Download as pdf or txt
Download as pdf or txt
You are on page 1of 6

Curr Cardiol Rep (2017) 19: 75

DOI 10.1007/s11886-017-0894-2

PERIPHERAL VASCULAR DISEASE (CJ COOPER AND R GUPTA, SECTION EDITORS)

Renal Artery Stenosis: New Findings from the CORAL Trial


Rajesh Gupta 1 & Salem Assiri 1 & Christopher J. Cooper 1

Published online: 27 July 2017


# Springer Science+Business Media, LLC 2017

Abstract ment of RAS has been controversial for many years. The
Purpose of Review The goal of this review is to summarize CORAL trial provides important insights into the optimal
recent advances in the field and highlight important new in- management of subjects with RAS.
sights from the Cardiovascular Outcomes in Renal
Atherosclerotic Lesions (CORAL) trial regarding the optimal
management of patients with renal artery stenosis (RAS). Keywords Renal artery stenosis . CORAL . Stent .
Recent Findings The CORAL trial demonstrated that subjects Hypertension
with RAS had similar outcomes whether randomized to opti-
mal medical therapy alone or optimal medical therapy plus
renal artery stenting. Subgroup analyses have failed to dem-
onstrate that baseline blood pressure or lesion gradients can Introduction
predict which subjects may have improved response after stent
intervention. Importantly, urine albumin to creatinine ratio Atherosclerotic renal artery stenosis (RAS) is a manifestation
appears to different subjects that may benefit from stent inter- of atherosclerosis affecting the renal arteries, typically at the
vention versus subjects that are unlikely to achieve any bene- ostium of the vessel. In some patients, it may lead to disturbed
fit. In addition, there was a trend toward increase benefit in intrarenal hemodynamics and subsequent sequelae including
subjects with greater percent stenosis. hypertension and chronic kidney disease. The prevalence or
Summary Atherosclerotic RAS is a frequent finding and is RAS can vary from 1 to 5% in unselected populations with
often seen in patients with resistant hypertension, congestive hypertension, and up to 15–40% in populations with other
heart failure, chronic kidney disease, and rarely those who manifestations of atherosclerosis, such as peripheral vascular
need renal replacement therapy. Risk factors for RAS overlap disease (PVD) and coronary artery disease (CAD) [1–4].
with those of generalized atherosclerosis including hyperlip- RAS is oftentimes found in patients with hypertension,
idemia, smoking, hypertension, and diabetes. Patients with chronic kidney disease, congestive heart failure, and other
CAD or PVD frequently have co-existing RAS. The manage- sequelae of hypertension, including myocardial infarction
and stroke. The Cardiovascular Outcomes in Renal
Atherosclerotic Lesions (CORAL) trial was the largest ran-
domized controlled trial of medical therapy alone versus med-
This article is part of the Topical Collection on Peripheral Vascular ical therapy plus stent intervention for subjects with RAS. The
Disease CORAL trial was designed to overcome limitations of prior
RCTs in this field and provide definitive evidence regarding
* Rajesh Gupta the optimal treatment strategies for subjects with RAS and
rajesh.gupta@utoledo.edu hypertension or chronic kidney disease (CKD).
The goals of this review are to provide an overview of the
1
University of Toledo Medical Center, 3000 Arlington Ave, MS# evidence base for the management of RAS and to summarize
1118, Toledo, OH 43614, USA the best practices for treating this condition.
75 Page 2 of 6 Curr Cardiol Rep (2017) 19: 75

