Our review addresses problems and unresolved issues that have developed as a result of new diagnostic and therapeutic technologies, which follow from data on the prevalence and natural history of atherosclerotic renovascular disease, the effects of ACE inhibitors, and the outcome of available therapies. We do not provide answers to all of the questions raised because the available data are fragmentary and come mainly from retrospective analyses in selected populations.

Diagnosis of Renovascular Disease

Detection of renovascular disease depends on demonstrating the presence of stenotic lesions in the renal arteries and, ideally, proving that these lesions are responsible for impaired renal function. The benchmark diagnostic procedure is arteriography performed with the intra-arterial injection of iodinated contrast media. Papers cited in this review use arteriography to define the presence of renal artery stenosis. Severity is generally graded in quartiles, with 0% to 25% stenosis being defined as minimal; 26% to 50% as mild; 51% to 75% as moderate; and 76% to 100% as severe. Progression is defined, in most instances, as moving from a less severe to a more severe quartile. No completely satisfactory test exists to determine the functional significance of an anatomic lesion.

Because detection of renal artery stenosis requires arteriography, the risk of this procedure is a significant issue in deciding whether to pursue this diagnosis in patients with renal disease. A transient reduction in renal function is common following iodinated contrast administration when renal insufficiency is present [9]. The incidence of acute renal failure increases as renal function declines, averaging perhaps 10% in nondiabetic persons with mild to moderate pre-existing renal disease, and as high as 90% in diabetic patients with severe renal insufficiency [9]. Although institution of chronic dialysis therapy after arteriography has been reported, premature precipitation of end-stage renal disease (ESRD) appears to be uncommon. The use of nonionic iodinated compounds or CO₂ as contrast media may reduce or eliminate contrast nephropathy [10, 11], but the small and as yet undefined risk for atheroembolic disease and other consequences of catheter manipulation will remain.

The usefulness of alternative tests for detecting renal artery stenosis in patients with hypertension has recently been reviewed [12, 13], but it cannot be assumed that these data are applicable to the diagnosis of renal artery stenosis as a cause of renal insufficiency. Duplex Doppler ultrasonography [14] and magnetic resonance imaging [15] appear to be effective noninvasive tests to demonstrate stenoses in large proximal renal vessels. Both tests are operator dependent and their utility with more widespread use is yet to be determined. Functional tests that examine the renal handling of isotopically labeled drugs before and after ACE inhibition appear to decrease in accuracy as renal function declines [16]. Comparison of the cortical accumulation of isotope during furosemide washout of agents such as I-131-orthoiodohippurate [17] or Tc-99m-mercaptoacetyltriglycine [18], with and without ACE inhibition, may be useful, but these newer tests have not been extensively evaluated.

Prevalence of Renovascular Disease

Many textbooks state that 5% of persons with hypertension have renovascular disease [19]. If this prevalence figure is used, and 20% to 30% of the population of the United States has hypertension, then 2 to 4 million people are predicted to have renovascular disease. Assessment of the true prevalence, of course, would require arteriographic screening of a representative cross-section of the population and is unlikely to be ascertained. In attempts to quantify the prevalence, several approaches have been used. All such studies have been carried out in relatively high-risk subgroups, and all demonstrate a prevalence much greater than 5%. In addition, they show that renovascular disease is not confined to persons with hypertension.

One approach to assess the prevalence of renal artery stenosis is to review unselected autopsies. In one report, all autopsies performed during a 10-month period were checked to determine whether renal artery stenosis was present [20]. The cases were divided into those with diastolic blood pressure greater than 100 mm Hg and those with diastolic pressures of 100 mm Hg or less. Of 221 patients older than 50 years, 85% were in the latter group. In this group, 22.5% were found to have at least one renal artery that was more than 50% stenosed. In the group with diastolic blood pressure greater than 100 mm Hg who were older than 50 years, 53% had a renal artery stenosis of more than 50%. The overall prevalence of stenosis more than 50% in these two groups was 27%. The patients in this report were clearly selected because death was required for examination.

