Hypertension is a common, modifiable risk factor of cardiovascular mortality. At present, 30% to 40% of adults in developed countries suffer from hypertension.¹ When hypertension was first recognized, control was difficult due to limited treatment options and often regarded as a fruitless endeavor. In 1931, Dr. Paul Dudley White wrote, “Hypertension may be an important compensatory mechanism which should not be tampered with, even were it certain that we could control it.”² Starting in the 1930s, surgical sympathectomy to reduce sympathetic tone was observed to significantly lower blood pressure at the expense of significant procedural morbidity and long-term disability. It was not until the late 1960s that pharmacotherapy for hypertension became available and widespread. Pharmacologic treatment for hypertension has proven generally effective, spurred development of many classes of medications, and assisted in the reduction in mortality from cardiovascular disease. Despite this success, many patients have hypertension that remains uncontrolled. Many cases of uncontrolled hypertension may be attributed to inaction or lack of awareness on the part of patient or provider, but the prevalence of true treatment-resistant hypertension is increasing. In recent years, interest in modifying the sympathetic nervous system has reemerged with the advent of novel minimally invasive methods such as catheter-based renal artery denervation that can potentially restore more balanced autonomic nervous system physiology and offer alternative treatment options for systemic conditions such as hypertension. In this chapter, we will review what is known about the efficacy of catheter-directed autonomic modulation as a novel treatment of hypertension, discuss best practices to achieve desired results, and outline what the future may hold for this controversial area of endovascular medicine.

Pathophysiology

Although the pathophysiology of hypertension is complex and incompletely characterized, the sympathetic nervous system is known to play an important role. In essential hypertension, prior studies have demonstrated increased systemic sympathetic nerve firing as well as excessive sympathetic drive to the kidneys. In the 1850s, French physiologist Claude Bernard discovered that disrupting innervation of the greater splanchnic nerve resulted in diuresis and that electrical stimulation resulted in antidiuresis.³ Early work on the pressor nerves built the foundation for describing the sympathetic nervous system’s effect on regulation of blood pressure. In animal studies, stimulation of chemosensitive renal afferent nerves has been shown to increase systemic efferent sympathetic nerve activity and increase blood pressure.⁴ Renal efferent nerves increase systemic blood pressure by stimulating renin production, tubular reabsorption of sodium, and renal arterial vasoconstriction.⁵ Furthermore, dense sympathetic innervation of the renal tubules helps regulate pressure natriuresis and sodium excretion in the setting of hypertension, although this process is impaired in long-standing hypertension.

Direct measurement of renal sympathetic activation by sampling renal norepinephrine spillover confirms increased sympathetic drive to the kidneys in the setting of long-standing hypertension. Norepinephrine spillover is defined as the amount of norepinephrine escaping neuronal uptake and local metabolism and “spilling over” into the venous outflow of an organ.⁶ By infusing radiolabeled norepinephrine, the outward flow of norepinephrine can then be measured by renal vein sampling. Measurements of renal norepinephrine spillover suggest elevated spillover in hypertensives compared to the normotensive population. Resistant hypertensives have even higher norepinephrine spillover than the general hypertensive population likely due to excessive underlying sympathetic stimulation and concurrent medications (eg, vasodilators, diuretics) that stimulate sympathetic activation.⁷

Surgical Sympathectomy and Rise of Modern Pharmacotherapy

Surgical sympathectomy was the first intervention investigated for treatment of uncontrolled hypertension. In the 1930s to 1940s, several surgeons described different methods of sympathectomy resulting in significant and sustained reductions in systolic and diastolic pressure.^8,9 Splanchnicectomy and lumbar sympathectomy were 2 more common procedures for malignant hypertension. Studies of surgical sympathectomy demonstrated a high risk of morbidity such as prolonged hospitalization, orthostatic hypotension, impotence, and gait disturbances.¹⁰ Despite associated morbidities, these trials showed a mortality benefit in selected subjects with malignant hypertension.^11,12 These results were not generalizable to all who suffered from hypertension, and many clinicians remained skeptical of surgical therapeutics to reduce blood pressure.² The first effective pharmacologic antihypertensives were ganglionic blockers (eg, tetraethylammonium) modeled after the initial surgical experience, although these too carried a high risk of side effects similar to surgical therapy.¹³

