Joseph V. Nally, Jr., MD

The balance between dietary intake and excretion of phosphorus can be impaired in patients with decreased renal function, leading to hyperphosphatemia. Many patients with end-stage renal disease on dialysis require phosphorus-binding drugs to control their serum phosphorus levels.

See related editorial and article

In this review, we discuss the pathophysiology of hyperphosphatemia in kidney disease, its consequences, and how to control it, focusing on the different classes of phosphorus binders.

ROLE OF THE INTERNIST

With kidney disease common and on the increase,¹ nephrologists and internists need to work together to provide optimal care.

Further, many internists in managed care plans and accountable care organizations now handle many tasks previously left to specialists—including prescribing and managing phosphorus binders in patients with kidney disease.

PATHOPHYSIOLOGY OF HYPERPHOSPHATEMIA

The pathophysiology of bone mineral disorders in kidney disease is complex. To simplify the discussion, we will address it in 3 parts:

• Phosphorus balance
• The interplay of hormones, including fibro­blast growth factor 23 (FGF23)
• The mechanism of hyperphosphatemia in kidney disease.

Phosphorus balance

Phosphorus is a macronutrient essential for a range of cellular functions that include structure, energy production, metabolism, and cell signaling. It exists primarily in the form of inorganic phosphate.

An average Western diet provides 20 mg of phosphorus per kilogram of body weight per day. Of this, 13 mg/kg is absorbed, and the rest is excreted in the feces.²

Absorption of dietary phosphorus occurs mainly in the jejunum. It is mediated by both a paracellular sodium-independent pathway (driven by high intraluminal phosphorus content) and by active sodium-dependent cotransporters. It is also influenced by diet and promoted by active vitamin D (1,25 dihydroxyvitamin D₃, also called calcitriol).³

Absorbed phosphorus enters the extracellular fluid and shifts in and out of the skeleton under the influence of parathyroid hormone.

Phosphorus excretion is handled almost entirely by the kidneys. Phosphorus is freely filtered at the glomerulus and reabsorbed mainly in the proximal tubule by sodium-phosphate cotransporters.

Normally, when phosphorus intake is adequate, most of the filtered phosphorus is reabsorbed and only 10% to 20% is excreted in the urine. However, the threshold for phosphorus reabsorption in the proximal tubule is influenced by parathyroid hormone, FGF23, and dietary phosphorus intake: low serum phosphate levels lead to an increase in the synthesis of sodium-phosphorus cotransporters, resulting in increased (nearly complete) proximal reabsorption and an increase in the serum phosphorus concentration.⁴ Conversely, both parathyroid hormone and FGF23 are phosphaturic and decrease the number of phosphorus transporters, which in turn leads to increased phosphorus excretion and a decrease in serum phosphorus concentration.⁵

Interplay of hormones

FGF23 is a phosphaturic glycoprotein secreted by osteoblasts and osteocytes. It acts by binding to fibroblastic growth receptor 1 in the presence of its coreceptor, the Klotho protein.⁶

FGF23 is regulated by serum phosphorus levels and plays a major role in the response to elevated serum phosphorus. It causes a direct increase in urinary phosphorus excretion, a decrease in intestinal phosphorus absorption (indirectly via inhibition of calcitriol), and decreased bone resorption via a decrease in parathyroid hormone production.⁷

Mechanism of hyperphosphatemia in kidney disease

In chronic kidney disease, phosphorus retention can trigger secondary hyperparathyroidism, as rising phosphorus levels stimulate FGF23. In the early stages of chronic kidney disease, this response can correct the phosphorus levels, but with several consequences:

• Decreased calcitriol due to its inhibition by FGF23⁹
• Hypocalcemia due to decreased calcitriol (leading to decreased intestinal calcium absorption) and calcium binding of retained phosphorus
• Elevated parathyroid hormone due to low calcitriol levels (lack of inhibitory feedback by calcitriol), hyperphosphatemia, and hypocalcemia (direct parathyroid hormone stimulation).

As the elevated phosphorus level is likely to be the triggering event behind secondary renal hyperparathyroidism, it needs to be controlled. This is accomplished by restricting dietary phosphorus and using phosphorus binders.

HYPERPHOSPHATEMIA MAY LEAD TO VASCULAR CALCIFICATION

Elevated serum phosphorus levels (normal range 2.48–4.65 mg/dL in adults¹¹) are associated with cardiovascular calcification and subsequent increases in mortality and morbidity rates. Elevations in serum phosphorus and calcium levels are associated with progression in vascular calcification¹² and likely account for the accelerated vascular calcification that is seen in kidney disease.¹³

Hyperphosphatemia has been identified as an independent risk factor for death in patients with end-stage renal disease,¹⁴ but that relationship is less clear in patients with chronic kidney disease. A study in patients with chronic kidney disease and not on dialysis found a lower mortality rate in those who were prescribed phosphorus binders,¹⁵ but the study was criticized for limitations in its design.

Hyperphosphatemia can also lead to adverse effects on bone health due to complications such as renal osteodystrophy.

However, in its 2017 update, the Kidney Disease: Improving Global Outcomes (KDIGO) program only “suggests” lowering elevated phosphorus levels “toward” the normal range in patients with chronic kidney disease stages G3a through G5D, ie, those with glomerular filtration rates less than 60 mL/min/1.73 m², including those on dialysis. The recommendation is graded 2C, ie, weak, based on low-quality evidence (https://kdigo.org/guidelines/ckd-mbd).

DIETARY RESTRICTION OF PHOSPHORUS

Diet is the major source of phosphorus intake. The average daily phosphorus consumption is  20 mg/kg, or 1,400 mg, and protein is the major source of dietary phosphorus.

In patients with stage 4 or 5 chronic kidney disease, the Kidney Disease Outcomes Quality Initiative recommends limiting protein intake to 0.6 mg/kg/day.¹⁶ However, in patients on hemodialysis, they recommend increasing protein intake to 1.1 mg/kg/day while limiting phosphorus intake to about 800 to 1,000 mg/day. This poses a challenge, as limiting phosphorus intake can reduce protein intake.

Sources of protein can be broadly classified as plant-based or animal-based. Animal protein contains organic phosphorus, which is easily absorbed.¹⁸ Plant protein may not be absorbed as easily.

Moe et al¹⁹ studied the importance of the protein source of phosphorus after 7 days of controlled diets. Despite equivalent protein and phosphorus concentrations in the vegetarian and meat-based diets, participants on the vegetarian diet had lower serum phosphorus levels, a trend toward lower 24-hour urinary phosphorus excretion, and significantly lower FGF23 levels than those on the meat-based diet. This suggests that a vegetarian diet may have advantages in terms of preventing hyperphosphatemia.

Another measure to reduce phosphorus absorption from meat is to boil it, which reduces the phosphorus content by 50%.²⁰

Processed foods containing additives and preservatives are very high in phosphorus²¹ and should be avoided, particularly as there is no mandate to label phosphorus content in food.

PHOSPHORUS AND DIALYSIS

Although hemodialysis removes phosphorus, it does not remove enough to keep levels within normal limits. Indeed, even when patients adhere to a daily phosphorus limit of 1,000 mg, phosphorus accumulates. If 70% of the phosphorus in the diet is absorbed, this is 4,500 to 5,000 mg in a week. A 4-hour hemodialysis session will remove only 1,000 mg of phosphorus, which equals about 3,000 mg for patients undergoing dialysis 3 times a week,²² far less than phosphorus absorption.

In patients on continuous ambulatory peritoneal dialysis, a daily regimen of 4 exchanges of 2 L per exchange removes about 200 mg of phosphorus per day. In a 2012 study, patients on nocturnal dialysis or home dialysis involving longer session length had greater lowering of phosphorus levels than patients undergoing routine hemodialysis.²³

Hence, phosphorus binders are often necessary in patients on routine hemodialysis or peritoneal dialysis.

PHOSPHORUS BINDERS

Phosphorus binders reduce serum phosphorus levels by binding with ingested phosphorus in the gastrointestinal tract and forming insoluble complexes that are not absorbed. For this reason they are much more effective when taken with meals. Phosphorus binders come in different formulations: pills, capsules, chewable tablets, liquids, and even powders that can be sprinkled on food.

The potency of each binder is quantified by its “phosphorus binder equivalent dose,” ie, its binding capacity compared with that of calcium carbonate as a reference.²⁴

Phosphorus binders are broadly divided into those that contain calcium and those that do not.

Calcium-containing binders

The 2 most commonly used preparations are calcium carbonate (eg, Tums) and calcium acetate (eg, Phoslo). While these are relatively safe, some studies suggest that their use can lead to accelerated vascular calcification.²⁵

According to KDIGO,²⁶ calcium-containing binders should be avoided in hypercalcemia and adynamic bone disease. Additionally, the daily elemental calcium intake from binders should be limited to 1,500 mg, with a total daily intake that does not exceed 2,000 mg.

The elemental calcium content of calcium carbonate is about 40% of its weight (eg, 200 mg of elemental calcium in a 500-mg tablet of Tums), while the elemental calcium content of calcium acetate is about 25%. Therefore, a patient who needs 6 g of calcium carbonate for efficacy will be ingesting 2.4 g of elemental calcium per day, and that exceeds the recommended daily maximum. The main advantage of calcium carbonate is its low cost and easy availability. Commonly reported side effects include nausea and constipation.

A less commonly used calcium-based binder is calcium citrate (eg, Calcitrate). It should, however, be avoided in chronic kidney disease because of the risk of aluminum accumulation. Calcium citrate can enhance intestinal absorption of aluminum from dietary sources, as aluminum can form complexes with citrate.²⁷

Calcium-free binders

There are several calcium-free binders. Some are based on metals such as aluminum, magnesium, iron, and lanthanum; others, such as sevelamer, are resin-based.

Aluminum- and magnesium-based binders are generally not used long-term in kidney disease because of the toxicity associated with aluminum and magnesium accumulation. However, aluminum hydroxide has an off-label use as a phosphorus binder in the acute setting, particularly when serum phosphorus levels are above 7 mg/dL.²⁸ The dose is 300 to 600 mg 3 times daily with meals for a maximum of 4 weeks.

Sevelamer. Approved by the US Food and Drug Administration (FDA) in 1998, sevelamer acts by trapping phosphorus through ion exchange and hydrogen binding. It has the advantage of being calcium-free, which makes it particularly desirable in patients with hypercalcemia.

The Renagel in New Dialysis²⁵ and Treat-To-Goal²⁹ studies were randomized controlled trials that looked at the effects of sevelamer vs calcium-based binders on the risk of vascular calcification. The primary end points were serum phosphorus and calcium levels, while the secondary end points were coronary artery calcification on computed tomography and thoracic vertebral bone density. Both studies demonstrated a higher risk of vascular calcification with the calcium-based binders.

Another possible benefit of sevelamer is an improvement in lipid profile. Sevelamer lowers total cholesterol and low-density lipoprotein cholesterol levels without affecting high-density lipoprotein cholesterol or triglyceride levels.³⁰ This is likely due to its bile acid-binding effect.³¹ Sevelamer has also been shown to lower C-reactive protein levels.³² While the cardiovascular profile appears to be improved with the treatment, there are no convincing data to confirm that those properties translate to a proven independent survival benefit.

The Calcium Acetate Renagel Evaluation³³ was a randomized controlled study comparing sevelamer and calcium acetate. The authors attempted to control for the lipid-lowering effects of sevelamer by giving atorvastatin to all patients in both groups who had a low-density lipoprotein level greater than 70 mg/dL. The study found sevelamer to be not inferior to calcium acetate in terms of mortality and coronary calcification.

Further studies such as the Brazilian Renagel and Calcium trial³⁴ and the Dialysis Clinical Outcomes Revisited trial failed to show a significant long-term benefit of sevelamer over calcium-based binders. However, a secondary statistical analysis of the latter study showed possible benefit of sevelamer over calcium acetate among those age 65 and older.³⁵

To understand how sevelamer could affect vascular calcification, Yilmaz et al³⁶ compared the effects of sevelamer vs calcium acetate on FGF23 and fetuin A levels. Fetuin A is an important inhibitor of vascular calcification and is progressively diminished in kidney disease, leading to accelerated calcification.³⁷ Patients on sevelamer had higher levels of fetuin A than their counterparts on calcium acetate.³⁷ The authors proposed increased fetuin A levels as a mechanism for decreased vascular calcification.

In summary, some studies suggest that sevelamer may offer the advantage of decreasing vascular calcification, but the data are mixed and do not provide a solid answer. The main disadvantages of sevelamer are a high pill burden and side effects of nausea and dyspepsia.

Lanthanum, a metallic element, was approved as a phosphorus binder by the FDA in 2008. It comes as a chewable tablet and offers the advantage of requiring the patient to take fewer pills than sevelamer and calcium-based binders.

Sucroferric oxyhydroxide comes as a chewable tablet. It was approved by the FDA in 2013. Although each tablet contains 500 mg of iron, it has not been shown to improve iron markers. In terms of phosphorus-lowering ability, it has been shown to be noninferior to sevelamer.³⁹ Advantages include a significantly lower pill burden. Disadvantages include gastrointestinal side effects such as diarrhea and nausea and the drug’s high cost.

Ferric citrate was approved by the FDA in 2014, and 1 g delivers 210 mg of elemental iron. The main advantage of ferric citrate is its ability to increase iron markers. The phase 3 trial that demonstrated its efficacy as a binder showed an increase in ferritin compared with the active control.⁴⁰ The study also showed a decrease in the need to use intravenous iron and erythropoesis-stimulating agents. This was thought to be due to improved iron stores, leading to decreased erythropoietin resistance.⁴¹

The mean number of ferric citrate tablets needed to achieve the desired phosphorus-lowering effect was 8 per day, containing 1,680 mg of iron. In comparison, oral ferrous sulfate typically provides 210 mg of iron per day.⁴²

Disadvantages of ferric citrate include high pill burden, high cost, and gastrointestinal side effects such as nausea and constipation.

Chitosan binds salivary phosphorus. It can potentially be used, but it is not approved, and its efficacy in lowering serum phosphorus remains unclear.⁴³

CHOOSING THE APPROPRIATE PHOSPHORUS BINDER

The choice of phosphorus binder is based on the patient’s serum calcium level and iron stores and on the drug’s side effect profile, iron pill burden, and cost. Involving patients in the choice after discussing potential side effects, pill burden, and cost is important for shared decision-making and could play a role in improving adherence.

Phosphorus binders are a major portion of the pill burden in patients with end-stage renal disease, possibly affecting patient adherence. The cost of phosphorus binders is estimated at half a billion dollars annually, underlining the significant economic impact of phosphorus control.¹¹

Calcium-based binders should be the first choice when there is secondary hyperparathyroidism without hypercalcemia. There is no clear evidence regarding the benefit of correcting hypocalcemia, but KDIGO recommends keeping the serum calcium level within the reference range. KDIGO also recommends restricting calcium-based binders in persistent hypercalcemia, arterial calcification, and adynamic bone disease. This recommendation is largely based on expert opinion.

Noncalcium-based binders, which in theory might prevent vascular calcification, should be considered for patients with at least 1 of the following⁴⁴:

• Complicated diabetes mellitus
• Vascular or valvular calcification
• Persistent inflammation.

Noncalcium-based binders are also preferred in low bone-turnover states such as adynamic bone disease, as elevated calcium can inhibit parathyroid hormone.

However, the advantage of noncalcium-based binders regarding vascular calcification is largely theoretical and has not been proven clinically. Indeed, there are data comparing long-term outcomes of the different classes of phosphorus binders, but studies were limited by short follow-up, and individual studies have lacked power to detect statistical significance between two classes of binders on long-term outcomes. Meta-analyses have provided conflicting data, with some suggesting better outcomes with sevelamer than with calcium-based binders, and with others failing to show any difference.⁴⁵

Because iron deficiency is common in kidney disease, ferric citrate, which can improve iron markers, may be a suitable option, provided its cost is covered by insurance.

SPECIAL CIRCUMSTANCES FOR THE USE OF PHOSPHORUS BINDERS

Tumor lysis syndrome

Tumor lysis syndrome occurs when tumor cells release their contents into the bloodstream, either spontaneously or in response to therapy, leading to the characteristic findings of hyperuricemia, hyperkalemia, hyperphosphatemia, and hypocalcemia.⁴⁶ Phosphorus binders in conjunction with intravenous hydration are used to treat hyperphosphatemia, but evidence about their efficacy in this setting is limited.

Hypocalcemia in tumor lysis syndrome is usually not treated unless symptomatic, as the calcium-phosphorus product can increase, leading to calcium phosphate crystallization. When the calcium-phosphorus product is greater than 60, there is a higher risk of calcium phosphate deposition in the renal tubules that can lead to acute renal failure in tumor lysis syndrome.⁴⁷ To lower the risk of calcium phosphate crystallization, calcium-based binders should be avoided in tumor lysis syndrome.

Total parenteral nutrition

Since patients on total parenteral nutrition do not eat, phosphorus binders are considered ineffective; there are no concrete data showing that phosphorus binders are effective in these patients.⁴⁸ In patients with kidney disease, the phosphorus content in the parenteral nutrition formulation must be reduced.

Pregnancy

Data on phosphorus binders in pregnancy are limited. Calcium can cross the placenta. Calcium carbonate can be used in pregnancy, and fetal harm is not expected if calcium concentrations are within normal limits.⁴⁹ Calcium acetate, sevelamer, and lanthanum are considered pregnancy category C drugs. Patients with advanced chronic kidney disease and end-stage renal disease who become pregnant must receive specialized obstetric care for high-risk pregnancy.

FUTURE DIRECTIONS

Future therapies may target FGF23 and other inflammatory markers that are up-regulated in renal hyperparathyroidism. However, trials studying these markers are needed to provide a better understanding of their role in bone mineral and cardiovascular health and in overall long-term outcomes. Additionally, randomized controlled trials are needed to study long-term nonsurrogate outcomes such as reduction in cardiovascular disease and rates of overall mortality.

The balance between dietary intake and excretion of phosphorus can be impaired in patients with decreased renal function, leading to hyperphosphatemia. Many patients with end-stage renal disease on dialysis require phosphorus-binding drugs to control their serum phosphorus levels.

See related editorial and article

In this review, we discuss the pathophysiology of hyperphosphatemia in kidney disease, its consequences, and how to control it, focusing on the different classes of phosphorus binders.

ROLE OF THE INTERNIST

With kidney disease common and on the increase,¹ nephrologists and internists need to work together to provide optimal care.

Further, many internists in managed care plans and accountable care organizations now handle many tasks previously left to specialists—including prescribing and managing phosphorus binders in patients with kidney disease.