A second approach is to review unselected arteriograms. This approach is, of course, clearly biased toward patients with renovascular disease. In an early report, 500 consecutive aortograms were reviewed [21]. Although 91 of these were done specifically to examine the renal vessels, most were done to evaluate disease of the aorta or iliofemoral system. This report also divided patients into hypertensive and nonhypertensive groups, using a diastolic pressure of more than 90 mm Hg as the dividing line. There were many different descriptions of vascular disease, but almost all had some variant of atherosclerotic renovascular disease. In the normotensive group of 301 persons, 32% had some form of renovascular disease. In the hypertensive group, 62% were found to have renovascular disease. The degree of stenosis was not generally specified. The range in age was from 31 to more than 80 years; 80% of the patients were older than 50 years old.

In a more recent study, 79% of 1651 consecutive patients undergoing routine cardiac catheterization also had a technically adequate single-plane renal angiogram to screen for renovascular disease [22]. In this group of patients, 30% were found to have some degree of renal artery stenosis. Of these, one half (15% of all patients) had stenosis of more than 50%. As in the studies cited earlier, a large fraction of these patients (47%) were not hypertensive. A notable feature in the patients with significant renal artery stenosis was the high incidence of generalized vascular disease. Only 15% had no evidence of concomitant coronary artery or peripheral vascular disease.

From studies such as these, it is apparent that renal artery stenosis is much more common than 5% in persons older than 50 years with generalized vascular disease and that hypertension is not a particularly sensitive indicator of this process. The prevalence in the general population, however, remains unknown.

Progression of Renal Atherosclerosis

An understanding of the natural history of atherosclerotic renovascular disease is central to the approach to diagnosis and management. Whether renal artery stenosis is a progressive disease and, if it is progressive, how frequently and how rapidly it worsens, are critical questions. The first of these questions has been answered by studies in which serial angiograms have been done in patients with renovascular disease. In an early review in a small group of patients, Dustan and coworkers [23] provided the first clear evidence of the progressive nature of this disease. Combining urographic and angiographic observations in 18 patients with atherosclerotic renovascular disease, they found that 61% progressed over 6 years. Wollenweber [24] and colleagues from the Mayo Clinic reported a larger series in 1968, all followed by angiography. In addition to giving an estimate of the natural history of renal artery stenosis, this report again documents the high prevalence of generalized atherosclerotic vascular disease in patients with renovascular disease. At the time of diagnosis, 33% of the patients also had underlying heart disease, 11% had cerebrovascular disease, and 32% had peripheral vascular disease. Of the patients without coronary artery disease, signs or symptoms of heart involvement within 5 years developed in 47%. The presence of vascular disease, which threatens several organs, is a constant in reports of renovascular disease and complicates diagnostic and therapeutic considerations.

The second question, concerning the frequency of progression, is answered much less clearly. Table 1 summarizes the results in five reports in which serial angiograms have been done in 237 patients with renovascular disease. These studies include four retrospective analyses [24-27] and one prospective randomized study (patients randomized to medical treatment) [28]. Because the patients in these studies often had repeated arteriography because of signs and symptoms of worsening vascular disease, the results should be interpreted with caution. Nonetheless, all reports showed progression of disease. The incidence of progression, however, varied widely. An average of 49% of patients had progression of disease with a range from 29% to 71%. Part of the variation relates to the reasons for follow-up angiogram, which vary from study to study. It is noteworthy that the lowest frequency of progression is found in the one prospective study [28]. This report includes four patients in whom it was thought that the artery responsible for hypertension became occluded, three in whom a stenosis in the contralateral artery progressed, and three in whom a previously normal contralateral artery developed significant stenosis. In the four studies that noted whether the stenosis progressed to complete occlusion, the average prevalence of occlusion was 14%. Where specifically noted, control of hypertension appeared to have little effect on progression [26, 28], and the serum creatinine level did not accurately reflect progressive anatomic disease [26-28]. All studies preceded the era of aggressive control of lipids.

