British Journal of Radiology 74 (2001),213-218 © 2001 The British Institute of Radiology
Increasing the diagnostic yield of renal angiography for the diagnosis of atheromatous renovascular disease
A E Shurrab, MD, MRCP1,
H Mamtora, MBChB, FRCR2,
D J O'Donoghue, MBChB, FRCP1,
S Waldek, MBChB, FRCP1 and
P A Kalra, MD, FRCP1
1 Departments of Renal Medicine
2 Radiology, Hope Hospital, Stott Lane, Salford M6 8ES, UK
 |
Abstract
|
|---|
Atheromatous renovascular disease (ARVD) is a common cause of hypertension and chronic renal failure (CRF). In this unit, intravenous digital subtraction angiography (DSA) (or intraarterial DSA if indicated) is used as a screening angiographic study when ARVD is suspected. However, increased use of these investigations has resulted in a longer waiting time for angiography. As the majority of studies are negative for ARVD, clinical features and results of investigations of patients undergoing angiography were reviewed to identify those having the greatest likelihood of ARVD. The clinical notes were reviewed for all 249 patients undergoing angiography over an 18-month period. Primary indications for investigation were: hypertension 71 (28.5%), CRF 156 (62.7%) and CRF with severe hypertension 22 (8.8%). 12 of the CRF patients had end-stage renal failure. 166 (66.7%) patients had no evidence of ARVD, while only 83 (33.3%) patients showed some degree of ARVD, 29 (35%) of which had bilateral renal artery disease. There was no significant difference between the ARVD group and the non-ARVD group for mean age (69.0 years vs 63.3 years), male to female ratio, history of smoking (68.7% vs 55.4%), severe hypertension (10.8% vs 9.0%), hypercholesterolaemia (61.4% vs 47.0%), diabetes mellitus (28.6% vs 25.3%) or angiotensin converting enzyme inhibitor-related renal dysfunction (9.6% vs 6.1%). More patients in the ARVD group were investigated for CRF than in the non-ARVD group, as reflected by the higher serum creatinine level and the lower creatinine clearance in the ARVD group. 55 (33.1%) of the non-ARVD patients had no comorbid vascular disease, vascular bruits or ultrasound discrepancy in the size of the two kidneys, whereas all ARVD patients had at least one of these features (negative predictive value 100%). All three features were present in 19.3% of ARVD patients but in only 3.0% of the non-ARVD patients (positive predictive value 76.2%, specificity 97%). We plan to rationalize the criteria for angiography in the light of these findings, anticipating an increase in the diagnostic yield of renal angiography from its current 33.3% to above 42%.
|
Introduction
|
|---|
Atheromatous renovascular disease (ARVD) accounts for 2–5% of all cases of secondary hypertension [1] and is a common cause of both chronic renal failure (CRF) and end-stage renal failure (ESRF) [2–4]. More than 15% of all patients with ESRF have ARVD, and the prevalence approaches 25% in the elderly dialysis population [5, 6]. ARVD is frequently associated with extrarenal vascular disease, most patients having generalized arteriopathy [7–9]. ARVD may also contribute to the morbidity of many elderly patients with congestive cardiac failure [10]. As the condition is accompanied by a high mortality as well as a high likelihood of requiring dialysis, there is justification to detect ARVD at the earliest possible stage.
Several non-invasive screening techniques are currently available for the diagnosis of ARVD. In addition to definitive angiography, these include duplex ultrasound, isotope scintigraphy, spiral CT and magnetic resonance angiography (MRA). There has been much debate as to the most applicable screening test for clinical use. Renal artery duplex ultrasound is limited as it is time consuming and highly operator dependent [11, 12], whereas captopril renography, although sensitive for the detection of renal artery stenosis (RAS) in hypertensive patients with normal renal function, is far less accurate in patients with renal impairment [13, 14]. Spiral CT is limited by requiring large volumes of contrast medium, as well as by artefacts from calcification and a diminished field of view [15]. MRA is developing as perhaps the most accurate non-invasive investigation for ARVD in both hypertensive and CRF patients. It has the added potential of providing functional information rather than simply two-dimensional images of the proximal renal arteries [16–18]. Intravenous digital subtraction angiography (IV-DSA) is another screening technique and is the one favoured in our centre.
