Progress in the treatment of atherosclerotic renovascular disease: the conceptual journey and the unanswered questions

Abstract

Over the past decades, management of atherosclerotic renovascular disease (ARVD) has undergone significant progress, in parallel with increased knowledge about the complex pathophysiology of this condition. Modern multi-targeted medical management of atherosclerosis has driven a change in both the natural history and the clinical outcomes of ARVD. Progression to total renal artery occlusion is a much less common occurrence and while early studies have reported that up to 41% of patients reached renal end-points over a mean follow-up of 44 months, the latest randomized controlled trials have shown that progressive renal impairment occurs in 16–22% of patients, with <8% of patients reaching end-stage kidney disease (ESKD) over a similar time-frame. However, the results of the latest large ARVD trials investigating the effect of renal stenting upon clinical outcomes have been influenced by selection bias as high-risk patients with clinically significant renal artery stenosis (RAS) have largely been excluded from these studies. Although the neutral results of these trials have shown uncertainty about the role of revascularization in the management of patients with ARVD, there is evidence that revascularization can optimize outcomes in selected patients with a high-risk clinical phenotype. Future challenges lie in identifying important subgroups of patients with critical RAS and viable kidneys, while continuing to develop strategies to protect the renal parenchyma and hence improve clinical outcomes.

INTRODUCTION

The management of atherosclerotic renovascular disease (ARVD) has undergone significant changes over the years, in parallel with increased understanding of this condition. Neutral results of recent large randomized controlled trials (RCTs) have dampened interest in revascularization and there is emerging evidence that contemporary management of atherosclerosis has improved both the natural history and clinical outcomes of ARVD. There is, however, ample evidence that revascularization still plays an important role in the management of patients with high-risk phenotypes who may not have been adequately represented in RCTs [1]. Accurate identification of these individuals remains a considerable challenge.

Here we critically review selected studies related to management of ARVD that have been performed over the past few decades and illustrate how the clinical outcomes of ARVD have evolved in tandem with significant progress in both the medical treatment of atherosclerosis and revascularization techniques. The progress is depicted with key historic milestones as shown in Table 1.

RENOVASCULAR HYPERTENSION: SURGERY AND EARLY ANTI-HYPERTENSIVE AGENTS (1960–70S)

Landmark animal studies performed by Goldblatt [2] in 1934 stimulated interest in the study of hypertension by showing that experimental renal artery stenosis (RAS) in dogs led to systemic hypertension and this could potentially be reversed by removing unilateral diseased kidneys [3]. Almost three decades later, pharmacological treatment of hypertension became available but this was fraught with side-effects and hence diagnosing potential surgically correctable forms of hypertension was a priority. Increased use of renal angiography, however, revealed that atherosclerosis often involved both kidneys and could be a multi-organ disease process. In 1968, Wollenweber et al. [4] showed that patients with ARVD had a poor prognosis, with a 5-year survival of 66.7% compared with 91.7% in a normal age-adjusted population. Despite development of complex surgical revascularization techniques to allow sparing of renal tissue, surgery was often hazardous and unsuccessful, especially in patients with long-standing hypertension. On the other hand, anti-hypertensives such as guanethidine, hydralazine and thiazide diuretics improved blood pressure control in up to 65% of patients who were not fit to undergo surgery [5] (Table 2). Blood pressure targets during this period were higher than current recommendations, a diastolic blood pressure of 100 mmHg or less being considered satisfactory, but it was remarkable to note that better blood pressure control was associated with improved survival in this cohort of patients, whereas revascularization generally had no overall effect on survival [4].

ISCHAEMIC NEPHROPATHY—SURGERY AND NEWER ANTI-HYPERTENSIVE AGENTS (1970–90S)

Serial angiographic studies in the 1970–80s shed light on the natural history of ARVD, raising the concern that this disease could progress in up to 44% of affected arteries, especially within the first two years of follow-up and in arteries with more severe stenosis at baseline [23]. Progressively declining renal function in patients with compromised blood flow to either both kidneys or to a solitary functioning kidney was termed ‘ischaemic nephropathy’ [24]. This was thought to be the primary cause of end-stage kidney disease (ESKD) in ∼14% of all haemodialysis patients [25] and the focus of revascularization shifted from blood pressure control to preservation of functional renal tissue.