Background Clinical Trial Overview stenting and medical therapy versus medical therapy alone
(35.1 and 35.8%, respectfully, HR with stenting 0.94; 95%
In the 1990s, several uncontrolled studies showed that renal CI, 0.76 to 1.17; p = 0.58) (Fig. 1) [9••].
artery angioplasty with or without stenting resulted in blood
pressure reductions [5, 6]. Subsequently, randomized con-
trolled trials were performed to assess medical therapy versus
percutaneous intervention for RAS. The Angioplasty and Renal Artery Stenosis: Contrast with Other Forms
Stenting for Renal Artery Lesions (ASTRAL) trial and the of Atherosclerosis
Stent Placement and Blood Pressure and Lipid-Lowering for
the Prevention of Progression of Renal Dysfunction Caused Importantly, although atherosclerosis is a diffuse process, the
by Atherosclerotic Ostial Stenosis of the Renal Artery (STAR) particular manifestations and treatment response may differ
trial assessed the usefulness of renal artery stenting with re- based on the specific aspects of the local organ or vascular
spect to kidney function and showed no significant difference bed. For example, CAD may present symptomatically with
in this key measure [7, 8]. The DRASTIC trial also evaluated angina or acute coronary syndromes. Similarly, lower extrem-
the effect of renal artery angioplasty without stenting on kid- ity PVD may present symptomatically with claudication or
ney function but was unable to demonstrate any significant critical limb ischemia. Extracranial cerebrovascular disease
difference in the primary endpoints [7]. Some criticisms of is usually asymptomatic until it causes a stroke or TIA, and
these trials included that they may have enrolled some indi- the mechanism is typically by atheroembolization, rather than
viduals with non-significant RAS and they did not have core stenosis leading to reduction in blood flow. In addition, the
lab to confirm the severity of RAS. treatment response and indications for interventional treat-
ment vary by each organ system and vascular bed.
RAS poses a number of difficulties for the clinician. First, it
CORAL Study Overview is usually asymptomatic, except for hypertension which is both
a highly prevalent condition and usually asymptomatic.
The purpose of the CORAL trial was to measure the useful- Second, it often co-exists with other co-morbid conditions
ness of renal artery stenting for the prevention of major ad- which also lead to chronic kidney disease, especially hyperten-
verse renal and cardiovascular events in subjects with RAS. sion and diabetes mellitus. Therefore, it becomes difficult to
Previous trials had focused on blood pressure response or “untangle” the role of RAS versus these other conditions in
measures of renal function, but the CORAL trial was the first the clinical situations such as worsening CKD. Third, it can
RCT in the field of RAS that was powered for important be difficult to accurately diagnose RAS by non-invasive means.
clinical endpoints including mortality, myocardial infarction, Vascular duplex ultrasound requires a highly skilled
stroke, congestive heart failure, worsening renal failure, and
progression to renal replacement therapy [9••].
The CORAL trial screened 5322 subjects and enrolled 947
in the randomized controlled trial. Subjects were enrolled be-
tween 2005 and 2010. The study inclusion criteria included
evidence of severe RAS by angiography, duplex ultrasound,
CT, or MRI AND either HTN with systolic blood pressure of
155 mmHg or higher while on at least two anti-hypertensive
agents, OR CKD with which was defined as an estimated
glomerular filtration rate (GFR) of less than 60 ml/min/
1.73 m2 of body surface area, as calculated with the use of
the modified Modification of Diet in Renal Disease (MDRD)
formula. Most subjects (about 70%) were enrolled based on
angiography with about 25% enrolled based on non-invasive
imaging including duplex ultrasound, computed tomography
angiography (CTA), or MRA [9••]. Of note, for all entry im-
aging, a core lab review of each study was required. Fig. 1 Kaplan-Meier curves for the primary endpoint in the CORAL trial
The primary end points were death from cardiovascular or (a composite of death from cardiovascular or renal causes, stroke,
renal causes, stroke, myocardial infarction, hospitalization for myocardial infarction, hospitalization for congestive heart failure,
progressive renal insufficiency, or the need for permanent renal
congestive heart failure, progressive renal insufficiency, or replacement therapy)
permanent renal replacement therapy. The rate of the primary (From: Cooper CJ, et al., Copyright © 2014 Massachusetts Medical
end point was similar between participants who underwent Society. Reprinted with permission) [9••].
Curr Cardiol Rep (2017) 19: 75 Page 3 of 6 75