The increasing awareness of the need for early detection of ARVD has resulted in a substantial increase in the number of requests to our radiology department for renal angiography. As a consequence, the waiting list has lengthened for this particular investigation. Many of these studies prove negative for ARVD and, although referring clinicians have usually based their suspicion of ARVD upon clinical findings as wellas indicative biochemical and simple radiological features, it is felt that a number of patients might undergo unnecessary investigations. We reviewed the results of all renal angiograms performed at our institution during an 18-month period. The results were considered in the context of the patients' basic indicative clinical, biochemical and radiological features. It was hoped that a combination of features with a high positive or negative predictive value for the presence of ARVD might be defined, both to rationalize the requests for renal angiography and to improve the diagnostic yield from future studies.
|
Patients and methods
|
|---|
This retrospective study included all patients who had undergone renal angiography at this hospital as part of an investigation for hypertension or CRF in the 18-month period between January 1997 and August 1998.
Patient characteristics
The case notes of all patients were reviewed, and basic clinical features and investigations were recorded. The clinical details included age, sex, history of smoking, diabetes mellitus, hypercholesterolaemia (defined as serum cholesterol of >5.2 mmol l-1 or patient receiving lipid-lowering therapy), history of angiotensin converting enzyme inhibitor (ACE-I)-related renal dysfunction and history of severe hypertension (defined as blood pressure that is difficult to control with three or more antihypertensive medications). Evidence of comorbid vascular disease, with a history of ischaemic heart disease (IHD), peripheral vascular disease (PVD) or cerebrovascular disease (CVD), and/or the presence of vascular bruits (renal, aortic or femoral) was also noted. The investigations considered included serum creatinine level, creatinine clearance and a discrepancy in the bipolar length of the two kidneys on ultrasound (a significant discrepancy was defined as a difference of >1.5 cm).
Angiographic protocol
IV-DSA has been long established in our centre as a screening test for ARVD. IV-DSA has some limitations, which include poor image resolution in the obese or very elderly as well as in patients with poor cardiac output, excessive bowel gas or overlying blood vessels. These factors result in technically uninterpretable results in about 7% of cases [19]. Intraarterial digital subtraction angiography (IA-DSA) is used in these circumstances. IA-DSA requires much smaller volumes of contrast medium compared with conventional intraarterial angiography [20]. This policy was adopted because IV-DSA is cheaper, has a lower complication rate and is less invasive than IA-DSA, and it is available on an outpatient basis.
IV-DSA was performed by injection of 35 ml iopamidol-370 with 25 ml normal saline as a chaser through a 16 G cannula into an anticubital fossa vein. The injection pressure was 450 lb sq.in-1 at a rate of 20 ml s-1 and the acquisition frame rate was 1 per s. IA-DSA was performed through the transfemoral route using the Seldinger technique. A 3 F gauge pigtail catheter was positioned immediately above the origin of the renal arteries at the L1 level. 35–45 ml iopamidol-300 was then injected at 650 lb sq.in-1 at 15 ml s-1. Left anterior oblique and anteroposterior acquisitions were obtained as well as occasionally a right anterior oblique view. A patient was considered to have positive evidence of ARVD if the angiogram showed any detectable RAS or renal artery occlusion (RAO). For the purpose of further analysis, all degrees of RAS severity were included in the ARVD group. This was because the natural history of ARVD is known to be that of a progressive disease [21, 22].
Statistical analysis
Patients were divided into two groups based upon the presence or absence of ARVD. Their clinical features and results of investigations were compared in the context of these two groups. Mean, standard deviation (SD) and median values were calculated where appropriate. Variables were analysed by two-way analysis of variance (ANOVA). When the ANOVA indicated significant differences between the groups for a particular parameter, differences in categorical data were analysed with 
2 tests. For continuous data, means were compared by t-tests. A p-value of less than 0.05 was considered a significant difference. Positive or negative predictive values for the presence of ARVD were calculated for particular parameters, either alone or in combination.
|
Results
|
|---|
During the 18-month period, 249 patients underwent renal angiography (52 IA-DSA, 197 IV-DSA); case notes were obtained in all patients. All patients were suspected of having ARVD. They were subdivided into 71 (28.5%) patients primarily with hypertension, 156 (62.7%) patients with CRF and 22 (8.8%) patients with CRF and severe hypertension. 12 of the patients in the CRF category had ESRF. 83 (33.3%) patients had evidence of ARVD (Table 1
), whereas angiography was considered negative in the remaining 166 (66.7%) patients. None of the patients reviewed had evidence of fibromuscular dysplasia.
