Intra-renal Doppler wave-form analysis as a screening test for renal artery stenosis

Australasian Radiology (1 996) 40, 276-282

Intra-renal Doppler wave-form analysis as a screening test for renal artery stenosis Philip Lucas, Steven Blome and James Roche Department of Diagnostic Radiology, Royal North Shore Hospital, St Leonards, Sydney, Australia

SUMMARY

A double-blinded trial was performed comparing Doppler interrogation of intra-renal arteries with findings at renal angiography in the assessment of renal artery stenosis. Fifty-three consecutive patients were studied.

Previously reported parameters and cut-off values were used for predicting the presence of renal artery stenosis with Doppler. Doppler wave-form analysis proved relatively specific (91 %) but insensitive (68%) t o the presence of renal artery stenosis (of greater than 60%) if acceleration parameters were used alone. The technique of using a difference in mean resistive indices was found to be both insensitive (61%) and non-specific (54%).

Patients with fibromuscular dysplasia, branch vessel stenoses, multiple renal arteries and chronic renal impairment proved to cause the most false negatives. Young patients with normal arteries were the source of the few false positive results.

The interrogation of intra-renal arteries does not have appropriate sensitivity t o be useful as a screening test for renal artery stenosis.

Key words: angiography; Doppler ultrasound; intra-renal wave-form analysis; renal artery stenosis.

INTRODUCTION In the Australian community, hypertension is a significant problem affecting approximately 20% of the adult population.‘ Although the majority will have benign essential hypertension, a small number will have secondary causes which may be amenable to intervention and reduce the need for anti- hypertensive medication. In particular, renovascular hypertension secondary to stenosis of the renal artery or its branches accounts for 1-4% of cases of hypertension, with a prevalence of 30-40% in a selected subgroup with malignant hypertension and renal The underlying pathology in approximately two-thirds is atherosclerosis and in about one third, fibromuscular dysplasia.2 Intervention, by way of per- cutaneous transluminal angioplasty or re-vascularisation surgery, has been shown to be useful in treating hypertension and improving renal function in patients with significant renal artery ~tenosis.~

The gold standard for the diagnosis of renal artery stenosis (RAS) is angiography. However, in view of the inherent risks and cost of this relatively invasive procedure, it is not an ideal screening test for renovascular disease. Other less invasive tests have been used including rapid sequence intravenous

urogram and radio-isotape renograms with or without admin- istration of Captopril.

Doppler interrogation of the renal artery itself has been utilized over the last decade and can be successful but it is often technically very difficult, operator-dependent and time-con~uming.~~~

The tardus-parvus phenomena observed on renal artery spectral Doppler has recently been describeds-B and refers to the change in spectral wave-form seen distal to a tight stenosis. It reflects loss of the high frequency Doppler shifts early in systole with a dampened wave-form downstream.’ This is analogous to the ‘anachrotic’ pulse felt in the carotid arteries in a patient with severe aortic stenosis. The proposed mecha- nism for this effect is a pressure drop over the segment of stenosis which dampens the high-frequency component of the Doppler spectrum.7 More recent work suggests that the change is not just dependent on the trans-stenotic pressure drop but also related to the compliance in the downstream vascular bed.15

Several recent s t ~ d i e s ~ - ~ have evaluated this phenomenon in the detection of significant RAS, by interrogating distal segmental, inter-lobar or arcuate arteries within the kidney. By

P Lucas MB, BS; S Blome MB, BS FRACR; J Roche MB, ChB, FRACR. Correspondence: Dr P Lucas, Department of Diagnostic Radiology, Royal North Shore Hospital, St Leonards, NSW 2065, Australia. Submitted 25 October 1995; accepted 1 February 1996.

DOPPLER WAVE-FORM ANALYSIS 277

analysing the acceleration parameters and early systolic peak or the difference in mean intra-renal resistive indices

(Rl)g they have claimed relatively high sensitivities and specificities for the detection of upstream stenosis.