sonographer, a well-prepared patient, and a knowledgeable Laird and Rocha-Singh used an embolic protection system
interpreting physician to perform accurately and varying in patients undergoing angioplasty and stenting for atheroscle-
criteria have been used to define severe RAS [10]. Criteria that rotic RAS which showed feasibility, no clinical evidence of
may be used are sensitive (example systolic velocity >180 cm/ peripheral atheroembolism, favorable trends in blood pressure
s), so as to detect RAS when it is present, or more specific control in 6-month follow-up, and no temporary or permanent
(example systolic velocity >300 cm/s), when an affirmative renal replacement therapy [16].
diagnosis needs to be made. CTA and MRA both involve con- Cooper et al. studied renal artery stenting and the role of
trast agents with relative or absolute contraindications in sub- platelet inhibition with abciximab (Reopro, GP IIb/IIIa inhib-
jects with chronic kidney disease and both frequently overesti- itor) and embolic filter protection using the Angioguard de-
mate the severity of RAS. Fourth, the “gold standard” diagnos- vice (Cordis) [17]. The trial was a 2 × 2 factorial design study-
tic measurement, angiographic percent stenosis, has high levels ing the use of abciximab and the Angioguard filter during
of interobserver variability and poor correlation with other renal artery stenting in 100 patients. Surprisingly, abciximab
measures of stenosis severity such as pressure gradients [11]. alone and the Angioguard filter alone were not different than
Lastly, there is poor correlation between different diagnostic the negative control (neither abciximab nor Angioguard fil-
modalities and “severe” RAS is difficult to define and identify ter), however, the group that received both abciximab and the
[12]. All of these problems pose difficulties for the interven- Angioguard filter did have an improvement [17].
tional treatment of RAS. It appears that renal artery stenting significantly activates
In addition to all of these difficulties, there are unique phys- platelets and may also be associated with significant
iological aspects of the kidneys. Relative to other organs, the atheroembolization. Other investigators have also assessed
kidneys receive high blood flow due to their role in filtration various embolic protection devices and mechanisms and some
of the blood supply, yet oxygen extraction in the healthy kid- have documented significant atheroembolic debris during re-
ney is relatively low [13, 14]. nal artery stenting. The situation in the renal arteries may be
The kidney has multiple mechanisms to maintain adequate more similar to that of carotid artery stenting or stenting of a
renal perfusion and oxygenation. The seminal work of saphenous vein bypass graft, which are both situations where
Goldblatt demonstrated that reduction in blood flow to the embolic protection has been demonstrated to be beneficial.
kidney results in increased systemic blood pressure [15]. Certain locations of atherosclerotic plaque are more vulnera-
The tubuloglomerular feedback mechanism involves sensing ble to atheroembolic injury, either due to the nature of the
of sodium by the macula densa and subsequent release of atherosclerotic plaque or due to the unique nature of the vas-
renin which leads to activation of the renin-angiotensin- cular beds in those organ systems. With regard to RAS, it
aldosterone system and increased blood flow to the kidneys remains unknown if embolic filter protection is beneficial
as well as systemic hypertension. In addition, the sympathetic since no properly powered trial has compared stent interven-
nervous system also regulates renal blood flow. RAS leading tion alone to stent intervention with embolic protection. The
to impairment of renal arterial blood flow results in activation role of embolic protection remains uncertain and it seems
of the renin-angiotensin-aldosterone axis with sequelae that unlikely that an adequately powered clinical trial will be per-
include vasoconstriction, sodium and water retention, aldoste- formed to answer this question.
rone secretion, sympathetic nervous system activation, vascu-
lar remodeling, and hypertension. Interrupting this pathogenic
cascade provides the rationale for renal artery stenting; how- What Constitutes Optimal Medical Therapy?
ever, the reality is more complex.
The CORAL trial represents the best standard for optimal
medical therapy in subjects with RAS. In the CORAL trial,
all subjects were treated with aspirin, a long-acting angioten-
Varying Response to Stent Intervention sin receptor blocker (candesartan), a long-acting calcium
channel antagonist (amlodipine), and a statin, (atorvastatin.)
Multiple studies have demonstrated a varying response to Hydrochlorothiazide was added if additional blood pressure
stent intervention with some patients having improved renal control was needed. This combination of anti-platelet therapy,
function, some unchanged, and some developing worsening cholesterol lowering with a potent statin, and effective anti-
renal function after stent intervention. Post procedural renal hypertensive medical therapy constitutes a robust medical in-
function deterioration has been reported to be as high as 20– tervention for subjects with RAS.
40%. Contrast nephropathy and atheroembolism may play a With regard to the choice of the angiotensin II receptor
role in this observed deterioration. blocker (ARB), there are some basic science findings indicat-
Attempts have been made to assess embolic protection in ing unique efficacy with candesartan [18, 19], and although
stent intervention for RAS. this drug was not compared with ACE-I or other ARB, it was
75 Page 4 of 6 Curr Cardiol Rep (2017) 19: 75