Despite development of better anti-hypertensive agents in the mid-1980s, these were not thought to mitigate ARVD progression. Angiotensin-converting enzyme inhibitors (ACEi) were thought to decrease perfusion pressure across a stenosis, precipitating episodes of significant acute kidney injury (AKI) especially in patients with bilateral RAS or RAS in a solitary kidney [6]. A prospective study performed by Dean et al.  [17] in 1981 looked at 41 patients with presumed renovascular hypertension (RVH) randomly allocated to medical management and showed that 17 patients (41%) had significant deterioration in renal function or loss of renal size that required referral for surgery despite adequate blood pressure control in the majority (88%).

In 1991, the same group looked retrospectively at 58 patients with presumed ‘ischaemic nephropathy’ based on a diagnosis of underlying ARVD and serum creatinine of at least 1.8 mg/dL (158 µmol/L), who underwent renovascular surgery. Pre-operative glomerular filtration rate (GFR) ranged between 0 and 46 mL/min, and eight patients were dialysis-dependent at time of surgery. This cohort of patients had rapid loss of renal function pre-operatively which significantly slowed down post-operatively (P = 0.0462). The effect of revascularization was heterogeneous but six out of the eight patients on dialysis at time of operation regained independent renal function immediately after revascularization [7].

Enthusiasm for surgical revascularization continued to increase. An observational study performed by Novick et al. looking at revascularization outcomes over an average follow-up of 46 months now showed that surgery could achieve a 5-year survival rate of 96%, almost equivalent to that expected in a normal population. However, these patients were undoubtedly selected, as those with symptomatic coronary artery disease or cerebrovascular occlusive disease underwent angiographic screening with pre-emptive correction of vascular occlusive lesions. Novick speculated that although revascularization was ineffective in patients with severe renal impairment and irreversible parenchymal damage, it could potentially improve outcomes in highly selected patients with viable parenchyma (Table 2) [18].

Although, as a result of improved cardiovascular survival, increasingly older patients with more severe renal impairment were now being considered for revascularization, renovascular surgery in these high-risk populations was more hazardous and had poorer clinical outcomes, with a 30-day mortality of 15.5% compared with 5.6% in lower risk patients, and a 5-year survival of ∼64% [26].

DECREASING SURGICAL MORBIDITY AND MORTALITY—PERCUTANEOUS TRANSLUMINAL ANGIOPLASTY (1980–90S)

Percutaneous transluminal angioplasty (PTRA) was first applied to RAS in 1978 [8], triggering interest in using this non-invasive and inexpensive technique in patients with a high comorbid burden. Mortality rates were documented to be lower than for renovascular surgery, but there was a disappointingly high technical failure rate of ∼50% for atherosclerotic RAS lesions due to elastic recoil post-intervention (Table 2) [19, 27], especially following dilatation of ostial lesions. A critical review of studies involving the use of PTRA for the management of RVH concluded that there was no evidence that the procedure was superior to medical therapy in patients with ARVD, given that none of these studies were controlled or randomized [27].

IMPROVING TECHNICAL OUTCOMES—PTRA WITH STENTING (LATE 1990S)

The development of balloon-expandable intraluminal stents aimed to overcome the challenges involved in treating ostial atheromatous disease. In a randomized study comparing PTRA against PTRA with stenting (PTRAS) published in 1999, van de Ven established that PTRAS was technically more successful, with a restenosis rate of 14% compared with 48% for PTRA over 6 months follow-up. Most cases of restenosis were silent and detected only on repeat angiography. The technical success of PTRAS did not necessarily translate into ‘clinical’ superiority (Table 3) [9].

Despite this, van de Ven's study fuelled enthusiasm for PTRAS and several observational case series were carried out throughout the 1990s. Although PTRAS appeared to improve or stabilize the decline in renal function in the short term in ∼75%, and also allowed the reintroduction of ACEi [20], a quarter of patients still progressed to ESKD. However, in a small selected population of ARVD patients with pre-procedure decline in renal function, Harden showed that stenting decreased the rate of decline in renal function around four-fold in 18 out of 23 patients studied, even in those with severe, progressive deterioration in renal function [10] (Table 2). Again, in keeping with the contemporary approach to research, no control group was utilized in the study.