The purpose of this study was to review prospectively the use of intra-renal Doppler wave-form analysis as a screening test for RAS.

METHOD We prospectively examined 53 patients with a total of 104 kidneys using colour flow Doppler and duplex sonography. All of these patients were recruited sequentially from those who were to undergo angiography to rule out renovascular disease between February 1993 and November 1993. The scans were performed either before or after angiography due to availability of equipment, although both the sonographers (PL, SB) and angiographers (JR) were blinded to the results of the oppos- ing modality and the sonographers were also blind to previous imaging and diagnostic tests.

Duplex and colour Doppler examinations were performed on an ACUSON 128 unit (Acuson, Mountain View, California), with a 5 MHZ vector phased array transducer in the majority. In a small minority, a curved linear array 3.5 MHZ transducer was required for adequate penetration.

Spectral wave-forms over at least three cardiac cycles were obtained from intra-renal segmental, inter-lobar or arcuate arteries in both the upper and lower poles and often from the mid/hilar region. Care was taken to have an angle of interrogation less than 60 degrees to the vessel in question. Colour Doppler was used to locate suitable vessels for interrogation and it was usually necessary for patients to hold their breath during the recording of the Doppler signal.

Maximum Doppler power for both colour and spectral Doppler was utilized. The display was set with the lowest velocity scale which still prevented aliasing. A sweep speed of 50 mm/s and a relatively large Doppler gate were utilised.

The patient was scanned supine or decubitus with a postero- lateral window to the kidney. The examinations were timed and were on average 20-30 m in duration.

Spectral analysis included measurement of resistive indices (RI) acceleration the acceleration of the early systolic upstrokes and the presence or absence of an early systolic peak (ESP). Display calipers on the monitor's spectral display and the machine software were utilized. ESP is the transient peak of the initial upstroke of systolelo - detected qualita- tively. Acceleration time is the time between onset of systole and the ESP or the highest point of the systolic upstroke if there was no ESP (Fig. 1). The acceleration of early systole is the gradient of the initial upstroke of systole and is similar to the acceleration index described by Hanna.'j The RI were calculatedg for each vessel interrogated and a mean RI was calculated for each kidney (sum of the RI divided by number of vessels examined) and the difference between the mean RI of the two kidneys calculated and re~orded.~ If there was only one kidney or if no suitable wave-form could be obtained from a kidney then this could not be calculated.

Digital subtraction angiography was performed on the same day in all patients. A standard Seldinger technique was used via the femoral artery with a 4 or 5 French pigtail catheter for aortic injection a 5 French catheter for selective renal artery injection. Midstream AP and obliques were always obtained and selective renal artery injections performed as required. The degree of stenosis was calculated with a handheld ruler by a single angiographer (JR). Significant stenosis was con-

Flg. 1 Normal intra-renal Dop- pler spectral pattern. There is a sharp systolic upstroke with an early systolic peak (arrowhead). Acceleration index is calculated by change in velocity (4V) divided by change in time (AT).

278 P LUCAS ETAL.

sidered 360% narrowing of the artery’s diameter. This degree of stenosis was chosen as it correlated with previous reports of this technique and there is evidence in the literature that there is benefit to the patient with intervention with 60% stenoses or mom3 The number of renal arteries and the site of any stenoses (is. main renal artery or branch artery) were recorded.

The purpose of the study was to determine which of the patients screened with Doppler examination would or would not require angiography to further evaluate suspected RAS, if Doppler was to be used as a screening tool. This was achieved in three ways.

screening cut-off parameters were chosen as AT 20.07 s, acceleration of early systole ~ 3 . 0 m/s2 and in particular absence of the early systolic peak. The presence of two or more of these criteria was regarded as probable RAS and considered as positive screening result for Doppler.

The method used by Schwerk et a/.@ based on the mean RI difference was also analysed. A mean RI difference of 25% was considered a positive Doppler screen result. Analysis of a mean RI difference of 210% was also made.