PATHOPHYSIOLOGY OF HYPERPHOSPHATEMIA

The pathophysiology of bone mineral disorders in kidney disease is complex. To simplify the discussion, we will address it in 3 parts:

• Phosphorus balance
• The interplay of hormones, including fibro­blast growth factor 23 (FGF23)
• The mechanism of hyperphosphatemia in kidney disease.

Phosphorus balance

Phosphorus is a macronutrient essential for a range of cellular functions that include structure, energy production, metabolism, and cell signaling. It exists primarily in the form of inorganic phosphate.

An average Western diet provides 20 mg of phosphorus per kilogram of body weight per day. Of this, 13 mg/kg is absorbed, and the rest is excreted in the feces.²

Absorption of dietary phosphorus occurs mainly in the jejunum. It is mediated by both a paracellular sodium-independent pathway (driven by high intraluminal phosphorus content) and by active sodium-dependent cotransporters. It is also influenced by diet and promoted by active vitamin D (1,25 dihydroxyvitamin D₃, also called calcitriol).³

Absorbed phosphorus enters the extracellular fluid and shifts in and out of the skeleton under the influence of parathyroid hormone.

Phosphorus excretion is handled almost entirely by the kidneys. Phosphorus is freely filtered at the glomerulus and reabsorbed mainly in the proximal tubule by sodium-phosphate cotransporters.

Normally, when phosphorus intake is adequate, most of the filtered phosphorus is reabsorbed and only 10% to 20% is excreted in the urine. However, the threshold for phosphorus reabsorption in the proximal tubule is influenced by parathyroid hormone, FGF23, and dietary phosphorus intake: low serum phosphate levels lead to an increase in the synthesis of sodium-phosphorus cotransporters, resulting in increased (nearly complete) proximal reabsorption and an increase in the serum phosphorus concentration.⁴ Conversely, both parathyroid hormone and FGF23 are phosphaturic and decrease the number of phosphorus transporters, which in turn leads to increased phosphorus excretion and a decrease in serum phosphorus concentration.⁵

Interplay of hormones

FGF23 is a phosphaturic glycoprotein secreted by osteoblasts and osteocytes. It acts by binding to fibroblastic growth receptor 1 in the presence of its coreceptor, the Klotho protein.⁶

FGF23 is regulated by serum phosphorus levels and plays a major role in the response to elevated serum phosphorus. It causes a direct increase in urinary phosphorus excretion, a decrease in intestinal phosphorus absorption (indirectly via inhibition of calcitriol), and decreased bone resorption via a decrease in parathyroid hormone production.⁷

Mechanism of hyperphosphatemia in kidney disease

In chronic kidney disease, phosphorus retention can trigger secondary hyperparathyroidism, as rising phosphorus levels stimulate FGF23. In the early stages of chronic kidney disease, this response can correct the phosphorus levels, but with several consequences:

• Decreased calcitriol due to its inhibition by FGF23⁹
• Hypocalcemia due to decreased calcitriol (leading to decreased intestinal calcium absorption) and calcium binding of retained phosphorus
• Elevated parathyroid hormone due to low calcitriol levels (lack of inhibitory feedback by calcitriol), hyperphosphatemia, and hypocalcemia (direct parathyroid hormone stimulation).

As the elevated phosphorus level is likely to be the triggering event behind secondary renal hyperparathyroidism, it needs to be controlled. This is accomplished by restricting dietary phosphorus and using phosphorus binders.

HYPERPHOSPHATEMIA MAY LEAD TO VASCULAR CALCIFICATION

Elevated serum phosphorus levels (normal range 2.48–4.65 mg/dL in adults¹¹) are associated with cardiovascular calcification and subsequent increases in mortality and morbidity rates. Elevations in serum phosphorus and calcium levels are associated with progression in vascular calcification¹² and likely account for the accelerated vascular calcification that is seen in kidney disease.¹³

Hyperphosphatemia has been identified as an independent risk factor for death in patients with end-stage renal disease,¹⁴ but that relationship is less clear in patients with chronic kidney disease. A study in patients with chronic kidney disease and not on dialysis found a lower mortality rate in those who were prescribed phosphorus binders,¹⁵ but the study was criticized for limitations in its design.

Hyperphosphatemia can also lead to adverse effects on bone health due to complications such as renal osteodystrophy.

However, in its 2017 update, the Kidney Disease: Improving Global Outcomes (KDIGO) program only “suggests” lowering elevated phosphorus levels “toward” the normal range in patients with chronic kidney disease stages G3a through G5D, ie, those with glomerular filtration rates less than 60 mL/min/1.73 m², including those on dialysis. The recommendation is graded 2C, ie, weak, based on low-quality evidence (https://kdigo.org/guidelines/ckd-mbd).

DIETARY RESTRICTION OF PHOSPHORUS

Diet is the major source of phosphorus intake. The average daily phosphorus consumption is  20 mg/kg, or 1,400 mg, and protein is the major source of dietary phosphorus.

In patients with stage 4 or 5 chronic kidney disease, the Kidney Disease Outcomes Quality Initiative recommends limiting protein intake to 0.6 mg/kg/day.¹⁶ However, in patients on hemodialysis, they recommend increasing protein intake to 1.1 mg/kg/day while limiting phosphorus intake to about 800 to 1,000 mg/day. This poses a challenge, as limiting phosphorus intake can reduce protein intake.

Sources of protein can be broadly classified as plant-based or animal-based. Animal protein contains organic phosphorus, which is easily absorbed.¹⁸ Plant protein may not be absorbed as easily.

Moe et al¹⁹ studied the importance of the protein source of phosphorus after 7 days of controlled diets. Despite equivalent protein and phosphorus concentrations in the vegetarian and meat-based diets, participants on the vegetarian diet had lower serum phosphorus levels, a trend toward lower 24-hour urinary phosphorus excretion, and significantly lower FGF23 levels than those on the meat-based diet. This suggests that a vegetarian diet may have advantages in terms of preventing hyperphosphatemia.

Another measure to reduce phosphorus absorption from meat is to boil it, which reduces the phosphorus content by 50%.²⁰

Processed foods containing additives and preservatives are very high in phosphorus²¹ and should be avoided, particularly as there is no mandate to label phosphorus content in food.

PHOSPHORUS AND DIALYSIS

Although hemodialysis removes phosphorus, it does not remove enough to keep levels within normal limits. Indeed, even when patients adhere to a daily phosphorus limit of 1,000 mg, phosphorus accumulates. If 70% of the phosphorus in the diet is absorbed, this is 4,500 to 5,000 mg in a week. A 4-hour hemodialysis session will remove only 1,000 mg of phosphorus, which equals about 3,000 mg for patients undergoing dialysis 3 times a week,²² far less than phosphorus absorption.

In patients on continuous ambulatory peritoneal dialysis, a daily regimen of 4 exchanges of 2 L per exchange removes about 200 mg of phosphorus per day. In a 2012 study, patients on nocturnal dialysis or home dialysis involving longer session length had greater lowering of phosphorus levels than patients undergoing routine hemodialysis.²³

Hence, phosphorus binders are often necessary in patients on routine hemodialysis or peritoneal dialysis.

PHOSPHORUS BINDERS

Phosphorus binders reduce serum phosphorus levels by binding with ingested phosphorus in the gastrointestinal tract and forming insoluble complexes that are not absorbed. For this reason they are much more effective when taken with meals. Phosphorus binders come in different formulations: pills, capsules, chewable tablets, liquids, and even powders that can be sprinkled on food.

The potency of each binder is quantified by its “phosphorus binder equivalent dose,” ie, its binding capacity compared with that of calcium carbonate as a reference.²⁴

Phosphorus binders are broadly divided into those that contain calcium and those that do not.

Calcium-containing binders

The 2 most commonly used preparations are calcium carbonate (eg, Tums) and calcium acetate (eg, Phoslo). While these are relatively safe, some studies suggest that their use can lead to accelerated vascular calcification.²⁵

According to KDIGO,²⁶ calcium-containing binders should be avoided in hypercalcemia and adynamic bone disease. Additionally, the daily elemental calcium intake from binders should be limited to 1,500 mg, with a total daily intake that does not exceed 2,000 mg.

The elemental calcium content of calcium carbonate is about 40% of its weight (eg, 200 mg of elemental calcium in a 500-mg tablet of Tums), while the elemental calcium content of calcium acetate is about 25%. Therefore, a patient who needs 6 g of calcium carbonate for efficacy will be ingesting 2.4 g of elemental calcium per day, and that exceeds the recommended daily maximum. The main advantage of calcium carbonate is its low cost and easy availability. Commonly reported side effects include nausea and constipation.

A less commonly used calcium-based binder is calcium citrate (eg, Calcitrate). It should, however, be avoided in chronic kidney disease because of the risk of aluminum accumulation. Calcium citrate can enhance intestinal absorption of aluminum from dietary sources, as aluminum can form complexes with citrate.²⁷

Calcium-free binders

There are several calcium-free binders. Some are based on metals such as aluminum, magnesium, iron, and lanthanum; others, such as sevelamer, are resin-based.

Aluminum- and magnesium-based binders are generally not used long-term in kidney disease because of the toxicity associated with aluminum and magnesium accumulation. However, aluminum hydroxide has an off-label use as a phosphorus binder in the acute setting, particularly when serum phosphorus levels are above 7 mg/dL.²⁸ The dose is 300 to 600 mg 3 times daily with meals for a maximum of 4 weeks.

Sevelamer. Approved by the US Food and Drug Administration (FDA) in 1998, sevelamer acts by trapping phosphorus through ion exchange and hydrogen binding. It has the advantage of being calcium-free, which makes it particularly desirable in patients with hypercalcemia.

The Renagel in New Dialysis²⁵ and Treat-To-Goal²⁹ studies were randomized controlled trials that looked at the effects of sevelamer vs calcium-based binders on the risk of vascular calcification. The primary end points were serum phosphorus and calcium levels, while the secondary end points were coronary artery calcification on computed tomography and thoracic vertebral bone density. Both studies demonstrated a higher risk of vascular calcification with the calcium-based binders.

Another possible benefit of sevelamer is an improvement in lipid profile. Sevelamer lowers total cholesterol and low-density lipoprotein cholesterol levels without affecting high-density lipoprotein cholesterol or triglyceride levels.³⁰ This is likely due to its bile acid-binding effect.³¹ Sevelamer has also been shown to lower C-reactive protein levels.³² While the cardiovascular profile appears to be improved with the treatment, there are no convincing data to confirm that those properties translate to a proven independent survival benefit.

The Calcium Acetate Renagel Evaluation³³ was a randomized controlled study comparing sevelamer and calcium acetate. The authors attempted to control for the lipid-lowering effects of sevelamer by giving atorvastatin to all patients in both groups who had a low-density lipoprotein level greater than 70 mg/dL. The study found sevelamer to be not inferior to calcium acetate in terms of mortality and coronary calcification.

Further studies such as the Brazilian Renagel and Calcium trial³⁴ and the Dialysis Clinical Outcomes Revisited trial failed to show a significant long-term benefit of sevelamer over calcium-based binders. However, a secondary statistical analysis of the latter study showed possible benefit of sevelamer over calcium acetate among those age 65 and older.³⁵

To understand how sevelamer could affect vascular calcification, Yilmaz et al³⁶ compared the effects of sevelamer vs calcium acetate on FGF23 and fetuin A levels. Fetuin A is an important inhibitor of vascular calcification and is progressively diminished in kidney disease, leading to accelerated calcification.³⁷ Patients on sevelamer had higher levels of fetuin A than their counterparts on calcium acetate.³⁷ The authors proposed increased fetuin A levels as a mechanism for decreased vascular calcification.

In summary, some studies suggest that sevelamer may offer the advantage of decreasing vascular calcification, but the data are mixed and do not provide a solid answer. The main disadvantages of sevelamer are a high pill burden and side effects of nausea and dyspepsia.

Lanthanum, a metallic element, was approved as a phosphorus binder by the FDA in 2008. It comes as a chewable tablet and offers the advantage of requiring the patient to take fewer pills than sevelamer and calcium-based binders.

Sucroferric oxyhydroxide comes as a chewable tablet. It was approved by the FDA in 2013. Although each tablet contains 500 mg of iron, it has not been shown to improve iron markers. In terms of phosphorus-lowering ability, it has been shown to be noninferior to sevelamer.³⁹ Advantages include a significantly lower pill burden. Disadvantages include gastrointestinal side effects such as diarrhea and nausea and the drug’s high cost.

Ferric citrate was approved by the FDA in 2014, and 1 g delivers 210 mg of elemental iron. The main advantage of ferric citrate is its ability to increase iron markers. The phase 3 trial that demonstrated its efficacy as a binder showed an increase in ferritin compared with the active control.⁴⁰ The study also showed a decrease in the need to use intravenous iron and erythropoesis-stimulating agents. This was thought to be due to improved iron stores, leading to decreased erythropoietin resistance.⁴¹

The mean number of ferric citrate tablets needed to achieve the desired phosphorus-lowering effect was 8 per day, containing 1,680 mg of iron. In comparison, oral ferrous sulfate typically provides 210 mg of iron per day.⁴²

Disadvantages of ferric citrate include high pill burden, high cost, and gastrointestinal side effects such as nausea and constipation.

Chitosan binds salivary phosphorus. It can potentially be used, but it is not approved, and its efficacy in lowering serum phosphorus remains unclear.⁴³

CHOOSING THE APPROPRIATE PHOSPHORUS BINDER

The choice of phosphorus binder is based on the patient’s serum calcium level and iron stores and on the drug’s side effect profile, iron pill burden, and cost. Involving patients in the choice after discussing potential side effects, pill burden, and cost is important for shared decision-making and could play a role in improving adherence.

Phosphorus binders are a major portion of the pill burden in patients with end-stage renal disease, possibly affecting patient adherence. The cost of phosphorus binders is estimated at half a billion dollars annually, underlining the significant economic impact of phosphorus control.¹¹

Calcium-based binders should be the first choice when there is secondary hyperparathyroidism without hypercalcemia. There is no clear evidence regarding the benefit of correcting hypocalcemia, but KDIGO recommends keeping the serum calcium level within the reference range. KDIGO also recommends restricting calcium-based binders in persistent hypercalcemia, arterial calcification, and adynamic bone disease. This recommendation is largely based on expert opinion.

Noncalcium-based binders, which in theory might prevent vascular calcification, should be considered for patients with at least 1 of the following⁴⁴:

• Complicated diabetes mellitus
• Vascular or valvular calcification
• Persistent inflammation.

Noncalcium-based binders are also preferred in low bone-turnover states such as adynamic bone disease, as elevated calcium can inhibit parathyroid hormone.

However, the advantage of noncalcium-based binders regarding vascular calcification is largely theoretical and has not been proven clinically. Indeed, there are data comparing long-term outcomes of the different classes of phosphorus binders, but studies were limited by short follow-up, and individual studies have lacked power to detect statistical significance between two classes of binders on long-term outcomes. Meta-analyses have provided conflicting data, with some suggesting better outcomes with sevelamer than with calcium-based binders, and with others failing to show any difference.⁴⁵

Because iron deficiency is common in kidney disease, ferric citrate, which can improve iron markers, may be a suitable option, provided its cost is covered by insurance.

SPECIAL CIRCUMSTANCES FOR THE USE OF PHOSPHORUS BINDERS

Tumor lysis syndrome

Tumor lysis syndrome occurs when tumor cells release their contents into the bloodstream, either spontaneously or in response to therapy, leading to the characteristic findings of hyperuricemia, hyperkalemia, hyperphosphatemia, and hypocalcemia.⁴⁶ Phosphorus binders in conjunction with intravenous hydration are used to treat hyperphosphatemia, but evidence about their efficacy in this setting is limited.

Hypocalcemia in tumor lysis syndrome is usually not treated unless symptomatic, as the calcium-phosphorus product can increase, leading to calcium phosphate crystallization. When the calcium-phosphorus product is greater than 60, there is a higher risk of calcium phosphate deposition in the renal tubules that can lead to acute renal failure in tumor lysis syndrome.⁴⁷ To lower the risk of calcium phosphate crystallization, calcium-based binders should be avoided in tumor lysis syndrome.

Total parenteral nutrition

Since patients on total parenteral nutrition do not eat, phosphorus binders are considered ineffective; there are no concrete data showing that phosphorus binders are effective in these patients.⁴⁸ In patients with kidney disease, the phosphorus content in the parenteral nutrition formulation must be reduced.

Pregnancy

Data on phosphorus binders in pregnancy are limited. Calcium can cross the placenta. Calcium carbonate can be used in pregnancy, and fetal harm is not expected if calcium concentrations are within normal limits.⁴⁹ Calcium acetate, sevelamer, and lanthanum are considered pregnancy category C drugs. Patients with advanced chronic kidney disease and end-stage renal disease who become pregnant must receive specialized obstetric care for high-risk pregnancy.

FUTURE DIRECTIONS

Future therapies may target FGF23 and other inflammatory markers that are up-regulated in renal hyperparathyroidism. However, trials studying these markers are needed to provide a better understanding of their role in bone mineral and cardiovascular health and in overall long-term outcomes. Additionally, randomized controlled trials are needed to study long-term nonsurrogate outcomes such as reduction in cardiovascular disease and rates of overall mortality.

The balance between dietary intake and excretion of phosphorus can be impaired in patients with decreased renal function, leading to hyperphosphatemia. Many patients with end-stage renal disease on dialysis require phosphorus-binding drugs to control their serum phosphorus levels.

See related editorial and article

In this review, we discuss the pathophysiology of hyperphosphatemia in kidney disease, its consequences, and how to control it, focusing on the different classes of phosphorus binders.

ROLE OF THE INTERNIST

With kidney disease common and on the increase,¹ nephrologists and internists need to work together to provide optimal care.

Further, many internists in managed care plans and accountable care organizations now handle many tasks previously left to specialists—including prescribing and managing phosphorus binders in patients with kidney disease.

PATHOPHYSIOLOGY OF HYPERPHOSPHATEMIA

The pathophysiology of bone mineral disorders in kidney disease is complex. To simplify the discussion, we will address it in 3 parts:

• Phosphorus balance
• The interplay of hormones, including fibro­blast growth factor 23 (FGF23)
• The mechanism of hyperphosphatemia in kidney disease.

Phosphorus balance

Phosphorus is a macronutrient essential for a range of cellular functions that include structure, energy production, metabolism, and cell signaling. It exists primarily in the form of inorganic phosphate.

An average Western diet provides 20 mg of phosphorus per kilogram of body weight per day. Of this, 13 mg/kg is absorbed, and the rest is excreted in the feces.²

Absorption of dietary phosphorus occurs mainly in the jejunum. It is mediated by both a paracellular sodium-independent pathway (driven by high intraluminal phosphorus content) and by active sodium-dependent cotransporters. It is also influenced by diet and promoted by active vitamin D (1,25 dihydroxyvitamin D₃, also called calcitriol).³

Absorbed phosphorus enters the extracellular fluid and shifts in and out of the skeleton under the influence of parathyroid hormone.

