	From Kaufman Center for Heart Failure, Department of Cardiovascular Medicine, Cleveland Clinic Foundation, Cleveland, Ohio

	Since its approval as a diagnostic test in the United States in 2001, plasma B-type natriuretic peptide (BNP) has been used extensively to diagnose heart failure in the clinical setting.  Studies have repeatedly demonstrated the unsurpassed success of plasma BNP and its N-terminal counterpart, NT-proBNP, to accurately diagnose congestive heart failure in patients presenting with acute dyspnea or edema.1 Meanwhile, the consistent demonstration of the prognostic value of plasma BNP and NT-proBNP measurements has led to the widespread belief that the rise in BNP or NT-proBNP levels represents the corresponding ventricular response to additional pressure or volume overload, thus directly corresponding to disease progression in heart failure.

	While it is true, for the most part, that plasma BNP levels do relate to clinical heart failure, there remains the fundamental question of why there are patients presenting with comparable clinical profiles of heart failure who display a wide range of plasma BNP values. Over the years, we have observed that there is a cohort of patients with chronic heart failure symptoms who have plasma BNP levels lower than what would be considered “diagnostic” for heart failure. We therefore conducted a retrospective analysis of the heart failure patients in our outpatient clinic to determine the prevalence and clinical characteristics of this subset.

	We performed a retrospective cross-sectional chart analysis. The patient cohort, who were seen as outpatients between November 2001 and February 2003 and who underwent plasma BNP testing, consisted of consecutive patients with a clinical diagnosis of chronic heart failure (longer than 3-month duration, on a stable medical regimen, and with left ventricular ejection fraction less than 50% by two-dimensional echocardiography). Excluded from the study were patients with a history of congenital heart disease, cardiac transplantation, primary valvular disease, or active ischemia requiring urgent revascularization. We performed a standardized chart re-view via our electronic medical records to extract clinical, demographic, and laboratory data at the time of plasma BNP testing. A plasma BNP level of less than 100 pg/mL was considered to be in the “normal” diagnostic range.

	Among 662 consecutive patients in the cohort, 104 were excluded. There were 109 asymptomatic patients (New York Heart Associa-

	tion [NYHA] functional class I) with plasma BNP levels ranging from 5 to 572 pg/mL (median 87 pg/mL). One hundred six patients of the remaining 449 symptomatic patients (NYHA functional class II—III) had plasma BNP levels less than 100 pg/mL (in the normal range). Patients who demonstrated normal plasma BNP levels were more likely to be younger and female, and to have a nonischemic pathogenesis, less depressed left ventricular ejection fraction, lower serum creatinine levels, and less atrial fibrillation (table). Age, cardiac and renal functions, and heart failure etiology were all independent predictors for normal plasma BNP levels.

	The main finding in our study was that a substantial proportion of stable, ambulatory patients with chronic systolic heart failure have plasma BNP levels within the range that is typically considered to be normal, despite the presence of chronic symptoms and underlying left ventricular systolic dysfunction. This finding, however, does not contradict the usually reliable diagnostic role of plasma BNP levels in the presence of acute dyspnea, or the consistently observed prognostic value of BNP.1

	It is reasonable to believe that patients with heart failure and relatively low plasma BNP levels represent a subgroup of patients who simply respond well to optimal therapy despite having persistent symptoms. What the data in our study point out, however, is the importance of interpreting plasma BNP levels in the appropriate clinical context. There is also the question of whether there is a subgroup of patients that may have symptomatic heart failure and relatively low plasma BNP levels. Indeed, we observed that patients with a nonischemic etiology tend to present with relatively lower plasma BNP levels.

	Because BNP is produced by cardiac myocytes (and, to some extent, surrounding fibroblasts), it is plausible that the underlying cause of myocyte dysfunction can in some ways affect the trigger, synthesis, release, or even clearance of BNP. Such a hypothesis should be tested in future mechanistic studies, as these processes may greatly affect the interpretation of plasma BNP levels in a clinical setting.

	There have also been recent reports of potential genetic heterogeneity in BNP production.2 We have always assumed, based on previous animal studies, that a direct correlation exists between ventricular loading and myocardial or plasma BNP production.3 While this load-dependent mechanism is likely, we should begin to explore other potential load-independent mechanisms that may regulate BNP release, or even different sites of plasma BNP production, in order to explain this somewhat unexpected observation.

	While there may be many underlying factors that determine the plasma BNP levels measured in our study cohort, our observations have identified what is true to all diagnostic tests—that results of the test can only be an adjunct to clinical judgment in making the diagnosis. A plasma BNP level less than 100 pg/mL or a plasma NT-proBNP level less than 125 pg/mL, therefore, may not necessarily exclude heart failure or left ventricular systolic dysfunction. While a low plasma BNP level may help to rule out heart failure as the underlying cause of acute symptoms, low plasma BNP levels in the case of chronic and stable symptomatic heart failure may not be as useful. Such variability also complicates use of a “cutoff” plasma BNP level to initiate or titrate medical therapies in order to lower plasma BNP levels.

	As we continue to enthusiastically embrace BNP and NT-proBNP levels as important new parameters in the diagnosis and management of heart failure, we must also strive to better understand how to interpret these values in the context of a general clinical setting as well as in specific patient populations. 

Articles

Cardiology Review Online

Plasma B-type natriuretic peptide levels in chronic heart failure

	B-type natriuretic peptide (BNP) is a 32-amino acid protein secreted primarily by the cardiac ventricles in response to increased intraventricular pressure.1 Physiologically, BNP functions as a counterregulatory hormone that opposes the effects of angiotensin II, norepinephrine, and endothelin, thus promoting vasodilation and natriuresis. BNP levels are increased in patients with acute or chronic heart failure, and the magnitude of BNP elevation correlates moderately with the left ventricular end diastolic pressure and symptom severity (eg, New York Heart Association class).2,3 In addition, BNP levels can be measured rapidly and relatively inexpensively (less than $50) using point-of-care devices approved by the Food and Drug Administration. As a result, BNP has achieved widespread acceptance as a valuable new tool to aid in the diagnosis of heart failure and, in particular, to distinguish dyspnea due to heart failure from that attributable to other causes (eg, lung disease). In the Breathing Not Properly Study, for example, BNP was found to be more accurate than all other clinical and laboratory findings in identifying heart failure among 744 patients presenting to the emergency department with dyspnea.4

	Despite the prevalent enthusiasm for BNP, both as a diagnostic tool and as an aid in managing heart failure patients, it must be recognized that it is not a perfect test. In the Breathing Not Properly Study, a BNP cutpoint of 100 pg/mL was associated with a 90% sensitivity, 76% specificity, and 83% overall

	predictive accuracy for diagnosing heart failure.4 This means that 10% of patients with acute heart failure had a normal BNP level, 24% of patients without acute heart failure had an elevated BNP level, and BNP was “wrong” in 1 of 6 patients (17%) presenting with worsening shortness of breath. It is thus evident that although BNP is a useful test, it must be interpreted in the context of other available data (ie, history, physical examination, chest x-ray, etc), and it should not be presumed that a normal BNP value excludes heart failure, nor that an elevated BNP value is “diagnostic” of heart failure.

	Tang and Francis (page 41) provide additional evidence indicating that BNP levels should not be considered the new “gold standard” for diagnosis of heart failure, particularly in the outpatient setting. In their study of 449 ambulatory patients with symptomatic (NYHA class II-III) systolic heart failure, 106 (23.6%) had “normal” BNP levels (< 100 pg/mL). Moreover, among 109 patients with asymptomatic left ventricular systolic dysfunction, well over half had normal BNP levels. The authors conclude that it is essential to consider the clinical context when interpreting the results of

	Although not discussed by Tang and Francis in their article, Bayes’ theorem provides some guidance on how to interpret the results of any less-than-perfect diagnostic test (of which BNP is a clear example).5

	According to Bayes’ theorem, the predictive value of any such test is largely dependent on the pretest likelihood of having the disease. Thus, an elevated BNP level in a patient with a high likelihood of having heart failure (eg, a patient with an ejection fraction of 20% and a chest x-ray showing pulmonary edema) is almost certainly a “true positive,” ie, the positive predictive value in such patients approaches 100%, whereas a “negative” test in the same patient should be regarded as a probable “false negative.” Similarly, in a patient with a low pretest probability of having heart failure (eg, a young asthmatic patient with active wheezing), a positive test is likely to represent a false positive, whereas a negative test is very likely to be a true negative. Note that a clinically important implication of Bayes’ theorem is that a diagnostic test such as BNP is most helpful in patients with an intermediate likelihood of having the disease (eg, patients with dyspnea of uncertain cause and nondiagnostic findings from other tests). It also follows that patients with classic pulmonary edema do not need a BNP to confirm a diagnosis of heart failure, nor do most patients with severe lung disease and a previously normal cardiac evaluation require a BNP to rule out heart failure.