ERA OF THE FIRST RCT (LATE 1990S TO 2000)

At the end of the 1990s the emergence of evidence-based medicine heralded the era of RCT in ARVD treatment (Table 3). Two separate small RCTs compared PTRA without stenting against medical therapy. The EMMA Study [29] included 49 hypertensive patients with presumed haemodynamically significant unilateral RAS (≥75% or ≥60% stenosis with a positive lateralization test) but well-preserved renal function while the Scottish and Newcastle study [30] looked at 55 patients with less severe unilateral or bilateral RAS (≥50% stenosis). Both studies involved a run-in period of standardized anti-hypertensive treatment for a few weeks prior to randomization. There was no difference in average blood pressure in patients with unilateral RAS at the end of either study, although Webster et al. detected improved systolic blood pressure control in patients with bilateral RAS. In the EMMA study, 7 out of 26 patients in the control group crossed over to receive intervention due to refractory hypertension, possibly because ACEi were not permitted in these patients. Both the initial qualifying angiography and subsequent angioplasty were associated with a significant complication rate hence these studies cast doubt on the role of revascularization, although there was no long-term clinical outcome data.

The DRASTIC study was a larger RCT involving 106 hypertensive patients again with well-preserved renal function and either unilateral or bilateral RAS ≥50%. These patients were randomized to anti-hypertensive medication only or to angioplasty with or without stenting [31]. Although DRASTIC showed no significant difference between the interventional and medical arm in terms of blood pressure control, renal function or drug doses at 12 months of follow-up, a significant proportion of patients (44%) failed medical therapy and required angioplasty due to refractory hypertension or deteriorating renal function. It was not specified whether these patients had underlying bilateral ARVD, but it was noteworthy that angioplasty was successful in controlling hypertension or decreasing drug doses in this subset of patients [31].

CHANGING NATURAL HISTORY AND IMPROVING CLINICAL OUTCOMES WITH ENHANCED VASCULAR PROTECTION (1990S TO 2000)

In parallel with progressive improvements in percutaneous techniques, new evidence was emerging about the value of vascular protection using aspirin and statins [34, 35]. Concurrent development of renal duplex ultrasonography enabled easier follow-up of ARVD progression and there was a suggestion that atherosclerotic RAS was progressing less commonly to total renal artery occlusion. Although earlier angiographic studies had shown a rate of occlusion of up to 39% over ∼13 months follow-up in arteries with initial RAS ≥75% [23], RAS assessment using novel Duplex ultrasound criteria, which correlated with renal angiography with an accuracy of ∼93% [36] showed a rate of occlusion of 3% over 3 years in renal arteries with baseline RAS ≥60% [37].

The results of two separate observational studies highlighted that ARVD outcomes largely depended on the degree of underlying renal parenchymal damage and overall atherosclerotic burden (Table 2). Chabova et al. retrospectively analysed a cohort of patients with significant ARVD who were treated exclusively medically. The higher average age of this population (71.8 years), compared with that of earlier studies, reflected improved cardiovascular survival. One-third of patients were receiving ACEi, and hypertension was successfully controlled in the majority of patients. There was no information about administration of statins. Only 2 patients out of 68 developed ESKD secondary to progressive vascular disease, while another 2 failed anti-hypertensive therapy [21]. This is in sharp contrast with earlier studies in which up to 41% of patients with ARVD were adjudged as requiring intervention due to progressive deterioration of renal function or loss of renal mass [17].

Losito's group in Italy exploited the state of clinical equipoise between revascularization and medical treatment to determine long-term outcomes such as cardiovascular mortality in two comparable patient groups. In this observational non-randomized study, 54 ARVD patients treated medically and 136 patients who underwent angioplasty were followed-up for an average of 54.4 months. Statins were only used in patients with high cholesterol levels and around a third of patients were on ACEi. Interestingly, ACEi did not increase the risk of loss of renal function but rather, were associated with improved survival in both groups (P = 0.002). Renin–angiotensin blockade became part of the standard of care for ARVD in bilateral as well as unilateral disease, as it conferred cardiovascular protection to this high-risk population. Revascularization on the other hand, again appeared to have no impact on survival or incidence of ESKD [11].

However, despite lack of conclusive evidence that revascularization improves outcomes in atherosclerotic RAS, the late 1990s and early 2000s were characterized by the phenomenon of ‘drive-by angioplasty’, where some interventional cardiologists performed routine renal angiography concurrently with coronary angiography, and carried out PTRAS if RAS was present, irrespective of its degree or clinical associations [38]. Medicare data suggested that 30 000 plus PTRAS were being performed each year in the US alone in this pro-revascularization era.