The next stage of antihypertensive development produced centrally acting sympathetic inhibitors as well as α- and β-adrenergic antagonists. Shortly afterward, diuretics and directly acting vasodilators like hydralazine became available. It was not until the Veterans Administration Cooperative Studies in the 1970s that pharmacologic therapy for hypertension became widely accepted.^14,15 Angiotensin-converting enzyme inhibitors and angiotensin receptor blockers as well as dihydropyridine calcium channel blocking agents became commonly used in the 1990s as these medications were more tolerable and provided longer lasting blood pressure reduction. Current guidelines from the Eighth Joint National Committee (JNC-8) recommend combination therapy with the following specific blood pressure targets: <150/90 mm Hg for patients >60 years old and <140/90 mm Hg for patients <60 years old or with chronic kidney disease or diabetes.¹⁶ Recently, new evidence suggests more aggressive antihypertensive control may provide further benefit in selected patients. The Systolic Blood Pressure Intervention Trial (SPRINT) showed a reduction in major adverse cardiovascular events compared to usual practice at 1 year with more aggressive antihypertensive therapy targeting systolic blood pressure <120 mm Hg in patients >50 years old with high-risk features for cardiovascular events. Existing guidelines and recent studies highlight the importance of antihypertensive therapy and the need for additional therapies.¹⁷

Resistant and Uncontrolled Hypertension

Despite advances in our understanding of resistant hypertension and in pharmacologic treatment, the prevalence of true treatment-resistant hypertension is as high as 15% to 30% of treated hypertensive patients.¹⁸ Patients with resistant hypertension, according to an American Heart Association consensus document, are defined as patients whose blood pressure remains above goal despite adequate treatment with 3 or more different classes of antihypertensive agents, ideally including a diuretic.¹⁹ Physician inertia, medication noncompliance, and nonadherence to lifelong pharmacologic therapy are all attributed to hypertensive pharmacologic pseudoresistance, but a sizable proportion of remaining patients can be classified as true treatment-resistant hypertensives. Resistant hypertensives are at higher risk for cardiovascular events. In patients with hypertension and coronary artery disease, the prevalence of resistant hypertension is greater than 30% and is associated with a higher risk of all-cause death and a higher risk of cardiovascular mortality.²⁰ An analysis of the Reduction of Atherothrombosis for Chronic Health (REACH) registry, a multinational database of patients with 3 or more atherosclerotic risk factors or with known coronary, cerebrovascular, or peripheral vascular disease, identified a resistant hypertension prevalence of 12.7%. These patients suffered an 11% higher hazard of a combined primary end point of cardiovascular death, myocardial infarction, and stroke at 4 years. A dose-response relationship was noted, with patients on more antihypertensive medications experiencing a significantly higher incidence of cardiovascular outcomes.²¹

The resistant hypertensive population can be difficult to separate from uncontrolled hypertensives as a whole. The spectrum of uncontrolled hypertension ranges from patients with high blood pressure due to inadequate medical regimens or nonadherence to true treatment resistance.¹⁹ In the Antihypertensive and Lipid-Lowering Treatment to Prevent Heart Attack Trial (ALLHAT) cohort, nearly half the patients required 1 to 2 medications, whereas the rest needed 3 or more antihypertensives. At 5 years, approximately one-third of patients remained uncontrolled despite 2 medications.²² The ALLHAT study demonstrates that resistant hypertension remains prevalent despite close monitoring and protocol-directed medication titration. Due to the limitations of current pharmacotherapy and the long-term morbidity associated with resistant hypertension, novel catheter-directed and device-based therapies for hypertension are of great interest in their potential to achieve guideline-defined blood pressure targets in patients struggling to respond.