	Given what is known about the physiology of BNP, how can we account for the fact that nearly one fourth of patients with symptomatic heart failure in the report by Tang and Francis had “normal” BNP levels? The authors propose several possible mechanisms. First, and perhaps most likely, patients with normal BNP levels may simply have been “well-compensated,” with essentially normal left ventricular diastolic pressures at rest. Such patients might still be symptomatic with activity, due to an increase in left ventricular pressure with exertion, but infrequent and transient elevations of left ventricular pressure may not be sufficient to increase the BNP level above the normal range. Alternatively, BNP levels may be dependent on multiple factors beyond left ventricular pressure. Indeed, the authors found that younger age, better cardiac and renal function, and nonischemic etiology of heart failure were predictive of normal BNP in their population. In addition, genetic factors could influence the synthesis, bioactivity, metabolism, and clearance of BNP.6

	An important limitation of the study by Tang and Francis is that the population was relatively young (mean age 59 years), predominantly male (71%), and restricted to patients with impaired left ventricular systolic function. In contrast, the median age of heart failure patients in the community is about 75 years, approximately half are women, and up to 50% have preserved left ventricular systolic function.7 Moreover, BNP levels are known to increase with age, they tend to be higher in women than in men (especially among the elderly),8 and although BNP levels are generally increased in patients with heart failure and preserved systolic function,9 the magnitude of elevation tends to be less than in patients with systolic heart failure.10 In light of these factors, the sensitivity and specificity of BNP for diagnosing heart failure decline with age,8 as a result of which both the false-positive and false-negative rates increase with age, thus mandating an even more cautious approach to the interpretation of BNP levels in elderly patients with suspected heart failure.

	In summary, the study by Tang and Francis confirms what most clinicians already know—there are very few silver bullets in the practice of medicine, and although our diagnostic tools continue to improve, there is still a pivotal role for good, sound clinical judgment.

BNP for the diagnosis of heart failure: A valuable tool, but no silver bullet 

	Despite several articles that enumerate the factors causing false positive elevations in troponin levels,1,2 this biochemical marker has been proven to be a reliable and sensitive indicator of myocardial damage.3,4 Utilizing this datum, the study by Landesberg and colleagues (page 33) confirms several of what should be intuitively obvious presumptions: first, that patients undergoing major vascular surgery have an increased incidence of coronary artery disease, and second, that exacerbation of this disease results in increased postoperative mortality.

	The authors studied 447 patients who underwent major vascular surgery (231 carotid endarterectomies, 77 aortic aneurysm repairs, and 192 lower extremity bypasses). Unfortunately, they did not stratify their patient population further based

	on the type of surgery performed. There is certainly a difference in cardiac risk in the three classes of surgical procedures cited. Carotid endarterectomies and infrainguinal bypasses create far fewer physiologic myocardial stresses than aortic procedures. Cross-clamping of the aorta creates a sudden increase in afterload resistance and may also have deleterious effects on renal function. After aortic revascularization is accomplished, the “unclamping” phenomenon also has several negative effects on the myocardium, including reperfusion into a dilat-

	ed lower extremity vascular bed (with resultant hypotension), sudden flushing of acidotic metabolites into the circulation from the previously ischemic limbs, and acute decreases in afterload resistance. All of these occurrences clearly contribute to the observed myocardial damage. It would be interesting to see whether this subset of patients is the most likely to develop untoward cardiac events. As the shift toward endovascular aortic aneurysm repair (as opposed to open repair) gains momentum, one would expect to see a significant reduction in cardiac risk in aortic surgery. In fact, recent publications have reported a reduction of cardiac complications in endovascular aortic surgery to nearly zero.5

	The real question the study by Landesberg and colleagues posits is whether or not this observation is clinically important. In other words, should the discovery of postoperative cardiac marker enzyme elevations change the management of these patients? If so, can it be done in a cost-effective manner?

	In my own practice (an academic-affiliated community practice), patients such as these would be discharged after the usual hospital stay with continued outpatient surveillance by their cardiologist. Further cardiologic evaluation is then done electively as needed. In the absence of additional evidence of myocardial infarction (diagnostic electrocardiogram changes or clinical symptoms), there appears to be no need for acute cardiac intervention, lengthened hospitalization, or change in therapy.

Should postoperative troponin elevation change patient management?

	A 72-year-old man was admitted to the hospital for elective carotid endarterectomy surgery due to an asymptomatic 95% internal carotid artery stenosis. His medical history was significant for multiple vascular operations; he had a carotid endarterectomy on the opposite side 2 years earlier and repeated lower-extremity arterial bypass procedures for intermittent claudication. He had insulin-dependent diabetes mellitus and ischemic heart disease status post, and a coronary artery bypass graft surgery 7 years earlier for left-main and triple-vessel coronary artery disease. He had grade IIa stable angina pectoris, essential hypertension, and mild renal failure, but no history of congestive heart failure.

	An electrocardiogram (ECG) revealed left ventricular hypertrophy, left anterior hemiblock, and right bundle branch block. The patient underwent carotid endarterectomy under general anesthesia. Just before the induction of anesthesia the patient was connected to the 12-lead ST-trend monitor, and a baseline, resting ECG was acquired. The operation itself was uneventful. At the end of the surgery, the patient was extubated, neurologically intact, and hemodynamically stable. In the recovery room, he reported pain in his neck, presumably from the sur-gery, and received repeated intravenous doses of morphine. Despite the absence of chest pain, myocardial ischemia was suspected and, although the patient was participating in a study and the monitoring was supposed to be blinded, the ST-trends were made available for the treating physician.

	The figure demonstrates the ST and heart rate trends of the patient before the induction of anesthesia and for the next 24 hours. The patient developed prolonged (190 minutes) ST-segment depression-type ischemia, which started at the end of surgery on emergence from anesthesia, in accordance with a moderate increase in heart rate. Although the ischemic ECG changes reverted to baseline at the end of the ischemic episode, 10 hours later, his serum troponin I increased to 10.2 ng/mL, and then to 32.1 ng/mL the next morning. The figure shows marked ST-segment depression in the anterior leads (367 microvolts) during peak ischemia compared with the patient’s baseline.

	This patient had prolonged silent ST-segment depression-type ischemia and subsequent myocardial infarction (MI). The MI could have gone completely unnoticed if we had not performed the continuous ECG monitoring and routine troponin I measurements used for the study protocol. The ECG changes that occurred during ischemia completely returned to baseline, despite the elevation of troponin I to 20 times the upper limit of normal. This case demonstrates how, unless one uses continuous ST-trend monitoring, diagnosing MI by single snapshots of 12-lead ECGs in patients with baseline ECG abnormalities can be difficult.

Case report: Value  of ECG monitoring and  troponin measurement 

	From the Departments of Anesthesiology, Critical Care Medicine, Vascular Surgery, Clinical Biochemistry, and Cardiology, Hadassah Medical Center, Jerusalem, Israel

	More than 4% of patients who are diagnosed with coronary artery disease (CAD) or have risk factors for CAD and undergo major noncardiac surgery have postoperative myocardial infarction (MI) and major cardiac complications.1,2 Perioperative MI is often difficult or even impossible to diagnose clinically based on only two of the classical triad of factors—cardiac symptoms, typical electrocardiographic (ECG) findings, and biochemical markers. The silent nature of perioperative infarction, the subtle and transient ST-segment depression ECG changes resulting in non—Q-wave infarction,3 and the claimed low specificity of CK-MB isoenzyme4 lead to inconsistencies in the diagnosis, making the long-term significance of perioperative markers of MI uncertain. Based on continuous 12-lead ECG monitoring and long-term survival, our study sought to determine whether elevated postoperative troponin has significant prognostic implications, and the plasma concentrations of cardiac troponin I and T and CK-MB that correlate with postoperative MI.

	The study included 447 consecutive patients who underwent 231 carotid endarterectomies, 77 abdominal aortic operations, and 193 lower-extremity bypass procedures at the Hadassah University Hospital, for a total of 501 major vascular surgical procedures. Before anesthesia was administered, patients were connected to a continuous 12-lead ECG monitor wired to a Cardiac Review station (ST-Guard, Marquette Electronics). All 12-lead ECG complexes were automatically stored every minute, and the absolute and relative ST-segment deviation at 60 milliseconds after the J point in all leads was measured compared with preoperative baselines.

	Patients were monitored for at least 48 hours and up to 72 hours. We measured cardiac troponin I or T, or both, and CK-MB in all patients immediately after surgery and every morning for the first three postoperative days. We examined three cutoff levels of cardiac troponin: (1) troponin I levels above 1.5 ng/mL and/or troponin T levels above 0.1 ng/mL, which were the receiver—operator characteristic curve medical decision cutoffs for MI defined by the manufacturers of these assays; (2) troponin I levels above 0.6 ng/mL and/or troponin T levels above 0.03 ng/mL, which corresponded to the lowest levels with less than 10% imprecision or coefficient variance (coefficient variation) for these assays5; and (3) troponin I levels above 3.1 ng/mL and/or troponin T levels above 0.2 ng/mL. The upper limit of normal for CK total was 170 IU, and the upper limit of normal for CK-MB/total CK was 10% in our lab. We also investigated the cutoff level of 5%.

	If troponin I levels above 1.5 ng/mL and/or troponin T levels above 0.1 ng/mL were associated with typical ischemic symptoms, ECG changes indicative of ischemia, or new pathological Q-waves, a diagnosis of clinical MI was made by the treating physicians independent of this study. The hospital’s information system, which is continuously updated by the Israeli Ministry of the Interior with all newly deceased individuals, was used to track long-term survival.

	Myocardial ischemia. Sixty-six patients had 104 ischemic episodes during the 25,622 patient-hours of continuous 12-lead ECG monitoring (51.4 ± 15.7 hours/patient). All episodes but one were denoted by ST-segment depression alone. The longest ischemic episode lasted over 30 minutes in 44 (9.8%) of the patients and over 60 minutes in 23 (5.1%) of the patients.