Investigators for two of the five reports attempted to address the third question and quantitate the rate of progression of renovascular disease [26, 27]. Unfortunately, neither study was prospective or studied unselected patients. Therefore the cautions regarding selection bias apply to these attempts to determine a progression rate. The report by Schreiber and colleagues [26] showed approximately 1.5% additional stenosis for each month of follow-up, and the report by Tollefson and associates [27] showed a rate of 0.38%. The variability around the mean rates was large in both studies. In the one prospective study [28], the investigators stated that it was not possible to quantify the rate of progression because the variability was so great. Some patients remained stable for long periods, whereas others progressed rapidly. They could not identify the specific factors that influenced progression in their small patient group.

Progression to End-Stage Renal Disease

Although the precise risk and frequency are not well defined, it is clear that renovascular disease can result in ESRD. This conclusion is not surprising given the high prevalence of bilateral renal artery involvement Table 2 and the known progressive nature of the disease in some patients (Table 1). Because of the uncertainties involved in estimating the rate of progression, however, it is difficult to predict the frequency with which renal failure develops in this population. Nonetheless, given the high prevalence of this disease in the growing number of people older than 50 years, it is likely that renovascular disease causes a clinically important fraction of ESRD. A major problem is establishing that renovascular disease is the cause among patients with ESRD. Many patients who are classified as having progressive hypertensive nephrosclerosis may well have had progressive renal artery stenosis or occlusion. Unless refractory hypertension develops in the course of this process, patients are unlikely to undergo diagnostic testing for renovascular disease. The sixfold increase in ESRD attributed to hypertension in the last decade [29], a time when new antihypertensive agents dramatically improved the treatment of high blood pressure, raises the question of whether many of these patients might have had progressive renovascular disease. Mailloux and coworkers [7] tried retrospectively to identify these individuals among their patients with ESRD. They implicated renovascular disease as the cause of renal failure in patients with clinical evidence of systemic atherosclerotic vascular disease who either had a positive renal angiogram at some time in the past or in whom a radionuclide angiogram suggested the diagnosis in the absence of heavy proteinuria or abnormal urinary sediment. Using these criteria, they found that 16.5% of the patients most recently entering their program (1982 to 1985) had renovascular disease as the primary cause of their renal failure. Strikingly, this percentage increased from 7% to 8% in the patients entering their program earlier (1970 to 1982). Not surprisingly, more than 90% of these patients were older than 50 years, and the increased incidence of renovascular disease paralleled the increase in the average age of their ESRD population. Scoble and colleagues [30] did a similar analysis in their ESRD program during an 18-month period in England and found that 6% of all patients entering their program and 14% older than 50 years had clinically significant renovascular disease that contributed to their renal failure.

Both reports are limited because they are retrospective and include predominantly white participants. It is difficult to be certain whether renovascular disease was incidental or actually caused ESRD and whether a similar prevalence would be found in a more representative cross-section of the U.S. population. In 1989, 29 019 patients older than 50 years started dialysis in the United States [31]. If we use a prevalence estimate of 15% based on the data cited above, then approximately 4400 patients would have progressed to ESRD that year with a coexisting diagnosis of atherosclerotic renovascular disease. If one estimates that renovascular disease causes ESRD in 50% of these patients, then this disease is responsible for more than 2000 cases of ESRD, equivalent to 5% of patients beginning dialysis each year. At an average cost of $37 000/y for medical care for each of these patients [32], this estimate suggests the total cost of this disease could be $75 million per year. If a large portion of the ESRD caused by this disease is treatable by early intervention, a significant reduction in morbidity and medical care cost is possible.