well tolerated in the CORAL population. Importantly, ACE-I this likely represents regression to the mean. Furthermore,
or ARB medical therapy is often withheld in subjects with there was no correlation between peak systolic pressure gra-
RAS due to fears of worsening renal function. However, in dient and response to either therapy arms in the CORAL trial.
the CORAL trial, candesartan was well tolerated. Given the Pressure gradients have been postulated as a means to confirm
importance of ACE-I/ARB therapy in the prevention of pro- “true severe” stenosis, but the utility of this metric to predict
gressive CKD, subjects with RAS should be treated with response to stent intervention was not confirmed in the
ACE-I or ARB (again, we favor candesartan since it is the CORAL trial. However, the number of subjects (199) that
best studied ACE-I or ARB in subjects with RAS). had this measurement performed was modest.
Physicians can closely monitor renal function after initiating There was a trend favoring stenting versus medical therapy
candesartan and a 25% rise in creatinine is acceptable, as long with the highest category for stenosis severity (stenosis
as the level then stabilizes. >77.05% measured by core lab), but this was not statistically
significant (p = 0.25) (Fig. 2). These subgroup analyses may
be underpowered and it remains possible that subjects with
Insights from CORAL Subgroup Analyses severe stenosis may benefit from stent intervention, but this
subgroup made up only 25% of the CORAL cohort and rep-
Murphy et al., in a post hoc subgroup analysis from the resents a small subset of the total population of patients with
CORAL study, found baseline blood pressure did not predict RAS. Lastly, it may be difficult to identify these patients in the
response to stent intervention [20••]. This is an important ob- real world given discrepancies between core lab and operator
servation since multiple investigators have shown that high classification of percent stenosis and failure of pressure gradi-
baseline blood pressure is associated with a greater reduction ent to provide meaningful discrimination between responders
in blood pressure after revascularization. Now, we understand and non-responders to stent intervention. It remains to be

Fig. 2 Baseline percent stenosis


and freedom from the primary
composite end point through
5 years in in the CORAL trial. a
Core lab percent stenosis divided
into quartiles (<60.22, 60.22 to
<68.46, 68.46 to <77.05,
≥77.05%). Log-rank test between
groups for first quartile
(<60.22) = 0.4895. Log-rank test
between groups for second
quartile = 0.4780. Log-rank test
between groups for third
quartile = 0.3065. Log-rank test
between groups for fourth
quartile = 0.1487. The p value is
for all available follow-up (not
limited to 5 years). b Core lab
percent stenosis divided into
clinically significant categories
(<60, 60 to <80, ≥80%). Log-rank
test between groups for the first
category (<60) = 0.3992. Log-
rank test between groups for
second category = 0.1023. Log-
rank test between groups for the
third category = 0.2487. The p
value is for all available follow-up
(not limited to 5 years).
(Reprinted from Murphy TP
et al., Copyright © 2015, with
permission from Elsevier) [20••]
Curr Cardiol Rep (2017) 19: 75 Page 5 of 6 75

Fig. 3 Freedom from the primary


composite end point through
5 years of follow-up by treatment
group and urine albumin/
creatinine ratio (uACR).
(Reprinted from: Murphy TP
et al., Copyright © 2016, with
permission from Wolters Kluwer)
[21••]

determined if additional measurements such as IVUS imag- The CORAL data did not identify baseline hypertension
ing, measures of renal perfusion, biomarkers, or measures of severity, peak systolic pressure gradient across the stenosis,
organ function can discriminate between responders and non- or angiographic stenosis severity as predictors of response to
responders to stent intervention. stent intervention. It is possible that subjects with severe ste-
Murphy et al. also performed a post hoc analysis of the nosis and subjects with low urine albumin to creatinine ratio
CORAL trial data based on urine albumin/creatinine ratio may benefit from stent intervention, but this may be a very
(UACR) [21••]. The cutoff value for UACR was 22.5 mg/g small subgroup of RAS patients. The role of embolic protec-
based on the median value for this measurement. Subjects tion and potent anti-platelet therapies remains to be
with lower than median UACR had fewer clinical events with determined.
renal artery stent treatment than those assigned to medical All patients with RAS should be treated with optimal med-
treatment (Fig. 3). However, subjects with higher than median ical therapy consisting of anti-platelet therapy, high-intensity
UACR did not demonstrate these benefits. statin medical therapy, and effective anti-hypertensive treat-
Urine albumin to creatinine ratio provides a simple mea- ments with long-acting drugs that may include amlodipine,
surement of organ function, and importantly, addresses one of candesartan, and the addition of a thiazide-type diuretic if
the main “competing risks” for RAS interventions, namely, needed to achieve goal blood pressure.
intrinsic renal diseases including hypertensive and diabetic
nephropathies. The hypothesis is that patients with RAS and
Compliance with Ethical Standards
healthy kidneys may benefit from renal artery stent interven-
tion. Subjects with poor organ function, measured as an ele-
Conflict of Interest Rajesh Gupta and Salem Assiri declare that they
vated UACR, may not benefit. Further studies are needed to have no conflict of interest.
investigate this observation from the CORAL trial. Christopher J. Cooper reports grants from NIH and Pfizer, grants and
other from AstraZeneca, and grants and other from Cordis.

