantage could be of potential economic interest, but subsequent studies have to confirm more significant differences.
PESDA was well tolerated and did not compromise the safety of US. This excellent patient tolerance has already been demonstrated in stress echocardiograph studies that used PESDA as contrast agent .
The small number of patients impose some limitations to the present study, However, the high prevalence of renal artery stenosis in this selected group of hypertensives counterbalance this limitation.
In conclusion, the detection of renal artery stenosis by Doppler US depends on the criteria used and infusion of PESDA contrast seems not to improve the accuracy, despite a reduction in the examination duration and an increase in specificity based on one Doppler criterion. Also, the feasibility of US is dependent of the quality of the machine, and the infusion of contrast does not add advantages if the performance of the US machine is excellent. However, it remains unknown if the PESDA infusion can improve feasibility if the machine does not have a good imaging quality. So, there is a need for establishing a consensus opinion regarding Doppler useful criteria and thresholds for the diagnosis of renal arterial stenosis, regardless of the US equipment used or infusion of ultrasonographic contrast.