Finally, a combined approach, where an abnormal early systole (i.e. decreased acceleration (~3.0 m/s2) or absence of the ESP) and/or a mean RI difference of 25% were considered a positive screening result, was analysed.

Calculation of sensitivities and specificities for each approach was performed.

Firstly, as described by Stavros et

RESULTS A total of 53 patients was studied prospectively (28 women and 25 men) with an age range of 25-82 years (mean 58 years). Two patients had had nephrectomies, so that there was a total of 104 kidneys.

The indications for angiography included: uncontrolled hypertension (n=42), hypertension with chronic renal impairment suggesting renovascular disease (n=5), recent renal infarct detected on CT scans with possible firbomuscular dysplasia (n=2), and known RAS prior to angioplasty (n=4).

Doppler examination failed to give an adequate diagnostic wave-form in three cases (all with normal renal arteries at angiography). In at least two of these patients, the inability to suspend respiration long enough for a suitable trace to be obtained was at least partially responsible.

Angiography was considered diagnostic in all cases. In 14 of the 104 kidneys (1 3%) there were two or more renal arteries present.

Twenty-two of the 104 kidneys (21 %) showed significant stenosis with 260% narrowing the diameter of either the renal artery or one of its major branches (i.e. first order branches). One complete occlusion was seen. Two patients had bilateral stenoses. Eighteen of the 22 were origin stenoses, the other four being situated in one of the major branches.

The pathology present, as judged from angiography, was presumed atherosclerosis in 17 arteries and fibromuscular dysplasia in five arteries. Branch stenoses were commonest in fibromuscular dysplasia, seen in three patients, whereas only two of the dysplasia group had proximal main renal artery stenosis. The pathology, however, has not been confirmed.

If Doppler parameters of acceleration time and index of early systole or ESP (as per Stavros) were used then 15 of the 22 RAS would have been correctly predicted (Table 1, Fig. 2). Seven other significant stenoses would have failed to be identified. Four of the 82 kidneys with no significant stenosis would have been incorrectly directed to angiography to confirm the Doppler results.

Patients in whom a wave-form trace was not able to be obtained were considered in the false positive group as they would require angiography if they were in a screening group. The sensitivity and specificity for Doppler wave-form analysis with this approach are 68% and 91% respectively.

The false negative group contained one patient with an accessory renal artery which was stenosed (presumably ath- erosclerotic), three patients with presumed fibromuscular dysplasia with either severe segmental artery narrowing (n=2) or stenosis of an accessory artery (n=1) and three patients with known longstanding chronic renal impairment who demonstrated high resistive indices, compared with the remainder of the study population, with sharp systolic peaks (Figs 3,4).

False positives occurred in four patients, three of whom

Table 1. Analysis of acceleration parameters alone

Angiography Significant stenosis No significant

Doppler wave-form analysis 2 60% stenosis

Negative screen

Positive screen

No adequate Doppler wave-form

AT < 0.07s; A>3 ms2; +ve EPS 7 75

AT 3 0.01; A s 3 ins2; EPS Neg 15 4

obtained 0 3

Total 22 82

Table 2. Analysis of mean RI difference (cut-off RI difference of 35%)


Doppler wave-form analysis 360% stenosis

Positive mean RI difference 35% 11 11

Negative mean RI difference <5% 7 19

No comparison available 1 4

Total 19 34


were relatively young (<40 years) with no known medical con- ditions apart from hypertension (Fig. 5).

