Q: Which is the most likely diagnosis?

• Idiopathic thrombocytopenic purpura
• Vitamin C deficiency (scurvy)
• Kaposi sarcoma not related to human immunodeficiency virus (HIV) infection
• Henoch-Schönlein purpura
• Polyarteritis nodosa

A: The correct answer is Henoch-Schönlein purpura.

Idiopathic thrombocytopenic purpura is an autoimmune disease caused by specific antibodies against platelet-membrane glycoproteins. It is characterized by thrombocytopenia not explainable by contact with toxic substances or by other causes. Along with nonpalpable purpura, other common signs are epistaxis, gingival bleeding, menorrhagia, and retinal hemorrhage.

Scurvy is an uncommon deficiency of ascorbic acid (vitamin C). The elderly and alcoholics are at higher risk, as they do not take in enough vitamin C in the diet. Patients usually show perifollicular hemorrhages of the skin and mucous membranes, typically petechial hemorrhage or ecchymosis of the gums around the upper incisors. Other cutaneous signs are follicular hyperkeratosis on the forearms, small corkscrew hairs, and sicca syndrome, which is more common in adults.

Non-HIV Kaposi sarcoma usually affects elderly patients, with pink, red, or brown papules or nodules on the legs and, less commonly, on the head and neck. Histopathologic examination shows newly formed irregular blood vessels with an inflammatory infiltrate of plasma cells and lymphocytes; immunohistochemical human herpes virus staining is usually positive.

Polyarteritis nodosa is a systemic vasculitis that affects medium or small arteries with necrotizing inflammation; renal glomeruli and arterioles, capillaries, and venules are unaffected. Skin manifestations include palpable purpura, livedo reticularis, ulcers, and distal gangrene. The condition also usually affects the kidneys, the heart, and the musculoskeletal and nervous systems.

A SYSTEMIC VASCULITIS

Henoch-Schönlein purpura is a systemic vasculitis affecting the skin, gastrointestinal tract, kidneys, and joints. Palpable purpura and joint pain are the most common and consistent presenting symptoms. The kidneys are affected in about one-third of children and in 60% of adults, and this is the major factor determining the long-term outcome.¹

In our patient, laboratory testing that included a complete blood cell count, biochemical testing (including IgA levels), urinary sediment, and coagulation studies showed no abnormalities except elevations of the erythrocyte sedimentation rate and the concentration of C-reactive protein (an acute-phase reactant). These can be normal in some patients. Renal involvement was also not present.

DIAGNOSIS

The diagnosis relies on clinical manifestations. Because Henoch-Schönlein purpura is less common in adults, biopsy plays a more important role in establishing the diagnosis in this age group, and it does this by demonstrating leukocytoclastic vasculitis with a predominance of IgA deposition under immunofluorescence. Recent studies in children showed that an elevated IgA concentration along with reduced IgM levels was associated with a higher rate of severe complications.² However, depending on the age of the biopsied lesion, IgA may not be detected.

TREATMENT DIRECTED AT SYMPTOMS

Our patient received oral corticosteroids 0.5 mg/kg per day for 20 days, and the lesions resolved by 4 weeks.

Management of Henoch-Schönlein purpura is mainly directed at the symptoms, with oral hydration and nonsteroidal anti-inflammatory drugs. For severe cases, a short course of corticosteroids (0.5–1 mg/kg) may be used.

Although no controlled clinical trial has proven that Henoch-Schönlein purpura responds to corticosteroids, colchicine, or other drugs, corticosteroids are used most often, especially in patients with renal disease. Patients with severe renal insufficiency, abdominal pain, joint involvement, or bleeding should be hospitalized. Plasmapheresis³ has been used in severe cases.

HENOCH-SCHÖNLEIN PURPURA AND MALIGNANCY

During a follow-up evaluation 1 month later, our patient was diagnosed with adenocarcinoma of the breast. This highlights the value of a workup for cancer in adults with cutaneous vasculitis.

Cutaneous vasculitis can represent a paraneoplastic syndrome associated with a malignant tumor. The pathophysiology of this association is unclear, but one proposed mechanism is the exaggerated production of antibodies that react against tumor neoantigens, leading to the formation of immune complexes, or that occasionally recognize endothelial cells because of similarities with tumor antigens. Another theory is that abnormally high levels of inflammatory cytokines are produced by neoplastic cells or in response to decreased immune complex clearance.

Yet another theory is that hyperviscosity of the blood, seen in some cancers, increases the contact time for deposition of immune complexes and causes endothelial damage. Drugs used to treat cancer have also been reported to produce Henoch-Schönlein purpura.⁴

Although hematologic malignancy is three to five times more common than solid tumors in patients with small-vessel vasculitis, the disease has been associated with solid tumors of the liver, skin, colon, and breast in adults over age 40.⁵ An evaluation for neoplasm is therefore reasonable in adults with Henoch-Schönlein purpura, as is an evaluation for tumor recurrence or metastasis if the patient has been previously treated for a malignant tumor.

Q: Which is the most likely diagnosis?

• Idiopathic thrombocytopenic purpura
• Vitamin C deficiency (scurvy)
• Kaposi sarcoma not related to human immunodeficiency virus (HIV) infection
• Henoch-Schönlein purpura
• Polyarteritis nodosa

A: The correct answer is Henoch-Schönlein purpura.

Idiopathic thrombocytopenic purpura is an autoimmune disease caused by specific antibodies against platelet-membrane glycoproteins. It is characterized by thrombocytopenia not explainable by contact with toxic substances or by other causes. Along with nonpalpable purpura, other common signs are epistaxis, gingival bleeding, menorrhagia, and retinal hemorrhage.

Scurvy is an uncommon deficiency of ascorbic acid (vitamin C). The elderly and alcoholics are at higher risk, as they do not take in enough vitamin C in the diet. Patients usually show perifollicular hemorrhages of the skin and mucous membranes, typically petechial hemorrhage or ecchymosis of the gums around the upper incisors. Other cutaneous signs are follicular hyperkeratosis on the forearms, small corkscrew hairs, and sicca syndrome, which is more common in adults.

Non-HIV Kaposi sarcoma usually affects elderly patients, with pink, red, or brown papules or nodules on the legs and, less commonly, on the head and neck. Histopathologic examination shows newly formed irregular blood vessels with an inflammatory infiltrate of plasma cells and lymphocytes; immunohistochemical human herpes virus staining is usually positive.

Polyarteritis nodosa is a systemic vasculitis that affects medium or small arteries with necrotizing inflammation; renal glomeruli and arterioles, capillaries, and venules are unaffected. Skin manifestations include palpable purpura, livedo reticularis, ulcers, and distal gangrene. The condition also usually affects the kidneys, the heart, and the musculoskeletal and nervous systems.

A SYSTEMIC VASCULITIS

Henoch-Schönlein purpura is a systemic vasculitis affecting the skin, gastrointestinal tract, kidneys, and joints. Palpable purpura and joint pain are the most common and consistent presenting symptoms. The kidneys are affected in about one-third of children and in 60% of adults, and this is the major factor determining the long-term outcome.¹

In our patient, laboratory testing that included a complete blood cell count, biochemical testing (including IgA levels), urinary sediment, and coagulation studies showed no abnormalities except elevations of the erythrocyte sedimentation rate and the concentration of C-reactive protein (an acute-phase reactant). These can be normal in some patients. Renal involvement was also not present.

DIAGNOSIS

The diagnosis relies on clinical manifestations. Because Henoch-Schönlein purpura is less common in adults, biopsy plays a more important role in establishing the diagnosis in this age group, and it does this by demonstrating leukocytoclastic vasculitis with a predominance of IgA deposition under immunofluorescence. Recent studies in children showed that an elevated IgA concentration along with reduced IgM levels was associated with a higher rate of severe complications.² However, depending on the age of the biopsied lesion, IgA may not be detected.

TREATMENT DIRECTED AT SYMPTOMS

Our patient received oral corticosteroids 0.5 mg/kg per day for 20 days, and the lesions resolved by 4 weeks.

Management of Henoch-Schönlein purpura is mainly directed at the symptoms, with oral hydration and nonsteroidal anti-inflammatory drugs. For severe cases, a short course of corticosteroids (0.5–1 mg/kg) may be used.

Although no controlled clinical trial has proven that Henoch-Schönlein purpura responds to corticosteroids, colchicine, or other drugs, corticosteroids are used most often, especially in patients with renal disease. Patients with severe renal insufficiency, abdominal pain, joint involvement, or bleeding should be hospitalized. Plasmapheresis³ has been used in severe cases.

HENOCH-SCHÖNLEIN PURPURA AND MALIGNANCY

During a follow-up evaluation 1 month later, our patient was diagnosed with adenocarcinoma of the breast. This highlights the value of a workup for cancer in adults with cutaneous vasculitis.

Cutaneous vasculitis can represent a paraneoplastic syndrome associated with a malignant tumor. The pathophysiology of this association is unclear, but one proposed mechanism is the exaggerated production of antibodies that react against tumor neoantigens, leading to the formation of immune complexes, or that occasionally recognize endothelial cells because of similarities with tumor antigens. Another theory is that abnormally high levels of inflammatory cytokines are produced by neoplastic cells or in response to decreased immune complex clearance.

Yet another theory is that hyperviscosity of the blood, seen in some cancers, increases the contact time for deposition of immune complexes and causes endothelial damage. Drugs used to treat cancer have also been reported to produce Henoch-Schönlein purpura.⁴

Although hematologic malignancy is three to five times more common than solid tumors in patients with small-vessel vasculitis, the disease has been associated with solid tumors of the liver, skin, colon, and breast in adults over age 40.⁵ An evaluation for neoplasm is therefore reasonable in adults with Henoch-Schönlein purpura, as is an evaluation for tumor recurrence or metastasis if the patient has been previously treated for a malignant tumor.

Q: Which is the most likely diagnosis?

• Idiopathic thrombocytopenic purpura
• Vitamin C deficiency (scurvy)
• Kaposi sarcoma not related to human immunodeficiency virus (HIV) infection
• Henoch-Schönlein purpura
• Polyarteritis nodosa

A: The correct answer is Henoch-Schönlein purpura.

Idiopathic thrombocytopenic purpura is an autoimmune disease caused by specific antibodies against platelet-membrane glycoproteins. It is characterized by thrombocytopenia not explainable by contact with toxic substances or by other causes. Along with nonpalpable purpura, other common signs are epistaxis, gingival bleeding, menorrhagia, and retinal hemorrhage.

Scurvy is an uncommon deficiency of ascorbic acid (vitamin C). The elderly and alcoholics are at higher risk, as they do not take in enough vitamin C in the diet. Patients usually show perifollicular hemorrhages of the skin and mucous membranes, typically petechial hemorrhage or ecchymosis of the gums around the upper incisors. Other cutaneous signs are follicular hyperkeratosis on the forearms, small corkscrew hairs, and sicca syndrome, which is more common in adults.

Non-HIV Kaposi sarcoma usually affects elderly patients, with pink, red, or brown papules or nodules on the legs and, less commonly, on the head and neck. Histopathologic examination shows newly formed irregular blood vessels with an inflammatory infiltrate of plasma cells and lymphocytes; immunohistochemical human herpes virus staining is usually positive.

Polyarteritis nodosa is a systemic vasculitis that affects medium or small arteries with necrotizing inflammation; renal glomeruli and arterioles, capillaries, and venules are unaffected. Skin manifestations include palpable purpura, livedo reticularis, ulcers, and distal gangrene. The condition also usually affects the kidneys, the heart, and the musculoskeletal and nervous systems.

A SYSTEMIC VASCULITIS

Henoch-Schönlein purpura is a systemic vasculitis affecting the skin, gastrointestinal tract, kidneys, and joints. Palpable purpura and joint pain are the most common and consistent presenting symptoms. The kidneys are affected in about one-third of children and in 60% of adults, and this is the major factor determining the long-term outcome.¹

In our patient, laboratory testing that included a complete blood cell count, biochemical testing (including IgA levels), urinary sediment, and coagulation studies showed no abnormalities except elevations of the erythrocyte sedimentation rate and the concentration of C-reactive protein (an acute-phase reactant). These can be normal in some patients. Renal involvement was also not present.

DIAGNOSIS

The diagnosis relies on clinical manifestations. Because Henoch-Schönlein purpura is less common in adults, biopsy plays a more important role in establishing the diagnosis in this age group, and it does this by demonstrating leukocytoclastic vasculitis with a predominance of IgA deposition under immunofluorescence. Recent studies in children showed that an elevated IgA concentration along with reduced IgM levels was associated with a higher rate of severe complications.² However, depending on the age of the biopsied lesion, IgA may not be detected.

TREATMENT DIRECTED AT SYMPTOMS

Our patient received oral corticosteroids 0.5 mg/kg per day for 20 days, and the lesions resolved by 4 weeks.

Management of Henoch-Schönlein purpura is mainly directed at the symptoms, with oral hydration and nonsteroidal anti-inflammatory drugs. For severe cases, a short course of corticosteroids (0.5–1 mg/kg) may be used.

Although no controlled clinical trial has proven that Henoch-Schönlein purpura responds to corticosteroids, colchicine, or other drugs, corticosteroids are used most often, especially in patients with renal disease. Patients with severe renal insufficiency, abdominal pain, joint involvement, or bleeding should be hospitalized. Plasmapheresis³ has been used in severe cases.

HENOCH-SCHÖNLEIN PURPURA AND MALIGNANCY

During a follow-up evaluation 1 month later, our patient was diagnosed with adenocarcinoma of the breast. This highlights the value of a workup for cancer in adults with cutaneous vasculitis.

Cutaneous vasculitis can represent a paraneoplastic syndrome associated with a malignant tumor. The pathophysiology of this association is unclear, but one proposed mechanism is the exaggerated production of antibodies that react against tumor neoantigens, leading to the formation of immune complexes, or that occasionally recognize endothelial cells because of similarities with tumor antigens. Another theory is that abnormally high levels of inflammatory cytokines are produced by neoplastic cells or in response to decreased immune complex clearance.

Yet another theory is that hyperviscosity of the blood, seen in some cancers, increases the contact time for deposition of immune complexes and causes endothelial damage. Drugs used to treat cancer have also been reported to produce Henoch-Schönlein purpura.⁴

Although hematologic malignancy is three to five times more common than solid tumors in patients with small-vessel vasculitis, the disease has been associated with solid tumors of the liver, skin, colon, and breast in adults over age 40.⁵ An evaluation for neoplasm is therefore reasonable in adults with Henoch-Schönlein purpura, as is an evaluation for tumor recurrence or metastasis if the patient has been previously treated for a malignant tumor.

A major placebo-controlled trial has found that a statin can reduce the risk of venous thromboembolism (VTE).¹

We do not recommend prescribing this class of drugs for this purpose until much more research has been done, and we certainly do not recommend substituting a statin for anticoagulant therapy in a patient at risk of VTE.

Nevertheless, we are excited by the latest findings, and we find comfort in knowing that if a patient is taking a statin for an approved indication, ie, reducing the risk of cardiovascular disease in a patient with hyperlipidemia or a previous cardiovascular event, the drug will also reduce the risk of VTE.

In the pages that follow, we describe and comment on what is known about the effect of statins on the risk of VTE.

ARTERIAL AND VENOUS THROMBOSIS: HOW ARE THEY LINKED?

The causes of arterial thrombosis may not be entirely distinct from those of deep vein thrombosis and pulmonary embolism, collectively referred to as VTE. Some studies have found that risk factors for arterial thrombosis overlap with those for VTE.^2–4 However, other studies have shown no association between venous and arterial events.^5–10

Hyperlipidemia, in particular, has been evaluated to see if it is a risk factor for VTE. As with other risk factors for arterial thrombosis, the data have been mixed, with some reports favoring an association with VTE and others not.^4,5,11 Even so, preventive strategies targeting arterial risk factors have shown promise in reducing VTE events.¹²

Although commonly used to treat hyperlipidemia, statins (3-hydroxy-3-methylglutaryl coenzyme A reductase inhibitors) are believed to reduce the incidence of thrombosis by a number of mechanisms¹³:

• Decreasing platelet aggregation
• Inhibiting expression of tissue factor and plasminogen activator inhibitor 1
• Increasing expression of tissue plasminogen activator
• Increasing expression of thrombomodulin, which can activate protein C and prevent thrombin-induced platelet and factor V activation and fibrinogen clotting.

STATINS AND VTE IN OBSERVATIONAL AND CASE-CONTROL STUDIES

In view of the multiple effects of statins, several studies have looked at whether these drugs reduce the occurrence of both arterial thrombosis and VTE.^14–19

Two prospective observational studies and four case-control studies found that statins reduced the risk of VTE by 20% to 60%.^14–19 Interestingly, two of the case-control studies found that antiplatelet therapy did not reduce the risk of VTE.^18,19

THE JUPITER STUDY

The Justification for the Use of Statins in Prevention: an Intervention Trial Evaluating Rosuvastatin (JUPITER) study primarily sought to determine if rosuvastatin (Crestor) 20 mg/day, compared with placebo, would reduce the rate of first major cardiovascular events.²² A prespecified secondary end point of the trial was VTE, making JUPITER the first randomized, placebo-controlled trial to specifically test whether statins prevent VTE.¹

Inclusion criteria: Normal LDL, high CRP

The study included men age 50 and older and women age 60 and older with no history of cardiovascular disease. In addition, their lowdensity lipoprotein (LDL) cholesterol levels had to be lower than 130 mg/dL (3.4 mmol/L), their triglyceride levels had to be lower than 500 mg/dL (5.6 mmol/L), and their highsensitivity C-reactive protein (hs-CRP) levels had to be 2.0 mg/L or higher.

Since high levels of hs-CRP, a marker of inflammation, predict cardiovascular events and since statins lower hs-CRP levels, the investigators hypothesized that people with elevated hs-CRP but without hyperlipidemia might benefit from statin treatment.²¹

Patients were excluded if they had received lipid-lowering therapy within 6 weeks of the trial screening, had diabetes mellitus or uncontrolled hypertension, were currently using postmenopausal hormone-replacement therapy, or had had cancer within the previous 5 years, except for certain skin cancers.

Candidates who complied well during a 4-week placebo run-in phase were randomly assigned to receive either rosuvastatin 20 mg daily (an intermediate dose) or a matching placebo. In all, 17,802 people were randomized. The two assigned groups appeared to be well matched.

Patients were to come in for visits twice a year for 60 months after randomization to be assessed for symptomatic deep venous thrombosis and pulmonary embolism. New cases of VTE were confirmed by imaging studies, by the initiation of anticoagulation therapy, or by death ascribed to pulmonary embolism.

Idiopathic VTE was classified as unprovoked if it occurred in the absence of trauma, hospitalization, or surgery within 3 months before the event, and in the absence of any diagnosed cancer within 3 months before and after the event. Provoked VTE events were those that occurred in a participant with cancer or when a precipitating event was associated with trauma, hospitalization, or surgery.

Rosuvastatin prevents heart attack, stroke

On the recommendation of the trial’s independent data and safety monitoring board, JUPITER was stopped early because the trial drug showed evidence of efficacy in preventing the combined primary end point of a first major cardiovascular event—ie, nonfatal myocardial infarction, nonfatal stroke, hospitalization for unstable angina, an arterial revascularization procedure, or confirmed death from a cardiovascular cause.²² (The cardiovascular outcomes of the JUPITER study were reviewed by Shishehbor and Hazen²³ in the January 2009 issue of the Cleveland Clinic Journal of Medicine; see doi:10.3949/ccjm.75a.08105).

Formal follow-up for the trial's primary and secondary efficacy end points ended then, but data on VTE continued to be collected until each patient’s closeout visit as part of a safety monitoring protocol. The last closeout visit occurred on August 20, 2008. The primary analysis focused on events occurring up to March 30, 2008, the date the study was stopped.

Secondary end point results: Rosuvastatin prevents VTE

At a median follow-up of 1.9 years, an episode of VTE had occurred in 94 (0.53%) of the 17,802 patients—34 in the rosuvastatin group and 60 in the placebo group.¹ This translates to 0.18 and 0.32 events per 100 person-years of follow-up in the rosuvastatin and placebo groups, respectively (hazard ratio for the rosuvastatin group 0.57, 95% confidence interval [CI] 0.37–0.86, P = .007).

Forty-four cases of VTE were classified as provoked and 50 cases were categorized as unprovoked. The risk reduction was statistically significant for provoked cases (hazard ratio 0.52, 95% CI 0.28–0.96, P = .03), but not for unprovoked events (hazard ratio 0.61, 95% CI 0.35–1.09, P = .09).

Subgroup analysis revealed no significant association between patient characteristics and the impact of rosuvastatin on the risk of a VTE event, but, as expected, more benefit was associated with higher baseline lipid levels.

STILL TOO SOON TO ADVISE ROUTINE STATIN USE TO PREVENT VTE

While the JUPITER trial data show an apparent benefit of statin use on the rate of VTE events, advising routine use of statins to prevent VTE is premature, for three main reasons.

Many must be treated to prevent one case of VTE. The number needed to treat (NNT) with rosuvastatin for 5 years to prevent either a case of VTE or a cardiovascular event was 21, and the NNT to prevent one cardiovascular event was 25. In a review of the two most recent case-control studies investigating the effects of statins on VTE,^18,19 Cushman²⁴ calculated that the NNT to prevent one VTE event each year was 333 for those age 75 and older. Though the Jupiter data did not provide the specific incidence of VTE at 1 year, except graphically, we can estimate that the NNT to prevent one VTE event at 1 year in the study is also very high.

Practically speaking, the perceived benefits of VTE prevention require large numbers to be treated, and the net clinical gain is still largely in preventing arterial events such as heart attack and stroke rather than VTE.