Table 2 shows the basic demography and some of the features observed in the investigated patients. The two groups were comparable for age and male to female ratio, but patients with ARVD were more likely to have been investigated because of CRF rather than for isolated hypertension. Consequently, patients with ARVD had a significantly greater mean serum creatinine level (282±216 µmol l-1 vs 192±144 µmol l-1; p<0.01) and a lower creatinine clearance (41.6±27.4 ml min-1 vs 63.2±39.4 ml min-1; p<0.01). There were no differences between the two groups in regards to smoking history, hypercholesterolaemia, diabetes mellitus or severe hypertension. 18 patients had ACE-I-related renal dysfunction; 10 of these patients had no evidence of ARVD and all of these patients had evidence of an intercurrent illness with some element of intravascular volume depletion. Ultrasound discrepancy in the bipolar renal length was present in 34 (41%) patients with ARVD and 24 (14.5%) patients in the non-ARVD group (p<0.001). The sensitivity of this investigation for the presence of ARVD was low (60%) due to the presence of bilateral RAS in 29 (35%) patients in the ARVD group (18 of these 29 patients had no difference in the size of the two kidneys). Nevertheless, the absence of ultrasound discrepancy had a useful negative predictive value of 85% [23].
The presence of comorbid vascular disease is shown in Figure 1. Patients in the ARVD group were significantly more likely to have IHD and PVD than patients in the non-ARVD group. Evidence of comorbid vascular disease was present in 83.1% of the ARVD patients compared with only 56.0% in the non-ARVD group (p<0.01). The positive predictive value of a history of extrarenal vascular disease was 83% and its specificity was 84% for the prediction of ARVD. Patients with femoral bruits often gave no history of PVD and the proportion of patients with vascular bruits is therefore displayed separately in Figure 2. Patients with ARVD were significantly more likely to have renal/aortic, femoral or any of these types of bruit. This feature had a specificity of 77%, a positive predictive value of 57% and a negative predictive value of 73% in the diagnosis of ARVD.
View larger version (12K):
[in this window]
[in a new window]
|
Figure 1. Comparison of atheromatous renovascular disease (ARVD) and non-ARVD groups for the presence of comorbid vascular disease. IHD, ischaemic heart disease; PVD, peripheral vascular disease; CVA/TIA, cerebrovascular accident/transient ischaemic attack; comorbid VD, comorbid vascular disease.
|
|
View larger version (11K):
[in this window]
[in a new window]
|
Figure 2. Comparison of atheromatous renovascular disease (ARVD) and non-ARVD groups for the presence of vascular bruit.
|
|
Table 3 shows an analysis of whether patients in either group had evidence of any of the three key discriminatory features, namely ultrasound bipolar length discrepancy, history of comorbid vascular disease and vascular bruits. Figure 3 clearly shows that the absence of all three parameters conferred a negative predictive value of 100% for the diagnosis of ARVD at renal angiography. Furthermore, all three features were present in 16 (19.3%) patients in the ARVD group but in only 6 (3.6%) patients in the non-ARVD group, indicating a positive predictive value of 76.2%, with specificity as great as 97%.
View this table:
[in this window]
[in a new window]
|
Table 3. Patients with atheromatous renovascular disease (ARVD) and non-ARVD patients divided into subgroups according to the presence of the following features: ultrasound discrepancy in renal size, comorbid vascular disease and vascular bruit
|
|
|
Discussion
|
|---|
ARVD is very common [24] and many affected patients, especially those who present with advanced renal failure and/or bilateral severe ARVD [9], have a poor prognosis. It is important to diagnose ARVD at the earliest possible stage so that optimal medical treatment and, when indicated, renal revascularization can be initiated with a greater chance of improving patient outcome. Requests for renal angiography escalated three-fold in our centre during the last decade, but resources are limited and a significant waiting list has developed. This prompted us to undertake the current study in the hope of identifying key features that would enable more appropriate requesting of renal angiograms, with a higher positive diagnostic yield for ARVD.