For the approach of Schwerk et a/.,9 using a mean RI difference of 35% as a cut-off between positive and negative Doppler screen, then Doppler predicted 11 out of the 19 patients with significant stenosis. This technique failed to iden- tify the other eight (one case because the patient had had a previous nephrectomy) (Table 2). Of the 34 patients with angiograms showing no significant stenosis, Doppler wave- form analysis incorrectly predicted 11 of these patients as probable RAS. In five patients no comparison of RI could be made due to either nephrectomy (n=2) or inability to obtain an adequate waveform from one of the kidneys. If these are regarded for screening purposes as false positives (as they would require angiography to further assess) then a sensitivity

Flg. 2 (a) Tardus-parvus phenomenon. Selective left renal angiogram in 22 year old female with hypertension demonstrates tight stenosis (>go%) at trifurcation of main renal artery. (b) Doppler wave-form from superior left kidney demonstrates marked reduction in wave-form height, slow systolic upstroke and loss of early systolic peak -+ 'Tardus-Parvus phenomenon'. This is a true positive case. (c) Selective right renal angiogram in same patient demonstrates beaded appearance of fibromuscular dysplasia without significant stenosis.

of 61% and a specificity of 54% were calculated for this approach.

Table 3 shows the results obtained if a mean RI difference of >lo% was used as a screening cut-off.

This approach yielded a sensitivity of 22% and a specificity of 86%.

If a combined approach was adopted utilizing an analysis of early systole acceleration parameters and/or a mean RI difference of 35% then a sensitivity of 68% was achieved with a specificity of 47% (Table 4).

False negatives using the mean RI difference approach also included two patients with fibromuscular dysplasia and branch stenosis, one accessory renal artery with atheromatous origin stenosis and two patients with chronic renal impairment with bilateral high resistive indices.

280 P LUCAS ETAL.

Table 3. Analysis of mean RI difference (cut-off RI difference 310%)


Doppler wave-form analysis 260% stenosis

Positive mean Ri difference 310%

Negative mean RI difference <lo%

No comparison available

Total

4

14 1

19

0

30 4

34

Table 4. Analysis of combined approach (As3 ms2/EPS negative and/or an RI difference of 35%)


Doppler wave-form analysis >60% stenosis

Positive screen combined approach 14 15 Negative screen combined approach 5 16

Unable to determine any trace 0 3

Total 19 34

Flg. 3 (a) Right renal artery from midstream aortogram of a 67 year old female with chronic renal impairment, hypertension and deteriorat- ing renal function. Tight irregular stenosis (approx. 80%) of main renal artery is demonstrated. (b) Intra-renal Doppler wave-form from right kidney inferior pole. High resistive wave-form but retention of steep systolic upstroke. False negative case.

Fig. 4 (a) Branch stenosis (false negative) in a 68 year old male with hypertension. Aortogram with views of right renal artery showing tight stenosis of major branch to the lower pole with poor perfusion. (b) Intra-renal Doppler wave-form from right kidney lower pole. Normal systolic upstroke with early systolic peak present. False negative case.


Fig. 5 (a) 34 year old male with hypertension. Right renal artery from midstream aortogram. No stenosis demonstrated. (b) Intra-renal Dop- pler wave-form of right kidney upper pole. Tardus-parvus phenomenon with decreased slope to systolic upstroke and loss of early systolic peak. False positive case.

DISCUSSION A suitable screening test for RAS should have a high sensitivity with a similar relatively high specificity, be non-invasive, cheap, technically straightfoward and reproducible. Screening for RAS may be useful in patients with difficult-to-control hypertension, those of a young age with no family history of hypertension, generalized atherosclerosis, renal bruit, hypocalaemia and in patients with a deterioration in renal function, particularly if precipitated by use of an angiotensin converting enzyme inhibitor.2

Several other tests besides angiography have been used in the diagnosis of RAS in the past several years. The Capto- pril test - utilizing peripheral venous sampling of plasma renin activity after a single oral dose of Captopril - has good specificity but is very insensitive and not useful in an unselected population.2

Renal vein renin sampling is an invasive procedure and is very expensive, time-consuming and relies on co-ordinated handling, labelling and processing of samples.2,8 It also has high false positive and negative rates.