Statins, though safe, can still have adverse effects. The JUPITER study found a trend (albeit nonsignificant) toward more muscle complaints and elevations on liver function testing in apparently healthy persons taking a statin.²² Although severe complications of statin therapy such as rhabdomyolysis and elevations of creatine phosphokinase are rare, patients taking a statin have a 39% higher risk of an adverse event, most commonly myalgias or abnormalities on liver function testing.²⁵ Were statins to be given routinely to even more people than they are now, more adverse outcomes would be likely.

More study is needed. The JUPITER study did not address a high risk of VTE. In fact, the investigators provided no information as to the VTE history of those enrolled.

Clearly, statins should not be substituted for proven prophylaxis and anticoagulation without further investigation, especially for patients with recurrent deep venous thrombosis, hospitalized patients, postoperative patients, and other patients prone to VTE.

OUR VIEW

The JUPITER study is an important leap forward in adding to our knowledge of how to prevent VTE. For people with another indication for taking a statin (eg, a previous cardiovascular event, hyperlipidemia), it is helpful to know that their risk of VTE may be reduced without exposure to the risks of other kinds of conventional thromboprophylaxis.

We look forward to additional studies to elaborate on the benefits of statins in both the prevention and treatment of VTE for averagerisk and VTE-prone populations.

A major placebo-controlled trial has found that a statin can reduce the risk of venous thromboembolism (VTE).¹

We do not recommend prescribing this class of drugs for this purpose until much more research has been done, and we certainly do not recommend substituting a statin for anticoagulant therapy in a patient at risk of VTE.

Nevertheless, we are excited by the latest findings, and we find comfort in knowing that if a patient is taking a statin for an approved indication, ie, reducing the risk of cardiovascular disease in a patient with hyperlipidemia or a previous cardiovascular event, the drug will also reduce the risk of VTE.

In the pages that follow, we describe and comment on what is known about the effect of statins on the risk of VTE.

ARTERIAL AND VENOUS THROMBOSIS: HOW ARE THEY LINKED?

The causes of arterial thrombosis may not be entirely distinct from those of deep vein thrombosis and pulmonary embolism, collectively referred to as VTE. Some studies have found that risk factors for arterial thrombosis overlap with those for VTE.^2–4 However, other studies have shown no association between venous and arterial events.^5–10

Hyperlipidemia, in particular, has been evaluated to see if it is a risk factor for VTE. As with other risk factors for arterial thrombosis, the data have been mixed, with some reports favoring an association with VTE and others not.^4,5,11 Even so, preventive strategies targeting arterial risk factors have shown promise in reducing VTE events.¹²

Although commonly used to treat hyperlipidemia, statins (3-hydroxy-3-methylglutaryl coenzyme A reductase inhibitors) are believed to reduce the incidence of thrombosis by a number of mechanisms¹³:

• Decreasing platelet aggregation
• Inhibiting expression of tissue factor and plasminogen activator inhibitor 1
• Increasing expression of tissue plasminogen activator
• Increasing expression of thrombomodulin, which can activate protein C and prevent thrombin-induced platelet and factor V activation and fibrinogen clotting.

STATINS AND VTE IN OBSERVATIONAL AND CASE-CONTROL STUDIES

In view of the multiple effects of statins, several studies have looked at whether these drugs reduce the occurrence of both arterial thrombosis and VTE.^14–19

Two prospective observational studies and four case-control studies found that statins reduced the risk of VTE by 20% to 60%.^14–19 Interestingly, two of the case-control studies found that antiplatelet therapy did not reduce the risk of VTE.^18,19

THE JUPITER STUDY

The Justification for the Use of Statins in Prevention: an Intervention Trial Evaluating Rosuvastatin (JUPITER) study primarily sought to determine if rosuvastatin (Crestor) 20 mg/day, compared with placebo, would reduce the rate of first major cardiovascular events.²² A prespecified secondary end point of the trial was VTE, making JUPITER the first randomized, placebo-controlled trial to specifically test whether statins prevent VTE.¹

Inclusion criteria: Normal LDL, high CRP

The study included men age 50 and older and women age 60 and older with no history of cardiovascular disease. In addition, their lowdensity lipoprotein (LDL) cholesterol levels had to be lower than 130 mg/dL (3.4 mmol/L), their triglyceride levels had to be lower than 500 mg/dL (5.6 mmol/L), and their highsensitivity C-reactive protein (hs-CRP) levels had to be 2.0 mg/L or higher.

Since high levels of hs-CRP, a marker of inflammation, predict cardiovascular events and since statins lower hs-CRP levels, the investigators hypothesized that people with elevated hs-CRP but without hyperlipidemia might benefit from statin treatment.²¹

Patients were excluded if they had received lipid-lowering therapy within 6 weeks of the trial screening, had diabetes mellitus or uncontrolled hypertension, were currently using postmenopausal hormone-replacement therapy, or had had cancer within the previous 5 years, except for certain skin cancers.

Candidates who complied well during a 4-week placebo run-in phase were randomly assigned to receive either rosuvastatin 20 mg daily (an intermediate dose) or a matching placebo. In all, 17,802 people were randomized. The two assigned groups appeared to be well matched.

Patients were to come in for visits twice a year for 60 months after randomization to be assessed for symptomatic deep venous thrombosis and pulmonary embolism. New cases of VTE were confirmed by imaging studies, by the initiation of anticoagulation therapy, or by death ascribed to pulmonary embolism.

Idiopathic VTE was classified as unprovoked if it occurred in the absence of trauma, hospitalization, or surgery within 3 months before the event, and in the absence of any diagnosed cancer within 3 months before and after the event. Provoked VTE events were those that occurred in a participant with cancer or when a precipitating event was associated with trauma, hospitalization, or surgery.

Rosuvastatin prevents heart attack, stroke

On the recommendation of the trial’s independent data and safety monitoring board, JUPITER was stopped early because the trial drug showed evidence of efficacy in preventing the combined primary end point of a first major cardiovascular event—ie, nonfatal myocardial infarction, nonfatal stroke, hospitalization for unstable angina, an arterial revascularization procedure, or confirmed death from a cardiovascular cause.²² (The cardiovascular outcomes of the JUPITER study were reviewed by Shishehbor and Hazen²³ in the January 2009 issue of the Cleveland Clinic Journal of Medicine; see doi:10.3949/ccjm.75a.08105).

Formal follow-up for the trial's primary and secondary efficacy end points ended then, but data on VTE continued to be collected until each patient’s closeout visit as part of a safety monitoring protocol. The last closeout visit occurred on August 20, 2008. The primary analysis focused on events occurring up to March 30, 2008, the date the study was stopped.

Secondary end point results: Rosuvastatin prevents VTE

At a median follow-up of 1.9 years, an episode of VTE had occurred in 94 (0.53%) of the 17,802 patients—34 in the rosuvastatin group and 60 in the placebo group.¹ This translates to 0.18 and 0.32 events per 100 person-years of follow-up in the rosuvastatin and placebo groups, respectively (hazard ratio for the rosuvastatin group 0.57, 95% confidence interval [CI] 0.37–0.86, P = .007).

Forty-four cases of VTE were classified as provoked and 50 cases were categorized as unprovoked. The risk reduction was statistically significant for provoked cases (hazard ratio 0.52, 95% CI 0.28–0.96, P = .03), but not for unprovoked events (hazard ratio 0.61, 95% CI 0.35–1.09, P = .09).

Subgroup analysis revealed no significant association between patient characteristics and the impact of rosuvastatin on the risk of a VTE event, but, as expected, more benefit was associated with higher baseline lipid levels.

STILL TOO SOON TO ADVISE ROUTINE STATIN USE TO PREVENT VTE

While the JUPITER trial data show an apparent benefit of statin use on the rate of VTE events, advising routine use of statins to prevent VTE is premature, for three main reasons.

Many must be treated to prevent one case of VTE. The number needed to treat (NNT) with rosuvastatin for 5 years to prevent either a case of VTE or a cardiovascular event was 21, and the NNT to prevent one cardiovascular event was 25. In a review of the two most recent case-control studies investigating the effects of statins on VTE,^18,19 Cushman²⁴ calculated that the NNT to prevent one VTE event each year was 333 for those age 75 and older. Though the Jupiter data did not provide the specific incidence of VTE at 1 year, except graphically, we can estimate that the NNT to prevent one VTE event at 1 year in the study is also very high.

Practically speaking, the perceived benefits of VTE prevention require large numbers to be treated, and the net clinical gain is still largely in preventing arterial events such as heart attack and stroke rather than VTE.

Statins, though safe, can still have adverse effects. The JUPITER study found a trend (albeit nonsignificant) toward more muscle complaints and elevations on liver function testing in apparently healthy persons taking a statin.²² Although severe complications of statin therapy such as rhabdomyolysis and elevations of creatine phosphokinase are rare, patients taking a statin have a 39% higher risk of an adverse event, most commonly myalgias or abnormalities on liver function testing.²⁵ Were statins to be given routinely to even more people than they are now, more adverse outcomes would be likely.

More study is needed. The JUPITER study did not address a high risk of VTE. In fact, the investigators provided no information as to the VTE history of those enrolled.

Clearly, statins should not be substituted for proven prophylaxis and anticoagulation without further investigation, especially for patients with recurrent deep venous thrombosis, hospitalized patients, postoperative patients, and other patients prone to VTE.

OUR VIEW

The JUPITER study is an important leap forward in adding to our knowledge of how to prevent VTE. For people with another indication for taking a statin (eg, a previous cardiovascular event, hyperlipidemia), it is helpful to know that their risk of VTE may be reduced without exposure to the risks of other kinds of conventional thromboprophylaxis.

We look forward to additional studies to elaborate on the benefits of statins in both the prevention and treatment of VTE for averagerisk and VTE-prone populations.

A major placebo-controlled trial has found that a statin can reduce the risk of venous thromboembolism (VTE).¹

We do not recommend prescribing this class of drugs for this purpose until much more research has been done, and we certainly do not recommend substituting a statin for anticoagulant therapy in a patient at risk of VTE.

Nevertheless, we are excited by the latest findings, and we find comfort in knowing that if a patient is taking a statin for an approved indication, ie, reducing the risk of cardiovascular disease in a patient with hyperlipidemia or a previous cardiovascular event, the drug will also reduce the risk of VTE.

In the pages that follow, we describe and comment on what is known about the effect of statins on the risk of VTE.

ARTERIAL AND VENOUS THROMBOSIS: HOW ARE THEY LINKED?

The causes of arterial thrombosis may not be entirely distinct from those of deep vein thrombosis and pulmonary embolism, collectively referred to as VTE. Some studies have found that risk factors for arterial thrombosis overlap with those for VTE.^2–4 However, other studies have shown no association between venous and arterial events.^5–10

Hyperlipidemia, in particular, has been evaluated to see if it is a risk factor for VTE. As with other risk factors for arterial thrombosis, the data have been mixed, with some reports favoring an association with VTE and others not.^4,5,11 Even so, preventive strategies targeting arterial risk factors have shown promise in reducing VTE events.¹²

Although commonly used to treat hyperlipidemia, statins (3-hydroxy-3-methylglutaryl coenzyme A reductase inhibitors) are believed to reduce the incidence of thrombosis by a number of mechanisms¹³:

• Decreasing platelet aggregation
• Inhibiting expression of tissue factor and plasminogen activator inhibitor 1
• Increasing expression of tissue plasminogen activator
• Increasing expression of thrombomodulin, which can activate protein C and prevent thrombin-induced platelet and factor V activation and fibrinogen clotting.

STATINS AND VTE IN OBSERVATIONAL AND CASE-CONTROL STUDIES

In view of the multiple effects of statins, several studies have looked at whether these drugs reduce the occurrence of both arterial thrombosis and VTE.^14–19

Two prospective observational studies and four case-control studies found that statins reduced the risk of VTE by 20% to 60%.^14–19 Interestingly, two of the case-control studies found that antiplatelet therapy did not reduce the risk of VTE.^18,19

THE JUPITER STUDY

The Justification for the Use of Statins in Prevention: an Intervention Trial Evaluating Rosuvastatin (JUPITER) study primarily sought to determine if rosuvastatin (Crestor) 20 mg/day, compared with placebo, would reduce the rate of first major cardiovascular events.²² A prespecified secondary end point of the trial was VTE, making JUPITER the first randomized, placebo-controlled trial to specifically test whether statins prevent VTE.¹

Inclusion criteria: Normal LDL, high CRP

The study included men age 50 and older and women age 60 and older with no history of cardiovascular disease. In addition, their lowdensity lipoprotein (LDL) cholesterol levels had to be lower than 130 mg/dL (3.4 mmol/L), their triglyceride levels had to be lower than 500 mg/dL (5.6 mmol/L), and their highsensitivity C-reactive protein (hs-CRP) levels had to be 2.0 mg/L or higher.

Since high levels of hs-CRP, a marker of inflammation, predict cardiovascular events and since statins lower hs-CRP levels, the investigators hypothesized that people with elevated hs-CRP but without hyperlipidemia might benefit from statin treatment.²¹

Patients were excluded if they had received lipid-lowering therapy within 6 weeks of the trial screening, had diabetes mellitus or uncontrolled hypertension, were currently using postmenopausal hormone-replacement therapy, or had had cancer within the previous 5 years, except for certain skin cancers.

Candidates who complied well during a 4-week placebo run-in phase were randomly assigned to receive either rosuvastatin 20 mg daily (an intermediate dose) or a matching placebo. In all, 17,802 people were randomized. The two assigned groups appeared to be well matched.

Patients were to come in for visits twice a year for 60 months after randomization to be assessed for symptomatic deep venous thrombosis and pulmonary embolism. New cases of VTE were confirmed by imaging studies, by the initiation of anticoagulation therapy, or by death ascribed to pulmonary embolism.

Idiopathic VTE was classified as unprovoked if it occurred in the absence of trauma, hospitalization, or surgery within 3 months before the event, and in the absence of any diagnosed cancer within 3 months before and after the event. Provoked VTE events were those that occurred in a participant with cancer or when a precipitating event was associated with trauma, hospitalization, or surgery.

Rosuvastatin prevents heart attack, stroke

On the recommendation of the trial’s independent data and safety monitoring board, JUPITER was stopped early because the trial drug showed evidence of efficacy in preventing the combined primary end point of a first major cardiovascular event—ie, nonfatal myocardial infarction, nonfatal stroke, hospitalization for unstable angina, an arterial revascularization procedure, or confirmed death from a cardiovascular cause.²² (The cardiovascular outcomes of the JUPITER study were reviewed by Shishehbor and Hazen²³ in the January 2009 issue of the Cleveland Clinic Journal of Medicine; see doi:10.3949/ccjm.75a.08105).

Formal follow-up for the trial's primary and secondary efficacy end points ended then, but data on VTE continued to be collected until each patient’s closeout visit as part of a safety monitoring protocol. The last closeout visit occurred on August 20, 2008. The primary analysis focused on events occurring up to March 30, 2008, the date the study was stopped.

Secondary end point results: Rosuvastatin prevents VTE

At a median follow-up of 1.9 years, an episode of VTE had occurred in 94 (0.53%) of the 17,802 patients—34 in the rosuvastatin group and 60 in the placebo group.¹ This translates to 0.18 and 0.32 events per 100 person-years of follow-up in the rosuvastatin and placebo groups, respectively (hazard ratio for the rosuvastatin group 0.57, 95% confidence interval [CI] 0.37–0.86, P = .007).

Forty-four cases of VTE were classified as provoked and 50 cases were categorized as unprovoked. The risk reduction was statistically significant for provoked cases (hazard ratio 0.52, 95% CI 0.28–0.96, P = .03), but not for unprovoked events (hazard ratio 0.61, 95% CI 0.35–1.09, P = .09).

Subgroup analysis revealed no significant association between patient characteristics and the impact of rosuvastatin on the risk of a VTE event, but, as expected, more benefit was associated with higher baseline lipid levels.

STILL TOO SOON TO ADVISE ROUTINE STATIN USE TO PREVENT VTE

While the JUPITER trial data show an apparent benefit of statin use on the rate of VTE events, advising routine use of statins to prevent VTE is premature, for three main reasons.

Many must be treated to prevent one case of VTE. The number needed to treat (NNT) with rosuvastatin for 5 years to prevent either a case of VTE or a cardiovascular event was 21, and the NNT to prevent one cardiovascular event was 25. In a review of the two most recent case-control studies investigating the effects of statins on VTE,^18,19 Cushman²⁴ calculated that the NNT to prevent one VTE event each year was 333 for those age 75 and older. Though the Jupiter data did not provide the specific incidence of VTE at 1 year, except graphically, we can estimate that the NNT to prevent one VTE event at 1 year in the study is also very high.

Practically speaking, the perceived benefits of VTE prevention require large numbers to be treated, and the net clinical gain is still largely in preventing arterial events such as heart attack and stroke rather than VTE.

Statins, though safe, can still have adverse effects. The JUPITER study found a trend (albeit nonsignificant) toward more muscle complaints and elevations on liver function testing in apparently healthy persons taking a statin.²² Although severe complications of statin therapy such as rhabdomyolysis and elevations of creatine phosphokinase are rare, patients taking a statin have a 39% higher risk of an adverse event, most commonly myalgias or abnormalities on liver function testing.²⁵ Were statins to be given routinely to even more people than they are now, more adverse outcomes would be likely.

More study is needed. The JUPITER study did not address a high risk of VTE. In fact, the investigators provided no information as to the VTE history of those enrolled.

Clearly, statins should not be substituted for proven prophylaxis and anticoagulation without further investigation, especially for patients with recurrent deep venous thrombosis, hospitalized patients, postoperative patients, and other patients prone to VTE.

OUR VIEW

The JUPITER study is an important leap forward in adding to our knowledge of how to prevent VTE. For people with another indication for taking a statin (eg, a previous cardiovascular event, hyperlipidemia), it is helpful to know that their risk of VTE may be reduced without exposure to the risks of other kinds of conventional thromboprophylaxis.

We look forward to additional studies to elaborate on the benefits of statins in both the prevention and treatment of VTE for averagerisk and VTE-prone populations.

The role of stenting for atherosclerotic renal artery stenosis is hotly debated among different specialties.^1,2 If we may generalize a bit, interventionalists (cardiologists, interventional radiologists, vascular surgeons, and vascular medicine specialists) have been in favor of liberal use of stenting, and nephrologists often favor medical therapy alone. And as with all controversial issues, each group feels rather strongly about its position.

Because few prospective randomized trials have been completed, the management of atherosclerotic renal artery stenosis has been guided by retrospective studies and case series. ³

See related article

In this issue of the Cleveland Clinic Journal of Medicine, Dr. James Simon⁴ provides an excellent overview of the prevalence, natural history, and clinical presentation of atherosclerotic renal artery stenosis. In addition, he does an admirable job of reviewing the available prospective randomized trials and providing editorial commentary about the role of the various specialists in the management of renal artery disease. And while the title of his paper says that it is “time to be less aggressive,” Dr. Simon ultimately comes to the same conclusions that we do⁵ on the indications for renal artery stenting (see Table 3 of Dr. Simon’s article), which are those of the multidisciplinary 2006 American College of Cardiology/American Heart Association guidelines on the management of peripheral artery disease.³

So what then is all the controversy about? We all agree that prospective randomized trials that provide class I, level A evidence impart the only unbiased scientific information on the best treatment strategy for patients with renal artery disease. The basic controversial issue is the interpretation of these trials. We contend that the three randomized trials of stenting vs medical therapy published so far^6–8 (see below) are so seriously flawed that it is impossible to make treatment decisions based on their results.

Since these trials were published in wellrespected journals, their results are often taken as gospel. However, careful review of each of these will reveal the flaws in study design and implementation.

THE DRASTIC TRIAL

In the Dutch Renal Artery Stenosis Intervention Cooperative (DRASTIC) trial,⁶ 106 patients with renal artery stenosis and hypertension (diastolic blood pressure > 95 mm Hg) despite treatment with two antihypertensive medications were randomly assigned to either renal angioplasty (n = 56) or drug therapy (n = 50).

Authors’ conclusions

“In the treatment of patients with hypertension and renal-artery stenosis, angioplasty has little advantage over antihypertensive-drug therapy.”⁶

Four serious problems

As we discussed in a letter to the editor of the New England Journal of Medicine on August 10, 2000, this study had four serious problems that invalidate its authors’ conclusions.⁹

The sample size was insufficient to detect a significant difference between treatment groups. In other words, the chance of a type 2 statistical error is high.

Balloon angioplasty without stenting was used as the method of revascularization. Experts now recognize that stenting is required for renal artery intervention to have a durable result.^3,5

Renal artery stenosis was defined as greater than 50% stenosis. This allowed a large number of patients to enter the trial who had hemodynamically and clinically insignificant lesions. Most clinicians believe that stenosis of less than 70% is not hemodynamically important.^5,10,11

Twenty-two of the 50 patients randomized to medical therapy crossed over to the angioplasty group because their blood pressure became difficult to control. In other words, 44% of the patients in the medical group underwent angioplasty, an astounding percentage in an intention-to-treat analysis comparing one therapy with another.

Despite these serious flaws, the results of DRASTIC influenced therapy for years after its publication.

THE STAR TRIAL

In the Stent Placement in Patients With Atherosclerotic Renal Artery Stenosis and Impaired Renal Function (STAR) trial,⁷ 140 patients with a creatinine clearance of less than 80 mL/min/1.73m², renal artery stenosis greater than 50%, and well-controlled blood pressure were randomized to either renal artery stenting plus medical therapy (n = 64) or medical therapy alone (n = 76). The primary end point was a 20% or greater decrease in creatinine clearance. Secondary end points included measures of safety and cardiovascular morbidity and mortality.