Angiotensin-converting Enzyme Inhibitors and Renovascular Disease

To the extent that renovascular hypertension is mediated by high circulating levels of angiotensin II, ACE inhibitors are effective and specific antihypertensive agents. At the same time, concerns have arisen about the use of these drugs in the presence of renal artery stenosis. These stem from well-described changes in glomerular hemodynamics induced by these agents when glomerular perfusion pressure is low because of renal artery stenosis [33]. Angiotensin-converting enzyme inhibitors cause a profound decrease in glomerular capillary pressure and, when perfusion pressure is already low, this reduction in pressure can lead to a complete cessation of glomerular filtration.

The possible effects of these agents on renal function in three clinical settings are illustrated in Figure 1. In the first setting, the entire renal mass is at ischemic risk. This occurs when there is clinically significant bilateral renal artery stenosis or when there is stenosis of the artery to a solitary kidney. The administration of ACE inhibitors in this setting is well known to cause renal failure. In the second situation, one kidney has renal artery stenosis and the contralateral kidney has parenchymal disease such as hypertensive nephrosclerosis. Use of ACE inhibition in this setting decreases the glomerular filtration rate in the ischemic kidney, but the filtration rate in the contralateral diseased kidney remains stable. Serum creatinine would be predicted to increase in this situation and to stabilize at some new elevated level. The increase in creatinine levels in situations 1 and 2 is a diagnostic clue to the presence of renovascular disease.

View larger version (20K): [in this window] [in a new window]   Figure 1. Potential effect of angiotensin-converting enzyme inhibition on glomerular filtration rate (GFR) and serum creatinine concentration (SCR) in patients with hemodynamically significant renal artery stenosis. (Left) The effect in the setting of bilateral stenosis. (Middle) The effect in the setting of unilateral stenosis with contralateral parenchymal disease (gray area). (Right) The effect in the setting of unilateral stenosis with a normal contralateral kidney.

The third condition depicted in Figure 1 is the presence of an ischemic kidney with a normal contralateral kidney. When a converting enzyme inhibitor is given in this context, renal function in the ischemic kidney is again markedly reduced, but glomerular filtration rate may increase in the normal kidney so that a clinically obvious change in the serum creatinine level does not occur. As a result, renal artery stenosis may be missed unless a functional study is done that uncovers the asymmetry. This third situation raises particular concern about the use of ACE inhibitors in patients with unsuspected renovascular disease. Several studies have compared the administration of ACE inhibitors with other antihypertensive agents in animal models of unilateral renal artery stenosis [34-36]. These studies show that treatment with an ACE inhibitor controlled hypertension better and slowed the progression of nephrosclerosis in the kidney without surgical renal artery constriction (unclipped). However, in this treatment group, severe fibrosis developed in the kidney on the side of the clipped renal artery. These observations suggest that ACE inhibitors induce a specific form of damage to kidneys with hemodynamically significant renal artery stenosis.

The suggestion that ACE inhibition is detrimental in renovascular disease can be countered by the observations suggesting that inhibition of angiotensin II effects may be necessary for the best treatment of this condition. Angiotensin II is not only a potent vasoconstrictor but also appears to stimulate cell hypertrophy and proliferation [37, 38]. Thus, high levels of angiotensin II could contribute to vascular and ventricular hypertrophy [39], proliferation associated with atherosclerosis [40], and progressive glomerular sclerosis [39] independent of the associated hemodynamic effects. It could be argued, therefore, that ACE inhibition may have important therapeutic benefits in patients with renovascular disease in whom these problems are accentuated, independent of the effect on blood pressure.

A clinical dilemma thus exists in which inhibition of the effects of angiotensin II is important for optimal therapy of renovascular disease, yet pharmacologic interruption of angiotensin II production risks loss of ischemic renal mass. If these observations are correct, it follows logically that detection and repair of anatomic abnormalities is more important than previously suspected.

"Conservative" versus "Aggressive" Therapy

Four general therapeutic approaches are available in patients who have renovascular disease. The first, not often considered, is to offer medical therapy to control hypertension but to discourage dialysis if and when ESRD develops. The second approach is to use medical therapy and to provide dialytic support if ESRD develops. The last two approaches involve attempts to correct the stenosis, either through percutaneous transluminal angioplasty or by surgical revascularization.