The features that were significantly more often present in patients with ARVD were a history of coexistent extrarenal vascular disease, the presence of vascular bruits and a significant disparity in bipolar renal length measured on ultrasound. These findings are not unexpected and are supported by the literature. Several studies have documented the frequent coexistence of ARVD and other atherosclerotic disease, especially PVD and IHD. In one series of 100 patients with PVD, bilateral ARVD was present in 24%, 7 of whom had unilateral RAO [8]. Other groups have found that ARVD frequently coexists with IHD. In a large series of over 1300 patients undergoing angiographic investigation for angina, concomitant abdominal aortography showed that 4% and 11% of patients had significant bilateral and unilateral RAS, respectively, whereas an additional 15% had insignificant RAS [7]. As in our study, the presence of femoral and aortic/renal bruits (the latter being considered together as they are often difficult to distinguish clinically) has previously been found to be useful in screening patients for ARVD [25].
Two important outcomes arose from the analysis of whether all or none of the above three features were present in patients with or without ARVD. The presence of all three features conferred a very high specificity (97%) for the detection of ARVD, but the positive predictive value of this combination was weakened (76.2%) because bilateral ARVD with accompanying lack of disparity in renal size was common. However, the most clinically applicable and potentially useful finding was that the absence of all three features conferred a negative predictive value of 100%. We recommend that, with the exception of those few patients with none of these features but who present with recurrent flash pulmonary oedema [26] or CRF with severe hypertension, such patients should not undergo investigation for ARVD. This policy could be implemented without the risk of failing to diagnose patients with ARVD. If this policy had been applied to our study population, renal angiography would have been avoided in 22% of all patients investigated and 33% of all those with negative angiography, with a potential saving of resources.
These findings have allowed the development of a simple patient screening protocol that should rationalize the requesting of renal angiograms. Once implemented, we anticipate that a major increase in the diagnostic yield of ARVD from renal angiography should result in our centre, and it is intended to re-evaluate this outcome in the future.
Received for publication August 17, 2000.
Revision received November 21, 2000.
Accepted for publication November 27, 2000.
|
References
|
|---|
-
Rudnick KV, Sackell DL, Hirst S, Holmes C. Hypertension in family practice. Can Med Assoc J 1997;117:492–7.
-
Preston RA, Epstein M. Ischemic renal disease: an emerging cause of chronic renal failure and end stage renal disease. Am J Hypertens 1997;15:1365–77.
-
Mailloux LU, Napolitano B, Belluci AG, Vernace M, Wilkes BM, Mossy RT. Renal vascular disease causing end-stage renal disease, incidence, clinical correlates and outcome: a 20 year experience. Am J Kidney Dis 1994;24:622–9.[Medline]
-
Rimmer JM, Gennari FJ. Atherosclerotic renovascular disease and progressive renal failure. Ann Intern Med 1993;118:712–20.[Abstract/Free Full Text]
-
O'Neil EA, Hansen KJ, Canzanello VJ, Pennell TC, Dean RH. Prevalence of ischemic nephropathy in patients with renal insufficiency. Am J Surg 1992;58:485–90.
-
Herrara AH, Davidson RA. Renovascular disease in older adults. Clin Geriatr Med 1998;14:237–53.[Medline]
-
Harding MB, Smith LR, Himmelstein SI, Harrison K, Philips HR, Schward SJ, et al. Renal artery stenosis: prevalence and associated risk factors in patients undergoing routine cardiac catheterization. J Am Soc Nephrol 1992;2:1608–16.[Abstract]
-
Choudhri AH, Cleland JG, Rowlands PL, Tran TL, McCarthy M, Al-Kutoubi MA. Unsuspected renal artery stenosis in peripheral vascular disease. BMJ 1990;301:1197–201.
-
United States Renal Data System, USRDS 1992. Annual data report. The National Institute of Health, National Institute of Diabetes, Digestive and Kidney Diseases. Bethesda, MD: USRDS, 1992.