Authors’ conclusions

“Stent placement with medical treatment had no clear effect on progression of impaired renal function but led to a small number of significant procedure-related complications. The study findings favor a conservative approach to patients with [atherosclerotic renal artery stenosis], focused on cardiovascular risk factor management and avoiding stenting.”⁷

Serious flaws

A number of serious flaws render this study uninterpretable.

Mild renal artery stenosis. At least 33% of the patients in the study had mild renal artery stenosis (50%–70%), and 12 (19%) of the 64 patients in the group randomized to stenting actually had stenosis of less than 50%. How can one expect there to be a benefit to stenting in patients with mild (and hemodynamically insignificant) renal artery stenosis? This is especially true when the primary end point is a change in renal function.

More than half of the patients had unilateral disease. It seems intuitive that if one were to plan a trial with a primary end point of a change in renal function, only patients with bilateral renal artery stenosis of greater than 70% or with stenosis of greater than 70% to a solitary functioning kidney would be included. One would not expect that patients with unilateral disease and a stenosis of less than 70% would derive any benefit from revascularization.

Not all “stent” patients received stents. All of the patients in the medical group received medication and there were no crossovers. However, only 46 (72%) of the 64 patients randomized to stenting actually received a stent, while 18 (28%) did not. There were two technical failures, and 12 patients should not have been randomized because they had less than 50% stenosis on angiography and thus were not stented. Yet all 64 patients were analyzed (by intention to treat) in the stent group. With these numbers, one could predict that the results would be negative.

Like DRASTIC, this trial was underpowered, meaning that the chance of a type 2 statistical error is high. In fact, the editors of the Annals of Internal Medicine, in a note accompanying the article, cautioned that the study “was underpowered to provide a definitive estimate of efficacy.”⁷ If the study was underpowered to answer the question at hand, why was it deemed worthy of publication?

High complication rates. The periprocedural complication and death rates were much higher than in many other reports on renal artery stenting (see details below).⁵

THE ASTRAL TRIAL

In the Angioplasty and Stenting for Renal Artery Lesions (ASTRAL) trial,⁸ the primary outcome measure was the change in renal function over time as assessed by the mean slope of the reciprocal of the serum creatinine. In this trial, 806 patients with atherosclerotic renal artery stenosis were randomized to either stent-based revascularization combined with medical therapy or medical therapy alone.

Authors’ conclusions

“We found substantial risks but no evidence of a worthwhile clinical benefit from revascularization in patients with atherosclerotic renovascular disease.”⁸

Despite size, flaws remain

Unlike the other trials, ASTRAL had a sample size large enough to provide an answer. However, numerous flaws in study design and implementation invalidate its results for the overall population of patients with renal artery stenosis. The major flaws in ASTRAL were:

Selection bias. For a patient to be enrolled, the treating physician had to be undecided on whether the patient should undergo revascularization or medical management alone. However, the treatment of atherosclerotic renal artery stenosis is so controversial that physicians of different specialties cannot agree on the most effective treatment strategy for most patients.^1,2 Therefore, to exclude patients when their physicians were sure they needed or did not need renal artery revascularization is incomprehensible and introduces considerable selection bias into the trial design.

Normal renal function at baseline. The primary outcome was a change in renal function over time. Yet 25% of patients had normal renal function at the outset of the trial. In addition, a significant number had unilateral disease, and 41% had a stenosis less than 70%. What made the investigators think that stent implantation could possibly be shown to be beneficial if they entered patients into a renal function study who had near-normal renal function, unilateral disease, and mild renal artery stenosis? These are patients whose condition would not be expected to worsen with medical therapy nor to improve with stenting. Most clinicians would not consider stenting a patient to preserve renal function if the patient has unilateral mild renal artery stenosis.

There was no core laboratory to adjudicate the interpretation of the imaging studies. To determine the degree of stenosis of an artery in an accurate and unbiased fashion, a core laboratory must be used.

The reason this is so important is that visual assessment of the degree of stenosis on angiography is not accurate and almost always overestimates the degree of stenosis.^12,13 In a study assessing the physiologic importance of renal artery lesions, the lesion severity by visual estimation was 74.9% ± 11.5% (range 50%–90%), which exceeded the quantitative vascular angiographic lesion severity of 56.6% ± 10.8% (range 45%–76%).¹³

Therefore, in ASTRAL, some patients in the 50%–70% stenosis group (about 40% of patients entered) actually had a stenosis of less than 50%. And some patients in the group with stenosis greater than 70% had stenosis of less than 70%. This further illustrates that, for the most part, the patients in ASTRAL had mild to moderate renal artery stenosis.

A high adverse event rate. The major adverse event rate in the first 24 hours was 9%, whereas the usual rate is 2% or less.^14–18 Of the 280 patients in the revascularization group for whom data on adverse events were available at 1 month, 55 (20%) suffered a serious adverse event (including two patients who died) between 24 hours and 1 month after the procedure. This is in contrast to a major complication rate of 1.99% in five reports involving 727 patients.⁵

The trial centers were not high-volume centers. During the 7 years of recruitment, 24 centers (42% of all participating centers) randomized between one and five patients, and 32 centers (61% of all participating centers) randomized nine patients or fewer. This means that many participating centers randomized, on average, less than one patient per year! This was not a group of high-volume operators.

WILL CORAL GIVE US THE ANSWER?

The CORAL (Cardiovascular Outcomes in Renal Atherosclerotic Lesions) trial is under way.¹⁹ Enrollment was to have ended on January 31, 2010, and it will be several years before the data are available for analysis.

CORAL, a multicenter study funded in 2004 by the National Institutes of Health, will have randomized more than 900 patients with greater than 60% stenosis to optimal medical therapy alone or optimal medical therapy plus renal artery stenting. Inclusion criteria are a documented history of hypertension on two or more antihypertensive drugs or renal dysfunction, defined as stage 3 or greater chronic kidney disease based on the National Kidney Foundation classification (estimated glomerular filtration rate < 60 mL/min/1.73 m² calculated by the modified Modification of Diet in Renal Disease [MDRD] formula) and stenosis of 60% or greater but less than 100%, as assessed by a core laboratory. The primary end point is survival free of cardiovascular and renal adverse events, defined as a composite of cardiovascular or renal death, stroke, myocardial infarction, hospitalization for congestive heart failure, progressive renal insufficiency, or need for permanent renal replacement therapy.

We hope this trial will give us a clear answer to the question of whether renal artery stenting is beneficial in the patient population studied. One note of caution: recruitment for this trial was difficult and slow. Thus, there were a number of protocol amendments throughout the trial in order to make recruitment easier. Hopefully, this will not be a problem when analyzing the results.

WE ALL AGREE ON THE INDICATIONS FOR STENTING

So, are we really so far apart in our thinking? And is it really “time to be less aggressive” if we follow the precepts below?

All renal arteries with stenosis do not need to be (and should not be) stented.

There must be a good clinical indicationandhemodynamically significant stenosis. This means the degree of stenosis should be more than 70% on angiography or intravascular ultrasonography.

Indications for stenting. Until more data from compelling randomized trials become available, adherence to the American College of Cardiology/American Heart Association guidelines on indications for renal artery stenting is advised³:

• Hypertension: class IIa, level of evidence B. Percutaneous revascularization is reasonable for patients with hemodynamically significant renal artery stenosis and accelerated hypertension, resistant hypertension, and malignant hypertension.
• Preservation of renal function: class IIa, level of evidence B. Percutaneous revascularization is reasonable for patients with renal artery stenosis and progressive chronic kidney disease with bilateral renal artery stenosis or a stenosis to a solitary functioning kidney.
• Congestive heart failure: class I, level of evidence B. Percutaneous revascularization is indicated for patients with hemodynamically significant renal artery stenosis (ie, > 70% stenosis on angiography or intravascular ultrasonography) and recurrent, unexplained congestive heart failure or sudden, unexplained pulmonary edema.

The role of stenting for atherosclerotic renal artery stenosis is hotly debated among different specialties.^1,2 If we may generalize a bit, interventionalists (cardiologists, interventional radiologists, vascular surgeons, and vascular medicine specialists) have been in favor of liberal use of stenting, and nephrologists often favor medical therapy alone. And as with all controversial issues, each group feels rather strongly about its position.

Because few prospective randomized trials have been completed, the management of atherosclerotic renal artery stenosis has been guided by retrospective studies and case series. ³

See related article

In this issue of the Cleveland Clinic Journal of Medicine, Dr. James Simon⁴ provides an excellent overview of the prevalence, natural history, and clinical presentation of atherosclerotic renal artery stenosis. In addition, he does an admirable job of reviewing the available prospective randomized trials and providing editorial commentary about the role of the various specialists in the management of renal artery disease. And while the title of his paper says that it is “time to be less aggressive,” Dr. Simon ultimately comes to the same conclusions that we do⁵ on the indications for renal artery stenting (see Table 3 of Dr. Simon’s article), which are those of the multidisciplinary 2006 American College of Cardiology/American Heart Association guidelines on the management of peripheral artery disease.³

So what then is all the controversy about? We all agree that prospective randomized trials that provide class I, level A evidence impart the only unbiased scientific information on the best treatment strategy for patients with renal artery disease. The basic controversial issue is the interpretation of these trials. We contend that the three randomized trials of stenting vs medical therapy published so far^6–8 (see below) are so seriously flawed that it is impossible to make treatment decisions based on their results.

Since these trials were published in wellrespected journals, their results are often taken as gospel. However, careful review of each of these will reveal the flaws in study design and implementation.

THE DRASTIC TRIAL

In the Dutch Renal Artery Stenosis Intervention Cooperative (DRASTIC) trial,⁶ 106 patients with renal artery stenosis and hypertension (diastolic blood pressure > 95 mm Hg) despite treatment with two antihypertensive medications were randomly assigned to either renal angioplasty (n = 56) or drug therapy (n = 50).

Authors’ conclusions

“In the treatment of patients with hypertension and renal-artery stenosis, angioplasty has little advantage over antihypertensive-drug therapy.”⁶

Four serious problems

As we discussed in a letter to the editor of the New England Journal of Medicine on August 10, 2000, this study had four serious problems that invalidate its authors’ conclusions.⁹

The sample size was insufficient to detect a significant difference between treatment groups. In other words, the chance of a type 2 statistical error is high.

Balloon angioplasty without stenting was used as the method of revascularization. Experts now recognize that stenting is required for renal artery intervention to have a durable result.^3,5

Renal artery stenosis was defined as greater than 50% stenosis. This allowed a large number of patients to enter the trial who had hemodynamically and clinically insignificant lesions. Most clinicians believe that stenosis of less than 70% is not hemodynamically important.^5,10,11

Twenty-two of the 50 patients randomized to medical therapy crossed over to the angioplasty group because their blood pressure became difficult to control. In other words, 44% of the patients in the medical group underwent angioplasty, an astounding percentage in an intention-to-treat analysis comparing one therapy with another.

Despite these serious flaws, the results of DRASTIC influenced therapy for years after its publication.

THE STAR TRIAL

In the Stent Placement in Patients With Atherosclerotic Renal Artery Stenosis and Impaired Renal Function (STAR) trial,⁷ 140 patients with a creatinine clearance of less than 80 mL/min/1.73m², renal artery stenosis greater than 50%, and well-controlled blood pressure were randomized to either renal artery stenting plus medical therapy (n = 64) or medical therapy alone (n = 76). The primary end point was a 20% or greater decrease in creatinine clearance. Secondary end points included measures of safety and cardiovascular morbidity and mortality.

Authors’ conclusions

“Stent placement with medical treatment had no clear effect on progression of impaired renal function but led to a small number of significant procedure-related complications. The study findings favor a conservative approach to patients with [atherosclerotic renal artery stenosis], focused on cardiovascular risk factor management and avoiding stenting.”⁷

Serious flaws

A number of serious flaws render this study uninterpretable.

Mild renal artery stenosis. At least 33% of the patients in the study had mild renal artery stenosis (50%–70%), and 12 (19%) of the 64 patients in the group randomized to stenting actually had stenosis of less than 50%. How can one expect there to be a benefit to stenting in patients with mild (and hemodynamically insignificant) renal artery stenosis? This is especially true when the primary end point is a change in renal function.

More than half of the patients had unilateral disease. It seems intuitive that if one were to plan a trial with a primary end point of a change in renal function, only patients with bilateral renal artery stenosis of greater than 70% or with stenosis of greater than 70% to a solitary functioning kidney would be included. One would not expect that patients with unilateral disease and a stenosis of less than 70% would derive any benefit from revascularization.

Not all “stent” patients received stents. All of the patients in the medical group received medication and there were no crossovers. However, only 46 (72%) of the 64 patients randomized to stenting actually received a stent, while 18 (28%) did not. There were two technical failures, and 12 patients should not have been randomized because they had less than 50% stenosis on angiography and thus were not stented. Yet all 64 patients were analyzed (by intention to treat) in the stent group. With these numbers, one could predict that the results would be negative.

Like DRASTIC, this trial was underpowered, meaning that the chance of a type 2 statistical error is high. In fact, the editors of the Annals of Internal Medicine, in a note accompanying the article, cautioned that the study “was underpowered to provide a definitive estimate of efficacy.”⁷ If the study was underpowered to answer the question at hand, why was it deemed worthy of publication?

High complication rates. The periprocedural complication and death rates were much higher than in many other reports on renal artery stenting (see details below).⁵

THE ASTRAL TRIAL

In the Angioplasty and Stenting for Renal Artery Lesions (ASTRAL) trial,⁸ the primary outcome measure was the change in renal function over time as assessed by the mean slope of the reciprocal of the serum creatinine. In this trial, 806 patients with atherosclerotic renal artery stenosis were randomized to either stent-based revascularization combined with medical therapy or medical therapy alone.

Authors’ conclusions

“We found substantial risks but no evidence of a worthwhile clinical benefit from revascularization in patients with atherosclerotic renovascular disease.”⁸

Despite size, flaws remain

Unlike the other trials, ASTRAL had a sample size large enough to provide an answer. However, numerous flaws in study design and implementation invalidate its results for the overall population of patients with renal artery stenosis. The major flaws in ASTRAL were:

Selection bias. For a patient to be enrolled, the treating physician had to be undecided on whether the patient should undergo revascularization or medical management alone. However, the treatment of atherosclerotic renal artery stenosis is so controversial that physicians of different specialties cannot agree on the most effective treatment strategy for most patients.^1,2 Therefore, to exclude patients when their physicians were sure they needed or did not need renal artery revascularization is incomprehensible and introduces considerable selection bias into the trial design.

Normal renal function at baseline. The primary outcome was a change in renal function over time. Yet 25% of patients had normal renal function at the outset of the trial. In addition, a significant number had unilateral disease, and 41% had a stenosis less than 70%. What made the investigators think that stent implantation could possibly be shown to be beneficial if they entered patients into a renal function study who had near-normal renal function, unilateral disease, and mild renal artery stenosis? These are patients whose condition would not be expected to worsen with medical therapy nor to improve with stenting. Most clinicians would not consider stenting a patient to preserve renal function if the patient has unilateral mild renal artery stenosis.

There was no core laboratory to adjudicate the interpretation of the imaging studies. To determine the degree of stenosis of an artery in an accurate and unbiased fashion, a core laboratory must be used.

The reason this is so important is that visual assessment of the degree of stenosis on angiography is not accurate and almost always overestimates the degree of stenosis.^12,13 In a study assessing the physiologic importance of renal artery lesions, the lesion severity by visual estimation was 74.9% ± 11.5% (range 50%–90%), which exceeded the quantitative vascular angiographic lesion severity of 56.6% ± 10.8% (range 45%–76%).¹³

Therefore, in ASTRAL, some patients in the 50%–70% stenosis group (about 40% of patients entered) actually had a stenosis of less than 50%. And some patients in the group with stenosis greater than 70% had stenosis of less than 70%. This further illustrates that, for the most part, the patients in ASTRAL had mild to moderate renal artery stenosis.

A high adverse event rate. The major adverse event rate in the first 24 hours was 9%, whereas the usual rate is 2% or less.^14–18 Of the 280 patients in the revascularization group for whom data on adverse events were available at 1 month, 55 (20%) suffered a serious adverse event (including two patients who died) between 24 hours and 1 month after the procedure. This is in contrast to a major complication rate of 1.99% in five reports involving 727 patients.⁵

The trial centers were not high-volume centers. During the 7 years of recruitment, 24 centers (42% of all participating centers) randomized between one and five patients, and 32 centers (61% of all participating centers) randomized nine patients or fewer. This means that many participating centers randomized, on average, less than one patient per year! This was not a group of high-volume operators.

WILL CORAL GIVE US THE ANSWER?

The CORAL (Cardiovascular Outcomes in Renal Atherosclerotic Lesions) trial is under way.¹⁹ Enrollment was to have ended on January 31, 2010, and it will be several years before the data are available for analysis.

CORAL, a multicenter study funded in 2004 by the National Institutes of Health, will have randomized more than 900 patients with greater than 60% stenosis to optimal medical therapy alone or optimal medical therapy plus renal artery stenting. Inclusion criteria are a documented history of hypertension on two or more antihypertensive drugs or renal dysfunction, defined as stage 3 or greater chronic kidney disease based on the National Kidney Foundation classification (estimated glomerular filtration rate < 60 mL/min/1.73 m² calculated by the modified Modification of Diet in Renal Disease [MDRD] formula) and stenosis of 60% or greater but less than 100%, as assessed by a core laboratory. The primary end point is survival free of cardiovascular and renal adverse events, defined as a composite of cardiovascular or renal death, stroke, myocardial infarction, hospitalization for congestive heart failure, progressive renal insufficiency, or need for permanent renal replacement therapy.

We hope this trial will give us a clear answer to the question of whether renal artery stenting is beneficial in the patient population studied. One note of caution: recruitment for this trial was difficult and slow. Thus, there were a number of protocol amendments throughout the trial in order to make recruitment easier. Hopefully, this will not be a problem when analyzing the results.

WE ALL AGREE ON THE INDICATIONS FOR STENTING

So, are we really so far apart in our thinking? And is it really “time to be less aggressive” if we follow the precepts below?

All renal arteries with stenosis do not need to be (and should not be) stented.

There must be a good clinical indicationandhemodynamically significant stenosis. This means the degree of stenosis should be more than 70% on angiography or intravascular ultrasonography.

Indications for stenting. Until more data from compelling randomized trials become available, adherence to the American College of Cardiology/American Heart Association guidelines on indications for renal artery stenting is advised³:

• Hypertension: class IIa, level of evidence B. Percutaneous revascularization is reasonable for patients with hemodynamically significant renal artery stenosis and accelerated hypertension, resistant hypertension, and malignant hypertension.
• Preservation of renal function: class IIa, level of evidence B. Percutaneous revascularization is reasonable for patients with renal artery stenosis and progressive chronic kidney disease with bilateral renal artery stenosis or a stenosis to a solitary functioning kidney.
• Congestive heart failure: class I, level of evidence B. Percutaneous revascularization is indicated for patients with hemodynamically significant renal artery stenosis (ie, > 70% stenosis on angiography or intravascular ultrasonography) and recurrent, unexplained congestive heart failure or sudden, unexplained pulmonary edema.

The role of stenting for atherosclerotic renal artery stenosis is hotly debated among different specialties.^1,2 If we may generalize a bit, interventionalists (cardiologists, interventional radiologists, vascular surgeons, and vascular medicine specialists) have been in favor of liberal use of stenting, and nephrologists often favor medical therapy alone. And as with all controversial issues, each group feels rather strongly about its position.

Because few prospective randomized trials have been completed, the management of atherosclerotic renal artery stenosis has been guided by retrospective studies and case series. ³

See related article

In this issue of the Cleveland Clinic Journal of Medicine, Dr. James Simon⁴ provides an excellent overview of the prevalence, natural history, and clinical presentation of atherosclerotic renal artery stenosis. In addition, he does an admirable job of reviewing the available prospective randomized trials and providing editorial commentary about the role of the various specialists in the management of renal artery disease. And while the title of his paper says that it is “time to be less aggressive,” Dr. Simon ultimately comes to the same conclusions that we do⁵ on the indications for renal artery stenting (see Table 3 of Dr. Simon’s article), which are those of the multidisciplinary 2006 American College of Cardiology/American Heart Association guidelines on the management of peripheral artery disease.³

So what then is all the controversy about? We all agree that prospective randomized trials that provide class I, level A evidence impart the only unbiased scientific information on the best treatment strategy for patients with renal artery disease. The basic controversial issue is the interpretation of these trials. We contend that the three randomized trials of stenting vs medical therapy published so far^6–8 (see below) are so seriously flawed that it is impossible to make treatment decisions based on their results.

Since these trials were published in wellrespected journals, their results are often taken as gospel. However, careful review of each of these will reveal the flaws in study design and implementation.

THE DRASTIC TRIAL

In the Dutch Renal Artery Stenosis Intervention Cooperative (DRASTIC) trial,⁶ 106 patients with renal artery stenosis and hypertension (diastolic blood pressure > 95 mm Hg) despite treatment with two antihypertensive medications were randomly assigned to either renal angioplasty (n = 56) or drug therapy (n = 50).

Authors’ conclusions

“In the treatment of patients with hypertension and renal-artery stenosis, angioplasty has little advantage over antihypertensive-drug therapy.”⁶

Four serious problems

As we discussed in a letter to the editor of the New England Journal of Medicine on August 10, 2000, this study had four serious problems that invalidate its authors’ conclusions.⁹

The sample size was insufficient to detect a significant difference between treatment groups. In other words, the chance of a type 2 statistical error is high.