	Myocardial infarction. Depending on the biochemical criteria used, 14 (2.9%) to 107 (23.9%) patients had a postoperative MI, but only 17.7% to 61.9% had signs or symptoms of MI (table 1). None of the patients had new Q-waves. Each of the biochemical criteria for MI was associated with prolonged postoperative ischemia, and a higher incidence of prolonged postoperative ischemia was found among those who had higher troponin and CK-MB levels (table 1). Based on the combination of elevated cardiac troponin with ECG findings, prolonged chest pain, or both, 16 patients (3.6%) were diagnosed by the treating physicians as having postoperative MI, 12 (2.8%) of whom had symptoms related to MI.

	Long-term survival. There were 82 deaths (18.3%) during the 5-year follow-up period (mean, 32.3 ± 13.8 months). Results of the multivariate Cox survival analysis of the pre- and postoperative predictors of survival are shown in table 2. The only preoperative predictors of long-term survival, according to the multivariate Cox regression analysis, were age, type of vascular surgery (lower-extremity bypass surgery), previous MI, and renal insufficiency. By univariate analysis (data not shown), all cutoff levels of CK-MB, cardiac troponin, and ischemia duration predicted long-term survival (figure). When all postoperative biochemical markers were included with the preoperative predictors in the multivariate analysis, only the CK-MB levels above 10% and troponin I levels above 0.6 ng/mL and/or troponin T levels above 0.03 ng/mL were independent predictors of mortality (odds ratio [OR] = 4.21 and 1.96; P = .002 and .005, respectively). Using only the lowest cutoff levels of CK-MB and troponin, both CK-MB above 5% and troponin I above 0.6 ng/mL and/or troponin T above 0.03 ng/mL independently predicted mortality (OR = 2.14 and 1.89; P = .018 and .01, respectively).

	The results of our study showed that even minor elevations in cardiac troponin or CK-MB during the first 3 postoperative days predict increased risk of long-term mortality after major vascular surgery. Those with higher levels of biochemical markers, indicating a larger volume of myocardial injury or infarction, had worse survival. Levels of CK-MB above 10% and troponin I levels above 1.5 ng/mL and/or troponin T levels above 0.1 ng/mL predicted a 3.75- and 2.06-fold increase in long-term mortality, respectively, independent of the preoperative predictors, such as the patient’s age, type of vascular surgery, previous MI, and renal failure. Levels of CK-MB above 5% and troponin I levels above 0.6 ng/mL and/or troponin T levels above 0.03 ng/mL independently

	predicted a 2.15- and 1.89-fold increase in mortality, respectively. As found in a previous study, there was a strong correlation between prolonged postoperative ischemia and elevated troponin levels (table 1). Prolonged postoperative ischemia longer than 30 minutes’ duration was associated with a 2.6-fold increase in adjusted long-term mortality, and ischemic episodes longer than 60 minutes were associated with a 3.7-fold increase (table 2).

	Our study supports the findings of two previous studies that inves-tigated the effect of postoperative cardiac troponin measurements on

	6-month outcome following noncardiac surgery.6,7 Our in-depth analysis of the biochemical markers, including low-level troponin elevations as well as perioperative ischemia, and longer follow-up expand on the findings of these studies.

	Debate over which cutoff levels of troponin should be used to define a clinically important MI continues. The conventional cutoff values, those obtained by titration of troponin to a population of patients with a medical diagnosis of MI using receiver—

	operator characteristic curve analyses, were troponin I levels above 1.5 ng/mL and troponin T above 0.1 ng/mL for the assays used in our study. On the other hand, in their consensus document for redefinition of MI, the American College of Cardiology/European Society of Cardiology (ACC/ESC) task force held that even minor increases to levels above the 99th percentile of the normal population, in the setting of documented myocardial ischemia, should be considered as MI because cardiac troponins are so specific to myocardial necrosis.8 Most troponin assays, however, are not precise enough at this low range. Therefore, somewhat higher cutoff levels are suggested based on less than 10% imprecision or coefficient variance. Troponin I and T levels above 0.6 and 0.03 ng/mL, respectively, were used for the assays in our study.

	Creatine kinase-MB predicted long-term mortality independent of cardiac troponin. Patients with both elevated CK-MB and troponin, however, had a 4.19-fold increase in mortality compared with a 2.0- to 2.2-fold increase in mortality among those with only one of these markers.

	Clinically evident infarction. Only 25 patients (5.6%) met the conventional World Health Organization (WHO) clinical definition of MI. The WHO definition requires at least two of the following criteria: prolonged chest pain, elevated CK-MB or cardiac troponin, and ischemic ECG changes. Even in these 25 patients, the diagnosis of MI could not be made without the routine postoperative biochemical markers and continuous ECG monitoring because only 16 (3.6%) had prolonged chest pain or congestive heart failure suspicious of overt MI. Our data support the findings of Kim and colleagues,7 who reported that 12% of the patients had elevated troponin I on routine postoperative surveillance, but only 3% had clinical infarction based on the WHO definition. Routine monitoring of cardiac troponin, CK-MB, and silent ischemia in the first postoperative day, therefore, may help predict the short- and long-term risks of those who undergo major vascular surgery.

	Our study findings are also important with regard to the pathophysiology of perioperative MI. Recent studies have stressed that even minor elevations of cardiac troponin have prognostic importance in patients with unstable coronary syndromes. Those undergoing elective major vascular surgery represent a different type of cardiac patient than those with unstable coronary syndromes. The former often have long-standing, yet stable CAD. Their postoperative cardiac events are usually the result of stress-induced, ST-segment depression-type ischemia leading to infarction.4

	Our findings showed that cardiac troponin, CK-MB, and prolonged postoperative ischemia predict long-term mortality after major vascular surgery. Elevated troponin and CK-MB levels independently predict long-term survival, and when combined, are associated with an even worse prognosis. Further studies are needed to determine how cost-effective routine measurement of postoperative cardiac markers is, and whether more aggressive diagnostic and treatment strategies will improve survival in patients with postoperative MI. 

Perioperative markers  of MI and survival after vascular surgery

	This month Cardiology Review opens with a continuing medical education feature in the hypertension section by Dr. Thomas Zeller and associates from Germany. They studied 215 consecutive patients with ostial renal artery stenosis, a cause of poorly controlled hypertension. These patients were treated with percutaneous stenting of the lesions with encouraging results, as Dr. Naveed N. Masani and colleagues emphasize in their commentary.

	Our second continuing medical education feature is a combined lipid and arrhythmia section report on yet another use for those wonder drugs, the statins. Can statins really protect against atrial fibrillation, as Drs. Yinong Young-Xu and Shmuel Ravid suggest in their study of 449 patients? Is the mechanism related to reduced serum cholesterol levels or an anti-inflammatory effect? Dr. Ernst A. Raeder ponders both of these questions in his commentary.

	In a combined article in the diabetes and stroke sections, Drs. John K. Ninomiya and Michael H. Criqui report on their study of the relationship between the metabolic syndrome (a major component of which is insulin resistance) and certain cardiovascular events, including stroke. Using a large database of 15,922 subjects (from the National Health and Nutrition Examination Survey [NHANES] III), the authors found that the prevalence of stroke was 2.9% and myocardial infarction and/or stroke 5.2% when weighted to the general US population. Their study confirms the clinical importance of the metabolic syndrome as a significant risk factor for cardiovascular disease, a conclusion with which our commentator, Dr. RoseMarie Pasmantier, concurs. She also emphasizes the need for developing strategies to educate physicians about controlling this syndrome.

	In the CAD/angina section, Dr. Giora Landesberg and colleagues from Israel studied 447 consecutive vascular surgery patients and found between 14 and 107 had some degree of perioperative myocardial damage. Cardiology Review asked its vascular surgery consultant, Dr. Louis Kozloff, to comment on this study, and he adds some useful caveats in his remarks.

	In our final article this month in the heart failure section, Drs. W.H. Wilson Tang and Gary S. Francis conducted a retrospective evaluation of 662 consecutive heart failure patients in order to determine how many had nondiagnostic (low) B-natriuretic peptide levels. Their provocative findings are commented on by Dr. Michael W. Rich.

The importance of the metabolic syndrome as a risk factor for CV disease

	Diabetes triples the risk of developing heart failure (HF) in postmenopausal women with heart disease, and kidney disease or poorly controlled diabetes confers a 6 to 10 times increased risk of HF compared with women without diabetes. These findings were gleaned from a new analysis of the Heart and Estrogen/progestin Replacement Study (HERS), which appears in Circulation (2004;110[11]:1424-1430).

	“We have identified several risk factors that could be modified and have the potential to reduce the burden of heart failure in women with established coronary disease,” write the study’s authors, led by Kirsten Bibbins-Domingo, PhD, MD, instructor in medicine, epidemiology, and biostatistics, University of California, San Francisco.

	The analysis included 2,391 postmenopausal women with known coronary disease but no HF at baseline. The mean follow-up time was 6.3 years, during which time 237 women were hospitalized or died of HF. After adjusting for age and medication use, nine factors were independently associated with the development of HF.

	The strongest risk factor for development of HF was diabetes at study entry, which conferred a 3.1 times greater risk, followed by atrial fibrillation, which was associated with a hazard ratio of 2.9. The other significant risk factors were more than one previous myocardial infarction, renal insufficiency (defined as a creatinine clearance < 40 mL/

	min), systolic blood pressure greater than 140 mm Hg, current smoking, body mass index (BMI) greater than 35 kg/m2, left bundle branch block, and left ventricular hypertrophy.