Phosphorus excretion is handled almost entirely by the kidneys. Phosphorus is freely filtered at the glomerulus and reabsorbed mainly in the proximal tubule by sodium-phosphate cotransporters.

Normally, when phosphorus intake is adequate, most of the filtered phosphorus is reabsorbed and only 10% to 20% is excreted in the urine. However, the threshold for phosphorus reabsorption in the proximal tubule is influenced by parathyroid hormone, FGF23, and dietary phosphorus intake: low serum phosphate levels lead to an increase in the synthesis of sodium-phosphorus cotransporters, resulting in increased (nearly complete) proximal reabsorption and an increase in the serum phosphorus concentration.⁴ Conversely, both parathyroid hormone and FGF23 are phosphaturic and decrease the number of phosphorus transporters, which in turn leads to increased phosphorus excretion and a decrease in serum phosphorus concentration.⁵

Interplay of hormones

FGF23 is a phosphaturic glycoprotein secreted by osteoblasts and osteocytes. It acts by binding to fibroblastic growth receptor 1 in the presence of its coreceptor, the Klotho protein.⁶

FGF23 is regulated by serum phosphorus levels and plays a major role in the response to elevated serum phosphorus. It causes a direct increase in urinary phosphorus excretion, a decrease in intestinal phosphorus absorption (indirectly via inhibition of calcitriol), and decreased bone resorption via a decrease in parathyroid hormone production.⁷

Mechanism of hyperphosphatemia in kidney disease

In chronic kidney disease, phosphorus retention can trigger secondary hyperparathyroidism, as rising phosphorus levels stimulate FGF23. In the early stages of chronic kidney disease, this response can correct the phosphorus levels, but with several consequences:

• Decreased calcitriol due to its inhibition by FGF23⁹
• Hypocalcemia due to decreased calcitriol (leading to decreased intestinal calcium absorption) and calcium binding of retained phosphorus
• Elevated parathyroid hormone due to low calcitriol levels (lack of inhibitory feedback by calcitriol), hyperphosphatemia, and hypocalcemia (direct parathyroid hormone stimulation).

As the elevated phosphorus level is likely to be the triggering event behind secondary renal hyperparathyroidism, it needs to be controlled. This is accomplished by restricting dietary phosphorus and using phosphorus binders.

HYPERPHOSPHATEMIA MAY LEAD TO VASCULAR CALCIFICATION

Elevated serum phosphorus levels (normal range 2.48–4.65 mg/dL in adults¹¹) are associated with cardiovascular calcification and subsequent increases in mortality and morbidity rates. Elevations in serum phosphorus and calcium levels are associated with progression in vascular calcification¹² and likely account for the accelerated vascular calcification that is seen in kidney disease.¹³

Hyperphosphatemia has been identified as an independent risk factor for death in patients with end-stage renal disease,¹⁴ but that relationship is less clear in patients with chronic kidney disease. A study in patients with chronic kidney disease and not on dialysis found a lower mortality rate in those who were prescribed phosphorus binders,¹⁵ but the study was criticized for limitations in its design.

Hyperphosphatemia can also lead to adverse effects on bone health due to complications such as renal osteodystrophy.

However, in its 2017 update, the Kidney Disease: Improving Global Outcomes (KDIGO) program only “suggests” lowering elevated phosphorus levels “toward” the normal range in patients with chronic kidney disease stages G3a through G5D, ie, those with glomerular filtration rates less than 60 mL/min/1.73 m², including those on dialysis. The recommendation is graded 2C, ie, weak, based on low-quality evidence (https://kdigo.org/guidelines/ckd-mbd).

DIETARY RESTRICTION OF PHOSPHORUS

Diet is the major source of phosphorus intake. The average daily phosphorus consumption is  20 mg/kg, or 1,400 mg, and protein is the major source of dietary phosphorus.

In patients with stage 4 or 5 chronic kidney disease, the Kidney Disease Outcomes Quality Initiative recommends limiting protein intake to 0.6 mg/kg/day.¹⁶ However, in patients on hemodialysis, they recommend increasing protein intake to 1.1 mg/kg/day while limiting phosphorus intake to about 800 to 1,000 mg/day. This poses a challenge, as limiting phosphorus intake can reduce protein intake.

Sources of protein can be broadly classified as plant-based or animal-based. Animal protein contains organic phosphorus, which is easily absorbed.¹⁸ Plant protein may not be absorbed as easily.

Moe et al¹⁹ studied the importance of the protein source of phosphorus after 7 days of controlled diets. Despite equivalent protein and phosphorus concentrations in the vegetarian and meat-based diets, participants on the vegetarian diet had lower serum phosphorus levels, a trend toward lower 24-hour urinary phosphorus excretion, and significantly lower FGF23 levels than those on the meat-based diet. This suggests that a vegetarian diet may have advantages in terms of preventing hyperphosphatemia.

Another measure to reduce phosphorus absorption from meat is to boil it, which reduces the phosphorus content by 50%.²⁰

Processed foods containing additives and preservatives are very high in phosphorus²¹ and should be avoided, particularly as there is no mandate to label phosphorus content in food.

PHOSPHORUS AND DIALYSIS

Although hemodialysis removes phosphorus, it does not remove enough to keep levels within normal limits. Indeed, even when patients adhere to a daily phosphorus limit of 1,000 mg, phosphorus accumulates. If 70% of the phosphorus in the diet is absorbed, this is 4,500 to 5,000 mg in a week. A 4-hour hemodialysis session will remove only 1,000 mg of phosphorus, which equals about 3,000 mg for patients undergoing dialysis 3 times a week,²² far less than phosphorus absorption.

In patients on continuous ambulatory peritoneal dialysis, a daily regimen of 4 exchanges of 2 L per exchange removes about 200 mg of phosphorus per day. In a 2012 study, patients on nocturnal dialysis or home dialysis involving longer session length had greater lowering of phosphorus levels than patients undergoing routine hemodialysis.²³

Hence, phosphorus binders are often necessary in patients on routine hemodialysis or peritoneal dialysis.

PHOSPHORUS BINDERS

Phosphorus binders reduce serum phosphorus levels by binding with ingested phosphorus in the gastrointestinal tract and forming insoluble complexes that are not absorbed. For this reason they are much more effective when taken with meals. Phosphorus binders come in different formulations: pills, capsules, chewable tablets, liquids, and even powders that can be sprinkled on food.

The potency of each binder is quantified by its “phosphorus binder equivalent dose,” ie, its binding capacity compared with that of calcium carbonate as a reference.²⁴

Phosphorus binders are broadly divided into those that contain calcium and those that do not.

Calcium-containing binders

The 2 most commonly used preparations are calcium carbonate (eg, Tums) and calcium acetate (eg, Phoslo). While these are relatively safe, some studies suggest that their use can lead to accelerated vascular calcification.²⁵

According to KDIGO,²⁶ calcium-containing binders should be avoided in hypercalcemia and adynamic bone disease. Additionally, the daily elemental calcium intake from binders should be limited to 1,500 mg, with a total daily intake that does not exceed 2,000 mg.

The elemental calcium content of calcium carbonate is about 40% of its weight (eg, 200 mg of elemental calcium in a 500-mg tablet of Tums), while the elemental calcium content of calcium acetate is about 25%. Therefore, a patient who needs 6 g of calcium carbonate for efficacy will be ingesting 2.4 g of elemental calcium per day, and that exceeds the recommended daily maximum. The main advantage of calcium carbonate is its low cost and easy availability. Commonly reported side effects include nausea and constipation.

A less commonly used calcium-based binder is calcium citrate (eg, Calcitrate). It should, however, be avoided in chronic kidney disease because of the risk of aluminum accumulation. Calcium citrate can enhance intestinal absorption of aluminum from dietary sources, as aluminum can form complexes with citrate.²⁷

Calcium-free binders

There are several calcium-free binders. Some are based on metals such as aluminum, magnesium, iron, and lanthanum; others, such as sevelamer, are resin-based.

Aluminum- and magnesium-based binders are generally not used long-term in kidney disease because of the toxicity associated with aluminum and magnesium accumulation. However, aluminum hydroxide has an off-label use as a phosphorus binder in the acute setting, particularly when serum phosphorus levels are above 7 mg/dL.²⁸ The dose is 300 to 600 mg 3 times daily with meals for a maximum of 4 weeks.

Sevelamer. Approved by the US Food and Drug Administration (FDA) in 1998, sevelamer acts by trapping phosphorus through ion exchange and hydrogen binding. It has the advantage of being calcium-free, which makes it particularly desirable in patients with hypercalcemia.

The Renagel in New Dialysis²⁵ and Treat-To-Goal²⁹ studies were randomized controlled trials that looked at the effects of sevelamer vs calcium-based binders on the risk of vascular calcification. The primary end points were serum phosphorus and calcium levels, while the secondary end points were coronary artery calcification on computed tomography and thoracic vertebral bone density. Both studies demonstrated a higher risk of vascular calcification with the calcium-based binders.

Another possible benefit of sevelamer is an improvement in lipid profile. Sevelamer lowers total cholesterol and low-density lipoprotein cholesterol levels without affecting high-density lipoprotein cholesterol or triglyceride levels.³⁰ This is likely due to its bile acid-binding effect.³¹ Sevelamer has also been shown to lower C-reactive protein levels.³² While the cardiovascular profile appears to be improved with the treatment, there are no convincing data to confirm that those properties translate to a proven independent survival benefit.

The Calcium Acetate Renagel Evaluation³³ was a randomized controlled study comparing sevelamer and calcium acetate. The authors attempted to control for the lipid-lowering effects of sevelamer by giving atorvastatin to all patients in both groups who had a low-density lipoprotein level greater than 70 mg/dL. The study found sevelamer to be not inferior to calcium acetate in terms of mortality and coronary calcification.

Further studies such as the Brazilian Renagel and Calcium trial³⁴ and the Dialysis Clinical Outcomes Revisited trial failed to show a significant long-term benefit of sevelamer over calcium-based binders. However, a secondary statistical analysis of the latter study showed possible benefit of sevelamer over calcium acetate among those age 65 and older.³⁵

To understand how sevelamer could affect vascular calcification, Yilmaz et al³⁶ compared the effects of sevelamer vs calcium acetate on FGF23 and fetuin A levels. Fetuin A is an important inhibitor of vascular calcification and is progressively diminished in kidney disease, leading to accelerated calcification.³⁷ Patients on sevelamer had higher levels of fetuin A than their counterparts on calcium acetate.³⁷ The authors proposed increased fetuin A levels as a mechanism for decreased vascular calcification.

In summary, some studies suggest that sevelamer may offer the advantage of decreasing vascular calcification, but the data are mixed and do not provide a solid answer. The main disadvantages of sevelamer are a high pill burden and side effects of nausea and dyspepsia.

Lanthanum, a metallic element, was approved as a phosphorus binder by the FDA in 2008. It comes as a chewable tablet and offers the advantage of requiring the patient to take fewer pills than sevelamer and calcium-based binders.

Sucroferric oxyhydroxide comes as a chewable tablet. It was approved by the FDA in 2013. Although each tablet contains 500 mg of iron, it has not been shown to improve iron markers. In terms of phosphorus-lowering ability, it has been shown to be noninferior to sevelamer.³⁹ Advantages include a significantly lower pill burden. Disadvantages include gastrointestinal side effects such as diarrhea and nausea and the drug’s high cost.

Ferric citrate was approved by the FDA in 2014, and 1 g delivers 210 mg of elemental iron. The main advantage of ferric citrate is its ability to increase iron markers. The phase 3 trial that demonstrated its efficacy as a binder showed an increase in ferritin compared with the active control.⁴⁰ The study also showed a decrease in the need to use intravenous iron and erythropoesis-stimulating agents. This was thought to be due to improved iron stores, leading to decreased erythropoietin resistance.⁴¹

The mean number of ferric citrate tablets needed to achieve the desired phosphorus-lowering effect was 8 per day, containing 1,680 mg of iron. In comparison, oral ferrous sulfate typically provides 210 mg of iron per day.⁴²

Disadvantages of ferric citrate include high pill burden, high cost, and gastrointestinal side effects such as nausea and constipation.

Chitosan binds salivary phosphorus. It can potentially be used, but it is not approved, and its efficacy in lowering serum phosphorus remains unclear.⁴³

CHOOSING THE APPROPRIATE PHOSPHORUS BINDER

The choice of phosphorus binder is based on the patient’s serum calcium level and iron stores and on the drug’s side effect profile, iron pill burden, and cost. Involving patients in the choice after discussing potential side effects, pill burden, and cost is important for shared decision-making and could play a role in improving adherence.

Phosphorus binders are a major portion of the pill burden in patients with end-stage renal disease, possibly affecting patient adherence. The cost of phosphorus binders is estimated at half a billion dollars annually, underlining the significant economic impact of phosphorus control.¹¹

Calcium-based binders should be the first choice when there is secondary hyperparathyroidism without hypercalcemia. There is no clear evidence regarding the benefit of correcting hypocalcemia, but KDIGO recommends keeping the serum calcium level within the reference range. KDIGO also recommends restricting calcium-based binders in persistent hypercalcemia, arterial calcification, and adynamic bone disease. This recommendation is largely based on expert opinion.

Noncalcium-based binders, which in theory might prevent vascular calcification, should be considered for patients with at least 1 of the following⁴⁴:

• Complicated diabetes mellitus
• Vascular or valvular calcification
• Persistent inflammation.

Noncalcium-based binders are also preferred in low bone-turnover states such as adynamic bone disease, as elevated calcium can inhibit parathyroid hormone.

However, the advantage of noncalcium-based binders regarding vascular calcification is largely theoretical and has not been proven clinically. Indeed, there are data comparing long-term outcomes of the different classes of phosphorus binders, but studies were limited by short follow-up, and individual studies have lacked power to detect statistical significance between two classes of binders on long-term outcomes. Meta-analyses have provided conflicting data, with some suggesting better outcomes with sevelamer than with calcium-based binders, and with others failing to show any difference.⁴⁵

Because iron deficiency is common in kidney disease, ferric citrate, which can improve iron markers, may be a suitable option, provided its cost is covered by insurance.

SPECIAL CIRCUMSTANCES FOR THE USE OF PHOSPHORUS BINDERS

Tumor lysis syndrome

Tumor lysis syndrome occurs when tumor cells release their contents into the bloodstream, either spontaneously or in response to therapy, leading to the characteristic findings of hyperuricemia, hyperkalemia, hyperphosphatemia, and hypocalcemia.⁴⁶ Phosphorus binders in conjunction with intravenous hydration are used to treat hyperphosphatemia, but evidence about their efficacy in this setting is limited.

Hypocalcemia in tumor lysis syndrome is usually not treated unless symptomatic, as the calcium-phosphorus product can increase, leading to calcium phosphate crystallization. When the calcium-phosphorus product is greater than 60, there is a higher risk of calcium phosphate deposition in the renal tubules that can lead to acute renal failure in tumor lysis syndrome.⁴⁷ To lower the risk of calcium phosphate crystallization, calcium-based binders should be avoided in tumor lysis syndrome.

Total parenteral nutrition

Since patients on total parenteral nutrition do not eat, phosphorus binders are considered ineffective; there are no concrete data showing that phosphorus binders are effective in these patients.⁴⁸ In patients with kidney disease, the phosphorus content in the parenteral nutrition formulation must be reduced.

Pregnancy

Data on phosphorus binders in pregnancy are limited. Calcium can cross the placenta. Calcium carbonate can be used in pregnancy, and fetal harm is not expected if calcium concentrations are within normal limits.⁴⁹ Calcium acetate, sevelamer, and lanthanum are considered pregnancy category C drugs. Patients with advanced chronic kidney disease and end-stage renal disease who become pregnant must receive specialized obstetric care for high-risk pregnancy.

FUTURE DIRECTIONS

Future therapies may target FGF23 and other inflammatory markers that are up-regulated in renal hyperparathyroidism. However, trials studying these markers are needed to provide a better understanding of their role in bone mineral and cardiovascular health and in overall long-term outcomes. Additionally, randomized controlled trials are needed to study long-term nonsurrogate outcomes such as reduction in cardiovascular disease and rates of overall mortality.

Editor’s note: This Medical Grand Rounds was presented as the 14th Annual Lawrence “Chris” Crain Memorial Lecture, a series that has been dedicated to discussing kidney disease, hypertension, and health care disparities in the African American community. In 1997, Dr. Crain became the first African American chief medical resident at Cleveland Clinic, and was a nephrology fellow in 1998–1999. Dr. Nally was his teacher and mentor.

African Americans have a greater burden of chronic kidney disease than whites. They are more than 3 times as likely as whites to develop end-stage renal disease, even after adjusting for age, disease stage, smoking, medications, and comorbidities. Why this is so has been the focus of much speculation and research.

This article reviews recent advances in the understanding of the progression of chronic kidney disease, with particular scrutiny of the disease in African Americans. Breakthroughs in genetics that help explain the greater disease burden in African Americans are also discussed, as well as implications for organ transplant screening.

ADVANCING UNDERSTANDING OF CHRONIC KIDNEY DISEASE

In the 1990s, dialysis rolls grew by 8% to 10% annually. Unfortunately, many patients first met with a nephrologist on the eve of their first dialysis treatment; there was not yet an adequate way to recognize the disease earlier and slow its progression. And disease definitions were not yet standardized, which led to inadequate metrics and hampered the ability to move disease management forward.

Standardizing definitions

The situation improved in 2002, when the National Kidney Foundation published clinical practice guidelines for chronic kidney disease that included disease definitions and staging.¹ Chronic kidney disease was defined as a structural or functional abnormality of the kidney lasting at least 3 months, as manifested by either of the following:

• Kidney damage, with or without decreased glomerular filtration rate (GFR), as defined by pathologic abnormalities or markers of kidney damage in the blood, urine, or on imaging tests

• 

  GFR less than 60 mL/min/1.73 m², with or without kidney damage.

A subsequent major advance was the recognition that not only GFR but also albuminuria was important for staging of chronic kidney disease (Figure 1).²

Developing large databases

Surveillance and monitoring of chronic kidney disease have generated large databases that enable researchers to detect trends in disease progression.

US Renal Data System. The US Renal Data System has collected and reported on data for more than 20 years from the National Health and Nutrition Examination Survey and the Centers for Medicare and Medicaid Services about chronic and end-stage kidney disease in the United States.

Cleveland Clinic database. Cleveland Clinic has developed a validated chronic kidney disease registry based on its electronic health record.³ The data include demographics (age, sex, ethnic group), comorbidities, medications, and complete laboratory data.⁴

Alberta Kidney Disease Network. This Canadian research consortium links large laboratory and demographic databases to facilitate defining patient populations, such as those with kidney disease and other comorbidities.