Balloon angioplasty without stenting was used as the method of revascularization. Experts now recognize that stenting is required for renal artery intervention to have a durable result.^3,5

Renal artery stenosis was defined as greater than 50% stenosis. This allowed a large number of patients to enter the trial who had hemodynamically and clinically insignificant lesions. Most clinicians believe that stenosis of less than 70% is not hemodynamically important.^5,10,11

Twenty-two of the 50 patients randomized to medical therapy crossed over to the angioplasty group because their blood pressure became difficult to control. In other words, 44% of the patients in the medical group underwent angioplasty, an astounding percentage in an intention-to-treat analysis comparing one therapy with another.

Despite these serious flaws, the results of DRASTIC influenced therapy for years after its publication.

THE STAR TRIAL

In the Stent Placement in Patients With Atherosclerotic Renal Artery Stenosis and Impaired Renal Function (STAR) trial,⁷ 140 patients with a creatinine clearance of less than 80 mL/min/1.73m², renal artery stenosis greater than 50%, and well-controlled blood pressure were randomized to either renal artery stenting plus medical therapy (n = 64) or medical therapy alone (n = 76). The primary end point was a 20% or greater decrease in creatinine clearance. Secondary end points included measures of safety and cardiovascular morbidity and mortality.

Authors’ conclusions

“Stent placement with medical treatment had no clear effect on progression of impaired renal function but led to a small number of significant procedure-related complications. The study findings favor a conservative approach to patients with [atherosclerotic renal artery stenosis], focused on cardiovascular risk factor management and avoiding stenting.”⁷

Serious flaws

A number of serious flaws render this study uninterpretable.

Mild renal artery stenosis. At least 33% of the patients in the study had mild renal artery stenosis (50%–70%), and 12 (19%) of the 64 patients in the group randomized to stenting actually had stenosis of less than 50%. How can one expect there to be a benefit to stenting in patients with mild (and hemodynamically insignificant) renal artery stenosis? This is especially true when the primary end point is a change in renal function.

More than half of the patients had unilateral disease. It seems intuitive that if one were to plan a trial with a primary end point of a change in renal function, only patients with bilateral renal artery stenosis of greater than 70% or with stenosis of greater than 70% to a solitary functioning kidney would be included. One would not expect that patients with unilateral disease and a stenosis of less than 70% would derive any benefit from revascularization.

Not all “stent” patients received stents. All of the patients in the medical group received medication and there were no crossovers. However, only 46 (72%) of the 64 patients randomized to stenting actually received a stent, while 18 (28%) did not. There were two technical failures, and 12 patients should not have been randomized because they had less than 50% stenosis on angiography and thus were not stented. Yet all 64 patients were analyzed (by intention to treat) in the stent group. With these numbers, one could predict that the results would be negative.

Like DRASTIC, this trial was underpowered, meaning that the chance of a type 2 statistical error is high. In fact, the editors of the Annals of Internal Medicine, in a note accompanying the article, cautioned that the study “was underpowered to provide a definitive estimate of efficacy.”⁷ If the study was underpowered to answer the question at hand, why was it deemed worthy of publication?

High complication rates. The periprocedural complication and death rates were much higher than in many other reports on renal artery stenting (see details below).⁵

THE ASTRAL TRIAL

In the Angioplasty and Stenting for Renal Artery Lesions (ASTRAL) trial,⁸ the primary outcome measure was the change in renal function over time as assessed by the mean slope of the reciprocal of the serum creatinine. In this trial, 806 patients with atherosclerotic renal artery stenosis were randomized to either stent-based revascularization combined with medical therapy or medical therapy alone.

Authors’ conclusions

“We found substantial risks but no evidence of a worthwhile clinical benefit from revascularization in patients with atherosclerotic renovascular disease.”⁸

Despite size, flaws remain

Unlike the other trials, ASTRAL had a sample size large enough to provide an answer. However, numerous flaws in study design and implementation invalidate its results for the overall population of patients with renal artery stenosis. The major flaws in ASTRAL were:

Selection bias. For a patient to be enrolled, the treating physician had to be undecided on whether the patient should undergo revascularization or medical management alone. However, the treatment of atherosclerotic renal artery stenosis is so controversial that physicians of different specialties cannot agree on the most effective treatment strategy for most patients.^1,2 Therefore, to exclude patients when their physicians were sure they needed or did not need renal artery revascularization is incomprehensible and introduces considerable selection bias into the trial design.

Normal renal function at baseline. The primary outcome was a change in renal function over time. Yet 25% of patients had normal renal function at the outset of the trial. In addition, a significant number had unilateral disease, and 41% had a stenosis less than 70%. What made the investigators think that stent implantation could possibly be shown to be beneficial if they entered patients into a renal function study who had near-normal renal function, unilateral disease, and mild renal artery stenosis? These are patients whose condition would not be expected to worsen with medical therapy nor to improve with stenting. Most clinicians would not consider stenting a patient to preserve renal function if the patient has unilateral mild renal artery stenosis.

There was no core laboratory to adjudicate the interpretation of the imaging studies. To determine the degree of stenosis of an artery in an accurate and unbiased fashion, a core laboratory must be used.

The reason this is so important is that visual assessment of the degree of stenosis on angiography is not accurate and almost always overestimates the degree of stenosis.^12,13 In a study assessing the physiologic importance of renal artery lesions, the lesion severity by visual estimation was 74.9% ± 11.5% (range 50%–90%), which exceeded the quantitative vascular angiographic lesion severity of 56.6% ± 10.8% (range 45%–76%).¹³

Therefore, in ASTRAL, some patients in the 50%–70% stenosis group (about 40% of patients entered) actually had a stenosis of less than 50%. And some patients in the group with stenosis greater than 70% had stenosis of less than 70%. This further illustrates that, for the most part, the patients in ASTRAL had mild to moderate renal artery stenosis.

A high adverse event rate. The major adverse event rate in the first 24 hours was 9%, whereas the usual rate is 2% or less.^14–18 Of the 280 patients in the revascularization group for whom data on adverse events were available at 1 month, 55 (20%) suffered a serious adverse event (including two patients who died) between 24 hours and 1 month after the procedure. This is in contrast to a major complication rate of 1.99% in five reports involving 727 patients.⁵

The trial centers were not high-volume centers. During the 7 years of recruitment, 24 centers (42% of all participating centers) randomized between one and five patients, and 32 centers (61% of all participating centers) randomized nine patients or fewer. This means that many participating centers randomized, on average, less than one patient per year! This was not a group of high-volume operators.

WILL CORAL GIVE US THE ANSWER?

The CORAL (Cardiovascular Outcomes in Renal Atherosclerotic Lesions) trial is under way.¹⁹ Enrollment was to have ended on January 31, 2010, and it will be several years before the data are available for analysis.

CORAL, a multicenter study funded in 2004 by the National Institutes of Health, will have randomized more than 900 patients with greater than 60% stenosis to optimal medical therapy alone or optimal medical therapy plus renal artery stenting. Inclusion criteria are a documented history of hypertension on two or more antihypertensive drugs or renal dysfunction, defined as stage 3 or greater chronic kidney disease based on the National Kidney Foundation classification (estimated glomerular filtration rate < 60 mL/min/1.73 m² calculated by the modified Modification of Diet in Renal Disease [MDRD] formula) and stenosis of 60% or greater but less than 100%, as assessed by a core laboratory. The primary end point is survival free of cardiovascular and renal adverse events, defined as a composite of cardiovascular or renal death, stroke, myocardial infarction, hospitalization for congestive heart failure, progressive renal insufficiency, or need for permanent renal replacement therapy.

We hope this trial will give us a clear answer to the question of whether renal artery stenting is beneficial in the patient population studied. One note of caution: recruitment for this trial was difficult and slow. Thus, there were a number of protocol amendments throughout the trial in order to make recruitment easier. Hopefully, this will not be a problem when analyzing the results.

WE ALL AGREE ON THE INDICATIONS FOR STENTING

So, are we really so far apart in our thinking? And is it really “time to be less aggressive” if we follow the precepts below?

All renal arteries with stenosis do not need to be (and should not be) stented.

There must be a good clinical indicationandhemodynamically significant stenosis. This means the degree of stenosis should be more than 70% on angiography or intravascular ultrasonography.

Indications for stenting. Until more data from compelling randomized trials become available, adherence to the American College of Cardiology/American Heart Association guidelines on indications for renal artery stenting is advised³:

• Hypertension: class IIa, level of evidence B. Percutaneous revascularization is reasonable for patients with hemodynamically significant renal artery stenosis and accelerated hypertension, resistant hypertension, and malignant hypertension.
• Preservation of renal function: class IIa, level of evidence B. Percutaneous revascularization is reasonable for patients with renal artery stenosis and progressive chronic kidney disease with bilateral renal artery stenosis or a stenosis to a solitary functioning kidney.
• Congestive heart failure: class I, level of evidence B. Percutaneous revascularization is indicated for patients with hemodynamically significant renal artery stenosis (ie, > 70% stenosis on angiography or intravascular ultrasonography) and recurrent, unexplained congestive heart failure or sudden, unexplained pulmonary edema.

Author’s note: Atherosclerosis accounts for about 90% of cases of renal artery stenosis in people over age 40.¹ Fibromuscular dysplasia, the other major cause, is a separate topic; in this paper “renal artery stenosis” refers to atherosclerotic disease only.

Renal artery stenosis is very common, and the number of angioplasty-stenting procedures performed every year is on the rise. Yet there is no overwhelming evidence that intervention yields clinical benefits—ie, better blood pressure control or renal function— than does medical therapy.

See related editorial

Earlier randomized controlled trials comparing angioplasty without stents and medical management showed no impressive difference in blood pressure.^2,3 The data on renal function were even more questionable, with some studies suggesting that, with stenting, the chance of worsening renal function is equal to that of improvement.⁴

Two large randomized trials comparing renal intervention with medical therapy failed to show any benefit of intervention.^5–7 A third study is under way.⁸

It is time to strongly reconsider the current aggressive approach to revascularization of stenotic renal arteries and take a more coordinated, critical approach.

RENAL INTERVENTIONS ON THE RISE

Renal angioplasty began replacing surgical revascularization in the 1990s, as this less-invasive procedure became more readily available and was shown to have similar clinical outcomes.⁹ In the last decade, stent placement during angioplasty has become standard, improving the rates of technical success.

As these procedures have become easier to perform and their radiographic outcomes have become more consistent, interventionalists have become more likely, if they see stenosis in a renal artery, to intervene and insert a stent, regardless of proven benefit. In addition, interventionalists from at least three different specialties now compete for these procedures, often by looking at the renal arteries during angiography of other vascular beds (the “drive-by”).

As a result, the number of renal interventions has been rising. Medicare received 21,660 claims for renal artery interventions (surgery or angioplasty) in 2000, compared with 13,380 in 1996—an increase of 62%. However, the number of surgeries actually decreased by 45% during this time, while the number of percutaneous procedures increased by 240%. The number of endovascular claims for renal artery stenosis by cardiologists alone rose 390%.¹⁰ Since then, the reports on intervention have been mixed, with one report citing a continued increase in 2005 to 35,000 claims,¹¹ and another suggesting a decrease back to 1997 levels.¹²

HOW COMMON IS RENAL ARTERY STENOSIS?

The prevalence of renal artery stenosis depends on the definition used and the population screened. It is more common in older patients who have risk factors for other vascular diseases than in the general population.

Renal Doppler ultrasonography can detect stenosis only if the artery is narrowed by more than 60%. Hansen et al¹³ used ultrasonography to screen 870 people over age 65 and found a lesion (a narrowing of more than 60%) in 6.8%.

Angiography (direct, computed tomographic, or magnetic resonance) can detect less-severe stenosis. Thus, most angiographic studies define renal artery stenosis as a narrowing of more than 50%, and severe disease as a narrowing of more than 70%. Many experts believe that unilateral stenosis needs to be more than 70% to pose a risk to the kidney.^14,15

Angiographic prevalence studies have been performed only in patients who were undergoing angiography for another reason such as coronary or peripheral arterial disease that inherently places them at higher risk of renal artery stenosis. For instance, renal artery stenosis is found in 11% to 28% of patients undergoing diagnostic cardiac catheterization. ¹⁶

No studies of the prevalence of renal artery stenosis have been performed in the general population. Medicare data indicate that from 1999 to 2001 the incidence of diagnosed renal artery stenosis was 3.7 per 1,000 patientyears. ¹⁷ Holley et al,¹⁸ in an autopsy series, found renal artery stenosis of greater than 50% in 27% of patients over age 50 and in 56.4% of hypertensive patients. The prevalence was 10% in normotensive patients.

WHO IS AT RISK?

Factors associated with a higher risk of finding renal artery stenosis on a radiographic study include¹⁴:

• Older age
• Female sex
• Hypertension
• Three-vessel coronary artery disease
• Peripheral artery disease
• Chronic kidney disease
• Diabetes
• 
  Tobacco use
• A low level of high-density lipoprotein cholesterol
• The use of at least two cardiovascular drugs.

The prevalence of renal artery stenosis in at-risk populations ranges from 3% to 75% (Table  1).^2,4,6,19,20

HOW OFTEN DOES STENOSIS PROGRESS?

The reported rates of progression of atherosclerotic renal artery lesions vary depending on the type of imaging test used and the reason for doing it.

In studies that used duplex ultrasonography, roughly half of lesions smaller than 60% grew to greater than 60% over 3 years.^21,22 The risk of total occlusion of an artery was relatively low and depended on the severity of stenosis: 0.7% if the baseline stenosis was less than 60% and 2.3% to 7% if it was greater.^21,22

In a seminal study in 1984, Schreiber and colleagues²³ compared serial angiograms obtained a mean of 52 months apart in 85 patients who did not undergo intervention. Stenosis had progressed in 37 (44%), and to the point of total occlusion in 14 (16%). In contrast, a 1998 study found progression in 11.1% over 2.6 years, with older patients, women, and those with baseline coronary artery disease at higher risk.²⁴

The the rates of progression differed in these two studies probably because the indications for screening were different (clinical suspicion²³ vs routine screening during coronary angiography²⁴), as was the severity of stenosis at the time of diagnosis. Also, when these studies were done, fewer people were taking statins. Thus, similar studies, if repeated, might show even lower rates of progression.

Finally, progression of renal artery stenosis has not been correlated with worsening renal function.

FOUR CLINICAL PRESENTATIONS OF RENAL ARTERY STENOSIS

Renal artery stenosis can present in one of four ways:

Clinically silent stenosis. Because renal artery stenosis is most often found in older patients, who are more likely to have essential hypertension and chronic kidney disease due to other causes, it can be an incidental finding that is completely clinically silent.^16,25

Renovascular hypertension is defined as high blood pressure due to up-regulation of neurohormonal activity in response to decreased perfusion from renal artery stenosis. Renal artery stenosis is estimated to be the cause of hypertension in only 0.5% to 4.0% of hypertensive patients, or in 26% of patients with secondary hypertension.³

Ischemic nephropathy is more difficult to define because ischemia alone rarely explains the damage done to the kidneys. Activation of neurohormonal pathways and microvascular injury are thought to contribute to oxidative stress and fibrosis.²⁶ These phenomena may explain why similar degrees of stenosis lead to varying degrees of kidney damage in different patients and why the severity of stenosis does not correlate with the degree of renal dysfunction.²⁷

Furthermore, stenosis may lead to irreversible but stable kidney damage. It is therefore not surprising that, in studies in unselected populations (ie, studies that included patients with all presentations of renal artery stenosis, not just those more likely to benefit), up to two-thirds of renal interventions yielded no clinical benefit.²⁵

As a result, if we define ischemic nephropathy as renal artery stenosis with renal dysfunction not attributable to another cause, we probably will overestimate the prevalence of ischemic nephropathy, leading to overly optimistic expectations about the response to revascularization.

Recurrent “flash” pulmonary edema is a less common presentation, usually occurring in patients with critical bilateral renal artery stenosis or unilateral stenosis in an artery supplying a solitary functioning kidney. Most have severe hypertension (average systolic blood pressure 174–207 mm Hg) and poor renal function.^28–30

The association between pulmonary edema and bilateral renal artery stenosis was first noted in 1998 by Pickering et al,³¹ who in several case series showed that 82% to 92% of patients with recurrent pulmonary edema and renal artery stenosis had bilateral stenosis, compared with 20% to 65% of those with other presentations. Later case series corroborated this finding: 85% to 100% of patients with renal artery stenosis and pulmonary edema had bilateral stenosis.^28–30

STENTING IS NOW STANDARD

Stenting has become standard in the endovascular treatment of renal artery stenosis.

Most atherosclerotic renal artery lesions are located in the ostium (ie, where the artery branches off from the aorta), and many are extensions of calcified aortic plaque.^26,32 These hard lesions tend to rebound to their original shape more often with balloon angioplasty alone. Stenting provides the additional force needed to permanently disrupt the lesion, leading to a longer-lasting result.

Rates of technical success (dilating the artery during the intervention) are higher with stents than without them (98% vs 46%– 77%).^33,34 If the lesion is ostial, this difference is even more impressive (75% vs 29%). In addition, restenosis rates at 6 months are lower with stents (14% vs 26%–48%).³⁴

GOALS: LOWER THE BLOOD PRESSURE, SAVE THE KIDNEY

Because endovascular procedures pose some risk to the patient, it is critical to intervene only in patients most likely to respond clinically. The decision to intervene depends largely on the clinical goal, which should depend on the clinical presentation.

However, if renal artery stenosis is clinically silent, most of the evidence suggests that intervention has no benefit. Furthermore, although retrospective studies have indicated that intervention may improve survival rates,^35,36 prospective studies have not. Similarly, studies have not shown that intervention generally improves cardiovascular outcomes, even though renal artery stenosis is associated with cardiovascular risk.

Hypertension plus stenosis is not necessarily renovascular hypertension

Essential hypertension and clinically silent renal artery stenosis often coexist, which is why blood pressure control often does not improve after stenting. Also, essential hypertension often coexists with renovascular hypertension.³⁷ In this situation, stenting may not eliminate the need for antihypertensive drugs, although it may improve blood pressure control and decrease the drug burden.

Before stents came into use, several randomized controlled trials found that blood pressure was no better controlled after angioplasty, ^2,3,38 except in cases of bilateral stenosis.² This may be because stenosis tended to recur after angioplasty without stents.

The 2000 Dutch Renal Artery Stenosis Intervention Cooperative (DRASTIC) study was the first randomized controlled trial to examine the effect of angioplasty on blood pressure control in renal artery stenosis.³⁸ It had significant design flaws. For example, many patients crossed over from the medical management group to the intervention group because their hypertension was resistant to medical therapy. Overall, intervention (balloon angioplasty without stents in 54 of 56 patients, with stents in the other 2) carried no benefit. However, in subgroup analysis, the patients who crossed over because of resistant hypertension (failure of a three-drug regimen) were more likely to benefit from angioplasty. This suggested that risk stratification should take place early on, before proceeding with revascularization.

With stents, Zeller,³⁹ in a prospective nonrandomized study, found that the mean arterial pressure decreased by 10 mm Hg. Randomized trials (see below) have failed to demonstrate such a benefit.

 

 

Stenting may not improve renal function

Coincidental renal artery stenosis in a patient with unrelated chronic kidney disease is very hard to differentiate from true ischemic nephropathy. Furthermore, most patients with ischemic nephropathy do not benefit from revascularization, making it challenging to identify those few whose renal function may respond.

Given that patients with chronic kidney disease tend to have a higher risk of cardiovascular disease, it is not surprising that 15% of them may have renal artery stenosis,⁴ most often incidental.

Chábová⁴⁰ examined the outcomes of 68 patients who had chronic kidney disease and a renal artery lesion larger than 70% who did not undergo angioplasty. In only 10 (15%) of the patients did the glomerular filtration rate (GFR) decline by more than 50% of its baseline value during the study period of 3 years. Given the retrospective nature of the study, it cannot be determined (and is rather unlikely) that ischemic nephropathy was the cause of the decline in kidney function in all 10 patients.

In a prospective cohort study in 304 patients with chronic kidney disease and renal artery stenosis who underwent surgical revascularization, Textor⁴ reported that the serum creatinine level showed a meaningful improvement afterward in 28%, worsened in 19.7%, and remained unchanged in 160 52.6%. (A “meaningful” change was defined as > 1.0 mg/dL.) Findings were similar in a cohort that underwent stenting.³³

Davies et al⁴¹ found that 20% of patients who underwent renal stenting had a persistent increase in serum creatinine of 0.5 mg/dL or more. These patients were nearly three times more likely (19% vs 7%) to eventually require dialysis, and they had a lower 5-year survival rate (41% vs 71%).

Zeller et al³⁹ found that renal function improved slightly in 52% of patients who received stents. The mean decrease in serum creatinine in this group was 0.22 mg/dL. However, the other 48% had a mean increase in serum creatinine of 1.1 mg/dL.

From these data we can conclude that, in an unselected population with renal artery stenosis, stenting provides no benefit to renal function, and that the risk of a worsening of renal function after intervention is roughly equal to the likelihood of achieving any benefit.

Other predictors of improvement in renal function have been proposed, but the evidence supporting them has not been consistent. For example, although Radermacher et al⁴² reported that a renal resistive index (which reflects arterial stiffness downstream of the stenosis) lower than 0.8 predicted a response in renal function, this finding has not been reliably reproduced.^43,44 Similarly, while several studies suggest that patients with milder renal dysfunction have a higher likelihood of a renal response,^45,46 other studies suggest either that the opposite is true³⁹ or that baseline renal function alone has no impact on outcome.⁴⁷

In addition, once significant renal atrophy occurs, revascularization may not help much, since irreversible sclerosis has developed. Thus, the goal is to identify kidneys being harmed by renal artery stenosis during the ischemic phase, when the tissue is still viable.