	Uncontrolled hypertension was present in about one third of the women in the study. Even systolic blood pressures in the range of 120 to 140 mm Hg were associated with a significantly increased risk of HF. The authors emphasize that blood pressure control “should be an integral part of heart failure prevention,” noting that antihypertensive therapy has been shown in clinical trials to reduce HF risk.

	“We found higher fasting blood glucose levels to predict a greater adjusted risk of heart failure among diabetic women with coronary disease,” state the authors. Compared with diabetic women with controlled fasting blood sugar (80 to 150 mg/dL), those with fasting blood glucose higher than 300 mg/dL had a threefold adjusted risk of incident HF. Clinical trials are needed to determine whether or not treatment of long-standing hyperglycemia will decrease the risk of developing HF, the authors add.

	The risk of developing HF increased steadily with the number of risk factors, such that women with no risk factors had an annual incidence of 0.4%, whereas women with three or more risk factors but no diabetes had a 3.4% annual incidence of HF. Diabetic women with three or more additional risk factors had a yearly HF incidence of 8.2%.

	Among the diabetic women, concomitant renal insufficiency was associated with the highest annual incidence of HF (12.8%). Diabetic women with morbid obesity (BMI > 36 kg/m2) had an annual incidence of 7.0%.

	Halting the progression of renal disease through the use of angiotensin-converting enzyme inhibitors and angiotensin receptor blockers may be important for modifying HF risk in women with diabetes and coronary disease, the authors state.

	According to Dr. Bibbins-Domingo, the take-home message for physicians is the need to recognize that

	(1) diabetes is a powerful risk factor for HF in women, independent of whether or not they have had a heart attack; (2) diabetic women can have an annual risk of HF as high as 7% to 13% per year if they also have obesity, renal insufficiency, or if their diabetes is poorly controlled; and (3)

	the focus should be on prevention, including glycemic control, control of blood pressure, weight control, and smoking cessation.

Risk factors for HF identified in older women with coronary disease

	Blood pressure—related cognitive decline occurs in younger as well as older persons, according to recent research published in Hypertension (2004;44:1-6). The results indicate that pathophysiologic processes that adversely affect the brain “may begin earlier in the adult lifespan than previously thought,” according to the researchers.

	They studied 529 participants from the Maine-Syracuse Longitudinal Study of Hypertension, each of whom was given the Wechsler Adult Intelligence Scale and a hypertension diagnostic examination at baseline. The subjects were followed over 20 years, during which time the relationship between blood pressure and cognitive decline was examined for two age groups (18 to 46 years and 47 to 83 years).

	Baseline blood pressure was unrelated to crystallized/verbal, memory, and speed performance measures of the intelligence scale. In both the younger and older age groups, however, systolic, diastolic, and mean arterial pressure at enrollment and the blood pressure categories created by the Seventh Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure (JNC 7) predicted the decline in visualization/fluid abilities. Visualization/

	fluid abilities were measured using picture completion, picture arrangement, block design, and object assembly tests.

	The levels of decline with higher levels of blood pressure were not such that they interfered with everyday functioning; “in fact, one can decline from a very high level of ability to a lower level ability and still perform above the average,” said the study’s lead author, Penelope K. Elias, PhD, from the Department of Psychology, University of Maine, Orono. She notes that moving from a “normal” blood pressure classification to “stage I hypertension” would result in an estimated 8.12% decrement in correct responses on the visual/fluid component of the intelligence scale over 20 years.

	“Younger adults were generally less affected by hypertension than older adults, but the higher the blood pressure, the greater the change in cognitive performance over time for both young adults and older adults,” she said. “Thus, youth is not a protection against very subtle but progressive decline in cognition in relation to blood pressure level.”

	She continued, “Our work is also important because it is now believed that some individuals continue beyond mild decline in cognitive performance to catastrophic change, such as dementia. Thus, prevention of high blood pressure is important at all ages.”

	At the population level, lowering blood pressure by 20 mm Hg systolic or 10 mm Hg diastolic, or from “hypertensive” to “normal” classification under the JNC definition, would have a “considerable beneficial effect on the preservation of cognitive abilities in the population as a whole,” the authors write.

	Whether or not cognitive changes can be reversed with antihypertensive drug therapy should be studied in clinical trials, they believe. “Our feeling is that preventing the rise in blood pressure over time is the critical aspect of protecting cognitive functioning from the ravages of sustained untreated high blood pressure,” said Dr. Elias.

Cognitive decline observed with higher blood pressure at all ages

	WASHINGTON, DC—Drug-eluting stents continue to outperform bare metal stents in patients treated for coronary stenosis, and now drug-eluting stents have been shown to reduce the risk of restenosis compared with their bare metal counterparts in a trial conducted exclusively in patients with diabetes. These findings were among those released during the late-breaking clinical trial sessions at the TCT 2004 conference.

	In a trial known as TAXUS IV, 1,314 patients undergoing elective stenting of single de novo coronary stenoses were randomly assigned predilatation to receive a bare metal stent (control group) or a paclitaxel-eluting stent (TAXUS). All patients received antiplatelet therapy (clopidogrel) for 6 months after the stenting procedure. The 2-year follow-up results were reported by Gregg W. Stone, MD, vice chairman and director of cardiovascular research and education, The Cardiovascular Research Foundation, Lenox Hill Heart and Vascular Institute, New York City.

	The rate of major adverse cardiac events (MACE) was 14.7% in the group assigned to the drug-eluting stent compared with 24.9% in those assigned to the bare metal stent. Late events, those that occurred between 1 and 2 years, occurred in 5.1% of patients receiving a drug-eluting stent and in 7.3% of the controls (P = .09). The rates of target lesion revascularization, target vessel revascularization, and target vessel failure between year 1 and year 2 were also lower with the drug-eluting stents, although the difference reached statistical significance only for target lesion revascularization (1.6% versus 3.6%; P = .03).

	“At 2 years’ follow-up, the incremental benefit of having received the TAXUS stent rather than a bare metal stent continues to increase, without evidence of late catch-up,” said Dr. Stone. The clinical advantage with the drug-eluting stent was realized across all patient subgroups, including those with small vessels, long lesions, or diabetes, he added.

	A sirolimus-eluting stent was associated with a significant reduction in in-segment late lumen loss and restenosis compared with a bare metal stent in patients with diabetes who had de novo coronary stenoses, according to Manuel Sabatè, MD, of the Hospital Clinico San Carlos, Madrid. In the Diabetes and Sirolimus-eluting Stent trial, 160 patients with 221 stenoses were randomly assigned to a sirolimus-eluting stent or a bare metal stent. To be eligible, patients had to have type 2 diabetes or type 1 diabetes with evidence of ischemia. Thirteen percent of treated lesions were total occlusions.

	At 9 months, in-segment late lumen loss was 0.08 mm in the patients assigned to a sirolimus-eluting stent compared with 0.44 mm—an 82% reduction (P < .001). This difference in in-segment late lumen loss resulted in a 76% reduction in in-segment restenosis, from 33% to 7% (P < .001) and an 84% reduction in in-stent restenosis, from 31% to 4.9% (P < .001), in patients receiving the drug-eluting stent.

	The 9-month rate of MACE was 11.3% in the drug-eluting stent arm and 36.3% in the bare metal stent arm (P < .001). The rate of target lesion revascularization was also significantly less in patients receiving the sirolimus-eluting stent (7.5% versus 31.3%; P < .001).

	In a third presentation, percutaneous coronary intervention (PCI) using a sirolimus-eluting stent was associated with a lower rate of MACE when compared with historical MACE rates of patients undergoing coronary artery bypass graft (CABG) surgery or PCI using a bare metal stent, said Patrick Serruys, MD, PhD, chief of interventional cardiology at Thoraxcenter-Erasmus University, Rotterdam, the Netherlands.

	In the second Arterial Revascularization Therapies (ARTS II) study, 607 patients with advanced coronary disease underwent PCI with a sirolimus-eluting stent. The MACE rate in ARTS II was compared with the MACE rate in the CABG and bare metal stent arms

	in the ARTS I trial, which included 600 patients.

	The rate of MACE at 6 months was 6.4% in the patients in whom the drug-eluting stent was implanted in ARTS II, compared with 9.0% in the CABG arm in ARTS I and 20.0% in the bare metal stent arm in ARTS I. The lower MACE rate was observed in the drug-eluting stent arm in ARTS II, even though ARTS II included patients with more complex stenoses and more patients who underwent triple-vessel stenting than in ARTS I, said Dr. Serruys.

Drug-eluting stents show long-term clinical value

	Atrial fibrillation is the most common arrhythmia afflicting elderly patients, particularly those with rheumatic or hypertensive heart disease, congestive heart failure, or coronary artery disease (CAD).* While there is persistent debate about the best management strategies for atrial fibrillation (eg, cardioversion versus rate control and chronic anticoagulation), there is also interest in preventive measures. In this issue of Cardiology Review, Young-Xu and Ravid (page 20) review an observational study they conducted in 449 patients with CAD. They hypothesized that because of the efficacy of HMG-CoA reductase inhibitors (statins) in primary and secondary prevention of CAD and because of shared risk factors for CAD and atrial fibrillation, statin therapy might reduce the incidence of atrial fibrillation.