Given the prevalence of resistant hypertension, there is intense academic and commercial interest in developing therapies that can predictably modulate autonomic tone for promotion of health. Ablating renal sympathetic innervation was originally thought to provide an antihypertensive effect by reducing tubular reabsorption of sodium and downregulating the renin-angiotensin-aldosterone system (RAAS) via interruption of signaling from the central nervous system (CNS) to the kidneys via renal efferent nerves. More recently, it has been recognized that disruption of renal afferent nerves also alters the feedback loop of sympathetic signaling from the kidneys back to the CNS, which may reduce central systemic sympathetic outflow as well. Studying the effects of total nephrectomy and sympathetic tone in patients with end-stage renal disease, Converse et al²³ demonstrated increased postganglionic sympathetic nerve discharge in those who had not undergone nephrectomy. Thus, excess sympathetic nervous activity in the setting of chronic renal insufficiency is associated with renal afferent innervation. The concept of a feedback loop of efferent and afferent signaling between the kidneys and brain may have far-reaching influence on other disease and metabolic activity.

From the initial surgical experience in the 1930s to 1940s, sympathetic denervation was established as a means to reduce blood pressure and mortality.¹² However, how to selectively target renal sympathetic nerves for blood pressure reduction while avoiding the morbidity of disrupting neighboring abdominal, pelvic, and lower extremity innervation remained a challenge. The anatomy of postganglionic renal sympathetic nerves is favorable for catheter-directed therapy as neuronal fibers extending from the sympathetic chain reside in the adventitia along the length of the renal arteries in route to the kidneys. The first patented minimally invasive method to disrupt renal innervation emerged in the early 2000s with development of a single-lead endovascular radiofrequency ablation catheter that could be navigated along the length of the renal artery (Fig. 53-1). With electrode wall contact on the intimal side, externally directed radiofrequency energy could thermally ablate neighboring sympathetic nerves residing in the adventitia. Initial studies applied radiofrequency energy every 5 mm along the length of bilateral main renal artery trunks.²⁴ Early swine catheter studies demonstrated a reduction in norepinephrine content in renal outflow greater than 85%, which was comparable to direct surgical renal denervation. A subsequent first-in-human report described successful renal denervation in a 59-year-old man with a history of 2 transient ischemic attacks, obstructive sleep apnea, and long-standing hypertension despite 7 antihypertensive agents. His mean blood pressure prior to invasive treatment was 161/107 mm Hg. At 30 days, his blood pressure was reduced to 141/90 mm Hg and then to 127/81 mm Hg at 1 year. Total-body norepinephrine spillover was reduced by 42% along with reductions in other measures of sympathetic excess.²⁵

The initial open-label, proof-of-concept pilot study, SYMPLICITY HTN-1,²⁴ enrolled 153 patients at 19 investigational sites in Australia, Europe, and the United States (Table 53-1). Patients with systolic office blood pressure (OBP) ≥160 mm Hg despite taking at least 3 antihypertensive drugs, including a diuretic, were eligible. Investigators performed catheter-directed radiofrequency ablation of the renal sympathetic nerves with the Symplicity Flex catheter (Medtronic, Santa Rosa, CA). The study achieved systolic and diastolic OBP reductions of 20 and 10 mm Hg at 1 month, respectively. The study investigators have reported data on 88 subjects out to 3 years with sustained OBP reduction (–32/–14 mm Hg).^26,27 A follow-up phase II study, SYMPLICITY HTN-2, enrolled 106 patients at 24 centers in Australia, New Zealand, and Europe in a multicenter, prospective, randomized controlled trial (see Table 53-1).²⁸ Patients with baseline systolic OBP ≥160 mm Hg (or ≥150 mm Hg in patients with type 2 diabetes mellitus) despite at least 3 antihypertensive drugs (diuretic was not mandated) were eligible for 1:1 randomization between catheter-based renal denervation (RDN) and control. The control arm of the trial was not blinded as patients continued preexisting antihypertensive therapy. The mean change in systolic and diastolic OBP at 6 months was –32/–12 mm Hg in the RDN group compared to an insignificant change from baseline (mean, 1/0 mm Hg) in the control group. Further follow-up data of 70 of the 89 patients in SYMPLICITY HTN-2 who received RDN demonstrated a mean 36-month (treatment) and 30-month (crossover) systolic and diastolic OBP reduction of –34 and –14 mm Hg, respectively.²⁹

