The first renal artery balloon angioplasties were performed by Felix Mahler in Berne and Andreas Grüntzig in Zurich in 1977 using self-made, bulky, coaxial balloon catheters.1,2 Until the beginning of the 1990s, balloon angioplasty was the only method of percutaneous treatment of renal artery stenosis (RAS) with satisfying acute and long-term results for angioplasty of stenoses caused by fibromuscular dysplasia (FMD) and atherosclerotic RAS of the renal artery trunk, with procedural success rates of 82–100% and restenosis rates of about 10%.3–7 Balloon angioplasty of ostial atherosclerotic lesions was limited by a low acute technical success rate of 50–62% and a high restenosis rate of up to 47% in the long term due to dissections, elastic recoil and rigidity of the lesion in calcified stenoses.3,8 The introduction of stenting has revolutionised percutaneous renal revascularisation. Following promising single-centre reports,8–15 two randomised studies proved the superiority of stenting over conventional balloon angioplasty in terms of acute treatment success and technical durability3,16 of the endovascular treatment of atherosclerotic ostial RAS, the most common manifestation of RAS. Nowadays, using pre-mounted low-profile stent devices, atherosclerotic RAS can be treated successfully in almost all cases, with one-year restenosis rates ranging from 0 to 23% depending on the diameter of the renal artery.16–21
Despite these technical improvements, the indications for renal artery stenting are still a matter of debate due to the unproven benefit of renal revascularisation compared with best medical therapy,3,22,23 including the recently presented results of the ASTRAL trial (Angioplasty and STent for Renal Artery Lesion)24 and the STAR trial.25 However, both trials, as well as the former POBA versus Medication trials,3,22,23 are hampered by inappropriate patient selection, including a significant number of patients having had only mild or moderate renal artery disease.
Renal Artery Stenting Technique
Indications for endovascular treatment of complex, potentially complication-associated lesions are summarised in Table 1. The potential risks of endovascular treatment of complex renal artery lesions are listed in Table 2. Even using contemporary devices and techniques, there is still a peri-procedural risk of death, as recently reported by the STAR investigators.25
Pre-interventional diagnostic work-up of patients with RAS consists primarily of colour duplex ultrasound. In experienced hands this technology allows the reliable diagnosis of severe RAS – defined as at least 70% stenosis.26 Furthermore, it is the ideal diagnostic tool to examine the morphology of the abdominal aorta in terms of detecting aneurysm formation and plaque burden. However, this technology frequently misses atypically located accessory renal arteries and usually does not allow the visualisation of the course of the renal artery in detail and the division into the side branches. For these purposes, magnetic resonance angiography (MRA) or computed tomography angiography (CTA) are the diagnostic tools of choice. These methods allow the 3D reconstruction of the renal artery anatomy and, in the case of CTA, evaluation of the plaque composition. This can be helpful for planning the interventional access route even if the majority of renal interventions can be managed via femoral access.
The most frequently used access routes are femoral and brachial, with few physicians preferring radial access. The main indication for brachial or, if preferred, radial access is an acute angle of the off-take of the renal artery. Another more rare indication is an untreated total occlusion of the infra-renal aorta (see Figure 1). The most frequently used guiding catheter configuration for this approach is the multipurpose shape – the Amplatz-right-1 and the Judkins-right-4 configurations.
In particular cases of an ulcerated and bent course of the renal artery in which the guidewire – usually a 0.014-inch stiff uncoated type – cannot be navigated through the lesion, the telescoping technique can manage to cross the lesion with wire. A 125cm-long 4Fr angled diagnostic catheter (e.g. Multipurpose) has to be introduced through a 6Fr guiding catheter (100cm in length) to navigate the guidewire through the lesion (see Figure 2).
In the majority of cases the femoral approach is feasible, even in unfavourable anatomical situations, as shown in Figure 3, using the telescoping technique.
Usually, a 6Fr guiding catheter is big enough to introduce stent devices up to a diameter of 7mm. However, in bifurcation lesions requiring kissing-balloon angioplasty or kissing stenting, 8Fr guiding catheters may be indicated. In small aortic diameters, the IMA or Judkins-right configuration is appropriate; in larger diameters (>20mm) the ‘renal double curve’ or ‘hockey stick’ configurations fit the best. In extremely caudal angulated off-takes of the renal artery the telescoping technique must be used to introduce the guidewire – a 5Fr Simmonds (SIM) 1 or SOS Omni diagnostic catheter has to be introduced through a 6Fr guiding catheter, such as an ‘IMA’ shaped catheter, to engage the renal artery origin with the guidewire. Once the wire is in place, the guiding catheter is advanced over the diagnostic catheter close to the ostium of the renal artery, and the diagnostic catheter can be withdrawn (see Figure 4). The standard guidewire is a 0.014-inch extra support wire with a floppy radio-opaque tip such as Galeo ES™ (Biotronik, Berlin, Germany). Hydrophilic-coated wires should be used only in difficult anatomies because of the increased risk of perforation of the intrarenal side branches. Pre-dilation is indicated only in subtotal occlusions or highly calcified lesions when it is uncertain if the stenosis can be dilated with a balloon. In situations of a non-dilatable stenosis, a Cutting Balloon™ (Boston Scientific, Natick, MA) or an Angiosculpt™ (Biotronik, Berlin, Germany) might be indicated.