In deciding whether aggressive investigation and therapy are appropriate, it is important to understand the natural history of the patient with atherosclerotic renovascular disease after reaching ESRD and beginning dialysis therapy. If one considers those patients reaching ESRD with a diagnosis of hypertension as the cause of renal failure, the age-adjusted mortality rate is 30% in 2 years and 60% in 5 years [41]. By age 65, the mortality rate for patients with ESRD as a result of hypertension is 20% to 30% a year. Because patients with atherosclerotic renovascular disease have a high incidence of cardiovascular disease, mortality rates in this group of patients with hypertension are likely to be even higher. This likelihood is supported by the report of Mailloux and coworkers [7]. The patients they identified as having renovascular disease had a median survival period of only 27 months and a 5-year survival rate of less than 10%.

The coexistence of systemic atherosclerotic vascular disease in patients with ESRD secondary to renovascular disease is clearly the major cause of the high mortality rate in these patients. Whether the dialysis treatment, per se, with its associated stresses of infection and access surgery, is a contributor remains controversial. Only two reports present data that may be relevant to this issue [42, 43]. Both are retrospective analyses that compare long-term survival in patients who had improvement in renal function after repair of renal artery stenosis with patients who did not. Both show dramatically improved survival rates in the former group, none of whom required dialysis during the period of follow-up. These two reports include small numbers of patients, and neither gives information about whether the two groups are really comparable, that is, had similar comorbidity. Nonetheless, they raise the possibility that successful repair of renal artery stenosis can not only keep patients from requiring dialysis but also improve long-term survival. This important question requires further study.

Outcome of Revascularization Procedures

Several reports have documented that either revascularization surgery or percutaneous angioplasty can restore renal function lost due to renal artery stenosis. Still impressive is the early report of Morris and colleagues [44], in which renal function was recovered in eight patients surgically revascularized at a time before the general availability of dialysis. Comparison of renal revascularization procedures is difficult because of differences in patient selection, technique, and in the way outcomes are reported. Our review yielded eight surgical reports and six angioplasty reports in which sufficient data are available to determine whether renal function improved, stabilized, or worsened in patients with renovascular disease and impaired renal function. All are uncontrolled retrospective analyses. In these 14 reports, 9 defined improvement or deterioration of renal function in fairly specific terms. In general, these definitions approximated a 20% change in serum creatinine. Three of the remaining reports simply categorized patients as improved, stable, or worsened. The final two reports, both presenting results of angioplasty, designated renal function into categories of improved/stable or worse or as improved or stable/worse.

Table 3 summarizes the results of seven of the eight surgical reports. The report by Hallett and associates [52] was excluded from our final analysis because in these 98 patients the results differed markedly from the remaining reports. The surgical approach also differed notably because more than 50% of the cases involved some sort of aortic replacement in addition to repair of renal artery stenosis. The mortality rate was similar in the Hallett report (8%), but only 22% of patients showed improvement in renal function after operation (53% had stable renal function and 25% had worse renal function). Novick and coworkers [6] have stressed that entering the aorta in patients with diffuse aortic atheromatous disease significantly worsens outcome. In many cases aortic replacement is required for processes other than renovascular disease, but this procedure should not be considered a primary approach to renal revascularization. Excluding Hallett's observations, the combined results of the surgical reports show that 55% of patients had improvement of renal function (Table 3). Although the first two show mortality rates of 17%, comparable to those reported in the cooperative study of renovascular hypertension [1] that left many disenchanted with surgical renovascularization, the remaining five show much lower rates. The combined mortality rate for all seven reports is 6%. Results of percutaneous balloon angioplasty are summarized in Table 4. In this group, improvement of renal function occurred in 41% to 43% of patients, with a similar mortality rate (5% to 6%).