	The key finding of the study was a reduction in the incidence of atrial fibrillation by more than half in statin users compared with nonusers (18% versus 37% per 1,000 person-years). There appears to be a modest correlation between serum cholesterol reduction and the relative risk of atrial fibrillation (see table 2, page 23), which does not hold up in a multivariate model, however. Analysis of total exposure to statins versus risk of atrial fibrillation suggests a weak dose-response relationship, although the confidence intervals for the four patient groups are of borderline significance. The mechanisms through which statins could affect the incidence of atrial fibrillation are unclear. The authors interpret their data as suggesting that it is independent of the cholesterol-lowering ability of statins. Whether anti-inflammatory effects, modulation of autonomic activity, or other as yet unidentified effects play a role remains to be determined. There are potential concerns in an observational study such as this, which the authors strive to identify and exclude. For example, the statin users were significantly younger and more likely to be taking beta blocking agents than nonusers, potentially influencing the outcome. In addition, it can be difficult to diagnose transient, asymptomatic atrial fibrillation. Because of these considerations, as well as the potential impact of statin therapy on risk of atrial fibrillation, a future prospective trial would be warranted. This could corroborate the findings of the present study and provide data on morbidity and mortality. Furthermore, a formal trial could be designed to elucidate the potential mechanism(s) whereby statins lower the risk of atrial fibrillation.

Does statin therapy reduce the  incidence of atrial fibrillation? 

	From the Lown Cardiovascular Center and Research Foundation, Brookline, Massachusetts; Departments of Epidemiology and Nutrition, Harvard School of Public Health; Department of Medicine, Brigham & Women’s Hospital, and Harvard Medical School, Boston, Massachusetts

	Atrial fibrillation is the most common sustained cardiac arrhythmia, affecting approximately 2.2 million Americans1 and responsible for up to 30% of all ischemic strokes.2 It has a higher prevalence among elderly, hypertensive, and diabetic patients and among men, all of whom also have a higher prevalence of coronary artery disease (CAD). Clinical trials have shown that HMG-CoA reductase inhibitors (statins) are effective for the primary and secondary prevention of CAD.3,4 We hypothesized that statin therapy might have a beneficial effect on the prevention of atrial fibrillation in patients with CAD because of shared risk factors.

	We studied patients with stable chronic CAD, without a history of atrial fibrillation, enrolled in an ongoing observational study. The exposure was recorded as time spent on statin therapy of any brand during follow-up. The outcome was the occurrence of atrial fibrillation during follow-up, verified through a review of records and confirmed by 12-lead electrocardiogram (ECG) or ambulatory ECG monitoring.

	We divided the study population into two mutually exclusive groups according to statin use: users and nonusers (of any cholesterol-lowering drugs). We compared the effects of statin use with no cholesterol therapy at all, as well as nonstatin cholesterol-lowering therapy, on the development of atrial fibrillation.

	We calculated the relative risks of developing atrial fibrillation and their accompanying 95% confidence intervals (CIs), which were estimated from a Cox proportional hazards regression model. Fisher’s exact test was used to compare proportions, and the Student t test was performed to compare means. All statistical tests were two-sided, and all analyses were carried out using Stata 7.0 for Windows (Stata Corporation, College Station, Texas).

	A total of 606 patients with at least 1 year’s worth of follow-up data, complete medication information, and information on history of atrial fibrillation were eligible for this study. Thirty-three patients with previous atrial fi-brillation were excluded, and 124 patients (20%) who had used nonstatin cholesterol-lowering drugs were excluded from the main analysis. The final study sample consisted of 449 men and women with stable CAD. Among these, 186 patients (41%) never used any cholesterol-lowering drugs during the study, and 263 patients (59%) used statins as their only cholesterol-lowering drugs. Average follow-up time was 5 years (range, 1—9 years). Completion rate for follow-up forms was 100%. Of 449 subjects, 18 (4%) voluntarily withdrew from the study during follow-up.

	Statin users were younger, but included a larger proportion of patients who were current or past smokers compared with nonusers. They also had a larger proportion of patients who were treated with beta blocking agents, but had a smaller proportion of patients who were taking aspirin (table 1).

	Among statin users, 9% developed atrial fibrillation. In contrast, 15% of nonusers developed atrial fibrillation. Corresponding incidence rates of atrial fibrillation per 1,000 person-years were 18 (95% CI, 11—26) for statin users and 37 (95% CI, 26–54) for nonusers (figure). The relative risk of developing atrial fibrillation for statin users compared with nonusers was 0.48 (95% CI, 0.28–0.83; table 2). The risk of developing atrial fibrillation was stratified for several known risk factors of atrial fi-brillation: advanced age, male sex, and hypertension (table 2). For all risk factors examined, statin use was consistently associated with a significant reduction in the risk of atrial fibrillation (figure and table 2).

	We investigated whether the association between statin use and atrial fibrillation depended on baseline serum cholesterol levels and on changes in these cholesterol levels during follow-up. Within each of the groups of total cholesterol and each of the groups of change in total cholesterol, relative risks consistently indicated a reduced likelihood of developing atrial fibrillation (table 2). In a multivariate Cox proportional hazards model, we entered baseline cholesterol level and percentage of change in cholesterol level separately and then together, and found that inclusion of these variables did not alter the association between statin use and atrial fibrillation.

	In addition, we calculated the total number of years a patient took statins during the study. We then divided patients into four groups according to the length of statin use (table 2). The relative risk of atrial fibrillation decreased as the total length of statin use increased (P = .03; table 2).

	Of the 124 patients who had used nonstatin lipid-lowering agents (gemfibrozil, cholestyramine, clofibrate, colestipol, and probucol) and who were excluded from the main analysis, 17% developed atrial fibrillation during follow-up compared with 15% of nonusers of any lipid-lowering drugs and 9% of statin users. The incidence rate of atrial fibrillation was not statistically different from the incidence rate of atrial fibrillation among nonusers (34 versus 37 per 1,000 person-years, respectively; relative risk = 0.92; 95% CI, 0.56—1.50).

	We examined the association between statin use and incidence of atrial fibrillation, adjusting for potential confounders, including age, systolic blood pressure, left ventricular ejection fraction (by echocardiogram), alcohol use (any versus none), history of heart failure, history of myocardial infarction, and total serum cholesterol level. The adjusted relative risk of atrial fibrillation according to statin use was 0.37 (95% CI, 0.18—

	0.76) after controlling for these potential confounders. Additional analyses were performed to test the impact of other potential covariates by including them in the main multivariate model one at a time. These included sex; height; weight; smoking (past or current); self-assessed health status; diastolic blood pressure; history of hypertension; diabetes; left ventricular dysfunction; left ventricular hypertrophy; serum creatinine level; potassium level; use of angiotensin-converting enzyme inhibitors, beta blocking agents, calcium channel blocking agents, diuretics, antiarrhythmic drugs, warfarin, or aspirin; changes in total serum cholesterol level; and high-density lipoprotein cholesterol level. Each time, after one of these variables was entered into our main regression model, the relative risk between statin use and atrial fibrillation changed less than 5% and remained significant.

	These findings support our hypothesis that statin use is protective against atrial fibrillation in a cohort of patients with chronic CAD. We used a multivariate Cox proportional hazards model that controlled for potential confounding factors, and statin use remained associated with a significant reduction in the onset of atrial fibrillation. The effect of statin therapy on the development of atrial fibrillation was independent of the changes in serum cholesterol levels over time and baseline cholesterol level.

	The findings do not appear to be spurious because different definitions of exposure to statins were used, and a dose-response relationship between length of statin use and reduction of atrial fibrillation was found. Confounding is unlikely, as the associations persisted after adjustment for many known risk factors for atrial fibrillation, including age, sex, clinical characteristics, concomitant medications, and relevant cardiac diagnoses.

	The mechanisms involved in the effects of statins on occurrence of atrial fibrillation are unclear. These effects might be related to the fact that mevalonic acid is a precursor not only of cholesterol synthesis but of other signaling intermediates, which are considered important for the cellular localization and intracellular trafficking of several membrane-bound proteins.5 One possible explanation for this effect might be related to the major role that the autonomic nervous system plays in the pathophysiology of atrial fibrillation.6,7 Recent reports suggest that statins can normalize autonomic tone and balance between the sympathetic and parasympathetic limbs of the autonomic nervous system, which in turn might decrease vagal tone and alter the electrophysiologic substrate, resulting in prevention of atrial fibrillation.8,9

	Another plausible explanation for the beneficial effects of statin use might be related to their anti-inflammatory effects. A recent report from the Cardiovascular Health Study indicated that C-reactive protein is increased in patients 65 years and older with atrial fibrillation on one hand, and increased C-reactive protein is strongly associated with future development of atrial fibrillation in patients without previous atrial fibrillation on the other hand.10 Another recent report suggested that increased C-reactive protein is also associated with the development of hypertension, a major risk factor for atrial fibrillation.11 C-reactive protein has also been shown to be a potent determinant of successful cardioversion of paroxysmal atrial fibrillation to sinus rhythm.12 All of these findings suggest that inflammation could play an important role in developing atrial fibrillation and that the preventive benefit of statins might be related to their anti-inflammatory effects and suppression of

	In a cohort of outpatients with stable CAD without prior atrial fi-brillation, we found that statin use was associated with a markedly reduced incidence of atrial fibrillation (50% to 60% reduction) over an average of 5 years of follow-up. This association, which appeared to be dose-dependent, remained significant after adjusting for potential confounders and was independent of serum total cholesterol level. The mechanisms behind this previously unreported effect of statins are not well defined and warrant further investigation. 