Kaiser Permanente Renal Registry. Kaiser Permanente of Northern California insures more than one-third of adults in the San Francisco Bay Area. The renal registry includes all adults whose kidney function is known. Data on age, sex, and racial or ethnic group are available from the health-plan databases.

DEATHS FROM KIDNEY DISEASE

The mortality rate in patients with end-stage renal disease who are on dialysis has steadily fallen over the past 20 years, from an annual rate of about 25% in 1996 to 17% in 2014, suggesting that care improved during that time. Patients with transplants have a much lower mortality rate: less than 5% annually.⁵ But these data also highlight the persistent risk faced by patients with chronic kidney disease; even those with transplants have death rates comparable to those of patients with cancer, diabetes, or heart failure.

Death rates correlate with GFR

After the publication of definitions and staging by the National Kidney Foundation in 2002, Go et al⁶ studied more than 1 million patients with chronic kidney disease from the Kaiser Permanente Renal Registry and found that the rates of cardiovascular events and death from any cause increased with decreasing estimated GFR. These findings were confirmed in a later meta-analysis, which also found that an elevated urinary albumin-to-creatinine ratio (> 1.1 mg/mmol) is an independent predictor of all-cause mortality and cardiovascular mortality.⁷

Keith et al⁸ followed nearly 28,000 patients with chronic kidney disease (with an estimated GFR of less than 90 mL/min/1.73 m²) over 5 years. Patients with stage 3 disease (moderate disease, GFR = 30–59 mL/min/1.73 m²) were 20 times more likely to die than to progress to end-stage renal disease (24.3% vs 1.2%). Even those with stage 4 disease (severe disease, GFR = 15–29 mL/min/1.73 m²) were more than twice as likely to die as to progress to dialysis (45.7% vs 19.9%).

Similar observations were made by Thompson et al¹⁰ based on the Alberta Kidney Disease Network database. They tracked cardiovascular causes of death and found that regardless of estimated GFR, cardiovascular deaths were most often attributed to ischemic heart disease (about 55%). Other trends were also apparent: as the GFR fell, the incidence of stroke decreased, and heart failure and valvular heart disease increased.

AFRICAN AMERICANS WITH KIDNEY DISEASE: A DISTINCT GROUP

African Americans constitute about 12% of the US population but account for:

• 31% of end-stage renal disease
• 34% of the kidney transplant waiting list
• 28% of kidney transplants in 2015 (12% of living donor transplants, 35% of deceased donor transplants).

In addition, African Americans with chronic kidney disease tend to be:

• Younger and have more advanced kidney disease than whites¹¹
• Much more likely than whites to have diabetes, and somewhat more likely to have hypertension

• 

  Adapted from Navaneethan SD, Schold JD, Arrigain S, Jolly SE, Nally JV Jr. Cause-specific deaths in non-dialysis-dependent CKD. J Am Soc Nephrol 2015; 26:2512–2520.

  More likely than whites to die of cardiovascular disease (37.4% vs 34.2%) (Figure 2).⁹

Overall, the prevalence of chronic kidney disease is slightly higher in African Americans than in whites. Interestingly, African Americans are slightly less likely than whites to have low estimated GFR values (6.2% vs 7.6% incidence of < 60 mL/min/1.73 m²) but are about 50% more likely to have proteinuria (12.3% vs 8.4% incidence of urine albumin-to-creatinine ratio ≥ 30 mg/g).

More likely to be on dialysis, but less likely to die

Although African Americans have only a slightly higher prevalence of chronic kidney disease (about 15% increased prevalence) than whites,¹² they are 3 times more likely to be on dialysis.

Nevertheless, for unknown reasons, African American adults on dialysis have about a 26% lower all-cause mortality rate than whites.⁵ One proposed explanation for this survival advantage has been that the mortality rate in African Americans with chronic kidney disease before entering dialysis is higher than in whites, leading to a “healthier population” on dialysis.¹³ However, this theory is based on a small study from more than a decade ago and has not been borne out by subsequent investigation.

African Americans with chronic kidney disease: Death rates not increased

African Americans over age 65 with chronic kidney disease have all-cause mortality rates similar to those of whites: about 11% annually. Breaking it down by disease severity, death rates in stage 3 disease are about 10% and jump to more than 15% in higher stages in both African Americans and whites.⁵

However, African Americans with chronic kidney disease have more heart disease and much more end-stage renal disease than whites.

Disease advances faster despite care

The incidence of end-stage renal disease is consistently more than 3 times higher in African Americans than in whites in the United States.^5,14

Multiple investigations have tried to determine why African Americans are disproportionately affected by progression of chronic kidney disease to end-stage renal disease. We recently examined this question in our Cleveland Clinic registry data. Even after adjusting for 17 variables (including demographics, comorbidities, insurance, medications, smoking, and chronic kidney disease stage), African Americans with chronic kidney disease were found to have an increased risk of progressing to end-stage renal disease compared with whites (subhazard ratio 1.38, 95% confidence interval 1.19–1.60).

We examined care measures from the Cleveland Clinic database. In terms of the number of laboratory tests ordered, clinic visits, and nephrology referrals, African Americans had at least as much care as whites, if not more. Similarly, African Americans’ access to renoprotective medicines (angiotensin-converting enzyme inhibitors, angiotensin II receptor blockers, statins, beta-blockers) was the same as or more than for whites.

Although the frequently attributed reasons surrounding compliance and socioeconomic issues are worthy of examination, they do not appear to completely explain the differences in incidence and outcomes. This dichotomy of a marginally increased prevalence of chronic kidney disease in African Americans with mortality rates similar to those of whites, yet with a 3 times higher incidence of end-stage renal disease in African Americans, suggests a faster progression of the disease in African Americans, which may be genetically based.

In 2010, two variant alleles of the APOL1 gene on chromosome 22 were found to be associated with nondiabetic kidney disease.¹⁵ Three nephropathies are associated with being homozygous for these alleles:

• Focal segmental glomerulosclerosis, the leading cause of nephrotic syndrome in African Americans
• Hypertension-associated kidney disease with scarring of glomeruli in vessels, the primary cause of end-stage renal disease in African Americans
• Human immunodeficiency virus (HIV)- associated nephropathy, usually a focal segmental glomerulosclerosis type of lesion.

The first two conditions are about 3 to 5 times more prevalent in African Americans than in whites, and HIV-associated nephropathy is about 20 to 30 times more common. 

African sleeping sickness and chronic kidney disease

Retrospective analysis of biologic samples from trials of kidney disease in African Americans has revealed interesting results.

From Parsa A, Kao WH, Xie D, et al; AASK Study Investigators; CRIC Study Investigators. APOL1 risk variants, race, and progression of chronic kidney disease. N Engl J Med 2013; 369:2183–2196. Reprinted with permission from Massachusetts Medical Society.

The Chronic Renal Insufficiency Cohort (CRIC) observation study¹⁸ enrolled patients with an estimated GFR of 20 to 70 mL/min/1.73 m², with a preference for African Americans and patients with diabetes. Nearly 3,000 participants had adequate samples for DNA testing. They found that African Americans with the double variant allele had worse outcomes, whether or not they had diabetes, compared with whites and African Americans without the homozygous gene variant.

Mechanism not well understood

The mechanism of renal injury is not well understood. Apolipoprotein L1, the protein coded for by APOL1, is a component of high-density lipoprotein. It is found in a different distribution pattern in people with normal kidneys vs those with nondiabetic kidney disease, especially in the arteries, arterioles, and podocytes.^19,20 It can be detected in blood plasma, but levels do not correlate with kidney disease.²¹ Not all patients with the high-risk variant develop chronic kidney disease; a “second hit” such as infection with HIV may be required.

Investigators have recently developed knockout mouse models of APOL1-associated kidney diseases that are helping to elucidate mechanisms.^22,23

EFFECT OF GENOTYPE ON KIDNEY TRANSPLANTS IN AFRICAN AMERICANS

African Americans receive about 30% of kidney transplants in the United States and represent about 15% to 20% of all donors.

Lee et al²⁴ reviewed 119 African American recipients of kidney transplants, about half of whom were homozygous for an APOL1 variant. After 5 years, no differences were found in allograft survival between recipients with 0, 1, or 2 risk alleles.

However, looking at the issue from the other side, Reeves-Daniel et al²⁵ studied the fate of more than 100 kidneys that were transplanted from African American donors, 16% of whom had the high-risk, homozygous genotype. In this case, graft failure was much likelier to occur with the high-risk donor kidneys (hazard ratio 3.84, P = .008). Similar outcomes were shown in a study of 2 centers²⁶ involving 675 transplants from deceased donors, 15% of which involved the high-risk genotype. The hazard ratio for graft failure was found to be 2.26 (P = .001) with high-risk donor kidneys.

These studies, which examined data from about 5 years after transplant, found that kidney failure does not tend to occur immediately in all cases, but gradually over time. Most high-risk kidneys were not lost within the 5 years of the studies.

The fact that the high-risk kidneys do not all fail immediately also suggests that a second hit is required for failure. Culprits postulated include a bacterial or viral infection (eg, BK virus, cytomegalovirus), ischemia or reperfusion injury, drug toxicity, and immune-mediated allograft injury (ie, rejection). 

Genetic testing for APOL1 risk variants is on the horizon for kidney transplant. But at this point, providing guidance for patients can be tricky. Two case studies^27,28 and epidemiologic data suggest that donors homozygous for an APOL1 variant and those with a family history of end-stage kidney disease are at increased risk of chronic kidney disease. Even so, most recipients even of these high-risk organs have good outcomes. If an African American patient needs a kidney and his or her sibling offers one, it is difficult to advise against it when the evidence is weak for immediate risk and when other options may not be readily available. Further investigation is clearly needed into whether APOL1 variants and other biomarkers can predict an organ’s success as a transplant.

The National Institutes of Health are currently funding prospective longitudinal studies with the APOL1 Long-term Kidney Transplantation Outcomes Network (APOLLO) to determine the impact of APOL1 genetic factors on transplant recipients as well as on living donors. Possible second hits will also be studied, as will other markers of renal dysfunction or disease in donors. Researchers are actively investigating these important questions.

KEEPING SCIENCE RELEVANT

In a recent commentary related to the murine knockout model of APOL1-associated kidney disease, O’Toole et al offered insightful observations regarding the potential clinical impact of these new genetic discoveries.²³

As we study the genetics of kidney disease in African American patients, we should keep in mind 3 critical questions of clinical importance:

Will findings identify better treatments for chronic kidney disease? The AASK trial found that knowing the genetics did not affect outcomes of routine therapy. However, basic science investigations are currently underway targeting APOL1 variants which might reduce the increased kidney disease risk among people of African descent.

Should patients be genotyped for APOL1 risk variants? For patients with chronic kidney disease, it does not seem useful at this time. But for renal transplant donors, the answer is probably yes.

How does this discovery help us to understand our patients better? The implications are enormous for combatting the assumptions that rapid chronic kidney disease progression reflects poor patient compliance or other socioeconomic factors. We now understand that genetics, at least in part, drives renal disease outcomes in African American patients.

Editor’s note: This Medical Grand Rounds was presented as the 14th Annual Lawrence “Chris” Crain Memorial Lecture, a series that has been dedicated to discussing kidney disease, hypertension, and health care disparities in the African American community. In 1997, Dr. Crain became the first African American chief medical resident at Cleveland Clinic, and was a nephrology fellow in 1998–1999. Dr. Nally was his teacher and mentor.

African Americans have a greater burden of chronic kidney disease than whites. They are more than 3 times as likely as whites to develop end-stage renal disease, even after adjusting for age, disease stage, smoking, medications, and comorbidities. Why this is so has been the focus of much speculation and research.

This article reviews recent advances in the understanding of the progression of chronic kidney disease, with particular scrutiny of the disease in African Americans. Breakthroughs in genetics that help explain the greater disease burden in African Americans are also discussed, as well as implications for organ transplant screening.

ADVANCING UNDERSTANDING OF CHRONIC KIDNEY DISEASE

In the 1990s, dialysis rolls grew by 8% to 10% annually. Unfortunately, many patients first met with a nephrologist on the eve of their first dialysis treatment; there was not yet an adequate way to recognize the disease earlier and slow its progression. And disease definitions were not yet standardized, which led to inadequate metrics and hampered the ability to move disease management forward.

Standardizing definitions

The situation improved in 2002, when the National Kidney Foundation published clinical practice guidelines for chronic kidney disease that included disease definitions and staging.¹ Chronic kidney disease was defined as a structural or functional abnormality of the kidney lasting at least 3 months, as manifested by either of the following:

• Kidney damage, with or without decreased glomerular filtration rate (GFR), as defined by pathologic abnormalities or markers of kidney damage in the blood, urine, or on imaging tests

• 

  GFR less than 60 mL/min/1.73 m², with or without kidney damage.

A subsequent major advance was the recognition that not only GFR but also albuminuria was important for staging of chronic kidney disease (Figure 1).²

Developing large databases

Surveillance and monitoring of chronic kidney disease have generated large databases that enable researchers to detect trends in disease progression.

US Renal Data System. The US Renal Data System has collected and reported on data for more than 20 years from the National Health and Nutrition Examination Survey and the Centers for Medicare and Medicaid Services about chronic and end-stage kidney disease in the United States.

Cleveland Clinic database. Cleveland Clinic has developed a validated chronic kidney disease registry based on its electronic health record.³ The data include demographics (age, sex, ethnic group), comorbidities, medications, and complete laboratory data.⁴

Alberta Kidney Disease Network. This Canadian research consortium links large laboratory and demographic databases to facilitate defining patient populations, such as those with kidney disease and other comorbidities.

Kaiser Permanente Renal Registry. Kaiser Permanente of Northern California insures more than one-third of adults in the San Francisco Bay Area. The renal registry includes all adults whose kidney function is known. Data on age, sex, and racial or ethnic group are available from the health-plan databases.

DEATHS FROM KIDNEY DISEASE

The mortality rate in patients with end-stage renal disease who are on dialysis has steadily fallen over the past 20 years, from an annual rate of about 25% in 1996 to 17% in 2014, suggesting that care improved during that time. Patients with transplants have a much lower mortality rate: less than 5% annually.⁵ But these data also highlight the persistent risk faced by patients with chronic kidney disease; even those with transplants have death rates comparable to those of patients with cancer, diabetes, or heart failure.

Death rates correlate with GFR

After the publication of definitions and staging by the National Kidney Foundation in 2002, Go et al⁶ studied more than 1 million patients with chronic kidney disease from the Kaiser Permanente Renal Registry and found that the rates of cardiovascular events and death from any cause increased with decreasing estimated GFR. These findings were confirmed in a later meta-analysis, which also found that an elevated urinary albumin-to-creatinine ratio (> 1.1 mg/mmol) is an independent predictor of all-cause mortality and cardiovascular mortality.⁷

Keith et al⁸ followed nearly 28,000 patients with chronic kidney disease (with an estimated GFR of less than 90 mL/min/1.73 m²) over 5 years. Patients with stage 3 disease (moderate disease, GFR = 30–59 mL/min/1.73 m²) were 20 times more likely to die than to progress to end-stage renal disease (24.3% vs 1.2%). Even those with stage 4 disease (severe disease, GFR = 15–29 mL/min/1.73 m²) were more than twice as likely to die as to progress to dialysis (45.7% vs 19.9%).

Similar observations were made by Thompson et al¹⁰ based on the Alberta Kidney Disease Network database. They tracked cardiovascular causes of death and found that regardless of estimated GFR, cardiovascular deaths were most often attributed to ischemic heart disease (about 55%). Other trends were also apparent: as the GFR fell, the incidence of stroke decreased, and heart failure and valvular heart disease increased.

AFRICAN AMERICANS WITH KIDNEY DISEASE: A DISTINCT GROUP

African Americans constitute about 12% of the US population but account for:

• 31% of end-stage renal disease
• 34% of the kidney transplant waiting list
• 28% of kidney transplants in 2015 (12% of living donor transplants, 35% of deceased donor transplants).

In addition, African Americans with chronic kidney disease tend to be:

• Younger and have more advanced kidney disease than whites¹¹
• Much more likely than whites to have diabetes, and somewhat more likely to have hypertension

• 

  Adapted from Navaneethan SD, Schold JD, Arrigain S, Jolly SE, Nally JV Jr. Cause-specific deaths in non-dialysis-dependent CKD. J Am Soc Nephrol 2015; 26:2512–2520.

  More likely than whites to die of cardiovascular disease (37.4% vs 34.2%) (Figure 2).⁹

Overall, the prevalence of chronic kidney disease is slightly higher in African Americans than in whites. Interestingly, African Americans are slightly less likely than whites to have low estimated GFR values (6.2% vs 7.6% incidence of < 60 mL/min/1.73 m²) but are about 50% more likely to have proteinuria (12.3% vs 8.4% incidence of urine albumin-to-creatinine ratio ≥ 30 mg/g).

More likely to be on dialysis, but less likely to die

Although African Americans have only a slightly higher prevalence of chronic kidney disease (about 15% increased prevalence) than whites,¹² they are 3 times more likely to be on dialysis.

Nevertheless, for unknown reasons, African American adults on dialysis have about a 26% lower all-cause mortality rate than whites.⁵ One proposed explanation for this survival advantage has been that the mortality rate in African Americans with chronic kidney disease before entering dialysis is higher than in whites, leading to a “healthier population” on dialysis.¹³ However, this theory is based on a small study from more than a decade ago and has not been borne out by subsequent investigation.

African Americans with chronic kidney disease: Death rates not increased

African Americans over age 65 with chronic kidney disease have all-cause mortality rates similar to those of whites: about 11% annually. Breaking it down by disease severity, death rates in stage 3 disease are about 10% and jump to more than 15% in higher stages in both African Americans and whites.⁵

However, African Americans with chronic kidney disease have more heart disease and much more end-stage renal disease than whites.

Disease advances faster despite care

The incidence of end-stage renal disease is consistently more than 3 times higher in African Americans than in whites in the United States.^5,14

Multiple investigations have tried to determine why African Americans are disproportionately affected by progression of chronic kidney disease to end-stage renal disease. We recently examined this question in our Cleveland Clinic registry data. Even after adjusting for 17 variables (including demographics, comorbidities, insurance, medications, smoking, and chronic kidney disease stage), African Americans with chronic kidney disease were found to have an increased risk of progressing to end-stage renal disease compared with whites (subhazard ratio 1.38, 95% confidence interval 1.19–1.60).

We examined care measures from the Cleveland Clinic database. In terms of the number of laboratory tests ordered, clinic visits, and nephrology referrals, African Americans had at least as much care as whites, if not more. Similarly, African Americans’ access to renoprotective medicines (angiotensin-converting enzyme inhibitors, angiotensin II receptor blockers, statins, beta-blockers) was the same as or more than for whites.