Importantly, neither SYMPLICITY HTN-1 nor HTN-2 reported any significant differences in adverse events or measures of safety between denervation treatment and control groups. At long-term follow-up, renal function was preserved in each study, and initial imaging substudy analysis showed no incidence of renal artery stenosis at the site of ablation. Postural hypotension was not observed, suggesting that venoconstriction remains intact after RDN. Finally, given persistent blood pressure reduction data and signs of an increasing responder rate over time, the Symplicity Flex catheter received CE (Conformité Européene) market approval as a treatment for resistant hypertension in Europe and subsequent similar approval in several areas around the globe.

The success of SYMPLICITY HTN-1 and HTN-2 generated great enthusiasm for RDN as a novel treatment for resistant hypertension. In order to pursue US Food and Drug Administration (FDA) approval, SYMPLICITY HTN-3, a prospective, single-blind, randomized, sham-controlled phase III trial, was conceived with hopes to definitively demonstrate clinical efficacy and safety of RDN (see Table 53-1).³⁰ A total of 535 patients with resistant hypertension taking maximally tolerated doses of at least 3 antihypertensive medications, including a diuretic, were randomized in 2:1 fashion, with 364 patients in the RDN group and 171 in the control group. The defined end points were as follows: a primary efficacy end point of change in OBP at 6 months, a secondary efficacy end point of change in 24-hour ambulatory blood pressure monitoring (ABPM) at 6 months, and a primary composite safety end point. All patients underwent baseline renal angiography, but after randomization, those in the control group remained on the table as a sham procedure with all randomized patients blinded to their treatment assignment. The intention was to identify patients with severe manifestations of resistant hypertension and to eliminate the bias of study subjects’ awareness of treatment assignment.

At baseline, patients in HTN-3 were taking an average of 5 antihypertensive medications, 4 of which were at maximal dosing. All patients enrolled were on diuretics, primarily of the thiazide class. Despite this aggressive medication regimen, the average baseline systolic OBP was 180 mm Hg. All patients enrolled in HTN-3 also underwent 24-hour ABPM with a mean baseline of 160 mm Hg. At 6 months, the adverse event rate was low and no different from control, demonstrating that RDN, as implemented in HTN-3, can be performed safely. Although OBP and 24-hour ABPM decreased by 14.13 mm Hg and 6.89 mm Hg, respectively, after RDN, similar declines were also seen in the control group. The mean difference in office-based systolic blood pressure was –2.4 mm Hg, and mean difference in 24-hour ABPM was –1.96 mm Hg. Both of these differences failed to meet predefined superiority margins and were statistically insignificant. After completing 6 months of follow-up, control group patients who qualified to crossover to RDN were treated and followed but failed to demonstrate any significant superiority in blood pressure lowering 6 months after RDN. Given prior data suggesting a long-term responder effect, 12-month OBP and 24-hour ABPM measures of patients initially randomized to denervation were examined but also failed to demonstrate any significant long-term difference from those in the control arm on medications alone.³¹

The results of SYMPLICITY HTN-3 greatly tempered prior enthusiasm that had developed around the concept of RDN. HTN-3 also launched a debate regarding the validity of all prior studies of RDN that had demonstrated such a stark difference in efficacy of blood pressure lowering while excluding sham procedure in a blinded control group. Questions surfaced as to whether previous demonstrations of blood pressure reduction after RDN were merely a manifestation of placebo effect. Given the many physiologic components driving elevated blood pressure in hypertensive patients, many also question whether denervation trial participants manifested the Hawthorne effect of behavior modification by virtue of being observed in a trial, thereby leading to blood pressure declines in both treatment and control groups. Another concern raised in the design of these RDN trials is the likelihood of regression to the mean given that qualifying patients were included primarily based on isolated OBP measurements above 160 mm Hg, and not 24-hour ABPM.³² Proponents of RDN have challenged that the design of HTN-3 underestimated the potential benefits of RDN. The following section will explore these questions.

Technologic Limitations of the Symplicity Flex Catheter: Was Denervation Underdosed?