Interventional Tools
In bent arteries, dedicated renal stent devices with a long so-called ‘antiflip’ tip of the balloon catheter facilitate the introduction of the stent into the lesion without pre-dilation. The Hippocampus™ 0.014 Balloon Expanding Rapid Exchange Renal Stent System device (Invatec Corp., Concesio Brescia, Italy – not yet approved in the US) has a long ‘non-flip’ tip with progressive flexibility and minimal entry profile. As the system shows progressive flexibility from the long tip, followed by the long balloon-cone, the guidewire will not be straightened and possibly flipped out of the origin (see Figure 3). As soon as the balloon segment with the crimped stent is advanced through the curve of the guiding catheter or introducer sheath, the long tip is inserted into the renal artery origin so the position cannot be lost.
Currently, no dedicated distal protection devices for renal use are on the market. The most useful device might be the Fibernet® (Lumen Biomedical, Plymouth, MN, US).27 However, the general use of such devices cannot yet be recommended, considering the fact that a small randomised trial recently failed to prove the concept.28 The most dangerous interventional step, potentially leading to renal embolism, is the engagement of the renal artery with the guiding catheter. During this manipulation, a lot of debris can be captured from the aortic wall with the tip of the guiding catheter. With the first selective contrast-agent injection into the renal artery, all of this captured debris is dislodged into the kidney before any distal protection device works. To avoid embolism during this crucial step of the intervention, the ‘no-touch’ technique29 (see Figure 5) should be used; alternatively, before the first dye injection the catheter should be cleaned of debris by aspirating blood through a Y-connector (see Figure 6).
In ostial atherosclerotic RAS, balloon-expandable stents are mandatory to resist vessel recoil at origin. Several stents show a progressive radial strength towards the proximal end, which covers the renal artery origin (e.g. Hippocampus). An interesting new design is the Bullseye® stent (SquareOne, Mountain View, CA, US), which can be secondarily flared at the proximal end of the aortic wall. In more distally located lesions, low-profile self-expanding stents fitting through a 6Fr guiding catheter (e.g. Xpert™, Abbott Vascular, Diegem, Belgium or Maris Deep™, Invatec) should be used in order to avoid stent fractures, which had been occasionally reported for rigid balloon-expandable stents. In some patients with hyper-mobile kidneys, respiration synchronic vessel bending can lead to stent fracture, with consecutive abrupt vessel occlusion resulting in functional kidney loss. No passive stent-coating17,18 or drug-eluting stent technology30 can, as yet, prove a superior technical outcome compared with bare-metal stent technology.
Summary of Technical Outcomes
Overall, including complex lesions, in experienced hands the technical success rates of endovascular renal artery interventions are close to 100%, and reported one-year restenosis rates, depending on the definition of restenosis, range from 5 to 20% using current stent devices.9,12,14,16,17,20,31–34 Restenosis rates inversely correlate with the vessel or stent diameter.17,18 Restenosis can become a challenging situation, especially in those cases where the stent protrudes more than 1mm into the aortic lumen.
The engagement of the proximal stent lumen can become time-consuming and might need several attempts with different kinds of diagnostic or guide catheters. There are two major signs that predict the correct positioning of the wire. The first is when the diagnostic or guide catheter easily slips into the proximal stent segment after the lesion is passed by the wire. If this is not possible, the wire is passed through the stent meshes from the outside. The second sign of correct wire position is when the balloon catheter, or in the case of intended direct stenting the stent device, can easily be placed in the restenosed stent. In some rare cases it can be helpful to use an orthogonal view for the engagement of the stent. Data are rare concerning the appropriate treatment strategy for restenosis. Plain-balloon angioplasty, cutting-balloon angioplasty and restenting are associated with recurrent restenosis rates of about 30%. Covered stents (Jostentgraft™, Abbott Vascular) or drug-eluting stents (TAXUS™, Boston Scientific) seem to considerably reduce this recurrence rate.35 However, no randomised data are yet available to prove either concept.