Although the frequently attributed reasons surrounding compliance and socioeconomic issues are worthy of examination, they do not appear to completely explain the differences in incidence and outcomes. This dichotomy of a marginally increased prevalence of chronic kidney disease in African Americans with mortality rates similar to those of whites, yet with a 3 times higher incidence of end-stage renal disease in African Americans, suggests a faster progression of the disease in African Americans, which may be genetically based.

In 2010, two variant alleles of the APOL1 gene on chromosome 22 were found to be associated with nondiabetic kidney disease.¹⁵ Three nephropathies are associated with being homozygous for these alleles:

• Focal segmental glomerulosclerosis, the leading cause of nephrotic syndrome in African Americans
• Hypertension-associated kidney disease with scarring of glomeruli in vessels, the primary cause of end-stage renal disease in African Americans
• Human immunodeficiency virus (HIV)- associated nephropathy, usually a focal segmental glomerulosclerosis type of lesion.

The first two conditions are about 3 to 5 times more prevalent in African Americans than in whites, and HIV-associated nephropathy is about 20 to 30 times more common. 

African sleeping sickness and chronic kidney disease

Retrospective analysis of biologic samples from trials of kidney disease in African Americans has revealed interesting results.

From Parsa A, Kao WH, Xie D, et al; AASK Study Investigators; CRIC Study Investigators. APOL1 risk variants, race, and progression of chronic kidney disease. N Engl J Med 2013; 369:2183–2196. Reprinted with permission from Massachusetts Medical Society.

The Chronic Renal Insufficiency Cohort (CRIC) observation study¹⁸ enrolled patients with an estimated GFR of 20 to 70 mL/min/1.73 m², with a preference for African Americans and patients with diabetes. Nearly 3,000 participants had adequate samples for DNA testing. They found that African Americans with the double variant allele had worse outcomes, whether or not they had diabetes, compared with whites and African Americans without the homozygous gene variant.

Mechanism not well understood

The mechanism of renal injury is not well understood. Apolipoprotein L1, the protein coded for by APOL1, is a component of high-density lipoprotein. It is found in a different distribution pattern in people with normal kidneys vs those with nondiabetic kidney disease, especially in the arteries, arterioles, and podocytes.^19,20 It can be detected in blood plasma, but levels do not correlate with kidney disease.²¹ Not all patients with the high-risk variant develop chronic kidney disease; a “second hit” such as infection with HIV may be required.

Investigators have recently developed knockout mouse models of APOL1-associated kidney diseases that are helping to elucidate mechanisms.^22,23

EFFECT OF GENOTYPE ON KIDNEY TRANSPLANTS IN AFRICAN AMERICANS

African Americans receive about 30% of kidney transplants in the United States and represent about 15% to 20% of all donors.

Lee et al²⁴ reviewed 119 African American recipients of kidney transplants, about half of whom were homozygous for an APOL1 variant. After 5 years, no differences were found in allograft survival between recipients with 0, 1, or 2 risk alleles.

However, looking at the issue from the other side, Reeves-Daniel et al²⁵ studied the fate of more than 100 kidneys that were transplanted from African American donors, 16% of whom had the high-risk, homozygous genotype. In this case, graft failure was much likelier to occur with the high-risk donor kidneys (hazard ratio 3.84, P = .008). Similar outcomes were shown in a study of 2 centers²⁶ involving 675 transplants from deceased donors, 15% of which involved the high-risk genotype. The hazard ratio for graft failure was found to be 2.26 (P = .001) with high-risk donor kidneys.

These studies, which examined data from about 5 years after transplant, found that kidney failure does not tend to occur immediately in all cases, but gradually over time. Most high-risk kidneys were not lost within the 5 years of the studies.

The fact that the high-risk kidneys do not all fail immediately also suggests that a second hit is required for failure. Culprits postulated include a bacterial or viral infection (eg, BK virus, cytomegalovirus), ischemia or reperfusion injury, drug toxicity, and immune-mediated allograft injury (ie, rejection). 

Genetic testing for APOL1 risk variants is on the horizon for kidney transplant. But at this point, providing guidance for patients can be tricky. Two case studies^27,28 and epidemiologic data suggest that donors homozygous for an APOL1 variant and those with a family history of end-stage kidney disease are at increased risk of chronic kidney disease. Even so, most recipients even of these high-risk organs have good outcomes. If an African American patient needs a kidney and his or her sibling offers one, it is difficult to advise against it when the evidence is weak for immediate risk and when other options may not be readily available. Further investigation is clearly needed into whether APOL1 variants and other biomarkers can predict an organ’s success as a transplant.

The National Institutes of Health are currently funding prospective longitudinal studies with the APOL1 Long-term Kidney Transplantation Outcomes Network (APOLLO) to determine the impact of APOL1 genetic factors on transplant recipients as well as on living donors. Possible second hits will also be studied, as will other markers of renal dysfunction or disease in donors. Researchers are actively investigating these important questions.

KEEPING SCIENCE RELEVANT

In a recent commentary related to the murine knockout model of APOL1-associated kidney disease, O’Toole et al offered insightful observations regarding the potential clinical impact of these new genetic discoveries.²³

As we study the genetics of kidney disease in African American patients, we should keep in mind 3 critical questions of clinical importance:

Will findings identify better treatments for chronic kidney disease? The AASK trial found that knowing the genetics did not affect outcomes of routine therapy. However, basic science investigations are currently underway targeting APOL1 variants which might reduce the increased kidney disease risk among people of African descent.

Should patients be genotyped for APOL1 risk variants? For patients with chronic kidney disease, it does not seem useful at this time. But for renal transplant donors, the answer is probably yes.

How does this discovery help us to understand our patients better? The implications are enormous for combatting the assumptions that rapid chronic kidney disease progression reflects poor patient compliance or other socioeconomic factors. We now understand that genetics, at least in part, drives renal disease outcomes in African American patients.

Editor’s note: This Medical Grand Rounds was presented as the 14th Annual Lawrence “Chris” Crain Memorial Lecture, a series that has been dedicated to discussing kidney disease, hypertension, and health care disparities in the African American community. In 1997, Dr. Crain became the first African American chief medical resident at Cleveland Clinic, and was a nephrology fellow in 1998–1999. Dr. Nally was his teacher and mentor.

African Americans have a greater burden of chronic kidney disease than whites. They are more than 3 times as likely as whites to develop end-stage renal disease, even after adjusting for age, disease stage, smoking, medications, and comorbidities. Why this is so has been the focus of much speculation and research.

This article reviews recent advances in the understanding of the progression of chronic kidney disease, with particular scrutiny of the disease in African Americans. Breakthroughs in genetics that help explain the greater disease burden in African Americans are also discussed, as well as implications for organ transplant screening.

ADVANCING UNDERSTANDING OF CHRONIC KIDNEY DISEASE

In the 1990s, dialysis rolls grew by 8% to 10% annually. Unfortunately, many patients first met with a nephrologist on the eve of their first dialysis treatment; there was not yet an adequate way to recognize the disease earlier and slow its progression. And disease definitions were not yet standardized, which led to inadequate metrics and hampered the ability to move disease management forward.

Standardizing definitions

The situation improved in 2002, when the National Kidney Foundation published clinical practice guidelines for chronic kidney disease that included disease definitions and staging.¹ Chronic kidney disease was defined as a structural or functional abnormality of the kidney lasting at least 3 months, as manifested by either of the following:

• Kidney damage, with or without decreased glomerular filtration rate (GFR), as defined by pathologic abnormalities or markers of kidney damage in the blood, urine, or on imaging tests

• 

  GFR less than 60 mL/min/1.73 m², with or without kidney damage.

A subsequent major advance was the recognition that not only GFR but also albuminuria was important for staging of chronic kidney disease (Figure 1).²

Developing large databases

Surveillance and monitoring of chronic kidney disease have generated large databases that enable researchers to detect trends in disease progression.

US Renal Data System. The US Renal Data System has collected and reported on data for more than 20 years from the National Health and Nutrition Examination Survey and the Centers for Medicare and Medicaid Services about chronic and end-stage kidney disease in the United States.

Cleveland Clinic database. Cleveland Clinic has developed a validated chronic kidney disease registry based on its electronic health record.³ The data include demographics (age, sex, ethnic group), comorbidities, medications, and complete laboratory data.⁴

Alberta Kidney Disease Network. This Canadian research consortium links large laboratory and demographic databases to facilitate defining patient populations, such as those with kidney disease and other comorbidities.

Kaiser Permanente Renal Registry. Kaiser Permanente of Northern California insures more than one-third of adults in the San Francisco Bay Area. The renal registry includes all adults whose kidney function is known. Data on age, sex, and racial or ethnic group are available from the health-plan databases.

DEATHS FROM KIDNEY DISEASE

The mortality rate in patients with end-stage renal disease who are on dialysis has steadily fallen over the past 20 years, from an annual rate of about 25% in 1996 to 17% in 2014, suggesting that care improved during that time. Patients with transplants have a much lower mortality rate: less than 5% annually.⁵ But these data also highlight the persistent risk faced by patients with chronic kidney disease; even those with transplants have death rates comparable to those of patients with cancer, diabetes, or heart failure.

Death rates correlate with GFR

After the publication of definitions and staging by the National Kidney Foundation in 2002, Go et al⁶ studied more than 1 million patients with chronic kidney disease from the Kaiser Permanente Renal Registry and found that the rates of cardiovascular events and death from any cause increased with decreasing estimated GFR. These findings were confirmed in a later meta-analysis, which also found that an elevated urinary albumin-to-creatinine ratio (> 1.1 mg/mmol) is an independent predictor of all-cause mortality and cardiovascular mortality.⁷

Keith et al⁸ followed nearly 28,000 patients with chronic kidney disease (with an estimated GFR of less than 90 mL/min/1.73 m²) over 5 years. Patients with stage 3 disease (moderate disease, GFR = 30–59 mL/min/1.73 m²) were 20 times more likely to die than to progress to end-stage renal disease (24.3% vs 1.2%). Even those with stage 4 disease (severe disease, GFR = 15–29 mL/min/1.73 m²) were more than twice as likely to die as to progress to dialysis (45.7% vs 19.9%).

Similar observations were made by Thompson et al¹⁰ based on the Alberta Kidney Disease Network database. They tracked cardiovascular causes of death and found that regardless of estimated GFR, cardiovascular deaths were most often attributed to ischemic heart disease (about 55%). Other trends were also apparent: as the GFR fell, the incidence of stroke decreased, and heart failure and valvular heart disease increased.

AFRICAN AMERICANS WITH KIDNEY DISEASE: A DISTINCT GROUP

African Americans constitute about 12% of the US population but account for:

• 31% of end-stage renal disease
• 34% of the kidney transplant waiting list
• 28% of kidney transplants in 2015 (12% of living donor transplants, 35% of deceased donor transplants).

In addition, African Americans with chronic kidney disease tend to be:

• Younger and have more advanced kidney disease than whites¹¹
• Much more likely than whites to have diabetes, and somewhat more likely to have hypertension

• 

  Adapted from Navaneethan SD, Schold JD, Arrigain S, Jolly SE, Nally JV Jr. Cause-specific deaths in non-dialysis-dependent CKD. J Am Soc Nephrol 2015; 26:2512–2520.

  More likely than whites to die of cardiovascular disease (37.4% vs 34.2%) (Figure 2).⁹

Overall, the prevalence of chronic kidney disease is slightly higher in African Americans than in whites. Interestingly, African Americans are slightly less likely than whites to have low estimated GFR values (6.2% vs 7.6% incidence of < 60 mL/min/1.73 m²) but are about 50% more likely to have proteinuria (12.3% vs 8.4% incidence of urine albumin-to-creatinine ratio ≥ 30 mg/g).

More likely to be on dialysis, but less likely to die

Although African Americans have only a slightly higher prevalence of chronic kidney disease (about 15% increased prevalence) than whites,¹² they are 3 times more likely to be on dialysis.

Nevertheless, for unknown reasons, African American adults on dialysis have about a 26% lower all-cause mortality rate than whites.⁵ One proposed explanation for this survival advantage has been that the mortality rate in African Americans with chronic kidney disease before entering dialysis is higher than in whites, leading to a “healthier population” on dialysis.¹³ However, this theory is based on a small study from more than a decade ago and has not been borne out by subsequent investigation.

African Americans with chronic kidney disease: Death rates not increased

African Americans over age 65 with chronic kidney disease have all-cause mortality rates similar to those of whites: about 11% annually. Breaking it down by disease severity, death rates in stage 3 disease are about 10% and jump to more than 15% in higher stages in both African Americans and whites.⁵

However, African Americans with chronic kidney disease have more heart disease and much more end-stage renal disease than whites.

Disease advances faster despite care

The incidence of end-stage renal disease is consistently more than 3 times higher in African Americans than in whites in the United States.^5,14

Multiple investigations have tried to determine why African Americans are disproportionately affected by progression of chronic kidney disease to end-stage renal disease. We recently examined this question in our Cleveland Clinic registry data. Even after adjusting for 17 variables (including demographics, comorbidities, insurance, medications, smoking, and chronic kidney disease stage), African Americans with chronic kidney disease were found to have an increased risk of progressing to end-stage renal disease compared with whites (subhazard ratio 1.38, 95% confidence interval 1.19–1.60).

We examined care measures from the Cleveland Clinic database. In terms of the number of laboratory tests ordered, clinic visits, and nephrology referrals, African Americans had at least as much care as whites, if not more. Similarly, African Americans’ access to renoprotective medicines (angiotensin-converting enzyme inhibitors, angiotensin II receptor blockers, statins, beta-blockers) was the same as or more than for whites.

Although the frequently attributed reasons surrounding compliance and socioeconomic issues are worthy of examination, they do not appear to completely explain the differences in incidence and outcomes. This dichotomy of a marginally increased prevalence of chronic kidney disease in African Americans with mortality rates similar to those of whites, yet with a 3 times higher incidence of end-stage renal disease in African Americans, suggests a faster progression of the disease in African Americans, which may be genetically based.

In 2010, two variant alleles of the APOL1 gene on chromosome 22 were found to be associated with nondiabetic kidney disease.¹⁵ Three nephropathies are associated with being homozygous for these alleles:

• Focal segmental glomerulosclerosis, the leading cause of nephrotic syndrome in African Americans
• Hypertension-associated kidney disease with scarring of glomeruli in vessels, the primary cause of end-stage renal disease in African Americans
• Human immunodeficiency virus (HIV)- associated nephropathy, usually a focal segmental glomerulosclerosis type of lesion.

The first two conditions are about 3 to 5 times more prevalent in African Americans than in whites, and HIV-associated nephropathy is about 20 to 30 times more common. 

African sleeping sickness and chronic kidney disease

Retrospective analysis of biologic samples from trials of kidney disease in African Americans has revealed interesting results.

From Parsa A, Kao WH, Xie D, et al; AASK Study Investigators; CRIC Study Investigators. APOL1 risk variants, race, and progression of chronic kidney disease. N Engl J Med 2013; 369:2183–2196. Reprinted with permission from Massachusetts Medical Society.

The Chronic Renal Insufficiency Cohort (CRIC) observation study¹⁸ enrolled patients with an estimated GFR of 20 to 70 mL/min/1.73 m², with a preference for African Americans and patients with diabetes. Nearly 3,000 participants had adequate samples for DNA testing. They found that African Americans with the double variant allele had worse outcomes, whether or not they had diabetes, compared with whites and African Americans without the homozygous gene variant.

Mechanism not well understood

The mechanism of renal injury is not well understood. Apolipoprotein L1, the protein coded for by APOL1, is a component of high-density lipoprotein. It is found in a different distribution pattern in people with normal kidneys vs those with nondiabetic kidney disease, especially in the arteries, arterioles, and podocytes.^19,20 It can be detected in blood plasma, but levels do not correlate with kidney disease.²¹ Not all patients with the high-risk variant develop chronic kidney disease; a “second hit” such as infection with HIV may be required.

Investigators have recently developed knockout mouse models of APOL1-associated kidney diseases that are helping to elucidate mechanisms.^22,23

EFFECT OF GENOTYPE ON KIDNEY TRANSPLANTS IN AFRICAN AMERICANS

African Americans receive about 30% of kidney transplants in the United States and represent about 15% to 20% of all donors.

Lee et al²⁴ reviewed 119 African American recipients of kidney transplants, about half of whom were homozygous for an APOL1 variant. After 5 years, no differences were found in allograft survival between recipients with 0, 1, or 2 risk alleles.

However, looking at the issue from the other side, Reeves-Daniel et al²⁵ studied the fate of more than 100 kidneys that were transplanted from African American donors, 16% of whom had the high-risk, homozygous genotype. In this case, graft failure was much likelier to occur with the high-risk donor kidneys (hazard ratio 3.84, P = .008). Similar outcomes were shown in a study of 2 centers²⁶ involving 675 transplants from deceased donors, 15% of which involved the high-risk genotype. The hazard ratio for graft failure was found to be 2.26 (P = .001) with high-risk donor kidneys.

These studies, which examined data from about 5 years after transplant, found that kidney failure does not tend to occur immediately in all cases, but gradually over time. Most high-risk kidneys were not lost within the 5 years of the studies.

The fact that the high-risk kidneys do not all fail immediately also suggests that a second hit is required for failure. Culprits postulated include a bacterial or viral infection (eg, BK virus, cytomegalovirus), ischemia or reperfusion injury, drug toxicity, and immune-mediated allograft injury (ie, rejection). 

Genetic testing for APOL1 risk variants is on the horizon for kidney transplant. But at this point, providing guidance for patients can be tricky. Two case studies^27,28 and epidemiologic data suggest that donors homozygous for an APOL1 variant and those with a family history of end-stage kidney disease are at increased risk of chronic kidney disease. Even so, most recipients even of these high-risk organs have good outcomes. If an African American patient needs a kidney and his or her sibling offers one, it is difficult to advise against it when the evidence is weak for immediate risk and when other options may not be readily available. Further investigation is clearly needed into whether APOL1 variants and other biomarkers can predict an organ’s success as a transplant.

The National Institutes of Health are currently funding prospective longitudinal studies with the APOL1 Long-term Kidney Transplantation Outcomes Network (APOLLO) to determine the impact of APOL1 genetic factors on transplant recipients as well as on living donors. Possible second hits will also be studied, as will other markers of renal dysfunction or disease in donors. Researchers are actively investigating these important questions.

KEEPING SCIENCE RELEVANT

In a recent commentary related to the murine knockout model of APOL1-associated kidney disease, O’Toole et al offered insightful observations regarding the potential clinical impact of these new genetic discoveries.²³

As we study the genetics of kidney disease in African American patients, we should keep in mind 3 critical questions of clinical importance:

Will findings identify better treatments for chronic kidney disease? The AASK trial found that knowing the genetics did not affect outcomes of routine therapy. However, basic science investigations are currently underway targeting APOL1 variants which might reduce the increased kidney disease risk among people of African descent.