Isotope renography with technetium labelled DTPA has been used with and without Captopril. It has reported sensitivities and specificities of 71 -94% and 75-95%, re~pectively.~ This procedure has largely replaced renal vein sampling as the determinant of functional significant stenosis. It carries the risk, however, of post-procedure hypotension and even acute renal failure.

Duplex ultrasound studies of the renal arteries proper has reported sensitivities of 79-91% especially if there is only a single renal artery.” Some reports have shown in large unselected groups much lower sensitivitie~.~,~ Technically, the examination is difficult (especially in obese patients) and the renal artery is often obscured by bowel gas. False negatives also occur in the presence of the multiple renal arteries. In the last two years, several authors have reported the technique of Doppler examination of intra-renal arteries and analysis of the Doppler wave-form, with the tardus-parvus phenomena6-8,12 or a difference in mean Rl,9.12 as an indication of upstream stenosis.

Lafortune and Patriquin7.l2 showed this phenomenon experi- mentally in dogs’ renal arteries which were progressively ‘stenosed’ with a snare to a critical stenosis of 75% or greater. They confirmed their results in a cohort study of children suspected of having RAS. Stavros et aL8 reported a sensitivity of 95% and a specificity of 97% in detection of RAS 360% with analysis of the Doppler wave-form, especially the presence or absence of an early systolic peak. Their study population, as they acknowledged, was a selected group with known atherosclerosis.

Schwerk et aL9 found a difference in mean RI of 35% had a sensitivity for RAS (greater than 50%) of 82% and a specificity of 92%. They suggested that applying this analysis would eliminate any uncontrolled factors up- and downstream such as vasomotor tone, by adding a contralateral control. They indicated major limitations, including the difficulty with assessment in patients with bilateral RAS and the inability to establish a difference if only one kidney is present.

Both methods of analysis of Doppler wave-form, described above, were assessed. This present study was unable to reproduce the sensitivities of these recent investigators, to select those individuals who required angiography to prove or exclude RAS.

The design was to assess spectral wave-form analysis as

P LUCAS ETAL.

a possible screening test. By using the acceleration parameters and early systolic peak detection, a sensitivity of 68% and a specificity of 91% were obtained to detect a stenosis 260%. If an analysis of the intra-renal difference in RI was made then a sensitivity of 61% and a specificity of 54% could be achieved. By changing the difference in RI to 10% the test was made more specific (86%) but extremely insensitive (22%).

Combining these two approaches showed some improve- ment in sensitivity (68%) but a similar low specificity of 47%.

The insensitivity of this test has previously been reported in the literature by KIiewerl4 who, by assessing acceleration parameters only, found a sensitivity of only 71 % and a specificity of 52% in an analysis of 1 18 kidneys.

The false negatives in both approaches were more likely to occur in those with multiple renal arteries and in those with fibrornuscular dysplasia and significant segmental stenosis. This may be partly due to the fact that although multiple arteries were sampled (upper, lower, mid-poles), those with the brightest colour Doppler signal tended to be selected for interrogation as they were easier to obtain a good wave-form envelope to analyse. This approach may unsuspectingly avoid branches from the stenosed artery or branch stenoses. The introduction of colour power Doppler may allow interrogation of these intra-renal arteries with decreased flow.

In addition, the vascular anatomy does not follow a simple upper mid- and lower pole distribution, with the distribution from multiple renal arteries being extremely variable. Kotval13 also found that the tardus-pawus phenomenon was not as obvious in the presence of well-developed collaterals. The presence of multiple arteries may act like a collateral network and therefore make the tardus-parvus phenomenon less obvious. It may be that such collaterals render these stenoses of little functional significance in causing renovascular hypertension.

False negatives were also common in those with chronic renal impairment with significantly increased resistive indices in which a sharp acceleration peak persisted despite marked proximal stenosis.