THE LATER RCT (2002–14)

In the wake of this increased interest in revascularization, which admittedly was partly financially driven, up to one in six patients with ARVD underwent intervention [39]. However, it was now established that the majority of patients did not gain benefit from intervention. There was a need to understand pathogenesis of renal injury in the post-stenotic kidney and how this could influence outcomes. Multi-targeted atherosclerotic risk factor control using anti-platelet therapy, renin–angiotensin blockade and statins became a central part of ARVD management, in an attempt to prevent progression of RAS as well as irreversible renal parenchymal damage.

The contemporary uncertainty paved the way for a head-to-head comparison of optimized medical therapy with revascularization combined with medical therapy in three RCTs [12,13,32]. (Table 3). Novel non-invasive imaging techniques such as computed tomography angiography (CTA) or magnetic resonance angiography (MRA) were employed together with conventional angiography to select patients for enrolment in both STAR and ASTRAL. However, at definitive on-table angiography in patients randomized to stenting, 19% of patients in STAR and 20% of patients in ASTRAL were found to have stenosis <50%, and so did not undergo revascularization. STAR excluded patients with uncontrolled hypertension and ASTRAL [12] specifically excluded patients if the clinician was certain that revascularization would be of benefit, thereby not denying the patient this treatment. Although CORAL investigators attempted to include patients with more significant RAS using a radiology ‘core lab’ to standardize RAS interpretation and severity, inclusion criteria were relaxed due to slow enrolment [13]. In addition, whereas blood pressure was the primary end-point for earlier RCTs, the pre-defined primary outcome measure for ASTRAL and STAR was change in renal function. In ASTRAL, 40% of patients had serum creatinine <150 μmol/L, and hence the likelihood of demonstrating that intervention could exert a beneficial change in renal function was limited. All three RCTs, encompassing almost 1900 randomized patients, demonstrated that revascularization did not confer any added benefit to optimal medical therapy in terms of renal, cardiovascular and mortality outcomes, but lack of selection of the recruited population was actually a major criticism of these studies as patients with ‘high-risk’ features were excluded.

Less patients required renal replacement therapy in both arms of each of these three RCTs compared with earlier studies [17], perhaps reflecting the renoprotective action of ACEi and statins. In ASTRAL, 8% of patients, irrespective of randomization arm, reached ESKD over 34 months while less than half this figure reached ESKD in CORAL after 43 months, but patients in CORAL had a higher GFR at baseline (58 mL/min) than patients in ASTRAL (40 mL/min) [13]. ASTRAL showed that revascularization may slow the rate of progression of renal impairment but this was not statistically or clinically significant. The much lower crossover rate from medical to interventional arms (e.g. 4.4% in ASTRAL) compared with that seen in DRASTIC (44%) was also a reflection of the efficacy of contemporary medical treatment in achieving target blood pressure control and stable renal function. However, 16–22% of patients in both ASTRAL and CORAL still reached adverse renal end-points (largely AKI or renal death in ASTRAL, doubling of creatinine in CORAL) irrespective of treatment arm, probably due to irreversible renal parenchymal injury [40].

The RAS-CAD trial provided insights into the cardioprotective effect of modern medical management. The study enrolled patients with underlying coronary artery disease and concomitant RAS to evaluate the effect of revascularization in addition to optimal medical therapy on progression of left ventricular hypertrophy (LVH). Medical therapy decreased the degree of LVH, while revascularization conferred no additional benefit, although again, patients with severe RAS were excluded from this trial [33]. This finding was replicated in the cardiac MR sub-study of ASTRAL which showed no differences in cardiac structure and function at 12 months after randomization in stented versus medically treated patients [41].

It is noteworthy how historic comparisons can be misleading. The mortality rate reported by ASTRAL (8% per year) was similar to that reported by Wollenweber in 1968. However, the earlier cohort had an average age of 54.5 years compared to 70 years in ASTRAL, and selection bias was manifest by only younger hypertensive patients being referred for angiography and potential surgical revascularization in that era.

THE 2015 VIEW ON REVASCULARIZATION

A recent Cochrane meta-analysis of eight RCTs comparing PTRA or PTRAS and medical therapy for management of ARVD concluded that revascularization had no significant effect on the incidence of cardiovascular or renal adverse events or mortality [42]. As a result, the number of revascularization procedures performed in the past few years has declined by 50–75% [43]. The American Heart Association guidelines reflect the uncertainty surrounding indications for revascularization; however, despite the lack of robust evidence, they advocate that revascularization may still have a role in ‘high-risk’ situations [44].