In summary, techniques and interventional tools for the treatment of RAS have further improved, resulting in excellent acute and long-term technical outcome results when in experienced hands. However, treatment of in-stent restenosis remains a challenge; the introduction of drug-coated balloons might solve this problem in the future, as has been shown for coronary in-stent restenosis.36
Limitations of Published Randomised Controlled Trials
The methodological quality of all published randomised controlled trials comparing either PTA3,22,23 or stenting24,25 with medical therapy, analysed according to the method of Jadad,37 was low to moderate.38 The risk of bias, defined according to the Cochrane Collaboration handbook,39 ranged from moderate to severe.37,38 Major points of criticism are the unblinded analysis of the study results, the high rate of cross-over to angioplasty – in particular in the largest of the three PTA trials (DRASTIC: 44%) – and the long enrolment periods – especially in the DRASTIC trial (seven years for 106 patients), but also in the ASTRAL trial (eight years). Moreover, the most severe bias of all of the studies, including the ongoing CORAL trial (but excluding the RADAR trial), is the definition of the severity of RAS.
Lesion severity started with a reduction of the luminal diameter of 50 and 60%, respectively. Up to 30% of the patients included in the trials had RAS with a diameter stenosis of less than 70%; some less than 50%. None of the studies asked for proof of the haemodynamic relevance of the RAS before enrolment. Experimental studies have shown that lumen diameter reduction does not lead to a drop of post-stenotic flow before exceeding 70%.40,41 The only tool to prove the haemodynamic relevance of RAS is analysis of the intrarenal Doppler flow curves with calculation of the resistance index and the acceleration time by duplex ultrasound.42,43 The bias created by including patients with an insignificant RAS is two-sided: patients undergoing revascularisation will have no clinical benefit from the intervention but will potentially be in danger of peri-procedural complications, as shown in the STAR trial, while the clinical outcome of patients assigned to medical treatment will not be adversely affected by an insignificant RAS. The RADAR trial started in 2008 and is the only trial demanding proof of the haemodynamic relevance of RAS based on duplex ultrasound before including the patient in the trial.
Another key problem of all RAS treatment randomised controlled trials is the high selection bias – on the one hand there is no evidence of a beneficial outcome from the revascularisation of RAS, but on the other hand some endovascular specialists and home-care physicians have experienced impressive improvements in clinical parameters following renal artery revascularisation in selected cases. This may lead to patients who have the potentially highest benefit of revascularisation not being enrolled in randomised trials that include a conservative treatment arm because physicians – and patients – fear harming patients’ health. This has led to the unusual study design of the ASTRAL trial, which randomises only those patients about whom the study’s physicians are uncertain whether they should be revascularised. The CORAL trial, which does not have these restrictive inclusion criteria, is currently facing the problem of relatively low enrolment rates and the inclusion of mostly moderate lesions.
Another question is: what is the appropriate study end-point? Blood pressure control was the primary end-point in the published PTA trials and renal functional outcome in the STAR and ASTRAL trials. Hypertension is not an appropriate end-point as the ‘cure’ of hypertension in patients with atherosclerotic RAS is rare3,9,22,23 due to the fact that, in most of these patients, RAS leads to secondary deterioration of the control of pre-existing essential hypertension. Thus, revascularisation of RAS may improve blood pressure control, which has been reported in randomised trials in terms of a significant reduction in the use of antihypertensive drugs.3,22,23
However, the most important potential benefits of renal artery revascularisation are the stabilisation or improvement of renal function and the avoidance of heart failure. Renal and myocardial dysfunction are the most powerful predictors of cardiovascular mortality.44–47 Thus, a benefit of renal revascularisation compared with conservative therapy is the improvement of renal and cardiac function affecting patient survival, the most important study end-point. There is a known link between the increased mortality of successfully revascularised patients with atherosclerotic RAS with pre-interventional advanced renal dysfunction and myocardial insufficiency. In a prospective multivariable analysis of a consecutive registry of renal artery stenting, the most powerful predictor for mortality was myocardial dysfunction, not renal dysfunction.48 On the other hand, registry data suggest that RAS revascularisation may lead to the regression of myocardial hypertrophy, the key component of diastolic and systolic dysfunction.49,50 The CORAL and RADAR trials will investigate this important link between RAS and cardiovascular events. A recently presented comparison of two consecutive registries with more than 500 patients – one including conservatively treated patients with atherosclerotic RAS from Salford, UK, and a second including patients who underwent stent angioplasty of significant RAS (Bad Krozingen, Germany) – analysed the one-year outcome of the two patient cohorts in terms of renal function and survival.51 In patients with chronic kidney disease (CKD) stages three to five, estimated creatinine clearance was highly significantly improved in the revascularisation group compared with a slight decline in the conservatively treated patient cohort. Moreover, one-year mortality was reduced by 70% in the revascularisation group.
Conclusion
The challenge of conducting a properly designed comparative trial between medical therapy and the revascularisation of atherosclerotic RAS is hampered by the fear of potentially harming the patient by withholding revascularisation in patients with a high risk of end-organ damage being randomised to conservative treatment. Thus, patients with the highest probability of a revascularisation benefit are frequently excluded from randomised trials. This underlines the need – besides that for well-designed randomised controlled trials – for properly controlled real-world registries.