Human and Animal Rights and Informed Consent This article does
not contain any studies with human or animal subjects performed by any
Conclusion of the authors.

RAS is a common condition in subjects with advanced ath-


erosclerosis. Insights from the CORAL trial have demonstrat-
ed that for most subjects with RAS, optimal medical therapy is References
the best treatment strategy.
Subgroup analysis from the CORAL trial has failed to iden- Papers of particular interest, published recently, have been
tify conventional baseline markers, such as blood pressure se- highlighted as:
verity, as good predictors of response to stent intervention. The •• Of major importance
one baseline variable that appears promising for predicting a
favorable response to stent intervention is the urine albumin to 1. Derkx FH, Schalekamp MA. Renal artery stenosis and hyperten-
sion. Lancet. 1994;344(8917):237–9.
creatinine ratio, with a low UACR portending a favorable 2. Harding MB, Smith LR, Himmelstein SI, Harrison K, Phillips HR,
response. Schwab SJ, et al. Renal artery stenosis: prevalence and associated
75 Page 6 of 6 Curr Cardiol Rep (2017) 19: 75

risk factors in patients undergoing routine cardiac catheterization. J 13. O'Connor PM. Renal oxygen delivery: matching delivery to meta-
Am Soc Nephrol. 1992;2(11):1608–16. bolic demand. Clin Exp Pharmacol Physiol. 2006;33(10):961–7.
3. Olin JW, Melia M, Young JR, Graor RA, Risius B. Prevalence of doi:10.1111/j.1440-1681.2006.04475.x.
atherosclerotic renal artery stenosis in patients with atherosclerosis 14. Bittl JA. Damage control for renal artery stenting. Circulation.
elsewhere. Am J Med. 1990;88(1N):46N–51N. 2008;117(21):2724–6. doi:10.1161/CIRCULATIONAHA.108.
4. Rihal CS, Textor SC, Breen JF, McKusick MA, Grill DE, Hallett 776732.
JW, et al. Incidental renal artery stenosis among a prospective co- 15. Goldblatt H, Lynch J, Hanzal RF, Summerville WW. Studies on
hort of hypertensive patients undergoing coronary angiography. experimental hypertension : I. The production of persistent eleva-
Mayo Clin Proc. 2002;77(4):309–16. doi:10.1016/S0025- tion of systolic blood pressure by means of renal ischemia. J Exp
6196(11)61782-5. Med. 1934;59(3):347–79.
5. Blum U, Krumme B, Flugel P, Gabelmann A, Lehnert T, Buitrago- 16. Laird JR, Tehrani F, Soukas P, Joye JD, Ansel GM, Rocha-Singh K.
Tellez C, et al. Treatment of ostial renal-artery stenoses with vascu- Feasibility of FiberNet® embolic protection system in patients un-
lar endoprostheses after unsuccessful balloon angioplasty. N Engl J dergoing angioplasty for atherosclerotic renal artery stenosis.
Med. 1997;336(7):459–65. doi:10.1056/NEJM199702133360702. Catheter Cardiovasc Interv. 2012;79(3):430–6. doi:10.1002/ccd.
6. Burket MW, Cooper CJ, Kennedy DJ, Brewster PS, Ansel GM, 23292.
Moore JA, et al. Renal artery angioplasty and stent placement: 17. Cooper CJ, Haller ST, Colyer W, Steffes M, Burket MW, Thomas
predictors of a favorable outcome. Am Heart J. 2000;139(1 Pt 1): WJ, et al. Embolic protection and platelet inhibition during renal
64–71. artery stenting. Circulation. 2008;117(21):2752–60. doi:10.1161/
7. van Jaarsveld BC, Krijnen P, Pieterman H, Derkx FH, Deinum J, CIRCULATIONAHA.107.730259.
Postma CT, et al. The effect of balloon angioplasty on hypertension
18. Chen S, Ge Y, Si J, Rifai A, Dworkin LD, Gong R. Candesartan