Should patients be genotyped for APOL1 risk variants? For patients with chronic kidney disease, it does not seem useful at this time. But for renal transplant donors, the answer is probably yes.

How does this discovery help us to understand our patients better? The implications are enormous for combatting the assumptions that rapid chronic kidney disease progression reflects poor patient compliance or other socioeconomic factors. We now understand that genetics, at least in part, drives renal disease outcomes in African American patients.

Can a percutaneous catheter-based procedure effectively treat resistant hypertension?

Radiofrequency ablation of the renal sympathetic nerves is undergoing randomized controlled trials in patients who have resistant hypertension and other disorders that involve the sympathetic nervous system. Remarkably, the limited results available so far look good.

See related editorial

This article discusses the physiologic rationale for renal denervation, the evidence from studies in humans of the benefits, risks, and complications of the procedure, upcoming trials, and areas for future research.

DESPITE MANY TREATMENT OPTIONS, RESISTANT HYPERTENSION IS COMMON

Hypertension is a leading reason for visits to physicians in the United States and is associated with increased rates of cardiovascular disease and death.^1,2 A variety of antihypertensive agents are available, and the percentage of people with hypertension whose blood pressure is under control has increased over the past 2 decades. Nevertheless, population-based studies show that the control rate remains suboptimal.³ Effective pharmacologic treatment may be limited by inadequate doses or inappropriate combinations of antihypertensive drugs, concurrent use of agents that raise the blood pressure, noncompliance with dietary restrictions, and side effects that result in poor compliance with drug therapy.

Resistant hypertension is defined as failure to achieve goal blood pressure in patients who are adhering to full tolerated doses of an appropriate three-drug regimen that includes a diuretic.^1,4,5 If we use these criteria, many patients labelled as having resistant hypertension probably do not truly have it; instead, they are nonadherent to therapy or are on an inadequate or inappropriate regimen. Although the true prevalence of resistant hypertension is not clear, estimates from large clinical trials suggest that about 20% to 30% of hypertensive patients may meet the criteria for it.⁴ For the subset of patients who have truly resistant hypertension, nonpharmacologic treatments such as renal sympathetic denervation are an intriguing avenue.

SURGICAL SYMPATHETIC DENERVATION: TRIED AND ABANDONED IN THE 1950s

More than a half century ago, a surgical procedure, thoracolumbar sympathectomy (in which sympathetic nerve trunks and splanchnic nerves were removed), was sometimes performed to control blood pressure in patients with malignant hypertension. This was effective but caused debilitating side effects such as postural hypotension, erectile dysfunction, and syncope.

Smithwick and Thompson⁶ reported that, in 1,266 hypertensive patients who underwent this procedure and 467 medically treated controls, the 5-year mortality rates were 19% and 54%, respectively. Forty-five percent of those who survived the surgery had significantly lower blood pressure afterward, and the antihypertensive effect lasted 10 years or more.

The procedure fell out of favor due to the morbidity associated with this nonselective approach and to the increased availability of drug therapy.

THE SYMPATHETIC NERVOUS SYSTEM IS A DRIVER OF HYPERTENSION

A variety of evidence suggests that hyperactivation of the sympathetic nervous system plays a major role in initiating and maintaining hypertension. For example, drugs that inhibit the sympathetic drive at various levels have a blood-pressure-lowering effect. Further, direct intraneural recordings show a high level of sympathetic nerve activity in the muscles of hypertensive patients, who also have high levels of cardiac and renal norepinephrine “spillover”—ie, the amount of this neurotransmitter that escapes neuronal uptake and local metabolism and spills over into the circulation.⁷

The kidneys are supplied with postganglionic sympathetic nerve fibers that end in the efferent and afferent renal arterioles, the juxtaglomerular apparatus, and the renal tubular system. Studies in animals and humans have shown that an increase in efferent signals (ie, from the brain to the kidney) leads to renal vasoconstriction and decreased renal blood flow, increased renin release, and sodium retention.^8,9 Afferent signals (from the kidney to the central nervous system), which are increased in states of renal ischemia, renal parenchymal injury, and hypoxia, disinhibit the vasomotor center (the nuclei tractus solitarii) in the central nervous system, leading to increased efferent signals to the kidneys, heart, and peripheral blood vessels (Figure 1).¹⁰

Enhanced sympathetic activity in patients with hypertension may play a role in subsequent target-organ damage such as left ventricular hypertrophy, congestive heart failure, and progressive renal damage.¹¹

Studies of renal denervation in animals, using surgical and chemical techniques, have further helped to establish the role of renal sympathetic nerves in hypertension.^12,13

CATHETER-BASED RENAL DENERVATION

Renal sympathetic nerves run through the adventitia of the renal arteries in a mesh-like pattern.

In the renal denervation procedure, a specially designed catheter is inserted into a femoral artery and advanced into one of the renal arteries. There, radiofrequency energy is applied to the endoluminal surface according to a proprietary algorithm, thereby delivering thermal injury selectively to the renal sympathetic nerves without affecting the abdominal, pelvic, or lower-extremity nerves. The energy delivered is lower than that used for cardiac electrophysiologic procedures.

The nerves are not imaged or mapped before treatment. The procedure is performed on both sides, with four to six sites ablated in a longitudinal and rotational manner in 2-minute treatments at each site, to cover the full circumference (Figure 1).

In the United States, the device (Symplicity Renal Denervation System; Medtronic, Inc, Mountain View, CA) is available only for investigational use.

Below, we briefly review the studies of renal denervation to date. SYMPLICITY HTN-1 Symplicity HTN-1 was a proof-of-principle study in 45 patients with resistant hypertension (Table  1).^14,15

Effect on blood pressure. Six months after renal denervation, blood pressure was significantly lower than at baseline (−22/−11 mm Hg, 95% confidence interval [CI] 10/5 mm Hg) in 26 patients available for follow-up. At 12 months, the difference from baseline was −27/−10 mm Hg (95% CI 16/11 mm Hg) in 9 patients available for follow-up (Table 2).¹⁴

Evidence of the durability of blood pressure reduction came from an expanded cohort of 153 patients followed for 2 years after denervation.¹⁶

Further follow-up data showed a sustained and significant blood pressure reduction through 3 years after denervation (unpublished results presented at the 2012 annual meeting of the American College of Cardiology). Notably, patients who were initially considered to be nonresponders (defined as failure of their blood pressure to go down by at least 10 mm Hg) were all reported to have a clinical response at 36 months.

Adverse events. In the initial and expanded cohorts combined, one patient suffered a renal artery dissection due to manipulation of the guiding catheter before the radiofrequency energy was delivered, and three patients developed a femoral pseudoaneurysm. No other long-term arterial complications were observed.

Comments. Limitations of this study included a small number of patients, no control group, and a primary outcome of a reduction in office blood pressure rather than in ambulatory blood pressure.

Additionally, although the authors concluded that there was no significant deterioration in renal function during the study period, we should note that in an additional follow-up period in this cohort, 10 patients with available 2-year data had a decrease in estimated glomerular filtration rate (eGFR) of −16.0 mL/min/1.73 m². In 5 patients who did not have spironolactone (Aldactone) or another diuretic added after the first year of followup, a lesser but significant decrease (−7.8 mL/min/1.73 m²) was noted. The investigators surmised that denervation may enhance diuretic sensitivity, leading to prerenal azotemia in some patients.¹⁷

SYMPLICITY HTN-2

The Symplicity HTN-2 trial was a larger, randomized, efficacy study that built on the earlier results, providing additional evidence of therapeutic benefit.¹⁵

An international cohort of 106 patients with resistant hypertension, defined as systolic blood pressure of 160 mm Hg or higher (or ≥ 150 mm Hg in patients with type 2 diabetes) despite the use of three or more antihypertensive medications, were randomly assigned to undergo renal denervation with the Symplicity device (n = 52) or to continue their previous treatment with antihypertensive medications alone (n = 54). The primary effectiveness end point was the change in seated office blood pressure from baseline to 6 months (Table 1).

Effect on blood pressure. In the denervation group, at 6 months, office blood pressure had changed by a mean of −32/−12 mm Hg (standard deviation [SD] 23/11 mm Hg) compared with a mean change of 1/0 mm Hg (SD 21/10 mm Hg) in the control group. Fortyone (84%) of the 49 patients who underwent denervation had a decrease in systolic blood pressure of 10 mm Hg or more at 6 months compared with baseline values, while five (10%) had no decline in systolic blood pressure. Nineteen patients had a reduction in systolic pressure to less than 140 mm Hg in the denervation group.

A subset of patients (20 in the denervation group and 25 in the control group) underwent 24-hour ambulatory blood pressure monitoring at 6 months. This showed a similar though less pronounced fall in blood pressure in the denervation group and no change in the controls. A subanalysis that censored all data for patients whose medication was increased during the follow-up period showed a blood pressure reduction of −31/−12 mm Hg (SD 22/11 mm Hg) in the renal denervation group.

Adverse events. Procedure-related adverse events included a single femoral artery pseudoaneurysm, one case of postprocedural hypotension requiring a reduction in antihypertensive medications, and 7 (13%) of 52 patients who experienced intraprocedural bradycardia requiring atropine.

Effect on renal function. No significant difference was noted between groups in the mean change in renal function at 6 months, whether assessed by eGFR, serum creatinine level, or cystatin C level. At 6 months, no patient had a decrease of more than 50% in eGFR, although two patients who underwent renal denervation and three controls had more than a 25% decrease in eGFR.

At 6 months, the urine albumin-to-creatinine ratio had changed by a median of −3 mg/g (range −1,089 to 76) in 38 patients in the treatment group and by 1 mg/g (range −538 to 227) in 37 controls.

Most patients (88%) undergoing renal denervation underwent renal arterial imaging at 6 months, on which a single patient showed possible progression of an underlying atherosclerotic lesion that was unrelated to the procedure and that did not require intervention.

Denervation and the normal stress response. Whether renal denervation negatively affects the body’s physiologic response to stress that is normally mediated by sympathetic nerve activity was addressed in an extended investigation of Symplicity HTN-2 using cardiopulmonary exercise tests at baseline and 3 months after renal denervation.¹⁸ In the denervation group, blood pressure during exercise was significantly lower at 3 months than at baseline, but the heart rate increase at different levels of exercise was not affected. Additionally, the resting heart rate was lower and heart rate recovery after exercise improved after the procedure, particularly in patients without diabetes.

Comments. The Symplicity HTN-2 trial benefited from a randomized trial design and strict inclusion criteria of treatment resistance, but it still had notable limitations. A pretrial evaluation for causes of secondary hypertension or white-coat hypertension was not explicitly described. The control group did not undergo a sham procedure, and data analyzers were not masked to treatment assignment. Although not analyzed as a primary end point, the use of home-based and 24-hour ambulatory blood pressure assessment—measures important for determining white-coat hypertension—revealed substantial differences in blood pressure changes relative to office measurements. Because nearly all the patients (97%) were white, the generalizability of treatment results to black patients with resistant hypertension may be limited. Isolated diastolic hypertension (defined as diastolic pressure ≥ 90 mm Hg with systolic pressure < 140 mm Hg), which is more common in younger patients, was not studied.

DOES RENAL DENERVATION REDUCE SYMPATHETIC TONE?

A subgroup of 10 patients in the Symplicity HTN-1 trial whose mean 6-month office blood pressure was reduced by 22/12 mm Hg underwent assessment of renal norepinephrine spillover. A substantial (47%) reduction in renal norepinephrine spillover was noted 1 month after the procedure.¹⁴

The investigators additionally described a marked reduction in renal norepinephrine spillover from both kidneys in one patient, with a reduction of 48% from the left kidney and 75% from the right kidney 1 month after the procedure. Whole-body norepinephrine spillover in this patient was reduced by 42%. This effect was accompanied by a 50% decrease in plasma renin activity and by an increase in renal plasma flow. Aldosterone levels were not reported.¹⁹

Thus, the decrease in renal norepinephrine spillover suggests a reduction of renal efferent activity, and the decrease in total body norepinephrine spillover suggests a reduction in central sympathetic drive via the renal afferent pathway.

Microneurography in this same patient showed a gradual reduction in muscle sympathetic nerve activity to normal levels, from 56 bursts per minute at baseline to 41 at 30 days and 19 at 12 months).¹⁹ Decreased renin secretion, via circulating angiotensin II, may affect central sympathetic outflow as well.

Comments. While these findings address some of the underlying mechanisms, the small number of patients in whom these studies were done limits the generalizability of the results. The impact of the procedure on renal hemodynamics will need to be studied, including possible direct effects of the procedure, and whether there are differences in different study populations or differences based on blood pressure levels.

WHICH PATIENTS RESPOND BEST TO THIS PROCEDURE?

Although the Symplicity HTN-2 investigators report some predictors of increased reduction in blood pressure on multivariate analysis, including increased blood pressure at baseline and reduced heart rate at baseline, these are not specific enough to enable patient selection.

Interestingly, results from the expanded cohort of the Symplicity HTN-1 study found that patients on central sympatholytic agents such as clonidine had a greater reduction in blood pressure, although the reason for this is unclear.¹⁶ Identifying specific predictors of treatment success at baseline will be essential in future studies.

The earlier Symplicity trials and the ongoing Symplicity HTN-3 trial are in patients who have high blood pressure not responding to three or more antihypertensive drugs. The mean baseline systolic blood pressure in the Symplicity HTN-1 and HTN-2 trials was 178 mm Hg, and patients were taking an average of five antihypertensive drugs (Table 1). It is not known whether denervation will produce similar blood-pressure-lowering results across the spectrum of hypertension severity.

WHAT ARE THE LONG-TERM RESULTS OF DENERVATION?

Enthusiasm for the results from the Symplicity trials is tempered by concerns about the durability of the effects of the procedure, the need for better understanding of the impact of renal denervation on a wide array of pathophysiologic cascades leading to hypertension, and the effect on renal hemodynamics.

Antihypertensive efficacy has been reported to persist up to 2 years after the procedure,¹⁶ with recent unpublished data suggesting efficacy up to 3 years, but longer follow-up is needed to address whether these effects are finite.

Although reinnervation of afferent renal nerves has not been described, transplant models have shown anatomic regrowth of efferent nerves; the impact of this efferent reinnervation on blood pressure remains unclear. Experience from renal transplantation also shows that implanted kidneys that are “denervated” can still maintain fluid and electrolyte regulation.

Follow-up renal imaging in the Symplicity trials did not indicate renal artery stenosis at the sites of denervation in patients who underwent the procedure. Animal studies using the Symplicity catheter system showed renal nerve injury as evidenced by nerve fibrosis and thickened epineurium and perineurium, but no significant smooth muscle hyperplasia, arterial stenosis, or thrombosis by angiography or histology at 6 months.²⁰

WHAT ARE THE RISKS?

Adverse effects that were noted in the short term are detailed under discussion of the trials and in Table 2.

Long-term adverse events in the Symplicity HTN-2 trial that required hospitalization were reported in five patients in the denervation group and three patients in the control group (Table 2). These included transient ischemic attacks, hypertensive crises, hypotensive episodes, angina, and nausea.

Renal function was maintained for the duration of both trials, and details regarding eGFR change have been described above under the discussion of the trials.

Diffuse visceral pain at the time of the procedure is reported as an expected occurrence, managed with intravenous analgesic medications.

DOES SYMPATHETIC DENERVATION HAVE A ROLE IN OTHER CONDITIONS?

Interestingly, other sympathetically driven diseases, such as diabetes mellitus and polycystic ovary syndrome, may prove to be targets for this therapy in the future.²¹

Mahfoud et al²² conducted a pilot study in 37 patients with resistant hypertension undergoing renal denervation and 13 control patients. Fasting glucose levels declined from 118 ± 3.4 mg/dL to 108 ± 3.8 mg/dL after 3 months in the intervention group (P = .039), compared with no change in the control group. Insulin and C-peptide levels were also lower in the intervention group. The reported improvement in glucose metabolism and insulin sensitivity suggests that the beneficial effects of this procedure may extend beyond blood pressure reduction.

Brandt et al²³ reported regression of left ventricular hypertrophy and significantly improved cardiac functional parameters, including increase in ejection fraction and improved diastolic dysfunction, in a study of 46 patients who underwent renal denervation. This findings suggests a potential beneficial effect on cardiac remodeling.

Witkowski et al²⁴ reported lowering of blood pressure in 10 patients with refractory hypertension and obstructive sleep apnea who underwent renal denervation, which was accompanied by improvement of sleep apnea severity.

Ukena et al²⁵ reported reduction in ventricular tachyarrhythmias in two patients with congestive heart failure who had therapy-resistant electrical storm.

A recent pilot study in 15 patients with stage 3 and 4 chronic kidney disease (mean eGFR 31 mL/min/1.73 m²) showed significantly improved office blood pressure control up to 1 year, restoration of nocturnal dipping on 24-hour monitoring, as well as a nonsignificant trend towards increased hemoglobin levels and decreased proteinuria. No additional deterioration of renal function was reported in these patients (2 patients had renal function assessed up to 1 year).²⁶

Thus, the benefits of this procedure may extend to other diseases that have a common underlying thread of elevated sympathetic activity, by targeting the “sympathorenal” axis.²⁷

GUARDED OPTIMISM AND FUTURE DIRECTIONS

Given the well-known cardiovascular risks and health care costs associated with uncontrolled hypertension and the continued challenge that physicians face in managing it, novel therapies such as renal denervation may provide an adjunct to existing pharmacologic approaches.

While there is certainly cause for guarded optimism, especially with the striking blood pressure-lowering results seen in trials so far, it should be kept in mind that the mechanisms leading to the hypertensive response are complex and multifactorial, and further understanding of this therapy with long-term follow-up is needed. A comparison study with spironolactone, which is increasingly being used to treat resistant hypertension (in the absence of a diagnosis of primary aldosteronism)^28,29 would help to further establish the role of this procedure.

Studies of carotid baroreceptor stimulation via an implantable device have shown sustained reduction in blood pressure in patients with resistant hypertension. A study comparing this technique with renal denervation for efficacy and safety end points could be considered in the future.^30,31

The planned Symplicity HTN-3 study in the United States will be the largest trial to date, with a targeted randomization of more than 500 patients using strict enrollment criteria, including the use of maximally tolerated doses of diuretics and more focus on the use of ambulatory blood pressure monitoring and on the blinding of participants. This study will help further analysis of this technology in a more diverse population.^32,33

Future studies should be designed to clarify pathophysiologic mechanisms, patient selection criteria, effects on target organ damage, and efficacy in patients with chronic kidney disease, obesity, congestive heart failure, and in less severe forms of hypertension.