Statin drugs protect against atrial fibrillation

	The metabolic syndrome is a cluster of risk factors associated with cardiovascular disease. Presently, there is an epidemic of this syndrome and obesity in the United States and worldwide. While there are various definitions of the syndrome, the National Cholesterol Education Program Adult Treatment Panel III has created a practical definition that can be easily implemented in clinical practice. It requires three of the following criteria: central obesity, hypertriglyceridemia, low high-density lipoprotein cholesterol, high blood pressure, and high fasting glucose.

	Data from the Third National Health and Nutrition Examination Survey (NHANES III), conducted in the United States from 1988 to 1994, demonstrated that the metabolic syndrome was present in almost 23% of men and women.1 The prevalence varied with older age and ethnic status. It was present in 4.6%, 22.4%, and 59.6% of normal weight, overweight, and obese men, respectively, with similar values in women.

	The Diabetes Epidemiology: Collaborative Analysis of Diagnostic Criteria in Europe (DECODE) study group found a 15% overall prevalence of the metabolic syndrome in nondiabetic European men and women.2 Its presence was associated with a 1.4-fold increase in all-cause mortality for both men and women, with a 2.3-fold increase in the risk of cardiovascular disease mortality for men and a 2.8-fold increase for women.

	The Botnia study evaluated the prevalence of cardiovascular morbidity and mortality associated with the metabolic syndrome in a high-risk insulin-resistant and/or diabetic Scandinavian population.3 In men and women, respectively, the metabolic syndrome was found in 15% and 10% of subjects with normal glucose tolerance, 64% and 42% of those with impaired fasting glucose or impaired glucose tolerance, and 84% and 78% of those with type 2 diabetes. The risk for coronary heart disease and stroke was increased threefold in subjects with the metabolic syndrome. Cardiovascular mortality was 12% over a median follow-up of 6.9 years compared with 2.2% in those without metabolic syndrome.

	Ninomiya and Criqui (page 27) report on the association of the metabolic syndrome with myocardial infarction (MI) and stroke in a nondiabetic population using the NHANES III database. After adjusting for age, sex, ethnicity, and smoking status, patients with metabolic syndrome had a twofold increase in the risk of MI or stroke. These data support the concept that the metabolic syndrome defines a cohort that is at increased risk for vascular disease morbidity and mortality.

	Physicians and patients must be educated about the metabolic syndrome so that high-risk individuals can be identified. An integrated approach to management can then be undertaken. Implementing preventive lifestyle interventions, such as diet modification, increased physical activity, weight management, and smoking cessation, is imper-ative. Some subjects may require pharmacologic intervention.

Metabolic syndrome:  Controlling the epidemic

	From the Department of Family and Preventive Medicine, University of California, San Diego

	The metabolic syndrome is a cluster of physiologic and metabolic risk factors linked to cardiovascular disease. Over the years, a number of different sets of criteria defining the metabolic syndrome have been proposed; the definition contained in the National Cholesterol Education Program Adult Treatment Panel III (ATP III) recommendations is the most recent and widely publicized of these.1 In the present study, we examine the association of the metabolic syndrome, as identified based on the ATP III criteria, with myocardial infarction (MI) and stroke in a representative sample of American adults.

	The data source for our analysis was the National Health and Nutrition Examination Survey III (NHANES III), a survey conducted by the National Center for Health Statistics of the Centers for Disease Control and Prevention from 1988 to 1994. NHANES III is notable for its large size, its extensive anthropometric, physiologic, and laboratory testing (including physician examinations), and its complex, multistage sampling design, which allows results to be extrapolated to the total US population. Details of the data collection protocol have been described elsewhere.2 Our study was based on a subset of 15,922 NHANES subjects aged 20 to 89 years for whom the relevant data were available. NHANES III is a cross-sectional study. The events of interest, MI and stroke, were identified based on subjects’ self-reports of history of physician-diagnosed disease.

	Using ATP III criteria, the metabolic syndrome was identified based on the presence of at least three of the following five component conditions: insulin resistance, hypertension, hypertriglyceridemia, low high-density lipoprotein (HDL) cholesterol, and abdominal obesity, defined by high waist circumference.1 Insulin resistance was identified as fasting glucose of 110 mg/dL or above; for the purposes of this analysis, insulin resistance could also be identified by a self-report of current use of insulin or oral hypoglycemic agents. Hypertension was defined as systolic blood pressure of 130 mm Hg or above or diastolic blood pressure of 85 mm Hg or above; for the purposes of this analysis, the hypertension could also be identified by a self-report of current use of antihypertension medication. Hypertriglyceridemia was identified based on triglyceride levels of 150 mg/dL or above. Low HDL cholesterol was identified as below 40 mg/dL in men or below 50 mg/dL in women. High waist circumference was defined as greater than 102 cm in men or greater than 88 cm in women.

	Logistic regression analysis was used to determine the association

	of the metabolic syndrome and its component conditions with MI, stroke, and combined MI and/or stroke (MI/stroke). We controlled for age, sex, race, and smoking status as possible confounders. Analyses were performed using the SUDAAN software, which takes into account the complex sampling design of NHANES III and allows weighting of results and computation of confidence intervals (CIs) and statistical tests appropriate to the total US population.3

	Of 15,922 subjects, 752 had a history of MI, 464 had a history of stroke, and 1,098 had a history of MI/stroke, with unadjusted prevalences of 4.7%, 2.9%, and 6.9%, respectively. When weighted to the general US population, the prevalences were 3.7%, 2.0%, and 5.2% for MI, stroke, and MI/stroke, respectively.

	Table 1 shows the results of logistic regression models estimating the association of MI and stroke with the full metabolic syndrome and its component conditions, adjusted for age, sex, race, and smoking status. Two models—one for the full metabolic syndrome and another for the five component conditions&mdash;are shown for each disease history. Table 2 shows sex-specific results limited to MI/stroke for power reasons. Odds ratios (ORs) and CIs for the covariates of age, sex, race, and smoking are shown for the full syndrome model; those from the individual component conditions models (not shown) were very similar in direction and magnitude. Results are presented graphically in the figure.

	Based on ATP III criteria, the metabolic syndrome was significantly associated with MI/stroke (OR, 2.05; 95% CI, 1.64—2.57), MI (OR, 2.01; 95% CI, 1.53–2.64), and stroke (OR, 2.16; 95% CI, 1.48–3.16; table 1). It was significantly associated with MI/stroke in both men (OR, 1.93; 95% CI, 1.34–2.78) and women (OR, 2.20; 95% CI, 1.56–3.11; table 2). This difference in ORs between men and women was not statistically significant based on interaction models.

	Among the individual metabolic syndrome components, low HDL cholesterol was significantly associated with MI and MI/stroke, but not with stroke alone. Low HDL cholesterol was significantly associated with MI/stroke in men, but not in women. Insulin resistance was significantly associated with MI/stroke (OR, 1.30; 95% CI, 1.03—1.66), but not with MI or stroke separately. Hypertension was of borderline significance for MI/stroke (OR, 1.44; 95% CI, 1.00–2.08; P = .051); OR for MI and stroke separately were similar in magnitude and of borderline significance (P  < .10). MI/stroke was significantly related to hyperten-sion in women (OR, 2.19; 95% CI, 1.15–4.14), but not in men. Triglycerides had the largest OR of any of the component conditions for MI, stroke, and MI/stroke (OR, 1.66; 95% CI, 1.20–2.30 for MI/stroke). This association was stronger in women; the association in men was only of borderline significance. Abdominal obesity (high waist circumference) was the only component that was not independently related to prevalent disease in any model.

	Among the covariates, age was highly significantly related to prevalent disease, with ORs of 1.08 to 1.09 per additional year of age in all models. Female sex was significantly protective for combined MI/ stroke and MI, but not for stroke. Both current and past history of smoking were associated with significantly elevated odds of combined MI/stroke and MI; only current smoking was significantly associated with stroke.

	The only statistically significant ethnic difference was seen in non-Hispanic blacks. In multivariate models, non-Hispanic blacks had significantly higher odds of MI/

	stroke as compared with non-Hispanic whites (OR, 1.36; 95% CI, 1.06—1.76). Non-Hispanic blacks had increased odds of disease that were significant for stroke (OR, 1.49; 95% CI, 1.03–2.17) but not for MI. In sex-specific models, a significantly elevated risk among non-Hispanic blacks was seen in women (OR, 1.87; 95% CI, 1.31–2.67) but not men (OR, 1.05; 95% CI, 0.73–1.51).

	The present study shows that the metabolic syndrome, as defined by ATP III criteria, is independently associated with a history of MI/

	stroke, MI, and stroke. All of the component conditions of the syndrome, with the exception of abdominal obesity, are also independently associated with prevalent disease.

	Previous studies have shown linear relationships between coronary heart disease and either body mass index (BMI) or waist-to-hip ratio independent of BMI after adjusting for other risk factors.4 The lack of an independent association of high waist circumference with prevalent disease in our study suggests that some of the effect of high waist

	circumference is mediated by its role as a risk factor for other components of the syndrome. If this were so, we would expect that a model excluding the other component conditions would reveal a significant relationship between high waist circumference and MI/stroke, which, in fact, it does (OR, 1.43; 95% CI, 1.15—1.78).

	One major limitation of our study is its cross-sectional design. Although NHANES III is a useful data source because of its size, scope, and level of detail, its cross-sectional nature makes it impossible to prove conclusively that the hypothesized causes (the metabolic syndrome) actually preceded the observed effects (MI and stroke). This type of analysis also excludes patients whose MI or stroke resulted in death.