There has been consistent evidence through the decades of particular subgroups of patients with specific clinical phenotypes who do seem to benefit from revascularization. These patients have tended not to be well-represented in RCTs. As mentioned before, two uncontrolled studies from the 1990s showed that revascularization improved the rate of decline in renal function particularly in patients with rapidly progressive renal impairment or bilateral RAS [7, 10]. There were <100 patients defined as having prior rapidly declining renal function in ASTRAL, and although there was a signal that revascularization impacted upon this, this was non-significant. In contrast, despite a belief that revascularization may be futile in patients with advanced renal failure [18, 45], a combined review of two separate UK and German ARVD cohorts showed that revascularization improved GFR in ∼50% of patients with CKD Stage 4 and 5 and that this was associated with a survival benefit [46].

There is also repeated, albeit non-randomized, evidence that revascularization can prevent recurrent flash pulmonary oedema (FPE) in patients with bilateral RAS or significant RAS to a solitary kidney and this is an accepted indication for revascularization [44]. A recent retrospective single-centre study of 237 patients with underlying ARVD and a high-risk clinical phenotype (i.e. presenting with FPE, refractory hypertension or rapidly declining kidney function), showed that revascularization was associated with improved survival in patients with FPE and in those with the combination of rapidly declining renal function and refractory hypertension [1], but not when the latter conditions presented alone. Other investigators have shown that improved clinical outcomes can be more predictable when the physiological significance of RAS is confirmed by comprehensive ultrasound Doppler studies [47].

FUTURE DIRECTIONS IN MANAGEMENT OF ARVD

Establishing a causal relationship between haemodynamically significant ARVD and high-risk clinical phenotypes would enable timely identification of patients likely to gain benefit from revascularization. There is currently controversy regarding whether renal angiography or Doppler ultrasound can identify haemodynamically significant stenosis with high specificity, and conventional cut-offs (50% stenosis for angiography and peak systolic velocities >2.0 m/s for Doppler ultrasound) overestimate severity of stenosis [48]. Although evidence is limited, transstenotic pressure gradients may correlate more closely with the physiological significance of RAS [49].

Functional characterization of renal parenchyma distal to RAS may help predict revascularization outcomes. High renal resistance indices (RI) (>0.8) obtained on Doppler ultrasound have been suggested to indicate irreversibly damaged renal parenchyma [50]. However, this has not been confirmed in other studies [51] hence RI should not dictate revascularization decisions [48]. Modern imaging techniques such as blood-oxygen level-dependent magnetic resonance imaging (BOLD-MRI) can identify critically ischaemic kidneys. A high level of hypoxia in relation to the GFR of individual kidneys can predict improvement in renal function post-revascularization and this might represent ‘hibernating parenchyma’ [14]. Recent research efforts have focussed on prevention of irreversible parenchymal injury in order to optimize clinical outcomes. Oxidative stress and ischaemia–reperfusion injury in animal experiments stimulate secretion of an adverse cytokine profile and macrophage infiltration, and eventually lead to loss of renal microvascular architecture. Novel strategies based on cell-based therapies are being explored to target this remodelling process [15, 16].

Despite the neutral results of the latest RCTs, it is clear that patients with ARVD constitute a very heterogeneous group, and hence indications for revascularization need to be individualized (Table 4). Multi-targeted treatment for atherosclerosis remains the undeniable cornerstone of ARVD management, but identifying the subset of patients who do respond to revascularization remains a priority. Although ideally another RCT should be performed to help define the role of revascularization in these high-risk subgroups, it seems unlikely that this will occur in the near future. However, an international registry study for patients with ARVD undergoing revascularization should be established to shed light on the responses of individual clinical phenotypes; this would also serve as a powerful research tool.

CONCLUSION

The past 80 years have witnessed significant progress in our understanding of the natural history, pathophysiology and management of ARVD. Contemporary therapy for atherosclerosis has decreased the rate of RAS progression to total occlusion after initial clinical presentation [17, 37] and has improved clinical outcomes. The role of revascularization needs to be redefined in light of the neutral results of recent large RCTs. Future research efforts need to focus on timely identification of subgroups of patients who can potentially gain benefit from revascularization, as well as on controlling the intra-renal inflammatory, fibrotic and microvascular processes which perpetuate ischaemic injury.

CONFLICT OF INTEREST STATEMENT

The authors have no conflicts of interest or relevant funding to disclose.

REFERENCES