A CALL FOR PARTICIPANTS IN A CLINICAL TRIAL

The Departments of Cardiology and Nephrology and Hypertension at Cleveland Clinic are currently enrolling patients in the Symplicity HTN-3 trial. For more information, please contact George Thomas, MD ([email protected]), or Mehdi Shishehbor, DO, MPH ([email protected]), or visit www.symplifybptrial.com.

Can a percutaneous catheter-based procedure effectively treat resistant hypertension?

Radiofrequency ablation of the renal sympathetic nerves is undergoing randomized controlled trials in patients who have resistant hypertension and other disorders that involve the sympathetic nervous system. Remarkably, the limited results available so far look good.

See related editorial

This article discusses the physiologic rationale for renal denervation, the evidence from studies in humans of the benefits, risks, and complications of the procedure, upcoming trials, and areas for future research.

DESPITE MANY TREATMENT OPTIONS, RESISTANT HYPERTENSION IS COMMON

Hypertension is a leading reason for visits to physicians in the United States and is associated with increased rates of cardiovascular disease and death.^1,2 A variety of antihypertensive agents are available, and the percentage of people with hypertension whose blood pressure is under control has increased over the past 2 decades. Nevertheless, population-based studies show that the control rate remains suboptimal.³ Effective pharmacologic treatment may be limited by inadequate doses or inappropriate combinations of antihypertensive drugs, concurrent use of agents that raise the blood pressure, noncompliance with dietary restrictions, and side effects that result in poor compliance with drug therapy.

Resistant hypertension is defined as failure to achieve goal blood pressure in patients who are adhering to full tolerated doses of an appropriate three-drug regimen that includes a diuretic.^1,4,5 If we use these criteria, many patients labelled as having resistant hypertension probably do not truly have it; instead, they are nonadherent to therapy or are on an inadequate or inappropriate regimen. Although the true prevalence of resistant hypertension is not clear, estimates from large clinical trials suggest that about 20% to 30% of hypertensive patients may meet the criteria for it.⁴ For the subset of patients who have truly resistant hypertension, nonpharmacologic treatments such as renal sympathetic denervation are an intriguing avenue.

SURGICAL SYMPATHETIC DENERVATION: TRIED AND ABANDONED IN THE 1950s

More than a half century ago, a surgical procedure, thoracolumbar sympathectomy (in which sympathetic nerve trunks and splanchnic nerves were removed), was sometimes performed to control blood pressure in patients with malignant hypertension. This was effective but caused debilitating side effects such as postural hypotension, erectile dysfunction, and syncope.

Smithwick and Thompson⁶ reported that, in 1,266 hypertensive patients who underwent this procedure and 467 medically treated controls, the 5-year mortality rates were 19% and 54%, respectively. Forty-five percent of those who survived the surgery had significantly lower blood pressure afterward, and the antihypertensive effect lasted 10 years or more.

The procedure fell out of favor due to the morbidity associated with this nonselective approach and to the increased availability of drug therapy.

THE SYMPATHETIC NERVOUS SYSTEM IS A DRIVER OF HYPERTENSION

A variety of evidence suggests that hyperactivation of the sympathetic nervous system plays a major role in initiating and maintaining hypertension. For example, drugs that inhibit the sympathetic drive at various levels have a blood-pressure-lowering effect. Further, direct intraneural recordings show a high level of sympathetic nerve activity in the muscles of hypertensive patients, who also have high levels of cardiac and renal norepinephrine “spillover”—ie, the amount of this neurotransmitter that escapes neuronal uptake and local metabolism and spills over into the circulation.⁷

The kidneys are supplied with postganglionic sympathetic nerve fibers that end in the efferent and afferent renal arterioles, the juxtaglomerular apparatus, and the renal tubular system. Studies in animals and humans have shown that an increase in efferent signals (ie, from the brain to the kidney) leads to renal vasoconstriction and decreased renal blood flow, increased renin release, and sodium retention.^8,9 Afferent signals (from the kidney to the central nervous system), which are increased in states of renal ischemia, renal parenchymal injury, and hypoxia, disinhibit the vasomotor center (the nuclei tractus solitarii) in the central nervous system, leading to increased efferent signals to the kidneys, heart, and peripheral blood vessels (Figure 1).¹⁰

Enhanced sympathetic activity in patients with hypertension may play a role in subsequent target-organ damage such as left ventricular hypertrophy, congestive heart failure, and progressive renal damage.¹¹

Studies of renal denervation in animals, using surgical and chemical techniques, have further helped to establish the role of renal sympathetic nerves in hypertension.^12,13

CATHETER-BASED RENAL DENERVATION

Renal sympathetic nerves run through the adventitia of the renal arteries in a mesh-like pattern.

In the renal denervation procedure, a specially designed catheter is inserted into a femoral artery and advanced into one of the renal arteries. There, radiofrequency energy is applied to the endoluminal surface according to a proprietary algorithm, thereby delivering thermal injury selectively to the renal sympathetic nerves without affecting the abdominal, pelvic, or lower-extremity nerves. The energy delivered is lower than that used for cardiac electrophysiologic procedures.

The nerves are not imaged or mapped before treatment. The procedure is performed on both sides, with four to six sites ablated in a longitudinal and rotational manner in 2-minute treatments at each site, to cover the full circumference (Figure 1).

In the United States, the device (Symplicity Renal Denervation System; Medtronic, Inc, Mountain View, CA) is available only for investigational use.

Below, we briefly review the studies of renal denervation to date. SYMPLICITY HTN-1 Symplicity HTN-1 was a proof-of-principle study in 45 patients with resistant hypertension (Table  1).^14,15

Effect on blood pressure. Six months after renal denervation, blood pressure was significantly lower than at baseline (−22/−11 mm Hg, 95% confidence interval [CI] 10/5 mm Hg) in 26 patients available for follow-up. At 12 months, the difference from baseline was −27/−10 mm Hg (95% CI 16/11 mm Hg) in 9 patients available for follow-up (Table 2).¹⁴

Evidence of the durability of blood pressure reduction came from an expanded cohort of 153 patients followed for 2 years after denervation.¹⁶

Further follow-up data showed a sustained and significant blood pressure reduction through 3 years after denervation (unpublished results presented at the 2012 annual meeting of the American College of Cardiology). Notably, patients who were initially considered to be nonresponders (defined as failure of their blood pressure to go down by at least 10 mm Hg) were all reported to have a clinical response at 36 months.

Adverse events. In the initial and expanded cohorts combined, one patient suffered a renal artery dissection due to manipulation of the guiding catheter before the radiofrequency energy was delivered, and three patients developed a femoral pseudoaneurysm. No other long-term arterial complications were observed.

Comments. Limitations of this study included a small number of patients, no control group, and a primary outcome of a reduction in office blood pressure rather than in ambulatory blood pressure.

Additionally, although the authors concluded that there was no significant deterioration in renal function during the study period, we should note that in an additional follow-up period in this cohort, 10 patients with available 2-year data had a decrease in estimated glomerular filtration rate (eGFR) of −16.0 mL/min/1.73 m². In 5 patients who did not have spironolactone (Aldactone) or another diuretic added after the first year of followup, a lesser but significant decrease (−7.8 mL/min/1.73 m²) was noted. The investigators surmised that denervation may enhance diuretic sensitivity, leading to prerenal azotemia in some patients.¹⁷

SYMPLICITY HTN-2

The Symplicity HTN-2 trial was a larger, randomized, efficacy study that built on the earlier results, providing additional evidence of therapeutic benefit.¹⁵

An international cohort of 106 patients with resistant hypertension, defined as systolic blood pressure of 160 mm Hg or higher (or ≥ 150 mm Hg in patients with type 2 diabetes) despite the use of three or more antihypertensive medications, were randomly assigned to undergo renal denervation with the Symplicity device (n = 52) or to continue their previous treatment with antihypertensive medications alone (n = 54). The primary effectiveness end point was the change in seated office blood pressure from baseline to 6 months (Table 1).

Effect on blood pressure. In the denervation group, at 6 months, office blood pressure had changed by a mean of −32/−12 mm Hg (standard deviation [SD] 23/11 mm Hg) compared with a mean change of 1/0 mm Hg (SD 21/10 mm Hg) in the control group. Fortyone (84%) of the 49 patients who underwent denervation had a decrease in systolic blood pressure of 10 mm Hg or more at 6 months compared with baseline values, while five (10%) had no decline in systolic blood pressure. Nineteen patients had a reduction in systolic pressure to less than 140 mm Hg in the denervation group.

A subset of patients (20 in the denervation group and 25 in the control group) underwent 24-hour ambulatory blood pressure monitoring at 6 months. This showed a similar though less pronounced fall in blood pressure in the denervation group and no change in the controls. A subanalysis that censored all data for patients whose medication was increased during the follow-up period showed a blood pressure reduction of −31/−12 mm Hg (SD 22/11 mm Hg) in the renal denervation group.

Adverse events. Procedure-related adverse events included a single femoral artery pseudoaneurysm, one case of postprocedural hypotension requiring a reduction in antihypertensive medications, and 7 (13%) of 52 patients who experienced intraprocedural bradycardia requiring atropine.

Effect on renal function. No significant difference was noted between groups in the mean change in renal function at 6 months, whether assessed by eGFR, serum creatinine level, or cystatin C level. At 6 months, no patient had a decrease of more than 50% in eGFR, although two patients who underwent renal denervation and three controls had more than a 25% decrease in eGFR.

At 6 months, the urine albumin-to-creatinine ratio had changed by a median of −3 mg/g (range −1,089 to 76) in 38 patients in the treatment group and by 1 mg/g (range −538 to 227) in 37 controls.

Most patients (88%) undergoing renal denervation underwent renal arterial imaging at 6 months, on which a single patient showed possible progression of an underlying atherosclerotic lesion that was unrelated to the procedure and that did not require intervention.

Denervation and the normal stress response. Whether renal denervation negatively affects the body’s physiologic response to stress that is normally mediated by sympathetic nerve activity was addressed in an extended investigation of Symplicity HTN-2 using cardiopulmonary exercise tests at baseline and 3 months after renal denervation.¹⁸ In the denervation group, blood pressure during exercise was significantly lower at 3 months than at baseline, but the heart rate increase at different levels of exercise was not affected. Additionally, the resting heart rate was lower and heart rate recovery after exercise improved after the procedure, particularly in patients without diabetes.

Comments. The Symplicity HTN-2 trial benefited from a randomized trial design and strict inclusion criteria of treatment resistance, but it still had notable limitations. A pretrial evaluation for causes of secondary hypertension or white-coat hypertension was not explicitly described. The control group did not undergo a sham procedure, and data analyzers were not masked to treatment assignment. Although not analyzed as a primary end point, the use of home-based and 24-hour ambulatory blood pressure assessment—measures important for determining white-coat hypertension—revealed substantial differences in blood pressure changes relative to office measurements. Because nearly all the patients (97%) were white, the generalizability of treatment results to black patients with resistant hypertension may be limited. Isolated diastolic hypertension (defined as diastolic pressure ≥ 90 mm Hg with systolic pressure < 140 mm Hg), which is more common in younger patients, was not studied.

DOES RENAL DENERVATION REDUCE SYMPATHETIC TONE?

A subgroup of 10 patients in the Symplicity HTN-1 trial whose mean 6-month office blood pressure was reduced by 22/12 mm Hg underwent assessment of renal norepinephrine spillover. A substantial (47%) reduction in renal norepinephrine spillover was noted 1 month after the procedure.¹⁴

The investigators additionally described a marked reduction in renal norepinephrine spillover from both kidneys in one patient, with a reduction of 48% from the left kidney and 75% from the right kidney 1 month after the procedure. Whole-body norepinephrine spillover in this patient was reduced by 42%. This effect was accompanied by a 50% decrease in plasma renin activity and by an increase in renal plasma flow. Aldosterone levels were not reported.¹⁹

Thus, the decrease in renal norepinephrine spillover suggests a reduction of renal efferent activity, and the decrease in total body norepinephrine spillover suggests a reduction in central sympathetic drive via the renal afferent pathway.

Microneurography in this same patient showed a gradual reduction in muscle sympathetic nerve activity to normal levels, from 56 bursts per minute at baseline to 41 at 30 days and 19 at 12 months).¹⁹ Decreased renin secretion, via circulating angiotensin II, may affect central sympathetic outflow as well.

Comments. While these findings address some of the underlying mechanisms, the small number of patients in whom these studies were done limits the generalizability of the results. The impact of the procedure on renal hemodynamics will need to be studied, including possible direct effects of the procedure, and whether there are differences in different study populations or differences based on blood pressure levels.

WHICH PATIENTS RESPOND BEST TO THIS PROCEDURE?

Although the Symplicity HTN-2 investigators report some predictors of increased reduction in blood pressure on multivariate analysis, including increased blood pressure at baseline and reduced heart rate at baseline, these are not specific enough to enable patient selection.

Interestingly, results from the expanded cohort of the Symplicity HTN-1 study found that patients on central sympatholytic agents such as clonidine had a greater reduction in blood pressure, although the reason for this is unclear.¹⁶ Identifying specific predictors of treatment success at baseline will be essential in future studies.

The earlier Symplicity trials and the ongoing Symplicity HTN-3 trial are in patients who have high blood pressure not responding to three or more antihypertensive drugs. The mean baseline systolic blood pressure in the Symplicity HTN-1 and HTN-2 trials was 178 mm Hg, and patients were taking an average of five antihypertensive drugs (Table 1). It is not known whether denervation will produce similar blood-pressure-lowering results across the spectrum of hypertension severity.

WHAT ARE THE LONG-TERM RESULTS OF DENERVATION?

Enthusiasm for the results from the Symplicity trials is tempered by concerns about the durability of the effects of the procedure, the need for better understanding of the impact of renal denervation on a wide array of pathophysiologic cascades leading to hypertension, and the effect on renal hemodynamics.

Antihypertensive efficacy has been reported to persist up to 2 years after the procedure,¹⁶ with recent unpublished data suggesting efficacy up to 3 years, but longer follow-up is needed to address whether these effects are finite.

Although reinnervation of afferent renal nerves has not been described, transplant models have shown anatomic regrowth of efferent nerves; the impact of this efferent reinnervation on blood pressure remains unclear. Experience from renal transplantation also shows that implanted kidneys that are “denervated” can still maintain fluid and electrolyte regulation.

Follow-up renal imaging in the Symplicity trials did not indicate renal artery stenosis at the sites of denervation in patients who underwent the procedure. Animal studies using the Symplicity catheter system showed renal nerve injury as evidenced by nerve fibrosis and thickened epineurium and perineurium, but no significant smooth muscle hyperplasia, arterial stenosis, or thrombosis by angiography or histology at 6 months.²⁰

WHAT ARE THE RISKS?

Adverse effects that were noted in the short term are detailed under discussion of the trials and in Table 2.

Long-term adverse events in the Symplicity HTN-2 trial that required hospitalization were reported in five patients in the denervation group and three patients in the control group (Table 2). These included transient ischemic attacks, hypertensive crises, hypotensive episodes, angina, and nausea.

Renal function was maintained for the duration of both trials, and details regarding eGFR change have been described above under the discussion of the trials.

Diffuse visceral pain at the time of the procedure is reported as an expected occurrence, managed with intravenous analgesic medications.

DOES SYMPATHETIC DENERVATION HAVE A ROLE IN OTHER CONDITIONS?

Interestingly, other sympathetically driven diseases, such as diabetes mellitus and polycystic ovary syndrome, may prove to be targets for this therapy in the future.²¹

Mahfoud et al²² conducted a pilot study in 37 patients with resistant hypertension undergoing renal denervation and 13 control patients. Fasting glucose levels declined from 118 ± 3.4 mg/dL to 108 ± 3.8 mg/dL after 3 months in the intervention group (P = .039), compared with no change in the control group. Insulin and C-peptide levels were also lower in the intervention group. The reported improvement in glucose metabolism and insulin sensitivity suggests that the beneficial effects of this procedure may extend beyond blood pressure reduction.

Brandt et al²³ reported regression of left ventricular hypertrophy and significantly improved cardiac functional parameters, including increase in ejection fraction and improved diastolic dysfunction, in a study of 46 patients who underwent renal denervation. This findings suggests a potential beneficial effect on cardiac remodeling.

Witkowski et al²⁴ reported lowering of blood pressure in 10 patients with refractory hypertension and obstructive sleep apnea who underwent renal denervation, which was accompanied by improvement of sleep apnea severity.

Ukena et al²⁵ reported reduction in ventricular tachyarrhythmias in two patients with congestive heart failure who had therapy-resistant electrical storm.

A recent pilot study in 15 patients with stage 3 and 4 chronic kidney disease (mean eGFR 31 mL/min/1.73 m²) showed significantly improved office blood pressure control up to 1 year, restoration of nocturnal dipping on 24-hour monitoring, as well as a nonsignificant trend towards increased hemoglobin levels and decreased proteinuria. No additional deterioration of renal function was reported in these patients (2 patients had renal function assessed up to 1 year).²⁶

Thus, the benefits of this procedure may extend to other diseases that have a common underlying thread of elevated sympathetic activity, by targeting the “sympathorenal” axis.²⁷

GUARDED OPTIMISM AND FUTURE DIRECTIONS

Given the well-known cardiovascular risks and health care costs associated with uncontrolled hypertension and the continued challenge that physicians face in managing it, novel therapies such as renal denervation may provide an adjunct to existing pharmacologic approaches.

While there is certainly cause for guarded optimism, especially with the striking blood pressure-lowering results seen in trials so far, it should be kept in mind that the mechanisms leading to the hypertensive response are complex and multifactorial, and further understanding of this therapy with long-term follow-up is needed. A comparison study with spironolactone, which is increasingly being used to treat resistant hypertension (in the absence of a diagnosis of primary aldosteronism)^28,29 would help to further establish the role of this procedure.

Studies of carotid baroreceptor stimulation via an implantable device have shown sustained reduction in blood pressure in patients with resistant hypertension. A study comparing this technique with renal denervation for efficacy and safety end points could be considered in the future.^30,31

The planned Symplicity HTN-3 study in the United States will be the largest trial to date, with a targeted randomization of more than 500 patients using strict enrollment criteria, including the use of maximally tolerated doses of diuretics and more focus on the use of ambulatory blood pressure monitoring and on the blinding of participants. This study will help further analysis of this technology in a more diverse population.^32,33

Future studies should be designed to clarify pathophysiologic mechanisms, patient selection criteria, effects on target organ damage, and efficacy in patients with chronic kidney disease, obesity, congestive heart failure, and in less severe forms of hypertension.