	Three prospective studies have examined the risks associated with the metabolic syndrome. Lakka and colleagues reported the association of the metabolic syndrome with cardiovascular disease and overall mortality in 1,209 Finnish men free of cardiovascular disease, diabetes, and cancer at baseline.5 They reported a lower prevalence of the syndrome (based on the ATP III definition) than in our study (8.8% versus 24.0%). Compared with the OR

	for nonfatal MI of 2.01 in our study (95% CI, 1.53—2.64), they reported

	a hazard ratio of 4.16 (95% CI, 1.60—10.8) for coronary mortality. In another analysis, McNeill and colleagues studied the association of the metabolic syndrome (based on ATP III criteria) with incident coronary artery disease in a population sample. They reported a hazard ratio of 2.58 for women and 1.74 for men, results that are also quite similar to our estimates.6 Finally, Malik and colleagues reported that the metabolic syndrome in NHANES II was associated with a significant increase in cardiovascular (OR, 2.29) and total (OR, 1.39) mortality.7 Thus, the limited available prospective data confirm our findings.

	Our study further supports the clinical importance of the metabolic syndrome as a significant risk factor for cardiovascular disease. The results emphasize the need

	for strategies to control this syndrome as well as its component conditions.

Metabolic syndrome, myocardial infarction, and stroke

	Zeller and colleagues (page 13) prospectively studied 215 consecutive patients with angiographically confirmed ostial atherosclero- tic renal artery stenosis of 70% or greater who underwent percutaneous stent placement and were followed up for 1 year. These subjects also had hypertension or impaired renal function (serum creatinine concentration of > 1.1 mg/dL for women and > 1.2 mg/dL for men), or both, but the initial clinical indications for suspecting the presence of renal artery stenosis were not defined. The primary end point was improvement in renal function at the end of 1 year, as defined by a reduction in serum creatinine concentration. Over the 1-year follow-up period, 52% of patients achieved a meager reduction in median serum creatinine concentration of 0.02 mg/dL that was equivalent to a mean increase in creatinine clearance of 2.3 mL/min. Patients with a baseline serum creatinine concentration of 1.5 mg/dL or less had no benefit from the procedure, compared with those with a baseline serum creatinine concentration greater than 1.5 mg/dL, who had the greatest benefit, with a median reduction of 0.33 mg/dL. Others, however, found a serum creatinine of 1.5 mg/dL or more to be a negative prognostic indicator.1 In the study by Zeller and colleagues, there was an overall procedural complication rate of 6.9%. Two percent of patients required long-term dialysis from acute postprocedure complications secondary to either contrast nephropathy or atheroembolic disease. In those patients without improvement in serum creatinine, the median creatinine increased by 0.11 mg/dL. Although there were no immediate procedure-related deaths, the 1-year mortality rate was 7.4%. Based on routine duplex sonographic screening at 6 and 12 months of the study, there was angiographic evidence of restenosis at the end of 1 year in 11.2% of patients, of which 85% occurred in the first 6 months.

	The secondary end point of blood pressure response to stenting of renal artery stenosis showed a reduction in blood pressure in 76% of patients, but the precise blood pressure values and medications were not delineated. Interestingly, Dorros and colleagues also showed a significant improvement in blood pressure and reduction in the number of medications required 4 years after stent placement.1 After stenting, hypertension is generally categorized as cured, improved, or unchanged, and reduced need for antihyper-

	tensive medications. Comparisons among studies have been difficult because of nonstandardized criteria, different target blood pressures, and variations in the medical regimen.2 In general, the reduction in blood pressure has been moderate, and resolution of hypertension has been the exception rather than the rule, suggesting the presence of an underlying essential hypertension rather than a primary renovascular etiology.1,2

	As mentioned by the authors, the lack of a control group is a sig-

	nificant limitation to the study. Chabova and colleagues followed 68 patients with atherosclerotic renal artery stenosis over an average of

	3 years and found that 85% had stable renal function and less than

	10% advanced to end-stage renal disease.3 The mortality rate was 29% over the 3 years, primarily due to cardiovascular disease.3 These data suggest that renal function remains stable for extended periods of time, despite the presence of atherosclerotic renal artery stenosis, and there is a relatively high mortality rate due to cardiovascular disease compared with end-stage renal disease.2 Yet, atherosclerotic renal artery stenosis is currently responsible for 5% to 15% of incident end-stage renal disease patients and is expected to increase as mean age increases, thus creating an unsolved dilemma for diagnosis and treatment.2

	The primary goal of renal artery stenting is preservation of renal function weighed against the risk of adverse events, particularly contrast nephropathy and atheroemboli.2 But the appropriate timing of any in-

	tervention is controversial. Most studies evaluate renal function by monitoring the serum creatinine concentration. Serum creatinine is an inaccurate marker in the early stages of renal dysfunction, how-ever, as changes may not be seen until 50% of total renal mass is lost. In the initial stages of atherosclerotic renal artery stenosis, the reduction in glomerular filtration of the affected kidney is often offset by increased glomerular filtration in the contralateral kidney that maintains a “normal” serum creatinine value.2 This might, in part, explain the absence of improved renal function after stenting in patients with a baseline creatinine concentration of 1.5 mg/dL or less. Moreover, an increased serum creatinine in unilateral atherosclerotic renal artery stenosis suggests the possibility of severe undetected bilateral renal artery stenosis, a nonfunctioning contralateral kidney, or parenchymal disease in the contralateral kidney. Determination of split renal function by nuclear imaging may detect hyperfiltration or reduced function of the contralateral kidney.2 Baseline proteinuria should be determined in diabetic patients because high-grade proteinuria indicates the presence

	of glomerular disease, which, along with a resistive index of 0.8 or greater by duplex Doppler, is considered a negative predictor of a favorable renal outcome.4

	With advances in noninvasive technologies, the diagnosis of renal artery stenosis has been made with greater regularity. The diagnosis should be entertained if there is deteriorating renal function, especially in elderly patients with coronary, peripheral, and cerebral vascular disease. Suspicion may also be raised in the presence of “flash” pulmonary edema, asymmetrical kidney sizes, or worsening blood pressure control. Depending on specific local expertise, user-dependent Doppler sonography or magnetic resonance angiography appear to be the studies of choice for initial screening. The indications for routine renal angiography or “drive-by” renal angiography have yet to be established. After the diagnosis of renal artery stenosis has been established, it is reasonable to consider stenting if the stenosis is 70% or more. Currently accepted indications for renal artery angioplasty with stenting include progressive azotemia, solitary kidney (functional or anatomical) with high-grade stenosis, flash pulmonary edema, and resistant hypertension. In the absence of accepted indications for intervention, the decision to stent a high-grade stenosis remains controversial. Caution should be exercised in the presence of severe atherosclerotic changes of contiguous arteries, advanced comorbid conditions, and short life expectancy. Future efforts should focus on prevention of atherosclerosis and improved methods of treatment of atherosclerotic renal artery stenosis with regard to prevention of contrast nephropathy and atheroemboli to the kidneys or periphery.

Percutaneous angioplasty  of renal artery stenosis

	A 54-year-old man with a diffuse type of atherosclerosis, including coronary three-vessel disease with severely impaired left ventricular function after posterior wall infarction, high-grade stenosis of the left subclavian artery, occlusion of the left common and superficial femoral arteries, occlusion of the right superficial femoral artery, and 90% stenosis of the left renal artery, was referred for interventional and surgical therapy. Known risk factors were smoking, hyperlipidemia, and hypertension. The serum creatinine concentration was elevated by 1.35 mg/dL, and the creatinine clearance was reduced by 58 mL/min. Average blood pressure measurements obtained by ambulatory 24-hour monitoring were systolic, 146 mm Hg; diastolic, 83 mm Hg; and mean, 102 mm Hg. The patient was taking three antihypertensive drugs. The difference of the intrarenal resistance index between the two kidneys, according to Pourcelot, was 0.09 (0.68 for the right kidney and 0.59 for the left kidney).

	Stent-supported angioplasty of the left subclavian artery and left renal artery (implantation of a Palmaz P 204 stent, 5-mm balloon) and recanalization of the left common and superficial femoral arteries were performed in the first session. Recanalization of the right superficial femoral artery was performed in the second session. Six weeks later, the patient successfully underwent surgical coronary artery revascularization.

	The resistance index of the left kidney was equalized immediately after the intervention (0.68 for both kidneys). After 6 and 12 months, serum creatinine concentration decreased to 1.00 mg/dL and 0.95 mg/dL, respectively, and creatinine clearance increased to 83 mL/min and 88 mL/min, respectively. Average blood pressure measurements dropped to 127/71 mm Hg with a mean of 92 mm Hg at 6 months and 121/75 mm Hg with a mean of 91 mm Hg at 12 months, despite an ongoing reduction of antihypertensive drugs by one drug.

	This patient, who had a diffuse type of atherosclerosis, underlines

	the results of our study, showing that patients benefit from stent-supported revascularization, even with unilateral severe renal artery stenosis, par-

	ticularly those with preinterventionally impaired renal and left ventricular function.

Case report: Improved renal function with stent-supported angioplasty

	Renal artery stenosis is a progressive disease involving the origin of the vessel, frequently occurring in patients with atherosclerosis.1 Kidney hypoperfusion causes poorly controllable hypertension and chronic renal failure due to hypertensive or ischemic nephropathy.2 Stent-supported percutaneous transluminal renal angioplasty has become an established treatment option that has been investigated in various studies.2-6 The effect of percutaneous renal revascularization on preservation of renal function, however, is still a matter of debate.2,5,7,8 The aim of this prospective study was to identify characteristics that could predict or preclude improvement of renal function during a 1-year follow-up period in a large cohort of patients undergoing stent placement for severe ostial renal artery stenosis.