in atherosclerotic renal-artery stenosis. Dutch Renal Artery Stenosis
suppresses chronic renal inflammation by a novel antioxidant action
Intervention Cooperative Study Group. N Engl J Med.
independent of AT1R blockade. Kidney Int. 2008;74(9):1128–38.
2000;342(14):1007–14. doi:10.1056/NEJM200004063421403.
doi:10.1038/ki.2008.380.
8. Bax L, Woittiez AJ, Kouwenberg HJ, Mali WP, Buskens E, Beek
FJ, et al. Stent placement in patients with atherosclerotic renal artery 19. Yu C, Gong R, Rifai A, Tolbert EM, Dworkin LD. Long-term, high-
stenosis and impaired renal function: a randomized trial. Ann Intern dosage candesartan suppresses inflammation and injury in chronic
Med. 2009;150(12):840–8. W150-1 kidney disease: nonhemodynamic renal protection. J Am Soc
9.•• Cooper CJ, Murphy TP, Cutlip DE, Jamerson K, Henrich W, Reid Nephrol. 2007;18(3):750–9. doi:10.1681/ASN.2006070770.
DM, et al. Stenting and medical therapy for atherosclerotic renal- 20.•• Murphy TP, Cooper CJ, Matsumoto AH, Cutlip DE, Pencina KM,
artery stenosis. N Engl J Med. 2014;370(1):13–22. doi:10.1056/ Jamerson K, et al. Renal artery stent outcomes: effect of baseline
NEJMoa1310753. An optimal medical therapy achieved blood pressure, stenosis severity, and translesion pressure gradient.
similar outcomes as optimal medical therapy plus stent J Am Coll Cardiol. 2015;66(22):2487–94. doi:10.1016/j.jacc.2015.
intervention in the CORAL randomized controlled trial of 09.073. Surprisingly, many of the conventional baseline
subjects with renal artery stenosis. variables thought to predict improved response to stent
10. AbuRahma AF, Srivastava M, Mousa AY, Dearing DD, Hass SM, intervention were not predictive of response in the CORAL
Campbell JR, et al. Critical analysis of renal duplex ultrasound trial. Baseline blood pressure and lesion gradients did not
parameters in detecting significant renal artery stenosis. J Vasc predict response to stent intervention. There was a trend
Surg. 2012;56(4):1052–1059, 60 e1; discussion 9-60. doi:10. toward greater benefit with stent intervention among subjects
1016/j.jvs.2012.03.036. with greater percent stenosis, but this was non-significant.
11. Mangiacapra F, Trana C, Sarno G, Davidavicius G, Protasiewicz M, 21.•• Murphy TP, Cooper CJ, Pencina KM, D'Agostino R, Massaro J,
Muller O, et al. Translesional pressure gradients to predict blood Cutlip DE, et al. Relationship of albuminuria and renal artery stent
pressure response after renal artery stenting in patients with reno- outcomes: results from the CORAL randomized clinical trial (car-
vascular hypertension. Circ Cardiovasc Interv. 2010;3(6):537–42. diovascular outcomes with renal artery lesions). Hypertension.
doi:10.1161/CIRCINTERVENTIONS.110.957704. 2016;68(5):1145–52. doi:10.1161/HYPERTENSIONAHA.116.
12. Drieghe B, Madaric J, Sarno G, Manoharan G, Bartunek J, 07744. This study demonstrated that albuminuria has a
Heyndrickx GR, et al. Assessment of renal artery stenosis: side- significant impact on likelihood of benefit from stent
by-side comparison of angiography and duplex ultrasound with intervention. Subjects with below median values of
pressure gradient measurements. Eur Heart J. 2008;29(4):517–24. albuminuria had improved benefit with renal artery stent
doi:10.1093/eurheartj/ehm631. intervention.

You might also like