Unfortunately, we still lack a good renal stress test—eg, analogous to the cardiac stress test—to diagnose reversible ischemia in the kidney. The captopril renal scan has that capability but is not accurate in patients with bilateral stenosis or a GFR less than 50 mL/min, severely limiting its applicability.²⁶ Newer technologies such as blood-oxygen-level-dependent (BOLD) magnetic resonance imaging are being investigated for such a role.⁴⁸

Cohort studies in patients with declining renal function

In several case series, patients whose renal function had been declining before intervention had impressive rates of better renal function afterward.^{33,39,47,49–54} In a prospective cohort study by Muray et al,⁴⁷ a rise in serum creatinine of more than 0.1 mg/mL/month before intervention seemed to predict an improvement in renal function afterward.

One would expect that, for renal function to respond to intervention, severe bilateral stenosis or unilateral stenosis to a solitary functioning kidney would need to be present. However, this was an inconsistent finding in these case series.^{33,39,47,52,53} The Angioplasty and Stent for Renal Artery Lesions (ASTRAL) trial,^6,7 discussed later, sheds a bit more light on this.

Stenting usually improves flash pulmonary edema

Acute pulmonary edema in the setting of bilateral renal artery stenosis seems to be a unique case in which improvement in clinical status can be expected in most patients after intervention. Blood pressure improves in 94% to 100% of patients,^28,31 renal function either improves or stabilizes in 77% to 91%,^28–31 and pulmonary edema resolves without recurrence in 77% to 100%.^28–30

NEW RANDOMIZED TRIALS: STAR, ASTRAL, AND CORAL

Despite the lack of evidence supporting revascularization of renal artery stenosis, many interventionalists practice under the assumption that the radiographic finding of renal artery stenosis alone is an indication for renal revascularization. Only three randomized controlled trials in the modern era attempt to examine this hypothesis: STAR, ASTRAL, and CORAL.

STAR: No clear benefit

The Stent Placement and Blood Pressure and Lipid-lowering for the Prevention of Progression of Renal Dysfunction Caused by Atherosclerotic Ostial Stenosis of the Renal Artery (STAR) trial⁵ was a European multicenter trial that enrolled 140 patients with ostial renal artery stenosis greater than 50%, blood pressure controlled to less than 140/90 mm Hg, and creatinine clearance 15 to 80 mL/min.

Patients were randomized to undergo stenting or medical therapy alone. High blood pressure was treated according to a protocol in which angiotensin-converting enzyme (ACE) inhibitors and angiotensin receptor blockers were relegated to second-line use. All patients received a statin, regardless of lipid levels.

At 2 years, the primary end point (a decline in creatinine clearance of 20% or greater) had been reached in 10 (16%) of the 64 patients in the stent group and 16 (22%) of the 76 patients in the medication group; the difference was not statistically significant (hazard ratio 0.73, 95% confidence interval 0.33–1.61). No difference was seen in the secondary end points of the degree of blood pressure control or the rates of cardiovascular morbidity and death.⁵

 

 

ASTRAL: Also no clear benefit

In the international, multicenter ASTRAL trial,^6,7 806 patients with at least one stenotic renal artery considered suitable for balloon angioplasty, stenting, or both⁷ were randomized to undergo intervention or medical management. Hypertension treatment was not specified by a protocol. The mean estimated GFR was 40 mL/min. Most patients (95%–96%) were on statin therapy. The primary outcome was the rate of decline of renal function over time. Secondary outcomes included blood pressure control, renal events, cardiovascular events, and death.

Results. At a mean follow-up of 33.6 months (range 1–4 years), no difference was noted between treatment groups in decline in renal function or blood pressure control at 1 year. Renal function worsened slightly in both groups.

The decline in renal function over time, calculated as the mean slope of the reciprocal of the serum creatinine level over time, was slightly slower in the revascularization group, but the difference was not statistically significant (−0.07 × 10⁻³ vs −0.13 × 10⁻³ L/μmol/year, P = .06). This difference did not appear until the last year of the study. There was no difference in the number of patients whose renal function improved or declined during the study.

There was no difference in the rate of any secondary outcome. The medical management group required a slightly higher number of antihypertensive drugs, reaching statistical but not clinical significance (2.97 vs 2.77 drugs, P = .03). More people in the revascularization group were taking ACE inhibitors or angiotensin receptor blockers. There was no difference in the number of patients on any antihypertensive therapy (97% vs 99%). Interestingly, amputations were more common in the revascularization group, occurring in 42 (12%) of the 386 patients in the revascularization group vs 29 (7%) of the 395 patients in the medical group (P = .04).

Seventeen percent of patients randomized to intervention did not have the procedure done. An as-treated analysis of the 317 (83%) patients randomized to revascularization who did receive it showed no differences in outcomes.

There were no differences in outcomes among specific, predefined subgroups based on severity of stenosis at baseline, renal length, baseline estimated GFR, baseline serum creatinine, and rate of progression of renal dysfunction before randomization.⁷

Comments. ASTRAL contradicts previous nonrandomized studies that suggested that rapidly declining renal function (loss of 20% in 1 year) predicts response to intervention. Considering the large number of patients with unilateral disease in the study, it would be interesting to see what effect stenting had on patients with both severe disease and declining renal function.

ASTRAL has been criticized because it lacked a central laboratory to interpret the severity of stenosis, it did not use a standardized intervention technique (5% of patients underwent angioplasty without stents, although this did not affect outcomes⁷), and patients were enrolled only if the clinician involved in the case was uncertain of the appropriate management.

This last issue raises the concern for selection bias toward inclusion of more difficult cases that may not respond to intervention. But these shortcomings are not serious enough to negate the fact that preliminary results from the largest randomized controlled trial to date confirm conclusions of other randomized trials, ie, that intervention in renal artery stenosis yields no benefits over medical management in most patients.

Based on the results of STAR and ASTRAL, the practice of indiscriminately revascularizing stenosed renal arteries without strong evidence that the procedure will provide a clinical benefit is no longer tenable. The challenge is to identify those few patients who will respond, and to intervene only on them. Unfortunately, none of the subgroups from ASTRAL helped characterize this population.

CORAL: A large trial is ongoing

The Cardiovascular Outcomes in Renal Artherosclerotic Lesions (CORAL) trial,⁸ an ongoing multicenter randomized controlled trial in the United States, may be of additional help.

Unlike ASTRAL, CORAL is studying patients who have difficult-to-control hypertension (systolic blood pressure ≥ 155 mm Hg on two or more drugs).⁸ Chronic kidney disease is not an exclusion criterion unless the serum creatinine concentration is greater than 3.0 mg/dL.

CORAL is using a standardized medical protocol to control blood pressure. In addition, use of embolic protection devices during stenting is encouraged. Hopefully, the large size (a goal of 1,080 patients) and the inclusion of patients with more marked hypertension will address the utility of intervention in higher-risk populations with renal artery stenosis.

RECOMMENDED APPROACH TO INTERVENTION IN RENAL ARTERY STENOSIS

As we wait for CORAL to be completed, we have two modern-era randomized controlled trials that leave us with fewer indications for renal intervention. Table 2 lists commonly cited indications for intervention in renal artery stenosis and the evidence to support them. As most of these are based on retrospective data or have conflicting support in the literature, their utility remains in question. At this point we can safely recommend:

• Patients with preserved or even decreased but stable renal function will not likely have a benefit in renal function after intervention.
• Patients with resistant hypertension may benefit.
• The best evidence supporting intervention is for bilateral stenosis with flash pulmonary edema, but the evidence is from retrospective studies.
• Stenting in bilateral disease without another indication has no apparent benefit.
• Declining renal function is not a guarantee of success.
• It is unclear if patients with severe bilateral stenosis or severe stenosis to a solitary functioning kidney with declining renal function will benefit. Anecdotally, they do respond more often, but as with many other indications for intervention that have gone by the wayside, this may not bear out when studied properly.

As the utility of intervention narrows, the scope of practice for such interventions should narrow accordingly. Attention should now be focusing on clinical, rather than radiographic, indications for intervening on renal artery stenosis.

Therefore, the decision to intervene must not be made solely by the interventionalist. A multidisciplinary approach should be adopted that at the very least includes the input of a nephrologist well versed in renal artery stenosis. In this way, the clinical risks and benefits of renal intervention can be discussed with the patient by providers who are likely to be involved in their care should renal function or hypertension fail to improve afterward.

RISK OF ATHEROEMBOLISM

While renal stenting yields improved technical success in the treatment of renal artery stenosis, it carries with it an increasingly common risk to kidney function: atheroembolism as the stent crushes the plaque against the vessel wall. This may lead to obstruction of the renal microvasculature, increasing the risk of irreversible damage to renal function.

Atheroembolic kidney disease can manifest as progressive renal failure occurring over weeks to months, commonly misdiagnosed as permanent damage from contrast nephropathy.⁵⁵

Embolic protection devices, inserted downstream of the lesion before stenting to catch any debris that may break loose, have been developed to help address this problem.

Holden et al ⁵⁷ prospectively studied 63 patients with renal artery stenosis and deteriorating renal function (undefined) who underwent stenting with an embolic protection device. At 6 months after the intervention, renal function had either improved or stabilized in 97% of patients, suggesting that many of the deleterious effects of stenting on renal function are related to atheroembolism.

The Prospective Randomized Study Comparing Renal Artery Stenting With or Without Distal Protection (RESIST) trial, in which renal dysfunction was mild and the GFR was not declining (average estimated GFR 59.3 mL/min), found contrary results.⁵⁷ In a two-by-two factorial study, patients were randomized to undergo stenting alone, stenting with the antiplatelet agent abciximab (ReoPro), stenting with an embolic protection device, or stenting with both abciximab and an embolic protection device. Interestingly, renal function declined in the first three groups, but remained stable in the group that received both abciximab and an embolic protection device.

ANTIPLATELET THERAPY AFTER RENAL STENTING: HOW LONG?

We have no data on the optimal duration of antiplatelet therapy after renal stenting, and guidelines from professional societies do not comment on it.⁵⁸ As a result, practice patterns vary significantly among practitioners.

While in-stent restenosis rates are acceptably low after renal stenting, the risks and side effects of antiplatelet therapy often lead to arbitrary withdrawal of these drugs. The effect on stent patency is yet to be determined.

FUTURE DEVELOPMENTS

Results of STAR and ASTRAL confirm the growing suspicion that the surge seen in the last decade in renal artery stenting should be coming to an end. We await results either from CORAL or possibly a post hoc analysis of ASTRAL that might identify potential high-risk groups that will benefit from renal intervention. And as embolic protection devices become more agile and suitable to different renal lesions, there remains the possibility that, due to lower rates of unidentified atheroembolic kidney disease, CORAL may demonstrate improved renal outcomes after stenting. If not, the search for the best means to predict who should have renal intervention will continue.

We know through experience that stenting does provide great benefits for some patients with renal artery stenosis. Furthermore, the clinical problem is too intriguing, and too profitable, to die altogether.

Author’s note: Atherosclerosis accounts for about 90% of cases of renal artery stenosis in people over age 40.¹ Fibromuscular dysplasia, the other major cause, is a separate topic; in this paper “renal artery stenosis” refers to atherosclerotic disease only.

Renal artery stenosis is very common, and the number of angioplasty-stenting procedures performed every year is on the rise. Yet there is no overwhelming evidence that intervention yields clinical benefits—ie, better blood pressure control or renal function— than does medical therapy.

See related editorial

Earlier randomized controlled trials comparing angioplasty without stents and medical management showed no impressive difference in blood pressure.^2,3 The data on renal function were even more questionable, with some studies suggesting that, with stenting, the chance of worsening renal function is equal to that of improvement.⁴

Two large randomized trials comparing renal intervention with medical therapy failed to show any benefit of intervention.^5–7 A third study is under way.⁸

It is time to strongly reconsider the current aggressive approach to revascularization of stenotic renal arteries and take a more coordinated, critical approach.

RENAL INTERVENTIONS ON THE RISE

Renal angioplasty began replacing surgical revascularization in the 1990s, as this less-invasive procedure became more readily available and was shown to have similar clinical outcomes.⁹ In the last decade, stent placement during angioplasty has become standard, improving the rates of technical success.

As these procedures have become easier to perform and their radiographic outcomes have become more consistent, interventionalists have become more likely, if they see stenosis in a renal artery, to intervene and insert a stent, regardless of proven benefit. In addition, interventionalists from at least three different specialties now compete for these procedures, often by looking at the renal arteries during angiography of other vascular beds (the “drive-by”).

As a result, the number of renal interventions has been rising. Medicare received 21,660 claims for renal artery interventions (surgery or angioplasty) in 2000, compared with 13,380 in 1996—an increase of 62%. However, the number of surgeries actually decreased by 45% during this time, while the number of percutaneous procedures increased by 240%. The number of endovascular claims for renal artery stenosis by cardiologists alone rose 390%.¹⁰ Since then, the reports on intervention have been mixed, with one report citing a continued increase in 2005 to 35,000 claims,¹¹ and another suggesting a decrease back to 1997 levels.¹²

HOW COMMON IS RENAL ARTERY STENOSIS?

The prevalence of renal artery stenosis depends on the definition used and the population screened. It is more common in older patients who have risk factors for other vascular diseases than in the general population.

Renal Doppler ultrasonography can detect stenosis only if the artery is narrowed by more than 60%. Hansen et al¹³ used ultrasonography to screen 870 people over age 65 and found a lesion (a narrowing of more than 60%) in 6.8%.

Angiography (direct, computed tomographic, or magnetic resonance) can detect less-severe stenosis. Thus, most angiographic studies define renal artery stenosis as a narrowing of more than 50%, and severe disease as a narrowing of more than 70%. Many experts believe that unilateral stenosis needs to be more than 70% to pose a risk to the kidney.^14,15

Angiographic prevalence studies have been performed only in patients who were undergoing angiography for another reason such as coronary or peripheral arterial disease that inherently places them at higher risk of renal artery stenosis. For instance, renal artery stenosis is found in 11% to 28% of patients undergoing diagnostic cardiac catheterization. ¹⁶

No studies of the prevalence of renal artery stenosis have been performed in the general population. Medicare data indicate that from 1999 to 2001 the incidence of diagnosed renal artery stenosis was 3.7 per 1,000 patientyears. ¹⁷ Holley et al,¹⁸ in an autopsy series, found renal artery stenosis of greater than 50% in 27% of patients over age 50 and in 56.4% of hypertensive patients. The prevalence was 10% in normotensive patients.

WHO IS AT RISK?

Factors associated with a higher risk of finding renal artery stenosis on a radiographic study include¹⁴:

• Older age
• Female sex
• Hypertension
• Three-vessel coronary artery disease
• Peripheral artery disease
• Chronic kidney disease
• Diabetes
• 
  Tobacco use
• A low level of high-density lipoprotein cholesterol
• The use of at least two cardiovascular drugs.

The prevalence of renal artery stenosis in at-risk populations ranges from 3% to 75% (Table  1).^2,4,6,19,20

HOW OFTEN DOES STENOSIS PROGRESS?

The reported rates of progression of atherosclerotic renal artery lesions vary depending on the type of imaging test used and the reason for doing it.

In studies that used duplex ultrasonography, roughly half of lesions smaller than 60% grew to greater than 60% over 3 years.^21,22 The risk of total occlusion of an artery was relatively low and depended on the severity of stenosis: 0.7% if the baseline stenosis was less than 60% and 2.3% to 7% if it was greater.^21,22

In a seminal study in 1984, Schreiber and colleagues²³ compared serial angiograms obtained a mean of 52 months apart in 85 patients who did not undergo intervention. Stenosis had progressed in 37 (44%), and to the point of total occlusion in 14 (16%). In contrast, a 1998 study found progression in 11.1% over 2.6 years, with older patients, women, and those with baseline coronary artery disease at higher risk.²⁴

The the rates of progression differed in these two studies probably because the indications for screening were different (clinical suspicion²³ vs routine screening during coronary angiography²⁴), as was the severity of stenosis at the time of diagnosis. Also, when these studies were done, fewer people were taking statins. Thus, similar studies, if repeated, might show even lower rates of progression.

Finally, progression of renal artery stenosis has not been correlated with worsening renal function.

FOUR CLINICAL PRESENTATIONS OF RENAL ARTERY STENOSIS

Renal artery stenosis can present in one of four ways:

Clinically silent stenosis. Because renal artery stenosis is most often found in older patients, who are more likely to have essential hypertension and chronic kidney disease due to other causes, it can be an incidental finding that is completely clinically silent.^16,25

Renovascular hypertension is defined as high blood pressure due to up-regulation of neurohormonal activity in response to decreased perfusion from renal artery stenosis. Renal artery stenosis is estimated to be the cause of hypertension in only 0.5% to 4.0% of hypertensive patients, or in 26% of patients with secondary hypertension.³

Ischemic nephropathy is more difficult to define because ischemia alone rarely explains the damage done to the kidneys. Activation of neurohormonal pathways and microvascular injury are thought to contribute to oxidative stress and fibrosis.²⁶ These phenomena may explain why similar degrees of stenosis lead to varying degrees of kidney damage in different patients and why the severity of stenosis does not correlate with the degree of renal dysfunction.²⁷

Furthermore, stenosis may lead to irreversible but stable kidney damage. It is therefore not surprising that, in studies in unselected populations (ie, studies that included patients with all presentations of renal artery stenosis, not just those more likely to benefit), up to two-thirds of renal interventions yielded no clinical benefit.²⁵

As a result, if we define ischemic nephropathy as renal artery stenosis with renal dysfunction not attributable to another cause, we probably will overestimate the prevalence of ischemic nephropathy, leading to overly optimistic expectations about the response to revascularization.

Recurrent “flash” pulmonary edema is a less common presentation, usually occurring in patients with critical bilateral renal artery stenosis or unilateral stenosis in an artery supplying a solitary functioning kidney. Most have severe hypertension (average systolic blood pressure 174–207 mm Hg) and poor renal function.^28–30

The association between pulmonary edema and bilateral renal artery stenosis was first noted in 1998 by Pickering et al,³¹ who in several case series showed that 82% to 92% of patients with recurrent pulmonary edema and renal artery stenosis had bilateral stenosis, compared with 20% to 65% of those with other presentations. Later case series corroborated this finding: 85% to 100% of patients with renal artery stenosis and pulmonary edema had bilateral stenosis.^28–30

STENTING IS NOW STANDARD

Stenting has become standard in the endovascular treatment of renal artery stenosis.

Most atherosclerotic renal artery lesions are located in the ostium (ie, where the artery branches off from the aorta), and many are extensions of calcified aortic plaque.^26,32 These hard lesions tend to rebound to their original shape more often with balloon angioplasty alone. Stenting provides the additional force needed to permanently disrupt the lesion, leading to a longer-lasting result.

Rates of technical success (dilating the artery during the intervention) are higher with stents than without them (98% vs 46%– 77%).^33,34 If the lesion is ostial, this difference is even more impressive (75% vs 29%). In addition, restenosis rates at 6 months are lower with stents (14% vs 26%–48%).³⁴

GOALS: LOWER THE BLOOD PRESSURE, SAVE THE KIDNEY

Because endovascular procedures pose some risk to the patient, it is critical to intervene only in patients most likely to respond clinically. The decision to intervene depends largely on the clinical goal, which should depend on the clinical presentation.

However, if renal artery stenosis is clinically silent, most of the evidence suggests that intervention has no benefit. Furthermore, although retrospective studies have indicated that intervention may improve survival rates,^35,36 prospective studies have not. Similarly, studies have not shown that intervention generally improves cardiovascular outcomes, even though renal artery stenosis is associated with cardiovascular risk.

Hypertension plus stenosis is not necessarily renovascular hypertension

Essential hypertension and clinically silent renal artery stenosis often coexist, which is why blood pressure control often does not improve after stenting. Also, essential hypertension often coexists with renovascular hypertension.³⁷ In this situation, stenting may not eliminate the need for antihypertensive drugs, although it may improve blood pressure control and decrease the drug burden.

Before stents came into use, several randomized controlled trials found that blood pressure was no better controlled after angioplasty, ^2,3,38 except in cases of bilateral stenosis.² This may be because stenosis tended to recur after angioplasty without stents.

The 2000 Dutch Renal Artery Stenosis Intervention Cooperative (DRASTIC) study was the first randomized controlled trial to examine the effect of angioplasty on blood pressure control in renal artery stenosis.³⁸ It had significant design flaws. For example, many patients crossed over from the medical management group to the intervention group because their hypertension was resistant to medical therapy. Overall, intervention (balloon angioplasty without stents in 54 of 56 patients, with stents in the other 2) carried no benefit. However, in subgroup analysis, the patients who crossed over because of resistant hypertension (failure of a three-drug regimen) were more likely to benefit from angioplasty. This suggested that risk stratification should take place early on, before proceeding with revascularization.

With stents, Zeller,³⁹ in a prospective nonrandomized study, found that the mean arterial pressure decreased by 10 mm Hg. Randomized trials (see below) have failed to demonstrate such a benefit.

 

 

Stenting may not improve renal function

Coincidental renal artery stenosis in a patient with unrelated chronic kidney disease is very hard to differentiate from true ischemic nephropathy. Furthermore, most patients with ischemic nephropathy do not benefit from revascularization, making it challenging to identify those few whose renal function may respond.

Given that patients with chronic kidney disease tend to have a higher risk of cardiovascular disease, it is not surprising that 15% of them may have renal artery stenosis,⁴ most often incidental.