The results of surgery and angioplasty are compared in Table 5. From this comparison, it is difficult to choose between the two modalities. When the two angioplasty studies that did not categorize outcomes into three groups are excluded, no significant differences are noted between the two approaches. However, when either of these two are added, the surgical outcome is superior (P < 0.05 for comparison with Martin's data included, and P < 0.01 for comparison with Canzanello's data included). Nonetheless, caution should be exercised in this interpretation because of undefined differences inpatient selection. The average mortality rate is similar for the two approaches, but the mortality data are incomplete, and the way in which death is defined (for example, only deaths due to complications of the procedure versus all deaths) differs among the studies.

The longevity of these repair procedures is also difficult to compare. Novick and coworkers' [47] report of 241 patients, 66% of whom were operated on for preservation of renal function, suggests excellent long-term results. During a mean follow-up period of 38.8 months, thrombosis or re-stenosis occurred in 4.3% of the 254 operated arteries, and the mortality rate was 11.2%. Unfortunately, long-term follow-up of renal function is not reported. The studies reporting the results of balloon angioplasty for preservation of renal function have follow-up periods of 2 years or less. Pickering and coworkers [43] report that renal function remained stable for 2 years in all patients whose function improved initially after angioplasty, whereas Canzanello and coworkers [55] report that renal function declined during an average of 22 months in 48% of 69 patients who had a successful anatomic result with angioplasty. A prospective, randomized trial in comparable patients is needed to answer the question of which approach offers the best outcomes.

Unresolved Problems and Future Directions

Several features of atherosclerotic renovascular disease are well documented. This entity is common, particularly as persons reach 50 years of age or more and especially in those patients with other evidence of atherosclerotic vascular disease. The disease clearly progresses in some patients, often to involve the contralateral renal artery. However, progression rate appears to be highly variable, and therefore it is difficult to predict what will happen in a given case. The disease can ultimately produce occlusion of both renal arteries and hence cause end-stage renal failure. Given the apparent high prevalence of renovascular disease in persons older than 50 years, it is possible that this disorder is the cause of ESRD in 5% to 15% of patients currently entering the dialysis population each year.

Surgery and percutaneous angioplasty can both improve renal function in patients with renal artery stenosis. In experienced hands, surgery appears to be more successful at preserving renal function, but the procedures have not been compared directly in a prospective, randomized study. On the basis of observations to date, the outcome for either surgery or angioplasty is acceptable, given the high mortality rate associated with dialysis treatment in patients with diffuse vascular disease. Thus, it seems reasonable to consider diagnostic and therapeutic intervention in patients with evidence of systemic atherosclerotic vascular disease and otherwise unexplained progression of renal insufficiency. The morbidity and mortality risks associated with these interventions, particularly the risk for atheroembolic disease, are poorly defined and obviously must be weighed in this decision. In the patient who is losing renal function rapidly, however, we believe the balance is often in favor of intervention because of the high mortality risk and poor quality of life that often ensue when ESRD occurs. In all cases, even if comorbidity precludes diagnosis and repair of renal artery stenosis, a straightforward presentation of the difficulties associated with hemodialysis is appropriate. This allows patients to appraise the value of proposed therapies more realistically.

In the patient with diffuse atherosclerotic vascular disease and stable mild to moderate renal insufficiency, the data presently available offer little guidance on how to proceed. The uncertainties regarding the frequency and rate of progression and the risk of diagnostic studies render decision making difficult, if not impossible. A small prospective study is underway in such patients [57], which we hope will shed some light on these questions.

This review has emphasized that renovascular disease is an important public health issue in our aging population. We are at a point at which there will be continued growth in aggressive diagnostic procedures and interventions related to this process. These can develop in a piecemeal fashion, as was the case with therapy for coronary artery disease, until many therapies become widespread and entrenched in medical practice before their usefulness is understood. By contrast, organization of careful multicenter studies to examine the place of various diagnostic and therapeutic procedures in advanced atherosclerotic renovascular disease could provide significant insight and cost savings. Finally, therapies that can cause regression of atherosclerosis need to be tested early in the course of renovascular disease to develop the optimal treatment for this growing cause of renal failure.