A CALL FOR PARTICIPANTS IN A CLINICAL TRIAL

The Departments of Cardiology and Nephrology and Hypertension at Cleveland Clinic are currently enrolling patients in the Symplicity HTN-3 trial. For more information, please contact George Thomas, MD ([email protected]), or Mehdi Shishehbor, DO, MPH ([email protected]), or visit www.symplifybptrial.com.

Can a percutaneous catheter-based procedure effectively treat resistant hypertension?

Radiofrequency ablation of the renal sympathetic nerves is undergoing randomized controlled trials in patients who have resistant hypertension and other disorders that involve the sympathetic nervous system. Remarkably, the limited results available so far look good.

See related editorial

This article discusses the physiologic rationale for renal denervation, the evidence from studies in humans of the benefits, risks, and complications of the procedure, upcoming trials, and areas for future research.

DESPITE MANY TREATMENT OPTIONS, RESISTANT HYPERTENSION IS COMMON

Hypertension is a leading reason for visits to physicians in the United States and is associated with increased rates of cardiovascular disease and death.^1,2 A variety of antihypertensive agents are available, and the percentage of people with hypertension whose blood pressure is under control has increased over the past 2 decades. Nevertheless, population-based studies show that the control rate remains suboptimal.³ Effective pharmacologic treatment may be limited by inadequate doses or inappropriate combinations of antihypertensive drugs, concurrent use of agents that raise the blood pressure, noncompliance with dietary restrictions, and side effects that result in poor compliance with drug therapy.

Resistant hypertension is defined as failure to achieve goal blood pressure in patients who are adhering to full tolerated doses of an appropriate three-drug regimen that includes a diuretic.^1,4,5 If we use these criteria, many patients labelled as having resistant hypertension probably do not truly have it; instead, they are nonadherent to therapy or are on an inadequate or inappropriate regimen. Although the true prevalence of resistant hypertension is not clear, estimates from large clinical trials suggest that about 20% to 30% of hypertensive patients may meet the criteria for it.⁴ For the subset of patients who have truly resistant hypertension, nonpharmacologic treatments such as renal sympathetic denervation are an intriguing avenue.

SURGICAL SYMPATHETIC DENERVATION: TRIED AND ABANDONED IN THE 1950s

More than a half century ago, a surgical procedure, thoracolumbar sympathectomy (in which sympathetic nerve trunks and splanchnic nerves were removed), was sometimes performed to control blood pressure in patients with malignant hypertension. This was effective but caused debilitating side effects such as postural hypotension, erectile dysfunction, and syncope.

Smithwick and Thompson⁶ reported that, in 1,266 hypertensive patients who underwent this procedure and 467 medically treated controls, the 5-year mortality rates were 19% and 54%, respectively. Forty-five percent of those who survived the surgery had significantly lower blood pressure afterward, and the antihypertensive effect lasted 10 years or more.

The procedure fell out of favor due to the morbidity associated with this nonselective approach and to the increased availability of drug therapy.

THE SYMPATHETIC NERVOUS SYSTEM IS A DRIVER OF HYPERTENSION

A variety of evidence suggests that hyperactivation of the sympathetic nervous system plays a major role in initiating and maintaining hypertension. For example, drugs that inhibit the sympathetic drive at various levels have a blood-pressure-lowering effect. Further, direct intraneural recordings show a high level of sympathetic nerve activity in the muscles of hypertensive patients, who also have high levels of cardiac and renal norepinephrine “spillover”—ie, the amount of this neurotransmitter that escapes neuronal uptake and local metabolism and spills over into the circulation.⁷

The kidneys are supplied with postganglionic sympathetic nerve fibers that end in the efferent and afferent renal arterioles, the juxtaglomerular apparatus, and the renal tubular system. Studies in animals and humans have shown that an increase in efferent signals (ie, from the brain to the kidney) leads to renal vasoconstriction and decreased renal blood flow, increased renin release, and sodium retention.^8,9 Afferent signals (from the kidney to the central nervous system), which are increased in states of renal ischemia, renal parenchymal injury, and hypoxia, disinhibit the vasomotor center (the nuclei tractus solitarii) in the central nervous system, leading to increased efferent signals to the kidneys, heart, and peripheral blood vessels (Figure 1).¹⁰

Enhanced sympathetic activity in patients with hypertension may play a role in subsequent target-organ damage such as left ventricular hypertrophy, congestive heart failure, and progressive renal damage.¹¹

Studies of renal denervation in animals, using surgical and chemical techniques, have further helped to establish the role of renal sympathetic nerves in hypertension.^12,13

CATHETER-BASED RENAL DENERVATION

Renal sympathetic nerves run through the adventitia of the renal arteries in a mesh-like pattern.

In the renal denervation procedure, a specially designed catheter is inserted into a femoral artery and advanced into one of the renal arteries. There, radiofrequency energy is applied to the endoluminal surface according to a proprietary algorithm, thereby delivering thermal injury selectively to the renal sympathetic nerves without affecting the abdominal, pelvic, or lower-extremity nerves. The energy delivered is lower than that used for cardiac electrophysiologic procedures.

The nerves are not imaged or mapped before treatment. The procedure is performed on both sides, with four to six sites ablated in a longitudinal and rotational manner in 2-minute treatments at each site, to cover the full circumference (Figure 1).

In the United States, the device (Symplicity Renal Denervation System; Medtronic, Inc, Mountain View, CA) is available only for investigational use.

Below, we briefly review the studies of renal denervation to date. SYMPLICITY HTN-1 Symplicity HTN-1 was a proof-of-principle study in 45 patients with resistant hypertension (Table  1).^14,15

Effect on blood pressure. Six months after renal denervation, blood pressure was significantly lower than at baseline (−22/−11 mm Hg, 95% confidence interval [CI] 10/5 mm Hg) in 26 patients available for follow-up. At 12 months, the difference from baseline was −27/−10 mm Hg (95% CI 16/11 mm Hg) in 9 patients available for follow-up (Table 2).¹⁴

Evidence of the durability of blood pressure reduction came from an expanded cohort of 153 patients followed for 2 years after denervation.¹⁶

Further follow-up data showed a sustained and significant blood pressure reduction through 3 years after denervation (unpublished results presented at the 2012 annual meeting of the American College of Cardiology). Notably, patients who were initially considered to be nonresponders (defined as failure of their blood pressure to go down by at least 10 mm Hg) were all reported to have a clinical response at 36 months.

Adverse events. In the initial and expanded cohorts combined, one patient suffered a renal artery dissection due to manipulation of the guiding catheter before the radiofrequency energy was delivered, and three patients developed a femoral pseudoaneurysm. No other long-term arterial complications were observed.

Comments. Limitations of this study included a small number of patients, no control group, and a primary outcome of a reduction in office blood pressure rather than in ambulatory blood pressure.

Additionally, although the authors concluded that there was no significant deterioration in renal function during the study period, we should note that in an additional follow-up period in this cohort, 10 patients with available 2-year data had a decrease in estimated glomerular filtration rate (eGFR) of −16.0 mL/min/1.73 m². In 5 patients who did not have spironolactone (Aldactone) or another diuretic added after the first year of followup, a lesser but significant decrease (−7.8 mL/min/1.73 m²) was noted. The investigators surmised that denervation may enhance diuretic sensitivity, leading to prerenal azotemia in some patients.¹⁷

SYMPLICITY HTN-2

The Symplicity HTN-2 trial was a larger, randomized, efficacy study that built on the earlier results, providing additional evidence of therapeutic benefit.¹⁵

An international cohort of 106 patients with resistant hypertension, defined as systolic blood pressure of 160 mm Hg or higher (or ≥ 150 mm Hg in patients with type 2 diabetes) despite the use of three or more antihypertensive medications, were randomly assigned to undergo renal denervation with the Symplicity device (n = 52) or to continue their previous treatment with antihypertensive medications alone (n = 54). The primary effectiveness end point was the change in seated office blood pressure from baseline to 6 months (Table 1).

Effect on blood pressure. In the denervation group, at 6 months, office blood pressure had changed by a mean of −32/−12 mm Hg (standard deviation [SD] 23/11 mm Hg) compared with a mean change of 1/0 mm Hg (SD 21/10 mm Hg) in the control group. Fortyone (84%) of the 49 patients who underwent denervation had a decrease in systolic blood pressure of 10 mm Hg or more at 6 months compared with baseline values, while five (10%) had no decline in systolic blood pressure. Nineteen patients had a reduction in systolic pressure to less than 140 mm Hg in the denervation group.

A subset of patients (20 in the denervation group and 25 in the control group) underwent 24-hour ambulatory blood pressure monitoring at 6 months. This showed a similar though less pronounced fall in blood pressure in the denervation group and no change in the controls. A subanalysis that censored all data for patients whose medication was increased during the follow-up period showed a blood pressure reduction of −31/−12 mm Hg (SD 22/11 mm Hg) in the renal denervation group.

Adverse events. Procedure-related adverse events included a single femoral artery pseudoaneurysm, one case of postprocedural hypotension requiring a reduction in antihypertensive medications, and 7 (13%) of 52 patients who experienced intraprocedural bradycardia requiring atropine.

Effect on renal function. No significant difference was noted between groups in the mean change in renal function at 6 months, whether assessed by eGFR, serum creatinine level, or cystatin C level. At 6 months, no patient had a decrease of more than 50% in eGFR, although two patients who underwent renal denervation and three controls had more than a 25% decrease in eGFR.

At 6 months, the urine albumin-to-creatinine ratio had changed by a median of −3 mg/g (range −1,089 to 76) in 38 patients in the treatment group and by 1 mg/g (range −538 to 227) in 37 controls.

Most patients (88%) undergoing renal denervation underwent renal arterial imaging at 6 months, on which a single patient showed possible progression of an underlying atherosclerotic lesion that was unrelated to the procedure and that did not require intervention.

Denervation and the normal stress response. Whether renal denervation negatively affects the body’s physiologic response to stress that is normally mediated by sympathetic nerve activity was addressed in an extended investigation of Symplicity HTN-2 using cardiopulmonary exercise tests at baseline and 3 months after renal denervation.¹⁸ In the denervation group, blood pressure during exercise was significantly lower at 3 months than at baseline, but the heart rate increase at different levels of exercise was not affected. Additionally, the resting heart rate was lower and heart rate recovery after exercise improved after the procedure, particularly in patients without diabetes.

Comments. The Symplicity HTN-2 trial benefited from a randomized trial design and strict inclusion criteria of treatment resistance, but it still had notable limitations. A pretrial evaluation for causes of secondary hypertension or white-coat hypertension was not explicitly described. The control group did not undergo a sham procedure, and data analyzers were not masked to treatment assignment. Although not analyzed as a primary end point, the use of home-based and 24-hour ambulatory blood pressure assessment—measures important for determining white-coat hypertension—revealed substantial differences in blood pressure changes relative to office measurements. Because nearly all the patients (97%) were white, the generalizability of treatment results to black patients with resistant hypertension may be limited. Isolated diastolic hypertension (defined as diastolic pressure ≥ 90 mm Hg with systolic pressure < 140 mm Hg), which is more common in younger patients, was not studied.

DOES RENAL DENERVATION REDUCE SYMPATHETIC TONE?

A subgroup of 10 patients in the Symplicity HTN-1 trial whose mean 6-month office blood pressure was reduced by 22/12 mm Hg underwent assessment of renal norepinephrine spillover. A substantial (47%) reduction in renal norepinephrine spillover was noted 1 month after the procedure.¹⁴

The investigators additionally described a marked reduction in renal norepinephrine spillover from both kidneys in one patient, with a reduction of 48% from the left kidney and 75% from the right kidney 1 month after the procedure. Whole-body norepinephrine spillover in this patient was reduced by 42%. This effect was accompanied by a 50% decrease in plasma renin activity and by an increase in renal plasma flow. Aldosterone levels were not reported.¹⁹

Thus, the decrease in renal norepinephrine spillover suggests a reduction of renal efferent activity, and the decrease in total body norepinephrine spillover suggests a reduction in central sympathetic drive via the renal afferent pathway.

Microneurography in this same patient showed a gradual reduction in muscle sympathetic nerve activity to normal levels, from 56 bursts per minute at baseline to 41 at 30 days and 19 at 12 months).¹⁹ Decreased renin secretion, via circulating angiotensin II, may affect central sympathetic outflow as well.

Comments. While these findings address some of the underlying mechanisms, the small number of patients in whom these studies were done limits the generalizability of the results. The impact of the procedure on renal hemodynamics will need to be studied, including possible direct effects of the procedure, and whether there are differences in different study populations or differences based on blood pressure levels.

WHICH PATIENTS RESPOND BEST TO THIS PROCEDURE?

Although the Symplicity HTN-2 investigators report some predictors of increased reduction in blood pressure on multivariate analysis, including increased blood pressure at baseline and reduced heart rate at baseline, these are not specific enough to enable patient selection.

Interestingly, results from the expanded cohort of the Symplicity HTN-1 study found that patients on central sympatholytic agents such as clonidine had a greater reduction in blood pressure, although the reason for this is unclear.¹⁶ Identifying specific predictors of treatment success at baseline will be essential in future studies.

The earlier Symplicity trials and the ongoing Symplicity HTN-3 trial are in patients who have high blood pressure not responding to three or more antihypertensive drugs. The mean baseline systolic blood pressure in the Symplicity HTN-1 and HTN-2 trials was 178 mm Hg, and patients were taking an average of five antihypertensive drugs (Table 1). It is not known whether denervation will produce similar blood-pressure-lowering results across the spectrum of hypertension severity.

WHAT ARE THE LONG-TERM RESULTS OF DENERVATION?

Enthusiasm for the results from the Symplicity trials is tempered by concerns about the durability of the effects of the procedure, the need for better understanding of the impact of renal denervation on a wide array of pathophysiologic cascades leading to hypertension, and the effect on renal hemodynamics.

Antihypertensive efficacy has been reported to persist up to 2 years after the procedure,¹⁶ with recent unpublished data suggesting efficacy up to 3 years, but longer follow-up is needed to address whether these effects are finite.

Although reinnervation of afferent renal nerves has not been described, transplant models have shown anatomic regrowth of efferent nerves; the impact of this efferent reinnervation on blood pressure remains unclear. Experience from renal transplantation also shows that implanted kidneys that are “denervated” can still maintain fluid and electrolyte regulation.

Follow-up renal imaging in the Symplicity trials did not indicate renal artery stenosis at the sites of denervation in patients who underwent the procedure. Animal studies using the Symplicity catheter system showed renal nerve injury as evidenced by nerve fibrosis and thickened epineurium and perineurium, but no significant smooth muscle hyperplasia, arterial stenosis, or thrombosis by angiography or histology at 6 months.²⁰

WHAT ARE THE RISKS?

Adverse effects that were noted in the short term are detailed under discussion of the trials and in Table 2.

Long-term adverse events in the Symplicity HTN-2 trial that required hospitalization were reported in five patients in the denervation group and three patients in the control group (Table 2). These included transient ischemic attacks, hypertensive crises, hypotensive episodes, angina, and nausea.

Renal function was maintained for the duration of both trials, and details regarding eGFR change have been described above under the discussion of the trials.

Diffuse visceral pain at the time of the procedure is reported as an expected occurrence, managed with intravenous analgesic medications.

DOES SYMPATHETIC DENERVATION HAVE A ROLE IN OTHER CONDITIONS?

Interestingly, other sympathetically driven diseases, such as diabetes mellitus and polycystic ovary syndrome, may prove to be targets for this therapy in the future.²¹

Mahfoud et al²² conducted a pilot study in 37 patients with resistant hypertension undergoing renal denervation and 13 control patients. Fasting glucose levels declined from 118 ± 3.4 mg/dL to 108 ± 3.8 mg/dL after 3 months in the intervention group (P = .039), compared with no change in the control group. Insulin and C-peptide levels were also lower in the intervention group. The reported improvement in glucose metabolism and insulin sensitivity suggests that the beneficial effects of this procedure may extend beyond blood pressure reduction.

Brandt et al²³ reported regression of left ventricular hypertrophy and significantly improved cardiac functional parameters, including increase in ejection fraction and improved diastolic dysfunction, in a study of 46 patients who underwent renal denervation. This findings suggests a potential beneficial effect on cardiac remodeling.

Witkowski et al²⁴ reported lowering of blood pressure in 10 patients with refractory hypertension and obstructive sleep apnea who underwent renal denervation, which was accompanied by improvement of sleep apnea severity.

Ukena et al²⁵ reported reduction in ventricular tachyarrhythmias in two patients with congestive heart failure who had therapy-resistant electrical storm.

A recent pilot study in 15 patients with stage 3 and 4 chronic kidney disease (mean eGFR 31 mL/min/1.73 m²) showed significantly improved office blood pressure control up to 1 year, restoration of nocturnal dipping on 24-hour monitoring, as well as a nonsignificant trend towards increased hemoglobin levels and decreased proteinuria. No additional deterioration of renal function was reported in these patients (2 patients had renal function assessed up to 1 year).²⁶

Thus, the benefits of this procedure may extend to other diseases that have a common underlying thread of elevated sympathetic activity, by targeting the “sympathorenal” axis.²⁷

GUARDED OPTIMISM AND FUTURE DIRECTIONS

Given the well-known cardiovascular risks and health care costs associated with uncontrolled hypertension and the continued challenge that physicians face in managing it, novel therapies such as renal denervation may provide an adjunct to existing pharmacologic approaches.

While there is certainly cause for guarded optimism, especially with the striking blood pressure-lowering results seen in trials so far, it should be kept in mind that the mechanisms leading to the hypertensive response are complex and multifactorial, and further understanding of this therapy with long-term follow-up is needed. A comparison study with spironolactone, which is increasingly being used to treat resistant hypertension (in the absence of a diagnosis of primary aldosteronism)^28,29 would help to further establish the role of this procedure.

Studies of carotid baroreceptor stimulation via an implantable device have shown sustained reduction in blood pressure in patients with resistant hypertension. A study comparing this technique with renal denervation for efficacy and safety end points could be considered in the future.^30,31

The planned Symplicity HTN-3 study in the United States will be the largest trial to date, with a targeted randomization of more than 500 patients using strict enrollment criteria, including the use of maximally tolerated doses of diuretics and more focus on the use of ambulatory blood pressure monitoring and on the blinding of participants. This study will help further analysis of this technology in a more diverse population.^32,33

Future studies should be designed to clarify pathophysiologic mechanisms, patient selection criteria, effects on target organ damage, and efficacy in patients with chronic kidney disease, obesity, congestive heart failure, and in less severe forms of hypertension.

A CALL FOR PARTICIPANTS IN A CLINICAL TRIAL

The Departments of Cardiology and Nephrology and Hypertension at Cleveland Clinic are currently enrolling patients in the Symplicity HTN-3 trial. For more information, please contact George Thomas, MD ([email protected]), or Mehdi Shishehbor, DO, MPH ([email protected]), or visit www.symplifybptrial.com.