Chábová⁴⁰ examined the outcomes of 68 patients who had chronic kidney disease and a renal artery lesion larger than 70% who did not undergo angioplasty. In only 10 (15%) of the patients did the glomerular filtration rate (GFR) decline by more than 50% of its baseline value during the study period of 3 years. Given the retrospective nature of the study, it cannot be determined (and is rather unlikely) that ischemic nephropathy was the cause of the decline in kidney function in all 10 patients.

In a prospective cohort study in 304 patients with chronic kidney disease and renal artery stenosis who underwent surgical revascularization, Textor⁴ reported that the serum creatinine level showed a meaningful improvement afterward in 28%, worsened in 19.7%, and remained unchanged in 160 52.6%. (A “meaningful” change was defined as > 1.0 mg/dL.) Findings were similar in a cohort that underwent stenting.³³

Davies et al⁴¹ found that 20% of patients who underwent renal stenting had a persistent increase in serum creatinine of 0.5 mg/dL or more. These patients were nearly three times more likely (19% vs 7%) to eventually require dialysis, and they had a lower 5-year survival rate (41% vs 71%).

Zeller et al³⁹ found that renal function improved slightly in 52% of patients who received stents. The mean decrease in serum creatinine in this group was 0.22 mg/dL. However, the other 48% had a mean increase in serum creatinine of 1.1 mg/dL.

From these data we can conclude that, in an unselected population with renal artery stenosis, stenting provides no benefit to renal function, and that the risk of a worsening of renal function after intervention is roughly equal to the likelihood of achieving any benefit.

Other predictors of improvement in renal function have been proposed, but the evidence supporting them has not been consistent. For example, although Radermacher et al⁴² reported that a renal resistive index (which reflects arterial stiffness downstream of the stenosis) lower than 0.8 predicted a response in renal function, this finding has not been reliably reproduced.^43,44 Similarly, while several studies suggest that patients with milder renal dysfunction have a higher likelihood of a renal response,^45,46 other studies suggest either that the opposite is true³⁹ or that baseline renal function alone has no impact on outcome.⁴⁷

In addition, once significant renal atrophy occurs, revascularization may not help much, since irreversible sclerosis has developed. Thus, the goal is to identify kidneys being harmed by renal artery stenosis during the ischemic phase, when the tissue is still viable.

Unfortunately, we still lack a good renal stress test—eg, analogous to the cardiac stress test—to diagnose reversible ischemia in the kidney. The captopril renal scan has that capability but is not accurate in patients with bilateral stenosis or a GFR less than 50 mL/min, severely limiting its applicability.²⁶ Newer technologies such as blood-oxygen-level-dependent (BOLD) magnetic resonance imaging are being investigated for such a role.⁴⁸

Cohort studies in patients with declining renal function

In several case series, patients whose renal function had been declining before intervention had impressive rates of better renal function afterward.^{33,39,47,49–54} In a prospective cohort study by Muray et al,⁴⁷ a rise in serum creatinine of more than 0.1 mg/mL/month before intervention seemed to predict an improvement in renal function afterward.

One would expect that, for renal function to respond to intervention, severe bilateral stenosis or unilateral stenosis to a solitary functioning kidney would need to be present. However, this was an inconsistent finding in these case series.^{33,39,47,52,53} The Angioplasty and Stent for Renal Artery Lesions (ASTRAL) trial,^6,7 discussed later, sheds a bit more light on this.

Stenting usually improves flash pulmonary edema

Acute pulmonary edema in the setting of bilateral renal artery stenosis seems to be a unique case in which improvement in clinical status can be expected in most patients after intervention. Blood pressure improves in 94% to 100% of patients,^28,31 renal function either improves or stabilizes in 77% to 91%,^28–31 and pulmonary edema resolves without recurrence in 77% to 100%.^28–30

NEW RANDOMIZED TRIALS: STAR, ASTRAL, AND CORAL

Despite the lack of evidence supporting revascularization of renal artery stenosis, many interventionalists practice under the assumption that the radiographic finding of renal artery stenosis alone is an indication for renal revascularization. Only three randomized controlled trials in the modern era attempt to examine this hypothesis: STAR, ASTRAL, and CORAL.

STAR: No clear benefit

The Stent Placement and Blood Pressure and Lipid-lowering for the Prevention of Progression of Renal Dysfunction Caused by Atherosclerotic Ostial Stenosis of the Renal Artery (STAR) trial⁵ was a European multicenter trial that enrolled 140 patients with ostial renal artery stenosis greater than 50%, blood pressure controlled to less than 140/90 mm Hg, and creatinine clearance 15 to 80 mL/min.

Patients were randomized to undergo stenting or medical therapy alone. High blood pressure was treated according to a protocol in which angiotensin-converting enzyme (ACE) inhibitors and angiotensin receptor blockers were relegated to second-line use. All patients received a statin, regardless of lipid levels.

At 2 years, the primary end point (a decline in creatinine clearance of 20% or greater) had been reached in 10 (16%) of the 64 patients in the stent group and 16 (22%) of the 76 patients in the medication group; the difference was not statistically significant (hazard ratio 0.73, 95% confidence interval 0.33–1.61). No difference was seen in the secondary end points of the degree of blood pressure control or the rates of cardiovascular morbidity and death.⁵

 

 

ASTRAL: Also no clear benefit

In the international, multicenter ASTRAL trial,^6,7 806 patients with at least one stenotic renal artery considered suitable for balloon angioplasty, stenting, or both⁷ were randomized to undergo intervention or medical management. Hypertension treatment was not specified by a protocol. The mean estimated GFR was 40 mL/min. Most patients (95%–96%) were on statin therapy. The primary outcome was the rate of decline of renal function over time. Secondary outcomes included blood pressure control, renal events, cardiovascular events, and death.

Results. At a mean follow-up of 33.6 months (range 1–4 years), no difference was noted between treatment groups in decline in renal function or blood pressure control at 1 year. Renal function worsened slightly in both groups.

The decline in renal function over time, calculated as the mean slope of the reciprocal of the serum creatinine level over time, was slightly slower in the revascularization group, but the difference was not statistically significant (−0.07 × 10⁻³ vs −0.13 × 10⁻³ L/μmol/year, P = .06). This difference did not appear until the last year of the study. There was no difference in the number of patients whose renal function improved or declined during the study.

There was no difference in the rate of any secondary outcome. The medical management group required a slightly higher number of antihypertensive drugs, reaching statistical but not clinical significance (2.97 vs 2.77 drugs, P = .03). More people in the revascularization group were taking ACE inhibitors or angiotensin receptor blockers. There was no difference in the number of patients on any antihypertensive therapy (97% vs 99%). Interestingly, amputations were more common in the revascularization group, occurring in 42 (12%) of the 386 patients in the revascularization group vs 29 (7%) of the 395 patients in the medical group (P = .04).

Seventeen percent of patients randomized to intervention did not have the procedure done. An as-treated analysis of the 317 (83%) patients randomized to revascularization who did receive it showed no differences in outcomes.

There were no differences in outcomes among specific, predefined subgroups based on severity of stenosis at baseline, renal length, baseline estimated GFR, baseline serum creatinine, and rate of progression of renal dysfunction before randomization.⁷

Comments. ASTRAL contradicts previous nonrandomized studies that suggested that rapidly declining renal function (loss of 20% in 1 year) predicts response to intervention. Considering the large number of patients with unilateral disease in the study, it would be interesting to see what effect stenting had on patients with both severe disease and declining renal function.

ASTRAL has been criticized because it lacked a central laboratory to interpret the severity of stenosis, it did not use a standardized intervention technique (5% of patients underwent angioplasty without stents, although this did not affect outcomes⁷), and patients were enrolled only if the clinician involved in the case was uncertain of the appropriate management.

This last issue raises the concern for selection bias toward inclusion of more difficult cases that may not respond to intervention. But these shortcomings are not serious enough to negate the fact that preliminary results from the largest randomized controlled trial to date confirm conclusions of other randomized trials, ie, that intervention in renal artery stenosis yields no benefits over medical management in most patients.

Based on the results of STAR and ASTRAL, the practice of indiscriminately revascularizing stenosed renal arteries without strong evidence that the procedure will provide a clinical benefit is no longer tenable. The challenge is to identify those few patients who will respond, and to intervene only on them. Unfortunately, none of the subgroups from ASTRAL helped characterize this population.

CORAL: A large trial is ongoing

The Cardiovascular Outcomes in Renal Artherosclerotic Lesions (CORAL) trial,⁸ an ongoing multicenter randomized controlled trial in the United States, may be of additional help.

Unlike ASTRAL, CORAL is studying patients who have difficult-to-control hypertension (systolic blood pressure ≥ 155 mm Hg on two or more drugs).⁸ Chronic kidney disease is not an exclusion criterion unless the serum creatinine concentration is greater than 3.0 mg/dL.

CORAL is using a standardized medical protocol to control blood pressure. In addition, use of embolic protection devices during stenting is encouraged. Hopefully, the large size (a goal of 1,080 patients) and the inclusion of patients with more marked hypertension will address the utility of intervention in higher-risk populations with renal artery stenosis.

RECOMMENDED APPROACH TO INTERVENTION IN RENAL ARTERY STENOSIS

As we wait for CORAL to be completed, we have two modern-era randomized controlled trials that leave us with fewer indications for renal intervention. Table 2 lists commonly cited indications for intervention in renal artery stenosis and the evidence to support them. As most of these are based on retrospective data or have conflicting support in the literature, their utility remains in question. At this point we can safely recommend:

• Patients with preserved or even decreased but stable renal function will not likely have a benefit in renal function after intervention.
• Patients with resistant hypertension may benefit.
• The best evidence supporting intervention is for bilateral stenosis with flash pulmonary edema, but the evidence is from retrospective studies.
• Stenting in bilateral disease without another indication has no apparent benefit.
• Declining renal function is not a guarantee of success.
• It is unclear if patients with severe bilateral stenosis or severe stenosis to a solitary functioning kidney with declining renal function will benefit. Anecdotally, they do respond more often, but as with many other indications for intervention that have gone by the wayside, this may not bear out when studied properly.

As the utility of intervention narrows, the scope of practice for such interventions should narrow accordingly. Attention should now be focusing on clinical, rather than radiographic, indications for intervening on renal artery stenosis.

Therefore, the decision to intervene must not be made solely by the interventionalist. A multidisciplinary approach should be adopted that at the very least includes the input of a nephrologist well versed in renal artery stenosis. In this way, the clinical risks and benefits of renal intervention can be discussed with the patient by providers who are likely to be involved in their care should renal function or hypertension fail to improve afterward.

RISK OF ATHEROEMBOLISM

While renal stenting yields improved technical success in the treatment of renal artery stenosis, it carries with it an increasingly common risk to kidney function: atheroembolism as the stent crushes the plaque against the vessel wall. This may lead to obstruction of the renal microvasculature, increasing the risk of irreversible damage to renal function.

Atheroembolic kidney disease can manifest as progressive renal failure occurring over weeks to months, commonly misdiagnosed as permanent damage from contrast nephropathy.⁵⁵

Embolic protection devices, inserted downstream of the lesion before stenting to catch any debris that may break loose, have been developed to help address this problem.

Holden et al ⁵⁷ prospectively studied 63 patients with renal artery stenosis and deteriorating renal function (undefined) who underwent stenting with an embolic protection device. At 6 months after the intervention, renal function had either improved or stabilized in 97% of patients, suggesting that many of the deleterious effects of stenting on renal function are related to atheroembolism.

The Prospective Randomized Study Comparing Renal Artery Stenting With or Without Distal Protection (RESIST) trial, in which renal dysfunction was mild and the GFR was not declining (average estimated GFR 59.3 mL/min), found contrary results.⁵⁷ In a two-by-two factorial study, patients were randomized to undergo stenting alone, stenting with the antiplatelet agent abciximab (ReoPro), stenting with an embolic protection device, or stenting with both abciximab and an embolic protection device. Interestingly, renal function declined in the first three groups, but remained stable in the group that received both abciximab and an embolic protection device.

ANTIPLATELET THERAPY AFTER RENAL STENTING: HOW LONG?

We have no data on the optimal duration of antiplatelet therapy after renal stenting, and guidelines from professional societies do not comment on it.⁵⁸ As a result, practice patterns vary significantly among practitioners.

While in-stent restenosis rates are acceptably low after renal stenting, the risks and side effects of antiplatelet therapy often lead to arbitrary withdrawal of these drugs. The effect on stent patency is yet to be determined.

FUTURE DEVELOPMENTS

Results of STAR and ASTRAL confirm the growing suspicion that the surge seen in the last decade in renal artery stenting should be coming to an end. We await results either from CORAL or possibly a post hoc analysis of ASTRAL that might identify potential high-risk groups that will benefit from renal intervention. And as embolic protection devices become more agile and suitable to different renal lesions, there remains the possibility that, due to lower rates of unidentified atheroembolic kidney disease, CORAL may demonstrate improved renal outcomes after stenting. If not, the search for the best means to predict who should have renal intervention will continue.

We know through experience that stenting does provide great benefits for some patients with renal artery stenosis. Furthermore, the clinical problem is too intriguing, and too profitable, to die altogether.

Author’s note: Atherosclerosis accounts for about 90% of cases of renal artery stenosis in people over age 40.¹ Fibromuscular dysplasia, the other major cause, is a separate topic; in this paper “renal artery stenosis” refers to atherosclerotic disease only.

Renal artery stenosis is very common, and the number of angioplasty-stenting procedures performed every year is on the rise. Yet there is no overwhelming evidence that intervention yields clinical benefits—ie, better blood pressure control or renal function— than does medical therapy.

See related editorial

Earlier randomized controlled trials comparing angioplasty without stents and medical management showed no impressive difference in blood pressure.^2,3 The data on renal function were even more questionable, with some studies suggesting that, with stenting, the chance of worsening renal function is equal to that of improvement.⁴

Two large randomized trials comparing renal intervention with medical therapy failed to show any benefit of intervention.^5–7 A third study is under way.⁸

It is time to strongly reconsider the current aggressive approach to revascularization of stenotic renal arteries and take a more coordinated, critical approach.

RENAL INTERVENTIONS ON THE RISE

Renal angioplasty began replacing surgical revascularization in the 1990s, as this less-invasive procedure became more readily available and was shown to have similar clinical outcomes.⁹ In the last decade, stent placement during angioplasty has become standard, improving the rates of technical success.

As these procedures have become easier to perform and their radiographic outcomes have become more consistent, interventionalists have become more likely, if they see stenosis in a renal artery, to intervene and insert a stent, regardless of proven benefit. In addition, interventionalists from at least three different specialties now compete for these procedures, often by looking at the renal arteries during angiography of other vascular beds (the “drive-by”).

As a result, the number of renal interventions has been rising. Medicare received 21,660 claims for renal artery interventions (surgery or angioplasty) in 2000, compared with 13,380 in 1996—an increase of 62%. However, the number of surgeries actually decreased by 45% during this time, while the number of percutaneous procedures increased by 240%. The number of endovascular claims for renal artery stenosis by cardiologists alone rose 390%.¹⁰ Since then, the reports on intervention have been mixed, with one report citing a continued increase in 2005 to 35,000 claims,¹¹ and another suggesting a decrease back to 1997 levels.¹²

HOW COMMON IS RENAL ARTERY STENOSIS?

The prevalence of renal artery stenosis depends on the definition used and the population screened. It is more common in older patients who have risk factors for other vascular diseases than in the general population.

Renal Doppler ultrasonography can detect stenosis only if the artery is narrowed by more than 60%. Hansen et al¹³ used ultrasonography to screen 870 people over age 65 and found a lesion (a narrowing of more than 60%) in 6.8%.

Angiography (direct, computed tomographic, or magnetic resonance) can detect less-severe stenosis. Thus, most angiographic studies define renal artery stenosis as a narrowing of more than 50%, and severe disease as a narrowing of more than 70%. Many experts believe that unilateral stenosis needs to be more than 70% to pose a risk to the kidney.^14,15

Angiographic prevalence studies have been performed only in patients who were undergoing angiography for another reason such as coronary or peripheral arterial disease that inherently places them at higher risk of renal artery stenosis. For instance, renal artery stenosis is found in 11% to 28% of patients undergoing diagnostic cardiac catheterization. ¹⁶

No studies of the prevalence of renal artery stenosis have been performed in the general population. Medicare data indicate that from 1999 to 2001 the incidence of diagnosed renal artery stenosis was 3.7 per 1,000 patientyears. ¹⁷ Holley et al,¹⁸ in an autopsy series, found renal artery stenosis of greater than 50% in 27% of patients over age 50 and in 56.4% of hypertensive patients. The prevalence was 10% in normotensive patients.

WHO IS AT RISK?

Factors associated with a higher risk of finding renal artery stenosis on a radiographic study include¹⁴:

• Older age
• Female sex
• Hypertension
• Three-vessel coronary artery disease
• Peripheral artery disease
• Chronic kidney disease
• Diabetes
• 
  Tobacco use
• A low level of high-density lipoprotein cholesterol
• The use of at least two cardiovascular drugs.

The prevalence of renal artery stenosis in at-risk populations ranges from 3% to 75% (Table  1).^2,4,6,19,20

HOW OFTEN DOES STENOSIS PROGRESS?

The reported rates of progression of atherosclerotic renal artery lesions vary depending on the type of imaging test used and the reason for doing it.

In studies that used duplex ultrasonography, roughly half of lesions smaller than 60% grew to greater than 60% over 3 years.^21,22 The risk of total occlusion of an artery was relatively low and depended on the severity of stenosis: 0.7% if the baseline stenosis was less than 60% and 2.3% to 7% if it was greater.^21,22

In a seminal study in 1984, Schreiber and colleagues²³ compared serial angiograms obtained a mean of 52 months apart in 85 patients who did not undergo intervention. Stenosis had progressed in 37 (44%), and to the point of total occlusion in 14 (16%). In contrast, a 1998 study found progression in 11.1% over 2.6 years, with older patients, women, and those with baseline coronary artery disease at higher risk.²⁴

The the rates of progression differed in these two studies probably because the indications for screening were different (clinical suspicion²³ vs routine screening during coronary angiography²⁴), as was the severity of stenosis at the time of diagnosis. Also, when these studies were done, fewer people were taking statins. Thus, similar studies, if repeated, might show even lower rates of progression.

Finally, progression of renal artery stenosis has not been correlated with worsening renal function.

FOUR CLINICAL PRESENTATIONS OF RENAL ARTERY STENOSIS

Renal artery stenosis can present in one of four ways:

Clinically silent stenosis. Because renal artery stenosis is most often found in older patients, who are more likely to have essential hypertension and chronic kidney disease due to other causes, it can be an incidental finding that is completely clinically silent.^16,25

Renovascular hypertension is defined as high blood pressure due to up-regulation of neurohormonal activity in response to decreased perfusion from renal artery stenosis. Renal artery stenosis is estimated to be the cause of hypertension in only 0.5% to 4.0% of hypertensive patients, or in 26% of patients with secondary hypertension.³

Ischemic nephropathy is more difficult to define because ischemia alone rarely explains the damage done to the kidneys. Activation of neurohormonal pathways and microvascular injury are thought to contribute to oxidative stress and fibrosis.²⁶ These phenomena may explain why similar degrees of stenosis lead to varying degrees of kidney damage in different patients and why the severity of stenosis does not correlate with the degree of renal dysfunction.²⁷

Furthermore, stenosis may lead to irreversible but stable kidney damage. It is therefore not surprising that, in studies in unselected populations (ie, studies that included patients with all presentations of renal artery stenosis, not just those more likely to benefit), up to two-thirds of renal interventions yielded no clinical benefit.²⁵

As a result, if we define ischemic nephropathy as renal artery stenosis with renal dysfunction not attributable to another cause, we probably will overestimate the prevalence of ischemic nephropathy, leading to overly optimistic expectations about the response to revascularization.

Recurrent “flash” pulmonary edema is a less common presentation, usually occurring in patients with critical bilateral renal artery stenosis or unilateral stenosis in an artery supplying a solitary functioning kidney. Most have severe hypertension (average systolic blood pressure 174–207 mm Hg) and poor renal function.^28–30

The association between pulmonary edema and bilateral renal artery stenosis was first noted in 1998 by Pickering et al,³¹ who in several case series showed that 82% to 92% of patients with recurrent pulmonary edema and renal artery stenosis had bilateral stenosis, compared with 20% to 65% of those with other presentations. Later case series corroborated this finding: 85% to 100% of patients with renal artery stenosis and pulmonary edema had bilateral stenosis.^28–30

STENTING IS NOW STANDARD

Stenting has become standard in the endovascular treatment of renal artery stenosis.

Most atherosclerotic renal artery lesions are located in the ostium (ie, where the artery branches off from the aorta), and many are extensions of calcified aortic plaque.^26,32 These hard lesions tend to rebound to their original shape more often with balloon angioplasty alone. Stenting provides the additional force needed to permanently disrupt the lesion, leading to a longer-lasting result.

Rates of technical success (dilating the artery during the intervention) are higher with stents than without them (98% vs 46%– 77%).^33,34 If the lesion is ostial, this difference is even more impressive (75% vs 29%). In addition, restenosis rates at 6 months are lower with stents (14% vs 26%–48%).³⁴

GOALS: LOWER THE BLOOD PRESSURE, SAVE THE KIDNEY

Because endovascular procedures pose some risk to the patient, it is critical to intervene only in patients most likely to respond clinically. The decision to intervene depends largely on the clinical goal, which should depend on the clinical presentation.

However, if renal artery stenosis is clinically silent, most of the evidence suggests that intervention has no benefit. Furthermore, although retrospective studies have indicated that intervention may improve survival rates,^35,36 prospective studies have not. Similarly, studies have not shown that intervention generally improves cardiovascular outcomes, even though renal artery stenosis is associated with cardiovascular risk.