-
MacDowall P, Kalra PA, O'Donoghue DJ, Mamtora H, Waldek S, Brown K. Risk of morbidity from renovascular disease in elderly patients with congestive cardiac failure. Lancet 1998;352:13–6.[Medline]
-
Robertson R, Murphy A, Dubbins PA. Renal artery stenosis: the use of duplex ultrasound as a screening technique. Br J Radiol 1988;61:196–201.[Abstract/Free Full Text]
-
Reihl J, Schmitt H, Bongart D, Bergmann D, Sieberth HG. Renal artery stenosis: evaluation with colour duplex ultrasonography. Nephrol Dial Transplant 1997;12:1608–14.[Abstract/Free Full Text]
-
Muller FB, Sealy JE, Case DB, Atlas SA, Pickering TG, Pecker MS, et al. The Captopril test for identifying renovascular disease in hypertensive patients. Am J Med 1986;80:633–44.[Medline]
-
Mann SJ, Pickering TG, Sos TA, Uzzo RG, Sarker S, Friend K, et al. Captopril renography in the diagnosis of renal artery stenosis: accuracy and limitations. Am J Med 1991;90:30–40.[Medline]
-
Rubin GD, Dake MD, Napel S. Spiral CT for renal artery stenosis: comparison of the three-dimensional rendering technique. Radiology 1994;190:181–9.[Abstract/Free Full Text]
-
Ghantous VE, Eisen TD, Sherman AH, Finkelstein FO. Evaluating patients with chronic renal failure for renal artery stenosis with gadolinium-enhanced magnetic resonance angiography. Am J Kidney Dis 1999;33:36–42.[Medline]
-
Hertz SM, Holland GA, Bauin RA, Haskal ZJ, Carpenter JP. Evaluating renal artery stenosis using magnetic resonance angiography. Am J Surg 1994;168:140–3.[Medline]
-
Miller S, Schick F, Duda SH, Nagel T, Hahn U, Teufl F, et al. Gd-enhanced 3D phase-contrast MR angiography and dynamic perfusion imaging in renal artery stenosis. Magn Reson Imaging 1998;16:1005–12.[Medline]
-
Albert FJ, Svetkey LP. Screening and diagnosis in renovascular hypertension. In: Novick A, Scoble J, Hamilton G, editors. Renal vascular disease. London: WB Saunders, 1996:216–7.
-
Kim D, Potter DH, Brown R. Renal artery imaging: a prospective comparison of intra-arterial digital subtraction angiography with conventional angiography. Angiography 1991;42:345–57.
-
Zierler RE, Bergelin RO, Davidson RC, Cantwell-Gab K, Polissar NL, Standness DE. Jr. A prospective study of disease progression in patients with atherosclerotic renal artery stenosis. Am J Hypertens 1996;9:1055–61.[Medline]
-
Crowley JJ, Conlon PJ, Santos RM, Kruse KR, Athirakul K, Philips HR, et al. Incidence andprogression of renal artery stenosis in patients undergoing routine cardiac catheterization. Circulation 1996;94:S299A
-
Guzman RP, Zieler RE, Isaacscon JA, Bergelin RO, Srandness DE Jr. Renal atrophy and arterial stenosis: a prospective study with duplex ultrasound. Hypertension 1994;23:346–50.[Abstract/Free Full Text]
-
Conlon PJ, O'Riordan E, Kalra PA. New insights into the epidemiologic and clinical manifestations of atherosclerotic renovascular disease. Am J Kidney Dis 2000;35:573–87.[Medline]
-
Svetky LP, Helms MJ, Dunnick NR, Klotman PE. Clinical characteristics useful in screening for renovascular disease. South Med J 1990;83:743–7.[Medline]
-
Bloch MJ, Trost TW, Pickering TG, Sos TA, August P. Prevention of recurrent pulmonary oedema in patients with bilateral renovascular disease through renal artery stent placement. Am J Hypertens 1999;12:1–7.[Medline]
This article has been cited by other articles:
|
|
|
|
 
S. R. Orth
Smoking and the Kidney
J. Am. Soc. Nephrol.,
June 1, 2002;
13(6):
1663 - 1672.
[Full Text]
[PDF]
|
|
|
Top
Abstract
Introduction