The catheter tested in the SYMPLICITY trials was a first-generation device with a single-lead electrode. Precise electrode positioning within the renal artery, proper wall apposition, and successful circumferential denervation are extremely operator dependent with this device. Further refinements and engineering advances have created “second-generation” catheter platforms that may better ensure adequate denervation and optimization of results. These catheters can deliver multidirectional circumferential nerve ablation, which is more efficient and reproducible in achieving complete denervation by maximizing “4-quadrant” renal arterial denervation. Ideally, the procedure should achieve a full 360° ablation by applying energy at the superior, inferior, anterior, and posterior arterial walls. A substudy analysis of HTN-3 suggested the number of successful renal arterial quadrants ablated is an independent predictor of blood pressure reduction after RDN. Of note, less than 25% of patients received circumferential 4-quadrant denervation in both right and left renal arteries. A dose effect trend was also documented between the number of full 4-quadrant ablations achieved during RDN and blood pressure lowering.³³ Although HTN-1 tested regional norepinephrine spillover after denervation to confirm the sympathetic modulation, HTN-3 did not include norepinephrine spillover measurements, rendering it difficult to compare the concept of dose effect between trials. The concept of a denervation dose effect requires further study for validation. Leaders in the field now debate that HTN-3’s failure to achieve its prespecified end points was attributable to inadequate denervation due to protocol limitations and overall operator inexperience. A wide range of operator volume was seen. Most operators had not previously performed RDN.⁷

How Generalizable and Homogeneous Was the SYMPLICITY HTN-3 Patient Population?

The severity of hypertension required to qualify for randomization was so extreme that it is difficult to generalize the HTN-3 study population to uncontrolled hypertensive patients typically seeking assistance in lowering their blood pressure. In one single-center study, only 0.8% of the entire hypertensive population of an academic cardiology patient practice would have met the general inclusion criteria of HTN-3.³⁴ Some have speculated that the underlying pathophysiology of resistant hypertension in many patients enrolled in HTN-3 may have been beyond alteration, even if adequate circumferential denervation had been achieved.

Ideally, in testing the efficacy of blood pressure lowering of a novel therapy, a trial would include patients with similar acuity of illness and minimize any confounding factors that would cloud the true treatment effect of the tested intervention. Although only patients thought to have severe true treatment-resistant hypertension were enrolled, there was large variance between patients in the number and type of medications taken. After enrollment, a substudy analysis of HTN-3 revealed that 38.2% and 42.1% of patients in the RDN and sham control groups, respectively, changed either the number of medications or dosing during the 6-month study period.³³ Such common medication changes confound the testing of a treatment effect. It is impossible to retrospectively determine the underlying nature of resistant hypertension in patients who qualified for randomization. Compliance with pharmaceutical regimens was confirmed only by patient interview and without an objective measure of drug ingestion.

The measurement of blood pressure in these and other antihypertensive efficacy trials has come under criticism in that almost all have historically used OBP for qualification and as a primary outcome measurement. Increasingly, however, the use of OBP limits assessment of blood pressure to a small window of time that is prone to confounders in the office setting such as white coat hypertension. Twenty-four–hour ABPM, with interventions such as witnessed medication ingestion prior to recording, has been suggested as a potentially more valid method to measure blood pressure in such trials.

Differences in treatment efficacy between African Americans and non–African Americans were also noted. Assessing outcomes by ethnicity, systolic OBP response in the control (sham) arm showed a significant and larger reduction in African Americans than in non–African Americans. Assessing prescribed medications, African Americans were prescribed aldosterone antagonists and vasodilators more often than non–African American participants. Furthermore, African Americans had significantly higher baseline diastolic blood pressure and more complex antihypertensive regimens (eg, at least 1 medication prescribed 3 times daily). Through interaction analysis, some experts hypothesize that improved medication adherence after randomization contributed to greater blood pressure reduction in the control (sham) arm, especially in the African American cohort.³⁵ In subgroup analysis, blood pressure response to RDN in non–African Americans was similar to that which has been described in prior RDN trials and global registries. These differences will need to be studied further to better understand what distinguishes RDN responders from nonresponders.