Hypertension plus stenosis is not necessarily renovascular hypertension

Essential hypertension and clinically silent renal artery stenosis often coexist, which is why blood pressure control often does not improve after stenting. Also, essential hypertension often coexists with renovascular hypertension.³⁷ In this situation, stenting may not eliminate the need for antihypertensive drugs, although it may improve blood pressure control and decrease the drug burden.

Before stents came into use, several randomized controlled trials found that blood pressure was no better controlled after angioplasty, ^2,3,38 except in cases of bilateral stenosis.² This may be because stenosis tended to recur after angioplasty without stents.

The 2000 Dutch Renal Artery Stenosis Intervention Cooperative (DRASTIC) study was the first randomized controlled trial to examine the effect of angioplasty on blood pressure control in renal artery stenosis.³⁸ It had significant design flaws. For example, many patients crossed over from the medical management group to the intervention group because their hypertension was resistant to medical therapy. Overall, intervention (balloon angioplasty without stents in 54 of 56 patients, with stents in the other 2) carried no benefit. However, in subgroup analysis, the patients who crossed over because of resistant hypertension (failure of a three-drug regimen) were more likely to benefit from angioplasty. This suggested that risk stratification should take place early on, before proceeding with revascularization.

With stents, Zeller,³⁹ in a prospective nonrandomized study, found that the mean arterial pressure decreased by 10 mm Hg. Randomized trials (see below) have failed to demonstrate such a benefit.

 

 

Stenting may not improve renal function

Coincidental renal artery stenosis in a patient with unrelated chronic kidney disease is very hard to differentiate from true ischemic nephropathy. Furthermore, most patients with ischemic nephropathy do not benefit from revascularization, making it challenging to identify those few whose renal function may respond.

Given that patients with chronic kidney disease tend to have a higher risk of cardiovascular disease, it is not surprising that 15% of them may have renal artery stenosis,⁴ most often incidental.

Chábová⁴⁰ examined the outcomes of 68 patients who had chronic kidney disease and a renal artery lesion larger than 70% who did not undergo angioplasty. In only 10 (15%) of the patients did the glomerular filtration rate (GFR) decline by more than 50% of its baseline value during the study period of 3 years. Given the retrospective nature of the study, it cannot be determined (and is rather unlikely) that ischemic nephropathy was the cause of the decline in kidney function in all 10 patients.

In a prospective cohort study in 304 patients with chronic kidney disease and renal artery stenosis who underwent surgical revascularization, Textor⁴ reported that the serum creatinine level showed a meaningful improvement afterward in 28%, worsened in 19.7%, and remained unchanged in 160 52.6%. (A “meaningful” change was defined as > 1.0 mg/dL.) Findings were similar in a cohort that underwent stenting.³³

Davies et al⁴¹ found that 20% of patients who underwent renal stenting had a persistent increase in serum creatinine of 0.5 mg/dL or more. These patients were nearly three times more likely (19% vs 7%) to eventually require dialysis, and they had a lower 5-year survival rate (41% vs 71%).

Zeller et al³⁹ found that renal function improved slightly in 52% of patients who received stents. The mean decrease in serum creatinine in this group was 0.22 mg/dL. However, the other 48% had a mean increase in serum creatinine of 1.1 mg/dL.

From these data we can conclude that, in an unselected population with renal artery stenosis, stenting provides no benefit to renal function, and that the risk of a worsening of renal function after intervention is roughly equal to the likelihood of achieving any benefit.

Other predictors of improvement in renal function have been proposed, but the evidence supporting them has not been consistent. For example, although Radermacher et al⁴² reported that a renal resistive index (which reflects arterial stiffness downstream of the stenosis) lower than 0.8 predicted a response in renal function, this finding has not been reliably reproduced.^43,44 Similarly, while several studies suggest that patients with milder renal dysfunction have a higher likelihood of a renal response,^45,46 other studies suggest either that the opposite is true³⁹ or that baseline renal function alone has no impact on outcome.⁴⁷

In addition, once significant renal atrophy occurs, revascularization may not help much, since irreversible sclerosis has developed. Thus, the goal is to identify kidneys being harmed by renal artery stenosis during the ischemic phase, when the tissue is still viable.

Unfortunately, we still lack a good renal stress test—eg, analogous to the cardiac stress test—to diagnose reversible ischemia in the kidney. The captopril renal scan has that capability but is not accurate in patients with bilateral stenosis or a GFR less than 50 mL/min, severely limiting its applicability.²⁶ Newer technologies such as blood-oxygen-level-dependent (BOLD) magnetic resonance imaging are being investigated for such a role.⁴⁸

Cohort studies in patients with declining renal function

In several case series, patients whose renal function had been declining before intervention had impressive rates of better renal function afterward.^{33,39,47,49–54} In a prospective cohort study by Muray et al,⁴⁷ a rise in serum creatinine of more than 0.1 mg/mL/month before intervention seemed to predict an improvement in renal function afterward.

One would expect that, for renal function to respond to intervention, severe bilateral stenosis or unilateral stenosis to a solitary functioning kidney would need to be present. However, this was an inconsistent finding in these case series.^{33,39,47,52,53} The Angioplasty and Stent for Renal Artery Lesions (ASTRAL) trial,^6,7 discussed later, sheds a bit more light on this.

Stenting usually improves flash pulmonary edema

Acute pulmonary edema in the setting of bilateral renal artery stenosis seems to be a unique case in which improvement in clinical status can be expected in most patients after intervention. Blood pressure improves in 94% to 100% of patients,^28,31 renal function either improves or stabilizes in 77% to 91%,^28–31 and pulmonary edema resolves without recurrence in 77% to 100%.^28–30

NEW RANDOMIZED TRIALS: STAR, ASTRAL, AND CORAL

Despite the lack of evidence supporting revascularization of renal artery stenosis, many interventionalists practice under the assumption that the radiographic finding of renal artery stenosis alone is an indication for renal revascularization. Only three randomized controlled trials in the modern era attempt to examine this hypothesis: STAR, ASTRAL, and CORAL.

STAR: No clear benefit

The Stent Placement and Blood Pressure and Lipid-lowering for the Prevention of Progression of Renal Dysfunction Caused by Atherosclerotic Ostial Stenosis of the Renal Artery (STAR) trial⁵ was a European multicenter trial that enrolled 140 patients with ostial renal artery stenosis greater than 50%, blood pressure controlled to less than 140/90 mm Hg, and creatinine clearance 15 to 80 mL/min.

Patients were randomized to undergo stenting or medical therapy alone. High blood pressure was treated according to a protocol in which angiotensin-converting enzyme (ACE) inhibitors and angiotensin receptor blockers were relegated to second-line use. All patients received a statin, regardless of lipid levels.

At 2 years, the primary end point (a decline in creatinine clearance of 20% or greater) had been reached in 10 (16%) of the 64 patients in the stent group and 16 (22%) of the 76 patients in the medication group; the difference was not statistically significant (hazard ratio 0.73, 95% confidence interval 0.33–1.61). No difference was seen in the secondary end points of the degree of blood pressure control or the rates of cardiovascular morbidity and death.⁵

 

 

ASTRAL: Also no clear benefit

In the international, multicenter ASTRAL trial,^6,7 806 patients with at least one stenotic renal artery considered suitable for balloon angioplasty, stenting, or both⁷ were randomized to undergo intervention or medical management. Hypertension treatment was not specified by a protocol. The mean estimated GFR was 40 mL/min. Most patients (95%–96%) were on statin therapy. The primary outcome was the rate of decline of renal function over time. Secondary outcomes included blood pressure control, renal events, cardiovascular events, and death.

Results. At a mean follow-up of 33.6 months (range 1–4 years), no difference was noted between treatment groups in decline in renal function or blood pressure control at 1 year. Renal function worsened slightly in both groups.

The decline in renal function over time, calculated as the mean slope of the reciprocal of the serum creatinine level over time, was slightly slower in the revascularization group, but the difference was not statistically significant (−0.07 × 10⁻³ vs −0.13 × 10⁻³ L/μmol/year, P = .06). This difference did not appear until the last year of the study. There was no difference in the number of patients whose renal function improved or declined during the study.

There was no difference in the rate of any secondary outcome. The medical management group required a slightly higher number of antihypertensive drugs, reaching statistical but not clinical significance (2.97 vs 2.77 drugs, P = .03). More people in the revascularization group were taking ACE inhibitors or angiotensin receptor blockers. There was no difference in the number of patients on any antihypertensive therapy (97% vs 99%). Interestingly, amputations were more common in the revascularization group, occurring in 42 (12%) of the 386 patients in the revascularization group vs 29 (7%) of the 395 patients in the medical group (P = .04).

Seventeen percent of patients randomized to intervention did not have the procedure done. An as-treated analysis of the 317 (83%) patients randomized to revascularization who did receive it showed no differences in outcomes.

There were no differences in outcomes among specific, predefined subgroups based on severity of stenosis at baseline, renal length, baseline estimated GFR, baseline serum creatinine, and rate of progression of renal dysfunction before randomization.⁷

Comments. ASTRAL contradicts previous nonrandomized studies that suggested that rapidly declining renal function (loss of 20% in 1 year) predicts response to intervention. Considering the large number of patients with unilateral disease in the study, it would be interesting to see what effect stenting had on patients with both severe disease and declining renal function.

ASTRAL has been criticized because it lacked a central laboratory to interpret the severity of stenosis, it did not use a standardized intervention technique (5% of patients underwent angioplasty without stents, although this did not affect outcomes⁷), and patients were enrolled only if the clinician involved in the case was uncertain of the appropriate management.

This last issue raises the concern for selection bias toward inclusion of more difficult cases that may not respond to intervention. But these shortcomings are not serious enough to negate the fact that preliminary results from the largest randomized controlled trial to date confirm conclusions of other randomized trials, ie, that intervention in renal artery stenosis yields no benefits over medical management in most patients.

Based on the results of STAR and ASTRAL, the practice of indiscriminately revascularizing stenosed renal arteries without strong evidence that the procedure will provide a clinical benefit is no longer tenable. The challenge is to identify those few patients who will respond, and to intervene only on them. Unfortunately, none of the subgroups from ASTRAL helped characterize this population.

CORAL: A large trial is ongoing

The Cardiovascular Outcomes in Renal Artherosclerotic Lesions (CORAL) trial,⁸ an ongoing multicenter randomized controlled trial in the United States, may be of additional help.

Unlike ASTRAL, CORAL is studying patients who have difficult-to-control hypertension (systolic blood pressure ≥ 155 mm Hg on two or more drugs).⁸ Chronic kidney disease is not an exclusion criterion unless the serum creatinine concentration is greater than 3.0 mg/dL.

CORAL is using a standardized medical protocol to control blood pressure. In addition, use of embolic protection devices during stenting is encouraged. Hopefully, the large size (a goal of 1,080 patients) and the inclusion of patients with more marked hypertension will address the utility of intervention in higher-risk populations with renal artery stenosis.

RECOMMENDED APPROACH TO INTERVENTION IN RENAL ARTERY STENOSIS

As we wait for CORAL to be completed, we have two modern-era randomized controlled trials that leave us with fewer indications for renal intervention. Table 2 lists commonly cited indications for intervention in renal artery stenosis and the evidence to support them. As most of these are based on retrospective data or have conflicting support in the literature, their utility remains in question. At this point we can safely recommend:

• Patients with preserved or even decreased but stable renal function will not likely have a benefit in renal function after intervention.
• Patients with resistant hypertension may benefit.
• The best evidence supporting intervention is for bilateral stenosis with flash pulmonary edema, but the evidence is from retrospective studies.
• Stenting in bilateral disease without another indication has no apparent benefit.
• Declining renal function is not a guarantee of success.
• It is unclear if patients with severe bilateral stenosis or severe stenosis to a solitary functioning kidney with declining renal function will benefit. Anecdotally, they do respond more often, but as with many other indications for intervention that have gone by the wayside, this may not bear out when studied properly.

As the utility of intervention narrows, the scope of practice for such interventions should narrow accordingly. Attention should now be focusing on clinical, rather than radiographic, indications for intervening on renal artery stenosis.

Therefore, the decision to intervene must not be made solely by the interventionalist. A multidisciplinary approach should be adopted that at the very least includes the input of a nephrologist well versed in renal artery stenosis. In this way, the clinical risks and benefits of renal intervention can be discussed with the patient by providers who are likely to be involved in their care should renal function or hypertension fail to improve afterward.

RISK OF ATHEROEMBOLISM

While renal stenting yields improved technical success in the treatment of renal artery stenosis, it carries with it an increasingly common risk to kidney function: atheroembolism as the stent crushes the plaque against the vessel wall. This may lead to obstruction of the renal microvasculature, increasing the risk of irreversible damage to renal function.

Atheroembolic kidney disease can manifest as progressive renal failure occurring over weeks to months, commonly misdiagnosed as permanent damage from contrast nephropathy.⁵⁵

Embolic protection devices, inserted downstream of the lesion before stenting to catch any debris that may break loose, have been developed to help address this problem.

Holden et al ⁵⁷ prospectively studied 63 patients with renal artery stenosis and deteriorating renal function (undefined) who underwent stenting with an embolic protection device. At 6 months after the intervention, renal function had either improved or stabilized in 97% of patients, suggesting that many of the deleterious effects of stenting on renal function are related to atheroembolism.

The Prospective Randomized Study Comparing Renal Artery Stenting With or Without Distal Protection (RESIST) trial, in which renal dysfunction was mild and the GFR was not declining (average estimated GFR 59.3 mL/min), found contrary results.⁵⁷ In a two-by-two factorial study, patients were randomized to undergo stenting alone, stenting with the antiplatelet agent abciximab (ReoPro), stenting with an embolic protection device, or stenting with both abciximab and an embolic protection device. Interestingly, renal function declined in the first three groups, but remained stable in the group that received both abciximab and an embolic protection device.

ANTIPLATELET THERAPY AFTER RENAL STENTING: HOW LONG?

We have no data on the optimal duration of antiplatelet therapy after renal stenting, and guidelines from professional societies do not comment on it.⁵⁸ As a result, practice patterns vary significantly among practitioners.

While in-stent restenosis rates are acceptably low after renal stenting, the risks and side effects of antiplatelet therapy often lead to arbitrary withdrawal of these drugs. The effect on stent patency is yet to be determined.

FUTURE DEVELOPMENTS

Results of STAR and ASTRAL confirm the growing suspicion that the surge seen in the last decade in renal artery stenting should be coming to an end. We await results either from CORAL or possibly a post hoc analysis of ASTRAL that might identify potential high-risk groups that will benefit from renal intervention. And as embolic protection devices become more agile and suitable to different renal lesions, there remains the possibility that, due to lower rates of unidentified atheroembolic kidney disease, CORAL may demonstrate improved renal outcomes after stenting. If not, the search for the best means to predict who should have renal intervention will continue.

We know through experience that stenting does provide great benefits for some patients with renal artery stenosis. Furthermore, the clinical problem is too intriguing, and too profitable, to die altogether.

Should patients with poorly controlled asthma be treated empirically for gastroesphageal reflux disease (GERD)?

Current guidelines¹ indicate that trying a proton pump inhibitor may be worthwhile. However, the results of a recent multicenter trial² indicate that this does not help control asthma symptoms and that we need to reevaluate the guidelines and focus on other factors that can worsen asthma control.

REFLUX DISEASE IS LINKED TO ASTHMA

GERD’s association with asthma has long been recognized. Asthma patients have a higher prevalence of GERD than the general population, with reported rates of 20% to 80%.^3–8

GERD may worsen asthma via several mechanisms. If stomach acid gets into the airway, it can induce bronchoconstriction, vagal reflexes, and chronic airway inflammation, all of which can increase airway reactivity.^9–16 Chronic reflux can also cause inflammation of the esophagus, which can exacerbate cough and possibly bronchospasm via neurogenic mechanisms.¹⁷

In turn, asthma may worsen GERD. Airway restriction can lead to hyperinflation and increased negative inspiratory pleural pressure, both of which may reduce the effectiveness of the lower esophageal sphincter.¹⁸ In addition, the beta-agonists and methylxanthines used to treat asthma may impair function of the lower esophageal sphincter and exacerbate reflux.^18–20

CURRENT GUIDELINES ARE BASED ON LIMITED INFORMATION

The symptoms of GERD and asthma are nonspecific and can be similar (chest tightness, chest discomfort), which can make it challenging for clinicians or patients to distinguish asthma from GERD.² Moreover, in asthma patients, GERD often presents without classic symptoms such as heartburn, and thus has been labeled “silent” GERD.

However, these studies all had significant limitations, such as small sample size. Also, the definitions of asthma and GERD differed from study to study. In some cases, the definition of GERD included self-reported GERD, which often fails to correlate with GERD documented with esophageal pH monitoring in asthma patients.¹ These limitations were highlighted in a Cochrane review,³⁰ which found that asthma patients with GERD showed no overall improvement in asthma after treatment of reflux. It concluded that small groups of patients may benefit, but that predicting who will respond is difficult.

Larger randomized controlled trials^28,29 attempted to address some of these limitations, with varying results.

Littner et al²⁹ gave lansoprazole (Prevacid) 30 mg twice daily or placebo to 207 patients with moderate to severe asthma and symptomatic GERD and saw no improvement in daily asthma symptoms, ie, asthma control in the active-treatment group. While these patients had an improvement in symptoms of severe reflux, their overall quality-of-life scores were similar to those of the placebo group. Of note, patients needing more than one type of drug for asthma control had a lower rate of asthma exacerbations.

Kiljander et al²⁸ gave esomeprazole (Nexium) 40 mg twice daily or placebo to 770 patients who had mild to moderate asthma and symptoms of nocturnal asthma with or without symptoms of GERD. The only benefit was a slight improvement in peak expiratory flow in those with symptoms of both GERD and nocturnal asthma, and this was most significant in patients taking long-acting beta-agonists. Other measures—eg, the forced expiratory volume in the first second (FEV₁), use of a beta-agonist, symptom scores, and nocturnal awakenings—did not improve.

In both of these studies,^28,29 patients reported symptoms of GERD, so they did not have silent GERD.

THE DESIGN OF SARA

In SARA, 412 patients age 18 and older with inadequately controlled asthma were randomized to receive esomeprazole 40 mg twice a day or placebo for 24 weeks. Inadequate control was defined as a score of 1.5 or higher on the Juniper Asthma Control Questionnaire³¹ despite treatment with inhaled corticosteroids. Patients had no symptoms of GERD. The 40-mg twice-daily dosage of esomeprazole was chosen because it is known to suppress more than 90% of acid reflux.^24,32

All patients completed a baseline asthma diary, recording peak expiratory flow rates, asthma symptoms, nighttime symptoms, and beta-agonist use. This information was collected every 4 weeks throughout the trial.

All participants also underwent esophageal pH monitoring for an objective confirmation of GERD. Patients were randomized independently of the results of the pH probe; in fact, investigators and patients were blinded to these results.

The primary outcome measure was the rate of episodes of poor asthma control, with poor control defined as any of the following:

• A decrease of 30% or more in the morning peak expiratory flow rate on 2 consecutive days, compared with the patient’s best rate during the run-in period
• An urgent visit, defined as an unscheduled health care visit, for asthma symptoms
• The need for a course of oral prednisone for treatment of asthma.

Asthma was defined as doctor-diagnosed, plus either a positive methacholine challenge test (a concentration of methacholine causing a 20% reduction in FEV₁ [PC₂₀] < 16 mg/mL) or a positive bronchodilator response (a 12% increase in FEV₁) to an inhaled beta-agonist. Participants had no other indication for acid suppression, including symptoms of GERD or previously diagnosed erosive esophageal or gastric disease.

Acid reflux was evaluated by ambulatory pH monitoring, which had to last at least 16 hours and span one meal and 2 hours in the recumbent position. Reflux was present if the pH was less than 4.0 for more than 5.8% of total time, 8.2% of time upright, or 3.5% of time lying down.³³ Episodes and severity were measured by the Gastroesophageal Reflux Disease Symptom Assessment Scale.³⁴

SARA RESULTS: NO IMPROVEMENT IN ASTHMA WITH GERD TREATMENT

The SARA treatment and control groups had similar baseline characteristics, with similar asthma symptoms. Most of the patients were women: 72% of the placebo group and 64% of the esomeprazole group. Most had baseline spirometric results at the lower end of normal (the mean FEV₁ was 76% ± 16 SD in the treatment group and 78% ± 15 in the placebo group) and had very poor asthma control, with an average Juniper Asthma Control Questionnaire score of 1.9 (> 1.5 is considered poor control).³¹ GERD was documented with esophageal pH monitoring in 40% of patients, showing that a significant number had silent GERD.

Episodes of poor asthma control occurred with similar frequency in the esomeprazole and placebo groups (2.5 vs 2.3 events per personyear, P = .66). Treatment made no difference in this end point regardless of the baseline results of pH monitoring. No treatment effect was noted in the individual components of the episodes of poor asthma control or in secondary outcomes, including pulmonary function, airway reactivity, asthma control, symptom scores, nocturnal awakening, or quality of life.

In addition, subgroup analysis failed to identify any group—including those with documented reflux on pH probe testing or those receiving a long-acting beta-agonist—who benefited from proton pump inhibitor therapy.

The investigators concluded that these data suggest treatment of silent GERD does not improve asthma control and thus that a reevaluation of current guidelines and clinical practice is warranted.²

ISSUES REMAIN

This large clinical trial, in which asthma and GERD were well defined and objectively measured, was robustly negative in terms of showing any benefit of treatment of silent GERD on asthma control. The study population was representative of those for whom such a treatment is recommended in the current NIH guidelines, which are based on data published prior to SARA.

However, while SARA was well designed and had clear results, it had some limitations, and some issues regarding GERD and asthma remain unanswered.