Recent work by Bude et a/.,15 with hydrodynamic experimental studies, found that the degree of dampening of the high frequency component of the wave-form (or tardus-parvus) decreased as the compliance of the post-stenotic segment of vessel increased and was relatively independent of the trans- stenotic pressure drop. This may explain the false negatives in those with chronic renal impairment and high intra-renal resistive indices whose intra-renal vessels may have very little compliance secondary to hardening and act like non-distensible tubing and therefore demonstrate minimal tardus-parvus effect.

It also may explain why false positives were common in young people where a tardus-parvus effect occurred despite no stenosis, as their vessels are very compliant and may need only a minimal pressure drop to give a dampened wave-form.

CONCLUSION The tardus-parvus phenomenon is a haemodynamic effect related to upstream stenosis and downstream vessel compli-

ance. As the study is performed on intra-renal vessels, it appears as an attractive, easy alternative to renal artery interrogation to assess indirectly for RAS. At our institution, however, the sensitivities and specificities of the test are inadequate to recommend it as a screening tool for RAS in an unselected population.

It may be that the stenoses which were identified angio- graphically but which did not demonstrate the tardus-paws phenomenon are not functionally significant as a cause for hypertension. Thus, a controlled study incorporating treatment outcomes may be of value to determine if these apparent false negatives on analysis of spectral wave-form have a reduction in blood pressure after angioplasty/surgery.

ACKNOWLEDGEMENT The authors would like to thank Miss Lindsey Drysdale for the preparation of the manuscript.

REFERENCES 1.

2.

3.

4.

5.

6.

7.

8.

9.

10.

11.

12.

13.

14.

Harrison's Principles of lnternal Medicine, 6th ed. McGraw Hill. 1970. McGrath B, Clarke K. Renal artery stenosis: Current diagnosis and treatment. Medical J. Aust. 1993; 158: 343-6. Lossino F, Zuccala A et a/. Renal artery angioplasty for renovascular hypertension and preservation of renal function. AIR 1994;

Berland LL, Koslin DB et a/. Renal artery stenosis: Prospective evaluation of diagnosis with colour duplex US compared with angiography. Radiology 1990; 174: 421 -3. Desberg AL, Paushter DM et a/. Renal artery stenosis: Evaluation with colour Doppler flow imaging. Radiology 1990; 177: 749-3. Handa N, Fukunaga R etal. Efficacy of echo-Doppler examination for the evaluation of renovascular disease. Ultrasound Med. Biol.

Lafortune M, Patriquin H etal. Renal artery stenosis: Slowed systole in the downstream circulation: Experimental study in dogs. Radiology 1992; 184: 474-8. Stavros AT, Parker SH. Segmental stenosis of the renal artery: Pattern recognition of the tardus parvus abnormalities with duplex sonography. Radiology 1992; 184: 487-92. Schwerk WB, Restrepo I. Renal artery stenosis: Grading with image- directed Doppler US evaluation of renal resistive index. Radiology

Tupler R, Kliewer M. Early systolic compliance peaWrefiectiie-wave complex (Letter). Radiology 1993; 188: 286-7. Hansen KJ, Tribble RW et a/. Renal duplex sonography: Evalu- ation of clinical utility. J. Vasc. Surg. 1990; 12: 227-36. Patriquin H, Lafortune M. Stenosis of the renal artery: Assessment of slowed systole in the downstream circulation with Doppler sonography. Radiology 1992; 184: 479-85. Kotval PS. Doppler waveform parvus and tardus: A sign of proximal flow obstruction. J. Ultrasound Med. 1989; 8: 435-40. Kliewer M, Tupler R. Renal artery stenosis: Analysis of Doppler wave-form parameters and tardus pawus pattern. Radiology 1993; 189: 779-87.

162: 853-7.

1988; 14: 1-5.

1994; 190: 785-90.

15. Bude R, Rubin J. Pulus tardus: Its cause and potential limitations in detection of arterial stenosis. Radiology 1994; 190: 779-84.

Documents

Intra-renal Doppler wave-form analysis as a screening test for renal artery stenosis