	The study included 215 consecutive patients with 277 ostial renal artery stenoses, representing 91% of all patients undergoing stent placement for arteriosclerotic renal lesions located within 1 cm of the origin during a 5-year period. In most cases, assessment of renal artery stenosis was primarily based on duplex ultrasound.6,9 Before intervention, duplex ultrasound was always confirmed by angiography, showing a percent diameter stenosis of at least 70% on visual estimation. Indications for renal percutaneous stenting were hypertension (World Health Organization grade 1 or higher regardless of concomitant therapy) or impaired renal function (serum creatinine concentration > 1.1 mg/dL in women and > 1.2 mg/dL in men), or both.

	We excluded patients who had been on a hemodialysis program for longer than 1 year and patients who had had interventions in chronically occluded arteries. The guiding catheter technique and the procedure for implantation of the various types of stents have been described previously.10 In 50 patients (23%), bilateral stenoses were treated simultaneously.

	The end point of the study was the proportion of patients with decreased serum creatinine within 1 year of study entry (serum creatinine at 1 year less than baseline serum creatinine). Analysis was by intention to treat. The last creatinine measurement before entering the hemodialysis program was taken as the baseline for patients who were included in a hemodialysis program during the study period. For those who were withdrawn from a hemodialysis program, follow-up creatinine concentration was recorded. The creatinine clearance calculated by the Cockroft-Gault formula was also analyzed.

	Logistic regression analysis was performed to identify independent predictors for the study end point. The following variables were en-

	tered into the model: age; sex; baseline serum creatinine concentration; baseline mean blood pressure; left ventricular function; diameter stenosis of  90% or greater; resistance index greater than 0.80 after inter-

	vention; parenchymal/pelvic ratio greater than 1 for the affected kidney; bilateral stenosis; diabetes mellitus; concomitant atherosclerotic disease, such as coronary disease, peripheral occlusive disease, and cerebral occlusive disease; dyslipidemia; obesity; and smoking. Two-tailed hypothesis testing was used. Statistical significance was considered P < .05.

	Intervention was successful for all 215 patients. Severe procedure-related adverse events occurred in six patients (2.8%), including progression from preterminal renal failure to terminal renal failure due to renal embolism and dye-induced nephropathy (in four patients). During the follow-up period, 16 patients died, resulting in a 1-year mortality rate of 7.4%. Twelve patients (73%) died from cardiac causes; cerebral incidents occurred in two patients (13.5%); and malignancy occurred in two patients (13.5%). Eight patients were lost to follow-up, leaving 191 patients with 249 treated lesions available for 1-year follow-up. One year after entry into the study, 28 (11.2%) of the 249 renal arteries developed restenosis.

	Mean serum creatinine concentrations decreased significantly from 1.21 mg/dL (quartiles: 0.92 mg/dL, 1.60 mg/dL) at baseline to 1.10 mg/dL (quartiles: 0.88 mg/dL, 1.50 mg/dL) at 1 year (P = .047). Reduction in serum creatinine concentration over the 1-year follow-up period was reached in 99 of 191 patients (52%). In seven patients hospitalized with flash pulmonary edema or acute renal failure requiring acute hemodialysis, or both, hemodialysis could be discontinued. The median reduction in serum creatinine concentration was 0.02 mg/dL (quartiles: —0.11 mg/dL, 0.23 mg/dL,

	P = .011) at the 1-year follow-up. Creatinine clearance was also significantly increased, by 2.3 ± 15.1 mL/min (P = .028). The same outcome, with a median decrease in serum creatinine concentration of 0.02 mg/dL (quartiles: —0.10 mg/dL, 0.25 mg/dL, P = .001), was obtained when the last available serum creatinine concentration determination for patients who were lost to follow-up or died was included.

	Patients with improved renal function showed a median decrease in serum creatinine concentration of 0.22 mg/dL (quartiles: 0.12 mg/dL, 0.39 mg/dL). There was a median increase of 0.11 mg/dL (quartiles: 0.05 mg/dL, 0.23 mg/dL) in those patients who did not have improved renal function.

	Patients with impaired renal function at study entry (creatinine > 1.5 mg/dL; n = 48) had a significant median decrease in serum creatinine concentration of 0.33 mg/dL (quartiles: —0.01 mg/dL, 0.67 mg/dL, P = .025), whereas those with normal renal function (creatinine > 1.5 mg/

	dL; n = 143) had no significant change in serum creatinine concentration (median, 0.01 mg/dL [quartiles: —0.11 mg/dL, 0.14 mg/dL]). This difference between the two groups defined by the 1.5 mg/dL cutoff for serum creatinine was statistically significant at P < .001. In general, the higher the decrease in serum creatinine concentration, the more severe impaired renal function was at baseline (figure 1).

	There was no significant difference in baseline characteristics, except baseline serum creatinine, between the two groups. Decreases in serum creatinine concentration similar to the decrease in the entire cohort were found in patients with diabetes mellitus (0.05 mg/dL [quartiles: —0.11 mg/dL, 0.30 mg/dL]), severe nephrosclerosis (0.09 mg/

	dL [quartiles: —0.09 mg/dL, 0.21 mg/dL]), and unilateral involvement (0.01 mg/dL [quartiles: –0.11 mg/dL, 0.21 mg/dL], P = .047). Baseline serum creatinine concentration (P = .004) and left ventricular function (P = .032) were significant independent predictors for the decrease in serum creatinine at 1 year (figure 2).

	Slowing the decline in renal function is sufficient to claim a benefit from renal artery angioplasty according to the guidelines of the American Heart Association.11 This study showed significant improvement in renal function 1 year after intervention. Thus, our findings provide a conservative estimate of the advantages of renal angioplasty using stents. Impaired left ventricular function and increased serum creatinine were independent predictors of improved renal function. Nephrosclerosis and diabetes mellitus were not associated with a poor outcome in renal function. Furthermore, unilateral and bilateral treatment were associated with improvement of renal function.

	Other studies have not shown marked improvement of renal function after angioplasty of renal artery stenosis. The Dutch Renal Artery Stenosis Intervention Cooperative (DRASTIC) study was the only randomized trial to compare medical treatment with balloon angioplasty of atherosclerotic renal artery stenosis.12 That study did not show improved renal function after angioplasty. The interpretation of the study, however, is hindered by an outmoded interventional procedure with a 44% restenosis rate and a large percentage of crossover patients. Plain balloon angioplasty has been replaced by stent-supported angioplasty for atherosclerotic ostial renal artery stenosis because stenting provides outstanding acute and long-term results.13 Several studies, however, have not shown any marked improvement in serum creatinine concentrations with stenting.3,7 These differing results may be because of a small sample size and because a considerable number of patients in earlier studies had 50% to 70% stenoses, which were not likely to cause hemodynamic compromise.7,8,12,13 The results of our study, which had a sufficient sample size, showed improved renal function comparable with that shown in other studies.4,5,14 The favorable outcome of the present study may have been the result of using percutaneous stenting as the treatment.

	This study shows that more patients can benefit from stent-supported renal angioplasty than originally believed, especially patients with nephrosclerosis and diabetes mellitus. Radermacher and colleagues, however, found no improvement from renal artery revascularization in patients with severe nephrosclerosis.8 The difference between their study and ours may be that the majority of patients in their study were treated with plain balloon angioplasty.

	Contrary to earlier studies, the results achieved from the intervention were not related to bilateral stenoses in our study.4,5,8,14 The patients who showed the greatest improvement in renal function had congestive heart failure and renal insufficiency. If there is a concomitant systemic prerenal component or if the hemodynamic compromise is serious enough to result in renal dysfunction, improvement in renal artery stenosis may be most effective.

	This study was limited by the lack of a control group. Although progressive renal failure is a result,of severe renal artery stenosis,2 our study showed that the process can be prevented by angioplasty with stenting. The advantages of endovascular treatment of renal artery stenosis over conservative treatment, however, require further study. The benefit of stent-supported angioplasty in patients without renal dysfunction may have been underestimated in our study. Angioplasty may halt decline in renal function in these patients, but it is a benefit that does not become evident without a control group.

	Percutaneous stenting for serious ostial renal artery stenosis improved blood pressure and renal function in a larger range of patients than was formerly believed. Patients with

	diabetes mellitus, nephrosclerosis, and unilateral involvement should not be prevented from undergoing stent-supported angioplasty. Patients with left ventricular dysfunction or those who already had renal dysfunction received the greatest improvement in renal function from stent-supported angioplasty.

November 2004

Plasma B-type natriuretic peptide levels in chronic heart failure

BNP for the diagnosis of heart failure: A valuable tool, but no silver bullet

Should postoperative troponin elevation change patient management?

Case report: Value of ECG monitoring and troponin measurement

Perioperative markers of MI and survival after vascular surgery

The importance of the metabolic syndrome as a risk factor for CV disease

Risk factors for HF identified in older women with coronary disease

Cognitive decline observed with higher blood pressure at all ages

Drug-eluting stents show long-term clinical value

Does statin therapy reduce the incidence of atrial fibrillation?

Statin drugs protect against atrial fibrillation

Metabolic syndrome: Controlling the epidemic

Metabolic syndrome, myocardial infarction, and stroke

Percutaneous angioplasty of renal artery stenosis

Case report: Improved renal function with stent-supported angioplasty

Percutaneous stenting of atherosclerotic ostial renal artery stenosis

Latest News