Is acid the only problem in GERD? SARA focused on acidic GERD. Aspiration of substances such as pancreatic enzymes, pepsin, and bile has also been shown to induce symptoms in asthma patients.^2,32,35 In addition, distention of the esophagus and stimulation of neurogenically mediated reflexes can cause symptoms or neurogenic airway inflammation that is not mitigated by drugs that target acid reflux.³²

Indirectly supporting this theory is evidence that surgical interventions such as fundoplication can improve asthma symptoms. ³⁶ However, this evidence is only from small studies with significant limitations.

Is proximal GERD worse than distal GERD? SARA did not address whether proximal and distal reflux may affect asthma differently. The importance of proximal reflux in asthma has not been clearly established, but there is evidence that patients with proximal GERD have a higher incidence of nocturnal cough than patients who have only distal reflux. ³⁷

Dimango et al³⁸ recently reported additional data from SARA in which patients with poorly controlled asthma underwent both proximal and distal esophageal pH monitoring to see if proximal GERD was associated with poor asthma control: 304 patients underwent dual pH-probe assessment and 38% of them had proximal reflux. The authors found no difference between those with and without proximal GERD with regard to nocturnal awakenings, need to use a rescue inhaler, inhaled controller medication dose, lung function, or airway reactivity by methacholine challenge. However, they did find that those with proximal GERD had worse asthma quality-of-life scores, and worse health-related quality-of-life scores and were more likely to complain of cough.

Thus, it appears that proximal GERD may worsen quality of life in asthmatic patients but does not worsen asthma control.

SARA RESULTS: IMPLICATIONS FOR MANAGEMENT

The SARA results suggest that patients with poorly controlled asthma who are on adequate controller medications should not be treated empirically for silent GERD in the expectation that the asthma will improve. Rather, they suggest that the focus should be on other factors that can worsen asthma control, such as the ability to properly use an inhaler, the ability to afford medications, compliance with drug treatment, and adequate control of other significant comorbidities such as allergic bronchopulmonary aspergillosis, sinusitis, allergic rhinitis, vocal cord dysfunction, and occult heart disease. The most recent NIH guidelines also suggest considering referral to an asthma specialist if symptoms persist despite adequate controller therapy.

Should patients with poorly controlled asthma be treated empirically for gastroesphageal reflux disease (GERD)?

Current guidelines¹ indicate that trying a proton pump inhibitor may be worthwhile. However, the results of a recent multicenter trial² indicate that this does not help control asthma symptoms and that we need to reevaluate the guidelines and focus on other factors that can worsen asthma control.

REFLUX DISEASE IS LINKED TO ASTHMA

GERD’s association with asthma has long been recognized. Asthma patients have a higher prevalence of GERD than the general population, with reported rates of 20% to 80%.^3–8

GERD may worsen asthma via several mechanisms. If stomach acid gets into the airway, it can induce bronchoconstriction, vagal reflexes, and chronic airway inflammation, all of which can increase airway reactivity.^9–16 Chronic reflux can also cause inflammation of the esophagus, which can exacerbate cough and possibly bronchospasm via neurogenic mechanisms.¹⁷

In turn, asthma may worsen GERD. Airway restriction can lead to hyperinflation and increased negative inspiratory pleural pressure, both of which may reduce the effectiveness of the lower esophageal sphincter.¹⁸ In addition, the beta-agonists and methylxanthines used to treat asthma may impair function of the lower esophageal sphincter and exacerbate reflux.^18–20

CURRENT GUIDELINES ARE BASED ON LIMITED INFORMATION

The symptoms of GERD and asthma are nonspecific and can be similar (chest tightness, chest discomfort), which can make it challenging for clinicians or patients to distinguish asthma from GERD.² Moreover, in asthma patients, GERD often presents without classic symptoms such as heartburn, and thus has been labeled “silent” GERD.

However, these studies all had significant limitations, such as small sample size. Also, the definitions of asthma and GERD differed from study to study. In some cases, the definition of GERD included self-reported GERD, which often fails to correlate with GERD documented with esophageal pH monitoring in asthma patients.¹ These limitations were highlighted in a Cochrane review,³⁰ which found that asthma patients with GERD showed no overall improvement in asthma after treatment of reflux. It concluded that small groups of patients may benefit, but that predicting who will respond is difficult.

Larger randomized controlled trials^28,29 attempted to address some of these limitations, with varying results.

Littner et al²⁹ gave lansoprazole (Prevacid) 30 mg twice daily or placebo to 207 patients with moderate to severe asthma and symptomatic GERD and saw no improvement in daily asthma symptoms, ie, asthma control in the active-treatment group. While these patients had an improvement in symptoms of severe reflux, their overall quality-of-life scores were similar to those of the placebo group. Of note, patients needing more than one type of drug for asthma control had a lower rate of asthma exacerbations.

Kiljander et al²⁸ gave esomeprazole (Nexium) 40 mg twice daily or placebo to 770 patients who had mild to moderate asthma and symptoms of nocturnal asthma with or without symptoms of GERD. The only benefit was a slight improvement in peak expiratory flow in those with symptoms of both GERD and nocturnal asthma, and this was most significant in patients taking long-acting beta-agonists. Other measures—eg, the forced expiratory volume in the first second (FEV₁), use of a beta-agonist, symptom scores, and nocturnal awakenings—did not improve.

In both of these studies,^28,29 patients reported symptoms of GERD, so they did not have silent GERD.

THE DESIGN OF SARA

In SARA, 412 patients age 18 and older with inadequately controlled asthma were randomized to receive esomeprazole 40 mg twice a day or placebo for 24 weeks. Inadequate control was defined as a score of 1.5 or higher on the Juniper Asthma Control Questionnaire³¹ despite treatment with inhaled corticosteroids. Patients had no symptoms of GERD. The 40-mg twice-daily dosage of esomeprazole was chosen because it is known to suppress more than 90% of acid reflux.^24,32

All patients completed a baseline asthma diary, recording peak expiratory flow rates, asthma symptoms, nighttime symptoms, and beta-agonist use. This information was collected every 4 weeks throughout the trial.

All participants also underwent esophageal pH monitoring for an objective confirmation of GERD. Patients were randomized independently of the results of the pH probe; in fact, investigators and patients were blinded to these results.

The primary outcome measure was the rate of episodes of poor asthma control, with poor control defined as any of the following:

• A decrease of 30% or more in the morning peak expiratory flow rate on 2 consecutive days, compared with the patient’s best rate during the run-in period
• An urgent visit, defined as an unscheduled health care visit, for asthma symptoms
• The need for a course of oral prednisone for treatment of asthma.

Asthma was defined as doctor-diagnosed, plus either a positive methacholine challenge test (a concentration of methacholine causing a 20% reduction in FEV₁ [PC₂₀] < 16 mg/mL) or a positive bronchodilator response (a 12% increase in FEV₁) to an inhaled beta-agonist. Participants had no other indication for acid suppression, including symptoms of GERD or previously diagnosed erosive esophageal or gastric disease.

Acid reflux was evaluated by ambulatory pH monitoring, which had to last at least 16 hours and span one meal and 2 hours in the recumbent position. Reflux was present if the pH was less than 4.0 for more than 5.8% of total time, 8.2% of time upright, or 3.5% of time lying down.³³ Episodes and severity were measured by the Gastroesophageal Reflux Disease Symptom Assessment Scale.³⁴

SARA RESULTS: NO IMPROVEMENT IN ASTHMA WITH GERD TREATMENT

The SARA treatment and control groups had similar baseline characteristics, with similar asthma symptoms. Most of the patients were women: 72% of the placebo group and 64% of the esomeprazole group. Most had baseline spirometric results at the lower end of normal (the mean FEV₁ was 76% ± 16 SD in the treatment group and 78% ± 15 in the placebo group) and had very poor asthma control, with an average Juniper Asthma Control Questionnaire score of 1.9 (> 1.5 is considered poor control).³¹ GERD was documented with esophageal pH monitoring in 40% of patients, showing that a significant number had silent GERD.

Episodes of poor asthma control occurred with similar frequency in the esomeprazole and placebo groups (2.5 vs 2.3 events per personyear, P = .66). Treatment made no difference in this end point regardless of the baseline results of pH monitoring. No treatment effect was noted in the individual components of the episodes of poor asthma control or in secondary outcomes, including pulmonary function, airway reactivity, asthma control, symptom scores, nocturnal awakening, or quality of life.

In addition, subgroup analysis failed to identify any group—including those with documented reflux on pH probe testing or those receiving a long-acting beta-agonist—who benefited from proton pump inhibitor therapy.

The investigators concluded that these data suggest treatment of silent GERD does not improve asthma control and thus that a reevaluation of current guidelines and clinical practice is warranted.²

ISSUES REMAIN

This large clinical trial, in which asthma and GERD were well defined and objectively measured, was robustly negative in terms of showing any benefit of treatment of silent GERD on asthma control. The study population was representative of those for whom such a treatment is recommended in the current NIH guidelines, which are based on data published prior to SARA.

However, while SARA was well designed and had clear results, it had some limitations, and some issues regarding GERD and asthma remain unanswered.

Is acid the only problem in GERD? SARA focused on acidic GERD. Aspiration of substances such as pancreatic enzymes, pepsin, and bile has also been shown to induce symptoms in asthma patients.^2,32,35 In addition, distention of the esophagus and stimulation of neurogenically mediated reflexes can cause symptoms or neurogenic airway inflammation that is not mitigated by drugs that target acid reflux.³²

Indirectly supporting this theory is evidence that surgical interventions such as fundoplication can improve asthma symptoms. ³⁶ However, this evidence is only from small studies with significant limitations.

Is proximal GERD worse than distal GERD? SARA did not address whether proximal and distal reflux may affect asthma differently. The importance of proximal reflux in asthma has not been clearly established, but there is evidence that patients with proximal GERD have a higher incidence of nocturnal cough than patients who have only distal reflux. ³⁷

Dimango et al³⁸ recently reported additional data from SARA in which patients with poorly controlled asthma underwent both proximal and distal esophageal pH monitoring to see if proximal GERD was associated with poor asthma control: 304 patients underwent dual pH-probe assessment and 38% of them had proximal reflux. The authors found no difference between those with and without proximal GERD with regard to nocturnal awakenings, need to use a rescue inhaler, inhaled controller medication dose, lung function, or airway reactivity by methacholine challenge. However, they did find that those with proximal GERD had worse asthma quality-of-life scores, and worse health-related quality-of-life scores and were more likely to complain of cough.

Thus, it appears that proximal GERD may worsen quality of life in asthmatic patients but does not worsen asthma control.

SARA RESULTS: IMPLICATIONS FOR MANAGEMENT

The SARA results suggest that patients with poorly controlled asthma who are on adequate controller medications should not be treated empirically for silent GERD in the expectation that the asthma will improve. Rather, they suggest that the focus should be on other factors that can worsen asthma control, such as the ability to properly use an inhaler, the ability to afford medications, compliance with drug treatment, and adequate control of other significant comorbidities such as allergic bronchopulmonary aspergillosis, sinusitis, allergic rhinitis, vocal cord dysfunction, and occult heart disease. The most recent NIH guidelines also suggest considering referral to an asthma specialist if symptoms persist despite adequate controller therapy.

Should patients with poorly controlled asthma be treated empirically for gastroesphageal reflux disease (GERD)?

Current guidelines¹ indicate that trying a proton pump inhibitor may be worthwhile. However, the results of a recent multicenter trial² indicate that this does not help control asthma symptoms and that we need to reevaluate the guidelines and focus on other factors that can worsen asthma control.

REFLUX DISEASE IS LINKED TO ASTHMA

GERD’s association with asthma has long been recognized. Asthma patients have a higher prevalence of GERD than the general population, with reported rates of 20% to 80%.^3–8

GERD may worsen asthma via several mechanisms. If stomach acid gets into the airway, it can induce bronchoconstriction, vagal reflexes, and chronic airway inflammation, all of which can increase airway reactivity.^9–16 Chronic reflux can also cause inflammation of the esophagus, which can exacerbate cough and possibly bronchospasm via neurogenic mechanisms.¹⁷

In turn, asthma may worsen GERD. Airway restriction can lead to hyperinflation and increased negative inspiratory pleural pressure, both of which may reduce the effectiveness of the lower esophageal sphincter.¹⁸ In addition, the beta-agonists and methylxanthines used to treat asthma may impair function of the lower esophageal sphincter and exacerbate reflux.^18–20

CURRENT GUIDELINES ARE BASED ON LIMITED INFORMATION

The symptoms of GERD and asthma are nonspecific and can be similar (chest tightness, chest discomfort), which can make it challenging for clinicians or patients to distinguish asthma from GERD.² Moreover, in asthma patients, GERD often presents without classic symptoms such as heartburn, and thus has been labeled “silent” GERD.

However, these studies all had significant limitations, such as small sample size. Also, the definitions of asthma and GERD differed from study to study. In some cases, the definition of GERD included self-reported GERD, which often fails to correlate with GERD documented with esophageal pH monitoring in asthma patients.¹ These limitations were highlighted in a Cochrane review,³⁰ which found that asthma patients with GERD showed no overall improvement in asthma after treatment of reflux. It concluded that small groups of patients may benefit, but that predicting who will respond is difficult.

Larger randomized controlled trials^28,29 attempted to address some of these limitations, with varying results.

Littner et al²⁹ gave lansoprazole (Prevacid) 30 mg twice daily or placebo to 207 patients with moderate to severe asthma and symptomatic GERD and saw no improvement in daily asthma symptoms, ie, asthma control in the active-treatment group. While these patients had an improvement in symptoms of severe reflux, their overall quality-of-life scores were similar to those of the placebo group. Of note, patients needing more than one type of drug for asthma control had a lower rate of asthma exacerbations.

Kiljander et al²⁸ gave esomeprazole (Nexium) 40 mg twice daily or placebo to 770 patients who had mild to moderate asthma and symptoms of nocturnal asthma with or without symptoms of GERD. The only benefit was a slight improvement in peak expiratory flow in those with symptoms of both GERD and nocturnal asthma, and this was most significant in patients taking long-acting beta-agonists. Other measures—eg, the forced expiratory volume in the first second (FEV₁), use of a beta-agonist, symptom scores, and nocturnal awakenings—did not improve.

In both of these studies,^28,29 patients reported symptoms of GERD, so they did not have silent GERD.

THE DESIGN OF SARA

In SARA, 412 patients age 18 and older with inadequately controlled asthma were randomized to receive esomeprazole 40 mg twice a day or placebo for 24 weeks. Inadequate control was defined as a score of 1.5 or higher on the Juniper Asthma Control Questionnaire³¹ despite treatment with inhaled corticosteroids. Patients had no symptoms of GERD. The 40-mg twice-daily dosage of esomeprazole was chosen because it is known to suppress more than 90% of acid reflux.^24,32

All patients completed a baseline asthma diary, recording peak expiratory flow rates, asthma symptoms, nighttime symptoms, and beta-agonist use. This information was collected every 4 weeks throughout the trial.

All participants also underwent esophageal pH monitoring for an objective confirmation of GERD. Patients were randomized independently of the results of the pH probe; in fact, investigators and patients were blinded to these results.

The primary outcome measure was the rate of episodes of poor asthma control, with poor control defined as any of the following:

• A decrease of 30% or more in the morning peak expiratory flow rate on 2 consecutive days, compared with the patient’s best rate during the run-in period
• An urgent visit, defined as an unscheduled health care visit, for asthma symptoms
• The need for a course of oral prednisone for treatment of asthma.

Asthma was defined as doctor-diagnosed, plus either a positive methacholine challenge test (a concentration of methacholine causing a 20% reduction in FEV₁ [PC₂₀] < 16 mg/mL) or a positive bronchodilator response (a 12% increase in FEV₁) to an inhaled beta-agonist. Participants had no other indication for acid suppression, including symptoms of GERD or previously diagnosed erosive esophageal or gastric disease.

Acid reflux was evaluated by ambulatory pH monitoring, which had to last at least 16 hours and span one meal and 2 hours in the recumbent position. Reflux was present if the pH was less than 4.0 for more than 5.8% of total time, 8.2% of time upright, or 3.5% of time lying down.³³ Episodes and severity were measured by the Gastroesophageal Reflux Disease Symptom Assessment Scale.³⁴

SARA RESULTS: NO IMPROVEMENT IN ASTHMA WITH GERD TREATMENT

The SARA treatment and control groups had similar baseline characteristics, with similar asthma symptoms. Most of the patients were women: 72% of the placebo group and 64% of the esomeprazole group. Most had baseline spirometric results at the lower end of normal (the mean FEV₁ was 76% ± 16 SD in the treatment group and 78% ± 15 in the placebo group) and had very poor asthma control, with an average Juniper Asthma Control Questionnaire score of 1.9 (> 1.5 is considered poor control).³¹ GERD was documented with esophageal pH monitoring in 40% of patients, showing that a significant number had silent GERD.

Episodes of poor asthma control occurred with similar frequency in the esomeprazole and placebo groups (2.5 vs 2.3 events per personyear, P = .66). Treatment made no difference in this end point regardless of the baseline results of pH monitoring. No treatment effect was noted in the individual components of the episodes of poor asthma control or in secondary outcomes, including pulmonary function, airway reactivity, asthma control, symptom scores, nocturnal awakening, or quality of life.

In addition, subgroup analysis failed to identify any group—including those with documented reflux on pH probe testing or those receiving a long-acting beta-agonist—who benefited from proton pump inhibitor therapy.

The investigators concluded that these data suggest treatment of silent GERD does not improve asthma control and thus that a reevaluation of current guidelines and clinical practice is warranted.²

ISSUES REMAIN

This large clinical trial, in which asthma and GERD were well defined and objectively measured, was robustly negative in terms of showing any benefit of treatment of silent GERD on asthma control. The study population was representative of those for whom such a treatment is recommended in the current NIH guidelines, which are based on data published prior to SARA.

However, while SARA was well designed and had clear results, it had some limitations, and some issues regarding GERD and asthma remain unanswered.

Is acid the only problem in GERD? SARA focused on acidic GERD. Aspiration of substances such as pancreatic enzymes, pepsin, and bile has also been shown to induce symptoms in asthma patients.^2,32,35 In addition, distention of the esophagus and stimulation of neurogenically mediated reflexes can cause symptoms or neurogenic airway inflammation that is not mitigated by drugs that target acid reflux.³²

Indirectly supporting this theory is evidence that surgical interventions such as fundoplication can improve asthma symptoms. ³⁶ However, this evidence is only from small studies with significant limitations.

Is proximal GERD worse than distal GERD? SARA did not address whether proximal and distal reflux may affect asthma differently. The importance of proximal reflux in asthma has not been clearly established, but there is evidence that patients with proximal GERD have a higher incidence of nocturnal cough than patients who have only distal reflux. ³⁷

Dimango et al³⁸ recently reported additional data from SARA in which patients with poorly controlled asthma underwent both proximal and distal esophageal pH monitoring to see if proximal GERD was associated with poor asthma control: 304 patients underwent dual pH-probe assessment and 38% of them had proximal reflux. The authors found no difference between those with and without proximal GERD with regard to nocturnal awakenings, need to use a rescue inhaler, inhaled controller medication dose, lung function, or airway reactivity by methacholine challenge. However, they did find that those with proximal GERD had worse asthma quality-of-life scores, and worse health-related quality-of-life scores and were more likely to complain of cough.

Thus, it appears that proximal GERD may worsen quality of life in asthmatic patients but does not worsen asthma control.

SARA RESULTS: IMPLICATIONS FOR MANAGEMENT

The SARA results suggest that patients with poorly controlled asthma who are on adequate controller medications should not be treated empirically for silent GERD in the expectation that the asthma will improve. Rather, they suggest that the focus should be on other factors that can worsen asthma control, such as the ability to properly use an inhaler, the ability to afford medications, compliance with drug treatment, and adequate control of other significant comorbidities such as allergic bronchopulmonary aspergillosis, sinusitis, allergic rhinitis, vocal cord dysfunction, and occult heart disease. The most recent NIH guidelines also suggest considering referral to an asthma specialist if symptoms persist despite adequate controller therapy.

We have previously discussed in the Journal the recrudescence of pertussis in adults and the challenges to its diagnosis, which include the misperception that it is a rare disease. ¹ Adult pertussis infection is usually due to the waning effect of childhood vaccination, and in adults it is more often an extreme annoyance than a life-threatening illness.

In this issue, Dr. Camille Sabella² discusses measles, an infection thought to be all but eradicated in the United States by vaccination, predominantly using the live-attenuated measles virus contained in the measles-mumps-rubella (MMR) vaccine.

But measles outbreaks are seemingly on the rise, and because measles is extremely contagious, it poses a real risk to closed communities such as college dormitories, churches, and health care facilities. Measles infection can have significant adverse outcomes, particularly in immunosuppressed patients.

Although outbreaks have been attributed to virus imported from locations outside the United States, the spread of infection has been blamed on an increased number of unvaccinated children and adults. The reasons for decreased vaccination rates are many, and include parental fears that the vaccine will cause problems such as autism.

This autism link is a goblin that refuses to go away, despite strongly worded debunking by the US Centers for Disease Control and Prevention, the Institute of Medicine,¹ and many peer-reviewed publications. The very recent retraction by the Lancet⁴—based on ethical and nondisclosure concerns—of the 1998 paper by Wakefield et al⁵ (which suggested a link in 12 children between MMR vaccine and chronic gastrointestinal problems and autism spectrum disorders) may further diffuse this concern. But I fear it will not.

So at present, we should reeducate ourselves on the clinical features, natural history, and potential complications of this eradicable disease, particularly if we treat patients who work in closed communities or have defects in cellular immunity.